Sustainable Urban Development in China A Literature Review on Issues, Policies, Practices, and Effects

Rui Wang

Assistant Professor School of Public Affairs

UCLA

3250 Public Affairs Building Box 951656

Los Angeles, CA 90095

ruiwang@ucla.edu Tel: (310) 367-3738 Fax: (310) 206-5566

November 2009

1 Introduction Nearly three decades of rapid economic growth in Mainland China has

fundamentally impacted its population and redefined international political, economic and environmental relationships. China is now the third largest economy (second if measured by purchasing power parity), the second largest energy consumer, the largest automobile producer and market, and the largest carbon dioxide emitter in the world. The rise of China as a heavyweight economic power is being accompanied by the grandest urbanization process in human history. Only about 200 million people lived in Chinese cities in 1980. Urbanization rate increased from less than 20% in 1980 to 45% in 2008. In the past decade more than ten million new urban dwellers were added annually. The UN World Urbanization Prospects (2007 Revision) projects close to a 900 million urban population by 2030.

Since 1993 China has been a net importer of oil. It continues to be the world’s largest consumer of coal, a resource that accounts for about 70 percent of China’s primary energy supply. According to Levine (2008), China’s rapid growth has been accompanied by an annual energy growth rate of greater than ten percent from 2001 to 2005. This in turn has led to 70 gigawatts (GW) of added generating capacity in 2005 alone, an amount on par with the scale of the entire British power grid (Steinfeld, 2008). The following year witnessed an additional 102 GW of capacity expansion. This recent growth in energy use has led to China matching the carbon consumption rate of United States, at six billion tons per year in 2006, far sooner than was expected. Chinese per capita energy consumption remains well below levels found in advanced industrial societies as of now. However, the per capita energy consumption of urban citizens is 3.5 times that of rural citizens (Steinfeld, 2008). Not surprisingly, China’s urbanization will produce substantially higher residential energy consumption.

Room for future growth in mobility seems enormous. As people, especially the urban residents, grow wealthier, China approaches a threshold where auto ownership rate accelerates. The most recent data (Haddock & Jullens 2009) show that China’s overall auto usage was just 18 cars per 1,000 people in 2008, compared to 104 in Brazil and 213 in Russia. China, together with India, another rapidly motorizing country registering 11 cars per 1,000 people in 2008, is likely producing an unprecedented global wave of motorization. Such a trend is reinforced by the infrastructure-led economic development of this country. While the national expressway system started to operate only in 1988, by the end of 2001, it became the second longest in the world.1 The construction of highways is still accelerating, thanks to the recent economic stimulus plan.

Economic development enabled cities in China to manage some environmental issues, such as sanitation and indoor air pollution, which are typical problems of urban households in low-income countries. However, China continues to struggle with issues such as air pollution, water scarcity, solid waste management, and greenhouse gas (GHG) emissions.

Currently, two-thirds of China’s cities face water scarcity. China uses six times more water per unit of GDP than the Republic of Korea and ten times more than Japan 1 Data from http://www.jttj.gov.cn, retrieved on Aug. 8, 2009.

Page 2 of 20

(Yusuf & Nabeshima, 2008). Additionally, underground water is being pumped at unsustainable rates. Although the rate of urban water treatment is increasing – up to 45 percent in 2005 – the absolute number of urban residents not linked to water treatment systems has also increased. On the other hand, China has yet to realize a substantial reduction in industry-based water pollution due to changes in industrial structure favoring cleaner downstream production (World Bank & SEPA 2007).

Nevertheless, cities in China have coped relatively well with rapid urbanization. China has been able to absorb more than 370 million people in its cities without the proliferation of urban slums, although sewage and waste disposal services have struggled to keep up with demand. Between 1990 and 2000, 130 million new urban dwellers gained access to improved sanitation facilities (Yusuf & Nabeshima, 2008). However, if we take the definition of sustainable urban development by Girardet (2004) – cities that enable all residents to meet their own needs and prosper without degrading the natural world or the lives of other people, now or in the future – we need to pay attention to all major aspects of environmental sustainability of the cities: air, water, solid waste, energy, transportation, and land use.

This paper reviews recent literature on issues, practices, and policies related to the sustainable development of Chinese Cities. It will focus on China’s reform era since 1978, especially since mid 1990s, when rapid urbanization started.

2 Urban air quality

2.1 Issues Urban air pollution is a major environmental issue threatening the health of the

Chinese citizens. According to He et al (2002), black smoke from factory stacks became commonplace in Chinese industrial cities during the 1970s. During the 1980s, many southern cities began to suffer serious acid rain pollution; and beginning in the 1990s, the air quality in large cities deteriorated due to nitrous oxides (NOx), carbon monoxide (CO), and photochemical smog, which are typical components of vehicle pollution. Moreover, total energy consumption in China has increased 70 percent between 2000 and 2005, with coal consumption increasing by 75 percent, which perhaps explains why China is the largest source of SO2 emissions in the world (World Bank & SEPA, 2007).

Due to the adoption of various control measures, Chen et al (2004) find that the ambient air quality in a number of large cities in China has improved in terms of ambient air total suspended particles (TSP) and SO2 levels. However, ambient air NOx level has been rising due to the increased number of motor vehicles. Ambient air pollution in large cities has changed from the conventional coal combustion type to the mixed coal combustion/motor vehicle emission type.

The high health cost of particulate matters (PM), and especially the fine particulates (PM2.5), started to attract attention in late 1990s. Peng et al (2002) study the emissions inventory of Shijiazhuang (a large North China city). The results reveal significant health costs associated with Shijiazhuang’s high concentration of sulphate, a fine particulate matter originating mainly from coal consumption. The use of cleaner coal

Page 3 of 20

was found to be the most cost effective in improving urban air quality and reducing human exposure. He et al (2001) analyze weekly PM2.5 samples in Beijing from 1999 to 2000. They find that much higher coal consumption coupled with weather was the likely causes of the high PM2.5 concentrations in the winter, while PM2.5 levels in the spring were clearly impacted by dust storms. With annual average PM10 concentrations of over 100µg/m3, several selected cities in both northern and southern China are among the most polluted cities in the world (World Bank & SEPA, 2007).

Air pollution can be more serious in smaller cities due to loose regulations. Wang and Mauzerall (2006) estimate the health impacts and damage costs of anthropogenic air pollutants in Zaozhuang, a mid-sized North China city, to be 10 percent of local GDP. They also project that advanced coal gasification technologies would be much more effective in controlling local air pollution than end-of-pipe controls and also provide opportunities to sequester CO2 underground.

2.2 Policies and practices The control of SO2 and acid rain has been focused on coal combustion in China.

Emission charges and technological standards were implemented starting in the 1990s, in particular for newly built thermal power plants and other large or medium-sized industrial furnaces. Cities in the Acid Rain Zones and SO2 Pollution Control Zones were required to adopt comprehensive protection plans for the control of SO2. However, He et al (2002) question the effectiveness of emission charges, primarily on grounds that the charges were significantly lower than the abatement costs. Other policies included research and development of clean coal and emissions control technologies.

Rapid motorization since the 1990s has turned many cities’ attention to mobile sources to control NOx, CO, and photochemical smog pollution. In the 1990s, the State Environmental Protection Agency (SEPA) enacted standards for vehicular emissions based on international control standards from the mid-1970s. Following successful experiences in Europe, new vehicle emission standards phased in quickly, with Euro 1 standards adopted in 2001, Euro 2 in 2004, and Euro 3 in 2007. Cities facing serious vehicle emissions pollution, such as Beijing and Shanghai, have been adopting more stringent local standards. Research and development (R&D) subsidies to cleaner vehicles have also been implemented. Combined with increased concerns about energy shortage and global climate change, national automotive R&D policy has been particularly favoring of electric vehicles recently. However, as He et al (2002) point out, Chinese cities face challenges in effectively implementing inspection/maintenance (I/M) programs for in-use vehicles, mainly due to poor cooperation between local Environment Protection Bureaus and city traffic administrations. Leaded gasoline was phased out in 2000. However, other aspects of fuel quality have not improved much, except in a few larger cities. The high sulfur content still exacerbates emissions and impedes implementation of more stringent standards (He et al, 2002).

Hao et al (2006) study vehicle emission control policies in Beijing during the decade following mid 1990s. They estimate that emission standards for new vehicles and I/M programs for in-use vehicles are the most significant contributors to vehicular

Page 4 of 20

emission reduction, followed by public transportation improvements and limiting the use of private cars.

However, comparing China and US urban air pollution regulations, NAE and NRC (2007) find that China has focused on directly emitted PM and SO2 emissions and concentrations, with less regulatory attention on secondary pollutants such as O3 or the sulfate, nitrate, and ammonium components of PM. They also suggest that China has made great progress over the last 25 to 30 years in reducing emissions per unit of fuel use or production. However, these were largely achieved by shuttering obsolete facilities, adopting modern engine designs and requiring cleaner fuels – the “low-hanging fruit” among all necessary regulatory measures. In addition, many cities benefited from relocating key industries away from urban centers, which cleaned urban air. Ultimately, however, this will not lower regional emissions. Greater reductions for a larger number of emitters and economic sectors will be needed to attain healthy air quality, especially given China’s rapid growth in all energy sectors.

No research has shown the extent to which the urban air quality change of Chinese cities can be attributed to industrial relocation. However, one can expect that with air pollutants increasingly emitted from the consumer rather than industrial sector, spatial relocation of pollution source will become less feasible.

3 Water scarcity and quality

3.1 Issues China’s available water per person is one-third the world average. Moreover, total

water availability in the north is about one-sixth that of the south and one-tenth the world average. More than 400 of China’s 600 cities are believed to be short of water, and about 100 face serious water shortage problems (Shalizi, 2008). Water scarcity may lead to depletion of groundwater. In some areas of China, the groundwater table has fallen 50 meters since 1960, and it continues to fall three to five meters annually (World Bank & SEPA, 2007). The problem is more much severe for larger cities. The groundwater table in Beijing is estimated to have dropped 100 to 300 meters (Shalizi, 2008), and the city’s aquifer is nearly dried up (World Bank & SEPA, 2007). The removal of underground water domes has many adverse consequences, including saltwater intrusion in coastal cities and subsidence of land mainly in the northern cities.

Separate from water scarcity, water pollution happens in both the northern and southern parts of China. According to Economy (2004), “more than 75 percent of the water in rivers flowing through China’s urban areas is unsuitable for drinking or fishing. Only 6 of China’s 27 largest cities’ drinking water supply meet State standards . . . [and] many urban river sections and some large freshwater lakes are so polluted that they cannot even be used for irrigation”.

The amount of wastewater discharged from larger industries has leveled off since the early 1990s, due to regulation. In small cities and rural areas, however, discharges from the numerous town and village enterprises are increasing rapidly and causing extensive pollution of water bodies (World Bank & SEPA, 2007). Although more than

Page 5 of 20

1,000 wastewater plants were built between 2000 and 2006, only about half of the discharges from municipal sources are treated. This is partially due to inadequate wastewater collection facilities, and partially because revenues collected from customers are transferred to the general city budget and not used to ensure that treatment plants have the resources needed to operate adequately (Shalizi, 2008)

Agriculture remains the largest source of water consumption in China, but growth in demand has been greatest in urban and industrial sectors. In 2005, China consumed about 563 billion cubic meters of water, of which 64 percent was used for agriculture, 23 percent for industry, and 12 percent for household purposes. Demand for water grew at an annual rate of 7.3 percent for urban households and 4.3 percent for factories between 1980 and 2002. During this time, water demand by rural households and farms remaining almost unchanged (Shalizi, 2008). Shalizi also projects that water demand from urbanization will be harder to accommodate than overall population growth, particularly if average demand for water by urban households continues to grow rapidly, and if there is not a substantial deceleration or decline in water demand by agriculture and industry.

3.2 Policies and practices The water quantity and quality crises in China has prompted many critiques of

China’s water policies. Xie and others (2009) urge fundamental reforms of the current water governance system, which is characterized by (1) an incomplete, ambiguous and conflicting legal system that lacks mechanisms and procedures; and (2) fragmented and uncoordinated institutions with limited transparency, and very low public participation. Other researchers are more optimistic. Duan (2005) believes that sustainable water resources management, water conservation, the completion of the south to north water diversion project will solve Beijing’s water shortage problem in the long run. However, Varis and Vakkilainen (2001) regard the planned diversion of water from the Yangtze River basin as a tiny drop to their solution. The real actions needed are an increase of water use efficiency and the abatement of environmental degradation in the North China Plain.

Raising water price has been argued by many as one effective measure to control the growth of water consumption. Zhong and Mol (2009) review the reforms of urban water price since 1985, and find that the sharp increase in urban water price in recent years was accompanied by a decrease in average urban domestic water use per capita from 220.2 l/day in 2000 to 178.4 l/day in 2007 and the decrease in the average water intensity from 554 m3 per 10,000 RMB (constant price of 2000) GDP in 2000 to 254 m3 in 2007. Nevertheless, Wang et al (2008) conclude that household water supplies are still heavily subsidized in China, which imposes an equity issue, because the vulnerable groups use less water than the richer groups. Through a survey of about 1,500 households in five suburban districts in Chongqing Municipality, Wang et al (2008) further suggest that a significant increase in the water price is politically feasible as long as the poorest households can be properly subsidized and certain public awareness and accountability campaigns are conducted.

Sustainable urban water use certainly involves reforming municipal water utilities. Browder and others (2007) review China’s accomplishments in providing urban water

Page 6 of 20

services, finding that most of China’s water utilities face high service standards set by the government. Government subsidies typically do not sufficiently cover the shortfall in user fees. In particular, many wastewater utilities operate at low levels of efficiency under the supervision of municipal governments, characterized by a lack of accountability, transparency, and customer orientation. A policy recommendation of Browder and others is to increase private sector participation in municipal water services. As Zhong et al (2008) find, China has started experiments allowing private participation in urban water service in the 1990s. Although it seems private sector involvement is more and more popular, most contracts are in early stages. The current private sector involvements in the Chinese water sector still face many legal and regulatory uncertainties.

Finally, public disclosure of information on water quality and community consultation could improve feedback and facilitate better monitoring. Research by Jiangsu Province, SEPA, and the World Bank on pilot versions of community consultation and feedback processes, as well as public disclosure of information, has shown that these methods are effective and have the potential to be scaled up (Shalizi, 2008).

4 Solid waste

4.1 Issues Even at the same income level, urban residents produce two to three times more

waste than their rural counterparts. Driven by urbanization and increasing affluence, China recently surpassed the United States as the world’s largest municipal solid waste generator (Hoornweg & Lam, 2005).

However, compared to air and water, the dearth of adequate solid waste management in the Chinese cities has attracted little attention. Where sanitary landfills, recycling programs and other properly managed means of solid waste disposal are not available, open dumping and burning are the norms. This leads to environmental and health hazards for urban residents – especially those who live closest to the dumping sites. The sprawling waste can contaminate ground and surface water, which eventually can result in ecosystem and health damages. With annual waste production growing at close to ten percent per year, sanitary landfills are still rare in China (Dong et al, 2001).

Hoornweg and Lam (2005) also notice significant and continuing changes in waste composition, which will affect the technological choice of solid waste treatment and disposal technologies. One important aspect is the decrease in household-level coal ash in the waste stream, resulting from the abandonment of coal usage for home heating by urban households. This lowers the cost of waste treatment, as coal ash mixed with organics degrades finished compost quality by introducing heavy metals, and coal ash mixed in the combusted waste stream reduces efficiency of incinerators.

4.2 Policies and practices Municipal policy and administrative efforts have mainly focused on improving

the treatment of solid waste. The government of China has issued a series of favorable policies to encourage investment in incinerators. These incentives include tax refund,

Page 7 of 20

prioritized commercial bank loans, state subsidy for loan interest, and guaranteed subsidized price for purchase of electricity (Hoornweg & Lam, 2005). As a result, numerous waste-to-energy facilities have been built in China in recent years, primarily in large and middle-sized cities (Raufer, 2007). Besides reducing the volume of waste, these facilities can provide both thermal and electrical energy services. However, Hoornweg and Lam (2005) offer the criticism that China might have more than optimal incineration capacity for a few reasons. Incineration is not environmentally beneficial because toxic and greenhouse gases are released in the process. It is also expensive and does not encourage waste reduction and recycling.

Besides the recent national prohibition on free plastic carrier bags by stores, little policy emphasis on waste reduction and reuse has been put forward in China. The recycling of solid waste has been marginalized in Chinese cities. Through surveying the junk buyers in central China’s biggest city, Wuhan, Li and Goss (2005) examine the socio-spatial activities of waste recovery that are typical in Chinese cities. They find that without rural migrants as door-to-door collectors, the current waste recovery system in China would break down. While junk-buyers play a role of linkage between waste sources and redemption depots, the low social status and even “illegal” status of rural migrants as junk-buyers in the cities determine that they have no rights to claim or fight for urban space as their territory, even though they have developed unofficial collecting territories.

5 Energy efficiency

5.1 Issues China’s industrialization, urbanization, and motorization have not only resulted in

domestic environmental problems, but also raised global concerns of energy security and climate change. The recent rocketing energy demand, which led to a series of electricity, fuel, and coal shortages, followed by massive energy supply capacity expansion, and the recent economic stimulus have further induced the growth. China’s intensity also declined from 1985 to 2000, but, since 2000, it has been increasing (NAE & NRC, 2007).

By the beginning of 2007, China had become the world’s largest construction market (Steinfeld, 2008). The building sector consumes from more than 15 to 45 percent (Steinfeld, 2008; Li & Colombier, 2009) of China’s total primary energy, depending on the exact definition of the sector. But all agree that this proportion is growing steadily. The Chinese Ministry of Construction estimates that around 15 to 20 billion m2 of urban zone housing will be built between 2005 and 2020 in order to accommodate newcomers to the cities – equivalent to the entire existing building stock in the EU-15 (Li & Colombier, 2009).

The growth of urban energy consumption has much to do with income growth and the improvement of life quality. A residential consumption survey in Shanghai shows that residential electricity varies significantly depending on household income and home size, with higher income households and larger apartments consuming over twice as much as lower income families in smaller houses (Li & Colombier, 2009).

Page 8 of 20

However, many researchers have argued that the efficiency of energy use in the building sector is very low. Less than half of the cities in China currently have a district heating supply infrastructure; electrical demand is also being driven by air conditioning use in urban areas, but much of the growth comes from small scale, relatively inefficient window units (Raufer, 2007). In Beijing, despite a significant reduction of building energy consumption after the implementation of the 1995 Energy Efficiency Standard, residential buildings are estimated to consume 50 to 100 percent more energy for space heating as compared to buildings in similar cold climates in Western Europe or North America, while commercial buildings consume 40 percent more energy than buildings of the same type in Japan (Zhu & Lin, 2004).

5.2 Policies and practices China is still at a relatively early stage of urbanization, with a great deal of long-

term investment in urban infrastructure and buildings ahead. Measures that can cut energy use and energy losses have a high payback. Moreover, increasing energy efficiency and development of clean energy also have positive effects on public health. Chen et al (2007) estimate the exposure level of residents of Shanghai to air pollution under various planned scenarios including energy efficiency improvement, expanding natural gas use for final sectors, and wind electricity generation. They find that implementing these policies will not only decrease the emission of greenhouse gases, but also reduce air pollutant emissions, improve air quality, and promote public health.

As the building of urban commercial and residential space ramps up, it is imperative that the government promotes energy efficient designs and construction materials. Some believe the design of construction and the incorporation of new energy- and materials-conserving technologies are essential. Yusuf and Nabeshima (2008) propose strict electricity consumption standards and codes for air conditioning and other appliances, smart meters that encourage off-peak electricity use, and adoption of other general “green” technologies and eco-friendly designs. At a large scale of urban design, the integration of combined heat and power and district heating/cooling systems, which offer a longer term means of addressing energy demand requirements in fast-growing urban areas, is recommended (Raufer, 2007).

However, many barriers to energy efficiency improvement have been identified. Four consecutive building energy efficiency standards have been stipulated in China since the mid-1980s. However, the enforcement and implementation of these regulations have encountered numerous difficulties at the local level, particularly in the small and medium cities where the inertia of builders and the prevalence of conventional building techniques and practices are major barriers (Li & Colombier, 2009). The 11th Five Year Plan (2006–10) calls for energy savings of 50 percent in new buildings, but local developers loathe paying the higher up-front costs for energy-efficient materials and building systems. (Steinfeld, 2008), probably due to the increased construction costs associated with energy-efficient measures, combined with a dearth of available advanced technologies (Li & Colombier, 2009). The lack of marginal cost pricing has been particularly criticized for heating services in the north. Heating consumption is billed on the basis of surface area instead of actual consumption (Li & Colombier, 2009). As a

Page 9 of 20

consequence, economic incentives to reduce consumption are nonexistent to consumers and housing developers.

Many call for broader policy and market reforms. Li and Colombier (2009) argue that supportive policy and appropriate regulatory framework, such as: building code updates; high-efficiency appliance labeling and certification; energy service corporate financing tools provided by the Clean Development Mechanism; carbon markets and energy pricing reform, etc., are prerequisites for facilitating the upgrade of technological innovation in the building sector. Yusuf and Nabeshima (2008) further prescribe four areas of actions necessary to ensure that urban development is not constrained by rising energy costs: energy pricing (especially of power); motorization strategy; green technology development and transfer; and a range of regulatory policies, including environmental and land use policies and policies defining building codes and standards for consumer appliances.

Finally, as argued by Yusuf and Nabeshima (2008), among the measures with the greatest energy consequences for urban areas, the design of urban transport is the most significant, because it determines the physical characteristics of the city and the degree of reliance on automobiles. The following sections will discuss the energy implications of both transportation and land use.

6 Transportation

6.1 Issues Besides its key role in facilitating economic growth, urban transportation affects

urban environmental quality, energy efficiency, and emissions of greenhouse gases. A notable characteristic of urban transportation in China is motorization, especially the growth in automobiles. Another aspect of urban motorization in China is the development of mass transit, in particular bus systems in large and medium-sized cities. There are only about ten cities currently operating rail transit systems, although several other cities are in stages of preparation or construction. Loo and Li (2006) describe the history of metro system building in China, emphasizing the strong control of the central government on metro projects.

Motor vehicle emission is a major contributor to urban air pollution. Liu et al (2007) estimate vehicle emission inventory in Beijing and Shanghai and find that emissions from on-road mobile sources in Beijing were much larger than in Shanghai because of the high vehicle-kilometers travelled in Beijing. The amount of particulate matter is higher in Shanghai, which they attribute partially to Beijing’s stringent single vehicle emission standards, compared to Shanghai’s focus on controlling the quantity of private cars.

Wang (2008) estimates the environmental and safety cost for major urban passenger modes in China. Assuming the modes are efficiently operated with up-to-date mainstream technologies, the study finds that while environmental costs per vehicle-kilometers travelled are higher for transit modes than automobiles, they are much cleaner on the passenger-kilometers travelled basis. In addition, the structure of environmental

Page 10 of 20

costs is found to be different in China from results in the US. Auto air pollution costs were indicated to be about 3.4 to 4.7 times as much as climate change costs in the United States (Small & Verhoef, 2007). In China, climate change costs are at least comparable and in many cases larger than air pollution costs of motor vehicles.

The bicycle was dominant in urban transport in the 1980s, a position maintained in the 1990s, even in the largest city of Shanghai (Zacharias, 2002). Bicycle’s relative decrease in importance in the recent past, unfortunately, has not caught much attention from neither researchers nor policy makers. Some observe that the traditional non-motorized bicycles are undergoing significant changes as more urban residents choose electric two-wheelers in order to travel with more comfort and for longer distances. Weinert et al (2007) document the detailed history of the development of the electric bicycle in China. Annual electric bike sales in China grew from 40,000 in 1998 to 10 million in 2005, even though the higher speed capabilities of electric bikes resulted in serious safety concerns. As they mention, due in part to regional variations in policy enforcement, electric bike penetration is noticeably different from city to city. Cherry et al (2009) find that from a life-cycle perspective, electric two-wheelers emit several times lower pollution per kilometer than motorcycles and cars, have comparable emission rates to buses and higher emission rates than bicycles. However, led is one pollutant on which electric two-wheelers perform poorly due to dependence on batteries.

6.2 Policies and practices Sustainable urban transportation can be tackled by several policies, such as

technological standards, demand management, and infrastructure investment.

Although Chinese fuel economy standards are already stricter than some western countries like the US, Canada, and Australia (An & Sauer, 2004), further tightening fuel economy standards seem important for China’s urban energy future. Using historical data on oil consumption and CO2 emissions from China’s road transport sector between 1997 and 2002, He et al (2005)’s forecasting model concludes that China’s road transportation will gradually become the largest oil consumer in China in the next two decades, but that improvements in vehicle fuel economy have potentially large oil-saving benefits. Their results suggest that in order to contain the dramatic growth in transport oil consumption, China needs to increase vehicle fuel economy immediately. This is especially important now because China is in a period of very rapid motor vehicle sales growth, and the vehicles currently on the market are relatively inefficient.

Through measuring emissions from sample on-road vehicles in Beijing, Zhou et al (2007) find that older vehicles are contributing substantially to the total fleet emissions, indicating the large-scale retrofit of older vehicles with three-way catalysts had far fewer impacts on controlling emissions than expected. They suggest that the adoption of more stringent emission standards remains the most effective available strategy to counteract the environmental pressure caused by the increase in private vehicle sales.

Besides national policies on standards, some regional policies also affect energy and environmental impacts of motor vehicles. Historically, there existed long-term regional discriminatory policies against lightweight vehicles because of the safety and efficiency concerns and implicit local industrial protectionism. The central government

Page 11 of 20

finally required that all the national or regional discriminatory restrictions on the use of economic cars be abolished by March 2006. These efforts should stimulate the consumption of economic cars for private use (Winebrake et al, 2008).

Relatively less research and policy practice focus on demand management in urban transportation. In fact, the national policy on fuel price almost exclusively focuses on its impact on macro economy. Fuel price was kept low and only became comparable to prices in the US when domestic demand and international oil price started to surge a few years ago. Because intercity highways and bridges often collect heavy tolls, Chinese policymakers are somewhat hesitant to add additional gas taxes (Winebrake et al, 2008). However, urban roads remain largely free, although a few cities recently started to consider congestion tolls.

Many (e.g., Gan, 2003; Winebrake et al, 2008) suggest that the government should actively discourage automobile ownership due to energy consumption and the environmental concerns. A national policy discouraging the ownership of automobiles, however, is not very likely, due to the strong incentives for economic growth to both national and local governments. Barth and Shaheen (2003) suggest that carsharing can be an effective way to reduce automobile ownership and use. Although carsharing may have a stronger market potential in China relative to developed countries, it is unclear whether policies encouraging carsharing will result in more or less driving. Wang (2008) suggests that a Pigovian tax on gasoline targeting the environmental externalities of automobiles is far from enough to deter the growth of motorization. Taxes that vary with vehicle characteristics and time and location of driving should have more potential. Overall, not much discussion has focused on evaluating policy alternatives. For example, Shanghai’s license plate quota policy has not received serious empirical examinations.

Not much work has examined the supply side of urban transportation either. The central government has been criticized for its controlling of major investment projects, like metro systems. Loo and Li (2006) point out that the existing official criteria (size of population and local economy) for approving the building of metro systems are insufficient and should be supplemented by more vigorous evaluation criteria. However, it is unclear whether the real issue here is the over centralization of decision power or the lack of sufficient objective project evaluation in decision process. Wang (2008) compares the full social costs of major passenger transport modes in representative Chinese urban commuting corridors, showing that commuting by bus and bicycle has lower social cost than by automobile or rail in large cities.

Finally, transportation interacts with land use in ways that have important environmental and energy implications. Kenworthy and Hu (2002) compare Chinese cities to a large sample of other cities from around the world in 1995. The Chinese cities are among the highest in population density, but the lowest in per capita road length and mode share of private motor vehicles. From a 1996 housing relocation survey, Yang (2006) observes that commuting time increases by 30 percent as households move away from previous housing locations and from central districts to suburban districts. However, we know little about quantified environmental and/or energy use impacts of transportation when the feedback of land use is incorporated. This perhaps is due to the analytical difficulty of quantifying causal effects between transportation and land use

Page 12 of 20

changes. This area is both difficult and important to making policies about urban transportation and land use, as both are believed by many to have long-term energy and environmental effects.

7 Land use

7.1 Issues Many have expressed concerns about low-density urban growth, or the

environmental and social costs of sprawl, and have recommended policies for sprawl control. Among major arguments against sprawl are energy and environmental costs of motorized transportation and food security impacts resulting from loss of agricultural land. The first has seemingly been widely accepted as evidence (e.g., Kenworthy 2003) that low urban population density is associated with high passenger transportation carbon emissions per capita across the world. The second is more of a natural concern for countries like China with limited arable land relative to its population. However, the question of the extent to which these concerns apply to China remains.

In China, limited evidence seems to produce mixed results. Using data from 45 core Chinese cities, Chen et al (2008) correlate cities’ compactness to a range of environmental variables. The findings suggest that urban population density is positively correlated with accessibility of services, infrastructure efficiency, and use of public transport, while negatively correlated with per capita domestic energy consumption. However, population density also comes with air pollution, noise, and loss of green space. In addition, the positive environmental effects generated by high urban density only hold up to the density level of approximately 168 persons per hectare, which is higher than most of the sample cities but not the largest ones. The authors also acknowledge that the influences of urban compactness on studied environmental attributes are not as significant as expected.

On the other hand, doubts have been cast on the consequences of food security. Lichtenberg and Ding (2008), together with many observers, find that evidence supports that a substantial share of farmland losses does not represent a reduction in food production capacity, and that increases in other factors of production can compensate for farmland losses. In other words, China can remain largely self-sufficient in terms of food production with current rates of population growth and farmland loss despite its scarcity of additional arable land. In fact, as pointed out by Lichtenberg and Ding, water shortage is perhaps a much more serious issue than sprawl, in terms of food security.

A potentially more controversial issue is whether Chinese cities have been sprawling, specifically if sprawl means low-density urban growth. Many researchers such as Zhang (2000) and Campanella (2008) and practitioners such as NRDC (2007) have implicitly assumed that urban sprawl prevails in China. It is often unclear, however, whether they refer specifically to urban spatial expansion when they use the term “sprawl.”

There is no question that urban spatial expansion is happening at an unprecedented pace in China, fueled mainly by growth in income, urban population size,

Page 13 of 20

and transportation infrastructure investment, as suggested by Deng et al (2008). Based on the observations of land use rights granted from 1993 to the first half of 2000 in the built-up area of Beijing, Ding (2004) suggests that the slope of the land rent curve declined over time. However, this does not mean urban growth in Beijing has been of low density. In fact, using cross-sectional land price data, Wang (2009) suggests that richer cities have larger land price gradients, which supports the argument of Li et al (2002) that employment decentralization and polycentricity are still underdeveloped in Chinese cities.

Sometimes rural housing, township and village industrial land use can be taken as evidence of urban sprawl. Studying satellite images, Ho and Lin (2004) find that in the south region of the coastal province of Jiangsu, rural settlements are scattered across a large amount of land, while township industries also cause dispersed agricultural land loss. However, these can at most be seen as land management issues in rural areas, rather than low-density urban growth. Even many other urban industrial land uses are difficult to be classified as sprawl. Wu (2008) studies how foreign direct investment (FDI) has been transforming Shanghai’s peri-urban area, where development zones were set up for manufacturing FDI. However, such areas are often self-contained, with a concentration of temporary migrant workers, who hardly have any relationship with the city of Shanghai. In addition, the living density of these workers is relatively high. As Wu finds, per capita housing is less than eight square meters.

Sprawl has a natural institutional barrier in China. Wang (2008) argues that China’s dual urban-rural land administrative system results in two important urban spatial characteristics: city-oriented infrastructure and social service; and government-led suburbanization. The former increases the amenity of central city infrastructure relative to the urban fringe, while the latter entails that urban residents cannot move to surrounding rural areas as residents in many western cities, except after rural land has been converted into urban uses by the government.

Another argument is that sprawl happens because local governments are “addicted” to urban growth and redevelopment – anything that will allow them to lease more land. However, so far there is no evidence to support that. Theoretically, such a motive does not automatically induce urban sprawl, even when local officials only focus on short-term fiscal revenue. This is because the density of urban land development is positively related to its leasing value, which is what truly motivates the local government. Moreover, if the government provides major infrastructure in urban areas, higher density actually will reduce the government’s costs. Empirical evidence found by Deng et al (2008) also suggests that the overall unobserved effect of current policy has greatly slowed the growth of the cities instead of speeding up sprawl.

7.2 Policies and practices The aspect of land use that the Chinese government can potentially have clear

influence at the city level is land use and migration policies. Optimal city size becomes more relevant to sustainable urban development in the Chinese context.

A study by Zhao and Zhang (1995) is one of the earlier reflections on city size control in China, which has not changed much since the study was conducted. After examining the correlations between city size and measures of economic, social, and

Page 14 of 20

environmental efficiencies of cities, they argue that restriction on the development of large cities in China seems to work in an opposite direction to sustainable industrialization and urbanization. The large cities fare better not only in terms of economic development, but also in terms of social and environmental conditions. Their conclusion is supported again recently by the economic modeling results of MGI (2008). Comparing different scenarios of future urban systems, given China’s urbanization trend, MGI suggests that a concentrated growth will result in not only higher per capita GDP, but also: more agglomerated talent as an engine of economic growth; higher energy efficiency; decreased loss of arable land; more efficient mass-transit; and more effective pollution control.

In summary, city size seems to be a huge benefit, given that China expects more than ten million rural migrants a year in the coming decades. The Chinese government needs to design appropriate policies that can lead to large cities that are cleaner, more energy efficient, and more competitive in the global economy.

8 Conclusion Will the rapidly growing cities of China live up to their potential as dynamic

engines of growth and social modernization, or will they be mired by pollution, congestion, and related socio-economic problems? The answer to this question depends on both national and local policy creation and implementation. Because China is a relatively late starter in terms of urbanization, and much building and renewal of urban physical capital lies in the future, there is unparalleled scope for designing efficient and livable cities (Yusuf & Nabeshima, 2008). In addition, the Chinese central and local governments have the authority to invest in infrastructure, regulate land use, and control the number of rural migrants, all of which can facilitate efficient decision-making and effective implementation.

The above review of issues, policies, and practices related to sustainable urban development provides a broad picture of our current knowledge. It also illustrated several areas that future research is needed. First of all, policy analysis should emphasize more demand management. Careful studies and policy simulations should be conducted to determine efficient and acceptable road toll, gas and vehicle taxes, and water prices. After significant amounts of investment in cities’ physical infrastructure, managing demand will be more important as the expansion of supply becomes more expensive.

Second, current research on water policy seems to hardly match the importance of water scarcity in China’s cities. For cities in North China, depending on water transfers from the south will be expensive and even infeasible for technical, ecological, or climate change reasons. The design of water and sewage systems, standards of water quality and appliance efficiency, and the pricing of water are essential components of a sustainable urban water strategy. In addition, strategic considerations of water resources should be a part of regional development plans in China. Despite growing water scarcity in the north, there has been no noticeable demographic shift in population from north to south (Yusuf & Nabeshima, 2008). Besides adjusting prices to reflect the real social cost of water, it

Page 15 of 20

may be important for national urbanization policies to consider the long-term distribution of population.

Third, more attention should be paid to policies on municipal solid waste, in particular its energy and climate implications. Reducing, reusing, and recycling domestic and commercial solid waste will become more energy rewarding as China urbanizes and citizens’ incomes and consumption grow. On the other hand, the high global warming potential of methane makes its capture and use in urban landfills an important part of the overall effort to reduce the urban sector’s climate impacts.

Fourth, urban transportation research should pay more attention to non-motorized transportation, its relationship to land us, and its integration with motorized transportation, especially mass transit. A common fallacy of sustainable urban transportation development is to emphasize the attractiveness of mass transit so that driving can be reduced. This omits the reality that transit draws ridership from both drivers and non-motorized travelers. This is particularly relevant in Chinese cities, where walking and cycling are often major modes of daily travel. Without enough attention to walking and cycling, our policy may create cities that consumes more energy and less healthy in the future.

Finally, very few studies look at the effects of information disclosure and public participation on sustainable urban development policies. It is not only useful to observe information disclosure and public participation’s impact on environmental policy enforcement, but also important to estimate how such policies can induce improvement in energy efficiency and voluntary behaviors.

It is clear that the United States and China are inextricably intertwined through global competition for scarce energy resources and their disproportionate impact on the Earth’s environmental health. For US policy makers, the literature on sustainable urban development in China offers two implications. On the one hand, China’s income growth, urbanization, and motorization will affect the global energy market and the global collective effort to reduce GHG emissions. Such “negative” effects are more likely to continue to escalate in the foreseeable future. On the other hand, there are a variety of actions the US can undertake to help China reduce its urban development’s energy and climate effects. China can learn a great deal from the US on environmental information disclosure, mechanisms of public participation in environmental and resource decision making, and enforcement monitoring at all levels. Encouraging transfer of green technologies such as alternative energy use and energy efficiency improvement can benefit both countries and the rest of the world.

Page 16 of 20

Reference An, Feng, and Amanda Sauer. 2004. Comparison of passenger vehicle fuel economy and

greenhouse gas emission standards around the world. The Pew Center on Global Climate Change.

Barth, Matthew, and Susan A. Shaheen. 2003. The Potential for Shared-Use Vehicle Systems in China. Institute of Transportation Studies, University of California, Davis, Research Report UCD-ITS-RR-03-11.

Browder, Greg J., & others. 2007. Stepping Up: Improving the Performance of China’s Urban Water Utilities. Washington, DC: the World Bank.

Campanella, Thomas J. 2008. The Concrete Dragon: China’s Urban Revolution and What It Means for the World. New York: Princeton Architectural Press.

Chen, Bingheng, Chuanjie Hong, and Haidong Kan. 2004. Exposures and health outcomes from outdoor air pollutants in China. Toxicology 198: 291-300.

Chen, Haiyan, Beisi Jia, and S. S. Y. Lau. 2008. Sustainable urban form for Chinese compact cities: Challenges of a rapid urbanized economy. Habitat International 32(1): 28-40.

Chen, Changhong, Bingheng Chen, Bingyan Wang, Cheng Huang, Jing Zhao, Yi Dai, and Haidong Kan. 2007. Low-carbon energy policy and ambient air pollution in Shanghai, China: A health-based economic assessment. Science of the Total Environment 373: 13-21.

Cherry, Christopher R., Jonathan X. Weinert, Xinmiao Yang. 2009. Comparative environmental impacts of electric bikes in China. Transportation Research Part D 14(5): 281-290.

Chinese Academy of Engineering (CAE), and National Research Council (NRC). 2003. Personal Cars and China. Washington, DC: The National Academies Press.

Deng, Xiangzheng, Jikun Huang, Scott Rozelle, and Emi Uchida. 2008. Growth, population and industrialization, and urban land expansion of China. Journal of Urban Economics 63: 96-115.

Ding, Chengri. 2004. Urban spatial development in the land policy reform era: Evidence from Beijing. Urban Studies 41(10): 1889-1907.

Ding, Chengri, Yan Song, Yang Zhang, Wu Cifang, Gerrit Knaap, and Terry Moore. 2009. Assessment on planning institutional arrangement for better managing urban growth in China. Lincoln Institute of Land Policy Working Paper.

Dong, Suocheng, Kurt W. Tong, and Yuping Wu. 2001. Municipal solid waste management in China: using commercial management to solve a growing problem. Utilities Policy 10(1): 7-11.

Duan, Wei. 2005. Beijing water resources and the south to north water diversion project. Canadian Journal of Civil Engineering 32(1): 159-163.

Gan, Lin. 2003. Globalization of the automobile industry in China: Dynamics and barriers in the greening of road transportation. Energy Policy 31(6): 537-551.

Girardet, H. 2004. Cities People Planet: Livable Cities for a Sustainable World. Chichester, UK: John Wiley and Sons.

Hao, Jiming, Jingnan Hu, and Lixin Fu. 2006. Controlling vehicular emissions in Beijing during the last decade. Transportation Research Part A 40(8): 639-651.

Page 17 of 20

He, Kebin, Fumo Yang, Yongliang Ma, Qiang Zhang, Xiaohong Yao, Chak K. Chan, Steven Cadle, Tai Chan, and Patricia Mulaw. 2001. The characteristics of PM2.5 in Beijing, China. Atmospheric Environment 35: 4959-4970.

He, Kebin, Hong Huo, and Qiang Zhang. 2002. Urban air pollution in China: Current status, characteristics, and progress. Annual Review of Energy and the Environment 27(1): 397.

He, Kebin, Hong Huo, Qiang Zhang, Dongquan He, Feng An, Michael Wang, Michael P. Walsh. 2005. Oil consumption and CO2 emissions in China’s road transport: current status, future trends, and policy implications. Energy Policy 33: 1499-1507.

Haddock, Ronald, and John Jullens. 2009. The best years of the auto industry are still to come. Strategy+Business 55: 2-12. Booz and Company Reprint No. 09204.

Ho, Samuel P. S., and George C. S. Lin. 2004. Converting land to nonagricultural use in China’s coastal provinces: Evidence from Jiangsu. Modern China 30(1): 81-112.

Hoornweg, D., and P. Lam. 2005. Solid waste management in China: Issues and options. Waste: The Social Context (2005) 261-271.

Kenworthy, Jeff, and Gang Hu. 2002. Transport and Urban Form in Chinese Cities: An International Comparative and Policy Perspective with Implications for Sustainable Urban Transport in China. DISP 151: 4-14.

Levine, Mark. 2008. Energy efficiency in China: Glorious history, uncertain future. American Institute of Physics Conference Proceedings 1044(1): 15-27.

Li, Jun and Michel Colombier. 2009. Managing carbon emissions in China through building energy efficiency. Journal of Environmental Management 90(8): 2436-2447.

Li, Shichao and Jon Goss. 2005. Socio-spatial solid waste recycling in post-Mao China—the case of Wuhan. Waste: The Social Context (2005) 402-411.

Lichtenberg, Erik, and Chengri Ding. 2008. Assessing farmland protection policy in China. Land Use Policy 25: 59-68.

Lichtenberg, Erik, and Chengri Ding. 2009. Local officials as land developers: Urban spatial expansion in China. Journal of Urban Economics 66: 57-64.

Liu, Huan, Kebin He, Qidong Wang, Hong Huo, James Lents, Nicole Davis, Nick Nikkila, Changhong Chen, Mauricio Osses, and Chunyu He. 2007. Comparison of vehicle activity and emission inventory between Beijing and Shanghai. Journal of the Air and Waste Management Association 57(10): 1172-1177.

Loo, Becky P. Y., and Dennis Y. N. Li. 2006. Developing metro systems in the People’s Republic of China: Policy and gaps. Transportation 33: 115-132.

McKinsey Global Institute (MGI). 2008. Preparing for China’s Urban Billion: Summary of Findings.

National Academy of Engineering (NAE), and National Research Council (NRC). 2007. Energy Futures and Urban Air Pollution: Challenges for China and the United States. Washington, DC: The National Academies Press.

NRDC. 2007. Smart Cities: Solutions for China’s Rapid Urbanization. Peng, Chaoyang, Xiaodong Wu, Gordon Liu, Todd Johnson, Jitendra Shah and Sarath

Guttikunda. 2002. Urban air quality and health in China. Urban Studies 39(12): 2283-2299.

Raufer, Roger K. 2007. Sustainable urban energy systems in China. NYU Environmental Law Journal 15(1): 161-204.

Page 18 of 20

Shalizi, Zmarak. Water and Urbanization, Chapter 7 in Yusuf, Shahid, and Tony Saich (Eds.). 2008. China Urbanizes: Consequences, Strategies, and Policies. Washington, DC: the World Bank.

Steinfeld, Edward S. Energy Policy, Chapter 6 in Yusuf, Shahid, and Tony Saich (Eds.). 2008. China Urbanizes: Consequences, Strategies, and Policies. Washington, DC: the World Bank.

Varis, Olli, & Pertti Vakkilainen. 2001. China’s 8 Challenges to Water Resources Management in the First Quarter of the 21st Century. Geomorphology 41: 93-104.

Wang, Hua, Jian Xie, & Honglin Li. 2008. Domestic water pricing with household surveys: A study of acceptability and willingness to pay in Chongqing, China. World Bank Policy Research Working Paper 4690.

Wang, Rui. 2008. Safety and environmental costs of Chinese urban transportation. Proceedings of the Transportation Research Board 87th Annual Meeting. Washington, DC.

Wang, Rui. 2008. Autos, Transit and Bicycles: Transport Choices in Chinese Cities. Harvard University Ph.D. Dissertation.

Wang, Rui. 2009. The structure of Chinese urban land prices: Estimates from benchmark land price data. Journal of Real Estate Finance and Economics 39(1): 24-38.

Wang, Xiaoping, and Denise L. Mauzerall. 2006. Evaluating impacts of air pollution in China on public health: Implications for future air pollution and energy policies. Atmospheric Environment 40(9): 1706-1721.

Weinert, Jonathan, Chaktan Ma, and Christopher Cherry. 2007. The transition to electric bikes in China: History and key reasons for rapid growth. Transportation 34: 301-318.

Winebrake, James, Sandra Rothenberg, Jianxi Luo, and Erin Green. 2008. Automotive transportation in China: Technology, policy, market dynamics, and sustainability. International Journal of Sustainable Transportation 2(4): 213-233.

World Bank, and State Environmental Protection Administration, P. R. China (SEPA). 2007. Cost of Pollution in China: Economic Estimates of Physical Damages. Washington, DC: The World Bank.

Wu, Jiaping. 2008. The peri-urbanisation of Shanghai: Planning, growth pattern and sustainable development. Asia Pacific Viewpoint 49(2): 244-253.

Xie, Jian, & others. 2009. Addressing China’s Water Scarcity: Recommendations for Selected Water Resource Management Issues. Washington, DC: the World Bank.

Yang, Jiawen. 2006. Transportation implications of land development in a transitional economy: Evidence from housing relocation in Beijing. Journal of Transportation Research Board 1954:7-14.

Yusuf, Shahid, and Kaoru Nabeshima. Optimizing urban development, Chapter 1 in Yusuf, Shahid, and Tony Saich (Eds.). 2008. China Urbanizes: Consequences, Strategies, and Policies. Washington, DC: the World Bank.

Zacharias, John. 2002. Bicycle in Shanghai: Movement patterns, cyclist attitudes and the impact of traffic separation. Transport Reviews 22(3): 309-322.

Zhang, Tingwei. 2000. Land market forces and government’s role in sprawl: The case of China. Cities 17(2): 123-135.

Zhao, Shuqing, Liangjun Da, Zhiyao Tang, Hejun Fang, Kun Song, and Jingyun Fang. 2006. Ecological consequences of rapid urban expansion: Shanghai, China. Frontiers in Ecology and the Environment 4:341–346.

Page 19 of 20

Page 20 of 20

Zhao, Xiaobin, and Li Zhang. 1995. Urban performance and the control of urban size in China. Urban Studies 32(4-5): 813-845.

Zhong, Lijin, & Arthur P. J. Mol. 2009. Water Price Reforms in China: Policy-Making and Implementation. Water Resource Management in press.

Zhou, Yu, Lixin Fu, and Linglin Cheng. 2007. Characterization of in-use light-duty gasoline vehicle emissions by remote sensing in Beijing: Impact of recent control measures. Journal of the Air and Waste Management Association 57(9): 1071-1077.

Zhu, Yingxin, and Borong Lin. 2004. Sustainable housing and urban construction in China. Energy and Buildings 36(12): 1287-1297.

Zhong, Lijin, Arthur P. J. Mol, & Tao Fu. 2008. Public-private partnerships in China’s urban water sector. Environmental Management 41:863-877.

Transportation Research Part A 40 (2006) 639–651

www.elsevier.com/locate/tra

Controlling vehicular emissions in Beijing during the last decade

Jiming Hao *, Jingnan Hu, Lixin Fu

Department of Environmental Science and Engineering, Tsinghua University, Beijing 100084, People’s Republic of China

Abstract

The vehicle population of Beijing is sharply increasing at an average annual rate of 14.5%, causing severe transportationand environmental problems. The Beijing municipal government and the public have worked hard to control vehicularemissions since 1995. Strategies and measures have been introduced to regulate land use and traffic planning, emissioncontrol of in-use vehicles and new vehicles, fuel quality improvement, introduction of clean fuel vehicle technology andfiscal incentives. New development plans for Beijing will change the transportation structure by encouraging public trans-portation. For in-use vehicles, the I/M program has employed ASM tests since early 2003 and the government has encour-aged the retirement of high-emission vehicles. For new vehicles, Beijing introduced Euro 1 and Euro 2 emission standardsin early 1999 and 2003, respectively. It is also confirmed that Euro 3 standards will be introduced in 2005. At the same time,the fuel quality in Beijing was improved significantly, by banning lead and reducing sulfur among other changes. CNG andLPG were introduced in 1999 and are used in buses and taxis. Today Beijing has the largest CNG bus fleet in the worldwith more than 2000 dedicated CNG buses. Beijing has also focused on fiscal incentives such as tax deductions for newvehicles meeting enhanced emission standards to encourage their sales. These strategies and measures have had an impacton the control of vehicular emissions. Despite the rapid increase of the vehicle population by 60% between 1998 and 2003,total vehicular emissions have not increased. With the enhancement of vehicular emission control, the air quality in Beijingis improving as the city strives to its goal for a ‘‘Green Olympics’’. 2005 Elsevier Ltd. All rights reserved.

Keywords: Air pollution; Vehicular emission control; Emission standards; Beijing

1. Introduction

During the past 20 years, the urbanization rate in China has increased rapidly from 20% in 1980 to 38% in2002, which is double the world average rate increase (World Bank, 2004). With urbanization, comes a moremobile society. Vehicle growth has risen as indicated by an average annual rate (AAR) of 8.7% in China since1985 (SSB, 2002). In metropolises such as Beijing, vehicle population is rising more sharply, especially for cars.Since 1990, vehicles in Beijing increased at an AAR of 14.4% and will soon exceed 2.4 million, causing heavytraffic congestions and serious air pollution. The annual rate of increase of the car population alone is 19.1%(BJTAB, 1995-2004).

0965-8564/$ - see front matter 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.tra.2005.11.005

* Corresponding author. Tel.: +86 10 6278 2195; fax: +86 10 6277 3650.E-mail address: hjm-den@tsinghua.edu.cn (J. Hao).

640 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

In 1999, the annual average concentrations of ambient PM10, NOx and CO were 0.180, 0.140 and 2.9 mg/m3, respectively, in Beijing. PM10 pollution has exceeded the limits of national air quality standards by 80%. Inaddition, there were 119 days and 777 h of ozone violation (BJEPB, 1999, 2000). Vehicular emissionsaccounted for 46%, 78% and 83% of the total NOx, CO and HC emissions, respectively, in the Beijing urbanarea. Of this, cars were the largest contributor, accounting for 35%, 56% and 44% of the total vehicular NOx,CO and HC emissions (Fu et al., 2000). Vehicular emissions comprised 68.4% and 76.5% of the ambient NOx

and CO concentrations (Hao et al., 2001a). For airborne PM10, 13.6% and 32.8% of the ambient concentra-tion came from vehicle tailpipe emissions and road fugitive dust, respectively (Tsinghua University, 2001).Fine particulate (PM2.5), which has larger adverse effects than PM10, and photochemical smog are also affectedby vehicular emissions, even more significantly than PM10 (Kittelson, 1998; Nevers, 2000). Modeling indicatesthat 80.2% of the PM10 from vehicular tailpipes is PM2.5 in Beijing, while the proportion of PM2.5 in ambientPM10 is 55% on average (Wu et al., 2002; Yang et al., 2002). Source apportionment of ambient PM2.5 in Bei-jing showed that the vehicular emissions are a major source of the elemental carbon and organic compounds inthe PM2.5 composition (Song et al., 2002).

Beijing’s vehicle population growth is not likely to be reduced soon. Effective strategies and measures tocontrol vehicle pollution have taken in two routes, one of which is to decrease the transportation demandand vehicle mileage traveled, and the other to reduce the emissions per vehicle. Most of the strategies and mea-sures have been initiated since 1999, and together they make up the enhanced control of total vehicularemissions.

2. Control strategies and measures

The strategies and measures carried out to control vehicular emissions in Beijing, involve six categories.They are land use and traffic planning, emission control of in-use vehicles, control of new vehicles, fuel qualityimprovement, clean fuel vehicle technology and fiscal incentives.

2.1. Land use and traffic planning

The average personal trips were 2.82 trips per person per day and the total personal trips were about 40million trips daily in Beijing in 2002. Public transportation and cars (private cars, government vehicles andtaxis) accounted for 31% and 16%, respectively. Bicycles were still selected for 50% of the personal trips. Sub-way and light rail only accounted for 9.8% of public transportation—the remainder was by bus (BJTRC,2004). The number of trips and the modes of travel heavily impact ground transportation, and the resultingsevere traffic jams causing a large proportion of idling thereby increasing vehicular emissions.

Most of the personal trips occur during the day. The rush hours peak at 7:00–8:00 in the morning and17:00–18:00 in the afternoon, and personal trips in these 2 h accounted for 28.6% of the total trips in the24-h period (BJTRC, 2004). Fig. 1 gives the hourly variation of personal trips during a day, in average, in2002.

Proper land use can decrease traffic demand, so as to abate the vehicular emissions by reducing vehicle mile-age traveled. It can also improve traffic conditions. Over the past half century, ‘‘urban sprawl’’ was the dom-inant form of urban growth for mega cities worldwide (Southworth, 2001). The Beijing municipal governmenthas paid particular attention to this phenomenon in recent years and the new city plans, finalized in 2004,reflect this. These plans anticipate converting the land use map of Beijing from a single-centered style to‘‘two axles, two corridors with multi-centered’’ style. Fig. 2 shows the current map and the new city planin Beijing (BMICPD, 2004). The new planning is expected to decrease traffic demand, especially for commut-ing trips. The total personal trips are expected to be from 52 to 55 million trips daily in Beijing in 2020, whichmeans the average personal trips would be about 2.90 trips per person per day as the total population of Bei-jing would increase from 14 million in 2002 to 18 million in 2020. Since the economy and vehicle ownership areexpected to keep increasing during the next decade, it would be significantly hard to keep the average personaltrips only 3% higher than the present.

Increasing the share of public transportation, especially subway and light rail, will lower stress of groundtransportation. To accelerate the conversion of the personal trip mode to public transportation, more subways

Fig. 2. Land use map up-to-date vs. the new city planning in Beijing.

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

0:00-5:00

6:00-7:00

8:00-9:00

10:00-11:00

12:00-13:00

14:00-15:00

16:00-17:00

18:00-19:00

20:00-21:00

22:00-23:00

Hour

Per

cent

age

Fig. 1. Hourly variation of personal trips in a typical day in Beijing in 2002.

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 641

and light railways are under construction in Beijing. Their total length will reach 300 km before 2008. Accord-ing to the development planning for the 2008 Olympic Games, the bus population per ten thousand persons inBeijing will increase to more than 12, and the total bus population would increase from 16 thousand in 2002(BSB, 2003) to 18–20 thousand (Liu, 2003). Bus rapid transit (BRT), a moderately rapid mass transit systemwith much lower infrastructure investment than subway or light rail, is also under development.

Besides, enhancement of the construction of minor arterial roads, road branches and intersections areimportant to distribute the traffic and improve the ground transportation. At present, much of the traffic con-gestion during the rush hours is due to traffic volume overwhelming the capacity of some exits of expresswaysand intersections in Beijing urban area.

2.2. Emission control of in-use vehicles

In-use vehicles are the sources of exhaust emission discharge. Although all the measures to control vehic-ular emissions are directly toward limiting or decreasing the emissions from in-use vehicles ultimately, thereare some specific measures to control in-use emissions that will be described.

Inspection/maintenance (I/M) program is the most effective means for emission reduction of in-use vehi-cles. For any I/M program, the quality of program implementation is most important to provide accurate

642 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

emission data and to prevent fraud. An I/M program without quality assurance wastes money and time.Beijing started pilot I/M programs with two speed idle tests in 1995, and the complete I/M programs havebeen enforced since 1999. According to the studies by Tsinghua University about these programs, CO andHC emissions from in-use vehicles can be significantly reduced by I/M programs performed in Beijing.Fig. 3 shows the reduction rates of CO and HC emissions at idle and for high speed idle tests, respectively(Hao et al., 2001b).

To strengthen NOx emission control, Acceleration Simulation Mode (ASM) testing, which more closelyrepresents real on-road conditions than idle, was initiated in early 2003 in Beijing. By the end of 2003, therewere 175 ASM test lanes in use. Fig. 4 shows the cut points of ASM standards in Beijing and start-up ASMstandards in the United States (BJEPB and BJBQTC, 2003; USEPA, 2004). For light-duty vehicles, the cutpoints are very stringent in Beijing because the control technologies applied on newer vehicles in Beijing since1999 were introduced in the US in early 1980s and mid-1990’s (corresponding to Tier 0 standards and Tier 1standards). According to the data from the Beijing Environmental Protection Bureau (BJEPB), the failure rateof light-duty vehicles exceeded 30% during the first month of required ASM testing. The very high failure rateand crowded testing and repair stations encouraged vehicle owners to seek shortcuts to pass the emission tests,while vehicle tampering reduced the effectiveness of ASM tests significantly.

Accelerating the retirement of older, high-emitting in-use vehicles is another effective way to reduce emis-sions. In Beijing, PM and NOx emissions of an old pre-Euro 1 diesel bus are more than 6 times and 1.5 times,respectively, the emissions of a new diesel bus meeting Euro 3 standards. The Beijing municipal governmenthas introduced 160 new diesel buses meeting Euro 3 standards in 2002 that operate on low sulfur diesel fuel,replacing the older diesel buses. Light-duty vehicles that cannot meet Euro 1 standards are required to beinspected biannually and have been banned since late 2003 from operating within the 2nd ring expressway,which represents the central core of the city. The high cost of scheduled inspection and the restrictions in usageencourage the owners to retire the old vehicles early. Taxis have also been targeted for special attention. Theemissions of a taxi are 9.5 times of that from a similar private car in Beijing because of the taxi’s higher annualmileage which averages 4.5 times the miles driven by a private car, and because of the higher deterioration rateof emission controls. Mandatory retirement of a taxi is required when it is 6–8 years old in Beijing, much morestringent than the nationwide requirement of mandatory retirement of vehicles at 15 years. A retrofit programto convert gasoline taxies to gasoline and LPG bi-fuel cars was carried out between 1999 and 2001 and will bedescribed in more detail below.

Applying retrofits to in-use vehicles has also been used to reduce emissions. There was a large retrofit pro-gram for light-duty vehicles in Beijing between 1998 and 2001. About 190,000 carbureted vehicles registeredafter 1995 were retrofitted with a three way catalyst (TWC), supplemental air and close-loop controls (BJEPB,2001). Test data indicated that CO emissions were reduced by 78–90% while HC + NOx emissions were low-ered by 71–88% immediately after the retrofit. Most of these retrofitted emission control systems, however,were only durable for a short time (warranted by the manufacturer for 2 years or 50,000 km). Many of theretrofitted vehicles become high emitters again on the road.

0%

20%

40%

60%

80%

100%

CO at HighSpeed Idle

CO at Idle HC at HighSpeed Idle

HC at Idle

Per

cent

age

Before Maintenance

After Maintenance

Fig. 3. Reduction rates of CO and HC emissions by I/M programs with two speed idle tests in Beijing. (Note: Accumulativeconcentrations of CO emissions and HC emissions of 2452 tested vehicles at idel tests are set as 100%, respectively.)

HC

0

100

200

300

400

500

1975 1979 1983 1987 1991 1995 1999 2003Model Year

HC

Cut

Poi

nt (p

pmC

6)

US

Beijing

Euro 1StandardsIntroducedin 1999

New StandardsIntroduced in 1980

CO

0

0.5

1

1.5

2

2.5

3

1975 1979 1983 1987 1991 1995 1999 2003Model Year

CO

Cut

Poi

nt (%

)

US

Beijing

Euro 1StandardsIntroducedin 1999

New StandardsPhased in 1981-1983

NOx

0

1000

2000

3000

4000

1975 1979 1983 1987 1991 1995 1999 2003

Model Year

NO

x C

ut P

oint

(ppm

NO

) US

Beijing

New StandardsIntroduced in 1981

Euro 1StandardsIntroducedin 1999

Fig. 4. Cut points for light-duty vehicles set in ASM standards in Beijing vs. US. Equivalent test weight is 3500 lbs (1588 kg) andASM2525 is used. (ASM5015 is used for pre-conditioning and first decision.)

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 643

2.3. Emission standards for new vehicles

Implementation of more stringent emission standards for new vehicles is the primary way to lower vehicleemissions in a city. The Beijing municipal government introduced Euro 1 emission standards (referred toas Phase 1 standards) in 1999 and Euro 2 standards (Phase 2) in 2003. These standards were implemented

Table 1Comparison of emission standards under different phases

Vehicle type Emission standards Cut points

CO HC + NOx HC NOx PM

Gasoline cars, GVW <2.5 t (g/km) Pre-Phase 1 (ECE15) 21.47 – 3.16 2.52 –Phase 1 (ECE15 + EUDC with 40 s idling start) 3.16 1.13 – – –Phase 2 (ECE15 + EUDC with 40 s idling start) 2.20 0.50 – – –Euro 3 (ECE15 + EUDC w/o 40 s idling start) 2.30 – 0.20 0.15 –Euro 4 (ECE15 + EUDC w/o 40 s idling start) 1.00 – 0.10 0.08 –

Heavy-duty diesel engines,>85 kW (g/kW h)

Pre-Phase 1 Only smoke tested at full load, filter smokenumber (FSN) <4.0

Phase 1 (ECE R-49) 4.5 – 1.1 8.0 0.36Phase 2 (ECE R-49) 4.0 – 1.1 7.0 0.15Euro 3 (ESC) 2.1 – 0.66 5.0 0.10Euro 4 (ESC) 1.5 – 0.46 3.5 0.02

644 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

nationwide in 2000 and 2004, respectively. Euro 3 and Euro 4 emission standards are going to be introduced inBeijing in 2005 and 2008, respectively, while Euro 3 standards may extended to all of China in 2007. On-boardDiagnosis (OBD) is required on vehicles meeting Euro 3 standards—it has the potential to play an importantpart in the I/M program of in-use vehicles. The Beijing municipal government is considering one-year waiverfor OBD installation at the beginning of Euro 3 emission standards.

Table 1 gives the emission limits for Pre-Phase 1, Phase 1, Phase 2 and Euro 3, Euro 4 emission standardsfor gasoline cars and heavy-duty diesel engines. Despite the difference between the driving cycles employed byPre-Phase 1 and Phase 1 standards, it can be seen that the absolute values of these emission limits for gasolinecars decrease significantly from Pre-Phase 1 to Phase 1 standards: about 85% for CO and 80% for HC + NOx

emissions. The emission limit comparison is confounded by a change in the driving cycles from ECE15 toECE15 + EUDC with 40 s idling start before emissions are collected. The maximum and average speeds ofECE15 are 50 km/h and 18.7 km/h, much lower than that of EUDC, which are 120 km/h and 62.6 km/h,respectively. Therefore, it is understandable that CO and HC emissions would fall and NOx emission wouldrise in EUDC compared to ECE15 when testing the same car. From Phase 1 to Phase 2 standards, the limits ofHC + NOx emissions decreases about 50% further, while CO is only 20% lower.

For heavy-duty diesel engines, before the implementation of Phase 1 standards only smoke was tested atfull engine load. There was no emission control of the gaseous pollutants. From Phase 1 to Phase 2 standards,the limits of PM decrease most significantly, by 58%, while that of CO and NOx by only about 12%. All of theabove indicates that there is a substantial reduction from Pre-Phase 1 to Phase 1, which is much greater thanthe reductions from Phase 1 to Phase 2.

Euro 3 and Euro 4 emission standards cannot be compared to Phase 1 and Phase 2 standards directlybecause of the difference in driving cycles. Euro 4 standards are much more stringent than Euro 3 standards.The emission limits for gasoline cars decrease by about 50%, while heavy-duty diesel engines’ limits for gas-eous pollutants decease by 30% and by 80% for PM.

At the beginning of Beijing Phase 1 standards, the emission durability distance was 50,000 km (raised to80,000 km in 2000), and 80,000 km for Phase 2 standards. The durability requirements for nationwide Phase1 and Phase 2 standards are both 80,000 km. For Euro 4 standards, the emission durability requirement willincrease to 100,000 km.

To promote the sales of lower emission vehicles, an excise tax deduction was provided for new light-dutyvehicles meeting Phase 2 standards during late 2001–2003 in Beijing. This is described further below.

2.4. Fuel quality improvement

Fuel quality has close relationship to emission control technologies. The more important indices reflectinggasoline quality are octane number, Reid vapor pressure (RVP), sulfur and lead content and the percentage ofolefins, aromatics and oxygenate. Diesel fuel quality is impacted by sulfur content, cetane number, and the

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 645

percentage of aromatics and additives. Generally, sulfur content in gasoline and diesel fuel affects vehicularemissions the most. For gasoline fuel, RVP also plays a very important role in HC emission control andthe lead emission is closely correlated to lead content. Tables 2 and 3 provide a simplified summary of theimprovement of Chinese fuel product specifications during the last decade. Gasoline specifications wererevised in 1999 to remove lead for Beijing and for other major cities of Shanghai and Guangzhou by July2000. This standard has been implemented nationwide at the beginning of 2003. For light diesel fuel, eachpour point grade of diesel fuel was further classified by allowable sulfur content into premium, first leveland qualified until the new specifications went into effect in 2002. These distinctions were abolished in thenew specifications and the more restrictive sulfur specification of the premium diesel was adopted for allpour-point grades (Fu et al., 2002).

The adoption of stringent vehicle emission standards requires fuel quality improvement. Fig. 5 gives thelimits of sulfur content in gasoline and diesel required by the vehicle technologies to meet the emission stan-dards. High sulfur content in the fuel would poison the catalyst causing it to lose its effectiveness. Fig. 5 showslarge reduction in sulfur from the base case to Euro 2 requirements and from Euro 3 to Euro 4 requirements.

In Beijing, sulfur content in gasoline ranges from 300 ppm to 500 ppm at present, and from 500 ppm to800 ppm for diesel fuel. Sulfur content will fall further with the coming implementation of Euro 3 standardsto 150 ppm for gasoline and 350 ppm for diesel. The amount of olefins in the gasoline will also decrease.

The control of hazardous components in the fuel decreased the adverse impacts on human health directly.Unleaded gasoline has been required nationwide for several years. In Beijing, leaded gasoline was successfullyphased out by 1998 by the local government.

2.5. Clean fuel vehicle technology

Clean fuel vehicles are those vehicles meeting the emission standards more stringent than currentlyrequired. In the US, vehicles meeting Low Emission Vehicle (LEV), Ultra-Low Emission Vehicle (ULEV)and Zero Emission Vehicle (ZEV) standards have qualified as clean fuel vehicles, without restricting the fuelsthat the LEVs and ULEVs used.

Table 2Fuel quality standards of unleaded gasoline in China

Index Start year

1993a 2000/2003b

Octane (RON) min 90 93 97 90 93 97Lead (g/L) max 0.013 0.013 0.013 0.005 0.005 0.005Sulfur (ppm) max 1500 1500 1500 800 800 800Olefins (%) max – – – 35 35 35Aromatics (%) max – – – 40 40 40Benzene (%) max – – – 2.5 2.5 2.5Oxygenates (%) max – – – 2.7 2.7 2.7

a Standard no. SH0041-93.b Standard no. GB17930-99.

Table 3Fuel quality standards of light diesel in Chinaa

Index Start year

1994b 2002c

Grade Premium First Qualified Premium First QualifiedCetane 45 45 45 45Sulfur (ppm) max 2000 5000 10,000 2000 (abolished)T95 (C) max 365 365 365 365

a Only the specifications of zero-degree-Celsius pour-point (the most popular) grade of light diesel are given.b Standard no. GB252-94.c Standard no. GB252-00.

Fig. 5. Sulfur content limitation in gasoline and diesel required by emission standards.

646 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

When developing clean fuel vehicle technology, cleaner gasoline and diesel powered vehicles are encouragedat first. Dedicated CNG/LNG heavy-duty vehicles have lower PM emissions than the diesel buses and trucksoperating in Beijing, and lower NOx emissions with some power loss and less emissions of smog-producingsubstances. Table 4 gives some estimated emission factors of dedicated CNG bus and diesel buses meetingPre-Phase 1 and Euro 3 emission standards in Beijing. Though PM emissions of Euro 3 diesel bus are 84%lower than Pre-Phase 1 bus, the PM is still 8 times of that of a CNG bus. Total HC of a CNG bus is 3.1 timesof that of a Euro 3 diesel bus, but more than 90% of the HC is non-reactive methane. Today Beijing has thelargest CNG bus fleet worldwide with more than 2000 dedicated CNG buses. The transit operators are satis-fied with these buses except for the higher costs and occasional lack of power when overloaded.

Dedicated LPG cars have almost same regular emissions as Phase-2 gasoline cars in Beijing, and loweremissions of toxics and smog-producing substances. There are 600 dedicated LPG taxis and about 32,000 gas-oline and LPG bi-fuel taxis in Beijing. About 22,000 of these were retrofitted from carbureted cars during1999–2001 and the remainders were purchased from the car manufacturers with electronic fuel injection(EFI) and a TWC system for using gasoline. The dedicate LPG taxis show good performance on the road,with HC, CO and NOx emissions close to those of gasoline cars meeting Phase 2 emission standards. The ret-rofitted taxis, however, show little emission reduction. Some conversions even had an adverse effect on theengine power. The performance of about 10,000 originally manufactured bi-fuel taxis was better than the ret-rofitted taxis, but using LPG did not lower the emissions for there were no EFI systems for LPG. Only lessthan 30% of the bi-fuel taxi drivers are currently using LPG.

Pure electric vehicles are Zero Emission Vehicles (ZEVs) if only vehicular emissions are considered; sincethe battery problem—limited range and short life—has not been resolved, the market has begun to shrink.

Table 4Estimated emission factors of diesel buses and CNG bus

Pollutant Emission factors (g/km)

Pre-Phase 1 diesel Euro 3 diesel CNG

PM 2.50 0.40 0.05NOx 30.0 20.0 9.4HC 4.5 2.0 6.2CO 7.8 5.0 5.3

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 647

It is not feasible developed ZEVs on a big scale—now there is a fleet of about 20 pure electric vehicles used fordemonstration in Beijing. Hybrid electric vehicles (HEVs) have lower emissions than gasoline vehicles and sig-nificantly better fuel economy. The light-duty HEVs are commercialized, but the cost is still high for heavy-duty HEVs. Light-duty HEVs may be promoted in Beijing, as well some pilot hybrid electrical buses. Fuel cellvehicles (FCVs) are also ZEVs, if only vehicular emissions are considered, but the current cost is still too high.At present, only several FCVs are planned for operation in Beijing before 2008, in a pilot program funded byGlobal Environmental Foundation.

2.6. Fiscal incentives

From the purchase of new vehicles to the retirement of older vehicles, fiscal incentives can be a part ofthe emission control strategy. Fig. 6 shows the fiscal incentives that can be employed during the life of acar (Holman, 2001). When people purchase new cars, financial subsidies can be applied to low-emissionvehicles to encourage their selection. For in-use vehicles, subsidies can be made available to promote theretrofit program when it is feasible and effective. An increase in local parking fees or the charge of a congestionfee can relieve the transportation stress in the central urban area and reduce the vehicular emissions. For oldervehicles with high emissions, the government can give the vehicle owners a financial bonus to encourage earlyscrappage.

In early 1999, the taxes on leaded gasoline were increased to make its price no lower than unleaded gasoline,as part of the accelerated phase-out of leaded gasoline.

In mid-2000, the State Ministry of Finance and Administration of Taxation issued a notice that a 30%reduction in the excise tax, which is 5% of the vehicle price for a typical car in China, would apply to the pur-chase of light-duty vehicles meeting Phase 2 emission standards which were not required in Beijing until 2003.It has been put into effect by the end of 2001 and was successful in promoting the early sales of low-emission

Fig. 6. Fiscal incentives to be employed during the life of a car.

648 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

vehicles. A similar excise tax deduction will be given to vehicles meeting Euro 3 emission standards in the nearfuture.

3. Impact assessment of the control activities

Table 5 provides the annual average concentration (AAC) of ambient NO2, CO and PM10 in Beijing from1998 to 2003. The AAC of ambient NO2 remained within 0.07–0.08 mg/m3 since 1998, just meeting the0.08 mg/m3 limit of the national standard. Contemporarily, the AAC of ambient CO fell gradually in Beijing,from 3.3 mg/m3 in 1998 to 2.4 mg/m3 in 2003. CO pollution has become less of a concern during the last 5years. PM10 is the most serious pollutant and also most difficult to control in Beijing. The AAC of ambientPM10 fell from 0.180 mg/m3 in 1999 to 0.141 mg/m3 in 2003, due, in part, to vehicular emission control. Fur-thermore, ozone violation hours decreased from 777 in 1999 to 151 in 2002, but rebounded to 255 in 2003,which indicates that ozone pollution remains serious (BJEPB, 2002, 2003).

When assessing the strategies and measures for controlling vehicular emissions, a vehicle emission inven-tory must first be developed. Tsinghua University in Beijing used basic traffic data and modified MOBILE5and PART5 models to account for Beijing conditions in estimating the emission factors by vehicle type. Then,combining the average annual mileage traveled and registration data by vehicle type and age, the emissioninventory can be established. Furthermore, to give the spatial and temporal variation, a bottom-up methodis used to estimate emission rates of line sources, and a top-down method to estimate emission rates of areasources. The emission inventory of mobile sources in Beijing, however, is not sufficiently accurate. It does notinclude the latest development of population centers, roads and traffic patterns and so must be improved andupdated.

Based on the prepared emission inventory, ISCST3 model developed by USEPA was employed to simulatethe spatial and temporal variation of primary pollutant concentrations in Beijing. ISCST3 is an Eulerianmulti-source pollution dispersion model. Different mathematical methods are used for different source types.The simulation result of the spatial distribution of CO concentrations in Beijing showed that CO concentra-tions were dominated by vehicular emissions. In 87.4% of the area, the contribution rates of vehicular emis-sions are from 70% to 90%. In the central urban area, the CO contribution rate has risen to 86.3% (Hao et al.,2001b).

When assessing the environmental impact of control strategies, a study carried out by Tsinghua Universitydesigned three scenarios for Beijing, including No Control Scenario, Business As Usual (BAU) Scenario andEnhanced Scenario. The vehicular emissions in the Beijing urban area were compared for different years underthe three scenarios. Table 6 lists the strategies and measures adopted or to be implemented in each scenario.They are sorted into four kinds. When estimating the emission reduction of these measures, we used differentmethods. Modified MOBILE5 and PART5 models can simulate the vehicular emission factors with the intro-duction of new vehicle emission standards in different years so its affect on emission reduction was estimated.These models can also give emission factors for a specific type of vehicles at different ages, which can be usedto estimate the emission reduction by controlling the older vehicles with high emissions. There is no specialmodel to simulate the emission factors of alternative fuel vehicles so these factors came from limited experi-mental data. Concerning the measures to improve public transportation and limit the use of private cars, we

Table 5Annual average concentrations of ambient pollutants in Beijing, 1998–2003

Year Annual average concentrations (mg/m3)

NO2 CO PM10 TSP

1998 0.074 3.3 – 0.3781999 0.077 2.9 0.180 0.3642000 0.071 2.7 0.162 0.3532001 0.071 2.6 0.165 0.3702002 0.076 2.5 0.166 0.3732003 0.072 2.4 0.141 0.252

Table 6Measures to control vehicular emissions under the three scenarios

Strategies and measures Scenario

No control BAU Enhanced

Emission standards for new vehicles andI/M programs for in-use vehicles

None Phase 1 in January 1999,Phase 2 in January 2003,Phase 3 in July 2007;ASM test in January 2003

Phase 1 in January 1999,Phase 2 in January 2003,Phase 3 in July 2005,Phase 4 in July 2008;ASM test in January 2003

Control of high emission vehicles None High emission vehicles restrictedand inspected biannually since 2003

High emission vehicles restrictedand inspected biannually since 2003;remote sensing since 2006

Application of alternative fuel vehicles None 2000 CNG buses and 800trolley buses in 2008

6000 CNG buses, 800 trolley busesand 3000 HEVs in 2008

Improving public transportation andlimiting the use of private cars

None 114 km of subway and light railand 15,000 buses in 2008

201 km of subway and light rail and20,000 buses in 2008;taxi management improved;congestion fee charged andparking fee increased

Table 7Vehicular emission reduction by the control measures in BAU and Enhanced Scenarios compared to No Control Scenario

Scenario Control measures 2002 2008

CO (t/yr) NOx (t/yr) CO (t/yr) NOx (t/yr)

BAU Emission standards and I/M programs 62.0 2.2 135.3 6.4Control of high-emission vehicles 0 0 2.6 0.4Application of alternative fuel vehicles 0.0 0.1 0.1 0.1Improving public transportation 0 0 1.5 0.2

Enhanced Emission standards and I/M programs 62.0 2.2 136.6 7.4Control of high-emission vehicles 0 0 5.8 0.6Application of alternative fuel vehicles 0.0 0.1 0.1 0.4Improving public transportation and limiting the use of private cars 0 0 12.2 1.2

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 649

used the vehicle kilometers traveled (VKT) by vehicle type, which were estimated by Beijing TransportationResearch Centre with TRIPS and VISUM models, and emission factors given by modified MOBILE5 andPART5 models with the vehicle average speed increased to estimate the vehicular emission reduction. Table7 gives the estimated vehicular emission reduction by the four kinds of measures in BAU and Enhanced Sce-narios compared to No Control Scenario, in which a negative number means an emission increase. It is seenthat emission standards for new vehicles and I/M programs for in-use vehicles have the most significant con-tribution to vehicular emission reduction in Beijing between 1998 and 2008, followed by improving publictransportation and limiting the use of private cars between 2002 and 2008. Summing up the effectiveness ofall the existing and proposed measures, Fig. 7 shows the total vehicular emission reduction under the scenar-ios. There is a large difference between no control and the control scenarios, with smaller differences amongthe control scenarios. The continued implementation of control strategies and measures are reversing the trendof vehicular emissions before 1998, which is consistent with the air quality improvement in Beijing during thelast 5 years.

4. Look into the future

As an international metropolitan area under rapid development, Beijing is facing a sharp rise of its vehiclepopulation, which is definitely determined by the increase of GDP per capita. Using Gompertz curve (S shape)

CO

0

50

100

150

200

250

1995 1998 2002 2008Year

No control

BAU

Enhanced

NOx

0.0

4.0

8.0

12.0

16.0

20.0

1995 1998 2002 2008Year

Em

issi

ons

(104

t)E

mis

sion

s (1

04t)

No control

BAU

Enhanced

a

b

Fig. 7. Potential reduction under different control scenarios of vehicular emissions.

650 J. Hao et al. / Transportation Research Part A 40 (2006) 639–651

(Dargay and Gately, 1999) and based on the development of population and per capita GDP in Beijing, thevehicle population has been forecasted by Tsinghua University. Fig. 8 shows this prediction; it can be seen thatthe vehicle population in Beijing may reach 3.3 million in 2008 and 5.0 million in 2020. The fastest increaseoccurs around 2005.

The quality of living in Beijing is dependent on whether effective strategies and measures for vehicular emis-sion control are implemented. Control of vehicular emissions was initiated in 1995, and it took 10 years for

0

100

200

300

400

500

600

1990 1995 2000 2005 2010 2015 2020

Year

Veh

icle

Pop

ulat

ion

(×10

4 )

Cars Other Vehicles

Fig. 8. Vehicle population in Beijing, past, present and future.

J. Hao et al. / Transportation Research Part A 40 (2006) 639–651 651

Beijing to put in place a vehicle emission control system that western countries developed in 20 years. Thisaccomplishment is due to an aggressive attitude of China’s leadership to lower vehicle emissions coupled withthe experience imported from the west. Existing strategies and measures have had an obvious affect on thecontrol of vehicular emissions. Despite the rapid increase of the vehicle population by 60%, the vehicular emis-sions in 2003 were not greater than in 1998. Because of the substantial effort to control vehicular emissions inrecent years, the air quality in Beijing is improving at the same time that the city undergoes rapid development.With persistent and enhanced efforts focused on vehicular emission control, the air quality in Beijing will con-tinue to improve and fulfill the objective of a ‘‘Green Olympics’’.

Acknowledgement

The authors thank Dr. Charles N. Freed, who retired from US EPA, for his great help to improve ourpaper.

References

Beijing Environmental Protection Bureau (BJEPB), Beijing Bureau of Quality and Technical Supervision (BJBQTC), 2003. Emissionstandard for exhaust pollutants from gasoline vehicles under steady-state loaded mode, DB 11/122-2003. China’s Standards Press,Beijing.

Beijing Environmental Protection Bureau (BJEPB), 1999–2003. Gazette of environmental conditions in Beijing, 1999–2003. BJEPB,Beijing. Available from: <http://www.bjepb.gov.cn>.

Beijing Traffic Administration Bureau (BJTAB), 1995–2004. Data collected by investigation (informally released), Beijing.Beijing Transportation Research Center (BJTRC), 2004. Beijing transport annual report, 2003. BJTRC, Beijing (in Chinese).Beijing Municipal Institute of City Planning and Design (BMICPD), 2004. Study of development strategies on city space of Beijing,

BMICPD, Beijing.Beijing Statistical Bureau (BSB), 2003. Beijing Statistical Yearbook, 2003. China Statistical Press, Beijing (in Chinese).Dargay, J., Gately, D., 1999. Income’s effect on car and vehicle ownership, worldwide: 1960–2015. Transportation Research Part A 33,

101–138.Fu, L.X., Hao, J.M., He, D.Q., He, K.B., 2000. The emission characteristics of pollutants from motor vehicles in Beijing. Environmental

Science 21 (3), 68–70 (in Chinese).Fu, L.X., Zhou, Y., Hao, J.M., 2002. Fuel quality in China and its implications for energy security and the environment. Paper for the

workshop on China’s oil security, auto fuel quality, and policy solutions, Hangzhou, China.Hao, J.M., Wu, Y., Fu, L.X., He, K.B., He, D.Q., 2001a. Motor vehicle source contributions to air pollutants in Beijing. Environmental

Science 22 (5), 1–6 (in Chinese).Hao, J.M., Fu, L.X., He, K.B., Wu, Y., 2001b. Pollution control of urban vehicular emissions. Environmental Science Press of China,

Beijing (in Chinese).Holman, C., 2001. Economic Instruments: EU Experience. Presentation in Conference of EU-SEPA Cooperation Project, Beijing.Kittelson, D.B., 1998. Engines and nanoparticles: a review. J. Aerosol Sci. 29 (5/6), 575–588.Liu, Q. (Ed.), 2003. Study of Beijing Olympics’ Economy. Beijing Press, Beijing (in Chinese).Nevers, N., 2000. Air Pollution Control Engineering, second ed. Tsinghua University Press & McGraw-Hill Publishers, Beijing.Song, Y., Tang, X.Y., Fang, C., Zhang, Y.H., Hu, M., Zeng, L.M., 2002. Source apportionment on fine particles in Beijing.

Environmental Science 23 (6), 11–16 (in Chinese).Southworth, F., 2001. On the potential impacts of land use change policies on automobile vehicle miles of travel. Energy Policy 29, 1271–

1283.State Statistical Bureau (SSB), 2002. China Statistical Yearbook, 2002. China Statistical Press, Beijing (in Chinese).Tsinghua University, 2001. The research report of comprehensive strategies for air pollution control in Beijing. Special project supported

by MOST, Beijing. (in Chinese).US Environmental Protection Agency (USEPA), 2004. Acceleration simulation mode test procedures, emission standards, quality control

requirements, and equipment specifications: final technical guidance, EPA420-B-04-011. US Government Printing Office, Washington,DC.

World Bank, 2004. 2004 World development indicators. World Bank, Washington, DC. Available from: <http://www.worldbank.org/data/wdi2004/>.

Wu, Y., Hao, J.M., Li, W., Fu, L.X., 2002. Calculating emissions of exhaust particulate matter from motor vehicles with PART5 model.Environmental Science 23 (1), 6–10 (in Chinese).

Yang, F.M., He, K.B., Ma, Y.L., Zhang, Q., Yu, X.C., 2002. Variation characteristics of PM2.5 concentration and its relationship withPM10 and TSP in Beijing. China Environmental Science 22 (6), 506–510 (in Chinese).

lable at ScienceDirect

Journal of Environmental Management 90 (2009) 2436–2447

Contents lists avai

Journal of Environmental Management

journal homepage: www.elsevier .com/locate/ jenvman

Managing carbon emissions in China through building energy efficiencyq

Jun Li a,*, Michel Colombier b

a CERNA, Ecole des Mines de Paris–Mines ParisTech, 60, boulevard Saint-Michel, 75272 Paris, Franceb Institut du developpement durable et des relations internationales, 13, rue de l’universite, 75007 Paris, France

a r t i c l e i n f o

Article history:Received 29 April 2008Received in revised form26 November 2008Accepted 21 December 2008Available online 3 April 2009

Keywords:Building energy efficiencyCarbon emissions mitigationPolicy instruments

q The authors thank three anonymous refereessuggestions and are grateful to Jean Lin who helpedauthors are alone responsible for any remaining erro

* Corresponding author. Tel.: þ33145 497 678; fax:E-mail addresses: junli@ensmp.fr (J. Li),

(M. Colombier).

0301-4797/$ – see front matter 2008 Elsevier Ltd.doi:10.1016/j.jenvman.2008.12.015

a b s t r a c t

This paper attempts to analyse the role of building energy efficiency (BEE) in China in addressing climatechange mitigation. It provides an analysis of the current situation and future prospects for the adoptionof BEE technologies in Chinese cities. It outlines the economic and institutional barriers to large-scaledeployment of the sustainable, low-carbon, and even carbon-free construction techniques. Based ona comprehensive overview of energy demand characteristics and development trends driven byeconomic and demographic growth, different policy tools for cost-effective CO2 emission reduction in theChinese construction sector are described. We propose a comprehensive approach combining buildingdesign and construction, and the urban planning and building material industries, in order to drasticallyimprove BEE during this period of rapid urban development. A coherent institutional framework needsto be established to ensure the implementation of efficiency policies. Regulatory and incentive optionsshould be integrated into the policy portfolios of BEE to minimise the efficiency gap and to realisesizeable carbon emissions cuts in the next decades. We analyse in detail several policies and instruments,and formulate relevant policy proposals fostering low-carbon construction technology in China.Specifically, Our analysis shows that improving building energy efficiency can generate considerablecarbon emissions reduction credits with competitive price under the CDM framework.

2008 Elsevier Ltd. All rights reserved.

1. Introduction

The next ten years will be critical for the world in light of thelong-term GHG emissions stabilisation scenario (490 ppmv CO2) inthe IPCC’s fourth report, requiring a reduction in global emissionsfrom 2015 onwards, and a further reduction amounting to less than50% of today’s level by 2050, in order to avoid catastrophic climatechange (Fisher et al., 2007, p. 172). The building sector presentsa great challenge to achieving this ambitious target, with itsintrinsic inertia and irreversibility in energy consumption. In China,more than half of the existing buildings in 2015 will be built after2000 (World Bank, 2001), and the Chinese Ministry of Constructionestimates that around 15–20 billion m2 of urban-zone housing willbe built between 2005 and 2020 in order to accommodatenewcomers to the cities – equivalent to the entire existing buildingstock in the EU-15. In the meantime, coal still accounts for two-thirds of China’s primary energy supply and this trend is likely to

for helpful comments andimprove the paper. But the

rs.þ33145 497 685.michel.colombier@iddri.org

All rights reserved.

continue in the next decades (IEA, 2007a). Therefore, effectivelyimplementing energy efficiency in China’s sizeable housing stockwill most likely play a central role in addressing global challenge ofclimate change.

The building sector consumes more than one-fifth of finalenergy consumption in China, and this proportion is growingsteadily (IEA, 2006; Jiang, 2007; Long, 2007a).1 The design andconstruction of long-lifespan urban infrastructuredof buildings inparticulardwill shape the energy perspectives for decades to come.The inertia of the construction sector is such that a house willremain untouched for at least 30–50 years before being demolishedor renovated. Inefficient constructions will result in tremendousenergy and climate implications and render climate and energysecurity more vulnerable. Implementing carbon emissions mitiga-tion options in buildings is associated with a wide range of co-benefits, including social welfare benefits for low-income house-holds, increased access to energy services, improved indoor andoutdoor air quality, as well as increased comfort, health and qualityof life, job creation and economic competitiveness (Levine et al.,2007, p.389). Furthermore, increasing energy efficiency is seen asthe most effective way of improving the security of energy supply,

1 Biomass is excluded here since it is considered as carbon neutral.

2 Energy used in manufacturing building materials and construction is included.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2437

reducing carbon emissions and increasing competitiveness (IEA,2007b). It is widely accepted that it is cheaper and more efficient toaddress the energy efficiency of buildings at the construction stage,as reconstruction and/or retrofit during the operational stage isvery costly and may also create social problems.

BEE offers promising prospects for the long-term GHG emis-sions mitigation, the Intergovernmental Panel on Climate Change(IPCC)’s Fourth Assessment Report considers that there is anopportunity to cost-effectively cut global CO2 emissions in existingbuildings by nearly 30% by 2020. More specifically, conceiving thedesign and the operation of buildings as a single integrated systemcan allow further cost-effective CO2 reduction, on the order of 75%or even more (Levine et al., 2007, p.389), Moreover, energy-efficientbuildings do not necessarily cost more than less-efficient oneswhen using a life-cycle analysis approach. Optimised buildingdesign from the start, incorporating passive ventilation and passivesolar cooling and heating, can allow significant reduction inmechanical equipment. In this way, the incremental cost of thebuilding envelope with respect to conventional practices can beoffset (for full technical details, see Bedel and Salomon, 2001;chapters 4 and 13 in Harvey, 2006 and section 3 in Salat, 2006.Although most actors in the building markets (e.g. developers,contractors, house owners, and tenants) are aware of sustainable orgreen buildings, only a few have ever been directly involved.Financiers and building developers are considered as the mainbarriers to more sustainable approaches in the building value chain(WBSCD, 2007).

The global effort to combat climate change will inevitably relyon improving the energy efficiency of buildings in the comingdecades (IEA, 2006), with buildings responsible for one-third ofglobal CO2 emissions (Urge-Vorsatz et al., 2007). A recently studypublished in Science demonstrates how a suite of existing techno-logical options could be used to reduce greenhouse gases (GHG)emissions to a level that is sufficient to avoid the dangerous effectsof climate change (Pacala and Socolow, 2004). GHG emissions canbe broken down into manageable (though still large) ‘wedges,’ eachof which relates to a technology set in a specific sector. According totheir projections, global carbon emissions could be cut by one-quarter in buildings and appliances by 2054. In addition, employ-ment of proactive policies aiming at improving BEE and drasticallyreducing CO2 emissions is justified by long-term risk adverseeconomic analyses recognizing that higher GHG concentration canresult in global catastrophe, irreversibility, and intergenerationalinequity (Arrow et al., 1996; Fisher et al., 2007; Stern, 2007;Halsnæset al., 2007). It is important to note that traditional marginalanalysis framework is inappropriate for the economic assessmentof climate change (Stern, 2008). Proactive actions should be takenurgently since a wait-and-see approach or a do it later approach thatwaits until the cost of mitigation ‘‘declines’’ may result in irre-versible and catastrophic climate change. As such, mainstreameconomic thinking about the problem has shifted from a single-discipline focus on cost-benefit analysis to a new inter- and multi-disciplinary risk analysis. This shift is more evidence of the failureof the traditional, marginal equilibrium approach in general toprovide an adequate understanding on both a micro- and macro-level (Barker, 2008).

The objective of this paper is to assess the long-term policyimplications of improving BEE and mitigating carbon emissionsassociated with energy consumption in buildings. Different policytools are discussed to address the barriers to deployment of BEEtechnologies and practices. It is argued that the market-basedeconomic instruments should be complementary to mandatorybuilding efficiency regulations, encouraging BEE with more co-benefits for society, and allowing the achievement of long-termclimate change mitigation objectives.

The paper is organised into five sections. Section two givesa comprehensive overview of energy demand characteristics andprojected scenarios in the Chinese building sector and reviews cost-effective technical solutions for dramatic improvement in BEE. Thethird section discusses the barriers to building efficiency imple-mentation, as well as the means to address these challenges.Section four focuses on the policy priorities aiming at achievingbetter energy efficiency and reducing energy-related GHG emis-sions to avoid carbon lock-in. Section five concludes the paper andoutlines future research directions.

2. The role of buildings in mitigating carbon emissionsin China

2.1. General context of China’s building sector

At present buildings, account for more than one-fifth of finalenergy consumption in China and this proportion is very likely toincrease to 35% by 2020 (Long, 2005),2 with this sector contributingto 25% of GHG emissions. Each year, about 1200–1400 millionsquare meters of residential buildings are constructed in China,with an investment of nearly 1000 billion Yuan (143 billion US$),approximately 20% of total fixed assets investment and 8–10% ofChina’s GDP (NBS, 2006). Housing construction in China consumes20% of total steel output and 17.6% of cement production each year.Average steel and cement consumption in housing construction inChina is 55 kg/m2 and 222 kg/m2, respectively, 10–25% higher thanin developed countries (Centre for Housing Industrialization,2007). Details of energy consumption in China in 2004 are reportedin Table 1.

2.2. Construction trends and potential for CO2 emissionin buildings

China is engaged in a vast programme of urban development.The urbanisation rate is expected to reach 55% in 2020 and 58% in2030, the population in urban zones is projected to increase from460 million in 2000 to 830 million in 2030, with an average annualgrowth rate of 2% (Toth et al., 2003). Some 300–400 million ruralresidents are expected to migrate to the cities in the next 20 years.The average per capita living area in cities is expected to reach30 m2 by 2020, roughly equivalent to that in developed countries inthe 1990s (MOC, 2003). More than 2 billion square meters ofbuildings in China have been constructed annually since 2000,making China the country adding the most building surface areaper year in the world.

The rapid rate of economic development and urbanisation posea formidable challenge to China’s building sector in terms of energysupply and carbon emissions. China will need to employ all meansto meet the ever-growing demand for energy services resultingfrom economic development and a rise in standard of living.Furthermore, households are likely to opt for modern energysources as income increases. Fig. 1 illustrates the evolution ofbuilding energy consumption by source since 1980. Two keymessages can be concluded: 1) Building energy consumptionincreased sharply after 1990, with total consumption more thandoubling during the period 1980–2005. 2) More specifically, theshare of coal use in buildings has been reduced significantly, fallingfrom nearly 88% in 1980 to 31% in 2005, essentially replaced byelectricity and gas. This confirms the assumption that more andmore Chinese households, particularly in rural areas, will graduallyswitch from solid biomass to electricity and gas.

Table 1Final energy consumption in buildings in China, 2004.

Area (million m2) Energy consumption

1. Rural residential (Biomass is included) 24 000 219 mtoe of solid energy90 TWh of electricity

2. District heating in northern urban area 6500 92.86 mtoe of coal/year3. Urban consumption (excluding district heating in 2) Residential electricity 9500 260 TWh/year of electricity

Commercial electricity 5500 240 TWh/year of electricitySubtotal 15 000 500 TWh/year of electricity

Total 312 mtoe of coal and biomass and 590 TWh of electricity were consumed on site (final consumption)in China’s building sector in 2004.

Sources: Jiang (2007). ‘Current status of energy use in buildings in China’, in 2007 Annual Report on China Building Energy Efficiency, Tsinghua University; China Energy StatisticalYearbook 2005 (NBS, 2005).Note: solid biomass is excluded.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–24472438

2.3. Life-cycle analysis of building energy efficiency

The economic analysis of BEE should go beyond analyses based onconventional design and construction, and adopt a life-cycle analysisapproach. Energy consumption in buildings includes both the opera-tional and embodied energy. The embodied energy of a building is theenergy used to acquire raw materials and manufacture, and to trans-port and install building products in the initial construction stage ofbuildings. The energy for materials production and buildingconstruction accounts for about 10–15% of energy use over the lifecycle of a building, with approximately 85% of energy use expendedduring the operational life of buildings (WBSCD, 2007).

Although efficient building material has more embodied energythan traditional material, a significant portion of the extra energyconsumed in manufacturing insulating material can be offset byenergy savings during the first years of building occupancy. Thusenergy efficiency measures permit consumers to reduce signifi-cantly energy consumption from a life-cycle analysis point of view.Embodied energy in production phase has a more important sharein the total energy needed in low-energy buildings, but materialrecycling and combustion following demolition offers the potentialof a reduction of 40–50% of embodied energy in northern Europeancountries such as Sweden (Thormark, 2002).

Fig. 2 shows three scenarios of life-cycle energy consumption ofa typical residential building in northern China throughout a lifespan of50 years, expressed in MJ/m2 (a full explanation is provided in theAppendix A). Life-cycle energy consumption may be reduced by one-third if current Chinese BEE policy is fully implemented, and more than60% energy savings can be achieved through adopting Swedish buildingcode, the most stringent mandatory building code in the world.

2.4. Techniques and cost of improving building energy efficiency

Cost-effectiveness and policy relevance should be consideredthe primary criteria in assessing the mitigation programmes in thebuilding sector. Table 2 outlines the major technical interventionsand their associated costs.3 From this table, it can be seen thatimprovement in building efficiency in China is the most cost-effective way to address CO2 emission mitigation.

A broad array of technical measures can be used to reduceenergy consumption and related CO2 emission with co-benefits(Urge-Vorsatz et al., 2007). Besides enhanced building envelopeinsulation, urban and architectural design also plays a fundamentalrole in reducing energy consumption in buildings. Various tech-niques allow energy consumption reduction in buildings, such as:sustainable urban planning; optimised site planning and design,including natural ventilation and suitable orientation; solar,

3 The cost reference comes from the existing literature.

geothermal and other renewable energy integration; bioclimaticarchitecture design; and enhanced mechanical ventilation withoptimised heat recovery system. All these design and operatingimprovements can lead to substantial reduction in building energyconsumption, all while providing much superior thermal comfortfor residents (Harvey, 2006; Salat, 2006).

However, incremental cost often constitutes the principalbarrier to widespread deployment of BEE techniques and practices,in spite of BEE’s superior thermal comfort and reduced energy billsfor energy users. As mentioned in WBSCD (2007), most buildingdevelopers and constructors are likely to overestimate the cost ofenergy efficiency in green buildings: Most constructors perceivethe extra cost to be 17%, instead of the 5% that corresponds to thereal additional cost. Conversely, energy efficiency does not neces-sarily entail higher costs. For example, highly energy-efficientbuildings incorporating passive heating and cooling allowsa significant reduction of mechanical equipment and related costs,as compared to inefficiently designed buildings with oversized heatdistribution networks (Harvey, 2006). PV-integrated buildings, byusing distributed energy, can help to reduce the transmissionbottleneck in conventional large centralised power infrastructure,thus improving grid efficiency. This is particularly the case incentral and southern provinces, where cooling in buildings isresponsible for nearly 40% of peak load in the summer, resulting inincreased security risk for the electric grid, and inefficiency in theelectricity infrastructure, for a considerable part of generatingcapacity would be left idle during baseload time.

With less than 10% of additional up-front cost, the German Pas-sivhaus and Swiss Minergie-P labelled houses can achieve ultra-lowenergy consumption with improved air-tightness (<0.6 vol./h) (lessthan 15 kWh/m2/year for space heating and 120 kWh/m2/year, allenergy uses included in terms of primary energy demand). The latestGerman know-how can allow constructing new buildings withvirtually no heating requirement with an extra cost of 5–12%compared to conventional practices. Inclusion of solar and otherrenewable energy, enhanced building envelope insulation, opti-mised architecture design (building shape, orientation, passiveheating and cooling), and other technical measures all allow cost-effective reduction in energy consumption. A recent energy-efficientbuilding pilot project in Tianjin (in Northern China, 100 km east ofBeijing), implemented in collaboration with World Bank-GEF,achieved 65% heating consumption reduction with less than 8%additional cost, as well as improved thermal comfort. The U-valueof the roofs and exterior walls is 0.42 W/m2k and 0.52 W/m2krespectively, and air tightness is maintained at 1.93 vol./h under50 Pa (Liu, 2006).4

4 The percentage of extra cost is dependent on types of buildings (low-rise orhigh-rise buildings, brick-laid or reinforced concrete-casting) and efficiency tech-niques employed.

Final energy demand in buildings in China

0

40

80

120

160

200

1980 1985 1990 1995 2005

Mto

e

electricity coalHeat

gas Petroleum products

Fig. 1. Final energy consumption in the building sector in China by energy source (source: International Energy Agency database). Note: Biomass is excluded because detailedinformation regarding biomass energy consumption before 1994 is not available from the IEA).

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2439

2.5. The forces driving the increase in energy demand in buildings

Past experiences in developed countries demonstrates a closerelationship between energy service demand (heating, air-condi-tioning, hot water, electric appliance etc.) and household incomelevel. For instance, in OECD countries, household heating intensity(useful energy per square metre per degree day) has continuouslydecreased after 1974, with considerable improvements in thethermal performance of buildings, while consumption of electricappliances has increased steadily over the past 30 years (Geller et al.,2006), as a result of technological innovation, and householdstructure and lifestyle changes. Structure change (floor area percapita, appliance ownership) also has a significant impact onhousehold energy consumption: Larger homes have greater energydemand than smaller ones, and an empirical study shows thatstructure change contributed 23–56% to an increase in householdappliance consumption in various OECD countries over the period1973–1992 (Schipper et al., 1996). It is not surprising to find thatrising incomes have enabled consumers to purchase more electronicappliances and larger houses (Zacarias-Farah and Geyer-Allely,2003). A residential consumption survey (SMRELRA and SMCC,2006) in Shanghai showed that residential electricity varied signif-icantly depending on household income and home size, with higherincome households and larger apartments consuming over twice asmuch as lower income families in smaller houses, 3247 kWh/a forhigh income households (homes greater than 100 m2) and1484 kWh/a in low-income households (homes less than 50 m2).5

It is projected that per capita income in China could reach US$10 000 (PPP) by 2020 (NDRC, 2004a). In light of this, householdelectric appliance ownership is expected to rise steadily, withassociated consumption expected to significantly increase withoutserious action on appliance efficiency improvement.

Furthermore, urban low-income households and rural residentswill gradually switch to modern energy sources such as natural gasand electricity, replacing the conventional biomass energy (mostlystraw and stalk) that currently enjoys widespread use for cookingand heating in rural areas. Meanwhile, the ageing population impliessignificant changes in urban household structure, such as anincreasing number of households on one hand, and reduced averagehousehold size on the other hand. Empirical studies show thatenergy demand per capita is much higher for small households

5 Higher income households had average monthly income of 6678 RMB (900US$) in 2006 price, while lower income families averaged 2497 RMB(340 US$) permonth in the survey sample.

(greater floor area per capita) than in large families (less floor areaper capita). For example, U.S. residential survey data (DOE, 1997)shows that average per capita energy consumption is 25 Mbtu in a 5-person household, 37 Mbtu in a 3-person household, and 75 Mbtu ina single-member household. The aforementioned Shanghai survey(SMRELRA and SMCC, 2006) provides similar results.

2.6. Perspectives on energy demand and CO2 emission reductionsin buildings

Li (2008) compares different scenario analyses investigating thepotentials for reducing energy consumption and associated carbonemissions in China’s buildings sector. Improvement of energy effi-ciency by tightening building regulations can lead to a reduction of112 mtoe energy use in 2020, avoiding nearly 100 Mt of CO2

emissions (ERI, 2003). The Demand-Side Management (DSM)measure is the most effective co-benefit means for mitigatingclimate change. The full implementation of the DSM Policy ofelectricity consumption in the Chinese building sector can allowa reduction of up to 347 million tons of CO2 in 2030 compared to thereference scenario, and the active development of renewableenergies and nuclear power in the electric production would allowa further reduction of 390 million tons of CO2 in China (IEA, 2006).Details of scenarios are provided in Tables A.1 and A.2 in theAppendix.

From a near-term analysis perspective, enforcement andimplementation of BEE regulations can also significantly contributeto carbon emissions reduction in buildings. Improved BEE can leadto significant reduction in electricity consumption of air-condi-tioning (AC) and space heating during peak hours in summer andwinter in the southern and central provinces. Long (2007b) showedthat sizeable potentials for CO2 emissions can be achieved byimproving the energy efficiency in both new and existing resi-dential buildings in the city of Shanghai. It is projected that over theperiod 2007–2010, more than 2.54 Mt CO2 emission could bereduced by implementing BEE (enhanced envelope insulation anddouble-glazing fenestration6) in all new buildings, and retrofitting10% of non-efficient houses each year. Note that fenestration energyefficiency improvement is extremely important in this climatezone, accounting for more than nearly 85% of the potential for

6 A software based dynamic energy simulation is conducted in this study to studythe impact of improved efficiency in different building component. For example,assumed double-glazed window has a U-value of 2.8 W/m2K and shape coefficientof 0.8, compared with 4.7 W/m2k and 0.9 in the conventional non-efficient design.

Fig. 2. Life energy consumption assessment in buildings in China.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–24472440

reduction in energy consumption and related CO2 emissions. Inaddition, improving the energy efficiency of new commercialisedroom air conditioners (RACs) on the market and replacing all RAColder than 10 years in Shanghai’s residential sector will allow foranother reduction of 2.6 Mt CO2 over the same period (Long,2007b).

3. Regulatory instruments and barriers to implementingenergy efficiency in buildings

3.1. Building efficiency design standards

Each year, about 130 million tonnes of standard coal equivalent(tce) are burned just for space heating in urban residential andcommercial buildings – approximately 52% of total energy

consumption in buildings (Jiang, 2006). Inappropriate design anda deficiency of regulation in centralised heating systems oftenresult in huge energy loss in urban district heating. China’s vastgeographic zones produce significant climate variations. Spaceheating is the primary energy demand in buildings in the north,whereas air-conditioning dominates electricity consumption insummer in the southern and central provinces. Four consecutiveBEE standards have been stipulated in China since the mid-1980s.Lang (2004) gave a comprehensive overview of these mandatoryBEE design standards. However, the enforcement and imple-mentation of these regulations have encountered numerous diffi-culties at the local level, particularly in the small and medium citieswhere the inertia of builders and the prevalence of conventionalbuilding techniques and practices constitute the major barriers. Inthese areas, the spread of new efficiency technologies by

Table 2CO2 emissions mitigation technical solutions and cost reductions in US$/tCO2.

Abatement options CO2 mitigation with net economic benefits Net cost of abatement (US$/tCO2)

2004 2020

Thermal insulation Yes Zero even negative idemFuel switch from coal to gas in boiler in district heatinga No 37a

DSM (electrical appliance improvement) Yes Zero or negative idemCarbon Capture and Storageb No 50–100 25–50

Source: Gielen and Podkanski (2004); IEA (2008); Li (2008).a Calculated based on the lower heating value (LHV) of fuels and price level in China in 2005; water and pumping system energy consumption in heating system are

excluded.b The cost of CO2 abatement is dependent on various technical parameters and geographical location.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2441

inadequate technical capacities, as well as a lack of information anda trained workforce pool.

3.2. Barriers

The inflexible energy efficiency gap in buildings has underminedthe implementation of energy conservation and GHG emissionsreduction. Substantial market barriers persist and need to be over-come through appropriate policies and programmes. These barriersinclude the high costs of gathering reliable information on energyefficiency measures, the ‘‘stickiness’’ of the techniques practiced bythe builders, lack of proper incentives between landlords who wouldpay for efficiency and tenants who would realise the benefits (aprincipal-agent problem), limited access to financing, energysubsidies, and also the fragmentation of the construction industryand of the design process into many professions, trades, work stagesand industries (Levine et al., 2007, p. 390). These barriers areparticularly varied and strong in the residential and commercialsectors; obstacles thus can only be overcome through a diverseportfolio of policy instruments. NDRC (2004b) devised a compre-hensive multi-sectoral energy conservation plan allowing thebuilding sector to save 50 million tons of standard coal and 290million kilowatt-hours over the period 2006–2010 if the specifiedmeasures are combined appropriately. However, realising the savingpotentials will confront strong resistance of actors in the construc-tion sector. Note that the actual enforcement and implementationrate is relatively low in many medium and small cities due to inad-equate technical and institutional capacities.

The increased construction costs associated with energy-effi-cient measures, combined with a dearth of available advancedtechnologies, give developers little incentive to comply withbuilding codes. Moreover, the current energy billing and pricingsystem is not consistent with the rules of a market economy. Fordecades, urban space heating has been considered a public servicein the Northern provinces. Heating consumption is billed on thebasis of surface area instead of actual consumption. Consumers arenot given any price signal allowing them to monitor their energyconservation, and no economic incentive is available for housingdevelopers to build more energy-efficient houses. Although thecentral government issued a guideline in 2003 urging localgovernments nationwide to implement heating reform as soon aspossible, the old billing and pricing system remains dominant. Thepricing reform lags behind the government’s reform schedule.More specifically, a number of energy suppliers and buildingconstructors remain reluctant to establish individual billing byinstalling heat meters and thermostatic valves; end-users are thusunable to regulate interior temperature. Many district heatingcompanies owned by the municipality have even encounteredfinancial difficulties because the price paid by the consumer issignificantly lower than the real cost. Significant energy losses arealso caused by a series of maintenance and management defi-ciencies stemming from the current heat billing system. Moreover,

the development of distributed energy systems such as small andmedium-sized high-efficiency combined heat and power (CHP)faces tremendous difficulties in selling electricity to the grid. Theyare in cutthroat competition with suppliers of low-cost coal-firedsmall boilers in absence of policy support.

4. Policy instruments for improving building energyefficiency and mitigating CO2 emissions

Policies and instruments play a stimulating role in climatechange mitigation. Examples of theses policies and instruments areregulations, taxes and charges, tradable permits; distribution ofinformation, and subsidies. However, the cost-effectiveness ofa policy is a key decision parameter in a world with scarce resources(Gupta et al., 2007). As mentioned earlier, improvement in BEE isthe most cost-effective way of mitigating CO2 emission in China’sbuilding sector and can make a significant contribution to emis-sions reduction through available technologies. A comprehensivepolicy portfolio is a prerequisite of the effective deployment of BEEtechnology. Economic instruments should complement regulatorystandards in order to maximise the probability of the uptake ofefficiency practices and techniques lowering the cost of imple-mentation. In China, energy efficiency policies are still largelyimplemented by a traditional command-and-control approach.Market-oriented measures have developed slowly. Besides theregulatory tools such as different economic and policy instrumentscan be put in place to foster transformation towards low-carbonand climate resilient buildings infrastructure, such as deployinghigh-efficient appliances by energy labelling and certification,promoting energy service companies (ESCOs), diffusing innovativefinancing tools, establishing internal carbon markets, and encour-aging energy pricing reform.

4.1. Improving building energy efficiency design andconstruction practices

Despite a significant reduction in heating intensity after theimplementation of the 1995 Energy Efficiency Standard for NewResidential Buildings, the average energy consumption for heatingan efficient house complying with MOC (1995) in Northern China(90–100 kWh/m2a) is still almost twice as high as in the mostefficient houses in Sweden, Denmark, the Netherlands and Finland(40–50 kWh/m2a). The national building code should thus berevised and updated, taking into account the state-of-the-art BEEdesign standards in northern European countries.

Moreover, sustainable and bioclimatic building design should bepromoted in order to revolutionise practices and to train architectsand building engineers. As mentioned earlier, optimised buildingdesign through passive heating and cooling can significantly reduceenergy consumption without incremental cost, and can even leadto a cost reduction in many cases by downsizing the mechanicalequipment and using recyclable and environmentally friendly

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–24472442

building materials. Scaled-up development of low-carbonconstruction will trigger fundamental transformation of the wholebuilding supply chain, channelling more resources (both financialand human capital) into efficient technologies and high-perfor-mance products.

From a climate perspective, CO2 emissions should be integratedinto energy-efficient design standards and performance assessmentcriteria in order to encourage the use of alternative heating and hotwater supply systems, in particular the development of renewableenergies. Heating system efficiency should also be upgraded in orderto minimise energy loss in the distribution network.

4.2. Scaling up energy service companies (ESCO)

Commercial ESCOs offer an alternative to bridge the technicaland financial gaps in the efficiency market. Three major types ofESCO contract terms can be identified: 1. Shared saving, which haskept the most common ESCO contract; the service supplier sharesthe saved energy cost with clients; 2. Guaranteed saving, which ismostly used in industry. 3. Outsourcing energy management,particularly adapted to commercial buildings but still underdevelopment in China. There may be various areas in which ESCOscan intervene based on integrated energy management experience.

Currently, most ESCO development projects in China focus onindustry and high-end commercial building efficiency improve-ment, but technical intervention in these projects remains limitedas the principal actions have concentrated on supply-side technicalimprovements such as electric machine efficiency improvement.ESCO service in China is often characterised by its short-termcontracting period, less than 3 years in general instead of 8–10years in developed countries. The ESCO market in most Chinesecities remains immature and short-sighted.7 Longer-term ESCOcontracts face a number of market barriers and unforeseeable riskssuch as the financial accountability of clients.8 Furthermore, manycontractors are more concerned about investment payback thanintegrating energy performance improvement proposals. There arealso institutional barriers to ESCO development, such as leasingregime constraints hindering the development of efficiencymanagement projects.

In addition, some ESCOs in China are almost becoming purelyprofessional efficient energy product promoters rather than energyservices supplier. From the client’s perspective, the hard investment(equipment, appliances) is more tangible than the soft investment(idea, organization; management). An ESCO is likely to offer a singlesolution such as installation of high-performing equipment, whileignoring the importance of comprehensive energy management inbuildings.

ESCO involvement in the building sector implies significantpreliminary diagnostic work on the energy performance of build-ings, including thermal quality, lighting, high-voltage alternatingcurrent (HVAC) system efficiency and electric machinery. Never-theless, only a few energy companies are capable of supplying sucha comprehensive provision in the ESCO market since it requirestechnical proficiency and high capacity.

7 According to Ye Wenbiao, director of Shanghai Energy Conservation ServiceCentre (SHECSC), one of the earliest ESCOs. It was once a government agency incharge of energy conservation management in the municipal administration,particularly in the industry sector. In 1998, the World Bank launched a series ofdemonstration projects to promote ESCO services in China in collaboration withother international organization such as UNIDO, in which the SHECSE was involved.

8 This matters in particular in the industry where a number of firms wentbankruptcy after signing the ESCO contract and the payment capacity ofenergy-efficient services crumbled.

The opening of the energy service market in China offers anunprecedented opportunity for the international energy utilitiesand their energy service provision subsidiaries specialised inbuilding energy performance improvement. Many well-knowninternational energy companies have stepped in China’s ESCOmarket such as Siemens, Schneider Electronics and so forth. Tech-nology transfer is one of the key elements of ESCO market devel-opment. Patent protection institutions should be established toprotect the IPR and make the technical know-how transfersmoother. Market regulation and legislative specifications are themost urgent issues to be addressed by the government if more andmore world-leading ESCOs become involved in China’s market,since the stakeholders require risk aversion guarantees in energyservice investment.

4.3. Energy efficiency and carbon labelling and certification

Some developed countries have successfully introduced ultra-low, zero-energy and/or ‘environment-friendly’ labelling schemesin buildings sector, such as the European passive Habitat label(Passivhaus in Germany, Minergie in Switzerland, the HQE label inFrance), BRREM in the United Kingdom, and the LEED label in theUnited States. The latter two integrate energy efficiency and envi-ronmental impact indicators in general housing quality assessmentfor new housing construction projects.

On the basis of experience in the US and in the EU, BEE certifi-cate and carbon labelling can be introduced to new constructionand retrofit programme in Chinese cities. In fact, the labellingapproach has already been in place in China for several years in thewhite goods sector (refrigerators and air-conditioners). The nextstep is to extend the national administrative standardizationprocess to the third-party assessment, and to the building energyand carbon efficiency certification.

4.4. Removal of financing bottlenecks

Financing constitutes one of the major challenges to thedeployment of low-carbon technologies in the building sector,contributing to a market functioning with reduced efficiency, andwith persistent barriers to capital formation. Technology risk,inflexible energy pricing structures, regulatory uncertainty andother obstacles keep borrowing costs for energy efficiency projectsrelatively high (Wellington et al., 2007). In addition to these highertransaction costs, most investors in China’s building sector areunfamiliar with many low-carbon technologies and thereforeperceive them as risky, thus increasing the cost of capital.

In addition, many building efficiency projects lack sustainedfinancial resources since the major financial actors are rarelyinvolved in investment in energy efficiency projects. Chinesefinancial institutions such as commercial banks typically do nothave expertise in energy efficiency technologies and thus lack theability to assess technical risks of BEE financing programs. Energyefficiency investments are typically small and come with hightransaction costs. The value of energy-efficient projects is relativelypoorly recognised (Cheng, 2005). Innovation and institutionalcapacity-building with regard to BEE financing will play a signifi-cant role in facilitating policy implementation.

It is urgent to remove the financial barriers to the widespreaddiffusion of the best- practice technologies in the installation ofhigh-performance or renewable energy systems such as photo-voltaic (PV), geothermal, and heat and cool storage in the buildingsector. New ideas are required and the institutional frameworkneeds to be reshaped. Policy instruments need to be put in place toincentivise commercial banks to get involved in the business ofhigh-efficiency construction projects, or in the installation of

India

6%

R. of Asia

5%

ECA

1%

Brazil

6%

R.of Latin

America

5%

China

73%

Fig. 3. Global CDM market (Capoor and Ambrosi, 2008)

HFC-2372%

wind6%

hydro4%

biomass 1%

NO 6%

Coal bedmethane 5%

EnergyEfficiency

4%

landfill gas2%

70Mt CERs in 2007

Fig. 4. Breakdown of CERs of CDM projects in China total (source: IEA, 2007; UNEPCD4CDM database).

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2443

renewable and innovative energy supply systems, by supplying lowinterest credits.

4.4.1. Linking Kyoto financing and building efficiencyThe Clean Development Mechanism (CDM) of the Kyoto Protocol

has paved the way to alternative financing portfolio regardingenergy efficiency improvement projects in this sector, in collabo-ration with the international financial actors. European and Japa-nese actors have been especially active in the global carbon marketsover the past years, and European buyers completely dominatedthe CDM market in 2006 with 86% of transaction volume. As can beseen from Fig. 3, China dominates the CDM market accounting formore than 70% of global CDM transaction volume (Capoor andAmbrosi, 2008). However, most of CDM projects in China arefocused on non-CO2 GHG reductions, at the expense of energyefficiency projects in particular (see Fig. 4). Institutional reforms arenecessary to bring CDM projects into sustainable infrastructuredevelopment such as the building sector in Chinese cities.

The stakeholders in the building sector will be interested inCDM if the adoption of non-mandatory higher BEE can yieldsufficient carbon emission credits tradable at competitive prices. Asimplified calculation can illustrate this financing mechanism.Assume that a house builder in the city of Tianjin would be ready toadopt the best available techniques (BAT) of building energy effi-ciency (equivalent to the Swedish building code9). He perceivesthat the CDM financing can be profitable for him, in other words,the Certified Emissions Reductions (CER) credits of CDM financingwill allow him to payback the initial extra cost. Put formally,a builder adopts higher efficiency practice (the Swedish codestandard equivalent is used here) if and only if:

PCER EAC

and

EAC ¼ ICR T

0

Etjt Eswt

$ertdt

(1)

where

9 Here we are concerned with the technical improvement of efficiency byenhanced envelope insulation, fenestration and ventilation. The passive heating isrelated to building design and thus difficult to quantify. However, it is by no meansto deny the importance of optimised architecture design.

PCER is the CER price of net CO2 removal in CDM project (US$/tCO2)EAC is the discounted emission abatement cost: (US$/tCO2)IC is the unitary incremental cost of building efficiency update:US$/m2

Etjt is the per floor space emission in baseline building (tCO2/year/m2)Eswt is the per floor space emission in buildings complying withBEE equivalent of Swedish buildings code (tCO2/year/m2)r is discount rate: 6% which is equal to the long-term mortgageinterest rate in China

The data from Liu (2006) is used here to estimate the cost of BATpractices and to calculate the emission reduction potential andrelevant costs. The overall envelope (wall, window, roof) cost ofa typical multi-storey apartment under the baseline scenario (builtin compliance with the current building code in Tianjin) is around280 Yuan per square metre of floor space in 2005 prices (36 US$ perm2). Updating to the equivalent Swedish efficiency performance10

in the same building entails an incremental cost of 61 Yuan /m2

(7.9 US$/ m2 in 2005 prices), including the optimised buildingdesign, and the cost of an improved ventilation and heat recoverysystem. Heating intensity of a house complying with the two BEEstandards is 49 kWh and 18 kWh per square metre, respectively. Itis assumed that the upstream heat supply system will switch fromcoal-fired district boilers with 65% efficiency to gas-fired CHP with80% overall efficiency under CDM package. The correspondingemission factor is 49.8 kg CO2/m2a and 10 kg CO2/m2a in the baseyear, and 59 kg CO2/m2a and 11.5 kg CO2/m2a in 20 years (theincrease in demand for heating service is taken into account tomirror the rise in living standard; full data for the calculations isavailable upon request). Thus the total estimated emissionsreduction over the 20 years of CDM crediting period (shorter thanthe actual building lifetime) is 873.4 kg CO2/m2. With equation (1),

10 It must be mentioned that the Swedish reference mentioned here refers to theequivalent performance in terms of U-value, rather than the actual constructiontechniques practiced in Sweden due to difference of building materials and localcharacteristics, also the labour training and know-how.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–24472444

it can be easily calculated that the EAC is 16 US$/tCO2 (at 2005prices), which is slightly higher than average CERs price of China’sCDM projects, but still significantly lower than the EU-ETS Allow-ance (EUA) price that is close to $30 US$/tCO2, (Point Carbon, 2008).The EUA price reached V30 per tonne CO2 in April 2006. In addition,the carbon price in the European market is likely to continue to riseover the next decades as the European Union is expected to tightenthe emissions reduction objective for the post-Kyoto regime.Recent literatures on the economics of climate change and energymodelling all suggest that the benchmark CO2 price such as the EUAprice and the marginal abatement cost (MAC) in the Europeancountries and other parts of the world are likely to range between30 and 40 US$/tCO2 in the 2020–2030 timeframe (Anderson, 2006;Fisher et al., 2007; Stern, 2007; PowerPoint, 2007; Blanco andRodrigues, 2008; Colombier et al., 2008). Thus, the CDM CER priceof residential BEE upgrading in Chinese northern cities is verycompetitive for carbon financiers and investors. Moreover, thetransaction costs of CDM are expected to decrease further with thecity-level and sector-based CDM models.

The potential demand for ERs in EU countries could increase to300 Mt CO2e per year to attain the European target of reducingGHG emissions further to a 30% below 1990 levels by 2020(Capoor and Ambrosi, 2008). China’s building sector offerstremendous potential for mitigating CO2 emissions through theimplementation of clean technologies under CDM’s framework,and Chinese policy-makers should move more quickly to takeadvantage of this win-win finance mechanism. Institutionalchange is needed to enlarge the scope of CER-EUA interchangeon the global carbon market and to facilitate the implementationof programmatic CDM based on a whole jurisdiction of localauthorities.

4.4.2. Emissions trading and carbon taxIndeed, the climate change is the foremost global externality

that the world has ever seen, a carbon tax is necessary to correct themarket failure and inefficiency along with other command-and-control governing instruments (Stern, 2007). Thus, clear propertyright of carbon emissions and appropriate carbon price are neededto correct the market failure. Emissions trading and carbon tax arethe two most common administrative approaches with economicincentives in carbon emissions regulation. The choice between taxand CO2 trade permit reflects the trade-off of policy makersbetween cost and quantities as in most environmental regulatoryinstruments implementation (Weitzman, 1974). Emissions tradingamong industrialised and transitional countries (former Sovietbloc) was included in the Kyoto Protocol to the UN FrameworkConvention on Climate Change. Carbon tax is an example of envi-ronmental tax, it has been considered in the EU prior to theEmission Trade Scheme (EU-ETS). Carbon taxes have been intro-duced in Denmark, Sweden, Norway, Finland, Italy, theNetherlands, and the United Kingdom, while Germany, Austria,Belgium, and Japan have adopted broader energy taxes (Rich,2004), Although an international emissions trading system doesnot necessarily preclude the use of carbon taxes (domestically orinternationally), the two are commonly seen as competing policyinstruments to reduce GHG (Baumert, 1998). Both taxes and trad-able permits put a price on emissions, and that price is equal to allparties involved. Therefore, emission reduction targets are met atminimum cost.

4.4.2.1. Emissions trading. A national CO2 emission allowancescheme based on the cap-and-trade (CAT) principal may be devisedin China. Under the framework of the scheme, the major energy-intensive industries, of which a large number are involved inbuilding materials and appliances manufacturing (cement, steel,

glass, concrete bloc, insulation materials, white appliances), as wellas property developers and builders will be subject of quantitativeCO2 emissions caps, determined beforehand in the national allo-cation plan on the basis of grandfathering. Design of such a nationalcarbon trading plan must involve comprehensive consultation withinternational experts across different areas (engineering,economics, finances, policy etc), taking into account the industrialdevelopment context. The carbon trading market will be regulatedby a special scrutinizing commission under the leadership of theNational Development and Reform Commission (NDRC). Proposalsfor the allocation plan should be reviewed thoroughly withextensive diagnostic evaluation of sectoral and industry perfor-mance. A national carbon emissions trading market should beestablished to allow industries to buy or sell emission rights.

4.4.2.2. Carbon tax. Contrarily to CAT system which aims at specificquantity objective, tax instrument leaves the quantity of emissionreduction uncertain. Economic theory tells that tax is more efficientthan subsidy instrument by reducing the free-rider and transactioncosts in competitive market. The well-known Pigovian tax which isequal to the marginal cost of emission abatement will give indus-trials clear price signal to reduce their emission such that the totalcost of abatement will be minimised.

Although sensitivity in sectoral energy demand scenarios inChina has been tested in response to the degree of global climateconstraints in previous modelling (IEA, 2006; Jiang and Hu, 2006),no explicit cost of externalities, i.e. carbon is considered beingimposed on buildings industrials and end-user in the policyscenario. Larson et al. (2003) build a MARKAL-China model of energysystem representing all sectors of the economy and including bothenergy conversion and end-use technologies. Strikingly, they foundthat even when significant limitations on carbon emissions werestipulated, an advanced energy technology strategy on the basis ofthe author’s technology-cost assumptions would not incur a highercumulative (1995–2050) total discount cost in energy system thanthe business-as-usual strategy. This result has a strong energy andenvironmental policy implications in China. Low-carbon technolo-gies like ultra-efficient buildings can be commercially viable withexplicit CO2 price incentive. The lack of carbon constraint in energydemand and CO2 emission scenario would render the imple-mentation of cost-effective building efficiency policies more difficultand low-carbon technologies less competitive. In this respect, a Statecarbon tax could be levied together with environmental tax andpollution charges by central and provincial governmental andredistributed to financing the investment gap in the disseminationof appropriate efficient technologies such as buildings integrated PV,high-efficiency boilers, renewables and geothermal, districtbiomass-fuelled CHP etc.

4.4.3. Pricing reform and integrated land use/fiscal instrumentsExperiences in the OECD countries show that energy efficiency

improvements since 1973 mainly resulted from ongoing techno-logical progress and also response to rising energy prices (Gelleret al., 2006). However, energy prices for end-users in China areoften biased and subsidised by the government. Transparent andeffective energy pricing mechanism must be set up to reflect theactual economic cost and internalise environmental externalities.Apart from government intervention and industry mobilisation inenergy efficiency programmes in the building sector, privateconsumer-side initiatives should also be encouraged throughenergy pricing mechanism combined with fiscal incentives.Consumers will not be interested in energy efficiency programmeswithout a clear price signal or other market incentives. In China,both central and local government revenues still rely heavily on thetax base from public and private corporations, whereas individual

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2445

income tax represents only a very small part of governmentrevenue under the current fiscal regime. Tax instruments should beintegrated into building efficiency policies to encourage urbanhouseholds to choose high-efficiency appliances in residentialbuildings.

Similarly, land use and fiscal incentives could also be introducedinto the property market to induce developers and/or constructorsto build more efficient houses and adopt low-carbon technology;property developers should benefit from land use-related taxreductions or increased plot ratio in property development projectprovided that more efficient buildings and low-carbon or carbon-free energy supply are considered. For example, the French UrbanPlanning Act stipulates that a developer can be partly exemptedfrom land use regulations or urban planning codes or other speci-fied technical requirements such as plot ratios when high-efficiency buildings are constructed, or when large-scale renewableenergy use is included in the property development project. Thefiscal advantages of fixed capital investment, applicable in the1990s could be duplicated into BEE programmes. In addition,property tax could be instituted according to clearly defined energyefficiency criteria.

4.5. Public procurement, renewable obligation and greenjob creation

Many public buildings use low-cost inefficient appliances andequipments due to historical purchasing policies. Promotingenergy-efficient products in public sector can contribute toreducing wasted energy and saving money, as well as encouragingefficiency product market and creating green jobs. The model ofAgreement on Government Procurement of the WTO can be used asreference model of efficient appliance purchasing in public andgovernment buildings. European countries have valuable experi-ences with regard to public procurement for energy efficiency, forexample, the Dutch government stipulated that by 2010, 100% ofnational public procurement will include sustainable procurementcriteria and 50% for local and regional government (EC, 2008).

Likewise, the policies of thermal solar energy obligation in newresidential buildings and the share of solar heating in the totalbuilding stock have been implemented in the building codes of Spainand Portugal since 2007. China could also introduce similar regula-tion stipulating that all new residential and commercial buildingshave obligation of installing solar thermal system provided thatthere is favourable climate conditions with available technologies.

The policies that encourage energy efficient, sustainable andlow-carbon buildings will be decidedly powerful driver of BEEmarket and subsequently creating green employment in the wholesupply chain and industry of energy efficiency and renewableproducts. The total market potential for BEE is expected at 2000billion Yuan (263 billion US$ in 2005 price) in 2020 in the scenarioof proactive implementation of BEE policy (CCN, 2008). With a totalturnover being projected to reach 100 billion Yuan (13 billion US$ in2005 price), more than 2 million jobs would be created in therenewable sector if the objective of the national renewable energyplan can be achieved (NDRC, 2007). BEE-related industries wouldbe the backbone of environmentally friendly economic develop-ment in China.

5. Summary and conclusions

In this paper, we have investigated the carbon emissionsreductions potential offered by efficiency improvements in China’sbuilding sector. Huge economic and environmental benefits couldbe generated through boosting BEE and carbon emission mitigationpolicies in China’s building sector. More than 700 Mt CO2 emissions

could be saved in 2030 in the scenario of full implementation ofappropriate BEE policies (Li, 2008), which is 10-fold of the globalCDM market today (70 Mt in 2007). Major technical and institu-tional barriers to scaled deployment of BEE techniques and prac-tices were also identified. China’s economic development andcontinuous opening policies offer a positive perspective on inter-national collaboration in tackling these concerns in the context offossil fuel depletion and global warming. Strong action should betaken promptly in order to reduce the avoidable cost of inaction assuggested in the Stern Review (Stern, 2007).

We assessed different policy and economic instruments andoutlined their perspective of development such as building codesupdate, high-efficient appliances labelling and certification, energyservice corporate (ESCOs), financing tools innovation Kyoto mech-anism, carbon markets and energy pricing reform etc. It was arguedthat supportive policy and appropriate regulatory framework areprerequisite for facilitating the scaling-up of technological inno-vation in the building sector. Technical, financial and institutionalbarriers to the large-scale application of highly efficient, low-carbon technology can be overcome by making significant changesin public policies.

Implementation of BEE involves fundamental transformation inthe whole supply chain of the building industry, from architecturaldesign to upstream material manufacturing. However, the currentbuilding market is extremely localised and fragmented, synergybetween actors and transversal policy framework need to beformulated to promote the BEE. Economic and policy instrumentsneed to be tailored appropriately to accompany the regulatory andcompulsory standards. Market-based instruments such as ESCOs,efficiency and carbon labelling and certification for Eco-design andrenewable-integrated buildings, land-use right premium in theproperty market should be promoted. Importantly, the illustrativeexample in Section 4.4.1 showed that BEE will offer great potentialsfor CDM development in China since sizeable CERs can be gener-ated. In particular, it was demonstrated that the price of emissionreduction credits generated from upgrading the Chinese BEEstandards in the programmatic CDM framework could becompetitive in the international carbon market. However, appro-priate institutions need to be set up to facilitate the implementationof programmatic CDM in buildings sector.

In summary a holistic approach should be adopted by inte-grating the quality of energy infrastructure, building design andefficiency optimisation and public policies to remove the barriers toimplementation of energy efficiency in buildings. The long-termintegration of BEE into regulatory and incentive scheme is of criticalimportance in the policy making. Moreover, interagency collabo-ration among relevant institutions and public-private partnershipshould be strengthened: the ministry of construction, ministry offinance, ministry of science and technology and national bureau ofstandardization should all be involved in the global project. Adynamic link needs to be created to incentivise both upstream anddownstream actors in the whole building supply chain to movetowards low-carbon buildings perspective.

Further research regarding the building energy efficiency willfocus on creation of specific mechanism in favour of extensivepartnership between public institutions and private actors(industry, consumers) to foster the large-scale diffusion anddeployment of energy-efficient, low-carbon technologies in thebuilding sector in China. An innovative institutional framework isrequired to scale up BEE deployment across the nation, a nationalBEE market should be established by linking low-carbon or zero-carbon buildings programme with international carbon market todiversify the investment portfolios with regard to BEE financing.Lastly, wide collaboration and consensus on BEE between differentactors in both public and private sectors are critically necessary to

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–24472446

ensure viability, availability and performance of the long-term GHGmitigation in China’s building sector.

Appendix

Table A1Comparison of final energy consumption projection in buildings in differentscenarios.

2015 mtoe 2020 mtoe 2030 mtoe

Without biomassERI (2003)Ordinary effort 448Green effort 327NDRC (2004a)A 396B 372C 331DOE (2006) Reference case 282 312 401

With biomassIEA (2006),a

Reference scenario 502 549 642Policy scenario 463 487 534Jiang and Hu (2006)Baseline scenario 417 666Policy scenario 347 479Zhou et al. (2007) (commercial

buildings)190

Source: NDRC (2004); ERI (2003); Jiang and Hu (2006); IEA (2006); DOE (2006),figures in 2020 of IEA (2006) are adapted from WEO 2006 by the authors.

a Values in 2020 of IEA (2006) ’s projection are interpolated by the authors.

Table A2Scenarios of energy-related CO2 emission from buildings sector in China in themodels.a (in Mt)

2010 2015 2020 2030

NDRC (2004a)A 946 (19.9 %) 1 800 (26.2%)B 902 (19.7%) 1 599 (25.3%)C 810 (19.8%) 1 379 (24.9%)

IEA (2006)Reference 1786 (23%) 2108 2751 (26.4%)Alternative Policy 1636 (22.4%) 1814 2170 (24.7%)

IEA (2007)Reference 1590 (18.4%) 1901 2523 (22%)Alternative Policy 1507 (21.3%) 1636 1893 (18.6%)

Note: in fact, CO2 emission values are not supplied in explicit way in some of thesestudies, figures in 2020 of IEA are derived from their energy modelling results, formore details see Li (2008).

a Note: Not every model we reviewed provided specific value of the CO2 emissionfrom buildings.

Appendix A. Methodology of life-cycle building energy consumption

Each of the scenarios represents an assumption with whicha specific building design standard is adopted. The estimation ofenergy consumption of upstream material manufacturing is basedon published data in China’s building industry. The average inten-sity of cement and steel production is around 820 and 114 koe pertonne, respectively. Insulating materials energy consumption isaround 1.5 toe per 1 tonne EPS (heat transfer coefficient l¼ 0.04 W/mK) manufactured. The volume of EPS employed depends upon thebuilding envelope type and thickness which in turn is determinedby the efficiency objective required. Construction (mechanical,electric etc) and demolition consume another 3800 Mj per m2 offloor space. Nevertheless, the energy consumption related to

materials transport is ignored. According to Jiang (2007), theaverage annual energy consumption other than space heating inthe residential sector is estimated at 27–30 kWh/m2. Consequently,all these energy uses are summed up to compare the life-cycleenergy consumption of a typical residential building located in theclimate zone of Beijing.

References

Anderson, D., 2006. Costs and Finance of Abating Carbon Emissions in the EnergySector. HM Treasury, UK.

Arrow, K.J., et al., 1996. Intertemporal equity, discounting, and economic efficiency.In: Bruce, J.P., Lee, H., Haites, E.F. (Eds.), IPCC (1996), Climate Change 1995:Economic and Social Dimensions of Climate Change, Contribution of WorkingGroup III to the Second Assessment Report of the Intergovernmental Panel onClimate Change. Cambridge University Press, Cambridge.

Barker, T., 2008. The economics of avoiding dangerous climate change. An editorialessay on The Stern Review. Climatic Change 89, 173–194.

Baumert, K., 1998. Carbon Taxes vs. Emissions Trading: What’s the Difference, andWhich is Better? Available at: http://www.globalpolicy.org/socecon/glotax/carbon/ct_et.htm.

Bedel, S., Salomon, T., 2001. La Maison des Negawatts. Terre Vivante, 154 p.Blanco, M., Rodrigues, G., 2008. Can the future EU ETS support wind energy

investments? Energy Policy 36, 1509–1520.Centre for Housing Industrialization, Ministry of Construction, 2007. http://www.

chinahouse.gov.cn/cyfz16/160002.htm.Cheng, C., 2005. Electricity Demand-side Management for an Energy Efficient

Future in China: Technology Options and Policy Priorities. Doctoral thesisdissertation. MIT.

Capoor, K., Ambrosi, P., 2008. State and Trends of the Carbon Market. World Bank,Washington.

Colombier, M., Crassous, R., Criqui, P., 2008. EPE Etude FONDDRI Scenarios souscontrainte Carbone[. Powerpoint presentation. Paris, May 2008.

China Construction News, 2008. http://www.newsccn.com/Detail.asp?id¼2534.Department of Energy (DOE), 2006. EIA. International Energy Outlook. In: DOE,

Washington Energy Issues, vol. 24 (No. 3/4).Department of Energy (DOE), 1997. Residential Energy Consumption Survey.

Washington.(Energy Research Institute ERI), 2003. China’s Sustainable Energy Future. Beijing.European Commission, 2008. Communication from the Commission to the Council

and the European Parliament on a first assessment of national energy efficiencyaction plans as required by Directive 2006/32/EC on energy end-use efficiencyand energy services – moving forward together on energy efficiency.

Fisher, B.S., Nakicenovic, N., Alfsen, K., Corfee Morlot, J., de la Chesnaye, F.,Hourcade, J.-Ch., Jiang, K., Kainuma, M., La Rovere, E., Matysek, A., Rana, A.,Riahi, K., Richels, R., Rose, S., van Vuuren, D., Warren, R., 2007. Issues related tomitigation in the long term context. In: Metz, B., Davidson, O.R., Bosch, P.R.,Dave, R., Meyer, L.A. (Eds.), Climate Change 2007: Mitigation. Contribution ofWorking Group III to the Fourth Assessment Report of the Inter-governmentalPanel on Climate Change. Cambridge University Press, Cambridge.

Geller, H., Harrington, P., Rosenfeld, A., Tanishima, S., Unander, F., 2006. Polices forincreasing energy efficiency: thirty years of experience in OECD countries.Energy Policy 34 (5), 556–573.

Gielen, D., Podkanski, J., 2004. Prospects for CO2 Capture and Storage. IEA, Paris.Gupta, S., Tirpak, D.A., Burger, N., Gupta, J., Hohne, N., Boncheva, A.I., Kanoan, G.M.,

Kolstad, C., Kruger, J.A., Michaelowa, A., Murase, S., Pershing, J., Saijo, T., Sari, A.,2007. Policies, instruments and co-operative arrangements. In: Metz, B.,Davidson, O.R., Bosch, P.R., Dave, R., Meyer, L.A. (Eds.), Climate Change 2007:Mitigation. Contribution of Working Group III to the Fourth Assessment Reportof the Intergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, United Kingdom and New York, NY, USA.

Harvey, D., 2006. A Handbook on Low-Energy Buildings and District-EnergySystems. Fundamentals, Techniques and Examples. Earthscan Publications,London, 720 p.

Halsnæs, K., Shukla, P., Ahuja, D., Akumu, G., Beale, R., Edmonds, J., Gollier, C.,Grubler, A., Ha Duong, M., Markandya, A., McFarland, M., Nikitina, E.,Sugiyama, T., Villavicencio, A., Zou, J., 2007. Framing issues. In: Metz, B.,Davidson, O.R., Bosch, P.R., Dave, R., Meyer, L.A. (Eds.), Climate Change 2007:Mitigation. Contribution of Working Group III to the Fourth Assessment Reportof the Intergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, United Kingdom and New York, NY, USA.

IEA, 2006. World Energy Outlook. International Energy Agency. OCED/IEA, Paris.IEA, 2007a. World Energy Outlook 2007, China and India Insights. OCED/IEA, Paris.IEA, 2007b. Scaling Up Energy Efficiency: Bridging the Action Gap. 2–3 April 2007

Workshop Background Paper. OCED/IEA, Paris.IEA, 2008. CO2 Capture and Storage: A Key Carbon Abatement Option. OCED/IEA,

Paris.Jiang, Y., 2006. Energy Efficiency Approaches in buildings heating system in China

(in Chinese). Heating Ventilation and Air Conditioning 36 (3).Jiang, Y. (ed.) 2007. Annual Report on China Building Energy Efficiency 2007 (in

Chinese). Tsinghua University Buildings Energy Efficiency Research Center,China Architecture and Building Press. Beijing.

J. Li, M. Colombier / Journal of Environmental Management 90 (2009) 2436–2447 2447

Jiang, K., Hu, X., 2006. Energy demand and emission in 2030 in China: scenarios andpolicy options. Environmental Economics and Policy Studies 7 (3), 232–250.

Lang, S., 2004. Progress in energy-efficiency standards for residential buildings inChina. Energy and Buildings 36 (12), 1191–1196.

Levine, M., Urge-Vorsatz, D., Blok, K., Geng, L., Harvey, D., Lang, S., Levermore, G.,Mongameli Mehlwana, A., Mirasgedis, S., Novikova, A., Rilling, J., Yoshino, H.,2007. Residential and commercial buildings. In: Metz, B., Davidson, O.R.,Bosch, P.R., Dave, R., Meyer, L.A. (Eds.), Climate Change 2007: Mitigation.Contribution of Working Group III to the Fourth Assessment Report of theIntergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, United Kingdom and New York, NY, USA.

Larson, E., Wu, Z., DeLaquil, P., Chen, W., Gao, P., 2003. Future implications of China’senergy-technology choices. Energy Policy 31, 1189–1204.

Li, J., 2008. Towards a low-carbon future in China’s building sectordA review ofenergy and climate models forecast. Energy Policy 36 (5), 1736–1747.

Liu, F., 2006. Economic Analysis of Residential Building Energy-Efficient Design Stan-dards in Northern China: A Case Study of Tianjin City. World Bank Interim Report.

Long, W, 2005. Proportion of Energy Consumption of Building Sector and Target ofBuilding Energy Efficiency in China. China Energy (in Chinese), 27(10).

Long, W., 2007a. On scientific point of view of building energy efficiency. BuildingScience. 23, 2 (in Chinese).

Long, W., 2007b. Research on energy efficiency of residential buildings in Shanghai.(in Chinese)

Ministry of Construction of China (MOC), 2003. http://www.cin.gov.cn/INDUS/speech/2003111801.htm.

MOC, 1995. Industrial Standard of the People’s Republic of China (JGJ 29-95).Energy Conservation Design Standard for New Heating Residential Buildings.Ministry of Construction Enforcement, July 1, 1996.

NationalBureau of Statistics (NBS),2005. China Energy Statistical Yearbook 2005. Beijing.National Bureau of Statistics (NBS), 2006. China Statistical Yearbook 2006. Beijing.NDRC, 2004a. China’s energy development strategies and policy implications

(Scenario Analysis on Energy Demand). Beijing.NDRC, 2004b. National Long and Medium Term Energy Efficiency Plan. Beijing.NDRC, 2007. Long and Medium-term plan of renewable energy development in

China. Bejing.Pacala, Stephen, Socolow, Robert, 2004. Stabilization wedges: solving the climate

problem for the next 50 years with current technologies. Science 305, 968–972.

Point Carbon, 2008. Available at: http://www.pointcarbon.com/.Rich, D., 2004. Climate Change, Carbon Taxes, and International Trade: An Analysis

of the Emerging Conflict between the Kyoto Protocol and the WTO. Universityof California, Berkeley.

Salat, S. (Ed.), 2006. The Sustainable Design Handbook China: High EnvironmentalQuality Cities and Buildings. CSTB, Paris, 399 p.

Schipper, L., Haas, R., Sheinbaum, C., 1996. Recent trends in residential energy use inOECD countries and their impact on carbon dioxide emissions: a comparativeanalysis of the period 1973–1992. Mitigation and Adaptation Strategies forGlobal Change 1, 167–196.

Shanghai Municipal Real Estate and Land Resource Administration (SMRELRA) andShanghai Municipal Commission of Construction (SMCC). 2006. Shanghaibuildings energy consumption survey 2006.

Stern, N., 2007. Stern Review on the Economics of Climate Change. HM Treasury,London, UK.

Stern, N., 2008. Richard ELY lecture, American Economic Association MeetingsJanuary 2008. American Economic Review 98.

Thormark, C., 2002. A low energy building in a life cycledits embodied energy,energy need for operation and recycling potential. Building and Environment37, 429–435.

Toth, F., Cao, G., Hizsnyik, E., 2003. Regional Population Projection for China. IIASA,Austria.

Urge-Vorsatz, D., Harvey, L.D., Mirasgedis, S., Levine, M., 2007. Mitigating CO2emissions from energy use in the world’s buildings’. Building Research andInformation 35 (4), 379–398.

WBSCD, 2007. Energy Efficiency in Buildings: Business Realities and Opportunities.Weitzman, M.L., 1974. Prices vs. Quantities. Review of Economic Studies 41 (4),

477–491.Wellington, F., et al., 2007. Scaling Up: Global Technology Deployment to Stabilize

Emissions. World Resources Institute, Washington, DC.World Bank, 2001. China: Opportunities to Improve Energy efficiency in Buildings.

Washington.Zacarias-Farah, Geyer-Allely, E.A., 2003. Household consumption patterns in OECD

countries: trends and figures. Journal of Cleaner Production 11 (8), 923–926.Zhou, N., McNeil, M.A., Fridley, D., Lin, J., Price, L., De la Rue du Can, S., Sathaye, J.,

Levine, M., 2007. Energy Use in China: Sectoral Trends and Future Outlook.LBNL, 96 p.

China UrbanizesConsequences, Strategies, and Policies

Shahid Yusuf

Tony Saich

China’s outward-oriented industrialization, spearheaded by the coastalprovinces, led to a quickening of urbanization from the start of reforms inthe early 1980s. In 1980 China, with an urbanization rate of 19.6 percent,was less urbanized than Indonesia (22.1 percent), India (23.1 percent), orPakistan (28.1 percent) (table 1.1). By 2005, 42.9 percent of China’spopulation lived in urban areas, still somewhat below the global averageof 50 percent but close to the average for East Asia (41 percent) and wellin excess of India’s 28.7 percent rate.1

Between 1980 and 2000, 268 million Chinese entered into the urbandomain, mainly through migration from rural areas. This figure wasalmost twice as large as the increase of the urban population in the restof East Asia (table 1.2). By 2020, urbanization could pass the 60 percentmark, with 200 million or more rural dwellers joining the ranks of theurban population. The scale of urbanization in China will dwarf thatoccurring elsewhere in East Asia.

This shift in the demographic center of gravity has seismic implicationsfor China and major spillover effects for the rest of the world. It will bea driver of economic growth. The urban economy should generateenough jobs to absorb the additions to the urban workforce at steadily

C H A P T E R 1

Optimizing Urban Development

1

Shahid Yusuf and Kaoru Nabeshima

1 The classification of the urban population has changed overtime (see Fang 1990; Zhouand Ma 2003).

rising wages if the economy continues to grow at least 8 percent a year(feasible given the elastic labor supply); capital accumulation is sustained;and the scope for enhancing technological capabilities increases.2,3

Meeting these conditions is crucial for urban residents, because urbanemployment opportunities and the median urban wage will determine liv-ing standards in cities (see chapter 8). These conditions will have a bearingon whether the influx of people leads to the formation of slums, as it has inmany Latin American cities. They will also affect income distribution,4

2 China Urbanizes: Consequences, Strategies, and Policies

2 Bosworth and Collins (2007) estimate that total factor productivity rose by 4 percent ayear between 1993 and 2004 and that its contribution to overall growth was only a littleless than that of capital. He and Kuijs (2007) estimate that TFP grew by 2.8 percent perannum during 1993–2005.

3 New York’s real manufacturing wages kept rising in the second half of the 19th century,even though a significant number of immigrants came to New York. In the early partof the 20th century, when immigration to the United States increased dramatically,real wages in New York (the entry point of many immigrants) started to fall(Glaeser 2005b).

4 Because of the widening gap between rural and urban incomes and interprovincial dispari-ties, China’s income distribution, as reflected by the Gini coefficient, has risen rapidly, from0.33 in 1980 to 0.49 in 2005 (“China’s Income Gap” 2006). Other estimates of theGini coefficient in 2005 are slightly lower. The distribution of net wealth in urbanareas, which is strongly influenced by property ownership, is also becoming moreskewed, although it remains relatively equal (Wu 2004; Gustafsson, Shi, and Zhong2006; Saich 2006).

Table 1.2. Urban Population in China and East Asia, 1960–2005

(millions)

Year China East Asia

1960 130.7 86.2

1970 144.2 125.2

1980 191.4 177.6

1990 302.0 241.1

2000 459.1 314.2

2005 562.1 352.4

Source: Data for East Asia are from World Bank 2006. Data for China are from Fang 1990, Pannell 2003, and NBS 2006.

Note: East Asia includes Hong Kong (China), Indonesia, Japan, the Republic of Korea, Malaysia, the Philippines,

Singapore, Thailand, and Vietnam.

Table 1.1. Percentage of Population Living in Urban Areas in Selected Asian

Countries, 1980–2005

Country 1980 1985 1990 1995 2000 2005

China 19.6 23.0 27.4 31.4 35.8 42.9

India 23.1 24.3 25.5 26.6 27.7 28.7

Indonesia 22.1 26.1 30.6 35.6 42 48.1

Pakistan 28.1 29.3 30.6 31.8 33.1 34.9

Source: NBS 2006; World Bank 2006.

overall energy and water consumption, and the quality of life in cities. Arising median wage rate and a relatively egalitarian income distributionwould be broadly advantageous, but they would also push up per capitaresource consumption. Chinese cities would become larger users of localand global resources, including global public goods. At the same time,prosperity and technological capability would provide the means to containthe resource costs and externalities associated with growth.A slow-growthscenario or a scenario in which average income rises but incomes becomemore unequal might lead to somewhat lower resource consumption, per-haps more than counterbalanced by sociopolitical tensions, which couldjeopardize economic performance.

A host of policies will collectively determine growth, income distri-bution, resource use, and the quality of life. From an urban perspective,five sets of policies are especially noteworthy:

• Policies affecting rural-to-urban migration and intersectoral differencesin average household incomes

• Policies affecting the size distribution of cities and the relative concen-tration of people in major metropolitan centers

• Policies affecting the development of urban infrastructure• Policies that impinge on the availability of and access to public services

and social safety nets• Policies and institutions that regulate energy and water use in cities

and help control urban pollution.5

These policies are the primary focus of this volume.The rest of this chapter is divided into five sections. The first section

briefly reviews the history of urbanization since ancient times. Thesecond section describes some of the positive and negative conse-quences of migration to cities. The third section explores factors thatwill define China’s urban development strategy. The fourth sectionexamines the policies that will guide urban change. The fifth sectionprovides some concluding observations.

Urbanization in China since Ancient Times

Urbanization in China began almost 4,000 years ago, although Neolithicvillages had begun sprouting in river valleys as early as 5000 BC

Optimizing Urban Development 3

5 China’s plans with regard to energy-use pricing, efficiency, technology, and regulationare presented in NDRC (2007). Rosen and Houser (2007) assess the demand and sup-ply situation and the implications of China’s consumption on global markets. Shalizi(2007) provides an assessment based on a comprehensive modeling framework.

(Ebrey 1996). Ho Ping–ti writes of the “large urban centers” that arosein Shang times (circa 1700–1100 BC) and of the high walls of packedearth that surrounded many settlements, including most notably, thecities near Cheng-chou and An-yang in Hunan (Ho 1975; Friedman2005). The number of cities proliferated during the Zhou dynasty(1122—221 BC). Created primarily to fulfill military and administrativeroles, these cities also took on other functions (Zhao 1994).

The curve tilts ever so gently upward as urban populations began togrow during the Qin (221–206 BC) and Han dynasties (202 BC–AD 220).By the time of the Southern Song dynasty (12th century), 10–13 percentof the Chinese population lived in cities, with Kaifeng, the capital of theSong, having a population of almost 1 million people (Bairoch 1991).6

Mote (1999) surmises that the number of urban dwellers in China duringthe middle years of the Song dynasty equaled those in the rest of theworld at that time.

For a few hundred years thereafter, the curve remains fairly flat.However, the scale of cities such as Hangzhou on the West Lake impressedMarco Polo, who had seen no comparable center in Europe (“the mostsplendid city in the world . . . [with] 13,000 bridges mostly of stone”)(Polo 1958: 213). By the time the Ming dynasty was entering its twilightyears in the 16th century, the curve of urbanization had inched up anotch to 11–14 percent (Bairoch 1991). Major centers such as Beijingand Nanjing housed almost 1 million people, a handful of cities hadpopulations of half a million or more, and “scores of urban places” hadpopulations of 100,000. “Urban life was rich, comfortable, and elegant . . .varied and lively [in the larger cities]” (Mote 1999: 763).

By the late 19th century, the urban share of China’s population hadfallen to 6.0–7.5 percent, although the absolute number of urbandwellers rose, because population growth accelerated in the 18th andearly 19th centuries (Bairoch 1991; Zhao 1994). By this time, the indus-trializing countries of Europe had pulled ahead, with urbanization ratesof 61 percent in Britain and 29 percent in Europe as a whole. This gapbetween China and Western Europe had widened further by 1949, whenthe communist regime took hold of the reins of government.

Initially, the new government allowed cities to grow. Since the 1960s,however, China has sought to tightly manage the course of urbanization(see Kwok 1981; Fang 1990). The intersectoral movement of peopleand, from the 1970s, fertility rates were controlled with considerable

4 China Urbanizes: Consequences, Strategies, and Policies

6 Zhao (1994) cites a much higher figure of 22 percent for the urban population underthe Song dynasty.

success through the combined efforts of the Communist Party andthe government bureaucracy.7

The main instrument used to regulate movement is the hukou system,which assigns every person in China a residence in a specific locality.8 Thissystem distinguishes urban from rural residents, with urban householdsenjoying far more benefits and privileges than rural ones (see Friedman2005). The one-child policy—which is still enforced, albeit more flexiblythan it once was—meanwhile checked population growth, pushingfertility down from 5.9 in 1970 and 2.9 in 1979 to 1.7 in 2004 (Hesketh,Lu, and Xing 2005). Urban fertility was 1.3 in 2005, while the rate in ruralareas was a little less than 2 (Hesketh, Lu, and Xing 2005). By 2005China’s population was growing at 0.59 percent a year (NBS 2006).

Together these two policies slowed the increase in the urban share ofChina’s population to a crawl until well into the 1980s (Fang 1990).Migration between sectors was not brought to a complete halt, but thehukou system reduced it to a trickle, by making it difficult to find hous-ing or gain access to essential services outside of one’s official place ofresidence. Changing one’s residence and, most important, obtaining anurban hukou required and still requires hard-to-obtain official approval,especially in larger cities.

Once industrial and trade reforms gathered momentum in the 1980s,demand for workers from urban enterprises began drawing more migrantsto the cities, increasing the pressure on municipal authorities to relaxhukou rules. Initially, many cities resisted these pressures, preferring avery gradual easing of the restrictions for fear that anything more wouldattract an unmanageable flood of migrants. This change in policy stimu-lated the multiplication of industry in small towns and villages in ruralareas, which by 1990 employed 93 million workers (see chapter 2) andwas responsible for 17 percent of China’s exports of manufactures.9 By

Optimizing Urban Development 5

7 China’s efforts to regulate population growth gathered momentum after the CulturalRevolution in the late 1960s. Fertility was already declining in the 1970s before theannouncement and subsequent implementation of the one-child policy in 1979(Baochang and others 2007). See Hesketh, Lu, and Xing (2005) regarding the impactof the policy over a quarter century.

8 During the second half of the 1960s and in the 1970s, the government also “sentdown” urban youth to rural areas and redeployed millions of urban workers to interi-or southwestern provinces in order to disperse industrial capabilities and reduceChina’s vulnerability to attacks from abroad. On these rustication and Third Frontinland industry development programs, see Gardner (1971), Bernstein (1977),Naughton (1988), Fang (1990), and Demurger and others (2002).

9 The government encouraged these former “commune and brigade” enterprises,because they raised rural incomes and stemmed migration (Zhu 2000; Wu 2005).

1996 township and village enterprises (TVEs) employed 135 million andaccounted for 46 percent of exports (Li 2006).10 Rural industrializationdrew on an unforeseen reservoir of entrepreneurship and was aided byfiscal decentralization that encouraged lower-level cadres to take thelead in developing industry (Oi 1992; Qian 1999).

By the mid-1980s, attitudes toward urbanization began to shift, withcities coming to be viewed as “growth poles” and the “city as leading thecountry.” Some Chinese researchers argued that “growth poles should bescattered through the country, each sending waves of economic growthin its hinterland” (Fan 1997: 630). In order to accelerate industrializationand meet the needs of construction and other services, cities had toabsorb more migrants. Moreover, the small towns that had becomeimportant foci for industrial development grew to become substantialurban centers with concentrations of industry and were reclassified ascities, a process known as in situ urbanization.11

China retained the hukou system, but by the mid-1990s theinevitability of rising urbanization was widely accepted, only its speedremained an issue. Differing views came to determine the enforcementof hukou requirements. Viewing migration as a means of expanding theirindustrial bases and using the fiscal revenue generated to build urbaninfrastructure, small- and medium-size cities began to welcome the flowof labor from the rural sector. In contrast, many larger cities, especially incoastal areas, remained wary. They absorbed large numbers of temporarymigrants to satisfy their for industrial and construction workers needs12

but continued to use hukou to limit permanent migration.13

Benefits and Challenges of Urban Migration

Urbanization is now perceived as intrinsic to the process of growthand modernization, and the role of rural migration in diversifyingsources of rural incomes and narrowing intersectoral disparities inhousehold incomes is better understood (Knight and Song 2003). But

6 China Urbanizes: Consequences, Strategies, and Policies

10 Employment in TVEs declined thereafter but has since recovered, reaching 143 millionin 2005 (NBS 2006).

11 Zhu (2000) describes this process in Jinjiang county, Fujian Province. See also chapter 2.12 Pannell (2003) describes the regional pattern of urbanization in China and the demo-

graphic structure of the urban population.13 In some cities, particularly in Guangdong and Fujian (for example, Dongguan), non-

residents account for up to half of the population.

the desire to manage migration and contain the costs of urban housingand social benefits provided to residents means that there is anunwillingness to dismantle the hukou system, although Beijing hasallowed local governments much greater discretion regarding how itis applied and enforced.

Per capita annual income disparities of 1:2.4 between rural andurban areas and vastly greater job opportunities in cities make it highlyattractive for rural people to migrate.14 The result is that migration isadding to the numbers of registered urban residents and swelling theranks of the so-called “floating population,” made up of people with ruralhukou who are temporarily living and working in cities (see chapter 3).The size of this transient (inter- and intracounty) population was almost148 million in 2005 (see chapter 3); they are most numerous in easternmetropolitan centers, such as Beijing, Guangzhou, Shanghai, and Shenzhen,which have plentiful jobs and the “bright lights” that draw migrants.15

The influx of migrants, permanent or floating, has had a number ofpositive and negative effects. The migration of mainly young people tocities drawn from the better-educated rural cohorts has promotedgrowth by enhancing the labor supply and by injecting an additionaldose of entrepreneurship and dynamism into the urban labor market(Bloom and Williamson 1997).16 The remittances migrants send to theirvillages have significantly bolstered rural household consumption, insome cases contributing as much as 40 percent of annual householdincomes (the average is closer to 20 percent). Migrants have helped bringliving standards in some of the poorest rural areas closer to urban levels.These and other positive outcomes outweigh some of the problemsassociated with migration.

Optimizing Urban Development 7

14 The unadjusted differential in rural and urban incomes is 1:3.5. See chapter 2 and tables10.8 and 10.18 in the China Statistical Yearbook (NBS 2006). The rural and urbanincome divide is the main cause of income inequality in China (Sicular and others2007). Tsui (2007) shows that inequality among provinces has arisen from the alloca-tion over time of capital and FDI and the influence that allocation has had on totalfactory productivity across provinces.

15 The provinces and cities that have attracted the largest number of migrants areGuangdong, Zhejiang, Jiangsu, Shandong, Beijing, Shanghai, and Guangzhou. About15–20 million migrants work in Guangdong (“Delta Dreams” 2006), the destination ofmigrants from Hunan, Jiangxi, Sichuan, Guangxi, and Hainan. Shandong has attractedmigrants from Heilongjiang and Liaoning (see chapter 3; Fan 2005).

16 Bloom and Williamson (1997) find that demographic shift, which affected labor forcegrowth, age structure, domestic savings, and domestic investment, was responsible for1.4–1.9 percent of the annual growth in GDP in East Asia between 1965 and 1990.

Perhaps the most serious concern centers on the risk migrants run ofbecoming part of the urban poor.17 Temporary migrants have limitedaccess to health and education services. Their age makes them healthierthan older people but leaves them more vulnerable to accidents andchildbirth. Migrants are less likely to visit a doctor when sick and morelikely to self-medicate. The cost of sending their children to school canbe a major burden. Some live in crowded conditions, although the evidenceon this problem is equivocal.18

While migrants’ income net of remittances can be meager, migrantsare less likely than elderly or disabled urban residents to fall below thepoverty line. The evidence reported in chapters 2, 3, and 4 suggeststhat only a small percentage of urban migrants can, strictly speaking, beclassified as poor.19

There are worries that the departure of many young educated workerswill denude the countryside of skills, know-how, and entrepreneurship.Thisis unlikely for some time to come, if ever. China’s farm population exceedsthe numbers needed; many workers return to their villages after a stint inthe cities and invest their earnings in farming or other rural activities.Moreover, the most educated are less likely to migrate (see chapter 3;Murphy 2002).20

The increase in China’s urban population from 191 million in 1980to 562 million in 2005 has called for massive investment in urbanhousing and infrastructure. Amazingly, China has been able to absorbmore than 370 million people in its cities without the proliferation ofurban slums, although sewerage and waste disposal services have struggledto keep up with demand. Between 1990 and 2000, 130 million newurban dwellers were provided access to improved sanitation facilities;

8 China Urbanizes: Consequences, Strategies, and Policies

17 Townspeople have traditionally displayed antipathy toward rural migrants, finding itdifficult to comprehend their dialects, complaining about their lack of culture, andblaming them for bringing crime and disease to urban areas. Some of this antipathypersists and is responsible for the continuing resistance to migration (Zhang 2001).According to Ravallion, Chen, and Sangraula (2007), urban poverty is just 4 percent ofthe rural rate, and it has remained low since the mid-1990s, even in the face of heavymigration to the cities.

18 Wu (2002) finds that about a third of all migrants in Shanghai live in dormitories pro-vided by employers and about half rent their accommodations. Overall, migrants insome of the larger coastal cities appear to be living in housing equivalent to that ofurban residents at similar levels of income (Jiang 2006).

19 The exclusion of dependants could introduce some bias.20 Murphy’s (2002) study of counties in Jiangxi describes the proactive approach taken

by officials to entice back migrants who have accumulated capital and skills while incities. Returning migrants are responsible for establishing numerous businesses in theirhome towns and villages.

nearly a third of the urban population still lacks these services, however(Mohan 2006).

Old workplaces, their housing compounds, and concentrations of smallbusinesses contribute to urban dilapidation in inner-city areas. On thefringes of major cities such as Beijing, migrants have created small enclaves,where housing quality can be variable. In Beijing’s Zhejiang village, forexample, living standards are relatively high, because migrants work insmall businesses that produce clothing and footwear. Migrants from Henanwho collect rubbish are much poorer (Wu 2004).

The relative smoothness of the urban transition has been made possi-ble by the availability of investment funds intermediated by the bankingsystem, the remarkable strides made by the construction sector, andacceptable growth in regulatory capacity in urban centers. But the roleof capital generated through high domestic savings has been paramount.

Urban investment started from a modest base in the mid-1980s andaccelerated throughout 1990s, although growth slowed in 1999 and2000. Beginning in 2001, urban investment picked up the pace again,registering double-digit growth, especially in 2003, when it rose 20 per-cent, and 2004, when it rose 31 percent. For 2004 the share of housingexpenditure is estimated to have been 10 percent of GDP, with totalinfrastructure spending estimated at almost 20 percent of GDP (Yusufand Nabeshima 2006b).

This investment has made an enormous difference, visible to any visitor.More important, it has enabled China to accommodate a far-reachinggeographic and intersectoral distribution of the population. Might theresources have been allocated more efficiently and through more-variedfinancial channels? Chapter 5 indicates how the allocative process hasbeen distorted by the need to sustain state-owned enterprises and bythe actions of local authorities, who are able to influence banks’ lendingdecisions. But the fact remains that no country has matched the scale ofChina’s achievement in mobilizing financial resources using the bank-ing system to funnel capital to urban development. This financialwidening is signified by the high rates of M3—a definition of the moneysupply that includes currency, demand deposits, savings and timedeposits, ODs, money market accounts, eurodollar deposits, and REPOSand the ratio of loans to GDP, which compare favorably with those ofJapan and exceed those of Brazil, the Republic of Korea, and Mexico.However, the low ratios of bonds and market capitalization of listedcompanies to GDP indicate a good deal of room for enhancing financialdepth (figure 1.1).

Optimizing Urban Development 9

Urban residents consume far more energy than their rural counterparts.21

The energy is used for transport, heating, cooling, for generating electricity,and industry. In 2005, agriculture, forestry, and fishery consumed only3.6 percent of total commercial energy. Urban households consumed3.63 times more commercial energy than their rural counterparts.22 Asmore of China’s population locates in cities, commercial energy use percapita is bound to rise significantly.

Other factors also contribute to the energy intensity of the urbaneconomy. They include motorization, space heating and cooling, and theproliferation of energy-using appliances, all of which have high-incomeelasticities.As urban households become more affluent, their demand for

10 China Urbanizes: Consequences, Strategies, and Policies

0

20

40

60

80

100

120

140

160

Japan

per

cen

tag

e o

f G

DP

Liquid liabilities (M3) aspercentage of GDP

Loans as percentage of GDP

Market capitalization of listedcompanies (percentage of GDP)

Public sector bonds(percentage of GDP)

Private sector bonds aspercentage of GDP

China Korea, Rep of. Brazil Mexico

Source: Data on M3 and market capitalization are from World Bank 2006; data on public and private sector bonds

as a percentage of GDP are from the World Bank’s internal financial data base; data on loans as percentage of

GDP are from International Fianancial Statistics (IFS) Chinese bond data are from Mu 2005.

Figure 1.1. Financial Development in Selected Countries, 2005

21 For energy production in China from renewable and nonrenewable sources between1980 and 2002, see Chen and Chen (2007a, 2007b).

22 This does not include energy consumption by rural enterprises or use of energy frombiomass (Pan 2002).

all three will continue to push up energy consumption.23 In 2005, industryaccounted for 70 percent, transport for 7 percent, and households for10 percent of energy consumed (NBS 2006).24 The share of transport issure to rise substantially, propelled by the trend toward private vehicleuse.25 Household consumption of electricity will also rise (see chapter 6).

Energy consumption creates negative externalities, in the form of carbondioxide pollutants, and acid rain. In 2006, China became the leadingemitter of carbon dioxide into the atmosphere (6.2 billion tons as against5.8 billion tons by the United States). Release of sulfur dioxide and particu-lates, particularly in the northern parts of the country, is exacerbated by theshortage of water.As a result, only a very small fraction of the coal used canbe washed to rid it of sulfur, ash, and impurities (Roumasset, Wang, andBurnett 2006).26 As energy use climbs, air pollution in Chinese cities fromnitrogen oxide, sulfur dioxide, and particulates—already among the mostsevere in the world—could become even more intense.27 Of the 20 citieswith the worst air pollution in the world, 16 are in China (Wu 2006; Hoand Nielsen 2007).28

China’s energy demand mirrors the unusually dynamic growth of itsurban economy. The increasing use of energy is a sign of economic vigorand rising incomes. Between 2000 and 2005, the elasticity of consumptionaveraged 0.93; in 2005, 69 percent of energy derived from coal and 21 percentfrom oil. Of the energy derived from petroleum, net imports accounted for

Optimizing Urban Development 11

23 Although shifts in consumption that increase the share of services will reduce energyconsumption, the Economist Intelligence Unit estimates that China’s energy consump-tion relative to that of the United States will rise from 39 percent in 2000 to 86 per-cent in 2011 (“The Health of a Nation” 2007).

24 Industry and transport absorb 80 percent of petroleum consumed, mostly in the formof middle distillates (CBO 2006).

25 Ownership of passenger vehicles increased from 9.9 million vehicles in 2001 to 21 mil-lion in 2005, propelled by an easing of consumer credit (Roumasset, Wang, and Burnett2006; China Statistical Yearbook 2006). Some cities, such as Suzhou, are taking a leadin improving air quality by encouraging the use of motorbikes that use battery power.But even in this “beautifully preserved” and “well-tended city,” the air is “almostunbreathable” and the “canals are filled with black bubbling water” (Cheng 2006: 1859).

26 The impurities present in the coal being shipped increases the burden on China’s rail-way system, which devotes 40 percent of its capacity to the transport of coal (“FreeFlow” 2005). Washing coal is not without complication, because the sludge and waste-water must be treated to avoid localized pollution.

27 The severity of air and water population was already evident in the early 1970s. Inresponse, a national conference was held in 1973 and a basic environmental law passedin 1979 (Kojima 1987). Acid rain falls on one-third of China; emissions from Chineseindustry and power plants also contribute to acid rain in Japan and the Republic ofKorea (Roumasset, Wang, and Burnett 2006).

28 The world’s most polluted city is Linzen, in Shanxi, which produces coke. Lanzhou,the capital of Gansu, is also among the top 10 (“Lanzhou to Walk” 2007).

44 percent, and the share is rising (CBO 2006; NBS 2006). Given thedepletion of petroleum resources and the threat of climate change inducedby greenhouse gases (which could also contribute to a significant reductionin national crop fields by mid-century) the implications of China’s urban-ization are disconcerting in the medium term and troubling over the longerrun. Limiting the energy intensity of urban development will be a struggle,but it is one that policy makers will find impossible to sidestep.

While fossil fuels can be imported, water in the quantities requiredcannot. Desalinating seawater consumes energy; pumping the waterinland adds to the energy costs. Looking ahead, urban development inthe drier regions of China is likely to be circumscribed by the availabilityof usable water. Currently, two-thirds of China’s cities are faced withwater scarcity, caused by the uneven geographical distribution of watersupplies, the diversion of water for agricultural purposes, and pollutionfrom industrial sources (especially organic material), which renders upto 70 percent of the water from five of China’s seven largest rivers unfitfor consumption. Research on the Pearl River delta area shows that urbanriver water is far more polluted than water in rural counties (Ouyang,Zhu, and Kuang 2006). At great cost, China is diverting water from theYangtze to the northeastern part of the country to meet the needs of theincreasingly water-stressed 3-H region (the basins of the Huai, Hai, andHuang [Yellow] Rivers). Once completed, this effort will provide somerelief, though for how long is uncertain, as are the ecological consequencesfor the Yangtze basin and the receiving region.

Per capita water availability in 2005 was just 2,152 cubic meters, only12 percent of which was used for household purposes. Urban per capitawater consumption in China is relatively modest compared with Japanand the United States (see chapter 7). But water is inefficiently utilized,because it remains underpriced. China uses 6 times more water per unitof GDP than the Republic of Korea and 10 times more than Japan. For thisreason, underground water is also being pumped at unsustainable rates,causing the water table to fall, increasing the mineral content of water, andresulting in the subsidence of the land in cities and the infiltration ofbrackish water into subterranean aquifers.29

12 China Urbanizes: Consequences, Strategies, and Policies

29 The North China Plain derives close to 60 percent of its water supplies from ground-water, and according to some projects, the aquifers could be largely depleted in 30 years(“Beneath Booming Cities 2007; Evans and Merz 2007). Groundwater tables havedropped by as much as 90 meters in the Hai plains and by 100–300 meters in Beijing;they are also dropping in Shanghai and Shijiazhuang, where many wells must be dugto a depth of 200 meters to find clean water (see chapter 7). See Pielou (1998) on theproblems caused by the unsustainable extraction of groundwater.

More than 680 million Chinese live in the drier northern region (whichhas just one-sixth of the per capita water supply available in the southand one-tenth of the world’s average), more than half of them are urbandwellers. As this ratio climbs to 60 percent in the next 15 years and percapita urban consumption rises, as is likely, the intersectoral allocation ofwater and the management of water use will require strategic thinking onthe cultivation of water-intensive cereals such as corn and wheat, the courseand shape of urban development and the effective coordination of basin-wide water management both surface and subterranean (see chapter 7).The policies described below will have to be applied with considerableforce, as there are no substitutes for water.

Urban development is a complex, multifaceted process; farsightedand entrepreneurial management is a key to success. The sheer pace atwhich Chinese cities are expanding and the decentralized structure ofgovernment puts a particularly high premium on the planning andmanagerial skills of local authorities. Cities in China have coped moreeffectively with rapid urbanization, the mobilization of resources, thebuilding of infrastructure, and the wooing of industry than cities in othermiddle- and lower-middle-income countries. Moreover, the country’slarger cities are better governed than its smaller ones (see chapter 8). Sofar, most cities have been able to arrest the spread of slums (Flavin andGardner 2006; Jiang 2006) and contain the spread of crime. Chinesecities are cleaner than average, and in the majority of cases, the combinedefforts of Street Offices and higher-level municipal departments ensure thatpolicies are competently executed. This is a considerable achievement, andthe fact that citizen satisfaction levels rose between 2003 and 2005 is agood sign (see chapter 8). On average, other large countries, such asBrazil, India, and Indonesia, lag well behind China in terms of effectivemunicipal functioning.

However, according to the World Values survey conducted in 2006,Chinese have much more confidence in their government than Americans,with 97 percent of Chinese and just 37 percent of Americans expressingconfidence. Moreover, 84 percent of Chinese but just 37 percent ofAmericans believe that the government is not in the grip of specialinterest groups (Shiller 2006).30 The findings reported in chapter 8indicate that urban Chinese place less trust in their local governments(67 percent) than in the central government (84 percent).

Optimizing Urban Development 13

30 A Lichtman/Zogby poll conducted in late 2006 found that only 3 percent of Americanssurveyed had trust in the U.S. Congress (“The Way We Were” 2006).

The dissatisfaction of China’s urban dwellers with local governmentsderives from four sources. First, corruption is a major concern. As inother countries, it is associated with land deals, construction projects,bank lending, social security funds, and other activities.31 TransparencyInternational ranked China 70th of 163 countries in 2006, but this typeof index provides only a partial perspective (Transparency Internationalhttp://www.transparency.org/).

Second, there is dissatisfaction with the provision of health services,more so than with education.32 Even for privileged urban residents, healthservices are becoming less accessible and costlier. More and more peoplehave to pay out of pocket for health care and medications, and the shift tocurative care is shortchanging more cost-effective preventive medicine.33

Third, as cities expand into periurban areas, the confiscation and saleof farm land to developers is being strongly condemned, especially bydisplaced farmers, who receive limited compensation and face difficultyfinding employment in the urban labor market. Others view these salesas evidence of corrupt dealings and inept fiscal management, becausecurrent expenditures are being offset by the proceeds from these salesrather than being aligned with revenues appropriately augmented byintergovernmental transfers.

Fourth, the urban safety net for the poor—a mix of the widowed eld-erly, the disabled, laid-off state enterprise workers lacking marketableskills, people working in the informal sector, and migrants—is inadequate(Wu 2004). Di Bao—a means-tested transfer that offers minimal assistanceto urban residents who satisfy the poverty criteria—is a bare-bones schemethat deserves to be augmented or supplemented by additional assistance.Beyond this, there is growing concern regarding unemployment compensa-tion for laid-off workers and the adequacy of pension benefits.34

These are not minor complaints, and they are rising in volume, despitethe efforts of the central government to root out corruption with fre-quent inspection tours by the Communist Party’s Central Commission for

14 China Urbanizes: Consequences, Strategies, and Policies

31 In September 2006 a number of officials from the Shanghai administration, includingthe mayor, were implicated in the misappropriation of US$400 million from themunicipal pension fund (“Anti-Graft Campaign” 2006; “Shanghaied” 2006).

32 Gan and Gong (2007) show how periods of morbidity before the age of 21 significantlyreduce an individual’s education status.

33 Medical expenses account for 11.8 percent of household consumption, more than edu-cation or transport (“China’s Income Gap” 2006).

34 Wu (2005) provides a detailed account of how China has developed the elements of asocial security system (pensions, medical insurance, and unemployment compensation)and reviews current reform options. Other proposale for reforming the pension systemare presented by Dunaway and Arora (2007).

Discipline Inspection, which meted out harsh punishments.Wu (2006) citesan official report indicating that 42,000 public officials were investigatedfor corrupt practices each year between 2002 and 2005 and that action wastaken against 30,000 every year.35 He notes that corruption was largelyresponsible for losses by the banking system equal to 6.25 percent of GDPbetween 1999 and 2001 and fraudulent public expenditures amounting to2.4 percent of GDP. Reforms of the bureaucratic structure and incentives,the health system, local taxation, revenue sharing with the central govern-ment, transfers from the central government to cover the costs of unfundedexpenditure assignments, and social security are all ongoing, but theybarely keep up with the problems. As a result, the clamor about urbangovernance is not subsiding; as China’s urban middle class grows andbecomes more aware, protests could become more widespread.36

Governance issues may be easier to resolve, if partially, in China thanelsewhere, because unlike many other countries, it has a vibrant urban econ-omy that is generating jobs and constantly adding to the pool of resources.China does not face entrenched problems of slums, urban decay, an impov-erished underclass, or low fiscal buoyancy, and so far it has been able toabsorb migrant flows (Flavin and Gardner 2006). China’s cities have per-formed relatively well, and many are governed by able and energetic lead-ers who are eager to improve economic circumstances and living conditions.

Crafting an Urban Development Strategy

Like many other countries, China is seeking a development path that tendsto equalize rural and urban per capita incomes over time.37 This objective,emphasized in the 11th Plan, calls for comparable growth rates acrosssectors (see Yusuf and Nabeshima 2006a). Barring that, rough paritybetween sectors can be maintained only by a decline in the population of

Optimizing Urban Development 15

35 By redoubling its efforts in 2006 and firing four high-level officials, the government hasmade some headway (“China’s Crackdown” 2006).

36 The number of protests rose tenfold between 1993 and 2005, to 87,000 (“In Face ofRural Unrest” 2006; Wu 2006). The spike in protests appeared after 1996, when thereform of state-owned enterprises began to add to the ranks of the urban unemployed.Many of those complaining are former state enterprise employees and displaced farm-ers. Some of these and other protests fall into the category of “rightful resistance,” inwhich protestors frame “their claims with reference to protections implied in ideologiesor conferred by policymakers” (O’Brien and Li 2006: 3).

37 Per capita rural incomes were below the national average in 21 of 31 provincial-levelunits in 2005. Rural per capita incomes were 20 percent below the national average inSichuan and Chongaing and 40 percent below in Gansu and Guizhou (“China: Doesthe Countryside?” 2007).

the slower-growing sector or income transfers from the higher-income sec-tor to the lower-income sector.

Narrowing Rural–Urban GapsCrop yields in China are high relative to China’s main comparators(table 1.3), leaving little scope for more than a very modest annualincrease. Rice yields are close to those of Japan and the United States andwell above those of Vietnam. Yields of wheat match those of the UnitedStates.These high yields are achieved through the use of improved seeds,the heavy application of fertilizers, and in the north through increasingreliance on groundwater. Farmers in China use 228 kilograms of plantnutrients per hectare—far more than the world average of 90 kilogramsin 2002 (FAO 2003). By using agricultural extension services effectively,Chinese farmers have introduced new varietals and exploited biogenetictechnologies, bringing themselves close to the technological frontier forfood grains (Jin and others 2002), especially in rice production.38 Thegap between potential and actual yield is only 15 percent. This gap ismuch larger in India (58 percent) and the Philippines (65 percent)

16 China Urbanizes: Consequences, Strategies, and Policies

Table 1.3. Rice, Wheat, and Maize Yields in Selected Countries and Regions, 1997–2002

(tons per hectare)

Rice (2002) Wheat (1998–2000) Maize (1997–99)

Country Yield Country Yield Country Yield

United States 7.4 China 3.8 United States 8.3

Japan 6.6 United States 2.9 Brazil 5.3

China 6.3 Argentina 2.5 China 4.9

Vietnam 4.6 Canada 2.4 Mexico 2.7

Thailand 2.6 Russian 1.6 Argentina 2.4

Federation

Asia 4.0 East Asia 3.8 East Asia 4.8

World 3.9 World 2.7 World 4.3

Source: Pingali 2001; Ekboir 2002; data on rice are from World Rice Statistics (http://www.irri.org/science/ricestat/).

38 Genetically modified crops are being widely researched and planted in China. Chinabegan research on genetically modified crops in the early 1980s and is now one of theleading countries in this field (Falkner 2006). Bt cotton is a transgenic strain of cottonthat incorporates the genes of a soil-dwelling bacterium, bacillus thuringiensis, hence thename. The added genes induce the cotton plant to secrete toxins, which reduces thedepradations of certain caterpillars, beetles, and flies that feed on the plant and candestroy it. Bt cotton was approved in 1997, and 3.7 million hectares were planted in2004. Genetically modified varieties of rice, wheat, soybean, potato, rapeseed, cabbage,and tomatoes are under development or being introduced (Huang and others 2007).

(Jin and others 2002). According to Liu and Wang (2005), between1991 and 1999, technological advances were responsible for more thanhalf of growth in agricultural productivity in China.

A continuing shift toward animal husbandry, horticulture, and off-farm activities should gradually raise farm incomes, but substantial gainsthrough a large increase in the prices of major grains, for example, wouldincur heavy fiscal costs, face resistance from urban interests, and be subjectto restrictions by the World Trade Organization (WTO). The possibilityof widening the scale of off-farm activities exists, but TVEs have passedtheir high-water mark, and industry thrives more in urban and periurbanlocations than in rural ones.

Rural development has been the objective of a succession of governmentprograms, including, most notably, the 8–7 program, which spannedmuch of the 1990s.39 Other programs are building infrastructure andattempting to improve the delivery of social services. Recently, thegovernment has taken steps to raise the disposable incomes of agriculturalhouseholds by eliminating the agricultural income tax. Despite thesemeasures, bringing rural incomes closer to the urban average is provingto be an uphill task.

The challenge of narrowing income gaps is similar to that experi-enced in more-advanced countries. In Japan, for example, income dif-ferentials between sectors narrowed only as a result of migration, whichsharply reduced rural populations; generous agricultural price supportprograms; and the increase in off-farm employment opportunities. In2003 per capita incomes in the leading rice-producing prefectures, suchas Niigata and Akita, were close to those in Osaka and 70 percent of percapita income in Tokyo (Japan Statistics Bureau 2005).

A mix of policies will be needed in China, but a significant narrowingwill depend mainly on migration plus remittances. Other policies willalso play roles, however. These include (a) continuing efforts to strengthenagricultural productivity through diversification into higher-value activities;(b) technological advances that raise yields and conserve land, water, andother inputs; (c) investment in rural infrastructure in areas where returnsover the longer term are high; (d) provision of secure, longer-term prop-erty rights over farmland;40 (e) provision of better social services for

Optimizing Urban Development 17

39 The program, announced in 1993, provided subsidized loans, supported public works,and offered budgetary grants (Park, Wang, and Wu 2002).

40 The recently passed property law strengthens ownership rights and allows farmers torenew land leases (“Caught between Right and Left” 2007).

rural households; (f) rural credit schemes; and (g) to the extent feasible,resource transfers via the price mechanism or fiscal channels.

From the perspective of a development strategy that seeks to maintainhigh aggregate growth and bring rural incomes closer to urban levels, amultistranded approach is warranted. In conjunction with pricing poli-cies, efforts to raise agricultural yields, conserve water, promote diversifi-cation, and strengthen the transport and marketing infrastructure canincrease rural incomes and temper the incentives to migrate to cities.Creation of infrastructure should focus on areas with long-term potential,however; other kinds of transfer and income support are better suited forrural communities in which the land is infertile and water scarce.Encouraging people to move out of fragile areas is the most-sensibleapproach from both economic and ecological perspectives.41 Attemptingto improve their livelihoods through costly investments is likely to havea modest payoff and only delay by a few years an exodus from these areas.Regional policies—in Brazil, Italy, and other parts of the EuropeanUnion—have a poor record (Sinn and Westermann 2001).

Directing Migrant Flows and Managing Urban Growth Migration should be to where jobs are going to be; it should supportgrowth in urban regions with the greatest longer-term promise.Directing migrants to high-growth areas would ensure that they areabsorbed by urban labor markets and increase their chances of beingassimilated into urban society. It is when migrants enter slow-growingor stagnating urban economies that problems of unemployment lead tosocial problems and the flaring of tensions between newcomers andlongtime residents.

Geographical location and city size have the greatest effect on whetherurban migration can contribute to a virtuous urban growth spiral. Migratingto a coastal location or a location on a major transport artery was favored inthe past and remains advantageous, even though great advances in surfaceand air transport should have diminished the relative attraction of suchlocations.42 Coastal cities in particular exert an unusually strong pull, whichis linked to the quality of their physical environment and their milder cli-mate. With sea levels set to rise, some coastal cities might be endangered

18 China Urbanizes: Consequences, Strategies, and Policies

41 Current policies are helping shape such a trend: by the end of 2005, 23 million hectaresof low-quality farmland had been converted to woodland or grassland (“Saying ‘No’”2006).

42 See Liu (1993) on the location of Chinese cities.

three and four decades from now, but for the moment, the pull they exertis undiminished.43

The availability of fresh water is emerging as an additional determinantof urban growth and livability, as cities grow very large and becomevoracious consumers of water.44 A location along a waterway can helplessen water-supply constraints, and waterfront development can enhancethe quality of urban life.

The size of cities is also important.45 Economies of scale and agglom-eration increase growth rates as cities expand (see, for example, Yusuf andothers 2003; Rosenthal and Strange 2004). Agglomeration contributes bydeepening labor markets, inducing technological spillovers, and encouraginga wide mix of activities. It enables firms and consumers to more easily accessinputs and services and allows networked clusters of firms to emerge.Agglomeration also supports innovation—sometimes at the intersectionof two or more activities or scientific disciplines—and the diversificationof goods and services (Bettencour, Lobo, and Strumsky 2007; Carlino,Chatterjee, and Hunt 2007). Such diversification is often the principalavenue for increasing sales in national or world markets. In a globalizingeconomy, agglomeration economies are a safety valve permitting urbanindustry to expand in new directions and to maintain both a diversifiedportfolio of outputs and the potential for adding new activities as someexisting ones die out.46

A large urban center also provides an environment in which firmshave an easier time achieving scale economies, because local markets arelarge and enable firms to move down the cost curve before venturing

Optimizing Urban Development 19

43 The likelihood that some coastal areas will be submerged as seawaters rise might call forplanning with regard to the development of coastal cities (“Cities Should Plan” 2005).The experience of the Dutch will become more and more relevant. Among China’smegacities, Shanghai confronts the greatest challenge, because of its limited elevationabove the current sea level and low-lying terrain; subsidence caused by groundwaterdepletion; scouring of coastline by strong currents; the presence of wetlands and flood-prone areas; and susceptibility to typhoons (Sherbinin, Schiller, and Pulsipher 2007).

44 China is home to 22 percent of the world’s population but just 8 percent of globalfresh water supplies (Flavin and Gardner 2006).

45 China has three megacities with populations of more than 10 million: Beijing,Shanghai, and Chongqing. A fourth, Shenzhen, probably falls into this category if thenonresident population is included.

46 Although China’s exports to the United States increasingly overlap with those fromOECD countries, these exports sell at a discount, because their quality and technologicalsophistication are lower (Schott 2006). Hummels and Klenow (2005) and Hausmannand Klinger (2006) suggest that the growth of export revenues depends on diversifica-tion into new products (many in product categories that are close to those of currentexports) and or improvements in quality as incomes rise.

into overseas markets.47 Large cities are more likely to offer environmentsthat are contestable, if not competitive, with low barriers to the entryand exit of firms and greater incentives for firms to be innovative.

Little in the empirical literature suggests that cities are subject todiminishing returns to scale, but poorly planned and managed cities canconfront serious issues of congestion, pollution, and high living expenses,particularly as a result of increasing rents. These problems can alsoaffect medium-size cities. Cross-country experience shows that goodland planning, regulation, and coordination by bureaucracies as well asadministrative subdivisions can enable cities to reap the benefits of sizeand avoid most of the pitfalls. In fact, as survey evidence presented inchapter 8 suggests, that larger Chinese cities tend to be better managedthan smaller cities.

Polycentric spatial development (which prevents the congestion arisingfrom a single downtown focus with the help of zoning regulations andthe use of floor area ratios to vary population densities and create mul-tiple foci) and a well-designed transport system are key to making largecities livable. Also important are land-use policies that conserve landthrough densification and mixed use without sacrificing essential greenspaces and recreational amenities conducive to livability. Legislating rulesis one key step; enforcing them firmly, but when needed, flexibly, isanother. Cities often fail to follow through with policies governing theuse of automobiles. As a result, they end up with severe congestion;urban sprawl, which increases energy consumption; and air and noisepollution.48 The capacity to implement policy is thus a hallmark of thesuccessful metropolitan area.

Large cities can encounter difficulties if they do not mobilize sufficientrevenues to defray current expenditures or fail the test of creditworthiness,which makes it hard for them to raise capital from capital markets forlong-lived investments.This problem is not limited to large cities, althoughthe bigger centers are more likely to be burdened with fiscal expenditures.

In summary, size is a plus. In a more-open and competitive global econ-omy, a large city gains an edge from agglomeration and urbanization

20 China Urbanizes: Consequences, Strategies, and Policies

47 Pannell (1992) reviews the history of Chinese cities through the early postreform period.He finds that large cities were more efficient than smaller ones.

48 Sprawl is a particular problem for secondary cities in China. Because of the lack ofinfrastructure financing, new urban development tends to take place along existinghighways or trunk roads, without much planning. This contributes to sprawl andincreases commute time, congestion, the cost of providing energy, water, and sanita-tion infrastructure, and pollution.

economies that impart industrial flexibility. Major urban centers also enjoythe advantages of global transport connections and are more likely to behooked into the international business networks for manufacturing, pro-ducer services, and research. These international links are sources of trade,capital, and ideas, the oxygen that gives life to urban dynamism.

Financing Urban Development Urban development is not possible on the cheap. Huge volumes of fundshave to be raised and committed to projects that can take many years tocome to fruition and the effective life of which can span decades oreven centuries.

Cities faced with the prospects of substantial in-migration canbecome caught in low-level traps if they fail to pour capital into suchinvestments in a fairly short period of time. A “big-push” investmentstrategy has obvious merits for putting in place axial transport, housing,commercial, energy and water, and sewerage infrastructures.

Building ahead of demand makes sense, so that industry is not hamstrungby capacity constraints and urban physical plant can accommodate theinflux of people without congestion and the creation of slums. An exampleis New York City, which was designed for 1 million residents when thepopulation was barely 100,000. Central Park was created 150 years ago, andthe subway system was built 100 years ago, well ahead of demand (“TheNew New York” 2006).49 Planning for long-term growth smoothed theexpansion of Tokyo in the postwar period.50

Several Chinese cities have followed this route. Throughout the1980s, Shanghai spent 5–8 percent of its GDP on urban infrastructureinvestment. In the 1990s it spent 11–14 percent of GDP, in a big pushto redevelop the city, including developing Pudong (figure 1.2); thiseffort is now winding down. Both Beijing and Tianjin spend more than10 percent of their GDP on urban infrastructure (Yusuf andNabeshima 2006b). To cope with the rising demand for electricity andto eliminate brownouts, China commissioned 80 gigawatts of generat-ing capacity in 2006 and will put an estimated 75 gigawatts on line in2007 (“What Shortage?” 2006).

Optimizing Urban Development 21

49 Omnibus services began in the 1820s (Glaeser 2005b).50 Until the 1960s, fewer than 100,000 housing units were constructed in Tokyo.

Beginning in the 1970s, in line with the rapid increase of population in Tokyo, the con-struction of new housing units accelerated dramatically, reaching more than 900,000units by 1990.

Where cities have approached infrastructure development piecemealand lagged behind demand—Lagos, Lima, Karachi, and Mumbai are primeexamples—the urban environment has deteriorated.51 As populationpressures have mounted, industry has struggled to grow, fewer jobs havebeen created, and cities have entered into a low-level growth syndromein which poverty, slums, and crime have become firmly entrenched.52

Conventional wisdom has resisted the big-push strategy, conflating itwith lumpy investments. But experience with migration-led urbanizationsuggests that a high level of investment in industry and infrastructure hasmultiplier and accelerator effects, which can stoke prolonged virtuousspirals that generate not just growth and employment but also the urbanfacilities and housing needed to accommodate a rising population.

Urban development—especially when it is driven by a big-pushstrategy—requires capital, lots of it. Although the global integration ofcapital markets has created channels for the circulation of capital, muchof this capital must come from domestic saving and financial entities

22 China Urbanizes: Consequences, Strategies, and Policies

0

2

4

6

8

10

12

14

16

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

per

cen

tag

e o

f G

DP

Source: Shanghai Municipal Statistical Bureau 2005.

Figure 1.2. Investment in Urban Infrastructure in Shanghai, as Percentage of GDP,

1985–2004

51 In 1950 fewer than 300,000 people lived in Lagos. The city’s population rose by anaverage rate of 6 percent a year in the second half of the 20th century. Every year morethan 600,000 people migrated to Lagos from West Africa. If current trends continue,by 2015 Lagos will have 23 million people, making it the third-largest city in the worldafter Tokyo and Mumbai (“The Megacity” 2006).

52 Informal transactions account for at least 60 percent of economic activity in Lagos(“The Megacity” 2006).

(Feldstein and Horioka 1980; Feldstein 2005). Rapid urban developmentthat can keep pace with large intersectoral transfers of workers demandsrising domestic savings and mechanisms for investing the resources. To becreditworthy enough to gain access to these resources, cities must be wellmanaged financially. Financial deepening can facilitate the process, but ittakes time to build institutions; train people; and create sophisticatedinstruments, risk-assessment skills, rating and monitoring agencies, andregulatory capabilities (Yusuf 2007). Late-starting countries that mustcope with urbanization rates of 3–5 percent a year or more have to relyon banks initially, but they need to move quickly to establish mortgage andbond markets; institutions for both securitizing instruments such as mort-gages and regulating the intermediates involved; and avenues for the sec-ondary trading of securities.53

Local-currency bond markets make it possible to diversify lendingaway from banks and to match long-term assets with debt of equivalentmaturity. They generate yields for a range of maturities and permit thehedging of exposures. Well-functioning bond markets not only lowerborrowing costs, they also impart greater stability to financial markets.54

If urbanization is to avoid the many pitfalls that lie in wait, it needs tobe supported by resource mobilization and fiscal transfers commensuratewith the desired rate of development. When cities are not “bankable”—that is, when resource mobilization is weak or insufficient capital findsits way into urban projects because public and private channels are inad-equate or transaction costs are too high—urbanization cannot bematched by the requisite urban development.The lack of financial depthand sophistication has not initially proven to be the binding constraint.It is the feeble supply of domestic capital for the urban sector (becausethe instruments, skills, and channels have not been created) and ineptmunicipal financial management that are frequently associated withweak economic growth.

Optimizing Urban Development 23

53 By the end of 2005, China’s mortgage market, which started in 1998, had grown toUS$227 billion (Y 1,777 billion), larger than the market in the Republic of Korea(US$200 billon). China’s market represents just 10 percent of GDP, however, whilethe market in the Republic of Korea represents 27 percent of that country’s GDP(“Mortgage Industry” 2006).

54 By abolishing a quota that limited the annual issuance of corporate bonds to Y 100 billion(US$13.2 billion), the Chinese authorities have encouraged listed companies to raisefunds by issuing bonds and to use the funds to pay off higher-interest bank debts. Doingso is particularly appealing to companies engaged in urban real estate, infrastructure,and urban development (“Chinese Companies” 2007).

Providing Urban Social ServicesIn addition to physical infrastructure, people require social services.Vulnerable people need a safety net to avoid sliding into poverty as theyage, become unemployed, or are affected by accidents or loss of property.55

From social as well as private perspectives, the most needed services andthe ones with the highest returns are health care and education. Theseservices build human capital, contribute to individual well-being, providea measure of insurance against poverty, and produce positive externalities(see chapters 2 and 3; Yusuf, Nabeshima, and Ha 2007a, 2007b).

Adequate access to health and education services for the entireurban population should be a central objective of urban development.It is an objective that is often not given the priority it deserves. Theshortfall is most serious during the critical stage when urban popula-tions are exploding. By making inadequate provisions for services, citiesfail to augment a resource that over the longer term is vital for growth,industrial diversification, and the quality of the business climate. Byfocusing on physical infrastructure, governments at both the nationaland subnational levels defer essential and complementary investmentsin human capital, which builds manufacturing and technological capa-bilities. Investing in human capital is also the best insurance againstunemployment and urban crime. Rather than waiting until shortagesbecome glaringly apparent—by which time it is often too late to mobi-lize sufficient resources—cities should view services as intrinsic to theirbig-push urban development strategies.

Health and education services also permit the gradual phasing in of anold-age safety net and unemployment insurance schemes. These schemesare expensive. They need to be backstopped by supporting institutionsand to evolve together with the financial sector.

Limiting Increases in Urban Energy ConsumptionThe lifeblood of the urban economy is energy (see chapter 6). Urbantransport, industry, and households dominate energy consumption inall middle- and high-income countries. Energy use fuels growth andenhances livability, but it is also the principal source of air pollutionand carbon emissions. For the foreseeable future, urban developmentwill remain dependent on ample supplies of energy, with transportdepending on petroleum and households relying mainly on electricity

24 China Urbanizes: Consequences, Strategies, and Policies

55 On the problems posed by shocks for individual households and the options for insuringagainst them, see Baeza and Packard (2006).

and gas. Barring an incident that leads to an interruption of supplies,there is no imminent shortage of petroleum in the near term.

The warning signs of tightening petroleum supplies are everywhere,however. The more-accessible major sources of petroleum are beingrapidly exploited, and even if new reserves are found offshore, productioncosts will be much higher, because the extraction must be from greatdepths. Thus, if economic growth rates worldwide remain healthy, therelative prices of energy could increase significantly.

Given the likely increase in energy prices and the deleterious effect ofthe consumption of fossil fuels on the environment, reducing the energycoefficient of urban development is essential. Especially for countries at arelatively early stage of urbanization, with a great deal of long-lived invest-ment in urban infrastructure and buildings ahead, measures that can cutenergy use and energy losses have a high payback. Among the measureswith the greatest consequences, the design of urban transport is the mostsignificant, because it determines the physical characteristics of the city56—how much it sprawls and encroaches into the surrounding agriculturalland—and the reliance on automobiles for intra- and intercity travel.Appropriate incentives can lead to the efficient utilization of public transitin large cities, which are much more energy efficient than other locales.57

Enforcing strict standards and codes are two additional measures thatcan limit the increase in electricity consumption for air conditioning andappliances, even as urban populations and urban incomes continue theirupward march. Another approach would be to install “smart” meters,which can assess variable charges based on the time of day to encourageenergy conservation, especially at peak load times (“Going Metric” 2006).Encouraging the adoption of “green” technologies and eco-friendly designscan also reduce energy use (see Yusuf and Nabeshima 2006a and thenumerous practical suggestions in Steffen 2006).58

An efficient energy-conserving big-push strategy that is also eco-friendly needs to incorporate the construction industry. The quality

Optimizing Urban Development 25

56 The spatial characteristics of a city, the occupational activities there, and the degree towhich people depend on cars for mobility profoundly influence the incidence of chronicdiseases associated with obesity, such as diabetes and cardiovascular disease, accordingto Frumkin, Frank, and Jackson (2004) and Monda and others (2006).

57 Phang (2000) describes Singapore’s techniques for controlling the ownership and useof cars.

58 In projecting the energy intensity of China’s economy, Wei and others (2006) find thatincomes are the principal source of rising utilization and technological change the mainmitigating factor.

and design of construction and the incorporation of new energy- andmaterials-conserving technologies will determine how much energy issaved, directly and indirectly (Fernandez 2007). The design, construc-tion, and maintenance of infrastructure also influence energy use bythe transport, water, and sanitation sectors. Among all of the industriescontributing to urban development and local multiplier effects, construc-tion is far and away the most important. Enhancing the productivity andtechnological capabilities of this industry should be a critical part of anyurban development strategy.

Dealing with the Scarcity of Water Full recognition of the trend in water scarcity is long overdue. For somecities, a crisis looms not too far in the future. Many others are likely toconfront severe shortages within a few decades. For late urbanizers, thereis a clear opportunity to design the water supply, wastewater collection,and sanitation systems so that they maximize the potential for recyclingwater, supplying water of different grades for different purposes, andminimizing the loss of water from leaking pipes.

As with energy, standards for appliances and sanitary systems can alsoreduce the amount of water and restrict the use of drinking-qualitywater for some purposes only. Standards for the purity of water releasedby commercial and industrial establishments can increase recycling andminimize pollution of water courses and aquifers.

Regulation, physical design, and technology are three strands of anurban water strategy. Pricing is a fourth. It complements the others andis critical to the success of any longer-term strategy to ensure that anurbanizing world will not run short of water.

Managing UrbanizationUrbanization, development, and a rising quality of life are difficult tocombine without sound planning and regulation and the implementationof a host of policies. Small cities and large ones must be well managed forbenefits to be fully realized and diseconomies kept in check. Achievinggood management—and governance—is subject to many factors. Theseinclude the autonomy to conduct policy and raise revenue; the quality oflocal leadership; and planning and administrative capacities. They alsoinclude the availability of policy instruments; the existence of institutionsfor mediating and implementing policies (such as private–public part-nerships and the legal system); and the efficacy of interjurisdictionalcoordination, where this counts. But urban development is arguablymost affected by the design and efficiency of regulations that incentivize

26 China Urbanizes: Consequences, Strategies, and Policies

industry; by land-use and transport policies; by fiscal management andhow it is reflected in credit ratings, as well as the provision of publicservices; and by environmental policies that intersect with and reinforcethose impinging on land use and urban transport (see chapter 8 on thedynamics of decision making at the local level).

From Strategy to Policy

Given the pace of urbanization and the numbers of people involved,the decision makers responsible for guiding China’s urban developmenthave their hands full. Fortunately, they are better placed than theircounterparts in other countries to achieve successful urban develop-ment, for several reasons:

• China is generating the resources to finance an urban big-push andto date has been able to channel these resources into urban indus-try and infrastructure, through the fiscal system, the banks, and newfinancial instruments.

• Because China is a relatively late starter and much building and renewalof urban physical capital lies in the future, there is unparalleled scopefor designing efficient and livable cities.

• Chinese municipalities have the autonomy and the authority to introduceand implement regulations governing land use, the transport system,and the urban environment.

• The hukou system enables municipalities to exercise some controlover the flow of permanent migrants. Industrial growth virtuallythroughout China is such that manufacturing, construction, and serv-ices are largely able to absorb inflows of migrants.

• Although many Chinese cities must cope with a backlog of air andwater pollution, the slums and endemic poverty that have taken rootin other countries are largely absent from Chinese cities thus far.

• As a late starter, China can draw on the experience of other countrieswith respect to urban design, the effects of private vehicle use, andpollution.At the same time, it can exploit advances in a host of technolo-gies that will conserve energy and water and curtail harmful emissions.

This is not to say that urban development will be trouble free. ForChina, however, the enormous intersectoral transfer of people can be aless-daunting process than it was and is for other countries.

The chapters in this volume delineate a number of policies that canpromote rapid urban development within a framework of a national

Optimizing Urban Development 27

strategy that seeks to achieve a balanced increase in incomes. From therange of policies presented, seven stand out.

Increase Human CapitalIn both urban and rural sectors, education and health care policies thatincrease human capital can stimulate growth of agricultural and industrialactivities and reduce the risks of unemployment (Glaeser and Saiz 2003;Berry and Glaeser 2005; Glaeser 2005a; Glaeser and Berry 2006). Greateraccess to these services can raise the incomes of rural households, whichfacilitates migration. Making these services available to urban migrantsdeepens the resource base in cities, promotes equity, and helps combatpoverty. From the standpoint of both growth and welfare, health policiesshould focus on preventive and primary care, and education policiesshould seek to enhance the quality and raise the level of education.

Manage the Flow of MigrationAlthough the hukou system should be dismantled over the longer run, itremains a useful tool for directing and managing the flow of migrants. Itshould be used to achieve two objectives. One is to try to contain thenumber of migrants individual cities absorb as permanent residents. Tothe extent that the hukou system can achieve this, migration can matchthe supply of affordable housing, infrastructure, social services and jobsin cities. A second is to try to direct migrant flows from areas with decliningagricultural potential and water shortages to urban areas with bettergrowth possibilities, in order to realize economies of agglomeration andscale.59 This matching needs to be combined with the planning anddesign of megacities in a way that achieves compactness and polycen-tricity. The coastal cities in the Yangtze delta and the south are likely tocontinue to attract migrants. These cities may need to extend urbanhukou privileges to some of these migrants and to invest in infrastructureand services to accommodate them.

Deepen Financial MarketsTo finance urban development, China needs financial markets that allocateresources more efficiently. It also needs a wider range of instruments, inorder to meet the needs of different kinds of borrowers and offer the spreadof maturities required by investors. The reliance on banks for financing and

28 China Urbanizes: Consequences, Strategies, and Policies

59 China’s WTO membership is likely to negatively affect wheat- and corn-growing farmersin the northeast, as noted in chapter 2.

the use of urban development investment corporations (UDICs) (createdby municipal authorities as semiautonomous vehicles that borrow fromthe banks) have advantages but also some risks (Su and Zhao 2007).Municipalities can limit their fiscal commitments to urban developmentand instead tap the banks. But the UDIC–banking nexus increases banks’exposure on long-lived investments through organizational channels thatmight create problems in the future.60 Changing the legal, tax, and account-ing rules to permit the emergence of secondary bond and mortgage marketswould be a step forward. Other measures include adoption of rules thatencourage securitization of mortgages; strengthening of mortgage insur-ance; issuance of general financial bonds; and trading, including forwardtransactions, in the interbank bond market. Financial innovations and insti-tutional developments, including refinancing arrangements and loanguarantees, would also facilitate urbanization.

Improve Cities’ Fiscal EffortsMany Chinese cities have been balancing their books by selling or leasingland or charging off-budget fees. The income from such transactionsaccounts for almost 25 percent of municipal revenues. Sooner or later,this process will come to an end. Urban governments need to create adurable fiscal system that can meet future and current capital needs,taking account of the anticipated growth rates of urban economies andpopulations. Doing so entails revenue-sharing and transfer agreementswith the central authorities and an elastic base for local revenues witha few robust tax instruments to satisfy local needs. It also calls for a firmagreement with the central government on expenditure assignmentsthat are equitable and sustainable given the anticipated flows of revenuefrom all sources.61

Contain Energy CostsEnsuring that urban development is not constrained by rising energycosts will depend on four factors. Arguably the most important are pricingpolicies (especially of power) that accurately communicate informationon relative scarcities and induce efficient utilization (see chapter 6).

Optimizing Urban Development 29

60 Chinese banks have a history of loan portfolio problems. Although reforms havereduced the scale of the problem, weak risk-assessment skills and poor governanceremain sources of vulnerability (Podpiera 2006).

61 See, for instance, World Bank (2003a, 2003b); Dabla-Norris (2005); and Su and Zhao(2007).

A second factor is how China proceeds with its motorization strategywithin the context of urbanization.62 One part of this strategy relates to theincentives for the automobile industry to produce fuel-efficient cars and toredouble efforts to innovate. Another part has to do with demand for cars,which will be a function of policies on taxes on cars, gasoline, licensing andregistration fees, and road user charges; car financing; research and develop-ment; urban land use; investment in road building; and public transport.Currently, many Chinese cities are subject to so-called “ribbon develop-ment” alongside major highways and floor area ratios are still quite low.Thistype of development saves developers and local governments from invest-ing in secondary feeder and access roads, but it leads to much greater sprawland raises the energy- and infrastructure-related costs of urbanization.

A third factor is advances in home-grown technology and technologytransfer from abroad. The degree to which policy accelerates theseadvances will influence the efficiency with which fossil fuels are utilizedas well as the diversification into renewable sources.63

A fourth factor is the raft of regulatory policies, including environ-mental and land use policies and policies defining building codes andstandards for consumer appliances. In conjunction with the perfecting ofmechanisms for enforcement, these policies will play a significant part indetermining energy demand.

Manage Water ResourcesPricing, regulatory, technology, sewage treatment, and wastewater recy-cling policies will also be decisive with respect to the utilization ofwater.64 Given the distribution of water, its low per capita availability,

30 China Urbanizes: Consequences, Strategies, and Policies

62 Between 1991 and 2005, the number of cars per 1,000 people in China rose from lessthan 2 to 10.

63 Considerable progress is being made in developing clean-coal technologies that Chinacould tap. These include supercritical boilers (which heat steam to 600ºC beyond the critical boiling point and therefore need 17 percent less coal than conventional coal-firedplants); integrated gasification combined cycles (which convert coal to gas); and techniquesfor capturing carbon (“Big Effort” 2006).An earmarked tax on electricity consumption nowfinances the development of renewable energy. The government is also investing innuclear power and ethanol. China has 9 operating nuclear reactors supplying electricity,2 more under construction, and another 20 planned (Flavin and Gardner 2006; Hunt andSawin 2006; “Saying ‘No’” 2006). Currently, among renewable sources of energy, onlyonshore wind turbines in certain locations are generally profitable. Solar energy, offshorewind turbines, and tidal power are still relatively costly (“On the Verge” 2006).

64 A little more than half of urban sewage is currently treated before being discharged intowater bodies, which party accounts for the low quality of the lake and river water(“Saying ‘No’” 2006).

and its deteriorating quality, the problem of water availability needs tobe tackled immediately. As noted in this chapter and in chapter 7, thegeography of future urbanization and the degree to which it is concen-trated in the relatively water-abundant parts of the country could playan important role in providing a solution to the problem.

A massive transfer of water could be a costly solution, on a number ofcounts. The distribution of water resources, the implications of impendingclimate change on future supplies, and policies affecting the productionof grain in the northeast should all be factored into the urbanizationstrategy. Ongoing climate change makes it desirable to take account ofwater availability across the country when making long-term plans forurban development.

Reduce PollutionPolicies on the conservation and consumption of water and energyresources will affect environmental pollution, which is a drag on GDPgrowth and degrades the quality of life. There are various estimates ofthe costs of indoor and outdoor pollution to the economy, ranging from3 to 6 percent of GDP. A redoubling of efforts by the government andindustry would appear to be desirable (“Green GDP” 2007; Ho andNielsen 2007; “China’s Green Accounting” 2007).

Concluding Observations

This chapter raises but a few of the issues and policies that will guideurbanization and urban development in China, but they are likely to beamong the most crucial. More than half of China’s population lives inrural areas; decades of urbanization lie ahead. Almost three-quarters ofBrazil’s population live in urban areas, and an even higher percentage ofthe population of the United States is urban (World Bank 2006).Possibly by the middle of this century, China could be approaching theselevels of urbanization. Between now and then, decisions will be madethat will affect the geographical distribution of the population and thebuilding of the urban, transport, and water supply infrastructure tohouse and support urban inhabitants. Enormous amounts of capital willbe committed if the big-push approach is continued.

With so much at stake, it is essential that decisions be closer to the opti-mal in the long-term sense. Markets alone cannot achieve the outcomesdesired, but efficient financial, energy, and water markets, for example,can help achieve good outcomes. Taking the institutional and policy

Optimizing Urban Development 31

steps to make markets work more efficiently should be a priority.Successful markets will need to be backed by good government planningand policies based on careful analysis, using the best information avail-able. Factoring in the systems of urban design and construction, as wellas innovation, could increase the likelihood that urban developmentachieves multiple objectives. The approaches to urbanization, urban/design and urban innovation systems will play major roles, especiallywhen it comes to making longer-term decisions about the distribution ofpeople geographically and across cities of different sizes.

The Chinese authorities have some instruments with which to influ-ence both of these outcomes, namely, the hukou system and the govern-ment’s role in allocating investment. These instruments are two among anumber of factors that determine the flows of people and capital. Thedesign of policies and the application of these instruments depend oncomplex negotiations by several levels of government as well as otherplayers, such as the banks. Such interactions among the various stakehold-ers are useful, because many points of view and a broad range of informa-tion can be factored in; the process also diminishes the risk of egregiousmistakes. By the same token, the negotiated approach increases thedegree of policy slippage and delays in implementation. Policy slippagesand the likelihood of delays are unavoidable, but urbanization will notwait. China must move forward with policies to contain the costs of rapidgrowth, narrow the gap between rural and urban incomes, raise the qual-ity life, and minimize negative externalities. To do so, urbanization andurban development should remain at the top of policy makers’ agendasuntil the pressing issues are resolved.

The potential gains to China from urbanization are substantial. So,too, are the costs. Striking the right balance between the two will bethe greatest challenge for Chinese policy makers over the next quartercentury and more.

References

“Anti-Graft Campaign Intensifies.” 2006. Business Asia, October 30.

Baeza, Cristian C., and Truman G. Packard. 2006. Beyond Survival. Stanford, CA:Stanford University Press.

Bairoch, Paul. 1991. Cities and Economic Development: from the Dawn of History tothe Present. Chicago: University of Chicago Press.

32 China Urbanizes: Consequences, Strategies, and Policies

Baochang, Gu, Wang Feng, Guo Zhigang, and Zhang Erli. 2007. “China’s Localand National Fertility Policies at the End of the Twentieth Century.”Population and Development Review 33 (1): 129–47.

“Beneath Booming Cities, China’s Future Is Drying Up.” 2007. New York Times,September 28.

Bernstein,Thomas P. 1977. Up to the Mountains and down to the Villages: The Transferof Youth from Urban to Rural China. New Haven, CT: Yale University Press.

Berry, Christopher, and Edward Glaeser. 2005. “The Divergence of HumanCapital Levels Across Cities.” NBER Working Paper 11617, National Bureauof Economic Research, Cambridge, MA.

Bettencourt, Luis, M. A., Jose Lobo, and Deborah Strumsky. 2007. “Interventionin the City: Increasing Returns to Patenting as a Scaling Function ofMetropolitan Size.” Research Policy 36: 107–120.

“Big Effort to Scrub Up a Dirty Image.” 2006. Financial Times, October 20.

Bloom, David E., and Jeffrey G. Williamson. 1997. “Demographic Transitions andEconomic Miracles in Emerging Asia.” NBER Working Paper 6268, NationalBureau of Economic Research, Cambridge, MA.

Bosworth, Barry P., and Susan M. Collins. 2007. “Accounting for Growth:Comparing China and India.” NBER Working Paper 12943, National Bureauof Economic Research, Cambridge, MA.

Carlino, Gerald A., Satyajit Chatterjee, and Robert M. Hunt. 2007. “UrbanDensity and the Rate of Innovation.” Journal of Urban Economics 61:389–419.

“Caught between Right and Left, Town and Country.” 2007. Economist, March 8.

CBO (Congressional Budget Office). 2006. China’s Growing Demand for Oil andIts Impact on U.S. Petroleum Markets. Washington, DC.

Chen, B., and G. Q. Chen. 2007a. “Resource Analysis of the Chinese Society1980–2002 Based on Energy. Part 2: Renewable Energy Sources and Forest.”Energy Policy 35 (4): 2051–64.

Chen, G. Q., and B. Chen. 2007b. “Resource Analysis of the Chinese Society1980–2002 Based on Energy. Part 1: Fossil Fuels and Energy Minerals.” EnergyPolicy 35 (4): 2038–50.

Cheng, Margaret Harris. 2006. “China’s Cities Get a Little Healthier.” Lancet368 (9550): 1859–60.

“China: Does the Countryside Hold the Seeds of Crisis?” 2007. Oxford Analytica,Global Strategic Analysis, May 30.

“China’s Crackdown on Corruption Still Largely Secret.” 2006. Washington Post,December 31.

Optimizing Urban Development 33

“China’s Green Accounting System on Shaky Ground.” 2007. Nature 448:518–19. August 2.

“China’s Income Gap Grows despite Pledges.” 2006. Financial Times, December 27.

“Chinese Companies Rush to Issue Bonds under Relaxed Regulations.” 2007.Financial Times, August 18.

“Cities Should Plan Now for Effects of Global Warming on Infrastructure.”February 25, 2005. http://www.newsdesk.umd.edu.

Dabla-Norris, Era. 2005. “Issues in Intergovernmental Fiscal Relations inChina.” IMF Working Paper WP/05/30, International Monetary Fund,Washington, DC.

“Delta Dreams.” 2006. Business China, April 24.

Demurger, Sylvie, Jeffrey D. Sachs, Wing Thye Woo, Shuming Bao, Gene Chang,and Andrew Mellinger. 2002. “Geography, Economic Policy, and RegionalDevelopment in China.” Asian Economic Papers 1 (1): 146–97.

Dunaway, Steven, and Vivek Arora. 2007. “Pension Reform in China:The Need fora New Approach.” IMF Working Paper WP/07/109. IMF, Washington, DC.

Ebrey, Patricia. 1996. China. Cambridge, U.K.: Cambridge University Press.

Ekboir, J. ed. 2002. CIMMYT 2000–2001 World Wheat Overview and Outlook:Developing No-Till Packages for Small-Scale Farmers. International Maize andWheat Improvement Center (CIMMYT), Mexico City.

Evans, Richard, and Sinclair K. Knight. 2007. “Groundwater Resource ManagementChallenges in North China.” National Academy of Sciences. Arthur M. SacklerColloquia, Washington, DC.

Falkner, Robert. 2006. “International Sources of Environmental Policy Change inChina: The Case of Genetically Modified Food.” Pacific Review 19 (4): 473–94.

Fan, C. Cindy. 1997. “Uneven Development and Beyond: Regional Developmenttheory in Post-Mao China.” International Journal of Urban and RegionalResearch 21 (4): 620–39.

–––––. 2005. “Modeling Interprovincial Migration in China.” Eurasian Geographyand Economics 46 (3): 165–84.

Fang, Shan. 1990.“Urbanization in Mainland China.” Issues & Studies 26 (2): 118–33.

FAO (Food and Agricultural Organization). 2003. Selected Indicators of Food andAgriculture Development in Asia-Pacific Region: 1992–2002. Regional Officefor Asia and the Pacific, Bangkok.

Feldstein, Martin. 2005.“Monetary Policy in a Changing International Environment:The Role of Global Capital Flows.” NBER Working Paper 11856, NationalBureau of Economic Research, Cambridge, MA.

Feldstein, Martin, and Charles Horioka. 1980. “Domestic Saving and InternationalCapital Flows.” Economic Journal 90 (358): 314–29.

34 China Urbanizes: Consequences, Strategies, and Policies

Fernandez, John E. 2007. “Materials for Aesthetic, Energy-Efficient, and Self-Diagnostic Buildings.” Science 315 (5820): 1807–10.

Flavin, Christopher, and Gary Gardner. 2006. “China, India, and the New WorldOrder.” In State of the World 2006, ed. Worldwatch Institute. New York:W. W. Norton.

“Free Flow: Big Challenge for Beijing: Moving Coal.” 2005. International HeraldTribune, August 18.

Friedman, John. 2005. China’s Urban Transition. Minneapolis: University ofMinnesota Press.

Frumkin, Howard, Lawrence Frank, and Richard Joseph Jackson, 2004. UrbanSprawl and Public Health: Designing, Planning, and Building for HealthyCommunities. Washington, DC: Island Press.

Gan, Li, and Guan Gong. 2007. “Estimating Interdependence between Healthand Education in a Dynamic Model.” NBER Working Paper 12830, NationalBureau of Economic Research, Cambridge, MA.

Gardner, John. 1971. “Educated Youth and Urban-Rural Inequalities, 1958–66.” InThe City in Communist China, ed. John W. Lewis. Stanford, CA: StanfordUniversity Press.

Glaeser, Edward L. 2005a. “Reinventing Boston: 1640–2003.” Journal of EconomicGeography 5 (2): 119–153.

———. 2005b. “Urban Colossus: Why Is New York America’s Largest City?”NBER Working Paper 11398, National Bureau of Economic Research,Cambridge, MA.

Glaeser, Edward L., and Christopher Berry. 2006. “Why Are Smarter PlacesGetting Smarter?” Policy Briefs PB-2006-2, Department of Economics,Harvard University, Cambridge, MA.

Glaeser, Edward L., and Albert Saiz. 2003. “The Rise of the Skilled City.”NBER Working Paper 10191, National Bureau of Economic Research,Cambridge, MA.

“Going Metric” 2006. Economist, December 13.

“Green GDP Will Be Expanded to Entire Mainland.” 2007. China Daily,January 18.

Gustafsson, Bjorn, Li Shi, and Wei Zhong. 2006. “The Distribution of Wealth inUrban China and in China as a Whole in 1995.” Review of Income and Wealth52 (2): 173–88.

Hausmann, Ricardo, and Bailey Klinger. 2006. “Structural Transformation andPatterns of Comparative Advantage in the Product Space.” Center forInternational Development, Harvard University, Cambridge, MA.

“The Health of a Nation.” 2007. Economist, February 26.

Optimizing Urban Development 35

He, Jianwu, and Louis Kuijs. 2007. “Rebalancing China’s Economy—Modeling aPolicy Package.” Beijing PRC. World Bank China Research Paper No. 7.

Hesketh, Therese, Li Lu, and Zhu Wei Xing. 2005. “The Effect of China’sOne-Child Family Policy after 25 Years.” New England Journal of Medicine353 (11): 1171–76.

Ho, Mun S., and Chris Nielsen. 2007. Clearing the Air: the Health and EconomicDamages of Air Pollution in China. Cambridge, Mass.: MIT Press.

Ho, Ping-ti. 1975. The Cradle of the East. Hong Kong (China): Chinese Universityof Hong Kong Press.

Huang, Jikun, Ruifa Hu, Scott Rozelle, and Carl Pray. 2007. “China: EmergingPublic Sector Model for GM Crop Development.” In Gene Revolution: GMCrops and Unequal Development, ed. Sakiko Fukuda-Parr. London: Earthscan,130–55.

Hummels, David, and Peter J. Klenow. 2005. “The Variety and Quality of aNation’s Exports.” American Economic Review 95 (3): 704–23.

Hunt, Suzanne C., and Janet L. Sawin, with Peter Stair. 2006. “CultivatingRenewable Alternative to Oil.” In State of the World 2006, ed. WorldwatchInstitute. New York: W. W. Norton.

“In Face of Rural Unrest, China Rolls Out Reforms” 2006. Washington Post,January 28.

Japan Statistics Bureau. 2005. Japan Statistical Yearbook 2005. Ministry of InternalAffairs and Communications, Tokyo.

Jiang, Leiwen. 2006. “Living Conditions of the Floating Population in UrbanChina.” Housing Studies 21 (5): 719–44.

Jin, Songqing, Jikun Huang, Ruifa Hu, and Scott Rozelle. 2002. “The Creation andSpread of Technology and Total Factor Productivity in China’s Agriculture.”American Journal of Agricultural Economics 84 (4): 916–30.

Knight, John, and Lina Song. 2003. “Chinese Peasant Choices: Migration, RuralIndustry or Farming.” Oxford Development Studies 31 (2): 123–47.

Kojima, Reeitsu. 1987. “Urbanization and Urban Problems in China.” IDEOccasional Paper 22, Institute for Developing Economies, Tokyo.

Kwok, Yin-Wang. 1981. “Trends of Urban Planning and Development in China.”In Urban Development in Modern China, ed. Laurence J. C. Ma and Edward W.Hanten. Boulder, CO: Westview Press.

“Lanzhou to Walk Off Pollution.” 2007. China Daily, January 17.

Li, Peter Ping. 2006. “The Puzzle of China’s Township-Village Enterprises: TheParadox of Local Corporatism in a Dual-Track Economic Transition.”Management and Organization Review 1 (2): 197–224.

Liu, Haiyan. 1993. “The Past and Future of Urban Histories Examining China’sModern Period (1840–1949).” Social Sciences in China 14 (Winter): 115–27.

36 China Urbanizes: Consequences, Strategies, and Policies

Liu, Yunhua, and Xiaobing Wang. 2005. “Technological Progress and ChineseAgricultural Growth in the 1990s.” China Economic Review 16 (4): 419–40.

“The Megacity.” 2006. The New Yorker, November 13.

Mohan, Rakesh. 2006. “Asia’s Urban Century: Emerging Trends.” Paper presentedat the conference “Land Policies for Urban Development,” Cambridge, MA,June 5.

Monda, Keri L., Penny Gordon-Larsen, June Stevens, and Barry M. Popkin. 2006.“China’s Transition: The Effect of Rapid Urbanization on Adult OccupationalPhysical Activity.” Social Science & Medicine 64 (4): 858–70.

“Mortgage Industry Untapped, Says BIS.” 2006. Shanghai Daily, December 11.

Mote, Frederick W. 1999. Imperial China: 900–1800. Cambridge, MA: HarvardUniversity Press.

Mu, Huaipeng. 2005. “China’s Bond Market: Innovation and Development.”Paper presented at the Second Annual Asia Pacific Bond Congress, Hong Kong(China), June 16.

Murphy, Rachel 2002. How Migrant Labor Is Changing Rural China. Cambridge,U.K.: Cambridge University Press.

NBS (National Bureau of Statistics). 2002. China Statistical Yearbook 2001.Beijing: China Statistics Press.

———. 2006. China Statistical Yearbook 2001. Beijing: China Statistics Press.

Naughton, Barry. 1988. “The Third Front: Defense Industrialization in the ChineseInterior.” China Quarterly 115: 351–86.

NDRC (National Development and Reform Commission). 2007. “11th Five-YearPlan on Energy Development.” Beijing.

“The New New York.” 2006. Economist, December 13.

O’Brien, Kevin J., and Lianjiang Li. 2006. Rightful Resistance in Rural China.Cambridge: Cambridge University Press.

Oi, J. C. 1992. “Fiscal Reform and the Economic Foundations of Local StateCorporatism in China.” World Politics 45 (1): 99–126.

“On the Verge of Viability.” 2006. Financial Times, October 20.

Ouyang, Tingping, Zhaoyu Zhu, and Yaoqiu Kuang. 2006. “Assessing Impactof Urbanization on River Water Quality in the Pearl River DeltaEconomic Zone, China.” Environmental Monitoring and Assessment 120(1–3): 313–25.

Pan, Jiahua. 2002. “Rural Energy Patterns in China:A Preliminary Assessment fromAvailable Data Sources.” Program on Energy and Sustainable DevelopmentWorking Paper 12, Stanford University, Stanford, CA.

Pannell, Clifton W. 1992. “The Role of Great Cities in China.” In Urbanizing China,ed. Gregory Eliyu Guldin. New York: Greenwood Press.

Optimizing Urban Development 37

–––––. 2003. “China’s Demographic and Urban Trends for the 21st Century.”Eurasian Geography and Economics 44 (7): 479–96.

Park, Albert, Sangui Wang, and Guobao Wu. 2002. “Regional Poverty Targeting inChina.” Journal of Public Economics 86 (1): 123–53.

Phang, Sock-Yong. 2000. “How Singapore Regulates Urban Transportationand Land Use.” In Local Dynamics in An Era of Globalization, ed. ShahidYusuf, Weiping Wu, and Simon Evenett, 159–63. New York: OxfordUniversity Press.

Pielou, E. C. 1998. Fresh Water. Chicago: University of Chicago Press.

Pingali, P. L., ed. 2001. CIMMYT 1999–2000 World Maize Facts and Trends:Meeting World Maize Needs: Technological Opportunities and Priorities for thePublic Sector. International Maize and Wheat Improvement Center (CIMMYT),Mexico City.

Podpiera, Richard. 2006. “Progress in China’s Banking Sector Reform: Has BankBehavior Changed?” IMF Working Paper WP/06/71, International MonetaryFund, Washington, DC.

Polo, Marco. 1958. The Travels of Marco Polo. Translated by Ronald E. Latham.New York: Penguin Classics.

Qian, Yingyi. 1999. “The Institutional Foundations of China’s Market Transition.”Paper presented at the Annual World Bank Conference on DevelopmentEconomics, Washington, DC, April 29.

Ravallion, Martin, Shaohua Chen, and Prem Sangraula. 2007. “The Urbanizationof Global Poverty.” World Bank Background Paper 2008, Washington, DC.

Rosen, Daniel H., and Trevor Houser. 2007. “China Energy: A Guide for thePerplexed.” Peterson Institute for International Economics, Washington, DC.

Rosenthal, Stuart S., and William C. Strange. 2004. “Evidence on the Nature andSources of Agglomeration Economics.” In Handbook of Regional and UrbanEconomics, ed. J. Vernon Henderson and Jacques-François Thisse, 2119–71.North Holland.

Roumasset, James, Hua Wang, and Kimberly Burnett. 2006. “EnvironmentalResources and Economic Growth.” World Bank, Washington, DC.

Saich, Tony. 2006. “Hu’s in Charge.” Asian Survey 46 (1): 37–48.

“Saying ‘No’ to Growing Filthy Rich.” 2006. Business China, November 20.

Schott, Peter K. 2006. “The Relative Sophistication of Chinese Exports.”NBER Working Paper 12173, National Bureau of Economic Research,Cambridge, MA.

Shalizi, Zmarak. 2007. “Energy and Emissions: Local and Global Effects of theGiants’ Rise.” In Dancing with Giants, ed. L. Alan Winters and Shahid Yusuf.Washington, DC: World Bank.

38 China Urbanizes: Consequences, Strategies, and Policies

Shanghai Municipal Statistical Bureau. 2005. Shanghai Statistical Yearbook 2005.Beijing.

“Shanghaied.” 2006. Economist, September 30.

Sherbinin, Alex de, Andrew Schiller, and Alex Pulsipher. 2007. “The Vulnerabilityof Global Cities to Climate Hazards.” Environment & Urbanization 19 (1):39–64.

Shiller, Robert J. 2006. “Thrifty China, Spendthrift America.” Project Syndicate.http://www.project-syndicate.org/commentary/shiller40.

Sicular, Terry, Yue Ximing, Bjorn Gustafsson, and Li Shi. 2007. “The Urban–RuralIncome Gap and Inequality in China.” Review of Income and Wealth 53 (1):93–126.

Sinn, Hans-Werner, and Frank Westermann. 2001. “Two Mezzogiornos.” NBERWorking Paper 8125, National Bureau of Economic Research, Cambridge, MA.

Steffen, Alex, ed. 2006. Worldchanging: A User’s Guide for the 21st Century. NewYork: Harry N. Abrams, Inc.

Su, Ming, and Quanhou Zhao. 2007. “China: Fiscal Framework and UrbanInfrastructure Finance.” In Financing Cities: Fiscal Responsibility and UrbanInfrastructure in Brazil, China, India, Poland and South Africa, ed. George E.Peterson and Patricia Clarke Annez. London: Sage Publications.

Tsui, Kai-yuen. 2007. “Forces Shaping China’s Interprovincial Inequality.” Reviewof Income and Wealth 53 (1): 60–92.

“The Way We Were in 2006.” 2006. The Week, December 29.

Wei, Yi-Ming, Qiao-Mei Linag, Ying Fan, Norio Okada, and Hsien-Tang Tsai.2006. “A Scenario Analysis of Energy Requirements and Energy Intensity forChina’s Rapidly Developing Society in the Year 2020.” TechnologicalForecasting & Social Change 73 (4): 405–21.

“What Shortage?” 2006. Business Asia, November 13.

World Bank. 2003a. China: Promoting Growth with Equity. Report 24169–CHA,Washington, DC.

———. 2003b. “Public Finance Reform and Macroeconomic Management.”World Bank Policy Note, Washington, DC.

———. 2006. World Development Indicators. Washington, DC: World Bank.

Wu, Friedrich. 2006. “What Could Brake China’s Rapid Ascent in the WorldEconomy?” World Economics 7 (3): 63–87.

Wu, Fulong. 2004. “Urban Poverty and Marginalization under Market Transition:The Case of Chinese Cities.” International Journal of Urban and RegionalResearch 28 (2): 401–23.

Wu, Jinglian. 2005. Understanding and Interpreting Chinese Economic Reform.Mason, OH: South-Western College Publishing.

Optimizing Urban Development 39

Wu, Weiping. 2002. “Temporary Migrants in Shanghai: Housing and SettlementPatterns.” In The New Chinese City: Globalization and Market Reform, ed. John R.Logan, 212–26. Oxford: Blackwell.

Yixing, Zhou, and Laurence J. C. Ma. 2003. “China’s Urbanization Levels:Reconstructing a Baseline from the Fifth Population Census.” China Quarterly(173): 176–96.

Yusuf, Shahid. 2007. “About Urban Mega Regions: Knowns and Unknowns.”World Bank Policy Research Working Paper 4252, Washington, DC.

Yusuf, Shahid, M. Anjum Altaf, Barry Eichengreen, Sudarshan Gooptu, KaoruNabeshima, Charles Kenny, Dwight H. Perkins, and Marc Shotten. 2003.Innovative East Asia: The Future of Growth. New York: Oxford University Press.

Yusuf, Shahid, and Kaoru Nabeshima. 2006a. China’s Development Priorities.Washington, DC: World Bank.

———. 2006b. Post Industrial East Asian Cities. Stanford, CA: Stanford UniversityPress.

Yusuf, Shahid, Kaoru Nabeshima, and Wei Ha. 2007a.“Income and Health in Cities:The Messages from Stylized Facts.” Journal of Urban Health 84 (S1): 35–41.

———. 2007b. “What Makes Cities Healthy?” World Bank Policy ResearchWorking Paper 4107, Washington, DC.

Zhang, Li 2001. Strangers in the City: Reconfigurations of Space, Power, and SocialNetworks within China’s Floating Population. Stanford, CA: StanfordUniversity Press.

Zhao, Gang. 1994. “Looking at China’s Urban History from a MacroscopicPerspective.” Social Sciences in China 15 (3): 171–79.

Zhou, Yixing, and Laurence J. C. Ma. 2003. “China’s Urbanization Levels:Reconstructing a Baseline from the Fifth Population Census.” China Quarterly173: 176–96.

Zhu, Yu. 2000. “In Situ Urbanization in Rural China: Case Studies from FujianProvince.” Development and Change 31 (2): 413–34.

40 China Urbanizes: Consequences, Strategies, and Policies

Having negotiated a tortuous path through sweeping reform and rapideconomic development for 25 years, Chinese society finds itself copingwith the relatively new phenomena of rapid urbanization and soaringdemands for energy. As with so many other challenges in recent Chinesehistory, these new challenges have emerged as by-products of earliersocietal achievements—the inevitable offshoots of phenomenal economicgrowth—accelerating at previously unimagined speed.This chapter exploresthe nature of an urbanizing China’s rapidly accelerating demand forenergy, the complexities involved in meeting that demand, and the evenbroader policy and institutional challenges surrounding long-termresource and environmental sustainability.

The Nature of the Challenge

In 1980, China’s urbanization rate hovered just below 20 percent, a ratelower than Pakistan (28.1 percent), India (23.1 percent), and Indonesia(22.1 percent) (see table 1.1). By 2005 China’s levels had surged to42.9 percent, outstripping Pakistan (34.9 percent) and India (28.7 percent)and almost reaching the level of Indonesia (48.1 percent).

During this period, Chinese energy consumption soared. In 1973China consumed 7.9 percent of the world’s energy; by 2005 the figurehad risen to 14.2 percent, making China the world’s second-largestconsumer, trailing only the United States (IEA 2007a). Since 1993 China

C H A P T E R 6

Energy Policy

125

Edward S. Steinfeld

has been a net importer of oil, and it continues to be the world’s largestconsumer (and producer) of coal, which accounted for more than 76.4 percent of China’s primary energy supply in 2004 (IEA 2007b).

Certainly since 2002, China’s electric power sector has been growingat a torrid pace. Total generating capacity increased by nearly a thirdbetween 2003 and 2006 (MIT 2007). In 2005, the system added about70 gigawatts (GW) of generating capacity, an amount on par with the scaleof the entire British power grid (MIT 2007). Many observers doubtedthat China could increase its capacity by another 70 GW. Nonetheless,the following year witnessed an additional 102 GW of capacity expansion(McGregor 2007). Concomitantly, and far more quickly than previouslypredicted, in 2006 China became the world’s largest global emitter of car-bon dioxide (Landsberg 2007).

In theory, the connection between urbanization and rising energyconsumption appears obvious. As people shift from rural lifestyles tohigh-density, multistory urban dwellings, demand for energy-intensiveclimate control and extensive lighting should surge. So, too, shoulddemand for energy-intensive appliances, automobiles, and the extensivelong-distance transportation networks needed to channel goods into urbanmarkets. Urban lifestyles presumably also generate demands for entirelynew, and decidedly energy-intensive, production systems, such as therefrigerated food supply chain, from upstream industrial-scale preparationto supermarket retailing.

China’s soaring energy consumption have not yet reflected these newdrivers of energy demand (Rosen and Houser 2007; Zheng 2007); thatis, the long-term consumption ramifications of urbanized lifestyles havenot yet begun to kick in. Chinese per capita energy consumptionremains well below levels found in advanced industrial societies: in2005, annual per capita energy consumption stood at 1.56 tons of oilequivalent (toe) (Zheng 2007), a fraction of levels in Europe (3.46 toe),Japan (4.12 toe), and the United States (7.88 toe) (Rosen and Houser2007). Within China, however, the per capita energy consumption ofurban citizens is 3.5 times that of rural citizens. Given the country’saccelerating pace of urbanization, it would be foolish to assume thatover the long run, residential energy consumption in China will not rise,in all likelihood substantially. There is every reason to believe that theChina of tomorrow will exhibit an energy-demand pattern similar tothat of urbanized societies throughout the world.

For the time being, industrial consumption drives Chinese energydemand—to a greater extent than virtually anywhere else in the world.In 2005, the industrial sector accounted for 71 percent of China’s energy

126 China Urbanizes: Consequences, Strategies, and Policies

demand, with the remainder split between transport (10 percent) andresidential, commercial, and agriculture use (19 percent). In India theindustrial sector accounts for 49 percent, transport accounts for 21 percent,and residential, commercial, and agriculture use accounts for 30 percent.Far at the other end of the spectrum, in the United States industryaccounts for only 25 percent of energy demand, while transport accountsfor 33 percent and residential, commercial, and agricultural uses for43 percent (Rosen and Houser 2007).

The fact that the effects of urbanization on energy consumption have yetto be felt has sweeping implications.At the very least, it means that China’saccelerating demand for energy—with all the pressures it is exerting onglobal resources and the global environmental commons—is unlikely to beanywhere close to peaking. Chinese energy demand rose steeply throughthe 1990s and will likely continue to do so in the coming decades, even ifreductions in energy intensity are achieved (tables 6.1–6.3).

Although residential energy consumption can be expected to rise, littlereason exists to believe that industrial consumption will fall substantially,either in the aggregate or as a portion of total consumption.The expansionof energy-intensive heavy industry in China, a phenomenon that beganin the 1990s, is related to the build-out of urban infrastructure on anational scale. Chinese firms are churning out the steel, aluminum, con-crete, and other basic building materials going into the nation’s new roads,mass-transit systems, and vast urban residential and commercial realestate development projects. Conceivably, this phenomenon could peterout over time once basic infrastructure is established, but the time framefor this will likely extend across decades.

Moreover, beyond just supplying domestic infrastructural needs,Chinese industry is increasingly producing for global markets. Indeed,urbanization has moved hand in hand with the development of tech-nology and energy-intensive manufacturing for the global market. It isnot just that the world’s electronics, automotive parts, and consumergoods are being assembled in China. The most energy-intensive compo-nents for these products—everything from steel and aluminum tosemiconductors—are being produced in China. The globalized supplychain now permits the most energy-intensive (and often lowest valueadded) production aspects of global products to be delinked from theless energy-intensive but often highest-value production aspects (design,research and development, marketing, and so forth) of those products.It is precisely these energy-intensive but often low-value productionactivities that are now concentrating so heavily in China. These activitiesmake China a critical link in global supply chains but also a repository

Energy Policy 127

128 Table 6.1. Energy Production and Consumption, 1991–2005

Item 1991 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Primary energy production

Raw coal (tens

of thousands of

metric tons) 108,741 136,073 139,670 137,282 125,000 104,500 99,800 116,078 138,000 166,700 199,232 220,473

Crude oil

(millions of

barrels) 14,099 15,004 15,733 16,074 16,100 16,000 16,300 16,396 16,700 16,960 17,587 18,135

Natural gas

(108 cubic meters) 161 179 201 227 233 252 272 303 327 350 415 493

Hydro

(108 kilowatts) 1,251 1,906 1,880 1,960 2,080 2,038 2,224 2,774 2,880 2,837 3,535 3,970

Nuclear

(108 kilowatts) n.a. 128 143 144 141 149 167 175 251 433 505 531

Total energy consumption

Raw coal

(tens of thousands

of metric tons) 110,432 137,677 144,734 139,248 129,492 126,365 124,537 126,211 136,605 163,732 193,596 216,723

Crude oil

(millions of barrels) 12,384 16,065 17,436 19,692 19,818 21,073 22,439 22,838 24,780 27,126 31,700 32,535

Natural gas

(108 cubic meters) 159 177 185 195 203 215 245 274 292 339 397 479

Hydro

(108 kilowatts) 1,251 1,906 1,880 1,960 2,080 2,038 2,224 2,774 2,880 2,837 3,535 3,970

Nuclear

(108 kilowatts) n.a. 128 143 144 141 149 167 175 251 433 505 531

Source: NBS various years.

n.a. Not available.

129

Table 6.2. Projected Demand for Primary Energy and Oil in Selected Countries in 2025

(millions of barrels of oil equivalent a day)

Primary energy Oil

Country 2001 2025 2001 2025

United States 96.3 132.4 19.6 27.3

China 40.9 109.2 4.9 14.2

Japan 21.9 24.7 5.4 5.3

India 13.8 29.3 2.2 4.9

World 403.9 644.6 78 119.2

Source: U.S. Department of Energy 2007.

for the indirect energy demands of global consumers. To the extent thatglobal demand for consumer products remains robust, so, too, willChina’s demand for energy.

That China’s energy demands for the foreseeable future are linked tothe dual phenomena of urbanization and globalization is important in tworespects. On the urbanization front, it means that nonproduction-relatedenergy consumption will likely increase significantly only in the future,particularly once the growing numbers of urban residents begin consumingat levels comparable to global norms. On the globalization front, itmeans that China’s appetite for energy is in many ways a reflection ofthe global appetite—particularly in advanced industrial markets—forconsumer goods, production of which is increasingly concentrating inChina. In this sense, the problem of Chinese energy demand, both in itsorigins and its potential solutions, must be understood as global innature. In essence, Chinese and global sustainability have become onein the same.

Energy Intensity and per Capita ConsumptionEnergy intensity (consumption per unit of GDP) in China is one of thehighest in the world: in 2002, it was more than 7.0 times that of Japan,3.5 times that of the United States, and 1.7 times that of Indonesia(Sun 2003). Historically, as economies shift from agriculture to industry,energy intensity rises steadily; peaks with the deepening of heavy industry;begins dropping as technological transformation occurs; and thencontinues to descend with the shift into the more service-oriented, lessmanufacturing-intensive activities typical of postindustrial economies.Energy intensity peaked in the United States in 1920 and globally in1955 (Sun 2003).

Energy Policy 129

Table 6.3. Alternative Projections of Growth in Final Energy Demand in China, by Sector

Asia Pacific Energy Research Centre International Energy Agency Tsinghua University

Share of

Annual Share of Projected Annual Share of Projected Annual total

Projected growth total demand in growth total demand growth demand

End-use demand in 1999–2020 demand 2030 (Mtoe) 2000–30 demand in 2030 1999–2030 in 2030

sector 2020 (Mtoe) (percent) (percent) (percent) (percent) in 2030 (Mtoe) (percent) (percent)

Industry 605 2.7 46 553 1.9 43 696 1.6 41

Transport 205 5.3 16 236 4.1 23 339 4.0 20

Residential 397 1.5 30 217 3.1 17 464 1.8 27

Commercial 73 4.7 5 111 4.8 9 97 2.8 6

Other 43 3.2 3 97 1.6 8 101 4.1 6

Total 1,322 2.7 100 1,264 2.6 100 1,697 2.2 100

Source: APERC 2004.

130

Energy Policy 131

Whether China has reached peak levels is debatable; the trends areambiguous. In the first two decades of reform, particularly in the late1990s, energy intensity declined, partly as a result of technologicalupgrading in heavy industry and power generation and partly as a resultof the shutting down of obsolete firms (table 6.4). The decline may alsohave been an artifact of statistical anomalies surrounding underreportedcoal production and consumption. Whatever the cause, energy intensityappeared to be on the rise by 2002, the point, not coincidentally, atwhich the growing gap between the supply of and demand for electricpower generation began resulting in more frequent service interruptionsin booming manufacturing centers.

The globalization of production and the fragmentation of industrialsupply chains—phenomena intimately linked to China’s economicdevelopment—may have substantially changed the traditional relationbetween development and energy intensity. Because energy-intensiveproduction activities can be geographically delinked from production-related services and management that are not energy intensive, coun-tries like China may end up with disproportionately high levels of

Table 6.4. Energy Intensity, 1991–2005

(tons coal equivalent / GDP)

Year Energy intensity

1991 5.12

1992 5.12

1993 4.42

1994 4.18

1995 4.01

1996 3.88

1997 3.53

1998 3.15

1999 2.90

2000 1.40

2001 1.33

2002 1.30

2003 1.36

2004 1.43

2005 1.22

Source: NBS various years.

Note: GDP for 1991–99 calculated at 1990 prices; GDP for 2000–04

calculated at 2000 prices; 2005 GDP calculated at 2005 prices.

132 China Urbanizes: Consequences, Strategies, and Policies

energy-intensive production, while advanced industrial societies continueto produce higher-value services that are not energy intensive. Wholeindustries need not move globally, only particular segments of thoseindustries. For prolonged periods, economies such as China’s are thereforenot likely to attract a full-package of industrial activities (services andproduction) but a package heavily tilted toward energy-intensive activities.This is true both regionally and globally: more advanced economies,particularly in northeast Asia, have moved both manufacturing assemblyoperations and their industrial-driven energy needs and energy externalitiesto China (Gaulier, Lemoine, and Unal-Kesenci 2006). Although China ismore energy intensive than advanced industrial economies, its per capitaenergy consumption is nevertheless relatively modest. Low per capitaconsumption figures, however, do not suggest that China, even if it wereto achieve its efficiency targets, could simply do with its energy crunchwhat it has in so many other areas of economic and institutional reform—that is, grow its way out of the problem. To the contrary, low per capitaconsumption suggests that energy demand in China is likely to rise sub-stantially. Although high U.S. consumption patterns might not presageChina’s future, the more modest patterns associated with Japan, theRepublic of Korea, or the European Union—already several times China’scurrent consumption levels—probably serve as indicators of the directionin which China is heading.

Moreover, because of several factors—some specific to China, othersrelated to broader changes in the global organization of production—energyintensity is unlikely to decline as quickly as that of previous modernizers.Because it enjoys the mixed blessing of vast domestic coal reserves, forthe foreseeable future China will probably continue to rely on coal asthe main source of energy. With its high carbon content, coal burns lessefficiently than other hydrocarbons (such as oil or natural gas).The morecarbon in a hydrocarbon fuel, the less energy it has (lower hydrogen tocarbon ratios entail lower efficiency of combustion). To the extent thatChina remains dependent on coal, it will have to forgo the efficiency gainsassociated with the switch even to alternative fossil fuels. In addition,higher quality coal is concentrated in the north and northwest, thusnecessitating energy-consuming (often oil-consuming) transport toindustrial centers along the eastern and southeastern coast (60 percentof railroad transport is powered by coal). China’s problem, therefore, isnot just that fuel has to be transported over great distances but that thematerial being transported is not energy dense.

China’s Unique Energy Security ChallengeThe nature of China’s “energy security” challenge goes beyond the fact thatgrowth and modernization alone are not solutions to the supply-demandgap. In the broadest sense, energy security involves the accommodation ofdifficult-to-reconcile objectives: adequate energy for long-term economicgrowth, energy that can be secured without exposure to undue geopoliticalrisk, energy supply and utilization consistent with long-term public health,and energy supply flexible enough to meet rising popular expectationsfor public and private goods.

Under normal circumstances, these demands would be difficult tomeet. China’s circumstances are not “normal,” however, for several reasons.First, on the domestic front, the variables feeding into the energy securitycalculus are shifting with extreme rapidity. China is simultaneouslyexperiencing an industrial revolution, an economic boom, a rapid phaseof urbanization, and, in many respects, an information revolution, partic-ularly at the level of the individual citizen. Citizens have increasinglycome to expect not only macroeconomic growth and the energy neces-sary to fuel that growth but also a wide array of goods associated withadvanced economies (consumer goods, ranging from refrigerators toautomobiles, and public goods, ranging from clean air to comprehensivehealth care). This expectation means that energy provision—in terms ofboth quantity and quality—has become central to the issue of good gov-ernance. Put simply, good governance in China today entails fueling anindustrial revolution as dramatic as anything experienced by 19th cen-tury England but doing so in a manner acceptable to a public whose liv-ing standard expectations are decidedly 21st century and cosmopolitan.

Second, these challenges must be resolved at a time when at least onekey global energy resource, petroleum, appears to be approaching depletionin the medium term. Optimistic forecasts suggest that peak global oilproduction (Hubbert’s Peak, or the point at which expansion of produc-tion ceases and a depletion curve ensues) will occur around 2035; morepessimistic views assert that this point has already been reached(Deffeyes 2005).The amount of oil recorded each year as known reservespeaked in 1961. Since then, technological advances have permitted com-mercially sustainable drilling in the North Sea, Africa, and the Arctic.Much of the “easy oil” appears to have been extracted, however, and newfinds are becoming smaller and smaller. The “easy oil” that does existremains primarily in the Persian Gulf and more broadly in member nationsof the Organization of the Petroleum Exporting Countries (OPEC).

Energy Policy 133

In 2006 Gulf countries accounted for 31.1 percent of global crude oilproduction. Saudi Arabia alone accounted for 12.9 percent of global pro-duction (IEA 2007a). Given political instability in the region, the secu-rity and reliability of these flows are uncertain. Precisely as China movestoward becoming a modern economy, the future availability of petroleumis in serious doubt.

Whether and when peak global oil production will be reached isuncertain. What is clear, however, is that China is viewed by many of theworld’s largest energy producers and consumers alike as putting a majornew strain on global energy resources and markets. China’s consumptionpatterns, and the choices China makes to secure the resources needed tomeet those consumption needs, have become matters of concern for anumber of countries. Geostrategically, “business as usual” on the energyfront for China may entail increased competition and conflict with othermajor consuming nations, particularly the United States. China has littlechoice, then, but to seek to redefine traditional developmental paths andchart an alternative energy course into the future.

Internalizing ExternalitiesCharting a path to the future involves complex decisions, ultimatelyabout price. In the case of energy, however, calculation of price entailsthe internalization of extensive and highly ambiguous externalities.Coal, for example, appears inexpensive in the near term for China.But if coal is burned without environmental cleanup mechanisms, fluegas desulfurization systems, and related technologies, it imposes acostly public health toll. To the extent that the public deems urbanenvironmental conditions unacceptable, such sentiments also havepolitical ramifications.

Factoring these costs in raises the cost of domestic coal. But replacingcoal with alternative energy sources, such as imported petroleum ornatural gas, also creates negative externalities, such as the need to investin military assets to protect sea lanes or in diplomatic relationships withsuppliers. Taking these considerations into account, coal—albeit coalproduced using sophisticated decarbonization, gasification, or liquefac-tion processes—may be the least costly fuel after all.

Some “clean-coal” technologies, while promising, are unproventechnologically and commercially. Development costs may be high, butthey may permit the realization of positive externalities in industrialinnovation and global competitiveness. Even in the relatively nearterm—the 5- to 10-year horizon—externalities make the calculation of

134 China Urbanizes: Consequences, Strategies, and Policies

cost in the energy sector exceedingly complex, enough so to force policymakers to consider all options.

Given the scope of its energy needs, and its centrality in global produc-tion networks, China appears likely to be the place where “new to theworld” energy-related innovations—in civilian nuclear power, clean-coaltechnologies, efficiency-related upgrades on the consumption side, and avariety of other areas—will be implemented for the first time. Whether itis foreign or domestic players who design and implement these innovationsis open to question; that China will be the venue is almost beyond doubt.How this emerging reality will then feed back into Chinese economicdevelopment and affect China’s position globally on the industrialinnovation front represents an important issue for policy makers andcommercial actors alike.

Trends in Energy Consumption

China accounted for 14.2 percent of the world’s total energy consumptionin 2005 (IEA 2007a).Virtually across the board in the energy sector, Chinarepresents the fastest-growing market in the world. Electric power genera-tion, 70–80 percent of which is consumed by industry (a range that hasremained relatively stable in the reform era) faces tremendous expansionpressures to meet the relatively conservative projections for industrialdemand growth. The industrial sector is the driver of outcomes today;demand in the transport, urban residential, and commercial sectors remainsrelatively small but will grow significantly in the future (table 6.5).

Transport and AutomobilesFor decades, primary energy consumption in China has been dominatedby the electric power sector. This trend continues today at steady growthlevels, predictably driving demand for domestic coal.

A newer, more dynamic, and less predictable phenomenon is the ris-ing demand from the transport sector, demand that involves liquidhydrocarbons—petroleum today, but possibly liquefied natural gas andcoal-based liquids in the future. As the government ramps up infrastruc-ture investment and continues to promote the automobile industry,transportation-related energy demand is projected to rise 4.0–5.5 percenta year in the medium term. Noteworthy is both the pace of growth andthe fact that the required fuels are domestically scarce.

The increase in demand for petroleum is already evident. By the startof 2004, China was just overtaking Japan as the world’s second-largest

Energy Policy 135

Table 6.5. Total Energy Consumption, by Sector, 1997–2005

(ten thousand tons coal equivalent)

Sector 1997 1998 1999 2000 2001 2002 2003 2004 2005

Industry 92,375.3 88,521.9 87,151.2 95,442.8 98,273.3 104,088.1 121,731.9 143,244.0 159,491.6

Total residential

consumption 16,368.0 14,392.7 15,213.9 15,964.6 16,567.5 17,527.4 19,827.2 21,281.0 23,449.5

Transport, storage,

and post 7,286.3 7,957.0 9,011.8 10,067.1 10,363.0 11,171.0 12,818.8 15,104.0 16,629.2

Farming, forestry, animal

husbandry, fishery, and

water conservancy 5,905.4 5,790.3 5,993.4 6,045.3 6,400.3 6,612.5 6,716.0 7,679.9 7,918.4

Wholesale and retail

trade and catering 2,394.4 2,552.1 2,901.5 3,038.8 3,265.0 3,520.3 4,179.6 4,820.3 5,031.1

Construction 1,179.0 1,612.1 1,979.4 2,142.5 2,234.0 2,543.7 2,859.6 3,258.6 3,411.1

Other 4,702.8 5,212.6 5,562.5 5,851.5 6,096.4 6,334.1 6,818.7 7,838.8 8,691.2

Total 137,799.0 132,213.9 133,831.0 138,552.6 143,199.2 151,797.3 174,951.6 203,227.0 224,682.0

Source: NBS various years.

Note: Totals may not sum correctly because of rounding errors.

13

6

consumer of petroleum products.Almost a decade earlier in 1993, Chinahad become a net importer of oil (U.S. Department of Energy 2006).China’s oil demand is projected to reach 14.2 million barrels a day by2025 (see table 6.2).

Chinese demand for oil imports rose steadily throughout the 1990s,at 4 percent a year; by 2005, domestically produced crude oil accountedfor only 55 percent of total Chinese oil consumption. Strong demand foroil has made China a significant enough oil importer to move markets.The spring 2004 spike in oil prices was at least partly related to China’ssurging demand for imports, particularly in the context of an increasinglyuncertain geopolitical situation in the Persian Gulf. Economic developmentnaturally increases demand for transportation- and transport-relatedfuels. Rapid industrialization drives demand for electric power, which drivesdemand for coal, which must be transported through an increasinglyextensive rail and road system. Similarly, expansion and integration ofmarkets for intermediate industrial and final consumer goods means thatincreasing amounts of material must be transported by air, rail, and road.Throughout the 20th century, modernization has entailed the expansion oftransport economies. And unlike electric power generation, transportdepends almost exclusively on oil: transport accounted for 60.3 percent ofworld oil consumption in 2005, the single-largest sector by far (IEA 2007a).

In China, this natural shift is being accelerated and encouraged bygovernmental policy. The automobile sector has been promoted as a key“pillar” industry, on the basis of a series of presumed spillover effects. Itsextensive network of supporting and related industries is expected toprovide employment, and its technology intensity is expected to promoteinnovation and global competitiveness. Its final product simultaneouslydrives the deepening of financial markets (through auto financing), stim-ulates growth (through personal consumption), and meets demands formobility and modernity on the part of an increasingly sophisticatedemerging middle class.

China is hoping that the automobile industry will do for it in the 21stcentury what the industry did for the United States and Japan in the20th century. The danger is that China is pursuing this industrial strategyat a time when petroleum resources globally are becoming stretched andpopular awareness of the potential impact on already strained domesticenvironmental and infrastructure conditions is growing. In this sense, Chinais on a trajectory comparable to other developing nations, such as Thailand,where rapid growth in personal transportation led to severe traffic conges-tion and severe environmental problems in urban areas.

Energy Policy 137

China’s automobile sector has boomed since 2001. In 2002, Chinaproduced and sold 1 million cars, up 50 percent from the previous year.In 2006, China surpassed Japan as the world’s second-largest auto market(behind the United States), with total sales of 7.2 million units (“China2007 Auto Output”). By 2030 the total number of vehicles, estimated at37 million vehicles in 2006, is expected to grow to 370 million (Rosen andHouser 2007). Between 2002 and 2012, Chinese purchases are expectedto account for one-fifth of all new car sales in the world (Rosen andHouser 2007).

Automobiles create a variety of negative externalities.Although use ofnewer vehicles tends to increase fuel efficiency on a vehicle-mile basis,the trend globally in recent years has been toward decreases in fuel effi-ciency on a passenger-mile basis, as rising levels of automobile ownershiphave increased the use of single-occupant vehicles, increasing trafficcongestion. Such conditions are already apparent in most major Chinesecities. Particularly when promoted officially as the anchor of a consumereconomy and socially as a key indicator of sophistication and modernity,automobiles encourage extremely inefficient utilization of energy, withsubstantial environmental costs. Advances in internal combustion enginetechnology, infrastructure, and “smart” traffic management systems willlead to efficiency gains in the future, but they are likely to be offset bythe inefficiencies of declining mass-transit use and the rising costs of pol-lution. China’s macroeconomic growth requires the expansion of thetransport economy, but automobiles need not be a primary mode oftransportation. That they have become one is a reflection of choicerather than necessity.

This choice induces energy-related externalities in urban planning.Promotion of automobiles necessitates massive road and infrastructureconstruction. Severe constraints on land in Chinese cities and limitedpublic funds mean that construction of this infrastructure comes at theexpense of mass-transit systems. At the individual consumer level, auto-mobile ownership has enabled movement, particularly by the wealthy, tosuburbs, where parking is available, larger homes (associated with moreenergy-intensive heating and cooling, more appliances, and so forth) arepossible, and commuting in a single-occupant vehicle is common.

The substantial investments being made in the extensive supportingenergy infrastructure for automobiles—petroleum distribution facilities,filling stations, and so forth—raise the costs of switching to alternativetransportation fuels in the future.This extensive supporting infrastructurecreates a variety of vested interests that also make it difficult to switch toalternative fuels and alternative modes of transportation.

138 China Urbanizes: Consequences, Strategies, and Policies

The decision to promote automobiles will have tremendous ramifica-tions for China’s ability to adapt to changing energy circumstances inthe future. Significant vulnerabilities (urban pollution and congestion,dependence on external and uncertain sources of oil, and so forth) andsubstantial opportunity costs (investment in a public transport infra-structure, investment in alternative fuels, and so forth) are beingincurred as a result.

Urban Residential and Retail Energy DemandThe second major shift in energy demand is coming from rising urbanresidential and commercial utilization. Urbanization and rising incomesare usually accompanied by steep increases in household electricityconsumption. Acquisition of energy-consuming durable goods (washingmachines, televisions, refrigerators, and PCs) becomes the norm, anddemand for energy-intensive heating and cooling rises. In 1990, therewere about 42 refrigerators and 59 color televisions and 0.34 air condi-tioners for every 100 urban households in China. By 2005, those figureshad grown to 91, 135, and 81, respectively (NBS 2006).

Globally, increasing urban demand for electricity has moved forward intardem with global information technology (IT) revolution. On the onehand, the proliferation of computers, routers, and related IT infrastructurehas permitted the realization of certain energy efficiencies. Lean produc-tion has led to efficiencies in transport and transport-related fuels; digitaltransmission of information has reduced the need for face-to-face interac-tion and related travel; and IT–related smart traffic management systemsease energy-wasting congestion. On the other hand, increases in efficiencyhave been outmatched by the even greater increases in aggregate energydemand as residential and commercial consumers around the world aresurrounding themselves with IT-related products and equipment. The netresult has been that in the context of the IT revolution, countries as diverseas the United States and China have experienced increased demand forelectricity in the urban household and retail sectors.

Urban populations are more directly exposed to the pollution effectsof power generation. Thus, clean power generation becomes a primaryconcern, as does the desire to move heavy industry outside cities, increasingthe need for energy-consuming transportation development. Pressure forclean power encourages the promotion of noncoal-fired power plants,increasing demand for fuels such as natural gas or liquefied natural gas,which, particularly in the east and southeast, increasingly come fromoverseas. Urban consumer electricity demand entails more-complexpower management than traditional industrial utilization. Consumer

Energy Policy 139

demand fluctuates on a seasonal and daily basis; it not infrequentlyexhibits significant surges. Variability and intermittency create pressuresfor movement toward more flexible fuels and generating facilities andmore-distributed modular power systems. Traditional large-scale coal-firedplants become far less attractive, whereas smaller-scale systems, oftenutilizing natural gas or other more energy-dense fuels, which can bebrought on and off line, gain in appeal. As distributed power systems(based on fossil fuels or renewable alternatives) proliferate, pressuresincrease to find an effective currency for energy, a storage fuel (liquefiedhydrogen, liquefied coal, coal-based syngas, or a variety of other options)that can be transported easily across complex networks of smallerpower-generation facilities and multiple utilizations.

Rising urban demand creates pressures for substantial change in urbanenergy infrastructure, energy management, and technological develop-ment. Concerns about energy consumption should force thoughtfulconsideration of public choices about urbanization strategy. Even withan effective push toward efficient distributed power systems, China willstill likely suffer stiff energy penalties if policies of dispersed urbanizationare pursued. This is particularly true in transport, because smaller-scale,more-dispersed locales are less suited than large compact settings toextensive intraurban public transportation development. At the sametime, dispersed urbanization creates pressures for more-extensive, energy-intensive interurban transport, whether by road, rail, ship, or air.

Trends in Energy Production and Supply

Domestic production and supply of all fuels have increased since 2001.Despite those efforts, supply has been outstripped by demand.

CoalIn 2005, 76.4 percent of China’s primary energy production came fromcoal, 12.6 percent from petroleum, 3.3 percent from natural gas, and lessthan 7.7 percent from nuclear, hydropower, and wind (NBS 2006). Interms of the narrow definition of cost, coal is the cheapest fuel for largepower plants. The power industry in China is by far the largest consumerof primary energy. Moreover, heavy industry—which is likely to remain asubstantial component of the Chinese economy, regardless of gradual shiftstoward services and more information-intensive sectors—is a massive con-sumer of crude coal.

While over the long run, coal’s share in overall national energy consump-tion will gradually fall, absolute demand for coal will continue to rise,

140 China Urbanizes: Consequences, Strategies, and Policies

and for the foreseeable future, coal will remain the mainstay of China’senergy supply (MIT 2007).

Efficiency gains can be realized at various stages, including in theprocessing and conversion, transportation, storage, and final consump-tion of coal. Several projects exist for the colocation of large coal-firedpower plants near large, high-quality, low-sulfur content mines. Oneadvantage of locating coal near these mines is that crude coal no longerneed be transported across great distances. Options for utilizing thepower that is generated include transmitting coal by wire across powerlines, with some loss resulting in the process; creating coal-based liquidfuels, which could be transported relatively cheaply and could substitutefor petroleum in the transport sector; producing coal slurry, whichcould be transported by pipeline; and, potentially in the future, producingliquefied hydrogen.

A number of experimental projects are under way, including theShenhua Group’s coal liquefaction facility in Inner Mongolia and a varietyof other efforts involving coal gasification and coalbed methane production(UNESCO 2007). China has also expressed interest in experimental de-carbonization and carbon dioxide sequestration technologies for coal-basedpower generation.

PetroleumUse of petroleum and natural gas, while still a small portion of China’stotal energy supply, has accelerated in recent years. This trend is consistentwith pressures associated with modernization and other policy-inducedfactors (particularly the emphasis on automobile production and owner-ship). Rising use of petroleum and natural gas increases dependence onoverseas energy resources (table 6.6).

In response to this growing dependence on imported oil, Chinesefirms have been acquiring interests in overseas upstream exploration andproduction. Concessions have been acquired in Azerbaijan, RépublicaBolivariana de Venezuela, Indonesia, Islamic Republic of Iran, Iraq,Kazakhstan, Peru, and Sudan (U.S. Department of Energy 2006). Thepotential geopolitical risks are obvious, as is the challenge of competingwith other import-dependent oil consumers in East Asia, namely, Japanand the Republic of Korea.

Natural GasNatural gas, which has never been an important fuel in China, beganto receive substantial attention in the mid- to late-1990s. Accounting for3 percent of total energy consumption in 2005 (NBS 2006), natural gas

Energy Policy 141

Table 6.6. Imports and Exports of Energy, by Type, 1991–2005

(ten thousand metric tonnes, unless otherwise noted)

Item 1991 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Imports

Coal 136.8 163.5 321.7 201.3 158.6 167.3 212.0 249.0 1081.0 1109.8 1861.4 2,617.1

Crude oil 597.3 3400.6 2261.7 3546.6 2732.0 3661.4 7027.0 6026.0 6941.0 9102.0 12272.0 12,681.7

Gasoline 11.2 15.9 7.9 8.4 1.5 0.0 n.a. n.a. n.a. n.a. n.a. n.a.

Diesel 319.6 612.3 465.1 742.8 310.8 30.9 25.9 27.5 47.7 84.9 274.9 53.2

Kerosene 2.6 76.1 65.9 138.1 129.1 211.2 255.5 201.9 214.5 210.3 282.0 328.3

Fuel oil 124.6 659.1 942.6 1,371.1 1,627.2 1,757.0 1,480.0 1,823.6 1,659.7 2,395.5 3,059.2 2,608.6

Liquefied

petroleum gas n.a. 232.6 355.0 358.2 476.6 322.3 481.7 488.9 626.2 636.7 641.0 617.0

Other petroleum

products 11.5 95.7 106.5 176.1 190.6 208.1 161.5 201.3 384.3 432.1 384.2 443.4

Natural gas

(108 cubic meters) n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

Electricity

(108 kWh) 31.1 6.4 1.2 0.9 0.2 3.7 15.5 18.0 23.0 29.8 34.0 50.1

Exports

Coal 2,000.1 2,861.7 3,648.4 3,073.0 3,229.7 3,743.9 5,505.0 9,012.0 8,384.0 9,402.9 8,666.4 7,172.4

Crude oil 2,259.8 1,822.7 2,040.3 1,982.9 1,560.0 716.7 1,031.0 755.0 766.0 813.3 549.2 806.7

Gasoline 250.2 185.5 131.4 178.2 182.0 413.8 455.2 572.5 612.0 754.2 540.7 560.0

Diesel 121.0 130.6 157.4 232.1 98.5 60.5 55.5 25.6 124.0 224.0 63.7 147.6

Kerosene 32.1 37.4 74.4 72.3 91.6 125.0 198.8 182.2 170.0 201.7 205.0 268.7

14

2

(continued)

14

3

Table 6.6. Imports and Exports of Energy, by Type, 1991–2005 (continued)

(ten thousand metric tonnes, unless otherwise noted)

Item 1991 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Fuel oil 69.5 27.8 36.6 51.7 57.5 25.5 33.4 44.1 64.0 76.1 181.7 230.0

Liquefied

petroleum gas 1.1 7.1 33.3 39.2 50.2 7.5 1.6 2.1 5.6 2.4 3.2 2.7

Other petroleum

products 148.8 131.1 117.3 155.7 202.5 221.0 280.5 325.5 246.0 261.8 360.7 473.0

Natural gas

(108 cubic meters) n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 24.4 29.7

Electricity

(108 kilowatts) 2.6 60.3 37.1 72.0 71.7 91.5 98.8 101.9 97.0 103.4 94.8 111.9

Coke 108.3 886.1 768.6 1,058.1 1,146.4 997.4 1,520.0 1,385.0 1,357.0 1,472.1 1,501.2 1,276.4

Source: NBS various years.

n.a. Not available.

is expected to become an increasingly important fuel in the future asChinese cities seek cleaner sources of energy (author interviews). Con-struction of the extensive infrastructure needed to support this fuel—pipelines to distribute gas; shipping trains, port terminals, and gasificationfacilities needed to handle imported liquefied natural gas—is well underway (Watts 2006).

Nuclear Power Nuclear power has been developing rapidly, albeit from a low base,particularly with respect to the electricity sector. Generally speaking,nuclear power is a more expensive means of generating electricity thancoal or natural gas.

In 2005, the government declared its goal of adding 40 GW of civiliannuclear power capacity by 2020. China’s nine civilian nuclear reactorshad a total generating capacity of roughly 7 GW in 2006 (“China’s Goal”2006). Nuclear power accounted for 2.3 percent of Chinese electricitygeneration and 0.85 percent of total Chinese energy production in 2004(IEA 2007b). Even with the most ambitious growth program, nuclearpower will likely account for little more than 5 percent of total energysupply in the coming decades.

HydropowerHydropower represents an important component of Chinese electricpower generation, although it accounts for a relatively small componentof total energy production. In 2004 hydropower accounted for 2 percentof total Chinese energy production and 16 percent of electric powergeneration (IEA 2007b).

Increasing hydropower’s contribution to China’s overall energy mix isdifficult, because the sources of hydropower tend to be in the center andwest of the country, far from the main areas of regional demand along thecoast. The costs and energy inefficiencies associated with large-scale nation-al transmission and distribution systems are immense and arguably prohib-itive. These inefficiencies are exacerbated by the significant sociopoliticaland environmental costs of large-scale hydropower projects.

Policy Directions for the Future

China’s overall energy strategy is somewhat confused and uncoordinated—not unlike that of the world’s other large consuming nations, includingthe United States. China has pursued a number of ambitious efficiency

144 China Urbanizes: Consequences, Strategies, and Policies

goals and conducted a variety of interesting local experiments. Theseinclude Beijing municipality’s establishment of coal-free zones, Shanghai’smaglev train, regional pollution-rights trading programs, a national taxon high-sulfur coal, and municipal efforts to shift public buses over tocleaner burning fuels. This multiplicity of approaches, however, particu-larly when combined with other national goals that impinge indirectlyon energy, creates confusion and unintended consequences. Not unlikeapproaches to other aspects of institutional reform in China, energypolicy has been fragmented, both horizontally and vertically. Numerousexperiments, competing standards, and alternative microlevel approacheshave been allowed to proliferate. At the same time, at the central level,as in most countries, various aspects of energy policy—or policy areasthat impinge on energy issues—end up spread in uncoordinated fashionacross a range of administrative organs. Such diffusion and fragmentationmake all the more difficult the internalization of the externalities asso-ciated with national energy choices.

Whether by default or design, national industrial policy is energy policy.The decision to promote automobile production and consumption hasimplications for energy demand and urban planning; it also diverts researchand development (R&D) resources away from alternative energy projects.

Macroeconomic growth policy is also energy policy. Policies thatpromote growth and urbanization not only increase demand for energy,they also alter the kinds of energy demanded.

Environmental regulatory policy is also energy policy, to the extent thatit shifts the relative costs of fuels and the availability of energy-efficientappliances and materials. Health care policy is also energy policy, for itultimately must cope with the impact of pollution on people. Finally,given increasing dependence on foreign energy sources, foreign policy isenergy policy because ultimately it must be directed toward guarantee-ing steady overseas supplies.

Meeting China’s energy needs does not necessarily require centralizedsolutions, such as large-scale regional power-generation projects ornationally integrated power grids. Quite to the contrary, distributed,modularized power arrangements are in many cases better suited toChina’s highly varied geographic, demographic, and developmental land-scape. The point is that given the centrality of energy policy to China’sdevelopment goals, that centrality demands concerted attention andcomprehensive cross-bureaucratic coordination.

By virtue of its market size and rapid rate of growth, China has theability to make markets. In setting and enforcing tough energy-efficiency

Energy Policy 145

standards for consumer appliances and vehicles, the government leavesforeign producers little choice but to comply and innovate. Similarly, to theextent that domestic producers are forced to meet these standards, theydevelop core competencies in the design, development, and production ofenergy-efficient products, competencies for which global markets will onlygrow as energy constraints become more binding on all nations in the future.

As they chart their way to a more sustainable national energy posture,Chinese policy makers face important choices over a wide range of tech-nologies and energy-related sectors. The areas and recommendations listedbelow are intended to outline the domains across which change is bothpossible and likely to proceed.

Improving Energy Efficiency Rather than promoting automobiles or semiconductors as drivers ofnational innovation, the government should direct industrial policy towardthe development of alternative-energy vehicles and renewable energytechnologies. A national effort on these fronts not only would addressdomestic energy supply issues, but also would set up Chinese producers tobecome key innovators in an increasingly energy-constrained world. Inshort, China should use its power as a global producer and global con-sumer to make energy efficiency and energy-related innovation the core ofits national industrial competitiveness.

China must deepen its commitment to end-use, energy-efficiencyimprovements. In many cases, regulations are already in place but notuniformly enforced. As the building of urban commercial and residentialspace ramps up, it is imperative that the government promote energy-efficient designs and construction materials. By the beginning of 2007,China had become the world’s largest construction market, addingroughly 2 billion square meters of floor space every year (WorldwatchInstitute 2007). As of the end of 2006, the manufacturing and transportof building materials, the construction of new residential and commer-cial space, and the heating and cooling of buildings consumed 45 percentof China’s total primary energy. The 11th Five Year Plan (2006–10) callsfor energy savings of 50 percent in new buildings, but local developersare loathe to pay the higher up-front costs for energy-efficient materialsand building systems. Given the potential long-term energy—and, byextension, cost-savings from more-efficient construction techniques andmaterials—to government needs to enforce its emerging building stan-dards and to educate the public at large about the overall economic andenvironmental benefits.

146 China Urbanizes: Consequences, Strategies, and Policies

It is also imperative that the government enforce the new fuel standardsfor automobiles that it promulgated in 2004. Particularly given theappeal of its automotive market for global producers, China has everyreason to become a global leader in pushing vehicle fuel and emissionsstandards. The first phase of the new standards went into effect in 2005,and the second phase will commence in 2008. Enforcement has been, andwill continue to be, a main challenge in this process. Lax enforcementthreatens to vitiate not just the standards but the credibility of thegovernment more broadly.

End-use efficiency enhancement must be coupled with measures toensure that efficiency gains do not lead to expanded usage, as they havein many countries. Achieving this goal will inevitably involve complexmanagement of domestic tariff structures. In transport, for example, thegovernment will almost certainly have to explore restrictions on auto-mobile access to urban areas (along the lines of London’s congestionpricing or Singapore’s road-use pricing).

Allowing Market Forces to OperateIt is critical that energy prices be permitted to reflect market forces ofsupply and demand. For the most part, coal prices in China do reflectcurrent domestic supply and demand conditions, but prices for oil andelectricity clearly do not. The current system of setting domestic oilprices based on international levels (through a formula based on monthlyaverages in Singapore, Rotterdam, and New York) insulates domesticprices from local market guidance and leads to shortages. Oil prices needto be freed up domestically, so that domestic suppliers and consumerscan adjust accordingly.

Similarly, retail electricity prices tend to be shielded from marketguidance. Prices are kept artificially low, facilitating even more rapidgrowth in household appliances and unprecedented high peak powerloads in major Chinese cities. What results are blackouts and brownouts.Electricity prices must be allowed to reflect basic fuel prices, for coal, oil,or natural gas.

Shifting to GasificationAlthough coal will remain the dominant primary energy source, empha-sis must shift from combustion technologies to gasification. Such tech-nologies permit the production of cleaner gas and liquid coal-based fuels,alternatives to imported natural gas and petroleum. Gasification andliquefaction also facilitate potentially commercially viable carbon dioxide

Energy Policy 147

capture and sequestration, thus addressing the emission not only of sulfurdioxide but also of carbon-related greenhouse gases. These technologiesare still experimental today; focused research and development effortsare required to bring down costs and attain commercial viability. Suchimprovements are arguably more important—and more globally applica-ble—than anything else China could do today in the area of nationalindustrial policy.

Integrating Renewable Energy Sources on a Large ScaleWith the development of more-modular, distributed power systems, theability to integrate renewable energy sources on a large scale becomesincreasingly feasible. As suggested by the government’s 2005 NationalRenewable Energy Law, there is potential for far greater use of wind andsolar energy. Using Japan’s example to craft a regulatory framework thatsupports photovoltaic use in urban residential and commercial buildingsor Germany’s Freiburg model to promote both wind and solar power atthe municipal level (“Germany Sets Shining” 2007), China could sub-stantially increase its use of alternative renewables. Large-scale windfarms in the west could be linked to urban centers by high-voltage DCtransmission lines. With or without a shift toward hydrogen, Chinashould aim to rely on alternative wind and solar power for 10 percent ofits total energy supply by 2020. This would involve using sizable tractsof land not too far from centers of consumption.

Price Reform and Marketization in the Power Sector

It is in the power sector that some of the most dramatic changes inChina’s energy posture are manifested today. Driven both by industrialand urban household consumption, demand for electricity is soaring inChina. As China rushes to meet this demand by building new generationfacilities, expanding transmission networks, and securing new sources forkey fuels, the ramifications for everything from living standards to overallnational security are vast. On the electricity supply side, China facesurgent decisions regarding types of generation technologies and fuelfeedstocks to invest in, the location of new generation facilities, and theupgrading of transmission networks to transport power regionally andnationally. On the demand side, equally substantial issues are associatedwith how, where, and when consumers use power.

The choices made today have monumental consequences for the future.Through these choices, China can launch itself on a path of sustainable

148 China Urbanizes: Consequences, Strategies, and Policies

energy utilization—a path that will at once foster growth, rising livingstandards, and stability, both within and beyond China’s borders.

Principles of Marketization and Pricing in the Power SectorTechnological innovation and efficiency-promoting regulation in thepower sector are important elements of a long-term strategy. But themost fundamental element—the one on which the success of furtherreforms will hinge—is the issue of price reform and marketization. Pricesin any market are essential not just for collecting revenue but also forensuring sufficient supply and efficient utilization.To the extent that pricesignals are clear and unrestricted, they indicate to consumers the cost ofproducing the goods or services consumed; they indicate to producers thewillingness of consumers to pay. In theory, the market-clearing priceshould settle at the intersection of the marginal cost of the last producerand the marginal value to the last consumer. It is through this price thatresource allocation should ultimately be determined.

Pricing for power is not so simple. Electricity consumption flows overtime in a pattern of wide peaks and troughs. Because electricity cannotbe effectively stored in low-demand periods, it must be generated whenneeded. This fact has several important ramifications for the prices ofelectricity generation.

First, it makes sense economically to build generating plants of varyingtechnologies and fuel types. Some plants should be able to run all thetime at low cost (without being easily be ramped up or down in theshort run); others should be able to start and stop on short notice.

Second, as demand rises and falls, certain generating plants will comeon- and offline. The determination of the order by which this takes place(“dispatch”) should be driven by short-run marginal cost. Through“merit-order dispatch,” plants with the lowest marginal costs are broughtonline first, with those with higher marginal costs brought online in suc-cession as demand rises. In this manner, short-run costs to the system asa whole are minimized.

Third, the price paid by the final customer should be set at the mar-ginal cost of the system as a whole. The marginal cost of a generatingsystem is the running cost of the last (most expensive) generating plantbrought online each hour plus the value to the consumer of electricityat times when the system is short of capacity. In other words, outputprices for generation should be set at the running costs of the marginalproducer for each hour plus—for peak hours—a charge that recovers theinvestment cost of a peaking plant.

Energy Policy 149

Transmission pricing, in its intermediary position between generationand distribution, has its own complexities. In a marketized electricitysector, one would expect to see a variety of competing generators thatare dispatched on merit order. Given current wire-based technologies,however, transmission tends to be a monopoly activity necessitatingsome sort of regulated price. If the goal were simply to keep the trans-mission company solvent, one could divide total needed revenues byall the electricity sold, thus creating a “postage stamp” for use of thetransmission system. Such a mechanism would ensure cash flow tocover the transmission company’s existing cost structure, but it wouldnot provide any pressures or incentives to shift the underlying tech-nologies or management practices driving that cost structure. Toachieve the incentive effects needed for efficient resource allocationand utilization, however, a more complex, market-oriented tariff isnecessary, for two main reasons.

First, for efficient real-time use of a transmission network, users whoat any given time are willing to pay more (and thus value the networkmore) need to be given priority over those who do not. Prices mustultimately manage congestion, a problem that if left unresolved leads topower outages and instability in the power system. The costs of suchstrain must be internalized. One mechanism for doing so involves theuse of “nodal pricing.” In any power system, unique “prices” for electricitycan be defined at each node of the transmission system. Such prices varylocationally, depending on the amount of congestion in the system at agiven point and the distance from generating plants (because distancedrives the amount of electricity lost through transmission). In a market-ized system, generators are paid the price at their location, while largeconsumers and distribution companies pay the price at their location.Congestion rents accumulate when nodal prices diverge. Regulation andsupervision is then required to ensure that the transmission company—alocal or regional monopoly in most cases—does not grab these rents andthus face incentives to increase congestion.

Second, marketized transmission prices are necessary to guide longer-term location and investment decisions, whether for electricity producersand consumers. Electricity generators generally like to be near their fuels,while major industrial consumers like to be near their markets and cus-tomers. Transmission tariffs need to reflect the systemic costs (caused byincreased congestion or increased electrical losses) imposed by suchdecisions. The combination of a “postage stamp” transmission access feeand nodal price transmission tariff can achieve this reflection.

150 China Urbanizes: Consequences, Strategies, and Policies

In summary, efficient, sustainable utilization of energy resources dependson myriad interconnected decisions by producers and consumers. Themarket, operating through the mechanism of price, is the most effectivemechanism for guiding these decisions. Given the unique features of theelectric power sector, however, marketization can proceed only if certainconditions are met. Because of the differing market structures of genera-tion, transmission, and distribution, these three areas must be separated outin terms of both pricing and ownership. It is not enough simply to aggre-gate a series of charges related to electricity production and delivery andthen divide them by a unit of electricity sold. Rather, to facilitate merit-order dispatch on the generation side—a critical underpinning of marketpricing—competition must be permitted among generators.

Moreover, to ensure that dispatch actually proceeds on the basis ofmarginal cost, ownership of generation must be separated from ownershipof transmission. To the extent that transmission entities are permitted toown generators, conflicts of interest inevitably arise, because transmittersfavor their own generators in the dispatch ordering process and block theentry of new generators. Furthermore, particularly given the key role ofregulation in the less competitive parts of the power sector (transmissionand distribution), regulatory power must be separated from ownership.

Although the application of competition and market pricing generallybegins in generation, it must not stop there. Particularly for systems facingimmediate pressures for physical expansion, transmission pricing must gobeyond mere access fees to ultimately reflect the costs of congestion anddistance-induced losses. It is only at that point that price will effectivelyguide the sort of longer-term investment decisions by electricity producersand consumers that deeply affect the efficacy and physical status of thepower system as a whole.

Reform and Marketization in the Chinese Power SectorMarket restructuring of the power sector has been a clear policy goal ofthe government since at least the mid-1990s. The 1996 Electricity Lawpermitted the entry of nonstate entities into the generating sector, recog-nized the need for electricity prices to cover producer costs, andacknowledged the need to separate the regulatory function of the govern-ment from the ownership role of power producers. This law wasfollowed in 1998 by State Council Document 146, which mandated theseparation of ownership of electricity generation from the transmissionnetwork, thus providing the means for an unbundling of generation andtransmission prices and the means for merit-order dispatch.

Energy Policy 151

State Council Document 5, issued in 2002, pushed the agenda sub-stantially forward by calling for full competition in the power sector,beginning with generation. Market trials permitting generators to sellpower directly to large customers were permitted. The document alsoidentified a series of longer-term goals, including (a) the formal separationof generation from transmission in terms of ownership and regulation;(b) the establishment of competitive regional markets for dispatchinggenerators; (c) the establishment of new pricing mechanisms, includingmechanisms that take into account environmental impacts; and (d) thedevelopment of market-oriented pricing mechanisms for all parts of theelectricity supply chain, including not just generation but also transmission,distribution, and retail pricing.

That the government committed itself to this highly ambitious andprogressive agenda is both extraordinary and commendable. In at least onearea—the freeing up of rules on power plant financing—the successes areindisputable. Changes appear to have been far less dramatic in other areas,although information is anecdotal. Diversification of financing for—andownership over—power plants has driven a substantial ramping up ofgenerating capacity since 2002.

At the same time, a vast gap remains between these goals and realityon the ground. Several reforms need to be made.

Improve the pricing of electricity. China’s system of electricity pricingremains rigid, inefficient, and nonmarket oriented. Those are basicallytwo types of tariffs: one for the purchase of power by provincial orregional power companies from independent power producers and onefor the purchase by final consumers from the power company. The firsttariff is determined contractually on a generator-by-generator basis. Thesecond tariff is fixed, varying only by class of consumer (industrial ver-sus household, high-voltage versus low-voltage, and so forth). The vastcomplexity in pricing is based not on time, place, or extent of usage—thefactors one would expect market pricing to be based on—but on thenature of the customer. Moreover, these tariffs are unresponsive to shiftsin supply and demand.

Unbundle generation and transmission pricing. No clear mechanismexists for passing along efficiency-related cost reductions on the part ofgenerators to consumers, and no clear mechanism has been set for raisingthe funds needed to construct and upgrade transmission and distributionnetworks. No mechanism exists for incorporating into the final retail price

152 China Urbanizes: Consequences, Strategies, and Policies

of electricity the costs arising from system congestion and electricity lossthrough transmission.

Implement a clear method of market-oriented, merit-based generatordispatch. Dispatch hours (running times) are currently allocated to plantsbased on the principle of “fair” distribution. Because merit-order dispatchdoes not occur—and indeed cannot occur, to the extent that generationand transmission prices remain bundled—electricity prices remain funda-mentally nonmarket oriented.

Simplify cross-subsidies and increase transparency of differential pricingsystems. A substantial portion of electricity consumers—namely, the urbanhousehold sector—pay a low price for electricity, which is subsidized byhigher-voltage industrial customers and the power generators themselves.Generators find themselves caught between liberalized, rising fuel costsand governmental restrictions on the amount that can be charged forelectricity production. Even some of the newer power projects that havepower purchase agreements (PPAs) with regional grid companies—whichmandate prices higher than national standards—have faced substantialproblems. The PPA mandated prices have been overridden by governmen-tal pricing bureaus in the name of fighting inflation. Moreover, in regionswhere surplus power exists, state grid companies in some cases have refusedto abide by the PPA take-or-pay clauses to which they initially agreed.

Implement regulatory restructuring. To the extent that the distinctionbetween transmission company and generator and between commercialoperator and governmental regulator remains blurry, real marketization isunlikely to occur. The problems of inflexible tariffs, bundled tariffs, non-market-based dispatch, price subsidies, and regulatory conflicts of interestare deeply intertwined. These problems fundamentally impede the sort ofmarket-oriented price reforms that are absolutely necessary to guide thebehavior of commercial producers and consumers as well as long-terminvestment decisions.

China can and should pursue a variety of means of ensuring energysecurity for the future. Among these means are regulatory regimes thatencourage energy conservation, diversification into new fuels, anddevelopment and dissemination of energy-saving technologies. In thenear term, however, price reform in the power sector stands out as notjust the single-greatest policy challenge but also the one that, if met,will yield the highest near-term returns and have the most profound

Energy Policy 153

impact on the behavior of energy producers and the growing numbersof urban consumers alike.

References

APERC (Asia Pacific Energy Research Centre). 2004. Energy in China:Transportation, Electric Power, and Fuel Markets. Institute of Energy Economics,Tokyo. http://www.ieej.or.jp/aperc/pdf/CHINA_COMBINED_DRAFT.pdf

“China 2007 Auto Output.” http://www.Forbes.com, September 9.

“China’s Goal to Increase Nuclear Power Difficult but Unchanged.” 2006. XinhuaNews Service, June 8.

Deffeyes, Kenneth S. 2005. Beyond Oil: The View from Hubbert’s Peak. New York:Farrar, Straus and Giroux.

Downs, Erica S. 2004. “The Chinese Energy Security Debate.” China Quarterly177 (March): 21–41.

EIA (Energy Information Administration). 2003a. China: Country Analysis.U.S. Department of Energy, Washington, DC. http://www.eia.doe.gov.

______. 2006. China: Country Analysis. U.S. Department of Energy, Washington,DC. http://www.eia.doe.gov.

———. 2003b. China: Environmental Issues. U.S. Department of Energy,Washington, DC. http://www.eia.doe.gov.

———. 2004a. China’s Nuclear Industry. U.S. Department of Energy,Washington,DC. http://www.eia.doe.gov.

———. 2004b. International Energy Outlook 2004. http://www.eia.doe.gov/oiaf/ieo/download.html

Gaulier, Guillaume, Françoise Lemoine, and Deniz Unal-Kesenci. 2006. “China’sEmergence and the Reorganisation of Trade Flows in Asia.” Centre d’EtudesProspectives et d’Informations Internationales Working Paper 2006-05, Paris.

“Germany Sets Shining Example in Providing a Harvest for the World.” 2007.Guardian, July 23.

Hunt, Sally. 2002. Making Competition Work in Electricity. New York: Wiley.

———. 2003. Guiding Principles of Pricing: Jiangsu Power Sector Tariff Study. WorldBank, Washington, DC.

IEA (International Energy Agency). 2003. Key World Energy Statistics.http://www.iea.org.

______. 2007a. Key World Energy Statistics. http://www.iea.org.

______. 2007b. “Statistics by Country/Region.” http://www.iea.org/Textbase/stats/index.asp.

154 China Urbanizes: Consequences, Strategies, and Policies

Kroeze, Carolien, Jaklien Vlasblom, Joyeeta Gupta, Christiaan Boudri, andKornelis Blok. 2004. “The Power Sector in China and India: Greenhouse GasEmissions Reduction Potential and Scenarios for 1990–2020.” Energy Policy32 (1): 55–76.

Landsberg, Mitchell. 2007. “The World: China May Lead in Greenhouse Gases.”Los Angeles Times, June 21.

Larson, Eric D., Wu Zongxin, Pat DeLaquil, Chen Wenying, and Gao Pengfei.2003. “Future Implications of China’s Energy-Technology Choices.” EnergyPolicy 31: 1189–1204.

Manning, Robert A. 2000. The Asian Energy Factor: Myths and Dilemmas of Energy,Security, and the Pacific Future. New York: Palgrave.

McGregor, Richard. 2007. “China’s Power Capacity Soars.” Financial Times.February 6.

MIT (Massachusetts Institute of Technology). 2007. The Future of Coal: Optionsfor a Carbon-Constrained World. Cambridge, MA. http://web.mit.edu/coal/.

NBS (National Bureau of Statistics). 2006. China Statistical Yearbook. Beijing:China Statistics Press.

National Research Council. 2004. The Hydrogen Economy: Opportunities, Costs,Barriers, and R&D Needs. Washington, DC: National Academies Press.

Ni, Weidou, and Thomas B. Johansson. 2004. “Energy for Sustainable Developmentin China.” Energy Policy 32 (10): 1225–29.

Roberts, Paul. 2004. The End of Oil. Boston: Houghton Mifflin.

Rosen, Daniel, and Trevor Houser. 2007. “China Energy: A Guide for thePerplexed.” In China Balance Sheet. Center for Strategic and InternationalStudies and Peterson Institute for International Economics, Washington, DC.

Shiu, Alice, and Pun-Lee Lam. 2004. “Electricity Consumption and EconomicGrowth in China.” Energy Policy 32 (1): 47–54.

Sinton, Jonathan E., and David G. Fridley. 2003. “Comments on Recent EnergyStatistics from China.” Sinosphere Journal 6 (2): 6–11.

Sun, J. W. 2003. “Three Types of Decline in Energy Intensity: An Explanation forthe Decline of Energy Intensity in Some Developing Countries.” Energy Policy31 (6): 519–26.

UNESCO/Shell Chair in Coal Gasification Technology. 2007. “Report onApplying Coal Gasification Technology in China’s Coal Chemical Industry.”UNESCO, Beijing.

U.S. Department of Energy. 2005. National Security Review of International EnergyRequirements. Washington, DC.

U.S. Department of Energy. 2007. “International Data Projections.” http://www.eia.doc.gov/oiaf/forecasting.htm.

Energy Policy 155

Victor, David G., Thomas C. Heller, Joshua C. House, and Pei Yee Woo. 2004.“Experience with Independent Power Projects (IPPs) in Developing Countries.”Working Paper 23, Center for Environmental Science and Policy, Program onEnergy and Sustainable Development, Stanford University, Palo Alto, CA.

Wald, Matthew L. 2004. “Questions about a Hydrogen Economy.” ScientificAmerican 290 (5): 66–74.

Wang, Xiaohua, and Zhenmin Feng. 2003. “Energy Consumption withSustainable Development in Developing Country: A Case in Jiangsu, China.”Energy Policy 31: 1679–84.

Watts, Jonathan. 2006. “Thirsty China Opens Huge LNG Terminal.” Guardian,June 28.

Worldwatch Institute. 2007. “China Pushing for Energy Efficient Buildings.”http://www.worldwatch.ore/node/4874.

Wu, Yanrui. 2003. “Deregulation and Growth in China’s Energy Sector: A Reviewof Recent Development.” Energy Policy 31 (15): 1417–25.

Yang, Hong, He Wang, Huacong Yu, Jianping Xi, Rongqiang Cui, and GuangdeChen. 2003. “Status of Photovoltaic Industry in China.” Energy Policy 31:703–07.

Zhang, Chi, and Thomas C. Heller. 2003. “Reform of the Chinese ElectricPower Market: Economics and Institutions.” Working Paper 3, Center forEnvironmental Science and Policy, Program on Energy and SustainableDevelopment, Stanford University, Palo Alto, CA.

Zheng, Li. 2007. “Energy and China.” Paper presented at the MIT EnergyInitiative Energy Seminar, Massachusetts Institute of Technology, Cambridge,MA, February 21.

156 China Urbanizes: Consequences, Strategies, and Policies

The geography, spatial characteristics, and pace of urbanization in Chinawill be powerfully affected by the availability of potable water to urbanresidents and industry. This chapter examines the evolving water supplysituation in China’s urban sector; the degree to which water could con-strain potential growth and urban development in parts of the country(Bao and Fang 2007); and the scope for enhancing the efficiency withwhich supplies are used and recycled.

The chapter is divided into six sections. The first section provides anoverview of the growing scarcity of water in China. The second sectiondescribes patterns and trends of water supply. The third section describesthe effect of pollution on water supply. The fourth section assesses thelikely trajectory of water demand and its distribution across sectors. Thefifth section draws some implications for investment. The final sectionprovides some policy recommendations.

Low per Capita Availability of Water

China’s total naturally available water flows (not stocks) from all surfaceand underground sources are estimated at about 2,812 billion cubic

C H A P T E R 7

Water and Urbanization

157

The author would like to acknowledge the able research assistance provided by Holly Li,Siyan Chen, and Tomoko Okano, as well as comments and inputs provided by Hua Wang,Shahid Yusuf, Kaoru Nabeshima, and anonymous reviewers of earlier drafts.

Zmarak Shalizi

meters a year, placing China fifth in the world, behind Brazil, the RussianFederation, Canada, and Indonesia (FAO 2007). However, on a per capitabasis, China’s naturally available annual water flow of 2,114 cubic metersper person in 2003–07 is one of the lowest levels in the world for a popu-lous country, next only to India’s 1,150 cubic meters per person (FAO2007). China’s available water per person is one-third the world average(6,794 cubic meters per person), and one-quarter the average for theUnited States (9,446 cubic meters per person) (World Bank 2007). Thus,in a global context, China’s per capita availability of water is exceedinglylow, suggesting the potential for water stress as demand for usable waterrises with growth in population and per capita income.

Despite the one-child policy introduced in 1979, China’s populationhas been growing steadily, from almost 1 billion in 1980 to 1.19 billion in1993 and 1.31 billion in 2005 (table 7.1). As a result, annual per capitawater availability dropped by 25 percent between 1980 and 2005, from2,840 to 2,147 cubic meters per person (table 7.2).

Regional Differences in Water Availability

China’s low natural availability of water per person masks substantialregional disparities in water availability.1 Demand for water is growingthroughout the country, but total water availability in the north isabout one-sixth that in the south (405 billion cubic meters versus2,406 billion cubic meters (see table 7.2) and one-tenth the worldaverage (Wang and Lall 2002). The 596 cubic meters per person in thenorth in 2005 qualifies the north as a whole as an area of water scarcity,a condition worse than one of water stress.2 The north is a very large areaand was home to 680 million people (more than the total population ofEurope or Latin America) in 2005. Although it accounts for roughly52 percent of China’s population, it has just 14 percent of China’swater resources (NBS 2006).

Water scarcity is most acute north of the Yangtze River, particularly inthe catchments of the Huai, Hai, and Huang (Yellow) Rivers (the 3-H

158 China Urbanizes: Consequences, Strategies, and Policies

1 Average annual rainfall is about nine times greater in the southeast (1,800 millimeters)than in the northwest (200 millimeters). More than 45 percent of China receives lessthan 400 millimeters of precipitation a year (Economy 2004).

2 Water scarcity is defined as an annual supply of water less than 1,000 cubic meters perperson. Water stress is defined as an annual supply of water of less than 2,000 cubicmeters per person.

15

9

Table 7.1. Population of China, 1980–2005, by Region

Population

1980 1993 2002 2005 Annual growth rate (percent)

Region Billion Percent Billion Percent Billion Percent Billion Percent 1980–93 1993–2005 1980–2005

Northa 0.52b 52.5 0.62 52.1 0.65 51.0 0.66 50.4 1.4b 0.5 1.0c

Southa 0.48b 47.5 0.57 47.9 0.63 49.0 0.65 49.6 1.3b 0.1 1.2c

Urban 0.19 19.2 0.33 27.7 0.50 39.1 0.56 42.8 4.3 4.5 4.4

Rural 0.80 80.8 0.85 71.4 0.78 60.9 0.75 57.3 0.5 –0.01 –0.3

Totalc 0.99 100.0 1.19 100.0 1.28 100.0 1.31 100.0 1.4 0.8 1.1

Source: NBS 1981, 1994, 2003, and 2006.

a. The north–south split is based on World Bank 2001a and IIASA 1993. North is defined as the Huai, Hai, and Huang River basin provinces (Beijing, Tianjin, Hebei, Shanxi, Inner Mongolia, Jiang-

su, Anhui, Shandong, Henan, Shaanxi, Gansu, Qinghai, and Ningxia) and the three provinces in the northeast (Liaoning, Jilin, Heilongjiang). South is defined as the rest of China.

b. Figures are from NBS 1981 for 1981, and the population of south China does not include Hainan Province.

c. Excludes Hong Kong, Macao, and Taiwan.

rivers).3 Since the 1980s, the magnitude and frequency of water shortageshave been growing, generating severe economic losses.4 Total water short-ages in 2000 were calculated at 38.8 billion cubic meters; unless measuresare taken to reduce demand and augment supplies, they are projected toreach 56.5 billion cubic meters by 2050. These shortages are estimated tocost the Chinese economy Y 5.0–Y8.7 billion a year (US$620 million–US$1.06 billion) (Economy 1997; Economy 2004).5

The problems in Beijing and the Hai River basin are well known butnot unique. In the relatively dry regions in the north, northwest, andnortheast, there are many large urban centers, including seven cities withpopulations of more than 2 million each and 81 cities with populationsof 200,000–500,000 each. In many of the major cities, urban water usehas increased, as mayors have embarked on beautification campaigns to

160 China Urbanizes: Consequences, Strategies, and Policies

Table 7.2. Gross Water Availability per Capita, in North and South, 1980–2005

Water availability per capita

Gross water availability a (cubic meters)

Item Billion cubic meters Percent of total b 1980 1993 2002 2005

Total 2,812 100 2,840 2,363 2,197 2,147

Surface 2,712 96 (76)

Aquifer 829 29 (23)

North 405 14 779 653 623 614

Surface 334 12 (10)

Aquifer 169 6 (5)

South 2,406 86 5,015 4,223 3,819 3,702

Surface 2,377 85 (67)

Aquifer 678 24 (19)

Source: IIASA 1993; table 7.1.

a. The sum of surface and aquifer water exceeds the total water resource by the amount of overlap between them.

b. The figures in parentheses are adjusted to account for the overlap.

3 In the densely populated Hai River basin, for example, industrial output is growing rap-idly, and the basin is intensively cultivated. However, water availability per capita isonly 343 cubic meters a year. Residents in the Pearl River basin in the south have ninetimes more water available per capita.

4 Agriculture is the most water-intensive activity, followed by food processing, paper, andtextiles (Guan and Hubacek 2007). It takes 1,000 tons of water to produce 1 ton ofgrain. The water-scarce north exports agricultural products to other regions, using7,340 million cubic meters of water, of which 4,284 million cubic meters is from sur-face water resources and the rest from rainfall. This amounts to a net export of 5 percentof water resources from the north to other regions of China. In contrast, Guangdong,relatively water-rich province, imports water-intensive goods (about 445 million cubicmeters) and produces or exports electric components and various commercial andsocial services, which are not water intensive (Guan and Hubacek 2007).

5 These numbers were calculated for all China as of 1997.

plant trees, shrubs, flowers, and grass along roadways and in municipalparks (USDA 2000), in part to attract new investments and skilled laborand in part to combat locally the effects of dust storms associated withthe depletion of surface and aquifer water elsewhere. These large citiescompete with agriculture for scarce water resources. The problems areemerging in an acute form in other metropolitan subregions experiencingvery rapid growth, because the elasticity of water demand with respect tourban population growth is greater than one (Bao and Fang 2007). Morethan 400 of China’s 600 cities are believed to be short of water, and about100 face serious water shortage problems (Wang and Lall 2002).

To compensate for surface water scarcity, China uses a growing relianceon groundwater in the north and desalinated water in coastal areas.6

Groundwater is being depleted at a faster rate than it is being replenished,leading to “mining” of aquifers. When aquifers are mined, they are notavailable as insurance in drought periods, compromising sustainable use ofthe resource for current as well as future generations. In 2006, 30 percentof arable land in Sichuan Province was expected to yield no outputbecause of the drought (“Still Poor” 2006).7 In some areas, the overuse ofunderground water is contributing to severe aridity and increasing migra-tion away from fragile lands.

The extent of the mining of groundwater is severe. Sustainablegroundwater flows in the Hai River basin have been estimated to be onthe order of 17.3 billion cubic meters a year, while 1998 withdrawalswere 26.1 billion cubic meters a year, indicating overextraction of asmuch as 8.8 billion cubic meters annually. As a result, groundwater

Water and Urbanization 161

6 China is also investing heavily in desalination plants. The second-largest plant in thecountry, which can process more than 100,000 tons of water a day, will be built inZhejiang, at a cost of Y 1.1 billion. Its production will enable it to supply industrialusers and 500,000 people across the coastal Xiangshan county. More than 20 desalina-tion plants process 120,000 cubic meters of seawater a day. By 2010 this will increaseto 800,000–1 million cubic meters a day. The State Development and ReformCommission forecasts that desalinated water will account for 16–24 percent of waterused in coastal areas in the future (“China Turns” 2007).

7 Rainfall in Guangdong Province was down 40 percent in 2005 (MacBean 2007). Theworst drought in 30 years hit Liaoning in 2007, drying up 88 small and medium-size reservoirs and leaving 1.2 million people short of drinking water (“DroughtLeaves” 2007). Water scarcity and climate change could reduce China’s agriculturaloutput by 5–10 percent by 2030. However, China is still aiming to achieve the targetof producing 95 percent of its grain consumption domestically (“China to Keep” 2007).In addition to water shortages, air pollution reduces agricultural productivity. Almost 70 percent of the crops planted in China cannot attain optimal yields, primarily becauseof the haze from pollution (MacBean 2007; Shalizi 2007).

tables have dropped by as much as 90 meters in the Hai plains (WorldBank 2001a). The groundwater table in Beijing is estimated to havedropped 100–300 meters. Anecdotal evidence suggests that some deepwells around Beijing now have to reach 1,000 meters to tap usable quan-tities of water, dramatically increasing the cost of water supply and therisk of contamination from arsenic and other contaminants.

The removal of underground water domes has many adverse conse-quences. It has resulted in saltwater intrusion along coastal provinces in72 locations, covering an area of 142 square kilometers, according to oneestimate (World Bank 2001a). It is also leading to subsidence in coastaland noncoastal areas. The subsidence is up to several meters in citiessuch as Beijing, Shanghai, Shijiazhuang, Taiyuan, and Tianjin, causingdamage to buildings and bridges and even leading to their collapse.Subsidence of land as water is extracted is also diminishing flood protec-tion and exacerbating water logging in urban areas, because drainage isless effective (World Bank 2001a).8

Contribution of Pollution to Water Shortages

Many of China’s water bodies are polluted, some heavily so. Surface andgroundwater pollution now represent a major problem for both publichealth and the environment. Pollution-degraded water exacerbates theshortage of water resources downstream. It also makes it difficult torecycle water where it is scarce. As such, pollution represents a growingconstraint on national development objectives in China.

In 2003, 38 percent of China’s river waters were considered to be pol-luted, up from 33 percent a decade earlier. According to the 2003 annualreport of the State Environmental Protection Administration (SEPA), morethan 70 percent of the water in five of the seven major river systems—theHuai, Songhua, Hai, Yellow, and Liao—was grade IV or worse, meaning itcould not be used for of any designated beneficial uses. In the Hai andHuai River systems, 80 percent of the water was unusable (EIA 2003;SEPA 2003). Even the majestic Yangtze River suffered a sharp decline inwater quality, more than doubling the percentage of its water not suitablefor human contact to 48.5 percent in 2002 (Economy 2004).

Half of all water pollution is caused by nonpoint sources in ruralareas, including fertilizer runoff (which increases the flow of nitrogen

162 China Urbanizes: Consequences, Strategies, and Policies

8 Such environmental damage can be reduced with better management of groundwaterextraction. See the example in Zhengzhou (Gong, Li, and Hu 2000).

and phosphorous into water bodies [see Palmer 2001]), pesticides; andwaste from intensive livestock production. These problems, especially incertain rural areas close to cities, can be expected to worsen in the nearfuture. With growing urban demand, livestock production has increasedits contribution to the gross value of agricultural output from 14 percentin 1970 to 31 percent in 1998. Horticultural production for urban centersis also rising steeply. These trends are expected to continue, as urbaniza-tion increases, disposable incomes rise, and food distribution systems inrural areas improve. Rural sources of pollution, such as livestock operations,rural industry, and towns and villages, remain essentially uncontrolledand unaccounted for by current government management programs(Wang 2004).

The remaining half of water pollution comes from industrial andmunicipal wastewater discharges and the leaching of pollutants fromunlined solid waste sites into surface or below-ground water bodies(World Bank 2004). The rapid growth in urban populations and indus-trial activities is adding to the pollution of China’s waterways fromphosphorous, indicator bacteria, metals, and solvents. In the absence ofsufficient water treatment plants, large volumes of raw sewage aredumped into local streambeds daily, and industrial water is oftenuntreated. Only 56 percent of urban wastewater was treated in 2006.The target is to reach 70 percent for cities with populations of morethan 500,000 (“Strong Growth” 2007). When upstream water isreturned to the stream polluted, water quality downstream is degraded.In some cases, polluted water in the streams has seeped into the ground-water (USDA 2000). Government monitoring and enforcement programsare having only limited impact, because of selective application of thelaws and low levels of fines at the provincial and central levels, combinedwith weak enforcement of rules at the local level, which diminish thedeterrence value of regulations. Regulations are also incomplete insofaras load-based standards are absent and the standards that are set are notachievable given China’s current technological capabilities. For thesereasons, “more than 75 percent of the water in rivers flowing throughChina’s urban areas is unsuitable for drinking or fishing. Only 6 ofChina’s 27 largest cities’ drinking water supply meet State standards . . .[and] many urban river sections and some large freshwater lakes are sopolluted that they cannot even be used for irrigation” (Economy 2004;see also ABS Energy Research 2006).

In 2000 the major water pollutant—chemical oxygen demand (COD)discharge—was split almost evenly between industrial and municipal

Water and Urbanization 163

sources.9 Industrial sources, mostly in urban areas, contributed about 7.40million tons, while municipal sources, from commercial, residential, andpublic amenities, contributed about 7.05 million tons of COD. Municipalwastewater and COD discharge has been growing (at 3.1 percent a yearin the 1990s) relative to industrial wastewater discharge; by 2000,municipal wastewater discharge (at 22.1 billion tons) was 14 percent(2.7 billion tons) more than industrial wastewater discharge (Wang2004). The costs of benefits forgone by not treating wastewater wereestimated at Y 4 billion in 2000 rising to Y 23 billion in 2050 in the Haiand Huai basins (World Bank 2001a).

Disaggregating the total industrial discharge into sectors shows that sixsectors (pulp/paper, food, chemicals, textiles, tanning, and mining) accountfor 87 percent of total industrial COD load but only 27 percent of thevalue of gross industrial output. Toxic pollution loads (principally metalsand solvents) are undocumented but estimated to be about 1.7 percentof total COD loads, representing a significant threat to public health andaquatic systems. High pollution loads in the water seriously affect thepollution of coastal zone waters, which do not meet coastal zone standardsfor marine aquatic life (Wang 2004).

Recent Trends in Water Demand

Water use in China is sometimes disaggregated into four categories. Tworeflect production-related demand by farms and factories (agricultureand industry), and two reflect consumption-related demand by house-holds (rural and urban) (table 7.3).

Production-Related Demand Agriculture remains the largest user of water in China, accounting forabout 64 percent of the total in 2005 (NBS 2006), even though annualwater use for agriculture decreased by 3 percent between 1980 and 2005(from 370 billion to 358 billion cubic meters), as industrial and urbanneeds preempted agricultural needs, and productivity (efficiency) ofwater use in agriculture increased.

Industry, which has sustained double-digit growth rates since the early1980s, is the second-most important source of demand for water.

164 China Urbanizes: Consequences, Strategies, and Policies

9 COD measures the oxygen needed to decompose organic matter. The United States usesthe five-day biochemical oxygen demand criterion (the amount of oxygen required bybacteria to break down organic matter over five days).The modeling of water quality andthe factors influencing the level of dissolved oxygen are described by Palmer (2001).

Table 7.3. Water Use, by Sector, 1980–2005

Water use

1980 1993 1997 2002 2005

Volume Volume Volume Volume Volume Annual growth rate

(billion (billion (billion (billion (billion (percent)

cubic Percent cubic Percent cubic Percent cubic Percent cubic Percent 1980– 1997– 1980–Sector meters) of total meters) of total meters) of total meters) of total meters) of total 1997 2005 2005

Production 416 94 471 91 504 91 488 89 487 87 1.1 –0.4 0.6

Agricultural 370 83 383 74 392 70 374 68 358 64 0.3 –1.1 –0.1

Industrial 46 10 89 17 112 20 114 21 129 23 5.4 1.8 4.2

Domestic 28 6 47 9.1 53 9 62 11 68 12 3.7 3.2 3.6

Urban 6.8 1.5 24 4.6 25 4.4 32 5.8 n.a. n.a. 7.9 5.3a 7. 3b

Rural 21.3 4.8 23 4.4 28 5 30 5.4 n.a. n.a. 1.6 1.4a 1.6b

Total 444 100 519 100 557 100 550 100 563 100 1.3 0.1 1.0

Source: IIASA 1999; NBS 2003, 2006.

n.a. Not available.

a. Figures are for 1997–2002.

b. Figures are for 1980–2002.

165

Between 1980 and 2005, water use in industry increased from 46 to 129billion cubic meters, an increase of 280 percent. In 2005 industryaccounted for 23 percent of total water consumption (NBS 2006).10

Together, the production sectors—agricultural and industry—are respon-sible for 87 percent of water demand in China.

Some observers believe that water demand by industry may bedecelerating, as industries are becoming more water efficient or shiftingtoward subsectors with lower water requirements (University of BritishColumbia 2004). The evidence for this, however, is still anecdotal. Evenif a shift is occurring, as recently as the late 1990s, industry in China wasconsuming 4–10 times as much water as industry in more-industrialcountries (Wang and Lall 2002). China uses six times more water perunit of GDP than the Republic of Korea and 10 times more than Japan(“Still Poor” 2006).

Consumption-Related Demand Urban residential water demand was insignificant in 1980, at 1.5 percentof the total. By 2005, the number of residents in China’s cities had morethan doubled, from 191 million in 1980 to an estimated 562 million in2005 (see tables 7.1 and 7.3), and their per capita income increased evenmore rapidly. As a result, between 1980 and 2002, urban residents’ shareof total water use quadrupled to almost 6 percent, with urban waterconsumption increasing from 7 billion to 32 billion cubic meters (seetable 7.3). This increase reflects the rising standard of living in urbanareas, which allowed urban residents to purchase washing machines andmove into apartments with flush toilets and individual showers.11 Urbanareas experienced the largest increase in water use of any sector in the pasttwo decades. The increase was accompanied by the rising discharge ofblack, yellow, and grey waters.

Per capita water use in cities varies greatly by region. Annual domesticdemand in Beijing rose from 552 million cubic meters a year in 1993 to829 million cubic meters in 2000. In contrast, in Tianjin, in the dry HaiRiver basin, residents still use only 135 liters of water a day—less than40 percent of the 339 liters a day used by residents in the wet urbanareas in the southern province of Guangdong (USDA 2000).

166 China Urbanizes: Consequences, Strategies, and Policies

10 The growth of industrial water use in China is commensurate with its stage of develop-ment: water withdrawals for industry average 59 percent of total water use in high-incomecountries and just 8 percent of total water use in low-income countries (UNESCO 2003).

11 Domestic household consumption per capita rose tenfold in the past five decades, to240 liters a day per person in 2000 (University of British Columbia 2004).

Future Demand ProjectionsAgriculture remains the largest consumer of water in China, but growth indemand has been greatest in urban and industrial use. In 2005, China con-sumed about 563 billon cubic meters of water, of which 64 percent wasused for agriculture, 23 percent for industry, and 12 percent for householdpurposes. Demand for water grew at an annual rate of 7.3 percent forurban households and 4.3 percent for factories between 1980 and2002, with water demand by rural households and farms remainingalmost unchanged.

In the absence of a detailed and calibrated China-wide simulationmodel, it is difficult to analyze the implications of different scenarios forwater demand. However, some aspects of future water demand can beanalyzed with the aid of a simple simulation model (Shalizi 2006). Themodel is used as a broad-brush illustrative exercise to understand the keydrivers of water demand. It cannot be used to identify location-specificpolicy priorities or planning targets.

The key variables in the model are population forecasts and changesin (a) the urban–rural composition of the population; (b) per capitawater demand by rural and urban households; and (c) the compositionof production by primary (agriculture), secondary (manufacturing), andtertiary (services) activities. These variables can be used to project thesensitivity of the aggregate demand for water to different average GDPgrowth rates through 2050, assuming that water demand per yuan ofoutput in the various subsectors does not change significantly. The pro-jections provide a backdrop for comparing actions to aggressivelyincrease the efficiency of water use in subsectors versus actions toaggressively increase water supplies, recognizing that the compositionof the portfolio of feasable actions will vary by region, river basin, andeven locality.

The projections use two scenarios for population in 2050. The first isthe United Nations’ medium-term projection, which assumes a growthrate of 0.2 percent a year, with the population peaking and then levelingoff at 1.4 billion people by 2050. The second is the figure of 1.6 billionin 2050, which was used in the World Bank’s water strategy for NorthChina (World Bank 2001a) and implicitly assumes a 0.45 percent annualgrowth rate. The model assumes the same urban–rural split and GDPstructure for 2050 used in the water strategy study of the World Bank.It also assumes that demand for water of rural and urban households willcontinue to grow at either a slow rate (similar to that during 1997–2002)or a fast rate (similar to that during 1980–97).

Water and Urbanization 167

Using these very simple assumptions, the model projects that a dou-bling of the population growth rate from 0.2 to 0.45 percent a year has anegligible impact on water demand. The increase in the urban share of thepopulation, however, is more significant, particularly if average demand forwater by urban households continues to grow rapidly. These changes indemand—not excessive in their own right—will be more difficult toaccommodate if there is not also a substantial deceleration, and possiblyeven a decline in water demand by agriculture and industry.

Even though the share of agriculture and industry in GDP is decreasing,if per unit water consumption patterns in the production sectors do notchange significantly, water shortages will continue to grow, constrainingthe economy’s ability to grow at an average rate of more than 5 percenta year over the next 50 years. Even a 4 percent annual growth ratethrough 2050 could generate the need to more than double water supplyin many areas.

To put this required increase in water in perspective, aggregate waterdemand grew only 27 percent during the explosive growth period of1980 to 2005, rising from 444 billion to 563 billion cubic meters. Thisincrease put acute strain on supplies in the 3-H river basin. Key metro-politan regions began experiencing so much water shortage that theyhad to resort to diverting water from downstream rural users and estuar-ies and to pumping aquifers at a rate faster than replenishment, a strategythat is unsustainable.

Policy and Investment Implications

Generalizations and “one size fits all” recommendations are likely to beinappropriate in China, because of its size and complexity. But solutionstailored to location-specific problems are difficult to summarize andtedious to enumerate. Moreover, some of the information necessary toevaluate proposed solutions is not easily obtainable in public documentsor consistent across sources (in part because information sources vary intheir definitions and coverage and are rarely complete).12

Many of the problems cited in this chapter are well known to Chineseauthorities, who have initiated a wide range of programs to cope with

168 China Urbanizes: Consequences, Strategies, and Policies

12 Many instances of water scarcity are highly localized and are not reflected in national sta-tistics. In addition, the accuracy and reliability of information vary greatly across subna-tional regions and categories of information, as does the year in which the informationwas gathered. As a result, establishing consistency between different variables within andacross time periods is difficult. All data should therefore be considered as estimates.

them.13 In addition to the comprehensive overview of China’s water needscompleted by the Ministry of Water Resources in 2002, the World Bankprepared a strategy document in 2002 that outlines key actions (WorldBank 2002a). Neither of these documents provides a quantitative assess-ment of how much of the various problems will be resolved by the actionsproposed, and neither fully costs or sequences the actions.The documentsnevertheless provide an excellent array of actions to be implemented.

China has been very successful in investing in physical infrastructure tocontrol flooding, restore forested watersheds, and improve water supplyand wastewater treatment. It has been far less successful in managingdemand through better pricing and conservation policies, or in achievingbetter institutional coordination of integrated water management pro-grams at different jurisdictional levels, although there have been somesuccesses that have not yet been generalized (as in the Tarim River basin[World Bank 2004]). Where water is not being used efficiently, expandingwater supply at increasing marginal costs will only increase the drain onpublic resources. Such a strategy is also not sustainable.

Expanding the role of markets and market price signals as a feedbackmechanism in allocating water would go a long way toward helping con-serve the resource and allocate it to the highest economic use; doing sowould also send signals on priority investment requirements. Expandingthe role of water markets and prices is a corollary of expanding the role ofmarkets in the production of private goods and services (Yaozhou 2000).However, the expansion of water markets and prices presupposes progressin establishing the institutional framework for water rights/entitlements,valuation, and appropriate measurement, efforts that are still incomplete.14

Water and Urbanization 169

13 The Water Resources Report (MWR 2003) summarizes the implementation status ofthese programs.

14 Tradable water rights are one potentially important route to improving institutions forthe allocation and use of water. This calls for designing and implementing mechanismsthat will facilitate the functioning of a system of tradable water rights. As noted in arecent World Bank report (2004: 3–4). “Such a system would make a major contribu-tion to increasing the value of production per unit of water consumed in irrigated agri-culture areas and to the reallocation of water from agriculture to priority uses. Asignificant amount of informal water trading already goes on in China. Chinese waterlaw includes provisions for the issuing of water licenses, but the issuing and enforcingof water licenses in irrigation areas is not widespread. At each point of water measure-ment, a corresponding water right should be issued that includes a flow rate, a totalvolume of allowable annual delivery per extraction, and a total volume of allowableannual consumptive use. The sum of all of the consumptive use rights for a river basinor aquifer should not exceed the allowable total consumptive use in the basin oraquifer in order to have sustainable water resources use and management. Once the

Imposing taxes and subsidies, as well as educating farmers, firm managers,and households in water conservation options, will be required to augmentthe role of water pricing where market prices provides insufficient infor-mation and incentives for the correct allocation of water.

Demand-management strategies, including conservation measures, areessential not just to reduce water wastage but also to reduce the need forcostly interbasin water transfers. With some important exceptions,encouraging urbanization (and new infrastructure development) in areasthat are not currently water scarce or likely to become water scarce maybe an efficient long-term strategy. There was a dramatic demographicshift from rural to urban areas between 1980 and 2005. Net-rural-tourban migration was about 310 million people in the 25-year period1980 and 2005. As a result the urban share of the national populationincreased from 19 to 43 percent (see table 7.1).15 This shift was associatedwith a quadrupling of the urban share in total water demand. However,during this 25-year period, there was a negligible net demographic shiftfrom north to south of approximately 20 million migrants.16 As a result,the relative shares of the two zones remained almost constant, at 52 per-cent for the north and 48 percent for the south, despite growing waterscarcity in the north.17 This anomaly requires further analysis, but lackof adequate price signals on the real economic costs of water could beone factor.

Allowing or encouraging continued urbanization also requires that thecollateral damage associated with expanding urban water demand be

170 China Urbanizes: Consequences, Strategies, and Policies

consumptive use water rights issued equal the allowable total consumptive use rights ina basin, no further water rights should be issued. No water diversion or well drilling andpumping should be permitted without a corresponding water right. Once a complete sys-tem of water rights per measurement is operational, then a system of tradable waterrights could begin to function. To ensure maintaining a water balance and no negativeimpacts on third parties, the consumptive use right is the right that should be traded. Allwater rights trades should be registered and approved by the government authoritiesensuring no effect on third parties.Without a complete system of consumptive use waterrights and measurement, tradable water rights will not be able to aid in the reallocationof water within a sustainable water resources management system; therefore, trading inwater rights should be restricted.”

15 If the urban share of the national population had been the same in 2005 as it was in1980, the urban population would have been 310 million less than it was.

16 If the northern share of China’s population had been the same in 2005 as it was in1980, the population in the north would have been 22 million less than it was.

17 One caveat is the possibility that official data underestimate the extent of migration ofthe population from the north to the south. This problem is analogous to the difficultyof measuring the population in urban areas, which include unregistered (non–hukou)migrants, as noted in chapter 3.

reduced. Using fresh water would be less consequential if water abstractedupstream could be returned to river flows in good condition to be usedagain downstream. This can be done only if water polluted throughurban and industrial use is treated appropriately first, which will requiresignificant new investments. For example, despite its 1,179 operationalindustrial wastewater primary treatment plants, which have the capacityto treat 1.13 billion cubic meters of industrial wastewater, Chongqing, acenter for heavy industry, treats only about 57 percent of its industrialwastewater and 54 percent of its household wastewater (Okadera,Watanabe, and Xu 2006). Because Chongqing is situated upstream ofthe Changjiang River, the polluted water flows through the Three GorgesDam to urban and rural areas downstream.An additional US$122 millionof investment is required to fully treat the water in Chongqing (Okadera,Watanabe, and Xu 2006).

Overall, about half of China’s urban wastewater is treated. Even inBeijing, only 50 percent of wastewater is treated, despite the increase inwastewater treatment capacity from 50 million tons in 1990 to 517 mil-lion tons in 2003; the goal is to treat 70 percent by 2010 (Yang andAbbaspour 2007). Some 278 cities have no treatment facilities (“Drip,Drip, Drip” 2006). Thus, in addition to higher water prices in urban areas,charges for wastewater discharges and other pollutants must rise. Bettermonitoring, information disclosure, and enforcement of appropriate stan-dards are also needed.

Pollution from urban municipal sources could be managed by the useof updated municipal sewerage systems, including collecting sewers andtreatment plants designed to receive industrial wastewater. This com-bined use would yield large savings, both to the municipality and toindustries, because of economies of scale in removing degradable organics(with the provision that participating industries would first remove toxicand other harmful substances using in-plant treatment before discharginginto the municipal system). To the extent practicable, the treated munici-pal effluent would be reused as water supply for irrigation and industry.Municipal systems would also include provisions for effective use of on-siteexcreta disposal units for homes and buildings not connected to municipalsewers (the same provision would apply to rural homes), so that thesewastes are not left unmanaged and subject to being flushed into waterwaysby surface runoff.

Public disclosure of information on water quality and community con-sultation could improve feedback and facilitate better monitoring.Research by Jiangsu Province, SEPA, and the World Bank on pilot versions

Water and Urbanization 171

172 China Urbanizes: Consequences, Strategies, and Policies

of community consultation and feedback processes, as well as public dis-closure of information (Wang and others 2004), have determined thatthey are effective and have the potential to be scaled up (Lu and others2003; Jiangsu Environmental Protection Bureau 2007).

Reducing the Cost of Wastewater Treatment and Improving Its MonitoringIncreasing wastewater charges (specially the rate by which charges increase)will be easier if there is better monitoring and information disclosureand if more wastewater is treated for reuse. Increasing the amount ofwastewater treated for reuse requires that increases in wastewater chargesbe complemented by declines in the cost of investment in wastewatertreatment plants.18

More than 1,000 wastewater plants were built between 2000 and2006 (“Strong Growth” 2007), but the utilization rate is only 60 percent.About 50 plants in 30 cities are operating at below 30 percent capacity,and some are left idle, mainly because of inadequate wastewater collectionfacilities (“Strong Growth” 2007;Yang and Abbaspour 2007) and becauserevenues collected from customers are transferred to the general citybudget and not used to ensure that treatment plants have the resourcesneeded to operate.19

Operational efficiency is also low, mainly because plants carry out onlyprimary treatment. Even in Shanghai the efficiency is only 10–30 percent(Okadera, Watanabe, and Xu 2006). Moreover, investment coordinationacross metropolitan regions is inadequate. There are economies of scaleand optimal sizes for wastewater treatment plants. Despite this approach,many small, adjacent municipalities respond to national directives by imple-menting their own suboptimal wastewater treatment plants, increasingthe overall national costs of wastewater treatment.

Widening the options for wastewater investment decisions can helpcontain costs. As noted in the World Development Report 2003 (WorldBank 2002b), New York City found it cheaper to repurchase land alongpart of its watershed that had been sold for development than to build

18 At the current low rate for wastewater collection, private sector participation is diffi-cult to imagine. Until these rates are increased, the public sector will have to shoulderthe needed investment costs associated with increasing wastewater treatment capacityto reduce water pollution in urban areas. One estimate puts the figure as US$30 billionbetween 2006 and 2010 (Zhong, Wang, and Chen 2006).

19 This information is based on informal communications with the author of an ongoing“City Development Strategy” in 11 cities in China.

Water and Urbanization 173

an expensive water treatment plant. Doing so was possible because thetopography provided for a natural filtration process that was very effec-tive. This option will not be an appropriate in all cases, but it may be insome cases. More important, widening the array of options to bereviewed and evaluated, when undertaking cost-benefit or feasibilitystudies, may help identify new cost-effective solutions. Restricting landdevelopment and restoring local watershed filtration capacities mayreduce the amount of water requiring treatment, thereby reducing thecost of wastewater treatment in China.

Augmenting the Water SupplyDemand-management policies through market prices or new institutionalarrangements may not be sufficient to deal with the full range of envi-ronmental water problems facing China. It is easier to introduce watercharges when scarcity has emerged but critical thresholds have not yetbeen crossed. It is more difficult to introduce water pricing and waste-water charges when implicit rights to subsidized or free water resourceshave been acquired and critical thresholds crossed.

Two such thresholds are important in China. The first is the lack of ade-quate water in rivers to ensure year-round flows to flush the rivers, trans-port silt to the delta, and avoid ecosystem damage downstream, all of whichare already occurring. The second is the need to restore some, if not all, ofthe groundwater that has been overpumped in the recent past. In bothcases, more water flow has to be restored to the ecosystem and aquifers.20

It may be possible to restore water flows to natural systems by settingthe price so high that existing users voluntarily renounce some of theirwater claims. But in the absence of water markets and well-defined waterrights, it may not be possible to fully restore the minimum requisite waterabstracted from rivers that now run dry. More important, ecosystemneeds are public goods and by definition difficult to include in water mar-kets (when flows in rivers are low, for example, water for environmental

20 In the case of aquifers, one study (Gunaratnam 2004) provides a clear set of actions tobe implemented: “The key actions required by the action plan are (a) definition ofgroundwater management units with determination of sustainable yields; (b) preparationof groundwater management plans; (c) allocation of licensing linked to sustainable yieldand undertaken by one department only; (d) licensing of well construction drillers; (e)development of a national groundwater database; and (f) preparation and implementa-tion of a groundwater pollution control strategy, including provision in selected cities forrecharging of groundwater by spreading of treated wastewater effluents or of floodwaterson permeable spreading areas, and for the injection of treated effluents to establishgroundwater mounds to prevent salinity intrusion into freshwater aquifers.”

174 China Urbanizes: Consequences, Strategies, and Policies

needs must come from a reduction in irrigation). Ensuring that publicgoods, such as ecosystem water requirements, are handled appropriatelywill require institutional reforms, as noted earlier.

Investment in catchment reservoirs and watershed management incatchment basins through reforestation, to stabilize damaging uneven-ness in water flows over time, are other examples of supply-orientedinterventions that need to complement demand-management policies,particularly where externalities and public goods are involved. For thefuture, some augmenting of existing supplies—through interbasin trans-fers and the reuse of wastewater, for example—will be required.21 Onesuch interbasin transfer is the South-to-North Water Transfer Project.22

This scheme entails transferring 19 billion cubic meters a year initiallyand eventually up to 45 billion cubic meters of water a year from theYangtze River, at a cost of US$60 billion. The crucial component of theproject, including that which supplies Beijing, was completed in July2006; the entire project will not be completed until 2050 (“China: AFive-Year Outlook” 2004; “Still Poor” 2006; Wu 2006).23 Depreciat-ing these costs and adding operating costs will likely require prices wellin excess of those currently prevailing of less than one yuan per cubicmeter. (World Bank 2001b) However, even with the south–north trans-fer, water use in irrigated agriculture in the 3–H basins will need to bereduced by 20–28 billion cubic meters from current levels by 2020.

Gunaratnam (2004) points out that in industrial countries, treatedmunicipal wastewater represents a very valuable source of supplementalwater for the industrial raw water supply and for irrigation of farming andurban green zones. The use of treated municipal wastewater for urbangreen zones may even be the preferred use, because of the lower qualityrequirements and relatively low infrastructure costs (Yang and Abbaspour2007).24 Hence, future plans for meeting urban needs should incorporateprovisions for municipal sewerage systems to facilitate reuse. Using theprice of fresh water in 2003 as an opportunity cost, the net economicbenefit of reusing wastewater in Beijing is Y 134–Y 298 billion a year

21 Interest in reusing water is not new. See the case study of water reuse project inChangzi city by Peng, Stevens, and Yiang (1995).

22 The concept of south–north water transfer was first aired by Mao Zedong, in 1952.Three channels (western, middle, and eastern routes) with a total length of 1,300 kilo-meters will link four major rivers of China: the Yangtze, Yellow, Huaihe, and Haihe(“China: Moving Water” 2003). See Gao and others (2006) and Wu and others (2006)for details on the western and middle routes of the South-to-North Transfer Project.

23 Initially, the capital costs of the south–north transfer of 18 billion cubic meters wasestimated to be Y 245 billion (World Bank 2001b).

24 Wastewater can also be used to replenish the groundwater (Yang and Abbaspour 2007).

Water and Urbanization 175

(19–43 percent of Beijing’s GDP in 2005). Beijing now requires new res-idential buildings with construction areas larger than 30,000 squaremeters to have on-site wastewater reuse facilities (Yang and Abbaspour2007).25 Such planned reuse would have to be subject to regulatory con-trol through permits to ensure that public health needs are protected.

Conclusions

China’s rapid urbanization increases the urgency of decisively tacklingthe growing scarcity of water, a constraint that can only tighten as theclimate warms and glaciers feeding the major river systems and aquifersdisappear in the coming decades. Urban development and the geographyof urbanization will need to be coordinated with policies aimed at man-aging demand in urban as well as rural sectors and by measures to aug-ment or recycle the usable supply of water.

Despite growing water scarcity in the north, there has been no notice-able demographic shift from north to south. This anomaly requiresfurther analysis; lack of adequate price signals on the real economic costsof water could be one factor. With some important exceptions, encour-aging urbanization (and new infrastructure development) in areas thatare not currently water scarce, or likely to become water scarce, may bean efficient long-term strategy.26

Among the actions discussed in this chapter, four stand out as mostsignificant from the perspective of urban development:

• Allocate water for public uses (such as estimated ecosystem waterneeds) first, before allocating it to private uses (industry, residential, andagriculture as residual claimants), through either markets or administra-tive arrangements.27 In either case, water use must be regulated to pro-tect public health and the environment.

• Shift from administrative to price-based allocation of water, initiallythrough better technocratic analysis, eventually complemented throughwater markets based on the fair and transparent allocation of propertyrights in water. Water markets cannot totally replace administrative

25 Onsite treatment is often economically unviable; treatment should be centralized(Yang and Abbaspour 2007).

26 This is a broader strategy than the informally discussed possibility of moving China’scapital out of Beijing to a less water-stressed area.

27 The South African Water Act of 1997 considers its water resources as a public good, aresource for all under state control and licensed (http: www.thewaterpage.com/SolanesDublin.html). China’s water laws have been revised recently to address some of thepublic goods issues raised above (Xiangyang 2004).

176 China Urbanizes: Consequences, Strategies, and Policies

(quota) allocations of water, for reasons that are well known and implic-it in the previous recommendation regarding public/private use of wa-ter. The balance between the two allocation mechanisms (administra-tive versus market) will be determined politically, though goodtechnocratic analysis can inform the political debate.

• Improve the institutions involved in water management, not only atthe metropolitan level but also at the river-basin level, includingthrough better coordination of water use through invigorated riverbasin/watershed management commissions, and greater involvementof communities in joint monitoring and enforcement through publicdisclosure schemes.

• Increase urban water recycling through more reliable and cost-effectivewastewater and sewerage treatment, and more appropriate sewerageand wastewater charges.

The greatest challenge is for national and local/urban governmentsto craft policies and rules within China’s complex cultural and legaladministrative system that provide incentives for users to increase effi-ciency of water use and for polluters to clean up the water they use andreturn clean water to stream flows. Using a standard public economicsframework, water requirements for public goods, such as ecosystemneeds, should be set aside first, before allocating property rights inwater (to enable water markets to function and generate efficient allo-cation signals). Even then, water markets will have to be regulated toensure that public goods, such as public health, are not compromised.Until water markets are implemented, staying the course on increasingthe water and sewerage or the wastewater prices administratively andencouraging water conservation is necessary to reduce the wasting ofcurrently scarce water resources as well as the new water supplies to beprovided in the future. Investments in supplying water for rapidly growingurban areas and treating urban sewerage and wastewater will be moreeffective when combined with more-vigorous demand-managementpolicies and institutional reform.

References

ABS Energy Research. 2006. Water and Waste Utilities of the World. London:

Bao, Chao, and Chuang-Lin Fang. 2007. “Water Resource Constraint Force onUrbanization in Water Deficient Regions: A Case Study of the Hexi Corridor,Arid Area of NW China.” Ecological Economics 62 (3–4): 508–17.

Water and Urbanization 177

“China: A Five-Year Outlook.” 2004. Oxford Analytica, March 31.

“China: Moving Water and People.” 2003. Business China, March 31.

“China to Keep Grains Goals despite Climate Change.” 2007. Reuters News, June 19.

“China Turns to Desalination to Keep Taps Flowing.” 2007. Straits Times, June 14.http://app.mfa.gov.sg/pr/read_content.asp?View,7452.

Chinese Academy of Science. 2003. China Country Report on SustainableDevelopment: Water Resources. Committee of Sustainable DevelopmentStrategy, Beijing.

“Drip, Drip, Drip.” 2006. Business China, September 27.

“Drought Hits Northern China” 2007. AP Newswires. June 19.

“Drought leaves 1.2 million short of drinking water in NE China.” 2007.Xinhua News. July 19. http//www.chinadaily.com.cn/china/2007-06/19/content_897742.htm.

EIA (Energy Information Administration). 2003. “Country Analysis Briefs.” U.S.Department of Energy, Washington, DC.

Economy, Elizabeth. 1997. “The Case Study of China.” http//www.library.utoronto.ca/pcs/state/china/chinasum.htm.

Economy, Elizabeth. 2004. The River Runs Black. Ithaca, NY: Cornell UniversityPress.

FAO (Food and Agricultural Organization). 2007. “AQUASTAT.” Rome.http://www.fao.org/nr/water/aquastat/dbase/index.stm.

Gao, Yanchun, Changming Liu, Shaofeng Jia, Jun Xia, Jingjie Yu, and Kurt PengLiu. 2006. “Integrated Assessment of Water Resources Potential in the NorthChina Region: Developmental Trends and Challenges.” Water International31 (1): 71–80.

Gong, Huili, Menlou Li, and Xinli Hu. 2000. “Management of Groundwater inZhengzhou City, China.” Water Research 34 (1): 57–62.

Guan, Dabo, and Klaus Hubacek. 2007. “Assessment of Regional Trade and VirtualWater Flows in China.” Ecological Economics 61 (1): 159–70.

Gunaratnam, Daniel. 2004. “China Water Resources Issues and Strategy.” WorldBank,Washington, DC.

IIASA (International Institute for Applied Systems Analysis). 1993. IIASA databases.http://www.iiasa.ac.at/Research/LUC/ChinaFood/data/water/wat_8.htm,http://www.iiasa.ac.at/Research/LUC/ChinaFood/data/water/wat_11.htm,http://www.iiasa.ac.at/Research/LUC/ChinaFood/data/water/wat_7.htm.

Jiangsu Environmental Protection Bureau. 2007. “Jiangsu Carried Out RoundTable Dialogue Pilot Project.” January. http://www.jshb.gov.cn/jshb/xwdt/showinfo.aspx?infoid=5e25a4db-df0b-474b-a34a-2e5fa3ee901e&categoryNum=002001&siteid=1.

178 China Urbanizes: Consequences, Strategies, and Policies

MacBean, Alasdair. 2007. “China’s Environment: Problems and Policies.” WorldEconomy 30 (2): 292–307.

MWR (Ministry of Water Resources). 2002. China Water Resources StatisticsReport Beijing. http//gks.chinawater.net.cn/cwsnet/CWSArticle_View.asp?CWSNewsID=14983.

MWR (Ministry of Water Resources). 2003. Water Resources Report. Bejjing:

NBS (National Bureau of Statistics). Various years. China Statistical Yearbook.Beijing: China Statistics Press.

Lu, Genfa, Yuan Wang, Yu Qian, Jun Bi, and Gangxi Xu. 2003. “Application ofCorporation Environmental Performance Disclosure in EnvironmentalManagement.” Chinese Practice. Vol. 1, No 1, 2003, March. Available at:http://www.chinaenvironment.net/pace/perspectives/print.php?id=7_0_6_0

Okadera, Tomohiro, Masataka Watanabe, and Kaiqin Xu. 2006. “Analysis of WaterDemand and Water Pollutant Discharge Using a Regional Input-OutputTable: An Application to the City of Chongqing, Upstream of the ThreeGorges Dam in China.” Ecological Economics 58 (2): 221–37.

Palmer, Mervin D. 2001. Water Quality Modeling. World Bank, Washington, DC.

Peng, Jian, David K. Stevens, and Xinguo Yiang. 1995. “A Pioneer Project ofWastewater Reuse in China.” Water Research 29 (1): 357–63.

SEPA (State Environmental Protection Agency). 2003. State of the Environment inChina. Beijing:

Shalizi, Zmarak. 2006. “Addressing China’s Growing Water Shortages andAssociated Social and Environmental Consequences.” Policy ResearchWorking Paper 3895, World Bank, Washington, DC.

———. 2007. “Energy and Emissions: Local and Global Effects of the GiantsRise.” In Dancing With Giants: China, India, and the Global Economy.Singapore: World Bank and Institute of Policy Studies.

“Still Poor and Now Thirsty, Too.” 2006. Business China, September 11.

“Strong Growth Seen in China’s Water Sector.” 2007. Business Times Singapore,June 18.

“The Case of China.” 1997.

UNESCO (United Nations Educational, Scientific, and Cultural Organization.2003. Water for People, Water for Life. http://unesdoc.unesco.org/images/0012/001295/129556e.pdf.

University of British Columbia. 2004. “Introduction to Beijing-Tianjin.” Center forHuman Settlement, Vancouver. http://www.chs.ubc.ca/china/introbeijing.htm.

USDA (U.S. Department of Agriculture). 2000. Agricultural Outlook January–February 2000. Washington, DC: Economic Research Service.

Wang, Hua. 2004. “China Water Issues.” World Bank, Development ResearchGroup, Washington, DC.

Water and Urbanization 179

Wang, H., J. Bi, D. Wheeler, J. Wang, D. Cao, G. Lu, and Y. Wang. 2004.“Environmental Performance Rating and Disclosure: China’s GreenwatchProgram.” Journal of Environmental Management 71 (2): 123–33.

Wang, Hua, and Somik Lall. 2002. “Valuing Water for Chinese Industries.” AppliedEconomics 34 (6):759–65.

Water Resources and Electric Power. 1997. Use of Water Resources in China. Beijing:

Woodrow Wilson Institute. 2004. “Water Conflict Resolution in China.” Summaryof Environmental Change and Security Program, January 28,Washington, DC:http://www.wilsoncenter.org.

World Bank. 1997. Clear Water, Blue Skies. New York: Oxford University Press.

———. 2001a. Agenda for Water Sector Strategy for North China, vol. 1.Washington,DC: East Asia and Pacific Region, World Bank.

———. 2001b. Agenda for Water Sector Strategy for North China, vol. 2.Washington, DC: East Asia and Pacific Region, World Bank.

———. 2001c. China Air, Land, and Water Environment Priorities for a NewMillennium. Washington, DC: East Asia and Pacific Region, World Bank.

———. 2002a. China: Water Resources Assistance Strategy. Washington, DC: EastAsia and Pacific Region, World Bank,

———. 2002b. World Development Report 2003. Washington, DC: World Bank.

———. 2004. “11th Plan Water Note.” World Bank, Washington, DC.

———. 2007. The Little Green Data Book. Washington, DC.

Wu, Xianfeng, Changming Liu, Guilian Yang, and Qing Fu. 2006. “AvailableQuantity of Transferable Water and Risk Analysis: Western Route Project forSouth-to-North Water Transfer in China.” Water International 31 (1): 81–86.

Xiangyang, Xiong. 2004. “The Revision of Water Law of P.R.C. and the Efforts forRemedying the Dry-Up Problems of Yellow River.” Beijing: Ministry of WaterResources.

Yang, Hong, and Karim C. Abbaspour. 2007. “Analysis of Wastewater ReusePotential in Beijing.” Desalination 212 (1–3): 238–50.

Yaozhou, Zhou, and Wei Bingcai. 2000. “Pricing Irrigation Water in China.” InPricing Irrigation Water, ed. Yacov Tsur. Washington, DC: Resources for theFuture.

Yusuf, Shahid, and Kaoru Nabeshima. 2006. “China’s Development Priorities.”World Bank, Development Research Group. Washington, DC.

Zhong, L., X. Wang, and J. Chen. 2006. “Private Participation in China’sWastewater Service under the Constraint of Charge Rate Reform.” WaterScience & Technology: Water Supply 6 (5): 77–83.

Autos, Transit and Bicycles: Comparing the Costs in Large Chinese Cities

Rui WangAssistant Professor, UCLA School of Public Affairs

3250 Public Affairs BuildingBox 951656Los Angeles, CA 90095

Tel: (310) 367-3738Fax: (310) 206-5566ruiwang@ucla.edu

AbstractThe study compares the full costs of major passenger modes in typical commuting corridors of a representative large Chinese city. The investment and operating plans for each mode have been designed to reduce the full cost using optimization techniques. The results show that in most circumstances commuting by bus and bicycle has lower social costs than by automobile or rail. Unlike results from similar studies conducted in the United States, automobile transportation does not cost less than bus transportation at low traffic volumes.

KeywordsChinese urban transport; alternative assessment; full social cost comparison; automobile; transit; bicycle

Funding SourceThe Harvard China Project, the Taubman Center and the Asia Programs at the Harvard Kennedy School, and the Harvard Graduate School of Arts and Sciences provided partial funds for this research.

Total number of words: 5,990 excluding title page.

July11, 2009

ManuscriptClick here to view linked References

1

1. Introduction

China’s rapid economic growth, urbanization and motorization have placed great strain on its

urban transportation system. Urban road systems and automobile ownership have been growing

so quickly that many are concerned that automobile-oriented development is unsustainable for

the country. Despite the high cost, urban mass rail transit is now often viewed by mayors and

planners of many large Chinese cities as the answer to urban transportation problems. At the

same time, services by less expensive bus transit are often inadequate to meet the demand of

travel, and the future of bicycle transportation is threatened by losses of street space and safety

caused by motorization. Urban transport planning and investment decisions are very important to

China and the world in that they affect efficiency and sustainability of the Chinese cities,

representing a big and growing share of the world’s population and economy. Unfortunately,

there has not been a systematic study of the social consequences of major contemporary

transportation technologies and systems in the Chinese cities.

This study assesses these alternatives by comparing total costs of major urban commuting

modes in a representative large Chinese city. The work builds on a series of studies developed

for the Western industrialized cities following the seminal work of Meyer, Kain and Wohl

(1965). Their research compared the capital and operating costs of rail, bus and automobile for

commuting travel in typical U.S. urban corridors. They found that the relative costs of these

modes depend on traffic volumes in the corridor with automobile being the least-cost mode in

corridors with low volumes, bus the least cost in corridors with intermediate or higher volumes,

and rail sometimes the least cost in corridors with high volumes. Their work was followed by a

number of others. For example, Boyd, Asher and Wetzler (1973, 1978) added traveler time cost

into the comparison between rail and bus, while Keeler and Small (1975) conducted a full-cost

2

comparison of rail, bus and auto with optimized modal investments and operations. Recent

studies on intermodal comparison, such as Pickrell (1990), Kain (1997) and GAO (2001), focus

more on actual vs. forecasted system performance rather than abstract comparison. In general,

previous studies agree that socially optimal modal planning depends on corridor volume, but the

exact volume thresholds between auto, bus and rail remain controversial.

This study extends the existing literature in three respects. First, it provides benchmark

cost estimation in the specific socio-economic and spatial context of the Chinese cities. In

addition to radial corridors, circumferential corridors are modeled for the first time. Second, the

present study includes a wider range of modes that were not considered previously, most notably

the bicycle, since it plays an important role in Chinese cities, but also the bus on arterial streets,

bus rapid transit (BRT) and light rail transit (LRT). This study also includes more categories of

costs, including those from noise and global warming. Third, this study adopts a more

sophisticated methodology than previous ones by optimizing vehicle size, train length and the

capacity of the highways all along the radial corridors.

Section 2 describes the analytical framework from three aspects: the representative large

Chinese city and corridors, characteristics of the alternative modes including optimal investment

and operation of the modes, and the allocation of fixed costs. Data sources and intermediary

estimates are discussed in Section 3. Section 4 presents the results of intermodal comparison,

including sensitivity analyses. Section 5 concludes with policy implications and discussions on

the difference in results between China and the U.S. and the limitations of this study.

2. Methodology

This study compares major commuting modes in a representative large Chinese city by

estimating their capital, operation, user time, safety and environmental costs. However, cost

3

variations due to service reliability and levels of comfort and convenience, as reflected by the

alternative-specific constants (ASCs) in discrete demand analysis, are difficult to measure. Due

to the lack of good estimates of ACSs from Chinese data and the well-known difficulty to

transfer estimates of ACSs (see, e.g. Small and Winston 1999), this study does not explicitly

account for such cost variations. Instead, this study approximates the full social cost comparison

through offsetting intermodal differences in comfort level and accounting for intermodal

differences in convenience, as described in Section 2.2.1

2.1. The representative city and commuting corridors

Large regional center cities with populations of approximately 1.5-3.5 million (there are about 20

of them in 2004) are of particular interest to this study because first, they are among the fastest

growing cities in terms of population and motorization; second, their rapid growth is forcing

them to invest massively in their transportation systems; and third, the feasible alternatives they

face are the most diverse and thus difficult to assess.

Besides high population density (see, e.g., the international comparison by Bertaud and

Malpezzi 2003), the Chinese cities share certain characteristics different from typical Western

industrialized cities due to their history and institutional background. These include flatter

density curve and less segregated land uses due to the socialist legacy (see, e.g., Gaubatz 1995

and Wang 2009), and distinct density and public service gap between urban areas and

surrounding rural areas due to the dual rural-urban economy and administrative system (see, e.g.,

1 Even if one could estimate the ASCs from real-world data in China, it might still be better not to use them in

intermodal comparison, as bizarre estimation of ASCs is a common outcome of the exclusive focus on statistical fit,

as we can see from the forecasting biases of transportation projects in the U.S. On the other hand, one can view this

study as a comparison of the measurable full social costs of the modes, which is nevertheless useful, especially when

one believes that the unmeasured costs are relatively small after measures described in Section 2.2 are taken.

4

Ho and Lin 2003). Although dominant mixed-use central city typifies the Chinese cities, limited

but growing amount of residential communities, often in the form of high-density “superblocks”,

have been growing at the fringe or outside of the central city, usually distributed along radial

corridors connecting to the central city (see, e.g., Monson 2008).

According to the unique spatial development pattern of the Chinese cities, the

hypothetical city in this study is a circular city developed on flat terrain with a mixed-use core

consisting of most jobs and a significant share of urban residents, surrounded by a suburban belt

consisting of residential areas developed along radial corridors, as shown in Figure 1.

(Figure 1)

Specific assumptions are made to the size of the city to give reasonable traffic volumes

and calculate costs precisely. The population of city is set at 2.36 million, which translates into a

total built-up area of 236 km2 and a central city radius of 5 km, given the typical Chinese urban

population density (also consistent with the national planning guideline) and assuming that on

average the area of central city is one third of the total built-up area. Figure 1 also depicts the

four typical commuting corridors, two radial and two circumferential. The radial corridors serve

suburban residents working within the central city, and the circumferential (“ring”) corridors

serve workers living in the central city. The two radial corridors examined are 10 and 20 km long

connecting the city center with suburban residents. The two circumferential corridors are an 15

km long inner ring approximately 2.4 km from the city center, and an 25 km long outer ring

about 4 km from the city center.

We assume that in the central city land use is completely mixed with jobs and residents

evenly distributed, while in the suburbs households are evenly distributed along radial corridors.

The average trip length on a radial corridor thus equals half of the corridor length, and the

5

inbound morning traffic volume along the corridor increases linearly from zero at its suburban

end to maximum at the boundary of central city, and then decreases linearly to zero at the city

center. The ring corridors have constant traffic volumes along the route during peak hours, and

the average trip length is assumed to be one fifth of the corridor length.

2.2. Characteristics and optimization of modal alternatives

Seven passenger modes are evaluated on each corridor, including (1) heavy rail transit (“metro”)

that uses electrically propelled multiple-unit trains operating on a fully grade-separated right of

way (ROW) that is underground within central city and elevated outside; (2) LRT that uses

electrically propelled multiple-unit trains operating on a ROW that is private but not fully grade

separated. Traffic signals give the LRT priority at intersections; (3) arterial bus (“bus”) that uses

diesel and operates in mixed motor vehicle traffic on urban arterial streets; (4) BRT that has

diesel buses operate at grade in a private ROW with signal priority at grade crossings similar to

that of LRT; (5) expressway bus flier (“flier”) that has diesel buses operate in mixed traffic on

limited-access urban expressway; (6) automobile (“auto”) that has gasoline cars use urban

expressways as trunk line; and (7) bicycle that has human-powered two-wheel vehicles operate

on urban arterial and local streets.

All transit modes drop off and/or pick up passengers over the entire route without short-

turn or express service. Because of the high density of the Chinese cities, commuters travel

between transit stations and their homes by either bicycling or walking, a choice determined by

comparing the total costs of the two access modes. Also because of the high job density in the

central city, it is assumed that all transit users walk between stations and their workplaces.

Two types of measures are taken to reduce the inequality of unmeasured costs (ASCs)

across the modes. First, vehicle specifications and loading levels are chosen so that the comfort

6

levels of the automobile and the transit modes are as close as possible. For example, all vehicles

are air-conditioned and about 70% of transit passengers have seats during peak hours (100% for

flier passengers). Table 1 summarizes the one-way maximum passenger capacities of all transit

modes based on realistic assumptions of vehicle capacity and frequency limits, making comfort

levels comparable across modes. Second, intermodal differences in convenience, such as the

walking distance within bus and rail stations, are quantified by detailed assumptions of typical

station size varying with capacity.

(Table 1)

Optimal investment and operation of the modes are modeled in a more realistic fashion

than the methods applied by the previous studies represented by Keeler and Small (1975). First,

transit vehicle size and train length are optimized within given ranges in addition to service

frequency, and a maximum service frequency is introduced to single-route operation of each

transit mode. Second, road widths at different locations of the radial corridors are optimized

according to the traffic volume of the respective section since the traffic volumes are not

constant along the radial corridors.

2.3. Allocation of fixed costs

For transit modes, capital costs are allocated to the peak-direction traffic during peak hours. This

follows the practice in MKW (1965) and Keeler and Small (1975), and is based on the argument

that the capacity of the transit system, and thus the required capital investment in transit plant

and equipment, are determined primarily by peak demand. An exception is the cost of road

capacity used by buses operating in mixed traffic on expressways or arterial streets. It is

calculated using the congestion cost caused by the buses.

7

For individual modes, as automobiles and bicycles are typically purchased for multiple

purposes and depreciate primarily with use rather than the passage of time, we allocate the

vehicle capital costs on the basis of the distance driven, assuming a fixed vehicle life in distance

driven and a typical distance driven per year. By contrast, automobile and bicycle parking spaces

depreciate little with use. The cost of the parking space at work is clearly attributable to

commuting, since it would not be needed unless the employee commuted by automobile or

bicycle. But the parking space at the residential end of the trip probably should be allocated to

either commuting or other uses depending on which type of trip was the primary or decisive

motive for owning an automobile or bicycle. Without knowing what the decisive motive is, we

compromise and allocate weekday residential parking costs to commuting trips, and weekend

parking costs to non-commuting trips. Sensitivity analyses will show how such assumption

affects results. Finally, road capacity costs of the automobile are calculated using congestion

costs, while those of the bicycle are allocated to major-direction peak travel, as the number of

bicycle lanes is determined by the peak demand.

3. Data

The different costs required are estimated from a variety of primary and secondary sources. Land

costs are calculated according to Wang (2009), which estimates the land price of Chinese cities

as a function of distance from urban center, population size and average per capita GDP based on

benchmark land prices published by the Chinese city governments. Capital and operating costs

are generally estimated based on actual costs for Chinese road and transit systems and

summarized in Tables 2 and 3. Environmental and safety costs are estimated largely by adjusting

estimates from the Western cities to account for higher densities and lower per capita incomes in

China, as calculated by Wang (2008) and summarized in Table 4.

8

(Tables 2,3 and 4)

Few empirical estimates of traveler’s value of time (VOT) in the Chinese cities exist, but

the international evidences on the relation between VOT and travelers’ income so far are largely

consistent. Following Small and Verhoef (2007), ALMEC and Chodai (2001) and others, we

estimate the VOT of riding in rail, bus and auto vehicles as 50% of a rider’s gross wage rate,

while walking and waiting time are valued 1.8 times of in-vehicle time. The typical urban gross

wage for large Chinese cities is roughly 12 Yuan per hour (2005 price). We also test the

sensitivity of results to a low of 30%, suggested by the World Bank (2007), and a high of 80%,

mentioned by Small and Verhoef (2007) as a recent evidence of higher VOT. There has been

very little research on the time value of commuting by bicycle. Given that bicycles are driven by

human power and that bicyclists, like pedestrians, are directly exposed to the ambient

environment, we assume per unit time spent riding bicycle has the same value as walking and

waiting.

Real social discount rates (SDRs) of 10-12% have been commonly used to evaluate

public projects in China over the past decade. A recent Chinese government study (Tongji and

RISN 2004), however, suggests that a lower rate of 8% might be more appropriate given the

prevailing saving rate and return on investment in China. This study uses 8% as the baseline

estimate of SDR. Alternative values are 4%, approximately the Chinese central bank benchmark

interest rate of a five-year deposit in 2005, and 12%, used in the past by the Chinese government

and the World Bank to evaluate public investments.

4. Results

Although the results for all four types of corridors will be discussed, the short radial corridor is

used as the primary example because of its increasing importance to commuters in large Chinese

9

cities and the fact that it is more comparable than the circumferential corridors to the ones

studied in the Western cities.

Different modes vary considerably in the composition of their costs, and these differences

affect intermodal cost comparison under different circumstances. Generally, capital costs are the

largest cost component followed by travel time costs. Automobile and rail modes are more

capital intensive than bus modes and bicycle, while the pattern reverses for time costs. Operating

costs are fairly modest, while safety and environmental costs are minimal for all modes except

the bicycle, as bicyclists are subject to significantly higher risk than other modes on a per-km

basis (Wang 2009). Of course, the exact composition of costs varies by corridor type, VOT and

SDR assumptions and particularly corridor volume, especially for modes like metro or BRT that

have heavy fixed capital investments in tunnels, viaduct or exclusive ROW.

(Figure 2)

Figure 2 simulates the cost curves of the seven modes for all four types of corridors with

demand volumes ranging from 1,000 to 50,000 pph at the one-way maximum-load point. Two

cost curves are shown for each of the automobile and bicycle modes to mark the difference of

allocating weekday home parking costs to commuting vs. to other purposes. The solid lines

represent cost estimates including home parking costs, while the dashed lines exclude those costs.

The costs per PKT decline with passenger volume for most modes due to indivisibilities

in capital investments. Such economies of traffic density are strongest for the modes with

exclusive structure and ROW, such as rail and BRT, and weakest for bicycle and arterial bus,

which need fewer fixed initial investments in ROW or equipment. The results show that auto is

never the least expensive mode regardless of corridor type or traffic volume. Among the transit

modes, rail is always more expensive than bus, especially at lower passenger volumes due to

10

large fixed capital investments. Among the three bus modes, arterial bus has the least cost at low

passenger volumes, while the BRT has the least cost at medium to high volumes. Considering all

seven modes, the least expensive is either bicycle or one of the bus modes, depending on

passenger volume and corridor type.

The circumferential corridors show results somewhat different from those of the radial

corridors. One important source of the differences is that the passenger loads carried are much

more balanced or even on the ring corridors than the radial corridors. On the ring corridors

passengers travel in both directions during both mornings and evenings, while on radial corridors

passengers travel inbound only in the morning and outbound only in the evening. In addition, on

ring corridors passengers can be assumed to get both on and off at uniform rates along the

corridor, while on radial corridors passengers board only in the suburbs and alight only in the

city core in the morning, with a reverse pattern in the evening. Such unbalanced loads increase

the costs of all modes on radial corridors, but they have stronger effect on the costs of rail and

bus than on the costs of the automobile or the bicycle because rail and bus are chauffeured

modes. The bicycle and automobile can start and stop where the passenger’s trip does, whereas

the railcar or bus often operates with many empty seats at the beginning or end of a radial

corridor and in the off-peak direction. As a result, the automobile is sometimes more cost-

effective than rail and BRT at very low volumes on radial corridors, but it is never less expensive

than rail or BRT on ring corridors. Similarly, bicycle is the least expensive mode at low to

medium volumes on radial but not on ring corridors.

To illustrate how the cost comparisons are affected by assumed values of SDR and VOT,

Figure 3 presents the results under three different SDRs (4%, 8% and 12%) and three different

11

VOTs (30%, 50% and 80% of wage rate for in-vehicle time, with a fixed ratio of 1.8 between

values of walking, waiting and bicycling time and in-vehicle time) for the short radial corridor.

(Figure 3)

As shown by the variation across the columns in Figure 3, the costs of the more capital

intensive modes, such as automobile and rail, are the most sensitive to changes in the SDR. The

cost curves of automobile and rail shift upward faster than those of the other modes when the

SDR increases from 4 to 12%, while the bicycle cost curve hardly moves at all. As a result, cost

advantage of the bicycle is more significant with high SDRs, while the automobile and rail

modes become more competitive relative to bicycle and bus under low SDRs. These changes are

not enough to affect the basic results of the analysis, however. Using the situation under the

baseline VOT (shown in the middle row) as example, a low discount rate reduces automobile and

rail costs but not by enough to overtake the bus or the bicycle. This applies to all types of

corridors, but the magnitudes of intermodal cost differences vary.

The results are less affected by VOT, as illustrated by the variation across rows. This

might be explained by the fact that modes with larger time cost components, such as bicycle and

bus modes, are typically also lower in total costs. The costs of all modes increase with VOT, but

VOT has the greatest effect on the costs of the slowest modes, principally bicycle and arterial

bus. As a result, with the increase of VOT, slower modes (bicycle and arterial bus) become less

cost-effective relative to faster modes (automobile, rail, BRT and flier). The variation of VOT is

not enough to affect the basic results under the baseline SDR as shown in the central column. If

one combines a high VOT (80% of wage rate) with a low SDR (4%), one can get situations

where automobile and rail are similar in cost to bus and bicycle, but still not enough to change

the overall intermodal cost relationship. Similarly, a low VOT (30% of wage rate) and a high

12

SDR (12%) are not enough to eliminate automobile’s cost advantage over BRT even at very low

volumes, but do make bicycle the most cost-effective mode across all volumes.

Besides VOT and SDR, assumptions about the allocation of the vehicle capital costs of

the automobile and bicycle and the design of the metro system ROW may affect intermodal cost

relationship. Capital costs are typically over 20% of the total costs of automobile transportation

in the model calculations here, a figure much higher than that reported in the comparable studies

in the United States. Thus, the change of automobile cost allocation from mileage based to fully

allocated to commuting trips, the way assumed for the transit modes, the vehicle capital costs

will at least triple, which will make the automobile even more expensive in our comparison. On

the other hand, even a change of metro design to be elevated for the entire length of the line will

lower its full cost by less than 10%. Such change basically won’t affect the cost-effectiveness of

the rail modes relative to other modes, except that the metro might turn out to be less expensive

than LRT under more situations.

Finally, it is important to note that the empirical cost data used in the cost comparisons

varies in quality, with the least reliable arguably being the estimates of the environmental and

safety costs. Fortunately, environmental and safety costs are very small components of total costs

for almost all the modes. For example, for the automobile mode, safety cost is about 2.4% of the

full cost, and environmental cost is about 1.0%. These are consistent to previous U.S. studies.

For example, in the Keeler and Small study, safety and environmental costs account for 3.5%

and 1.4% of total costs, respectively. Unless there is a reason to believe that these costs are

seriously underestimated (for example, by 5-10 times), the intermodal cost relationship is

unlikely to change significantly.

13

5. Conclusions and Discussions

Through comparing the full social costs (assuming the ASC residuals are not significant enough

to affect the results) of different urban passenger modes, this study provides a benchmark

evaluation of the social desirability of alternative passenger modes in the large Chinese cities.

For decision makers of those cities, this study shows: (1) the bicycle is the most cost-effective

option for radial commuting corridors, especially when trip length is relatively short. However,

the bicycle is not as competitive as the bus on ring corridors in the central city, due to its high

safety and parking costs. To some extent, this supports the policy adopted by some Chinese cities

to turn bicycle lanes on major urban streets into bus priority lanes in some busy streets, as such

policies may increase corridor capacity, control safety hazard, and save space occupied by

bicycle parking; (2) bus is more desirable than rail in the large Chinese cities, echoing the result

of Keeler and Small (1975). Bus has an advantage in part because it comes in several different

variations – arterial bus, BRT and expressway bus – with different speeds, frequencies and stop

spacing, so that cities can choose the right form of bus according to their circumstances.

Generally, arterial bus is more cost-effective under low traffic volumes, while BRT is the

cheapest mode under higher volumes. For radial corridors, expressway bus (flier) can be the best

choice under medium traffic volumes. This observation, like several other recent empirical

studies in the developed countries (for example, Pickrell 1990, Kain 1997 and GAO 2001), has

again confirmed that LRT hardly has any cost advantage to BRT in Chinese cities. However, it

should be mentioned that the difference between rail and bus is not always that significant,

especially when interest rate is low. On ring corridors, this study suggests that the costs of rail

and bus modes are much closer than on radial corridors, particularly under medium to high

volumes; and (3) it is difficult to justify the use of automobile as a major commuting mode in

14

Chinese cities even at volumes well below 5,000 pph. Automobile is not competitive to bus

running in mixed traffic of urban arterials. The result is robust to different SDRs and important

assumptions about auto occupancy, time value and number of peak hours.

The results of the present study are similar to the major conclusions in the U.S. in that

bus is less expensive than rail transit in most cases. The results differ, however, in that the

present study estimates that the automobile is almost never the least-cost mode in Chinese cities

while the earlier studies estimate that the automobile is least cost in the U.S. at very low traffic

volumes. A simulation of our short radial corridor model is conducted by simultaneously

changing the model settings to those assumed in Keeler and Small (1975). Three important

assumptions are changed to make the comparison more automobile-friendly: (1) using average

automobile occupancy of 1.5, which lowers auto costs per PKT; (2) raising values of walking,

waiting and bicycling time to be three times of in-vehicle time; and (3) allocating fixed costs to

1-hour rather than 2-hour peak transit operation, meaning doubling equipment and vehicle

capital costs for rail and all bus modes, and doubling right-of-way and infrastructure capital costs

for rail modes and BRT. The result, however, shows that the full costs of automobile remain

higher than arterial bus at low volumes. This suggest that socio-economic differences such as

urban density and structure of factor prices, rather than some technical assumptions in the

models, might be the real reasons behind the social attractiveness of the automobile. While the

urban density difference is obvious to explain, relative prices of tradable capital goods and land

vs. values of time, health and life of citizens also present vast difference between the two

countries. According to the cost estimates in Keeler and Small, the price of a car equals the value

of 900 hours’ in-vehicle time in the U.S. cities, but 25,000 hours in the Chinese cities; cost of

15

one square foot of land at the fringe of CBD or central city equals the value of 8-16 hours of in-

vehicle time in the U.S. while that of nearly 300 hours in the Chinese context.

It is necessary to note the limitations of this study. First, service reliability and operation

and investment flexibility are not accounted for in this study. Fortunately, it seems reliability and

flexibility offset each other at least among the transit modes – modes with exclusive ROW, such

as metro, LRT and BRT, are more reliable, but less flexible, than modes sharing ROW with the

general traffic, such as arterial and expressway buses. Second, using typical land use patterns,

predictability of the results presented above may diminish in the future as the assumed land use

pattern are no longer representative for the large Chinese cities. Of course, by that time it might

be necessary to redefine the question of modal alternative assessment in the large Chinese cities.

16

Reference

ALMEC Corp. and Chodai Co., ltd. (2001) Study for Public Transportation Improvement in

Chengdu City in the People's Republic of China: Final Report. Japan International

Cooperation Agency, Tokyo.

Bertaud, A. and Malpezzi, S. (2003) The Spatial Distribution of Population in 48 World Cities:

Implications for Economies in Transition. Center for Urban Land Economics Research

Working Paper, University of Wisconsin.

Boyd, J.H., Asher, N.J. and Wetzler, E.S. (1973) Evaluation of Rail Rapid Transit and Express

Bus Service in the Urban Commuter Market. Institute for Defense Analyses. Prepared for

U.S. Department of Transportation, Report No. DOT-P-6520.1. Washington, D.C.: U.S.

Government Printing Office.

Boyd, J.H., Asher, N.J. and Wetzler, E.S. (1978) ‘Nontechnological Innovation in Urban Transit:

A Comparison of Some Alternative’, Journal of Urban Economics, 5, pp. 1-20.

Government Accountability Office (GAO) (2001) Mass Transit: Bus Rapid Transit Shows

Promise. United States General Accounting Office Report to Congressional Requesters.

Ho, S.P.S. and Lin, G.C.S. (2003) ‘Emerging Land Markets in Rural and Urban China: Policies

and Practices’, The China Quarterly, 175, pp. 681-707.

Kain, J.F. (1997) ‘Cost-Effective Alternatives to Atlanta’s Costly Rail Transit System’, Journal

of Transport Economics and Policy, 31(1), pp. 25-50.

Keeler, T.E., and Small, K.A. (1975) The Full Costs of Urban Transport, Part III: Automobile

Costs and Final Intermodal Cost Comparisons. Monograph No. 21, Institute of Urban and

Regional Development, University of California, Berkeley.

17

Meyer, J.R., Kain, J.F. and Wohl, M. (1965) The Urban Transportation Problem, Harvard

University Press, Massachusetts.

Ministry of Construction (MOC), China (1990) Urban Land Classification and Land Use

Planning Standards (GBJ 137-90). (in Chinese)

Monson, K. (2008) ‘String Block Vs Superblock Patterns of Dispersal in China’, Architectural

Design, 78(1), pp. 46-53.

Pickrell, D.H. (1990) Urban Rail Transit Projects: Forecast versus Actual Ridership and Costs.

U.S. Department of Transportation.

Small, K.A. and Winston, C. (1999) ‘The Demand for Transportation: Models and Applications’,

in Gómez-Ibáñez, J.A., Tye, W.B. and Winston, C. (eds.), Essays in Transportation

Economics and Policy: A Handbook in Honor of John R. Meyer, Brookings Institution Press,

Washington, D.C.

Small, K.A., and Verhoef, E.T. (2007) The Economics of Urban Transportation, Routledge, New

York.

Tongji University and Research Institute of Standards and Norms, Ministry of Construction

(Tongji and RISN) (2004) Research on Governmental Investment Project Economic

Evaluation Method and Parameters, China Planning Press, Beijing. (in Chinese)

Wang, R. (2008) ‘Environmental and Safety Costs of Urban Passenger Transport Modes in

China’. Transportation Research Board 87th Annual Meeting Paper #08-1965.

Wang, R. (2009) ‘The Structure of Chinese Urban Land Prices: Estimates from Benchmark Land

Price Data’, Journal of Real Estate Finance and Economics, 39(1), pp. 24-38.

World Bank (1997) The Value of Time in Economic Evaluation of Transport Projects.

Infrastructure Note No. OT-5.

18

Tables

Table 1One-Way Capacity Limits of Transit Modes

Mode Speed(kph)

Frequency(per hour)

Vehicle Width(m)

Vehicle Length(m)

Vehicle Capacity(persons)

System Capacity(pph)

Metro 36 36 2.8 19.5*12 = 234 121*12 = 1,452 52,272LRT 30 30 2.6 19*8 = 152 110*8 = 880 26,400BRT 27 200 2.5 27 130 26,000Bus Varies 180 2.5 27 130 23,400Flier Varies 200 2.5 20 76 15,200

Note: specifications of metro, LRT and BRT are based on single track/running lane per direction; travel speeds of arterial and expressway buses vary with mainstream traffic, as well as station dwell time and station length. The multi-lane corridor capacities of the automobile and the bicycle are not provided but, in practice, are limited by reasonable urban expressway and arterial width.

19

Table 2Data and Assumptions of Operating Costs of Urban Passenger Transportation

Labor(per vehicle-km)

Energy(per vehicle-km)

Maintenance and Management(per vehicle-km)

BRT (wage rate/travel speed)*1.7

21 L/100km at 5 Yuan/L 1.3 times of labor costs for vehicle maintenance and management; road maintenance costs are counted as extra 16% of BRT lane capital costs

Bus (wage rate/travel speed)*1.7

25 L/100km at 5 Yuan/L 1.3 times of labor costs for vehicle maintenance and management; road maintenance costs are counted as extra 16% of arterial capital costs

Flier (wage rate/travel speed)*1.7

21 L/100km at 5 Yuan/L 1.3 times of labor costs for vehicle maintenance and management; road maintenance costs are counted as extra 16% of expressway capital costs

Auto n/a Arterial: 12 L/100km at 5 Yuan/L; expressway: 8 L/100km at 5 Yuan/L

0.15 Yuan/km plus expressway or arterial maintenance costs as extra 16% of capital costs

Bicycle n/a n/a 0.01 Yuan/km plus arterial bicycle lane maintenance costs as extra 12% of bicycle lane capital costs

Metro Total: 22 Yuan/car-km (underground); 17.6 Yuan/car-km (elevated)LRT Total: 15.84 Yuan/car-km

Table 3Capital costs of urban transit modes

Metro LRT BRT Flier BusSTRUCTUREStation Underground: 8,000

Yuan/m2 *77.9*train length (in m). Elevated: assume total construction area (assuming 30% at grade) as 85% of an underground station and a 3,000 Yuan/m2 unit cost.

Assuming unit structure cost equals that of arterial road. Total construction area assumed to be 75% of an underground station.

Station platform size is determined by peak passengers simultaneously using the station, at 0.5 m2

per passenger, plus dwell lane costs. Platform space savings are considered for radial routes.

Similar to BRT, plus additional elevated expressway structure costs of dwell lane and passing lane for multiple-berth station.

Platform size assumed to be 90% of BRT station with same capacity, plus additional arterial structure costs of dwell lane.

Pathway Underground: 80;Elevated: 30; Track: 8 (units: million/route-km).

10 million/km. Same track costs as metro.

400 Yuan/m2. Use congestion costs on urban expressway.

Use congestion costs on urban arterial.

Depot 1.86 million/car including equipments.

1.68 million/car including equipments (assuming cost is proportional to vehicle gross floor area).

255,000 Yuan per 9-m bus including equipments, assuming cost is proportional to bus length.

Same as BRT. Same as BRT.

LANDStation Underground: 200

m2/station. Elevated: 30% of total construction area.

Assuming total land area equals to total construction area.

Using platform and dwell lane area mentioned above.

Same as BRT, but 30% at grade and 70% elevated.

90% of BRT station area of same peak traffic volume.

Pathway Zero. 1/3 of pathway width by corridor length.

9-m route width by corridor length.

Included in congestion costs.

Included in congestion costs.

Depot 500 m2/car. 452 m2/car. 2.5*gross floor area/ bus. Same as BRT. Same as BRT.EQUIPMENTStation,pathway,dispatch, and depot

Underground: 83*(0.5+0.5*capacity/44,460) million Yuan/route-km. Elevated: 51.6*(0.5+0.5*capacity/44,460) million/route-km.

43.8*(0.5+0.5*capacity/44,460) million Yuan/route-km.

10*(0.5+0.5*capacity/44,460) million Yuan/route-km.

15*(0.5+0.5*capacity/44,460) million Yuan/route-km.

1*(0.5+0.5*capacity/44,460) million Yuan/route-km.

VEHICLE 7.5 million/car. 6.8 million/car (assuming cost is proportional to gross floor area).

Unit price assumed as function of bus length L (in m): 10,000*(10*L-45) Yuan.

Same as BRT. Same as BRT.

21

Table 4Estimated Urban Modal Environmental and Safety Costs in Chinese Cities

(in 2005 RMB cents per VKT)

ModeEnvironmental Costs

Safety CostsAir Pollution Noise Climate Change TOTAL

Auto (Average) 2.50 0.65 2.00 5.15 14.00Auto (expressway) 1.43 1.03 1.33 3.79 7.00Auto (arterial/local) 3.57 0.28 2.67 6.51 21.00Bus (expressway) 3.57 2.38 6.33 12.28* 21.00Bus (arterial street) 8.93 1.75 12.67 23.35* 63.00BRT 3.57 2.38 6.33 12.28* 31.50LRT 0.00 4.75 26.00 30.75 31.50Metro (underground) 0.00 0.00 26.00 26.00 15.80Metro (elevated) 0.00 4.75 26.00 30.75 15.80Bicycle 0.00 0.00 0.00 0.00 63.00Walking 0.00 0.00 0.00 0.00 75.60

*The presented environmental costs of bus modes (expressway bus, arterial bus and BRT) are those of 12-m bus.

22

Figures

Figure 1Representative Large Chinese City and Corridors

Figure 2Full-Cost Comparison across Peak-Hour Passenger Volumes on Different Corridors

Figure 3Cost Comparisons under Different SDRs and VOTs on the Short Radial Corridor

Figure_1Click here to download high resolution image

Figure_2Click here to download high resolution image

Figure_3Click here to download high resolution image

COST OF POLLUTIONIN CHINAECONOMIC ESTIMATES OF PHYSICAL DAMAGES

The World BankState Environmental Protection Administration, P. R. China

xiC H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O N

In recent decades, China has achievedrapid economic growth, industrializa-tion, and urbanization. Annual in-creases in GDP of 8 to 9 percent havelifted some 400 million people out ofdire poverty. Between 1979 and 2005,China moved up from a rank of 108thto 72nd on the World DevelopmentIndex. With further economic growth,most of the remaining 200 millionpeople living below one dollar per daymay soon escape from poverty. Al-though technological change, urban-ization, and China’s high savings ratesuggest that continued rapid growth is feasible, the resources that suchgrowth demands and the environmen-tal pressures it brings have raisedgrave concerns about the long-termsustainability and hidden costs ofgrowth. Many of these concerns areassociated with the impacts of air andwater pollution.

Rapid Economic Growth Has Had Positive EnvironmentalImpacts but Also Created New Environmental Challenges

Considering China’s strong economic growth over the last 20–25 years, thereis no doubt that it has had positive impacts on the environment. Along-side economic growth, technology improvements over this period have cre-ated much-improved resource utilization. Energy efficiency has improveddrastically—almost three times better utilization of energy resources in2000–02 compared to 1978. As a result of the changing industrial structure,the application of cleaner and more energy-efficient technologies, and pollu-tion control efforts, ambient concentrations of particulate matter (PM) andsulfur dioxide (SO2) in cities have gradually decreased over the last 25 years.Implementation of environmental pollution control policies—particularly command-and-control measures, but also economic and voluntarilymeasures—have contributed substantially to leveling off or even reducingpollution loads, particularly in certain targeted industrial sectors.

At the same time, new environmental challenges have been created. Fol-lowing a period of stagnation in energy use during the late 1990s, total energyconsumption in China has increased 70 percent between 2000 and 2005,with coal consumption increasing by 75 percent, indicating an increasinglyenergy-intensive economy over the last few years. Moreover, between 2000and 2005, air pollution emissions have remained constant or, in someinstances, have increased. The assessment at the end of the tenth five-yearplan (2001–05) recently concluded that China’s emissions of SO2 and sootwere respectively 42 percent and 11 percent higher than the target set at thebeginning of the plan. China is now the largest source of SO2 emissions inthe world. Recent trends in energy consumption, particularly increased coaluse, provide a possible explanation for the increase in SO2 emissions.

Water pollution is also a cause for serious concern. In the period between2001 and 2005, on average about 54 percent of the seven main rivers inChina contained water deemed unsafe for human consumption. This repre-

Executive Summary

sents a nearly 12 percent increase since the early1990s. The most polluted rivers occurred in thenortheast in areas of high population density.The trends in surface water quality from 2000 to2005 suggest that quality is worsening in themain river systems in the North, while improv-ing slightly in the South. This may partly be theresult of rapid urbanization (the urban popula-tion increased by103 million countrywide from2000 to 2005), which caused COD loads fromurban residents to increase substantially and,hence, surpass the planned targets for 2005.Rapid industrialization probably also plays a part.

Northern China Bears a DoubleBurden from Air and Water Pollution

While the most populous parts of China alsohave the highest number of people exposed to air

pollution, it is striking that the areas with thehighest per capita exposure are almost all locatedin northern China (Qinghai, Ningxia, Beijing,Tianjin, Shaanxi, and Shanxi). The exception isHunan, which is located in the South. In Fig-ure 1, the color of the provinces on the mapshows the percentage of the urban populationexposed to air pollution, while the bars indicatethe absolute number of people exposed.

Similarly, the most severely polluted waterbasins—of the Liao, Hai, Huai, and Songhuarivers—are also located in northern China (seefigure 2 for surface water quality). North Chinaalso has serious water scarcity problems. Someprovinces—including Beijing, Shanxi, Ningxia,Tianjin, and Jiangsu—seem to face the doubleburden of exposure to high levels of both air andwater pollution. However, while air pollutionlevels may be directly associated with population

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O Nxii

Xinjiang

Neimeng

Qinghai

Gansu

Sichuan

Jilin

Yunnan

Heilongjiang

Hebei

Hubei

Hunan

ShaanxiHenan

Guangxi

Shanxi

Anhui

JiangxiGuizhou

Liaoning

Fujian

Shandong

Guangdong

Jiangsu

ZhejiangChongqing

Ningxia

Hainan

BeijingTianjin

Shanghai

Pollution Exposure

0 - 10%

11 - 30%

31 - 45%

46 - 60%

61 - 70%

71 - 80%

81 - 90%

91 - 100%

200,000

Population Exposedto Pollution

F I G U R E 1 . Urban Population Exposed to PM10 levels, 2003

exposure, the same does not necessarily apply tosurface water pollution. This is because popula-tions generally have different drinking watersources that may allow them to escape high levelsof contamination. About 115 million people inrural China rely primarily on surface water astheir main source of drinking water. Surfacewater as a drinking water source is more vulner-able to possible pollution compared to other,safer drinking sources.

Air and Water Pollution have Severe Health Impacts

According to conservative estimates, the eco-nomic burden of premature mortality andmorbidity associated with air pollution was157.3 billion yuan in 2003, or 1.16 percent of

GDP. This assumes that premature deaths arevalued using the present value of per capita GDPover the remainder of the individual’s lifetime. If a premature death is valued using a value of a statistical life of 1 million yuan, reflectingpeople’s willingness to pay to avoid mortalityrisks, the damages associated with air pollutionare 3.8 percent of GDP. These findings differ intwo important ways from previous studies of theburden of outdoor air pollution in China. First,they are based on Chinese exposure-responsefunctions, as well as on the international litera-ture; and second, they are computed for indi-vidual cities and provinces. Previous estimates by WHO (Cohen et al. 2004) were based on the assumption that increases in PM beyond 100 g/m3 of PM10 caused no additional healthdamage.( In the base case considered by WHO,

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O N xiii

F I G U R E 2 . Water Quality Levels, 2004

relative risk does not increase beyond 50 g/m3of PM2.5, which is approximately equivalent to100 g/m3 of PM10.) This assumption impliesthat the WHO estimates cannot be used to eval-uate the benefits of specific urban air pollutioncontrol policies.

Two-thirds of the rural population is withoutpiped water, which contributes to diarrheal diseaseand cancers of the digestive system. The cost of thesehealth impacts, if valued using a VSL of 1 million,are 1.9 percent of rural GDP. Analysis of datafrom the 2003 National Health Survey indicatesthat two-thirds of the rural population does nothave access to piped water. The relationshipbetween access to piped water and the incidenceof diarrheal disease in children under the age of5 confirms this finding: the lack of access to

piped water is significantly associated with excesscases of diarrheal disease and deaths due to diar-rheal disease in children under 5 years of age.Although there are many indications that surfaceand drinking water pollution problems con-tribute to serious health impacts, the lack ofmonitoring data on specific pollutants and dataon household behavior regarding avoiding expo-sure to polluted drinking water make it difficultto quantify all of the health effects of water pol-lution. Specifically, the lack of exposure datamakes quantifying the relationship betweenchemical and inorganic pollution and the inci-dence of chronic diseases almost impossible. Pre-liminary estimates suggest that about 11 percentof cases of cancer of the digestive system may beattributable to polluted drinking water. More

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O Nxiv

Counties with no shading were categorized as 'Urban' or 'Urban Center with Rural Surroundings', which account

Rural HH NTW by County0 - 3458

3459 - 7800

7801 - 13574

13575 - 21886

21887 - 41341

Incidence of Diarrhea by Province0 - 72,061

72,062 - 208,769

208,770 - 393,469

393,470 - 633,312

633,313 - 893,222

F I G U R E 3 . Rural Households with No Access to Piped Water & Diarrhea Incidence

attention, however, needs to be given at the pol-icy level to reinforcing the surveillance capacityfor chronic exposures and disease incidence.

Health is Highly Valued by the People in China

The mortality valuation surveys conducted inShanghai and Chongqing as part of this studysuggest that people in China value improve-ments in health beyond productivity gains. Thevalue of a statistical life estimated in these surveys—the sum of people’s willingness to payfor mortality risk reductions that sum to one sta-tistical life—is approximately 1 million yuan.This number supports results of other studies,which suggest that the value of an avoided deathis greater than what is implied by the adjustedhuman capital approach, which is approximately280,000 Yuan in urban areas. Evaluation of thehealth losses due to ambient air pollution usingwillingness-to-pay measures raises the cost to 3.8 percent of GDP.

It is remarkable that the willingness to pay isabout the same in locations as different asShanghai and Chongqing, which differ greatlyin per capita GDP with a ratio as high as 5:1.(However, sample per capita incomes showed amore modest ratio of 2:1.) Furthermore, thesenew findings illustrate that the urban Chinesepopulation has a willingness to pay to reducemortality risk comparable in PPP terms to thelevels seen in several developed countries withmuch higher per capita incomes. This meansthat the Chinese people highly value their healthstatus and their longevity.

China’s Poor Are Disproportionately Affected byEnvironmental Health Burdens

Although the objective of this study was not tocompare the impacts of air and water pollutionon the poor versus the non-poor, the findingssuggest that environmental pollution falls dis-

proportionately on the less economically ad-vanced parts of China, which have a higher shareof poor populations. As shown in Figure 1,Ningxia, Xinjiang, Inner Mongolia, and otherlow-income provinces are more affected by airpollution on a per capita basis than high-incomeprovinces such as Guangdong and otherprovinces in the southeast.

From another perspective, analysis of the2003 National Health Survey showed that 75 percent of low-income households in ruralChina with children under 5 years of age have noaccess to piped water, compared to 47 percent inthe higher-income categories. This implies thatlow-income households rely more on otherdrinking water sources. In fact, about 32 percentof households within the lowest income quartilerely primarily on surface water as their primarysource of drinking water, compared to 11 per-cent in the highest income quintile. This means that the rural poor are at a substantiallyhigher risk from surface water pollution than the non-poor.

The fact that water quality in the North isworse than in the South may explain the slightlyhigher diarrheal prevalence seen in lower incomegroups in northern China (2.1 percent) com-pared to southern China (1.9 percent). How-ever, when focusing on differences betweenincome groups in the North, the data clearlyshow that the poor (lowest income quartile) havea much higher diarrheal prevalence (2.4 percent)in households using surface water compared tothe highest income groups, where no diarrheacases have been recorded.

Pollution Exacerbates WaterScarcity, Costing 147 Billion Yuan a Year

Water scarcity is a chronic problem, especially inthe North. It is closely related to problems ofwater pollution. Surface water pollution has putpressure on the use of groundwater for agricul-tural and industrial purposes. The depletion of

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O N xv

nonrechargeable groundwater in deep freshwateraquifers imposes an environmental cost, since itdepletes a nonrenewable resource and increasesfuture costs of pumping groundwater. It can alsolead to seawater intrusion and land subsidence.

Estimates of the cost of groundwater deple-tion suggest that it is on the order of 50 billionyuan per year, while estimates of the costs ofusing polluted water to industry are comparablein magnitude, bringing the overall cost of waterscarcity associated with water pollution to 147 billion yuan, or about 1 percent of GDP.These new findings indicate that the effects ofwater pollution on water scarcity are much moresevere than previous studies have estimated.

Air and Water Pollution Cause Significant Crop and Material Damage

This study makes clear that the impacts of airand water pollution on health are severe in bothabsolute and in economic value terms. Althoughwe acknowledge that not all non-health-relatedimpacts can be quantified, the impacts of pollu-tion on natural resources (agriculture, fish andforests) and manmade structures (e.g. buildings)are estimated to account for substantially lowerdamages in economic terms.

Acid Rain costs 30 billion yuan in crop damageand 7 billion in material damage annually. It is

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O Nxvi

The sum of groundwater depletion and polluted water supply (in 100 million cubic meters)0 - 1010 - 2020 - 3030 - 50>50

N

EW

S

F I G U R E 4 . Groundwater Depletion and Polluted Water Supply

Ground Water Depletion & Polluted Water Supply, 2003

estimated that acid rain, caused mainly byincreased SO2 emissions due to increased fos-sil fuel use—causes over 30 billion yuan in dam-ages to crops, primarily vegetable crops (about80 percent of the losses). This amounts to 1.8 percent of the value of agricultural output.Damage to building materials in the Southimposed a cost of 7 billion yuan on the Chineseeconomy in 2003. In addition to the humanhealth effects reported above, these damages pro-vide an additional impetus for controlling SO2.Damages to forests could not be quantified dueto lack of monitoring data in remote areas andadequate dose-response functions.

Six provinces account for 50 percent of acid raineffects. The burden of damages from acid rain is also unevenly distributed. Over half of the estimated damages to buildings occur in threeprovinces: Guangdong (24 percent), Zhejiang (16 percent), and Jiangsu (16 percent). Almost halfof the acid rain damage to crops occurs in threeprovinces: Hebei (21 percent), Hunan (12 per-cent), and Shandong (11 percent). However, theimpacts of acid rain extend across internationalboundaries and also affect neighboring countries.

Irrigation with polluted water costs 7 billionyuan per year. This study has quantified part ofthe damage caused by the use of polluted waterfor irrigation in agriculture and a portion of theimpact of water pollution on fisheries. Theimpact of irrigating with polluted water in desig-nated wastewater irrigation zones—consideringonly the impact on yields and produce quality,but not on human health—was estimated toreach 7 billion yuan in 2003.

The cost to fisheries is estimated at 4 billionyuan. The impact of acute water pollution inci-dents on commercial fisheries is estimated atapproximately 4 billion yuan for 2003. Theimpact of chronic water pollution on fisheriescould not be estimated for lack of exposure dataas well as adequate dose-response information.

Air Pollution Poses a Large Health Risk inUrban Areas and Water Pollution a SignificantHealth Risk in Rural Areas

The figures presented in the summary table atthe end of this chapter suggest that outdoor airpollution poses a very serious problem in urbanareas. This is not surprising when one comparesthe levels of ambient PM10 in Chinese cities withother large cities across the world. With annualaverage PM10 concentrations of over 100Ìg/m3,several selected cities in both northern andsouthern China are among the most pollutedcities in the world (see figure 5).

Although the health damages associated withwater pollution are smaller, in total, and as a per-cent of rural GDP, they are still 0.3 percent ofrural GDP if conservatively valued and 1.9 per-cent of rural GDP when valued using a 1 millionyuan VSL. Both figures ignore the morbidityassociated with cancer and therefore underesti-mate the health costs associated with water pol-lution. However, relative to other developingcountries, China’s diarrheal prevalence in ruralareas is quite low, actually lower than in coun-tries where a larger percentage of the rural pop-ulation has access to piped water supply (seefigure 6).

The Benefits of Sound PolicyInterventions May Exceed the Costs

This study report shows that the total cost of airand water pollution in China in 2003 was 362billion yuan, or about 2.68 percent of GDP forthe same year. However, it should be noted thatthis figure reflects the use of the adjusted humancapital approach, which is widely used in Chi-nese literature, to value health damages. If theadjusted human capital approach is replaced bythe value of a statistical life (VSL) based on stud-ies conducted in Shanghai and Chongqing, theamount goes up to about 781 billion yuan, orabout 5.78 percent of GDP.

Setting priorities for cost-effective interventions.Interventions to improve the environment inChina are likely to yield positive net benefits.Indeed, one of the advantages of the environ-

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O N xvii

mental cost model developed in this project isthat it can be used to evaluate the benefits ofspecific pollution-control policies and assist indesigning and selecting appropriate targetedintervention policies. Once the impact onambient air quality of a policy to reduce partic-ulate emissions has been calculated, the toolsused to calculate the health damages associatedwith particulate emissions can be used to com-pute the benefits of reducing them. To illus-trate, researchers have examined the costs andimpacts on ambient air quality of measures tocontrol SO2 emissions and fine particles in Shijiazhuang, the capital of Hebei Province(Guttikunda et al. 2003). The monetized valueof the health benefits associated with each mea-

sure could be calculated, using the techniquesdeveloped in this study, and compared with thecosts.

Targeting high-risk areas. The findings fromthis project suggest that a focus on northernChina is essential, particularly the North ChinaPlain and areas located northeast and northwestof the plain, where the study shows that there isa double burden from both air and water pollu-tion. This problem is further magnified by thepresence of disparities between the poor andnon-poor. On this basis, it seems relevant thatstronger policy interventions should be de-veloped to address air and water pollutionproblems. In addition, these efforts should becomplemented with emphasis on improving

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O Nxviii

F I G U R E 5 . Annual average PM10 concentrations observed in selected cities worldwide, 2004, 2005

Source: China Environmental Yearbook 2005 and WHO 2005.

access to clean water, with a specific focus on thelowest income groups.

Responding to people’s concerns. This studysuggests that the Chinese value the avoidance ofhealth risks beyond productivity gains. Thisimplies that people’s preference for a clean envi-ronment and reduced health risks associatedwith pollution are stronger than past policiesappear to have acknowledged. Growing con-cerns about the impacts of pollution are increas-ingly expected to guide national policies as wellas local actions. Public disclosure of envi-ronmental information such as emissions bypolluting enterprises, as well as ambient envi-ronmental quality data by local authorities,could be an important tool for responding topeople’s concerns and creating incentives forimproving local conditions.

Addressing the information gap. Past policiesand decisions have been made in the absence of

concrete knowledge of the environmental im-pacts and costs. By providing new, quantitativeinformation based on Chinese research underChinese conditions, this study has aimed toreduce this information gap. At the same time,it has pointed out that substantially more infor-mation is needed in order to understand thehealth and non-health consequences of pollu-tion, particularly in the water sector. It is criti-cally important that existing water, health, andenvironmental data be made publicly availableso the fullest use can be made of them. Thiswould facilitate conducting studies on theimpacts of water pollution on human and ani-mal health. Furthermore, surveillance capacity atthe local and national levels needs to beexpanded to improve the collection of environ-mental data, especially data on drinking waterquality. These efforts will further improve theanalysis begun in this project.

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O N xix

0

10

20

30

40

50

60

70

80

90

100

Cambo

dia 2

000

Moz

ambiq

ue 2

003

Indo

nesia

200

2/20

03

Ghana

200

3

Mad

agas

car 2

003/

2004

Kenya

200

3

Mor

occo

200

3-20

04

China

2003

Philipp

ines 2

003

Peru

2000

Bolivia

200

3

Colom

bia 2

005

Egypt

200

0

Jord

an 2

002

Per

cent

age

of r

ural

hou

seho

lds

w/ n

o pi

ped

wat

er

Piped water Diarrhea Prevalence last 2 weeks

Figure 6. Diarrheal Prevalence and Access to Piped Water Supply

Source: ORC Macro, 2006. MEASURE DHS STATcompiler. http://www.measuredhs.com, July 3 2006.

Developing an environmental-health actionplan. At present, an environmental-health actionplan is being jointly drafted by the State Envi-ronmental Protection Administration (SEPA)and the Ministry of Health (MoH). This planshould take into consideration the mortality andmorbidity impacts from water and air pollution

presented in this report. The plan should includea focus on the geographical areas identified innorthern China, where there is a double burdenof both air and water pollution. Furthermore,particular focus should be put on areas wherepoor populations are adversely affected fromlack of access to clean water and sanitation.

E X E C U T I V E S U M M A R Y

C H I N A – E N V I R O N M E N T A L C O S T O F P O L L U T I O Nxx

Addressing China’s Water Scarcity

Recommendations for Selected Water Resource Management Issues

Jian Xie

with

Andres Liebenthal, Jeremy J. Warford, John A. Dixon, ManchuanWang, Shiji Gao, Shuilin Wang, Yong Jiang, and Zhong Ma

For years, water shortages, water pollution, and flooding have constrainedgrowth and affected public health and welfare in many parts of China. North-ern China is already a water-scarce region, and China as a whole will soonjoin the group of water-stressed countries. The combined impact of thewidening gap between water demand and limited supplies and the deterio-rating water quality caused by widespread pollution suggests that a severewater scarcity crisis is emerging.

China’s leadership is aware of the worsening water situation, and is com-mitted to transforming China into a water-saving society. The 11th Five-Year Plan (2006–10) sets a number of policy goals and priorities for waterresource management, such as (a) adopting a more unified or better coordi-nated management system; (b) shifting from supply-side to demand-sidemanagement; (c) integrating river basin management with regional manage-ment; and (d) establishing a preliminary system of water rights trading.

To date, however, the increasing scarcity of water resources has not beeneffectively managed. Many national and local water resource managementand water pollution control plans have not been fully implemented and manytargets, including water pollution investment targets, have not been achieved.The economic costs of water pollution and scarcity are high. Water pollutionposes a serious threat to public health and causes major economic and envi-ronmental losses, estimated by the Chinese government at the amount equiv-alent to about 1.7 percent of GDP or more in 2004.

OBJECTIVES AND SCOPE OF THIS REPORT

This report reviews China’s water scarcity situation, assesses the policy and insti-tutional requirements for addressing it, and recommends key areas for strength-ening and reform. It is a synthesis of the main findings and recommendationsfrom analytical work and case studies prepared under the World Bank Analy -tical and Advisory Assistance (AAA) program entitled “Addressing China’s

Executive Summary

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xix

Water Scarcity: From Analysis to Action.” Thesestudies focus on several strategically importantthematic areas for China where additional researchwas needed, as identified by the research team andadvisory group based on a review of pressing issues.These areas are governance, water rights, pricing,ecological compensation, pollution control, andemergency response.

The approach has been to evaluate Chineseand international experience to identify policyand institutional factors that have proven effec-tive in promoting the adoption of water conser-vation and pollution reduction technologies. Theresearch was based on literature reviews, qualita-tive and quantitative policy analyses, householdsurveys, field trips, and case studies to developfeasible recommendations for a plan of actionbased on realities on the ground.

CHINA’S EMERGING WATER CRISIS

China’s water resources are scarce and unevenlydistributed. China’s renewable water resourcesamount to about 2,841 km3/year, the sixth largestin the world. Per capita availability, however—estimated at 2,156 m3/year in 2007—is onlyone-fourth of the world average of 8,549 m3/yearand among the lowest for a major country. WhileChina as a whole is facing serious water stress, itsproblems are made more severe by the fact thatits water resources are unevenly distributed, bothspatially and temporally.

China’s water resources availability variesgreatly over space. The South, with averagerainfall of over 2,000 mm/year, is more water-abundant than the North, where rainfall onlyaverages about 200–400 mm/year. Per capitawater availability in northern China is only 757 m3/year, less than one-fourth that in south-ern China, one-eleventh of the world average,and less than the threshold level of 1,000 m3/yearcommonly defined as “water scarcity.”

The temporal pattern of precipitation fur-ther intensifies the uneven spatial distribution ofwater resources. With a strong monsoonal cli-

mate, China is subject to highly variable rainfallthat contributes to frequent droughts and floods,often simultaneously in different regions. Whileprecipitation generally declines from the south-eastern coast to the northwestern highlands, itvaries greatly from year to year and from seasonto season. In the Hai and Huai basins, for exam-ple, river flows fall to 70 percent of their averagesone year in four and to 50 percent one year intwenty. Dry years tend to come in succession,accentuating the water problem.

China’s Water Productivity Is Low

China’s water productivity of $3.60/m3 is low incomparison with the average of middle-income($4.80/m3) and high-income ($35.80/m3) coun-tries. This gap is largely due to differences inthe sectoral structure and efficiency of waterconsumption.

Water productivity in agriculture, whichaccounted for 65 percent of total water with-drawals, is the lowest of all sectors, due to exten-sive waste in irrigation systems, and suboptimalallocation among crops and between differentparts of the same river basin. Only about 45 per-cent of water withdrawals for agriculture areactually used by farmers on their crops. In indus-try, which accounts for 24 percent of total waterwithdrawals, the water recycling level is 40 per-cent on average compared to 75–85 percent indeveloped countries.

A major contributor to China’s low water pro-ductivity is its very inefficient water allocationsystem. A recent study of the Hai basin has foundthat water productivity, as reflected by the eco-nomic value of water (EVW) in different uses,ranges from 1.0 yuan/m3 in paddy irrigation to12.3 yuan/m3 in vegetable fields, 21.3 yuan/m3

in manufacturing, and 33.7 yuan/m3 in the ser-vices sector. The magnitude of these differencesin an extremely water-short region is indicativeof a serious lack of market consciousness in thewater allocation process.

China’s water scarcity is aggravated by ex -tensive pollution. Over the past three decades,

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxx

despite great efforts to control it, water pollu-tion has increased, spreading from the coast toinland areas and from the surface to under-ground water resources. Total wastewater dis-charges have steadily risen to 53.7 billion tonsin 2006. Domestic wastewater discharges havesurpassed industrial discharges since 2000, andhave become the most important pollutionsource. It was not until 2007 that the risingtrend of water pollution discharges began toshow a sign of reverse, as total 2007 COD dis-charges were reported to be 3.14 percent lessthan in 2006. However, the water pollution sit-uation is still very serious. A major contributingfactor is that only 56 percent of municipalsewage is treated in some form, versus 92 per-cent of industrial discharges.

Water pollution incidents also represent aserious threat. They overwhelm the already frag-ile water environment, contaminate downstreamdrinking water for millions of people, and severelythreaten public health and the quality of life.

As a result of continuing pollution, the waterquality of most of China’s water bodies has beenextensively degraded. In 2004, of all 745 moni-tored river sections, 28 percent fell below theGrade V standard (that is, unsafe for any use),and only 32 percent met Grade IV–V standards(that is, safe for industrial and irrigation usesonly). Of 27 major monitored lakes and reser-voirs, fully 48 percent fell below Grade V stan-dards, 23 percent met Grade IV–V standards,and only 29 percent met Grade II–III standards(safe for human consumption after treatment).

The extent of pollution aggravates the scarcityof water. At present, approximately 25 km3 ofpolluted water are held back from consumption,contributing to unmet demand and ground-water depletion. As much as 47 km3 of water thatdoes not meet quality standards are neverthelesssupplied to households, industry, and agricul-ture, with the attendant damage costs. A further24 km3 of water beyond rechargeable quantitiesare extracted from the ground, which results ingroundwater depletion.

Water Scarcity and Pollution EntailSubstantial Costs

The most important costs relate to the healthrisks associated with polluted drinking watersources. Over 300 million people living in ruralChina have no access to safe drinking water. Theeconomic cost of disease and premature deathsassociated with the excessive incidence of diar-rhea and cancer in rural China has been esti-mated, based on 2003 data, at 66.2 billion yuan,or 0.49 percent of GDP.

Water scarcity is also undermining the ca -pacity of water bodies to fulfill their ecologi-cal functions. Due to excessive withdrawals, evena minimum of environmental and ecologicalflows cannot be ensured for some major riversin North China. To compensate for the deficitof surface water, North China has increasinglyrelied on groundwater withdrawals, often inexcess of sustainable levels. Such overexploitationhas resulted in the rapid depletion of ground -water reservoirs, leading to the lowering of watertables, the drying up of lakes and wetlands, andland subsidence in many cities.

The World Bank’s Cost of Pollution in Chinastudy estimated that the water crisis is alreadycosting China about 2.3 percent of GDP, ofwhich 1.3 percent is attributable to the scarcityof water, and 1 percent to the direct impacts ofwater pollution. These estimates only representthe tip of the iceberg. They do not include thecost of impacts for which estimates are unavail-able, such as the ecological impacts associatedwith eutrophication and the drying up of lakes,wetlands, and rivers, and the amenity loss fromthe extensive pollution in most of China’s waterbodies. Thus, total costs are undoubtedly higher.

A PLAN OF ACTION FORADDRESSING WATER SCARCITY

As outlined above, the major factors underlyingthe emerging water crisis point to the need forChina to reform and strengthen its water re -source management framework. In line with the

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxi

broad strategy of developing a market economy,the focus of the reform needs to be on clarifyingthe role of and relationships between govern-ment, markets, and society; improving the effi-ciency and effectiveness of water managementinstitutions; and fully embracing and usingmarket-based instruments as much as possible.

On this basis, the following thematic areaswere selected for attention: (a) improving watergovernance; (b) strengthening water rights admin-istration and creating water markets; (c) improv-ing efficiency and equity in water supply pricing;(d) protecting river basin ecosystems throughmarket-oriented eco-compensation instruments;(e) controlling water pollution; and (f ) improvingemergency response and preventing pollution dis-asters. The main findings and recommendationsare summarized below. The combined set of rec-ommendations, summarized in a table in the finalchapter of this report, represents an action plan foraddressing China’s water scarcity.

IMPROVING WATER GOVERNANCE

To address the growing complexity of waterresource management, China is moving from atraditional system with the government as themain decision-making entity toward a modernapproach to water governance that relies on (a) asound legal framework, (b) effective institutionalarrangements, (c) transparent decision makingand information disclosure, and (d) active pub-lic participation.

An Effective Water GovernanceSystem Has to Be Built on a Sound Legal Basis

China has made much recent progress in im -proving its legal framework. Even so, the effec-tiveness of the legal framework for water resourcemanagement is unsatisfactory, as evidenced bythe growing seriousness of water-related problems,including rampant water pollution nationwide. Itsmain weaknesses and areas for improvement are:

Lack of mechanisms and procedures

Existing laws and regulations are usually focusedon principles and lack mechanisms and proce-dures for enforcement, such as supervision, mon-itoring, reporting, evaluation, and penalties fornoncompliance.

Incomplete legal system

The coverage of the existing legal framework isstill limited. For example, the Water PollutionPrevention and Control Law requires that thestate establish and improve the compensationmechanism for ecological protection of the waterenvironment, but there are no national laws orregulations to support it. Neither is there a lawon water rights and trading.

Ambiguities and conflicts between legal provisions

Some laws contain ambiguities. For example, theWater Law does not clearly define the authorityof local governments and river basin manage-ment commissions (RBMCs). Such ambiguityin the provisions causes a vacuum of authorityand weakens the effectiveness of the legal system.

Existing Institutions Are Fragmentedand Uncoordinated

China’s water resource management system ischaracterized by extensive fragmentation, bothvertical and horizontal. Horizontally, at everylevel of government several institutions are in -volved in water management, with frequentoverlaps and conflicts of responsibilities. Thisunwieldy system has increased the administra-tive cost of coordination among different insti-tutions and undermined the effectiveness ofwater management.

The water management system is also verti-cally fragmented. It is mainly built upon theadministrative boundaries of different levels ofgovernment rather than at the river basin level.Each level of government has its own focal pointsand priorities. This makes the management of

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxii

transboundary rivers—most of China’s rivers—very difficult.

China has established RBMCs for its sevenlarge rivers as subordinate organizations of theMinistry of Water Resources. However, thesecommissions have limited power and have norepresentatives from the affected local govern-ments in the basin. As a result, it is difficult forthe RBMCs to coordinate with the provinces/municipal administrations and other stake-holders in river basin management.

Transparency Is Limited

Transparency means that the public can havebetter access to information on water resources,policies, and institutions on water-related issuesand water-related behaviors of various stake-holders. The Chinese government has been awareof the importance of transparency and madeefforts to increase the openness of public admin-istration, but as of now, information on waterquality and quantity, water users, and pollutersremains inaccessible to the public and to govern-ment agencies outside of the sector.

The legal definition of what informationshould be disclosed to the public is not clear, sothat many organizations or enterprises refuse todisclose water-related information in the guiseof protecting state or business secrets. Finally,the citizens’ right of access to information is notemphasized in the laws, so that although severalregulations on information disclosure have beenpromulgated, they are not yet implemented wellbecause of weak supervision by the governmentand the public.

Public Participation Is Very Low

Public participation is helpful to tailor policy tolocal situations, to maximize the social welfareand utility of resources use, and to protect vulner-able groups. Major forms of public participationin water management in China are (a) publicopinion surveys; (b) public hearings; (c) experts’assessment/reviews of development plans and

programs; and (d) stakeholder coordination. Butactual public participation is still very low, whichis attributable to limited awareness by govern-ment agencies and the general public regardingthe potential for public participation in watermanagement, lax legal requirements and super -vision, and legal barriers to the registration andparticipation of NGOs, which should be expectedto play a very active role.

Recommended Actions

Amend and improve existing water-relatedlaws and regulations

Given the vagueness and even contradictions ofexisting laws and regulations, the NPC shouldreview and revise existing water-related laws, withparticular attention to the enforcement issue andintegrated water management.

Improve law enforcement

Improving law enforcement is the number onepriority to make the legal framework useful andeffective. A series of actions need to be taken:

Implementation procedures: Detailed imple-mentation procedures should be stipulated inall water-related laws and regulations to makeexisting laws and regulations operational andenforceable.

Strengthened supervision and inspection: Super-vision and inspection by the national and lo -cal congresses and administrative branchesshould be strengthened.

Public participation: The public should be em -powered to help monitor and track down vio-lators and supervise local agencies responsiblefor law enforcement, and public-private part-nerships should be encouraged by laws andregulations.

Establish a national-level organization for integrated water management

One option is to establish a State Water Re -sources Commission as a coordinating and steer-ing organization on water-related affairs across

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxiii

the country at the highest level of government.This commission will serve as a high-level waterpolicy-making body, much like the newly estab-lished State Energy Commission headed by thepremier. Another option would be to merge majorwater-related duties currently under different gov-ernment agencies (namely MWR, MEP, MOA,MHURC, and MLR) and establish a new superministry to implement unified management ofwater quantity and quality, surface water andgroundwater, water resource conservation anduse, and water environmental protection.

Convert RBMCs into intersectoral commissions

The existing RBMCs for the seven major riversshould be converted into true intersectoral andintergovernmental “commissions” with repre-sentatives from relevant line agencies and localgovernments, instead of being subordinates ofMWR. In the long run, RBMCs should be madeindependent of MWR and accountable to theState Council directly.

Make public information disclosure acompulsory obligation of the government,companies, and relevant entities

Public information disclosure requirementsshould be incorporated into all major develop-ment strategies, policies, regulations, and opera-tional procedures. The information must beaccessible for the public and concerned groups/communities and be made available throughmultiple channels.

Build a strong legal foundation for public participation

The rights of public participation should beemphasized in relevant laws to empower the pub-lic. In such laws as the Water Law and the Envi-ronmental Protection Law, articles should beadded to explicitly grant rights of participationin water management to the public. Three rightsneed to be clearly defined: (1) the right of accessto information, (2) right of participation in deci-

sion making, and (3) right to challenge water-related decisions by the government.

STRENGTHENING WATER RIGHTSADMINISTRATION AND CREATINGWATER MARKETS

The allocation of water rights and the establish-ment of water markets can improve the eco-nomic efficiency of water use in China and helpresolve water shortages. China has been estab-lishing a water rights administration since 2000,and has made remarkable progress in some pilot-ing areas. A preliminary framework of laws, reg-ulations, and institutions on water rights hasbeen developed at the national level. Additionalactions are needed to deepen water rights admin-istration and develop water markets.

Water Allocations ExceedSustainable Levels

At present, there is a general lack of conside -ration and provision for environmental waterrequirements, with the result that for many sur-face water bodies and underground aquifers,water withdrawals are far in excess of sustain-able levels. In some instances, water has beenset aside for the environment, but these vol-umes are not allocated on a sound scientificbasis. This poses a threat to the long-termhealth and sustainability of the water resourcesin question.

Water Rights Are Still Unclear and Unenforceable

Establishing clear, enforceable rights is an essen-tial first step toward the creation of water mar-kets. At present it is not always clear who holdsthe right and what the right entitles the holderto. There are few rules in place that protectagainst changes to water rights and no clearprovisions dealing with what happens when aright is adversely affected.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxiv

Water Rights and Allocations Need to Be Based on the Evapo transpiration Approach

Past water management in China, based onwater abstraction only, has encountered onlylimited success because the saved water was usedto irrigate more land; that is, more water wasconsumed and less water returned to the surfaceand underground water systems. Recent advance-ments in remote sensing and geographic infor-mation system (GIS) technologies have made itfeasible to manage water resources in terms ofthe amounts of water actually consumed throughevapotranspiration (ET). The portion consumedthrough ET is the consumptive use that is lostand not available for users downstream. In con-trast, the portion that returns to the surface orunderground water systems is still available forother users downstream. ET technology thusmakes it feasible for China to adopt a more sci-entific approach for its water rights allocationand administration.

Water Rights Administration andTrading Need to Be Strengthened

China still has a distance to go in establishing awell-functioning water rights administrationsystem. First, water rights and water rights trad-ing represent a relatively new concept for waterresources management in China, and requirereforms in institutions and policies that havebeen traditionally based on “command-and-control” regulation. Second, implementing trad-able water rights requires improvements in themonitoring and information system for decisionmaking and the enforcement of regulations.Third, there is no precedent for implementingtradable water rights in a large developing coun-try like China, with its unique physical, eco-nomic, and social background. It is a challenge,but international experience and pilot projectssupport tradable water rights as a promisingapproach for China to pursue.

Recommended Actions

Use water resources allocation plans as the basis for water rights

Water resources allocation plans should bedeveloped—first at the basin level, then at theregional level—as the basis for allocating waterwithin a basin. Plans should set a cap for totalwater abstraction permits in the plan area andclearly identify the water available for abstrac-tion, the amounts of water consumed, and theamounts that must be returned to the local watersystem.

Recognize ecological limits of water resources

Water resource allocation plans should recognizethe requirements of the in-stream environmentfor water. Water should be set aside for this pur-pose, recognizing the importance of differentparts of the flow regime for different parts of theecosystem.

Water withdrawal permits need to beclearly specified and implemented

Permits must be specified in volumetric termsand need to be linked to the initial allocation ofwater established in the water resource plan.The total amount of water withdrawal permitsshould be limited to the maximum allowableamount based on sustainable water use with suf-ficient consideration of environmental uses andnew water uses.

Strengthen water rights administration andprovide certainty and security for holdersof water rights

Water rights administration needs to be strength-ened, with the conditions, procedures, rights,and obligations for water withdrawal, consump-tion, and return flows clearly specified. Theprocess for granting water rights, and in partic-ular for allocating water on an annual basis,should be clear and consistent. This will providecertainty and security for holders of water rights.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxv

Adequate monitoring, reporting, and enforce-ment are part of effective water rights adminis-tration. Public participation, such as grouppar ticipation through water user associations inrural areas, is critical to the success of waterrights management.

Where feasible, an ET-based waterresource management should be promoted

The ET approach focuses on actual water con-sumption and hence encourages more efficientuse of water, increased return flows, and theadoption of more water-saving technologies.The ET approach can thus help improve thesustainability of the water resource system inboth agricultural and urban areas. Governmentsshould promote the ET-based water resourcemanagement, especially in water-stressed areas.

Adopt a step-wise approach to water trading

The sale or lease of water rights can be an effec-tive approach for raising the productivity ofand returns to water within and between sec-tors. But before trading starts, the entitlementsof users under different levels of resource avail-ability must be clearly defined. Once all stake-holders have gained experience in managing,monitoring, and observing rights, trading optionscan be explored, often starting with temporarytrading in well-defined systems where infra-structure for delivery and monitoring is alreadyin place.

IMPROVING EFFICIENCY ANDEQUITY IN WATER SUPPLY PRICING

Traditionally, China’s policies have focused onmeeting the demand for water by increasing thesupply rather than managing demand. An impor-tant factor contributing to the current water-scarcity crisis is the lack of effective water resourcepolicies that focus on demand management andencourage efficient water use.

Water Pricing Can Be an EffectiveMeans to Reduce Demand for andImprove the Economic Efficiency of Water Use

The central and local governments in Chinahave recognized this, and allowed water tariffsto gradually rise since the early 1990s. Even so, repeated studies have shown that water and sewerage prices in China are still below the requirements for financial cost recoveryand take little account of environmental anddep letion costs. This has made it difficult forthe water and sewerage utilities to adequatelymaintain their infrastructure, expand their ser-vice to outlying and poorer areas, and operatetheir infrastructure in a manner that meetsenvironmental standards. Thus, the first steptoward setting prices right should be to at least meet the utilities’ financial performancerequirements.

To Promote Efficient Water Use,Water Prices Also Need to Reflectthe Marginal Opportunity Cost of Supply

Prices based on marginal opportunity cost(MOC)—which includes production, environ-mental, and depletion costs—would signal thefull scarcity value of water to the consumer andinduce the appropriate adoption of water-savingand efficiency technologies. Current tariff regu-lations in China already allow all of the com -ponents of marginal opportunity cost to berecognized and signaled to the consumer. Pro-duction costs are contained in the water devel-opment fee, environmental and depletion costsin the water resource fee, and waste disposal inthe sewerage fee. But local authorities have beenslow to fully implement the necessary tariffincreases allowed by regulation, mainly as aresult of concerns about the impact on the low-income population.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxvi

Equitable and Efficient Tariff ReformIs Feasible

Although often stemming from concern for thewell-being of poorer households, low water tar-iffs have perverse consequences for income dis-tribution. Results from household surveys showthat the social impact of low water pricing on thepoor is negative; they receive little or no benefitfrom the water price subsidies, but pay a highprice for poor water supply services in terms ofhealth impacts and the high cost of alterna-tives. On the other hand, tariff reforms can bedesigned to at least partially protect the poorfrom the impact of higher rates. Provided theincreased revenues are used to extend the serviceinfrastructure and improve the quality of service,a win-win solution can result. In China and othercountries, three such approaches have been used:(1) increasing block tariffs (IBT), as alreadyenshrined in Chinese regulations; (2) incomesupport; and (3) price waivers for the pooresthouseholds.

Recommended Actions

Given the low efficiency of and high demand forwater use, China should aggressively use pricingpolicy to manage water demand. This means thatwater tariffs, including wastewater treatmentfees, have to continue increasing in the years tocome. For pricing reform to be successful, the fol-lowing recommendations are important.

Adopt a step-wise approach to tariff reform

The public should be fully informed of the prob-lems of low service quality, indirect costs, ineffi-ciency caused by underpricing or subsidizationof water, and the importance of water tariffincreases. Public hearings, consumer education,and transparency are necessary to overcomeresistance to price reform, especially when exist-ing service quality is poor.

Raise water tariffs to fully reflect its scarcity value

While the first step in price reform must be tofully achieve financial cost recovery, pricing ofwater and sewerage should follow the MOCapproach and reflect the incremental costs ofwater and its disposal, including the costs ofenvironmental damage in production and con-sumption and the opportunity costs of deple-tion. A system should be devised in which MOCestimates can be integrated into regional andnational water management and economic plan-ning systems so as to enhance the market con-sciousness of the allocation process while thetrading system is being developed.

Address the social impact of tariff increases

The increasing block tariff approach, especially atwo-tier tariff structure, is recommended for resi-dential consumers. The first block should followthe WHO-recommended 40 liters per capita perday (i.e., about 5 m3 per month for a householdof four), with the second block gradually increas-ing to full MOC. Other pricing or income sup-port methods for the poor are encouraged to beadopted based on local political and economic cir-cumstances. Water tariffs for commerce andindustry should cover the full MOC.

Convert the water resource fee to a tax

The water resource fee, which is currently re -tained by local governments, provides little incen-tive for sustainable water resource developmentbasin-wide or at the national level. The fee shouldbe converted into a tax, the proceeds of which willbe transferred to and appropriated by the centralgovernment. Such a conversion will provide afinancial basis for the central government to facil-itate more efficient water resource planning basedon national priorities for water resource develop-ment and management. The funds of local waterresource management programs should be de-linked from the revenue of water resource fees and

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxvii

directly provided by central and local govern-ments through their annual budgets.

PROTECTING RIVER BASINECOSYSTEMS THROUGH MARKET-ORIENTED ECO-COMPENSATIONINSTRUMENTS

Addressing water scarcity requires protecting thesources of the water, especially the ecosystems inthe upper reaches of river basins, such as forests,wetlands, and even agricultural lands. Both cen-tral and local governments are increasingly inter-ested in the use of government transfers frompublic funds—under the name of ECMs—toprotect ecosystems in the upper reaches of riverbasins. But the current approach relies on pub-lic financial transfers (mainly from the centralgovernment), and lacks a direct link betweenecosystem service providers and ecosystem ser-vice beneficiaries. This raises some doubts aboutthe long-term financial sustainability and effi-ciency of ECMs.

Payment for Ecosystem Services(PES) Offer a More Market-OrientedApproach

In a PES system, a market for environmental ser-vices is created whereby money is collected orreallocated from the beneficiaries who use envi-ronmental services (water consumers) and pay-ments are made directly to those who providethese services (such as watershed land managers).PES offers a more market-oriented and self-financing alternative to the government-fundedECMs currently used in China.

PES Has Been Tested in OtherCountries and Has Great Potential in China

PES has been developed and implemented inother countries with encouraging results and canbe applied in China, as illustrated by the case

study of the Lashihai Nature Reserve in LijiangCity, Yunnan Province. While PES schemes arenot a universal panacea, nor always easy to intro-duce, they should be treated as one step forwardto enhance and complement existing efforts ofecosystem conservation in China.

Recommended Actions

Continue to expand the application of ECM

Given the urgency of protecting ecosystems inthe upper reaches of river basins for water sup-ply, China should continue to expand its ECMprograms, especially when the ecosystem serviceproviders and beneficiaries are distant from oneanother and their links cannot be explicitlydefined, or where there are obvious poverty alle-viation benefits.

Promote the piloting of PES

To improve the efficiency and effectiveness ofecological compensation and reduce the finan-cial burden on governments, China should vig-orously pilot more market-oriented approachesfor ecological compensation, such as PES. Ithas much appeal in China and should be pilot -ed and promoted, beginning with some smallwatersheds.

CONTROLLING WATER POLLUTION

The government has acknowledged the serious-ness of water pollution and placed it at the topof pollution problems facing the country. Sincethe mid-1990s, COD reduction has been oneof two major nationwide total emission controltargets (the other is SO2). Even so, total CODemissions have increased since the early 1990s,largely due to an increase in emissions of un -treated municipal wastewater. In spite of over adecade of effort, it was not until 2007 that therising trend in total COD discharges appears tohave finally been reversed.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxviii

There Is Inadequate Investment in Water Pollution Control and a Large Amount of Wastewater Is Still Untreated

The investment shortfall contributed to the fail-ure to meet pollution control targets—such asreducing COD discharge by 10 percent by theend of 2005—and to environmental deteriora-tion. With insufficient investment, wastewatertreatment capacity, including sewerage networks,has not expanded adequately, especially in smallcities and townships. As a result, only 56 per-cent of the 53.7 billion tons of domestic waste-water discharged is treated in some form; therest is still discharged without any treatment,offsetting the significant reduction in industrialpollution.

Many Water Pollution Preventionand Control Plans Have Failed to Achieve Their Objectives

China has prepared water pollution preventionand control plans at the national, local, and riverbasin levels. So far, many of these plans havefailed to achieve their targets. For example, theHuai River basin was the first river basin inChina to undertake a major planning effort forwater pollution control. Evaluation of the firsttwo five-year plans (1996–2005) found that thewater quality and total emission control targetswere not achieved. For instance, the 9th FYP’s(1996) water quality target for 2000 was toachieve Class III for the entire main stream.However, by 2005, the water quality at 80 per-cent of monitoring sites in the basin was still atClass IV or worse.

Serious Water Pollution IsAttributable to Institutional and Policy Shortcomings

The effectiveness of pollution control is con-strained by several issues: (a) poor law enforce-

ment and compliance; (b) failure to implementwater pollution prevention and control plans;(c) lack of incentives for wastewater treatment;(d) a wastewater discharge control system under-mined by problems with the issuance of permits,and their monitoring and enforcement; (e) lackof integrated river basin management and weaklocal commitment to pollution control underthe influence of local and sectoral interests; (f ) increasing and unchecked pollution fromtownships and nonpoint sources; and (g) insuf-ficient and spatially imbalanced investment inwastewater treatment.

A Number of Issues Deserve Greater Scrutiny

Some of these issues include carefully defining theobjectives of the Water Pollution Prevention andControl Law; providing more reliable and com-plete information on pollution sources; emphasiz-ing the linkage between water pollution andunsafe drinking water sources; integrating pollu-tion control measures, especially the use of permitsystems; strengthening the deterrent function ofcurrent legislation and enforcement systems formanaging water pollution; promoting routinepollution prevention over after-incident treat-ment; and addressing the relationship betweenthe polluter-pays principle and governmentresponsibility at the regional and national level,especially in those areas where governments havesome responsibility due to their past activities.

Recommended Actions

The key to controlling and solving serious waterpollution in China is the strengthening of lawenforcement to improve compliance by indus-tries and other polluters. The government has touse all available means—legal, institutional, andpolicy—and, through them, mobilize the publicand motivate the private sector to ensure full com-pliance with all pollution control requirements.Specific recommendations are provided below.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxix

Improve pollution control planning

Water pollution control planning in river basinsshould be improved, with the introduction ofmore realistic and tangible targets. Pollution con-trol should not be regarded as the final target, butthe way to achieve a clean and healthy water envi-ronment. This requires a long-term, integrated,but progressively targeted strategy for the protec-tion of water quality. The financing, implemen-tation, monitoring, and evaluation mechanismsshould be well-embedded in the plans.

Unify and strengthen the pollutionmonitoring system

Better monitoring capability is required for thewhole range of measures required for effectivepollution control. The current segmented watermonitoring system—involving MEP, MWR,and MHURC—has to be reformed. In the shortterm, the systems should be better coordinated,with a unified set of monitoring criteria and pro-cedures for releasing water quality informationin one channel. In the medium term, the differ-ent monitoring systems can be consolidated andmanaged by a third entity independent of anysingle ministry.

Strengthen the wastewater discharge permit system

To be effective, the wastewater discharge permitsystem should be built on a more solid legalbasis, with a special administrative regulationissued by the State Council. The issuing of per-mits has to be technically sound and based onenvironmental quality, with daily maximumlevels of discharge specified in order to achieveambient targets. It should target key pollutantsfirst and aim to control the total pollution loadwithin the allowed pollution carrying capacity ofthe environment.

Increase reliance on market-based instruments

Pollution control efforts should take full advan-tage of market mechanisms to overcome market

failures in pollution reduction. Economic incen-tive measures (such as the pollution levy andfines) have to be rigorously enforced to providea strong incentive for polluters to comply withemissions standards and other environmentalrequirements. The upper limits of maximumfines specified in current laws should be increased.Furthermore, the system of trading of water dis-charge permits should be gradually introduced inwatersheds to improve the economic efficiency ofwastewater treatment.

Enable litigation for public goods

The litigation system should be used to givemore protection to the public interest. The lawshould encourage or require local governmentson behalf of the public to initiate lawsuitsagainst polluters and demand full compensa-tion for damage to public goods—for example,to ecosystems—where damage to individuals ishard to identify. For significant cases, MEP itselfmight be the plaintiff.

Control rural pollution

Attention should be given to addressing risingwater pollution in small towns and rural areas.The regulation of industrial and municipal sourcesin small towns and rural areas should be carriedout by local EPBs and supervised by MEP. Withregard to wastewater, sewage treatment in smalltowns should be promoted through the intro-duction of cost recovery policies, selection ofefficient technologies, and the reuse of treatedwater for irrigation.

Increase financing for market gap areas

There are several areas where market-basedapproaches cannot be expected to effectivelyaddress, for which the central government needsto earmark special budgets with which to financewater pollution prevention and control. Theseareas include: (1) transprovincial pollution con-trol and management, (2) important ecologicalregions and water sources, (3) dealing with acci-dents affecting international water bodies, and

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxx

(4) other issues with a national dimension thatcannot be properly managed at the local level.

IMPROVING EMERGENCY RESPONSE AND PREVENTINGPOLLUTION DISASTERS

Despite some successful recent cases of envi-ronmental emergency response, the high fre-quency of serious pollution incidents and theirassociated costs in China support the need forcontinued reform and strengthening of exist-ing institutions for environmental pollutionemergency prevention and response. Currentpractice in emergency management still sug-gests that the main focus of local govern-ments has been on mitigation after an incident.But prevention of incidents by strict enforce -ment of appropriate policies and regulations istypically a more cost-effective approach andshould be emphasized. A situation analysisshows that the problem is attributable to vari-ous factors, ranging from weak legal and insti-tutional arrangements, lack of incentives, andpoor chemical management systems to inade-quate on-site coordinating, monitoring, andreporting.

Based on lessons from the international expe-rience, the basic elements of an effective preven-tion and response system, as already developedand implemented in many developed coun-tries, include (a) a shift from mitigation to afocus on risk assessment, prevention, and plan-ning; (b) enhancing the preparedness of firstresponders; (c) rigorous implementation of thepolluter-pays principle to shift financial res pon -sibility for the costs of potential disasters topol luters, (d) the establishment of chemicalinfor mation management sys tems to track theflow of toxic chemicals and provide the necessaryinformation for a quick and effective response ifan accident occurs, and (e) effective public infor-mation systems to provide timely information inthe event of an emergency.

Recommended Actions

Shift from mitigation to prevention and planning

Environmental protection and work safety agen-cies should be the competent authorities to ap -prove the adequacy of environment and safetyrisk assessment, applying a thorough risk man-agement approach that focuses on both preven-tion and mitigation of the impacts of chemicalincidents. All high-hazard plants—regardless ofage—should be subject to risk assessment and berequired to prepare an emergency response plan.

Enhance preparedness

First responders should be well trained for han-dling chemical incidents and equipped with themandate and resources to contain pollutionreleases. The National Chemical RegistrationCenter and its regional offices should establish aunit, independent from enforcement divisions,to provide 24-hour technical support to theemergency services on the properties and appro-priate responses to specific chemical releasesfrom a safety and environmental perspective.

Establish an environmental disaster fundthrough the implementation of thepolluter-pays principle

An environmental disaster fund with sufficientrevenue to support such activities as informationmanagement, training, and clean-up for envi-ronmental incidents should be established. Fundscould be raised through an increase in the pollu-tion levy and/or the introduction of environ-mental taxes on toxic chemicals. In addition,increased fines for pollution accidents to coverthe cost of clean-up and compensation should beconsidered another source for the fund.

Establish a chemical managementinformation system

The central government should establish andmaintain comprehensive inventories of all chemi-cals and pollution sources containing information

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxxi

consistent with international standards. The func-tion and effectiveness of the two existing systemsdeveloped by SAWS and MEP separately shouldbe reviewed. Inventories should be consistent,comprehensive, and easily used in public emer-gency prevention and response. A comprehensivelabeling system for chemicals should be estab-lished and applied to all parts of the production,storage, and transportation process.

Strengthen monitoring and public information

In the event of an incident, local environmentand safety authorities should establish appropri-ate additional monitoring to assess the impacton the health and safety of the local communi-ties and the environment. Investigation findingsshould be reported to the central authorities,and a mechanism established to share lessonslearned and introduce new legally binding prac-tices and procedures if necessary. The publichas the right to be informed of the final investi-gation results.

ISSUES FOR THE FUTURE

While this report has addressed a number of crit-ically important issues relating to water resourcemanagement in China, much work remains tobe done. The various studies identified a num-ber of areas where further work is required.Some of these issues for the future—relating toagricultural water, climate change, and strategicassessment and economic analysis for river basinplans and programs—are highlighted below.

Water Efficiency, Food Security, and Rural Development

The case studies have revealed great variationsin the economic value of water by sector andby region, low economic efficiency of agri -cultural water use, and poor cost-effectiveness of underground water withdrawal in North

China. Although the general direction shouldbe to raise water-use efficiency by reducingdemand for water by the agricultural sector,progress is complicated by and associated withvarious issues involving the rights and well-being of the rural population, national foodsecurity, agricultural sector protection, andpoverty alleviation. The central issue is how toreduce rural poverty and secure the nation’sfood supply while at the same time improvingthe efficiency of water use. This issue will requirefurther study.

Climate Change Adaptation

Global warming caused by human activities canbe one of the biggest threats to the natural envi-ronment and human well-being. The scarcityand vulnerability of China’s water system can benegatively affected by climate change, and reme-dial and adaptation measures need to be taken toameliorate these effects. How to fully take intoaccount climate change impacts and mainstreamadaptation measures in the institutional and pol-icy reform of water resource management inChina is an issue for further investigation.

Ecological and Economic Studies of River Basins

Effective application of water managementmeasures—such as water pricing, water alloca-tion and water rights administration, ecologicalcompensation, and water quality managementin a river basin—all depend on good analysisand understanding of the ecosystems and theeconomic value of competing water uses, suchas agriculture, energy, municipal water supply,and flood control in the river basin. In manycases, the important analytical work remains to be done. Developing a sophisticated analy -tical system—using advanced economic, geo-graphic, and ecological tools—is required forsound policy making.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Yxxxii

Development Strategies, Policies,and Plans and Their Long-TermImpacts on Water Scarcity

China’s Environmental Impact AssessmentLaw, effective in 2003, required strategic envi-ronmental assessments (SEAs) for regional andsectoral development plans. These include landuse, water resource management, and waterpollution control plans for river basins. How-ever, such SEAs have rarely been carried outdue to the lack of knowledge, expertise, andcapacity of planning agencies and local envi-ronmental bureaus and research institutes. As aresult, the long-term impacts of these plans onwater scarcity conditions and the natural envi-ronment are in question. This situation has tobe changed.

CONCLUDING REMARKS

There is no doubt that China is facing a majorchallenge in managing its scarce water resourcesto sustain economic growth in the years ahead.This is a daunting task for the Chinese leader-ship, but past experience in China and in othercountries provides some lessons as to the wayahead. Of course, China is unique in many ways,and will have to adapt techniques and policiesdeveloped elsewhere to suit its own circum-stances. But there are grounds for optimism; theChinese, who have demonstrated immense inno -vative capacity in their successful program of eco-nomic reform, can and should take another boldmove in reforming the institutional and policyframework to make it become a world leader inwater resource management.

E X E C U T I V E S U M M A R Y

A D D R E S S I N G C H I N A ’ S W A T E R S C A R C I T Y xxxiii

CHINA’S GREEN CITIES: HOUSEHOLD CARBON DIOXIDE

EMISSIONS AND URBAN DEVELOPMENT

by

Siqi Zheng

Tsinghua University,

Rui Wang

UCLA

and

Matthew E. Kahn

UCLA and NBER

September 14th 2009

Abstract

While there remains considerable debate about the expected costs of global warming, a growing scientific consensus believes that greenhouse gas emissions create significant risks of climate change. It is well understood that China’s population size and ongoing economic development poses major challenges for capping global greenhouse gas emissions. Much of the growth of China’s greenhouse gas production is tied to urbanization. We provide new rankings of China’s major cities with respect to a standardized household’s level of carbon dioxide production. We find that the “greenest” cities based on this criterion are Huaian and Xuzhou while the “dirtiest” cities are Daqing and Mudanjiang. In addition to ranking 74 Chinese cities with respect to their carbon emissions from transportation, electricity consumption, and household heating, we also seek to explain cross city differences in carbon emissions. We find that average January temperature is strongly negatively correlated with a city’s household carbon footprint. We discuss in detail how our results are useful for predicting how ongoing urbanization trends are likely to affect China’s overall per-capita greenhouse gas production and for studying the unintended environmental consequences of China’s current regional growth policies.

* We acknowledge the financial support of Lincoln Institute of Land Policy, and the CIDEG Center of Tsinghua

University. We thank Yi Huo, Yue Liu and Rongrong Ren for their valuable assistance.

2

Introduction

Urbanization in China is rapidly creating a system of diverse cities spanning different regions

of the nation and specialized in different industries. Export-oriented manufacturing concentrates

on the east and south coast, while most resource-dependent industries locate in the north and

west inland. Through facilitating trade, specialization, and knowledge creation and diffusion,

urbanization raises per-capita income. This income growth has offered a huge benefit as

millions of households have exited poverty and begun to enjoy the material well being that

people in the United States and Europe take for granted.

Economic development often has environmental consequences. In the absence of a world

greenhouse gas emissions treaty, neither the United States nor China has an incentive to reduce

its greenhouse gas emissions. In both nations, up to this point, urbanization has raised per-capita

income which in turn has increased greenhouse gas emissions which raises the prospects of more

several global warming.

Cities differ with respect to their per-capita carbon emissions. Today, the majority of

China’s greenhouse gas emissions are produced by its industrial sector but this sector’s share of

total emissions will decline as further economic development takes place. Around the world, the

residential sector’s share of emissions has grown as per-capita income has risen. To date, this

sector’s contribution to China’s total emissions has been under-researched.

In this paper, we rank China’s major cities based on this household level carbon emissions

criterion. We present new evidence on the income elasticity of household demand for

electricity, transportation, and home heating services. These estimates are of direct use for

3

judging how the ongoing economic development will affect natural resource demand. Overall,

electricity use accounts for about 40% of the total household carbon emission, and winter central

heating accounts for another 40%. Residential transportation only produces 10% of total carbon,

which is much lower than that in the US (around 50%). The remaining 10% comes from

domestic fuel (coal, LPG and coal gas). We find that richer cities have significantly higher

household carbon emissions. Carbon emission per household also has a strong negative

correlation with January temperature. This is largely due to the compulsory winter heating policy

which limits the heating entitlement to areas located in northern China. We also find that smaller

cities tend to be greener. In our ranking, most of the top “greenest” cities are located just south to

the compulsory heating geographical border in the coastal provinces (see Figure Four). Those

cities are not entitled to winter heating services while their summers are not that hot. Daqing,

China’s oil capital, is the “brownest” city because of its outstanding energy consumption.

Shanghai (ranks No. 33) is much greener than Beijing (ranks No. 72). Beijing’s relatively high

electricity consumption and carbon from home heating raise its carbon footprint.

By ranking China’s cities on the household carbon production criteria, we offer new insights

concerning the unintended consequences of government policies that facilitate growth in

different regions. Between 1978 and 2006, some major Chinese cities have grown more than

others. Beijing’s population grew by 34% between 1970 and 2000 and Shanghai’s population

grew by 15% over this time period. In contrast, Chongqing grew by 134% between 1970 and

2000. We seek to study what are the greenhouse gas implications of urban growth in major

Chinese cities. We conclude that the new effort to encourage the economic development of

northern cities such as Shenyang and Daqing will have significant carbon implications because

4

such cold regions rely on dirty coal for home heating and coal fired power plants produce these

cities’ electricity.

Urban Development and Environmental Externalities

Why is it useful to know the carbon dioxide emissions that households produce within

different cities throughout China? Glaeser and Kahn (2009) emphasize two major points. First,

if emissions are actually taxed at the appropriate rates then there is no global warming-related

reason for government policy to favor one place over another and no policy-related reason to

care about emissions in different places. Their second point is that these conditions are unlikely

to hold in the real world, and as a result there is an environmental externality associated with

moving into different locales. The size of the externality associated with moving into place A

rather than place B equals the expected emissions in A minus the expected emissions in B times

the social cost of carbon emissions minus the current carbon tax. This result motivates our

empirical strategy of estimating the carbon emissions for a typical family in different geographic

settings throughout China and multiplying those emissions by a plausible estimate of the social

cost of carbon dioxide emissions.

Throughout this paper, we will focus solely on carbon dioxide emissions as our measure of

city “greenness”. In recent work (see Zheng, Kahn and Liu 2009), we have examined how

ambient particulate levels and sulfur dioxide levels vary across 35 major Chinese cities as a

function of city per-capita income and FDI. We are optimistic that more and more of China’s

major cities have been growing past the Environmental Kuznets Curve peak such that increases

in per-capita income are associated with pollution reductions. For example, China is now

5

phasing in Euro IV new vehicle emissions standards in Beijing. This will likely cause smog

reductions with improved transit service and more effective travel demand management.

Below, we will report the correlation of our city ranking based on greenhouse gas emissions with

city rankings based on local pollution criteria.

Measuring Household Greenhouse Gas Emissions in China’s Major Cities

In this paper we will mainly focus on the household consumption sector. Our justification

for this “narrow” focus hinges on two points. Given the importance of the industrial sector, it

has received the bulk of attention. In contrast, the household’s sector has been under-emphasized

in carbon accounting calculations. The household’s share of total per-capita carbon emissions

will grow as China transitions from being a manufacturing economy to being a service economy

and as domestic households become richer and consume more electricity and private

transportation services. In Figure One, we report the time trends in sectoral shares at the national

level (SIQI).

We seek to estimate how much carbon dioxide emissions a Chinese household produces per

year if it lives in one of China’s 74 cities, including all the 35 major cities (all municipalities

directly under the federal government, provincial capital cities, and quasi provincial capital

cities) plus some cities that have enough sample observations. We focus on four major household

sources of carbon dioxide emissions: transportation, residential electricity consumption,

6

residential heating and domestic fuel. The following equation provides an accounting framework

for organizing our empirical work.

Carbon Dioxide Emissions = γ1*Transportation + γ2 *Electricity + γ3*Heating + γ4*Domestic

Fuel (1)

In this equation, transportation represents a vector of activities including liters of annual

gasoline consumed for households who own a car. Gasoline consumption could be high if a

household has a high probability of buying a car, if the household drives a car a lot or if it owns a

fuel inefficient vehicle. A household is more likely to own a car or drive more if it is wealthy or

if it lives in a low population density area far from the public transit network. A second

dimension of the transportation vector is miles traveled on cabs, a third dimension is that on

buses, and a fourth dimension could be that on subways. Note that each of these physical

activities is multiplied by an emissions factor vector defined as γ1. For example, in the case of

gasoline consumption each litter of # 93 gasoline consumed produces 2.226 Kg of carbon

dioxide.

The second term in this equation represents carbon dioxide emissions from residential

electricity consumption. For example, a household who lives in a humid city featuring very

warm summers may run air conditioning both at the house and at work. Such a household would

consume more electricity than a household who lives in a milder climate. The key factor for

determining how much carbon dioxide is produced from a given level of electricity consumption

is the regional area power plants’ average emissions factor (γ2) defined as carbon dioxide

emissions per megawatt hour of power. We know that coal fired power plants have a higher

7

emissions factor than natural gas fired power plants or power plants run on renewable power

such as wind , hydro or solar power. Major Chinese cities differ with respect to their geography.

Some are located in regions that receive more of their power from power plants with a lower

emissions factor. Such cities are likely to be “greener” cities.

The third term in the equation represents emissions from centralized home heating. In the

northern areas of China, the winters are colder and households must use more energy to provide

basic comforts. Most households depend on centralized heating facilities that mostly use coal.

Burning this coal source creates extra carbon dioxide emissions.

The fourth term in the equation is emissions from domestic fuels. This term includes

three components: coal, liquefied petroleum gas (LPG) and coal gas. Coal is the most abundant

but the dirtiest energy source in China due to its carbon intensity and accompanied sulfur and

other heavy metals. LPG and coal gas are distracted from petroleum oil and coal, and are much

cleaner and less carbon intensive by themselves.

As this accounting equation highlights, there are four key pieces to measuring each major

Chinese city’s household level greenhouse gas contribution. Ideally, we would like to estimate

how much transportation services, electricity, household heating fuels and household domestic

fuels a standardized household would consume if it lived in each of the 74 cities in China. Once

we estimate each of these entries and we have information on each city’s average power plant

emissions factor, then we have the raw ingredients to plug into the equation (1) to rank China’s

cities.

This estimation strategy builds on Glaeser and Kahn’s (2009) ranking 66 major U.S cities

with respect to their household carbon dioxide emissions. In their ranking, San Francisco is one

8

of the top cities while Houston is one of the “brownest” cities. San Francisco ranks high because

it is compact, public transit friendly city whose cool summer climate means that households do

not use much air conditioning. Home prices are high in San Francisco so households live in

smaller homes that require less energy use. California’s electric utilities rely on natural gas (a

relatively clean fossil fuel) and this means that San Francisco’s power plant emissions factor is

lower than Houston’s. In contrast, Houston is a car-dependent, humid, sprawled city. Its

residents drive a lot, live in large homes that require plenty of electricity and the power comes

from dirty sources.

Data

We use three separate data sets in this project. The first data set is the high quality micro

data from the Chinese Urban Household Survey (UHS) in the year 2006. This survey is

conducted annually by the Urban Survey Department of the State Statistic Bureau of China. The

survey targets households living in cities and towns for more than half a year. The data collected

from the survey is primarily used for estimating the urban consumer expenditure in GDP and

CPI. The total sample size of the annual UHS we use in this paper is around 25,300 across the 74

cities. We compute the carbon emissions from electricity use, private car, taxi, and three

domestic fuels based on this micro data set. It also provides information on city economic and

demographic variables such as per household income, household size, and age of household head.

Unfortunately, we are unable to obtain enough information from this micro data set to estimate

carbon emissions from buses, city subways and central heating service.

9

The second data set we use is city level data from the China Urban Statistic Yearbooks.

These data provide us with energy consumption information on buses, as well as many city-level

variables as our explanatory variables.

The third data set includes internal data and information on energy consumptions of

central heating and city subways, and various carbon emission factors needed to transform

energy consumptions to carbon dioxide emissions. This data set comes from various sources.

The carbon emission factors of regional power grid come from the Department of Climate

Change of the National Development Research Center of the State Council. The energy

consumption of central heating comes from the Department of Environmental Engineering and

Department of Building Science in Tsinghua University. The authors collected each of the ten

cities’ rail transit system electricity consumption in 2006.

Table One lists the names, definitions, means and standard deviations of key variables.

Table Two reports the summary statistics. The average household of the 74 cities has an annual

income of 40 thousand Yuan. It consumes 1,700 kWh of electricity and spends 130 Yuan on taxi

in 2006. 16.4% of the 25,300 households own cars. Auto ownership in Chinese cities is

undergoing rapid escalation, mainly due to income growth. In 2002, there were 1.5 million

private autos in Beijing, while in May, 2007, this number grew to 3 million, with an annual

growth rate of 12%. Beijing has a car ownership rate of 23% in our data set, which is higher than

that of Shanghai (16.4%). Shanghai’s relatively low car ownership may be the result of its higher

density, parking policy and also its plate quota policy, all of which increase the cost of owning

and/or using private cars. Some small and medium-sized coastal cities in the south have quite

high car ownership rates due to strong private-sector economic activities and household wealth

level, which may not be adequately reflected by the income data. Only 10 Chinese cities have

10

subways. Beijing and Shanghai have more lines. We can see that central heating is non-existent

to southern cities. In Chinese heating system, northern Chinese cities received fixed heating

between November 15 and March 15. Individual households are not able to control the indoor

temperature when central heating is provided, so the energy consumption from heating is mainly

correlated with winter temperature. It is well known that the Chinese heating system is coal-

based and technically inefficient (Almond et. al., 2009).

Using household level micro data for residents of major Chinese cities, we seek to

understand travel behavior, household electricity use and domestic fuel consumption. The unit

of analysis is household j in city k.

Results on Pooled Cross-City Regressions

Our first empirical goal is to study using the UHS micro data, how household

consumption of various carbon intensive resources (such as coal and gasoline) varies with

household demographics. In Table Three, we report six OLS regressions of the form:

Energy Consumption = City Fixed Effects + b1*Income + b2*Household Size + b3*Age of

Household Head + U (2)

For private car and three domestic fuels, we employ a two-stage strategy. This is because that the

use rates of the above four types are significantly different from 100%. In the first stage, we run

a logit choice model (private car as an example):

Own A Car = b1*Income + b2*Household Size + b3*Age of Household Head + U (3)

11

The above equation estimates the probability of car ownership. In the second stage, we

estimate an OLS equation only for the car owners, with the same explanatory variables as in

equation (2), using gas consumption as the independent variable.

It is important to note that these regressions impose the same marginal impact of

household demographics across cities. Below, we will relax this assumption. The payoff of

imposing this coefficient restriction is that it reduces the number of estimated coefficients down

to a manageable set.

As shown in Table Three, some interesting Engel curve estimates emerge. Not

surprisingly, richer households consume more electricity, and they are more likely to take taxi, to

own a vehicle and to consume more gasoline. The income elasticity of electricity consumption is

0.29, while the income elasticities of taxi spending and gasoline consumption are both above one.

Larger households tend to consume more electricity, have higher probability of owning cars, but

spend less on taxi. For domestic fuels, richer people switch from coal and LPG to coal gas. This

is quite understandable as coal gas is transmitted through pipes directly into households, while

the use of LPG and coal are much less convenient and in the case of coal, much dirtier.

We also find evidence that richer urban Chinese households move up the energy ladder

by substituting away from dirty home heating fuels such as coal and increasing consumption of

cleaner fuels such as electricity and coal gas, as one can find in Table Three. These urban China

results are in accord with past household Environmental Kuznets Curve (EKC) work by Pfaff et.

al. (2004). While many environmentalists focus on the fact that richer people in a free market

economy consume a greater quantity of goods, this “energy ladder” fact supports the optimistic

12

claim that richer households also consume a greater quality of products. This has direct

implications for the environment.

Results on the City Specific Consumption Regressions

Similar to Glaeser and Kahn (2009), we use the UHS data to estimate city specific

regressions for household consumption of gasoline, electricity, coal, LPG and coal gas. Each of

these regressions has the same form as those reported in Table Three but in this case, the

regression coefficients are estimated separately for each of the 74 cities.

This estimation yields 74*3 coefficient estimates for each of the six energy use types. To

spare the reader an examination of all of these coefficients, we suppress many of them and focus

on a subset. All of these results are available on request.

In Table Four, we report the income coefficient “b1” from equation (2) for each of the 74

cities for each of our dependent variables. Given that economic growth will continue to take

place in China, the income coefficients in this table are useful for judging which cities will

experience significant resource growth and consumption growth because of income growth.

Richer people use more electricity, drive more, spend more on taxi, and more often use coal gas

instead of coal or LPG, these different margins of “living well” have important environmental

implications.

We document large differences in the role of household income on resource consumption

across cities. For example, the marginal effect of income on electricity consumption in Beijing

is a relatively small 0.163. In contrast, in Zibo this income effect is three times larger (.445).

13

This indicates that income growth in Zibo will have a much larger impact on household

electricity consumption than income growth in Beijing. The carbon impact of such growth

hinges on whether Zibo’s local electric utilities are much cleaner than Beijing’s. If these two

cities have roughly the same power plant emissions factor, then Table Four’s column (1)

provides an estimate of the marginal carbon impact of income growth due to increased electricity

demand. In general, the table highlights the cross-city heterogeneity with respect to income

effects. Shanghai has a similar income elasticity (0.171) with Beijing, while Shenzhen has a

relatively higher elasticity (0.275). Shanghai’s income elasticity of private car fuel consumption

(conditioning on ownership) is 30% larger than Beijing.

An ongoing environmental economics literature has used cross-national data to document the

empirical relationship between per-capita income and carbon dioxide emissions (the

Environmental Kuznets Curve hypothesis). Schmalensee, Stoker and Judson (1998) document

that the marginal effect of real per-capita income on carbon dioxide emissions is always positive

but the slope is steeper for poor nations than for richer nations. Households in richer nations

are consuming more fossil fuels for transportation and are living in large housing structures and

working in buildings that require electricity to operate.

Within nations, geographical areas differ with respect to their resource consumption and

consequent pollution output. Auffhammer and Carson (2008) create a panel data set for 30

Chinese provinces covering the years 1985 to 2004. They find that a province/year’s log of

greenhouse gas emissions is an increasing and concave function of province/year log of per-

capita income.

14

Building Blocks for Ranking City Greenness

To reach an overall ranking of Chinese cities based on carbon dioxide emissions, we then

take the coefficients of city-specific energy consumption regressions and predict carbon dioxide

emissions for a standardized household for each of the seven energy use types. The standardized

household is defined as a household with an annual income of 40,000 Yuan, 3 members and a

household head of 45 years old. Predicting the carbon dioxide emission of a standardized

household, instead of just computing the simple average of that per household, we are able to

correct for household demographics. Our approach allows us to examine the variation of

emissions the same household with a fixed set of demographics will induce if it lives in different

cities, for example, in Beijing or in Huaian. When doing this, we do not correct for housing

characteristics. After all, we are not attempting to estimate emissions assuming that people in

Beijing live in Huaian’s “Southern-Huai -River” small town style homes. Our approach captures

the fact that a standardized household will live its life differently depending on the relative prices

that it faces in different cities.

Household Electricity

We use the city-specific electricity consumption regressions do make predictions. In the

case of household electricity in Shanghai, for example, we estimate:

Log(Electricity Use)= 3.58 + 0.33*Log(Income) + 0.10*Household Size - 0.0005*Age (12.4***) (12.1***) (5.5***) (-0.4)

T-statistics are in parentheses. In this regression, there are 1,018 observations and the R-

squared is 0.199. We then use the above coefficients to predict the annual electricity

15

consumption for a household living in Shanghai, with an income of 40,000 Yuan, 3 members and

a household head of 45 years old. The result is 1494.9 kWh. We then multiply this number with

the electricity conversion factor in Shanghai (0.8154 tCO2/mWh), which is γ2 in Equation (1), to

get the standardized household’s carbon dioxide emission of 1.219 ton from electricity use. This

procedure applies to all the 74 cities.

The electricity conversion factor (power plant emission factor, γ2) is a key parameter

here. It varies by region in China. Seven electricity grids (six regional grids on the Mainland plus

one on the Hainan Island) support most of China’s power consumption. The baseline emission

factors (at both operating margin and build margin) for regional power grids are estimated for

recent years by the Office of National Coordination Committee on Climate Change, a department

within the National Development and Reform Commission.

Taxi Usage

We employ the same procedure to predict a standardized household’s emissions from taxi

use in each of the 74 cities. The dependent variable in the city-specific regression is the

logarithm of gas consumption from taxi use, which is transformed from household taxi spending.

We consider the variations of taxi prices and vacant-cruising rates across cities. After obtaining

the predictions of the standardized household’s taxi gas consumptions in each of the 74 cities, we

use the corresponding conversion factor to get the carbon emissions.

Car usage

For private car, the city-specific two-stage model is used to first predict the probability

that a standardized household owns a car, and then estimate its fuel consumption assuming that

the standardized household owns a car. The overall fuel consumption is:

16

Car fuel consumption = Probability of owning car*Litters of fuel consumed given owning a car

In the case of car usage in Beijing, for example, we first estimate a logit model:

Ln (Prob(Owning a car)/(1- Prob(Owning a car))) =

-15.57 + 1.43*Log(Income) + 0.005*Household Size - 0.025*Age (-10.1***) (10.2***) (0.06) (-4.8***)

Z-statistics are in parentheses. In this regression, there are 2,081 observations. From the

regression we can predict that the standardized household has the probability of 17.9% to own a

car. Then we use car owner observations to run the second-stage OLS regression:

Log( Car Fuel Use)= -10.41+ 1.46*Log(Income) + 0.12*Household Size - 0.02*Age (-5.3***) (8.2***) (1.1) (-3.1***)

T-statistics are in parentheses. In this regression, there are 479 observations and the R-

squared is 0.157. Using the above coefficients, we can predict that the annual car fuel

consumption for a car owner living in Beijing, with standardized demographics, is 86.67 litters.

Thus we can obtain the car fuel consumption of a standardized household in Beijing:

Car fuel consumption = 17.9% *86.67 = 15.5 Liters

We then use the conversion factor to transform fuel consumption to carbon emission.

Bus and Subway

For subway and bus energy consumptions, the UHS expenditure data cannot give us

reliable estimates of the mileage and energy consumed by the sample households. So we have to

turn to aggregate data in China Urban Statistic Yearbooks and from other sources.

17

For bus, from the Yearbooks we know the total numbers of standard buses, LPG buses

and CNG buses. We assume that the bus operating rate is 90%, and every bus travel around 150

kilometers per day. The fuel consumption of a standard bus is 25 liters per 100 km. A LPG (or

CNG) bus is greener and it consumes 75% of the fuel a standard bus does for the same distance.

Thus we can calculate the total fuel consumption of the buses in a city. Divide this by the total

number of households in the city we reach the average fuel consumption per household from bus

usage. We use conversion factor to transform per household fuel consumption to per household

carbon emission.

There are only 10 Chinese cities having subway lines. They are: Beijing, Shanghai,

Guangzhou, Shenzhen, Tianjin, Dalian, Changchun, Nanjing, Wuhan and Chongqing. There is

no public data available on the electricity use of subways, so we rely on internal information

collected by the authors. We use the following formula to obtain the carbon emission per

household from subway usage:

Per household carbon emission from subway in city i =

(Total electricity use from subway in city i* Conversion factor) / Total number of household in

city i (4)

The total carbon emission from transportation sector is the sum of the above four sub-

categories: private car, taxi, bus and subway.

Coal, LPG and coal gas usages and emissions

We apply the same two-stage procedure as private car to predict a standardized

household’s carbon emissions from domestic fuel uses. We have three types: coal, LPG and coal

18

gas. For each type, we first estimate the probability the standardized household uses this fuel

type, then predict the consumption quantity of the fuel users. Multiplying these two together and

then times the corresponding conversion factor yields the carbon emission from this fuel type of

a household with standardized demographics.

Winter Central Heating

China has a unique compulsory heating system. Prior to 1980s, heating was considered a

basic right and the government provided free heating (which is called “central heating”) for

homes and offices, either directly or through state-owned enterprises. The legacy of this system

remains today as many homes and offices continue to receive free heat. Due to budgetary

limitations, the Chinese government limited the heating entitlement to areas located in northern

China. The border is defined by the Huai River and Qinling Mountains. Northern cities receive

unlimited heating between November 15 and March 15, while heating is non-existent to the

south (see more details in Almond, et. al., 2009).

We are not able to obtain a direct measure of heating consumption from the UHS micro

data. First, since many households in northern cities still receive free heating services, there is no

record of heating expenditure in the UHS. Second, due to the unlimited heating supply, no

heating meter has been installed in individual housing units so there is no record of heating

consumption quantity in the UHS either.

We employ an indirect way to estimate the standardized household’s heating energy

consumption in each of the 74 cities. Under the compulsory heating system in China, a

household’s energy consumption from central heating is a simple function of winter temperature

and home floor area size. We obtained the province-specific winter heating conversion factors

19

from Tsinghua University’s Department of Building Science and Department of Environmental

Engineering. The conversion factor gives how much carbon dioxide emission a square meter of

heating space produces in the corresponding province. We have household-level housing unit

size data in the UHS. We employ the same procedure as electricity and taxi to predict a

standardized household’s housing size consumption in each of the 74 cities. The dependent

variable in the city-specific regression is the logarithm of housing size (square meter). The

explanatory variables include household income, household size, and household head’s age. The

standardized household’s carbon dioxide emission from central heating is calculated by:

The standardized household’s carbon emission from central heating

= Its housing size consumption * province-specific conversion factor (5)

Till now, we can plug the predictions of carbon dioxide emissions from all energy use

types into equation (1) to obtain the total carbon emission of a standardized household in each of

the 74 cities. The results are shown in Table Five.

China’s Greenest Cities Based on the Household CO2 Metric

Table Five’s first 9 columns report our estimates for this standardized household in each

of the 74 cities. They are tons of CO2 from electricity consumption, coal consumption, LPG

consumption, coal gas, private vehicle use, tax, bus, rail, and home heating.

The first fact that jumps out is how small China’s major cities carbon emissions are

relative to U.S cities. Glaeser and Kahn (2009) report that the cleanest cities (San Diego and San

Francisco) have average around 26 tons of CO2 per-household, with roughly 50% of the

20

emissions coming from private transportation. Contrast these “green” U.S cities with Beijing

and Shanghai. Shanghai’s average household produces 1.7 tons of carbon and Beijing’s estimate

is 3.5 tons.

Table Five presents our ranking in order from Greenest to Brownest. The top ten cities

are:

Rank City Name Rank City Name 1 Huaian 6 Nanchang 2 Xuzhou 7 Zhenjiang 3 Suqian 8 Cangzhou 4 Taizhou 9 Shuozhou 5 Nantong 10 Haikou

And the bottom ten are:

Rank City Name Rank City Name 65 Taiyuan 70 Yinchuan 66 Huhehaote 71 Shenyang 67 Haerbin 72 Beijing 68 Tianjin 73 Baotou 69 Chifeng 74 Daqing

Figure Four shows the per household carbon dioxide emissions in each of the 74 cities in

a GIS map. Some patterns emerge. Northern cities generally have larger numbers of per

household CO2 emissions. Figure One also shows that a clear negative correlation exists

between a city’s average January Temperature and its carbon footprint. This can be easily

understood since CO2 emission from home heating is a major component of total emission for

the cold winter cities. In our ranking, most of the top “greenest” cities are located just south to

the compulsory heating geographical border in the coastal provinces (see Figure Four). Those

21

cities are not entitled to winter heating services while their summers are not that hot. Daqing,

China’s oil capital, is the “brownest” city because of its outstanding energy consumption.

Home heating in Northern cities may produce excessive CO2 emission because of its

inefficiency. The Chinese heating system is coal-based and highly-subsidized. Most heat is

provided by coal-fired heat-only boilers or combined heat and power generators, which are

inefficient in energy usage compared to electric, gas and oil heating systems in industrial

countries (T.J. Wang et al., 2000; Yi Jiang, 2007).

In Figure Four we can also see that on average the standardized household will produce

higher carbon emissions if it lives in coastal cities versus inland cities. This is the income effect.

Coastal cities are richer and the households there consume more residential energy.

The results reported in Table Five are measured in tons of carbon dioxide per household.

To judge the magnitude of these differences, we use an estimate of $35 per ton as the marginal

social cost of a ton of carbon dioxide. It is conservative relative to the Stern report (2008),

which suggests a cost of carbon dioxide that is twice this amount, but it is considerably more

aggressive than the numbers used by Nordhaus (2007). Tol (2005) is one meta-study that also

suggests that this number may be somewhat too high while our number is in the middle of the

range in Metcalf (2007).1

Given China’s current development, moving the average household from the greenest

city to the brownest would cause a social externality of $144 (35*(5.2-1.1)) per year. As a

fraction of average household income this equals roughly 5% of income which is fairly large.

1 It is relevant to note that carbon tax policy proposals have suggested taxes per ton of carbon dioxide roughly in this range. Metcalf (2007) proposes a bundled carbon tax and a labor tax decrease. As shown in his Figure Six, he proposes that the carbon tax start at $15 per ton (in year 2005 dollars) now and rise by 4% a year. Under this proposal, the carbon tax per ton of carbon dioxide would equal $60 per ton (in year 2005 dollars) by 2050.

22

As household consumption increases with more cars and more air conditioning, we expect these

cross-city differentials in CO2 emissions to rise over time. If the northern cities substitute away

from coal for home heating, or if the richer cities invest more in subways or other forms of

transit, this gap could narrow.2

Zheng, Kahn and Liu (2009) rank China’s major 35 cities with respect to hedonic quality

of life. A major component in their QOL index calculation is city air quality, measured by small

particulate matter, PM10. We calculate the correlation between the 35 cities’ PM10 levels and

our per-household carbon emission. These two sets of rankings have a positive correlation

coefficient of 0.33. This is because that in the cold northern cities, people burn the coal to

produce home and office heating which creates both particulate matter and carbon dioxide

emission.

Understanding Cross-City Differences in the Household Carbon Footprint

Table Six documents significant cross-city differences with respect to carbon production

for a standardized household. We now turn to explaining this variation using city attributes.

Table Six uses the 74 estimates of total carbon emissions and estimates how population, income,

temperature, and urban form associate with the projected total carbon emissions of a standard

household and its major components in different cities. Overall, total carbon emissions of a

standard household will increase in a city with lower January temperature (Figure One) or higher

income level (Figure Two).

2 Northern cities should be aware of the local ambient pollution problems caused by household coal use. After the horrific deaths in the great 1952 London Fog, the city banned home coal use. While households have little incentive to curb their greenhouse gas emissions, the cost of local pollution (caused by coal burning) provides a direct incentive to consider encouraging substitution to cleaner fuels.

23

City average temperature in January is negatively associated with the total emissions of a

standard household, which means a household in the northern cities emits more than its

emissions if it is located in the south, as shown in Figure One. One degree higher in average

January temperature corresponds to 0.03 ton less CO2 emission per household-year. January

temperature’s effect comes primarily from its impact on household heating emissions – one

degree higher in January temperature means 0.107 ton less CO2 emissions from heating. January

average temperature varies considerably across China. The standard deviation of January

temperature of the 74 cities is 8.66 degrees, which means total emissions per household decrease

by 0.26 ton with a standard deviation increase in January temperature.

As one would expect, higher income level increases emissions from use of electricity,

driving, and use of rail transit, but lowers emissions from use of taxi. A standard deviation

increase from the mean of the average household income of the cities is associated with an

increase of 0.25 ton of CO2 in the annual emissions of a standard household. With respect to the

components of the total emissions, a standard deviation in city average income level is associated

with an increase 0.20 ton CO2 in emissions from use of electricity, a much larger 0.96 ton

increase from use of private car and an even higher 1.46 ton from use of rail transit. On the

contrary, increasing average household income by one standard deviation from the mean is

associated with 0.63 ton fewer emissions from use of taxi.

The size of city population does not show a significant impact on the total carbon

emissions from a standard household (Figure Three), but it is positively correlated with

emissions from use of taxi, bus and electricity, and negatively correlated with emissions from use

of private car. A standard deviation increase in the city population size on average will decrease

the annual emissions of a standard household by 0.28 ton from use of car, but increase the

24

emissions by 0.35 ton from use of bus, 0.11 ton from use of taxi, and a much 0.05 ton from use

of electricity. This seems suggest that controlling for income and other variables, larger cities are

more transit oriented and less car dependent.

Urban form has significant impact on transportation emissions. To the surprise of some,

higher average population density decreases emissions from use of taxi and bus. An increase of

1,000 people per square km (about 19% of the sample standard deviation) on average will lower

the emissions per standard household by 0.424 ton from use of taxi and 0.837 ton from use of

bus. This may indicate a higher share of use of non-motorized modes, shorter average travel

distance and/or higher energy efficiency of urban public transportation system per passenger km.

V. Judging the Unintended Environmental Consequences of China’s Regional

Development Policy

By ranking major cities with respect to their household level carbon dioxide production,

our study provides new insights concerning how regional policies are affecting China’s overall

carbon dioxide production. For example, if regions with high per-capita footprints are receiving

substantial government incentives for growth (in population and/or income level), then the

government’s regional policy is working to increase the nation’s greenhouse gas production.

The Chinese government has not translated the importance of regional equity to coherent

policy approaches until recently. Three regional development policies are worth to note here.

The first is the Western Development Program launched in 1999 to give preference to west and

25

inland provinces. Two region-specific components of the strategy are infrastructure investment

and industrial structure adjustment from heavy and defense industries to consumer goods

industries, as pointed out by Chow (2002: 174). The second is the Northeast Revitalization

Program initiated in 2003. The most prosperous provinces of the Mao years that benefited from

the emphasis on heavy industry (Liaoning, Jilin and Heilongjiang) have suffered significant

relative decline under the reforms, struggling with high unemployment, aging industry and

infrastructure, and social welfare bills that are increasingly difficult to meet (Saich 2001: 149). If

the Western Development Program focuses on both urban and rural sectors, the Northeast

Revitalization Program is mostly focused on reinventing the declined cities. And the third and

most recent one is the further integration and development of Beijing- Tianjin and the

surrounding region in Hebei Province – a strategic move to speed up the development of this

northern mega-region led by the national capital to catch up the Yangtze and Pearl River Deltas

in the south. The Olympics certainly helped in terms of massive public investment in

infrastructure and environmental improvement, while the location of national political power will

continue to draw physical and human capital to grow the cities in this region.

However, our results indicate that northern colder cities impose larger CO2 costs when

they grow. As one can see from the above description on three major regional policies, the

central government’s subsidy toward the northern parts of China will certainly help boost income

levels in these regions and attract more rural migrants, both of which can increase China’s total

urban households’ carbon emissions compared to a regional policy less favoring this region. This

may become the hidden cost of China’s policy aiming at regional balance of economic

26

development level as the northern part of China are less developed.3 How to trade the economic

development equity against GHG emissions equity could be a new question facing the Chinese

leaders.

VI. Conclusion

We have used high quality micro data to rank China’s major cities with respect to

household carbon dioxide emissions. We find that the “greenest” cities based on this criterion

are Huaian and Xuzhou while the “dirtiest” cities are Daqing and Mudanjiang. The cross city

differential in the carbon externality is “large”. At $40 per ton of damage from carbon dioxide,

moving a standardized household from Daqing to Huaian would reduce the externality by

roughly $160 per year. This differential is mainly generated by cross city differences in climate,

central heating policy, regional electric utility emissions factors and urban form. Our results are

useful for predicting how ongoing urbanization trends are likely to affect China’s overall per-

capita greenhouse gas production. The income elasticities we report in Table Four are useful for

forming better projections of the likely dynamics of future emissions increases associated with

ongoing Chinese urbanization. We have also documented the relevant point concerning the

unintended environmental consequences of China’s current regional growth policies that favor

3 Such hidden cost may be further increased if urbanization leads to more reliance on local and regional energy sources. Fossil fuels are predominantly in the north, which has 90% of the oil and 80% of the coal reserves. Hydropower remains the vast majority of renewable power (roughly 17% of the total electricity) generated in China. Roughly two thirds of the hydropower is located in the south west region of China. In contrast to the distribution of fossil and hydro energy, the east and south coastal areas have very little energy resources. Of course, the northern part of the country has some potential in increasing its small but increasing share of renewable energy. Wind power is concentrated in the northern provinces and the east and south coasts. The seasonal fluctuation of wind power are complementary to hydropower, but the geographical distribution of land areas with rich wind power potential is to a large extent different from that of the demand for power. In addition, international energy trade may help reduce the northern cities’ carbon footprint. If the northern cities can import natural gas from Russia to substitute their coal use to a significant level, the geography of urban carbon foot print will be different.

27

the northern region. Future research should consider how to design incentives to reduce this

externality effect.

28

References

Almand, Douglas, Yuyu Chen, Micheal Greenstone and Hongbin Li. 2009. Winter Heating or Clean Air? Unintended Impacts of China’s Huai River Policy. Working paper. MIT Department of Economics.

Auffhammer, M., Carson, R T. 2008. Forecasting the Path of China's CO2 Emissions Using Province Level Information, Journal of Environmental Economics and Management 55(3), 229-247.

Brown, M. A., Logan, E., 2008. The Residential Energy and Carbon Footprints of the 100 Largest Metropolitan Areas, Georgia Insittute of Technology Scool of Public Policy, Working Paper 39.

Chow, Gregory C. 2002. China’s Economic Transformation. Malden, Mass: Blackwell.

Glaeser, Edward L, and Matthew E. Kahn 2008. The Greenness of Cities. NBER Working Paper #14238.

Golob, T. F., Brownstone, D., 2005. The Impact of Residential Density on Vehicle Usage and Energy Consumption, University of California Energy Institute, Policy & Economics Paper EPE-011

Holtz-Eakin, Douglas and Thomas M. Selden, Stoking the fires? CO2 emissions and economic growth, Journal of Public Economics, Volume 57, Issue 1, May 1995, Pages 85-101.

Jiang, Yi. 2007. “Promoting Chinese Energy Efficiency.” China and the World Discuss the Environment, http://www.chinadialog.net, June 25, 2007.

Kahn, M. E., 2006 Green Cities: Urban Growth and the Environment, Brookings Institution Press, Washington, DC.

Metcalf, G., 2007. A Proposal for a U.S. Carbon Tax Swap. Brookings Institution. Hamilton Project Working Paper.

Naughton, Barry. 2007. The Chinese Economy: Transition and Growth. Cambridge, Mass: the MIT Press.

Nordhaus, W. D., 2007. A Review of the Stern Review on the Economics of Climate Change, Journal of Economic Literature 45(3), 686–702.

Pfaff, Alexander. S. P., Shubham. Chaudhuri, and H. Nye. 2004. “Household Production and Environmental Kuznets Curves: Examining the Desirability and Feasibility of Substitution.” Environmental and Resource Economics 27, no. 2: 187–200.

Saich, Tony. 2001. Governance and Politics of China. New York, NY: Palgrave.

Stern, N., 2008. The Economics of Climate Change, American Economic Review 98(2), 1-37.

29

Tol, R., 2005. The Marginal Damage Costs of Carbon Dioxide Emissions: An Assessment of the Uncertainties of Energy Policy.

Wang, T.J., L.S. Jin, Z.K. Li, and K.S. Lam. 2000. “A modeling study on acid rain and recommended emission control strategies in China.” Atmospheric Environment, 34(26): 4467-4477.

Weitzman, M. L., 2007. A Review of the Stern Review on the Economics of Climate Change, Journal of Economic Literature 45(3), 703–724.

Zheng, Siqi, Matthew E. Kahn and Hongyu Liu. Towards a System of Open Cities in China: Home Prices, FDI Flows and Air Quality in 35 Major Cities , NBER Working Paper #14571)

30

Table One: Definitions of key variables

Variable Name Definition Unit Mean Std.dev.

Household level variables

ELECQ Household’s electricity consumption in 2006 kWh 1,699 1,089

CAR_USE Binary: 1=own a car, 0=otherwise. In 2006. 0.164 0.370

CARQ Household’s fuel consumption by driving car in 2006 Liter 178.8 202.9

TAXIQ Household’s fuel consumption by taking taxi in 2006 Liter 13.2 21.2

COAL_USE Binary: 1=use coal as domestic fuel, 0=otherwise. In 2006.

0.092 0.289

COALQ Household’s coal consumption in 2006 kg 760.4 654.7

LPG_USE Binary: 1=use LPG (liquefied petroleum gas) as domestic fuel, 0=otherwise. In 2006.

0.419 0.493

LPGQ Household’s LPG consumption in 2006 kg 82.9 55.4

COALGAS_USE Binary: 1=use coal gas as domestic fuel, 0=otherwise. In 2006.

0.582 0.493

COALGASQ Household’s coal gas consumption in 2006 m3 252.9 189.5

HHSIZE Household size person 2.9 0.8

AGE Household head’s age year 50.5 11.9

INCOME Annual household income yuan/household

39,639 23,056

HSIZE Housing unit size square meter 74.271 33.789

City-level variables

CINCOME City average household income yuan 37977 10273

POP City population 1000 persons 2,556 2,652

DENSITY City population density 1000persons/km2

13.4 5.3

JAN_TEMP Average temperature on January 0.46 8.66

31

JULY_TEMP Average temperature on July 27.21 2.65

Table Two: Summary of city-level data in 2006

City Obs Avg. income (Yuan)

Avg. electricity use (kWh)

Car ownership (%)

Avg. gas use (L)

Avg. taxi expenditure (Yuan)

Std. bus mileage (1e3 km)

LPG/CNG bus mileage (1e3 km)

Rail electricity use (1e3 kWh)

Heated floor space ( m2)

coal use rate (%)

Avg. coal use (kg)

LPG use rate (%)

Avg. LPG use (kg)

Coal gas use rate (%)

Coal gas use (m3)

Housing unit size (m2)

All 25330 40,058 1699.0 16.4 178.8 130 99,926 9.2 760.4 41.9 82.9 58.3 252.9 64.9 Beijing 2081 55,718 2286.2 23.0 309.3 255 758,835 174,926 244,907 222,180 7.4 1159.9 22.1 97.2 74.8 233.6 65.8 Tianjin 1554 40,441 2151.4 11.0 204.1 176 356,505 2,119 54,741 109,680 8.8 808.2 7.9 73.9 90.9 148.5 76.1 Shijiazhuang 301 32,201 1470.7 7.3 126.5 120 75,785 29,466 38,280 5.3 1150.6 53.2 85.4 53.2 340.0 68.4 Tangshan 200 37,647 1137.7 16.0 199.4 165 117,915 21,640 0.0 0.0 0.0 100.0 406.6 68.9 Qinhuangdao 200 29,472 1132.1 23.5 141.5 103 39,075 9,120 4.5 1411.1 65.0 76.2 39.0 368.3 73.4 Handan 200 28,633 1121.4 12.0 88.3 75 63,023 10,030 9.5 396.4 10.0 28.8 95.0 361.8 93.0 Cangzhou 150 30,080 1152.5 18.0 69.8 134 19,217 5,830 4.0 1200.0 92.0 83.1 38.0 112.0 70.4 Taiyuan 310 32,039 1293.2 8.1 159.7 102 99,683 37,850 3.2 528.0 7.1 83.8 92.9 447.6 85.3 Shuozhou 150 31,747 853.6 26.0 74.4 95 4,977 2,470 22.0 1597.0 48.7 60.2 57.3 131.1 72.9 Huhehaote 400 37,383 1293.2 11.8 188.8 189 10,200 44,200 12,030 9.0 1377.0 45.8 72.4 53.0 247.3 75.5 Baotou 400 40,109 1361.4 19.0 111.9 209 39,962 13,945 23,680 10.8 1215.0 37.5 67.2 42.5 242.3 85.9 Wuhai 150 34,596 1162.8 65.3 64.4 135 20,548 2,600 31.3 1567.9 27.3 54.4 13.3 214.8 75.9 Chifeng 200 26,572 1141.7 16.5 59.2 103 16,803 1,183 12,210 8.5 1386.5 89.0 46.5 1.5 42.7 72.9 Tongliao 150 28,275 1680.5 26.7 75.7 193 7,983 4,660 8.7 1923.8 78.0 85.1 5.3 60.8 65.8 Shenyang 502 31,190 1343.2 4.0 241.7 180 196,410 54,695 117,776 0.4 175.0 5.8 69.3 95.6 175.7 68.8 Dalian 508 37,514 1242.8 3.0 147.2 239 207,842 17,987 60,350 0.2 1275.0 11.4 56.9 94.1 344.0 71.2 Liaoyang 200 27,259 1182.8 5.5 60.7 124 18,774 12,630 3.5 62.7 56.0 102.7 63.0 87.6 77.5 Changchun 322 33,444 1224.8 6.8 206.8 210 86,823 100,028 14,680 49,480 5.3 1517.6 5.6 58.4 92.5 236.8 78.7 Jilin 300 29,074 1366.9 3.0 196.4 226 289,195 138,364 21,610 0.7 177.0 88.0 92.6 17.7 159.8 65.7 Haerbin 541 31,125 1571.0 3.9 132.7 162 124,666 102,837 55,750 1.3 426.9 16.5 108.2 88.5 403.5 64.8 Qiqihaer 300 22,989 1012.3 7.7 62.1 108 46,466 20,300 1.0 1666.7 18.3 106.1 90.3 198.8 86.1 Daqing 200 37,427 1708.0 4.0 72.1 228 91,011 29,610 0.0 0.0 74.0 133.9 7.5 80.2 64.6 Mudanjiang 200 22,349 1254.9 6.5 100.3 165 6,701 15,226 13,590 6.0 1612.5 74.0 48.9 3.5 104.1 61.1 Shanghai 1018 56,717 1778.3 16.4 183.6 242 818,852 13,846 427,302 0.0 0.0 3.4 64.1 97.0 403.6 69.9 Nanjing 821 47,448 1960.7 15.1 197.9 116 207,349 45,234 67,180 1.3 462.7 54.8 77.6 50.5 204.0 75.2

33

Wuxi 301 48,705 2057.3 18.9 194.8 127 130,283 8,968 2.7 259.4 35.2 76.0 65.4 350.0 66.0

301 35,331 1278.4 9.6 93.1 82 84,162 28.6 583.8 33.2 66.9 61.1 288.4

81.0

Changzhou 301 42,536 1746.6 42.5 142.8 104 65,240 13,994 7.0 458.8 52.8 81.4 59.8 161.6 83.1 Suzhou 300 49,096 1960.2 22.3 173.6 117 101,309 4.0 581.3 32.7 86.1 69.3 231.9 82.3 Nantong 205 38,890 1524.7 5.4 179.0 61 37,498 3.4 284.0 29.3 81.4 82.4 311.7 79.6 Huaian 200 29,379 1289.0 16.0 51.2 60 28,974 42.0 403.5 70.5 56.6 27.0 87.4 91.9 Yangzhou 200 36,080 1655.5 27.5 70.1 71 36,562 10.5 250.5 49.0 64.5 58.0 278.7 81.4 Zhenjiang 200 40,896 1593.5 10.0 52.1 98 36,119 24.0 532.0 40.5 54.2 69.0 149.6 87.9 Taizhou 200 33,580 1468.5 40.5 88.8 49 11,727 22.5 539.0 60.5 62.1 28.0 71.7 117.9 Suqian 207 26,626 1038.6 21.3 54.0 42 16,162 43.5 503.0 87.4 48.3 1.4 26.0 78.4 Hangzhou 614 51,432 2286.5 18.1 243.7 96 234,795 4.6 262.1 66.9 81.1 42.3 128.3 69.2 Ningbo 406 48,805 1768.0 15.0 262.4 90 130,776 2.2 351.7 78.1 112.3 27.6 60.9 87.3 Wenzhou 204 54,042 2836.0 43.1 335.7 195 83,620 0.5 50.0 84.8 115.4 16.2 34.4 76.0 Jiaxing 150 44,866 1716.0 38.0 186.4 87 32,374 5.3 197.3 80.0 98.0 25.3 55.0 80.7 Huzhou 200 42,087 1727.5 21.0 97.4 90 29,910 5.5 456.4 87.0 89.9 18.5 163.5 74.2 Shaoxing 200 49,815 1568.6 9.0 150.6 68 36,119 27.5 224.7 42.5 77.5 71.0 119.5 86.5 Jinhua 153 43,932 1514.1 37.3 137.5 119 33,162 24.2 238.2 85.0 71.7 16.3 37.3 96.4 Quzhou 150 37,848 1415.4 16.7 128.1 88 26,461 10.7 753.4 76.0 101.8 28.7 389.7 104.5 Taizhou 150 52,123 1914.5 39.3 263.3 86 17,591 12.7 88.9 84.0 118.3 18.0 50.8 88.1 Lishui 150 44,803 1881.5 48.7 152.8 50 4,583 14.0 187.5 92.7 90.8 4.0 30.1 71.1 Hefei 410 31,293 1624.5 5.6 106.4 222 96,776 21,927 27.8 417.3 58.8 83.7 48.5 177.6 68.3 Huainan 411 31,410 1255.3 6.3 84.9 232 42,623 28.0 575.4 46.7 51.5 64.7 325.6 82.0 Fuzhou 303 44,596 2804.7 24.8 144.8 76 91,750 0.3 30.0 68.6 117.8 36.3 120.3 87.0 Xiamen 201 52,711 2776.8 25.9 172.1 121 126,883 6.5 584.3 67.7 93.1 34.8 106.3 70.8 Nanchang 300 28,905 1537.2 1.7 50.0 38 108,159 0.3 150.0 75.0 85.4 38.3 250.8 75.6 Jinan 416 43,605 1573.0 31.3 133.2 185 138,561 46,269 23,680 25.7 1140.4 52.9 50.3 39.4 143.8 67.6 Qingdao 407 43,263 1668.9 10.3 236.3 172 177,291 20,646 18,370 20.4 1050.2 48.6 55.9 63.9 222.6 93.9 Zibo 150 38,050 1299.7 42.7 79.6 149 82,339 4,977 18,300 11.3 1205.3 66.7 81.2 29.3 148.9 72.8 Yantai 200 40,448 1248.6 26.0 72.6 184 73,962 296 15,970 8.5 1273.5 55.0 61.4 52.0 205.5 81.0

34

Rizhao 102 31,736 1412.1 60.8 136.6 139 35,527 5,670 22.5 1163.2 59.8 47.9 8.8 67.3 80.4 Zhengzhou 462 35,124 1508.3 7.4 97.6 52 87,857 70,365 10,610 14.7 991.9 30.3 87.2 79.9 241.4 82.6 Luoyang 307 32,683 1402.5 22.5 87.7 72 40,948 3,200 25.1 612.3 75.9 82.4 22.5 325.2 78.2 Wuhan 531 34,558 2092.8 5.5 222.0 135 201,091 83,225 11,284 7.7 503.4 50.5 109.9 63.8 253.7 79.5 Changsha 409 38,758 1918.3 16.4 207.9 227 131,564 2,562 8.6 604.3 81.4 97.1 31.5 258.3 77.2 Guangzhou 304 59,751 2361.1 27.0 242.3 166 209,271 316,444 185,417 0.7 66.0 69.1 112.6 50.3 260.7 95.5 Shenzhen 101 82,429 2893.5 42.6 466.6 219 100,485 0.0 0.0 53.5 140.4 59.4 79.2 87.8 Zhuhai 101 55,577 2039.5 36.6 295.0 162 57,504 5.0 118.8 97.0 137.3 12.9 48.4 74.2 Nanning 202 32,663 1591.5 39.1 136.2 65 107,567 12.4 319.3 92.6 107.4 6.4 70.0 94.8 Haikou 306 35,693 1580.0 33.7 193.4 64 42,377 4.2 577.3 85.6 97.3 15.0 139.4 75.5 Chongqing 308 35,571 2051.0 3.2 155.7 119 40,504 349,113 26,120 1.0 783.3 2.6 67.9 98.4 341.2 76.4 Chengdu 430 36,138 1821.8 20.9 262.3 116 3,597 32,669 2.1 1596.7 5.1 103.9 94.4 348.4 80.6 Mianyang 200 27,587 1394.5 9.5 188.8 100 2,661 33,556 0.5 150.0 4.5 71.3 97.5 279.6 63.5 Guiyang 316 33,034 2155.4 15.5 128.0 99 77,559 29.1 848.1 19.0 58.7 69.0 292.2 92.4 Kunming 600 30,445 1522.3 14.3 181.1 45 123,582 77,756 10.3 372.8 41.7 74.9 51.3 399.0 62.5 Xi'an 366 31,172 1396.1 6.6 50.2 108 135,309 131,466 18,390 20.8 678.6 38.5 58.2 53.3 247.2 61.9 Lanzhou 321 25,819 911.7 3.1 47.4 74 4,139 91,159 22,140 6.5 667.2 44.9 61.6 59.5 166.5 73.8 Xining 300 27,781 1507.6 4.7 111.8 113 86,379 130 17.0 1273.6 14.3 63.7 5.0 392.7 72.0 Yinchuan 314 27,870 1334.8 15.3 71.7 191 40,455 12,713 21,230 10.5 467.9 66.2 42.7 20.7 233.5 69.4 Wulumuqi 402 29,294 1053.5 4.0 37.4 136 12,812 196,755 29,000 0.7 500.0 26.9 54.1 61.7 233.5 64.9

35

Table Three: Cross City Regressions with City Fixed Effects Using Micro Data

Dependent variable

log(ELECQ) log(TAXIQ) CAR_USE log(CARQ) COAL_USE log(COALQ)

Model OLS OLS Logit OLS Logit OLS

beta t-stat beta t-stat beta z-stat beta t-stat beta z-stat beta t-stat

log(INCOME) 0.289 39.21*** 1.929 54.95*** 1.161 14.18*** 1.084 22.44*** -0.891 -7.86*** -0.728 -1.6

HHSIZE 0.06 11.77*** -0.287 -11.83*** 0.084 2.43** -0.019 -0.61 0.295 8.57*** 0.08 3.0***

AGE 0.0009 2.62*** -0.018 -11.37*** -0.039 -14.53*** -0.021 -9.53*** 0.021 6.39*** 0.01 5.12***

constant 3.988 51.1*** -13.642 -36.75*** -12.257 -13.38*** -6.418 -12.29*** 4.942 4.41*** 6.085 13.36***

City fixed effects

yes yes clustered by dcode yes clustered by dcode yes

Obs 25328 25328 25328 4146 25328 2338

R2 0.22 0.234 0.094 0.256 0.047 0.327

Dependent variable

LPG_USE log(LPGQ) COALGAS_USE log(COALGASQ) log(HSIZE) (for estimating heating)

Model Logit OLS Logit OLS OLS

beta z-stat beta t-stat beta z-stat beta t-stat beta t-stat

log(INCOME) -0.392 -3.04*** -0.028 -2.02** 0.584 4.66*** 0.11 8.97*** 0.265 61.36***

HHSIZE 0.065 1.5 0.097 10.26*** -0.051 -1.19 0.124 14.89*** 0.025 8.26***

AGE -0.007 -2.13** 0.008 12.75*** 0.013 4.22*** 0.011 19.77*** 0.0003 1.53

constant 3.903 3.06*** 3.75 25.58*** -6.246 -4.86*** 3.146 24.17*** 1.367 29.83***

City fixed effects

clustered by dcode yes clustered by dcode yes yes

Obs 25328 10601 25328 14753 25328

R2 0.01 0.161 0.023 0.308 0.222

36

Table Four: Income Effects from Individual City Regressions

Dependent variable log(elecq) log(taxiq) Car_use log(carq) log(hsize)

Model OLS OLS Logit OLS OLS

beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat

Beijing 0.163 3.631*** 1.432*** 10.230*** 1.463 8.213*** 1.073 14.489** 0.218 11.158***

Tianjin 0.402 9.811*** 1.499*** 8.470*** 1.526 6.044*** 1.067 16.112*** 0.416 19.992***

Shijiazhuang 0.321 2.605** -0.493 -1.011 1.894 3.155*** 1.242 7.824*** 0.312 6.087***

Tangshan 0.197 2.147** 2.227 4.043*** 1.53 2.915*** 1.071 5.659*** 0.328 6.456***

Qinhuangdao 0.138 1.59 0.532 1.542 1.174 2.396 0.832 5.412*** 0.137 4.175***

Handan 0.144 0.944 -0.657 -0.986 1.792 1.486 0.873 3.536*** 0.356 6.546***

Cangzhou 0.333 2.734*** 1.013 1.820* 0.567 0.879 1.113 4.995*** 0.162 3.139***

Taiyuan 0.08 1.221 1.681 3.365*** 1.194 1.929* 0.433 4.250*** 0.110 4.948***

Shuozhou 0.312 1.857* 0.476 1.195 1.054 2.398** 0.705 3.169*** 0.038 0.706

Huhehaote 0.163 1.636 1.173 3.398*** 2.521 4.780*** 1.129 9.523*** 0.221 7.348***

Baotou 0.167 1.488 1.215 3.253*** 0.574 1.218 1.267 8.728*** 0.191 5.684***

Wuhai 0.507 1.835* 0.411 0.804 0.719 1.712* 0.65 2.052** 0.487 6.074***

Chifeng 0.243 1.487 1.354 2.375** 2.023 3.581*** 0.823 3.879*** 0.247 4.507***

Tongliao 0.427 3.540*** 0.696 2.089** 1.172 3.082*** 1.122 7.579*** 0.325 8.107***

Shenyang 0.242 3.925*** 2.004 3.945*** 1.1 1.800* 1.113 11.125*** 0.322 10.696***

Dalian 0.344 6.375*** 1.571 2.905*** 0.095 0.161 1.362 12.211*** 0.264 8.803***

Liaoyang 0.041 0.512 1.675 2.132** 1.908 1.551 0.731 3.405*** 0.247 5.762***

Changchun 0.274 2.814*** 2.026 3.938*** 1.454 2.830*** 1.11 8.575*** 0.135 3.718***

Jilin 0.239 3.641*** 0.352 4.504*** 0.127 5.168***

Haerbin 0.373 4.859*** 0.269 0.648 0.882 1.341 1.141 11.373*** 0.358 12.992***

Qiqihaer 0.119 2.471** 1.14 2.385** -0.105 -0.127 0.399 5.145*** 0.123 5.649***

Daqing 0.219 1.379 0.558 3.378*** 0.097 3.881***

Mudanjiang 0.346 3.151*** 0.823 1.697* 1.253 2.245** 1.042 8.037*** 0.193 5.898***

Shanghai 0.171 4.932*** 1.34 6.738*** 1.983 8.036*** 1.387 15.294*** 0.485 15.975***

Nanjing 0.242 5.548*** 1.388 6.707*** 0.839 2.810*** 1.028 14.572*** 0.282 16.963***

37

Wuxi 0.279 4.084*** 1.382 4.137*** 2.328 4.884*** 1.058 8.493*** 0.304 7.811***

Xuzhou 0.251 3.674*** 1.182 3.339*** -0.31 -0.473 0.955 9.608*** 0.264 7.468***

Changzhou 0.376 4.943*** 0.597 2.466** 0.321 1.306 0.639 5.190*** 0.191 4.728***

Suzhou 0.32 4.264*** 0.927 3.089*** 0.968 2.899*** 1.001 7.267*** 0.416 9.257***

Nantong 0.213 2.156** 1.291 2.053** 3.545 2.282** 0.696 4.790*** 0.214 5.536***

Huaian 0.127 1.159 0.869 2.392** 1.269 2.181** 0.951 6.898*** 0.242 5.290***

Yangzhou 0.355 3.224*** 1.342 3.468*** 0.642 1.302 0.936 5.830*** 0.284 5.234***

Zhenjiang 0.375 3.895*** 1.438 2.750*** -0.007 -0.008 1.228 8.364*** 0.396 9.079***

Taizhou 0.29 2.835*** 0.431 1.551 0.928 3.182*** 0.553 4.361*** 0.254 4.479***

Suqian 0.238 2.071** 1.473 4.453*** 0.22 0.458 0.591 6.935*** 0.093 2.006**

Hangzhou 0.336 7.124*** 1.228 5.246*** 1.122 3.579*** 0.732 7.142*** 0.285 10.060***

Ningbo 0.13 2.942*** 1.231 4.391*** 0.678 2.368** 0.591 6.012*** 0.217 8.958***

Wenzhou 0.241 2.984*** 1.362 4.383*** 0.853 5.133*** 0.974 6.028*** 0.340 5.679***

Jiaxing 0.222 2.463** 1.381 3.214*** 0.383 1.056 0.752 3.862*** 0.227 4.645***

Huzhou 0.232 2.726*** 1.313 3.412*** 1.121 2.991*** 0.676 4.873*** 0.321 7.426***

Shaoxing 0.348 3.754*** 1.445 2.435** 0.684 0.471 0.847 5.516*** 0.326 6.775***

Jinhua 0.276 3.316*** 1.133 2.998*** 0.128 0.351 0.713 4.370*** 0.256 4.802***

Quzhou 0.2 2.214** 0.165 0.422 1.323 3.180*** 0.664 3.952*** -0.043 -0.620

Taizhou 0.326 3.100*** 0.992 2.519** 1.028 2.783*** 0.633 3.107*** 0.242 3.676***

Lishui 0.235 2.536** 1.676 4.326*** 1.03 3.196*** 0.389 3.480*** 0.202 4.431***

Hefei 0.097 1.308 2.597 4.425*** 0.439 0.42 1.198 8.307*** 0.225 6.235***

Huainan 0.221 2.425** 0.922 1.598 0.75 0.686 1.475 10.256*** 0.245 5.470***

Fuzhou 0.151 2.390** 1.275 3.781*** 0.645 1.511 0.748 4.691*** 0.250 6.845***

Xiamen 0.161 2.415** 0.538 1.623 1.638 3.406*** 0.956 4.853*** 0.328 5.836***

Nanchang 0.022 0.197 0.58 3.691*** 0.205 5.044***

Jinan 0.122 1.433 0.894 4.157*** 1.494 5.489*** 1.055 9.067*** 0.303 8.586***

Qingdao 0.374 4.117*** 2.168 4.960*** 1.56 2.666*** 0.928 7.000*** 0.275 8.525***

Zibo 0.445 2.112** 0.062 0.14 0.521 0.842 1.257 5.365*** 0.302 4.871***

Yantai 0.563 4.053*** -0.135 -0.28 1.549 2.338** 0.929 4.198*** 0.198 4.663***

Rizhao 0.403 2.511** 0.951 2.236** 1.578 4.102*** 0.891 4.782*** 0.190 5.014***

38

Zhengzhou 0.259 3.951*** 0.975 2.391** 0.457 0.742 0.492 4.413*** 0.219 6.540***

Luoyang 0.127 1.314 1.268 3.527*** 1.28 2.845*** 0.785 4.945*** 0.315 6.934***

Wuhan 0.246 5.273*** 2.823 5.385*** 1.779 2.882*** 0.933 7.340*** 0.269 8.347***

Changsha 0.345 5.948*** 1.206 4.120*** 1.226 3.336*** 1.146 10.947*** 0.322 10.448***

Guangzhou 0.185 2.213** 1.853 4.645*** 0.935 1.963** 1.196 6.611*** 0.333 6.615***

Shenzhen 0.275 4.076*** 1.581 2.691*** 0.144 0.682 0.472 1.639 0.158 3.364***

Zhuhai 0.098 1.001 1.711 3.263*** 1.416 3.761*** 0.886 3.423*** 0.085 1.747*

Nanning 0.167 2.916*** 0.829 3.300*** 0.653 2.232** 0.842 7.360*** 0.316 9.342***

Haikou 0.256 3.917*** 1.388 5.474*** 1.145 4.332*** 0.76 7.882*** 0.294 5.875***

Chongqing 0.229 4.134*** 1.255 1.705* -0.77 -0.534 1.055 6.016*** 0.336 7.458***

Chengdu 0.291 8.148*** 1.635 6.881*** 1.915 6.694*** 0.971 9.661*** 0.340 12.646***

Mianyang 0.287 3.775*** 2.458 3.610*** 2.837 2.349** 0.93 7.323*** 0.296 6.634***

Guiyang 0.203 3.138*** 1.868 5.220*** 1.344 3.588*** 0.875 6.822*** 0.358 10.284***

Kunming 0.307 5.270*** 1.077 4.538*** 0.599 2.531** 0.457 6.287*** 0.315 8.327***

Xi'an 0.395 5.133*** 1.238 2.894*** 1.134 1.677* 1.281 11.093*** 0.300 8.616***

Lanzhou 0.249 3.051*** 0.322 0.553*** 0.269 0.305 0.89 8.051*** 0.218 6.831***

Xining 0.285 2.423** 1.419 2.323** 2.075 1.812* 1.084 9.025*** 0.140 5.254***

Yinchuan 0.167 2.045** 0.473 1.641 0.804 2.019** 0.825 7.339*** 0.137 6.053***

Wulumuqi 0.349 3.068*** 0.965 1.684* -0.832 -0.745 1.215 10.001*** 0.136 5.441***

Dependent variable Coal_use log(coalq) lpg_use log(lpgq) coalgas_use log(coalgasq) Model Logit OLS Logit OLS Logit OLS

beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat

Beijing -0.344 -1.806 -0.058 -0.292 -0.447 -3.679*** -0.208 -2.727*** 0.514 4.415*** 0.144 2.918*** Tianjin -1.427 -7.558** 0.127 0.551 -0.959 -5.155*** 0.251 1.675* 0.516 3.018*** 0.164 4.410***

Shijiazhuang -1.361 -2.523** -0.606 -1.28 -0.387 -1.424 -0.088 -0.632 0.552 1.996 0.06 0.414 Tangshan 0 0 -0.072 -1.026

Qinhuangdao -0.293 -0.487 0.578 1.742* -1.095 -3.385*** -0.385 -3.291*** 1.023 3.296*** 0.282 1.431 Handan -0.3 -0.42 -0.766 -0.644 -1.192 -1.692* 0.876 1.809* -0.109 -0.129 0.055 0.431

Cangzhou -0.444 -0.632 0.047 0.379 0.335 0.882 0.114 0.451

Taiyuan -0.223 -0.893 1.152 1.091 -0.225 -1.108 -0.194 -0.412 0.093 0.361 -0.014 -0.272

39

Shuozhou -1.816 -4.031*** -0.003 -0.01 -0.934 -2.630*** -0.099 -0.468 1.646 4.030*** 0.53 2.298** Huhehaote -0.978 -2.986*** 0.41 1.16 -0.223 -1.164 0.093 0.835 0.235 1.236 0.103 0.886

Baotou -0.51 -1.421 -0.323 -1.065 -0.105 -0.442 0.03 0.159 0.98 3.788*** -0.042 -0.259 Wuhai -0.963 -1.803 -0.04 -0.116 1.593 2.498** 0.228 0.707 1.604 1.923 0.047 0.042

Chifeng -0.88 -1.456 0.431 0.511 -0.619 -1.016 -0.25 -2.121**

Tongliao -1.846 -3.529*** 0.023 0.279 0.276 0.85 0.113 1.059

Shenyang -0.733 -1.834* 0.185 0.489 0 0 0.093 1.546

Dalian -0.525 -1.775* 0.136 0.549 0.744 1.89 0.006 0.113

Liaoyang 0.811 2.381** 0.205 0.975 -0.803 -2.288 -0.202 -1.173

Changchun -3.219 -4.001*** -0.906 -1.353 -1.463 -2.386** -0.118 -0.287 -0.406 -0.876 0.189 2.194** Jilin 0.007 0.031 -0.011 -0.208 0.216 0.682 0.226 0.819

Haerbin -0.213 -0.977 -0.223 -1.469 0.51 2.002 0.135 2.438**

Qiqihaer 0.1 0.406 0.036 0.177 -0.317 -0.811 0.035 0.831

Daqing -0.454 -1.453 -0.074 -0.733 0.769 1.323 0.292 1.03

Mudanjiang -0.879 -1.721 0.433 1.283 -0.142 -0.533 0.034 0.358

Shanghai -0.651 -1.813 -0.393 -1.244 0 0 0.105 2.544** Nanjing -0.36 -0.682 0.541 1.002 -1.143 -8.306*** -0.137 -2.098** 1.368 9.538*** 0.197 2.806***

Wuxi -1.332 -5.051*** -0.068 -0.598 1.435 5.319*** 0.072 0.819 Xuzhou -1.174 -4.481*** -0.142 -0.68 -0.836 -3.809*** 0.137 1.039 1.149 5.090*** 0.04 0.391

Changzhou -0.819 -1.551 0.837 1.001 -0.045 -0.2 0.12 1.004 0.113 0.5 0.236 2.386** Suzhou -1.335 -1.985** 0.145 0.169 -1.057 -3.905*** -0.173 -1.143 0.861 3.221*** -0.093 -0.742 Nantong -1.124 -3.401*** -0.503 -2.507** 1.741 3.996*** 0.394 3.612*** Huaian -1.206 -4.156*** -0.188 -0.994 -0.614 -2.212** -0.117 -1.259 1.281 3.913*** -0.417 -1.609

Yangzhou -0.601 -1.202 -0.035 -0.058 -0.044 -0.149 -0.334 -2.053** 0.522 1.696 0.018 0.107 Zhenjiang -2.182 -4.894*** -0.404 -1.06 -0.935 -3.089*** -0.244 -1.148 1.923 4.942*** 0.276 2.145** Taizhou -1.017 -2.912*** 0.116 0.446 -0.416 -1.609 0.045 0.384 1.356 4.090*** 0.242 1.125 Suqian -0.43 -1.836 0.018 0.138 0.436 1.305 0.233 2.991***

Hangzhou -1.198 -3.298*** -1.052 -3.234*** -0.512 -3.047*** 0.01 0.16 0.563 3.530*** 0.208 2.082** Ningbo -0.042 -0.079 -1.474 -1.2 -0.233 -1.168 -0.057 -0.957 0.617 3.119*** -0.057 -0.519

Wenzhou -1.692 -3.741*** -0.151 -1.716* 2.086 4.302*** 0.203 0.582

Jiaxing -1.222 -2.580** -0.056 -0.366 0.466 1.102 0.146 0.644 Huzhou -0.992 -1.725 0.495 0.611 -1 -2.366** 0.069 0.757 1.318 3.440*** 0.61 1.573

Shaoxing -1.319 -3.346*** -0.171 -0.55 -1.258 -3.664*** -0.238 -1.082 1.61 4.102*** 0.368 3.406***

40

Jinhua -1.34 -3.299*** -0.381 -1.441 -0.935 -1.981** -0.015 -0.161 1.285 2.686*** 0.782 1.242 Quzhou -2.381 -3.706*** 0.191 0.253 -0.198 -0.589 0.246 2.154** 0.507 1.562 -0.106 -0.387 Taizhou -1.281 -2.207** 0.339 0.384 -2.094 -3.520*** 0.014 0.109 1.695 3.186*** -0.092 -0.255 Lishui -0.054 -0.138 -0.675 -1.534 -0.276 -0.525 -0.055 -0.547 Hefei -1.16 -3.776*** -0.305 -1.203 -0.693 -2.737*** -0.099 -0.789 1.081 4.151*** 0.185 1.506

Huainan -0.922 -2.872*** -0.371 -1.476 0.363 1.322 -0.11 -0.589 0.305 1.062 -0.406 -2.029** Fuzhou -0.594 -2.112** 0.025 0.239 0.877 3.148*** 0.123 1.052

Xiamen -2.627 -4.191*** -2.142 -2.017** -0.575 -1.871* -0.099 -0.769 1.162 3.509*** 0.054 0.371 Nanchang -0.397 -1.297 -0.168 -1.193 0.367 1.29 0.175 0.8

Jinan -1.398 -5.927*** 0.154 1.088 -0.865 -4.379*** 0.031 0.343 0.954 4.666*** 0.077 0.599 Qingdao -1.023 -4.018*** 0.123 0.624 -0.262 -1.295 -0.068 -0.602 0.528 2.486 0.119 0.91

Zibo -2.06 -3.002*** 0.371 1.037 -1.163 -2.509** 0.268 1.39 0.879 1.892 1.232 2.922*** Yantai -1.97 -2.470** 0.381 0.478 -0.685 -1.662* 0.16 0.614 0.789 1.884 0.388 2.039** Rizhao -1.598 -3.044*** -1.003 -2.231** -0.587 -1.519 0.174 0.947

Zhengzhou -0.876 -3.072*** 0.304 1.46 -0.298 -1.363 0.191 1.114 0.055 0.221 0.095 1.104 Luoyang -0.945 -3.033*** 0.28 1.004 -0.554 -1.727* -0.179 -1.456 0.747 2.237 -0.601 -1.529 Wuhan -1.425 -4.245*** -0.549 -1.536 -0.39 -2.158** -0.092 -0.916 0.91 4.638 -0.052 -0.602

Changsha -1.538 -4.916*** -0.105 -0.258 -0.98 -3.860*** -0.056 -0.846 1.06 5.001*** 0.311 1.730*

Guangzhou -1.357 -3.870*** -0.059 -0.472 1.007 3.298*** 0.476 2.702***

Shenzhen -1.275 -2.439** -0.316 -1.656* 2.16 3.567*** 0.168 0.751

Zhuhai 1.839 1.585 0.079 0.671 -0.341 -0.518 -0.527 -0.874 Nanning -1.796 -3.884*** -1.223 -1.777* -1.292 -2.712*** -0.047 -0.564 2.161 3.491*** 0.103 0.165 Haikou -2.218 -3.653*** 0.145 0.22 -0.546 -1.842* 0.016 0.194 1.456 4.305*** 0.31 1.834*

Chongqing 0 0 0.156 2.424** Chengdu -2.166 -3.118*** 0.545 0.454 -1.5 -3.574*** -0.164 -0.436 1.442 3.612*** 0.098 2.467**

Mianyang -1.116 -2.123** -0.366 -0.288 0 0 0.145 2.788*** Guiyang -0.859 -3.374*** 0.019 0.088 -0.813 -2.896*** -0.376 -1.683* 0.909 3.707*** 0.222 1.650* Kunming -0.025 -0.105 -0.156 -0.506 -0.576 -3.668*** 0.035 0.315 0.355 2.343 0.051 0.694

Xi'an -1.659 -5.612*** -0.337 -1.313 -1.651 -6.504*** 0.1 0.708 1.706 6.720*** 0.068 0.676 Lanzhou -1.116 -2.581** -0.031 -0.077 -0.382 -1.855* -0.094 -0.781 0.127 0.62 0.083 1.152 Xining -1.281 -4.680*** 0.889 1.861* 0.24 0.796 -0.19 -0.955 1.664 2.687*** 0.159 0.224

Yinchuan -0.334 -1.298 -0.268 -0.551 -0.652 -2.973*** 0.026 0.384 0.703 2.605 0.253 1.807* Wulumuqi -0.002 -0.007 -0.397 -2.040** 0.494 2.196 0.012 0.126

41

Table Five: Overall green city ranking by CO2 emissions per standard household (ton per household per year)

Rank City Electricity Coal LPG Coal gas Car Taxi Bus Rail Heating Total

1 Huaian 0.879 0.098 0.082 0.016 0.013 0.011 0.023 1.123

2 Xuzhou 0.946 0.070 0.046 0.112 0.008 0.010 0.040 0.006 1.237

3 Suqian 0.865 0.218 0.117 0.015 0.006 0.026 1.246

4 Taizhou 1.069 0.041 0.076 0.016 0.058 0.006 0.005 1.271

5 Nantong 1.062 0.036 0.164 0.002 0.007 0.012 1.283

6 Nanchang 0.978 0.141 0.048 0.007 0.130 1.305

7 Zhenjiang 1.098 0.067 0.036 0.064 0.004 0.009 0.027 1.306

8 Haikou 0.983 0.007 0.176 0.015 0.089 0.006 0.065 1.341

9 Yangzhou 1.123 0.033 0.063 0.083 0.029 0.009 0.019 1.359

10 Shuozhou 0.594 0.255 0.046 0.060 0.017 0.016 0.015 0.357 1.360

11 Shaoxing 1.170 0.048 0.066 0.052 0.006 0.006 0.021 1.369

12 Xining 0.878 0.250 0.020 0.012 0.005 0.019 0.175 0.016 1.376

13 Kunming 1.003 0.033 0.068 0.106 0.052 0.005 0.138 1.405

14 Huainan 1.008 0.144 0.056 0.085 0.006 0.063 0.058 1.420

15 Mianyang 1.157 0.001 0.209 0.015 0.012 0.027 1.420

16 Jinhua 1.154 0.046 0.167 0.002 0.056 0.008 0.016 1.449

17 Xi'an 0.871 0.072 0.037 0.101 0.005 0.018 0.104 0.246 1.454

18 Luoyang 0.905 0.155 0.127 0.027 0.026 0.010 0.038 0.189 1.477

19 Quzhou 1.115 0.030 0.189 0.068 0.033 0.007 0.037 1.478

20 Nanning 1.079 0.001 0.220 0.002 0.086 0.009 0.097 1.492

21 Changzhou 1.224 0.009 0.106 0.053 0.071 0.010 0.041 1.514

22 Chengdu 1.243 0.016 0.005 0.232 0.040 0.012 0.007 1.555

23 Nanjing 1.293 0.003 0.097 0.051 0.020 0.009 0.096 0.032 1.600

24 Changsha 1.204 0.028 0.193 0.044 0.029 0.021 0.088 1.607

25 Lishui 1.308 0.018 0.197 0.071 0.005 0.013 1.612

26 Jiaxing 1.286 0.187 0.009 0.100 0.007 0.028 1.617

27 Huzhou 1.330 0.014 0.194 0.008 0.017 0.007 0.059 1.629

42

28 Ningbo 1.328 0.004 0.213 0.011 0.029 0.006 0.050 1.641

29 Suzhou 1.424 0.016 0.068 0.077 0.034 0.008 0.033 1.661

30 Shanghai 1.219 0.007 0.235 0.009 0.014 0.118 0.074 1.675

31 Chongqing 1.396 0.229 0.006 0.014 0.039 0.004 1.687

32 Rizhao 1.060 0.092 0.065 0.121 0.013 0.060 0.318 1.730

33 Wuxi 1.461 0.071 0.123 0.013 0.010 0.060 1.737

34 Taizhou 1.359 0.008 0.256 0.004 0.095 0.007 0.009 1.738

35 Zhuhai 1.197 0.345 0.002 0.049 0.010 0.148 1.750

36 Zhengzhou 0.984 0.185 0.053 0.109 0.008 0.006 0.057 0.363 1.765

37 Hefei 1.360 0.069 0.101 0.064 0.009 0.044 0.138 1.785

38 Lanzhou 0.573 0.029 0.047 0.067 0.003 0.016 0.077 0.976 1.788

39 Guangzhou 1.315 0.213 0.052 0.026 0.008 0.127 0.055 1.796

40 Tangshan 0.865 0.232 0.054 0.017 0.058 0.625 1.851

41 Wuhan 1.526 0.016 0.133 0.092 0.010 0.011 0.069 0.003 1.860

42 Zibo 0.998 0.169 0.119 0.024 0.031 0.021 0.062 0.441 1.866

43 Guiyang 1.433 0.201 0.016 0.118 0.020 0.010 0.073 1.872

44 Hangzhou 1.650 0.006 0.132 0.026 0.023 0.006 0.087 1.930

45 Yantai 0.969 0.142 0.069 0.067 0.021 0.019 0.022 0.629 1.938

46 Wulumuqi 0.509 0.027 0.086 0.001 0.024 0.177 1.128 1.953

47 Handan 0.998 0.029 0.008 0.222 0.024 0.013 0.068 0.633 1.994

48 Shenzhen 1.491 0.261 0.012 0.139 0.010 0.112 2.025

49 Qingdao 1.205 0.248 0.060 0.067 0.012 0.020 0.053 0.388 2.053

50 Jinan 1.099 0.373 0.062 0.030 0.027 0.017 0.085 0.436 2.128

51 Cangzhou 0.868 0.185 0.020 0.018 0.029 0.014 1.087 2.221

52 Qinhuangdao 0.841 0.096 0.076 0.089 0.040 0.017 0.096 0.977 2.233

53 Wuhai 0.536 0.632 0.045 0.017 0.046 0.014 0.093 1.008 2.392

54 Xiamen 2.035 0.001 0.152 0.021 0.031 0.007 0.171 2.418

55 Fuzhou 2.124 0.201 0.025 0.027 0.006 0.054 2.438

56 Taiyuan 0.939 0.027 0.012 0.237 0.017 0.013 0.086 1.107 2.439

57 Tianjin 1.551 0.063 0.014 0.070 0.015 0.018 0.087 0.017 0.690 2.526

43

58 Wenzhou 2.057 0.286 0.001 0.163 0.015 0.051 2.572

59 Huhehaote 0.747 0.105 0.066 0.073 0.016 0.034 0.077 1.468 2.586

60 Shijiazhuang 1.110 0.044 0.091 0.099 0.025 0.019 0.048 1.313 2.749

61 Tongliao 1.448 0.063 0.162 0.036 0.058 0.019 0.972 2.758

62 Jilin 0.983 0.198 0.016 0.030 0.204 1.512 2.944

63 Chifeng 0.873 0.161 0.085 0.030 0.020 0.031 1.802 3.003

64 Liaoyang 0.962 0.139 0.024 0.013 0.026 0.028 1.885 3.077

65 Changchun 0.914 0.010 0.003 0.126 0.019 0.024 0.056 0.006 1.938 3.095

66 Baotou 0.698 0.102 0.054 0.053 0.017 0.021 0.072 2.134 3.152

67 Yinchuan 0.675 0.019 0.059 0.036 0.016 0.034 0.095 2.287 3.221

68 Dalian 0.904 0.015 0.191 0.004 0.040 0.071 0.007 2.143 3.376

69 Beijing 1.558 0.145 0.049 0.084 0.035 0.018 0.138 0.049 1.306 3.381

70 Haerbin 1.157 0.027 0.236 0.008 0.021 0.057 2.009 3.516

71 Shenyang 0.974 0.009 0.099 0.017 0.028 0.082 2.337 3.544

72 Qiqihaer 0.765 0.054 0.115 0.006 0.018 0.041 2.620 3.620

73 Mudanjiang 1.047 0.136 0.081 0.014 0.040 0.017 3.154 4.488

74 Daqing 0.998 0.233 0.003 0.026 0.137 3.719 5.115

Mean 1.122 0.093 0.102 0.077 0.032 0.016 0.067 0.036 1.263 2.073

44

Table Six: Explaining cross-city variation in carbon footprint*

Electricity Heating Car Taxi Rail Bus TOTAL beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat beta t-stat Log(CINCOME) 0.44 3.4*** 1.065 1.08 2.096 4.1*** -1.377 -4.96*** 3.188 2.00* -0.455 -1.22 0.535 4.18*** Log(POP) 0.067 1.95* -0.028 -0.13 -0.398 -2.71*** 0.153 1.79* 0.535 1.08 0.491 4.47*** 0.051 1.48 JAN_TEMP -0.111 -4.41*** -0.03 -8.24*** JULY_TEMP 0.031 2.64*** DENSITY 0.483 1.46 -0.424 -2.75*** -0.66 -0.7 -0.837 -4.01*** constant -5.898 -5.06*** -11.72 -1.22 -24.193 -4.33*** 10.085 3.41*** -41.273 -2.58** 0.238 0.06 -5.371 -4.09*** obs 74 35 71 74 10 73 74 R2 0.436 0.394 0.226 0.27 0.91 0.317 0.504

* Regression on emissions from use of coal, LPG and coal gas are omitted here due to their relative small contributions to total household emissions.

Figures Lo

g H

ouse

hold

Car

bon

Em

issi

ons

January Temperature (Celcius)-20 -10 0 10 20

0

.5

1

1.5

2

Beijing

TianjinShijiazh

Tangshan

Qinhuang

Handan

CangzhouTaiyuan

Shuozhou

Huhehaot

Baotou

Wuhai

ChifengTongliao

ShenyangDalian

LiaoyangChangchuJilin

HaerbinQiqihaer

Daqing

Mudanjia

ShanghaiNanjingWuxi

Xuzhou

ChangzhoSuzhou

Nantong

Huaian

YangzhouZhenjianTaizhouSuqian

Hangzhou

Ningbo

Wenzhou

JiaxingHuzhou

ShaoxingJinhuaQuzhou

TaizhouLishui

Hefei

Huainan

FuzhouXiamen

Nanchang

JinanQingdaoZiboYantai

RizhaoZhengzho

Luoyang

Wuhan

Changsha

Guangzho

Shenzhen

Zhuhai

Nanning

Haikou

ChongqinChengduMianyang

Guiyang

KunmingXi'an

Lanzhou

Xining

Yinchuan

Wulumuqi

Figure 1: The Cross-City Relationship between Winter Temperature and Household Carbon Emissions

46

Log

Hou

seho

ld C

arbo

n E

mis

sion

s

Log Household Income10 10.5 11 11.5

0

.5

1

1.5

2

Beijing

TianjinShijiazh

Tangshan

Qinhuang

Handan

CangzhouTaiyuan

Shuozhou

Huhehaot

Baotou

Wuhai

ChifengTongliao

ShenyangDalian

Liaoyang ChangchuJilin

HaerbinQiqihaer

Daqing

Mudanjia

ShanghaiNanjing

Wuxi

Xuzhou

ChangzhoSuzhou

Nantong

Huaian

YangzhouZhenjianTaizhouSuqian

Hangzhou

Ningbo

Wenzhou

JiaxingHuzhou

ShaoxingJinhuaQuzhou

TaizhouLishui

Hefei

Huainan

Fuzhou Xiamen

Nanchang

JinanQingdaoZiboYantai

RizhaoZhengzho

Luoyang

Wuhan

Changsha

Guangzho

Shenzhen

Zhuhai

Nanning

Haikou

ChongqinChengdu

Mianyang

Guiyang

KunmingXi'an

Lanzhou

Xining

Yinchuan

Wulumuqi

Figure 2: The Cross-City Relationship between City Income and Household Carbon Emissions

Log

Hou

seho

ld C

arbo

n E

mis

sion

s

Log City Population6 7 8 9 10

0

.5

1

1.5

2

Beijing

TianjinShijiazh

Tangshan

Qinhuang

Handan

CangzhouTaiyuan

Shuozhou

Huhehaot

Baotou

Wuhai

ChifengTongliao

ShenyangDalian

Liaoyang ChangchuJilin

HaerbinQiqihaer

Daqing

Mudanjia

ShanghaiNanjing

Wuxi

Xuzhou

ChangzhoSuzhou

Nantong

Huaian

YangzhouZhenjianTaizhou Suqian

Hangzhou

Ningbo

Wenzhou

Jiaxing Huzhou

ShaoxingJinhuaQuzhou

TaizhouLishui

Hefei

Huainan

FuzhouXiamen

Nanchang

JinanQingdaoZiboYantai

Rizhao Zhengzho

Luoyang

Wuhan

Changsha

Guangzho

Shenzhen

Zhuhai

Nanning

Haikou

ChongqinChengdu

Mianyang

Guiyang

Kunming Xi'an

Lanzhou

Xining

Yinchuan

Wulumuqi

47

Figure 3: The Cross-City Relationship between City Population Size and Household Carbon Emissions

Figure 4: Carbon Dioxide Emissions Per Household in 74 Chinese Cities