S E A N E H R L I C H , E C N 1 9 4 H

THE ENVIRONMENTAL POLICY AND POLLUTION INTENSITY IMPACTS OF

INTERNATIONAL TRADE

LITERARY REVIEW

• Is Trade Good or Bad for the Environment? Sorting Out the Causality Jeffrey A. Frankel and Andrew K. RoseNBER Working Paper No. 9201, September 2002, JEL No. Q2, F1

• Environmental Effects of International Trade Jeffrey A. Frankel EXPERT REPORT NO. 31 TO SWEDEN’S GLOBALISATION COUNCIL

• Trade, CO2, and the EnvironmentJoseph S. ShapiroSSRN Abstract No. 2374883, July 1, 2014, JEL No. F18; H23; Q56

• “Measuring Environmental Policy Stringency in OECD Countries: A Composite Index Approach”Botta, E. and T. Koźluk (2014) OECD Economics Department Working Papers, No. 1177, OECD

• Jbara '07, Brian W., "Exploring the Causality between the Pollution Haven Hypothesis and the Environmental Kuznets Curve" (2007). Honors Projects. Paper 21.http://digitalcommons.iwu.edu/econ_honproj/21

THE ENVIRONMENTAL KUZNETS CURVE

INTERNATIONAL TRADE’S ROLE IN POLLUTION EMISSIONS

CO2 Emissions from Fuel Combustion between OECD and non-OECD Countries. Adapted from OECD’s Carbon Dioxide Emissions Embodied in

International Trade, by OECD’s Inter-Country Input-Output Database.

IMPLICATIONS: POLLUTION HAVEN HYPOTHESIS

• Pollution Haven Hypothesis

• Countries with a lower demand for environmental quality (high demand for trade and FDI) will adopt lax environmental standards to attract multinational corporations and export pollution intensive goods

• Countries with high demand for environmental quality (developed countries) will import pollution-intensive goods

• Increased trade interdependence between developed and developing countries can predictably explain the increase in pollution levels in non-OECD countries.

CONVERSELY: THE GAINS FROM TRADE HYPOTHESIS

• Gains from Trade Hypothesis• Trade allows countries to attain more of what they want,

which includes environmental goods in addition to market-measured output

• International ratcheting up of environmental standards• The California Effect

• Technological and managerial innovation

THE QUESTIONS

• Does the Environmental Kuznets Curve hold, meaning developing countries will have a high pollution impact while developed countries will eventually decrease pollution levels?

• If the EKC holds, which theory supports the reduction of pollution in affluent countries, the Pollution Haven Hypothesis or Gains from Trade?

1

2

35

64

Frankel and Rose, 2002

ACCOUNTING FOR ENDOGENOUS RELATIONSHIP BETWEEN INCOME AND

TRADE• Controlling for Population • Shows evidence for impact of increased population on

pollution levels• Isolates the effect of population on the environment from

income’s effect on the environment• Controlling for GDP per capita• Shows evidence for the Environmental Kuznets Curve• Isolates the effect of development on the environment from

the effects of trade on the environment.• Isolated Openness Variable • Shows evidence for/or against the Gains from Trade

Hypothesis

TESTING FOR THE ENVIRONMENTAL KUZNETS CURVE: EMPIRICAL MODEL

• Ln(Pollution Level)ti or z-score of Policy Indexti= α+ Countryi + Timet + β0ln(Population)ti + β1ln(GDP/population)ti + β2ln(GDP/population)2

ti + ξ

• Coefficients of interest, β1 & β2

• All data for variables have been collected for 145 countries from 2000 to 2013. Countries vary in income level, location and stage of development.

• A Fixed-Effect Regression model is used to control for the country and year.

POLLUTION INTENSITY DATA

• CO2 Emissions (kiloton per year per country) (2000-2011)

• Particulate Matter 2.5 mean annual exposure (Micrograms per cubic meter) (2000-2013)

• SO2 Emissions (Gigagrams per year per country) (2000-2005)

CARBON DIOXIDE DATA COLLECTION

• Emissions data (kilatons per year) collected from fossil fuel combustion and cement manufacturing• Primary sources of anthropogenic carbon emissions

• Calculated by U.S. Dept. of Energy’s Carbon Dioxide Information Analysis Center (CDIAC)• Uses data from the United Nation’s Statistics Division’s World

Energy Data Set and the U.S. Bureau of Mines’ Cement Manufacturing Data Set.

• Presented carbon dioxide emissions are converted from reported elemental carbon emissions, by multiplying by the ratio of mass of carbon to that of CO2: 3.667.• This technique of estimating global carbon dioxide emissions

are accurate within 10, having smaller error bounds than country estimates.

PARTICULATE MATTER DATA COLLECTION

• The mean annual exposure (Micrograms per cubic meter) is the estimated annual ambient concentration in residential areas away from pollution “hotspots”: Industrial districts and transport corridors• Data are estimates are taken from cities with greater than

100,000 people by the World Bank’s Agriculture and Environmental Services Department

• Major source of urban particulate matter is industrial emissions• Should be thought of as a general indication of

local air quality

SULFUR DIOXIDE DATA COLLECTION

• Sulfur dioxide emissions (Gigagrams per year) recorded through unique method by NASA’s Socioeconomic Data and Application Center (SEDAC)

• Emissions calculated using bottom-up mass balance approach: emissions = sulfur content of fuel (or ore) minus the amount of sulfur removed or retained in bottom ash or in products.• Calculations are calibrated to known, and trusted, reported

country emissions data• Estimated annually by country for the following sources:

coal combustion, petroleum combustion, natural gas processing and combustion, petroleum processing, biomass combustion, shipping bunker fuels, metal smelting, pulp and paper processing, other industrial processes, and agricultural waste burning (AWB)

SIGNIFICANCE OF DATA

• All pollutants shown are present in the manufacturing intensive sectors.• Carbon Dioxide emissions calculated this way show the impact

of manufacturing on an otherwise incalculable global pollutant.

• Particulate Matter 2.5 data shows the local air quality in regions with high manufacturing.

• Sulfur Dioxide emissions data gives a measure of a local pollution levels stemming from fuel combustion and other industrial processes.

• Changes in pollution levels can allude to changes in composition of manufacturing, trade patterns or policy

POLLUTION INTENSITY DATA CONT.

• Environmental Performance Index • An aggregation of 20 indicators reflecting national-level

environmental data. Indicators are combined into nine issue categories, each of which fit under one of two overarching objectives, Environmental Health and Ecosystem Vitality.

• Years 2006, 2008, 2010, 2012*Data collected by Yale Center for Environmental Law & Policy

-1.4E+08

-1.2E+08

-1E+08

-80000000

-60000000

-40000000

-20000000

0

20000000

0 10000 20000 30000 40000 50000 60000 70000

Cha

nge

in P

ollu

tion

Emiss

ions

(%

)

GDP per capita (2011)

Development Environmental Kuznets Curve

EKC of SO2

EKC of PM 2.5

EKC of Carbon Dioxide

• With a lower income, the willingness for a country to reduce emissions is low, hence a low change in developing country’s emissions.

• As income increases, there tends to be a reduction in pollution emissions, supporting the Environmental Kuznets Curve.

EMPIRICAL MODEL

• Testing linear relationship between trade and pollution:• ln(Env. Quality Indicator)ti= α + Countryi + Timet +β0ln(Pop)ti

+β1ln(Y/Pop)ti +β2ln(Y/Pop)2ti β3ln([X+M]/Y)ti + ξ

• Testing quadratic relationship between trade and pollution:• ln(Env. Quality Indicator)ti= α + Countryi + Timet + β0ln(Pop)ti

+β1ln(Y/Pop)ti + β2ln(Y/Pop)2ti + β3ln([X+M]/Y)ti +

β4ln([X+M]/Y)2ti + ξ

• No linear relationship found between trade openness and any environmental quality indicator, controlling for country’s income

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cha

nge

in P

ollu

tion

Emiss

ions

(%

)

Openness (Total Trade/GDP) (2011)

Trade Environmental Kuznets Curve

EKC of Carbon Dioxide

EKC of PM 2.5

• A marginal reduction in pollution emissions is shown with increased trade.

• Controlling for country’s income, the effect of trade openness on pollution is positive, meaning the Gains from Trade Hypothesis does not hold.

• Rather, we can conclude the Pollution Haven Hypothesis must hold to account for a reduction in developed countries’ pollution levels.

EVIDENCE FOR POLLUTION HAVEN HYPOTHESIS

-1.8E+08

-1.6E+08

-1.4E+08

-1.2E+08

-1E+08

-80000000

-60000000

-40000000

-20000000

0

20000000

0 10000 20000 30000 40000 50000 60000 70000

Cha

nge

in C

arbo

n Di

oxid

e Em

issio

ns (

%)

GDP per capita (2011)

Environmental Kuznets Curve, Grouped by Openness Level

Openness level >1

Openness level < .5

Openness level between .5 and 1

• As income increases in medium openness countries, there is a relatively lower reduction in emissions than high openness countries.

• This leads me to believe that high openness countries are dramatically reducing their emissions while also increasing GDP per capita. How can this be?

• Evidence points to the Pollution Haven Hypothesis.

CHANNELS FOR TRADE TO IMPACT POLLUTION

• Production Specialization• A method of production where a country focuses on the

production of a limited scope of products or services in order to gain greater degrees of productive efficiency within the entire globalized economic system.

• For the different income groups we will look at the share of manufacturing versus services to see if there is a substantial difference in specialization that can support differences in emissions

• Environmental Policy• I will examine differences in environmental policy to see if

there is a substantial difference in policy decision-making across income groups

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cha

nge

in C

arbo

n Di

oxid

e Em

issio

ns (

%)

Openness (Total Trade/ GDP)

The Effect of Openness on CO2 Emissions (2011), by World Bank Income Group

Low Income Countries

High Income Countries

Lower Middle Income

Upper Middle Income

• Lower Middle Income Countries have the largest impact on carbon dioxide emissions given increases in trade.

• What separates lower income countries from the other countries?

0

5

10

15

20

25

30

35

40

45

50

1998 2000 2002 2004 2006 2008 2010 2012 2014

Man

ufac

turin

g, v

alue

add

ed (%

of G

DP)

Manufacturing Value Added in Lower Middle Income Countries

China

Phillipines

Swaziland

Thailand

LMC Average

Average

SPECIALIZATION REGRESSION

• Ln(Pollution Level)ti = α+ Countryi + Timet + β0ln(Population)ti + β1ln(Y/pop)ti x (Manufacturing share) + β2ln(Y/pop)2

ti x (Manufacturing share) + β3ln([X+M]/Y)ti + β4ln([X+M]/Y)2

ti +ξ

• Coefficients of interest, β1 & β2• Comparing to original Environmental Kuznets Curve will

show the effect of manufacturing on pollution levels.

• There is no real significant evidence that manufacturing share has a dramatic impact on pollution emissions, as a result of increased trade.

• What other channels can we look at to support the Pollution Haven Hypothesis?

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0 0.5 1 1.5 2

Cha

nge

in P

artic

ulat

e M

atte

r 2.5

(%)

Openness (Total Trade/GDP) (2011)

Trade Environmental Kuznets Curve

EKC of PM 2.5 with Manufacturing Share

EKC of PM 2.5

ENVIRONMENTAL POLICY DATA

• Environmental Sustainability Index- A measure a country’s overall progress towards environmental sustainability.• Based on composite indicators developed by the Yale

Center for Environmental Law and Policy• 5 core components: Environmental Systems, Reducing

Environmental Stresses, Reducing Human Vulnerability, Social and Institutional Capacity, Global Stewardship

• Years: Pilot 2000, 2001, 2002, 2005

ENVIRONMENTAL POLICY DATA CONT.

• European Bank for Reconstruction and Development’s “Climate Laws, Institutions and Measures Index” (CLIMI) (2006)

Measures the country’s environmental adaptation and mitigation levels.

ENVIRONMENTAL POLICY DATA CONT.

• Esty and Porter’s Environmental Regulatory Regime Index (EPPI)• A measure that summarizes the overall environmental

regulatory system.1. Standards2. Sophistication of regulatory structure and the extent of

subsidization of natural resources.3. Enforcement4. Quality of environmental institutions

POLICY EMPIRICAL QUESTION

• Z-score of Policy Indexti = α+ Countryi + Timet + β0ln(Population)ti + β1ln(Y/pop) + β2ln(Y/pop)2

ti + β3ln([X+M]/Y)ti + β4ln([X+M]/Y)2

ti +ξ

• Coefficients of interest, β3 & β4

• By controlling for population, income and income group, we can isolate different income group’s trade impact on policy.

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cha

nge

in E

nviro

nmen

tal S

usta

inab

ility

Inde

x

Openness (Total Trade/ GDP)

The Effect of Openness on Environmental Sustainability Index (2011), by World Bank Income Group

Lower Middle Income

Lower Income Countries

High Income Countries

Upper Middle Income

• Evidence shows that trade will have the smallest impact on lower middle income countries in reducing environmental policy.

• Trade will however have the largest income on high income countries.

• Further evidence for the Pollution Haven Hypothesis

Thank You

Any Questions?