Page 1 of 46

Cost and affordability of nutritious diets at retail prices: Evidence from 744 foods in 159 countries

Robel Alemu, Yan Bai, Steven A. Block, Derek Headey, and William A. Masters*

Tufts University Department of Economics Working Paper 10 November 2019

* Contact author:

W.A. Masters, Friedman School of Nutrition Science & Policy and Department of Economics Tufts University, 150 Harrison Avenue, Boston MA 02111 USA Phone +1.617.636.3751, email william.masters@tufts.edu, https://nutrition.tufts.edu/profile/william-masters

Author notes and acknowledgements:

First authorship is joint between Robel Alemu and Yan Bai (in alphabetical order). Robel Alemu is a doctoral student in the Economics and Public Policy Program at Tufts University, and Yan Bai is a doctoral student in the Friedman School of Nutrition at Tufts University. Steven A. Block is a Professor in the Fletcher School at Tufts University, Derek Headey is a Senior Research Fellow at the International Food Policy Research Institute (IFPRI), and William A. Masters is a Professor in the Friedman School and the Department of Economics at Tufts University. We are grateful to Nada Hamadeh and William Vigil Oliver for access to ICP price data, and to Doug Gollin and many other seminar participants for helpful feedback on previous versions of this paper. The work presented here was supported by a project led at IFPRI on Advancing Research in Nutrition and Agriculture (ARENA) funded by the Bill & Melinda Gates Foundation as OPP1177007, together with a project led at Tufts on Changing Access to Nutritious Diets in Africa and South Asia (CANDASA) funded jointly by UKAid and the Bill & Melinda Gates Foundation as OPP1182628, with additional support for data analysis from the Feed the Future Policy Impact Study Consortium as a subaward from Rutgers University under USDA Cooperative Agreement TA-CA-15-008, and the Feed the Future Innovation Lab for Nutrition under USAID grant contract AID-OAA-L-10-00006. Model code and data for replication of results will be available on the project’s website at http://sites.tufts.edu/candasa.

JEL codes: Q11, I15 Key words: Food Prices, Diet Costs, Nutrient Adequacy, Cost of Subsistence, Poverty Running head: The cost of nutritious diets across countries

Page 2 of 46

Cost and affordability of nutritious diets at retail prices:

Evidence from 744 foods in 159 countries

Abstract

Many policies and programs aim to bring nutritious diets within reach of the poor. This paper

uses retail prices and nutrient composition for 744 widely-consumed food and beverage items to

compute the daily cost of meeting nutrient requirements for an active and healthy life, for

comparison with the subsistence cost of dietary energy, measurement of affordability relative to

per-capita earnings, and analysis of stylized facts about how cost and affordability relates to

nutrition outcomes and economic development around the world. We find that the minimum cost

of adequate nutrition exceeds the minimum cost of adequate energy intake by a factor of 2.5-3.8,

with substantial variation both across and within national income groups. Relative to average

household expenditures, adequate nutrition is the least affordable in Sub-Saharan Africa, and the

most affordable in North America. In absolute terms, the minimum cost of adequate nutrition in

many poor countries exceeds the global $1.90/day poverty line. Costs are lower in higher-income

countries, which we explain as a function of infrastructure and the efficiency of agricultural

marketing and retail food systems.

Page 3 of 46

Cost and affordability of nutritious diets at retail prices:

Evidence from 744 foods in 159 countries

1. Introduction

Poor diets contribute to 1 in 5 adult deaths, through both insufficient intake of healthy foods

and excess intake of unhealthy items (Afshin et al., 2019). Multiple burdens of malnutrition

typically coexist, with symptoms of insufficiency (stunting, underweight, wasting, and

micronutrient deficiencies) observed alongside the consequences of excess intake such as

cardiovascular diseases and diabetes (WHO, 2003). Diverse types of food are needed to sustain a

healthy and active life, and food prices differ in systematic ways that might contribute to poor

diet quality around the world (Darmon and Drewnowski 2015, Headey and Alderman 2019).

This study uses retail prices and food composition data to compute the overall cost per day of

nutritious diets, for comparison with household incomes and the cost of subsistence diets, as a

measure of whether each country’s agricultural and food policies bring nutrient adequacy within

reach for low-income consumers. Our focus on retail costs of entire diets extends previous

analyses of food prices that use farmgate or wholesale cost of bulk commodities to guide

agricultural production and trade policy (FAO 2018), using the same rural and urban consumer

prices that underlie national accounts and global statistics on living standards and poverty rates

(World Bank 2019). We aim to generalize previous work on affordability that focuses on

particular reference diets (Hirvonen et al. 2019) or specific geographic settings such as Fan et al.

(2018) or Deptford et al. (2017), so as to identify stylized facts and about global patterns and

drivers of change in farm-to-retail food systems (Reardon and Timmer 2012).

Page 4 of 46

The focus of our study is the most affordable diet in each location that would deliver all

essential nutrients in the proportions needed for a healthy and active life, based on universal

human requirements that can be met with different foods based on local prices and availability in

each country. Balanced intake of macronutrients (protein, fats and carbohydrates) and

micronutrients (vitamins and minerals) is a well-documented necessary but not sufficient

requirement for lifelong health, complementing other food attributes such as phytochemicals or

probiotics as well as non-dietary contributors to nutritional well-being such as water quality and

physical activity. We use the latest medical specifications of lower and upper bounds for each

essential nutrient, combined with nationally representative retail prices and the nutrient

composition of each item, and apply linear programming to allow for substitution among foods

as each nutrient is obtained from a variety of foods.

Focusing on affordability for the least-cost sources of all essential nutrients allows us to

compare food prices in a systematic manner, based on their ability to deliver balanced nutrients

at an affordable cost per day. We can then test for links between modifiable aspects of a

country's agriculture and food systems and their cost of nutrients, and ask whether nutrition

outcomes vary with the cost of nutrients as opposed to other influences on food choice and

human health. Our focus on affordability explicitly aims for a lower bound on the cost of

essential nutrients, which in turn is a lower bound on the cost of maintaining health in general.

Diets balanced in terms of nutrients are just a first step towards overall health, but our approach

allows us to focus on the cost of nutrients relative to an even lower cost of a mere subsistence

diet, as well as the cost of nutrients relative to other household expenditures and total income.

Focusing on the lower-bound cost of nutrient adequacy reveals the degree to which the available

foods in each country deliver necessary nutrients at affordable prices. We then use data

Page 5 of 46

visualizations and regression analysis to test hypotheses regarding how the cost of nutritious

diets varies with economic development and structural factors including sectoral composition,

urbanization, rural infrastructure and access to international trade, and then explore links

between diet costs and the prevalence of nutrient deficiencies across countries.

Our work is driven by concerns that agricultural policies and market developments have

primarily focused on lowering the cost of starchy staples needed for daily energy, while

neglecting supply-demand balances and high prices of the diverse foods needed for lifelong

health (Global Nutrition Report 2018). Previous work reviewed by Darmon and Drewnowski

(2015) as well as recent results from Headey and Alderman (2019) or Hirvonen et al. (2019)

focus on specific foods and food groups, whereas our approach allows for a location-specific

choice of foods to meet nutrient requirements, comparing the cost of the most affordable

nutritious diet to the cost of ‘empty’ calories as well as actual incomes, household food

expenditure and poverty lines to shed light on affordability.

The analysis presented here has four principal objectives: First, we update existing methods

for measuring the cost of nutritious diets, adding macronutrient balance and upper limits for

potentially toxic micronutrients as well as minimum requirements to identify the premium to be

paid above the subsistence cost of daily energy for the specific mix of 21 essential nutrients1 at

levels associated with long-term health (Institute of Medicine 2006). Second, we apply those

criteria to compute least-cost diets for 159 countries using retail prices of 744 standardized

global and region-specific food and beverage items assembled by the International Comparison

Project (ICP 2018), matched with their nutrient composition (USDA 2013) to determine the cost

1 Tables A1 and A2 in the supplementary document presents these 21 essential nutrients, their daily requirements for a representative adult woman of reproductive age, their basic health function, and a list of some prominent foods dense in these nutrients.

Page 6 of 46

and affordability of the least-cost nutritionally adequate diet. Third, we describe international

variation in least-cost diets, using data visualizations and regression results to reveal stylized

facts about how the cost and affordability of nutritious diets relates to economic development

and structural transformation based on a variety of measures from World Bank (2019). Fourth,

we explore associations between the cost and affordability of the least-cost diet and nutritional

outcomes. All prices and diet costs are measured at purchasing power parity (PPP) prices for

2011, allowing direct comparison to the World Bank's international poverty line of US$1.90/day.

Results discussed here build on Allen (2017) and also Masters et al. (2018) and Hirvonen et

al. (2019), addressing global patterns in which foods provide the required nutrients at lowest total

cost in each country, and the degree to which each nutrient requirement influences the cost of an

overall nutritious diet. We also build on Headey et al. (2017), Headey, Hoddinott and Hirvonen

(2018) and Headey and Alderman (2019), who compare food groups using the same ICP data

and find systematic differences in the prices of nutrient-dense vegetables and animal sourced

foods relative to starchy staples. Those patterns could be explained by a model of price formation

in which calorie-dense staples are likely to be tradable commodities whose prices largely depend

on access to world markets, while nutrient-dense vegetables and animal-source foods are less

easily traded so their prices are more sensitive to the efficiency of local supply chains and retail

services. Since the nutrient profiles of different foods vary immensely, we expect the cost of a

nutritionally adequate diet to vary across countries for reasons including perishability, transport

costs, trade distortions and imperfect substitution.

Hypothesized market mechanisms that could lead to systematic patterns in retail prices for

different kinds of food are illustrated in Figure 1. When foods are easily transported and traded,

whether they are exportable (Panel 1a) or importable (Panel 1b), this model reveals how

Page 7 of 46

competitive markets link local availability to world market prices (Pworld), which depends in part

on export taxes or imports tariffs denoted t. When foods are too perishable or bulky for

international trade, Figure 1 shows how their local supply-demand balances drive local retail

prices (Pretail) which depends in part on farm-to-market services and transaction costs which may

be large (Panel 1c) or small (Panel 1d).

Figure 1. Models of price formation influencing the cost of a nutritious diet

Source: Authors’ illustration of hypothesized mechanisms affecting consumer prices (Pretail), based on differences in agricultural policy and food systems across countries and types of food.

In this paper, we hypothesize that relative prices among different kinds of food might vary

systematically with per-capita income and a country's economic structure, driven by the market

forces shown in Figure 1 for both tradable and non-tradable items. Bennett's law that demand for

diverse foods beyond starchy staples expands with income faster than demand for other foods

Page 8 of 46

(Clements and Si 2017) will influence the fraction of foods that are exportable or importable in

any given country and will influence the price of non-tradables directly through supply-demand

balances. More nutrient-dense fruits, vegetables and animal-source foods may be especially

sensitive to the cost of post-harvest services along the farm-to-market supply chain, compared to

cereal grains and other starchy staples that are easier to store and transport (Maestre et al. 2017).

From those observations we draw on Reardon and Timmer (2012) and the model in Figure 1 to

hypothesize that, at each level of per-capita income, countries might have a relatively lower cost

of essential nutrients when they have:

1. A larger service sector, offering more horizontal competition but also more vertical

integration in post-harvest handling;

2. Greater urbanization, which concentrates consumers in space and allows for scale

economies in farm-to-market supply chains;

3. Easier rural transportation and access to electricity, thereby improving the efficiency of

transport and storage from farm to market; and

4. Easier access to international markets, including lower import tariffs, for tradable items

that enter local food systems.

These four hypotheses refer to stylized facts about long-run equilibria as shown in Figure 1.

In the short run and for any particular food, many diverse factors would intervene to shift supply

and demand, and those factors would also influence our macroeconomic variables such as

urbanization and service orientation of the economy, roads and electrical infrastructure, and trade

policy.

Besides structural economic development indicators, we further hypothesize that the cost

of nutrient adequacy is positively associated with prevalence of under-nutrition including micro-

Page 9 of 46

nutrient deficiency and negatively associated with prevalence of diet-related obesity and diet-

related communicable disease. Our aim in this paper is not to isolate a causal relationship with

any one variable, but to provide an initial test of whether systematic patterns exist across

countries in a single year. Other work within countries and across time can help identify the

causes and consequences of food price changes, addressing questions such as whether

infrastructure can moderate seasonality in the prices of nutritious foods (Bai et al. 2018), or the

role of specific kinds of supply chains for particular food groups (Headey et al. 2018).

2. Methods

To compare the cost of a nutritious diet across countries, we use retail prices of the least

expensive foods needed to meet estimated requirements for a median healthy woman of

reproductive age. This builds on the concept of least-cost diets pioneered by Stigler (1945),

which has long been used to recommend combinations of foods that meet health needs for low-

income people in industrialized countries (Cofer et al. 1962, Gerdessen and De Vries 2015,

Parlesak et al 2016, Maillot et al. 2017) as well as the general population in lower-income

settings (Optifood 2012, Deptford et al. 2017, Vossenaar et al. 2017). Our application compares

least-cost diets across countries as a metric of each food system’s ability to deliver balanced

meals with all essential nutrients at low cost, extending O'Brien-Place and Tomek (1983) to

international comparisons. For this purpose, we include upper limits on some nutrients to avoid

excesses associated with chronic diseases, in addition to the lower bounds needed to avoid

undernutrition in low-income settings as in Chastre et al. (2007), Omiot and Shively (2017) and

Masters et al. (2018).

Page 10 of 46

To address cross-country differences in access to nutritious foods, our principal measure is

the Cost of Nutrient Adequacy (CoNA), defined as the minimum cost of foods that meet all

known requirements for essential nutrients and dietary energy requirements for a woman of

reproductive age. We then compare this to the Cost of Caloric Adequacy (CoCA), defined as the

price of the least-cost foods that are required to meet the caloric needs. To measure CoNA, we

use the price of each food and its nutrient content relative to lower bounds and upper limits

needed for daily energy and long-term health:

(1) CoNA = min. { C = Σipi × qi }

Subject to:

(2) Σiaij × qi ≥ EARj

(3) Σiaij × qi ≤ ULj

(4) Σiaij × qi ≤ AMDRj,upper × E / ej

(5) Σiaij × qi ≥ AMDRj,lower × E / ej

(6) Σiaie × qi = E

(7) q1 ≥ 0, q2 ≥ 0, q3 ≥ 0,…, qi ≥ 0

In this notation, the quantity of the jth nutrient in food i is denoted aij, which multiplied by its

quantity consumed (qi) must meet the population’s estimated average requirement (EAR) for

nutrient j, while remaining below upper levels (UL) for micronutrients and within a range for

macronutrients determined by acceptable macronutrient distribution ranges (AMDRlower and

AMDRupper) as percentages of daily energy needs (E), at lowest total cost given all prices (pi)

within the further constraint of overall energy needs (E). The reference number ej is the energy

density of macronutrients, equal to 4 kcal per gram of protein or carbohydrate, and 9 kcal per

Page 11 of 46

gram of lipid. This provides a lower bound on the cost of meeting all nutrient constraints, which

we contrast with the cost of meeting only the daily energy constraint (2107.63 kcal/day) in

equation (6), which we call the cost of caloric adequacy (CoCA). We then compute the CoNA to

CoCA ratio which represents the premium required above daily energy cost to meet all nutrient

requirements. To estimate the affordability of nutritious diet, we also compute ratios of CoNA to

average household food and total expenditure. The different price indexes are expressed either in

USD dollars per day at market exchange rates, absolute terms in US dollars per day at real

purchasing power parity (PPP) prices, as a ratio or in logarithmic forms.

For both CoNA and CoCA we report the foods needed in each country to meet nutritional

needs at lowest cost. A key feature of our approach is to constrain nutritious diets to meet not

only the EARs needed to avoid undernutrition, but also a balanced diet in terms of the three

macronutrients through the AMDR, and upper bounds on micronutrients for which excess intake

could be harmful, within energy balance. The resulting CoNA and CoCA values provide the

lowest costs of meeting nutritional and caloric requirements, respectively. These least-cost diets

will differ from actual consumption patterns, which often fall below or above required levels of

each nutrient. Least-cost diets are also less expensive than national dietary guidelines and

recommendations such as the EAT-Lancet reference diet (Hirvonen et al. 2019), because those

diets aim for greater palatability to meet cultural norms in addition to nutrient requirements.

Computation of least-cost diets reveals the extent which each nutrient constraint influences

total costs, as the shadow price associated with altering each constraint by one unit of that

nutrient:

(8) ��� = ¶�∗

¶(�,���,��,����)��

Page 12 of 46

here SPj is the shadow price of nutrient j (or daily dietary energy) computed as ¶�∗, the change

in minimum cost of meeting all constraints for each ¶(�, ���, ��, ����)�� change in energy

requirement e, the EAR or UL for micronutrients, or the AMDR for macronutrients. Since the

units of measure for these constraints vary widely, to compare across constraints we report all

nutrient costs as semi-elasticities denoted SP’, defined as the increment of cost in dollars per day

when each constraint is altered by 1%:

(9) ���� =

¶�∗

%D(���,��,����)��

Calculations for all equations were completed in R and resulting index values exported to

Stata 15 or Excel for visualization purposes, with model code and data for replication posted

online at the project website referenced in this paper’s acknowledgements.

3. Data

Our food price data comes from the World Bank’s International Comparison Program (ICP),

an initiative associated with the United Nations Statistical Commission to compare purchasing

power and living standards across countries (ICP-World Bank, 2018). The mandate of the ICP

includes collecting retail prices for a list of highly standardized goods and services that are

widely consumed across countries. For the 2011 round of ICP data, this list includes a total of

823 food items from 180 countries, although not all items are found in every country and not all

ICP countries have data that can be compared internationally. The ICP data contained two food

item lists i.e. global items list which contained items more or less common for most countries

and regional items list which consisted of foods that are specific to the four geographic regions

(Africa, Asia, West Asia and Latin America). Out of the combined 823 items in the ICP data,

nearly 90% (744) were matched with items in the USDA’s food composition database. Using

Page 13 of 46

individual item prices in local currency units from the World Bank, we generated an estimation

sample with 20,972 observations over the 159 countries for which we have both local food prices

and the national income accounting data needed for global comparisons.2 Descriptive statistics

on these prices are provided in our online annex of supplemental information, along with links to

model code for replication of our results with other datasets.

A key limitation of ICP data for measuring the cost of nutritious diets is that only

standardized foods observed in multiple countries are included, omitting country-specific items

such as the Ethiopian false banana (enset) and teff3, or foods that are sold in diverse forms of

different quality at specific locations such as local fish, fruits, leguminous grains and some dark

green leafy vegetables. A second limitation of ICP data is that 2011 was a somewhat unusual

year for food commodities, as the cost of some internationally traded items such as rice was

higher than in proceeding and subsequent years. Both concerns make our CoNA and CoCA

estimates an upper bound on the true measure, which would be lower if these foods actually

provided essential nutrients at lower cost than the results we obtain.

To calculate the nutritional content of each item, we match its description in the ICP data

with test results recorded in the U.S. National Nutrient Database for Standard Reference (USDA,

2013). Out of the 823 food items found in the ICP price data, 79 were dropped either due the

absence of clear correspondence to any USDA item or for being deemed to have little nutritional

value. Among the items dropped by our procedure include mineral water, since it does not

contribute nutrients for which there is a lower bound or upper limit in our least-cost diets and

stimulant items (coffee, tea and cocoa) given that they have little nutritional significance. A

2 Detailed lists of all countries and food items are provided in the annex of supplemental materials. 3 Teff is a cereal grain cultivated and consumed mainly in Ethiopia and Eritrea used to make a flat thin bread called “injera” or “keyta”.

Page 14 of 46

detailed list of all 744 food items with their respective nutrient compositions is provided in the

annex of supplemental information.

The third kind of data needed to calculate CoNA and CoCA are nutrient requirements, for

which we use the estimated average requirements (EARs) of a typical adult woman of

reproductive age (19-30) with tolerable upper level (UL) for micronutrients and acceptable

macronutrient distribution ranges (AMDR) for macronutrients, as specified in the Dietary

Reference Intake (DRI) developed by the U.S. National Academies of Sciences, Engineering and

Medicine of the United States (Institute of Medicine, 2006). All three types of constraint (EAR,

UL and AMDR) refer to usual daily intake. EAR is defined as the average daily nutrient intake

level estimated to meet requirements at least half of all individuals in an otherwise healthy

population of a given age, sex, height and weight, physical activity level or during pregnancy or

lactation. The main alternative to EARs is the DRI's recommended dietary allowance (RDA),

which adds two standard errors of the estimated uncertainty or biological variation to meet

estimated needs for 97.5 percent of an otherwise healthy population. RDAs are used primarily to

advise individuals or set food rations to ensure that a given person's needs are met, whereas

EARs are preferred for population-level analyses regarding the whole distribution as

characterized by the median person at each place and time. Both refer to the lower bounds on

essential nutrients, defined as compounds that cannot be synthesized in the body but are needed

for human health. Some of these nutrients also have an upper level (UL) of toxicity beyond

which further intake is associated with adverse effects. Also, among the macronutrients that

supply dietary energy (carbohydrates, protein and fats), the DRIs provide an average

macronutrient distribution range (AMDR) for the fraction of energy from each source associated

with reduced risk of metabolic conditions such as diabetes and other conditions linked to

Page 15 of 46

macronutrient imbalances. The annex of supplemental information provides a complete list of all

EAR, UL and AMDR constraints used as criteria for a nutritious diet.

This study also seeks to establish stylized facts about how the cost of nutrient adequacy

varies across countries, testing for associations with income and other characteristics of a

country's development path. For this we draw on the World Development Indicators database

compiled by the World Bank (2019), plus a geographic database maintained by IFPRI that

matches rural population density at each location with spatial data on rural infrastructure (IFPRI

2018a) and another IFPRI database on international trade (Bouët et al., 2017). To test the specific

hypotheses described in our motivation, the variables we use are gross national income (GNI)

per capita, measured in US dollars at purchasing power parity (PPP) prices in 2011, and four

indicators for each of our principal hypotheses: urbanization, defined here as the share of the

population living in urban areas as defined by national authorities, from World Bank (2019);

service orientation, defined as the fraction of the country's gross domestic product derived from

its services sector as opposed to agriculture, mining or manufacturing, also from World Bank

(2019); rural transportation infrastructure, defined as average travel time for rural people to reach

the nearest city with more than 50,000 people, from IFPRI (2018a) and rural electrification,

defined as the share of the rural population with access to an electricity grid, also from IFPRI

(2018a); and finally the country's access to international trade, defined as the country's average

duty applied on food imports from Bouët et al. (2017). These variables were chosen primarily for

their a priori correspondence to the hypotheses that motivate our study, narrowed further to limit

reductions in sample size caused by data availability. This specific list of variables results in a

final estimation sample of 134 countries. Our annex of supplemental information provides a

detailed set of summary statistics for them and for all 159 countries with price data.

Page 16 of 46

The last aim of this study is to examine associations between the cost of nutritious diets and

nutrition outcomes such as obesity, stunting and micronutrient deficiencies. For obesity

prevalence we use WHO’s Global Health Observatory Data Repository on the percent of adult

population whose body mass index (BMI) is 30 kg/m2 or higher, and for stunting rates we use

WHO’s Global Database on Child Growth and Malnutrition for the percent of under-five

children whose height-for-age z-score is more than 2 standard deviations below the median of

the international reference population. For micronutrient deficiencies, we use variables identified

by Harding et al. (2018): Anemia prevalence is measured as a hemoglobin concentration  less

than 110 g/dL for under-five children, and less than 120 g/dL for non-pregnant women; zinc

deficiency is as calculated by Wessells and Brown (2012) who extrapolated country-level

prevalence of inadequate intake using FAO’s food balance sheets; and vitamin A deficiency

(VAD) prevalence is from Stevens et al. (2015) who estimated VAD among children based on

serum retinol concentrations using a Bayesian hierarchical model. Due to data availability, the

estimation sample for these association studies is reduced to 134 countries for most malnutrition

indicators, with summary statistics for these variables in our annex of supplemental material.

Page 17 of 46

4. Results and discussion

4.1 Descriptive statistics and stylized facts

Least-cost diets in each country use a subset of all available items, selected based on their

nutrient composition and price. To describe the overall food price levels from which each diet is

chosen, Figure 2 summarizes the average price of all food and non-alcoholic beverage items

relative to the prices of other goods and services using local polynomial regression to show their

global means at each level of national income. Each sector’s average price level is normalized

within each country to its expenditure-weighted sum of all goods and services which is the

horizontal line in Figure 2, and normalized across countries to equal 1 in a reference location

which for these data is the United States.

Figure 2 reveals that the average price level for all kinds of food are systematically higher in

poorer than in richer countries, in contrast to the average prices of other sectors such as home

furnishings and transport which are less expensive in poorer countries. This higher price level

compounds the quantity effect of Engel’s Law, by which even if prices were equal people in

poorer countries would spend a larger fraction of their income on food. Higher food prices in

poorer countries can be traced to the causal framework in Figure 1, which reveals how higher

prices could be due to items being importable or non-traded with higher transaction costs, rather

than exportable, with lower supply relative to demand among the non-traded goods.

Page 18 of 46

Figure 2. Price level relative to all consumer expenditure by sector and income level in 2011

Note: Data shown are the average price of all goods and services in each sector at each level of gross national income per person (GNI), estimated by local polynomial regression over 159 countries, as a multiple or fraction of prices for all household expenditure, The horizontal line at 1.0 shows the price level for each sector in a reference country, which is the United States.

How does the cost of different foods vary by income level and regions?

Panels A and B of Figure 3 shows that there is little gradient in the average price of different

food groups by both income and geographic location. As expected, the most nutrient-dense foods

such as meat, vegetables and legumes, fruits, nuts, fish and seafood, dairy and eggs have the

highest costs per calorie while starchy staples, oils, fats and sugar were the least expensive. In

this context, foods are grouped using the United Nations system of national accounts rather than

the nutritional function of each food, whereas research using nutritional food groups would

typically separate eggs from dairy, and distinguish among various fruits, vegetables and legumes.

Page 19 of 46

Figure 3. Cost per calorie of different food groups by income levels and geographic location Panel A: Variation by level of national income

Panel B: Variation by geographic location

Note: Data shown are means and standard deviations in each region of prices by food group, as defined using the United Nations’ Classification of Individual Consumption According to Purpose (COICOP).

Page 20 of 46

How does the cost and affordability of the least cost nutritionally adequate diet vary by income

and geographic region?

Table 1 below summarizes the mean levels of CoNA, CoCA, CoNA/CoCA ratio and

CoNA/ total household expenditure ratio by income and geographic location categories as

defined World Bank’s ICP program.

As shown in Table 1, regional CoNA average is generally lower than the World Bank’s

$1.90/ day poverty line derived from actual expenditure pattern for all income categories except

for high income countries where it slightly exceeds this threshold. On the other hand, CoCA

which is associated with survival is much cheaper across all income levels ranging between $0.5-

0.6/day. The premium for nutrients ranged from 2.5-3.84 with substantial variation both across

and within income groups. With regard to the affordability of nutrients, there is a clear negative

income gradient. Looking across regions, we see considerable variation in the premium for

nutrients with the highest observed in South Asia [4.66 (1.26)] and the lowest in Middle East and

North Africa [2.19 (0.47)]. Nutrients were least affordable in SSA as evidenced by the highest

CoNA to household expenditure ratio [0.49 (0.28)] while it was the cheapest in North America

[0.02 (0.01)].

Page 21 of 46

Table 1. Diet costs per day, by income category and geographic region

N

Cost of nutrient adequacy (CoNA)

Cost of caloric adequacy (CoCA)

Premium for nutrients (CoNA /CoCA)

Affordability of nutrients (CoNA/ total expenditure)

Income cate-gories

Low income 32 1.54 (0.23) 0.53 (0.17) 3.05 (0.66) 0.54 (0.26)

Lower middle income

38 1.48 (0.29) 0.50 (0.29) 3.84 (1.58) 0.18 (0.07)

Upper middle income

39 1.62 (0.24) 0.66 (0.15) 2.50 (0.43) 0.13 (0.05)

High income 50 1.92 (0.74) 0.57 (0.24) 3.69 (1.27) 0.03 (0.02)

Geo-graphic regions

East Asia & Pacific

20 1.73 (0.55) 0.69 (0.13) 2.56 (0.86) 0.16 (0.06)

Europe & Central Asia

45 1.59 (0.27) 0.45 (0.15) 3.79 (0.96) 0.05 (0.05)

Latin America & Caribbean

22 1.79 (0.44) 0.81 (0.27) 2.33 (0.71) 0.10 (0.07)

Middle East & North Africa

16 1.90 (0.27) 0.81 (0.19) 2.19 (0.47) 0.07 (0.06)

North America 2 1.99 (0.03) 0.79 (0.09) 2.54 (0.26) 0.02 (0.01)

South Asia 7 1.33 (0.14) 0.33 (0.18) 4.66 (1.26) 0.18 (0.05)

Sub-Saharan Africa

47 1.57 (0.26) 0.55 (0.15) 3.02 (0.66) 0.49 (0.28)

Worldwide 159 1.61 (0.41) 0.57 (0.24) 3.25 (1.28) 0.18 (0.17) Note: Data shown are population weighted means, with standard deviations in parentheses, for the number of countries indicated in reach region. Worldwide, these 159 countries represented 95.9 percent of the world population. Income categories are from the World Bank, geographic regions are as defined by the UN statistical agencies for the ICP. Diet costs per day are computed as described in the text, and total household expenditure refers to the average expenditure per day on all goods and services as computed by the ICP.

To test for patterns pertaining how the different price indexes vary across countries, we

use locally weighted scatterplot smoothing (LOWESS) non-parametric regressions which fit

slopes of local means at each national income level. Figure 4 reveals that CoNA clusters around

or above $1.90/day in most poor countries and is generally lower in countries with higher levels

of national income. Outliers are clearly identifiable, revealing the specific countries that account

Page 22 of 46

for regional differences shown in Table 1, with notably high cost of nutrients in Eastern Asia

(Korea and Japan). On the other hand, we detect a less steep negative income gradient for CoCA.

Comparing both CoNA and CoCA to the World Bank's poverty line of $1.90, it is clear that in

the poorest countries caloric adequacy alone would require roughly half the budget of a

household living at that poverty line, while nutrient adequacy generally takes up nearly all of a

household’s budget who lives at the World Bank’s poverty line. Nutrient adequacy can be

obtained for less than $1.90/day in some poor countries although it is generally available at that

price only in richer countries where very few people live at such low-income levels (Ferreira et

al. 2016).

Figure 4. Cost per day for nutritious diets and daily energy by level of national income

Note: CoNA_PPP(all) refers to cost of nutritional adequacy in purchasing power parity terms, converting local currency prices to international dollars at the average price level of all household expenditure, as calculated by ICP for cost-of-living adjustments. CoCA_PPP(all) refers to cost of caloric adequacy, meeting daily energy requirements using only starchy staple foods. All nutrient requirements were calculated for a representative woman of reproductive age, not pregnant or lactating.

Page 23 of 46

Figure 5 explores the “nutrient premium,” that is the ratio of CoNA to CoCA. We find

that the nutrient premium is relatively higher for the world’s poorest countries and falls slightly

with small increase in income but gradually increases up to an income level of about $35,000 per

capita only to fall back again for the richest countries. Moreover, variation within regions can be

seen in Panel A of Figure 5 where the premium for nutrients is relatively higher in most of SSA,

India, Europe, Australia, Japan and New Zealand. Within Africa; Chad, Central African

Republic, Cote D’ivoire, Togo, Zambia, Malawi and Zimbabwe had the highest premium for

nutrients. Comparing the relative caloric prices (RCPs) of healthy and unhealthy foods across

176 countries, Headey and Alderman (2019) report similar patterns as RCPs of most nutrient-

dense foods were found to be relatively high in most SSA countries.

Page 24 of 46

Figure 5. Premium in cost of nutrient adequacy over caloric adequacy (CoNA/CoCA ratio) Panel A: Variation by geographic location

Panel B: Variation by level of national income

Note: Data shown are the ratio between cost of nutrient adequacy (CoNA) and the cost of caloric adequacy (CoCA), for 159 countries in 2011, as detailed in the text. CoNA is a least-cost diet using all available foods to meet 37 lower or upper limits for energy, 3 macronutrients and 17 essential micronutrients. CoCA is the least-cost source of subsistence energy from starchy staples only.

Page 25 of 46

CoNA as a fraction of average total household expenditure, used as a proxy for

affordability of nutritious diets, was the highest among most SSA countries while China and

most South Asian countries also had moderately high levels (Figure 6: Panel A). In addition,

there is a clear negative income gradient in the affordability of nutritious diets with CoNA to

household expenditure ratio quickly declining with income and approaching zero for most rich

countries (Figure 6: Panel B). We also see a similar pattern when looking at CoNA as a fraction

of household food expenditure (Figures shown in the online supplementary annex).

Page 26 of 46

Figure 6. Cost of nutrient adequacy as a fraction of mean household expenditure Panel A. Variation by geographic location

Panel B. Variation by income level

Note: Data shown are ratios of CoNA per day to total household expenditure on all goods and services from national accounts, as detailed in the text, for 159 countries in 2011.

Page 27 of 46

The specific patterns in the cost and affordability of nutritious diets that we see so far

(Figures 4-6) could be artifacts of our methodology and can also be influenced by some of the

data limitations explained in our data section such as absence of ICP prices for some local foods

that are not internationally comparable. If omitted products like local beans or vegetables were

locally available at sufficiently low prices relative to their nutrient density, the true CoNA would

be lower while the pattern of the other indexes may also be different. Other studies address this

question using a variety of location-specific datasets in Africa, as described in Masters et al.

(2018). Our focus here is on access to the specific list of 744 internationally-comparable foods in

the ICP data to investigate cross-country patterns associated with economic development.

Which food combinations provide complete nutrition at the lowest cost?

To answer this question, we compute the average energy contribution (kcal/day) of all the

foods that make it into the least cost nutritionally adequate diet disaggregated by income level

and geographic location. Figure 7 shows most of the energy in the least cost diet comes from

starchy staples across all income levels and regions. Furthermore, we see that the contribution of

starchy staples, oils and fats, dairy and eggs to daily energy needs is smaller in lower income

countries. Low income countries make up for this difference via higher calorie contributions

from fruits, vegetables and nuts, which are lower-cost sources of those nutrients in lower-income

countries. Interestingly, meat plays a negligible role in least-cost diet at all income levels, while

there is a clear positive income gradient for daily calories from oils and fats while the reverse is

true for fruits and nuts, indicating systematic variation in pricing across food groups.

Page 28 of 46

Figure 7. Foods included in least-cost diets, by income level and geographic location

Panel A: Variation by national income level

Panel B: Variation by geographic location

Note: Data shown are means and standard deviations in each region of least-cost dietary energy from each food category, using the UN Classification of Individual Consumption According to Purpose (COICOP).

Page 29 of 46

Figure 8 summarizes the average per calorie cost (Int.$/kcal) of different foods by national

income levels and geographic location. Overall, the most nutrient dense foods such as meat,

vegetables, legumes, fruits and nuts are more expensive relative to other food groups. Moreover,

we find that costs of dairy and eggs show a clear positive income gradient while the cost of fish

and seafood is relatively higher in the nutritionally adequate least cost diet of low-income

countries. With regards to geographic variation, Panel B of Table 8 shows that the cost of fruits

and nuts is relatively high in Asia, the Pacific and Europe whereas the cost of fish and seafood is

lower in Africa, Middle East and the Pacific region.

Page 30 of 46

Figure 8. Cost of food groups in least-cost diets, by income level and geographic location Panel A: Variation by level of national income

Panel B: Variation by geographic location

Note: Data shown are means and standard deviations for the cost per calorie of foods selected for least-cost diets in each country, in each region and food category.

Page 31 of 46

The heat map in Table 2 shows the extent to which each of the 19 required micronutrients is

provided by items from each of the 8 major food groups. For six of these micronutrients, more

than half of daily intake in the least cost diet comes from starchy staples. For another four

micronutrients, the majority of intake comes from vegetables and legumes, while remaining

requirements are filled by other food groups. For example the small quantities of meat in the

least-cost diets supply a majority of the required retinol, while oils and fats bring most vitamin E.

These results highlight the importance of considering the entire diet across diverse food groups

needed to meet all requirements at least cost in each food environment.

Table 2. A heat map of nutrient contributions (percent share) by major food groups

Nutrient All foods

Starchy staples

Vegetables & legumes

Milk, cheese & eggs Fish

Fruits & nuts Meat

Oils & fats Sugar

Calcium 100 24% 31% 36% 6% 3% 0% 0% 1%

Iron 100 45% 45% 1% 2% 5% 2% 0% 0%

Magnesium 100 49% 37% 6% 1% 7% 0% 0% 0%

Phosphorus 100 50% 22% 17% 5% 5% 1% 0% 0%

Potassium 100 27% 53% 12% 2% 5% 0% 0% 0%

Sodium 100 65% 10% 17% 7% 0% 0% 0% 0%

Zinc 100 54% 27% 10% 2% 6% 2% 0% 0%

Copper 100 37% 39% 1% 1% 9% 13% 0% 0%

Selenium 100 77% 8% 5% 7% 1% 2% 0% 0%

Vitamin C 100 8% 77% 3% 0% 11% 0% 0% 0%

Thiamin 100 58% 28% 5% 1% 7% 0% 0% 0%

Riboflavin 100 32% 26% 29% 3% 3% 8% 0% 0%

Niacin 100 64% 13% 1% 5% 13% 3% 0% 0%

Vitamin B6 100 56% 30% 5% 2% 4% 2% 0% 0%

Folate 100 19% 68% 2% 0% 8% 2% 0% 0%

Vitamin B12 100 1% 0% 21% 31% 0% 47% 0% 0%

Vitamin A 100 1% 73% 6% 0% 0% 19% 0% 0%

Retinol 100 0% 0% 24% 2% 0% 74% 0% 0%

Vitamin E 100 17% 11% 0% 2% 11% 0% 58% 0% Note: Data shown refers to the percent of nutrient quantities in the least cost diet coming from each food group relative to total quantities of each nutrient in all foods. Starchy staples include all cereals and white root vegetables. Green color show lower contributions while red indicate higher shares for each nutrient.

Page 32 of 46

The nutrients whose constraints add the most to total cost are shown in Figure 9 and 10

which present the nutrient shadow price semi-elasticities and frequency of the most binding

nutrients, respectively. Figure 9 show that costs would change primarily with variation in the

need for calcium and total energy, while a smaller role is played by requirements for vitamin E,

riboflavin and vitamin C. From Figure 10 we see that calcium and total energy are the most

frequently binding constraints followed by vitamins (C, A, B12 and E).

Figure 9. Sensitivity of diet cost to percentage changes in each nutrient constraint

Note: Data shown are population-weighted global means and standard deviations for the change in cost per day, in 2011 US$ at PPP prices, associated with a one percentage point change in each nutrient requirement. Changes in requirements would be increases in the EAR or the AMDR lower bounds or decreases in the AMDR upper bounds. Constraints not shown are never binding or impose negligible costs below 0.02 cents for each percentage point change ($0.0002/pp).

-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Protein_AMDRLower

VitB6_EAR

Folate_EAR

VitB12_EAR

Selenium_EAR

Lipid_AMDRUpper

Carb_AMDRUpper

VitC_EAR

Riboflavin_EAR

VitE_EAR

Energy_EAR

Calcium_EAR

Page 33 of 46

Figure 10. Frequency of binding nutrients across all countries

Note: Data shown are the number of countries where each nutrient constraint listed is a binding influence on the least-cost diet, using available foods and retail prices in 2011. Constraints denoted EAR are lower bounds on intake, while AMDRs may be either upper or lower bounds.

0 20 40 60 80 100 120 140 160 180

Energy

Calcium_EAR

Vit_C_EAR

Vit_A_EAR

Vit_B12_EAR

Vit_E_EAR

Carb_AMDR_Upper

Folate_EAR

Niacin_EAR

Vit_B6_EAR

Protein_AMDR_Lower

Lipid_AMDR_Upper

Magnesium_EAR

Riboflavin_EAR

Lipid_AMDR_Lower

Selenium_EAR

Iron_EAR

Zinc_EAR

Thiamin_EAR

Carb_AMDR_Lower

Retinol_UpperLimit

Page 34 of 46

4.2 Hypothesis tests

Is the cost of nutritious diets associated with structural indicators of economic development?

The patterns shown in Figures 4-6 suggest that a wide variety of factors may affect the

cost of adequate nutrients across different geographic locations and national income levels. To

identify potential links between these factors and a country's economic development, we test for

associations between cost of nutritious diets and some structural and market development

indicators.

The central hypothesis motivating our work is that post-harvest food systems, for both

internationally traded commodities and non-tradable goods and services, play an important role

in the cost of more nutritious foods. Using a standard economic model of price formation

illustrated in Figure 1, structural factors such as urbanization, service-sector development and

rural infrastructure as well as access to imported commodities could all drive retail prices and the

cost of meeting nutrient needs, in addition to regional geographic factors affecting agricultural

supply and consumers' food preferences. The core intuition for these hypotheses is that nutrient-

dense foods are often perishable, so their retail prices are more sensitive to variation in post-

harvest services than calorie-dense cereal grains and other staples. Marketing systems for dairy,

eggs and other animal sourced foods, as well as fruits and vegetables or other nutrient-dense

foods may require cold storage and more rapid distribution, implying lower relative costs in

countries whose economic development path favors access to efficient post-harvest services.

Agriculture is the source of both nutrient-dense foods and starchy staples, but a central

feature of structural transformation is how increasing productivity in any sector shifts activity

away from agriculture towards other sectors, including agricultural marketing systems (storage,

transportation, processing). Greater concentration of consumers in urban centers may further

Page 35 of 46

increase the density of agricultural marketing systems, lowering the cost of nutritional versus

caloric adequacy. For these reasons, at each level of national income, for a given set of

agroecological conditions and food preferences, countries with more structural transformation

out of farm production and towards post-harvest handling and other sectors may offer lower

prices for nutrient-dense foods.

With this foundation, we run ordinary least squares (OLS) regression to examine

potential associations between the cost and affordability of nutritious diets and key predictor

variables that are summarized in Table 3. We present regression results for three outcome

variables: log of CoNA, log of CoNA to household food expenditure ratio and log of CoNA to

all household expenditure ratio. Out of concern for potentially influential outliers in our data, we

employ the rreg routine in Stata (version 15) for robust standard errors. Our regression models

control for national income and region fixed effects to absorb the differences in agroecology,

culture and data-collection systems associated with ICP regions. The indicator service sector

labor share allows us to see potential effects of shifts to service sector while travel time to mid-

sized cities is used as a proxy for the density of agricultural value chains. Access to electricity

provides a broad indicator of the potential for cold chain formation, while average import duty

provides a broad indicator of the effect of trade policy.

Page 36 of 46

Table 3. Structural variables used for hypothesis tests N Mean Std.

Dev. Min Max

Income (log GNI per capita, PPP adjusted, 2011 Int. $)

134 9.05 1.20 6.51 11.32

Service sector size (share of labor in services, %)

134 50.26 19.28 6.10 85.41

Urbanization (share of population in urban areas, %)

134 56.07 21.35 10.91 100.00

Rural transport (log travel time to nearest city of > 50k pop.)

134 5.86 0.87 4.16 8.12

Rural electrification (share of rural pop. with access in 2011, %)

134 69.36 37.79 0.29 100.00

Trade access (average duty applied on imports, ad valorem)

132 0.15 0.10 0.01 0.64

Note: Trade access data is adopted from Bouet et al., (2017) while the rest of the variables are from the World Bank’s Global Development Database. GNI per capita (NY.GNP.PCAP.PP.KD) is obtained from World Bank’s International Comparison Program Database, service sector size (SL.SRV.EMPL.ZS); urbanization (SP.URB.TOTL.IN.ZS); rural electrification (EG.ELC.ACCS.RU.ZS) come from World Bank’s World Development Indicators (WDI) database.

Structural transformation is an inherently circular process in which directions of causality

cannot be identified from these data. With only a single cross section we cannot control for

unobservable differences associated with each country and this time period, and even if we had

many time periods the total number of countries in the world may not be large enough to detect

small associations. We make no claim that the correlations we find are causal, or that the absence

of statistical significance indicates that no correlation exists. Our aim is only to test for any

correlations that might be significant across the world’s countries in this one year, to compare

existing food systems in terms of least-cost diets that reach nutritional and caloric adequacy.

Table 4 presents results for CoNA, in logarithmic form. Our results show that rural travel

time to cities is the only significant correlate of CoNA, providing suggestive evidence that

CoNA is linked to the remoteness of rural populations (indicated by shorter travel times to

cities). Doubling such travel times is associated with 5.8 percent higher CoNA. Results for rural

Page 37 of 46

population with access to electricity, average import duties and service sector labor share are not

statistically different from zero.

Table 4. Structural transformation and the minimum cost of nutrient adequacy (1) (2) (3) (4) (5) (6) (7) lnGNI per capita

-2.599 -2.608 -2.535 -1.552 -3.624* -2.485 -1.429 (2.132) (2.154) (2.165) (2.118) (2.089) (2.127) (2.240)

lnGNI per cap., sq.

0.361 0.362 0.353 0.235 0.487** 0.347 0.219 (0.241) (0.243) (0.245) (0.239) (0.237) (0.241) (0.255)

lnGNI per cap., cu. -

0.016* -0.016* -0.016* -0.011 -0.021** -

0.016* -0.011

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009) (0.010) Services share of labor force

-0.000

0.001

(0.002)

(0.002)

Urban share of population

0.000

-0.000

(0.001)

(0.001)

Rural travel time to city >50k (log)

0.057***

0.058***

(0.019)

(0.019)

Rural electricity access (pop share)

-0.001

-0.000

(0.001)

(0.001)

Import tariffs (ave. duty applied)

-0.128 -0.062 (0.144) (0.150)

Constant 6.774 6.800 6.613 3.586 9.493 6.507 3.284 (6.210) (6.272) (6.303) (6.177) (6.064) (6.187) (6.501) N 134 134 134 134 134 132 132 R2 0.461 0.458 0.451 0.465 0.489 0.497 0.475 F 11.784 10.394 10.116 10.702 11.760 11.962 7.570

Note: Dependent variable is the natural log of CoNA in purchasing power parity (PPP) terms for all goods and services consumed by households, which is the same deflator as GNI per capita. Standard errors in parentheses, with significance levels denoted *** p<0.01, ** p<0.05, * p<0.1, from robust regressions (rreg). All specifications include indicator variables for ICP regions (not shown). Table 5 repeats these specifications replacing the dependent variable with log CoNA to

household food expenditure ratio. Our results show that both access to electricity and rural travel

time are significantly associated with the affordability of nutritious diets. We detect that a

doubling of travel time to the nearest city is associated with an 11.5 percent higher ratio of

Page 38 of 46

CoNA to household food expenditure, while a doubling of the share of the population with

access to rural electricity is associated with an 0.4 percent lower ratio of CoNA to household

food expenditure.

Table 5. Structural transformation and the affordability of nutritious diets (1) (2) (3) (4) (5) (6) (7) lnGNI per capita

-4.474 -5.189 -4.549 -3.265 -5.682 -4.485 -5.558 (4.069) (4.334) (4.252) (3.987) (4.137) (4.290) (4.129)

lnGNI per cap., sq.

0.375 0.455 0.384 0.243 0.545 0.374 0.534 (0.458) (0.489) (0.480) (0.450) (0.469) (0.485) (0.469)

lnGNI per cap., cu. -0.011 -0.014 -0.011 -0.006 -0.018 -0.011 -0.018 (0.017) (0.018) (0.018) (0.017) (0.018) (0.018) (0.018) Services share of labor force

0.002 0.004

(0.003) (0.003)

Urban share of population

-0.000 -0.001

(0.002) (0.002)

Rural travel time to city >50k (log)

0.115*** 0.117***

(0.033) (0.033)

Rural electricity access (pop share)

-0.005*** -0.004***

(0.002) (0.002)

Import tariffs (ave. duty applied)

-0.147 0.010

(0.264) (0.250)

Constant 13.467 15.571 13.667 9.291 16.367 11.982 13.663 (11.947) (12.654) (12.415) (11.660) (12.042) (12.506) (12.000) N 130 129 129 129 129 127 127 R2 0.929 0.922 0.923 0.932 0.928 0.918 0.932 F 174.833 140.176 141.495 162.852 152.145 129.799 109.183

Note: Dependent variable is the natural log of the ratio of CoNA to per-capita household expenditure on food and non-alcoholic beverages. Standard errors in parentheses, with significance levels denoted *** p<0.01, ** p<0.05, * p<0.1, from robust regressions (rreg). All specifications include indicator variables for ICP regions (not shown). Two countries (Comoros and Liberia) were removed from our regression model as the food expenditure data from the national accounts data was implausibly high. Results in Table 5 are robust to replacing the outcome variable with the log CoNA to all

household expenditure, as shown in the Annex Table A10, suggesting that nutritious diets may

Page 39 of 46

be more affordable in countries with more rural electricity and less rural remoteness. Regression

results for other outcome variables such as CoCA and CoNA/CoCA ratio showed no significant

association with any of the structural and market development indicators

Is the cost of nutritious diets associated with nutrition outcomes?

The last aim of this study is to explore linkages between the cost of nutritious diets and

nutrition outcomes, using all countries with observations regarding prevalence of adult obesity,

child stunting, female and child anemia as well as vitamin A and zinc deficiencies. To illustrate

variation in functional forms, we use semi-parametric regressions of these outcomes as functions

of the CoNA share of average household expenditures, controlling for a quadratic function of

GNI per capita, as well as indicator variables for ICP regions.

Figure 11. Obesity, stunting and the affordability of nutritious diets

Note: Data shown are residuals and semi-parametric estimates of the mean and its 95% confidence interval after controlling for a quadratic function of log GNI and indicators for ICP regions. U5 stands for children under 5 years of age.

Page 40 of 46

Starting with anthropometric outcomes, Figure 11 suggests that adult obesity is lower and child

stunting is greater in countries where nutritious diets cost a larger fraction of total spending.

Figure 12 applies the same approach to the prevalence of anemia in women and children. Here,

too, we find increasing rates as the cost of adequate nutrition increasingly strains household

budgets. This result is consistent with the finding of Block, et. al. (2004), that increasing pressure

on household food budgets leads first to decreasing consumption of micronutrient-rich foods.

Headey and Alderman (2019) also find positive associations between the prices of nutrient-rich

foods and child stunting. The greater vulnerability of children under 5 is reflected here, as well.

Figure 12. Anemia prevalence and the affordability of nutritious diets

Note: Data shown are residuals and semi-parametric estimates of the mean and its 95% confidence interval after controlling for a quadratic function of log GNI and indicators for ICP regions. U5 stands for children under 5 years of age.

Figure 13 provides further suggestive evidence with regard to micronutrient deficiencies as a

function of the cost of adequate nutrition relative to average household expenditure. The

Page 41 of 46

prevalence of Vitamin A deficiency is revealed here to be much more clearly a positive function

of the cost of nutrition than is the case for zinc deficiency. This result is consistent with our

previous finding (Figure 10) that the cost of Vitamin A is among the most frequently binding

constraints across countries, while zinc is rarely a binding constraint in the cost of adequate

nutrition.

Figure 13. Vitamin A and zinc deficiency and the affordability of nutritious diets

Note: Data shown are residuals and semi-parametric estimates of the mean and its 95% confidence interval after controlling for a quadratic function of log GNI and indicators for ICP regions.

As a robustness check on the semi-parametric regressions in Figures 11-13, we provide OLS

regression results using quadratic functions,4 and add the urban population share as an additional

regressor. These results in Table 6 are consistent with the non-parametric figures.

4 Hardle and Mammen tests failed to reject quadratic approximations of the non-parametric results in Figures 11 – 13.

Page 42 of 46

Table 6. Nutritional outcomes and the affordability of nutritious diets

(1) (2) (3) (4) (5) (6)

Obesity Stunting Anemia (women)

Anemia (children U5)

VitADef ZincDef

CoNA/hhld exp -18.072* 10.974 49.711** 54.754** 0.125 -6.769 (9.819) (14.701) (20.043) (25.156) (0.166) (13.910)

CoNAhhldexp_sq 10.469* -7.996 -30.130*** -34.678*** -0.076 6.777 (5.672) (7.927) (10.673) (13.050) (0.087) (8.930)

log of gnipc_ppp 11.004* -17.439 30.407** 15.668 -0.247* -18.179 (6.258) (11.073) (13.491) (15.129) (0.131) (11.691)

log_gnipc_sq -0.629* 0.811 -1.851*** -1.209 0.009 0.852 (0.340) (0.609) (0.685) (0.754) (0.007) (0.597)

Urban Pop Shr 0.100*** -0.300*** 0.099 0.027 0.002*** -0.065 (0.033) (0.078) (0.072) (0.074) (0.001) (0.063)

Constant -27.625 131.356** -104.145 -9.259 1.538** 116.356** (30.942) (51.348) (69.246) (79.430) (0.627) (57.912)

Number of observations

142 92 143 143 97 144

R2 0.697 0.733 0.629 0.810 0.820 0.527

F 76.741 40.680 31.167 98.268 181.054 17.008

Note: *** p<0.01, ** p<0.05, * p<0.1. Robust regressions. All specifications include indicator variables for ICP regions.

Limitations of this study are highlighted throughout the text and are briefly summarized here.

First, all our computations are specifically made for a healthy woman of reproductive age.

Variation in diet costs over the life course for different demographic groups are addressed in

other research. Second, our least cost diet is designed to provide an affordability metric for just

nutrient adequacy and does not reflect the additional costs that people will incur for other

attributes of desirable foods. Third, our methods do not consider differences in disease state or

bioavailability and potential interaction among nutrients which would alter the composition and

quantities of a least cost diet. For instance, it is widely reported that the bioavailability of dietary

iron and zinc is much lower in most plant-based diets due to the effects of anti-nutrients such as

phytate. Lastly, although our price data consists of the largest available list of global and regional

Page 43 of 46

food items, other foods may be available at lower prices that are not observed by national food

price monitoring systems. Ongoing research aims to overcome these limitations to the extent

possible, in partnership with national statistical services and international development agencies.

5. Conclusions

This paper extends previous literature along several dimensions. Previous studies of food

costs and nutrition have concentrated on starchy staples as the primary source of adequate energy

intake. In contrast, we construct internationally comparable estimates of the minimum cost of a

nutritionally adequate diet, based on retail prices of 744 standardized global and region-specific

food and beverage items collected from 159 countries. This exercise allows us to identify the

least cost food sources of all required nutrients and energy at the national level. The most

nutrient-dense foods such as meat, vegetables and legumes, fruits, nuts, fish and seafood, dairy

and eggs have the highest per calorie costs while starchy stapes, oils, fats and sugar were the

cheapest. In addition, we find that most of the energy in the least cost diet comes from starchy

staples across all income levels and regions.

Our internationally comparable estimates of the minimum cost of nutrient adequacy also

enable us to compare these costs relative to the global threshold of $1.90 per day for severe

poverty. We find that in absolute terms, the minimum cost of adequate nutrition in many poor

countries exceeds the global $1.90/day threshold for severe poverty, and this cost declines as

national incomes increase. Relative to average household expenditures, adequate nutrition is the

least affordable in Sub-Saharan Africa, and the cheapest in North America. Based on a simple

model of price formation, we further extend our results to develop stylized facts regarding cross-

country differences in the cost of diets and their implications for food policy. While future work

Page 44 of 46

will be needed to identify causality, our initial findings are that access to electricity and rural

travel time are both significantly associated with greater affordability of nutritious diets. We also

find that variation in the cost of nutritious diets might help to explain adverse nutritional

outcomes.

References

Allen, R. C. (2017). Absolute Poverty: When Necessity Displaces Desire. American Economic Review, 107(12): 3690–3721.

Bai, Y., Naumova, E., and W.A. Masters (2018). Seasonal differences in the affordability of nutritious diets in Tanzania. Selected presentation at the annual meetings of the American Society for Nutrition (ASN), 12 June 2018.

Block, S., et. al. (2004). Macro Shocks and Micro Outcomes: Child Nutrition During Indonesia’s Crisis. Economics and Human Biology, Vol. 2, No. 2.

Bouët, A., Cosnard, L. and Laborde, D. (2017). Measuring trade integration in Africa. Journal of Economic Integration, 32(4): 937-977.

Chastre, C.; Duffield, A.; Kindness, H.; LeJeune, S.; Taylor, A. (2007). The Minimum Cost of a Healthy Diet. London: Save the Children UK.

Cofer, E. ; Grossman, E. ; Clark, F. (1962). Family food plans and food costs: for nutritionists and other leaders who develop or use food plans. Home Economics Research Report No. 20. Washington, DC: USDA, Agricultural Research Service.

Clements, K.W. and Si, J., 2017. Engel’s Law, Diet Diversity, and the Quality of Food Consumption. American Journal of Agricultural Economics, 100(1), pp.1-22.

Deaton, A., 2010. ‘Price Indices, Inequality, and the Measurement of World Poverty’, American Economic Review, Vol. 100, No. 1, pp5-34.

Deptford, A., Allieri, T., Childs, R., Damu, C., Ferguson, E., Hilton, J., Parham, P., Perry, A., Rees, A., Seddon, J. and Hall, A. (2017). Cost of the Diet: a method and software to calculate the lowest cost of meeting recommended intakes of energy and nutrients from local foods. BMC Nutrition, 3(1): 26.

FAO (2018), Food Price Index. Rome: FAO, www.fao.org/worldfoodsituation/foodpricesindex.

Fan, L., Baylis, K., Gundersen, C., & Ploeg, M. V. (2018). Does a nutritious diet cost more in food deserts? Agricultural Economics, 49(5), 587-597.

Ferreira, F.H., Chen, S., Dabalen, A., Dikhanov, Y., Hamadeh, N., Jolliffe, D., Narayan, A., Prydz, E.B., Revenga, A., Sangraula, P. and Serajuddin, U., 2016. A global count of the extreme poor in 2012: data issues, methodology and initial results. Journal of Economic Inequality, 14(2), pp.141-172.

Gerdessen, J. C., & De Vries, J. H. M. (2015). Diet models with linear goal programming: impact of achievement functions. European Journal of Clinical Nutrition, 69(11): 1272-1278.

Page 45 of 46

International Food Policy Research Institute (IFPRI). (2014). Global Nutrition Report. Global Nutrition Report 2018: Actions and accountability to accelerate the world’s progress on nutrition. Washington, DC.

Global Nutrition Report (2018). Global Nutrition Report 2018: Shining a light to spur action on nutrition. Bristol, UK: Development Initiatives. Online at https://globalnutritionreport.org.

Headey D., Alderman H., Maitra C., Rao P., 2017. The Relative Prices of Healthy and Unhealthy Foods in 177 Countries. Paper presented at Agriculture for Nutrition and Health Academy Week. Kathmandhu, Nepal, 10–13th July, 2017.

Headey, D., & Alderman, H. (2019). The relative caloric prices of healthy and unhealthy foods differ systematically across income levels and continents. The Journal of nutrition, 149(11), 2020-2033.

Headey, D., Hirvonen, K., Hoddinott, J. (2018); Animal Sourced Foods and Child Stunting, American Journal of Agricultural Economics, 100 (5): 1302–1319.

Hirvonen, K., Bai, Y., Headey, D. and Masters, W.A. (2019). Affordability of the EAT–Lancet reference diet: a global analysis. Lancet Global Health, 7 November 2019.

ICP (2018). International Comparison Program. Washington, DC: World Bank, www.worldbank.org/en/programs/icp.

Institute of Medicine (2006). Dietary Reference Intakes: The Essential Guide to Nutrient Requirements. Washington, DC: The National Academies Press.

Afshin, A. et al. (2019). Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. The Lancet 393.10184: 1958-1972.

Maestre, M., Poole, N. and Henson, S., 2017. Assessing food value chain pathways, linkages and impacts for better nutrition of vulnerable groups. Food Policy, 68, pp.31-39.

Maillot, M., Vieux, F., Delaere, F., Lluch, A. and Darmon, N., 2017. Dietary changes needed to reach nutritional adequacy without increasing diet cost according to income: An analysis among French adults. PloS One, 12(3): e0174679.

Masters, W.A., Bai, Y., Herforth, A., Sarpong, D.B., Mishili, F., Kinabo, J. and Coates, J.C., 2018. Measuring the Affordability of Nutritious Diets in Africa: Price Indexes for Diet Diversity and the Cost of Nutrient Adequacy. American Journal of Agricultural Economics, 100(5), pp.1285-1301.

O’Brien-Place, P. M. and Tomek, W. G. (1983) Inflation in food prices as measured by least-cost diets, American Journal of Agricultural Economics, 65(4): 781-784.

Omiat, G. and Shively, G. (2017) ‘Charting the cost of nutritionally-adequate diets in Uganda, 2000-2011’, African Journal of Food, Agriculture, Nutrition and Development, 17(1): 11571–11591.

Optifood, An Approach to Improve Nutrition (2012). NCT01646710. ClinicalTrials.gov, https://clinicaltrials.gov/ct2/show/NCT01646710.

Parlesak, A., Tetens, I., Jensen, J.D., Smed, S., Blenkuš, M.G., Rayner, M., Darmon, N. and Robertson, A. (2016). Use of Linear Programming to Develop Cost-Minimized Nutritionally Adequate Health Promoting Food Baskets. PloS One, 11(10): e0163411.

Reardon, T., and P. Timmer. 2012. "The Economics of the Food System Revolution." Annual Review of Resource Economics 4.

Page 46 of 46

Stigler, G. J. (1945). The Cost of Subsistence, American Journal of Agricultural Economics, 27(2): 303–314.

USDA (2013). National Nutrient Database for Standard Reference Release 28, https://ndb.nal.usda.gov/ndb/search/list.

Vossenaar, M., Knight, F. A., Tumilowicz, A., Hotz, C., Chege, P., & Ferguson, E. L. (2017). Context-specific complementary feeding recommendations developed using Optifood could improve the diets of breast-fed infants and young children from diverse livelihood groups in northern Kenya. Public Health Nutrition, 20(6): 971-983.

Walter W., Rockstrom J., Loken B., Springmann M., Lang T., Vermeulen S., et al. (2019). Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet Commission. 393(10170):447-492.

World Bank (2019). World Development Indicators. Washington, DC: World Bank, http://data.worldbank.org/products/wdi.

WHO (2003). Diet, Nutrition and the Prevention of Chronic Diseases. Report of a joint WHO/FAO Expert Consultation. WHO Technical Report Series; 916.

Supplemental Information, page 1

Annex of supplementary information for

Cost and affordability of nutritious diets at retail prices: Evidence from 744 foods in 159 countries

Robel Alemu, Yan Bai, Steven A. Block, Derek Headey, and William A. Masters*

* Contact author:

W.A. Masters, Friedman School of Nutrition Science & Policy and Department of Economics Tufts University, 150 Harrison Avenue, Boston MA 02111 USA Phone +1.617.636.3751, email william.masters@tufts.edu, https://nutrition.tufts.edu/profile/william-masters

Supplemental tables and charts Table A1. Essential nutrients included in least-cost diets, their functions and food sources Table A2. Nutrient constraints used for computation of least-cost diets Table A3. Number of countries and food items included in ICP data and in least-cost diets Table A4. Number of price observations and example items by food category Table A5. Number of foods included in least-cost diets, by food category Table A6. Item names, USDA codes and price variation in ICP 2011 data Table A7. Cost of nutrient adequacy (CoNA) and caloric adequacy (CoCA) by country Table A8. Nutritional outcomes used to test for associations with diet costs Table A9. Regression results for structural transformation and nutrient shadow prices Table A10. Regression results for structural transformation and the cost of nutritious diets as a

share of all household expenditure Figure A1. Spatial variation in the cost of nutrient adequacy at PPP prices Figure A2. Spatial variation in the cost of nutrient adequacy at PPP(food) prices Figure A3. Spatial variation in the cost of caloric adequacy at PPP prices Figure A4. The cost of nutrient adequacy as a fraction of national poverty lines Note: Data and model code for replication will be posted on the project website, at sites.tufts.edu/candasa.

Supplemental Information, page 2

Table A1. Essential nutrients included in least-cost diets, their functions and food sources Nutrient Daily

Requirementsa Health functions Nutrient-dense foods

Calcium 800 mg Bone growth and health; blood clotting; nerve impulse transmission, muscle contractions, enzyme regulation

Dairy products (milk, yoghurt, cheese, etc.), Chinese cabbage, kale, broccoli

Iron 8.1 mg Functional component of hemoglobin and other key compounds used in respiration, immune function, cognitive development and energy metabolism

The most bioavailable (heme) iron comes from meat, poultry and fish. Other less readily absorbed (non-heme) sources of iron include fortified plant-based foods such as breads, cereals and breakfast bars

Magnesium 255 mg Bone formation, enzyme function, nerve and heart function

Green leafy vegetables, whole grains, nuts, chocolate and legumes

Phosphorous 580 mg Bone/ teeth growth and health; plays a role in maintaining a normal pH and tissue growth; integral part of several metabolic processes

Dairy products, processed foods, fish, soft drinks, bakery products, meats

Zinc 6.8 mg Required for many enzymes; immune function; growth and development; regulation of gene expression; stabilizes cell membranes and body proteins

Meat, shell fish, legumes, fortified cereals and whole grains

Copper 0.7 mg Plays role in enzyme function, growth, cardiovascular integrity, lung elasticity, neovascularization, neuroendocrine function, and iron metabolism

Organ meats, sea foods, nuts, seeds, wheat-bran cereals and whole grain products

Selenium 45mcg Serves as antioxidant and catalyst for the production of active thyroid hormone; needed for proper functioning of immune system

Meat, sea food, grains, dairy products, fruits and vegetables

Vitamin C 60mg Serves as an antioxidant and a cofactor in enzymatic and hormonal processes; biosynthesis of carnitine, neurotransmitters, collagen; modulates the absorption, transport, and storage of iron

Fruits and vegetables including citrus fruits, tomatoes, potatoes, strawberries, spinach, and cruciferous vegetables

Vitamin B1 (Thiamin)

0.9 mg Serves as a coenzyme in the metabolism of carbohydrates and energy release

Grain product, pork, ham and fortified meat substitutes

Supplemental Information, page 3

Vitamin B2 (Riboflavin)

0.9 mg Functions as a coenzyme in numerous oxidation–reduction reactions in several metabolic pathways and in energy production

Dairy product, bread products, and fortified cereals

Vitamin B3 (Niacin)

11 mg Coenzyme in reduction-oxidation reactions such as intracellular respiration, the oxidation of fuel molecules, and fatty acid and steroid synthesis

Meat, liver, poultry, fish, whole grain breads, and fortified cereals

Vitamin B6 1.1 mg Coenzyme in the metabolism of amino acids, heme synthesis, lipid metabolism; homocysteine metabolism

Highly fortified cereals, beef liver and other organ meats, and highly fortified, soy-based meat substitutes

Folate 320 mcg Coenzyme in DNA synthesis, homocysteine metabolism

Dark green vegetables, beans and legumes and fortified grain products.

Vitamin B12 2 mcg Serves as a cofactor in DNA synthesis and in both amino acid and fatty acid metabolism; plays a role in normal functioning of the nervous system and development of red blood cells.

Animal products such as meat, milk, eggs and fish; fortified plant-based foods (cereals)

Vitamin A 500 g RAEb Vision, gene expression, reproduction, embryonic development, growth and immune function.

Liver, dairy products, fruits and vegetables (carrots, broccoli, squash, peas, spinach, etc.), fortified grains, etc.

Vitamin E 12 mg Functions as a chain-breaking antioxidant in the body by preventing the spread of free-radical reactions.

Vegetable oils and spreads, unprocessed cereal grains, nuts, fruits, vegetables, and meats (especially the fatty portion)

aThe daily requirements refer to the estimated average requirement (EAR) for a representative woman of reproductive age based on the dietary reference intake (DRI). bRAE = Retinol activity equivalent, 1g RAE=1g Retinol 12 g -carotene, and 24 g -

carotene or -cryptoxanthin. Sources: Wardlaw’s Perspectives in Nutrition (2016), and Dietary Reference Intakes: The Essential Guide to Nutrient Requirements, National Academy of Sciences (2006).

Supplemental Information, page 4

Table A2. Nutrient constraints used for computation of least-cost diets

Nut No. Nutrient unit EAR

AMDR lower

AMDR upper UL UL Note

1 Energy kcal 2107.63 2 Protein g 37.62 52.69 184.42 3 Lipids g

46.84 81.96

4 Carbohydrate g 237.11 342.49 5 Calcium mg 800 2500 6 Iron mg 8.1 45 7 Magnesium mg 255 350 supplements 8 Phosphorous mg 580 4000 9 Zinc mg 6.8 40 10 Copper mg 0.7 10 11 Selenium mcg 45 400 12 Vitamin C mg 60 2000 13 Thiamin mg 0.9 14 Riboflavin mg 0.9 15 Niacin mg 11 35 supplements 16 Vitamin B6 mg 1.1 100 17 Folate mcg 320 1000 supplements 18 Vitamin B12 mcg 2 19 Vitamin A mcg 500 20 Retinol mcg

3000 preformed vit.A only

21 Vitamin E mg 12 1000 supplement

Supplemental Information, page 5

Table A3. Number of countries and food items included in ICP data and in least-cost diets

Number of countries

using each food list

Number of items on each food list

Number of items with

nutrient composition

Number of items ever

included in any least-cost diet

Number included in significant

quantities (≥50g/day)

All foods 180 823 744 139 23

Global list 180 201 199 58 16 Regional lists

Africa 50 203 175 29 4 Asia 23 167 144 27 1 W. Asia 12 177 162 13 1 LAC 16 75 64 12 1

Note: ICP data include 79 countries with food prices for only the global list of internationally-

standardized items, outside the four continents with region-specific food lists.

Supplemental Information, page 6

Table A4. Number of price observations and example items by food category

Food category using ICP classification (COICOP)

Number of price observations

Typical examples in each category

Fish and seafood 2,461 Canned sardine with skin, Dried small fish, Fishball

Fruits and nuts 2,965 Orange, Banana, Grapes, Roasted groundnuts Meat 4,073 Whole chicken, Pork liver, Mutton/goat liver,

Beef liver, Mutton liver Milk, cheese and eggs 2,534 Milk (unskimmed pasteurized), Powdered milk,

Sour cream, Ghee Oils and fats 658 Sunflower oil, Olive oil, Corn oil, Palm oil,

Soybean oil, Peanut oil, Vegetable oil Starchy staples (cereals & white root vegetables)

4,195 White rice, Brown rice, Wheat flour, oats, Maize flour, Millet, Sorghum, Baguette, White bread, Roll, Short pasta, Cream crackers, Dried noodles, Brown potatoes, Cassava, Tinned sweet corn

Sugar 551 White sugar, Brown sugar Vegetables and legumes 3,499 Bean curd, Spinach Chinese, Bell pepper, Carrots,

Onion, Green cabbage, Cassava leaves, Sweet potato leaves, Rape leaves, Chives, Taro leaves, Dried white beans, Dried black beans, Lentils, Green/Mung beans, Pigeon peas

Supplemental Information, page 7

Table A5. Number of foods included in least-cost diets, by food category

Food category using ICP classification (COICOP)

Pct. of countries

Ave number of items

Ave # where included

Starchy Staples (cereals & white root veg.) 87% 2.28 2.32 Vegetables & legumes 86% 1.65 1.71 Oils and fats 39% 0.53 1.20 Meat 18% 0.21 1.06 Milk, cheese and eggs 68% 1.04 1.36 Fruits and nuts 57% 0.94 1.46 Fish and seafood 43% 0.60 1.23 Sugar 7% 0.08 1.00

Total number of items included in least-cost diets needed for nutrient adequacy 7.34 7.34

Supplemental Information, page 8

Table A6. Item names, USDA codes and price variation in ICP 2011 data

Food item USDA Code Regional/ Global

N Mean Std. Dev.

CV Min Max

Anchovy 15001 Reg. 20 3.701 1.817 0.491 1.52 8.28 Banana, short finger length 09040 Reg. 43 1.617 0.721 0.446 0.71 4.13 Beef Feet/Trotters (Uncleaned)

10102 Reg. 37 3.918 1.928 0.492 1.28 8.79

Beef Merguez (spiced) 07956 Reg. 31 13.768 7.988 0.580 5.47 38.97 Beef ham 07030 Reg. 34 17.078 7.135 0.418 4.79 42.97 Beef prepacked 13498 Reg. 44 8.716 2.998 0.344 3.81 17.84 Beef without bones 13065 Reg. 45 7.174 2.891 0.403 3.43 19.14 Beetroots 11080 Reg. 40 3.788 2.195 0.579 0.84 10.64 Bream 15101 Reg. 21 4.301 2.290 0.532 1.16 9.51 Broad Beans 16052 Reg. 28 2.053 0.665 0.324 1.06 3.74 Broccoli 11090 Reg. 28 11.563 6.722 0.581 2.93 31.86 Brown sugar cubes 19334 Reg. 27 3.645 2.242 0.615 1.32 9.76 Butter, sold loose 01001 Reg. 29 8.065 3.505 0.435 2.93 16.90 Capitaine 15060 Reg. 27 5.748 2.545 0.443 2.45 14.23 Caramel groundnuts 19148 Reg. 21 4.103 2.192 0.534 1.64 11.27 Cassava Leaves 11134 Reg. 29 1.097 0.648 0.591 0.25 3.04 Catfish 15010 Reg. 20 4.197 2.057 0.490 2.26 10.29 Celery 11143 Reg. 30 4.123 2.696 0.654 1.17 11.66 Chicken wings 05100 Reg. 36 7.252 2.251 0.310 2.68 13.11 Chives 11156 Reg. 28 2.994 2.309 0.771 0.82 10.34 Clementine 09433 Reg. 24 5.744 3.834 0.667 1.43 18.69 Couscous (millet) 20031 Reg. 22 3.522 2.100 0.596 0.97 9.53 Crème fraîche 01017 Reg. 40 11.865 4.662 0.393 3.32 22.45 Dried Machoiron 15233 Reg. 18 8.028 4.161 0.518 2.93 19.86 Dried Okra 11279 Reg. 25 4.430 2.100 0.474 1.88 10.67 Dried apricots 09032 Reg. 23 10.880 7.544 0.693 2.63 28.37 Dried bonga 15215 Reg. 17 7.208 3.862 0.536 2.60 16.47 Dried plums 09291 Reg. 26 14.103 8.213 0.582 3.47 37.97 Dried sardines 15088 Reg. 26 7.428 3.769 0.507 1.98 14.89 Dried small fish 15088 Reg. 29 7.859 4.201 0.535 3.18 17.68 Duck - Dressed 05139 Reg. 29 13.587 7.944 0.585 4.92 47.44 Eggs, trad. production 01123 Reg. 43 7.228 3.496 0.484 3.68 21.91 Flatbread 18973 Reg. 25 4.152 2.178 0.525 1.79 9.74 Fresh Okra 11278 Reg. 38 2.827 1.544 0.546 0.89 9.10 Fresh Small Sardines 15039 Reg. 36 2.843 1.391 0.489 0.70 8.62 Fresh cheese edam 01018 Reg. 45 22.128 9.445 0.427 8.97 49.50 Fresh cheese emmental 01040 Reg. 42 21.840 8.847 0.405 11.22 53.02 Frozen Capitaine in Sea Water

15060 Reg. 22 6.678 3.951 0.592 2.99 18.15

Frozen Nile Perch 15060 Reg. 14 6.003 2.424 0.404 3.49 12.68

Supplemental Information, page 9

Frozen Sea-bream 15091 Reg. 29 6.047 3.693 0.611 1.99 17.42 Frozen Shrimps 15149 Reg. 36 16.262 7.281 0.448 5.64 39.19 Frozen Whiting 15065 Reg. 11 7.253 3.600 0.496 3.59 12.93 Ghee 01145 Reg. 34 9.023 3.778 0.419 3.40 15.85 Gherkins 11941 Reg. 41 10.596 3.933 0.371 3.83 25.04 Giant shrimps 15149 Reg. 31 21.710 13.134 0.605 5.50 62.48 Grapes, red 09132 Reg. 43 8.998 5.139 0.571 2.22 28.46 Green Asparagus 11011 Reg. 11 10.091 8.005 0.793 2.69 29.54 Green Beans 11052 Reg. 48 2.609 1.061 0.407 1.01 6.55 Green Plantain 09277 Reg. 36 2.008 1.209 0.602 0.54 5.93 Grouper 15031 Reg. 23 5.014 3.069 0.612 2.35 17.33 Lamb 17224 Reg. 37 9.082 4.371 0.481 4.66 23.18 Large Mango (Grafted) 09176 Reg. 40 1.957 1.190 0.608 0.70 7.17 Lasagne (sheets) 20420 Reg. 42 9.767 3.447 0.353 2.73 24.16 Liquid Yoghurt 01116 Reg. 42 4.058 1.750 0.431 1.65 12.07 Live Goat 17168 Reg. 40 3.256 1.300 0.399 1.27 8.10 Live Sheep 17168 Reg. 39 3.780 2.055 0.544 1.68 11.17 Live Turkey 05181 Reg. 23 8.983 4.102 0.457 3.54 23.72 Lobster, chilled 15154 Reg. 25 26.547 16.847 0.635 6.21 78.82 Long-grained rice 20444 Reg. 44 1.203 0.364 0.302 0.69 2.41 Macadamia nuts 12131 Reg. 8 66.052 64.400 0.975 22.44 183.84 Mackerel in vegetable oil 83110 Reg. 35 9.956 5.650 0.567 3.16 31.61 Maize Flour Yellow 20016 Reg. 41 1.184 0.487 0.412 0.43 2.59 Maize oil 04518 Reg. 32 5.443 2.238 0.411 2.50 13.18 Maizena 20019 Reg. 40 6.033 2.613 0.433 1.93 14.08 Millet Flour 20647 Reg. 26 1.298 0.368 0.283 0.73 2.18 Millet Whole Grain 20031 Reg. 31 0.949 0.329 0.347 0.43 1.72 Mix Frozen Chicken Parts 05006 Reg. 35 7.875 3.107 0.395 2.83 15.32 Mushroms 11266 Reg. 36 10.235 4.900 0.479 3.66 27.52 Mutton Tripes 13341 Reg. 29 3.805 1.279 0.336 1.36 6.42 Mutton chop 17074 Reg. 40 13.137 5.847 0.445 6.28 35.30 Natural Groundnuts 13341 Reg. 41 2.534 1.218 0.481 0.99 7.06 Nile Perch 15060 Reg. 18 4.597 1.248 0.271 2.70 6.63 Oxtail 10174 Reg. 41 7.734 3.309 0.428 2.82 16.73 Packed Brown sugar 19334 Reg. 36 2.437 1.379 0.566 1.06 7.95 Packed Peas 11306 Reg. 45 4.903 1.916 0.391 1.64 11.13 Packed White sugar 19336 Reg. 42 2.092 0.730 0.349 0.98 4.48 Palm oil unrefined 04055 Reg. 35 2.835 1.024 0.361 1.24 5.34 Passion fruit 09231 Reg. 23 3.966 1.525 0.384 1.71 7.90 Peas 11300 Reg. 77 2.437 0.901 0.370 1.02 5.02 Peeled Tomatoes 11529 Reg. 34 4.164 1.721 0.413 1.90 10.14 Pigeon peas 16101 Reg. 24 2.066 1.173 0.568 0.50 4.94 Pork meat 10005 Reg. 38 7.641 4.500 0.589 2.94 26.76 Powdered Glucose 19336 Reg. 26 9.475 5.279 0.557 2.55 27.12

Supplemental Information, page 10

Powdered milk 01091 Reg. 46 14.359 5.299 0.369 7.13 33.47 Pumpkin leaves 11418 Reg. 26 3.060 1.605 0.524 0.99 6.51 Radish 11429 Reg. 34 2.959 1.388 0.469 0.81 7.81 Rape Leaves 11233 Reg. 15 2.192 1.049 0.479 0.94 4.85 Red Snapper 15101 Reg. 30 5.795 3.762 0.649 2.55 19.49 Red mullet 15055 Reg. 21 4.382 2.765 0.631 2.40 15.24 Roasted groundnuts 16390 Reg. 39 4.052 2.295 0.566 1.90 15.22 Round Onions, red 11282 Reg. 47 1.595 0.491 0.308 0.61 2.63 Round bread 20075 Reg. 42 2.253 0.954 0.423 0.79 4.49 Sardines in tomato sauce 15089 Reg. 45 10.331 4.226 0.409 3.19 20.37 Sesame oil 04058 Reg. 20 6.514 3.226 0.495 2.90 15.23 Sirloin Steak 13919 Reg. 38 12.382 4.318 0.349 5.74 25.84 Sliced brown bread 18075 Reg. 34 3.598 2.703 0.751 1.15 12.98 Smoked carp 15009 Reg. 23 7.819 4.671 0.597 3.45 23.91 Smoked kapenta 15088 Reg. 17 6.371 5.344 0.839 1.23 23.42 Smoked kingfish 15200 Reg. 16 7.566 5.308 0.702 2.70 23.83 Smoked mboto 15014 Reg. 14 8.234 4.146 0.504 3.52 18.08 Smoked shrimps/prawns 15271 Reg. 22 18.114 9.889 0.546 6.09 46.30 Sole fish 15028 Reg. 26 6.798 4.462 0.656 2.03 18.67 Sorghum Red Grains 20067 Reg. 30 0.897 0.484 0.540 0.45 3.01 Sorghum White Grains 20067 Reg. 33 0.902 0.375 0.416 0.36 1.96 Sorrel Leaves 11616 Reg. 22 1.850 0.875 0.473 0.70 4.52 Sour (clotted) milk 01056 Reg. 37 2.280 0.881 0.386 0.87 4.16 Spinach 11457 Reg. 35 7.509 4.143 0.552 1.65 21.70 Spotted beans 16001 Reg. 42 1.810 0.574 0.317 0.90 3.48 Spring Onions 11291 Reg. 37 3.116 2.003 0.643 0.94 9.26 Sweet Bread 20075 Reg. 38 3.699 1.967 0.532 1.43 9.04 Sweet Potatoe Leaves 11505 Reg. 23 1.192 0.647 0.543 0.45 2.99 Taro Leaves 11520 Reg. 18 2.301 1.009 0.438 0.78 4.17 Tinned peaches 09238 Reg. 31 5.632 2.645 0.470 2.43 12.27 Trad. bred live chicken 05006 Reg. 43 8.762 2.232 0.255 4.97 17.83 Tuna 15117 Reg. 35 6.183 3.910 0.632 1.63 18.12 Tuna in vegetable oil 15119 Reg. 43 14.405 5.900 0.410 5.18 33.01 Turkey breast 05191 Reg. 28 12.000 7.321 0.610 5.18 38.82 Turnips 11564 Reg. 35 2.411 1.112 0.461 0.58 5.11 Veal without offals 17128 Reg. 31 13.394 5.805 0.433 6.14 31.70 White Maize grains 20314 Reg. 40 0.837 0.371 0.444 0.27 1.83 Whole Cassava 11134 Reg. 36 0.942 0.520 0.552 0.26 2.20 Yellow Broken Maize Gr. 20014 Reg. 28 0.945 0.424 0.449 0.43 2.28 Yellow Maize Grains 20014 Reg. 35 1.004 0.483 0.481 0.34 2.35 Yellow Maize Gr., Branless 20014 Reg. 33 1.175 0.505 0.429 0.53 2.63 Yoghurt with natural fruits 01294 Reg. 49 5.267 1.512 0.287 2.83 9.29 100% Beef, minced 13498 Glob. 145 8.639 2.420 0.280 4.68 16.56 All-butter croissant 18239 Glob. 134 14.479 5.521 0.381 4.47 40.42

Supplemental Information, page 11

Almonds, Unhusked 12061 Reg. 10 29.170 4.926 0.169 23.17 36.61 Animal Fats 04609 Reg. 8 10.798 4.810 0.445 5.35 20.80 Apple 09016 Reg. 10 2.038 0.496 0.243 0.85 2.60 Apple, Red Delicious 09003 Glob. 97 3.099 0.938 0.303 1.13 6.35 Apple, Typical Local Variety 09003 Glob. 108 2.728 1.401 0.514 0.79 9.53 Apricots 09021 Reg. 5 4.802 2.082 0.434 2.03 6.94 Avocado 09038 Glob. 81 3.962 2.620 0.661 0.81 13.98 Bacon, pork 10123 Reg. 15 16.716 3.117 0.186 10.26 20.76 Bacon, smoked 10123 Reg. 100 17.190 6.092 0.354 6.38 47.59 Baguette 18349 Glob. 120 3.321 1.370 0.412 0.89 7.19 Bajra Flour 20647 Reg. 2 0.963 0.047 0.049 0.93 1.00 Banana, Standard 09040 Glob. 156 2.300 0.999 0.434 0.63 6.27 Basmati Rice 20444 Glob. 124 4.199 1.618 0.385 1.41 9.53 Bean Curd - Tofu 16427 Glob. 31 8.960 8.882 0.991 1.64 34.57 Beaten rice, Chira 20450 Reg. 5 1.589 0.443 0.278 1.26 2.37 Beef liver 13325 Glob. 98 6.242 2.862 0.459 1.70 16.54 Beef with bones 23005 Glob. 102 8.678 3.315 0.382 3.25 19.64 Beef, Center brisket 13803 Glob. 122 8.313 2.611 0.314 3.42 16.25 Beef, Fillet 13065 Glob. 144 18.658 11.940 0.640 4.91 57.76 Beef, Fillet, frozen 13065 Reg. 8 12.021 1.979 0.165 9.51 15.14 Beef, Rump steak 13070 Glob. 141 12.519 5.333 0.426 4.77 38.98 Beef, for stew or curry 13019 Glob. 72 8.315 2.990 0.360 3.93 18.23 Beef, w bones, non-sp cut 23005 Reg. 14 9.979 3.539 0.355 6.35 18.48 Beef, wo bones, non-sp cut 13065 Reg. 15 8.252 2.060 0.250 5.47 12.04 Beetroot 11080 Reg. 13 2.759 0.992 0.360 1.61 5.19 Bell pepper 11821 Glob. 143 3.642 1.858 0.510 0.66 9.84 Black Pomfret 15068 Glob. 29 8.170 3.392 0.415 3.20 16.75 Breakfast sausage, chicken 07976 Reg. 18 7.987 2.808 0.352 3.38 12.72 Brown Potatoes 11352 Glob. 149 1.543 0.626 0.406 0.39 3.75 Brown rice - Family Pack 20036 Glob. 48 2.343 1.023 0.437 0.71 5.72 Brown rice - small pack 20036 Glob. 14 2.322 0.825 0.355 1.30 3.95 Brown sugar 19334 Glob. 73 2.845 1.902 0.668 0.68 9.24 Buffalo milk, not pasteurized

01108 Reg. 5 1.542 0.555 0.360 1.14 2.52

Buffalo milk, pasteurized 01108 Reg. 3 1.807 0.403 0.223 1.52 2.27 Butter, unsalted 01145 Glob. 147 11.875 3.947 0.332 4.32 25.08 Buttercup squash 11643 Reg. 12 1.383 0.445 0.322 0.65 2.15 Canned Tuna/Water 15119 Reg. 10 9.474 1.219 0.129 6.72 11.13 Canned beef, chunks 22905 Reg. 12 8.277 2.625 0.317 4.61 14.00 Canned chicken 05311 Glob. 47 10.552 5.419 0.514 3.46 29.08 Can mackerel fillet in veg oil 15048 Reg. 110 14.330 6.979 0.487 4.64 38.04 Canned peach halves 09238 Reg. 15 6.093 1.346 0.221 4.33 9.01 Canned sardine with skin 15088 Reg. 144 10.334 4.563 0.442 3.32 32.59 Canned sardines with skin, 15088 Reg. 13 11.998 6.449 0.538 3.69 26.61

Supplemental Information, page 12

in tomato sauce Canned tuna without skin 15119 Glob. 144 14.331 5.139 0.359 6.10 34.77 Carp 15008 Glob. 90 6.032 2.575 0.427 1.79 13.36 Carrots 11124 Glob. 155 1.644 0.695 0.423 0.55 4.40 Cassava - Manioc - Yuka 11134 Glob. 36 1.485 0.774 0.521 0.47 3.46 Catfish 15234 Reg. 13 4.851 1.529 0.315 2.81 8.71 Cauliflower 11135 Glob. 139 6.035 3.324 0.551 1.53 20.22 Caviar 15012 Reg. 7 265.678 325.398 1.225 82.68 998.75 Celery 11143 Reg. 11 2.613 1.090 0.417 0.89 4.91 Cheese spread 43276 Reg. 18 23.657 5.668 0.240 13.62 36.49 Cheese, Camembert Type 01007 Glob. 96 24.206 10.982 0.454 7.34 59.74 Cheese, Cheddar 43278 Glob. 93 18.481 6.768 0.366 9.02 44.67 Cheese, Cottage (halloum) 01012 Reg. 7 8.449 3.158 0.374 5.19 14.50 Cheese, Feta 01019 Reg. 10 5.353 0.896 0.167 4.33 7.25 Cheese, Gouda Type 01022 Glob. 113 16.527 7.477 0.452 7.31 43.97 Cheese, Haloumi 01157 Reg. 10 14.407 2.018 0.140 10.27 17.33 Cheese, Kashkaval 01030 Reg. 7 14.333 4.573 0.319 10.40 23.56 Cheese, Mozzarella 01026 Reg. 9 14.108 1.245 0.088 12.03 16.51 Cheese, mozarella type 01026 Reg. 12 15.029 5.681 0.378 9.24 27.43 Cheese, processed 01005 Glob. 144 14.830 6.204 0.418 6.51 56.66 Chicken br w skin &bones 05057 Glob. 83 8.080 3.679 0.455 3.65 24.48 Chicken breast w/o skin 05039 Glob. 135 15.801 6.217 0.393 6.43 40.95 Chicken drumsticks 05075 Reg. 18 8.161 1.894 0.232 5.60 13.31 Chicken egg, 1 01123 Reg. 21 5.124 1.586 0.310 3.58 10.70 Chicken eggs, 10, loose 01123 Reg. 21 5.072 1.580 0.312 3.59 10.50 Chicken legs 05075 Glob. 144 6.835 2.450 0.358 1.07 15.60 Chicken soup 06015 Glob. 9 14.848 1.715 0.115 12.43 17.18 Chicken wings 05100 Glob. 27 8.063 2.298 0.285 4.92 15.77 Chicken, non-sp. cuts, froz. 05001 Reg. 14 8.252 2.197 0.266 4.54 11.61 Chicken, non-sp. cuts, not fr 05001 Reg. 14 7.902 1.457 0.184 5.04 10.49 Coarse #2 20036 Reg. 3 0.717 0.365 0.509 0.30 0.95 Coarse #3 20036 Reg. 3 0.771 0.146 0.189 0.61 0.90 Coarse #5 20036 Reg. 5 1.002 0.078 0.078 0.90 1.11 Coarse #6 20036 Reg. 4 1.064 0.179 0.168 0.95 1.33 Coconut oil 04047 Reg. 8 7.263 1.721 0.237 4.66 9.40 Cod (Gadus morhua) 15015 Glob. 26 6.767 3.569 0.527 2.85 19.48 Corn 20014 Reg. 7 1.893 0.418 0.221 1.49 2.59 Corn oil 04518 Reg. 18 5.616 1.875 0.334 3.71 11.89 Corn tortilla 18363 Reg. 8 5.282 3.169 0.600 2.18 11.61 Corn/Maize Flour, loose 20316 Reg. 9 1.460 0.608 0.416 0.95 2.84 Corned beef 07020 Glob. 63 9.665 2.989 0.309 4.33 18.76 Couscous 20028 Glob. 46 3.848 2.195 0.570 1.07 10.58 Cream cheese 01017 Glob. 132 13.830 5.776 0.418 3.53 44.13 Cream crackers 18962 Reg. 22 5.977 1.155 0.193 4.00 8.48

Supplemental Information, page 13

Cucumber 11206 Glob. 155 2.126 0.970 0.456 0.69 6.01 Cucumber Pickles 11937 Reg. 7 3.276 1.064 0.325 1.82 5.17 Dark raisins 09298 Reg. 12 10.556 2.887 0.273 6.30 17.55 Dhal, Khesari 16085 Reg. 6 1.978 0.519 0.262 1.18 2.66 Dhal, Musur 16085 Reg. 8 2.626 0.578 0.220 2.01 3.94 Dhal, Split Peas 16085 Reg. 9 1.871 0.502 0.268 1.19 2.76 Domestic Apricots 09021 Reg. 5 2.916 0.527 0.181 2.27 3.52 Domestic Beetroot 11080 Reg. 8 2.207 0.632 0.286 1.48 3.19 Domestic Broad beans

16052 Reg. 7 2.253 0.402 0.178 1.77 2.86

Domestic Carrots 11124 Reg. 8 1.245 0.273 0.219 0.92 1.81 Domestic Cherries 09063 Reg. 3 3.924 1.331 0.339 2.84 5.41 Domestic Figs 09089 Reg. 4 3.358 1.575 0.469 1.93 5.56 Dom Fillet Steak(exc round &sirloin)

13065 Reg. 10 12.050 2.307 0.191 8.98 15.49

Domestic Grapes 09129 Reg. 3 3.242 0.527 0.162 2.80 3.82 Domestic Green beans (Pulses)

11052 Reg. 7 2.242 0.677 0.302 1.40 3.31

Domestic Green bell peppers

11333 Reg. 9 1.992 0.403 0.203 1.53 2.64

Domestic Ground mutton (Fresh)

17224 Reg. 8 13.582 4.473 0.329 8.35 20.00

Domestic Guava 09139 Reg. 5 3.078 0.747 0.243 1.97 3.94 Domestic Mango 09176 Reg. 3 4.064 2.284 0.562 2.58 6.69 Domestic Okra 11278 Reg. 7 4.052 1.192 0.294 2.49 5.27 Domestic Peanuts 16390 Reg. 4 22.044 9.579 0.435 13.96 33.74 Domestic Pears 09252 Reg. 5 2.692 0.806 0.299 1.76 3.48 Domestic Peas 11306 Reg. 7 2.828 1.087 0.385 1.43 4.59 Domestic Plums 09279 Reg. 5 2.404 0.490 0.204 2.04 3.10 Domestic Pomegranate 09286 Reg. 5 2.916 1.069 0.367 1.73 4.46 Domestic Red Dates

09421 Reg. 9 2.680 0.702 0.262 1.43 3.47

Domestic Red bell peppers

11821 Reg. 7 3.309 1.328 0.401 1.85 5.31

Domestic Round red raddish

11429 Reg. 5 1.206 0.232 0.192 0.91 1.50

Domestic Strawberries 09316 Reg. 6 10.444 6.768 0.648 4.47 20.46 Domestic Tomatoes 11529 Reg. 9 1.334 0.580 0.435 0.76 2.72 Domestic White onion

11282 Reg. 6 1.149 0.373 0.324 0.80 1.78

Domestic Zucchini

11477 Reg. 8 1.750 0.569 0.325 1.01 2.62

Dometic Dates 09421 Reg. 7 3.107 0.576 0.186 2.35 3.84 Dried Noodles 20409 Glob. 94 4.145 2.125 0.513 1.59 16.77

Supplemental Information, page 14

Dried Shrimp 15151 Glob. 50 34.302 28.511 0.831 3.46 106.28 Dried almonds 12061 Glob. 58 48.633 22.709 0.467 13.76 116.62 Dried black beans 16015 Reg. 11 2.696 1.045 0.388 0.98 4.80 Dried dates 09421 Glob. 98 7.481 3.183 0.426 2.13 23.45 Dried red beans 16145 Reg. 14 3.326 0.850 0.255 2.22 5.39 Dried white beans 16049 Glob. 87 3.139 1.486 0.473 1.36 10.07 Duck, whole 05139 Reg. 12 6.834 1.324 0.194 4.68 8.96 Egg noodles 20109 Reg. 17 3.479 1.150 0.331 2.22 6.47 Eggplant (aubergine) 11209 Reg. 147 2.179 1.094 0.502 0.35 5.37 Emperor fish 15101 Reg. 6 9.877 4.342 0.440 5.73 18.19 Figs 09089 Reg. 7 6.252 2.795 0.447 2.95 10.91 Fillet (Round or sirloin) 13919 Reg. 9 12.667 2.372 0.187 9.14 15.50 Fishball 15027 Reg. 11 4.975 1.568 0.315 2.03 6.93 Flank or skirt beef, for shredding

13065 Reg. 8 7.180 0.911 0.127 5.77 8.55

Flat (Iranian) Bread 18973 Reg. 8 2.751 1.409 0.512 1.09 5.01 Fresh Milk, 0% (import) 01151 Reg. 7 1.841 0.422 0.229 1.52 2.75 Fresh Milk, 1.5-2.5% 01174 Reg. 8 1.755 0.413 0.235 1.40 2.69 Fresh Milk, 3-4% 01078 Reg. 9 1.765 0.274 0.155 1.42 2.27 Fresh Milk, 3-4% 01078 Reg. 8 1.564 0.202 0.129 1.23 1.78 Fresh cheese 01228 Reg. 13 8.918 2.257 0.253 5.82 12.73 Fresh rice noodles 20134 Reg. 12 1.281 0.303 0.237 1.01 2.13 Fresh wet cheese 01228 Reg. 9 7.574 2.027 0.268 4.83 9.81 Fresh whole chicken 05001 Reg. 19 7.773 1.755 0.226 5.10 12.29 Fruit cocktail, Canned 09351 Reg. 6 3.044 0.877 0.288 1.76 4.14 Ghee 01145 Reg. 6 7.092 5.577 0.786 2.86 17.97 Ghee, cow/buffalo 01323 Reg. 8 13.121 2.830 0.216 8.79 17.07 Giant Shrimp 15149 Reg. 7 24.737 8.866 0.358 17.38 41.82 Glutinous Rice 20054 Reg. 13 1.707 0.304 0.178 1.31 2.17 Goat leg 17011 Reg. 13 11.874 2.762 0.233 4.54 14.89 Goat mix cut w bones (non-refrig.)

17168 Reg. 59 8.244 3.427 0.416 2.50 19.05

Goat, boneless. Non-specific cut

17168 Reg. 9 12.040 4.386 0.364 3.27 17.35

Grapefruit 09116 Glob. 100 3.891 1.762 0.453 0.80 9.94 Grapes, green 09129 Glob. 125 6.316 3.884 0.615 1.13 28.25 Grapes, violet, with seed 09132 Reg. 20 7.228 2.232 0.309 3.90 10.44 Green beans 11052 Reg. 8 4.122 1.078 0.261 2.64 5.93 Green cabbage 11109 Reg. 102 1.524 0.806 0.529 0.42 4.78 Green/Mung Beans, dried 16080 Reg. 29 2.903 0.392 0.135 2.10 3.59 Ground beef (Frozen) 13317 Reg. 7 6.227 3.048 0.490 3.71 12.40 Grouper (Hamour) fish 15031 Reg. 10 14.163 5.036 0.356 5.64 25.93 Hake fillet 15015 Reg. 7 8.066 2.849 0.353 4.73 11.71 Hard Loose Bulgur 20012 Reg. 7 1.888 0.330 0.175 1.45 2.49

Supplemental Information, page 15

Hazelnuts 12120 Reg. 8 41.222 6.723 0.163 31.31 51.71 Imported Apricots 09021 Reg. 10 4.874 2.301 0.472 2.29 10.83 Imported Beetroot 11080 Reg. 8 1.980 0.479 0.242 1.12 2.42 Imported Broad beans (Pulses)

16052 Reg. 7 1.826 0.517 0.283 1.09 2.76

Imported Cherries 09070 Reg. 9 6.944 3.442 0.496 4.11 14.98 Imported Coconut 12104 Reg. 9 4.422 0.924 0.209 2.75 5.48 Imported Dates 09421 Reg. 7 4.319 0.797 0.185 3.33 5.49 Imported Figs 09089 Reg. 7 3.716 1.576 0.424 1.61 6.52 Imported Fillet Steak(excl. round &sirloin)

13065 Reg. 10 10.418 2.735 0.263 4.91 15.24

Imported Grapes 09129 Reg. 9 5.148 1.772 0.344 2.32 7.74 Imp. Green beans (Pulses) 11052 Reg. 8 2.881 0.631 0.219 2.34 4.29 Imp. Grnd Mutton (Fresh) 17224 Reg. 9 10.873 2.000 0.184 7.65 13.57 Imported Guava 09139 Reg. 8 4.056 1.622 0.400 2.19 7.56 Imported Kiwi 09148 Reg. 7 4.052 0.977 0.241 2.69 5.24 Imported Melon 09181 Reg. 8 2.375 0.398 0.168 1.73 2.92 Imported Okra 11278 Reg. 8 3.028 0.769 0.254 2.23 4.05 Imported Peanuts 16390 Reg. 10 14.235 4.915 0.345 6.44 24.00 Imported Pears, Premium 09252 Reg. 10 3.350 0.710 0.212 2.46 5.03 Imported Peas 11306 Reg. 17 3.165 1.299 0.410 1.51 5.05 Imported Plums 09279 Reg. 10 4.118 0.989 0.240 2.68 5.70 Imported Pomegranate 09286 Reg. 8 6.537 2.293 0.351 2.71 9.15 Imported Red Dates 09421 Reg. 8 3.530 1.008 0.285 1.87 5.03 Imported Red bell peppers 11821 Reg. 8 2.399 0.801 0.334 1.62 3.93 Imported Tomatoes 11529 Reg. 9 1.468 0.502 0.342 0.63 2.42 Imported White onion 11282 Reg. 9 1.266 0.282 0.223 0.83 1.61 Imported Zucchini 11477 Reg. 9 1.533 0.364 0.237 1.14 2.05 Instant Noodles 06583 Glob. 133 9.134 4.228 0.463 3.30 25.54 Instant noodles 06583 Reg. 20 12.428 4.668 0.376 4.40 24.89 Jasmine Rice 20444 Glob. 70 3.060 1.635 0.534 1.21 10.28 Kiln Bread 18075 Reg. 9 1.196 0.467 0.390 0.36 1.74 Labneh 01294 Reg. 9 5.488 1.050 0.191 4.00 7.40 Lamb (Fresh) with bones , Non-specific cut

17074 Reg. 9 17.394 6.508 0.374 6.15 27.50

Lamb (Fresh), boneless, Non-specific cut

17224 Reg. 9 14.171 5.263 0.371 5.79 21.85

Lamb chops 17074 Glob. 115 18.904 9.015 0.477 6.59 65.85 Lamb whole leg 17011 Glob. 120 15.705 6.152 0.392 5.56 33.41 Large size chicken eggs 01123 Glob. 140 5.112 1.226 0.240 2.13 9.15 Lasagne 20420 Reg. 9 7.349 1.851 0.252 5.56 11.20 Lemon 09150 Glob. 146 3.889 1.716 0.441 1.07 10.23 Lentils, Dry 16069 Glob. 117 3.658 1.663 0.455 1.30 10.12 Lettuce 11253 Glob. 141 3.196 1.397 0.437 0.87 9.00

Supplemental Information, page 16

Light whipping or whipping 01052 Reg. 7 8.682 2.906 0.335 5.19 13.56 Lime 09159 Reg. 18 3.110 1.416 0.455 1.55 7.34 Live chicken 05001 Glob. 65 7.647 1.963 0.257 3.28 12.83 Live lamb 17074 Reg. 8 18.939 5.889 0.311 10.27 25.54 Live mutton 17074 Reg. 9 20.779 6.355 0.306 11.18 31.14 Local cheese 01017 Reg. 6 13.912 8.065 0.580 4.67 26.83 Local curd 01012 Reg. 2 5.763 3.729 0.647 3.13 8.40 Local hard, dry cheese 01227 Reg. 7 7.163 2.305 0.322 3.30 11.03 Long gr rice - Family Pack 20444 Glob. 62 1.847 0.573 0.310 0.73 3.64 Long gr rice - Non-Parboiled 20444 Glob. 91 1.972 0.907 0.460 0.78 4.91 Long grain rice - Parboiled 20446 Glob. 107 2.622 1.284 0.490 0.77 8.39 Long grain rice - loose 20036 Reg. 6 1.400 0.300 0.214 0.88 1.70 Macaroni 20420 Glob. 108 3.178 1.207 0.380 0.77 7.19 Mackerel, un-cleaned 15046 Glob. 91 6.012 3.548 0.590 1.23 23.61 Maigre 15020 Reg. 7 5.615 1.725 0.307 2.88 8.60 Maigre fillet 15020 Reg. 7 13.202 3.560 0.270 7.24 17.12 Maize 20014 Glob. 97 1.373 0.784 0.571 0.28 3.74 Maize Flour White 20316 Glob. 79 1.303 0.524 0.402 0.52 2.97 Maize oil 04518 Reg. 9 4.116 0.954 0.232 3.10 5.95 Maize semolina 20466 Reg. 3 3.872 1.901 0.491 2.22 5.95 Malanga/yautia/tannia 11991 Reg. 10 2.437 0.921 0.378 1.27 4.04 Mango 09176 Reg. 85 3.003 2.218 0.739 0.39 14.06 Medium size chicken eggs 01123 Glob. 95 5.049 1.285 0.255 2.88 10.00 Melon 09181 Glob. 83 3.454 1.825 0.528 1.14 11.27 Melon, Honeydew 09184 Reg. 5 5.672 3.300 0.582 3.06 11.05 Milk, low - fat , pasteurized in plastic bag

01174 Glob. 10 1.208 0.311 0.257 0.59 1.66

Milk, low-fat, Pasteurized 01174 Glob. 135 1.738 0.797 0.458 0.66 5.45 Milk, low-fat, UHT 01174 Reg. 14 1.535 0.194 0.127 1.31 1.91 Milk, not pasteurized 01078 Reg. 12 1.518 0.581 0.383 0.96 2.73 Milk, powdered 01091 Glob. 98 15.172 5.068 0.334 7.73 31.67 Milk, un-skimmed Pasteurized

01078 Glob. 138 1.691 0.684 0.404 0.10 4.18

Milk, un-skimmed UHT 01078 Glob. 144 1.826 0.711 0.389 0.71 5.50 Millet, Sorghum 20031 Reg. 6 1.267 0.604 0.477 0.71 2.11 Moong dahl, loose 16080 Reg. 10 3.355 0.653 0.195 2.05 4.31 Mortadella, loose 07050 Reg. 12 6.352 1.071 0.169 4.84 7.88 Mortadella, prepacked 07050 Reg. 12 7.324 1.181 0.161 5.30 8.80 Mud Crab 15143 Reg. 14 10.831 4.484 0.414 3.58 18.32 Mullet 15055 Glob. 33 5.648 2.155 0.381 1.84 10.15 Mushrooms, dried 11268 Reg. 17 22.655 6.180 0.273 14.46 41.04 Mustard oil 04582 Reg. 8 4.707 1.277 0.271 3.52 7.23 Mutton Liver (Chilled) 17199 Reg. 8 10.529 2.283 0.217 8.65 15.58 Mutton Liver (Fresh) 17199 Reg. 9 11.798 2.601 0.220 8.41 16.32

Supplemental Information, page 17

Mutton chops 17074 Reg. 10 13.995 3.787 0.271 5.95 19.71 Mutton mixed cut 17224 Glob. 64 9.346 3.457 0.370 4.24 19.95 Mutton with bones (Refrigerated) , Non-specific cut

17074 Reg. 9 15.806 5.740 0.363 7.46 26.42

Mutton/goat liver 17199 Reg. 7 8.649 2.715 0.314 4.48 12.89 Native house chicken 05001 Reg. 17 11.289 2.266 0.201 7.71 15.20 Natural honey, Mixed blossoms

19296 Glob. 140 11.936 6.240 0.523 2.67 41.78

Oats, rolled 20038 Glob. 130 4.786 2.638 0.551 1.15 17.77 Okra 11278 Reg. 6 4.492 0.800 0.178 3.61 5.73 Olive Oil 04053 Glob. 10 9.211 2.084 0.226 4.83 11.75 Olive oil 04053 Glob. 149 13.912 5.046 0.363 4.74 30.33 Olive oil, standard 04053 Reg. 19 18.013 5.372 0.298 9.80 26.99 Onion 11282 Glob. 153 1.313 0.511 0.389 0.36 3.30 Orange 09200 Glob. 174 2.278 1.151 0.505 0.46 7.63 Palm oil 04055 Glob. 61 3.628 1.675 0.462 1.60 11.29 Papaya 09226 Glob. 84 2.403 1.263 0.526 0.80 6.76 Passion fruit 09231 Reg. 12 4.803 2.400 0.500 1.83 11.33 Peach 09236 Glob. 103 4.830 3.014 0.624 1.22 17.66 Peanut oil 04042 Glob. 45 4.508 1.984 0.440 1.89 12.23 Peanuts in shell 16087 Reg. 16 2.670 0.722 0.270 1.34 3.78 Peas, Tinned 11306 Reg. 7 2.398 0.587 0.245 1.82 3.42 Pineapple 09266 Glob. 145 4.011 2.401 0.599 0.74 17.29 Pita bread 18413 Glob. 80 3.169 1.524 0.481 0.93 8.13 Plum tomatoes 11529 Reg. 12 2.026 0.717 0.354 1.18 3.36 Pork and beef sausages 07004 Reg. 14 8.173 3.415 0.418 4.40 16.97 Pork ham, pressed 10150 Glob. 99 20.662 8.062 0.390 4.88 58.92 Pork ham, pressed, bulk or loose

10146 Reg. 12 11.314 3.506 0.310 7.64 17.77

Pork liver 10110 Reg. 13 5.866 1.205 0.205 3.75 7.40 Pork loin, without bones 10224 Reg. 16 9.273 1.451 0.157 7.64 12.72 Pork thigh, with bones 10028 Reg. 15 9.042 3.076 0.340 5.05 18.93 Pork, fillet 10036 Glob. 123 14.887 6.524 0.438 6.27 47.43 Pork, loin chop 10036 Glob. 123 10.489 3.587 0.342 3.76 25.30 Pork, ribs 10204 Glob. 135 11.995 4.788 0.399 4.56 44.23 Pork, shoulder 10080 Glob. 60 10.389 3.558 0.342 4.76 21.76 Pork, with bones, non-specific cut

10204 Reg. 13 11.214 2.988 0.266 7.58 17.03

Pork, without bones, non-specific cut

10005 Reg. 16 8.144 1.636 0.201 5.46 11.17

Poulty sausages (chicken or turkey)

07976 Reg. 11 6.897 3.009 0.436 4.29 15.17

Powdered milk, box 01091 Reg. 17 14.499 5.227 0.360 6.10 24.70

Supplemental Information, page 18

Powdered milk, in bag or box

01091 Reg. 12 11.860 1.881 0.159 9.01 14.53

Prawn/Shrimp, medium 15149 Reg. 18 12.527 2.535 0.202 8.18 18.15 Prawn/Shrimp, small 15149 Reg. 18 8.644 2.481 0.287 6.11 15.88 Premium rice #1 20444 Reg. 9 1.878 0.491 0.262 1.19 2.69 Premium rice #2 20444 Reg. 16 2.284 0.773 0.338 1.16 4.45 Premium rice #3 20444 Reg. 7 1.844 0.423 0.229 1.06 2.23 Processed honey, pure 19296 Reg. 7 13.436 4.012 0.299 7.78 19.26 Processed honey, pure 19296 Reg. 10 12.719 2.139 0.168 8.89 16.43 Processed shrimp 15149 Reg. 5 9.758 3.627 0.372 4.99 13.25 Pumpkin 11422 Reg. 22 1.654 0.785 0.475 0.86 4.52 Pure Sesame Oile 04058 Reg. 8 11.651 2.666 0.229 7.30 14.98 Radish, white 11637 Reg. 20 1.713 0.724 0.423 0.77 3.11 Red porgy 15101 Reg. 7 10.183 3.773 0.371 6.67 17.10 Red porgy fillet 15101 Reg. 4 13.482 6.490 0.481 8.50 23.01 Red snapper 15101 Reg. 38 6.821 3.018 0.442 2.46 14.48 Rice [Specified brand] 20450 Reg. 9 2.879 0.851 0.296 2.09 4.30 Rice flour 20061 Reg. 16 1.830 0.353 0.193 1.21 2.55 Roasted groundnuts/peanuts

16390 Glob. 103 9.260 3.708 0.400 3.99 23.78

Roll 28288 Glob. 96 3.688 1.891 0.513 1.29 8.78 Roll or bun, loose 28288 Reg. 16 3.340 1.135 0.340 2.05 6.18 Roll or bun, prepacked 28288 Reg. 20 4.445 1.031 0.232 2.65 6.03 Round steak 13891 Reg. 16 10.008 2.829 0.283 5.37 14.81 Round tomato, loose 11529 Glob. 152 2.054 0.875 0.426 0.67 4.46 Salted & semi-dried fish 83110 Reg. 15 7.292 2.033 0.279 3.75 11.14 Salted Butter 01001 Glob. 143 12.536 7.682 0.613 4.08 91.31 Salted crackers 18232 Glob. 133 8.954 3.252 0.363 3.08 18.68 Salted dry cod 15018 Reg. 2 10.560 2.043 0.193 9.11 12.00 Salted duck egg 01255 Reg. 13 4.576 0.746 0.163 3.22 6.09 Sardines 15089 Reg. 2 2.455 1.327 0.541 1.52 3.39 Sattu 20130 Reg. 6 2.184 0.714 0.327 1.29 3.01 Sea Bass 15004 Glob. 64 11.318 5.427 0.480 2.96 20.64 Sea Crab 15139 Glob. 44 6.813 3.122 0.458 2.58 13.86 Sea Lobster 15154 Reg. 11 22.112 5.197 0.235 12.22 30.52 Semolina, Suji 20466 Reg. 5 1.298 0.181 0.139 1.17 1.60 Sesame oil or oil Alserg 04058 Reg. 7 11.637 1.437 0.123 9.49 14.23 Short past with eggs 20420 Reg. 2 2.646 0.164 0.062 2.53 2.76 Short pasta 20420 Glob. 126 2.367 1.129 0.477 0.62 6.62 Short-grained rice 20452 Glob. 67 1.432 0.502 0.350 0.71 3.12 Short-grained rice - prepacked

20452 Reg. 3 2.413 1.212 0.502 1.44 3.77

Shrimps 15149 Glob. 117 17.373 7.058 0.406 6.51 53.49 Sirloin steak 13919 Reg. 17 11.252 4.265 0.379 5.41 21.41

Supplemental Information, page 19

Sliced White bread 20075 Glob. 148 3.133 1.358 0.433 1.16 10.72 Sliced ham, pork 10150 Reg. 16 19.922 6.466 0.325 8.40 36.63 Small fresh fish 15039 Reg. 15 3.858 1.099 0.285 2.61 6.16 Smoked fish 15131 Reg. 10 8.374 2.326 0.278 4.55 11.75 Smoked salmon 15077 Glob. 75 37.247 15.078 0.405 8.77 83.40 Snack crackers 25071 Reg. 21 8.169 2.342 0.287 3.78 12.43 Soft/ Loose Bulgur 43060 Reg. 6 2.720 0.545 0.200 1.98 3.64 Sole 15028 Reg. 12 8.075 2.860 0.354 4.89 12.28 Sour cream 01056 Glob. 92 6.439 4.116 0.639 0.74 29.61 Soybean oil 04044 Glob. 81 4.101 1.690 0.412 2.11 12.18 Spaghetti 20420 Glob. 155 2.992 1.225 0.409 0.77 7.73 Spaghetti, with eggs 20420 Reg. 3 3.473 1.451 0.418 2.46 5.13 Spanish Mackerel 15051 Reg. 13 7.027 1.701 0.242 3.59 9.06 Spinach 11457 Glob. 121 3.072 2.108 0.686 0.66 10.89 Spinach Chinese 11457 Reg. 17 2.026 0.691 0.341 0.94 3.37 Squid 15175 Glob. 83 11.298 4.607 0.408 3.45 25.59 Squid, small 15175 Reg. 15 7.493 1.638 0.219 5.35 10.14 Sunflower oil 04060 Glob. 140 3.657 1.522 0.416 1.44 10.06 Surubi fillet 15234 Reg. 2 13.913 11.407 0.820 5.85 21.98 Sweet Potatoes 11516 Glob. 93 1.139 0.621 0.545 0.29 2.75 Taro 11518 Reg. 19 1.716 0.696 0.406 0.77 3.62 Thailand Rice 20444 Reg. 10 1.943 0.709 0.365 0.84 3.10 Tilapia 15261 Glob. 66 4.649 1.455 0.313 1.88 9.98 Tilapia fillet 15261 Reg. 11 10.976 3.638 0.331 5.45 14.81 Tinned Button Mushrooms 11260 Glob. 72 7.191 3.113 0.433 2.44 18.01 Tinned green peas 11306 Glob. 112 4.189 1.718 0.410 1.60 11.28 Tinned pineapple 09267 Glob. 119 3.941 2.100 0.533 1.13 13.03 Tinned sweet corn/Maize 11176 Glob. 126 4.055 2.261 0.558 1.08 14.45 Tinned white beans in tomato sauce

16051 Glob. 111 4.405 1.700 0.386 1.30 9.33

Toast petit beurre brown crunchy

28351 Reg. 6 1.885 0.918 0.487 1.09 3.48

Tuna 15117 Reg. 7 5.908 1.898 0.321 2.83 8.29 Tuna Steak 15117 Reg. 11 9.664 3.141 0.325 6.18 17.72 Tuna fish fresh 15117 Reg. 6 7.399 3.177 0.429 4.13 12.00 Tuna steaks 15117 Reg. 13 10.983 5.462 0.497 4.92 20.75 Veal breast (non-refrigerated), with bones

17271 Glob. 74 9.086 3.421 0.376 4.22 18.70

Veal chops 17128 Glob. 61 12.326 5.256 0.426 5.79 38.00 Veal with bones (Refrigerated)

17128 Reg. 9 13.304 1.999 0.150 10.24 15.81

Veal, Boneless (Fresh unchilled)

17271 Reg. 9 12.876 2.018 0.157 10.57 15.51

Veal, Boneless 17271 Reg. 9 11.799 2.405 0.204 8.52 15.16

Supplemental Information, page 20

(Refrigerated) Veal, with bones 17128 Reg. 7 9.542 2.035 0.213 6.92 12.35 Vegetable oil 04513 Glob. 131 3.184 1.047 0.329 0.43 7.33 Vegetarian vegetable soup 06067 Reg. 9 17.526 3.459 0.197 12.31 22.17 Vermicelli (Angel Hair) 43114 Glob. 126 3.158 1.903 0.603 0.66 12.17 Vermicelli (angel hair), w eggs 20420 Reg. 1 2.714 . . 2.71 2.71 Walnuts 12155 Reg. 7 34.465 6.048 0.175 29.48 46.73 Water Spinach 11503 Reg. 14 1.739 0.512 0.294 0.81 2.69 Watermelon 09326 Glob. 146 2.370 1.144 0.483 0.54 5.96 Wheat Semolina (Suji) 20466 Glob. 66 2.678 1.730 0.646 0.86 7.66 Wheat flour, loose 20481 Reg. 15 1.162 0.362 0.312 0.72 1.83 Wheat flour, not self-rising 20481 Glob. 155 1.173 0.548 0.467 0.44 3.92 White Pomfret 15068 Reg. 14 11.367 4.240 0.373 5.23 18.00 White bread 20075 Glob. 120 2.408 1.116 0.463 0.49 7.89 White bread, unsliced loaf 20075 Reg. 13 2.511 0.797 0.317 1.33 3.59 White or brown chicken eggs, Small size [domestic]

01123 Reg. 9 5.166 1.159 0.224 3.90 7.26

White potato 11354 Reg. 19 1.602 0.690 0.431 0.44 3.11 White rice #1 20450 Reg. 15 1.255 0.303 0.242 0.89 1.99 White rice #10 20450 Reg. 11 1.705 0.495 0.290 0.98 2.56 White rice #3 20450 Reg. 15 1.306 0.342 0.262 0.80 2.17 White rice #4 20450 Reg. 7 1.124 0.084 0.075 0.98 1.21 White rice #5 20450 Reg. 8 1.242 0.329 0.265 0.87 1.99 White rice #6 20450 Reg. 6 1.365 0.461 0.338 0.83 2.21 White rice #7 20450 Reg. 7 1.560 0.321 0.205 1.00 1.98 White rice #9 20450 Reg. 8 1.603 0.298 0.186 1.14 1.97 White rice, 20% broken 20452 Reg. 6 1.704 0.846 0.496 1.09 3.40 White rice, 25% broken 20452 Glob. 61 1.527 0.593 0.389 0.47 3.37 White rice, Medium Grain 20450 Glob. 66 1.534 0.555 0.362 0.75 3.48 White rice, medium grain - prepacked

20450 Reg. 8 1.727 0.582 0.337 0.80 2.88

White sugar 19336 Reg. 146 1.604 0.421 0.262 0.46 3.22 White sugar, bulk 19336 Reg. 8 1.582 0.263 0.166 1.13 1.86 White sugar, family size 19336 Reg. 12 1.406 0.405 0.288 0.77 2.32 White sugar, granulated, cane 19336 Reg. 7 1.626 0.300 0.185 1.31 2.15 White sugar, loose 19336 Reg. 17 1.547 0.398 0.257 0.44 2.05 White wheat 20087 Reg. 7 0.894 0.414 0.463 0.61 1.79 Whole Shrimps 15149 Glob. 64 13.772 7.064 0.513 5.62 48.09 Whole chicken 05001 Glob. 134 6.817 2.343 0.344 1.24 17.29 Whole chicken (Frozen) 05001 Reg. 10 5.887 0.622 0.106 5.18 7.10 Whole chicken - Broiler 05006 Glob. 129 7.350 2.512 0.342 3.81 16.92 Whole wheat bread 18075 Glob. 128 3.134 1.085 0.346 0.61 5.99 Wholemeal flour, Atta 20080 Glob. 12 1.301 0.562 0.432 0.68 2.27 Yellow maize flour 20016 Reg. 6 1.737 0.521 0.300 1.27 2.66

Supplemental Information, page 21

Yoghurt drink 01116 Reg. 16 4.448 1.291 0.290 2.64 7.28 Yoghurt, fruit 01294 Reg. 17 6.113 1.450 0.237 4.01 10.19 Yoghurt, plain 01116 Glob. 142 4.336 1.608 0.371 0.93 9.75 Yoghurt, with flavor 01116 Reg. 13 4.552 1.133 0.249 2.52 6.49 Yogurt drink 01116 Reg. 11 3.433 0.745 0.217 2.35 4.42 Zubaida Fish 15068 Reg. 5 15.457 5.746 0.372 10.84 22.74 Zucchini 11477 Reg. 10 1.837 0.640 0.348 1.02 3.08 N 20,972

Note: Prices shown are in international dollars at 2011 prices, converted from local currency units at the food sector PPP exchange rate to adjust for cross-country differences in average prices of all food and non-alcoholic beverages, for a total of 159 countries with complete ICP data. Prices are converted from original quantity units to costs per kilogram of product as purchased, and winsorized at 1% and 99% to eliminate unrepresentative extremes. Source: Computed from file data obtained by permission from the International Comparison Program (ICP)

Supplemental Information, page 22

Table A7. Cost of nutrient adequacy (CoNA) and caloric adequacy (CoCA) by country

Country CoNA CoCA Country CoNA CoCA

Albania 2.02 0.62 Lesotho 1.67 0.75 Algeria 1.75 0.70 Liberia 1.39 0.74 Angola 2.02 0.78 Lithuania 1.79 0.69 Armenia 2.27 0.98 Luxembourg 1.22 0.25 Australia 1.43 0.36 Macao SAR, China 2.10 0.88 Austria 1.56 0.36 Macedonia, FYR 2.50 0.63 Azerbaijan 2.31 0.74 Madagascar 1.78 0.67 Bahrain 1.29 0.63 Malawi 1.45 0.31 Bangladesh 1.57 0.55 Malaysia 1.65 0.73 Belarus 1.54 0.72 Maldives 1.77 0.88 Belgium 1.33 0.26 Mali 1.74 0.48 Benin 1.40 0.27 Malta 1.79 0.71 Bhutan 1.61 0.66 Mauritania 1.53 0.82 Bolivia 2.21 0.93 Mauritius 1.47 0.40 Bosnia and Herzegovina

1.92 0.57 Mexico 1.97 0.57

Botswana 1.69 0.66 Moldova 1.80 0.85 Brazil 1.28 0.68 Mongolia 1.45 0.64 Brunei Darussalam 1.52 0.53 Montenegro 2.67 0.64 Bulgaria 1.90 0.69 Morocco 1.43 0.55 Burkina Faso 1.54 0.48 Mozambique 1.92 0.55 Burundi 1.23 0.35 Myanmar 1.71 0.75 Cambodia 1.89 0.81 Namibia 1.37 0.61 Cameroon 1.59 0.68 Nepal 1.45 0.45 Canada 1.93 0.97 Netherlands 1.30 0.26 Cape Verde 1.47 0.50 New Zealand 1.63 0.41 Central African Republic 1.60 0.30 Nicaragua 2.60 1.35 Chad 1.64 0.37 Niger 1.48 0.54 Chile 1.80 0.81 Nigeria 1.80 0.63 China 1.53 0.62 Norway 1.79 0.34 Colombia 2.02 0.87 Oman 1.91 1.10 Comoros 2.32 0.67 Pakistan 1.53 0.68 Congo, Dem. Rep. 1.23 0.36 Palestinian Territory 1.52 0.84 Congo, Rep. 1.84 0.62 Panama 2.21 1.29 Costa Rica 1.84 0.90 Paraguay 1.59 0.72 Côte d'Ivoire 1.17 0.29 Peru 1.96 1.23 Croatia 1.69 0.49 Philippines 1.93 0.91 Cuba 1.96 1.88 Poland 1.40 0.46 Cyprus 1.92 0.64 Portugal 1.14 0.29 Czech Republic 1.33 0.30 Qatar 1.43 0.87 Denmark 1.25 0.33 Romania 1.60 0.60 Djibouti 0.94 0.60 Russian Federation (EUO) 1.75 0.53

Supplemental Information, page 23

Dominican Republic 2.28 1.00 Rwanda 1.24 0.29 Ecuador 2.00 1.19 São Tomé and Principe 1.39 0.73 Egypt, Arab Rep. (AFR) 2.17 0.98 Saudi Arabia 1.06 0.59 El Salvador 2.24 1.25 Senegal 1.25 0.48 Equatorial Guinea 2.16 1.37 Serbia 1.86 0.59 Estonia 1.56 0.44 Seychelles 1.73 0.61 Ethiopia 1.40 0.43 Sierra Leone 1.49 1.00 Fiji 1.41 0.71 Singapore 1.51 0.67 Finland 2.16 0.29 Slovakia 1.72 0.48 France 1.37 0.27 Slovenia 1.49 0.42 Gabon 1.71 0.72 South Africa 1.60 0.62 Gambia, The 2.02 0.63 Spain 1.27 0.45 Germany 1.50 0.31 Sri Lanka 1.70 0.76 Ghana 1.18 0.78 St. Kitts and Nevis 2.08 1.12 Greece 1.76 0.59 Sudan (AFR) 2.04 0.67 Guatemala 2.21 1.08 Suriname 2.47 1.11 Guinea 1.80 1.11 Swaziland 1.59 0.66 Guinea-Bissau 1.87 0.86 Sweden 1.28 0.34 Haiti 2.01 1.11 Switzerland 1.11 0.35 Honduras 2.14 0.82 Taiwan, China 1.82 1.19 Hong Kong SAR, China 1.72 0.94 Tajikistan 2.57 0.79 Hungary 1.88 0.58 Tanzania 1.52 0.48 Iceland 1.56 0.42 Thailand 1.83 0.98 India 1.26 0.24 Togo 1.34 0.44 Indonesia 1.52 0.82 Trinidad and Tobago 1.90 1.10 Iraq 1.81 0.79 Tunisia 1.73 0.52 Ireland 1.40 0.36 Turkey 1.72 0.61 Israel 1.50 0.41 Uganda 1.14 0.45 Italy 1.59 0.33 Ukraine 1.95 0.54 Jamaica 2.31 1.21 United Arab Emirates 1.37 0.89 Japan 3.24 0.67 United Kingdom 1.27 0.27 Jordan 0.78 0.53 United States 1.99 0.77 Kazakhstan 1.44 0.62 Uruguay 1.46 0.63 Kenya 1.67 0.57 Venezuela, RB 2.73 0.91 Korea, Rep. 3.98 0.67 Vietnam 1.57 0.90 Kuwait 0.84 0.51 Yemen 1.69 1.00 Kyrgyzstan 1.94 1.06 Zambia 1.62 0.24 Lao PDR 1.78 0.78 Zimbabwe 1.60 0.40 Latvia 1.82 0.58

Note: Data shown are costs per day in 2011 for a representative adult woman, converted from local currency units to international dollars at PPP exchange rates for all household expenditure. Analysis of affordability excludes 14 countries with incomplete national accounts data (Jordan, UAE, Kuwait, Yemen, Saudi Arabia, Qatar, Jamaica, Trinidad and Tobago, Bahrain, Oman, Palestinian Territory and Iraq).

Supplemental Information, page 24

Table A8. Nutritional outcomes used to test for associations with diet costs

N Mean Std. Dev. Min Max

Anemia prevalence in non-pregnant women (%)

134 28.20 13.18 8.10 64.40

Anemia prevalence in children under 5 (%)

134 36.62 21.09 6.40 87.40

Zinc deficiency prevalence in 2005 (%)

134 17.18 10.31 3.60 48.40

Obesity prevalence in men (%)

132 12.28 8.28 1.00 31.90

Obesity prevalence in women (%)

132 18.22 8.88 1.90 43.60

Vit. A deficiency prevalence in children under 5 (%)

91 0.28 0.19 0.03 0.67

Stunting prevalence in children under 5 (%)

26 29.33 12.27 4.00 46.60

Note: Data on anemia prevalence is from Stevens et al. (2013) where anemia for under 5 children is defined as those with hemoglobin concentration < 110 g/dL and among non-pregnant women defined as those with hemoglobin concentration < 120 g/dL. Data on zinc deficiency prevalence is adopted from Wessells and Brown (2012) who estimated the prevalence of inadequate zinc intake using FAO food balance sheet data and average zinc requirements. Data on obesity prevalence is from WHO’s Global Health Observatory Data Repository. Data on vitamin A deficiency (VAD) prevalence is from Stevens et al. (2015) which estimated VAD among children based on serum retinol concentrations using a Bayesian hierarchical model. Data on stunting (under-5 children whose height for age is more than two standard deviations below the median for the international reference population) prevalence is from WHO’s Global Database on Child Growth and Malnutrition.

Supplemental Information, page 25

Table A9. Regression results for structural transformation and nutrient shadow prices

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Energy Calcium Vit E Vit B2 Vit C Selenium Vit B12 Folate Vit B6

lnGNI -0.042 0.044 -0.005 -0.127** -0.004 -0.032 -0.000 0.031 0.138

(0.048) (0.031) (0.059) (0.055) (0.009) (0.029) (0.013) (0.036) (0.164)

lnGNI squared 0.006 -0.005 0.000 0.014** 0.001 0.003 -0.000 -0.004 -0.014

(0.005) (0.004) (0.007) (0.006) (0.001) (0.003) (0.001) (0.004) (0.017)

lnGNI cubed -0.000 0.000 0.000 -0.001** -0.000 -0.000 0.000 0.000 0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.001)

Services share of labor force

-0.000 -0.000** 0.000 0.000** 0.000** -0.000 -0.000 0.000 0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

Urban share of population

-0.000** 0.000 0.000*** -0.000 -0.000 -0.000 0.000 0.000*** 0.000*

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

Rural travel time to city >50k (log)

0.000 -0.000 0.000 -0.000 -0.000 0.001* 0.000 0.000 -0.000

(0.000) (0.000) (0.001) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

Rural electricity access (pop share)

-0.000 -0.000 0.000 0.000 -0.000 0.000 0.000 0.000 -0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

Import tariffs (ave. duty applied)

-0.003 0.003* 0.007* -0.003 -0.001 0.002 0.000 -0.003 0.001

(0.003) (0.002) (0.004) (0.003) (0.001) (0.002) (0.001) (0.002) (0.003)

Constant 0.102 -0.123 0.022 0.373** 0.007 0.087 0.002 -0.086 -0.453

(0.140) (0.091) (0.172) (0.164) (0.028) (4.023) (0.037) (0.106) (0.526)

Number of observations

132 132 132 132 132 132 132 132 132

Note: Dependent variable from model 1-9 are the natural log of shadow price semi-elasticities for nutrients with the most binding EAR constraints. Given that EAR constraints of these nutrients are often non-binding with zero shadow prices, models 1-9 represent tobit censored regressions which yield more consistent coefficient estimates than robust OLS regressions. Standard errors in parentheses, with significance levels denoted *** p<0.01, ** p<0.05, * p<0.1, from tobit regressions. All specifications include indicator variables for ICP regions (not shown).

Supplemental Information, page 26

Table A10. Regression results for structural transformation and the cost of nutritious diets as a share of all household expenditure (1) (2) (3) (4) (5) (6) (7)

lnGNI per capita

-3.881 -4.027 -3.365 -2.257 -5.277 -4.953 -4.088 (3.345) (3.363) (3.407) (3.179) (3.297) (3.301) (3.268)

lnGNI per cap., sq.

0.403 0.419 0.344 0.219 0.590 0.522 0.445 (0.380) (0.382) (0.387) (0.361) (0.376) (0.375) (0.373)

lnGNI per cap., cu. -0.017 -0.018 -0.015 -0.010 -0.025* -0.022 -0.019 (0.014) (0.014) (0.014) (0.014) (0.014) (0.014) (0.014) Services share of labor force

-0.001 -0.001

(0.003) (0.003)

Urban share of population

0.002 0.001

(0.002) (0.002)

Rural travel time to city >50k (log)

0.098*** 0.088***

(0.028) (0.029)

Rural electricity access (pop share)

-0.004*** -0.003**

(0.001) (0.001)

Import tariffs (ave. duty applied)

-0.047 0.005

(0.219) (0.216)

lnGNI per capita

Constant 12.876 13.329 11.370 7.613 16.174* 16.042* 12.503 (9.725) (9.775) (9.903) (9.258) (9.540) (9.578) (9.463)

N 126 126 126 126 126 125 125 R2 0.949 0.949 0.948 0.954 0.952 0.951 0.956 F 312.236 270.190 266.181 304.688 290.084 284.171 204.080

Note: Dependent variable is the natural log of the ratio of CoNA to per-capita household expenditure on food and non-alcoholic beverages. Standard errors in parentheses, with significance levels denoted *** p<0.01, ** p<0.05, * p<0.1, from robust regressions (rreg). All specifications include indicator variables for ICP regions (not shown).

Supplemental Information, page 27

Figure A1. Spatial variation in the cost of nutrient adequacy at PPP prices

Note: Data shown are the cost per day of a least-cost diet meeting all nutrient constraints for a healthy adult woman across 159 countries in 2011, converted to international dollars at purchasing power parity (PPP) price levels for all household goods and services in each country. Methods are as detailed in the text. Higher income countries often have higher prices in PPP terms, due to higher wages and other costs.

Figure A2. Spatial variation in the cost of nutrient adequacy at PPP(food) prices

Note: Data shown are the cost per day of a least-cost diet meeting all nutrient constraints for a healthy adult woman across 159 countries in 2011, converted to international dollars at purchasing power parity (PPP) price levels for all food and non-alcoholic beverages in each country. Food prices are often relatively high in lower-income countries, compared to average prices in other sectors and the price level of each sector in higher-income countries.

Supplemental Information, page 28

Figure A3. Spatial variation in the cost of caloric adequacy at PPP prices

Note: Data shown are the cost per day of a least-cost diet meeting just daily calorie for a healthy adult woman across 159 countries in 2011, converted to international dollars at purchasing power parity (PPP) price levels for all household goods and services in each country. Methods are as detailed in the text. Higher income countries often have higher prices in PPP terms, due to higher wages and other costs.

Supplemental Information, page 29

Figure A4. The cost of nutrient adequacy as a fraction of national poverty lines Panel A. Variation by geographic location

Panel B. Variation by income level

Note: Data shown are ratios of CoNA per day to national poverty lines in local currency terms, as detailed in the text, for 159 countries in 2011. Two outliers not shown are Gambia (2.98) and Madagascar (2.14).