Cost and affordability

of healthy diets across

and within countries
Background paper for The State of Food Security
and Nutrition in the World 2020
9 Cost and affordability
of healthy diets across
and within countries
Background paper for The State of Food Security
and Nutrition in the World 2020

By

Anna Herforth
Independent consultant and researcher

Yan Bai and Aishwarya Venkat

Doctoral students, Friedman School of Nutrition Science and Policy, Tufts University

Kristi Mahrt
Senior Research Analyst, International Food Policy Research Institute (IFPRI)

Alissa Ebel
Master’s student, Friedman School of Nutrition Science and Policy, Tufts University

William A. Masters
Professor, Friedman School of Nutrition Science and Policy, Tufts University

Food and Agriculture Organization of the United Nations

Rome, 2020
Required citation:
Herforth, A., Bai, Y., Venkat, A., Mahrt, K., Ebel, A. & Masters, W.A. 2020. Cost and affordability of
healthy diets across and within countries. Background paper for The State of Food Security and Nutrition
in the World 2020. FAO Agricultural Development Economics Technical Study No. 9. Rome, FAO.
https://doi.org/10.4060/cb2431en

The designations employed and the presentation of material in this information product do not imply
the expression of any opinion whatsoever on the part of the Food and Agriculture Organization of the
United Nations (FAO) concerning the legal or development status of any country, territory, city or area
or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dashed lines on maps
represent approximate border lines for which there may not yet be full agreement. The mention of specific
companies or products of manufacturers, whether or not these have been patented, does not imply that
these have been endorsed or recommended by FAO in preference to others of a similar nature that are
not mentioned.
The views expressed in this information product are those of the author(s) and do not necessarily reflect
the views or policies of FAO.

ISSN 2521-7240 [Print]

ISSN 2521-7259 [Online]
ISBN 978-92-5-133725-7
© FAO, 2020

Some rights reserved. This work is made available under the Creative Commons Attribution
NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/
licenses/by-nc-sa/3.0/igo).
Under the terms of this licence, this work may be copied, redistributed and adapted for non-commercial
purposes, provided that the work is appropriately cited. In any use of this work, there should be no
suggestion that FAO endorses any specific organization, products or services. The use of the FAO logo
is not permitted. If the work is adapted, then it must be licensed under the same or equivalent Creative
Commons license. If a translation of this work is created, it must include the following disclaimer along
with the required citation: “This translation was not created by the Food and Agriculture Organization
of the United Nations (FAO). FAO is not responsible for the content or accuracy of this translation. The
original [Language] edition shall be the authoritative edition.
Any mediation relating to disputes arising under the licence shall be conducted in accordance with the
Arbitration Rules of the United Nations Commission on International Trade Law (UNCITRAL) as at present
in force.

Third-party materials. Users wishing to reuse material from this work that is attributed to a third
party, such as tables, figures or images, are responsible for determining whether permission is needed
for that reuse and for obtaining permission from the copyright holder. The risk of claims resulting from
infringement of any third-party-owned component in the work rests solely with the user.
Sales, rights and licensing. FAO information products are available on the FAO website (www.fao.org/
publications) and can be purchased through publications-sales@fao.org. Requests for commercial use
should be submitted via: www.fao.org/contact-us/licence-request. Queries regarding rights and licensing
should be submitted to: copyright@fao.org.

Cover photo: ©Mint Images

 Contents
Preface vii
Acknowledgements viii
Acronyms ix
Executive summary xi

1 Introduction 1
1.1 Affordability of healthy diets as an aspect and indicator of food security 1
1.2 Affordability of diets goes beyond income and budget share 2
1.3 Methods for measuring least-cost diets 3
1.4 Diet quality criteria: energy, nutrients and food groups 4
1.5 Metrics for the cost of each diet 5
2 Methods 7
2.1 Food price data 7
2.2 Nutritional criteria for diet quality 8
2.3 Cost of diet indicators 9
2.4 Affordability indicators 16
3 Results: Global analysis 19
4 Results: Country case studies 31
4.1 United Republic of Tanzania 31
4.2 Malawi 34
4.3 Ethiopia 37
4.4 Ghana 38
4.5 Myanmar 41
5 Global simulation of policy impacts on the cost of a nutrient adequate diet 47
5.1 Impacts of trade policy and farm-to-market transport costs on the cost of nutrient
adequate diets 47
5.2 Impacts of transport cost reduction on the cost of nutrient adequate diets 52
6 Applications of new food price metrics 55
6.1 Implications for food price monitoring 55
6.2 Implications for poverty estimation 56
7 Policy options for improving affordability of healthy diets 59
8 Conclusions 63
References 65
Annexes 69
Annex 1. Food-based dietary guidelines used for calculating the cost of a healthy diet 71
Annex 2. Cost of the nutrient adequate diet using RDA compared to using H-AR 76
Annex 3. Average cost of each diet and national percentage of people who cannot afford
each diet, by region and income level (simple average across countries) 77
Annex 4. Lower- and upper-bound estimates of the number of people who cannot afford
each diet type 79
Annex 5. Nutrient adequacy of the selected diets 82
Annex 6. Methodological note on constructing a nutritious food poverty line in Myanmar 83
Annex 7. Data and methods to simulate impacts of policy change on diet costs 84

iii
Figures
Figure 1 Three increasing levels of diet quality 5
Figure 2 The cost of energy sufficient, nutrient adequate and healthy diets by region
and country income group 20
Figure 3 Premiums required to reach nutrient adequate and healthy diets, by region 21
Figure 4 Worldwide average costs of ten food-based dietary guidelines diets and four
EAT-Lancet reference diets 22
Figure 5 Affordability of diets relative to the global poverty line of USD 1.90/day 23
Figure 6 of
Affordability diets relative to actual food expenditure per capita 24
Figure 7 Proportion of people unable to afford each level of diet quality in 2017 25
Figure 8 Cost per person per day by food group, region and country income group
(2017 USD) 28
Figure 9 Average percentage contribution of each food group to the average global
cost of a healthy diet (2017 USD) 29
Figure 10 Cost contribution (percentage share of total cost) of each food group in a
healthy diet, by region in 2017 29
Figure 11 Average cost of a healthy diet by region in the United Republic of Tanzania,
2011–2015 32
Figure 12 Food group shares in the cost of a healthy diet by region in the
United Republic of Tanzania, 2011–2015 33
Figure 13 National and regional average diet costs per month by component in Malawi 35
Figure 14 Average cost of a nutrient adequate diet by demographic group in Malawi 36
Figure 15 of
Cost a nutrient adequate diet and wages in Ethiopia, 2002–2016 37
Figure 16 Price indexes by food group in Ethiopia, 2002–2016 38
Figure 17 Monthly cost of a healthy diet in Ghana, 2017–2019 39
Figure 18 Cost of a healthy diet by food group and region in Ghana, 2018 40
Figure 19 Frequency of missing price observations for each food group in Ghana 40
Figure 20 Cost of each food group for three food baskets in Myanmar 42
Figure 21 Calorie shares by food group in two Myanmar food baskets: food poverty line
and healthy diet with food preferences 42
Figure 22 Cost shares by food group in two Myanmar food baskets: food poverty line
and healthy diet with food preferences 43
Figure 23 Costs of items in each food group in Myanmar 43
Figure 24 Affordability of food poverty line, healthy diet and healthy diet with food
preferences baskets: % of the population living in households with food
expenditure below the cost of each basket in Myanmar 45
Figure 25 Official poverty line headcounts compared to healthy diet with food
preferences poverty line headcounts (%), by national, urban/rural and
expenditure quintile in Myanmar 45
Figure 26 Worldwide average nominal rates of protection, 2008–2011 and 2014–2017 48
Figure 27 Average nominal rates of protection by food group and country income group 49
Figure 28 Change in cost of a nutrient adequate diet due to agricultural trade policy,
2008–2011 51
Figure 29 Change in cost of a nutrient adequate diet due to agricultural trade policy,
2014–2017 51

iv
Figure 30 Change in cost of the nutrient adequate diet from lower transport costs,
2008–2011 54
Figure 31 Change in cost of the nutrient adequate diet from lower transport costs,
2014–2017 54
Figure A2.2 Regional mean cost by macroregion for all cost of diet indicators
(2017 international USD) 76

Tables
Table 1 Nutrient reference values for a representative adult woman 10
Table 2 Regions and populations represented by each FBDG 12
Table 3 Units of measure used for food-based dietary guidelines, by country
Table 4 Number
(plus of items required
EAT-Lancet diet) in healthy diet baskets, by food group 13
15
Table 5 Percentage of people who cannot afford each diet, by region and
Table 6 country income
Number group
of people who (population
cannot afford each diet,
weighted by region and
average) 26

country income group 27

Table 7 Cost of protein-rich foods in two Myanmar food baskets: healthy diet and
healthy diet with food preferences 44
Table 8 Official and healthy diet with food preferences poverty lines in Myanmar 46
Table 9 Nutrient adequacy of least-cost food baskets in Myanmar 46
Table 10 Transport cost changes used for simulation, by food group in 2008–2011
and 2014–2017 53
Table A1.1 Computing the cost of the healthy diet 71
Table A2.1 Regional mean cost by macroregion for all cost of diet indicators
Table A3.1 cost of eachUSD)
(2017 international
Average diet, by region and country income group 76

Table A3.2 (simple nationalacross

Averageaverage percentage of people who cannot afford each diet,
countries) 77

Table A4.1 by region and estimate

Lower-bound country income group (simple
of the number of people
average across who
(in millions) cannot
countries) 78

afford the cost of each diet, by region and country income group 79
Table A4.2 Upper-bound estimate of the number of people (in millions) who cannot
afford the cost of each diet, by region and country income group 81
Table A5.1
A7.1 Countries
Percent of by
nutrient
incomeneeds
group
met
and
bynumber
healthy of nominal
diets (meanrate
adequacy
of protection
ratio) 82

observations, 2008–2011 and 2014–2017 85

Table A7.2 Summary of nominal rates of protection data by food group and product 88

v
 Preface
This study aims to help bring affordability of healthy diets into focus for The State of
Food Security and Nutrition in the World 2020. Policymakers around the world have long
acknowledged that focusing on access to adequate calories, as reflected by the prevalence
of undernourishment, is only one part of the information needed to assess food security.
Beyond that perennial statistic, we need to know much more about access to healthy diets.
Food security, as defined by the Food and Agriculture Organization of the United Nations
(FAO) in 1996, is “when all people, at all times, have physical and economic access to
sufficient, safe, nutritious food that meets their dietary needs and food preferences for an
active and healthy life” (FAO, 1996). This definition embraces a vision of access to diverse,
healthy diets. The recent global measurement of the Food Insecurity Experience Scale
has broadened our understanding of the situation, by asking people directly about their
experience. However, no analysis to date has focused on access to nutritious food that meets
dietary needs.
This study and the metrics it provides are designed to answer the question: Can people
actually access nutritious food to meet dietary needs? In other words, do their food
environments provide nutrient adequate and healthy diets? More specifically, we ask:
In any country, when people go to the market, can they find a diet that meets dietary
recommendations? How much does it cost? Can people afford it?
Our findings highlight the high cost of healthy diets, which cost close to five times more
than energy sufficient diets. Even nutrient adequate diets cost more than the international
poverty line, without ensuring that recommended food group proportions are met. We find
that in total, about 3 billion people can’t afford the minimum cost of a healthy diet. There is
a vast difference in cost between the type of diets people can afford, and the type of diets
that will protect against malnutrition in all its forms and allow for a healthy and active life.

These results reveal that cost is an enormous barrier to the consumption of healthy diets.
Nutrition education that encourages consumption of balanced, healthy diets, including those
promoted in food-based dietary guidelines, cannot successfully lead people to consume diets
that are out of reach.
Nearly 25 years after the FAO definition of food security was published, this report
supports The State of Food Security and Nutrition in the World to include, for the first time,
a metric squarely focused on access to nutritious food to meet dietary needs. Many actions
are still needed to increase access to diverse, balanced diets to achieve our global shared
vision of food security. This report lays bare the need to do so.

vii
 Acknowledgements
This study uses indicators refined in the Changing Access to Nutritious Diets in Africa and
South Asia (CANDASA) project funded by UKAid and the Bill & Melinda Gates Foundation
(OPP1182628), initially developed in the Indicators of Affordability of Nutritious Diets in
Africa (IANDA) project, also funded by UKAid through a grant from Innovative Methods
and Metrics for Agriculture and Nutrition Actions (IMMANA). We thank our collaborators,
notably Kate Schneider at Tufts University and Harold Alderman at IFPRI for their input
and suggestions, and Daniel Sarpong (University of Ghana), John Nortey (Ghana Ministry
of Food and Agriculture), Stevier Kaiyatsa (Malawi Department of Economic Planning and
Development), Fulgence Mishili (Sokoine University of Agriculture), and Fantu Bachewe
and Derek Headey (IFPRI) for providing data that were used in the country case studies in
this study.
We also thank Cindy Holleman and Valentina Conti (FAO Agrifood Economics Division
[ESA]), Ramani Wijesinha-Bettoni (FAO Nutrition and Food Systems Division [ESN]),
Christian Derlagen, Emiliano Magrini and Valentina Pernechele (FAO, ESA), as well as other
members of The State of Food Security and Nutrition in the World 2020 Writing Team
for their technical contributions and guidance, and to Máximo Torero (FAO Economic and
Social Development Stream) for his support.
We thank Andrew Park for copy editing and Daniela Verona for the design and publishing
coordination.

viii
 Acronyms
AI Adequate Intake

AMDR Acceptable Macronutrient Distribution Range

AR Average requirement

CBN Cost of basic needs

CDRR Chronic Disease Risk Reduction Intake

CoCA Cost of Calorie Adequacy (energy sufficient diet)

CoNA Cost of Nutrient Adequacy (nutrient adequate diet)

CoRD Cost of Recommended Diet (healthy diet)

CoRD-FP Cost of Recommended Diet (healthy diet) with food preferences

CPI Consumer price index

DGLV Dark green leafy vegetable

EAR Estimated average requirement

EER Estimated energy requirement

FAFH Food away from home

FBDGs Food-based dietary guidelines

H-AR Harmonized average requirement

H-UL Harmonized upper level of intake

IANDA Indicators of Affordability of Nutritious Diets in Africa

ICP International Comparison Program

IOM Institute of Medicine (United States of America)

LCU Local currency units

MAFAP Monitoring and Analysing Food and Agricultural Policies (FAO)

MAR Mean adequacy ratio

MIS Market information systems

MoFA Ministry of Food and Agriculture (Ghana)

MPLCS Myanmar Poverty Living Conditions Survey

NPI Nutritional Price Index

NRP Nominal rate of protection

PPP Purchasing power parity

ix
RDA Recommended dietary allowance

UL Tolerable Upper Intake Level

UN United Nations

VAM Vulnerability Assessment and Mapping system (World Food Programme)

WHO World Health Organization

x
 Executive summary
Price and affordability are key barriers to accessing sufficient, safe, nutritious food to meet
dietary needs and food preferences for an active and healthy life. In this study, we identify
the least-cost items available in local markets to estimate the cost of three diet types: energy
sufficient, nutrient adequate, and healthy (meeting food-based dietary guidelines). For price
and availability we use the World Bank’s International Comparison Program (ICP) dataset,
which provides food prices in local currency units (LCU) for 680 foods and non-alcoholic
beverages in 170 countries in 2017. In addition, country case studies use national food price
datasets in United Republic of Tanzania, Malawi, Ethiopia, Ghana and Myanmar. In each
case we match the available items with food composition data to find the least-cost sources
of daily energy and nutrient adequacy, and match items to their food group for the least-cost
sources of a healthy diet.
We find that the global average cost of meeting daily energy needs using the most
affordable starchy staple at each time and place is USD 0.79 per day. The average cost of
meeting all essential nutrient requirements using the most affordable foods is USD 2.33
per day, and the average cost of a healthy diet as defined by national food-based dietary
guidelines is even higher. Using ten different definitions of a healthy diet published by
United Nations (UN) Member States, the range of the cost of healthy diets globally is between
USD 3.27 and USD 4.57 per day, with a point estimate based on median costs of USD 3.75.
The data reported here refer to the cost of purchasing the most affordable foods available
in each country. Consideration of food preferences and the time required to obtain and
prepare each food would raise daily costs, but our estimates provide a useful lower bound
on the affordability of healthy diets in each country and for the world as a whole.

Our findings reveal that healthy diets by any definition are far more expensive than the
entire international poverty line of USD 1.90, let alone the upper bound portion of the poverty
line that can credibly be reserved for food of USD 1.20. The cost of healthy diets exceeds
food expenditures in most countries in the Global South. These diets are unaffordable for
over 57 percent of the population in sub-Saharan Africa and Southern Asia, as well as high
proportions of people in South-eastern Asia (> 45 percent), Melanesia (> 40 percent) and
Latin America (> 20 percent).
In all, we estimate that 3 billion people globally lack sufficient income to purchase the
least-cost form of healthy diets recommended by national governments. The majority of
these reside in Southern Asia (1.3 billion) and sub-Saharan Africa (829 million), with high
numbers also in South-eastern Asia (326 million) and Eastern Asia (230 million). Our findings
show that 186 million people, mostly concentrated in Africa (149 million), cannot afford the
cheapest form of daily energy in their country, and 1.5 billion people cannot afford diets with
adequate levels of all essential nutrients. These numbers are comparable to the estimated
812–822 million people counted by the FAO prevalence of undernourishment indicator,
and the approximately 2 billion people who experience moderate or severe food insecurity
as measured by the FAO Food Insecurity Experience Scale (FAO, IFAD, UNICEF, WFP and
WHO, 2019).
Protein-rich foods including dairy, and fruits and vegetables, together make up more than
80 percent of the cost of healthy diets. Starchy staples and oils account for only 16 percent
of that cost, while fruits and vegetables account for 40 percent, and dairy and other protein
rich foods combined account for 44 percent. These proportions vary somewhat by region,
with dairy being notably more expensive in low-income countries.

xi
 Local prices vary significantly by region within countries, particularly for the most
expensive and nutrient-rich foods such as highly perishable fruits, vegetables, and animal
source foods. Country case studies in the United Republic of Tanzania and Malawi show that
the cost of nutrient-rich foods is more variable subnationally and by season than calories or
starchy staples. The Ethiopian case study shows that the cost of nutrient-dense food groups
has risen faster over time than starchy staples and oils. And the Myanmar case study shows
that the international standard method for calculating existing poverty lines is based on
food baskets that are nutrient deficient and imbalanced in terms of food groups. Hence,
we demonstrate the use of an alternative means of constructing a poverty line that would
meet dietary needs.
The findings presented here imply that meeting international development goals for food
and nutrition security would require: (a) lower prices for healthy, nutrient-rich foods including
fruits and vegetables, and protein-rich foods including dairy; and (b) greater support for
nutrition assistance and social safety net programmes. In addition, poverty lines may need
to rise, as current international and national poverty lines are insufficient for the purchase of
nutritious food to meet dietary needs. Existing food systems and assistance programmes do
not enable all people to access even the least-cost versions of healthy diets which would meet
dietary needs, and thereby do not fulfil longstanding aspirations for global food security.

Food-based dietary guidelines are official definitions of a healthy diet published by

governments to serve their citizens. They aim to help people consume healthy diets that
meet essential nutrient requirements and protect against diet-related non-communicable
diseases through a diet that is balanced among food groups, and are widely used in health
education and nutrition programmes throughout many countries. Our findings reveal that
education and individual behaviour change would be insufficient to achieve these dietary
objectives, because even the least expensive items from the required food groups are out of
reach for low-income people. For people to choose healthy diets, prices of those diets need
to decline, particularly from the most nutrient-rich food groups.

A variety of policy levers are needed to improve access to healthy diets. Our results show
that the cost of either nutrient-adequate or healthy diets in the market is more than many
people can afford; we do not account for food access via cultivation or wild harvesting. For
people and places with sufficient local resources, production and harvesting of vegetables,
legumes, fruits, dairy and eggs, fish and other foods can be important to provide access
to nutrient adequate and healthy diets where the market does not. Agriculture and rural
development should prioritize cost reductions for vegetables and fruits, and protein-rich
foods including dairy. More broadly, reducing the year-round cost of acquiring sufficient
quantities to meet dietary needs will require big changes in production and distribution.
The public and private actions needed to lower costs will vary by location and type of food.
Access to supplies from diverse sources within and between countries is also important to
overcome local resource constraints and gain resilience to shocks at any one place. Finally,
actions to improve storage and trade, combined with actions to improve production and
distribution, can sustain a rapid shift in agriculture and food systems that bring healthy diets
within reach.

xii
1 Introduction
Price and affordability are key barriers to accessing sufficient, safe, nutritious food to
meet dietary needs and food preferences for an active and healthy life. Previous literature
(e.g. Drewnowski and Darmon, 2005; Chastre et al., 2007; Masters et al., 2018; Headey
and Alderman, 2019) has shown clearly that more nutritious foods and diets cost more
than basic staples and energy sufficient diets. For the poorest people, acquiring sufficient
quantities of essential nutrients and nutritious food groups would consume a very large
proportion of their total income, or even exceed it. In such situations, affordability imposes
an insurmountable obstacle, so price and income constraints would need to be addressed
before nutrition knowledge and behaviour change could be effective drivers of food choice.

1.1 Affordability of healthy diets as an aspect and indicator of

food security
The longstanding, widely accepted definition of food security is “when all people, at all times,
have physical and economic access to sufficient, safe, nutritious food that meets their dietary
needs and food preferences for an active and healthy life” (FAO, 1996). Economic access
is measured here by affordability, defined as having enough income or other entitlement
to obtain sufficient food at each time and place. Affordability of diets reflects food prices
in comparison to an objective standard of income. We do not attempt to capture perceived
affordability, which would encompass an individual’s valuation of whether a product is worth
its cost. Our results omit the time cost of acquiring and preparing food. Physical barriers can
be very important for people with disabilities or other limits on access to the marketplaces
where prices are measured. Likewise, social barriers may also be important where certain
groups are prohibited from undertaking particular activities. All of these concerns ensure
that affordability is a necessary but insufficient condition for food security, making our
results a conservative lower bound on the gap between existing food systems and what
would be needed to achieve global development goals.
Until now, a limitation in measuring food security has been that indicators of economic
access to food have been limited to either income, or food price indexes that are defined in
ways that do not clearly relate to healthy diets. This study provides new metrics for the cost
and affordability of healthy diets, using newly available food price data to estimate economic
access to adequate food at the population level. Economic barriers based on prices and
income are not always the most important factor in food choice. Preferences, convenience
and other aspects of the food environment also matter, and may be more significant factors
to explain differences or changes in consumption patterns. Diet cost and affordability
measure one specific dimension of food security which differs from what is measured in
national food balance sheets or the prevalence of undernourishment, and also differs from
what is reported in the Food Insecurity Experience Scale. The cost and affordability data
reported here focus on diet quality, regarding sufficient quantities of nutritious foods for an
active and healthy life. Measuring the cost and affordability of safe food would require other
data regarding contaminant levels and associated disease burdens. We provide only one
application of a metric aimed at meeting food preferences.

1
Cost and affordability of healthy diets across and within countries

1.2 Affordability of diets goes beyond income and budget share

Income as such, without comparison to food prices and diet costs, has sometimes been
used as a measure of food security. Other income-related metrics include budget shares
that generally decline with higher income, such as the fraction of total expenditure spent on
food (known as Engel’s Law), and the fraction of food expenditure spent on starchy staples
(known as Bennett’s Law). The level of income and share of budget spent on food or starchy
staples successfully capture one dimension of affordability, but does not identify how market
prices for each type of food affect affordability. As shown in this study, the most affordable
items needed for nutrient adequate and healthy diets differ greatly in market price around
the world, ensuring that income alone cannot guarantee affordability of a nutritious or
healthy diet.
This study focuses on affordability, and whether food systems bring healthy diets within
reach of the poor. Which of the affordable items are actually chosen for consumption has
been studied elsewhere, including by economists using demand systems to estimate the
magnitude of change in household consumption or individual intake associated with changes
in price or income. These demand elasticities vary greatly for different types of food and
different aspects of diet quality.
Empirical evidence on how price and income affect diet quality is driven by the
observation that total energy consumption (kcal/person/day) can fall sharply in famine
situations, fluctuate seasonally, and vary with body size as well as physical activity levels;
however, it is known to change very little in response to changes in food prices or incomes
(Finaret and Masters, 2019). Behavioural responses to price and income changes typically
concern diet composition, as people substitute among foods to meet their daily energy needs.
The own- and cross-price elasticity of demand for any particular set of foods is a combination
of substitution effects (for example, eating more potatoes when the price of rice increases)
and income effects (for example, eating more preferred foods when the prices of all basic
staples decrease). The income effect of price changes is particularly important for foods
that occupy a large budget share among low-income people, as reducing their cost frees up
resources to buy other foods.
Price elasticities of demand and substitution effects are larger within food groups than
between them, and are larger for many nutrient-dense food groups than for basic staples
(Green et al., 2013; Cornelsen et al., 2015). Many previous studies have used observed
food choices to estimate demand systems, measuring a full set of own- and cross-price
elasticities in response to observed changes in a given population. This work has shown
that lowering prices of basic staples would have a much smaller impact on diet quality (via
income effects and low cross-price elasticities) than lowering prices of nutrient-rich foods
(via high own-price elasticities).
In summary, lowering food prices is not likely to affect demand for calories, but
which food prices are lowered will have a large impact on consumption choices. That is,
the cost of healthier foods specifically is important in understanding economic access to
healthy diets.
Our focus in this study is on a) the number of people for whom healthier diets are
out of reach, and b) the types of food that are most expensive, which makes healthy diets
unaffordable in existing food systems around the world. Where and when prices are low
enough for healthy diets to be affordable, then preferences become important and policy
interventions can focus on other drivers of food choice such as nutrition knowledge,
convenience and household decision-making, in addition to relative prices of food types.

2
 1 Introduction

1.3 Methods for measuring least-cost diets

In this study, we measure cost and affordability in terms of least-cost food baskets for three
dietary standards: energy sufficient, nutrient adequate, and healthy in terms of meeting
minimum quantities and proportionality among food groups.

Our estimates provide a conservative lower bound on the cost per day of meeting each
dietary standard. Least-cost diets allow for substitution among locally available items,
based on the most affordable combination of foods that meets each definition of diet
quality. Most marketplaces offer a variety of items to meet other needs, such as taste and
convenience, which consumers with higher incomes can afford to buy. Our aim is to
measure the lowest cost at which a country’s food systems deliver the calories and essential
nutrients and food groups required for each dietary standard, so as to identify the income
level required to afford that level of diet quality.
A key feature of the least-cost diets for each dietary standard is that the food items chosen
may vary over time and place, drawing on locally available or seasonal items as needed
to meet each dietary requirement. The least-cost energy sufficient diet substitutes among
the starchy staples based only on the energy content of each food. The least-cost nutrient
adequate diet recognizes substitution among alternative sources of each essential nutrient,
for example allowing month-to-month variation in vitamin A sources when different fruits
and vegetables are in season. For healthy diets, there is also similar substitution within each
food group, as defined by dietary guidelines.
The least-cost diets used to measure affordability are based on food prices and
availability in local markets, omitting information on the time cost of acquiring and preparing
meals at home. Higher levels of convenience and desirability, beyond what is reflected in
national dietary guidelines, would involve higher costs in money and time. By definition,
the thresholds used for this study are lower bounds, describing cost and affordability for the
least expensive of all available diets meeting each dietary standard. In most markets a wider
range of more expensive foods are also available, offering additional attributes that some
consumers are willing and able to acquire. Adding food preferences and convenience would
raise the estimated costs of reaching each standard, and raise the number of people who
cannot afford that level of diet quality.
In summary, our aims in this study are to quantify the cost of each diet, in absolute terms
and compared to the poverty line and typical food expenditures as standards of affordability,
and to quantify the number of people for whom even the cheapest nutritious or healthy diet
available on local markets is out of reach. To accomplish these aims:
♦ Income in low- and middle-income countries is measured as total household expenditure,
meaning the value of all goods and services reported to have been consumed. This includes
the value of food consumed from own production, but typically excludes the value of time
spent on household maintenance and childcare. In industrialized countries, income is
more readily measured using reported wages, salaries and other earnings. For poverty
and income distribution we use household surveys, and for a country’s total income per
person we use national accounts.
♦ Prices are measured at retail marketplaces, defined as the locations where people typically
acquire their food. These locations range from open markets with multiple vendors to
small neighbourhood shops and grocery stores of all sizes. Retail markets may offer
thousands of distinct items at different prices that vary over time and space. To compare
prices across and within countries, national statistical agencies identify representative
items at widely used marketplaces, and observe their price at regular intervals. We use
all prices reported by those national agencies, counting items with missing prices as not
available (or equivalently, having an infinitely high price).

3
Cost and affordability of healthy diets across and within countries

♦ Diets are defined in two ways: (1) in terms of their essential nutrients, meaning the
food composition in terms of total energy from carbohydrates, protein and fats plus all
known vitamins and minerals; (2) in terms of their food group classification in functional
terms such as fruits and vegetables, nuts and seeds, dairy, etc. Each level of diet quality
is defined in terms of different nutrient requirements or food group recommendations,
as explained below.

1.4 Diet quality criteria: energy, nutrients and food groups

The cost and affordability analysis conducted for this study focuses on three levels of diet
quality, defined successively in terms of energy sufficiency, nutrient adequacy, and overall
balance in terms of food groups. By definition, none of these diets reflects current
consumption patterns. They are benchmarks against which to compare incomes and current
food expenditure, which may be inadequate for nutrition and health and in any case would
be chosen based on other criteria such as cultural preferences, taste and convenience.
Some ultrapoor people may consume diets that are similar to the least-cost energy sufficient
diet, and preferences lead higher-income people to choose more nutritious foods, but many
people consume too little of some types of food for their long-term health, while others
consume more than they need.
The three types of diet referred to in this study are defined as follows:

1. “Energy sufficient” diets provide adequate calories for energy balance at a given level
of physical activity and body size, using only the least-cost starchy staple in each country.
For example, such a diet could consist of only the lowest cost type of rice in that country,
or only maize porridge.
2. “Nutrient adequate” diets provide not only adequate calories but also adequate levels
of all essential nutrients – namely, carbohydrates, protein, fat, vitamins and minerals,
within the upper and lower bounds needed to prevent deficiencies and avoid toxicity.
3. “Healthy” diets meet a set of dietary recommendations intended to provide nutrient
adequacy and long-term health. There are many definitions of a “healthy” diet pattern at
national, regional and global levels. In this case, we select the national food-based dietary
guidelines (FBDGs) of several countries from diverse regions, in order to represent a
range of dietary recommendations which have been articulated by UN Member States.
Dietary patterns have been studied extensively in the nutrition epidemiology literature,
relating specific foods and proportionality of different food groups to disease incidence
and prevention. Nutrients alone do not explain the relationship of food to health,
as there are many non-nutrient components of food, including but not limited to fibre,
phytochemicals, the food matrix, and interactions between these. FBDGs focus on foods
rather than nutrients, and typically concentrate on proportionality of food group intake.
Furthermore, proportionality in food group intake ensures a culturally acceptable diet
meeting at least a minimum standard for palatability and cultural norms, so the healthy
diet is closer to actual food preferences, in terms of dietary pattern, than the energy
sufficient or nutrient adequate diets.
A diet that meets calorie needs alone may be sufficient for short-term survival, but not
long-term health or well-being. It does not meet the definition of food security: adequate
food to meet dietary needs and food preferences for an active and healthy life. A nutrient
adequate diet meets calorie and nutrient needs (defined by a specific standard for specific
populations) but does not necessarily meet dietary guidelines (proportionality between food
groups), and does not necessarily satisfy food preferences. Healthy diets are protective of
long-term health, and FBDGs are also designed to meet general cultural food preferences.
Thus, ensuring access to healthy diets meets the full UN definition of food security.

4
 1 Introduction

Figure 1 illustrates the conceptual foundation of our work, which is to measure how big
of an increase in cost is needed to reach each level of diet quality. The height of each step in
the figure is determined by global and local food systems that determine the price of locally
available items needed to reach each dietary standard. Food systems differ in how steep
this stairway of affordability is for each population, and in the degree of assistance provided
along the way.

FIGURE 1 Three increasing levels of diet quality

THE STAIRWAY OF AFFORDABILITY, FROM SUBSTISTENCE TO HEALTH

HEALTHY DIET
includes foods from several food groups
and has greater diversity within food groups

NUTRIENT ADEQUATE DIET

meets required levels of all essential nutrients

ENERGY SUFFICIENT DIET

meets needs for short-term subsistence

Source: Authors’ own elaboration.

1.5 Metrics for the cost of each diet

We use the following new metrics to measure the cost and affordability of the three diets:

♦ The Cost of Calorie Adequacy (CoCA) and Cost of Nutrient Adequacy (CoNA) are metrics
of least-cost diets based on food composition and nutrient requirements. In terms of this
study and the terminology we use, CoCA refers to the cost of an energy sufficient diet,
and CoNA refers to the cost of a nutrient adequate diet.

♦ The Cost of Recommended Diet (CoRD) is a metric of a least-cost diet that meets food
based dietary guidelines, based on food group classifications. The Cost of Recommended
Diet with Food Preferences (CoRD-FP) is a variant which is the cost of a diet meeting
food-based dietary guidelines that accounts for local food preferences within groups.
In terms of this study and the terminology we use, CoRD refers to the cost of a healthy
diet, and CoRD-FP refers to the cost of a healthy diet with food preferences.

♦ The Nutritional Price Index (NPI) includes all foods from the country’s consumer price
index (CPI) but adjusts the weight on items in each food group to meet dietary guidelines,
while preserving the relative quantities of items within each group. This is similar to the
CoRD-FP, but expressed in price index terms.
Development of these metrics1 was based on the observation that current food price
measurements do not reflect foods or food baskets that would meet dietary needs.
For many decades, food prices have been collected and reported to meet a variety of needs.
For example, the FAO global food price index is designed to track the cost of the most widely
traded agricultural commodities on international markets, categorized into five groups:

1 The new metrics used to measure the cost and affordability of the three diets were created by projects led by
Tufts University with funding from UKAid and the Bill & Melinda Gates Foundation.

5
Cost and affordability of healthy diets across and within countries

grains, oils, meat, sugar and milk. Within countries, price reporting for agricultural market
information systems (MIS) focuses on commodities of greatest interest to producers and
traders, typically reporting farm-gate or wholesale prices of the most widely sold products.
Rural consumer prices may be collected to guide nutrition assistance, for example through
the USAID-funded Famine Early Warning System network (FEWS NET) and the World Food
Programme’s Vulnerability Assessment and Mapping system (VAM), and several of these
sources have been combined in the FAO Global Information and Early Warning System on
Food and Agriculture (GIEWS). However, these track prices for only a small number of items
in each location, typically staple foods. While these systems collect copious amounts of food
price data, the data has not yet been used to understand the cost of diets.
In principle, prices for the full range of foods consumed in each country are collected at
nationally representative locations for each country’s CPI to measure inflation and national
poverty lines. Across countries, nationally representative prices for similar items are also
reported every few years through the International Comparison Program (ICP) to compute
national price levels, the purchasing power parity (PPP) exchange rates between currencies,
and global poverty lines. Most of our analyses use prices collected for those purposes,
applying nutritional criteria to compute diet costs rather than the observed expenditures
used in standard price indexes.

Our metric of diet costs that most fully meets longstanding definitions of food security
is the cost of a healthy diet (CoRD), because it is based on UN Member States’ official
characterization of dietary needs for healthy and active lives. The CoRD reflects preferences
for balance between food groups, as the definition and quantities of foods in each group is
included among the criteria used by national authorities in defining their dietary guidelines.
Selecting the most affordable of the available items within each food group also reflects
preferences within groups, but only to the extent that available items for which price is
recorded in each dataset are actually purchased and consumed fairly commonly in that
country. For diet costs that reflect observed preferences and other constraints on food
choice, we introduce the healthy diet with food preferences (CoRD-FP) metric that reflects
consumption patterns within each food group, in a given population2 (Mahrt et al., 2019).
For example, in low-income countries the CoRD-FP will typically include less of all starchy
staples than people actually consume (because people consume a higher share of dietary
energy in starchy staples than recommended), but maintain the observed ratio among
maize, rice and wheat. By definition, taking account of preferences in the CoRD-FP leads
to a higher cost than the CoRD, which uses only the lowest-priced items within each group.
In this study we calculate daily costs for energy sufficient (CoCA), nutrient adequate (CoNA)
and healthy (CoRD) diets for all countries of the world. We also provide additional analyses
and address within-country variation through thematic case studies. We cannot calculate
the CoRD-FP for all countries because expenditure share weights are often not available, but
we apply the method for the Myanmar case study (see footnote 1) as a demonstration of the
price premium associated with maintaining observed consumption patterns of low-income
households within each food group.

2 In creating a CoRD-FP, one can select national average consumption patterns, or consumption patterns among
a specific group, if sufficient household survey data are available. In the Myanmar country case study in this
report, we demonstrate the use of the CoRD-FP for constructing a poverty line that would meet dietary needs,
and adhere to the standard methodology in poverty line measurement by characterizing consumption patterns
among the reference poor population.

6
2 Methods

KEY MESSAGES

We calculate the most affordable combination of locally available items needed for
an overall healthy diet at each time and place. This reveals food system performance
in bringing the required mix of foods within reach of low-income people.

For global comparisons, we use the World Bank’s ICP dataset of national average
prices for 2017 from 170 countries. For country studies, we use local data on
spatial and temporal variation from national statistical agencies.
We begin with the least-cost items for energy sufficient diets, then move on
to the cost of nutrient adequate diets, staying within upper and lower bounds
for 23 essential nutrients including acceptable macronutrient ranges. Finally,
we consider the cost of a healthy diet, which delivers those nutrients through
items selected from particular food groups in proportions specified by national
food-based dietary guidelines or other reference diets.
Using least-cost diets to measure food system performance brings together
available data on food items’ availability, price and nutrient composition at
each time and place, for comparison to universal benchmarks of human needs.
In combination with other kinds of data presented in The State of Food Security
and Nutrition in the World 2020, the method described here provides actionable
information to guide intervention towards global development goals.

2.1 Food price data

International Comparison Program food prices

To compare diet costs among all countries of the world we use retail prices reported by each
national statistical agency through the International Comparison Program (ICP). The ICP,
which is headquartered at the World Bank, works through regional bodies to obtain prices
for standard items across multiple countries, for the purpose of computing PPP exchange
rates. For this study we use ICP prices from 2017, the latest available round, published
in May 2020. These data provide an annual average, nationally representative price in
local currency units (LCU) for 680 foods and non-alcoholic beverages in 170 countries.
We excluded items that were non-caloric, ingredients, condiments, baby food, and items of
an unclear composition; these exclusions resulted in a final dataset of 552 items for the cost
of diet analysis. Prices were converted to United States dollars at 2011 PPP exchange rates
for comparison with the existing global poverty lines, because updated 2017 PPP levels were
not yet available at the time of writing. The individual item prices in ICP data are used by
permission and available to other researchers on request (details here: www.worldbank.
org/en/programs/icp).

7
Cost and affordability of healthy diets across and within countries

National sources for within-country variation in food prices

For country case studies of variation within United Republic of Tanzania, Malawi and
Ethiopia, we use price data collected by national statistical agencies for their official CPI.
In the United Republic of Tanzania, the dataset covers 21 regions and 71 food items between
2011 and 2015. In Malawi, the monthly prices of 55 food items were collected from 29 market
locations between 2007 and 2016. In Ethiopia, the data cover 97 food items collected from
120 markets from 2002 to 2016. Food prices in Ghana are from the Ministry of Food and
Agriculture’s monitoring information system, which covers all ten regions in Ghana and
74 items between 2017 and 2019. Food prices in Myanmar are derived from the 2015
Myanmar Poverty Living Conditions Survey (MPLCS), a nationally representative survey of
3 648 households implemented by the Government of Myanmar and the World Bank with
the purpose of monitoring socio-economic conditions. The survey collects data on 152 food
items consumed in the past seven days.

2.2 Nutritional criteria for diet quality

The quantities of each food required to meet each standard of diet quality are based on
nutritional criteria from sources detailed below. For energy sufficient and nutrient adequate
diets, these refer to the needs of a representative adult woman, not pregnant or lactating.
To calculate the estimated energy requirement (EER) of people in this reference group,
we use median heights and weights (163 cm and 57 kg) of the World Health Organization
(WHO) reference population at 30 years of age (Schneider and Herforth, 2020), and we
assume an “active” physical activity level which is recommended by the Institute of Medicine
of the United States of America for maintaining energy balance (IOM, 2006).3 This leads to
an EER of 2 329 kcal per day, which is used in all three diets (energy sufficient, nutrient
adequate, healthy). For the nutrient adequate diet, requirements for other essential nutrients
as detailed below are also specific to adult women; for the healthy diet, requirements for
balanced intake across food groups is generally applicable to the population as a whole.
To calculate energy and nutrient requirements we use the WHO reference population’s
median woman of reproductive age, for two reasons. First, their requirements fall roughly
at the median of the entire population distribution, in the sense that least-cost diets to meet
energy and nutrient requirements for people in this reference group are approximately the
median level of least costs for all sex-age groups over the entire life cycle (Bai, Herforth
and Masters, forthcoming). This reference group is therefore a good representation of the
population as a whole. Second, women of reproductive age are typically a nutritionally
vulnerable population group, as seen in their increased risk of dietary inadequacies (due to
social practices and norms that often disadvantage them in terms of access to food), which
have important consequences for themselves and their children. Previous studies have also
based their analyses on this reference group (Masters et al., 2018; Hirvonen et al., 2019).

3 “An average of 60 minutes per day of moderately intense physical activity (e.g. brisk walking or jogging at
3–4 mph) or shorter periods of more vigorous exertion (e.g. jogging for 30 minutes at 5.5 mph), in addition to
activities identified with a sedentary lifestyle, was associated with a normal BMI range and is the amount of
physical activity recommended for normal-weight adults.” (IOM, 2006, p. 94).

8
 2 Methods

2.3 Cost of diet indicators

Cost of an energy sufficient diet

The cost of an energy sufficient diet (CoCA) is defined as the minimum cost to meet energy
requirements using the least-cost available starchy staple food in each country. As detailed
above, the energy requirement we use for all indicators in this study is 2 329 kcal.

Cost of a nutrient adequate diet

The cost of a nutrient adequate diet (CoNA) is defined as the lowest-cost set of items available
at each time and place that would stay within the lower and upper bounds for dietary energy
and all essential nutrients. Operationally, the CoNA is measured as the minimum cost to
meet EER and relevant daily nutrient reference values of 23 macro- and micronutrients for
a representative person in a reference population. We apply global harmonized average
requirements (H-ARs), which are the levels of nutrients that meet the needs of 50 percent
of the healthy population of each age and sex; harmonized upper levels of intake (H-ULs),
the highest level likely to avoid risk of adverse health effects (Allen, Carriquiry and Murphy,
2019); and the Chronic Disease Risk Reduction Intake (CDRR) for sodium (National Academies,
2019; Schneider and Herforth, 2020). To calculate the CoNA, a linear program selects foods
to provide nutrient content above the H-ARs and below the H-ULs and the CDRR for sodium,
while specifying that the macronutrient intakes are within the Acceptable Macronutrient
Distribution Range (AMDR) set by the Institute of Medicine (IOM, 2006), and meeting the
energy content of exactly 2 329 kcal. These requirements are detailed in Table 1.
The CoNA is the daily cost of meeting all requirements listed in Table 1 using the most
affordable combination of foods available at a given time and place. The foods in the CoNA
offer the lowest-cost way to meet all known energy, macronutrient and micronutrient
needs. Each AR is defined as the level of usual intake needed to avoid deficiency diseases
for the median person in an otherwise healthy population. Thus, for half of that population,
requirements are lower so the true CoNA would be lower; likewise, for the other half,
requirements are higher so the true CoNA would be higher. For people who are less
physically active, energy requirements and diet costs are lower, and for people who are
more physically active, energy needs and diet costs are higher. The aim of the requirement
levels listed in Table 1 is to provide the best estimate of the average cost of meeting energy,
macro- and micronutrient needs within the population.
In a sensitivity analysis, we calculate the CoNA using the IOM recommended dietary
allowances (RDAs), or Adequate Intakes (AIs) if the latter is not larger than the H-ARs, in order
to determine the CoNA that would cover 97.5 percent of nutrient needs of the population.

9
Cost and affordability of healthy diets across and within countries

TABLE 1 Nutrient reference values for a representative adult woman

RDAs or AMDR AMDR

Nutrient Unit ARs upper UL
AIs* lower
1 Energy kcal 2 329 2 329

g g g
2 Protein 37.6 46.0 58.2 203.8
3 Lipids 51.8 90.6
4 Carbohydrates 262.0 378.5
5 Calcium mg 750 1 000 2 500
6 Iron2 mg 22.4, 11.2 22.4, 18 45

7 Magnesium1 mg 265 310 350

8 Phosphorous mg 580 700 4 000
9 Zincb,3 mg 8.9 10.2 25
10 Copper mg 0.7 0.9 5
11 Selenium mcg 45 55 300
12 Vitamin Cc mg 80 80 2 000
13 Thiamin mg 0.9 1.1
14 Riboflavinc mg 1.3 1.3

15 Niacin1 mg 11 14 35
16 Vitamin B6c mg 1.3 1.3 25

17 Folate1 mcg 250 400 1 000

18 Vitamin B12 mcg 2.0 2.4

19 Vitamin A4 mcg 490 700 3 000

20 Vitamin E mg 12 15 300
21 Sodium mg 2 300
22 Vitamin B5a mg 4.0 5.0
23 Cholinea mg 320 425 3 500
24 Manganesea,c mg 2.4 2.4 11
Notes: Values shown are for a 30-year-old, non-pregnant, non-lactating woman. Average requirements (ARs)
and Tolerable Upper Intake Levels (ULs) are taken from Allen, Carriquiry and Murphy (2019). *The values in
this column are recommended dietary allowances (RDAs – IOM) except where noted:
a. The value is an adequate intake (AI) value.
b. The value for zinc takes the assumption of an undefined diet.
c. The same values are used for both AR and RDA because the RDA/AI is not larger than the harmonized
average requirements (H-ARs).
1. The upper levels only refer to the supplement intakes, and therefore are not considered in the CoNA
calculation.
2. The H-AR of iron takes the assumption of a low-absorption diet for the AR value for the CoNA, and a
moderate-absorption diet for assessing nutrient content of the CoRD.
3. The H-AR of zinc takes the assumption of a semi-undefined diet for the AR value.
4. The upper level of vitamin A refers to the intake of retinol.
Source: Authors' own elaboration.

10
 2 Methods

Cost of a healthy diet

The cost of a healthy diet (CoRD) is defined as the lowest-cost set of items available at
each time and place that would meet requirements for each food group specified in food
based dietary guidelines (FBDGs). The CoRD can use any quantitative dietary standard,
such as the dietary guidelines issued by national governments and public health agencies.
Such guidelines provide expected quantities with upper and lower limits for the number of
servings and size of each serving from various food groups.
To meet each guideline’s requirements for dietary diversity within and across food
groups, the CoRD method selects the two cheapest items for starchy staples, protein-rich
foods (however defined), and fruits; the three cheapest items for vegetables (to meet
minimum diversity criteria articulated by some guidelines, and sometimes necessitating
subgroup inclusion, e.g. for dark green leafy vegetables [DGLVs]); and the one cheapest item
for oils and dairy, respectively. This method thus selects 11–13 individual food items specific
to each country to satisfy most FBDGs. This is aligned with the number of foods generally
recommended in guidelines; for example, China's FBDG explicitly aim for 12 different
individual foods or more in a day (Wang et al., 2016). The reason we choose more than
one cheapest food for most groups is indeed to meet the general “variety” requirement of
FBDGs. By definition all food items in the ICP list for a country are culturally acceptable,
because they are identified as having a significant level of expenditure share.
In this study, for each country, ten different definitions of a healthy diet are applied,
because there is no single definitive definition of a healthy diet. The definitions we use
are based on a range of FBDGs). We use several quantitative definitions from UN Member
States in diverse regions in this sensitivity analysis, and take the median and range of cost
in each country.
It is important to note that our purpose is to estimate the cost of healthy diets, not to
recommend what people should eat in any country. In order to find the range and best
estimate of what it would cost to eat a healthy diet, we apply a range of healthy diet definitions,
through recent, quantifiable FBDGs in diverse regions throughout the world. We choose a
range of FBDGs so that no one region would be more heavily weighted in the range and point
estimate of what it costs to consume a healthy diet.

Selection and quantification of FBDGs

Over 90 countries have national FBDG, most published within the last 20 years. Each FBDG
includes a variety of key messages and visual representations. A global review of FBDGs
has highlighted the similarities and differences of these guidelines (Herforth et al., 2019).
The review found that all FBDGs discuss proportionality and include a recommendation to
consume abundant fruits and vegetables, and almost all include recommendations to limit
sugar, fat and salt consumption. There is less convergence among FBDGs in other areas,
including whether to make dairy a separate category (as it is in 64 percent of countries),
and in the guidance on red meat, fats and oils, and nuts (Herforth et al., 2019). Not all
guidelines are quantitative; for example, Sweden’s guidelines focus on a short list of food
types to eat more or less of, and Brazil’s guidelines focus on consuming minimally processed
foods and enjoying food. Of those guidelines that appear quantifiable, there is a range of
quantifiability: some are missing critical information for one or more food groups, such as a
definition of serving size.
We selected a purposive sample of FBDGs to include a set of recent, fully quantifiable
FBDG from diverse regions globally. Country FBDGs were selected for having the most
clearly quantifiable guidelines within each region, and for recency of publication (within the
last ten years). In Asia and the Pacific, three countries were selected: one from Eastern Asia
(China), one from Southern Asia (India), and one from South-eastern Asia (Viet Nam), due to

11
Cost and affordability of healthy diets across and within countries

different diet patterns. In Europe, two were selected: one from Southern Europe (Malta) and
one from Western Europe (Netherlands). In Latin America and the Caribbean, two countries
were selected: one from Latin America (Argentina), and one from the Caribbean (Jamaica).
Each FBDG was downloaded from the FAO FBDG repository (FAO, 2020b). For Viet Nam,
the current FBDG is unavailable on the FAO repository and was obtained via personal
communication with one of the people involved in its development (Annex 1). These ten
FBDGs are not the only quantifiable FBDGs; we chose a purposive sample in order to
prioritize more recent guidance, because newer guidelines are more likely to incorporate
newer evidence on diet-health relationships; and in order to choose diets from a sample
of countries representing the majority of the global population (Table 2). Averaging all
quantifiable FBDGs would have resulted in overemphasis of guidance and diet patterns from
Europe and from Latin America and the Caribbean, where most countries have national
FBDG, and underemphasis of regions with few quantifiable FBDGs. Most notably, at the time
of writing this study, Africa has only one country with a quantified FBDG.

TABLE 2 Regions and populations represented by each FBDG

Subregion Year
FAO FBDG Rationales
% of global country FBDG
subregion population published for inclusion

Southern 24 India 2011 Most populous country.

Asia
Eastern Asia 23 China 2016 Most recent and most
populous country.
Sub-Saharan 12 Benin 2015 Most recent, only quantitative
Africa choice.
Europe* 11 Netherlands, 2015 Netherlands: most recent
Malta quantitative choice available at
time of writing, and incorporates
sustainability considerations;
Malta 2015 (Southern Europe)
is equally recent, included
to represent a distinct
(Mediterranean) diet pattern.
South 9 Viet Nam 2019 Most recent and clearly
eastern Asia quantified.
Latin 8 Argentina, 2015 Most recent, quantitative;
America Jamaica two chosen to represent
and the Latin America and the
Caribbean Caribbean, including one
Small Island Developing State.
Western 6 Oman 2009 Most recent quantitative choice
Asia and available at time of writing.
Northern
Africa*
Northern 5 USA 2015 Only quantitative choice,
America most populous country.

Notes: * Subregions below 5 percent of the global population are combined into a macroregion. Subregions
with less than 1 percent of world population are not represented, including Central Asia (0.4 percent of world
population), Melanesia (0.01 percent), and Australia and New Zealand (0.4 percent).
Source: Authors' own elaboration.

12
 2 Methods

Across the ten national FBDGs listed in Table 2, there are broad similarities in food
groupings:
♦ Six guidelines use exactly the same six food groupings (starchy staples, protein-rich foods
including legumes/flesh/eggs, dairy, vegetables, fruits, and fats/oils); (two in Asia, one in
Europe, one in Africa, one in Northern America).
♦ Two use the same food groupings, except that legumes are grouped with starchy staples
instead of protein-rich foods (one in Asia, one in Latin America and the Caribbean).

♦ One (in Europe) uses the same food groupings, except that nuts are an additional food
group recommended daily.
♦ One (in Western Asia and Northern Africa) uses the same food groupings, except that
legumes and flesh/eggs are both required subgroups of the protein category.
♦ One (in Latin America and the Caribbean) uses the same food groupings, except that
dairy and flesh/eggs are grouped together, and legumes are a separate required group.

In contrast to these national FBDGs, the EAT-Lancet reference diet (Willett et al., 2019) has
12 food groups (including requirements for an exact amount of consumption of red meat,
poultry, fish, eggs, legumes, and starchy roots; food groups vary within four diet patterns).
In most cases the least-cost EAT-Lancet diet pattern is vegan. The EAT-Lancet diet was
formulated expressly for both health and environmental impact, while most FBDGs do not yet
incorporate environmental sustainability (the exception being the FBDG from Netherlands).
The groupings in these ten FBDGs represent only some of the possible ways foods can be
grouped, primarily based on how foods are used culinarily. Globally, approximately half of
FBDGs use six food groups; other food groupings are reviewed in Herforth et al., 2019 and
van’t Erve et al., 2017. The specifications of each FBDG are shown in Annex 1.
The ten national dietary guidelines used for this study were quantified based on the units
of measurement shown in Table 3. Where recommended quantities were stated in terms
of weight or volume, prices were adjusted to correct for the water added to cooked foods
(for example, boiled rice or pasta) or the water removed from dehydrated foods (for example,
dried fruits and powdered milk).

TABLE 3 Units of measure used for food-based dietary guidelines,

by country (plus EAT-Lancet diet)

Country Units of measure for each food group

Jamaica, Argentina, Kilocalories

EAT-Lancet
India Grams
Viet Nam Grams of macronutrients (carbohydrates, protein, lipids),
mg of calcium

Benin, Malta, USA, China, Servings (variously defined)

Oman, Netherlands

Source: Authors' own elaboration.

The number of grams, calories or servings is typically given as a range. We take the
mean of that range to calculate the cost of a healthy diet, unless servings are specified for an
active female (as in India) or a specific calorie target equal to approximately 2 300 (as in the
United States of America and Oman).

13
Cost and affordability of healthy diets across and within countries

Dataset preparation
Each unique food in the ICP dataset was classified into food categories according to each
of the ten FBDGs. Broadly, these categories included: starchy staples, protein-rich foods
(flesh foods, fish, seafood, and sometimes including legumes, eggs, nuts, and dairy), dairy
(sometimes including soy, small fish and crustaceans), fruits, vegetables, and fats. In addition
to the above-mentioned exclusions (mixed foods/dishes, infant foods, alcoholic beverages,
non-caloric items, ingredients, and foods with unclear food composition information),
we excluded foods that were expressly not recommended (trans fats, processed meats) and
one food for which there was no guidance regarding inclusion or recommended amount
(tomato paste). We used country-specific recommended amounts of each food group,
and country-specific inclusion/exclusion criteria for foods (e.g. fruit juice is included in three
countries, excluded in seven). Each unique food in the ICP dataset was also matched to
food composition information for edible portion, energy, protein, carbohydrates, lipids and
calcium. Price per gram edible portion and price per calorie edible portion were calculated
for each food item.

Food selection
The items selected for the least-cost healthy diet food basket in each country are the most
affordable items that meet the requirements of each FBDG. The FBDGs from Jamaica and
Argentina as well as EAT-Lancet reference diets specify quantities in calories, so the items
selected are those with the lowest price per calorie. The FBDG from India is specified
in terms of grams, so the items selected have the lowest price per gram. In Viet Nam,
the target quantities for several food groups are based on their macronutrient composition:
i.e. the requirement for starchy staples is given in terms of carbohydrates; protein-rich
foods in terms of protein; fats in terms of lipids; and dairy in terms of calcium. Fruits and
vegetable requirements are specified in terms of grams. The FBDGs from Benin, Malta,
United States of America, China, Oman and the Netherlands describe requirements in terms
of servings, using specific examples (e.g. grams of rice). Because the examples did not cover
all food items in the food group, finding equivalent amounts for each food item required
an additional step. First, we found the relevant macronutrient content of the quantified
examples: carbohydrates for starchy staples, protein for protein-rich foods, lipids for fats,
and calcium for dairy (e.g. grams of carbohydrates in 90 g of rice). Next, we calculated
the mean macronutrient amount across the food examples specified. Finally, we specified
the target amount of each food item as the amount containing the mean macronutrient
amount relevant to each food group. Fruits and vegetable requirements are specified in
terms of grams.
FBDGs require diversity, which is particularly emphasized for some food groups,
generally those required in greater quantities. China’s FBDG specifies that at least 12 food
items should be consumed each day (Wang et al., 2016). We operationalized guidance on
diversity by specifying a number of food items from each food group, as shown in Table 4,
and including equal amounts of each item within a food group.

Finally, to ensure isocaloric comparability across all FBDGs, the healthy diet basket
as a whole was scaled to 2 329 kcal, based on the energy requirements of a 30-year old
active woman, consistent with the energy sufficient and nutrient adequate diets. For FBDGs
that include an allowance for discretionary foods or free sugars, these were included in
the final diet but not scaled, as the language of the FBDG indicates that the amounts of
these foods are a maximum not to be exceeded. For all other items, scaling preserves
the proportions between foods. In some cases, as suggested in India’s FBDG and also the
EAT-Lancet reference diets, variation in total energy needs could be met through variation
in starchy staple consumption without scaling the other food groups, but FBDGs from

14
 3 Results: Global analysis

the United States of America and Oman indicate proportional scaling across food groups,
and other FBDGs are not explicit on how to scale to different calorie needs. Neither method
(only starches, or all groups proportionally) is perfect, but we chose to scale proportionally,
because distorting the proportionality of FBDGs (by scaling starches more than other food
groups) is a larger problem than increasing the nutrient-rich food groups only slightly more
than necessary. Proportionality is a very important aspect of FBDGs, so it is important to
maintain proportionality when (as in most cases) guidelines are not explicit about how to
scale calories otherwise.

TABLE 4 Number of items required in healthy diet baskets, by food group

Food group Number of items Notes

Starchy 2 In USA and Oman, at least one starchy staple
staples must be whole grain.
Legumes/ 2
flesh/eggs
Dairy 1
Vegetables 3 In India and Benin, one vegetable is a dark green
leafy vegetable (DGLV).
In USA, at least one vegetable is a DGLV or
red/orange.
Fruits 2
Oils/fat 1

Discretionary 1 Only included in Argentina, USA and

foods EAT-Lancet diet.

Source: Authors' own elaboration.

Cost of a healthy diet extension: food preferences

A useful variant of the cost of a healthy diet (CoRD) retains the observed dietary patterns of
the population and scales each food group as a whole to meet FBDGs. This healthy diet with
food preferences metric (CoRD-FP) was first applied to the case of Myanmar (Mahrt et al.,
2019). It tracks the cost of meeting each FBDG with the recommended quantity of each food
group, using products purchased by the population of interest in their observed proportions
within each group. That mix of foods has a higher monetary cost than the CoRD, but reflects
other influences on food choice that may include culinary practices, taste and convenience,
or the time required to prepare each meal. The CoRD retains only the two least-cost items in
equal quantity, whereas the CoRD-FP retains items in the observed dataset in the observed
ratio to each other. For food groups such as starchy staples where observed consumption
exceeds dietary guidelines, all observed items are scaled down by the same proportion,
while items in other groups where observed consumption does not meet guidelines, such as
fruits and vegetables, are scaled up.
In the Myanmar case, the CoRD-FP was constructed directly from household survey data,
which shows how the ratio among items within each food group differs among households.
In some applications, it might be appropriate to remove households in upper expenditure
quintiles or non-poor households when estimating food consumption patterns.

15
Cost and affordability of healthy diets across and within countries

Cost of a healthy diet extension: Nutritional Price Index

Following the logic of the CoRD-FP, the Nutritional Price Index (NPI) is designed to monitor
changes in the cost of a healthy diet while capturing food preferences. It is designed to be
an indicator monitored in parallel with the food CPI, which monitors changes in the cost
of a typically purchased food basket, each food item weighted according to expenditure
share. Using the CPI food list, we assign a predetermined weight to each food group aligned
with a selected FBDG to ensure proportionality, and then use the relative CPI weight for
each food item within the group to capture food preferences. Weights of discretionary
foods not recommended in the FBDG (such as sugar-sweetened beverages) become zero,
and weights for all vegetables typically increase, each item in proportion to its expenditure
share. To weight each food group, we assign 50 percent share to fruits (20 percent) and
vegetables (30 percent), and weight starches, proteins, fats and dairy by calorie share of the
remainder. This is the approach to proportionality taken within the EAT-Lancet reference
diet (Willett et al., 2019). By definition, the NPI, like any CPI, tracks only the percentage
change or difference in diet costs. To address questions of affordability, we compare the
CoCA, CoNA and CoRD metrics of diet cost per person each day.

2.4 Affordability indicators

To determine affordability, diet costs need to be compared to a standard of income or
expenditures. In this study, we use two standards for the global analysis:
1. Affordability indicator 1: We compare the cost of the diets to 63 percent of the international
poverty line of USD 1.90 a day, which is equal to USD 1.20. We use 63 percent as
the estimated proportion of expenditures on food, because it is the mean proportion of
expenditures on food among the bottom consumer segment in low-income countries
(calculated from the World Bank Global Consumption Database [World Bank, 2020]).
2. Affordability indicator 2: We compare the cost of the diets to typical daily food expenditures
in each country. The source of national average food expenditures per capita used in
affordability indicator 2 is from the ICP.
3. Affordability indicator 3: We also calculate the percent of people in each country who
would not be able to afford a given diet using the World Bank PovcalNet online tool,
based on the estimated 2018 income distributions across 164 countries.4 All data from
PovcalNet is in 2011 PPP. We adjust 2017 PPP costs to 2011 PPP for PovcalNet calculations,
using American CPI inflation for each year between 2012 and 2017 (FRED, 2020). In the
main results, we calculate this number and percent based on 63 percent of income, which
is the proportion of expenditures on food by the bottom consumer segment in low-income
countries (calculated from the World Bank Global Consumption Database). We apply this
percentage as an estimate of the proportion of expenditures that could possibly be spent
on food by the poor globally. In the main results, we show the proportion of the total
population of each region, and proportion of the total population of the world, who cannot
afford the diets.5 In Annex 3, we provide the average percent who cannot afford the diets
across countries, regionally and globally. In Annex 4, we show a lower-bound estimation
of the number of people who cannot afford the diets, calculated assuming 100 percent
of income spent on food; that is, the lower bound is the percent of people who have total
daily income lower than the cost of the diet. The upper bound counts those for whom

4
2018 income distributions are used because those for 2017 are not available. In one country, India, we apply
the 2015 income distribution, which is the most recent available.
5
Population data are from the World Bank World Development Indicators; food expenditure data are from
national accounts.

16
 3 Results: Global analysis

the cost of diet exceeds the mean proportion of expenditures on food, by national income
level classification: food expenditures account for 15 percent, 28 percent, 42 percent,
and 50 percent of expenditures on average in high-, upper-middle-, lower-middle- and
low-income countries, respectively6 (see Annex 4).
4. Affordability indicator 4: To find the number of people who cannot afford a given diet,
we used affordability indicator 3 multiplied by the 2017 population in each country.
We show the lower-bound estimation in the main section, and the upper-bound estimation
in Annex 4. Population data are from the World Bank World Development Indicators;
food expenditure data are from national accounts.

Additionally, in one of the country case studies (Ethiopia), we compare the cost of diets
to rural wages. This was also done in an analysis in India by Raghunathan, Headey and
Herforth, 2020.
This study uses complete data on cost of energy sufficient, nutrient adequate and healthy
diets available for 170 countries. Affordability indicators 1 and 2 are available for all
170 countries; affordability indicators 3 and 4 are available for 143 countries.
A limitation of the ICP dataset is that it does not include prices on all relevant foods,
many of which might be cheaper than the ones in the dataset (such as certain vegetables that
are typically harvested and consumed locally, and not common across countries).

6 The median food expenditure shares are 14 percent, 25 percent, 41 percent and 51 percent for the four income
levels, which are quite close to the mean.

17
3 Results: Global analysis

KEY MESSAGES

The range of the minimum cost of healthy diets globally is between USD 3.27 and
USD 4.57 per day, with a point estimate of USD 3.75 as the global median among
170 countries.

There are multiple definitions of a healthy diet; in each country we calculated

the cost of ten different definitions of a healthy diet published by UN Member
States, to find a range of the cost of healthy diets, and a median point estimate.
In addition, we calculated the cost of the EAT-Lancet reference diet patterns in
each country.

Healthy diets by any definition are far more expensive than the entire international
poverty line of USD 1.90, let alone the upper-bound portion of the poverty line of
USD 1.20 that can credibly be reserved for food.

The cost of healthy diets is nearly five times as expensive as the cost of energy
sufficient diets.
In all, we estimate that 3 billion people globally cannot afford the least-cost form
of healthy diets. Moreover, 1.5 billion people cannot afford a nutrient adequate
diet. The majority live in Southern Asia and sub-Saharan Africa.

The global average cost of meeting calorie needs using the most affordable starchy staple
at each time and place is USD 0.79, while the average cost of meeting all nutrient needs is
USD 2.33.7 Looking across different regions, the median cost of a nutrient adequate diet is
2–5 times more expensive than that of an energy sufficient diet, and the cost of healthy diets
is 1.5–2 times more expensive than the least-cost nutrient adequate diet, and 3–8 times more
expensive than the least-cost energy sufficient diet (Figures 2 and 3).
For the cost of healthy diets, over the ten different definitions of a healthy diet published
by UN Member States, the range of the cost of these diets globally is between USD 3.27 and
USD 4.57 per day, with a point estimate based on median costs of USD 3.75 (Figure 4).
This compares to a range of between USD 3.31 and USD 3.61 for the least-cost versions of
the EAT-Lancet diet, including four specific diet variants (vegan, vegetarian, pescatarian,
flexitarian).8 FBDGs are generally designed to ensure nutrient adequacy, and our findings
show that the least-cost diets that meet FBDGs do satisfy 94 percent of nutrient needs on
average (Annex 5). The cost of healthy diet metric allows for substitution within each food
group to meet the recommended quantities for a healthy diet. Adding additional requirements
would impose additional costs, for example to meet personal or culturally typical food
preferences and time savings/convenience.

7 Annex 2 shows the cost of a nutrient adequate diet when using RDAs rather than H-ARs.
8 The estimate of USD 2.84 per day for the overall EAT-Lancet reference diet found in Hirvonen et al., 2019 used
a different dataset and some different assumptions, and is not directly comparable.

19
Cost and affordability of healthy diets across and within countries

This finding shows that the cost of healthy diets by any definition far exceeds the entire
international poverty line of USD 1.90, let alone the upper-bound portion of the poverty
line of USD 1.20 that can credibly be reserved for food (Figure 5). The cost of healthy diets
exceeds food expenditures in most countries in the Global South (Figure 6), and 87 percent
of people in low-income countries cannot afford these diets, including over 75 percent
of the population throughout sub-Saharan Africa and Southern Asia (Figure 7, Table 5).
High proportions of people in South-eastern Asia (> 50 percent), Melanesia (> 40 percent)
and Latin America (> 15 percent) also cannot afford these diets.9

FIGURE 2 The cost of energy sufficient, nutrient adequate and healthy

diets by region and country income group

A. BY REGION
6

4
D
S
U
7
1
0
2
2

0
Africa Asia Latin America Northern America Oceania
and the Caribbean and Europe

B. BY COUNTRY INCOME GROUP

6

4
D
S
U
7
1
0
2
2

0
Low-income Lower-middle Upper-middle High-income
countries countries countries countries

Energy sufficient diet Nutrient adequate diet Healthy diet

Source: Authors’ own elaboration.

9
Table A3.2 in Annex 3 reports the percentage of population who could not afford the three reference diets in
each region and country income group, expressed as unweighted averages.

20
 3 Results: Global analysis

FIGURE 3 Premiums required to reach nutrient adequate and healthy diets,

by region

A. RATIO OF THE COST OF A NUTRIENT ADEQUATE DIET TO THE COST OF AN ENERGY

SUFFICIENT DIET
10

6 o
it
a
4 R

B. RATIO OF THE COST OF A HEALTHY DIET TO THE COST OF AN ENERGY SUFFICIENT DIET
12

o
it
6 a
R

C. RATIO OF THE COST OF A HEALTHY DIET TO THE COST OF A NUTRIENT ADEQUATE DIET
3

2
o
it
a
R
1

Notes: Data shown are box plots for each ratio indicated, showing the median, 25th and 75 percentile,
1.5 times that interquartile range, and outlier values for the increase in cost associated with each
increment of diet quality in each region and national income group.
Source: Authors’ own elaboration.

21
Cost and affordability of healthy diets across and within countries

FIGURE 4 Worldwide average costs of ten food-based dietary guidelines

diets and four EAT-Lancet reference diets
5.00

4.50

4.00
)
3.50 D
y
SU(

3.00 soC
rept
nosrep
rep
ad

2.50

2.00

1.50

1.00

0.50

National food-based dietary guideline

EAT-Lancet reference diet
Global median cost across all ten FBDGs

Notes: Data shown are global means (n=170) for the least-cost set of locally available items meeting
the diet quality standard shown. The horizontal line at USD 3.75 is the median of the ten national
food-based dietary guidelines (FBDGs).
Source: Authors’ own elaboration.

In all, we estimate that 3 billion people globally cannot afford the least-cost form of
healthy diets (Table 6). The majority of these reside in Southern Asia (1.3 billion) and
sub-Saharan Africa (829 million), with high numbers also in South-eastern Asia (326 million)
and Eastern Asia (230 million).10 Additionally, 186 million people cannot afford the cheapest
form of energy sufficient diets, mostly concentrated in Africa (149 million). Last, 1.5 billion
people cannot afford the cheapest form of nutrient adequate diets, almost entirely in Asia
(754 million), Africa (680 million), and the Americas (72 million) (Table 6).
These numbers compare to an estimated 812–822 million people worldwide
who are undernourished (using the FAO prevalence of undernourishment indicator),
and approximately 2 billion people who experience moderate or severe food insecurity
(FAO, IFAD, UNICEF, WFP and WHO, 2019).

10 Annex 3 shows upper-bound results of the number of people who cannot afford these diets.

22
 3 Results: Global analysis

FIGURE 5 Affordability of diets relative to the global poverty line of

USD 1.90/day
A. COST OF AN ENERGY SUFFICIENT DIET COMPARED WITH THE INTERNATIONAL POVERTY LINE

B. COST OF A NUTRIENT ADEQUATE DIET COMPARED WITH THE INTERNATIONAL POVERTY LINE

C. COST OF A HEALTHY DIET COMPARED WITH THE INTERNATIONAL POVERTY LINE

No data < USD 1.20 USD 1.20−USD 1.90 > USD 1.90
Notes: Dotted line represents approximately the Line of Control in Jammu and Kashmir agreed
upon by India and Pakistan. The final status of Jammu and Kashmir has not yet been agreed upon
by the parties. Final boundary between the Republic of Sudan and the Republic of South Sudan
has not yet been determined. Final status of the Abyei area is not yet determined. A dispute exists
between the Governments of Argentina and the United Kingdom of Great Britain and Northern
Ireland concerning sovereignty over the Falkland Islands (Malvinas).
Source: Authors’ own elaboration. Conforms to Map No. 4170 Rev. 19 UNITED NATIONS October 2020.

23
Cost and affordability of healthy diets across and within countries

FIGURE 6 Affordability of diets relative to actual food expenditure per capita

A. RATIO OF THE COST OF AN ENERGY SUFFICIENT DIET AND AVERAGE NATIONAL
FOOD EXPENDITURES PER CAPITA

B. RATIO OF THE COST OF A NUTRIENT ADEQUATE DIET AND AVERAGE NATIONAL

FOOD EXPENDITURES PER CAPITA

C. RATIO OF THE COST OF A HEALTHY DIET AND AVERAGE NATIONAL FOOD EXPENDITURES
PER CAPITA

No data < 0.5 0.5–1 1–2 2–4 >4

Notes: Dotted line represents approximately the Line of Control in Jammu and Kashmir agreed
upon by India and Pakistan. The final status of Jammu and Kashmir has not yet been agreed upon
by the parties. Final boundary between the Republic of Sudan and the Republic of South Sudan
has not yet been determined. Final status of the Abyei area is not yet determined. A dispute exists
between the Governments of Argentina and the United Kingdom of Great Britain and Northern
Ireland concerning sovereignty over the Falkland Islands (Malvinas).
Source: Authors’ own elaboration. Conforms to Map No. 4170 Rev. 19 UNITED NATIONS October 2020.

24
 3 Results: Global analysis

FIGURE 7 Proportion of people unable to afford each level of diet quality

in 2017
A. PERCENTAGE OF POPULATION WHO CANNOT AFFORD AN ENERGY SUFFICIENT DIET

B. PERCENTAGE OF POPULATION WHO CANNOT AFFORD A NUTRIENT ADEQUATE DIET

C. PERCENTAGE OF POPULATION WHO CANNOT AFFORD A HEALTHY DIET

No data < 10% 10–25% 25–50% 50–75% 75–100%

Notes: Unaffordability is defined as the cost of the diet exceeding 63 percent of total expenditures,
which is the proportion of food expenditure for the bottom consumer segment in low-income
countries, reserving 37 percent for non-food expenditures. Dotted line represents approximately the
Line of Control in Jammu and Kashmir agreed upon by India and Pakistan. The final status of Jammu
and Kashmir has not yet been agreed upon by the parties. Final boundary between the Republic of
Sudan and the Republic of South Sudan has not yet been determined. Final status of the Abyei area is
not yet determined. A dispute exists between the Governments of Argentina and the United Kingdom
of Great Britain and Northern Ireland concerning sovereignty over the Falkland Islands (Malvinas).
Source: Authors’ own elaboration. Conforms to Map No. 4170 Rev. 19 UNITED NATIONS October 2020.

25
Cost and affordability of healthy diets across and within countries

TABLE 5 Percentage of people who cannot afford each diet, by region and
country income group (population weighted average)

Energy sufficient Nutrient adequate Healthy

diet diet diet
Macroregion
Africa 12.3% 56.4% 80.0%
Asia 0.5% 18.2% 46.7%
Latin America and
1.8% 11.7% 18.2%
the Caribbean
Northern America
and Europe 0.4% 1.0% 1.7%

Oceania 0.2% 0.8% 2.1%

Subregion

Australia and New Zealand 0.2% 0.5% 0.7%

Central Asia 0.3% 7.4% 22.4%
Eastern Asia 0.1% 0.8% 14.7%
Eastern Europe 0.1% 1.0% 2.4%
Latin America and
1.8% 11.7% 18.2%
the Caribbean
Melanesia 0.0% 9.6% 41.3%
Northern Africa 1.3% 37.3% 60.3%
Northern America 0.9% 1.4% 1.6%
Northern Europe 0.1% 0.3% 0.4%
South-eastern Asia 1.0% 23.3% 52.1%
Southern Asia 0.7% 33.4% 76.1%
Southern Europe 0.6% 1.9% 3.6%
Sub-Saharan Africa 14.9% 60.8% 84.6%
Western Asia 0.2% 4.8% 21.3%
Western Europe 0.0% 0.2% 0.2%
Country income group

High-income countries 0.5% 1.1% 1.7%

Low-income countries 8.3% 60.9% 87.0%
Lower-middle countries 3.9% 36.3% 72.8%
Upper-middle countries 0.8% 4.3% 17.0%
World 2.6% 21.6% 43.2%
Notes: Data shown are the percentage of people in each region whose household income is below the total
cost of the most affordable locally available items needed to meet each standard of diet quality, assuming
that they can spend no more than 63 percent of their income on food. Methods and data sources are detailed
in the text.
Source: Authors' own elaboration.

26
 3 Results: Global analysis

TABLE 6 Number of people who cannot afford each diet, by region and
country income group

Energy sufficient Nutrient adequate Healthy

diet diet diet
Macroregion
Africa 148.64 680.59 964.84
Asia 21.57 754.48 1933.93
Latin America and
10.46 66.77 104.24
the Caribbean
Northern America
and Europe 4.76 10.95 17.97

Oceania 0.06 0.21 0.55

Subregion

Australia and New Zealand 0.06 0.12 0.18

Central Asia 0.08 2.45 7.43
Eastern Asia 1.96 13.05 230.43
Eastern Europe 0.31 2.49 5.86
Latin America and
10.46 66.77 104.24
the Caribbean
Melanesia 0.00 0.08 0.36
Northern Africa 2.87 84.27 136.06
Northern America 3.34 5.05 5.96
Northern Europe 0.14 0.29 0.40
South-eastern Asia 6.29 145.40 325.53
Southern Asia 12.92 586.14 1 337.37
Southern Europe 0.95 2.81 5.40
Sub-Saharan Africa 145.76 596.32 828.78
Western Asia 0.32 7.45 33.16
Western Europe 0.03 0.31 0.36
Country income group

High-income countries 5.99 12.12 19.22

Low-income countries 48.31 354.94 506.56
Lower-middle countries 112.24 1 041.46 2 087.45
Upper-middle countries 18.95 104.49 408.30
World 185.49 1 513.01 3 021.53
Notes: Data shown are the number of people in each region whose household income is below the total cost
of the most affordable locally available items needed to meet each standard of diet quality, assuming that
they can spend no more than 63 percent of their income on food. Methods and data sources are detailed in
the text.
Source: Authors' own elaboration.

27
Cost and affordability of healthy diets across and within countries

Micronutrient-rich non-staples (fruits and vegetables, dairy, and protein-rich foods) are
the highest-cost food groups per day globally (Figure 8). There are regional differences,
with fruits being notably more expensive in Asia, and dairy being notably more expensive
in Africa and cheaper in Europe and Oceania. Starchy staples and oils account for only
16 percent of the cost of healthy diets, while fruits and vegetables account for 40 percent
of the cost, with dairy and other protein-rich foods combined accounting for 44 percent
(Figure 9). These proportions vary somewhat by region, with dairy being progressively more
expensive in low-income countries (Figure 10).

FIGURE 8 Cost per person per day by food group, region and country
income group (2017 USD)

A. BY REGION
1.2

1.0

0.8

D
0.6 S
U

0.4

0.2

0
Starchy Protein-rich Dairy Fruits Vegetables Fats
staples foods

Africa Asia Latin America and the Caribbean

Northern America and Europe Oceania World

B. BY COUNTRY INCOME GROUP

1.2

1.0

0.8

D
0.6 S
U

0.4

0.2

0
Starchy Protein-rich Dairy Fruits Vegetables Fats
staples foods

World High-income countries Upper-middle-income countries

Lower-middle-income countries Low-income countries

Source: Authors’ own elaboration.

28
 3 Results: Global analysis

FIGURE 9 Average percentage contribution of each food group to the

average global cost of a healthy diet (2017 USD)

4%
12%

21%
Starchy staples
Protein-rich foods
23% Dairy

Fruits
Vegetables
Fats
19%

21%

Note: Data shown are the average over all countries (n=170).
Source: Authors’ own elaboration.

FIGURE 10 Cost contribution (percentage share of total cost) of each food

group in a healthy diet, by region in 2017
100

80

60 egatnecreP

40

20

0
World Africa Asia Latin America Northern Oceania
and the America and
Caribbean Europe

Starchy staples Protein-rich foods Dairy

Fruits Vegetables Fats

Note: Data shown are the average over all countries in each geographic region for the healthy diet,
defined as the median cost in each country of meeting ten national food-based dietary guidelines.
Source: Authors’ own elaboration.

29
Cost and affordability of healthy diets across and within countries

These findings imply that (a) the cost of nutritious foods needs to go down, particularly
nutrient-rich foods including fruits and vegetables, dairy, and protein-rich foods; and
(b) poverty lines may need to be raised, as they are the basis for programme targets and
social safety net programmes, and currently do not support the ability of humans to access
even the least-cost versions of healthy diets that meet dietary needs. That is, they do not
support food security.
FBDGs are state-published definitions of a healthy diet appropriate for and targeted
toward its citizens, and are used as a basis for nutrition education throughout many countries.
These findings indicate that nutrition education and behaviour change is not sufficient to
shift population consumption toward healthy diets, particularly for the most nutritionally
vulnerable, because those diets are out of reach for the majority of the world’s poor. In order
to enable all people to shift behaviour toward healthy diets, prices of those diets, particularly
of the most nutrient-rich food groups, need to decrease.

30
4 Results: Country case studies

KEY MESSAGES

National price datasets allow subnational analysis of the cost of diets. These food
price datasets are collected by national statistical organizations or agriculture
market information systems.

The cost of healthy diets varies by region, as seen in the United Republic of
Tanzania, and is driven by different high-cost food groups in different regions.
The cost of healthy diets and nutrient adequate diets varies by season, as seen in
Malawi. Food price data can also be used to examine low availability of specific
food groups in specific times and places, as seen in Ghana. The Ethiopia case
study shows that the cost of the most nutrient-rich non-staple food groups has
gone up more rapidly than starchy staples.

Poverty lines are typically insufficient to support the purchase of least-cost

healthy diets. The standard cost of basic needs (CBN) approach to determining
food poverty lines selects a food basket that meets neither nutrient needs nor
dietary recommendations for healthy diets, as demonstrated in Myanmar. Instead,
we demonstrate a feasible method for generating food poverty lines that would
satisfy dietary guidelines, as well as taking into account typical consumption
patterns of the poor.

Each country case study is intended to showcase a different aspect of how national and
subnational food price data shed light on the cost of the nutrient adequate diet and the
healthy diet. Each country’s unique circumstances reveal important aspects of how food
prices relate to nutrition, showing how the cost of these diets can vary by region, by season,
by life stage, and when accounting for food preferences.

4.1 United Republic of Tanzania

The Tanzanian case study focused on regional variation in the cost of food groups. To calculate
the cost of a healthy diet, we used the Benin FBDG in this case, because it is the only country
in Africa currently with quantifiable FBDG. The United Republic of Tanzania, Ethiopia and
others are currently developing FBDGs. Those may be quantifiable and could then be used to
compare with the costs shown here. The main result of our Tanzanian study is the striking
regional variation in the cost of healthy diets around the country. The average cost of a
healthy diet between 2011 and 2015 in the southeast coast was USD 2.83 in Lindi, USD 2.77
in Mtwara, USD 2.59 in Pwani which includes the largest city of Dar es Salaam (USD 2.75),
and USD 2.54 in the east-coast region of Kilimanjaro. The average cost of a healthy diet in
these regions was about 35 percent higher than the average cost in Mbeya and Iringa in
the southwest, bordering Zambia and Malawi, and above the national average of USD 2.33
(Figure 11).

31
Cost and affordability of healthy diets across and within countries

In the high-cost regions, vegetable prices are very high, contributing about a third of
the total diet cost. In the Mara region, however, the biggest driver of cost is oils, which are
3–5 times more expensive than in most regions in the United Republic of Tanzania. In the
low-cost regions, starchy staples are predominant, and nutrient-rich non-staples are much
more affordable.

FIGURE 11 Average cost of a healthy diet by region in the United Republic of

Tanzania, 2011–2015

Mara

Kagera
Mwanza

Shinyanga Kilimanjaro

Kigoma
Arusha

Tabora Tanga
Singida
Zanzibar
Dodoma
Dar es
Rukwa Salaam
Morogoro

Iringa
Pwani
Mbeya

2011 USD
1.8 Lindi

2.1
2.4
2.7 Ruvuma Mtwara
3.0
Missing

Note: Data shown are the average cost of a healthy diet over all months in each region.
Source: Authors’ own elaboration. Conforms to Map No. 3667 Rev. 6 UNITED NATIONS January 2006.

32
 Looking deeper into subnational variation in the cost of food groups, Figure 12 shows
variation in the cost of each food group. We see the smallest magnitude of subnational
variation in fruits and protein-rich foods, and the most in vegetables and oils. Drivers such
as remoteness and market access, transportation costs, and perishability may be affecting
the cost patterns observed. Overall, vegetables are the most expensive in Lindi, Mtwara and
Pwani regions. In these regions, the average cost of vegetables in a healthy diet is USD 0.76,
which is 72 percent higher than the national average of USD 0.44.

FIGURE 12 Food group shares in the cost of a healthy diet by region in the
United Republic of Tanzania, 2011–2015

A. HEALTHY DIET COST CONTRIBUTION B. HEALTHY DIET COST CONTRIBUTION

FROM STARCHY STAPLES FROM PROTEIN-RICH FOODS

Mara Mara

Kagera Kagera
Mwanza Mwanza

Kilimanjaro Shinyanga Kilimanjaro

Shinyanga

Kigoma Kigoma
Arusha Arusha

Tabora Tanga Tabora Tanga

Singida Singida
Zanzibar Zanzibar
Dodoma Dodoma
Dar es Dar es
Rukwa Salaam Rukwa Salaam
Morogoro Morogoro

Iringa Iringa
Pwani Pwani
Mbeya Mbeya

2011 USD 2011 USD

0.1 Lindi 0.1 Lindi
0.2 0.2
0.3 0.3
0.4 0.4
0.5 Ruvuma Mtwara 0.5 Ruvuma Mtwara
> 0.6 > 0.6
Missing Missing

C. HEALTHY DIET COST CONTRIBUTION D. HEALTHY DIET COST CONTRIBUTION

FROM DAIRY FROM FRUITS

Mara Mara

Kagera Kagera
Mwanza Mwanza

Kilimanjaro Shinyanga Kilimanjaro

Shinyanga

Kigoma Kigoma
Arusha Arusha

Tabora Tabora
Tanga Tanga
Singida Singida
Zanzibar Zanzibar
Dodoma Dodoma
Dar es Dar es
Rukwa Salaam Rukwa Salaam
Morogoro Morogoro

Iringa Iringa
Pwani Pwani
Mbeya Mbeya

2011 USD 2011 USD

0.1 Lindi 0.1 Lindi
0.2 0.2
0.3 0.3
0.4 0.4
0.5 Ruvuma Mtwara 0.5 Ruvuma Mtwara
> 0.6 > 0.6
Missing Missing

►►

33
Cost and affordability of healthy diets across and within countries

FIGURE 12 Food group shares in the cost of a healthy diet by region in the
United Republic of Tanzania, 2011–2015 (cont.)

E. HEALTHY DIET COST CONTRIBUTION F. HEALTHY DIET COST CONTRIBUTION

FROM VEGETABLES FROM OILS

Mara Mara

Kagera Kagera
Mwanza Mwanza
Kilimanjaro
Kilimanjaro Shinyanga
Shinyanga

Kigoma Kigoma
Arusha Arusha

Tabora Tabora
Tanga Tanga
Singida Singida
Zanzibar Zanzibar
Dodoma Dodoma
Dar es Dar es
Rukwa Salaam Rukwa Salaam
Morogoro Morogoro

Iringa
Iringa
Pwani Pwani
Mbeya Mbeya

2011 USD 2011 USD

0.1 Lindi 0.1 Lindi
0.2 0.2
0.3 0.3
0.4 0.4
0.5 Ruvuma Mtwara 0.5 Ruvuma Mtwara
> 0.6 > 0.6
Missing Missing

Note: Data shown are the average share of each food group shown in that region’s cost of a healthy
diet over all months in each region.
Source: Authors’ own elaboration. Conforms to Map No. 3667 Rev. 6 UNITED NATIONS January 2006.

4.2 Malawi
The Malawi case study focused on variation in the cost of a nutrient adequate diet in two
ways: over seasons, and over life-cycle stages.
The average monthly cost of healthy diets and nutrient adequate diets show a greater
seasonal gap compared to the average monthly cost of energy sufficient diets in Malawi
(Figure 13). The seasonal gap refers to the highest-cost month minus the lowest-cost month.
The seasonal gap for the national cost of a healthy diet is about USD 0.15 (a 7 percent rise
from the lowest- to the highest-cost month), and for the cost of a nutrient adequate diet is
about USD 0.13 (an 11 percent rise), compared to about USD 0.06 (a 15 percent rise) for the
national cost of an energy sufficient diet. However, the seasonality of these indicators can
vary significantly across different regions. For example, Dowa, to the north of the capital city
of Lilongwe, is the district with the highest seasonality in cost of a healthy diet and cost of
a nutrient adequate diet (Bai, Naumova and Masters, 2020). In Dowa, the seasonal gap for
the cost of a healthy diet is USD 0.84 (a 32 percent rise), for the cost of a nutrient adequate
diet it is more than USD 0.50 (a 38 percent rise), and for the cost of an energy sufficient diet
about USD 0.20 (a 45 percent rise).
In Figure 13, panels B and C, we illustrate the average monthly cost of a nutrient
adequate diet and components by different food groups. Fruits and vegetables display a
greater seasonal variation compared to other food groups. The trend of monthly change for
fruits and vegetables is also in line with the trend of the national costs of a healthy diet and
a nutrient adequate diet. The synchronized price rises for nutrient-dense foods is the key
driver of the seasonal variations of the cost of a nutrient adequate diet in several Eastern
African countries (Bai, Naumova and Masters, 2020).

34
 4 Results: Country case studies

FIGURE 13 National and regional average diet costs per month by component
in Malawi

A. AVERAGE COST OF HEALTHY, NUTRIENT ADEQUATE AND ENERGY SUFFICIENT DIETS,

BY MONTH
3 Dowa cost of energy
sufficient diet

Dowa cost of
nutrient adequate diet
2
D Dowa cost of
S
U healthy diet
1
1
0 National cost of
2
1 energy sufficient diet

National cost of
nutrient adequate diet

0 National cost of
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec healthy diet

B. AVERAGE COST OF NUTRIENT ADEQUATE DIET FOOD GROUP COMPONENTS, BY MONTH

0.4 Animal source foods

Fruits and vegetables

0.3 Oils and fats
D Pulses, nuts and seeds
S
U
0.2 1
1 Starchy staples
0
2
Sweets
0.1

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

C. AVERAGE COST OF HEALTHY DIET FOOD GROUP COMPONENTS, BY MONTH

0.6 Dairy

Fruits

Oils and fats

0.4
D Protein-rich foods
S
U
1 Starchy staples
1
0
2
0.2 Vegetables

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Notes: Data shown are for a representative woman aged 19–30 years, at 67 kg and 163 cm, using
nutrient requirements specified by the IOM Dietary Reference Intakes (DRIs) – specifically, the
estimated average requirement (EAR), Tolerable Upper Intake Level (UL), Acceptable Macronutrient
Distribution Range (AMDR) and Chronic Disease Risk Reduction Intake (CDRR). All costs were
converted to 2011 USD at annual average purchasing power parity (PPP) exchange rates, smoothed
over months by the Denton method.
Source: Authors’ own elaboration.

35
Cost and affordability of healthy diets across and within countries

The cost of a nutrient adequate diet also varies across sex-age groups over the life cycle.
In Malawi, as shown in Figure 14 (panel A), pregnant and lactating women and adolescent
boys face the highest cost of a nutrient adequate diet. The average cost of a nutrient adequate
diet of these groups is more than USD 1.5 per day, which is much higher than 70 percent of
international poverty line and the food expenditure per capita per day in Malawi.
In terms of the cost per 1 000 kcal, females in general face a higher cost compared to
males, due to their relatively lower required energy intake and therefore their need for more
nutrient-dense foods (Figure 14, panel B). This trend is the same at a global level, showing
adolescent girls and older females facing particular challenges in terms of the need for highly
nutrient-dense diets (Bai, Herforth and Masters, forthcoming).

FIGURE 14 Average cost of a nutrient adequate diet by demographic group

in Malawi

A. COST OF NUTRIENT ADEQUATE DIET BY DEMOGRAPHIC GROUP

2.0

USD 1.29
1.5 (2011 food expenditure
per capita)
D
S
U
1.0 1
1
0
2
0.5

0
4−8 9−13 14−18 19−30 31−50 51−70 > 70
years years years years years years years

Female Pregnancy Lactation Male

B. COST OF NUTRIENT ADEQUATE DIET PER 1 000 KCAL BY DEMOGRAPHIC GROUP
0.8

0.6

D
S
U
0.4 1
1
0
2

0.2

0
4−8 9−13 14−18 19−30 31−50 51−70 > 70
years years years years years years years

Female Pregnancy Lactation Male

Note: Data shown are for a representative person of each age, sex and reproductive status, using
nutrient requirements specified by the IOM Dietary Reference Intakes (DRIs): Tolerable Upper
Intake Level (UL), Acceptable Macronutrient Distribution Range (AMDR) and Chronic Disease Risk
Reduction Intake (CDRR).
Source: Authors’ own elaboration.

36
 4 Results: Country case studies

4.3 Ethiopia
In recent years, Ethiopia has had rapid economic growth at about 10 percent per year,
although starting from a low base level. There have been significant road infrastructure
investments in addition to an African Green Revolution focus on fertilizer and seeds for
cereal grains. Yet Ethiopia has persistent high rural poverty, is dependent on food aid and
safety nets, and is vulnerable to food price shocks (see for example Bachewe and Headey,
2017). These and other health challenges are reflected in the high rates of child stunting
(38 percent), wasting (11 percent), micronutrient deficiencies, and underweight among
women (25 percent in rural areas).
This case study used food price and wage data to display the overall cost of all foods
needed for nutrient adequacy, as well as the subsistence cost of an energy sufficient diet,
and their affordability relative to daily wages for unskilled workers.
The study shows that the cost of a nutrient adequate diet has increased over time by
an annual rate of 3 percent, but not as much as wage increases at an annual rate of about
5 percent. Therefore, the affordability of nutrient adequate diets has improved due to wage
increases rather than improvements in the cost of food. The cost of a nutrient adequate
diet as a percentage of wages decreased from 32 percent in 2008 to 22 percent in 2016
(Figure 15).

FIGURE 15 Cost of a nutrient adequate diet and wages in Ethiopia, 2002–2016

8
50

6 40
stnio

D p 30
S
U e
4 1 atnece
rg
P
1
0
2 20

2
10

0 0
2002 2004 2006 2008 2010 2012 2014 2016

Cost of nutrient adequate diet Cost of nutrient adequate diet as % of wage Wage

Note: Data shown are the average national level for cost of nutrient adequate diet and wages on the
left axis, and cost of nutrient adequate diet as a percentage of daily wages on the right axis.
Source: Bachewe et al., 2019.

Price changes in different food groups reveal that nutrient-rich foods are the main
contributors to the rise in costs of a nutrient adequate diet and healthy diet over time.
The average nominal prices in LCU of all food groups increased between 2002 and 2016 in
Ethiopia. Prices of pulses and fruits and vegetables increased by about 9.4 and 7.9 times,
while the price of starchy staples showed the lowest growth of 5.6 times over the same
period (Figure 16). Looking at the real cost of different food groups considering inflation,
starchy staples have become cheaper while nutrient-rich food groups have become more
expensive over time (Bachewe and Headey, 2019).

37
Cost and affordability of healthy diets across and within countries

FIGURE 16 Price indexes by food group in Ethiopia, 2002–2016

1 000

800

s
600 exedni

e
c
400 irP

200

0
2002 2004 2006 2008 2010 2012 2014 2016

Animal source foods Fruits and vegetables Oils and fats

Pulses, nuts and seeds Starchy staples Sweets

Note: Data shown are the simple average of prices (local currency units [LCU]/kg) of all items in each
food group, indexed to January 2002=100.
Source: Authors’ elaboration and Bachewe et al., 2019.

While wage increases are positive for real affordability of diets, the increasing cost of
healthy diets and nutrient-rich food groups is a concern, because those food groups have the
highest elasticities of demand, so consumers are less likely to turn wage gains into nutrient
rich food purchases. Policy attention is needed not just on the traditional Green Revolution
crops, but on horticultural crops, legumes, and animal source foods to reduce their prices.

Researchers at the FAO Regional Office for Asia and the Pacific have also found that the
cost of fruits and vegetables has gone up over time more than the cost of other food groups
in South-eastern Asia (Dawe and Lee, 2017).
Effects might differ by gender, but gender-specific wage data were unavailable. Here,
we used nutrient needs for a reference female, and unskilled wage data not disaggregated
by gender. In a similar analysis of the affordability of healthy diets in India, there were
gender-specific dietary recommendations and gender-specific wage data. The diet was
more unaffordable for women than for men, because women’s wages were systematically
lower; therefore the cost of a healthy diet required a higher percentage (80–90 percent) of
women’s wages, compared to 50–60 percent of men’s wages (Raghunathan, Headey and
Herforth, 2020).

4.4 Ghana
The cost of healthy diet metric was developed in consultation with food price data collectors
within the governments of Ghana and the United Republic of Tanzania.11 During this
consultation led by the IANDA (Indicators of Affordability of Nutritious Diets in Africa) project
at Tufts, it became clear that one important source of food price information, the Ministry of

11 See workshop reports accessible at https://ianda.nutrition.tufts.edu/media

38
 4 Results: Country case studies

Food and Agriculture (MoFA) monitoring and information system (MIS), had data on many
foods but insufficient food diversity to be able to calculate the cost of a diet pattern that
included all recommended food groups. As a result of this discovery, MoFA requested input
on foods to include in its MIS. Input on commonly consumed nutritious foods was provided
by several agencies, nutrition specialists, and scholars in Ghana, in a process coordinated
by IANDA. MoFA piloted this expanded list in early 2017, and then rolled out the updated
food price monitoring list nationwide throughout 2017. The intent of this expanded list was
to provide information about potential investments in foods where market opportunities
may exist in certain areas and seasons, and to enable the tracking of the cost of healthy diet
metric and its component food groups.
The cost of a healthy diet in Ghana is shown in Figure 17, computed from the new MoFA
data using the expanded food list. These results show that the cost of a healthy diet rose in
late 2017 into 2018, and remained relatively stable throughout most of 2018–2019, possibly
with seasonal rises in May–June of both years.

FIGURE 17 Monthly cost of a healthy diet in Ghana, 2017–2019

4.5

yad
s
rep

4.0 n
idec

laianahG

3.5
m
ani

N
o

3.0
t t v c n b rMrpA y n l t t v c n b ra rpA y n l t t
p c o e a e a u g p c o e a u g p c v c
e
eS O N D J F
a u
M J
J u
A e O N D J
e
F M
a u
M J
J u
A e O o
N D
S S
2017 2018 2019

Note: Data shown use national average prices to compute the cost of a healthy diet, in nominal
Ghanaian cedis per day.
Source: Authors’ elaboration and the Ministry of Food and Agriculture (Ghana).

One of the main policy-relevant reasons for examining the cost of a healthy diet is to see
where and when healthy diets may be out of reach, and the specific foods that are missing
or expensive. Figure 18 reveals that the cost of a healthy diet is highest in the Northern and
Western regions of Ghana; in the Northern region, vegetables are much more expensive than
in other regions, while in the Western, Brong Ahafo and Upper East regions, dairy (including
small fish and crustaceans) is more expensive. There are many time points where the cost of
a healthy diet could not be computed, however, because no prices were observed for one or
more food groups. The food groups most often missing were dark green leafy vegetables, dairy,
oils and fruits (Figure 19). The regions where these food groups were most often unobserved
were the Ashanti, Northern, and Upper West regions. It is possible that the specific markets
captured in the MIS do not have oils or any dairy/small fish for sale, and that those items
may be purchased from other markets. The absence of dark green leafy vegetables and fruits,
however, is a concern because these would be present in the types of markets visited, if they

39
Cost and affordability of healthy diets across and within countries

were available. If these items are absent from markets at certain time points, then it may not
be feasible to purchase a complete recommended diet from the market at any price. The use
of the cost of healthy diet metric for monitoring unavailability is an additional application to
be explored in partnership with price data collectors and end users.

FIGURE 18 Cost of a healthy diet by food group and region in Ghana, 2018

5 yad
r
4 s
ep
n
idec
3
aianah
G
2

0
l it g l r a tU tU
a n norfo a
rt
n
r etaer r n
r saE se ta
l
n
r
n a a e c e o e
ito h h n ts c h W V ts
s B A e a A tr r pp
re e
a A C E G o p
e
N N W

Starchy staples Protein-rich foods Dairy

Fruits Vegetables Fats

Note: Data shown use national average prices to compute the cost of a healthy diet, in nominal
Ghanaian cedis per day.
Source: Authors’ elaboration and the Ministry of Food and Agriculture (Ghana).

FIGURE 19 Frequency of missing price observations for each food group

in Ghana
50

40

egatnecreP
30

20

10

0
Starchy Protein-rich Dairy Fruits Vegetables Dark Green Fats
staples foods Leafy Vegetables

Note: Data shown are the percent of districts with zero price observations for any item in each
food group.
Source: Authors’ elaboration and the Ministry of Food and Agriculture (Ghana).

40
 4 Results: Country case studies

4.5 Myanmar
The Myanmar country study drew upon the 2015 MPLCS household survey to gain insight
into the cost of diets in Myanmar. In particular, this study expanded upon the cost of healthy
diet metric’s estimation of the bare minimum cost of achieving dietary recommendations
with a variant that takes food preferences into account.
This case study also compared both cost of healthy diet metrics to Myanmar’s official
food poverty line (MoPF and World Bank, 2017a, 2107b). The country’s food poverty
line is derived from a basket of foods determined by food consumption quantities of poor
households in the 2015 MPLCS. Unlike the nutritious foods outlined in recommended diets
that form the basis of the costs of a healthy diet and of a healthy diet with food preferences,
the Myanmar food poverty line basket includes all foods and non-alcoholic beverages in the
2015 MPLCS household food survey. In order to satisfy the most basic nutrition requirement,
the basket is scaled to meet energy sufficiency based on the caloric needs of the Myanmar
population (2 238 per capita). The food poverty line is simply the cost of acquiring the energy
sufficient food basket at median household prices. To achieve consistency between the two,
both healthy diet food baskets are also scaled to meet the 2 238 calorie target.12 (For a more
detailed methodology, see Annex 6.)
The healthy diet with food preferences method is analogous to the determination of food
poverty lines in that within-food group costs are based on consumption quantity shares
observed in the poorest households (within 10 percentage points of the poverty line) in the
household survey. We take advantage of this similarity to show how nutritional needs can be
taken into account in determining poverty lines and poverty estimation. Unlike current food
poverty line estimation methods, the healthy diet with food preferences method additionally
requires that minimum food group-level proportional criteria are met so that diets used in
the estimation align with FBDG.

Myanmar does not have quantitated FBDG, and therefore guidelines from neighbouring
Bangladesh provide the recommended diet used in this analysis. The Bangladesh FBDG is
generally consistent with the food groupings and messages presented in a preliminary FBDG
developed by the Myanmar Government (Shaheen et al., 2013; MoHS, no date).
Figure 20 presents total and food group costs of the three food baskets: food poverty line,
healthy diet, and healthy diet with food preferences. Differences in costs between the three
are driven by two factors: basket composition and the costs of items within each food group.
Basket composition is the primary factor driving differences in food group costs between the
food poverty line and healthy diet with food preferences baskets, as food group costs in both
baskets are based on the same actual consumption patterns. Differences between the costs
of the cheapest foods and the preferred foods in each food group drive the differences in
costs between the healthy diet and healthy diet with food preferences baskets. For the latter,
a preference for higher-priced animal source foods is the main driver of its higher cost.
Figure 21 highlights the considerable difference in the caloric content between the food
poverty line and healthy diet with food preferences baskets. Not surprisingly, nearly three
quarters (72 percent) of calories in the former are derived from starchy staples compared to
just over half (53 percent) in the latter, which allocates a greater share of the diet to more
nutrient-dense foods. Figure 22 illustrates cost shares of the two baskets. The cost share of
starchy staples in the food poverty line basket is more than double that of the healthy diet
with food preferences basket (30 percent versus 13 percent). Notable in both Figures 21
and 22 is the absence of dairy from actual consumption which stands in sharp contrast

12 This calorie target is 91 kcal lower than the calorie target for an active 30-year-old woman used in the global
analyses of this report.

41
Cost and affordability of healthy diets across and within countries

to its 17 percent cost share in the healthy diet with food preferences basket. These large
differences in the composition of both baskets are evident in the total costs of nearly every
food group (Figure 20).

FIGURE 20 Cost of each food group for three food baskets in Myanmar
3.5

3.0

2.5
P
P
2.0 P
1
1
0
2
1.5 D
S
U
1.0

0.5

0
Starchy Pulses Fish, meat Dairy Fruits Vegetables Fats Other Total
staples and eggs and oils foods basket

Food poverty line Healthy diet Healthy diet with food preferences

Notes: Data shown are 2011 USD purchasing power parity (PPP) per day and refer to the items in the
food baskets used to compute the country’s food poverty line, healthy diet and healthy diet with food
preferences baskets. All diet costs exclude food away from home, and are scaled to 2 238 kcal/day.
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

FIGURE 21 Calorie shares by food group in two Myanmar food baskets:

food poverty line and healthy diet with food preferences
FOOD HEALTHY DIET
POVERTY LINE WITH FOOD PREFERENCES
3%
Starchy
staples
1% 16% Pulses
13%
2%
4% Fish, meat
6% and eggs
6%
54% Dairy
3%
72% 6% Vegetables

9% Fruits

5% Fats and oils

Other foods

Note: The food poverty line basket excludes food away from home and is scaled to meet the 2 238
calorie target.
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

42
 4 Results: Country case studies

FIGURE 22 Cost shares by food group in two Myanmar food baskets:

food poverty line and healthy diet with food preferences
FOOD HEALTHY DIET
POVERTY LINE WITH FOOD PREFERENCES

Starchy
6% staples
12% 13% 4%
10% Pulses
7% 30% Fish, meat
4% and eggs
20% Dairy
13% 30%
Vegetables
3% Fruits
31%
17%
Fats and oils
Other foods

Note: The food poverty line basket excludes food away from home and is scaled to meet the 2 238
calorie target.
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

Figure 23 highlights the other important factor driving differences in the cost of diets –
the price premium of current consumption patterns compared to least-cost food selection.
The cost of starchy staples, pulses, and oils varies little among the most commonly consumed
food items. In contrast, the costs of animal source, protein-rich foods, and to a lesser extent
vegetables, have a much wider distribution. This variation in costs is reflected in the large
differences in the cost of fish/meat/eggs and vegetables between the healthy diet and healthy
diet with food preferences baskets.

FIGURE 23 Costs of items in each food group in Myanmar

2.5

2.0 yad
P
rep
1.5
D
PP

1.0 1
SU

102
0.5

0
Starchy Pulses Fish, meat Dairy Fruits Vegetables Fats
staples and eggs and oils

Note: Data shown are 2011 USD purchasing power parity (PPP) per day, as box plots for the cost of
each food group in Myanmar. Box plots show the median, 25th and 75 percentile and 1.5 times that
interquartile range for acquiring the recommended diet quantity in each food group. Food items are
limited to those in the healthy diet with food preferences basket, using the median price of each food
item observed in local markets. 2015 kyat converted to 2011 USD PPP using the World Bank's PPP
conversion factor for private consumption (local currency units per international USD).
Source: Authors’ own elaboration.

43
Cost and affordability of healthy diets across and within countries

Table 7 provides greater detail on the costs of protein-rich foods in the healthy diet and
healthy diet with food preferences baskets. The least-cost, animal source protein-rich foods
are fish and eggs. Though the cost share of fish, legumes, and meat/eggs is quite similar
between the two baskets, the addition of meat, particularly chicken, results in a diet cost that
is nearly double that of a healthy diet (USD 1.38 PPP versus USD 0.78 PPP).

TABLE 7 Cost of protein-rich foods in two Myanmar food baskets:

healthy diet and healthy diet with food preferences

Healthy diet basket Healthy diet with food

preferences basket
Protein group
food item or Cost per day Protein group Cost per day Protein group
subgroup (2011 USD PPP) cost share (%) (2011 USD PPP) cost share (%)

Fish/seafood 0.20 37 0.28 35

Legumes 0.06 13 0.08 10
Eggs 0.27 50 0.29 12
Chicken 0.80 22
Pork 0.55 14
Beef 0.77 7
Notes: Data shown are cost levels in 2015 kyat converted to 2011 USD purchasing power parity (PPP) using
the World Bank's PPP conversion factor for private consumption (local currency units per international USD).
Source: Authors' own elaboration.

The three food baskets exceed reported household food expenditure for a large share of
the population (Figure 24). Nationally, 40 percent and 41 percent of the population live in
households that cannot afford the food poverty line or healthy diet basket without shifting
non-food expenditure to food expenditure. Sixty-five percent cannot afford a healthy diet
that aligns with food preferences. Furthermore, the majority of the population in the three
lowest wealth quintiles cannot easily afford the healthy diet with food preferences basket.
In order to explore the impact of a more nutritionally complete diet that is also consistent
with food group consumption patterns of poor households, we use the healthy diet with food
preferences basket in place of the food poverty line basket to estimate a nutrition-sensitive
poverty line. To facilitate comparison with Myanmar’s official poverty rates, we calculate
this poverty line as the sum of the healthy diet with food preferences poverty line and the
official non-food poverty allowance. The non-food poverty allowance is an estimate of
non-food expenditures by households with total consumption expenditures near the poverty
line. Table 8 shows the composition of the official and healthy diet with food preferences
poverty lines. Figure 25 shows the percentage of the population living in households with
total expenditure below each poverty line. The healthy diet with food preferences poverty
line indicates the share of the population that faces insecurity in attaining a healthy diet, and
suggests that the share of the population facing such nutrition insecurity is 17 percentage
points higher than the poverty rate, nationally, and 46 points higher (compared to zero) for
the third wealth quintile.
Finally, we compare the nutrient composition of the baskets to the estimated average
requirement (EAR) (IOM, 2006) of a 19–30 year old woman (Table 9). The costs of the healthy
diet and healthy diet with food preferences baskets meet or exceed the EAR of most key
nutrients, with the exception of vitamin E. In contrast, the food poverty line basket meets

44
 4 Results: Country case studies

approximately half of the EAR for calcium (53 percent), vitamin A (55 percent) and vitamin E
(56 percent), and less than 80 percent of the EAR for vitamin C (74 percent), folate (78 percent)
and vitamin B12 (71 percent). These serious micronutrient shortfalls call into question whether
a calorie standard provides a sufficient nutrient standard for poverty line calculations.

FIGURE 24 Affordability of food poverty line, healthy diet and healthy diet
with food preferences baskets: % of the population living in
households with food expenditure below the cost of each basket
in Myanmar
100

80

P
P
60 P
1
1
0
2
40 D
S
U

20

0
National Urban Rural Q1 Q2 Q3 Q4 Q5

Food poverty line Healthy diet Healthy diet with food preferences
Note: 2015 kyat converted to 2011 USD purchasing power parity (PPP) using the World Bank's PPP
conversion factor for private consumption (local currency units per international USD). Q1 to Q5
denote wealth quintiles.
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

FIGURE 25 Official poverty line headcounts compared to healthy diet with

food preferences poverty line headcounts (%), by national,
urban/rural and expenditure quintile in Myanmar
100

80

P
P
60 P
1
1
0
2
40 D
S
U

20

0
National Urban Rural Q1 Q2 Q3

Official poverty line Healthy diet with food preferences poverty line

Note: Q1 to Q3 denote wealth quintiles.

Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

45
Cost and affordability of healthy diets across and within countries

TABLE 8 Official and healthy diet with food preferences poverty lines
in Myanmar

Official Healthy diet with food preferences

poverty lines poverty lines
Food poverty line 2.34 3.27
Non-food poverty line 1.26 1.26
Total poverty line 3.60 4.53
Notes: For comparability, the official non-food poverty line is combined with the healthy diet with food
preferences poverty line to estimate the total healthy diet with food preferences poverty line. 2015 kyat
converted to 2011 USD purchasing power parity (PPP) using the World Bank's PPP conversion factor for
private consumption (local currency units per international USD).
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

TABLE 9 Nutrient adequacy of least-cost food baskets in Myanmar

Percentage of EAR
Food Healthy diet with
Healthy diet
Nutrient EAR poverty line food preferences

Protein (g) 38 144 163 177

Carbohydrates (g)
Calcium (mg) 750 53 112 119
Iron (mg) 11 105 123 137
Phosphorus (mg)
Magnesium (mg) 265 91 126 129
580 139 176 185
Zinc (mg) 9 89 99 105
Copper (mg) 1 280 209 384
Selenium (mcg) 45 137 171 153
Vitamin C (mg) 80 74 153 190
Thiamin (mg) 1 87 103 122
Riboflavin (mg) 1 51 122 100
Folate B6 (mg)
(mg)
Niacin (mcg)
Vitamin 11 112 97 128
1 106 135 135
250 78 169 149
Vitamin B12 (mcg) 2 71 118 154
Vitamin A (mcg) 490 55 107 116
Vitamin E (mg) 12 56 89 84
Notes: Data shown are adequacy levels by nutrient of the food baskets shown for a representative woman
aged 19–30, with a 2 238 calorie diet. Light shading indicates above 90 percent, medium 80–89 percent,
and dark below 80 percent.
Source: Authors’ elaboration; Ministry of Planning and Finance (Myanmar) and World Bank, 2017a.

46
5 Global simulation of
policy impacts on the cost
of a nutrient adequate diet
KEY MESSAGES

Public investment and government services can lower food prices and diet costs
through innovation and market infrastructure, but governments also often cause
prices to rise when they impose trade restrictions. Restricting imports protects a
country’s producers of a particular item, at the expense of others in that society.
Trade restrictions arise most often where the burden of higher prices is spread
among many dispersed consumers, while the protected group is well-organized
and influential.
Analysing price data by food group reveals that protectionism raises price the
most for poultry and eggs, and vegetables. These items have economies of scale
and are grown on larger farms, near cities, whose owners have more influence
than other farmers, while consumers of these items are dispersed and may not
know that prices are raised by trade restrictions. Protection of influential groups
also raises prices for other foods, especially in middle- and high-income countries.

The overall rise in cost of the most affordable diets was in the range of USD 35–
USD 70 per year (10–20 cents/day) in middle- and high-income countries in
2011, with the same level in high-income countries in 2017, and wide variation
elsewhere. This increase in diet costs could be a significant factor in diet quality,
especially for low-income people in middle-income countries.

Easier access to some imported commodities could lower consumer prices, as could
reducing transport costs within countries. Using Monitoring and Analysing Food
and Agricultural Policies (MAFAP) data we find impacts on total diet cost in Africa
of possible domestic transport cost reductions in the range of USD 2–USD 11
per year. Greater price reductions might be possible through a combination of
interventions that raise farm productivity and lower marketing costs in various
ways, tailored to local needs for each type of food.
5.1

Impacts of trade policy and farm-to-market transport costs on the

cost of nutrient adequate diets
A wide range of policies can lower the cost and improve affordability of nutritious foods,
including government support for research and development of new technology, public
irrigation infrastructure, land development, rural transport, electrification and markets,
as well as the institutions and regulations needed to maintain product standards and quality
assurance. These policies offer highly cost-effective ways to help both farmers and food

47
Cost and affordability of healthy diets across and within countries

consumers by raising productivity, typically lowering diet cost and improving access to
nutrient adequate and healthy diets through delivery of public services tailored to each place
and time (Norton, Alwang and Masters, 2014).
In this section, we focus on trade restrictions at each country’s borders, as a type of
agricultural policy that allows governments to help specific groups without need for public
expenditure or service delivery. We describe the available data on how these restrictions
change prices in low- and middle-income countries, and the effect of those price changes
on diet costs. For comparison, we also show how food prices and diet costs would change
if governments aimed to reduce the cost of transporting foods by investing in improved
infrastructure and institutional reforms. Governments impose trade restrictions, and invest
less in transport-cost reductions than consumers might want, so as to protect the influential
local producers who benefit from higher prices without having to incur any fiscal cost. Some
trade barriers actually raise government revenue through tariffs and the sale of quotas or
licenses. The price changes that result from trade restrictions typically deliver gains that are
concentrated among few members of well-organized groups that advocate for restriction,
while their costs are spread among all consumers in the country so that each consumer
bears a small share of the burden and is unlikely to be aware of that cost. Which sectors
receive this protection depends on their relative influence in government, as detailed in
Anderson (2016).
Our data on trade policies consist of percentage nominal rates of protection (NRPs),
obtained by comparing a country’s observed farm-gate commodity prices with the best
available estimate of what that price would be if government policies did not restrict
international trade. These estimates are available for 58 food commodities in 62 countries,
based on prices observed in farm surveys and rural wholesale markets, as well as commodity
prices observed for international trade adjusted for estimates of inland transport and handling
costs, reported as farm-gate equivalent NRPs. To smooth fluctuations and match with our
retail prices, we use all values observed in the four years leading up to and including each
round of ICP price collection, then aggregate them into food groups and national income
levels as shown in Figure 26 and Figure 27.

FIGURE 26 Worldwide average nominal rates of protection, 2008–2011 and

2014–2017
70
60
lega n
an oit
50 imon cetorp
revA
40
30 s
fo

20 etar

10
0 ,R t
yrtl sel sggE s s
sto dnasrniatn ae rs y s tis dnastu sd
e ria in
uo bat ls o t
e
n a u
r
e
e
u m e D r F sli
P ege P d te G
ebalp e o
R e N
V u w
S
2008–2011 2014–2017

Notes: Data shown are worldwide average percentage effects of trade policy on the wholesale price
of commodities from each food group computed from a total of 3 581 observations over 58 food
products in 62 countries. Sources are detailed in the methodological Annex 7.
Source: Authors’ own elaboration.

48
 Figure 26 and Figure 27 are sorted by food group, in decreasing order of support for
farmers producing the traded commodities in that food group. Poultry producers are the
most highly protected worldwide in both time periods shown, and vegetable farmers are
highly protected in high-income countries. Both groups, however, saw a decline in average
protection from the 2008–2011 period to the 2014–2017 period. Most but not all categories
are more protected in higher-income countries in both time periods. Pulses, often a
significant contributor to least-cost diets, are less protected in the later period for all income
groups except upper-middle-income countries. The average NRP is sometimes negative,
which could be sustained over time only when the government restricts exports to help the
country’s buyers of that product. While countries in each income level observe some export
restrictions on specific commodities, only for low- and lower-middle-income countries is
this common enough to result in a negative average protection value at the food group level.

FIGURE 27 Average nominal rates of protection by food group and country

income group

A. 2008–2011
120
100
l n
80 a oi
n
i tc
60 me
o to
n r
40 e p
g fo
20 a
r s
e te
0 v
A ar
-20
-40 ,s t
y s s s d s a sr y s st dnastuN sd
rt e
l g e
sl t n n
i e e ri n
i i
l b g o a a n a a u e
u a E u o sr t m e D r r e
sli
o t P R n d t G F
P e e a
l e ee o
g b
e p R w
V tu S

B. 2014–2017
100
80
l n
60 a oi
n
i tc
40 me
o to
n r
20 e p
g fo
0 a
r s
e e
-20 v t
A ar
-40
-60 ,stoo t
yrtlu s s s R dnas sniatn a sr y s st d s
elb ggE e
sl e e ria n
i i n d
n a u a e
ate u m e D r r st e
sl
o P rt ap d t G F i
P ebl e e u o
g e N
e u R w
V S
Low-income countries Lower-middle-income countries
Upper-middle-income countries High-income countries
Notes: Data shown are average percentage effects of trade policy on the wholesale price of
commodities from each food group in countries at each level of national income, computed from
a total of 2 154 observations over 58 food products in 62 countries in 2008–2011 and from a total
of 1 427 observations over 56 food products in 53 countries in 2014–2017. Sources are detailed in
Annex 7.
Source: Authors’ own elaboration.

49
Cost and affordability of healthy diets across and within countries

For all negative NRP observations, we check to ensure that the country is actively
exporting that product or could do so competitively. When this is not the case, we classify
these observations as measurement errors or temporary outliers and trim the value to zero.
The criterion we apply is whether the country’s FAO food balance sheet reports net exports
of that product in any of the four years leading up to and including the year of observation.
For these exportable foods, negative NRPs could be sustained by any policy that limits the
quantity exported, whereas negative NRP observations in other settings could be sustained
only through government subsidy payments for the entire quantity consumed in that country,
which is implausible even in countries with large government outlays. We assume these
values are caused by differences in quality, context and timing of comparisons between
items whose prices are compared in the NRP.
To compute the consequences of agricultural protection for the cost and affordability of a
nutrient adequate diet, first we estimate the impact of these commodity NRPs on the purchase
price of retail items, taking account of variation in the wholesale product’s share of each
item’s retail price. Our upper bound on price effects represents a scenario where farm-gate
commodity prices account for one-half of retail prices paid, and our lower bound represents
a scenario where that fraction is one-fourth. For each scenario we identify the quantities
of items needed to meet nutrient requirements at the lowest total cost per day, and show
the added expense imposed by the country’s agricultural trade restrictions (see Annex 7 for
detailed methodology).
Figure 28 and Figure 29 present the range of upper and lower bounds across countries
in 2008–2011 and 2014–2017, showing that agricultural trade policies raise the cost of
a nutrient adequate diet anywhere from 0 to 30 cents per day at the median, or roughly
USD 0–USD 108 per year per person, globally. In both time periods, cost increases are
highest in upper-middle-income and high-income countries. These groups see roughly the
same median increase in 2008–2011. In 2014–2017, upper-middle-income countries see
about 1.5 times the increase that high-income countries do, although with greater variation
in effect. While lower-middle-income countries see a USD 28–USD 57 increase per annum
due to trade policy in 2008–2011, this effect is much less in 2014–2017, at USD 7–USD
17. In low-income countries, the median change is approximately zero in 2008–2011, but
in 2014–2017 this group sees potential cost savings of USD 15–USD 32 per person per
annum, were trade restrictions to be lessened. For these countries, the food group for which
agricultural protection most raises price is grains, followed by pulses, staple root vegetables,
and other vegetables. In all of these categories, trade restrictions to protect farm producers
lead to retail price rises that could place nutrient adequate diets out of reach for many
consumers, especially the poorest.
In general, the largest plausible impact on consumer prices is twice the smallest plausible
impact in percentage change terms. The actual range of impacts on consumer costs also
depends on substitution among items. When an item’s price increase (or decrease) is large
enough, its quantity in the most affordable diet decreases (or increases) and is replaced by
other items that can meet nutrient needs at a lower cost. For that reason, the actual impact
of a given policy change on consumer prices at the upper bound is not always twice the
lower bound, and a paradoxical reversal can occur when substitution shifts consumption
from highly affected to less affected foods. In the cases of Benin, Sierra Leone, Ethiopia and
Zimbabwe in 2008–2011, the change in diet cost at the upper bound is slightly smaller in
absolute value than the change in diet cost at the lower bound of price transmission, because
the upper bound triggers a large change away from items that are highly affected by trade
policy towards items that are less affected.

50
 5 Global simulation of policy impacts on the cost of a nutrient adequate diet

FIGURE 28 Change in cost of a nutrient adequate diet due to agricultural

trade policy, 2008–2011
0.6

)
D
0.4 S
U
1
1
0
2
(
ts
0.2 o
c
n
ie
sa
e
rc
0
n
I

-0.2
Low-income Lower-middle Upper-middle High-income
countries countries countries countries

Upper bound Lower bound

Notes: Data shown are median, 25th and 75th percentile range, and whiskers of 1.5 times that
range for the effect of trade policy on daily retail cost of a nutrient adequate diet for an adult
woman, in 2011 USD, under two scenarios: the upper bound applies to settings where wholesale
costs affected by trade policy account for a large fraction of retail price, such as generic items
sold in open markets; while the lower bound applies to settings where trade policy and wholesale
costs are a smaller fraction of retail prices, such as supermarkets (as explained in the text and the
methodological appendix). The following outliers are not shown: Japan (upper: 0.53) and Republic
of Korea (upper: 0.74) in the high-income group.
Source: Authors’ own elaboration.

FIGURE 29 Change in cost of a nutrient adequate diet due to agricultural

trade policy, 2014–2017
0.6

)
D
S
U
0.4 7
2
10
t(s
n
oc
i
e
0.2 saercnI

0
Low-income Lower-middle Upper-middle High-income
countries countries countries countries

Upper bound Lower bound

Notes: Same as for Figure 28. Here, the following outliers are not shown: Bhutan (-0.2, -0.01),
Mauritania (-0.01, 0), El Salvador (-0.15, -.08) and Sudan (-0.2, -0.12) in the lower-middle-income
group; and Guinea (-0.07, -0.03), Comoros (-0.1, -0.08) and Chad (lower: -0.02) in the low-income group.
Source: Authors’ own elaboration.

51
Cost and affordability of healthy diets across and within countries

5.2 Impacts of transport cost reduction on the cost of

nutrient adequate diets
This section considers the effect of within-country transport costs on the cost of nutrient
adequate diets in 14 sub-Saharan African countries for which MAFAP collects agricultural
commodity transport data. Estimates of potential decreases in farm-gate to wholesale
transportation costs use as a reference point the country infrastructure of South Africa,
the most efficient country in the region. Using the ratio of the infrastructure dimension of the
World Bank’s Logistics Performance Index in each country to that of South Africa, the MAFAP
team calculated an adjusted transport cost for 21 commodity value chains (FAO, 2020a).
For this study, we compute an estimated transport cost reduction (TCRij) for each of the
21 commodities indexed i, in each of the 14 countries indexed j, in each year of observation
t. Other models would be needed to take account of how a country’s food system might
adjust to changes in its farm-to-market transport costs; here we isolate only the transport
cost change, holding all else constant, and apply the TCR for each product in each country to
identify plausible changes in retail prices if farm-gate and traded product prices remained
unchanged. For a price shock that can be compared directly to MAFAP’s tariff-equivalent
nominal rate of protection from trade policy, we express TCRijt as a fraction of the observed
farm-gate price (Pijt), focusing on the cost difference between farm-to-market transport costs
with reference-country infrastructure (TiRt) minus the observed farm-to-market transport
costs (Tijt):

TiRt –Tijt
TCRijt = (1)
Pijt

To apply farm-gate commodity-level shocks to retail prices, we average the shocks for
ten food groups over the four years leading up to and including 2011 and 2017 – the ICP
price collection years. Food groups vary in the number of countries and years for which
transport cost estimates are available. Cereal grains have the most frequent transport cost
data, with 128 observations across all 14 countries in each four-year period. Observations of
nutrient-dense food groups such as dairy, fruits, poultry and eggs, red meat, and vegetables
are relatively sparse, and often comprised of only one or two commodities from one or
two countries.
As shown in Table 10, average shocks range from -0.68 percent for dairy in 2014–2017
to 6.75 percent for sweeteners in 2008–2011. In almost all cases, the shock in 2011 is
greater, which would be consistent with countries in this region improving transportation
infrastructure over time, so they are closer to cost levels in the reference country. Those
groups for which the shock is based on only one product in one country are indicated by
an asterisk.
We assume that the entirety of the transport cost reduction is passed on to the commodity
price. Following the methodology of our trade policy simulation, we apply farm-gate-level
food group price shocks to one quarter and one half of the retail prices, representing lower
and upper bounds of impact (for details, see Annex 7). One key difference between the
2011 and 2017 simulations is the diversity of foods represented in the ICP retail data for
the 14 countries studied here. In 2011, there were 259 foods in these countries compared
to only 161 in 2017. This change in diversity may limit the sensitivity of the least-cost diet
metric to individual country contexts.

52
 TABLE 10 Transport cost changes used for simulation, by food group in
2008–2011 and 2014–2017
Food group 2008–2011 2014–2017
Dairy -1.81%* -0.68%*
Fruits -5.98%* -5.99%*
Grains -5.16% -3.12%
Nuts and oilseeds -2.42% -1.38%
Poultry and eggs -3.32%* -2.96%*
Pulses -2.23% -1.94%
Red meat -3.09% -2.49%
Roots, tubers and plantains -6.32% -3.03%
Sweeteners -6.75% -4.89%
Vegetables -1.74%* -2.06%*
Note: Data shown are averages over all observations in each food group and time period.
Source: Authors' own elaboration.

Figure 30 and Figure 31 show the upper and lower bounds of annual cost savings for a
least-cost nutrient adequate diet due to transport cost adjustment. In 2011, annual savings
range from USD 2.80 to USD 11.06 per person, measured in 2011 USD at purchasing power
parity (PPP). For Burkina Faso, Mozambique and Benin, the upper-bound effect is more than
USD 10 per person per year, while Uganda and Malawi see relatively small savings of less
than USD 6 at the upper bound. In 2011, we also see two alterations in the composition of
the least-cost diet. The high transport shock adjustments for fruits and for roots, tubers and
plantains make oranges a cost-efficient source of micronutrients in Burkina Faso and shift
the least-cost diet in Burundi to include sweet potatoes rich in vitamin A.
In 2017, effects are more modest, ranging from USD 2.24 to USD 11.15 per person per
annum, measured in 2017 USD at PPP. Potential savings amount to USD 7 per capita per
year, on average, across the countries analyzed. Assuming an average household size of five
members, these savings could amount on average to USD 35 per household, and up to USD
50 per household on an annual basis. Burkina Faso sees markedly higher savings compared
to other countries, with savings per household amounting to USD 55 per year, while at the
lower end of the effect range, Rwanda, Burundi, Malawi and Senegal all see annual cost
savings of less than USD 6 per person at the upper bound. There are no changes in the
composition of the least-cost diet in 2017. The lessened effect is consistent with the lower
average shocks in each food group in 2017 and suggests that while infrastructure continues
to improve in this region, there are still non-negligible costs of transportation inefficiency
that may affect poorer consumers who struggle to access nutritious foods.

53
Cost and affordability of healthy diets across and within countries

FIGURE 30 Change in cost of the nutrient adequate diet from lower

transport costs, 2008–2011
12

nosrep
10
D
raey/
8 re )
ast
sg
niv
p
soC
1
SU
6 102(

2 i l i il
cilbup a e
adnagU
w
a
ipoihE a
n i
n
agene d
n
a
iregi a
d a
a
yne n
i uqibmazoM so
a
l a n n a
h a
z u a M e F
a r K B
M t G eR n u N w a
a S B R n
i
T k
de fo r
tin
U u
B

Notes: Data shown are upper and lower bounds of estimated savings in annual cost of a nutrient
adequate diet due to increased efficiency in transportation infrastructure, in 2011 USD. Upper
bounds apply to settings where farm-gate to wholesale transportation costs account for a larger
fraction of retail price, such as generic items sold in open markets, while lower bounds apply to
settings where farm-gate to wholesale transportation costs are a smaller fraction of retail prices,
such as supermarkets.
Source: Authors’ own elaboration.

FIGURE 31 Change in cost of the nutrient adequate diet from lower

transport costs, 2014–2017
12
nosrep
10
D
raey/
8 r
so
niva
ep
st 7
sg SU
)
6
102(

4
C

2
a i i lageneS a a ci a ai n il e a a o
dna d
n w dnag yne lbup inazn
U poih
i
n a u n iregi s
a
l e M qibm a a
u
r a h F
w u K B G a
M U e a
de t N n
R B RT E azo i
f k
r
tin o M u
B

Notes: Data shown are upper and lower bounds of estimated savings in annual cost of a nutrient
adequate diet due to increased efficiency in transportation infrastructure, in 2017 USD. Upper
bounds apply to settings where farm-gate to wholesale transportation costs account for a larger
fraction of retail price, such as generic items sold in open markets, while lower bounds apply to
settings where farm-gate to wholesale transportation costs are a smaller fraction of retail prices,
such as supermarkets.
Source: Authors’ own elaboration.

54
6 Applications of new
food price metrics

KEY MESSAGES

The metrics we use in this study are designed for use in current food price
monitoring systems, to enable better use of food prices for understanding access
to healthy diets. The metrics can be calculated and used by anyone (i.e. in
government, international agencies, or research) with food price data covering a
sufficient diversity of foods that constitute a healthy diet.
The international poverty line of USD 1.90 per day is insufficient to maintain a
healthy diet. This basic observation at the international level is echoed at the
national level, as shown in the Myanmar case study. It calls for a re-evaluation of
how food poverty lines are determined, and how they could be constructed in a
way that accounts for nutritional needs.

6.1 Implications for food price monitoring

Food price data collection infrastructure

Measurement of cost and affordability in this study is made possible by the standardized
collection of food prices by national governments and international agencies. Most of the
price data we have used here, both through the ICP and national datasets, is collected by
central statistical agencies for CPI calculations, using protocols developed for the United
Nations System of National Accounts. Other price data are collected by agricultural agencies
to track the prices of specific commodities, in market information systems (MIS) used to
understand commercial opportunities. MIS data are collected at both retail and wholesale
markets, targeting a variable and often limited number of high-volume commodities; they are
not standardized across countries or uses.

How food prices are currently used

♦ CPI data is collected frequently (bimonthly at least) at retail markets to help countries
track inflation, determine poverty lines, and measure economic activity over time.

♦ MIS data are collected frequently and primarily used to understand commercial
opportunities. They sometimes contain data on diverse foods.
♦ Prices of staple foods are also used for vulnerability assessment and mapping, to monitor
food price spikes and warn against potential calorie inadequacy.

♦ FAO or other agencies track global food prices, but these monitor only internationally
traded commodities, which omits information about the many diverse foods on retail
markets needed for a healthy diet.

55
Cost and affordability of healthy diets across and within countries

♦ The ICP collects data from national governments on a standardized list of items that
can be compared across countries, once every three to seven years, for the purpose of
producing purchasing power parities (PPPs) and comparable price level indexes (PLIs).

Opportunities for using existing data better

Currently, food prices from all these sources are not tracked in a way that describes the price
of adequate or healthy diets. However, the underlying data can be used for that purpose:

♦ Consumer price index (CPI) data are collected by almost all UN Member States. They aim
to collect prices for a sufficient range of foods to represent national average food
expenditures, typically including between 40 and 200 (and sometimes more) distinct
foods and beverages which is usually enough diversity to measure the cost of nutrient
adequate and healthy diets. To reliably capture the most affordable options, CPI data
should include a range of fruits, vegetables, legumes, fish, eggs, dairy products and other
nutrient-rich items that are locally available in markets used by low-income households.
Prices for each food are typically collected monthly at multiple locations, then averaged
to obtain annual or regional totals. The underlying prices for individual items are often
treated as confidential but may be available to specialized researchers.
♦ Market information system (MIS) data are available for a limited range of countries and
time periods, but in some countries can be diverse, frequent, and sufficiently high quality
to be useful for calculating the cost of healthy diets. It is also sometimes possible to
update MIS to include a wider range of more nutrient-dense items.
♦ The food price data contained in the ICP datasets have been underutilized for
understanding the cost of diets; this analysis as well as Hirvonen et al. (2019) and
Bai et al. (2020) reveal new insights across countries.
The metrics we use in this study are designed for use by a wide range of actors at diverse
scales, from national governments and international agencies to programme implementers
and academic researchers. Diet costs can be computed for people at any time and place where
market prices are available for a sufficient diversity of foods. To reflect the cost of a healthy diet,
a convenient rule of thumb would be to have prices for at least 60 items, including legumes,
nuts and seeds, dark green leafy vegetables, other vegetables, deep orange vitamin A-rich
fruits, other fruits, meat, fish, dairy, eggs and poultry as well as the most commonly consumed
starchy staples. To reflect the most affordable healthy diets, price data should be collected for
the lower-cost items in each food group at marketplaces that serve low-income people.

Governments and international agencies can use their food price data better to
understand access to nutritious food by adopting metrics such as the cost of the healthy diet
and NPI, as the Government of Ghana (Ghana Statistical Service) has announced they intend
to do. Information on the cost and affordability of healthy diets, and identification of their
most costly components, can then inform policies and interventions to improve food access.
Interventions vary widely, and may include improving markets, on-farm own production,
cash transfers, and other context-specific solutions. Projects/interventions can also use food
price data in specific locations for programme design, monitoring and evaluation.

6.2 Implications for poverty estimation

It is clear from this analysis that the international poverty line of USD 1.90 per day is
insufficient to allow access to a healthy diet. This basic observation at the international
level is echoed at the national level, as shown in the Myanmar case study. It calls for a
re-evaluation of how food poverty lines are determined, and how they could be constructed
in a way that accounts for nutritional needs.

56
 6 Applications of new food price metrics

Existing food poverty lines typically use what is called a “cost of basic needs” (CBN)
approach. Our analysis suggests the potential to use healthy diets instead, following the
healthy diet with food preferences method that rescales poverty line food baskets to attain
levels of consumption needed for a healthy diet. This is important because for many actors in
governments (specifically in national statistical organizations), one of the main purposes for
household food consumption, price and quantity data is to determine poverty lines. The CBN
approach for poverty line calculation uses food consumption data to determine the cost of a
typical diet consumed by poor households scaled to meet basic energy needs, which may be
quite different from a nutritional standard.13
CBN poverty lines are based on the actual consumption patterns of poor or nearly poor
households. This approach depends on the concept that poor households are best able to
determine how to allocate their resources and as such is an approach that is decidedly not
paternalistic (Ravallion, 2016).14 In other words, the CBN approach hinges on preferences
of poor households. More specifically, the food poverty line is commonly constructed from a
basket of foods, including associated average quantities, consumed by poor or nearly poor
households. This food basket is scaled to meet energy requirements and then evaluated at
median prices, which yields the food poverty line. The total poverty line is the sum of the
food poverty line and an allowance to meet essential non-food needs.

Attaining sufficient energy intake is the most fundamental purpose of food consumption,
and thus the CBN poverty line provides a measure of severe deprivation. Food poverty lines
reflect the dietary preferences of households striving to meet their basic food needs given
limited resources. However, relatively poor households are likely to consume disproportionate
quantities of low-cost per calorie staple foods such as rice (see for example, Headey and
Alderman, 2019). The over-representation of staples in the food basket relative to a healthy
diet is apparent in the Myanmar case study (see Figure 21). As a result, the Myanmar poverty
line food basket fails to meet the nutritional standards of key micronutrients (see Table 9).
Consequently, poverty lines tied only to energy requirements underestimate the cost required
to access a nutrient adequate diet. Food policy in low- and middle-income countries is shifting
from meeting energy needs with a focus on staple crop production to diverse diets and
food systems. As policy increasingly prioritizes meeting an array of nutrient requirements
essential for good health, a food poverty line designed to satisfy nutrient requirements would
provide a useful additional policy tool.
The healthy diet with food preferences method has a number of features that make it
a convenient choice for a food poverty line that meets nutritional needs. First, as with the
poverty line food basket, the cost is calculated using consumption patterns of a reference
population and is in alignment with actual dietary norms. This method is analogous to
the determination of food poverty lines in that the costs within food groups are based on
consumption quantity shares observed in household surveys among poor households. But
unlike poverty line food basket estimation, the healthy diet with food preferences method
additionally requires that minimum food group-level proportional criteria be met in order
to adhere to the FBDG. Thus, the resulting poverty line is consistent with consumption
patterns among poor households within food groups while realigning consumption between
food groups to meet nutritional standards. This realignment of consumption patterns
between food groups is an important conceptual deviation from the CBN approach which,
as noted, is designed to allow for food poverty lines to meet nutritional standards without
being prescriptive.

13 See Ravallion (2016) for a comprehensive discussion of CBN poverty line estimation.
14 In some cases, such as Myanmar, food baskets exclude alcohol (MoPF and World Bank, 2017b).

57
Cost and affordability of healthy diets across and within countries

Second, the healthy diet with food preferences method is based on FBDG which are
national policy and education tools and thus provide a relevant and policy-coherent basis for
setting food poverty lines that are consistent with cultural norms as well as existing policy.
Developing a methodology that results in a higher poverty rate can be politically contentious.
Aligning the methodology with existing and accepted government guidelines has a clear
advantage. Finally, the healthy diet with food preferences approach is straightforward and
requires no more computational ability or tools to implement than the existing CBN poverty
line methodology.
One challenge with this new method relates to the imposed reallocation of food
consumption between food groups in accordance with FBDG, which introduces a hypothetical
scenario without observable household food preferences. A reallocation from the relatively
cheap staple food group to nutrient-rich yet more expensive food groups would likely be
accompanied by a reallocation toward cheaper items within those food groups. For example,
a poor household striving to meet FBDG might shift some consumption within an animal
source protein-rich food group away from relatively expensive meats toward eggs or cheaper
varieties of fish. As a result, the cost of attaining a healthy diet given food preferences is
likely overestimated.

However, poverty lines tied only to energy requirements clearly underestimate the cost of
attaining a diet that meets nutritional needs. Combined with the existing non-food allowance
of the CBN poverty line, a healthy diet with food preferences total poverty line can be
constructed and compared to total household expenditure to calculate the nutrition-sensitive
poverty headcount. Used in conjunction with the traditional CBN poverty line, this poverty
line provides a complementary poverty measure that considers nutritional needs. In a sense,
both poverty lines taken together provide bounds rooted in household consumption patterns.
The latter results in a nutrition-sensitive poverty line higher than the CBN poverty line,
with a greater proportion of expenditure required for food. This suggests a higher share of
the population faces insecurity in attaining a healthy diet than suggested by the poverty rate.

Others have attempted to construct nutrition-based poverty lines based on nutrient

needs. Allen (2017) introduces a linear programming approach to estimating international
poverty lines based on three least-cost diets that satisfy three levels of nutrient requirements.
Allen argues that a food poverty line should be based on the cheapest foods, as a poverty
line “represents the cost of meeting basic needs, not a level of satisfaction, and should be
set accordingly” (Allen, 2017, p. 3708). This is a considerable departure from the CBN
approach which strives to estimate the cost of a bundle of foods that poor households would
be willing to consume. Ravallion (2016) argues that attaining adequate nutrition is not the
only purpose of food consumption and that ignoring established food culture risks setting
a food poverty line at an expenditure level unrelated to what poor households actually
spend. Importantly, the linear programming approach ignores the proportionality needed in
healthy diets, which is reflected in FBDGs that are government policy documents intended
for all citizens. Furthermore, the least-cost nutrient-based approach results in diets that are
even less proportional than the current CBN method. Nutrient-only standards do not present
an equitable and dignified approach to meeting nutritional needs.
In the Myanmar case study, we have demonstrated that food baskets used to construct
standard CBN poverty lines fall short in meeting both nutrient and food group proportionality
standards. We argue that it is time for nutritional needs to be considered as basic needs.

58
7 Policy options for improving
affordability of healthy diets

KEY MESSAGES

Public funding for agricultural research and development has made grains and
starchy staples relatively more abundant and cheaper relative to nutrient-rich
items such as vegetables, fruits, and pulses, which have received much less
public support for research and development. That imbalance is a key factor
explaining the relatively high cost of nutrient adequate and healthy diets,
because productivity growth driven by public services that support private-sector
innovation has been the primary driver of cost reduction over time. Research and
investment have not sufficiently reduced barriers to productivity and profitability
of non-staple foods.

Our results show that trade restrictions are likely to be harmful in that they raise
the cost of foods needed for nutrient adequate and healthy diets.

The cost of nutritious food should be lowered by diversification of public investment

and increased market access, not by externalizing true costs by subsidizing
inefficient or environmentally harmful production methods. Agricultural and
trade policies should align with diet quality goals, as well as environmental,
social and economic sustainability.

There is currently enough food produced for all people to meet and exceed their dietary
energy needs, but even if this food were to be equally distributed it would be impossible for
all people to meet dietary recommendations. For example, most countries lack adequate
supplies of fruits and vegetables for all people to meet the WHO recommendations of 400 g
or more of fruits and vegetables per day (Siegel et al., 2014). Other micronutrient-rich
food groups, such as beans, nuts, and animal source foods are also far less available and
affordable than starchy staples (Herforth, 2015).
We see this reality play out on a global scale in the results of this analysis, showing that
a staggering number of people cannot afford even the lowest-cost form of healthy diets.
The main cause is that nutrient-dense foods are the most expensive components of the diet
and make up a large share of the cost required. The purpose of the least-cost diet calculation
is to establish a bottom floor: i.e. the lowest possible cost that someone would need to spend
to achieve the dietary guidelines – whereas we know (and demonstrate in the Myanmar case
study) that adding in food preferences only increases the cost, and thus the number of people
who cannot afford the diet.
The inability of people to afford healthy diets results in food insecurity and poor diets.
Poor diet quality has major impacts on malnutrition on all its forms. In relation to overweight,
obesity and diet-related non-communicable disease, the consumption of ultraprocessed foods
of minimal nutritional value, such as sodas, instant noodles, and packaged sweet and salty

59
Cost and affordability of healthy diets across and within countries

snacks, has been rising in many countries experiencing nutrition transitions. The rise in
consumption of these foods is based partly on the abundant supply and low prices of starchy
staples, sugars and oils that constitute their main ingredients and make them relatively
inexpensive, in addition to marketing which promotes these products as aspirational foods
compared to traditional, minimally processed foods. The result is that basic starches make
up too much of the diets of many people, whether in the form of monotonous, nutritionally
inadequate diets or diets that promote obesity and non-communicable diseases.
Increasing access to healthy and sustainable diets will ultimately come from an interplay
between supply-side and demand-side actions. Markets are a key arena where opportunities
exist to affect both supply and demand of nutrition foods. A systems approach is needed to
support increased supplies of vegetables, fruits, pulses, and sustainably produced animal
source foods, including improvements in diversified and sustainable production, storage and
transformation, and marketing.

Supply-side factors
The food environment – i.e. the kinds of food that are most available, affordable and convenient
– is a major determinant of diets (Herforth and Ahmed, 2015). While consumer demand
certainly influences what is produced, multiple supply-side factors do as well: for example,
subsidies, standards at collection/trade points, quality of seed supply, pest resistance,
transport and storage, and perishability. Supply-side policies and measures should align
with diet quality goals, as well as environmental, social and economic sustainability.

Staple grains have been the focus of public investment in agriculture throughout human
history, and particularly during the Green Revolution when more calories were urgently
needed to avert famine. In almost all countries today, food and nutrition problems no longer
call for just more calories, but for greater diversity in the food supply. Demand is increasing
for diversified diets, but the supply response has been surprisingly low, partly due to risk
for farmers and the other private-sector actors who supply each type of food (Pingali, 2015).
Low productivity and high risk leave nutritious non-staple foods, such as fruits, vegetables
and pulses, to remain out of reach for vast numbers of people.
Technological research and subsidies are needed to make more nutrient-rich, non-staple
crops accessible. Growth in agricultural productivity is important and needed, but its traditional
focus on agricultural commodities and export-led growth is not enough to improve nutrition
or sustainability. Efforts and investments must be made in diversifying production with a
focus on nutrient-dense foods such as fruits, vegetables, pulses, fish, dairy and eggs or other
animal source foods. Furthermore, production methods need to be environmentally, socially
and economically sustainable. It is critical to understand the ecological impact of production
in choosing policies that support both human and environmental health. Integrated crop
livestock systems, including agroforestry, can help to ensure more sustainable use of natural
resources (soil, water, biological diversity) as well as a supply of fuel for cooking and fodder
for animals. The most sustainable solutions, considering human and environmental health
as well as economics, will be context specific.

Nutrient-dense foods tend to be more perishable; therefore, improving post-harvest

handling is a key area for action to improve their availability and affordability. Post-harvest
measures can also improve the safety, quality and nutritional value of food. Development of
pro-poor technologies and inputs can facilitate the transport, storage and safe preservation
of more diverse foods, thereby reducing risk and loss which are barriers to increased
production. Examples include drying of fruits, and cold chains for vegetables and dairy.
Drying or freezing are processes that also protect the nutritional value of foods, while other
processes such as food fortification and fermentation help improve their nutritional value.

60
 7 Policy options for improving affordability of healthy diets

Demand-side factors
Markets are where consumers interact with the food supply. They are a key arena for actions
both in terms of promoting production (ensuring that production of nutrient-rich foods is
economically viable for farmers and traders) and consumption (promotion of high-quality,
nutritious, safe foods for consumers).
Nutrition-focused marketing can be a powerful tool to promote knowledge of and
demand for nutritious foods. How retail outlets are organized physically and how foods are
presented in the market can also significantly influence consumer choices. Geographical and
sustainability indications on labels are another way that producers can leverage demand for
diverse products and expand markets for traditional local foods that are sustainably produced.
There can be risks from food marketing as well, in particular with reference to children.
The Set of recommendations on the marketing of foods and non-alcoholic beverages to
children (WHO, 2010) aim to guide efforts by Member States in designing and/or strengthening
policies to reduce the impact on children of marketing of highly processed foods that are
high in saturated and trans fats, sugars and salt.

Nutrition education can be helpful in terms of selecting least-cost foods in order to improve
access to healthy diets; often least-cost diets require substantial labour and knowledge to
prepare. However, it is important to recognize that nutrition education will not result in the
purchase of healthy diets among the poor until prices of those diets go down.

Removing distortions and other inefficiencies of food systems to increase

affordability of healthy diets
Almost all people purchase some portion of their food, and the majority of smallholders are
net buyers of food (that is, they buy more than they sell). Therefore, it is necessary to focus
on what markets provide and the environments in which consumers live, in order to reach
all people and reduce malnutrition in all its forms. A key area for action is linking producers
to markets and improving rural infrastructure. Linking producers to markets is sometimes
needed to ensure that economic opportunities are inclusive; for example, linking smallholder
producers to larger markets that they have not been able to access on their own. Well
functioning market linkages can also serve to reduce food losses, for example of perishable
fruits and vegetables, thereby increasing the overall supply of nutritious food. Finally, there
is a need for trade policies that have co-benefits for nutrition, sustainability and equitable
economic growth. Our results show that trade distortions are likely to be harmful in that they
tend to raise costs of healthy diets.

61
8 Conclusions

Main three general takeaways from the analysis

1. Healthy diets are unaffordable for many people. The high cost of nutritious foods in
places where low-income people live is a major obstacle to the achievement of global
development goals.
2. Unaffordability of healthy diets is concentrated in Africa and Southern Asia. While these
are known to be hot spots for malnutrition, insufficient attention has been paid to diet
quality as a cause of malnutrition in all its forms. As the double burden intensifies in
these regions, diet quality is becoming even more paramount.
3. Supporting nutrient adequate and healthy diets requires a combination of higher
incomes and lower prices, particularly of diverse nutritious items, making a variety
of healthier foods more widely available at lower cost. Where countries have national
food-based dietary guidelines, there is often a lack of policy coherence on how to ensure
the affordability of those diets recommended for nutrition and health. We found no
definition of a healthy diet that would be globally affordable; all definitions result in
similar conclusions.

Main three policy takeaways

1. Tracking the overall cost of a healthy diet highlights the disparity between dietary needs
and what food systems actually produce. Food price data can be better used for nutrition,
both globally and within countries:
a. To understand where and when intervention is most needed, including which
requirements are most costly and which foods can most effectively reduce the overall
cost of healthy diets;
b. To shift attention towards consumer prices and availability of all items required for
healthy diets, in both rural and urban areas;
c. To construct poverty lines that account for the cost of food in terms of meeting dietary
needs, beyond only energy sufficiency.
2. Nutrition education and behaviour change will not substantially improve dietary
consumption where nutrient adequate and healthy diets, even in their cheapest form,
are unaffordable for the majority of the poor. A combination of social protection and food
systems policies are needed to reduce prices and improve access to and consumption of
healthy diets.
3. To make healthy diets cheaper, agricultural policies, research, and development need to
shift toward a diversity of nutritious foods. Prices should not be reduced by discounting
or externalizing real costs (such as reducing workers’ wages, clearing forests, or
intensive animal production that discounts animal welfare and water quality) or through
trade protectionism. Rather, prices should be reduced through policies that support
diversification and through market access that allows the flow of diverse products
into markets.

63
Cost and affordability of healthy diets across and within countries

Top ten take-home messages

1. People cannot live on bread alone. Food systems can and should focus on access to all
elements of a healthy diet.
2. Nutrition education will not solve the problem of poor diets. Systemic changes in the food
environment are also needed.
3. Healthy diets are often unaffordable for low-income people. Bringing healthy diets within
reach requires higher incomes and expanded safety nets as well as lower prices for a
variety of nutritious items.
4. Poverty lines need to be adjusted upward if they are meant to cover access to nutritious
food that meets dietary needs. Poverty lines that are defined to include the cost of healthy
diets would help align anti-poverty programmes with other global development goals,
linking policies in agriculture and food security to health and human development.
5. Farm production and food markets complement each other. Homestead production
of vegetables, legumes, dairy, poultry, small fish and fruits can be important forms of
social protection and provide nutritious food in some settings, while markets can provide
access to foods beyond what can be grown at each time and place.
6. Diversification in agriculture is needed. The innovations needed to lower diet costs differ
by type of food. Moving beyond starchy staples to legumes, vegetables, fruits, nuts and
seeds, as well as dairy, eggs, fish, and livestock calls for a wide range of actions including
access to higher quality seeds, biotic and abiotic stress resistance (e.g. pests, drought),
disease control, and management of natural resources around both crops and livestock,
as well as better storage and transport, including cold chains and market infrastructure
with product-specific steps to improve and maintain quality.
7. Diet costs and affordability vary significantly by region within countries, revealing
geographic hotspots that are poorly served by the existing food system. The variability
suggests that transport and storage networks are needed to stabilize prices and incomes
over space and time, and to provide access to certain foods in places and times where
they may be unavailable.
8. Protectionism in agricultural trade policy raises costs of nutrient adequate and healthy
diets. It often helps influential interest groups while reducing job creation and wage
growth for other people in the food sector. Reducing barriers between producers and
consumers will lower diet costs while raising farm incomes.
9. Nutrient adequate and healthy diets can be achieved most affordably with small
quantities of animal source foods, including dairy, eggs and small fish that complement
nutrient-rich plant-based foods. It is important that efforts to reduce the cost of diets
also internalize environmental costs, and therefore focus on both animal source and
plant-based foods with the lowest environmental impact.
10. The highest-priority regions are Southern Asia and Africa, but poor affordability of
healthy diets as well as heavy marketing of unhealthy options everywhere are driving
malnutrition in all its forms all around the world.

64
 6 Applications of new food price metrics

References
Allen, L.H., Carriquiry, A.L. & Murphy, S.P. 2019. Proposed harmonized nutrient reference
values for populations. Advances in Nutrition, 11(3): 469–483.
Allen, R.C. 2017. Absolute poverty: when necessity displaces desire. American Economic
Review, 107(12): 3690–3721.
Anderson, K. 2016. Agricultural trade, policy reforms, and global food security. Amsterdam,
Springer.

Bachewe, F. & Headey, D. 2017. Urban wage behaviour and food price inflation in Ethiopia.
The Journal of Development Studies, 53(8): 1207–1222.

Bachewe, F., Headey, D., Bai, Y. & Masters, W. 2019. Selected paper: Nutritional poverty
in economic transition: measuring the affordability of nutrient adequacy and dietary
energy in Ethiopia, 2001–2017. 23 July 2019, Atlanta, USA, Agricultural & Applied
Economics Association.

Bai, Y., Alemu, R., Block, S.A., Headey, D. & Masters, W.A. 2020. Cost and affordability
of nutritious diets at retail prices: evidence from 177 countries. Food Policy, in press.
https://doi.org/10.1016/j.foodpol.2020.101983

Bai, Y., Naumova, E. & Masters, W.A. 2020. Seasonality of diet cost reveals food system
performance in East Africa. Science Advances, 6(49): eabc2162.

Bai, Y., Herforth, A. & Masters, W.A. (forthcoming). Global variation in the cost of a
nutrient adequate diet by population group.

Chastre, C., Duffield, A., Kindness, H., LeJeune, S. & Taylor, A. 2007. The minimum cost
of a healthy diet: findings from piloting a new methodology in four study locations.
London, Save the Children.
Cornelsen, L., Green, R., Turner, R., Dangour, A.D., Shankar, B., Mazzocchi, M. & Smith,
R.D. 2015. What happens to patterns of food consumption when food prices change?
Evidence from a systematic review and meta-analysis of food price elasticities globally.
Health Economics, 24(12): 1548–1559.
Dawe, D. & Lee, J. 2017. Fruit and vegetable prices are rising in Asia and the Pacific.
Bangkok, FAO Regional Office for Asia and the Pacific.

Drewnowski, A. & Darmon, N. 2005. The economics of obesity: dietary energy density and
energy cost. The American Journal of Clinical Nutrition, 82(1): 265S–273S.
FAO. 1996. Rome Declaration on World Food Security and World Food Summit Plan of
Action. Rome.
FAO. 2020a. Database. In: Monitoring and Analysing Food and Agricultural Policies
(MAFAP) [online]. Rome. [Cited January 2020]. www.fao.org/in-action/mafap

FAO. 2020b. Food-based dietary guidelines. In: FAO [online]. Rome. [Cited January 2020].
www.fao.org/nutrition/education/food-dietary-guidelines/home/en/
FAO. 2020c. Trade, Crops and Livestock Products. In: FAOSTAT [online]. Rome. [Cited May
2020]. www.fao.org/faostat/en/#data/TP

6565
CostCost andand affordabilityaffordability ofof healthyhealthy dietsdiets acrossacross andand withinwithin countriescountries

FAO, International Fund for Agricultural Development (IFAD), United Nations

International Children's Emergency Fund (UNICEF), World Food Programme (WFP)
& World Health Organization (WHO). 2019. The State of Food Security and Nutrition
in the World 2019. Safeguarding against economic slowdowns and downturns. Rome,
FAO. (also available at www.fao.org/3/ca5162en/ca5162en.pdf).
FAO, IFAD, UNICEF, WFP & WHO. 2020. The State of Food Security and Nutrition in
the World 2020. Transforming food systems for affordable healthy diets. Rome, FAO.
(also available at https://doi.org/10.4060/ca9692en).

Finaret, A.B. & Masters, W.A. 2019. Beyond calories: the new economics of nutrition.
Annual Review of Resource Economics, 11: 237–259.

Federal Reserve Bank of St. Louis (FRED). 2020. Inflation, consumer prices for the United
States [online]. [Cited 29 May 2020]. https://fred.stlouisfed.org/series/FPCPITOTLZGUSA
Green, R., Cornelsen, L., Dangour, A.D., Turner, R., Shankar, B., Mazzocchi, M. & Smith,
R.D. 2013. The effect of rising food prices on food consumption: systematic review with
meta-regression. British Medical Journal (Clinical Research Edition), 346: f3703.
Headey, D. & Alderman, H. 2019. The relative caloric prices of healthy and unhealthy foods
differ systematically across income levels and continents. Journal of Nutrition, 149(11):
2020–2033.
Herforth, A. 2015. Access to adequate nutritious food: new indicators to track progress
and inform action. In D.E. Sahn, ed. The fight against hunger and malnutrition: The
role of food, agriculture, and targeted policies, pp. 139–162. Oxford, UK, Oxford
University Press. (also available at www.oxfordscholarship.com/view/10.1093/acprof
:oso/9780198733201.001.0001/acprof-9780198733201-chapter-7).
Herforth, A. 2016. Supply side policies and measures for diversifying food production and
for increasing availability and affordability of healthy diets. Paper commissioned by
FAO for the International Symposium on Healthy and Sustainable Diets. Rome.

Herforth, A. & Ahmed, S. 2015. The food environment, its effects on dietary consumption,
and potential for measurement within agriculture-nutrition interventions. Food Security,
7(3): 505–520.
Herforth, A., Arimond, M., Álvarez-Sánchez, C., Coates, J., Christianson, K. & Muehlhoff,
E. 2019. A global review of food-based dietary guidelines. Advances in Nutrition, 10(4):
590–605.
Hirvonen, K., Bai, Y., Headey, D. & Masters, W.A. 2019. Cost and affordability of the
EAT-Lancet diet in 159 countries. Lancet Global Health, 8(1): e59–e66.
Institute of Medicine (IOM). 2006. Dietary reference intakes: the essential guide to nutrient
requirements. Washington, DC, National Academies Press.
Mahrt, K., Mather, D., Herforth, A. & Headey, D. 2019. Household dietary patterns and
the cost of a nutritious diet in Myanmar. IFPRI Discussion Paper 1854. Washington,
DC, IFPRI (International Food Policy Research Institute). (also available at https://doi.
org/10.2499/p15738coll2.133344).
Masters, W.A., Bai, Y., Herforth, A., Sarpong, D., Mishili, F., Kinabo, J. & 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): 1285–1301.
6666
 6 REFERENCES
Applications of new food pricemetrics

Ministry of Health and Sports of Myanmar (MoHS). (no date). Food-based dietary guidelines
for Myanmar. Nay Pyi Taw.

Ministry of Planning and Finance of Myanmar (MoPF) & World Bank. 2017a. An analysis
of poverty in Myanmar (Part 2). Poverty Profile. Nay Pyi Taw.

MoPF & World Bank. 2017b. Myanmar poverty and living conditions survey 2017. Technical
Poverty Estimate Report. Nay Pyi Taw.

National Academies of Sciences, Engineering and Medicine. 2019. Dietary reference

intakes for sodium and potassium. V.A. Stallings, M. Harrison & M. Oria, eds. Washington,
DC, National Academies Press.
Norton, G.W., Alwang, J. & Masters, W.A. 2014. Economics of agricultural development:
world food systems and resource use, 3rd edition. New York, USA, Routledge.
Pernechele, V., Balié, J. & Ghins, L. 2018. Agricultural policy incentives in sub-Saharan
Africa in the last decade (2005–2016). FAO Agricultural Development Economics
Technical Study 3. Rome, FAO.

Pingali, P. 2015. Agricultural policy and nutrition outcomes – getting beyond the
preoccupation with staple grains. Food Security, 7(3): 583–591.

Raghunathan, K., Headey, D. & Herforth, A. 2020. Affordability of nutritious diets in rural
India. Food Policy, in press. https://doi.org/10.1016/j.foodpol.2020.101982

Ravallion, M. 2016. The economics of poverty: history, measurement, and policy. Oxford,
UK, Oxford University Press.
Schneider, K. & Herforth, A. 2020. Software tools for practical application of human
nutrient requirements in food-based social science research. Gates Open Research.
Scott, J. 2019. Nutrient database for fish in Myanmar. Yangon, Myanmar, WorldFish
(unpublished).
Shaheen, N., Rahim, A.T.M.A., Mohiduzzaman, M., Banue, C.P., Bari, M.L., Tukun, A.B.,
Mannan, M.A., Bhattacharjee, L. & Stadlmayr, B. 2013. Food Composition Table for
Bangladesh. Dhaka, University of Dhaka, Institute of Nutrition and Food Science.
Siegel, K., Ali, M.K., Srinivasiah, A., Nugent, R.A. & Venkat Narayan, K.M. 2014. Do
we produce enough fruits and vegetables to meet global health need? PLOS One, 9(8):
e104059.
United States Department of Agriculture (USDA). 2016. National Nutrient Database
for Standard Reference, Release 28. Version Current: May 2016. In: USDA [online].
Washington, DC. [Cited July 2020]. www.ars.usda.gov/ba/bhnrc/ndl

USDA. 2020. Global Agricultural Trade System (GATS). In: USDA [online]. Washington, DC.
[Cited May 2020]. https://apps.fas.usda.gov/gats

van’t Erve, I., Tulen, C.B.M., Jansen, J., van Laar, A.D.E., Minnema, R., Schenk, P.R.,
Wolvers, D., van Rossum, C.T.M. & Verhagen, H. 2017. Overview of elements within
national food-based dietary guidelines. European Journal of Nutrition & Food Safety,
7(1): 1–56.
Wang, S.S., Lay, S., Yu, H.N. & Shen, S.R. 2016. Dietary guidelines for Chinese residents
(2016): comments and comparisons. Journal of Zhejiang University SCIENCE B, 17(9):
649–656.

67
CostCost andand affordabilityaffordability ofof healthyhealthy dietsdiets acrossacross andand withinwithin countriescountries

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

World Bank. 2020. PovcalNet. In: World Bank [online]. Washington, DC. [Cited July 2020].
http://iresearch.worldbank.org/PovcalNet
World Bank. 2020. Global Consumption Database – Food and Beverages. In: World Bank
[online]. Washington, DC. [Cited 18 November 2020]. https://datatopics.worldbank.org/
consumption/sector/Food-and-Beverages
WHO. 2010. Set of recommendations on the marketing of foods and non-alcoholic beverages
to children. Geneva, Switzerland.

6868
Annexes

69
 Pacific
the
and
Asia Africa Region
FAO

FBDG
country
m( oderate
woman
)India China Benin
-L
)EAT
diet
(orancet

2011 2016 2015 FBDG

of
Date
Annex 1.

Number
of
6 *
7 6 i
( ncluding
groups
TABLE A1.1

subgroups
)

s ( dult
woman
3 –5
ervings
a;),3–6
9–20
m
woman
s
;)1 oderate
,( ervings
1 maize
,gsize
serving
c
:1ooked
85
paste
250–4
g 00 staple
Starchy
1d
:3 kcal
(/u
size00
) ry
ncooked
serving
g 0 c
220
1
riceooked
60
pasta
,g
b
1
cgread
,87.5
85–200
assava
0
6ari

2–4
woman
s oderate
(mervings
;)2.5 s servings
2-3
:7erving
;,gsize
m5eat Legumes
/fg
cm hicken
eat
0
:5ish
size
serving fish
100
,8
e
22
c
g
),(3w
shell
rabs
ggs
ith
0
00
120-200
g
1ry
5
e
(8
),3 00
0
5
ggs
kcal
d
g d
ifish
,g
),5
shell
ncluding
0
sried
(100
hrimp
kcal
1
)( 00
pulses b
140
)(c
p ,5
g eans
ooked
cheese
s oya
eanut
0
foods

Fle
e sh
/ ggs
Protein -rich

71
1
;s
servings
:1-2
size
yerving
2
p
,g 25
ogurt
0owder
30
g 0 local
50
,8
cheese
c
g
(u
).milk
oncentrated
5
nsweetened
the cost of a healthy diet

1
(7
sg
kcal 0
size
erving
:;3) 00
ervings milk
fluid ,they
diet
the
of
part
not
are
products
milk
If
be
can Dairy
equivalent such
finfish
as
rich
,in
calcium
other
foods
replaced
by
fish
dried
a
, nd
crustaceans
Computing the cost of the healthy diet

i
o
s
;,2(1Dthers
GLV
3ncludes
ervings
5
l
,g 0
;seaves
servings
:4-6
size
erving
and
tubers
);/ oots
r
potatoes 300–5
g 00 Vegetables
o
c100
,6
gther
arrots
0
kcal
2
)g
1 8
00
(:size
serving

s servings
erving
:;2–3
size
average
1
o
,g n
00
kcal
)(4
s :1
g
size 0
00
;1 erving 200–3
g 50 Fruits
3/4
or
cup
juice

s 4-8
oderate
(mervings
;,5)woman
erving
25–
g 30 ,15
tablespoons
2-3
p
g
tablespoon
er Fat
:5g
(45size
)kcal

additional
in
discussed
as
( uts
n
–
are
one 25–35
n
guts
)foods
r
protein
n
-(iich seeds
and
Nuts
clear
fat
n ot
);and
protein and
tofu
Food - based dietary guidelines used for calculating

Discretionar
/ y
ANNEXES

sugars
 Europe Pacific
the
and
Asia Regi
FAO on

FBD
counG
try
Netherlands Malta Nam
Viet
L diet
-(o
EAT
) rancet

2017 2015 2016 FBDG

of
Date

Number
of
*
7 6 6 i
( ncluding
groups
subgroups
)
TABLE A1.1 (cont .)

breakfast
o
g
:4
s f0
size
erving
; ervings 12–15
servings
serving
;each
size
;s
servings
4-5
: erving
examples
;8f0–100
o
g
,raw
cereals carbohydrate
g
of
equivalent
20
to s
sb rown
bread
sandwich
,1 erving staple
Starchy
,preferably
rice
and
pasta
or
wholegrain p:r(eice
,b otato
xamples
read
products
wholegrain
of
spoon
potatoes
or whole
p
g
;80otatoe meal
s sweet
potato
)

servings
Approx
= –12
week
.9pprox
a
.per Legumes
size
;s
day
per
servings
:1.5
erving serving
;e
servings
5-6 ach
equivalent
p1 ulse
fish
/m
:;s
size
erving
eat
gfish
115
),7
(r
),legumes
aw
aw
0 oish
7g
to
(e
:fprotein
,mfxamples
eat
mish
g
/f100
eat seafood
nuts
21
1e
seeds
and
,1
w
g hite
gg
00 e
s ggs
oybeans
),,tofu
foods

,9ed
r
g
(r
)meat 0
aw Fl
e
/ gg
essh
Protein - rich

72
milk
,ml
size
s:2 50
erving
; ervings
,ml
milk
150
size
;s erving
servings
2-3 ;each
servings
3–4
equivalent
serving
1tub
(1
)y
ml
3
c
,g 0–40
50
heese
ogurt Dairy
I
cncludes
g
.40
heese
drinks
soy calcium
of
mg
100
to
45-50
g
/irkotta
bejna
Computing the cost of the healthy diet

25
g 0 ;serving
servings
3-5
g
:8
size
0 3servings
;e
g
80
serving
ach Vegetables
CosT AND AFFORDABILITY OF HEALTHY DIETS ACROSS AND WITHIN COUNTRIES

20
g 0 ;serving
servings
2–3
:8
size
g 0 3servings
;e
g
80
serving
ach Fruits

size
s
:;1 erving equivalent
;each
servings
5-6
serving
4
g0 Fat
)(1
t ml 5
1 ablespoon to
l5gipids

25
g foods
)r
protein
-(inich )(in
fat
and
oil seeds
and
Nuts

Discretionar
/ y
sugars
 East
Near Caribbean
the
and
America
Latin Regi
FAO on

FBD
counG
try
(at00
Oman
3
)2
level
kcal Jamaica Argentina
L diet
-(o
EAT
) rancet

2009 2015 2016 FBDG

of
Date

Number
of
7 6 *
6 i
( ncluding
groups
)
subgroups
TABLE A1.1 (cont . )

,3size
whole
servings
0.95
;s
:refined
.7
erving kcal
:7
s 0
size
erving
;14
servings
staple
Starchy
cereal
flakes
pasta
,o
1c
dr
g
28
orup
rice
ry )total
(980
kcal +
total
kcal
6
;4 06
servings
foods
o"270
" ptional
kcal
)(876
total
kcal
7 3
kcal
size
s erving
:;3ervings
lentils
cooked
cup
0.75 Legumes
)(219
total
kcal

;soultry
91
:3
size
l
g
,p
,meat
fish
or
ean
erving
0 s
1 erving
foods

Fl
e
/ gg
essh
seeds
or
nuts
o
g
,1 5
gg
1ez total
kcal
)(224
Protein -rich

meat
if
kcal
7 5
size
s erving
:;5ervings

73
if
skim
milk
,4
kcal
0
whole
or
;serving
servings
:0.6
size )(total
consumed
skim
no
if
kcal
374
of 3servings
1catural
=
equivalent
,4
yogurt
or
milk
n
g up
5 Dairy
total
kcal
)(310
cheese

raw
;s
servings
3.4
:1c
,size
vegetables
erving
up
kcal
3 6
size
s
Computing the cost of the healthy diet

:; erving
ervings
,c
chopped
ooked
,12/2
greens
salad
leafy
cup
up 40
g 0 Vegetables
)(108
total
kcal
vegetables
1canned
,or
/2
vegetable
cup
juice

cup
;s
servings
3.95
:c
size
,1
fruits
raw/2
erving
up kcal
:4
s 0
size
erving
;3ervings
300
g Fruits
fruits
canned
,c
1or ooked
chopped
cup
, /2
juice
fruit )(120
total
kcal

coconut
);and
avocado
i
s ncluding
(6ervings s
2ervings
66.
g 5 Fat
total
kcal
2
)(4 5
: 70
size
serving total
kcal
)(270

,a(inlso
seeds
and
fats
foods
rprotein
n
)-(iich (in
)legumes ) seeds
and
Nuts
dried
fruits
including

Discretionar
/ y
foods
optional
kcal
270
ANNEXES

sugars
 America
Northern Regi
FAO on

V
USA
( egetarian
style M
USA
style
( editerranean A
USA
( merican
style FBD
counG
try
level
2
at
300
)kcal kcal
3
2
)at
level
00 kcal
3
2
)at
level
00 L diet
-(o
EAT
) rancet

2015 2015 2015 FBDG

of
Date

Number
of
6 6 6 i
( ncluding
groups
)
subgroups
TABLE A1.1 (cont .)

be
grain
whole
)should
8oalf
-e
(h z
q whole
grain
)should
-ealf
oz
7.5
(h
beq whole
grain
)should
-ealf
oz
7.5
(h
beq staple
Starchy

Legumes
3.75
i -encluding
eggs
,(ozq
n uts
),,tofu
legumes
-eish
oz
7.25
f(i
,meat
qncluding -eish
oz
6.25
,f(i
,meat
ncluding
q
foods

Fl
e
/ gg
essh
)t, ofu
n
e uts
ggs
poultry n uts
ggs
),teofu
poultry
Protein -rich

74
s
,milk
c
e up
q
;1-3oymilk ,s
-eoymilk
cup
2.25
;1c
milk
up
q s
milk
,;1c
e up
q
-3 oymilk
Dairy
cheese
oz
=.5
1yogurt
or cheese
oz
1=.5
yogurt
or cheese
oz
1=.5
yogurt
or

;3 q
e
1c
=
-o
reg
up
range ;3 q
e
1c
=
-o
reg
up
range ;3 q
e
1c
=
-o
reg
up
range
Computing the cost of the healthy diet

vegetables
other
legumes
c up
,1or other
vegetables
legumes
c
,1orup other
vegetables
corup
,1legumes Vegetables
,12c.5
DGLV
potatoes
cup
up DGLV
cupup
potatoes
,12c.5 DGLV
up
potatoes
,12c.5
cup
CosT AND AFFORDABILITY OF HEALTHY DIETS ACROSS AND WITHIN COUNTRIES

,fresh
c=
1-e q
;2 up fresh
,c
1-e
=
cup
;2.5up
q fresh
c
,=
1-e q
;2 up
Fruits
/4
,3uice
dried
j1/2
%
100
cup ,3uice
dried
j1/2
%
100
cup
/4 /4
,3uice
dried
j1/2
%
100
cup

30
g 3
g0 30
g Fat

foods
protein
)-r(inich )foods
r
protein
n
-(iich foods
r
protei
n
)-(iich n seeds
and
Nuts

Discretionar
/ y
kcal
230 kcal
230 kcal
230
sugars
 N/A Regi
FAO on

-Lancet
EAT EAT
L
- ancet EAT
L
- ancet EAT
L
- ancet FBD
counG
try
v
)( egan )(vegetarian p
() escatarian )(flexitarian L diet
-(o
EAT
) rancet

2019 2019 2019 2019 FBDG

of
Date

Number
of
7 9 10 12 i
( ncluding
groups
subgroups
)
TABLE A1.1 (cont . )

6
:kcal
g
2 78 s
roup kcal
6
:2 78 s
group kcal
6
g
2 78 s
: roup
+
grains
cereal
kcal
:6
g 78
2 roups
cerea
grai
+ l
ns cerea
grai
+ nsl 8 1 ns
kcal
cerea
grai
+ l staple
Starchy
starchy
kcal
81
roots
star
81
roots
kcal chy kcal
star
roots
81 chy roots
starchy

Source: Authors' own elaboration .

kcal
387 kcal
317 kcal
254 kcal
254 Legumes

+
eggs
2g4
:1
kcal
roups +
fish
kcal
3
1
:4
g48
eggs
roups
14 Fl
foods

kca
eggls e es
/ ggsh
fish
kcal
76 29
+
poultry
28
meat
red
kcal
Protein -rich

75
kcal
90 kcal
90 kcal
90 Dairy
Computing the cost of the healthy diet

kcal
133 kcal
114 kcal
107 kcal
96 Vegetables

kcal
124 kcal
108 kcal
103 kcal
95 Fruits

kcal
405 kcal
405 kcal
405 kcal
405 Fat

two required subgroups ; and " protein -rich foods" has three subgroups : protein, dairy and nuts .
kcal
180 kcal
180 kcal
180 kcal
180 seeds
and
Nuts

Discretionar
/ y
kcal
95 kcal
95 kcal
95 kcal
95

Notes : * Argentina: fruits and vegetables are required subgroups . China : fruits and vegetables are subgroups ;
dairy and nuts are required subgroups . Netherlands : nominally four groups , but " fruits and vegetables" has
ANNEXES

sugars
Cost and affordability of healthy diets across and within countries

Annex 2. Cost of the nutrient adequate diet using RDA

compared to using H-AR
TABLE A2.1 Regional mean cost by macroregion for all cost of diet indicators
(2017 international USD)

Energydiet Nutrient
adequate Healthy diet
sufficient adequate
Nutrient
Macroregion diet
diet (median)
(RDA)
Africa 0.73 2.15 2.56 3.87
Asia 0.88 2.18 2.48 3.97
Latin America and
1.06 2.83 3.30 3.98
the Caribbean
Northern America
and Europe 0.54 2.29 2.61 3.21

Oceania 0.55 2.07 2.39 3.06

World 0.79 2.33 2.71 3.75
Notes: To provide a sensitivity analysis, the Table compares the cost of the nutrient adequate diet calculated
in this study using harmonized average requirements (H-ARs), harmonized upper levels of intake (H-ULs)
and Acceptable Macronutrient Distribution Ranges (AMDRs) set by IOM (2006), with the cost of the nutrient
adequate diet calculated using the IOM (2006) recommended dietary allowances (RDAs), or Adequate Intakes
(AIs) if the latter is not larger than the H-ARs. For the methodology, see subsection 2.3.2.
Source: Authors' own elaboration.

FIGURE A2.2 Regional mean cost by macroregion for all cost of diet
indicators (2017 international USD)
4

D
S
U
2 7
1
0
2

0
World Africa Asia Latin AmericaNorthern America Oceania
and the Caribbean and Europe

Cost of energy sufficient diet Cost of nutrient adequate diet

Cost of nutrient adequate diet (RDA) Cost of healthy diet (median)

Source: Authors' own elaboration.

76
 Annexes

Annex 3. Average cost of each diet and national percentage

of people who cannot afford each diet, by region
and income level (simple average across countries)
TABLE A3.1 Average cost of each diet, by region and country income group
(simple average across countries)
Energy Nutrient
Regions Healthy diet
sufficient diet adequate diet

World 0.79 2.33 3.75

Africa 0.73 2.15 3.87
Northern Africa 0.75 2.90 4.12
Sub-Saharan Africa 0.73 2.06 3.84
Eastern Africa 0.61 1.98 3.67
Middle Africa 0.73 2.09 3.73
Southern Africa 0.86 2.29 3.99
Western Africa 0.80 2.05 4.03
Asia 0.88 2.18 3.97
Central Asia 0.84 2.04 3.39
Eastern Asia 1.27 2.63 4.69
South-eastern Asia 0.92 2.42 4.20
Southern Asia 0.80 2.12 4.07
Western Asia 0.74 1.87 3.58
Latin America and the Caribbean 1.06 2.83 3.98
Caribbean 1.12 2.89 4.21
Latin America 1.00 2.78 3.75
Central America 1.13 3.04 3.81
South America 0.91 2.61 3.71
Oceania 0.55 2.07 3.06
Northern America and Europe 0.54 2.29 3.21
Country income group

Low-income countries 0.70 1.98 3.82

Lower-middle income countries 0.88 2.40 3.98
Upper-middle income countries 0.87 2.52 3.95
High-income countries 0.71 2.31 3.43
Notes: Data shown are the average cost (USD) per person per day of each diet. Methods and data sources
are detailed in the text.
Source: Authors’ own elaboration.
77
Cost and affordability of healthy diets across and within countries

TABLE A3.2 Average national percentage of people who cannot afford

each diet, by region and country income group (simple average
across countries)

Energy sufficient Nutrient adequate Healthy

diet diet diet
Macroregion
Africa 11.3% 51.0% 73.8%
Asia 0.4% 11.7% 36.6%
Latin America and
3.7% 18.1% 26.5%
the Caribbean
Northern America
and Europe 0.3% 1.7% 3.7%

Oceania 0.1% 5.0% 21.0%

Subregion

Australia and New Zealand 0.2% 0.5% 0.7%

Central Asia 0.3% 11.0% 33.2%
Eastern Asia 0.3% 1.8% 15.6%
Eastern Europe 0.3% 1.7% 3.5%
Latin America and
3.7% 18.1% 26.5%
the Caribbean
Melanesia 0.0% 9.6% 41.3%
Northern Africa 1.4% 29.2% 46.0%
Northern America 0.6% 1.0% 1.2%
Northern Europe 0.2% 0.4% 0.8%
South-eastern Asia 0.7% 20.7% 46.2%
Southern Asia 0.5% 17.9% 57.6%
Southern Europe 0.4% 3.8% 8.7%
Sub-Saharan Africa 12.5% 53.4% 76.9%
Western Asia 0.3% 3.8% 21.7%
Western Europe 0.1% 0.2% 0.3%
Country income group

High-income countries 0.3% 0.9% 2.0%

Low-income countries 12.7% 61.4% 86.2%
Lower-middle countries 6.3% 33.1% 58.9%
Upper-middle countries 2.1% 11.5% 24.2%
World 4.6% 23.3% 38.3%
Notes: Data shown are the percentage of people in each region whose household income is below the total
cost of the most affordable locally available items needed to meet each standard of diet quality, assuming
that they can spend no more than 63 percent of their income on food. Average percentages are reported as in
Table 8 of The State of Food Security and Nutrition in the World 2020 (FAO, IFAD, UNICEF, WFP and WHO,
2020). Methods and data sources are detailed in the text.
Source: Authors' own elaboration.

78
 Annexes

Annex 4. Lower- and upper-bound estimates of the number of

people who cannot afford each diet type
The lower-bound estimate for the number of people who cannot afford each diet type is
calculated assuming 100 percent of income spent on food; that is, the lower bound counts
the number of people who have total daily income lower than the cost of a given diet
(Table A4.1). By this calculation, 1.86 billion people would be unable to afford healthy diets.
Because people do not spend all of their income on food, our main estimate assumes
37 percent of income needs to be spent on non-food items such as housing and agriculture.
However, many people need to spend more than this amount on non-food items, and would
not be able to spend 63 percent of income on food, particularly in higher-income countries
where food expenditures are typically lower than 30 percent and housing costs are high.
Therefore, we calculate upper-bound estimates as the mean food expenditure share across
countries within World Bank income groupings. The mean share of expenditure on food is
14.9 percent, 27.9 percent, 42.3 percent, and 50.2 percent for high-, upper-middle-, lower
middle- and low-income countries, respectively.15 We define income needed as the income
that would be required to afford both a healthy diet and other non-food needs, based on the
cost of the healthy diet and holding the food/non-food share constant. We calculate income
needed using the following formula: [Cost of the diet / food expenditure share in World
Bank country income classification]. For example, if the cost of a healthy diet is USD 3 in
a given low-income country, where food expenditures are on average 50 percent of total
expenditures, income would need to be USD 6 to afford both the healthy diet and non-food
needs. This upper-bound estimation is shown in Table A4.2. By this calculation, 4.75 billion
people would not be able to afford healthy diets.

TABLE A4.1 Lower-bound estimate of the number of people (in millions)

who cannot afford the cost of each diet, by region and
country income group

Energy sufficient Nutrient adequate Healthy

diet diet diet
Macroregion
Africa 59.62 413.55 754.34
Asia 2.81 184.05 1 050.71
Latin America and
4.24 31.02 49.41
the Caribbean
Northern America
and Europe 4.49 7.25 9.60

Oceania 0.06 0.13 0.24

Subregion

Australia and New Zealand 0.06 0.12 0.12

Central Asia 0.01 0.44 2.43
Eastern Asia 1.52 3.14 53.93
Eastern Europe 0.24 1.23 2.24
►►
15 The median food expenditure shares are 14 percent, 25 percent, 41 percent and 51 percent for the four income
levels, which are quite close to the mean.

79
Cost and affordability of healthy diets across and within countries

TABLE A4.1 (cont.) Lower-bound estimate of the number of people (in millions)
who cannot afford the cost of each diet, by region and
country income group

Energy sufficient Nutrient adequate Healthy

diet diet diet
Latin America and
4.24 31.02 49.41
the Caribbean
Melanesia 0.00 0.01 0.12
Northern Africa 0.42 43.70 86.44
Northern America 3.25 4.15 4.24
Northern Europe 0.12 0.20 0.24
South-eastern Asia 0.28 43.13 182.28
Southern Asia 0.98 136.13 800.23
Southern Europe 0.86 1.60 2.58
Sub-Saharan Africa 59.20 369.85 667.90
Western Asia 0.01 1.21 11.84
Western Europe 0.03 0.07 0.31
Country income group

High-income countries 5.71 8.36 11.00

Low-income countries 17.13 207.05 403.74
Lower-middle countries 40.82 373.21 1 316.69
Upper-middle countries 7.56 47.38 132.87
World 71.22 636.01 1 864.30
Notes: Data shown are the number of people in each region whose household income is below the cost of the
most affordable locally available items needed to meet each standard of diet quality, assuming that they can
spend 100 percent of their income on food. Methods and data sources are detailed in the text.
Source: Authors' own elaboration.

80
 Annexes

TABLE A4.2 Upper-bound estimate of the number of people (in millions)

who cannot afford the cost of each diet, by region and
country income group

Energy sufficient Nutrient adequate Healthy

diet diet diet
Macroregion
Africa 267.12 866.74 1 081.54
Asia 199.94 1 684.12 3 033.55
Latin America and
39.14 224.31 304.14
the Caribbean
Northern America
and Europe 10.01 67.50 154.33

Oceania 0.16 0.88 1.43

Subregion

Australia and New Zealand 0.12 0.36 0.68

Central Asia 0.27 6.68 19.31
Eastern Asia 26.00 258.23 886.97
Eastern Europe 0.94 23.20 68.02
Latin America and
39.14 224.31 304.14
the Caribbean
Melanesia 0.04 0.52 0.75
Northern Africa 10.38 134.50 182.95
Northern America 6.77 17.16 32.69
Northern Europe 0.28 2.09 4.60
South-eastern Asia 40.01 283.15 463.27
Southern Asia 124.66 1 088.04 1 583.58
Southern Europe 1.93 21.48 39.15
Sub-Saharan Africa 256.74 732.23 898.59
Western Asia 8.99 48.02 80.42
Western Europe 0.09 3.57 9.87
Country income group

High-income countries 23.35 101.88 212.17

Low-income countries 81.88 423.79 533.76
Lower-middle countries 327.06 1 755.80 2 497.58
Upper-middle countries 84.09 562.07 1 331.47
World 516.38 2 843.54 4 574.98
Notes: Data shown are the number of people in each region whose household income is below the cost of the
most affordable locally available items needed to meet each standard of diet quality, assuming that they can
spend the national average expenditure share of their income on food (14.9 percent, 27.9 percent, 42.3 percent
and 50.2 percent for high-, upper-middle-, lower-middle- and low-income countries, respectively). Methods
and data sources are detailed in the text.
Source: Authors' own elaboration.

81
Cost and affordability of healthy diets across and within countries

Annex 5. Nutrient adequacy of the selected diets

The healthy diet results were analysed for mean adequacy ratio (MAR) across 20 nutrients
that have a lower bound. For included nutrients, please refer to Table 1.

TABLE A5.1 Percent of nutrient needs met by healthy diets

(mean adequacy ratio)

MAR_HAR MAR_RDA
mean sd mean sd
Macroregion
Africa 0.93 0.03 0.88 0.04
Asia 0.94 0.04 0.89 0.06
Latin America
0.94 0.04 0.89 0.06
and the Caribbean
Northern America
0.94 0.03 0.89 0.04
and Europe
Oceania 0.94 0.03 0.89 0.04
Food-based dietary guidelines
Argentina 0.91 0.04 0.84 0.05
Benin 0.93 0.03 0.89 0.04
China 0.95 0.03 0.90 0.05
India (moderate woman) 0.93 0.04 0.87 0.05
Jamaica 0.93 0.04 0.88 0.05
Malta 0.91 0.04 0.85 0.05
Netherlands 0.94 0.03 0.89 0.04
Oman (at 2 300 kcal level) 0.93 0.04 0.88 0.04
USA (American style
0.96 0.03 0.92 0.04
at 2 300 kcal level)
USA (Mediterranean style
0.96 0.03 0.93 0.04
at 2 300 kcal level)
USA (Vegetarian style
0.97 0.02 0.93 0.03
at 2 300 kcal level)
Viet Nam 0.93 0.03 0.88 0.04
World 0.94 0.04 0.89 0.05

Source: Authors' own elaboration.

82
 Annexes

Annex 6. Methodological note on constructing a nutritious

food poverty line in Myanmar
In order to construct a healthy diet (CoRD) and healthy diet with food preferences (CoRD-FP)
comparable to the 2015 Myanmar food poverty line and total poverty line, we closely adhere
to the methodology and the following details outlined in the World Bank’s poverty estimation
technical report (MoPF and World Bank, 2017b): the reference poor population (households
with expenditure within 10 percentage points of the poverty line); the food poverty line
calorie target (2 238 calories); the composition of the food basket; the caloric content of
foods; and wastage factors.
The CoRD is designed to reflect the cost of the cheapest foods within each food group
and is typically constructed using data comprising commonly available foods. The CoRD-FP
reweights food consumption patterns to meet dietary guidelines and as such places greater
weight on non-staple food groups. The fish/meat/eggs, fruits, and vegetable food groups tend
to have a wide variety of foods with a wide distribution of prices. To reduce the influence of
atypically costly and/or rarely consumed foods, we limit the CoRD and CoRD-FP food baskets
to items within the top 85 percent of total food expenditure within each food group.
For consistency with the 2015 Myanmar total poverty line, the World Bank non-food
allowance of 436 kyat per capita per day is added to the CoRD-FP to yield a CoRD-FP poverty
line (MoPF and World Bank, 2017a). Affordability of the World Bank, CoRD and CoRD-FP food
baskets is assessed by comparing spatially adjusted per capita household food expenditure to
the cost of each food basket. Official poverty and CoRD-FP poverty headcounts are assessed
by comparing spatially adjusted per capita household total expenditure to each poverty line.
The World Bank poverty line food basket includes prepared foods purchased for
consumption at home and food away from home (FAFH). Because it is not possible to
classify prepared foods and FAFH into food groups, the CoRD-FP basket must exclude these
items. In order to compare the composition of the poverty line food basket to the CoRD and
CoRD-FP food baskets, a new poverty line basket is constructed, which excludes FAFH and
is then rescaled to hit the 2 238 calorie target.

83
Cost and affordability of healthy diets across and within countries

Annex 7. Data and methods to simulate impacts of policy

change on diet costs

Data sources
Agricultural market distortions are modelled using estimates of nominal rates of protection
for 62 countries and 58 food products. The nominal rate of protection (NRP) is calculated
using the difference between observed border price and farm-gate price, after accounting
for market access costs. This difference primarily represents international trade barriers,
though other policies such as price interventions or exchange rate management may also
play a role (Pernechele, Balié and Ghins, 2018). Nominal rates of protection are compiled,
harmonized and published by the AgIncentives Consortium with data and input from the
World Bank, Agrimonitor at the Inter-American Development Bank, the Monitoring and
Analysing Food and Agricultural Policies (MAFAP) unit at FAO, and the Organisation for
Economic Co-operation and Development.
Retail price data is from the 2011 round of the World Bank’s International Comparison
Project (ICP). The NRP values used to match 2011 and 2017 ICP retail prices are drawn
from the four years leading up to and including each year. We have updated the publicly
available NRP data using MAFAP file data, resulting in the addition of data for 2017 and
several observations for poultry meat and sesame seed in Mozambique from earlier years.
Additionally, there are several significant revisions to existing publicly available NRP data for
sub-Saharan African countries. The final dataset includes 3 581 total observations across all
eight years, with 2 154 observations in 2008–2011 and 1 427 observations for 2014–2017.
One limitation of the 2017 simulation is that the AgIncentives NRP data extends only through
2016. 2017 observations come only from the MAFAP file data, which covers 14 sub-Saharan
African countries.
After compiling all available data, we impose a validity check on any negative NRP values
in the dataset to retain only those values which could plausibly be sustained over time by
export restrictions. We check that for each negative NRP for a given product in a given
country, that country has had positive net exports of the product in that year or any one of
the three years prior. Products are matched by FAO commodity list code to FAO export and
import data (FAO, 2020c). In seven cases, where a product in the AgIncentives NRP dataset
had no exact match in the FAO trade data, the closest match from that commodity group
was chosen. In three cases where trade data was missing for a product with negative NRP
values, products were matched to the corresponding HS-6 codes and UN Comtrade data was
used to assess export and import levels (accessed through USDA, 2020). We assume that any
negative NRPs where the country did not competitively export the product within the four
years leading up to or including that year to be temporary outliers or measurement errors,
and we trim these values to zero. In total, we trimmed 185 negative observations across the
eight years considered in the study.

Countries and number of observations are listed in Table A7.1, which also shows 2011
and 2017 income classifications for results reported in the main text. We use World Bank
country income classifications as the basis for four income groups. Income group is chosen
over regional group because region classifications vary considerably in the heterogeneity
of their components. Some, like Southern Asia, are relatively alike, while others, like East
Asia and Pacific, are very diverse. Table A7.1 lists the 62 countries by income group
and shows the total number of NRP observations for each time period for each country.
Two-thirds (65.1 percent) of the data are from high-income and upper-middle-income
countries. There are relatively few observations for low-income countries (13.1 percent),

84
 Annexes

pointing to the need for better data on the implications of agricultural policies in these
countries. Observations are particularly sparse for nutrient-dense food groups such as fruits,
vegetables, and dairy.
Items in the price dataset are matched to one of nine functional food groups listed in
Table A7.2. Table A7.2 lists all of the 58 foods in the dataset by food group classification.
Analysing nutrient adequacy requires the use of nutritionally functional food groups.
For example, vegetables are typically grouped into leafy green vegetables, red-orange
vegetables, and other vegetables based on their differing micronutrient profiles. Palm oil,
vegetable oils, and leguminous crops like peanuts and soybean would be treated separately
as sources of saturated fat, unsaturated fat and protein, respectively. This analysis cannot
capture such levels of nuance due to the limited number of foods in the NRP dataset. We
classify foods based broadly on nutritional function, with some aggregations: all vegetables
are grouped together, and oil crops that may be used to produce either oil or other food
items, such as groundnuts and soybean, are grouped together as nuts and oilseeds. Though
there are relatively few observations of eggs compared to other food groups, we treat them
as a separate food group both due to nutritional differences between eggs and poultry meat
and because the value chains for these products have significant differences. For low-income
countries, we do not observe any NRPs for eggs, so we instead use the average NRP for eggs
in the sub-Saharan Africa region. The low-income country group has a majority of countries
from this region and may have similar policies due to regional trade agreements.

TABLE A7.1 Countries by income group and number of nominal rate of

protection observations, 2008–2011 and 2014–2017

2011 2017
countries obs. countries obs.
High-income Australia 60 Argentina 27
countries
Bahamas 16 Australia 45
Barbados 8 Bahamas 8
Canada 60 Barbados 8
European Union 68 Canada 45
Iceland 24 Chile 30
Israel 60 European Union 51
Japan 76 Iceland 18
Republic of Korea 40 Israel 45
New Zealand 40 Japan 57
Norway 36 Republic of Korea 30
Switzerland 44 New Zealand 30
Trinidad and Tobago 22 Norway 27
USA 52 Panama 20
Switzerland 33
Trinidad and Tobago 22
Total 606 565
►►

85
Cost and affordability of healthy diets across and within countries

TABLE A7.1 (cont.) Countries by income group and number of nominal rate of
protection observations, 2008–2011 and 2014–2017

2011 2017
countries obs. countries obs.
Lower Belize 10 Ghana 21
middle
Bolivia (Plurinational
income 16 India 57
State of)
countries
El Salvador 28 Indonesia 24
Ghana 24 Kenya 31
Guatemala 52 Nigeria 8
Guyana 14 Philippines 33
Honduras 9 Sri Lanka 7
India 76 Ukraine 39
Indonesia 48 Viet Nam 24
Nicaragua 18
Nigeria 16
Pakistan 31
Paraguay 26
Philippines 44
Senegal 12
Sri Lanka 28
Ukraine 52
Viet Nam 32
Total 536 244
Upper Argentina 36 Belize 10
middle
Brazil 36 Brazil 27
income
countries Chile 40 China 42
China 56 Colombia 33
Colombia 44 Costa Rica 27
Costa Rica 36 Dominican Republic 30
Dominican Republic 40 Ecuador 18
Ecuador 24 Guyana 5
Jamaica 48 Jamaica 12
Kazakhstan 48 Kazakhstan 36
Mexico 56 Mexico 42
Panama 20 Russian Federation 39
Peru 32 South Africa 42
Russian Federation 52 Suriname 9
►►

86
 Annexes

TABLE A7.1 (cont.) Countries by income group and number of nominal rate of
protection observations, 2008–2011 and 2014–2017

2011 2017
countries obs. countries obs.
Upper South Africa 56 Turkey 42
middle
Suriname 36
income
countries Turkey 56
Uruguay 30
Total 746 414
Low-income Benin 8 Benin 10
countries
Burkina Faso 28 Burkina Faso 25
Burundi 18 Burundi 18
Ethiopia 36 Ethiopia 31
Haiti 32 Malawi 12
Kenya 36 Mali 25
Malawi 11 Mozambique 21
Mali 28 Rwanda 15
Mozambique 18 Senegal 16
United Republic of
Rwanda 15 13
Tanzania
United Republic of Uganda
16 18
Tanzania
Uganda 20
Total 266 204

Total observations across all countries 2008–2011: 2 154

Total observations across all countries 2014–2017: 1 427

Source: Authors' own elaboration.

87
Cost and affordability of healthy diets across and within countries

TABLE A7.2 Summary of nominal rates of protection data by food group

and product

Nominal rate of protection Nominal rate of protection

2008–2011 2014–2017

Food group Product Obs. Median Min. Max. Obs. Median Min. Max.
Dairy Milk 160 6.1 -68.5 204.6 98 3.4 -56.0 361.7
Eggs Eggs 110 8.2 -48.4 309.6 73 0.0 -13.3 346.2
Fruits Bananas 56 0.0 -72.8 72.6 29 0.0 -58.8 109.7
Grapes 26 0.0 0.0 199.5 18 0.0 0.0 140.6
Apples 24 7.2 0.0 113.8 18 0.0 0.0 69.7
Pineapples 20 0.0 -69.8 77.1 13 0.0 -59.8 113.7
Oranges 19 0.0 0.0 235.5 10 0.0 0.0 287.7
Mangoes 0.0 -23.6 16.4 0.0 -49.6 0.0
16 7
Avocados 10 9.6 0.0 339.1 7 72.4 0.0 161.5
Coconuts 10 -12.6 -32.5 38.0 5 -17.5 -32.3 42.4

Grapefruit
(including 6 0.0 0.0 34.8 4 0.0 0.0 0.0
pomelos)

Other melons
(including 4 -30.0 -37.1 -27.0 N/A N/A
cantaloupes)

Pears 4 129.9 79.5 187.8 3 72.7 69.0 86.1

Strawberries 4 15.5 6.0 24.9 3 11.7 8.1 11.8
Tangerines,
mandarins, 4 88.5 31.2 117.7 3 41.8 29.6 42.9
clementines

Papayas 2 4.2 -27.4 35.9 2 18.0 -0.5 36.5

Grains Maize 165 0.0 -76.1 351.1 115 0.0 -67.9 189.2
Rice 157 11.7 -54.6 258.6 107 29.8 -71.4 216.3
Wheat 106 0.0 -43.9 95.9 76 0.0 -52.4 145.8
Barley 63 0.0 -43.3 146.2 48 0.0 -30.1 248.1
37 0.0 -85.0 156.0 27 0.0 -37.5 72.5
Sorghum
0.0 -45.0 100.5 21 0.0 -25.2 83.7
Oats 28
Rye -37.0 13.5 -22.9 6.3
8 -7.5 6 -9.6
Millet 4 3.4 -6.8 8.9 4 -2.5 -10.4 5.6
Teff 4 -24.1 -50.4 -12.6 4 69.7 60.0 97.4

Nuts and Soybeans 57 0.0 -36.3 990.5 38 0.0 -40.4 478.7

oilseeds
Groundnuts 36 0.0 -56.0 180.9 32 0.0 -47.5 108.0
Sunflower seed 32 0.0 -43.1 19.6 24 0.0 -44.9 75.1
Rapeseed 24 0.0 -24.2 159.2 18 0.0 -9.7 86.2
Palm oil 18 0.0 -22.7 25.7 12 0.7 -11.5 41.8
Cashew nuts 16 0.0 -22.6 34.6 13 -26.4 -54.2 71.2
Sesame seed 8 3.0 -19.2 18.1 10 21.9 -33.7 118.7

►►

88
 Annexes

TABLE A7.2 (cont.) Summary of nominal rates of protection data by food group
and product

Nominal rate of protection Nominal rate of protection

2008–2011 2014–2017

Food group Product Obs. Median Min. Max. Obs. Median Min. Max.
Pulses Beans, dry 45 27.4 -12.7 186.3 26 2.6 -46.3 330.4
Chickpeas, dry 8 0.4 0.0 1.9 3 8.0 0.0 10.7
Lentils, dry 8 0.0 -13.0 97.5 7 0.0 0.0 90.7
Peas, dry 4 0.0 0.0 0.0 3 0.0 0.0 0.0
Pulses, not
elsewhere 4 0.0 0.0 0.0 3 0.0 0.0 0.0
specified
Poultry Poultry meat 152 30.2 -19.2 512.2 99 15.1 0.0 484.5
Red meat Bovine meat 166 0.0 -48.9 152.7 101 1.5 -41.7 231.5
Pig meat 133 16.0 0.0 293.3 82 11.8 0.0 235.3
Sheep meat 59 0.0 0.0 69.0 44 5.3 0.0 170.3
Roots, Potatoes 51 0.1 -48.7 599.5 33 0.0 -71.8 325.8
tubers and
plantains Cassava 35 0.0 -28.8 153.6 24 10.7 -16.5 127.4
Yams 13 0.0 -37.0 62.5 5 0.0 -71.3 112.5
Plantains 12 7.1 -10.6 38.2 8 28.8 0.0 54.8
Sweet potatoes 11 36.0 0.0 137.0 4 145.8 0.0 173.4
Sweeteners Sugar 116 12.4 -71.6 208.8 77 26.1 -61.8 147.5
Honey 6 -16.9 -40.3 108.8 2 206.4 179.4 233.4
-48.0 129.1
Vegetables Tomatoes 37 0.0 21 0.0 -57.8 29.6
Onions 22 18.5 -47.3 342.3 15 5.0 -42.3 239.4
Cabbages and
8 126.2 27.0 314.1 6 126.5 27.0 245.2
other brassicas
Chillies and 403.9
8 125.2 0.0 4 183.8 -32.3 187.8
peppers
Beans, green 4 -40.5 -65.6 -26.1 2 -54.5 -55.1 -53.8
Cucumbers and
4 17.5 3.0 52.6 3 3.0 3.0 3.0
gherkins
Spinach 4 3.0 3.0 3.0 3 3.0 3.0 3.0
Asparagus 2 0.0 0.0 0.0 N/A N/A
Pumpkins,
squash and 2 42.3 -43.0 127.7 2 -25.9 -38.3 -13.6
gourds
Vegetables:
fresh, not 136.9 122.7 186.6
2 223.6 310.3 2 58.7
elsewhere
specified

Source: Authors' own elaboration.

89
Cost and affordability of healthy diets across and within countries

Data analysis
We first collapse the NRP observations from 2008–2011 and 2014–1207 for all countries and
products into a set of mean protection rate by income level for each food group. Because the
NRP is measured at the farm-gate level and our price data is at the retail level, we assume
100 percent as a lower bound of markup from farm gate to retail and 300 percent as an
upper markup bound. Assuming that the rate of protection applies only to the farm-gate
portion of the retail price, we compute a reference price for each retail food item as follows:

Pr Pr
Pr’ = m + (1)
1+m (1+τ)(1+m)

where:
m = assumed markup from wholesale to retail
τ = nominal rate of protection (NRP)
Pr = observed ICP retail price.

We calculate a wholesale price (Pw) based on observed retail price and assumed markup:

Pr
Pw = (2)
1+m

Assuming the additional cost due to markup remains constant as a cost of retail supply,
Ps = mPw , we can adjust the Pw to reflect a zero-distortion scenario as follows:

Pw
Pw’ = (3)
1+ τ

then add the constant cost of retail supply to obtain an adjusted retail price, which we label
a reference price:

Pr’ = Ps + Pw’ (4)

Substituting equations 2 and 3 into equation 4, we obtain equation 1. Because we

do not have an observed wholesale price, we keep the component of the price attributed
to the markup from wholesale to retail (Ps) constant for the purposes of this simulation.
We calculate two reference prices based on an income-group NRP imputation at 100 percent
and 300 percent markup for each of the 159 and 170 countries in the ICP datasets for 2011
and 2017, respectively. Reference prices are then used to calculate a new cost of nutrient
adequacy for each country.
90
Price and affordability are key barriers to accessing sufficient, safe, nutritious food to
meet dietary needs and food preferences for an active and healthy life. In this study,
the least-cost items available in local markets are identified to estimate the cost of
three diet types: energy sufficient, nutrient adequate, and healthy (meeting food
based dietary guidelines). For price and availability the World Bank’s International 9
Comparison Program (ICP) dataset is used, which provides food prices in local currency
units (LCU) for 680 foods and non-alcoholic beverages in 170 countries in 2017.
In addition, country case studies are developed with national food price datasets in
United Republic of Tanzania, Malawi, Ethiopia, Ghana and Myanmar.
The findings reveal that healthy diets by any definition are far more expensive than the
entire international poverty line of USD 1.90, let alone the upper bound portion of the
poverty line that can credibly be reserved for food of USD 1.20. The cost of healthy diets
exceeds food expenditures in most countries in the Global South. The findings suggest
that nutrition education and behaviour change alone will not substantially improve
dietary consumption where nutrient adequate and healthy diets, even in their cheapest
form, are unaffordable for the majority of the poor. To make healthy diets cheaper,
agricultural policies, research, and development need to shift toward a diversity of
nutritious foods.

The FAO Agricultural Development Economics Technical Study series collects technical
papers addressing policy-oriented assessments of economic and social aspects of food
security and nutrition, sustainable agriculture and rural development.
The series is available at www.fao.org/economic/esa/technical-studies

FOR FURTHER INFORMATION

Agrifood Economics - Economic and Social Development
♦ www.fao.org/economic/esa
ISBN 978-92-5-133725-7 ISSN 2521-7240
♦ ESA-Director@fao.org
Food and Agriculture Organization
of the United Nations (FAO)
9 789251
Rome, Italy 337257
CB2431EN/1/12.20