POLICY SEMINAR
What is the Cost of a Healthy Diet? New Price Indexes Reveal Changes in Affordability of Nutritious Foods
MAY 29, 2019 - 12:15 PM TO 01:45 PM EDT
The Economic and Organised Crime Office (EOCO) has been advised by the Office...
The Cost of Nutritious Diets: New Price Indexes Reveal Changes in Affordability of Foods and Nutrients
1. Results from
Changing Access to Nutritious Diets in Africa and South Asia (CANDASA)
funded by UKAid and the Bill & Melinda Gates Foundation
http://sites.tufts.edu/candasa
IFPRI Policy Seminar, 29 May 2019
The Cost of Nutritious Diets:
New Price Indexes Reveal Changes in
Affordability of Foods and Nutrients
2. Why measure the cost of nutritious diets?
• Food environments are changing rapidly,
in both urban and rural areas
– from own-production to use of markets
– from generic commodities to branded, packaged foods
– shifts in cost of starchy staples, milk, eggs, meat, fruit & veg
– shifts in earnings, towards more choice among affordable foods
• Food prices are collected to measure inflation and poverty
– economists have used prices for the cost of calories & food demand
– aid programs have used prices to recommend a least-cost diet
– we use prices to study food systems, for access & affordability
• Analysis of diet costs can guide intervention
– each CANDASA researcher will share findings from their work
– discussion today is at project midpoint, from Jan. 2018 to June 2020
Photo by Anna Herforth
at Nsawam market, Ghana
3. The CANDASA project (Jan. 2018 - June 2020)
builds on past work at Tufts, IFPRI and elsewhere
Our research team:
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
4. Why are CANDASA price indexes needed?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
Motivation for this project:
5. Standard food price data say nothing about nutrition
World food prices
(traded commodities only)
Source: www.fao.org/worldfoodsituation, April 2019
National food prices
(weighted by share of total expenditure)
Components:
Cereals
Vegetable oils
Sugar
Meat
Dairy
Vulnerability monitoring
(most commonly consumed
staples)
6. For the cost of nutritious diets, need data on diverse foods
Photo: A. Herforth (Ghana), 2016
Cereals White R&T Pulses Nuts and Seeds DGLV Seed Oil
Maize Yam Cowpea Groundnut Coconut oil
Millet Cocoyam Soya bean Groundnut (red) Palm oil
Sorghum Cassava Groundnut oil
Rice Gari
Sweet potato
Cassava dough
Dried cassava
Meat, Poultry and Fish Vegetables Fruits Egg
Beef Tomato Mango Egg
Pork Garden Egg Pineapple
Salted dried fish Okro Palm fruit
Live chicken bird Onion Watermelon
Smoked herring Ginger Orange
Anchovy Pepper Banana
Predecessor project (IANDA) helped Ghana MoFA expand price monitoring
Source: John Nortey, Ghana Ministry of Food and Agriculture (2017)
7. For the cost of nutritious diets, need data on diverse foods
Photo: A. Herforth (Ghana), 2016
Cereals White R&T Pulses Nuts and Seeds DGLV Seed Oil
Maize Yam Cowpea Groundnut Nkontommire Coconut oil
Millet Cocoyam Soya bean Groundnut (red) Jute mallow Palm oil
Sorghum Cassava Bambara Melon Seeds Alefu (Amaranth) Groundnut oil
Rice Gari
Sweet potato
Cassava dough
Dried cassava
Meat, Poultry and Fish Vegetables Fruits Egg
Beef Tomato Mango Egg
Pork Garden Egg Pineapple
Salted dried fish Okro Palm fruit
Live chicken bird Onion Watermelon
Smoked herring Ginger Orange
Anchovy Pepper Banana
Fresh fish Cabbage Coconut
Chicken meat Lettuce Avocado
Snail Carrot Pawpaw
Goat meat
Mutton
Predecessor project (IANDA) helped Ghana MoFA expand price monitoring
Source: John Nortey, Ghana Ministry of Food and Agriculture (2017)
Ghana MoFA
institutionalized the
updated food list
nationwide
8. New price indexes account for nutritional needs
Cost of Recommended Diets
• Based on food-based dietary guidelines
• Food groups & servings (g) per day
• Choose foods that meet recommendations
at lowest total cost
Cost of Nutrient Adequacy
• Lowest cost of meeting calorie, macronutrient, and
micronutrient needs
• Specify quantity of each nutrient
• Choose foods that meet needs at lowest total cost
Benin Food Guide India Food Guide
U.S. Food Guide
Compare to
Cost of
Caloric
Adequacy
9. What data are needed for the cost of nutrient adequacy?
Nutrient requirements, nutrient composition and food prices
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
10. For nutrient adequacy, diets must meet requirements
Source: Dietary Reference Intakes (DRI) values for the US & Canada, last revised 2011.
Available online at https://www.nal.usda.gov/fnic/dietary-reference-intakes
Energy balance 1 calorie constraint
Macronutrient ranges and requirements
For protein, fat and
carbohydrates
3 lower bounds
3 upper bounds
2 average requirements
Micronutrient requirements
For 7 minerals (calcium,
copper, iron, magnesium,
phosphorous, selenium,
zinc) and also for 9
vitamins (niacin, riboflavin,
thiamin, folate, and
Vitamin A, B6, B12, C, E)
16 average requirements
9 upper limits for toxicity
U.S. and Canada dietary reference intakes specify minimum and maximum levels
for up to 20 nutrients, totaling 34 constraints, for each demographic group
11. For nutrient adequacy, we consider all individuals’ needs
Nutrient needs vary by age, sex & status
0
100
200
300
400
500
600
700
800
900
1000
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Estimated Average Requirements for Vit. A
by age and status
w/Pregnancy w/Lactation
Female Male
Our base case
is an adult woman
We use average requirements,
to meet needs of a median
healthy person
-- We also compare to recommended
daily allowances (RDAs), to meet
needs of almost all people (97.5%)
Calculated from Institute of Medicine (2011), Dietary Reference Intakes.
https://www.ncbi.nlm.nih.gov/books/NBK56068
12. Can we compute price indexes facing households?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
Using household data to infer individual adequacy and cost
13. Cost per day depends on age, sex & maternity status
Women and children need fewer calories
than men, but typically need more
micronutrients per calorie.
• Adolescent girls need more calcium and
phosphorus;
• Breastfeeding calls for more folate,
protein, vit A, especially for older women
(>31);
• Older men need more vitamins B6, B12,
C and E;
• Children 1-3 years need the most lipids.
Different needs translate to different costs to meet all
daily nutrient needs.
14. Households are consuming a diet that, at best,
meets 70-80% of total nutrient needs.
Do households have enough food to meet nutrient needs?
We use household survey data to compare
consumption to needs.
• To aggregate needs to the household
level, we use two sharing scenarios and
individual nutrient density needs.
Sharing food: everyone eats the same diet.
• Nutrient density per nutrient, defined by
the member with the highest nutrient-per-
calorie need;
• Eaten per person in proportion to energy
needs.
Targeted diets: each person eats a separate
diet tailored to individual needs.
• Households are a sum of individual
needs.
15. How do prices vary across types of rural markets?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
Testing for differences between more vs. less remote markets
16. Do prices vary by type of market, as well as location?
Market in a
district capital
Market in a
more remote town
Mitundu market
(44 km from district capital )
Thyolo
(Thyolo district capital)
Map and photos: S. Kaiyatsa, April 2019
Are prices lower in district capitals, or in more remote markets?
17. In Malawi, diet costs are lower in more remote markets
Contrary to our expectations,
costs per day are lower in more remote towns than in district capitals
t-statistic
More remote markets
(n=12)
District capitals
(n=17)
Mean Std. Dev Mean Std. Dev
Cost of Nutrient Adequacy (CoNA) 1.48 0.81 1.61 0.82 -4.429***
Cost of Caloric Adequacy (CoCA) 0.75 1.29 1.01 2.41 -4.054***
Cost per day, 2007-17
(US$ at PPP prices)
Cost per item
(Malawi kwacha / kg)
t-statisticMore remote markets District capitals
Mean Std. Dev Mean Std. Dev
Vegetables 186 156 196 157 -5.0277***
Fruits 157 122 162 121 -1.6857**
Staples (incl. plantains) 167 150 170 157 -1.5342*
Animal foods 714 1139 719 1139 -0.3406
Fats and oils 377 278 372 252 0.7466
Legumes and nuts 356 194 347 190 2.3259**
Prices in more remote markets are lower for vegetables, fruits and starchy staples
Source: CANDASA results, from Kaiyatsa et al. (2019)
18. How do prices vary over time, in different regions?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
In Tanzania, a major factor in diet cost is seasonality
19. We have price data for 64 items
from 21 regions, 2011-15
Seasonal price rises are largest for fruits & vegetables,
and in southern regions
Source: CANDASA results, from Bai et al. (2019)
Seasonal rises in the cost of nutrient adequacy
are greatest in the southern regions
Width shows estimated period of higher prices
Dark dots show statistically significant peaks
Colors show magnitude of seasonal rise
Seasonality in diet costs varies among foods
in the timing, duration and magnitude of seasonal peaks
Starchy
staples
Pulses
Animal
foods
Fruits and
Vegetables
Oils and Fat
20. Seasonality is greater for nutrients than for energy
Seasonal variation is significantly greater
for nutrients than for daily energy
Source: CANDASA results, from Bai et al. (2019)
Seasonality is driven by fruits and vegetables,
even though a large fraction of diet costs
would be animal foods, beans and grains
21. How do prices relate to earnings, for affordability?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
In Ethiopia, the big change is higher wages
22. Ethiopia invested heavily in
cereals and jobs growth
• Cost of starchy staples and daily
energy has fallen since 2008
• Cost of nutrient-rich foods
(vegetables, meat, eggs) has been
stable or rising
• Real wages have grown sharply
Not all Ethiopians earn wages
• But unskilled wages are a proxy for
incomes of the poor:
➢ Correlated with HH spending
➢ A “reservation” or fallback wage
Food system transformation & nutritious diets
in Ethiopia: 2001-2017
0
1
2
3
4
5
6
7
8
9
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Trends in the costs of nutrient adequacy, caloric adequacy
and real wages in Ethiopia, 2001-2017 (2011 PPP $/day)
Real wages for unskilled workers
(right scale)
Cost of nutrient adequacy
Cost of caloric adequacy
Diet costs
($/day)
Daily wages
($/day)
Source: CANDASA results, from Bachewe et al. (2019)
23. Cost of
nutrient
adequacy
Cost of
calorie
adequacy
0%
5%
10%
15%
20%
25%
30%
35%
What’s the net result?
• Since 2011, the cost of nutritious diets
has fallen from >30% to <20% of wages
• But the cost of calories has fallen more,
from over 10% to only 5% of wages
• Attention is shifting from calories to
nutrients, and the relative cost of
healthier diets
Counterfactual:
• If nutrient costs had stayed constant as
wages rose, nutritious diets would now
cost just 11% of wages
Improved affordability shifts attention to food choice
Cost per day of nutrients and calories as a share of
unskilled wages in Ethiopia, 2001-2017
Source: CANDASA results, from Bachewe et al. (2019)
24. How do prices differ by food group,
for the cost of a recommended diet?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
In India, the big concern is diet quality
25. What does India’s officially recommended diet cost?
National recommendations specify
food groups and servings
that allow for a lacto-vegetarian diet
Servings Common items
Starchy
staples
9-20
Maize, Millet, Rice,
Wheat, Bread
Proteins 2-4
Peas, Gram, Pea
dal, Khesari dal
Dairy 3
Milk (cow), Milk
(buffalo), Curd
Fruit 1
Banana, Guava,
Papaya, Pineapple
Vegetables 2
Radish, Onion,
Gourd, Pumpkin
Green leafy
veg
1
Palak, Bhaji sag
leaves, Amranth
Oils fats 4-8
Mustard,
Groundnut, Palm
• Least-cost diet was $0.50 per day (2011 $)
• Largest cost was dairy, then staples
Starchy
staples
Proteins
Dairy
Fruit
Other
vegetables
Green leafy
vegetables
Oils & fats
$0.00
$0.10
$0.20
$0.30
$0.40
$0.50
$0.60
Source: CANDASA results, from Raghunathan et al. (2019)
+
26. How costly is the recommended diet relative to wages?
Diet costs are a larger
fraction of wages for women
• 70-80% of women’s wage
• 50-60% of men’s wage
Highly unaffordable:
• Costs for non-food expenses
and other family members?
• Volatile!
But some improvement:
• wages ↑ & diet cost ↓ since
2007, but more so for men
Trends in the cost of the recommend diet as a
percent of wages for rural unskilled labor
Women
Men
No data
Source: CANDASA results, from Raghunathan et al. (2019)
27. Data by region and food group reveal specific challenges
LOTS of variation in seasonality by food
group, especially fruits and vegetables
Fruit
Vegetables
Leafy-veg
-$0.10
-$0.05
$0.00
$0.05
$0.10
$0.15
$0.20
$0.25
$0.30
Seasonality in food group prices
LOTS of variation in trends for the cost of
a recommended diet relative to wages
-30 -20 -10 0 10 20 30
Uttaranchal
Odisha
Punjab
Andhra Pradesh
Tamil Nadu
Uttar Pradesh
Haryana
Karnataka
Assam
Himachal Pradesh
Rajasthan
Maharashtra
West Bengal
Kerala
Bihar
Chhattisgarh
Jharkhand
Madhya Pradesh
Gujarat
Change in diet cost, 2001 to 2011
Source: CANDASA results, from Raghunathan et al. (2019)
28. How can price indexes be used to inform policy?
Will Masters
(PI)
At Tufts: Anna Herforth
(Co-Investigator)
Daniel Sarpong
(U. of Ghana)
Fantu Bachewe
(IFPRI-Ethiopia)
Yan Bai
(PhD student)
Kate Schneider
(PhD student)
Derek Headey
(Co-PI)
Kalyani Raghunathan
(IFPRI-India)
At IFPRI:
Fulgence Mishili
(Sokoine U.-Tanzania)
Stevier Kaiyatsa
(Gvt of Malawi)
Individual consultants:
Ghana now has the most interest among government officials
29. • There is government interest,
but support is needed
Policy dialogue for use of new price indexes in Ghana
• To help introduce new metrics, an informal “Nutritious Food Prices Analysis Support Group” to be hosted at
the University of Ghana’s Institute for Statistical, Social & Economic Research (ISSER)
• This is to allow government statistical officers to meet regularly with local economists and other stakeholders,
before and after release of new data, to discuss the significance and interpretation of each indicator
• Meetings will begin mid-2019
• Since 2016 IANDA/CANDASA
has worked with Ghanaian
authorities (GSS, MoFA) to
help officials use the new price
indexes within their routine
food price monitoring
30. Conclusions and implications of CANDASA research
At Tufts:
Derek Headey
(Co-PI)
At IFPRI:
Individual consultants:
31. The cost of nutritious diets: what have we learnt so far?
What do price indexes for nutritious diets reveal?
1) Total cost per day depends on the combination of all needed foods and nutrients
a) Total cost exceeds earnings for the poorest, who need aid to afford a healthy diet
b) Relative cost among affordable foods varies widely, influencing food choice
2) Food system changes could make healthier diets more affordable for all
a) Reduce seasonality in cost of perishable foods
b) Reduce spatial variation by region and type of market
3) Price monitoring could help guide policies and programs
a) Data collection: product range, frequency, representativeness, quality
b) Results dissemination: raise the profile of nutritional affordability metrics
c) Policy & program decision-making: use of data to guide intervention
32. The cost of nutritious diets: what can we do in the future?
There is so much more for CANDASA & others to do…
1. Use additional dietary guidelines, for more countries & population groups
2. Refine cost of nutrient adequacy, with more on food composition & needs
3. Improve price information, for more diverse foods, locations and time periods
4. Expand analysis to cost of unhealthy foods and dietary patterns
5. Expand use of wage & income data (as in India & Ethiopia)
6. Compare least-cost to actual diets, to understand divergence
7. Analyze impact of local production, storage & trade on diet costs
8. Analyze impact of diet costs on nutrition and health outcomes
33. The cost of nutritious diets: what can we do in the future?
Encourage uptake of nutritious-food price indices
• Every national statistical agency collects food price data
• Many agricultural and food agencies collect food price data
• Yet almost none of them monitor the cost of nutrition!
How do we catalyze that change?
• Capacity strengthening projects
• Institutional collaboration: nutrition/health ministries and stats agencies/agricultural
ministries (e.g. get prices in SUN movement, multisectoral nutrition strategies)
• Make price data publicly available!