Transcript slides - Tufts University

Measuring the cost of nutritious diets:
Results for diet diversity
in Ghana and Tanzania
Will Masters
Friedman School of Nutrition Science & Policy, Tufts University
www.nutrition.tufts.edu | http://sites.tufts.edu/willmasters
Presentation at the FAO Expert Consultation on Trade and Nutrition
15-16 November 2016
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
New price indexes
• Indicators of Affordability of Nutritious Diets in Africa (IANDA)
– Project funded by UKAid through its program on Innovative
Methods and Metrics for Ag-Nutrition Actions (IMMANA),
managed by LCIRAH (LSHTM, SOAS) and Tufts University
– Project led by Jennie Coates and Anna Herforth,
with Rebecca Heidkamp (Johns Hopkins University),
Daniel Sarpong (University of Ghana),
Fulgence Mishili and Joyce Kinabo (Sokoine University)
– Today’s results are with Yan Bai (PhD student at Tufts)
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
What do food price indexes
tell us about the cost of nutritious diets?
But is
nutritious food
also becoming
less expensive?
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Existing food price indexes
are weighted by market value
To measure “world prices”, the FAO aims to include traded foods
The FAO Food Price Index is a measure of the monthly change in
international prices of a basket of food commodities. It consists of
the average of five commodity group price indices, weighted with
the average export shares of each of the groups for 2002-2004.
Total of 23 commodities (73 prices), in 5 groups:
• Cereals
-- wheat (11), maize (1), rice (16)
• Oils/Fats
-- soybean, sunflower, rapeseed, groundnut, cottonseed,
copra, palm kernel, palm, linseed, castor (1 each)
• Dairy
-- whole milk powder, skim milk powder, cheese (2 each),
cheese (1)
• Meat
-- poultry (13), beef (7), pork (6), sheep (1)
• Sugar
-- sugar (1)
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Food price trends and fluctuations
differ by level of value added
To measure cost of living (or output),
national accounts aim to include all goods & services
Consumer price indexes,
Food away from home
weighted by expenditure shares
Food at home
Processed food & feed
Unprocessed food and feed
Producer price indexes, weighted by market sales
Source: US. Bureau of Labor Statistics, downloaded 12 November 2016.
Definitions and chart data are available at http://myf.re/g/aPV1
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
What is the cost of a nutritious diet?
• Need nutritional values, rather than market values
– Previous subsistence indexes focus on dietary energy
• undernourishment, to maintain weight and physical activity
 cost of staple foods or calories
– Today will focus on cost of dietary diversity
• beyond staples, foods with more nutrients and other attributes
=> cost of specific foods or food groups
– Next step will be cost of nutrient adequacy
• account for quantities, to reach recommended dietary intake
=> cost of nutrients
– Then include cost of other healthful or harmful attributes
• dietary guidelines, with upper as well as lower limits
=> cost of attributes (+ for beneficial, - for harmful)
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Possible indexes of nutritional value
•
For hunger and dietary energy, a long history and wide use
– Playfair (1821) chart of wheat prices and wages for 1565 to 1821
– Sukhatme (1961) and FAO’s Prevalence of Undernourishment in calories
– Drewnowski (2004) measure of energy cost ($/kcal) and density (kcal/kg)
•
For nutrient adequacy, a long history and many specific uses
– Stigler (1945) linear programming to compute least-cost diets
– USDA Thrifty Food Plan for US nutrition assistance (1975, 1983, 1999, 2006)
– SCUK Cost of Diet tool (2009) and FANTA et al. Optifood (2012) for aid programs
•
But diet diversity in terms of food groups of particular interest today
– DD takes account of all attributes in each food group, not just nutrients
– Grouping foods by attributes facilitates dietary surveys, where frequency of intake
over 24hr (or 7d) is feasible to recall, whereas quantities are usually unknown
– Number of food groups included in the diet has been linked to child height
(Arimond & Ruel 2004) and to women’s nutrient adequacy (Arimond et al. 2010)
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Dietary diversity has overtaken nutrients
in popularity among nutritionists
Relative frequency of use in English-language books, 1950-2008
Source: Calculated from https://books.google.com/ngrams, 12 November 2016
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Definitions of dietary diversity
• DD measures have evolved rapidly in recent years
– Nonlinearity in response
• From number of food groups (DDS) to minimum thresholds (MDD)
– Nonunitary households
• From household scores (HDDS) to individuals (children 0-5, women 15-49)
• Focus now is on Minimum Dietary Diversity for Women (MDD-W)
– Defined as eating at least 5 out of 10 food groups in previous 24 hrs
•
•
•
•
•
•
•
•
•
•
Starchy staples (Grains, white roots/tubers, plantains)
Pulses (beans, peas and lentils – includes soybeans)
Nuts and seeds (higher fat than pulses, includes groundnuts)
Meat, poultry and fish
Dark green leafy vegetables
Other vitA-rich fruits & vegetables
Other vegetables
Other fruits
Eggs
Dairy
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Interpretation of dietary diversity
• MDD-W is primarily seen as a validated proxy for nutrient adequacy
– DD might also protect against excess intake of some foods
– DD can also be seen as a measure of real income and well-being
• MDD-W has a direct economic interpretation
–
–
–
–
–
–
–
Within groups, all foods are perfect substitutes
Each group provides a different mix of attributes
Every group also contributes towards energy balance
Groups can be ranked by cost towards total daily energy needed
The least costly energy source (typically starchy staples) can meet few other needs
People will include higher-cost energy sources in their diet to meet additional needs
People who include at least five groups are likely to reach adequacy thresholds
• MDD-W food groups reflect common dietary substitutions
•
•
•
•
Low-cost starchy staples:
Plant-based protein & fat:
Animal-sourced foods:
Micronutrient-rich foods:
Grains, white roots/tubers, plantains
Pulses; Nuts and seeds
Eggs; Dairy; Meat, poultry and fish
Dark green leafy vegetables,
Other vitamin A-rich fruits & vegetables
Other vegetables, Other fruits
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Computing the cost of dietary diversity
1.Obtain market prices
-- today’s data are from national statistical offices
(prices collected for national accounts, based on budget shares & uniformity)
-- another source could be agricultural market information services
(prices collected for market integration, based on role in long-distance trade or storage)
-- eventually sources could include humanitarian agencies
(prices collected for early warning (eg VAM), based on presence in remote markets)
2.Convert units of measure
-- all prices deflated to 2011 PPP prices based on PWT 8.1 data
-- quantities converted to dietary energy (kcal) or weight (kg)
3.Classify into MDD-W food groups
-- select the one lowest-cost item in each group
-- identify the fifth lowest-cost group
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Results for Ghana
• Monthly averages across multiple markets, Jan. 2005 – Dec. 2014
• No prices available for dark green leafy vegetables or dairy
• The five lowest-cost are usually, in rank order:
• Starchy staples:
• Least costly is cassava or maize
• Pulses
• Least costly is usually soyabeans, occasionally cowpea
• Nuts and seeds
• Least costly is groundnuts
• Vit A-rich fruits and vegetables
• Least costly is mangoes, occasionally oranges, rarely tomatoes
• Other fruit
• Least costly is banana, occasionally pineapple
• The higher-cost groups are usually:
• Meat, poultry and fish
• Least costly is usually smoked herring, occasionally salted dried tilapia
• Other vegetables
• Least costly is usually onions, occasionally garden eggs
• Eggs
• Usually the most expensive food group
• Key questions:
• Is it OK to use cost/kcal to reach energy needs, vs. cost/kg as meal ingredients?
• Is it OK to count only five lowest-cost groups, vs. all food groups?
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
The least-cost starchy staple is usually cassava,
but sometimes maize
Sometimes maize can replace cassava
Note: if we ranked foods by weight ($/kg), the least cost starchy staple would always be cassava
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
The least-cost vegetable is onions in season,
otherwise garden eggs
Garden eggs:
Note: if we ranked foods by weight ($/kg), the least cost other veg. would usually be garden eggs
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
The least-cost fruit is almost always banana,
and occasionally orange or pineapple
orange
orange
pineapple
Note: if we ranked foods by weight ($/kg), the least cost other fruit would usually be pineapple
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Aggregating all food groups,
fluctuations 2009-11 then rise 2012-14
The average of 8 foods,
one from each group
The price of just one food from the 5th group
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
We would get different results when
ranking foods and food groups by weight
When we choose the heaviest (least cost in $/kg)
food from each group, the average of 8 foods
is more stable and has a different trend
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Results for Tanzania
• Monthly averages across multiple markets, Jan. 2011 – Dec. 2015
• No prices available for dark green leafy vegetables or eggs
• The five lowest-cost are usually, in rank order:
• Starchy staples:
• Least costly is always maize
• Pulses
• Least costly is always soyabeans
• Nuts and seeds
• Least costly is always groundnuts
• Other fruits
• Least costly is usually avocado, occasionally green banana
• Meat, poultry and fish
• Least costly is always pork meat
• The higher-cost groups are usually:
• Dairy
• Least costly is fresh cow milk
• Vit A-rich fruits and vegetables
• Least costly is usually mangoes, sometimes papaya
• Other vegetables
• Least costly is onions
• Key questions:
• Is it OK to use cost/kcal to reach energy needs, vs. cost/kg as meal ingredients?
• Is it OK to count only five lowest-cost groups, vs. all food groups?
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
The least costly VitA-rich fruit or vegetable
is mangoes in season, otherwise papaya
Papaya is occasionally less costly than mango
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Aggregating across all food groups,
rise in 2011-12 then relative stability to 2015
The price of just one food from the 5th group
The average of 8 foods, one from each group
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
In the TZ data, the cost of the 5th group
is far above the average of all 8 groups
3
4
5
Prices
the least-cost
food and
in each
group (energy
basis)
Figure 15.of
Minimium
Prices by Foodstuffs
the MCDD/ACOG
Indicators (per
1,000kcal)
0
1
2
The average of 8 foods, one from each group
2011m1
2011m7
2012m1
2012m7
2013m1
2013m7 2014m1
Year/Month
2014m7
2015m1
2015m7
White Maize Grains
Round Onions
Green Banana
Avocado
Mangoes
Papaya
Fresh Cow Milk
Soya Beans
Groundnuts
Pork Meat
MCDD
ACOG
2016m1
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Conclusions and next steps
• Creating a price index for the cost of a nutritious diet is challenging
– Need clear definition of “nutritious”: a quantity index
• Here, we use MDD-W
• Next, we will use RDIs
• Eventually, include attributes other than nutrients
– Need new kinds of data
• Here, we use prices from GH & TZ national accounts
• This is good enough for a pilot, not yet ready to guide policy or programs
• Pilot reveals data-collection priorities
– For this particular aspect of nutrition policy and programs
» Missing groups:
green leafy vegetables, dairy, eggs
» Misfit definitions of groups:
pulses vs. nuts/seeds
» Misfit foods within groups:
meat/fish/poultry (in TZ case)
(For this index, would need least-cost item such as dried fish)
– For this and many other uses in nutrition policy and programs
» Food composition data, for items with market prices
» Dietary intake data, to test how intake responds to prices
Measuring the cost of nutritious diets in Africa
motivation | method | ghana | tanzania | next steps
Conclusions and next steps
(continued)
• How can economists meet demand for cost-of-nutrition data?
– Need new understanding of what a cost-of-nutrition index measures
• Not a cost-of-living index!
– A benchmark for least costly way to meet nutrition needs
• Relevant to actual choices, but neither descriptive nor normative
– Should fit only the envelope of actual behavior
» How close a fit? (e.g. purchases by weight, or for culinary reasons)
» Fit to whose needs? (what location, age/sex or other category)
– Need clear agreement on definition of nutrition needs
• So far we are focusing on MDD-W and RDIs, but…
– may need to measures that count gains above/below clinical thresholds
– may need measures of other attributes beyond energy & nutrients
=> a long and complex research agenda !
Thank you!
Data analysis:
Yan Bai (PhD student)
Data sourcing:
Daniel Sarpong (Ghana)
Fulgence Mishili (Tanzania)
Project leadership:
Jennie Coates (PI)
Anna Herforth (Co-PI)
Rebecca Heidkamp (JHU)
Joyce Kinabo (Sokoine)
Funding:
DFID, through IMMANA