Presentation at IRRI

1. Food Systems for Healthier Diets
Impact analysis at farm-household & market level
Ruerd Ruben

2. Personal introduction
 Research Coordinator Wageningen Economic Research
(Food & Nutrition Security; Value Chains ; M&E)
 Chair Impact Assesment (DEC-WU)
(commodity certification & labelling)
 Director Evaluation Division – Min. Foreign Affairs
 Chair Development Effectiveness – Radboud University
 Program leader Food Security & Sustainable Land Use
 Research on peasant economics, cooperatives, policies
 Links to CGIAR - A4NH (FP1), CCAFS (CoA 4.1), PIM (FP3)
2

3. Key Issues
 Yield gaps
● Decomposition analysis
● Policy analysis
● Yield gaps & Nutrition gaps
 Food Systems (FSHD – A4NH)
● Methods & Metrics
● Supply & Demand-side interventions
● Upscaling & Anchoring
 Impact analysis
● Household level (DiD, RCTs)
● Value chain level (MAS)
3

4. Yield gaps
 Global food production has to increase with 60% to fulfil
demand in 2050 (FAO, 2012).
 Production can be increased if yield gap can be closed.
4
Mueller et al. (2012)

5. Ambition
 Integrate economic and agronomic approaches to
assess yield gaps at the micro level (maize)
● Decomposition approach (Van Dijk et al., 2016)
● African Maize Database combining plot,
household, farm and spatial data for 6 countries.
 systematically link yield gaps components with
agricultural policies to increase smallholder
productivity and yield.
 Based on collaboration between WEcR and PPS
 Information to policy-makers
 Deliver scientific output
5

6. Yield gap programme
 Yield Gap Benchmarking atlas (2012-2015)
 CIMMYT Africa maize yield gap assessment (2015-2016)
 DFID IMAGINE (2015-2017)
 Countries: Ethiopia, Ghana, Uganda, Tanzania
 Researchers: Michiel van Dijk, Roel Jongeneel, Tom Morley
(part time research assistant), Ruerd Ruben, interns,
 Partners: PPS – Martin van Ittersum, Pytrik Reidsma; EEPRI –
Ethiopia; University of Ghana/ISSER – Ghana, CIMMYT.
6

7. Linking plot, farm and spatial data (TZA)
7
Global Yield Gap Atlas LSMS-ISA

8. Theoretical yield
response function
Frontier yield
response function
Potential (yp)
Input (ton/ha)
Yield (ton/ha)
Px/Py
Actual (ya) F1
xeo
Economically
feasible (yeo)
Technically
efficient (yte)
xte0 xp/xeu
F2FYG
TEYG
EFYG
YG
TYG
Feasible (yf)
Yield gap: Conceptual framework

9. Potential to increase maize output (Uganda)
9

10. Policies to close the yield gap
1. Extension services/education (e.g. diffusion of best practice)
2. Input subsidies (e.g. lower price of fertilizer)
3. Credit and insurance (i.e. lower risk of using inputs)
4. Diffusion of market information (e.g. mobile phones)
5. Applied agricultural research (e.g. hybrid seeds, precision
agriculture)
6. Producer organisations (i.e. increasing bargaining power)
7. Land rights (e.g. higher investment)
8. Infrastructure (e.g. roads and storage)
9. Gender empowerment (e.g. fair allocation of subsidies)
10.Regional trade (e.g. removing trade barriers)
10

11. Linking policies and yield gap framework
11
Yield gap Problem Policy solution Note
Technical efficiency
yield gap (TEYG)
 Lack of knowledge on best-
practice farm management, use
and combination of inputs.
 Extension services
 Knowledge transfer from best
practice to average farmers
 Farmer field schools
 Gender empowerment
Economically feasible
yield gap (EYG)
 Farmers cannot obtain credit to
purchase inputs.
 Farmers are risk averse and do
not purchase inputs because of
high risk of crop failure
 Less effort because of missing
property rights
 Credit facilities
 Insurance market
 Property right system
Feasible yield gap
(FYG)
 High costs of inputs because of:
o Transport costs
o Limited number of dealers
 Low price of outputs because of:
o Thin markets
o Limited storage
 Road infrastructure
 Irrigation
 Dealer network
 Smart input subsidies
 Storage facilities
 Market information (mobile phones)
 Marketing boards
 Farmer organisation
 Regional integration
Cannot be
fully closed
because of
high costs
Technology yield gap
(TYG)
 No knowledge, information and
enabling environment to use
advanced techniques and
technology (precision farming).
 No appropriate technology (e.g.
small scale tractors, drought
resistant seeds)
 Applied research programs Cannot be
fully closed
because of
random
shocks

12. From yield gaps to Nutrition gaps
 Does the increase of productivity increase diversify
nutrition of smallholders?
● Does productivity increase production diversity?
● Does production diversity increase nutrition
diversity?
 Research funded by CIMMYT
● Yield gap analyses
● Nutrition gap analyses (UGA, TZA, ETH)
● Use LSMS data
12

13. Analytical Framework
13
Based on Kanter et al. (2015)
Agricultural household model

14. Food and nutrition security indicators
 Nutrition/Diet diversity:
● Dietary diversity index (DDS) ;12 food groups, 7-
day recall period
● Food consumption score (FCS); 12 food groups,
weights no. consumption days in 7-day recall period
● Caloric intake; Consumption of 69 food items and
nutritional contribution
 Production diversity:
● Crop count ; 12 crop groups
● Simpson’s index : 1 − 𝑠𝑗
2
with 𝑠𝑗 =
𝑎 𝑖𝑗
𝐴 𝑖
and 𝑗 =
𝑎 𝑖𝑗
𝐴 𝑖
● Own crop count :Crops grown and consumed in 7-
day recall period (12 crop groups)
14

15. Data & Methods
 LSMS data for Uganda
● 2009/2010 (3,123 obs.); 2010/2011 (2,716 obs.)
● 2011/2012 (2,716 obs.)
 Panel set (max. 1,722 observations)
 Production data (plot level); Consumption (household
level); Socio demographic data (individual level)
 Panel analysis (3x3 models)
● 3 dependent variables household nutrition diversity
● 3 production diversity indicators
● Fixed-effects estimations (based on Hausman test)
15

16. Nutrition diversity Uganda
16
Source: LSMS Uganda, authors calculations
0.0
2.0
4.0
6.0
8.0
10.0 Dietary Diversity Score
2009
2010
2011 0
500
1000
1500
2000
Caloric intake per capita
per day
2009
2010
2011

17. Production diversity Uganda
17
0.0
1.0
2.0
3.0
4.0
5.0
6.0
Crop count
2009
2010
2011
0.00
0.20
0.40
0.60
0.80
Simpson index
2009
2010
2011
0.00
0.10
0.20
0.30
0.40
0.50
East West North Central Uganda
Own crop count share
2009
2010
2011

18. Regression results DDS
18
(1) (2) (3)
Production diversity
Crop count 0.046
Simpson's index 0.364
Own crop ratio 0.059
Household size 0.008 0.016 0.009
Household head gender - Male -0.316 -0.449 -0.332
Age of the household head -0.006 -0.001 -0.005
Education level of the household head 0.004 0.008 0.004
Food expenditure 0.001 0.002 0.001
Incomes 0.007 0.006 0.007
Total cropped area 0.001 0.001 0.001
Proportion of own production -0.303 -0.345 -0.275
# non-agricultural income sources 0.011 0.008 0.011
Agriculture Decision - Household Head 0.048 0.154 0.056
Year 2009-10 7.414 7.069 7.416
Year 2010-11 7.227 6.890 7.229
Year 2011-12 7.494 7.136 7.484

19. Summary of results
19
 Nutrition indicators differ across regions and have a mixed
influence: DDS/FCS ↑ and Caloric intake ↓
 Overall production diversity shows a negative trend
 Positive relationship between production diversity and
nutrition indicators
 Food expenditures increase food and nutrition diversity
 Household size is positively correlated to caloric intake
DDS FCS Caloric intake
Crop count 0.046*** 0.668*** 1.599**
Simpson' s index 0.364** 3.585** -2.688
Own production ratio 0.059* 0.844** 0.971

20. 20

21. Triple burden of malnutrition
500 billion

22. KB: Global Food & Nutrition Security
Pathways for
Sustainable
Agricultural
Development
Robust value
chains &
Resilient food
markets
Food
systems for
healthier
diets

23. Anchoring and
scaling
Food system
innovations
Diagnosis
&
Foresight
Multi-
stakeholder
Platformss
Co-innovation &
Lab-in-the Field
Experiments
Interdisciplinary
Problem Analysis
Food systems for healthier diets

24. Multiple drivers
Rapid Diet
Transitions
Supermarkets
&
Wet Markets
Processed
Foods
& Fast Food
Food
Environment
Healthy
Food
Choices

25.  Multi-disciplinary analysis
 Multiple scales (‘food web’): local ….. global
 Multiple stakeholders: public, private, civic
 Multiple incentives: market – governance - information
Convergence of approaches

26. How to understand the dynamics of food systems?
How can nutrient-rich food lead to healthier diets?
How can healthy food be delivered through sustainable &
resilient value chains?
How can food choices be nudged towards healthier diets?
Which incentives for food system innovations ? (RCTs)
Key research questions

27. 27

28. Impact analysis (Diff-in-diff)
Before After
Intervention
Control
Coffee
Tea
Cocoa
Palm Oil
Bananas

29. 0
.5
1
1.5
2
2.5
0 .2 .4 .6 .8 1
Propensity Score
Control
Treated
Before Matching
0
.5
1
1.5
2
Density
0 .2 .4 .6 .8
Propensity Score
Control
Treated
After Matching
ON-CS • OFF-CS • Total
FT 42 6 48
Organic 97 6 103
FT 39 9 48
Conventional 30 6 36
Number of observations on Common Support
Matching (PSM)

30. Commodity Standards: decreasing returns
Contested Areas:
• Over-certification
• Value added distribution inside chain
• Multi-annual contracts (trust)
Organisation
strengthening
Production
techniques
Farmers
income
Power

31. Commodity standards: comparative results

32. Value added distribution
32

33. Fair Chain
33
-€ 1.00
€ 0.00
€ 1.00
€ 2.00
€ 3.00
€ 4.00
€ 5.00
€ 6.00
€ 7.00
Figure 2: Value Chain Restructuring
Series1 Series2

34. Value chain analysis: gaming & simulation
34

35. Behavioural linkages
35
Standard for
business
support
Adoption of
GAP & GBP
Reliable &
sustainable
supply
Quality of
procurement
Improved
welfare &
sustainability
Risk
perception
Mutual
trust
Transaction
costs

36. VC Gaming Outcomes (trust)
36
Treatment Control
Descriptive aspect of trust yes no yes no
Keeps promises 24 2 7 22
Offers good price 29 1 13 34
Good timing of sorghum collection and payment 6 10 1 6
Certainty of the market 22 0 2 5
Provided training on farming 33 0 3 0
Provided inputs (seeds, loans, other) 14 0 2 0
Trader is honest 6 0 1 7
No other buyer available 1 0 5 0

37. VC Contract Game design
37

38. VC Simulation outcomes (dashbord)
38

39. Possible cooperation WEcR/DEC - IRRI
 Resource use efficiency (land, water, labour)
 Yield & Nutrition linkages in Rice systems
 Rice marketing (prices; margins; post-harvest mgt)
 Rice Value chains (contracts, trust, relability)
 Rice in Food Systems: Processed rice consumption
(additives; health)
39

40. Thanks for your attention
RRuerd
Ruerd.Ruben@wur.nl