General Equilibrium Effects of PSNP in the Small and in the Large
1. ETHIOPIAN DEVELOPMENT
RESEARCH INSTITUTE
General Equilibrium Effects of PSNP
in the Small and in the Large
Mateusz Filipski1 ; Getachew Ahmed Abegaz 1; Tadele Ferede2 ; J. Edward Taylor 3
; Xinshen Diao1 ; Alemayehu Seyoum Taffesse1
1: IFPRI
2 : Addis Ababa University
3: UC-Davis
Ethiopian Economics Association
13th International Conference on the Ethiopian Economy
July 23-25, 2015
Addis Ababa
1
2. 2
Motivation
• PSNP has two components: Cash Transfer (CT) and Public Works
(PW)
– Soil & water conservation (SWC), irrigation, roads, schools, clinics…
• Most evaluation focuses on recipient households
• But impacts may spread far beyond recipients:
– Cash Recipients spend their money within their economy
– Public works affect agro-ecological and economic environment
Need to evaluate the FULL impact
We evaluate full impacts locally, and nationwide
6. Economy-wide Modeling
• System of equations defining all economic flows
– Production output, factor and input demands
– Household incomes and expenditures
– Trade flows
– Taxes and transfers
– Etc.
• Can model the full impact of a shock
• Applicable to economy of any scale
– Single country
– Several countries, Region, Village, Household
– Kebele
6
7. 7
3-step evaluation
• Step 1: Econometrics
– Estimates the average impacts of PW projects on production
– Grounds simulations in reality
• Step 2: Local impacts: LEWIE model
– Kebele scale: Local Economy-Wide Impact Evaluation
– Provides detail
• Step 3: National impacts: CGE model
– Computable General Equilibrium
– Provides the big picture
8. 8
PSNP
Cash transfers
Productivity
Direct impact
Yield estimates,
Transfers
Econometrics
and Statistics
National
CGE model
LEWIE
models
Local GE Impacts
Production, Incomes,
Consumption, Wages,
Market sales
etc.
National GE Impacts
Production, Incomes,
Consumption, Wages,
Trade, GDP
etc.
9. LEWIE models CGE Model
Kebele level
(8 such models)
National
15 commodities
5 factors
3 household types
69 commodities
20 factor types
20 household types
4 agro-climatic zones
Distinguishes PSNP recipients
(Public works, Direct support, non-
recipients)
Distinguishes PSNP areas
(land accounts, activity accounts,
household accounts)
Bottom up calibration from
household data
Top-down calibration from
National accounts
Static, 1-year Static, 1-year
9
Comparing models
11. 11
Econometric Estimations
• (Relates to the work presented by Dr. Alemayehu)
• Figure out impact of project on yields
– NB: other impacts will be future work
– ex: transport costs, education, health
• Empirical strategy:
– Regress yield on number of projects (Fixed effects, GMM-IV)
• Results:
– Grain yields increase by 2.8% per year (SWC projects)
– Vegetable yield increase by 12% per year (irrigation projects)
– Other crops not significantly affected
12. 12
From econometrics to simulations
• Take econometric results and use them to simulate PSNP
in an economy-wide model
• Change parameters of the model
• Here, I report results for joint simulations
– Increase in grain yields
– Increase in vegetable yields
– Cash transfer
=> Model solves for “full impact”
15. 15
Shocks we simulate
• 2.8 % increase in grain yields (from SWC)
• 12% increase in vegetable yields (from irrigation)
• 18% transfer income for PSNP recipient households
• Model simulates the impact in a single year of PSNP
16. 16
Results – Production Output
0
1
2
3
4
5
6
7
8
Lege Fate Fura Leml Kole Joro Rame Fele
% change in total output in 8 kebeles
17. 17
0
1
2
3
4
5
6
7
8
0.25 0.3 0.35 0.4 0.45 0.5 0.55
Market Integration Index
% change in total output
Economic structure shapes results
19. 19
Results at local scale
• PSNP increased production in all Kebeles
– But size of total impact differs
– Structure of the economy matters
• Positive spillovers
– Non-recipients also benefit
– Pathways: increased yields and increased demand
21. 21
Shocks we simulate
• 2.8 % increase in grain yields (from SWC)
• 12 % increase in vegetable yields (from irrigation)
• 18 % transfer income for PSNP recipient households
(=3.7% for the PSNP area)
Model simulates the impact in a single year of PSNP
22. 22
Results – National Level
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
All Agriculture Grains
% change in output
0
2
4
6
8
All PSNP areas Non-PSNP
areas
% change in Household
Income
24. 24
Conclusions
• Local impacts are positive but differ across area
– Depend critically on the structure of the local economy
– We have the LEWIE tool to analyze those impacts
• Nationwide economic impact of PSNP is far from trivial
– 0.88% real GDP growth (bounds at 0.61% – 1.22%)
• PSNP has far-reaching impacts
– Benefits from combining Protection + Production
– Stimulates supply and demand simultaneously
– Need General Equilibrium framework to reveal full benefits
• There are also long-term impacts: needs more research