Mechanism Experiments, presented by Jan Duchoslav (IFPRI Malawi), 12 Feb 2020: BRACC resilience learning event, 11-12 Feb 2020, Lilongwe, Malawi

1. MECHANISM EXPERIMENTS
Existing and proposed mechanical experiments in BRACC
Jan Duchoslav
Malawi Strategy Support Program
International Food Policy Research Institute
Lilongwe | 12 February 2020

2. Components of BRACC IE
 WHAT? Was there an impact? How big was it?
 Randomized Control Trial
 HOW? How was the impact achieved?
 Qualitative investigations
 Ask people why they behave in a certain way
 Mechanism experiments
 Manipulate aspects of an intervention to verify/quantify particular mechanisms in the theory of change

3. Mechanism experiments in BRACC

4. Mechanism experiments in BRACC

5. Mechanism experiments in BRACC
 In place
 Testing the demand for funeral and health insurance
 Proposed
 Testing the demand for weather index insurance
 Does predictability matter for humanitarian assistance?

6. Demand for health and funeral insurance
0
50
100
150
200
250
300
350
400
450
1200 900 600 300 0
Coupon value
Offered
Purchased

7. Demand for health and funeral insurance
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

8. Demand for health and funeral insurance
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

9. Demand for health and funeral insurance
[CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE]
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

10. Demand for health and funeral insurance
[CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE]
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

11. Demand for health and funeral insurance
[CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE]
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

12. Demand for health and funeral insurance
[CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE]
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

13. Demand for health and funeral insurance
[CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE] [CELLRANGE]
0
50
100
150
200
250
300
350
400
450
0 30 600 900 1200
Price of annual insurance policy
Offered
Purchased

14. Proposed mechanism experiments
 Testing the demand for weather index insurance
 Same idea as funeral and health insurance
 Does predictability matter for humanitarian assistance?
 Use UBR to preselect HHs eligible for transfers should the need arise
 Randomly split intervention sample into two
 One half: Households find out just before transfers begin who was pre-selected (status quo)
 One half: Households are told in advance who was preselected
 Effects on:
 Consumption smoothing
 Food security
 On-farm and business investment
 Savings

15. Other possible mechanism experiments
 Implementation timelines
 Timing of insurance premium payments
 Information spread
 Farmer field schools vs lead farmers
 Framing
 Technology adoption in return for ‘incentives’ vs because of its benefits
 Watershed management in return for ‘incentives’ vs because of its benefits
 Pitching
 Health insurance in its own right vs compared to funeral insurance
 Different explanations of weather index insurance