Agricultural transformation and value

chain development:

Lessons from Randomized Control

Trials

2

3

The Problem

Understand

market failures

Develop

innovations

Use frontier

knowledge from

economics,

political science

and psychology

Impact

evaluations

(Randomized

control trials)

Product design

tests

Operational

research

Test in multiple

contexts

Learn when to do

what

Conferences, global

and local

Workshops with policy

makers and

practitioners

Policy memos and

practitioner briefs

Direct

implementation

Hands-on

technical

assistance

Practitioners’

toolkits

Our solution

• In our evaluations we only use randomized control trials (RCTs)

– Random samples

– Randomized groups: treatment, control groups

– “Treatments” (e.g. different discounts on fertilizer, different microcredit products)

• Why use RCTs?

– Reduces selection bias

– Measures what we want to measure and isolates “disturbing” factors

– Most reliable method of evaluation

So why are we interested in agricultural technology?

How is agriculture related to poverty?

5

Over 450 projects in 51

countries

6

7

IPA’s comparative advantage

• Research for impact: We make research results actionable.

• Experts in evaluation: We are the largest implementer of randomized

evaluations in the development field.

• Deep local presence: Allows us to cultivate deep relationships needed

for policy and research oversight.

• Long term partnerships: We bring together researchers and decision

makers to create and utilize innovative solutions that are relevant to the

needs of the poor and based on evidence.

8

Our 10 years have taught us …

• Need local research

infrastructure.

• Providing direct services can

conflict with our role.

• Influencing policy requires

more than communications.

• Requires deep relationships

and local understanding and

context.

• Need to replicate results to

different contexts and

understand operational

challenges.

9

Our Impact

• Large-scale influence

• Evaluation partners make

decisions based on study

results

• Other organizations adapt

evidence

• Global debates are

influenced

• Evidence-based decision

making culture or institutions

10

Over 40 million children dewormed

Randomized Controlled Trials

11

Agricultural Transformation:

Seeking Solutions

12

Agricultural yields are

higher in some regions

than in others.

Sub-Saharan Africa

almost stagnating.

World Development Report (2008)

Need for Evidence… Value

Chain Breakdown?

13

The Organic Business Guide

A Well-Timed Nudge

• Increasing fertilizer use: are high-cost subsidy

programs the only way to go?

14

A Well-Timed Nudge

• IPA Researchers (Duflo et al, 2010) studied the

effect of the Savings and Fertilizer Initiative

(SAFI) in Western Kenya.

15

A Well-Timed Nudge

• Enabling

farmers to

prepay for

fertilizer when

they had cash

on hand was

effective in

promoting

fertilizer

adoption.

16

Impact of Friends

• Impaction Technology Adoption: can social

networks work?

• Working with the Ministry of Agriculture and Food

Security IPA Researches (Beaman, et al) helped

identify seed farmers (local farmers trained by

Agricultural Extension Development Officers)

based upon their connections in the community

17

Impact of Friends

• Significant adoption of new agricultural technology (pit planting) in seasons two and three in network-targeted villages (61.4-79.5%) verse comparison villages (4.4%).

18

Fingerprinting to Reduce Risky

Borrowing

• If we see something that

works, why not implement

that program fully?

19

Fingerprinting to Reduce Risky

BorrowingA 2008 study (Yang, et al, 2012) in Malawi found that

fingerprinting identification for microloans caused high-

risk borrowers to take out smaller loans and improve their

repayment behavior (lowered default)

20

References for Studies Cited

• A Gentle Nudge: – Policy Brief: http://www.povertyactionlab.org/publication/well-

timed-nudge

– Paper: http://economics.mit.edu/files/6170

• Impact of Social Networks:– Overview: http://www.povertyactionlab.org/evaluation/making-

networks-work-policy-evidence-agricultural-technology-adoption-malawi

• Fingerprinting to Reduce Risky Borrowing:– Policy Brief:

http://www.povertyactionlab.org/publication/fingerprinting-reduce-risky-borrowing

– Paper: http://sites.lsa.umich.edu/deanyang/wp-content/uploads/sites/205/2014/12/gine-goldberg-yang-fingerprinting.pdf

21

More to come

• IPA is continuously conducting new evaluations and studies that add

to our knowledge about agricultural development

• Our latest work and ongoing research projects can be found online:

22

Agricultural Technology

Adoption Initiative (ATAI)

• Our partner organizations J-PAL and CEGA are conducting

a range of studies to identify barriers to technology

adoption and potential solutions

23