Transcript of Winning Hearts and Minds through Development: Evidence from a Field Experiment in Afghanistan Andrew...
- Slide 1
- Winning Hearts and Minds through Development: Evidence from a
Field Experiment in Afghanistan Andrew Beath, Fotini Christia,
Ruben Enikolopov
- Slide 2
- Motivation Internal conflicts play important role in
underdevelopment more than half of the world affected by civil war
in last 50 years Development aid used as a counterinsurgency tool
both in Iraq and Afghanistan Limited empirical evidence on aids
effectiveness in winning over the support of an embattled
population. Recent research has produced conflicting results.
Methodological Challenge: Non-random assignment of aid.
- Slide 3
- Testing the Hearts and Minds Strategy Randomized field
experiment deals with selection bias. Examine effects both at the
interim and final phase. Examine both perceptions and objective
measures. Providing basic public goods through development projects
is considered part of an effective counterinsurgency strategy. H1.
Public goods provision improves peoples economic well-being, H2.
which leads to improved attitudes towards the government, H3. and
in turns reduces the number of people willing to support the
insurgency, leading to fewer security incidents.
- Slide 4
- Preview of Results Development aid can positively effect
economic wellbeing, attitudes toward the government, and security
perceptions and conditions. Timeframe of Implementation and Initial
Violence Matter Effects on economic welfare and attitudes towards
the government stronger during program implementation, though still
observed almost a year after completion. Effect observed in secure
and insecure areas. Violence goes down during program
implementation but the effect diminishes after the flow of
resources stops. Effect observed only in areas with low levels of
initial violence.
- Slide 5
- Context: National Solidarity Program (NSP) NSP is the largest
development program in Afghanistan. Over 32,000 of Afghanistans
38,000 villages have received NSP. Sponsored by international
donors and run by the Ministry of Rural Rehabilitation and
Development. Implemented by NGOs in two main stages: Election of
Community Development Councils (CDCs) through secret-ballot
election Size of CDC proportional to size of community with equal
number of males and females. Project Selection $200 per household;
max community grant $60,000
- Slide 6
- Evaluation Covers 10 Districts across 6 Provinces Herat Ghor
Daykundi Balkh Baghlan Nangarhar
- Slide 7
- Slide 8
- Slide 9
- Slide 10
- Slide 11
- Similarity of Treatment and Control Villages Create CDCs
Treatment Villages (NSP) Select Projects Implement Projects
Projects Finished Control Villages (Non-NSP) May Oct. 2009May-Nov.
2011 Midline EstimatesEndline Estimates Baseline Survey Aug. Sep.
2007 1 st Follow- Up Survey 2 nd Follow- Up Survey The evaluation
estimates impacts by collecting data over four years in 500
villages: 250 NSP (treatment) & 250 non-NSP (control) Structure
of Evaluation and Data Collection
- Slide 12
- Slide 13
- Slide 14
- Security Events Data ISAF data on security incidents (almost
exclusively IEDs-) with exact location, date and time for the
period 01/2006-12/2011. Look at three different periods Before the
start of the program (Jan 2006 Sept 2007) Midline Period (Oct 2007-
Sept 2009) Endline Period (Oct 2009- Dec 2011)
- Slide 15
- VariableTreatment Effect Midline Treatment Effect Endline P
value for equality of coefficients N Male Respondent Expects
Household's Situation to Improve Next Year 0.052 (0.011)***
[0.012]*** 0.021 (0.019)* [0.013] 0.0358,946 Female Respondent
Expects Household Situation to Improve Next Year 0.046 (0.014)***
[0.024]* 0.039 (0.010)*** [0.023]* 0.6728,017 Summary Measure of
Objective Wellbeing 0.030 (0.015)** [0.023] 0.015 (0.017) [0.018]
0.4258,990 Economic Welfare and Governance Attitudes Summary
Measure of Governance Attitudes 0.096 (0.021)*** [0.043]** 0.056
(0.018)*** [0.031]* 0.114 8,982
- Slide 16
- Perceptions of Security and Security Incidents
VariableTreatment Effect Midline Treatment Effect Endline P value
for equality of coefficients N Security In and Around Village has
improved in Past Two Years 0.050 (0.013)** [0.022] ** 0.044
(0.015)*** [0.026]* 0.7298,962 Compared to Two Years Ago Teenage
Girls feel less Safe when Traveling to School and Socializing
-0.036 (0.016)** [0.011]*** -0.012 (0.014) (0.020) 0.2317,128
Attacks based on ISAF measure -0.016 (0.041) [0.037]** -0.021
(0.031) [0.047] 0.2251,000 Occurrence of a Security Incident -0.051
(0.041) [0.029]* 0.010 (0.037) [0.039] 0.301,000
- Slide 17
- VariableTreatment Effect Mid*Insecure Treatment Effect Mid*
Secure Treatment Effect End* Insecure Treatment Effect End* Secure
Male Respondent Expects Household's Situation to Improve Next Year
0.056 (0.021)*** [0.023]** 0.051 (0.013)*** [0.012]*** 0.043
(0.023)* [0.021]** 0.014 (0.010) [0.021] Female Respondent Expects
Household Situation to Improve Next Year 0.013 (0.024) [0.003]***
0.072 (0.015)*** [0.021]*** 0.060 (0.027)** [0.010]*** 0.046
(0.013)*** [0.027]* The Role of Initial Level of Insecurity Summary
Measure for Objective Wellbeing 0.034 [0.036] [0.011]*** 0.029
[0.017]* [0.026] -0.040 [0.040] [0.039] 0.032 [0.018]* [0.028]
Summary Measure for Governance Attitudes 0.038 (0.043) [0.031]
0.111 (0.023)*** [0.048]** 0.103 (0.037)*** [0.007]*** 0.042
(0.021)]** [0.035]
- Slide 18
- VariableTreatment Effect Mid*Insecure Treatment Effect Mid*
Secure Treatment Effect End* Insecure Treatment Effect End* Secure
Security Around Village has Improved in Past Two Years 0.012
(0.022) [0.023] 0.061 (0.016)*** [0.026]** 0.039 (0.034) [0.015]
0.045 (0.016)** [0.032] Teenage Girls Feel Less Safe Going to
School and Socializing Compared to Two Years Ago -0.002 (0.033)
[0.049] -0.044 (0.018)** [0.012]*** 0.013 (0.026) [0.014] -0.017
(0.028) [0.013]** Attacks based on ISAF measure 0.170 (0.122)
[0.155] -0.073 (0.033)** [0.028]** 0.034 (0.105) [0.132] -0.037
(0.019)* [0.030] Occurrence of a Security Incident 0.184 (0.164)
[0.167] -0.091 (0.036)** [0.034]*** 0.148 (0.095) [0.154] -0.037
(0.042) [0.029] The Role of Initial Level of Insecurity
- Slide 19
- Conclusion We find that the provision of development aid to
villages can: 1. Positively affect economic welfare of population
2. Improve attitudes towards the government 3. Improve security But
effects are conditional on initial levels of security and timeframe
of implementation of development program suggesting that
development aid: Can prevent spread of insurgency in areas with low
initial violence, but cannot contain it in areas with notable
violence. Need for continual provision of aid rather than one-shot
projects.
- Slide 20
- Specification We estimate the following OLS regression: where Y
tvi is the outcome of interest for household i in village v in the
midline (1) or endline (2), so that t {1,2}; T v is the village
treatment dummy, t is the dummy for period t, pt is the
village-pair*period fixed effect. Standard errors clustered at
village-cluster level; also account for spatial correlation(
Conley, 1999 ).