War and Local Institutions in Sierra Leone

1
War and Local Institutions in Sierra Leone

• John Bellows, UC Berkeley
• Edward Miguel, UC Berkeley and NBER
• April 2006

2
Conflict and Economic Development

• War and conflict could be critical for
  understanding cross-country economic development
  patterns (Drèze 1999, World Bank 2003)
• Conflict may be particularly important for
  Sub-Saharan Africa, where two-thirds of countries
  experienced armed conflicts in the 1980-1990s

3
War and Local Institutions in Sierra Leone

• We examine local institutional, political and
  social outcomes in Sierra Leone
• Sierra Leone experienced a brutal civil conflict
  from 1991 to 2002
• We have nationally representative survey data on
  both conflict experiences and local institutions

4
The Sierra Leone Civil War (1991-2002)

• Fighting started in the eastern Liberian border
  but eventually affected the entire country
• Collusion between rebel Revolutionary United
  Front (RUF) and Sierra Leone Army (SLA) meant
  that most violence was directed against civilians

5
Consequences of the Sierra Leone war

• RUF attackers massacred chiefs and elders,
  destroyed public buildings, displaced civilians
• Recruits often attacked their home villages
• Civilian Defense Forces (CDF) were locally
  organized and financed militias that successfully
  fought off the RUF/SLA in some areas

6
This talk

• (1) We use chiefdom level data (N152) on local
  conflict-related violence, attacks and battles to
  estimate localized impacts of the conflict
• (2) Sierra Leone is currently undergoing
  extensive local institutional reform. I will
  discuss the resilience of existing local
  institutions (chiefdoms) in the face of these
  changes

7
Empirical approach in main analysis

• Investigate association of conflict with postwar
  socioeconomic outcomes, and with local
  institutional and social outcomes (village
  meetings, voter registration, community groups,
  religiosity) in 2004-5
• We condition on geographic controls, location of
  diamond mines, district fixed effects, prewar
  socioeconomic characteristics

8
Empirical approach in main analysis

• Investigate association of conflict with postwar
  socioeconomic outcomes, and with local
  institutional and social outcomes (village
  meetings, voter registration, community groups,
  religiosity) in 2004-5
• We condition on geographic controls, location of
  diamond mines, district fixed effects, prewar
  socioeconomic characteristics
• Caveat 1 the location of violence is not random
• Caveat 2 nation-wide effects of the war cannot
  be estimated using this identification strategy

9
(No Transcript)
10
Where was the fighting in Sierra Leone?

• Local diamond mines, higher prewar incomes are
  both positively associated with local attacks and
  battles (Table 2), greed as a driving factor
• Prewar 1989 school enrollment (5-18 year olds) is
  negatively correlated with conflict victimization
• Organization of CDFs better there?

11
Conflict intensity and local institutions

• Conflict victimization is positively associated
  with several measures of postwar political
  mobilization including attendance at community
  meetings, voter registration, community group
  membership
• But there is no significant relationship between
  either conflict measure and postwar 2004
  socioeconomic outcomes (per capita consumption,
  schooling, child nutrition)

12
(No Transcript)
13
(No Transcript)
14
Discussion

• We find no adverse effects of civil war violence
  in Sierra Leone on local living standards or
  institutions
• Institutions, norms, and organizations forged
  during war persist into the postwar period
• The rise of Community Defense Forces may be part
  of a broader trend picked up in our data

15
Discussion (2)

• More speculatively did the Sierra Leone civil
  war generate new and possibly stronger
  institutions?
• Keen (2005) claims the war increased political
  awareness and mobilization, especially for youths
• Ferme (2001, p. 228) Sierra Leoneans have
  sometimes turned social instability into a
  creative, though violent, opportunity to
  refashion themselves vis-à-vis their own
  institutions

16
Competing local institutions in Sierra Leone

• We document increased popular mobilization
  postwar, and there is continued CDF influence
• The Local Councils were revived in 2004, after a
  30 year gap. But Chiefs remain dominant in rural
  areas
• 44 of male youths work on the chiefs farm
• 81 can name the Chief, 48 a Local Councilor

17
Competing local institutions in Sierra Leone

• We document increased popular mobilization
  postwar, and there is continued CDF influence
• The Local Councils were revived in 2004, after a
  30 year gap. But Chiefs remain dominant in rural
  areas
• 44 of male youths work on the chiefs farm
• 81 can name the Chief, 48 a Local Councilor
• Community driven development (CDD) interventions
• We are conducting a randomized evaluation of a
  government CDD project across 240 villages

18
Existing research

• Recent studies have not found persistent local
  effects of war damage on population (Davis and
  Weinstein 2002, Brakman et al. 2004) or economic
  performance (Miguel and Roland 2005)
• But existing empirical work does not study the
  impact of war on institutions the theoretical
  channel most likely to have persistent economic
  growth effects (e.g., by shifting the aggregate
  production function scale parameter, A)

19
Causes of the war

• Diamond smuggling financed group leaders and arms
  purchases, and armed groups often fought over
  control of the mines
• Deep underlying grievances against the state also
  fed the violence Sierra Leone was second to
  last in the 1990 UN Human Development Index

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
Data sources

• 2005 Institutional Reform and Capacity Building
  Project (IRCBP) Survey
• 2005 No Peace Without Justice (NPWJ) Report
• 2004 Integrated Household Survey (SLIHS)
• 1989 Household Survey (SLHS)
• Geographic, transport, mining controls

24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Remaining identification issues

• Selection could migration of low types out of
  conflict areas, or lower mortality of educated
  people in conflict areas explain our (positive)
  results?

32
Remaining identification issues

• Selection could migration of low types out of
  conflict areas, or lower mortality of educated
  people in conflict areas explain our (positive)
  results?
• Results are robust to controlling for average
  chiefdom education postwar (2004), but we cannot
  completely rule out unobserved variables
  correlated with migration
• Future work will use detailed data from the 2004
  SL Census to map migration patterns from
  1990-2004
• Others?