Political Violence and Underdevelopment

1
Political Violence and Underdevelopment

• Cristina Bodea
• Department of Political Science
• Michigan State University

Ibrahim Elbadawi Development Economic Research
Group World Bank Group 2008 GDN
Conference Brisbane, Australia Lunchtime
Presentation Jan 30, 1245-215
2
What does the paper do?

• Stochastic growth model
• Instrument multinomial model of political
  violence
• GMM dynamic panel regressions on economic growth
  outcomes
• Results and implications for Sub-Saharan Africa
  (SSA)

Analyze the development and economic growth
impact of organized political violence
3
Preview of main findings

• Overall effect of organized political violence
  likely higher than the direct capital destruction
  impact
• Democratization, social fractionalization and
  history important predictors of the three forms
  of conflict
• The risk of civil war and violent riots reduce
  growth
• Ethnic fractionalization has negative and direct
  effect on economic growth, mediated by democracy
• Sub-Saharan Africa disproportionately affected by
  risk of civil war widening income differential
  relative to East Asia
• Estimated cost of civil war in Sudan 23 billion

4
Context on political violence

• In 2000 30 major conflicts, of which 23 civil
  wars, with 10 civil wars in Sub-Saharan Africa
• Cost of wars in 2000 138 billion and 4 million
  deaths
• Displaced people as result of war between 1980
  and 1992 from 16 to 40 million
• Most violent regions in terms of number of battle
  deaths by decade
• - Before 1970s East Asia, South East Asia
• - 1980s Middle East and North Africa, Central
  and East Asia, Sub-Saharan Africa
• -1990s Sub-Saharan Africa

5
Growth model
6
Growth and violent conflict

• Direct impact of political violence of human,
  natural and physical capital
• General equilibrium costs of the risk violence
• - Individual and community physical security
• - Social capital
• - Change preferences
• - Shorten agents planning horizons
• - Misallocation of capital
• - Short circuited democratic politics

7
Stochastic economic growth model

• Follow Abadie and Gardeazabal (2006)
• Simplified one-country version of the AG model
• Derive the steady-state growth following onset of
  political violence
• Stochastic AK technology, with onset of political
  violence as innovations from a Poisson process,
  p(t), with parameter , and the onset
  (arrival) of political violence is assumed to
  destroy a fraction of the capital stock

• Economic growth is a complex function

8
Political violence
9
Received wisdom

• Civil war literature
• - War as an isolated phenomenon
• - War robustly associated with
• Older literature on violence not distinguish
  among types of violence and different causal
  factors

Low and stagnating income High dependence on
natural resources Large and highly dispersed
populations Rough geographical terrain
10
Our take on political conflict

• Subsume the study of civil war into the larger
  phenomenon of organized political violence
• Forms of violent contestation of political power
• - Civil war
• - Coup detat
• - Violent riots / demonstrations
• Reassess the role of grievance factors
• - Lack of democratically responsive political
  regimes
• - Social fractionalization
• Multinomial model

11
Democracy
Dark orange full autocracy Red partial
autocracy Light blue factional partial
democracy Green Non-factional partial
democracy Dark blue full democracy. Based on
Executive Recruitment (EXREC) and Competitiveness
of Political Participation (PARCOMP) variables in
the Polity IV data set. Table is from Goldstone
et al. 2005.
12
Social fractionalization ethnicity and language
13
Multinomial model (1)
The sample has 125 countries and go from 1950 to
1999. The table shows coefficients (standard
errors in parentheses below coefficients) from a
multinomial logit regression in which the
reference outcome is periods with no new civil
war onsets, no riots and no coups.
14
Multinomial model (2)
15
Multinomial predictions (1)
Goodness of fit of the violence model
Outcome
predicted probabilities versus actual outcomes
16
Multinomial predictions (2)
17
Civil war and human development (1)
18
Civil war and human development (2)
19
GMM estimations
20
GMM and instrumented political violence

• GMM Economic outcomes (growth and income per
  capita) endogenous to economic and political
  institutions
• Instrumented political violence from the
  multinomial model since Violence and per capita
  growth endogenous
• Our model

21
GMM dynamic panel estimation

• 68 countries 15 from Sub-Saharan Africa and 37
  from other developing regions
• Six 5-year non-overlapping averages spanning the
  period 1970 until 1999
• GMM implemented here uses lagged value of the
  dependent and independent variables as instrument
  combines regressions in differences with the
  regressions in levels
• Under assumed moment conditions, system GMM
  accounts for the problems of endogeneity and
  unobserved country effects

22
GMM estimation results (1)

• Dependent variable Growth of real GDP per capita
• Standard control variables significant
  expected sign
• Initial GDP per Capita Log
• Initial GDP per Capita Cyclical Component
• Inflation Log of Inflation 100
• Government Expenditures / GDP Log
• Human Capital Investment Secondary enrollment,
  in logs
• Governance International Country Risk guide
• Trade openness Trade volume / GDP, in logs

23
GMM estimation results (2)
24
Growth deceleration effect of risk of violence
25
Growth elasticity for risk of civil war and riots
26
Impact of civil war (1)
East Asia-Sub-Saharan Africa Income Gap (log
y(EA)/y(SSA))
27
Impact of civil war (2)

The Costs of Civil War in Sudan (as a ratio to
initial GDP per capita in 1975-79)
for t0, and
28
Impact of civil war (3)

The Forgone Income Per Capita Due to the Sudanese
Civil War (1983-2002)
29
Conclusion

• General equilibrium growth effects of political
  violence
• Instrument for organized political violence
  Multinomial analysis of coups, violent riots and
  civil war
• Risk of violent riots and civil war reduces
  economic growth
• Ethnic fractionalization reduces growth but is
  mediated by non-factional democracy
• Sub-Saharan Africa future risks and management
  of ethnic diversity

30
Backup slides
31
Multinomial predictions (3)
Manifestations of organized political violence by
regions
32
Stochastic economic growth model

• Follow Abadie and Gardeazabal (2006)
• Simplified one-country version of the AG model
• Derive the steady-state growth following onset of
  political violence
• Stochastic AK technology, with onset of political
  violence as innovations from a Poisson process,
  p(t), with parameter

where
33
How does violence affect growth?

• Direct destruction of capital
• Disruption and diversion channels
• Dis-saving channel

34
More on how violence affects growth

• Portfolio-substitution channel

35
Growth model Derivation
Now agents choose c(t) to maximize lifetime
discounted utility, subject to the law of motion
governing capital accumulation
36
Net marginal growth effect of violence