ISLAMS BLOODY INNARDS Religion and Political Terror, 19802000

1
ISLAMS BLOODY INNARDS?Religion and Political
Terror, 19802000

• Indra de Soysa Ragnhild Nordås
• Meeting in the Environmental Factors in Civil War
  Working Group,
• 21 September 2006

2
Outline of paper

• Background
• Culturalists claim that political outcomes are
  culturally determined
• Huntingtons claim that Islam has bloody
  innards cultural values make Islam prone to
  violence
• A Western public image of Islam as particularly
  repressive and violent
• Islam is apparently antithetical to individual
  rights and the respect for personal liberties
• Main findings
• Countries with larger shares of Catholics fare
  far worse than all others on repression
• Political and economic factors matter a whole lot
  more than religion

3
Why repress?

• Parsimonious model
• HR practices is one option for leaders to respond
  to a threat Policy decisions a function of
  strength (S) and threats (T)
• When decision-makers recognize that the
  governments strength (S) is weaker than the
  threat (T) Snt1ltTnt1, or that threat (T) is
  increasing relative to state strength (S)
  Snt1/Tnt1 lt Snt0/Tnt0, then the state will be
  willing to take action to increase its strength
  (S) or decrease the threat (T)
• Religion in this framework
• 1) Particular aspects of a religion may increase
  the level of dissent (threat)
• 2) Aspects of religion could affect responses to
  threat by legitimizing some actions over others
  (strength)

4
Why Islam?

• Claims made in the literature/public debate to
  validate the view that Islam is particularly
  repressive
• Islamic teachings do not show tolerance and
  discourage the separation of religious and
  secular life
• Muslims prefer to live under sharia law rather
  than accept liberal values
• Preference for order over chaos, which
  legitimizes the use of harsh methods against
  dissent
• Socially sanctioned right of Jihad high levels
  of dissent that often forms as violent protest
  (threat to regimes from below)

5
Main variables

• Dependent Violations of physical integrity
  rights (PIR) and the extent of political terror
  (PTS)
• Scale of reported extrajudicial killings/unlawful
  and arbitrary deprivation of life,
  disappearances, torture/inhumane and degrading
  treatment, and political imprisonment because of
  political activism and nonviolent opposition to
  government.
• Explanatory Religion
• The percentages of the major religions
  Protestant, Muslim, Catholic and Other that make
  up a society (state)
• Alternative specifications tested
• Dummy for majority Muslim
• Membership in the IOC

6
Controls

• Democracy (dummy) Polity2 scoregt6
• Oil export 1/3 of GDP (FL)
• Population size (ln)
• Economic standing
• GNI p/c in PPP terms (ln)
• Growth rate of income p/c
• British/ Socialist legal tradition
• Heterogeneity ethnic fractionalization (FL)
• Involvement in international and civil wars (and
  peaceyrs)
• Past levels of repression (LDV)
• Time trends in the data yr dummies
• Regional dummies (for robustness checks)

7
Sample / Estimations

• PIR scale estimates using
• Linear regression models (OLS) with robust SE
  (clustering on country)
• Ordered probit
• PCSE with correction for auto-correlation
• GEE (as robustness test)
• 131 countries, 1980-2000

8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
Findings

• Both Catholic and Muslim countries tend to have
  less respect for physical integrity rights than
  largely Protestant societies
• Islamic societies tend to have higher levels of
  PIR when the reference category is the share of
  Catholics
• It is the regional effects of the Middle East
  (not the share of the population Islamic) that
  helps us understand the nature of human rights
  violations
• Findings also supported with alternative
  dichotomous variables of Muslim-dominance or IOC
  membership
• Dropping oil and demo. from the model
  respectively does not change the result neither
  do alternative specifications
• The effect of Muslim societies becomes stronger
  against the share of Protestants when a dummy for
  MENA is added.

12
Substantive effects

• At cut-off point 4 (strong repression) increasing
  the share of the population by one s.d. (other
  var.s at mean)
• Muslim lowers the average predicted probability
  by 37
• Protestant reduces it by 30
• Raising income by one s.d. reduces the average
  pred. prob. by 73, (more than double the impact
  of the religious variables).
• The baseline predicted probability decreases by
• 70 going from non-democracy to democracy
• 73 by increasing income with one s.d.
• The baseline predicted probability increases by
• 413 going from peace to civil war
• 94 going from being a non-oil producer to an
  oil producer

13
Conclusion

• Islamic societies do better than majority
  Catholic ones
• These results do not change when we include a
  MENA dummy
• The MENA region has an independent effect on
  political terror net of oil wealth and democratic
  deficit
• Good news There is a lot of room for policy
  (e.g. directed at economic and political factors)
  
• Islam does not seem to be what matters