• Recap: ethnic diversity unrelated to conflict?
• How might ethnicity be related to conflict?
• How to measure ethnicity?

Fearon and Laitin (2003) find no relationship between ethnic diversity and conflict…

Correlates of Civil War:

\(1.\) Ethnic diversity: no relationship - measured as ELF

\(2.\) Linguistic/religious discrimination: no relationship

\(3.\) GDP per capita: negative relationship

\(4.\) Population size: positive relationship

\(5.\) Newly Independent: positive relationship

\(6.\) Mountainous terrain: positive relationship

Implications: Theory

• Suggests that motive matters less than opportunity in civil wars
• Ethnic diversity as such does not induce civil war

But…

• Theory suggests ethnic diversity produces conflict under specific conditions?
  • ethnic status differences, ethnicity more salient, more resources at stake, ethnic extremists
• Does Fearon and Laitin’s measurement address these points?

“Ethno-linguistic fractionalization”: “ELF”

\[ELF = 1 - \sum_{i=1}^{n} s_i^2\]

Where \(s_i\) is fraction of population for group \(i \in \lbrace 1 \ldots n \rbrace\)

Interpret as: probability that any two random individuals belong to different ethnic groups

ELF treats these two situations as identical

• Country A: Group 1 50%, Group 2 50%
  • \(ELF = 0.5\)
• Country B: Group 1 66.6%, Group 2 16.6%, Group 3 16.6%
  • \(ELF = 0.5\)

• Ignores where ethnic groups are in space
  • mixed? living in different regions?
• Ignores the content of the ethnic boundary
  • are there power disparities? institutional separation?
• Permits only one dimension of ethnicity
  • is this dimension activated or even operative?
• Assumes stability over time

50–75% of civil wars in the post 1945 period involve ethnicity

• ethnic civil wars are longer; more deadly (Doyle and Sambanis 2006)

If ethnic diversity as such does not lead to civil war, why do civil wars involving ethnicity start?

Address why ethnic civil wars occur. Measurement differs in key ways from Fearon and Laitin:

Measuring Conflict:

• Examine only conflicts coded as “ethnonationalist”
• based on ethnic aims of armed groups, structure of alliances, recruitment patterns of armed groups

Measuring Ethnicity:

• move away from “diversity” as such
• Ethnic Power Relations database
  • data on \(>800\) “politically relevant ethnic groups”: activated identities
  • indicates level of group access to power at national level: power/status disparity
  • geographic location of ethnic population within state
  • politically relevant groups and power/status change over time

Research Question:

comparing ethnic groups (not countries), similar in terms of group size and country size, GDP per capita, and past conflict…

\(1.\) Are ethnic groups that are excluded from political power more likely to be involved in civil war?

\(2.\) Are ethnic groups that experience a fall from political power more likely to be involved in civil war?

Research Question:

\(1.\) Are ethnic groups that are excluded from political power more likely to be involved in civil war?: YES

\(2.\) Are ethnic groups that experience a fall from political power more likely to be involved in civil war?: YES

In line with Horowitz/Sambanis and Shayo: shifts in ethnic group power and status lead to ethnic civil war.

In another paper, Cederman, Weidmann, and Gleditsch ask

Research Question: Do economic inequalities between groups drive ethnic civil war?

\(1.\) Are groups that experience greater economic inequality more likely to be involved in civil war?

\(2.\) Are groups that experience both political exclusion and economic inequality more likely to be involved in civil war?

\(3.\) Are groups that are particularly wealthy or particularly poor relative to other groups more likely to be involved in civil war?

\(\mathrm{Inequality} = [log(g/G)]^2\)

where \(g\) is is the GDP per capita of the group; \(G\) is the average GDP per capita of all groups.

Research Question:

\(1.\) Are groups that experience greater economic inequality more likely to be involved in civil war?: YES

\(2.\) Are groups that experience both political exclusion and economic inequality more likely to be involved in civil war? YES

\(3.\) Are groups that are particularly wealthy or particularly poor relative to other groups more likely to be involved in civil war? YES

Cederman, Gleditsch, and Hug investigate:

Research Question:

\(1.\) Do recent elections induce ethnic civil war?

• On average, more likely to be more ethnic civil war after elections, but not significant

\(2.\) Do recent competitive/non-competitive elections induce ethnic civil war?

• On average, ethnic war more likely to follow competitive elections. Only a significant difference for wars over territory (secession/autonomy)

1. ELF has limitations in matching our theories of ethnic conflict
2. Ethnic diversity as such does not produce conflict
3. Political mobilization of ethnic groups can produce conflict when…
   • there are power / status differences between groups
   • elections are more recent (maybe)

Evidence of psychological (maybe instrumental?) logic leading to ethnic civil war.