A guest post by Nils-Christian Bormann.
On a recent Daily Show episode, former FBI special agent Ali Soufan discussed Iranian internal politics with regard to its ethno-nationalist minorities and their linkages in the wider Middle Eastern & Caucasian region. Manuel Vogt, Lars-Erik Cederman and I have discussed the challenge of ethno-nationalism in the context of the Arab Spring in a new NCCR working paper. Our analysis of ethnic group conflict risk takes into consideration factors such as government inclusion of ethnic elites, a recent downgrade in power status, ethnic group size, a prior history of conflict, GDP and population size. Our model predicts that most groups in Iran have a conflict probability that is above the sample median (the sample average of all conflict probabilities is heavily influenced by actual conflicts – our model predicts an almost 100% Probability of insurgency for the Kurds within the decade from 2010-2019 – and Kurdish rebels were actually fighting the Iranian government in 2010). Interestingly enough, the Azeris – the Turkic group that Soufan is talking about the most have a much smaller conflict probability than most of Iran’s minorities (about a third). This is due to fact that in our data set Azerian elites are coded as having access to executive power in Iran and the Azeris have never experienced a civil war with the Iranian government. However, if Azeri elites were denied to the Iranian government in, for example, 2014 their 10-year cumulative probability of civil war onset would rise to 34% – a rather high probability in civil war studies. The probabilities underline that it is not ethno-nationalism per se that increase the likelihood of civil war but its meaning in the political system. Put differently, when a government can accommodate ethnic elites by offering them de facto access to executive positions, civil war risk is greatly reduced.
Soufan, however, refers explicitly to regional dynamics and ethno-national linkages of Turkic groups in the Middle East, the Caucusus and all the way to China. I therefore estimated a new model with a different sample (including the Middle East and the Caucasus but no Northern African countries) and three new dummy variables that indicate whether an ethnic group has any kin group in a neighboring country that has access to governmental power, that is excluded from executive power, or that is currently fighting a war (these data were just collected by Seraina Ruegger at ETH Zurich but are not yet published). The model (figure below) again underlines the importance of ethnic politics and its dynamic changes in explaining civil war onset. Moreover, having a kin group that is excluded from government power in the region is likely to increase conflict risk while having a kin group that is included does not (at least not significantly). Having a kin group that is involved in conflict raises the risk of civil war as well. The cases doing most of the work here will be the Kurds in Iran, Iraq, and Turkey. Indeed, if the Kurds are dropped from the sample, the excluded kin and conflict parameters turn out to be insignificant and the included kin variable raises the risk of conflict as Soufan suggested they would for Turkish kin groups.
However, neither model 1 nor 2 predict a higher risk of civil war in Iran. Arabs and Baloch along with the Kurds are still quite likely to be involved in a civil war in the next decade but Azeris and other Iranian minorities actually have a lower risk of conflict onset in these models. Model 2 does not change the dynamics in Iran very much. Indeed, the Turkish Kin connection hardly affects the conflict risk for Azeris in Iran. Even in the counterfactual case of a domestic war in which one of the Turkish kin groups of the Azeri were involved, conflict risk would not rise to levels that the Iranian leadership would have to worry about much according to the model.
All of these results have to be taken with a grain of salt. The current coding of the kin connection is solely based on ethnic affinity (shared language or religious adherence would result in a kin link) and does not capture Turkey’s geo-strategic interests. Moreover, it is also not geared towards prediction. However, it is suggestive of the fact that domestic ethnic power relations seem to play a bigger role than international linkages. Very different dynamics could open up if Iran’s army was involved in a costly conflict with Israel and opportunities for rebellion would open up. However, as long as Iranian Azeri elites have de facto executive access, a domestic uprising seems unlikely.