Fearon and Laitin, Sons of the Soil, page 1 Sons of the Soil, Immigrants and Civil War James D. Fearon David D. Laitin Stanford University This is a draft chapter of a book in progress. It assumes the reader is acquainted with the cross-sectional results and theory of earlier chapters. These results emphasize the importance of country poverty and slow economic growth as predictors of civil war. These are country-level variables. The only group-level variables that consistently come out as significant predictors are those associated with the geographical concentration of the group population and the inaccessibility of its terrain. Meanwhile, group grievances, such as economic inequality or cultural suppression, are poor predictors of civil war. In light of these findings, we developed a theoretical model of ethnic rebellion that was motivated by propitious conditions for insurgency rather than the existence of grievances. One purpose of this chapter is to make sense of the group concentration finding that is more consistent with an insurgency than a grievance model. The current version was redrafted on October 24, 2001. This is not for citation, but comments are welcome. Populations that are concentrated regionally, who populate not only cities but the surrounding rural areas, and who constitute the predominant population in their region, are far more likely to be engaged in sustained rebellion against the state than dispersed or urban-based groups that have no regional base in which they are the predominant group demographically. For purposes of this chapter, groups whose ancestors lived in the region in which they now live, and who constitute the predominant population of that region, have a regional base. When groups that have a regional base face demographic pressure through internal migration, they become likely candidates for rebellion. We call these groups “sons of the soil” (Weiner 1978). Meanwhile, those minority groups whose ancestors arrived in the country since the nineteenth century will be called immigrants. Immigrants, we shall see, are not likely candidates for rebellion. Nor are internal migrants who settle in the regions populated by sons of the soil. Our state-decade dataset includes all 124 civil wars fought throughout the world since 1945. Sixteen of them have been motivated, at least in part, by sons of soil insurgents. Possible cases of sons of soil insurgencies in this dataset include the Chakma peoples in the Chitttagong Hills of Bangladesh, the Nagas and other “tribals” in Northeast India, the Moros in the Philippines, the Tamils in the North and East in Sri Lanka, the Uighurs in Xinjiang province in China, a variety of ethnic minorities in Burma, the Sindhis against the Mohajirs around Karachi in Pakistan, the Bougainvilleans in Papua New Guinea, both the West Papuans and Achenese in Indonesia, and the Tuaregs in Mali. The estimated median and mean durations for sons of soil are 29 and 43 years, as compared to 6.2 and 9.2 for all the civil wars in the dataset.1 Therefore, these wars, while not ubiquitous, are certainly dangerous. This paper will explain why sons of the soil, but not the immigrants who settle in regions populated by ethnic others, are likely to be engaged in rebellion against the state. We will also suggest why these rebellions go on for so long. 1 . See James Fearon “Why do Some Civil Wars Go on for a Very Long Time?” Paper presented at the World Bank, DECRG-sponsored conference, “Civil Wars and Post-Conflict Transition” at the University of California, Irvine, May 18-20, 2001. To avoid right censoring of continuing civil wars, Fearon used maximum likelihood to fit a Weibull distribution to the data, and then used the estimated parameters to produce estimates of median and mean duration.Fearon and Laitin, Sons of the Soil, page 2 To give a clearer picture of the relationship of sons of soil to rebellion, we rely on our enhanced version of the Minorities at Risk (MAR) data set and the complementary narratives provided at the project’s web site.2 The MAR(SU) dataset allows us to portray group residence and its propensity to rebellion in a variety of ways. First, it distinguishes groups that migrated to their present country of residence since 1800 (n=50), from those groups whose ancestors lived in their current population centers pre-1800 (n=298). A second way to look at the relationship of residence and rebellion is to ask the degree to which the group is concentrated in a single region. We ask if there is a spatially contiguous region, larger than an urban area, that is part of the country, in which a substantial fraction (25 percent or more) of the group resides and in which the group constitutes the predominant proportion of the population. If the answer is yes, we consider the group to have a “regional base.”3 A third way to examine the relationship of residence and rebellion is to take those groups that have a regional base, and separate out those that faced, according to MAR coders, competition for vacant land in the 1980s. If there were such competition, we categorize the concentrated group as “sons of the soil.” Table 1 provides comparative means on maximum rebellion scores for groups differentiated by the time in which they arrived in the country and whether they have a regional base. In a regression with maximum rebellion of the group since 1980 as the dependent variable, with country wealth controlled for, each of these variables is independently significant (See Table 2). Long-term connection with the land, high levels of geographic concentration, and competition with other groups for vacant land within the concentrated areas are all are linked to a propensity to rebel. Table 1 Residence and Rebellion Mean Maximum Rebellion Scores by Category of Residence YES NO Has the Group Been in the Country Since 1800? 2.9 (n=248) 1.0 (n=50) Does the Group have a Regional Base? 2.9 (n=276) 1.1 (n=123) 2 . The unit of analysis in the MAR dataset, originally developed by Ted Gurr and associates at the University of Maryland, is group/country. For the original data set and the commentaries on particular groups, see http://www.bsos.umd.edu/cidcm/mar/. We retain the MAR’s 8-point ordinal scale of group rebellion, going from 0=”None Reported” to 7=”Protracted Civil War.” Our criterion for “significant rebellion” is having a score on the MAR rebellion index of at least “4” for