Territorial arrangements and ethnic conflict management: The paradox that isn’t
系统比较了十个数据集对领土自治的测量方式，发现不同指标会导致关于自治与族群内战关系的统计结果截然不同，因此所谓“领土自治悖论”被夸大了。
• The use of territorial self-governance seems to decrease the risk of ethnic civil war in some cases and to increase it in others. • We ask whether empirical indicators influence statistical results on the link between self-governance and ethnic civil war. • We review how ten datasets capture elements of self-rule, shared rule and the legal codification of self-governance. • Binary time-series-cross-section analyses reveal that different indicators yield different results. • We conclude that the notion of a ‘paradox’ of territorial self-governance is exaggerated. Ethnic civil war, the most common type of war in the 21st century, is one of the biggest challenges for development practitioners and scholars. Like other types of armed conflict, it impedes countries’ economic, social and political development, and there is no consensus on how ‘best’ to solve it. Territorial self-governance has received much attention in efforts to reduce the risk of ethnic civil war, but the academic and policy debates over its effects remain inconclusive. This has reinforced the notion that territorial self-governance is a ‘paradoxical’ institution, which either increases or mitigates the risk of ethnic civil war. In this article, we argue that claims of a ‘paradox’ of territorial self-governance are exaggerated, as they stem from differences in empirical operationalization. We present a systematic overview of the underlying definitions, geographic and temporal scope of quantitative indicators from ten datasets, and compare how they capture aspects of self-rule, shared rule and their legal codification. Using a series of binary time-series-cross-section analyses, we illustrate that different measures of territorial arrangements lead to different results, both regarding the significance and direction of statistical effects. Our findings highlight the need to pay greater attention to the deceptively simple yet empirically fundamental question of which data are being used and why.