Explaining Autonomy Success and Failure: An Identity-Based Approach

Abstract

This chapter deals with the central explanandum of the study and finds empirical evidence for the high variance in consolidation outcomes of post-conflict autonomy arrangements. To explain this variance, the chapter builds on Social Identity Theory and develops two multi-causal models, which explain autonomy consolidation as a successful process of ethnic recognition that proceeds via elite acceptance for institutional reforms, elite cooperation efforts, and rapprochement at the societal level. Since the initiation and maintenance of this process is obviously highly context dependent, the chapter identifies a number of potentially necessary conditions in a second step. From an identity-theoretical point of view, a combination of a peaceful and non-secessionist conflict strategy, a high scope of self-rule, weak social and economic inequalities, democracy, and a high degree of internationalization is favorable for post-conflict autonomy consolidation, whereas the absence of these conditions increases the chances of autonomy failure and continued separatist violence.

Appendix

Horizontal Inequalities

POL_HU and CUL_HU are expected to be no longer present in consolidation phases, as they are, at least de jure, abolished with the establishment of territorial self-government as a managing institution. However, both inequality dimensions are considered here in order to conduct a complete test of the interdependency and multidimensionality of intergroup asymmetries. The following indicators are used to operationalize the four dimensions (Table 2.5).

Table 2.5 Indicators for inequality dimensions

From an empirical point of view, inequality implies a significant deviation from the national average on the respective dimension. The independent variables are coded dichotomously (discriminated against/not-discriminated against) on group-year level on the basis of the respective indicators for all relevant groups in all states with more than 500,000 inhabitants. The dataset consists of 1209 ethnic groups in 156 countries.

A group is relevant if its interests are represented by at least one actor in the national political arena, or if members of the group are systematically discriminated against.²⁷ All titular nations are coded as “not-discriminated against”. In most cases, only one indicator per dimension is taken into account. For example, if a significantly lower enrollment rate indicates social inequality, the group´s health status is not further reviewed. The dependent variable for the regression analysis is the conflict intensity variable of the Heidelberg conflict approach.²⁸

For the political dimension, data from the EPR dataset is used.²⁹ Politically excluded are groups without executive influence opportunities (powerless) and those who are discriminated against (discriminated). The economic, social, and cultural dimension is coded on the basis of qualitative data from the World Directory of Minorities and Indigenous Peoples, the International Work Group on Indigenous Affairs, the United Nations as well as the NGO Saferworld.³⁰ In addition, data for the economic and cultural dimension are synchronized with the Minorities at Risk (MAR) and the All Minorities at Risk (AMAR) datasets.³¹ A dummy variable for cultural inequality is coded on the basis of the MAR variables CULPO1 (restrictions on the exercise of a religion) and CULPO2 (restrictions on the exercise of a language) and the value labels “informally restricted”, “somewhat restricted”, and “sharply restricted”.