What Makes and Breaks Trust: A Logistic Regression Approach to the Determinants of Social Trust in Kurdistan

Abstract

Social trust felt by individuals in a country, towards the overall society, is deemed as a key direct and indirect driver of sustainable economic growth and development in the existing trust literature. Although encouraging social trust can constitute an indispensable tool for developing country governments to utilize for promoting natural growth, a challenge is posed to this tool due to the lack of scholarly consensus on the systematic determinants of social trust. Three competing hypotheses emerge in the literature to explain what primarily drives social trust: heterogeneity hypothesis, which advocates for ethno-racial heterogeneity to be the most prevalent determinant; equality hypothesis, which claims economic inequality to be the main driver, and fairness hypothesis, which conversely points to institutional fairness to have the greatest explanatory power over social trust. Additionally, all three branches of the literature generally show minorities to be the least socially-trusting groups per country. In order to investigate which of the hypotheses pins down the key systematic determinants of social trust, especially in the context of mistrusting minorities, this paper analyzes the roots of social trust for an ethnic minority in a cross-country context: the Kurds. Kurds allow for maximizing variation in macro-level variables by virtue of existing in all of the sample countries of Turkey, Iran, and Iraq and over time, while controlling for any cultural component of social trust.

I employ a logistic regression approach in this study on panel data to test whether social trust levels of Kurds, as self-reported in the World Values Survey (WVS), can be explained by ethnic heterogeneity (proxied by the Historical Index of Ethnic Fractionalization), economic inequality (proxied by the Gini coefficient), or institutional fairness (proxied by ICRG valuations for chosen institutions). Marginal effects of logit models uncover the fairness hypothesis to have the greatest explanatory power for Kurdish social trust in the sample countries for the present WVS years, while the equality hypothesis emerges as a key determinant of social trust for the general population (but not for Kurds in particular). Surprisingly, contrary to the heterogeneity hypothesis, more ethnic heterogeneity appears to increase the odds of social trust for Kurds and for the overall population. For government intervention, my decomposition of institutional fairness into its components marks bureaucracy quality, corruption, and government stability to be the most exigent institutions to address to encourage more trust among discriminated ethnic minorities like Kurds, while economic inequality emerges as a close second in importance to boost the social trust of the overall population.