Ethnic switching: Longitudinal evidence on prevalence, correlates, and implications for measuring ethnic segregation
利用印尼、美国和印度的个体面板数据，发现族群身份并非固定不变，族群转换普遍存在且可预测，并会严重高估官方族群隔离数据，对研究族群多样性和隔离有重要启示。
Ethnic divisions are prominent in many countries and associated with various societal ills, ranging from discrimination to violence and intra-state conflict. Commonly seen as exogenous, individuals' ethnicity is not necessarily fixed or unmalleable, however. We first use individual-level panel data for Indonesia, the U.S. and India, to document the prevalence of generic, undirected intra-individual ethnicity change. Second, we focus on potential correlates of ethnic switching and find that ethnic switching is predictable and follows logical patterns. By zooming in on individuals’ choices to switch to a specific ethnicity, we similarly find strong evidence that such directed ethnic switching occurs systematically and is guided by deliberate individual decision making. Third, considering potential implications of endogenous ethnicity, we find that ethnic switching can severely inflate official numbers on ethnic segregation, specifically on marital endogamy. We conclude that ethnic switching is real and highly relevant for the study of ethnic diversity and segregation.