When Measure Matters : Coresidency, Truncation Bias, and Intergenerational Mobility in Developing Countries

Emran, M. Shahe ; Greene, William ; Shilpi, Forhad

World Bank, Washington, DC

关键词: CORRELATIONS;    DEVELOPMENT RESEARCH;    VARIABILITY;    STUDY;    EXPLORATION;   

DOI  :  10.1596/1813-9450-7608
RP-ID  :  WPS7608

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

Biases from truncation caused bycoresidency restriction have been a challenge for researchon intergenerational mobility. Estimates ofintergenerational schooling persistence from two data setsshow that the intergenerational regression coefficient, themost widely used measure, is severely biased downward incoresident samples. But the bias in intergenerationalcorrelation is much smaller, and is less sensitive to thecoresidency rate. The paper provides explanations for theseresults. Comparison of intergenerational mobility based onthe intergenerational regression coefficient acrosscountries, gender, and over time can be misleading. Muchprogress on intergenerational mobility in developingcountries can be made with the available data by focusing onintergenerational correlation.