NONAGRICULTURAL UNEMPLOYMENT IN BEIJING: A MULTILEVEL ANALYSIS.

In this paper, we undertake a multilevel investigation of nonagricultural unemployment in Beijing. We use data for the employed and unemployed resident population of Beijing from the 1995 1 Percent Population Sample Survey of China. The dependent variable is the person's unemployment status. The individual level independent variables are the worker's sex, age, his or her years of education, and whether he or she is a temporary (floating) resident of Beijing. The macro-level characteristics are the area's level of labor force diversity, and its degree of concentration of persons who have attended college. The higher the labor force diversity and the higher the concentration of graduates, the less the likelihood of unemployment. We use a multilevel model, specifically a hierarchical generalized linear model, to assess the likelihood of nonagricultural unemployment of Beijing residents. We determine the extent to which the human capital characteristics of the persons themselves, as well as the contextual characteristics of their residential regions, influence their likelihood of being unemployed. We find that both micro and macro factors are important in predicting the probability of being unemployed among Beijing residents. We address these issues and draw out their implications. [ABSTRACT FROM AUTHOR]