UNOBSERVED HOUSEHOLD AND COMMUNITY HETEROGENEITY AND THE LABOR-MARKET IMPACT OF SCHOOLING - A CASE-STUDY FOR INDONESIA

This study presents estimates of wage functions for Indonesia without and with fixed assets to see if controls make a difference in the estimated impact of schooling on wages. It examines a parallel question about the impact of schooling on hours suplied to the paid labor market. For both wage rates and hours supplied to the paid labor market it considers relations separately for males and females. For both wage rates and hours supplied to the paid labor market, it explores the robustness of the results to sample selectivity and to the presence of random effects instead of fixed effects. For Indonesia, estimates of the impact of schooling on wage rates and on hours supplied to the paid labor market that do not control for household fixed effects are substantially misleading. But the estimated directions and magnitudes of the biases due to the failure to control for such fixed effects differ substantially between males and females, between wage rates and hours supplied to the paid labor force, and among schooling levels. For the wage rate estimates for lower schooling levels the failure to control for such fixed effects causes substantial upward biases in the estimated impact of schooling at lower schooling levels. For the hours supplied to the paid labor market estimates for both males and females the bias tends to be upward, though with lesser relative differences once again for more years of schooling. -from Authors