Determinants of Earnings and Selection into Labor Market by Controlling Selection Bias in Bangladesh

Abstract

This paper uses the data of Household Income and Expenditure Survey (HIES) 2005 to estimate the determinants of earnings and selection into labor market in Bangladesh. We deal with the selection bias in earnings by using a maximum likelihood system of equations, and a multinomial selection approach is used modeling for selection into the labor market. By instrumenting years of schooling in both the multinomial selection approach and the earnings equations, we deal with reverse causality between educational attainment and earnings. We find that the estimated parameters of the earnings equation under multinomial selection approach differs from ordinary least square (OLS) estimates and a binomial selection procedure. The estimated parameters that vary the most with those related variables have the strongest impact on multinomial selection into the different labor-market statuses. We also find that workers with higher educational attainment are more likely to search out a salaried employees’ job, which non-salaried work is as another to inactivity.