This study explores the impact of sample selection bias on estimates of gender wage gaps among young workers in South Korea. It uses data from the Korean Labor and Income Panel Study (KLIPS) and applies two wage imputation methods to correct for selection into full-time, regular employment. Specifically, it estimates the wage offers of those not in full employment using wages from proximate years and a semi-parametric reweighting method. The findings show that ignoring sample selection bias causes wage offers to be overestimated, particularly for women, as those with lower wage offers are more likely to opt out full employment. Consequently, selection adjustments increase wage gaps in both 2001 and 2017. Positive selection weakens for women during this period which is associated with changes in education and marriage. Therefore, adjustments increase gaps more in 2001 than in 2017, leading to larger decreases in the wage gaps over time. This implies that the relative progress in labor market opportunities for women is larger than previously estimated. Finally, unadjusted gaps suggest gender disparities were largest in low-paid positions in 2001 but high-paid positions 2017, whereas adjusted gaps suggest they were largest in medium-paid positions in both years. Overall, the findings highlight the importance of accounting for sample selection bias when assessing gender wage disparities.
扫码关注我们
求助内容:
应助结果提醒方式:
