Interpretation of estimates based on OLS(Ordinary Least Squared) method includes the concept of an average. Therefore, if the income equation is estimated by OLS, the difference in income between individuals is caused by the difference in explanatory variables, and the age effect controlled the effects of other explanatory variables is the same regardless of individual characteristics such as education level, living region, birth year etc. However, in related to empirical studies, the age effect differs depending on the level of education or birth year, and the age when income peaks varies. This implies that the age effect differs depending on individual characteristics and this motivation leads me to analyze age effect based on the quantile regression. For this study, I use workplace-based insured men with income information during more than 21 years in the national pension history data during 1988-2020. From the comparison on the age effect by income level, first, the higher income class, the higher age premium. Second, the age when income peaks is higher as the income level increases. Specifically, income peaks at age 40 for the bottom 10% of the income bracket, however, income peaks at age 50 for the upper 10% of the income bracket.
Read full abstract