Ols Regression Estimates That The Return Of Schooling

1715 Words7 Pages
OLS regression estimates that the return to schooling is 9.06% for men and 10.62% for women. The first stage results, reported in the second column of Table1, indicate that the RoSLA to 16 in the UK increased the average number of years of education by 0.12 for men and 0.74 for women. Our second stage IV shows that an additional year of education increases returns to the log of wages of 13.21% and 22.39% for men and women respectively. Once again, our IV estimates are higher than OLS which could be to a few reasons such as measurement error, causing OLS to suffer downward bias, and LATE. In the case of both men and women there are many observable factors such as ability and skills on the impact of the years of schooling which can also affect the individual’s earnings. Moreover, IV captures the average return of an extra year of education induced of the compliers between the two reforms, while OLS captures average return across all people and all years. IV is thus higher due to higher marginal return to education at relatively low education level. Whilst I didn’t run regressions for the effects of the first reform on women, previous literature has shown that women typically had returns of around 8.3% (OLS) and 9.3% (IV) from the first reform (Dearden, 1998). From my results, it is evident that women experience significantly greater returns to education from the second reform than men did in both the first and second reforms. Men actually experience a decline in the returns
Open Document