Just finished my paper for the labor economics class. It was on the determinant of wage and gender discrimination on wage in the Indonesian urban areas. Some of the results are obvious. Some not-so-obvious ones:
Does education matter? The coefficient of education was not significant for male (very significant for female). My guess because the labor market in 1999 (I was using the 1999 Sakernas data) was still affected by the economic crisis. And the biggest casualties of the crisis were workers in the urban and modern sector (due to the nature of the crisis), as well as skilled male workers (because they are less flexible). This situation might have affected the data very much.
Vocational education. Everything else constant, having secondary vocational education raised hourly wage a little for both genders. But that was not the case for tertiary vocational education (DI/II/III).
Wage discrimination. A large portion of the female-male wage gap was attributed not to difference in skills, but to the 'unobserved' factors -- discrimination. About 68 percent of the gap was due to discrimination. Even after controlling for regional and sectoral characteristics, discrimination factor still accounted for 53 percent.
Selection bias. The selection bias coefficient is negative. Meaning that if those who are not working or not in the labor market - for any reasons - were working, they would receive higher wage than those who are already in. This is a bit surprising, although not impossible. If this is correct, then one possible explanation is most of those that did not work in 1999 went out of the labor market because they did not want to receive low(er) wage.