|dc.description.abstract||This dissertation includes two essays that analyze the wage gap in twelve developing countries, and another essay focuses on residential water demand in China. Chapter 1 decomposes the gender wage differentials in twelve developing countries and examines the variations of gender wage inequality with respect to different levels of education. The wage equations are extended from the Mincer earnings function and estimated for males and females separately by the OLS and Heckman selection models. We found that sample selection bias mainly exists in the female wage equations. The decomposition results suggest that all these developing countries exhibit significant discrimination effects on gender wage inequality from the OLS estimates. And only in Ghana, is discrimination estimated to be insignificant after the selection bias adjustment. The discrimination effect on the gender wage gap is actually offset by the advantages to human capital for women in China. Meanwhile, a decreasing trend in both total male/female wage inequality and discrimination effects against women is examined with increasing education.
Chapter 2 analyzes the wage differentials by both gender and public-private sector. With the multinomial logit estimation, we found that women who are married, have more children under 6 years old or have health problems are not likely to participate in the labor market, especially in private companies. The returns to education in wage equations are larger for women in both the public and private sectors in most of the countries. After the decomposition, most of the countries show a positive public-private pay gap, and the sector wage differential for women is generally larger than that for men. Meanwhile, the gender wage differential is generally smaller in the public sector, and female employees in the private sector experience more gender discrimination than those in the public sector.
Chapter 3 displays a residential water demand analysis based on panel data covering 31 provinces of China from 2004 to 2013. Two models are employed in this study: Traditional log-linear model and the EDM model. The estimates from the log-linear model show that different levels of income do not impact water price elasticities significantly but the fixed effects estimator gives a more appropriate estimate for the water demand system. With the EDM model estimations, the results reveal that the water price is elastic for residential water demand in the short run, and the partial price-supply elasticity is negative due to estimates of total elasticities.||en_US