Over-education And Wage Inequality

With the further reform of the economic system and the rapid growth of our economy, the education in China has gained rapid development in recent years. But at the same time, the problems of over-education and wage inequality have come into surface. Study on the relationship between over-education and wage inequality can help us better understand the current situation and causes of the widening urban income distribution gap, and it can also help in the evaluation and design of education and income distribution policies.This paper first reviews the economic meaning of over-education and presents the recent development of research in the impact of over-education on wage and income inequality. This paper also reviews the current research on over-education in China and the advantages of quantile regression estimation.Second, in order to present the current situation of the widening wage inequality in China, we calculate the GINI index and the ratio between wages at the 9th(W9) and the 1st deciles(W1) based on the whole sample as well as three different educational groups in the year 1995 and 2002. The GINI index of the whole sample rises 18.23%. The GINI index of tertiary workers rises 19.88%. The GINI index of secondary workers rises 13.31%. The GINI index of primary workers rises 11.71%. W9/W1 of the whole sample and three educational groups all rises. This indicates that there has been an overall deterioration in wage inequality in the year 2002, compared with the year 1995. At the same time, the gap between the rich and the poor widens.Then, we estimate the incidence of over-education. Our research indicates that the incidence of over-education is different in both occupations and industries at different quantiles of the conditional wage distribution. The incidence of over-education tends to stabilize in fluctuation with the increase of the quantiles of the conditional wage distribution. In different occupations, the group of professional workers has the lowest incidence of over-education with male by 11.26% and female by 6.68%. And the group of clerical and office staff has the highest incidence of over-education with male by 18.95% and female by 20.07%. Females working in government and party organizations have the lowest incidence of over-education, 5.89% and males working in finance and insurance have the lowest incidence of over-education, 9.09%.Fourth, we employ the quantile regression estimation to estimate the return to over and adequate education of different educational groups. OLS estimation results indicate that the wage penalty of tertiary male workers is 8.74% and the wage penalty of tertiary female workers is 11.7%, which means that female workers lose more income than male workers when over-education happens. The estimation results of quantile regression shows that the pay penalty of over-education to male decreases to zero with the increase of quantiles of conditional wage distribution but the pay penalty of over-education to tertiary female workers doesn't change much and the pay penalty of over-education to secondary female workers tends to increase. This indicates that college education increase the wage inequality of workers with low income and secondary education increases the wage inequality of female workers with high income.Fifth, variables that represent individual characteristics may have non-negligible impact on individual income. So we add in the model the control variables that represent individual characteristics and estimate the returns to over-education and adequate education of different educational groups in different areas. The estimation results indicate that there is an obvious decrease in return to adequate education after we control these variables and the extent of the decease is larger than the decrease in return to over-education, which means that personal characteristics and occupation match have significant impacts on workers' wage. This is also the conclusion of the definition of over-education in terms of occupation match. Estimation results of returns to education in different areas indicate that the return to over-education and adequate education in developed areas are smaller than in moderately developed and less developed areas. It is a remarkable fact that the returns to over-education and adequate education of secondary workers in moderately and less developed areas are higher than the corresponding group in the developed areas by 3800.11% and 2200.31% respectively.At last, we make comparisons of pay penalties in the year 1995 and 2002. From the comparisons of the two-year returns to education and pay penalties, we see that the returns to education of different educational levels all increase. In the sub group of college education, pay penalties for male and female all increase with male increased by 52.66% and female increased by 34.14%. This means that after we have excluded the impact of increase of average wage as a result of economic growth, workers with college education suffered more from over-education. This also verifies the implication of GINI index. But, pay penalties of both male and female workers with secondary education decrease. This means that individual wage loss deceases when over-education happens in this group, which conflicts with the implication of GINI index. This may due to the estimation bias caused by the defect of the measurement of over-education. Except that, we cannot figure out more convincing reasons. We see from the estimation results of quantile regression that pay penalty to workers with secondary education decreases with the increase of the quantiles of the conditional wage distribution, which means that the higher the workers' wages, the less workers may suffer from over-education when over-education happens. At the same time, it also indicates that over-education widens the wage inequality of male workers with low incomes. Pay penalty to female workers with college education fluctuates a lot in the year 2002, but the basic trend is constant. In year 1995 there is an obvious trend of decrease. This means that the dispersion of wage of the high income group increases and wage inequality deteriorates at the same time. For the group with secondary education, the trend of pay penalty in 1995 and 2002 is in different directions. In 1995, there is a trend of decrease and in 2002 there is trend of increase. In combination of the analysis of tertiary workers, we see that there is an overall deterioration of wage inequality in both the female workers of the two educational levels. This means that, when over-education happens, wage loss of this proportion of individual increases as time goes on.The research in this paper can be of valuable reference to the analysis of the urban education structure and the design of education policies.