Abstract
Using annual urban household survey data from 6 provinces in different regions of China, we analyze the rapid increase in earning inequality of urban China from 1988 to 2003.
We describe overall and residual inequality trends and use quantile regression to address the
relationship between education and earning inequality. We find that returns to education are
higher for the low earning individuals in the first half of this period conditional on their
observable characteristics. This suggests that education has a negative impact upon
within-group earning inequality. However, the situation is reversed during the recent half
period. Using the Quantile-JMP decomposition technique we partition the observed
distribution of earnings into ‘price’ components (earning coefficients) and ‘quantity’
components (labor force composition) and calculate, through simulation, the impact of
education on changes in overall earning dispersion. The decomposition shows that the rise of
earning dispersion between 1988 and 2003 is almost entirely accounted for by prices rather
than quantities, and it attributes a large proportion to the overall effect of education. From
1988 to1997 education serves as the equalizing force to decrease earning inequality but it is
the primary driving force which increases the earning inequality between 1997 and 2003. The
empirical analysis also reveals that the overall effect of college and above education category
on the growth of earning inequality is the most pronounced one.
Key words: Return to education; earning inequality; quantile regression
