| Abstract: |
Estimating the rate of return to a university degree has always been difficult
due to the problem of omitted variable biases. Benefiting from a special
feature of the University Admission system in China, which has clear cutoffs
for university entry, combined with a unique data set with information on
individual National College Entrance Examination (NCEE) scores, we estimate
the Local Average Treatment Effects (LATE) of university education based on a
Regression Discontinuity design. To the best of our knowledge, this is the
first study to use RD design to estimate the causal effect of a university
education on earnings. Our results show that the rates of return to 4-year
university education relative to 3-year college education are 40 and 60 per
cent for the compliers in the male and female samples, respectively, which are
much larger than the simple OLS estimations revealed in previous literature.
Since in our sample a large proportion of individuals are compliers (45 per
cent for males and 48 per cent for females), the LATEs estimated in this paper
have a relatively general implication. In addition, we find that the LATEs are
likely to be larger than ATEs, suggesting that the inference drawn from
average treatment effects might understate the true effects of the university
expansion program introduced in China in 1999 and thereafter. |