Estimates of the returns to schooling in Taiwan: evidence from a regression discontinuity design

Using micro-data from the Manpower Utilization Survey 1995-2015 of Taiwan, this paper exploits exogenous variation in years of schooling caused by the 9-year compulsory education reform to estimate the economic returns to schooling. The results of fuzzy regression discontinuity design indicate that the causal effect of education on wage is approximately 5.6%, smaller than the corresponding ordinary least squares estimates. Our results are robust to various specifications and placebo tests.