Abstract Most prior research on labor market mismatch was constrained by the unavailability of data on skill mismatch and also the absence of panel data which would provide controls for unmeasured heterogeneity. This paper makes use of the panel element of Korea Labor & Income Panel Survey (KLIPS) data and identifies the wage effects of educational mismatch and skill mismatch both separately and jointly. It clearly shows that only a small proportion of the wage effect of educational mismatch is accounted for by skill mismatch, suggesting a relatively weak relation between educational mismatch and skill mismatch. In the analysis appropriate panel methodology produces much weaker estimates of the relevant coefficients than the pooled OLS model. This result indicates that unobserved individual-specific characteristics play a substantial role in the way in which mismatch effects are determined.