Working Paper: NBER ID: w3693
Authors: Mckinley L. Blackburn; David Neumark
Abstract: Over the 1980s there were sharp increases in the return to schooling estimated with conventional wage regressions. We use both a signaling model and a human capital model to explore how the relationship between ability and schooling could have changed over this period in ways Chat would have increased the schooling coefficient in these regressions. Our empirical results reject the hypothesis that an increase in the upward bias of the schooling coefficient, due to a change in the relationship between ability and schooling, underlies the observed increase in the return to education over the 1980s. We also find that the increase in the return to education has occurred largely for workers with relatively high levels of academic ability.
Keywords: Return to schooling; Wage differentials; Ability; Human capital; Signaling model
JEL Codes: J31; I21
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| ability (G53) | wages (J31) |
| changes in relationship between ability and schooling (I24) | increase in return to schooling (I26) |
| increase in education-related earnings differentials (I24) | higher academic ability (I23) |
| distribution of workforce across education classes (J24) | increase in estimated return to schooling (I26) |
| change in relationship between ability and schooling (I24) | upward bias of schooling coefficient (I24) |