By Adam Blandin
The human capital literature is largely split between two models of human capital investment: Learning By Doing (LBD) and Ben-Porath (BP). Given the importance of human capital investment for a host of policy issues, I ask whether observable macroeconomic moments are informative about the relative importance of LBD investment versus BP investment. A life-cycle human capital model is constructed which nests both LBD and BP as extreme special cases. I find: (1) Both the BP and LBD versions of the model are consistent with the aggregate distribution of earnings, hourly wages, and hours worked for men in the PSID. (2) Conditional on matching these aggregate levels facts, the BP version of the model is more consistent with the variance in earnings growth rates in the data. (3) Policies which decrease the return to human capital investment, such as a progressive earnings tax, decrease aggregate human capital investment and earnings substantially more in a BP world than in a LBD world. Taken together my findings suggest that within a plausibly parametrized model of human capital accumulation, government policies which reduce the return to human capital investment will generate large decreases in human capital investment and earnings.
There is surprisingly little discipline in how to model human capital accumulation. In large part, this is due to the fact that it is not directly observable. Hence, this paper is looking at some indirect evidence, and the results is of course going to be model-dependent. But at least some effort is spent determining which process is more likely to be correct.
Here is the new version.