This dissertation consists of three chapters. The first two explore the impact of government policies on human capital accumulation.

Chapter one makes two novel contributions related to the two workhorse models in the human capital literature: Learning by Doing (LBD) and Ben-Porath (BP).

First, I show that BP is much more consistent with empirical life-cycle patterns related to individual earnings growth rates relative to LBD.

Second, I show that the same model features that generate different life-cycle predictions between models also generate different policy implications. In particular, increasing the top marginal labor tax rate, relative to the current US level, generates much larger reductions in lifetime human capital accumulation in the BP model versus the LBD model.

Chapter two examines reforms to the Social Security taxable earnings cap in the context of a human capital model. Old age Social Security benefits in the US are funded by a 10.6% payroll tax up to a cap of $118,500. There has been little work examining the likely outcomes of such a policy change. I use a life-cycle BP human capital model with heterogeneous individuals to investigate the aggregate and distributional steady state impacts of several policy changes the earnings cap. I find that when I eliminate the cap: (1) aggregate output and consumption fall substantially; (2) the role of endogenous human capital is first order; (3) total federal tax revenues are lower or roughly unchanged; (4) about 1/3 of workers are made worse off.



The final chapter studies the existence and optimality of equilibria in the presence of asymmetric information. I develop an equilibrium concept which corresponds to the presence of mutual insurance organizations for a class of adverse selection economies which includes the Spence (1973) signaling and Rothschild-Stiglitz (1976) insurance environments. The defining features of a mutual insurance organization are that policy holders are also the owners of the organization, and that the organization can write policies for which the terms depend on the experience of the mutual members. In general the equilibrium exists and is weakly Pareto optimal. Further, all equilibria have the same individual type utility vector.