Chapter 1 analyzes Pareto optimal non-linear taxation of profits and labor income in a private information economy with endogenous firm formation. Individuals differ in both their skill and their cost of setting up a firm, and choose between becoming workers and entrepreneurs. I show that a tax system in which entrepreneurial profits and labor income must be subject to the same non-linear tax schedule makes use of general equilibrium effects through wages to indirectly achieve redistribution between entrepreneurs and workers. As a result, constrained Pareto optimal policies can involve negative marginal tax rates at the top and, if available, input taxes that distort the firms' input choices. However, these properties disappear when a differential tax treatment of profits and labor income is possible, as for instance implemented by a corporate income tax. In this case, redistribution is achieved directly through the tax system rather than "trickle down" effects, and production efficiency is always optimal. When I extend the model to incorporate entrepreneurial borrowing in credit markets, I find that endogenous cross-subsidization in the credit market equilibrium results in excessive (insufficient) entry of low-skilled (high-skilled) agents into entrepreneurship. Even without redistributive objectives, this gives rise to an additional, corrective role for differential taxation of entrepreneurial profits and labor income. In particular, a regressive profit tax may restore the efficient occupational choice. In chapter 2, which is joint work with Nick Netzer, we show that, in the presence of a time-inconsistency problem with optimal agency contracts, competitive markets can implement allocations that Pareto dominate those achieved by a benevolent planner, and they induce more effort. In particular, we analyze a model with moral hazard and two-sided lack of commitment. After agents have chosen a hidden effort and the need to provide incentives has vanished, firms can modify their contracts and agents can switch firms, resulting in an adverse selection problem at the ex-post stage. As long as the ex-post market outcome satisfies a weak notion of competitiveness and sufficiently separates individuals who choose different effort levels, the market allocation is Pareto superior to a social planner's allocation with a complete breakdown of incentives. In addition, even when a planner without commitment is able to sustain effort incentives, competitive markets without commitment implement more effort in equilibrium under general conditions. We illustrate our findings with standard market equilibrium concepts. Chapter 3 studies Pareto-optimal risk-sharing arrangements in a private information economy with aggregate uncertainty and ex ante heterogeneous agents. I show that any such arrangement has to be such that ratios of expected inverse marginal utilities across different agents are independent of aggregate shocks. I use this condition to show how to implement Pareto-optima as equilibria when agents can trade claims to consumption contingent on aggregate shocks in financial markets. If aggregate shocks affect individual outputs only, the implementation of optimal allocations does not require interventions in financial markets. If they also affect probability distributions over idiosyncratic risk, however, transaction taxes need to be introduced that are higher for claims to consumption in states with a more volatile distribution of likelihood ratios in the sense of second-order stochastic dominance. Two implementation results are provided. If transaction taxes are constrained to be linear, they need to condition on individual outputs in addition to aggregate shocks. To prevent double-deviations, they induce additional risk for agents who buy financial claims and provide additional insurance to those who sell them. Finally, an implementation with non-linear transaction taxes that do not depend on idiosyncratic shocks is constructed.
