Optimal Taxation in Asset Markets with Adverse Selection

Consider markets for assets traded over the counter such as mortgage-backed securities and corporate bonds. Sellers in these markets may have more information on the value of their assets and their liquidity needs than buyers do. Also, sellers and buyers must search for trade partners, which is time-consuming and costly. During the 2007–09 financial crisis, activity in some of these markets declined to close to zero, and governments and central banks undertook various policies to encourage trading.

One of the many policy questions that have arisen since then is whether government interventions are a good policy tool from a social perspective. To address this question, I study the optimal sales tax schedule in an environment where the planner faces the same information asymmetries that market participants face. In this environment, the planner does not know the type of assets that the sellers try to sell or the sellers' liquidity needs.

I show that an economy without government intervention—a laissez-faire economy—is always constrained inefficient. That is, the planner can always find a tax schedule to improve the outcome of the laissez-faire economy. However, the optimal tax schedule is often non-monotonic. For example, trading of some high-price assets should be subsidized, and trading of some low-price assets should be taxed. The implementation of a non-monotonic tax schedule is difficult in practice because it is sensitive to the planner’s knowledge of the distribution of asset types. I show that tax schedules will become monotonic if an entry tax can be imposed on agents in addition to sales tax. The entry tax is levied on agents who enter the market regardless of whether they trade.