Effects of Bounded Rationality in Financial Markets

Abstract

The first chapter studies the historical growth of active asset management despite the industry's poor track record and its coincidence with net fund entry. Our quantitative model of demand for active management features learning about heterogeneity in skills and about fund-level decreasing returns to scale. Investors are uncertain about parameters governing fund returns and returns to scale, and they learn about them from realized returns. After observing a fund's negative performance, investors infer that the manager's skill is lower than expected rather than that the value of active management more generally is lower than expected. Optimism about the industry as a whole persists at the expense of disappointment about existing individual funds. On the other hand, fund-level decreasing returns will imply that the average unit cost associated with investing in active management is lower as the number of funds increases and, ceteris paribus, make the industry grow bigger over time.

The second chapter, coauthored with Delwin Olivan, studies the effect of monetary policy and ow-performance relationship on risk taking by active equity mutual funds. First, we document that the past decade provided conditions that encouraged these funds to ``reach for yield’’. Low interest rate periods are associated with both higher measures of performance and risk taking. We then utilize discrete Fed announcements providing forward guidance to inform event studies analyzing these factors. Our results are broadly consistent with these funds reaching for yield, and provide evidence of a strong interaction between unconventional low-rate policy and mutual fund behavior.

The third chapter studies the effects of rational inattention to public information in the ``beauty contest’’ model à la Morris and Shin (2002). If the preference for action complementarity is strong, for intermediate levels of the public signal precision, multiple equilibria, indexed by the level of attention allocation, emerge. I also show that equilibrium attention allocation is inefficient for a wide range of model parameters. Key to these results is the fact that strategic complementarity translates into attention complementarity, when the information cost is positive, but sufficiently small (compared to zero). Quintessentially, ``almost perfect attention’’ is very different from perfect attention assumption.