Capital gains overhang and the closed-end fund puzzle, and, Economic significance and arbitrage of idiosyncratic risk

The dissertation is divided into two chapters. Chapter I examines the impact of capital gains distribution rules on a closed-end fund (CEF) manager's incentive to collect and trade on information. The central hypothesis is that the treatment of realized capital gains embodied in the U.S. tax code, combined with the form of compensation contract offered CEF mangers, creates an "overhang" effect on CEF managers' incentives to optimally trade the fund portfolio. We model the overhang effect and show that (1) there exists an equilibrium in which CEF managers collect private information in the early stages of the fund (when no lock-in effect exists) but then choose to be uninformed in later stages (when the lock-in effect exists), and (2) rational-expectations pricing of the expected cash flows and risk resulting from the information equilibrium results in CEFs initially trading at a premium and subsequently trading at a discount. In addition, we examine the empirical implications from the model and find, consistent with the model, a significant negative relation between future risk adjusted performance and current unrealized capital gains as well a significant positive relation between NAV premiums and future risk adjusted performance. In addition we show there are two separate effects from unrealized capital gains: (1) the Malkiel (1977) tax effect on CEF investors and (2) the manager incentive-capital gain related effect. Chapter II examines the economic significance of idiosyncratic risk in the context of arbitrage profits as well the robustness of idiosyncratic risk estimates relative to additional known systematic risks. We estimate idiosyncratic risk using both rolling and single time-series EGARCH methods and form high and low idiosyncratic risk portfolios. After controlling for liquidity risk and momentum, we conclude that abnormal returns to idiosyncratic risk arbitrage strategies are not statistically and/or economically significant. Furthermore we find no evidence of priced idiosyncratic risk in closed-end funds. Finally, we find no relation between idiosyncratic risk and stock returns in Fama-MacBeth cross-sectional regressions.