Predictive Ability and Market Pricing of Fair Value Earnings Components for Closed-End Funds
50 Pages Posted: 3 Feb 2016 Last revised: 4 Jul 2017
Date Written: July 4, 2017
This study uses U.S. closed-end funds to investigate whether the realized component of fair value earnings conveys information about future fund and benchmark market performance and whether the market impounds this predictive information into fund share prices. We find that the realized gain/loss component of fair value earnings is strongly negatively related with future fund performance. This finding is consistent with two competing explanations for the fund managers’ trading behavior: the disposition effect behavioral bias hypothesis and the market timing hypothesis. Consistent with the market timing hypothesis, we find that the relative frequency of realized gains and the average magnitude of net realized gains in period t is higher (lower) conditional on benchmark market returns in t 1 being negative (positive). We find little evidence that fund managers’ asymmetric selling of winners versus losers is due to disposition effect behavioral biases. Market pricing tests reveal that investors in closed-end funds do not fully impound the signaling information about future fund performance and future benchmark returns that is conveyed by net realized gains/losses. Similar to historical cost earnings components, investors appear to focus on aggregate fair value earnings and fail to fully impound into fund share prices the signaling ability with respect to future earnings of the realized gain/loss component of fair value earnings.
Keywords: realized and unrealized gains/losses, closed-end funds, market timing, disposition effect, pricing of fair value earnings components
JEL Classification: M41, G11, G14
Suggested Citation: Suggested Citation