The closed-end fund anomaly: A review and novel tests

We study the persisting pricing anomaly affecting closed-end fund (CEF) stock prices, whereby on average CEF shares are overly discounted in the market, given their underlying net asset value (NAV). We present the theoretical explanations which have been proposed to explain the persistence of the anomaly. Empirical evidence concerning each of these explanations is presented and discussed. As a whole, the review of the literature demonstrates that while financial and economic theory can explain cross-sectional CEF discount differences, it offers little explanation as to the time-variation of discounts, and no explanation as to the links between CEF price, trading volume, and discount level. The two papers which constitute our thesis add to the literature by addressing those two unexplored aspects of the CEF anomaly.; In the first paper, we study the impact of domestic and foreign news on closed-end country fund stock returns, premia variations, and their conditional volatilities. News are defined as the unexpected component in observed information variables. Information variables include domestic and foreign stock market capital returns and dividend yield variations, and exchange rate variations. Stock returns and premia variations are modeled as a linear function of news, while their conditional volatility is modeled using an asymmetric GARCH formulation. Results indicate that news from foreign stock market capital returns and exchange rate variations are the main influences behind country fund stock return and premia variations. Results also indicate that premia volatility overreacts to bad news, but country fund stock returns volatility overreacts to good news.; In the second paper, we relate current CEF trading volumes and discount levels to past trading volumes at differential price levels. In doing so, we compare the predictions from the tax-loss selling effect and the disposition effect as competing explanatory hypotheses of CEF trading volume. Results indicate that neither effect seems to be the main influence behind CEF trading volume. Furthermore, there does not seem to be any relationship between high past volume, larger past prices changes, and current discount levels.