Gordon Gemmill

Noise Trading, Costly Arbitrage, and Asset Prices: Evidence from Closed-end Funds

If arbitrage is costly and noise traders are active, asset prices may deviate from fundamental values for long periods of time. We use a sample of 158 closed-end funds to show that noise-trader sentiment, as proxied by retail-investor flows, leads to fluctuations in the discount. Nevertheless, we reject the hypothesis that noise-trader risk is the cause of the long-run discount. Instead we find that funds which are more difficult to arbitrage have larger discounts, due to: (1) the censoring of the discount by the arbitrage bounds, and (2) the freedom of managers to increase charges when arbitrage is costly. Copyright The American Finance Association 2002.

What drives Credit Risk in Emerging Markets? The Roles of Country Fundamentals and Market Co-movements

This paper investigates how the creditworthiness of Argentina, Brazil, Mexico and Venezuela as reflected by their bond prices, is influenced by both global factors and country-specific fundamentals. We use an extended structural model suggested by Cathcart and El-Jahel (2003) and a Kalman Filter to obtain the distance-to-default implicit in prices of each countrys Brady bonds. We find that a small set of country fundamental variables and external factors, including a variable that measures market sentiment, are able to explain up to approximately 80% of the variance of the distance-to-default of each country. Whereas country specific factors are statistically significant in explaining the distance-to-default, external factors (such as the US stock market index, interest rates and market interdependence across countries) are much more important in explaining the dynamics of this variable. Using principal component analysis we find that there is a common factor for all of the countries which explains approximately 60% of the remaining variance (of the residuals). This common factor is therefore systematic and purely related to the bond market.

Political risk and market efficiency: Tests based in British stock and options markets in the 1987 election

Abstract This paper examines the behaviour of the stock and options markets in London during the 1987 election. We find an extremely close relationship between opinion polls and the FTSE 100 Index of share prices. However, in the last week of the election the options prices showed evidence of gross inefficiency: they implied a decreasing probability of a Conservative win while the polls indicated the opposite. The inefficiency was sufficiently large for a volatility arbitrage to be feasible, net of transactions costs. These two tests appear to indicate that a speculative bubble was present in the options.

The Impact of Corporate Governance on Closed-end Funds

This study uses a large sample of UK-listed closed-end funds to examine whether governance has an impact on two indicators of fund performance: the level of fund-management fees and the discount at which a fund trades. Fees are under the control of the directors, and we find that they are inversely related to fund returns, even after allowing for differences across investment sectors. Fees are, on average, higher if a fund has a large board, few directors from outside the fund-family, many directors from within the fund-family, and low ownership by the management company. Discounts for funds are wider if the management company or any blockholder has a significant long-term stake, suggesting that investors are wary of entrenched management. The results suggest that boards are frequently compromised in their duty to shareholders by their dependence on fund-management companies.

Warrants on the London Stock Exchange: Pricing Biases and Investor Confusion

This study of warrants on the London Stock Exchange examines whether they display particular pricing biases and whether investors understand how to value them at the time of issue. In a sample of 72 warrants on closed-end funds (investment trusts) over the 1985–94 period, more than one third of the 12,673 prices are anomalously low. The other two thirds behave like stock options, with lower volatility when they are in-the-money or have a long time until maturity. Despite their frequent undervaluation, it is rational to add warrants to a new equity issue. An examination of 127 new equity issues (95 with warrants) reveals that attaching warrants significantly increases market value. The reason for this appears to be investor confusion: they do not seem to understand that the more the warrants are worth, the less the value of the ordinary shares.

The Agency Effect of Repurchases on Closed‐End Funds

Share prices rise after companies announce repurchases, but there are differing views as to why this happens. Repurchases are announced by closed-end funds when their discounts are widening (market-to-book is falling). The immediate post-announcement effect is a small jump in a funds share price, but the main effect occurs over the next four years during which time there is significant outperformance both of the funds price and of its investment portfolio. Liquidity of the shares does not change. Repurchases, if executed, reduce the size of a fund and therefore the managers fees. Our findings are consistent with directors using the threat of repurchases to discipline managers whose investment performance has been poor, leading to a closer alignment of pay and performance.

Are IPO Investors Rational? Evidence from Closed-End Funds

Why buy a closed-end fund at IPO, when it is likely to trade at a discount in a few months’ time? One theory suggests that buying a new fund is justified by an initial period of investment outperformance. A second theory is that new funds are launched to provide access to assets that are temporarily illiquid and to exploit the subsequent liquidity gain while a third theory asserts that buyers of new issues are not fully rational but are influenced by time-varying sentiment. This paper tests the three theories using data from UK-traded closed-end equity-fund IPOs over 1984 to 2006. The empirical results provide strong support for the influence of sentiment but provide little or no support for the two other theories.

Fund Mortality and Fair Discounts on Closed-End Funds

We develop and test a rational model of discounts that takes account of conditional expectations about fund lives. Previous research has assumed that lives are very long, implying unrealistically large discounts. We find that expected lives are short: 10 years or less up to the age of five, then rising slowly to a plateau of 18 years. If dividends are paid, weighted-average lives are even shorter. The model is calibrated to the first 20 years of a fund’s life and tested with cross-section data from the UK and US. We conclude that life-expectancy plays a central role in explaining discounts, but it is not sufficient (by itself) to explain the premia on new issues.

Explaining CDS Prices Before and After the Lehman Default with a Simple Structural Model

We examine what determines CDS prices over 2005-2012. To do this, we calibrate Mertons model in a novel way that allows for deviations from lognormality. The model works well in cross-section and time-series, both within and out-of sample. It confirms that systematic equity volatility is the major determinant of CDS prices. Before the Lehman default, all firms have CDS prices that are close to those set by the model (with small variations due to illiquidity and earnings-uncertainty). After the default, some firms continue to have CDS prices at model-predicted levels, but others are now more-strongly influenced by idiosyncratic factors and have much higher prices.

Why are Some Issues of Convertibles Callable and Others Not

We develop a theoretical model to explain why some issues of convertible bonds include a call feature whereas others do not and then test the predictions of the model empirically. The reasoning in the model is that good firms do not need a call feature and therefore issue straight debt or non-callable convertibles. Bad firms issue callable convertibles, because a call might help them to avoid bankruptcy costs if their prospects deteriorate in a later period. A separating equilibrium for the model is demonstrated. The model predicts that the announcement of a non-callable convertible will have a smaller negative impact on the share price than that of a callable convertible; this is confirmed by the data, for which the price effects are -3.97% and -6.15% respectively. A second prediction is that announcing a call (and forcing conversion) conveys bad news to the market; we find that there is a significant short-term effect on the share price, but only if the call is made relatively quickly. Finally, we check whether issuing or calling a convertible is associated with abnormal subsequent performance over three years and find that it is not.