Performance Evaluation and Conditioning Information. Evidence From the UK Investment Trusts Industry

We are investigating whether conditioning information affect the performance of UK investment trusts to detect superior performance. Our results suggest that simple unconditional constant models that are used are not valid and are misspecified and biased to measure and evaluate the performance of the investment trusts. Constant expected returns lead to underestimation of the performance of investment trusts. By employing conditioning information or time – varying expected returns models to measure changes in the return and risk increase the statistical significance of the performance evaluation. We take into consideration that investment managers trade on public information and trading strategies. Our sample consists of 16 categories with a total number of 210 funds including the “dead” ones. We will use a comprehensive data set of NAV including both alive and “dead” funds from January 1990 to January 2006. The defunct funds that will be included have an incomplete dataset due to the limited period of their survival. They will be used for further tests to detect performance persistence bias and how they affect the whole sample