Abstract
GLS estimation with correction for autocorrelation is used to examine the performance of 72 UK Investment Trusts in four non-overlapping five-year samples, 1974 to 1993, using measures of ‘timing’, and ‘selectivity’. Abnormal performance on both indicators is found to be weak, with very little persistence. Patterns of significant selectivity and timing are found to reverse after the first period. Negative timing performance is found for around one third of trusts in the last three sample periods, and strong negative correlations are observed between timing and selectivity for all periods. Various interpretations of these findings are given.