Benchmark Changes Matter: The Case of Mutual Fund Relative Performance:

Nedanstående innehåll är skapat av Mimers Brunns besökare. Kommentera arbete

From data on Sweden-registered mutual funds in the period 1999-2004 we construct a data set free of survivorship bias which we use to test the hypotheses that, on average, a) among funds that change benchmarks, the new benchmark selected ex post has a worse performance than the original benchmark, b) benchmark changes are affected by funds’ performance relative to their original benchmarks, and c) benchmark changes affect fund flows. Using simple t-, binomial -, and Wilcoxon tests as well as multinomial logit and fixed effects regression models, we find robust evidence for all three hypotheses. This suggests that at least some funds take relative performance into account when changing benchmarks and that these funds face a tradeoff between reputation and presenting a favorable view to investors. In addition, we observe a high frequency of benchmark changes overall: the majority of funds chang...