The paper focuses on asymmetric fund performance and compares the performance characteristics of European and US large-cap mutual equity funds. For this specific category of funds, the risk-factor adjusted performance is expected to be quite robust. For our sample of 31 European and 35 U.S. large-cap mutual equity funds, however, the performance is found to be sensitive to the empirical estimation approach applied. Furthermore, the performance alphas exhibit asymmetry as they are not robust across the conditional return distribution. A large part of the U.S. individual funds significantly underperforms the benchmark, especially in the lower tail of the conditional distribution. A few of the European funds, on the other hand, exhibit significant and positive performance alphas in the lower tail of the conditional return distribution. From a risk-averse investor’s point of view, investing in European large-cap funds, the performance of an equally weighted fund of funds is more comforting. On average, the performance alphas are positive and highest in the lower part of the conditional distribution. Unfortunately, this result does not hold for a corresponding equally weighted fund of U.S. large-cap equity funds. The U.S. fund of large-cap funds exhibits the lowest alphas in the left tail of the conditional return distribution.
Method:
We partly follow the approach taken by Högholm et al. (2011b) and start the empirical investigation by first comparing the results of three different approaches for estimating the performance alpha. As a benchmark we estimate the traditional Fama-French four-factor model using an unconditional ordinary least squares (OLS) regression
Data:
The data sample consists of monthly fund returns for the time period September 1998 to December 2012 sampled from Thomson Reuters Datastream. For the European funds the sample period partly overlaps the sample period used in Högholm et al. (2011b). For comparison purposes we try to sample the same large-cap funds even though they focused on daily returns. For the empirical analysis we only use funds with data available for the entire sample period.