Abstract
We extend the analysis of Fama and French (2010, Journal of Finance) by allowing for time-variation in the parameters of Carhart’s (1997) Four Factor Capital Asset Pricing Model (4F-CAPM). As such, we are able to deal with the thorny problems highlighted by Fama and French (2010, p.1925). Using recently developed statistical procedures, we test whether subsets of the parameters in time-series regression models have changed over
time. We also use fixed-design bootstraps that are valid even in the presence of this time variation. Our results reveal that time variation has significant affects on our bootstrap simulation of measures of fund performance. When returns are measured gross of trading costs and management fees, we find actively managed funds generate returns no better than those from passive benchmark portfolio. However, the statistical strength of this result is weakened in the presence of time variation.
time. We also use fixed-design bootstraps that are valid even in the presence of this time variation. Our results reveal that time variation has significant affects on our bootstrap simulation of measures of fund performance. When returns are measured gross of trading costs and management fees, we find actively managed funds generate returns no better than those from passive benchmark portfolio. However, the statistical strength of this result is weakened in the presence of time variation.
| Original language | English |
|---|---|
| Number of pages | 23 |
| Publication status | Published - 12 Feb 2018 |