This paper produces multiple fund-of-funds’ portfolios based on sorting and a novel non-parametric approach to lend in-depth insights into the extent hedge fund returns display higher order (non-linear) persistence patterns. By exploiting monthly data from Hedge Fund Research Database between 1999:1 to 2015:12, an out-of-sample exercise provides robust evidence of performance persistence over the time horizon. However, up to 80% of hedge funds appear to be ‘lucky’ performers, confirmed upon undertaking a battery of robustness measures. We name those ‘lucky’ hedge funds, which pass the classical persistence selection procedure and are selected ‘by accident’ with regard to funds with persistent returns. We argue that these funds are unlikely to experience significant outperformance in the future. Our non-parametric identification mechanism allows for a reduced number of persistent funds, enabling practitioners to focus on a few with meaningful qualitative scrutiny. Moreover, we expand the debate in the literature by analysing post-crisis data and demonstrating that after the 2008 crisis, the proportion of genuinely persistent funds got significantly reduced.
|Number of pages||25|
|Publication status||Published - 4 Aug 2022|
|Event||9th International Conference on|
Computational and Financial Econometrics (CFE 2017) - Senate House, University of London, London, United Kingdom
Duration: 16 Dec 2017 → 18 Dec 2017
- Hedge fund
- Persistence performance