Hedge fund factorisation and benchmarking: Understanding hedge fund performance, benchmarking and the reward system for hedge fund managers

Govender, Nishlen

Hedge fund factorisation and benchmarking: Understanding hedge fund performance, benchmarking and the reward system for hedge fund managers

Master Thesis

2021

Abstract

Hedge funds give portfolio managers access to more tools to aid in better portfolio construction. The introduction of tools such as leveraging and shorting provided managers with the ability to augment exposures to different asset classes. The result is the ability to create portfolios with uncorrelated returns without having to invest in a plethora of asset classes thus providing better risk adjusted returns. This paper tests whether hedge funds in fact contain less exposure to individual asset classes than their long-only counterparts. In particular, the perception of uncorrelated returns has led to hedge funds being benchmarked against absolute return targets while charging performance fees higher than the typical long-only fund. If hedge fund returns are correlated with those of the asset class in which they invest, the available risk premia available in that asset class may drive returns more than manager skill. In circumstances where hedge fund returns are in fact correlated with asset class returns, then the benchmarks used to measure hedge fund performance ought to capture the perpetual risk premia of the asset classes for better performance measurement and performance fee rewards. This dissertation closely follows Hasanhodzic and Lo (2007) who sought to find the quantum of hedge fund returns attributable to asset class returns; that information was then used to create low-cost clones of typical hedge fund strategies. This dissertation also tested the strength of the relationship between hedge fund and asset class returns but used the result to build linear clones for benchmarking rather than as an alternative to hedge funds. What also distinguishes this dissertation is the jurisdiction: Hasanhodzic and Lo (2007) examined global hedge funds while this dissertation focusses on the South African hedge fund industry. The HedgeNews Africa database is the data source for South African hedge fund returns (from some 412 funds though only 160 of those are currently active). Database returns existed for the period June 1998 to June 2020. The regression assessment conducted regressed the returns of various hedge fund strategies against the returns of the relevant asset classes. The result of the regressions reveals significant coefficients relating to different asset class independent variables. The significant relationships accord with the logical association of certain hedge fund strategies with particular asset classes. For instance, equity long-short funds had a large and significant coefficient relative to equity market returns. Based on the regressions, clone benchmark portfolios were created which performed similarly to the various strategies in the ex-ante period from January 2019 to 2020. This lends credence to the idea that better benchmarks can be specified for hedge fund managers.

Keywords

Financial Management

Reference:

Collections

Masters

Full item page