Hedge Fund Performance During and After the Financial Crisis

The Carhart Four Factor Model was used to analyse performance over all three periods.

Over the entire period (Table 6), all models estimated were statistically valid with a p-value of the F statistic < 0.05. Looking at the performance over the whole data set (June 2007–January 2017) in Table 6, it is obvious that first almost all strategies had a significant positive degree of market risk, the highest being the Emerging Markets (0.47) and the Long/Short Equity (0.46); while Short Bias had a significant negative amount of market risk (−0.69) and Managed Futures were the exception. Only two strategies were able to generate a significant alpha over the whole period. These strategies were Global Macro and Multi Strategy. Both of them have a low significant positive correlation to the market and low significant negative correlation with the HML factor, which measures the return of value stocks, minus that of growth stocks. The negative sign of the HML coefficient shows that the portfolios were short in value stocks and long in growth stocks.

Table 6. Carhart's Four factor model estimations whole period: (6/2007–1/2017).

In the crisis period (Table 7), seven models estimated were found to be statistically valid (p-value < 0.05), while three were not (Equity Market Neutral, Global Macro and Managed Futures). These strategies show a lack of relationship with the entire market, in this time. During the financial crisis (June 2007–March 2009) period, six strategies (Convertible Arbitrage, Emerging Markets, Event Driven, Fixed Income Arbitrage, Global Macro, Long/Short Equity and Multi Strategy) continued to bear a significant positive amount of market risk, whilst Short Bias had a significant negative amount of market risk (−0.55), and Equity Market Neutral and Managed Futures continued to have an insignificant amount of market risk. None of the ten strategies analysed were able to produce a significant alpha. Moreover, the relationship with the HML factor was significantly negative for many of the strategies (Convertible Arbitrage, Emerging Markets, Event Driven, Fixed Income Arbitrage, Global Macro and Long/Short Equity) showing the negative relationship to the return from Value stocks, compared to the Growth stocks.

Table 7. Carhart's Four Factor Model estimations during the Financial Crisis (06/2007–03/2009).

In the after-crisis period (Table 8), all models estimated were statistically valid (p-value of F statistic < 0.05). The period after the financial crisis (April 2009–January 2017) was characterised by rising market prices. At this stage, almost all strategies had a significant positive amount of market risk (highest being the Long/Short (0.47), followed by the Emerging Markets (0.41)), with Short Bias continuing to have a significant negative amount of market risk (−0.81). Four out of ten strategies were able to generate a significant alpha, during this period (Convertible Arbitrage, Fixed Income Arbitrage, Global Macro and Multi Strategy) with Fixed Income Arbitrage being the highest (0.54%). Two of the strategies with significant positive alphas (Convertible Arbitrage and Fixed Income Arbitrage) had a significant negative UMD coefficient, showing that these strategies had a negative momentum factor – short in winners and long in losers.

Table 8. Carhart's Four Factor Model estimations after the Financial Crisis (04/2009–01/2017).

Analysis with the Carhart model, yielded several insights. Only two strategies were able to generate significant alphas, over the whole period. These strategies were the Global Macro and Multi Strategy. During the crisis, none of the strategies yielded positive alphas. After the crisis, Global Macro and Multi Strategy continued to show positive alphas. The strategies which were the most successful had low but significant betas. Over the whole period, almost all strategies showed significant amounts of market risk to different degrees (except Short Bias which was significantly negative). The HML factor showed significant negative correlation to the strategies that were able to generate positive alphas, over the whole period, and also during the crisis, implying the underlying portfolio of these strategies were short in value stocks and long in growth stocks. The relationship with the HML factor was significantly negative for many of the other strategies, as well at different times. Although the SMB and UMD factors did not have an explanatory value for the better performing strategies, the SMB factor had a significant negative correlation with Short Bias, over the whole period and after the financial crisis, suggesting this portfolio was short on small stocks and long in large stocks; the UMD factor had a significant positive value for Managed Futures, over the whole period and after the financial crisis, suggesting that these portfolios invest in momentum stocks.

Consistent with the findings of Capocci (2002) and Stoforos et al. (2016), the study showed no fund manager was able to generate a significant alpha over the crisis period. However, market beta was significant for all strategies, except Equity Market Neutral and Managed Futures, in this period. It meant that the other strategies were not able to completely hedge against the market risk in the crisis period. In the after-crisis period, all strategies had a significant positive market risk, except Short Bias; although the ones with significant alphas also had the lowest market risk.

While the estimations using Carhart's model afford several insights, they do not provide an overall evaluation of the return versus the risk on the various strategies. Due to this, it is relevant to compare the strategies' performances using alternative measures, such as persistence in performance and reward-risk payoffs.