The Long-Short Score Card: 2007 to 2009
How did hedge fund managers handle the peaks and troughs of one of the toughest market environments most of them will have experienced. Richard Johnston, of Albourne Partners in Hong Kong, assesses the numbers.
With such extreme moves in markets characterised by large draw downs and dramatic recoveries, we have reviewed the overall Asia Pacific Long Short space to assess overall performance. However, hedge funds and manager styles can differ radically. Therefore, we have first tried to cluster managers in a few style buckets:
Long Bias High Net Net averaged over 50% for last 3 years
Long Bias Low Net Net averaged 20% to 50% for last 3 years
Variable Long Bias Variable long, tend to de-risk in time of stress
Neutral Try to operate to a largely market neutral mandate
Variable Will mostly operate within plus or minus 30%
Aggressive variable Will operate in a very broad range of net exposures
In trying to assess overall alpha, one of the first problems we identified was a significant “Japan effect”. Some funds had Japan in their mandates but heavily avoided having much net exposure there so they performed more like ex. Japan funds. The broad benchmarks are significantly different dependent on whether Japan is included:
1yr return pa 2yr return pa 3yr return pa
MSCI AC TR Asia Pacific (AP) 37.59% -10.55% -2.94%
MSCI AC TR ex-Japan (APxJ) 73.22% -8.75% 4.35%
For our analysis, we picked the funds on our platform and some additional funds we also regarded as representative. This population is more geared towards the large, more institutional funds with 3 year plus track records. Note that there is a survivor bias to the population. We focused on funds with broad Pan-Asian mandates. Using a simple regression based beta we calculated the following average betas and annualised alpha for different styles for the period from 2007 to 2009. In summary terms we estimate:
Av. Beta AP Av. Alpha AP Beta APxJ Alpha APxJ
High net long bias 0.83 8.9% 0.65 3.2%
Low net long bias 0.30 5.5% 0.24 3.4%
Variable long bias 0.31 10.0% 0.26 7.8%
Neutral and variable bias 0.21 4.5% 0.18 3.0%
The Japan effect can be seen as compared against indices which include Japan, there is more beta and alpha. The biggest source of alpha for funds that could invest in Japan was simply not investing it. Using the ex-Japan benchmarks is probably more broadly representative of the strategy.
What may be surprising is our observation that the funds with more beta actually make more alpha. However, the best group has been the variable long bias. These tend to be single star manager funds with limited capacity. They tend to be tactical in running to cash, and the alpha driver is limiting the negative returns that damage long term compounded returns.
This also indicates a very clear split in our hedge fund population. The group of ‘high net long bias’ managers really does have a very different beta profile. We now regard them as much more of a proxy for long-only investment. It is debatable whether they belong in an absolute return focused portfolio.
The most obvious observation is that alpha is lower based on either of the short term beta measures. What this tells us is that most managers did add alpha by adjusting net and, therefore, there was less alpha from stock selection. September and October 2008 was a very brutal period and many generated negative alpha. The average manager’s net was 21%, yet their returns behaved in line with the longer term average beta of 48%.
It is fair to say that a reasonable manager, with a beta of around 0.2, producing around 5% alpha per year and a volatility of 15-20% lower than the market volatility of 31%, still suffered a 20-30% draw down in 2008.
While we are unable to discuss specific manager issues we can make a few broad comments about the main style groups:
As stated above, the stars of this period are the variable long bias managers with a handful of managers producing the best overall risk reward. In simple terms it comes down mostly to a single individual’s judgment on the timing of de-risking and re-risking. The most favorable outlier on the scatter plots did it better this time but there is no guarantee that this can be repeated.
A few of the more aggressive funds had problems and produced some of the less favorable outlier returns. The problem with a period like 2008 was that being one month wrong, in timing a major change in net, could and did have disastrous consequences.
The more institutional managers, which typically are ‘low net long bias’, produced relatively consistent returns, though they weren’t tactical enough to avoid the significant draw downs of 2008.
Of the ‘high net long bias’ managers, the ones that labeled themselves long only were typically poor at de-risking to any extent. Some of the more long biased hedge funds did a better job. Other managers where we had expected them to do better, were hurt by India which proved one of the worst areas of negative alpha.
Many of the variable net managers did not do as well as we may have expected. Most got whiplashed to some degree, and few managed to stay short through the key period in 2008.
There are not any true market neutral managers running for the full three years.
In the last three years, the Asian ex-Japan markets did make modest positive returns, and this population of hedge fund managers did average around 4-5% alpha pa. However, markets did drop over 50% in 2008 and around 70% of managers had draw downs in excess of 20%. Around 50% of assets in the strategy were withdrawn at levels close to the bottom. Yet we would argue most managers behaved in line with expectation. Most intelligently de-risked, but at a high cost in an environment where liquidity came at the wrong price.
Over a longer term cycle, we believe Asia is a stock pickers market. If there is a region we would argue you don’t want to own the beta via an ETF, or a tightly benchmarked long manager, it would be Asia. Too many large cap stocks are run for the agenda of governments and families. We don’t believe they are good proxies for the Asian growth story. Remember, the broad Asian index was below the 1993 peak at the end of 2008. If Japan is included it is still down over 20 years.
Therefore, it is not uncommon, in a 10 year cycle, to see minimal market return despite strong GDP growth. Yet we believe a good stock picker can achieve an average of 10% alpha pa, so in that sense 2007 to 2009 was a disappointment. Unfortunately this alpha is often bundled with beta and too often investors chase in and out of the region at the wrong time. For some investors Asia has become a series of bad experiences.
Over the cycle, we think the best long term return comes from ‘variable long bias’ managers, but there aren’t many good ones around and they are not scalable. The most stable institutional model is the ‘low net long bias’ group and with betas around 25%, stock selection still dominates longer term. We also now have an emerging group of managers running much closer to neutral portfolios.
The long net bias group is dominated, short term, by beta but over the longer term does produce the highest levels of alpha. We think the ones that are longer net bias hedge funds actually produce better long term returns than many which brand themselves as long-only, where they tend to stay close to fully invested and therefore drifting close to benchmark returns.
However, the ones which are branded long-only often succeed in getting more stable capital than the ones branded as hedge funds. It is likely that the former group is included in an equity portfolio while the latter group gets included in a hedge fund portfolio. This does not make complete sense, that the label matters more than the return profile.