After a decade of dramatic growth, the hedge fund industry is experiencing the greatest shock of its history. First estimates indicate that hedge funds assets under management shrunk by at least 20% in 2008. With massive redemptions and global deleveraging on the one hand, liquidity squeeze and extreme market conditions on the other, many hedge funds have experienced large drawdowns and have placed liquidity restrictions (side pockets, gates, suspension of redemptions).

While it is too early to estimate precisely today how many funds will be able to recover and/or overcome this crisis, this article aims at giving a first idea of the surge in hedge fund mortality in 2008.

Estimating the hedge fund mortality rate is everything but straightforward since hedge fund managers report to databases on a voluntary basis. As a consequence, managers naturally tend to stop and delay their reporting when performance turns really ugly. As a corollary, delays in reporting do not necessarily mean that the funds are in the process of liquidation. Either managers might wait several months for the performance to improve before reporting again, or they can exit the database for reasons other than liquidation. Hence, two types of measures need to be distinguished:

• the attrition rate, which is the percentage of funds exiting the database, and
• the liquidation rate, which is the percentage of funds exiting the database for liquidation reasons only.

Academic research has shown that historically, the attrition rate was twice as great as the liquidation rate. This article provides an early estimate for the attrition rate, not for the mortality rate, computed from the HFR database. We found that about 28% of the funds in the HFR database at the beginning of the year 2008 are no longer there today. Unsurprisingly, this corresponds to a great surge in the attrition rate as compared to the past: recent research has estimated an attrition rate of 8.7% on an annual basis for the period 1995-2004 (Liang and Park, 2008). This article is organised as follows: first we describe the data used for the analysis and the methodology followed to clean the dataset. Then, we explain how we estimate the attrition rate. This allows us to perform the calculation of this rate by month and by strategy in 2008. Detailed results appear in Table 1: Attrition Rate in 2008.

Dataset
This study is based on the HFR database as of 2 February 2009. We consider only USD denominated funds. Funds of hedge funds and indices are disregarded. We use the OCM strategy classification. This classification, as defined by OCM, divides hedge fund managers according to five main strategies, namely Directional Trading, Relative Value, Credit, Long/Short Equity, and Systematic Equity Trading.In order to reference eachfund in a unique way, we need to clean the dataset. All of the different series or share classes of a hedge fund are considered as a unique fund. In order to detect duplicates, we perform several tests, comparing names (cleaned beforehand), track record, and strategies. A duplicate is detected if two funds have the same cleaned or if two funds have the same strategy with a correlation superior or equal to 99%. If two funds are found identical, the one with the largest AuM is selected for the study.

Methodology
The attrition rate that we calculate here is derived from the number of funds that have ceased to report their returns to the database. For each month (from January to November 2008), we determine how many funds have stopped reporting since the previous month. Two points are worth mentioning:

1. As pointed out in the introduction, this is an estimation of the attrition rate, which is very different from the liquidation rate. Indeed, hedge funds can have several reasons (other than liquidation) for not reporting anymore: merger, closing to new investors, and important drawdowns, among others. Historically, the liquidation rate represents between 40% and 50% of the attrition rate (Baba and Goko, 2006, and see Table 2: Attrition rate estimate by Baba and Goko). In the current environment, one can suppose however that the recent attrition is primarily down to liquidation.

2. The reporting lag has an important impact on the last months of the year. Since managers can wait several months before reporting, our results probably over-estimate the number of exiting funds over the last few months. However, the surge in attrition rate observed since September is in line with the turmoil measured by the indices (HFRX or HFRI FoF).

Results
According to our estimates, 28% of the funds listed in the database as of January 1 2008 have stopped reporting between January and November last year. This was probably the most difficult year to date for the hedge fund industry. Not surprisingly, hedge funds with the lowest AuM have been the most severely impacted. If we only consider funds that manage more than $50 million, the attrition rate falls to 22%. It decreases to 19% for funds with more than $200 million under management.

Nevertheless, we have to interpret these results with care due to the lag effect mentioned above. Indeed, the number of funds that have stopped reporting has increased considerably since September, from 60 funds per month on average for the period preceding September to 138 per month on average. As this effect is particularly strong at the end of the year, we have decided to exclude December from our analysis. Figure 1 displays the attrition rates by month and by main strategy. The 2008 attrition rates by main strategy are reported in Fig.2.

As argued by Getmansky, Lo and Mei (2004), we can see that the attrition rates significantly differ by investment strategy. If we do not consider the ‘Other’ category (which only includes non-classified funds), the following hierarchical order can be made, in terms of the percentage of exiting funds:

• Relative Value strategy: 37% of ‘exiting’ funds
• Credit: 33%
• Long/Short Equity: 26%
• Global Macro and CTA: 20%

These findings are coherent with both the level of returns delivered by these strategies and with the type of trades they traditionally use: i) Directional funds (Global Macro and CTAs) were the only positive performers in 2008 and most trade liquid instruments; ii) Relative Value funds were severely hit by the liquidity squeeze. We can draw the same conclusions for funds having more than $50 million under management (with the exception of the Systematic Equity Trading strategy).

Conclusion
Although these results are not final yet (December is not included), they draw a first picture of the hedge fund industry as of the end of 2008. This year is characterised by a very important attrition rate of 28%, which rises to 37% for Relative Value funds. This is significantly larger than the historical attrition rate observed between 1995 and 2004.

Corentin Christory is an analyst with the risk and quantitative department of Olympia Capital Management, Paris. Prior to this, he worked in the Alternative Investment Products team with Crédit Agricole.

REFERENCES

Baba N., and H. Goko, “Survival Analysis of Hedge Funds.” Bank of Japan Working Paper Series, No 06-E-05, 2006
Getmansky, M., A. Lo, and S. Mei, “Sifting Through The Wreckage: Lessons From Recent Hedge Fund Liquidations.” Journal of Investment Management, 2, pp.6-38, 2004.
Liang B. and H. Park, “Predicting Hedge Fund Failure: A Comparison of Risk Measures.” Working paper, 2008.