Академический Документы
Профессиональный Документы
Культура Документы
April 2001
CROSSBORDER CAPITAL
Many investors believe that young hedge funds1 perform better than long-established hedge funds. Several reasons are provided, such as the existence of a temporary niche that is ultimately arbitraged away; the higher energy devoted to performance by a new entrepreneur, and then the adoption of a more conservative investment approach once the managers massive performance fees have been invested in the fund. However, a worrying counterargument focuses on survivor2 bias in the data samples. It suggests that there is likely to be a higher failure rate among new funds, and a fund that fails, by definition, falls out of the database. Some studies have shown that survivor bias may add or subtract as much as 2% to aggregate hedge fund statistics3. If young funds have a still higher failure rate, their survivor bias is likely to be greater. Consequently, returns from investing in younger funds may prove unattractive once this risk of failure is factored in. Clearly, this is a proposition that is worth testing. Using the TASS/Tremont database for the 1994-2000 period, we first screened according to age decile. The results for 3,733 funds, shown in Figures 1 and 2, reveal a median return of 13.4% for all deciles and a median return of a whopping 23.2% for the youngest decile. In other words, there is a massive spread of around 980 basis points in favour of young funds. Figure 2. Hedge Fund Returns (Unadjusted) By Age Decile, 1994-2000
24% 22% 20% 18% 16% 14% 12% 10% 0 20 40 60 80 100 120 140 160 y = -0.0482Ln(x) + 0.3457 R2 = 0.9262
Source: CrossBorder Capital, TASS/Tremont Usually defined as those offering a track record of less than three years. Also known as survivorship bias. 3 A negative or downward bias to future returns occurs when a very successful fund closes to new money and so fails to maintain its voluntary reporting. A positive or upward bias takes place when a poorly performing fund closes down.
2 1
CROSSBORDER CAPITAL
The above data make a strong case for investing in young funds. However, it has not been adjusted for survivor bias. Sceptics suggest that positive survivor bias unfairly flatters young funds because their allegedly higher failure rate is not captured in these crude figures. The sceptics are wrong. Figures 3 and 4 show the incidence of failure4 by age (i.e. the point when reporting stopped). Of the original 3,733 funds that reported in 1999, 341 or 9.1%, for some reason, failed to report in 2000. The breakdown of this subset of funds by age reveals a noticeable difference in failure rates between young and seasoned funds. Failure in the first year was, not surprisingly, low: managers were able to struggle on. Failure in year two was correspondingly higher. In each subsequent year, the failure rate fell. Overall, the probability of failure during the first three years at 4.2% was only slightly less than the 4.9% probability of failure for years four onwards. The large proportion of failures reported in year six and above (25.1%) is simply an aggregation of fund failures aged six years, seven years, eight years, etc.
Months
CROSSBORDER CAPITAL
The regression line fitted to the data in Figure 4 shows that the failure rate of hedge funds rises to a peak at 28 months and then decays at a roughly constant rate of 2-3% points per annum. These probabilities of failure (i.e. failure to report) can be applied to surviving funds to create a return series that shows the true costs and benefits of investing in funds of specific maturities. The first two columns of Figures 5 exclude all funds that reported in 1999 but failed to report in 2000. This adjustment covers funds that only reported in 1999, but not in 2000, as well as funds that reported from 1994-99 and not 2000. It should capture both negative (i.e. good funds that close) and positive (i.e. bad funds that fail) survivor bias. Both types of bias are likely to affect young funds. Managers that kick-off on the wrong foot may be forced to close down in the first year-or-two, thereby dragging down the true performance of the median young fund relative to older funds. Equally, successful new managers may quickly attract funds. If they then close to new money, it is probable that they will stop reporting their good performances. Young funds suffer most from negative survivor bias. Yet there is surprisingly little difference between the two data sets. The data for surviving funds slightly reinforce rather than dilute the case for young funds: around % per annum is added to the performance of the youngest three fund deciles after this first adjustment. Across the entire sample of 3392 funds the median ex post return only rises a tad to 13.6% from 13.4%. However, these ex post numbers do not show the expected returns from investing in different aged hedge funds. These are reported in the final column of Figure 5. Expected or ex ante returns are calculated by multiplying the survivors returns by the probability of future survival by age decile. We have assumed two things: (1) all failures are bad funds that fail, and (2) 30% of initial capital is lost5. These major constraints exaggerate the underperformance of failed funds6. Nonetheless, even with these heavy constraints, the expected returns shown in Figure 5 make a persuasive case for young funds. The youngest decile generates an expected return of 21.5%, or
Anecdotal evidence suggests that a fund that loses 30% of starting capital will likely close 6 Expected returns are constructed as survivors return times survival rate in decile a less initial capital times failure rate in decile a, e.g. for decile 1: 23.8% x 0.991 - 100% x 0.021 = 23.6% - 2.1% = 21.5%.
CROSSBORDER CAPITAL
760 basis points above the (estimated) median7 of 13.9%. This compares to 980 basis points for the crude data. The spread between the youngest and oldest deciles in the adjusted sample is 970 basis points, compared to 1,150 basis points in the crude data. For the second youngest decile, the excess return over the median drops to 290 basis points, against 710 basis points according to the crude data.
Decile 1 2 3 4 5 6 7 8 9 10 Median
A sample median does not exist. Consequently, we have estimated a median value by inserting the median age of the survivors sample into a regression equation that calculates expected returns by age.
CROSSBORDER CAPITAL
Figure 6. Hedge Fund Returns (Adjusted For Survivor Bias) By Age Decile, 1994-2000
26% 24% 22% 20% 18% 16% 14% 12% 10% 8% 0 20 40 60 80 100 120 140 160 y = -0.0356Ln(x) + 0.2729 R2 = 0.6542
CROSSBORDER CAPITAL
CROSSBORDER CAPITAL LTD, 2001. ALL RIGHTS RESERVED REGULATED BY THE SFA MARCOL HOUSE 289-293 REGENT STREET LONDON W1B 2HJ TELEPHONE 44 20 7535 0400 FACSIMILE 44 20 7535 0435 E-MAIL: HEDGEFUND@LIQUIDITY.COM CROSSBORDER CAPITAL BERMUDA LIMITED A SUBSIDIARY OF CROSSBORDER CAPITAL LIMITED REGULATED BY THE BMA 129 FRONT STREET PO BOX 1916 HAMILTON BERMUDA TELEPHONE 441 295 3294 FACSIMILE 441 292 2239 E-MAIL: JBARTON@IBL.BM This document researches a fund not authorised or recognised as a collective regulated investment scheme for the purposes of the UK Financial Services Act 1986 and is therefore for private circulation only and is not intended and must not be distributed to private investors. It is for information purposes only and does not offer any specific investment advice. Under no circumstances should it be used or considered as an advisory or offer to sell or a solicitation of any offer or advisory to buy any securities. The information in this document has been obtained from sources believed reliable, but we do not represent that it is accurate or complete, and it should not be relied upon as such. In addition any information contained herein does not imply any knowledge of the entire universe of such funds and CrossBorder Capital may or may not receive a fee from the fund manager for the promotion of such funds. Deduction for charges and expenses may not be made uniformly through the life of the investment. As a high volatility investment this fund may be subject to sudden falls in value and these could lead to a large loss on realisation which could equal the amount invested. An investor in the fund will not be liable to compensation for any losses. CrossBorder Capital points out that the value of all investments and the income derived therefrom can decrease as well as increase (this may be partly due to exchange rate fluctuations in investments that have an exposure to currencies other than the base currency of the fund). In addition that fund may from time to time use options, futures and warrants which are highly specialised activities and entail greater than ordinary investment risks. Thus a relatively small movement in the price of a security to which these relate may result in a disproportionately large percentage movement, unfavourable a well as favourable, in their price. The fund may gear itself by s other means as well. Past performance is no guide to future performance. Whilst given in good faith neither we nor any officer, employee, or agent of ours shall be liable for loss or damage, whether direct or indirect, which may be suffered by using or relying on the information, research, opinions, advice or recommendations contained herein or in any prior or subsequent written or verbal presentations. The employees of CrossBorder Capital Limited may have a position or otherwise be interested in funds mentioned in this report. This report may not be reproduced, distributed or published by any recipient for any purpose. CrossBorder Capital Limited is regulated by the SFA for the conduct of investment business in the UK.