RESEAR CH
Methodology for Evaluation of Hedge Fund Decisions Dealing with the challenges MARAT MOLYBOGA, EFFICIENT CAPITAL MANAGEMENT and CHRISTOPHE L’AHELEC, ONTARIO TEACHERS’ PENSION PLAN
A
s the hedge fund industry has grown to about $3 trillion in assets under management (AUM) in the first quarter
of 2015 and represents a significant portion of institutional portfolios, academic and industry publications have provided valuable insights into certain aspects of portfolio management, particularly in the area of fund evaluation and portfolio construction. Yet there is an apparent lack of a robust and flexible methodology that is capable of evaluating whether those insights can benefit a specific institutional investor once implemented with real world constraints.
There are several important challenges that need to be carefully dealt with. One – investors have their own objectives that vary substantially depending on the type of institution. For example, a family office or an asset management firm might seek to maximise Sharpe ratio, a university endowment attempts to target returns that exceed the university’s spending rate over a market cycle, and a pension fund pursues maximisation of risk- adjusted return within an asset-liability framework.
Two – sophisticated investors often impose rigorous filtering criteria or show stoppers such as the length of a track record and level of AUM. Most academic studies either completely ignore these show stoppers or selectively incorporate some of them with the purpose of accounting for certain biases such as small fund bias or incubation bias.
While accounting for biases is important, an institutional investor ultimately wants to know whether he will be able to benefit from a portfolio management technique given his own set of preferences and constraints. Three – most academic papers often compare portfolios that include hundreds of funds, which might be irrelevant to an investor who plans to hire three to five hedge funds. By contrast, such investor would be interested in knowing the impact of his manager selection and portfolio construction decisions on the outcome distribution.
Generating out-of-sample results for multiple subsets of five manager portfolios within in a simulation framework gives that information. Four – the investor cares about the marginal impact of the hedge fund investment on his existing portfolio but that is often ignored. Finally, hedge fund databases provide returns with a delay of about one month that is ignored in academic papers, as reported in Molyboga, Baek and Bilson (2015), creating a significant barrier to implementing results of most studies.
In our research paper, “A simulation-based methodology for evaluating hedge fund
38 Fig.1 Distribution of Sharpe 1.0 0.8 0.6 0.4 0.2 0.0 Source: Efficient
BLEND
Fig.2 Distribution of Calmar
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
Source: Efficient
BLEND
investments,” we introduce a methodology that is designed to evaluate hedge funds’ investments subject to the realistic constraints of institutional investors. The methodology is customisable to the real-life preferences and constraints of investors, including investment objectives, performance benchmarks, desired number of funds in a portfolio, and rebalancing frequency.
We illustrate the methodology by imposing the framework on a dataset of Commodity Trading Advisors with 604 active and 1,323 defunct funds over the period 1993-2014. We then measure the out-of-sample performance of three hypothetical risk-parity portfolios and two hypothetical
minimum risk portfolios and their marginal contributions to a typical 60-40 portfolio of stocks and bonds.
We find that an investment in CTAs improves performance regardless of the choice of the portfolio construction approach. For the out- of-sample period between January 1999 and December 2014, a 10% allocation to managed futures improves the Sharpe ratio of the original 60-40 portfolio of stocks and bonds from 0.376 to 0.399-0.416 on average, depending on the portfolio construction methodology. Blended portfolios have higher Sharpe ratios in at least 89% of simulations and higher Calmar ratios in at least 89.5% of
OUT-OF-SAMPLE 1/1999 - 12/2014, 10,000 SIMULATIONS, 5 CTAs
OUT-OF-SAMPLE 1/1999 - 12/2014, 10,000 SIMULATIONS, 5 CTAs
95.5
95.50
90.10
90.10
80.20
80.20
70.30
70.30
60.40
60.40
50.50
50.50
40.60
40.60
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73