Fundamental Factors in Hedge
Fund Returns
Stan Beckers
Simon Weinberger
Barclays Global Investors
Spitalfields Day
Cambridge 10 March 2005
0 BARCLAYS GLOBAL INVESTORS
Overview
The Raw Data: Issues
Skewness, Kurtosis and Autocorrelation
Communality in Hedge Fund Returns
Systematic and Residual Factors
1 BARCLAYS GLOBAL INVESTORS
1. Hedge Fund Returns : Data Issues
Return data only (no transparency)
Bias in Pricing/ Returns
• Survivorship Bias
• Instant History Bias
• Self-Reporting Bias
Short histories, low frequency data
Fund size ignored
2 BARCLAYS GLOBAL INVESTORS
HFR Database : Fund inception date,
Reporting start and end date
Number of Hed ge Funds laun ch ed 600
500
400
300
200
100
0
5% 3 6 5 5 4 5
24 BARCLAYS GLOBAL INVESTORS
Selected Findings from the Hedge Fund Literature
Equity Strategies tend to have market exposure and exposure to Fama/French
Factors SMB and HML (among others Fung and Hsieh, 2003)
30% of market-neutral funds have market risk exposure (Patton, 2004)
Option Strategies have explanatory power for non-directional strategies
(Agarwal and Naik, 2000)
Trend Following Strategy exhibits payoff similar to by lookback straddle (Fung
and Hsieh, 2001)
Merger Arbitrage exhibits payoff like uncovered put on Equity Index (Mitchell
and Pulvino, 2000)
25 BARCLAYS GLOBAL INVESTORS
Mean/Median Forecast Error at successive steps
Equity Hedge Funds
60
Mean
50 Median
40
Forecast Error
30
20
10
0
STP1
STP2
STP3
STP4
STP5
STP6
STP7
STP8
STP9
STP10
STP11
STP12
STP13
STP14
STP15
STP16
STP17
STP18
STP19
26 BARCLAYS GLOBAL INVESTORS
Explanatory power of a factor risk model
(Equity Funds)
Median : 35% in sample, 23% out of sample
Truncated Median: 41% in sample, 34% out of sample
15% Out-of-Sample
In-Sample
12%
9%
Frequenc y
6%
3%
0%
More
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
Adjusted Fit
27 BARCLAYS GLOBAL INVESTORS
28
Alpha p.a. (36m rolling estimation)
-5%
0%
5%
10%
15%
20%
25%
30%
199701
199707
199801
199807
199901
Median
199907
200001
200007
Top Quartile 200101
200107
200201
200207
Bottom Quartile
200301
200307
200401
Equity Hedge Fund In-Sample Alpha
BARCLAYS GLOBAL INVESTORS
Mean/Median Forecast Error at successive steps
Fixed Income Hedge Funds
25
Median
Mean
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
29 BARCLAYS GLOBAL INVESTORS
Explanatory power of a factor risk model
(Fixed Income Funds)
25%
Median Median*
Out-of-Sample
20%
in-sample 21.97% 31.19%
out-of-sample 10.26% 30.48% In-Sample
Frequency
15%
10%
5%
0%
More
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.05
0.15
0.25
0.35
0.45
0.55
0.65
0.75
Adjusted Fit
Histogram of in-sample (199401-200406) and out-of-sample fit (regression of fund returns on return
explained by risk model (product sum of prevailing exposure estimate and realised factor return)
Median* disregards observations with extreme fit (70%).
30 BARCLAYS GLOBAL INVESTORS
Fixed Income Fund In Sample Alpha
25%
20%
15%
Alpha p.a.
10%
5%
0%
199701
199707
199801
199807
199901
199907
200001
200007
200101
200107
200201
200207
200301
200307
200401
Median Q3 Q1
Fund-level rolling 36m estimation
31 BARCLAYS GLOBAL INVESTORS
HFR Hedge Fund Index Returns: Where is the Hedge?
T-Stat of Systematic Factors
LONG/ RELATIV FIXED GLOBAL FUND OF
SHORT E VALUE INCOME MACRO FUNDS
Adjusted R Squared 0.85 0.63 0.53 0.32 0.64
Intercept 2.29 1.54 2.91 2.77 2.02
Inflation % change 0.29 1.42 0.88 0.23 1.21
Inflation level -0.19 0.23 0.81 -0.45 0.39
VIX % change -0.19 -1.9 -0.62 0.35 -0.52
Vix level -0.73 0.74 -2.8 -2.13 -1.46
S&P 500 return 11.92 4.73 -1.32 3.44 5.1
Credit spread % change -0.47 -4.03 -2.9 -0.92 -2
US small - large return 9.43 2.8 0.19 3.36 4.9
US Value-growth return 6.22 -0.56 0.49 1.92 3.74
Slope of yield curve % Change 0.8 -2.5 -6.01 -0.41 -0.03
Slope of yield curve -3.78 -3.49 0.07 -0.62 -1.56
Risk appetite 1.97 2.99 -0.5 -0.97 1.54
32 BARCLAYS GLOBAL INVESTORS
Summary
Given the Quality of the Data, all Hedge Fund Empirical Research has to be
taken with a pinch of salt
Skewness, Kurtosis and Autocorrelation are less of an issue than some people
would lead you to believe
Hedge Fund Styles are not clearly delineated and somewhat arbitrary
Even so, Common factors can be identified within broad hedge fund style
classifications
Significant systematic factors are present in most hedge fund returns
Alpha does remain after taking systematic factors into account
Hedge Fund is somewhat of a misnomer
33 BARCLAYS GLOBAL INVESTORS