CHAPTER
4
CTA Performance,
Survivorship Bias,
and Dissolution Frequencies
Daniel Capocci
U
sing a database containing 1,892 funds (including 1,350 dissolved funds),
we investigate CTA performance and performance persistence to deter-
mine if some CTAs consistently and significantly outperform their peers over
various time periods. To test the persistence hypothesis, we use a methodol-
ogy based on Carhart’s (1997) decile classification. We examine performance
across deciles and across CTA strategies to determine if some deciles are
more exposed to certain strategies over time. We also analyze survivorship
bias and its evolution over time. We conclude the study by analyzing the dis-
solution frequencies across deciles and their evolution over time.
INTRODUCTION AND LITERATURE REVIEW
Unlike hedge funds, which appeared in the first academic journal in 1997,
commodity trading advisors (CTAs) have been studied for a longer time.
Many studies were published in the late 1980s and in the early 1990s
(see, e.g., Elton, Gruber, and Rentzler 1987, 1989, 1990; Edwards and Ma
1988). More recently, Billingsley and Chance (1996) and Edwards and Park
(1996) showed that CTA funds can add diversification to stocks and bonds
in a mean-variance framework. According to Schneeweis, Savanayana, and
McCarthy (1991) and Schneeweis (1996), the benefits of CTAs are similar
to those of hedge funds, in that they improve and can offer a superior risk-
adjusted return trade-off to stock and bond indices while acting as diversi-
fiers in investment portfolios.
Fung and Hsieh (1997b) showed that a constructed CTA style factor
persistently has a positive return when the Standard & Poor’s (S&P) has a
49

drawbacks of CTAs is that during bull markets, their performance is gener-
ally inferior to those of hedge funds.
Brorsen and Townsend (2002) show that a minimal amount of per-
formance persistence is found in CTAs, and there could exist some advan-
tages in selecting CTAs based on past performance when a long time series
of data is available and accurate methods are used.
This chapter aims to detect performance persistence of CTAs. We want
to determine if some CTAs consistently outperform their peers over time. In
50 PERFORMANCE
1
A lookback call is a normal call option, but the strike depends on the minimum
stock price reached during the life of the option. A lookback put is a normal put
option, but the strike depends on the maximum stock price reached during the life
of the option.
c04_gregoriou.qxd 7/27/04 11:05 AM Page 50
the next section, we describe the database, reporting the descriptive statis-
tics of the funds and analyzing the correlation between the various strate-
gies reported. The following section focuses on survivorship bias. We
analyze the presence of this bias over the whole period studied but also over
different time periods, including a bull and a bear market period. Further,
we report the methodology used to analyze CTA performance and per-
formance persistence before reporting the results of the performance analy-
sis in the next section. The next section reports the results of the persistence
analysis and analyzes the exposure of the deciles constructed on previous
year’s performance to the individual strategies. Then we report the complete
analysis of monthly and yearly dissolution frequencies.
DATABASE
In this section, we present our database and analyze the descriptive statistics
of the data before reporting the correlation between the various strategies.
Descriptive Statistics

issue, we report each fund in one strategy only.
2
Before entering the body of the study, we analyze the composition of
the database. Table 4.2 reports the descriptive statistics of the database.
Funds are classified according to strategy. The last line reports the statistics
for the whole database.
52 PERFORMANCE
TABLE 4.1 Grouping of Barclay Trading Group Strategies
Grouped CTA Barclay Trading
Strategies Group Strategy
Technical Diversified Technical Diversified
Technical Financial/Metals Technical Financial/Metals
Technical Currency Technical Currency
Other Technical Technical Interest Rate
Technical Energy
Technical Agricultural
Fundamental Fundamental Diversified
Fundamental Interest Rate
Fundamental
Financial/Metals
Fundamental Energy
Fundamental Currency
Fundamental
Agricultural
Discretionary Discretionary
Systematic Systematic
Stock Index Stock Index
Arbitrage Arbitrage
Option Strategies Option Strategies
No Category No Category

Min = minimum; Max = maximum. The Sharpe ratio is calculated with a 5 percent risk-free rate.
Note: The other technical strategy funds exist only for the August 1985–May 1995 period and for the October 1998–April 2001
period. Option strategy funds exist since September 1990.
53
c04_gregoriou.qxd 7/27/04 11:05 AM Page 53
Table 4.2 indicates that the systematic strategy is the most represented
strategy (with 897 funds) followed by total technical funds (416 funds) and
discretionary funds (299 funds). Other technical funds, option strategy
funds, and fundamental funds count only 8, 9, and 19 funds respectively.
The database contains 611 dissolved funds as a whole, 350 of which follow
the systematic strategy. Note that all the other technical funds and option
strategy funds are dissolved over the period studied. The median returns
indicate the same patterns.
Regarding the statistics, the highest mean monthly return is achieved
by the other technical funds (with 3.18 percent per month) followed by
the option strategy funds and discretionary funds (with 2.62 percent and
2.03 percent per month). Many strategies offer a monthly return of between
1.6 percent and 1.9 percent per month. The lowest returns are those of the
arbitrage funds (with 1.25 percent) followed by the technical currency
funds (with a monthly return of 1.58 percent). All the monthly returns are
significantly different from zero over the period studied.
The fundamental funds and the other technical funds are the more
volatile funds with a standard deviation of 7.60 and 7.25 percent. Because
there are few funds applying these strategies, there is no diversification
effect, which can explain why the returns of these strategies are so volatile.
The strategies that offer the most stable returns are the discretionary funds
(with a standard deviation of 3.01 percent) and the arbitrage funds (with a
standard deviation of 3.19 percent).
As one could expect, the strategies that are the most volatile also have
the lowest minimum return and the highest maximum return. The monthly

Tecdiv 0.93 −0.13 0.42 0.22 0.03 0.14 0.89 0.56 1.00 0.66 0.05 0.73
Tecfin 0.73 −0.05 0.27 0.20 0.11 0.13 0.71 0.56 0.66 1.00 0.10 0.50
Tecoth 0.14 0.24 0.00 −0.02 0.02 0.01 0.18 0.12 0.05 0.10 1.00 0.9
Nocat 0.81 −0.01 0.32 0.12 0.12 0.29 0.79 0.56 0.73 0.50 0.09 1.00
AllCTA = CTA Global Index; Arb = arbitrage; Discret = discretionary; Funda = fundamental; Stock = stock index; System =
systematic funds; Teccur = technical currency; Tecdiv = technical diversified; Tecfin = technical financial/metals; Tecoth = other
technical; Nocat = no category.
c04_gregoriou.qxd 7/27/04 11:05 AM Page 55
56 PERFORMANCE
global index is almost exactly correlated with the systematic funds. This can
be partly explained by the fact that this strategy contains the greatest num-
ber of funds. Forty-four coefficients out of sixty-six (66 percent of the co-
efficients) are under 0.5, indicating that most of the strategies are not
correlated. The lowest coefficient is the one between arbitrage and system-
atic funds at −0.21. There are nine negative coefficients in total represent-
ing 14 percent of the coefficients.
SURVIVORSHIP BIAS
Performance figures are subject to various biases. One of the most impor-
tant is the survivorship bias that appears when only surviving funds are
taken into account in a performance analysis study. The common practice
among suppliers of CTA databases is to provide data on investable funds
that are currently in operation. When only living funds
4
are considered, the
data suffer from survivorship bias because dissolved funds tend to have
worse performance than surviving funds.
Survivorship bias has already been studied. Fung and Hsieh (1997b)
precisely analyzed this bias and estimated it at 3.4 percent per year. They
also concluded that survivorship bias had little impact on the investment
styles of CTA funds. Returns of both surviving and dissolved CTA funds

0.80
0.90
1.00
88 88 89 89 90 90 91 91 92 92 93 93 94 94 95 95 96 96 97 97 98 98 99 99 00 00 01 01 02 02
FIGURE 4.1 Evolution of the Survivorship Bias (3-year Rolling Period)
Our database contains 1,899 CTAs (611 survived funds and 1,288 dissolved funds
as of December 2002). Numbers on the vertical axis are monthly percentages.
TABLE 4.4 Survivorship Bias Analysis over Different Periods
Bias 1985–2003 0.5 per Month
5.4 per Year
Bias 1985–1989 0.5 per Month
5.5 per Year
Bias 1990–1994 0.6 per Month
7.3 per Year
Bias 1995–1999 0.5 per Month
6.2 per Year
Bias 2000–2003 0.4 per Month
4.4 per Year
Our database contains 1,899 CTAs (611 survived funds and
1,288 dissolved funds as of December 2002).
c04_gregoriou.qxd 7/27/04 11:05 AM Page 57
58 PERFORMANCE
5
We take a month as a positive month if the whole database has a positive per-
formance. We consider a month as negative if the whole database does not reach
positive returns.
The figure indicates that the monthly bias ending January 1985
increases from around 0.7 percent at the beginning of the year to 0.85 per-
cent after summer before reaching the bottom of 0.9 percent at the begin-
ning of 1989. Afterward, it increases until January 1993 (0.9 percent) and

METHODOLOGY
The aim of this study is to determine if some CTAs consistently and per-
sistently outperform their peers. To achieve this objective, we construct a
CTA Global Index that contains all the funds present in our database and
c04_gregoriou.qxd 7/27/04 11:05 AM Page 58
one index per CTA strategy. To test if some funds significantly outperform
the indices, we use the following regression.
R
pt
= a
P
+ b
p1
R
It
+ e
pt
(4.1)
p = 1 to 1,899 and t = 1 to 216
where R
Pt
= return of CTA p at period t
R
It
= return of the index considered at period t
We run this analysis for each fund compared to the whole CTA data-
base index but also for each fund compared to its strategy index. Once we
obtain results, we want to determine if momentum is present in CTA
returns. Active CTA selection strategies could increase the expected return
on a portfolio if CTA performance is really predictable. We define the

(4.2)
P = 1 to 10 and t = 1 to 216
where R
Dt
= return of decile P at period t
R
It
= return of the 12 indexes (CTA Global Index, technically
currency, technically diversified, technically financial/metals,
technically others, stock index, options, systematic, arbitrage,
discretionary, fundamental, no category) at period t
We regress each decile against the CTA Global Index and each strategy
index. Doing so, we determine if some deciles are exposed to some strate-
gies, which indicates that that strategy is particularly present in the corre-
sponding decile.
PERFORMANCE ANALYSIS
Here we apply the model just discussed to our database to determine if
some strategies significantly outperform the CTA Global Index over differ-
ent time periods. In the next section we investigate whether momentum
exists in CTA performance.
Table 4.6 indicates some interesting results. First, we see that results are
different across strategies, indicating that the classification in substrategies
seems to be relevant. Second, the first column of the table reports the alpha
of the different strategies once the performance of the CTA database con-
sidered as a whole is taken into account through the CTA Global Index.
This is the performance not explained by the global CTA index. Seven out
of the 11 strategies are significantly positive at the 5 or 1 percent signifi-
cance level (technically financial/metals, technically currency, technically
other, discretionary, stock index, arbitrage, and option strategies); two are
not significantly different from zero (fundamental and no category); and

the CTA Global Index.
Finally, the R
2
column reports very different numbers. The R
2
ranges
from 0.00 for stock funds to 0.95 for systematic funds. As we could have
expected, the highest R
2
are obtained when the alphas are the lower and
particularly low when the beta is not significant.
Table 4.7 reports the same results over different subperiods. We divide
the analysis in three six-year periods (January 1985 to December 1990, Jan-
CTA Performance, Survivorship Bias, and Dissolution Frequencies 61
TABLE 4.6 Relative Performance Analysis of Strategy Indices
Alpha CTA Index R
2
Technically diversified −0.28
***
1.14
***
0.92
Technically financial and metals 0.65
**
0.64
***
0.38
Technically currency 0.92
***
0.38

0.53 0.03
No category 0.16 0.83
***
0.74
This table reports the results of the regression from the strategy subindices to the
whole database for the January 1985 to December 2002 period except for techni-
cally others (August 1985 to May 1995 and October 1998 to April 2001) and for
option strategies (September 1990–December 2002).
t-stat are heteroskedasticity consistent.
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*Significant at the 10 percent level.
Numbers in the table are monthly percentages.
c04_gregoriou.qxd 7/27/04 11:05 AM Page 61
TABLE 4.7 Subperiod Performance Analysis of the Various CTA Strategies
Panel 1: Subperiod analysis
Jan 1985–Dec 1990 Alpha CTA Index R
2
Jan 1991–Dec 1996 Alpha CTA Index R
2
Tech divers. −0.52
**
1.20
***
0.94 Tech divers. −0.08 1.04
***
0.89
Tech fin/met 1.55
**
0.58

Systematic −1.25
***
1.30
***
0.96 Systematic −0.50
***
1.48
***
0.97
Stock 3.66
***
−0.28 0.03 Stock 0.54
*
0.44
**
0.17
Arbitrage 2.49
***
0.15
**
0.10 Arbitrage 0.54
***
0.08 0.01
Option NA NA NA Option 1.29 0.79 0.04
No category 0.50
*
0.90
***
0.78 No category 0.04 0.62
***

*
−0.29 0.01 Tech other NA NA NA
Fundamental −0.35 0.47 0.03 Fundamental 0.72 0.07 0.00
Discretionary 0.62
***
0.24
***
0.13 Discretionary 0.81
**
0.37
**
0.15
Systematic −0.26
***
1.37
***
0.98 Systematic −0.37
***
1.4
***
0.98
Stock 1.21
***
0.20 0.03 Stock 2.67
***
−0.20 0.02
62
c04_gregoriou.qxd 7/27/04 11:05 AM Page 62
TABLE 4.7 (continued)
Jan 1997–Dec 2002 Alpha CTA Index R

***
0.85
Tech fin/met 0.15 0.49
***
0.40 Tech fin/met 0.14 0.71
***
0.51
Tech currency 0.22 0.51
***
0.39 Tech currency 0.36
***
0.58
***
0.33
Tech other 3.43
*
−0.50 0.04 Tech other 1.72
*
−0.23 0.01
Fundamental −1.81
*
0.93
*
0.11 Fundamental 0.62 0.34 0.02
Discretionary 0.56
***
0.17
**
0.11 Discretionary 0.70
***

0.18 0.00
No category 0.28 0.54
***
0.45 No category 0.33
***
0.48
***
0.49
t-stat are heteroskedasticity consistent.
Tech. divers. = technical diversified; tech. fin/met = technical financial/metals; tech. cur = technical currency; tech. other = other
technical; stock = stock index.
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*Significant at the 10 percent level.
Numbers in the table are monthly percentages.
63
c04_gregoriou.qxd 7/27/04 11:05 AM Page 63
uary 1991 to December 1996, January 1997 to December 2002) in Panel 1
before isolating bull and bear market periods in the last subperiod in Panel
2. These periods are January 1998 to March 2000 for the bull market and
April 2000 to December 2002 for the bear market. This last analysis is par-
ticularly interesting because we can determine how the strategies perform
compared to their peers during a bull and a bear market. For information
purposes we also include a 10-year analysis in Panel 3.
Results reported in Panel 1 indicate that few alphas change sign over the
subperiods, and no alpha that was significantly positive or negative for the
whole period becomes significantly negative or positive over the subperiods.
The first line indicates that technically diversified funds underperform
the CTA Global Index over each subperiod, but this underperformance is
significant only over the first and last subperiods. Panel 2 indicates that this

64 PERFORMANCE
c04_gregoriou.qxd 7/27/04 11:05 AM Page 64
perform during the whole January 1985 to December 2002 period do not
significantly deviate from the index over the last 10 years.
7
Regarding the exposure to the index, some strategies (technically diver-
sified, technically financial/metals, technically currency, discretionary, sys-
tematic) are always significantly exposed whereas others (technically other
and arbitrage funds) are exposed over some subperiods without always
being exposed. Fundamental, stock, and options funds are never or almost
never exposed to the index.
The adjusted R
2
does not change heavily over the subperiods analyzed.
The biggest variations in this coefficient occur for technically financial/metals
from 0.30 for the January 1985 to December 1990 period to 0.62 for the
January 1991 to December 1996 period, for technically currency funds
from 0.08 over the January 1985 to December 1990 period to 0.48 for the
January 1991 to December 1996 period and for the no-category funds from
0.78 for the January 1991 to December 1996 period to 0.39 over the
December 1997 to December 2002 period.
Individual Fund Results
In this subsection we determine if the results obtained for the whole data-
base are confirmed for individual funds. We will not report the results
obtained for all the funds, but we will summarize. The first step in this analy-
sis is to apply a filter on the database. To be included in the database, each
fund must have at least 24 months of data. We delete 385 funds to reach a
total of 1,508 funds. Then we apply the model to each individual fund
regressed over the CTA Global Index. Results are summarized in Table 4.8.
The table indicates that 13.7 percent of the funds significantly outper-

at the 5 percent significance level.
These results are interesting because they indicate that, as a whole, 21.7
percent of the funds significantly outperform the CTA Global Index while
15.4 percent significantly underperform. Outperformance is one thing; per-
sistence is another. It will be interesting to determine if this outperformance
is persistent and predictable or not. It is not surprising that most funds are
significantly exposed to the index. However, there are some funds that
are significantly negatively exposed to the index.
Table 4.9 reports descriptive statistics on the estimated coefficients. The
average alpha is 0.14 percent (median 0.107 percent) with a standard devi-
TABLE 4.9 Descriptive Statistics of the Individual Performance Estimation,
January 1985 to December 2002
Mean Std. Dev. Median Min Max
Alpha 0.14% 1.84 0.11% −8.06% 22.09%
CTA Global Index 0.89% 1.07 0.69% −6.24% 5.45%
R
2
0.18 0.21 0.09 −0.04 0.87
Min = minimum; Max = maximum.
Std. Dev. = standard deviation; t-stat are heteroskedasticity consistent.
Numbers in the table are monthly percentages.
c04_gregoriou.qxd 7/27/04 11:05 AM Page 66
ation of 1.84 percent. The average beta (in our case the beta is measured
relative to our CTA Global Index) is 0.89. This means that the average CTA
is not completely exposed to the market. This number can be compared to
the beta of a portfolio with an equity index like the S&P 500. The only dif-
ference is the reference index.
The average R
2
is 0.18 percent with a standard deviation of 0.21 per-

in magnitude. The highest ratios are those of poorly performing funds. This is
explained by the fact that the standard deviation is higher among the well-
performing funds.
CTA Performance, Survivorship Bias, and Dissolution Frequencies 67
c04_gregoriou.qxd 7/27/04 11:05 AM Page 67
Table 4.11 contains the results of the persistence analysis. The alpha
indicates that all deciles but decile 10 underperform relatively to the index.
Underperformance is significant only for D2, D4, D5, D6, D7, and D9.
These results indicate that when the performance of the index is taken into
account, most funds do not add value (they destroy value) over the January
1985 to December 2002 period. Interestingly, D10 (containing previous
year’s best-performing fund) has a positive but not significant alpha. All
deciles are positively exposed to the CTA Global Index, although D1, D6,
D8, and D9 are the only ones that are significantly exposed. The adjusted
R
2
obtained is quite high for each decile. However, for subdeciles (especially
those for D1), the R
2
is relatively low.
8
68 PERFORMANCE
TABLE 4.10 Decile Descriptive Statistics Based on Previous Year’s Performance
Mean Std. Sharpe
Return Dev. Median Min Max Skewness Kurtosis Ratio
D1 1.24 4.71 0.39 −8.37 30.38 1.69 7.07 0.17
D2 1.02 3.34 0.51 −5.70 20.67 1.74 6.39 0.25
D3 1.07 3.00 0.51 −4.34 15.21 1.66 4.79 0.27
D4 1.10 3.22 0.61 −6.97 19.86 1.97 7.84 0.25
D5 1.05 3.22 0.56 −6.07 24.42 2.59 14.65 0.26

positive. These interesting results indicate that the previous year’s best-
performing funds (around 10 percent of the whole database) significantly
outperform their peers over the bull market period. The results of subdecile
CTA Performance, Survivorship Bias, and Dissolution Frequencies 69
TABLE 4.11 CTA Persistence in Performance, January 1986 to December 2002
Mean Std. Dev. Alpha CTA Index R
2
adj
D1 1.24 4.71 −0.33 0.97
***
0.57
D2 1.02 3.34 −0.20
**
0.76
***
0.70
D3 1.07 3.00 −0.09 0.71
***
0.77
D4 1.10 3.22 −0.19
**
0.80
***
0.84
D5 1.05 3.22 −0.25
***
0.80
***
0.85
D6 1.35 3.79 −0.18

0.11
D10a 3.17 19.99 −0.39 1.04
***
0.49
D10b 1.90 14.77 −0.27 0.99
***
0.51
D10c 1.31 7.72 0.07 1.17
***
0.52
This table reports the performance analysis of the performance decile regressed
against the CTA Global Index.
t-stat are heteroskedasticity consistent.
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*Significant at the 10 percent level.
Numbers in the table are monthly percentages.
c04_gregoriou.qxd 7/27/04 11:05 AM Page 69
TABLE 4.12 Persistence in Performance Subperiod Analysis
Jan 1998– Bull Market Period Apr 2000– BearMarket Period Jan 1993– Ten-Year Period
Mar 2000 Alpha Index R
2
adj
Dec 2002 Alpha Index R
2
adj
Dec 2002 Alpha Index R
2
adj
D1 −1.05

*
0.85
***
0.82
D4 −0.31
**
0.98
***
0.86 D4 −0.01 0.90
***
0.91 D4 −0.05 0.81
***
0.85
D5 0.00 0.91
***
0.89 D5 −0.17
**
0.87
***
0.95 D5 −0.12
***
0.79
***
0.89
D6 −0.26
***
1.05
***
0.94 D6 0.08 1.03
***

***
1.29
***
0.86
D10 0.79
**
1.34
***
0.64 D10 −0.06 1.07
***
0.75 D10 0.38
**
1.21
***
0.71
D1a −0.02 2.13
*
−0.02 D1a −0.28 2.27
**
0.18 D1a −0.21 0.99
***
0.44
D1b 2.17 0.02 −0.04 D1b −1.73 0.88
**
0.00 D1b 0.04 0.44
***
0.20
D1c −0.78 0.94
**
0.03 D1c 1.15 0.32 -0.03 D1c 0.74

70
c04_gregoriou.qxd 7/27/04 11:05 AM Page 70
CTA Performance, Survivorship Bias, and Dissolution Frequencies 71
analyses are less significant.
9
The table also indicates that each decile is sig-
nificantly exposed to the CTA Global Index. The R
2
is particularly high,
especially for the upper deciles, but is generally low for the subdeciles.
The central part of Table 4.12 reports the decile analysis over the April
2000 to December 2002 period. This period corresponds to a bear market
since the technology bubble exploded in March 2000. It indicates that all
the deciles but D6 have negative alphas. The only one significantly negative
is D5. This result indicates that no group of funds offers persistent returns
during the bear market that began in the first half of 2000. As expected, the
top-performing subdecile (D10c) yields a positive (but not significant)
alpha. Nevertheless, each decile is significantly positively exposed to the
CTA Global Index.
The right-hand part of Table 4.12 reports the analysis for the 10-year
period ending December 2002. In this last case, all deciles but D10 are neg-
ative, and most of them significantly destroy value (D1, D2, D5, D7, D8,
and D9 have all significantly negative alphas). As in the bull period ana-
lyzed before, D10 has a significantly positive alpha. This indicates that the
funds in this particular decile persistently create value compared to their
peers. The exposure to the market is significantly positive for all deciles,
and as in all the other cases, R
2
is high for each decile.
Strategies Analysis

0.03 0.50
**
0.02 −0.03 0.21 0.00 0.38 0.56
D2 −0.49
*
0.00 −0.09 0.05 0.00 0.08 0.37 0.28
***
0.38
**
−0.02 0.05
**
0.34 0.74
D3 −0.27
**
0.03 0.03 0.03 0.00 0.07 0.58
***
0.13
**
0.10 0.08 −0.01 −0.01 0.84
D4 0.22
*
−0.03 0.10
***
0.00 0.00 −0.06
*
0.71
***
−0.03 −0.05 −0.04 0.00 −0.05 0.87
D5 0.17 −0.06
**

**
0.00 −0.01 −0.02 0.80
***
−0.13
*
0.08 −0.03 0.00 −0.05 0.87
D9 −0.14 0.05 0.17
*
0.01 0.01 0.01 0.66
***
−0.06 0.16 0.16 0.00 −0.11 0.76
D10 −0.27 −0.02 0.31
**
0.03 −0.01 0.29 0.16 0.06 0.17 0.26
**
0.01 0.31
**
0.59
D10a −0.74
*
0.05 0.21 0.11
**
0.00 −0.03 0.56
**
−0.08 0.21 −0.08 0.04 0.56
*
0.34
D10b −0.03 −0.02 −0.21 0.01 0.01 0.05 0.30
*
−0.05 0.08 0.01 −0.05

c04_gregoriou.qxd 7/27/04 11:05 AM Page 72
CTA Performance, Survivorship Bias, and Dissolution Frequencies 73
and D10c are significantly negative. All these figures are different from the
ones obtained in the performance or performance persistence analysis.
The other columns report the exposition of each decile to the strategies
defined earlier. We analyze the table horizontally, then vertically, but first we
want to underline the fact that negative significant exposure of a decile to a
strategy means that the decile negatively contributes to the creation of alpha.
Decile D1 (the worst-performing funds) is significantly positively exposed to
discretionary and systematic funds and significantly negatively exposed
to option funds. The mean return for decile D1 is 1.24 percent (see Table
4.11). Once we take the strategy performance into account, the alpha is
−0.58 (See Table 4.13). The difference between these two numbers comes
mainly from the exposure to fundamental and systematic funds.
11
D2 is sig-
nificantly positively exposed to technical currency, technical diversified, and
technically other funds. Interestingly, this decile is not significantly exposed
to systematic funds. D3 is significantly positively exposed to systematic
funds and to technical currency funds. D4 is positively exposed to discre-
tionary funds and to systematic funds. D5 is significantly negatively
exposed to arbitrage funds and significantly positively exposed to discre-
tionary, option strategies, and systematic funds.
D6 is significantly positively exposed to systematic funds and techni-
cally diversified funds and negatively exposed to technical financial/metal
funds. D7 is significantly positively exposed to systematic funds and nega-
tively exposed to technical financial/metals, whereas D8 is positively ex-
posed to discretionary funds and systematic funds. In this particular case,
the strategies reported cannot completely explain the alpha (since it is still
weakly significantly positive). D9 is significantly positively exposed to dis-