1 Keith C. Brown Keith C. Brown The University of Texas The University of Texas W. Van Harlow W. Van Harlow Fidelity Investments Fidelity Investments Federal Reserve Bank of Atlanta Financial Federal Reserve Bank of Atlanta Financial Markets Conference Markets Conference April 15, 2004 April 15, 2004 Staying the Course: Mutual Fund Investment Style Consistency and Performance Persistence
37
Embed
Keith C. Brown The University of Texas W. Van Harlow Fidelity Investments
Staying the Course: Mutual Fund Investment Style Consistency and Performance Persistence. Keith C. Brown The University of Texas W. Van Harlow Fidelity Investments Federal Reserve Bank of Atlanta Financial Markets Conference April 15, 2004. Research Premise. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Keith C. BrownKeith C. BrownThe University of TexasThe University of Texas
W. Van HarlowW. Van HarlowFidelity InvestmentsFidelity Investments
Federal Reserve Bank of Atlanta Financial Markets Federal Reserve Bank of Atlanta Financial Markets ConferenceConference
April 15, 2004April 15, 2004
Staying the Course:Mutual Fund Investment Style Consistency
and Performance Persistence
2
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
R1 SP500
W4500
R2
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
R1 SP500
W4500
R2
Research PremiseResearch Premise
Lower Style Consistency
Does Investment Style Consistency Impact Performance?
Fund Outflows Due to Style DriftFund Outflows Due to Style Drift Inability of Plan Sponsors to Identify Manager’s StyleInability of Plan Sponsors to Identify Manager’s Style
Higher Consistency = Lower Turnover?Higher Consistency = Lower Turnover? Possibility of Lower Transaction Costs and Expense RatiosPossibility of Lower Transaction Costs and Expense Ratios
Style Timing Might be a “Loser’s Game”Style Timing Might be a “Loser’s Game” Analog to Difficulty of Successful Tactical Asset AllocationAnalog to Difficulty of Successful Tactical Asset Allocation
Style Consistency as a Possible “Signal” of Style Consistency as a Possible “Signal” of Superior Manager PerformanceSuperior Manager Performance
Vishny (JF, 1994); Fama and French (JF, 1998); Chan and Lakonishok (FAJ, Vishny (JF, 1994); Fama and French (JF, 1998); Chan and Lakonishok (FAJ, 2004); Phalippou (Working Paper, 2004)2004); Phalippou (Working Paper, 2004)
Style DefinitionsStyle Definitions: Roll (HES, 1995); Brown and Goetzmann (JFE, 1997): Roll (HES, 1995); Brown and Goetzmann (JFE, 1997) Style RotationStyle Rotation: Barberis and Shleifer (JFE, 2003): Barberis and Shleifer (JFE, 2003)
Fund Performance PersistenceFund Performance Persistence Classic Study: Classic Study: Jensen (JF, 1968)Jensen (JF, 1968) Hot & Icy HandsHot & Icy Hands: Grinblatt and Titman (JF, 1992); Hendricks, Patel, : Grinblatt and Titman (JF, 1992); Hendricks, Patel,
Zeckhauser (JF, 1993); Brown and Goetzmann (JF, 1995); Elton, Gruber, Zeckhauser (JF, 1993); Brown and Goetzmann (JF, 1995); Elton, Gruber, Blake (JB, 1996), Ibbotson and Patel (Working Paper, 2002)Blake (JB, 1996), Ibbotson and Patel (Working Paper, 2002)
Accounting for MomentumAccounting for Momentum: Jegadeesh and Titman (JF, 1993); Carhart (JF, : Jegadeesh and Titman (JF, 1993); Carhart (JF, 1997); Wermers (2001)1997); Wermers (2001)
Conditioning InformationConditioning Information: Ferson and Schadt (JF, 1996), Christopherson, : Ferson and Schadt (JF, 1996), Christopherson, Ferson, and Glassman (RFS, 1998)Ferson, and Glassman (RFS, 1998)
Grinblatt, Titman, and Wermers (JF, 1997)Grinblatt, Titman, and Wermers (JF, 1997) Pros: Pros: Direct Assessment of Manager’s Selection and Timing Direct Assessment of Manager’s Selection and Timing
Skills; Benchmark Construction Around Security CharacteristicsSkills; Benchmark Construction Around Security Characteristics Cons: Cons: Unobservable or Observed with Considerable Lag; Unobservable or Observed with Considerable Lag;
Returns-Based Measures: Returns-Based Measures: Sharpe (JPM, Sharpe (JPM, 1992)1992) Pros: Pros: Direct Observation of “Bottom Line” to Investor; Direct Observation of “Bottom Line” to Investor;
Measured More Frequently and Over Shorter Time Intervals Measured More Frequently and Over Shorter Time Intervals than Holdingsthan Holdings
Cons: Cons: Indirect Measure of Managerial Decision-MakingIndirect Measure of Managerial Decision-Making
10
Model BasedModel Based::
Define a style factor model:Define a style factor model:
[1 – [1 – RR22] represents portion of return not related to style] represents portion of return not related to style
Benchmark BasedBenchmark Based::
Active Net Returns: Active Net Returns:
TE =TE = where P is the return periods per yearwhere P is the return periods per year
Returns-Based Measures of Returns-Based Measures of Investment Style ConsistencyInvestment Style Consistency
Rjt = [ bj0 + ΣbjkFkt ]+ ejt
K
K=1
Δjt = Σ xji Rjit - Rbt = Rjt - Rbt
N
i=1
σΔ√P
11
Testable HypothesesTestable Hypotheses
Hypothesis #1Hypothesis #1: Style-consistent (i.e., high R: Style-consistent (i.e., high R22, low TE) funds , low TE) funds have lower portfolio turnover than style-inconsistent (i.e., have lower portfolio turnover than style-inconsistent (i.e., low Rlow R22, high TE) funds., high TE) funds.
Hypothesis #2Hypothesis #2: Style-consistent funds have higher total : Style-consistent funds have higher total and relative returns than style-inconsistent funds.and relative returns than style-inconsistent funds.
Hypothesis #3Hypothesis #3: There is a positive correlation between the : There is a positive correlation between the consistency of a fund’s investment style and the consistency of a fund’s investment style and the persistence of its future performancepersistence of its future performance
12
DataData
Survivorship-bias free database of monthly returns for Survivorship-bias free database of monthly returns for domestic diversified equity funds for the period 1988-2000domestic diversified equity funds for the period 1988-2000
Mutual Fund characteristics for the period 1991-2000 Mutual Fund characteristics for the period 1991-2000 (e.g., expense ratio, turnover, total net assets)(e.g., expense ratio, turnover, total net assets)
Require three years of prior monthly returns to be included Require three years of prior monthly returns to be included in the analysis on any given datein the analysis on any given date
No sector funds; analyze with and without index funds (i.e., No sector funds; analyze with and without index funds (i.e., active vs. passive management)active vs. passive management)
13
Number of Funds withNumber of Funds withThree Years of Returns (Table Three Years of Returns (Table
1)1)
YearLargeValue
LargeBlend
LargeGrowth
MidValue
MidBlend
MidGrowth
SmallValue
SmallBlend
SmallGrowth
1991 135 163 118 60 47 79 25 29 42
1992 140 172 120 60 49 78 28 30 44
1993 156 184 126 65 54 78 31 30 49
1994 169 203 139 67 54 82 38 37 59
1995 215 245 178 69 62 106 47 52 78
1996 273 314 233 87 71 150 62 71 113
1997 350 382 297 102 99 183 79 97 152
1998 410 446 355 127 104 221 97 123 206
1999 504 584 425 167 125 289 121 147 262
2000 564 729 549 199 138 333 162 194 309
14
Average Fund Characteristics: Average Fund Characteristics: 1991-20001991-2000(Table 2)(Table 2)
PeerGroup
AverageTurnover
AverageExpense
Ratio
Average Fund Firm Size ($mm)
Large Value 67.57% 1.38% 25,298
Large Blend 69.14% 1.22% 44,611
Large Growth 92.93% 1.45% 45,381
Mid Value 84.73% 1.43% 5,731
Mid Blend 79.39% 1.45% 6,782
Mid Growth 132.96% 1.55% 4,917
Small Value 61.43% 1.48% 643
Small Blend 82.17% 1.50% 1,283
Small Growth 119.89% 1.64% 1,057
15
MethodologyMethodology
Use two alternative returns-based definitions of style Use two alternative returns-based definitions of style consistencyconsistency Goodness-of-fit from a multivariate factor model (i.e., RGoodness-of-fit from a multivariate factor model (i.e., R22))
Tracking error relative to peer-group specific benchmarksTracking error relative to peer-group specific benchmarks
Evaluate the impact of style consistency on performance by Evaluate the impact of style consistency on performance by using a tournament-based methodology (Brown, Harlow, using a tournament-based methodology (Brown, Harlow, Starks (JF, 1996))Starks (JF, 1996)) Relative performance within a peer group is the focusRelative performance within a peer group is the focus
Avoids the usual model specification issuesAvoids the usual model specification issues
Controls for cross-sectional differences in consistency measuresControls for cross-sectional differences in consistency measures
16
MethodologyMethodology Multivariate Performance Attribution ModelMultivariate Performance Attribution Model
Factor ModelsFactor Models EGB Four Factor - Elton, Gruber and Blake (JB, 1996)EGB Four Factor - Elton, Gruber and Blake (JB, 1996) Modified EGB with Five Factors (adding EAFE factor)Modified EGB with Five Factors (adding EAFE factor) FF Three Factors - Fama and French (1993)FF Three Factors - Fama and French (1993) FFC Four Factors - Carhart (1997)FFC Four Factors - Carhart (1997)
Use RUse R22 and alpha from the model and alpha from the model
t
kt
k
t
where
R
R
=
=
=
=
=
a
b
e
. . .
. . .
the risk-adjusted excess return (alpha);
the excess return of a fund in month t;
the excess return of factor k in month t (k = 1 … N);
the beta of factor k (k = 1 … N);
the tracking error in month t;
t t t Nt tR R R R= + + + + + a b b b e1 2 . . . , N1 2
17
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
R1 SP500
W4500
R2
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
R1V
R2V
R1G
R2G
R1V
R2V
R1G
R2G
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
Value to Growth (%)
Cap
: Sm
all t
o L
arge
(%
)
R1 SP500
W4500
R2
R2 = 0.92 R2 = 0.78
Methodology (Figure 1)Methodology (Figure 1)
Examples from Multivariate Factor Model
Cap
: S
mall t
o L
arg
e (
%)
Value to Growth (%)
Cap
: S
mall t
o L
arg
e (
%)
Value to Growth (%)
18
Methodology (Table 3)Methodology (Table 3)
Style Group Style Consistency Median R2 Median Tracking Error (%)
Large Value Low 0.86 5.23 (LV) High 0.93 3.75
Large Blend Low 0.88 4.93
(LB) High 0.96 2.85
Large Growth Low 0.83 7.44 (LG) High 0.92 4.94
Mid Value Low 0.77 7.46 (MV) High 0.87 5.07
Mid Blend Low 0.75 8.24 (MB) High 0.87 4.89
Mid Growth Low 0.80 8.72 (MG) High 0.88 5.92
Small Value Low 0.75 7.85 (SV) High 0.87 5.02
Small Blend Low 0.77 8.40 (SB) High 0.89 5.85
Small Growth Low 0.81 8.74 (SG) High 0.90 6.60
19
MethodologyMethodology
Use past 36 months of data to estimate model parametersUse past 36 months of data to estimate model parameters
Evaluate performance in tournamentEvaluate performance in tournament Standardized returns within each peer group on a give date to Standardized returns within each peer group on a give date to
allow for time-series and cross-sectional poolingallow for time-series and cross-sectional pooling
Peer rankingsPeer rankings
Above median performanceAbove median performance
Roll the process forward one quarter (one year) and estimate Roll the process forward one quarter (one year) and estimate all parameters again, etc.all parameters again, etc.
Use past 36 months of data to estimate model parametersUse past 36 months of data to estimate model parameters
Run a sequence of cross-sectional regressions of future Run a sequence of cross-sectional regressions of future performance against fund characteristics and model parameters performance against fund characteristics and model parameters (alpha and R(alpha and R2 2 ))
Average the coefficient estimates from regressions across the Average the coefficient estimates from regressions across the entire sample periodentire sample period
T-statistics based on the time-series means of the coefficientsT-statistics based on the time-series means of the coefficients
Style Consistency Implications forReturns of Low and High Expense Ratio and Alpha Quintiles
(1991-2000)
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
1990
12
1991
06
1991
12
1992
06
1992
12
1993
06
1993
12
1994
06
1994
12
1995
06
1995
12
1996
06
1996
12
1997
06
1997
12
1998
06
1998
12
1999
06
1999
12
2000
06
Date
Gro
wth
of
a $1
35
3.07 % 0.85 % 1.89 %
2.40 % 0.54 % 0.19 %
(1.80 %) 7.16 % 4.60 %
Consistency PremiumsConsistency Premiums
Consistency Premiums by Style Groups
36
ConclusionConclusion Funds with more style consistency within a peer group tend to Funds with more style consistency within a peer group tend to
have better performance, ceteris paribus, during the sample periodhave better performance, ceteris paribus, during the sample period
Findings robust with respect to two alternative definitions of Findings robust with respect to two alternative definitions of consistency (and four factor models for one definition of consistency)consistency (and four factor models for one definition of consistency)
Results are not related to active/passive management issuesResults are not related to active/passive management issues
Style consistency effect appears to be separate from past alpha Style consistency effect appears to be separate from past alpha and expense ratios in explaining future performanceand expense ratios in explaining future performance
Results are robust within sample period and across fund typesResults are robust within sample period and across fund types
Although not reported, analysis of performance back to 1981 (not Although not reported, analysis of performance back to 1981 (not entirely survivorship-bias free) produces identical results to the entirely survivorship-bias free) produces identical results to the 1991-2000 analysis1991-2000 analysis
37
Extensions and ImplicationsExtensions and Implications
Need to Extend Analysis through 2003Need to Extend Analysis through 2003: : Same Behavior in “Down” Markets?Same Behavior in “Down” Markets?
Consistency as a “Signal” of PersistenceConsistency as a “Signal” of Persistence: : Easier to Identify Good Managers?Easier to Identify Good Managers?
Consistency and GovernanceConsistency and Governance: Manager : Manager Evaluation Relative to Peer Group; Evaluation Relative to Peer Group; Manager Compensation; Single vs. Team-Manager Compensation; Single vs. Team-Managed FundsManaged Funds
Consistency and RegulationConsistency and Regulation: Easier to : Easier to Assess Whether Fund Prospectus Assess Whether Fund Prospectus Objectives and Constraints are Satisfied?Objectives and Constraints are Satisfied?