The Mutual Fund Industry Worldwide: Explicit and Closet Indexing

The Mutual Fund Industry Worldwide: Explicit and Closet Indexing,
Explicit and Closet Indexing, Fees, and Performance
Martijn Cremers, Miguel Ferreira, Pedro Matos and Laura Starks
1
The value of active management
Academic evidence regarding the value of active management in mutual funds
• Against
– Sharpe (1966), Jensen (1968), Gruber (1996), French (2008), Fama and French (2010), and others
• For
– Grinblatt and Titman (1989, 1993), Kacperzyck, Sialm and Zheng (2008), Cremers and Petajisto (2009), and others
2
But: the fund market speaks for itself
• What percentage of mutual fund assets worldwide is managed passively through index funds or exchange‐
traded funds?
16%
• Outside the U.S. the total is 7%
• In the U.S. the total is 20%
93% is actively managed!
80% is actively managed!
80% is actively managed!
3
Question 1
What is the degree of passive versus active g
p
fund management around the world?
•
•
•
Explicit indexing: (enhanced) index funds and Explicit
indexing: (enhanced) index funds and
passive, index‐tracking ETFs
Cl t I d i
Closet Indexing: funds claiming to be active but f d l i i t b
ti b t
whose holdings are ‘similar’ to their b h
benchmark, or with a low Active Share
k
ith l
A ti Sh
Active management: funds that are truly active in terms of distance from benchmark holdings
4
Question 2
How are the levels and nature of active management associated with shareholder outcomes?
associated with shareholder outcomes?
•
•
Fees: ‘total shareholder costs’ = total expense ratio + front‐end load / 5
total expense ratio + front end load / 5
Performance: relative to self‐declared benchmark
• Benchmark‐adjusted return
B h
k dj t d t
• Alpha of Benchmark‐adjusted return
• Information Ratio of Benchmark‐adj. return
f
i
i f
h
k dj
5
Question 3
What explains the levels and nature of active and passive management across countries? i
t
ti ?
•
•
•
Competition: industry concentration, foreign presence, presence + cost of explicit indexing, financially f
li i i d i fi
i ll
sophisticated investors
R l ti
Regulation: setup cost to start new fund, setup time to start t
tt t t
f d t ti
t t t
new funds
Investment opportunities: number of stocks concentration in
Investment opportunities: number of stocks, concentration in benchmark, liquidity, market cap, volatility
6
These are important questions
As of June 2010, worldwide
,
>68,000 mutual funds >$21 trillion in AUM
7600 U.S.
11%
60,400 non‐U.S.
89%
$10.5 trillion U.S.
<50%
$10.5 trillion non‐U.S.
>50%
7
Overview of Results
• Existence of passive management:
– Little explicit indexing around the world
Little explicit indexing around the world
– Sizeable amount of closet indexing • Implications for shareholder outcomes:
–P
Passive management is related to fees
i
ti l t dt f
• More explicit indexing exists ↔ lower fees
• More closet indexing exists ↔ higher fees
– Active management can outperform
• A 1 σ increase in Active Share associated with increase of
0.94%
0
94% per year in future benchmark‐adjusted returns and
per year in future benchmark adjusted returns and
0.50% per year in alpha.
• Active management is associated with indicators of a country’ss competitive
country
competitive and regulatory
and regulatory environment, but also environment but also
with more investment opportunities
8
Data
• Lipper Hindsight database – Characteristics and returns of open‐end equity mutual funds (called UCITS in Europe)
– Unit of observation: each fund’s primary share class
U it f b
ti
h f d’ i
h
l
UCITS = Undertakings for Collective
– 21,684 funds ($10.1 trillion in assets under management)
Investment
in Transferable
Securities
• Includes both active and dead funds
Includes both active
and dead funds
• Factset/Lionshares
F
/Li h
d b
database
– Mutual funds’ portfolio holdings – Coverage for > 80% of TNA of Lipper database
9
Sample
ƒ
10 145 funds ($7 9 trillion TNA) with holdings
10,145 funds ($7.9 trillion TNA) with holdings
ƒ
30 countries 30
countries
20 in Europe & North America + 10 in Asia‐Pacific
ƒ
Global funds, regional funds, country funds, sector funds
ƒ
2002‐2007 (being updated)
10
Benchmarks
Benchmark type (77 different benchmarks):
• COUNTRY (specific country – all stocks, industry or style, further divided into MAIN COUNTRY or OTHER))
• DOMESTIC Market => 4,136 funds, US$4.5 trillion
• FOREIGN
O G Market => 2,069 funds, US$ 0.6 trillion
k
2 069 f d
S$ 0 6 illi
• WORLD => 1,506 funds, US$ 1.4 trillion
1 506 f d US$ 1 4 t illi
• REGIONAL => 2,434 funds, US$1.5 trillion
2 434 f d US$1 5 t illi
11
International Sample TNA (2007, bln)
Cross‐section in 2007, sum of TNA of all funds with holdings: $ 6 700 billi
$ 6,700 billion
$536
$1,309
Global
R i
Regional
l
$1,218
Main Country
Main Country
$3,640
Other
12
Explicit Indexing
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
Poland
Perrcentage
e of TNA
A
Extent of Explicit Indexing
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
13
Truly Active versus Closet Index Funds
ƒ Decompose portfolio into two parts:
ƒ Distinguish truly active funds from those that are y
,
y
not very active, but only advertise themselves as active, i.e. closet index funds
14
Active Share Cremers and Petajisto, 2009
j ,
15
Tracking Error (Volatility)
ƒ Consider the same portfolio decomposition:
ƒ Expressed in terms of returns:
ƒ Tracking error is defined as the standard deviation of the active return:
ti
t
16
Active Share and Tracking Error and Passive Management
Passive Management
From Cremers-Petajisto, 2009
17
Example: Fidelity Magellan’s Period of Closet Indexing
Period of Closet Indexing
Peter Lynch
Morris J. Smith
Jeffrey N. Vinik
Robert E. Stansky
Harry Lange
100%
90%
A
Active Shar
e
80%
70%
60%
50%
40%
30%
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
20%
Now available in Morningstar:
g
18
18
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Active
e Share
Active Share
100%
90%
80%
70
70%
60
60%
50
50%
40
30
30%
20%
Total Net A
Assets ($bn)
Fidelity Magellan’s Active Share & Total Assets
Active Share & Total Assets
Assets
100
90
80
40%
20
10
0
19
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Active Share
Active Share
100%
90%
80%
140
120
70%
100
80
60%
60
50%
40%
TTotal Net Asssets ($bn)
More Assets, Less Active Management?
Growth Fund of America
Assets
200
180
160
40
20
0
20
Active Share percentage cut‐off
• Closet Indexer = Active Share below 60% – At least 40% of the fund portfolio holdings overlap with the benchmark index
• Subjective / arbitrary cutoff, but
– ex post, exactly half the holdings in any portfolio will beat the portfolio’s (average) return
– … thus a truly active fund manager should have an Active Share of clearly more than 50%
21
Explicit Indexing vs. Active Management in Worldwide Funds
Management in Worldwide Funds
explicit indexing
16%
active
i
84%
22
Explicit Indexing, Closet Indexing, and Truly Active % in Worldwide Funds
Truly Active % in Worldwide Funds
explicit indexing
16%
closet indexer*
indexer
18%
truly active
66%
*Active share < 60%
23
Explicit Indexing, Closet Indexing, and Truly Active % in Non‐U.S.
Truly Active % in Non
U.S. Funds
Funds
explicit
indexing
7%
closet indexer*
indexer
35%
truly active
58%
*Active share < 60%
24
Explicit Indexing
Closet Indexing
Truly Active
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
Poland
Perrcentage
e of TNA
A
Individual Country Analyses 100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
25
Explicit Indexing
Closet Indexing
Truly Active
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
Poland
Percentage
e of TNA
A
Global Funds Only
100%
90%
80%
70%
60%
50%
0%
40%
30%
20%
10%
0%
26
Explicit Indexing
Closet Indexing
Truly Active
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
Poland
Percentage
e of TNA
A
Regional Funds Only
100%
90%
80%
70%
60%
50%
0%
40%
30%
20%
10%
0%
27
Domestic Country Funds Only
(Fund domiciled in same country as the firms held)
(Fund domiciled in same country as the firms held)
100%
80%
70%
60%
50%
40%
30%
20%
10%
Explicit Indexing
Closet Indexing
Truly Active
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
0%
Poland
Perrcentage
e of TNA
A
90%
28
Foreign Country Funds Only
(Fund domiciled in a different country than the firms held)
(Fund domiciled in a different country than the firms held)
100%
80%
70%
60%
50%
40%
30%
20%
10%
Explicit Indexing
Closet Indexing
Truly Active
Netherlands
Ireland
Luxeembourg
Norway
USA
Denmark
D
Portugal
Finland
Asia Pacific
Canada
Germany
G
UK
Austria
Spain
Dublin
Belgium
Sweden
Italy
France
Switzerland
0%
Poland
Perrcentage
e of TNA
A
90%
29
Fund Fees (Total Shareholder Costs)
Calculated as
Total Expense Ratio + front‐end load/5
Table 3 shows (unsurprising) results across countries:
• Index funds and ETFs have lower costs
Index funds and ETFs have lower costs
• U.S. lowest costs • Differences across countries generally larger than Diff
i
ll l
h
differences within countries (across fund types) 30
Fees by Fund Type
ACTIVE Funds - TSC
ETF Funds - TSC
INDEX Funds - TSC
Austria
Austria
Austria
Belgium
Belgium
Belgium
Canada
Canada
Canada
Denmark
Denmark
Denmark
Dublin
Dublin
Dublin
Finland
Finland
Finland
France
France
France
Germany
Germany
Germany
Ireland
Ireland
Ireland
Italy
Italy
Italy
Luxembourg
Luxembourg
Luxembourg
Netherlands
Netherlands
Netherlands
Norway
Norway
Norway
Poland
Poland
Poland
P
Portugal
l
Portugal
Portugal
Spain
Spain
Spain
Sweden
Sweden
Sweden
Switzerland
Switzerland
Switzerland
UK
UK
UK
USA
USA
USA
Asia Pacific
Asia Pacific
Asia Pacific
0%
1%
2%
3%
4%
5%
31
0%
1%
2%
3%
4%
5%
0%
1%
2%
3%
4%
5%
Hypotheses on relation between fees and explicit or closet indexing
explicit or closet indexing
Indexing in a country is related to fees (i.e., competition)
– Presence of explicit index funds: • low cost alternatives to actively managed funds Implication: More explicitly indexed funds associated with lower fees for actively managed funds
– Presence of substantial closet indexing:
• less choice for investors in actively managed funds Implication: More closet indexing related to higher fees for actively managed funds
32
Slightly Lower Fees for Active Funds in Countries with more Explicit Indexing
in Countries with more Explicit Indexing
Average Fees and Explicit Indexing per Country
al Sharehold
der Costs (A
Active Funds))
Tota
.01
.02
.03
.04
Poland
Luxembourg
Austria
Italy
Canada
Ireland
UK
Dublin
Germany
France
Portugal
S it l d
Switzerland
Spain
Finland
Norway X_Other Asia Pacific
Denmark Belgium
Sweden
USA
Netherlands
0
.1
.2
% of TNA ((Index+ETF))
(mean) tsc_ter_mgt
.3
Fitted values
33
Much Higher Fees in Countries with more Closet Indexing
with more Closet Indexing
al Sharehold
der Costs (A
Active Funds
s)
Tota
.01
.02
.03
.04
4
Average Fees and Closet Indexing per Country
Poland
Luxembourg
Ireland
Austria
Canada
UK y
Germany
Dublin
Portugal
Spain
Finland
Norway
Belgium
Denmark X_Other Asia Pacific
USA
Italy
France
Switzerland
Sweden
Netherlands
.2
.4
.6
g ((% TNA Active Funds with AS<0.6))
% of Closet Indexing
(mean) tsc_ter_mgt
.8
Fitted values
34
Summary of Results for Shareholder Costs
Shareholder Costs • Lower Fees for –
–
–
–
•
•
•
•
Larger funds
Younger funds
Investment strategy of funds with a domestic focus
Funds with lower flows
Fees increase in a fund’s Active Share
Fees decrease with the level of explicit indexing in a country
Fees increase with the level of closet indexing in a country
Relation between fees and Active Share stronger for U.S. than for Non‐U.S. 35
Table V. Active Share and Indexing
Explicit indexing (by country)
Closet indexing (by country)
Explicit indexing (by country/type)
Closet indexing (by country/type)
R-squared (logistic)
Total shareholder cost
TNA (l
(log))
Fund age
y
Domestic dummy
Flows
Year dummies
Geographic focus dummies
Observations
R-squared
(1)
0.1782***
(3.13)
(7)
0.0608
(1.56)
0.0226
(0.71)
-0.1984*
0.1101**
0.0951*
(-1.97)
(2.39)
(1.99)
-0.2660*** -0.2881*** -0.2902***
((-18.54)
18.54)
((-14.58)
14.58)
((-12.80)
12.80)
-0.0912*** -0.0877*** -0.0876*** -0.0905*** -0.0854*** -0.0854*** -0.0857***
(-12.72)
(-12.25)
(-12.11)
(-12.84)
(-12.20)
(-12.12)
(-12.00)
0.0084
0.0091*
0.0092*
0.0076
0.0072
0.0075
0.0074
(1.67)
(1.68)
(1.69)
(1.47)
(1.49)
(1.52)
(1.53)
-0.0037**
0 0037** -0.0042***
0 0042*** -0.0042***
0 0042*** -0.0036**
0 0036** -0.0039**
0 0039** -0.0040**
0 0040** -0.0039**
0 0039**
(-2.18)
(-2.78)
(-2.81)
(-2.02)
(-2.35)
(-2.48)
(-2.39)
-0.0010*** -0.0010*** -0.0010*** -0.0009*** -0.0008*** -0.0008*** -0.0008***
(-5.98)
(-5.48)
(-5.59)
(-5.03)
(-5.34)
(-5.42)
(-5.36)
0.0801
0.0642**
0.0401
0.1206**
0.0146
-0.0013
-0.0054
(1.55)
(2.13)
(1.12)
(2.54)
(0.60)
(-0.05)
(-0.20)
0.0067**
0.0051*
0.0053*
0.0061*
0.0026
0.0025
0.0026
(2.12)
(1.84)
(1.92)
(1.97)
(0.97)
(0.95)
(1.01)
Yes
Yes
28198
0.580
(2)
(3)
0.1713***
(3.74)
-0.2266*** -0.2252***
(-8.26)
(-8.24)
Yes
Yes
28198
0.589
Yes
Yes
28198
0.591
(4)
(5)
(6)
Yes
Yes
28198
0.583
Yes
Yes
28198
0.628
Yes
Yes
28198
0.629
Yes
Yes
28198
0.630
36
Summary: where managers are more active
Active Share • Increasing in explicit indexing in country
I
i i
li it i d i i
t
– More explicit indexing means more low‐costs alternatives, thus more competition?
alternatives, thus more competition?
– Effects strongest at country/type ‐ level
• Decreasing in closet indexing in country
– Partly mechanical
• Decreasing in a fund’s R2, TNA, and age
• Increasing in investment opportunities
– Liquidity, more stocks, larger stocks, volatility, less concentrated benchmark larger stock market
concentrated benchmark, larger stock market
37
Example: funds benchmarked to the S&P 500 versus Madrid SE index • Domestic funds benchmarked to Madrid SE index
– About 75 domestic active funds in sample
– Average Active Share of 37%, medium of 32%
– Average (per year): Turnover 79%, Volatility 24%, Market cap (US$) 630 million, about 150 stocks
• Domestic funds benchmarked to S&P 500 index
– Over 200 domestic active funds in sample
– Average Active Share of 70%, medium of 70%
– Average (per year): Turnover 207%, Volatility g (p y )
y
30%, Market cap (US$) 14.6 billion, over 9,000 stocks
38
U.S. Equity Sample: Evolution of Active Share over Time
Evolution of Active Share over Time
39
Active Share and Performance?
• Closet Index Funds
• Similar holdings to benchmark funds
Si il h ldi
b h
kf d
• Higher costs than passive, explicit index funds
• Expected to underperform benchmarks
• Most Active Funds
• Holdings very different from their benchmarks
• Strong conviction / guts
• Could measure skill if competitive environment allows only most skillful managers to survive with a truly active portfolio
40
Does Active Share Pay Off in Terms of Future Performance?
in Terms of Future Performance?
Previous evidence from U.S. (Cremers‐Petajisto, 2009)
Excess Net Returns next year
(relative to benchmark, after expenses)
3.00%
2.70%
Perfo
ormance per year
2.50%
1.87%
2.00%
1.50%
1.00%
1.00%
0.50%
0.00%
0.08%
1
2
3
-1.10%
-1.07%
0.00%
0 50%
-0.50%
-1.00%
-1.50%
-0.96%
4
5
-0.44%
-0.91%
-2.00%
PRIOR YEAR RETURN Quintile
Stock Pickers (highest 20% Active Share)
Closet Indexers
41
Active Share & Batting Average:
% of funds beating the benchmark (1 year)
% of funds beating the benchmark (1 year)
65%
60%
55%
50%
45%
<60% Active Share
40%
All Funds
All Funds
>90% Active Share
35%
30%
25%
20%
All Styles
Global
Regional
Main Country
Country or Sector
42
Active Share & Annual Performance:
Raw performance net of the benchmark
Raw performance net of the benchmark
4.00%
3.00%
2.00%
1.00%
<60% Active Share
All Funds
All Funds
>90% Active Share
0.00%
All Styles
‐1.00%
Global
Regional
Main C
Country
t
Country or S t
Sector
‐2.00%
%
Per year
‐3.00%
43
Where is the predictability strongest?
• Active Share predicts future performance
– For all fund styles, but strongest for global funds
For all fund styles but strongest for global funds
• Predictability strongest in markets with
– Less foreign competition
– Less financially sophisticated investors
– Higher mutual fund setup costs
– Less explicit indexing (less competition?)
– More investment opportunities
• All of these may make it easier for active managers y
g
to outperform
44
Active Share & Annual Performance:
Benchmark‐aadjusted
d Annuall Perform
mance
Step‐wise
Step
wise linear regression evidence
linear regression evidence
4%
2%
0%
‐2%
‐4%
‐6%
Active Share
All Fund Styles
Global
Regional
g
Main Country
Country or Sector
45
Conclusions
ƒ Little explicit indexing, but lots of closet indexing
ƒ Shareholder Costs
Sh h ld C
ƒ negatively related to existence of low‐cost explicit index funds
ƒ positively related to prevalence of closet indexing
ƒ Returns to active management
ƒ “truly active” funds (high Active Share) outperform their benchmarks
ƒ “closet indexers” (low Active Share) underperform
… both depending on country factors (investment both depending on country factors (investment
opportunities, competition, regulation)
46