Playing the Market:Turnover of Institutional
Ownership and Stock Returns
Valentin DimitrovRutgers University
Vladimir GatchevUCF
February 5, 2010
Institutional Investors
Institutional investors play a dominant role in U.S. equity markets
According to the Conference Board, in 2006 the market value of total institutional equity holdings was $12.9 trillion, accounting for 66.3% of total equity
Institutions are active traders Boehmer and Kelley (2009) report that institutional
investors account for the majority of trading volume on NYSE
We find that institutional investors turn over 37% of their ownership per quarter
There is substantial variation in institutional turnover rates across stocks (0% to 100% per quarter)
Motivation
What are the implications of turnover rates by institutions for stock prices?
Share turnover rates and disagreement Karpoff (1986) Harris and Raviv (1993)
Disagreement and stock valuation Miller (1977)
A more comprehensive theory Harrison and Kreps (1978) Scheinkman and Xiong (2003)
Scheinkman and Xiong (2003)
Two classes of investors disagree about fundamentals
The relative valuations of the two classes fluctuate over time
When valuations cross, the optimists buy the shares from the pessimists, leading to share turnover
With short-sales constraints, share ownership comes with an option to sell to more optimistic investors
Share prices include a premium for the option to sell Investors buying the shares pay that premium Investors selling the shares require that premium
In equilibrium, higher turnover rates are accompanied by a higher premium in prices – investors “Pay to Play”
Do Institutions Pay to Play?
Is high turnover of ownership by institutions associated with a premium in stock prices?
If, on average, institutions expect to profit from trading, a premium in prices may exist
Would institutions’ expectations of trading profits depend on who they trade with?
Individuals Commonly held view is that individuals are less informed that
institutions If institutions hold that view, they will expect to make profits when
trading with individuals Other institutions
If institutions believe that other institutions are also well informed, then they may not expect to make profits from trading with other institutions
Hypotheses Turnover of institutional ownership is negatively related to
future stock returns
The negative relation between turnover of institutional ownership and future stock returns is:
More pronounced when turnover is due to trading of institutions and individuals than when turnover is due to trading among institutions
The negative relation between turnover of institutional ownership and future stock returns is more pronounced for firms in which differences of opinion are more likely to be high:
High return volatility stocks High total trading activity stocks High growth opportunities stocks
Related Literature Institutional ownership
Gompers and Metrick (2001) Yan and Zhang (2009)
Changes in institutional ownership Nofsinger and Sias (1999) Cohen, Gompers, Vuolteenaho (2002) Cai and Zheng (2004) Campbell, Ramadorai, Schwartz (2009)
Institutional herding Wermers (1999) Sias (2004) Dasgupta, Prat, Verardo (2009)
Changes in breadth of ownership Chen, Hong, Stein (2002)
Snapshot of Results (1) Turnover of institutional ownership is
negatively related to future returns Based on 10 portfolios, the hedge return
for the two extreme portfolios is 8.6% The relation is due to trading of institutions
with individuals Robust to controls for size, B/M, past
returns, level and change of institutional ownership, return volatility, and overall trading activity
Snapshot of Results (2) Results are more pronounced for
stocks with high overall trading activity 15.0% (vs 2.1%) hedge return
stocks with high stock return volatility 10.2% (vs 0.3%) hedge return
stocks with low B/M 13.6% (vs 2.5%) hedge return
Not subsumed by additional ownership-related variables change in breadth of institutional ownership persistence in institutional buying
Sample and Turnover Measures
Main sample from CDA/Spectrum database Common stocks between 1983:Q4 to 2007:Q4 A total of 416,384 observations for an average of 4,293 per
quarter A minimum of 3,263 and a maximum of 5,765 stocks per quarter
Turnover of institutional ownership Due to total trading (1):
Absolute value of quarterly change in shares held by each institution, summed up over all institutions, divided by average shares held by all institutions, over the past 8 quarters
Due to trading of institutions with individuals (2):Absolute value of quarterly change in shares held by all institutions, divided by average shares held by all institutions over the past 8 quarters
Due to trading among institutions: (1) minus (2)
Additional Variables Additional data sources
CRSP, Compustat Additional variables
Stock returns for the past 24- and 12-months Stock returns for the future 3-, 6-, and 12-months Market capitalization of equity Book-to-market of equity Stock return volatility for the past 24 months Total share turnover for the past 24 months Level of institutional ownership Average change in institutional ownership for the past 8
quarters Change in breadth of ownership (Chen, Hong, and Stein (2002)) Persistence in institutional buying (Dasgupta, Prat, Verardo
(2009))
Summary Statistics (1)Panel A: Summary statistics (416,384 observations)
Variable Mean Min 25th
percentile Median
75th percentile
Max
InstTurnTotal 0.367 0.000 0.237 0.333 0.461 1.087
InstTurnIndiv 0.159 0.000 0.058 0.106 0.200 0.906
InstTurnInst 0.207 0.000 0.108 0.201 0.291 0.557
InstOwn 0.342 0.000 0.114 0.295 0.539 1.000
InstOwn 0.006 – 0.270 – 0.002 0.003 0.012 0.414
MktCap 1.724 0.001 0.038 0.135 0.600 602.433
BM 0.701 0.010 0.330 0.566 0.884 9.905
Turnover 0.062 0.000 0.021 0.043 0.081 0.326
Volatility 0.025 0.000 0.007 0.014 0.029 0.218
Return-24 0.356 – 0.998 – 0.227 0.169 0.652 4.885
Return-12 0.165 – 0.991 – 0.191 0.073 0.371 2.981
Return12 0.143 – 1.000 – 0.217 0.061 0.357 53.663
Summary Statistics (2)Panel B: Time-series means of correlation coefficients
InstTurnTotal InstTurnIndiv InstTurnInst
InstTurnIndiv 0.743 – 0.121
InstTurnInst 0.564 – 0.121
InstOwn – 0.102 – 0.463 0.415
InstOwn 0.135 0.054 0.140
ln(MktCap) – 0.058 – 0.413 0.422
ln(BM) – 0.203 – 0.087 – 0.196
ln(Turnover) 0.392 – 0.051 0.645
ln(Volatility) 0.385 0.350 0.142
Return-24 0.085 0.063 0.054
Return-12 0.023 0.054 – 0.029
Return12 – 0.062 – 0.054 – 0.026
Future 12-month Hedge Returns Based on Total Turnover of Institutional Ownership
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
1984 1985 1987 1988 1990 1991 1993 1994 1996 1997 1999 2000 2002 2003 2005 2006 2008
Hed
ge re
turn
Portfolio Formation Date
Future 12-month Hedge Returns Based on Turnover of Institutions with Individuals
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
1984 1985 1987 1988 1990 1991 1993 1994 1996 1997 1999 2000 2002 2003 2005 2006 2008
Hed
ge re
turn
Portfolio Formation Date
Future 12-month Hedge Returns Based on Turnover Among Institutions
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
1984 1985 1987 1988 1990 1991 1993 1994 1996 1997 1999 2000 2002 2003 2005 2006 2008
Hed
ge re
turn
Portfolio Formation Date
Portfolio Analysis (1)Panel A: Means of portfolio stock returns -- portfolios based on InstTurnTotal
Port InstTurnTotal Return-24 Return-12 Return3 Return6 Return12
1 0.106 0.261 0.145 0.038 0.078 0.165
2 0.190 0.274 0.146 0.039 0.080 0.162
3 0.238 0.291 0.146 0.039 0.078 0.156
4 0.276 0.303 0.151 0.037 0.076 0.155
5 0.312 0.319 0.150 0.037 0.073 0.152
6 0.351 0.342 0.164 0.035 0.070 0.148
7 0.397 0.364 0.162 0.031 0.064 0.132
8 0.455 0.410 0.174 0.030 0.062 0.131
9 0.541 0.472 0.179 0.022 0.046 0.100
10 0.769 0.531 0.210 0.016 0.033 0.079
P10 – P1 0.663 0.270 0.065 – 0.022 – 0.046 – 0.086 (t-stat) (4.3) (1.7) (– 2.2) (– 2.6) (– 2.6)
Portfolio Analysis (2)Panel B: Means of portfolio stock returns -- portfolios based on InstTurnIndiv
Port InstTurnIndiv Return-24 Return-12 Return3 Return6 Return12
1 0.028 0.313 0.145 0.036 0.073 0.152
2 0.046 0.309 0.141 0.038 0.077 0.153
3 0.061 0.293 0.136 0.038 0.078 0.159
4 0.077 0.291 0.140 0.037 0.076 0.157
5 0.097 0.308 0.141 0.038 0.078 0.162
6 0.120 0.327 0.147 0.034 0.068 0.147
7 0.152 0.355 0.151 0.033 0.067 0.139
8 0.196 0.393 0.160 0.028 0.059 0.126
9 0.272 0.464 0.200 0.023 0.046 0.101
10 0.527 0.514 0.264 0.019 0.039 0.084
P10 – P1 0.499 0.201 0.119 – 0.017 – 0.035 – 0.067 (t-stat) (2.8) (3.4) (– 1.8) (– 2.1) (– 2.4)
Portfolio Analysis (3)Panel C: Means of portfolio stock returns -- portfolios based on InstTurnInst
Port InstTurnInst Return-24 Return-12 Return3 Return6 Return12
1 0.014 0.290 0.172 0.033 0.066 0.132
2 0.061 0.307 0.174 0.036 0.074 0.158
3 0.107 0.306 0.162 0.038 0.076 0.156
4 0.147 0.317 0.162 0.034 0.070 0.149
5 0.183 0.331 0.164 0.033 0.067 0.140
6 0.216 0.356 0.169 0.036 0.072 0.146
7 0.250 0.381 0.175 0.032 0.066 0.136
8 0.289 0.402 0.174 0.032 0.063 0.130
9 0.340 0.435 0.164 0.027 0.057 0.120
10 0.443 0.442 0.110 0.023 0.049 0.114
P10 – P1 0.429 0.152 – 0.062 – 0.010 – 0.018 – 0.018 (t-stat) (3.2) (– 2.0) (– 1.1) (– 1.3) (– 0.6)
Regression Analysis:Base Specification
Variable Model 1 Model 2 Model 3 Model 4
ln(MktCap) 0.001 – 0.002 0.002 – 0.002 (0.2) (– 0.4) (0.4) (– 0.4) ln(BM) 0.059 0.059 0.063 0.059 (7.0) (6.4) (7.0) (6.9) InstOwn 0.001 – 0.023 0.008 – 0.022 (0.04) (– 1.4) (0.4) (– 1.2) InstTurnTotal – 0.087 (– 2.5) InstTurnIndiv – 0.149 – 0.149 (– 5.0) (– 5.4) InstTurnInst – 0.020 – 0.003 (– 0.3) (– 0.1) Adjusted R2 0.59% 0.62% 0.54% 0.62%
Regression Analysis:Additional Controls (1)Panel A: Dependent variable is future 12-month return (Return12)
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
InstTurnIndiv – 0.149 – 0.141 – 0.137 – 0.158 – 0.126 – 0.120 (– 5.4) (– 5.1) (– 5.6) (– 6.7) (– 5.4) (– 5.7) InstTurnInst – 0.003 – 0.004 0.025 0.008 0.081 0.107 (– 0.1) (– 0.1) (0.6) (0.1) (2.0) (3.1) InstOwn – 0.330 – 0.538
(– 2.0) (– 3.6) ln(Volatility) – 0.010 – 0.007 (– 0.8) (– 0.6) Return-24 – 0.013 – 0.010 (– 1.5) (– 1.2) Return-12 0.036 0.039 (2.0) (2.3) ln(Turnover) – 0.021 – 0.020 (– 1.9) (– 2.8) Adjusted R2 0.62% 0.63% 0.63% 0.67% 0.66% 0.74%
Regression Analysis:Additional Controls (2)Panel B: Dependent variable is future 3-month return (Return3)
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
InstTurnIndiv – 0.029 – 0.029 – 0.027 – 0.041 – 0.023 – 0.030 (– 2.1) (– 2.1) (– 2.5) (– 3.6) (– 2.3) (– 3.6) InstTurnInst – 0.022 – 0.022 – 0.016 – 0.010 0.002 0.019 (– 0.8) (– 0.8) (– 0.9) (– 0.4) (0.1) (1.2) InstOwn 0.007 – 0.142
(0.1) (– 2.0) ln(Volatility) – 0.002 – 0.004 (– 0.4) (– 0.7) Return-24 – 0.004 – 0.004 (– 1.1) (– 0.9) Return-12 0.027 0.028 (2.9) (3.1) ln(Turnover) – 0.006 – 0.005 (– 1.1) (– 1.4) Adjusted R2 0.26% 0.26% 0.26% 0.49% 0.28% 0.53%
Conditional Portfolio Analysis (1)
Panel A: Future 12-month returns for portfolios based on Turnover and InstTurnIndiv
Turnover portfolio
InstTurnIndiv portfolio
1 2 3 4 5
1 0.163 0.165 0.141 0.146 0.141
2 0.167 0.157 0.137 0.143 0.147
3 0.161 0.177 0.145 0.140 0.153
4 0.171 0.169 0.166 0.160 0.139
5 0.162 0.184 0.160 0.136 0.111
6 0.163 0.176 0.160 0.137 0.092
7 0.167 0.189 0.149 0.113 0.075
8 0.160 0.169 0.165 0.104 0.074
9 0.137 0.169 0.133 0.071 0.029
10 0.141 0.136 0.093 0.064 – 0.009
P10 – P1 – 0.021 – 0.029 – 0.048 – 0.082 – 0.150 (t-stat) (– 1.5) (– 1.2) (– 1.2) (– 1.7) (– 3.5)
Conditional Portfolio Analysis (2)
Panel B: Future 12-month returns for portfolios based on Volatility and InstTurnIndiv
Volatility portfolio
InstTurnIndiv Portfolio
1 2 3 4 5
1 0.155 0.141 0.154 0.152 0.164
2 0.154 0.150 0.148 0.171 0.137
3 0.146 0.155 0.166 0.172 0.110
4 0.154 0.147 0.147 0.172 0.100
5 0.152 0.154 0.151 0.137 0.084
6 0.154 0.155 0.166 0.137 0.097
7 0.155 0.155 0.167 0.123 0.068
8 0.162 0.144 0.166 0.115 0.070
9 0.162 0.158 0.130 0.096 0.056
10 0.153 0.151 0.131 0.091 0.063
P10 – P1 – 0.003 0.009 – 0.023 – 0.061 – 0.102 (t-stat) (– 0.1) (0.5) (– 1.2) (– 3.0) (– 3.2)
Conditional Portfolio Analysis (3)
Panel C: Future 12-month returns for portfolios based on BM and InstTurnIndiv
BM portfolio
InstTurnIndiv Portfolio
1 2 3 4 5
1 0.122 0.140 0.154 0.154 0.186
2 0.126 0.139 0.149 0.162 0.172
3 0.119 0.135 0.157 0.177 0.190
4 0.107 0.141 0.154 0.162 0.195
5 0.060 0.159 0.148 0.183 0.201
6 0.056 0.138 0.159 0.175 0.192
7 0.041 0.124 0.146 0.168 0.198
8 0.018 0.098 0.155 0.168 0.199
9 – 0.010 0.089 0.131 0.156 0.209
10 – 0.014 0.090 0.116 0.146 0.162
P10 – P1 – 0.136 – 0.051 – 0.038 – 0.008 – 0.025 (t-stat) (– 3.2) (– 1.4) (– 1.6) (– 0.3) (– 1.1)
Conditional Regression Analysis
Model 1 Model 2 Model 3
InstTurnIndiv – 0.418 – 0.317 – 0.103 (– 4.4) (– 2.8) (– 4.0) InstTurnInst 0.071 0.016 – 0.003 (1.8) (0.4) (– 0.1) ln(BM) 0.055 0.055 0.045 (7.5) (8.8) (5.4) ln(Turnover) – 0.004 (– 0.5) ln(Volatility) – 0.002 (– 0.1) InstTurnIndiv × ln(Turnover) – 0.081 (– 3.5) InstTurnIndiv × ln(Volatility) – 0.049 (– 1.8) InstTurnIndiv × ln(BM) 0.065 (3.5) Adjusted R2 0.70% 0.65% 0.64%
Changes in Turnover and Returns
2IndivtInstTurn portfolio
InstTurnIndiv Portfolio
1 2 3 4 5 P5 – P1 (t-stat)
1 0.317 0.315 0.303 0.412 0.617 0.300 (2.3) [330] [192] [104] [62] [41] 2 0.301 0.297 0.344 0.416 0.561 0.261 (3.0) [191] [210] [169] [108] [50] 3 0.281 0.289 0.327 0.421 0.555 0.274 (3.3) [109] [168] [193] [164] [94] 4 0.235 0.226 0.256 0.330 0.492 0.257 (3.5) [64] [110] [166] [208] [181] 5 0.136 0.174 0.165 0.205 0.379 0.243 (4.2) [34] [49] [96] [187] [363] P5 – P1 – 0.181 – 0.141 – 0.138 – 0.206 – 0.238 (t-stat) (– 5.1) (– 4.3) (– 5.5) (– 4.6) (– 2.1)
Institutional OwnershipPanel A: Means of future 12-month returns for portfolios based on InstOwn and InstTurnIndiv
InstOwn portfolio
InstTurnIndiv portfolio
1 2 3 4 5
1 0.160 0.175 0.159 0.147 0.142
2 0.152 0.165 0.156 0.148 0.150
3 0.143 0.169 0.178 0.149 0.138
4 0.132 0.175 0.156 0.155 0.147
5 0.130 0.158 0.166 0.159 0.147
6 0.108 0.163 0.160 0.165 0.151
7 0.090 0.149 0.143 0.156 0.146
8 0.105 0.136 0.129 0.150 0.143
9 0.091 0.111 0.113 0.123 0.115
10 0.062 0.083 0.082 0.091 0.083
P10 – P1 – 0.098 – 0.092 – 0.078 – 0.056 – 0.059 (t-stat) (– 3.5) (– 3.1) (– 3.0) (– 1.7) (– 2.3)
Change in Institutional Ownership
Panel B: Means of future 12-month returns for portfolios based on InstOwn and InstTurnIndiv
InstOwn portfolio
InstTurnIndiv portfolio
1 2 3 4 5
1 0.149 0.155 0.158 0.142 0.139
2 0.169 0.162 0.152 0.149 0.151
3 0.171 0.160 0.147 0.149 0.151
4 0.183 0.153 0.158 0.156 0.141
5 0.186 0.155 0.155 0.155 0.122
6 0.172 0.143 0.162 0.148 0.110
7 0.160 0.145 0.148 0.147 0.099
8 0.135 0.140 0.130 0.131 0.099
9 0.154 0.120 0.096 0.108 0.063
10 0.124 0.090 0.069 0.064 0.073
P10 – P1 – 0.025 – 0.065 – 0.089 – 0.078 – 0.066 (t-stat) (– 0.8) (– 2.2) (– 2.9) (– 2.3) (– 2.6)
Change in Breadth of Ownership
Panel C: Means of future 12-month returns for portfolios based on Breadth and InstTurnIndiv
Breadth portfolio
InstTurnIndiv portfolio
1 2 3 4 5
1 0.138 0.156 0.155 0.157 0.152
2 0.135 0.147 0.169 0.158 0.154
3 0.151 0.160 0.161 0.159 0.163
4 0.143 0.166 0.178 0.156 0.169
5 0.138 0.155 0.157 0.168 0.173
6 0.146 0.155 0.155 0.150 0.164
7 0.131 0.135 0.153 0.142 0.134
8 0.102 0.118 0.124 0.123 0.136
9 0.102 0.104 0.109 0.098 0.115
10 0.047 0.063 0.090 0.099 0.092
P10 – P1 – 0.092 – 0.093 – 0.065 – 0.059 – 0.059 (t-stat) (– 2.7) (– 3.5) (– 2.6) (– 2.0) (– 1.7)
PersistencePanel D: Means of future 12-month returns for portfolios based on Persistence and InstTurnIndiv
Persistence portfolio
InstTurnIndiv portfolio
1 2 3 4 5
1 0.159 0.162 0.149 0.152 0.134
2 0.186 0.153 0.151 0.150 0.135
3 0.172 0.161 0.158 0.148 0.139
4 0.174 0.161 0.159 0.159 0.126
5 0.191 0.164 0.163 0.151 0.142
6 0.184 0.155 0.157 0.146 0.104
7 0.157 0.160 0.144 0.140 0.104
8 0.124 0.133 0.131 0.110 0.094
9 0.139 0.108 0.112 0.109 0.050
10 0.135 0.087 0.095 0.086 0.058
P10 – P1 – 0.025 – 0.075 – 0.054 – 0.066 – 0.077 (t-stat) (– 0.6) (– 2.2) (– 1.8) (– 2.5) (– 2.4)
Robustness Measuring turnover rates
6 quarters, 4 quarters Dollars vs shares
Robust across different time periods 1983-1997; 2000-2007
Other controls 8 lags of change in institutional ownership
Data selection Winsorizing future returns; truncating
variables Using stocks below $1/share
Conclusions (1)
High turnover of institutional ownership is associated with a premium in stock prices
The premium is driven by trading of institutions with individuals
The premium is more pronounced for: stocks with high stock return volatility stocks with high overall trading activity stocks with low B/M
Conclusions (2)
Results consistent with disagreement-based models
Harrison and Kreps (1978) Scheinkman and Xiong (2003)
Risk-based explanations? Liquidity? Adverse selection costs?
What differences between institutions and individuals drive our findings?
Predictable individual investor sentiment Agency issues of institutional investors