Common Ownership, Competition, and Top
Management Incentives
Miguel Antón†
Florian Ederer‡
† IESE
‡ Yale
SOM
Mireia Giné†
§ Michigan
Martin Schmalz§
Ross
XXIV Simposio Anual del CIF
IESE, Madrid, 26 de Abril de 2017
1 / 25
Scandal
2 / 25
Scandal
April 11th stock fell 4% after news - $1bn loss.
2 / 25
Scandal
April 11th stock fell 4% after news - $1bn loss.
After CEO apologized, down 1%, $250m loss.
2 / 25
Scandal
April 11th stock fell 4% after news - $1bn loss.
After CEO apologized, down 1%, $250m loss.
Disaster! What would you do if you were the owner of United?
2 / 25
But Who Owns United?
3 / 25
But Who Owns United?
3 / 25
But Who Owns United?
Buffett (CEO of Berkshire) owns 9.1%, and lost $24m.
McNabb (CEO of Vanguard) owns 6.8%, and lost $18m.
3 / 25
But Who Owns United?
Buffett (CEO of Berkshire) owns 9.1%, and lost $24m.
McNabb (CEO of Vanguard) owns 6.8%, and lost $18m.
But they declined to comment...
3 / 25
But Who Owns United?
Buffett (CEO of Berkshire) owns 9.1%, and lost $24m.
McNabb (CEO of Vanguard) owns 6.8%, and lost $18m.
But they declined to comment... Because they were happy?
3 / 25
Net gain/loss of largest shareholders
4 / 25
Net gain/loss of largest shareholders
With this event Buffett won overall $105m.
And Vanguard won overall $78m...
4 / 25
Net gain/loss of largest shareholders
With this event Buffett won overall $105m.
And Vanguard won overall $78m...
Because they also own Uniteds’ competitors!
4 / 25
Net gain/loss of largest shareholders
With this event Buffett won overall $105m.
And Vanguard won overall $78m...
Because they also own Uniteds’ competitors!
4 / 25
Who are these common owners?
5 / 25
In the banking sector?
6 / 25
In the banking sector?
JP Morgan Chase
%
Bank of America
%
Citigroup
%
BlackRock
Vanguard
State Street
Fidelity
Wellington
6.4
4.7
4.5
2.7
2.5
Berkshire Hathaway∗
BlackRock
Vanguard
State Street
Fidelity
6.9
5.3
4.5
4.3
2.1
BlackRock
Vanguard
State Street
Fidelity
Capital World Investors
6.1
4.4
4.2
3.6
2.4
Wells Fargo
%
US Bancorp
%
PNC Bank
%
Berkshire Hathaway
BlackRock
Vanguard
State Street
Fidelity
8.8
5.4
4.5
4.0
3.5
BlackRock
Vanguard
Fidelity
State Street
Berkshire Hathaway
7.4
4.5
4.4
4.4
4.3
Wellington
BlackRock
Vanguard
State Street
Barrow Hanley
8.0
4.7
4.6
4.6
4.0
* Warrants without voting rights.
6 / 25
In the Technology Sector?
7 / 25
In the Healthcare Sector?
8 / 25
Motivation
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from?
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from? Big increase of Index Funds.
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from? Big increase of Index Funds.
Does Common Ownership affect competition?
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from? Big increase of Index Funds.
Does Common Ownership affect competition? Yes, airlines and
banking studies.
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from? Big increase of Index Funds.
Does Common Ownership affect competition? Yes, airlines and
banking studies.
What is a possible mechanism?
9 / 25
Motivation
What happens when same industry firms are owned by the same
guys?
Where does it come from? Big increase of Index Funds.
Does Common Ownership affect competition? Yes, airlines and
banking studies.
What is a possible mechanism? Through Executive Compensation.
9 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)!
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
When firms strategically interact, common owners want to
encourage cooperation, not competition!
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
When firms strategically interact, common owners want to
encourage cooperation, not competition! ⇒ Use less RPE.
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
When firms strategically interact, common owners want to
encourage cooperation, not competition! ⇒ Use less RPE.
We examine Holmstrom (1982) × Hart (1979), also empirically
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
When firms strategically interact, common owners want to
encourage cooperation, not competition! ⇒ Use less RPE.
We examine Holmstrom (1982) × Hart (1979), also empirically
I
Holmstrom: assume firms want to maximize own profit. Use RPE with
exogenous benchmark.
10 / 25
Motivation
Goal: incentivize good behavior, but filter out exogenous shocks
Relative performance evaluation (RPE)! But relative to whom?
I
I
RPE gives incentives to improve own-firm performance – but also to
sabotage/compete aggressively with performance peers.
When firms strategically interact, common owners want to
encourage cooperation, not competition! ⇒ Use less RPE.
We examine Holmstrom (1982) × Hart (1979), also empirically
I
I
Holmstrom: assume firms want to maximize own profit. Use RPE with
exogenous benchmark.
Hart: when firms interact (i.e., industry performance endogenous),
own-profit maximization no longer the unanimous objective.
10 / 25
Which questions do we ask?
“But more importantly ... to what extent will the conduct of firms be different from the
assumed profit maximization behavior in classical theory; and if it differs, what
ramifications does that have for market outcomes...”
(Hart & Holmstrom, 1987)
11 / 25
Which questions do we ask?
“But more importantly ... to what extent will the conduct of firms be different from the
assumed profit maximization behavior in classical theory; and if it differs, what
ramifications does that have for market outcomes...”
(Hart & Holmstrom, 1987)
1
In theory, what are the optimal managerial incentives when
I
I
firms interact strategically (Aggarwal & Samwick, 1999), and
are commonly owned → internalize externalities on each other?
11 / 25
Which questions do we ask?
“But more importantly ... to what extent will the conduct of firms be different from the
assumed profit maximization behavior in classical theory; and if it differs, what
ramifications does that have for market outcomes...”
(Hart & Holmstrom, 1987)
1
In theory, what are the optimal managerial incentives when
I
I
2
firms interact strategically (Aggarwal & Samwick, 1999), and
are commonly owned → internalize externalities on each other?
Is there evidence for the model predictions in the data?
11 / 25
Which questions do we ask?
“But more importantly ... to what extent will the conduct of firms be different from the
assumed profit maximization behavior in classical theory; and if it differs, what
ramifications does that have for market outcomes...”
(Hart & Holmstrom, 1987)
1
In theory, what are the optimal managerial incentives when
I
I
2
firms interact strategically (Aggarwal & Samwick, 1999), and
are commonly owned → internalize externalities on each other?
Is there evidence for the model predictions in the data?
I
I
Identification with panel; Antón & Polk (2014) mutual fund shock
Bonus: common ownership for all industries
11 / 25
Mutual funds support non-benchmarked comp
Big funds engage on pay in 45% of 1,000s of meetings per year
12 / 25
Mutual funds support non-benchmarked comp
Big funds engage on pay in 45% of 1,000s of meetings per year
Pay is high & not very sensitive to relative performance
12 / 25
Mutual funds support non-benchmarked comp
Big funds engage on pay in 45% of 1,000s of meetings per year
Pay is high & not very sensitive to relative performance
I
I
NYT: Why does BLK “wield outsize stick like a wet noodle”?
Answer: omitting RPE maximizes BLK’s economic interest
12 / 25
Mutual funds’ coordination on pay
The biggest funds coordinate their strategies on pay ...
... with the result of supporting high, performance-insensitive pay
13 / 25
Theory
13 / 25
Model objectives and ingredients
Objective: incentivize manager, in the cheapest possible way, such
that s/he sets the desired product market strategy
14 / 25
Model objectives and ingredients
Objective: incentivize manager, in the cheapest possible way, such
that s/he sets the desired product market strategy
Ingredients
1
Imperfect competition: managers can affect industry profits
F
F
2
Strategic complements (differentiated Bertrand)
Strategic substitutes (differentiated Cournot)
Diversified shareholders: incentivize managers to maximize shareholder
value, not own-firm profits in isolation
14 / 25
Setup of the baseline model
Two firms, symmetric marginal cost c
I Inverse demand: P (q , q ) = A − bq − aq
i i j
i
j
Two risk-neutral managers set prices / quantities
I Linear contract: w = k + α π + β π
i
i
i i
i j
I
I
Compensation ratio of αi and β i determines managerial behavior
What’s the optimal αi and β i as a function of ownership?
Two shareholders
I A owns x ≥ 1/2 of firm 1 and 1 − x of firm 2
I B owns 1 − x of firm 1 and x of firm 2
15 / 25
Shareholder’s problem
Shareholder A’s maximization problem is given by
max x (πi − wi ) + (1 − x )(πj − wj )
ki ,αi ,β i
subject to
0
wi ≥ wi
and pi∗ ∈ arg max wi
pi
or
qi∗ ∈ arg max wi
qi
16 / 25
Optimal incentive contract
C:
B:
p
a2 + 4b 2 x 2 − 4ab (2 − 3x ) ∗
αi
2a(1 − x )
p
−e − 2(d − e )x + e 2 + 4ed (2 − 3x ) + 4d 2 x 2 ∗
∗
βi =
αi
2e (1 − x )
β∗i
=
−a + 2(a + b )x −
Proposition (Common Ownership and Incentives)
An increase in common ownership 1 − x increases the inverse
compensation ratio
β∗
α∗
for both forms of competition.
17 / 25
Empirics
17 / 25
Data
1
ExecuComp (S&P1500 + 500)
I
Flow pay as baseline; capitalized stocks & options for robustness
2
Compustat
I Sales → market shares
3
CRSP
I
I
I
4
Industry definition (4-digit SIC)
Performance = market cap increase
Rival performance = VW market cap increase (Aggarwal & Samwick
1999)
13Fs: ownership, MHHI Delta.
18 / 25
Common ownership is rising, with no end in sight
19 / 25
Baseline regression
ωijt
= ki +
+[α1 + α2 F (HHIjt ) + α3 F (MHHIDjt )] ∗ πjto +
+[ β 1 + β 2 F (HHIjt ) + β 3 F (MHHIDjt )] ∗ πjtr +
+γ1 F (HHIjt ) + γ2 F (MHHIDjt ) + ε ijt
(1)
α1 is pay-performance-sensitivity, β 1 is pay-for-rival-performance
sensitivity
I
I
We are mainly interested in α3 and β 3 (Proposition 1) ...
H0 : α3 = β 3 = 0
20 / 25
Top management pay panel regressions
PANEL A
Dependent Variable: Top Management Pay
(A&S)
(No Ctrl)
(Ctrls)
(CEOs)
(Non-CEOs)
0.137***
(4.473)
-0.128***
(-3.345)
-74.42
(-0.815)
0.226***
(15.43)
0.325***
(18.65)
No
Yes
Yes
192,110
0.160
-0.117**
(-2.057)
0.148**
(2.451)
888.2***
(9.007)
0.0543
(1.117)
-0.0322
(-0.568)
484.1***
(4.535)
0.330***
(6.043)
0.182***
(3.089)
No
Yes
Yes
192,110
0.164
-0.0918**
(-2.145)
0.106**
(2.257)
99.80
(1.404)
-0.0604
(-1.544)
0.0676
(1.516)
-366.8***
(-4.830)
0.230***
(5.472)
-0.0183
(-0.391)
Yes
Yes
Yes
183,133
0.463
-0.178
(-1.525)
0.244*
(1.856)
467.1**
(2.503)
-0.132
(-1.214)
0.181
(1.456)
-638.6***
(-3.251)
0.546***
(4.847)
-0.0755
(-0.581)
Yes
Yes
Yes
33,053
0.445
-0.0823**
(-2.509)
0.108***
(2.967)
41.90
(0.742)
-0.0477
(-1.606)
0.0677*
(1.948)
-328.3***
(-5.438)
0.183***
(5.736)
-0.0283
(-0.786)
Yes
Yes
Yes
150,080
0.407
Own * MHHID
Rival * MHHID
MHHID
Own * HHI
Rival * HHI
HHI
Own
Rival
Controls
Industry FE
Year FE
Observations
R-squared
21 / 25
Hypothesis test
S=
(α1 β 3 − α3 β 1 ) + (α2 β 3 − α3 β 2 ) ∗ F (HHI )
∂ ( β/α)
=
∂F (MHHID )
(α1 + α2 F (HHI ) + α3 F (MHHID ))2
22 / 25
Hypothesis test
S=
(α1 β 3 − α3 β 1 ) + (α2 β 3 − α3 β 2 ) ∗ F (HHI )
∂ ( β/α)
=
∂F (MHHID )
(α1 + α2 F (HHI ) + α3 F (MHHID ))2
PANEL B
Hypothesis test at the median (F(HHI)=0.5 and F(MHHID)=0.5)
Inverse Comp. Ratio Test
0.242*** 0.147*** 0.306** 0.150***
P-Value
0.006
0.008
0.041
0.001
The more common ownership, the less RPE!
22 / 25
Interpretation of results
Fact: common ownership associated with top management incentives
that may encourage less competition and more cooperation
23 / 25
Interpretation of results
Fact: common ownership associated with top management incentives
that may encourage less competition and more cooperation
That does not imply that common owners set up pay packages with
the conscious goal of reducing competition
I Index funds’ presence → activists’ absence
I
Index funds don’t push pro-competitive policies like activists
23 / 25
Interpretation of results
Fact: common ownership associated with top management incentives
that may encourage less competition and more cooperation
That does not imply that common owners set up pay packages with
the conscious goal of reducing competition
I Index funds’ presence → activists’ absence
I
Index funds don’t push pro-competitive policies like activists
That’s the benign interpretation ... but we also know
23 / 25
Interpretation of results
Fact: common ownership associated with top management incentives
that may encourage less competition and more cooperation
That does not imply that common owners set up pay packages with
the conscious goal of reducing competition
I Index funds’ presence → activists’ absence
I
Index funds don’t push pro-competitive policies like activists
That’s the benign interpretation ... but we also know
I
I
I
45% of “engagement” meetings feature compensation
“Passive Investors, not passive owners” (Appel, Gormley & Keim 2016)
Several examples of “misbehavior”: secret meeting with executives
from several pharmaceutical companies about pricing strategy, actively
advising airlines to cut capacity, ...
23 / 25
Next Steps
Stronger focus on wealth-performance sensitivity
I
I
WPS is a better measure of incentives than PPS
No longer about RPE, but additional predictions about strength of
incentives and product market business stealing
Firm-level common ownership measure
I
MHHID used to be the only measure, but that’s no longer true!
ISS Incentive Lab Data
I
I
Direct observation of contract terms rather than estimation
But contracts leave a discretional amount (50%), hence important to
look (as we do) at ex-post compensation
24 / 25
Conclusions
24 / 25
Conclusions
Economic incentives rationalize why broadly diversified investors
endorse relative-performance-insensitive compensation
I
CEOs in more commonly owned industries are rewarded relatively less
for own performance and more for rival performance
25 / 25
Conclusions
Economic incentives rationalize why broadly diversified investors
endorse relative-performance-insensitive compensation
I
CEOs in more commonly owned industries are rewarded relatively less
for own performance and more for rival performance
Implications
I
Future research in organizational economics, finance & IO may benefit
from taking heterogeneous shareholder preferences seriously
25 / 25