Mil;
3
^OfiO
LIBRARIES
0074^35
vorking paper
department
of economics
Breach of Trust in Takeovers
and the Optimal Corporate Charter
Monlka Schnitzer
No.
92-10
Feb. 1992
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
Breach of Trust in Takeovers
and the Optimal Corporate Charter
Honlka Schnltzer
No.
92-10
Feb. 1992
Breach of Trust
in
Takeovers
and the Optimal Corporate Charter
Monika Schnitzer*
First draft:
May
1991
This version: February 1992
University of Bonn, visiting scholar at the Massachusetts Institute of Technology.
This paper is based on Chapter 4 of my dissertation which was written within the European Doctoral Program in Quantitative Economics at Bonn University. I am grateful to
participants of the Tilburg Conference on Information Economics and seminars at Boston
University, MIT, University of Pennsylvania, University of Princeton, University of California at San Diego and University of California at Berkeley. I want to thank in particular
David Canning, Oliver Hart, Bengt Holmstrom, Georg Noldeke, Klaus Schmidt and Urs
Schweizer for many helpful comments and discussions. The usual disclaimer applies.
*
Abstract: This paper analyzes
how takeovers and takeover defenses
target companies, using an incomplete contracts framework.
We
affect the value of
consider a raider
who
intends to improve the efficiency of production and to appropriate the rents enjoyed by
stakeholders of the company. Anticipating this rent shifting the stakeholders invest too
little in
focus
cies.
relationship specific assets which
is
We
may
the ex post value increase.
Our main
on how the corporate charter should be designed to mitigate these
inefficien-
offset
show that shareholders can use poison
pills
and golden parachutes to solve the
underinvestment problem without forgoing profitable takeovers.
JEL
Classification Nos.: 026, 611, 511.
Keywords: Takeovers, Incomplete Contracts, Poison
Address of the author:
Monika Schnitzer
Department
of
Economics, E52-371
Massachusetts Institute of Technology
Cambridge,
USA
MA
02139
Pills.
1
The takeover wave
Introduction
of the eighties has roused a very controversial debate on their pros
cons. Proponents of a liberal takeover policy argue that takeovers are important to
and
company more efficiently and
reorganize a
But a number
to discipline or even replace sluggish managers. 1
of critics have challenged this view. Their
main objection
is
that takeovers
are often used to appropriate rents enjoyed by stakeholders, like workers, managers or
suppliers of the firm.
2
This "breach of trust", so the argument, has negative implications
not only for the stakeholders, but also for the owners of the company. However,
unclear
why
The
rents.
make huge
exactly shareholders should be negatively affected
remains
stakeholders lose their
empirical evidence suggests that shareholders of target companies typically
when
gains
there have been
poison
if
it
many
their
company
is
taken over.
What
is
puzzling, though,
is
that
incidences where shareholders approved of takeover defenses like
thus apparently forgoing potential gains from takeovers.
pills,
Despite the attention the recent takeover activity has received the theoretical foundations for evaluating takeovers and takeover defenses are
of this paper
is
to develop a formal
made precise, and
We
framework
in
still
rather thin.
The purpose
which the arguments given above can be
to explore whether they can account for the puzzling empirical evidence.
argue that the stakeholders' rents
may have
a positive impact on their incentive to
undertake relationship specific investments, and we investigate the role of poison
pills in
protecting these rents and thus promoting investment incentives.
Poison
pills
would not be necessary
if
the shareholders could give the stakehold-
optimal investment incentives by writing complete contingent contracts. However,
ers
complete contracts are not
feasible, this
may be
impossible.
It is
well
known
if
that in a
world of incomplete contracts the choice of the governance structure (Williamson, 1985)
has an important impact on economic behavior and efficiency.
erature on governance structures has been very
much
The
influenced by
recent theoretical
lit-
Grossman and Hart's
(1986) seminal article on the optimal allocation of ownership rights in a vertical relationship.
x
Their major innovation was to define "ownership" of an asset as the "residual right
This view
reflects the
theory of the market for corporate control. In a seminal article
pointed out that the separation of ownership and control in modern companies gives
Manne
problems which can be overcome by the threat of takeovers as a form of external competition
2
See
e.g. Shleifer
and Summers (1988).
(1965)
rise to incentive
for control.
to control" this asset in
all
contingencies which have not been dealt with in an explicit
contract before. Almost
all
of the subsequent literature followed
However, a governance structure
focusing on the allocation of asset ownership.
more complex than
Grossman and Hart
just the institution of ownership.
We
in
may be
argue that also the corporate
charter should be seen as part of the governance structure of a company. In the corporate
charter the shareholders determine to what extent they delegate their residual rights of
control, e.g. the right to use poison pills as takeover defenses.
To
simplify the analysis
We
agers.
we
man-
use a standard principal agent framework to model explicitly the rent the
manager derives from
his
employment and how
We
lationship specific investments.
this rent,
focus on only one group of stakeholders, the top
show
this rent affects his incentive to
that
first
if
make
re-
takeovers lead to an expropriation of
then the anticipation of a takeover causes underinvestment in relationship spe-
cific assets.
Since takeovers
may
face a tradeoff between ex ante
increase the efficiency of production ex post, shareholders
and ex post
efficiency distortions.
This result highlights
that takeovers cannot be evaluated only on the basis of the value increase induced by
the actual takeover, because this value increase does not reflect what the potential value
might have been
Then we
vestment
if
the threat of a takeover had not induced the manager to underinvest.
investigate
effect.
We
how
to design the corporate charter to
demonstrate that poison
pills
overcome the underin-
together with golden parachutes can
prevent rent shifting without preventing ex post profitable takeovers. Suppose the manager
is
given the right to fight takeovers with poison
takeover to succeed, they have to
make
pills.
If
the shareholders want the
sure that the manager does not object to
they have to compensate him for the rents he
is
it, i.e.
going to lose with a golden parachute.
Since his rents are protected, the possibility of a takeover does not affect the investment
incentives.
We
show that from the point
of view of social welfare
it
is
always desirable to
design the corporate charter such that rent shifting and underinvestment are avoided.
However, from the shareholders' perspective
this
may be
different.
If
the shareholders
participate in the profits generated by rent shifting in form of a higher takeover price,
and
if
this
outweighs the losses due to underinvestment, then the shareholders
may
opt
not to protect the manager's rents. In fact, their incentive to participate in rent shifting
may be
so large that they
may
encourage even ex post
inefficient takeovers. In this case
rent shifting gives rise to both ex ante and ex post inefficiencies.
of the theoretical literature has focused on takeovers as an instrument to
Most
discipline
managers
in their production decision in order to
company (Grossman and Hart
enhance the value of the
In these models
1980, Scharfstein 1988).
it
is
not in
the interest of the shareholders to allow takeover defenses (unless they serve to increase
This view
the takeover price).
is
questioned by Laffont and Tirole's (1988) study of a
dynamic regulation problem which they apply
to a takeover context.
They
the incumbent's (manager's) incentives to invest in future cost saving
(shareholders) might switch to a second source (raider) later on.
regulator might prefer to
commit
to a switching rule
and Tirole
spirit of
in that
we
and
turns out that the
Thus takeover defenses may
Our approach
differs
from the one of Laffont
interprete takeover defenses as residual rights of control in the
the incomplete contracts approach rather than as an instrument with which to
manipulate the takeover probability.
poison
the regulator
which favors the incumbent (takeover
defenses) in order to improve his investment incentives.
actually enhance the value of the company.
It
if
investigate
pills,
If
the corporate charter allows the manager to use
then this right protects the manager's interests but, in contrast to Laffont
Tirole's model, poison pills are never used in equilibrium.
The
rest of this
paper
and discuss the assumption
is
organized as follows: In Section 2
of incomplete contracts. In Section 3
we introduce the model
we study how
anticipated
takeovers affect the manager's investment incentives. In Section 4 a model of the tender
procedure
is
In Section 5
introduced which endogenizes the takeover price and takeover probability.
we
discuss the optimal choice of the corporate charter from the point of view
of the shareholders
and of
social welfare. Section 6 concludes.
2
2.1
We
a
start with
The Model
The General Framework
an overview of the general framework and
little bit later.
Consider a company that
is
fill
owned by a
in the details of the
large
number
model
of shareholders
status
quo
subgame
no takeover
manager
possible
hired
•
takeover
corporate
manager
charter
invests
attempt
takeover
takeover
subgame
set
production phase
up phase
FIGURE
who need
in
1:
The time
life
Shareholders and manager are engaged
a manager to carry out production.
a dynamic principal-agent relationship.
the
structure of the model.
of the
company, a
set
We
distinguish two phases of activities within
up phase and a production phase. In the
set
up phase,
the manager decides on the level of relationship specific investments that reduce the
expected costs of production later on. For example he spends some
effort to set
firm, to design the product or to organize production. This investment
is
up the
assumed
to
be
observable, but not verifiable at the end of the set up phase.
In the production phase the
he invested in the
set
manager has
up phase the
less effort
to spend effort
on production. The more
he has to spend now for any given
of production. Before he chooses an effort level the
manager
level
receives private information
about the state of the world which determines the productivity of his ex ante investment.
The manager can
gives
him a
exploit his informational advantage to get an information rent
better incentive to invest in the
Between the two phases,
i.e.
first
after the
which
place.
manager's investment costs are sunk but
before the production phase starts, a raider can launch a hostile takeover attempt.
If
the
takeover succeeds the raider can expropriate the manager's quasi-rent but he
increase the efficiency of production.
that the raider
fires
The
simplest
The corporate
when the company
model both
is
charter
is
they
set up, before
e.g.
our model
The
is
illustrated in Figure
contracts approach initiated by
i.e.
pills
or by facili-
usually written by a few founding shareholders
shares to a large
sell
number
of other sharehold-
The time
5.
structure of
1.
central contractual assumption of our
Assumption
assume
3
by introducing poison
This corporate charter will be specified in detail in Section
ers.
effects is to
also
influence the takeover price and the takeover probability
through the design of the corporate charter,
4
to
the manager and becomes owner-manager himself.
The shareholders can
tating dilution.
way
may
model follows
Grossman and Hart
closely the incomplete
(1986).
1 In the set up phase, no contingent contracts can be written,
no contracts that condition on anything that happens in the production
phase.
How
can this assumption be justified? Clearly, the contract cannot condition on the state
of the world or the manager's effort level in the production phase, because these are private
information of the manager. Nor can the manager's investment level be contracted upon.
Although the investment
in a contract in such a
is
way
observable
that
it
courts. For the other variables the
it is
not verifiable,
i.e. it is
impossible to specify
can be verified unambiguously to an outsider
argument
is
slightly
more
subtle.
like
it
the
We explicitly assume
that contingent contracts on, say, the production level are feasible at the beginning of the
production phase, but that the same contracts could not be written in the set up phase.
The
idea
is
that in the set up phase the production technology and the type of product
to be produced in the future
may depend on
the world (not modelled explicitly) which
3
An
the realization of a rather complex state of
may be
very costly
alternative scenario would be that the raider keeps the
if
manager but
not impossible to specify
offers
a more
efficient contract
because he can observe the realization of the state of the world.
4
Poison
pills
may
are introduced with the intention to
make a company
"indigestable" for a raider.
A
example oblige the raider to make huge payments to the shareholders or the creditors
of the company in case of a hostile takeover, such that the takeover becomes too costly. Dilution is
poison
pill
for
explained in detail in Section
4.
in a contract.
This will be
design of the product are clear,
Let us
now
simpler once the production technology and the final
much
i.e.
In the set
some more
describe the model in
2.2
up phase the manager
stage the shareholders can offer
any contract that gives him
5
beginning of the production phase.
in the
The
Set
detail.
Up Phase
hired from a competitive market for managers. At this
is
him only a
fixed
wage contract. The manager
at least his outside option utility
his relationship specific investment
The manager then makes
of (non-monetary) disutility,
the investment level
(i.e.
i
will accept
which we normalize to
i
which
is
measured
zero.
in units
causes the manager a disutility
0-
At the end
Section 4
we
of the set
hostile takeover occurs with probability q.
up phase a
In
specify a tender procedure which determines endogenously takeover price
and takeover probability.
2.3
Status
Quo Subgame
The Production Phase
(S)
If
no takeover occurs, production
in
terms of "producing" a
profit y,
is
carried out by the manager.
costs except for the manager's wage.
e
which
is
The
the realization of a
shareholders. For simplicity
and a good state
02,
his effort decision
we assume
some
(non-monetary)
in units of
of as the
Grossman and Hart
but which
drawn by nature with probability
demand
contractible in the production phase
of "q" in
production
effort
disutility
(1986). For a
is
is
is
observed
not observable by the
that there are only two states of the world, a bad
situation, a
Note that the assumption that the production
may be
to spend
variable 6, called the state of the world, which
by the manager before he takes
5
measured
all
productivity of his effort depends on his ex ante investment and on
random
They may be thought
is
think of production
the total profit of the firm net of
is
To produce y the manager has
not observable. This effort
to the manager.
state 6\
y
i.e.
We
level
(1
— /i) and
/j,
respectively.
parameter of the firm's cost function
cannot be contracted upon
exactly the
more formal
same
as the
in the set up phase but
assumptions on the contractibility
justification see their footnote 14.
no
takeover
manager
manager
payoffs
observes 9
chooses e
realized
shareholders
offer
and reports 9
FIGURE
2:
The time
structure of the status quo subgame.
or any other industry specific parameter.
is
shown
The time
structure of the status quo
subgame
in Figure 2.
At the beginning
of the status
quo subgame, but before the manager learns about
the state of the world, the shareholders offer the manager a contract as a take-it-or-leaveit offer.
Using the revelation principle, we model
this contract as
a direct mechanism.
This means that the contract specifies a wage and a production level conditional on the
manager's report,
level y{9) are
about the state of the world. The wage w(9) and the production
6,
chosen such that the manager
world truthfully.
If
is
always induced to announce the state of the
the manager rejects this contract no production takes place and he
gets his outside option utility. If he accepts he observes the realisation of 6
9.
Then he chooses some unobservable
effort e
and receives w(9)
if
and announces
he produces y(9).
Takeover Subgame (T)
Let us
now turn
to the situation where a raider has taken over the company. In this case
we assume that the
raider fires the
manager and becomes owner-manager
himself,
i.e.
he observes 9 and carries out production himself. To save notation, we assume that the
raider has the
same function
structure of the takeover
All players are
of disutility of effort as the
subgame
is
assumed to be
given in Figure
risk neutral
In principle, the shareholders could overcome
incumbent manager. The time
3.
and to maximize their expected
all
utilities.
problems that arise from asymmetric
information by making the manager residual claimant and extracting his expected profit
as a
lump sum payment ex
manager
is
ante.
We exclude this possibility
with the assumption that the
wealth constrained, so that he cannot pay a lump
sum
transfer in advance.
m
takeover
raider
raider
observes
fires
raider
payoffs
chooses e
realized
manager
FIGURE
The time
3:
Furthermore, we assume that he
is
structure of the takeover subgame.
subject to only limited
the manager chooses, in any state of the world, not to
liability.
fulfill
6
This implies that
if
the contract he has signed
he can be punished only to a limited extend. For expositional convenience, we focus on
the special case of zero liability which means that he cannot be punished at
to leave the firm
manager
and
is
and take
receives a
wage
all
but
is
Thus,
if
he does not
of zero. His utility function
is
assumed to be additively separable
his outside option.
fulfill
his contract, the
given by
U = w—e—
The value
of the firm in the production phase
is
i
(1)
.
given by
V = y-w.
we make the
Finally
Assumption
>
0,
i
>
0,
(2)
following technological assumptions.
2 Let e(y,i.8) be the minimal effort necessary to produce the
profit y given the
y
6
investment level
€ {#i,#2},
i
and
Then for
the state of the world 6.
e(t/,z,0) is twice continuously differentiate
all
and
(2.0) e(y,i,0 1 )>e{y,i,6 2 ).
(2.1) e(y,i,0)
is strictly
e(y,0,6i)
<
oo
if
y
increasing and convex in y, with e(0,i,6)
<
oo.
There
may
be a discontinuity at y
fixed costs. Furthermore, there exists a y
(2.2) e y (y,t,0i)
5
free
>
See Sappington (1983).
e y (y,i,0 2 )
and
e
w (y,t',0i) >
>
such that y
e yy (y,i,0 2 ).
>
=
=
and
due to
e(y,i,0i).
(2.3) e(y,i,6) is decreasing in
(2.4) ei(y,i,Oi)
>
ei(y,i,6 2 ).
(2.5) e yi (y,i,6)
=
Q.
Assumptions
and
(2.0)
(2.1)
i.
guarantee that there
is
an interior solution for the
best production level in both states of the world. According to the
first
part of (2.2) not
only total costs but also marginal costs are higher in the bad state of the world.
"single crossing property"
The second
is
first
This
a standard assumption in the mechanism design literature.
part of (2.2) ensures that the second order conditions are satisfied. (2.3) and
the cost reducing effect of the investment
(2.4) say that
higher in the good state of
is
the world than in the bad state of the world. (2.5) together with (2.3) implies that the
investment reduces only fixed production costs but not marginal production costs. This
assumption ensures (together with assumption
(2.1)
and
(2.2)) that the
manager's rent
is
monotonically increasing in his investment. Without assumption (2.5) we would have to
make assumptions on
third derivatives of the cost function, which do not have an intuitive
interpretation in our set-up, in order to get unambiguous results.
3
In this section
Rent Shifting and Efficiency
we analyze the
subgame and the impact
status quo and the takeover
of
the takeover probability on the manager's investment incentives. As a point of reference
let
us determine the
B
FB
(yf ,y 2
)
first
best levels of production for a given investment
= axgrnax{W =
(l-fi)-(y 1 -e(y 1 ,i,e 1 ))
i.
+ (i{y 2 -e(y2 ,i,d2 ))}
(3)
yi.src
Given Assumption
following
first
2.1,
y[
B
B and
y£ are
and uniquely defined by the
order conditions:
ey
3.1
The
strictly positive
B
(yf
^0
Production
shareholders' problem
is
3)
=
1,
i€
in the Status
to design a contract that
subject to the constraint that
it
is
{1,2}
Quo Subgame
maximizes the value of the company
optimal for the manager to accept and
9
(4)
.
fulfill
the
contract.
7
They know
that at this stage the manager's investment costs are sunk and thus
not relevant for his decision to accept the contract. Since they observe the investment
they can take
its
impact on the manager's
account.
effort cost into
To determine the optimal contract we do not have
to investigate the class of
feasible contracts but can refer to the revelation principle (Myerson, 1979)
attention to the class of
"direct
all
a contract {w(0),y(0)} saying that
incentive compatible, that
manager
well
announce the state of the world
to
known
that the optimal direct
truthfully.
mechanism
may
is
By
direct
it
mechanism has to be
must be optimal
for the
the revelation principle
it is
truthfully implements the best possible
outcome the shareholders can attain from any other
it
A
given the scheme {w(0),y(0)}
is
restrict
the manager announces the state of the world to
he has to produce y{0) and gets the payment w(0).
be
and
all
mechanisms". In our context a direct mechanism
if
i,
contract, no matter
how
sophisticated
be.
Let
j/j
=
y(0j)
=
and Wj
w(0j) with j G {1,2}.
The maximization problem
of the
shareholders can then be stated as the following program:
V=
max
,V2
2/1
M
(1
-
fi)[yi
- wi] + fi[y 2 - w 2
]
(5)
,
!«"2
subject to
IC1:
u>i
IC2:
w2 -
-e(yi,i,0i)
e(y 2 ,i,0 2 )
>
u> 2
-e(y 2 ,i,0i)
>
u>i
-
e(yi,i,62 )
IR1:
wi-e(yi,t,0i)
>
,
IR2:
w 2 -e(y 2 ,i,0 2
>
.
The
first
truthfully.
necessary to
make
Since
,
two conditions (incentive compatibility constraints) guarantee that the manager
announces
7
)
,
all
The second two
sure the
manager
shareholders have the
think of this contract offer as being
conditions (individual rationality constraints) are
will accept
and
fulfill
the contract. 8
common objective to maximize the value
made by one representative shareholder.
8
of their
company we can
The assumption of limited liability implies that a contract that does not satisfy both conditions
cannot be enforced, even if it is accepted. If only one of the two conditions holds the manager might
accept with the intention to fulfill the contract only in that state of the world for which the condition is
10
Proposition
1
An
imization problem
77ie
interior solution (yf ,yf ,wf ,u;f ) of the shareholders
uniquely defined by the following four equations
is
wf
=
e(yf,i,^),
w%
=
e{yli,6 2 )
e„(yf,i,0i)
=
l-A*[l-e,(yf,i,«a)],
(8)
e y {yl,i,e 2 )
=
1
(9)
(6)
+
e(yS,i,0 1 )-e(yS,i,9 2
)
(7)
,
.
corresponding expected equilibrium profit of the firm,
Vs =
is
max-
increasing in
i,
-
(1
p)[yf
-
tof]
+
p[yf
-
(10)
«,f],
since
lf <0
"af
-
<0
1
-
= °-
1=
°-
<»)
Proof: See appendix.
In the following
we
focus on the case where this interior solution
is
indeed optimal. 9
Note that the optimal mechanism has the following properties.
•
The manager
is
fully
compensated
an information rent [e(yf
necessary to induce
i,
•
since [e,(yf ,i,#i)
The optimal
world (y 2
=
him
—
,
i,
to
#i)
for his
— e(t/f
,
i,
announce 9 2
ei(yf,i,6 2 )]
>
production costs. Additionally he receives
6 2 )] in the
good state of the world which
truthfully.
This information increases with
by Assumption
2.4.
incentive scheme induces efficient production in the good state of the
y2
B and
)
inefficiently
low production in the bad state
course, the shareholders could implement the
Of
is
may
first
(j/f
< yf B ). Of
best production level in the
be optimal. However,
bad
in this case the shareholders could have
no production and no payment in that state of the world, so
that the contract would be "fulfilled" by the manager and would yield the same payoff for shareholders.
9
If a corner solution yields a higher payoff to the shareholders than the interior solution then the
shareholders will offer no production and no wage if the bad state is announced and j/f and w^ =
e (j/| *i ^2 )
tne g°°d state. In this case the manager does not get an information rent in the good state.
However, in a more general model, in which 6 is drawn from a larger set {<?i,02,
,6n) the expected
information rent of the manager is always positive as long as there is more than one state of the world
with a positive level of production. See Sappington (1983).
satisfied.
course, such a contract
offered another contract, in which there
>
is
m
•
11
•
state as well.
rent in the
But then they would have
to
pay the manager a higher information
good state to prevent him from announcing the bad
state.
The optimal
contract trades off the efficiency loss in the bad state and the information rent
necessary to induce truthtelling in the good state.
The
•
shareholders are aware that the manager's investment reduces his effort costs
and opportunistically exploit the
fact that investment costs are sunk.
higher the investment the lower the wages they
offer.
Thus, the
However, they cannot
fully
appropriate the returns on the investment because the productivity of the investment
depends on the state of the world which they cannot observe. So the returns on the
investment are shared by shareholders and manager.
Production
3.2
Let us
now turn
to the case in
in the
Takeover Subgame
which a takeover occured
after the set
up phase. What
can the raider do better than the shareholders? To keep the model as simple as possible
we have assumed
that the raider
out production himself. 10
An
more
generally, he
is
the manager and, after having observed
alternative specification would be to
keeps the manager, but that he
(or,
fires
is
8, carries
assume that the raider
able to observe 6 after he has bought the
company
better informed about the relevant state of the world than
shareholders are). In both cases the manager loses his information rent after the takeover
(at least partially).
Both
specifications induce the following value increases: In the
implements the
profits.
In the
efficient
production level y^ B yielding an efficiency gain which increases
good state he can
manager would otherwise
subgame
bad state the raider
raise profits
by appropriating the information rent the
Thus the expected value
enjoy.
of the firm in the takeover
is
V T = V s + E + R,
(12)
where
E=
(1
-
B
^){[yf
- e(y™
•",*)]
-
[yf
-
e(yf,i,*i)]}
10
The investment stays in the company when the manager leaves.
investment in the firm's organization rather than in human capital.
12
It
(13)
should be thought of as an
denotes the expected efficiency gain from choosing the first-best production level in state
#i
and
R=
ti [e(yf,i,0 1
)-e(yf,i,02 )]
(14)
denotes the expected information rent the manager would have enjoyed in state 6 2
that
1
>
^j?
and
due to Assumption
=
eyi
by Assumption
now
problem
it
is
Note further that
^=
since
£=
Note
by Proposition
2.5.
The Underinvestment
3.3
Consider
2.4.
.
Effect
The
the manager's investment decision in the set up phase.
shareholders'
that they cannot directly compensate the manager for his investment because
is
In addition, any indirect compensation contract which conditions
not verifiable.
on future realization of the company's
profits
is
ruled out by the incomplete contracts
assumption. Therefore the manager's incentive to undertake relationship specific invest-
ments comes only from
his
expected information rent. Let i SB
=
axgma,x{R(i)
— i)
denote
the second best investment level given the incomplete contract setting.
Note that i SB
is
W(i)
By Assumptions
for all
that
i
>
2.5
too low. Given yf
and
2.3
—
i
-(l-
±[(R(i)-i)]
=
/
SB
Suppose
>
the
i
R(i FB )
1
))
first
+
best investment level
FB <
On
(yf,i,e 1 )-ne (yli,e 2 )-l
.
.
Since
i
FB G
i
ipates
what
+
i
is
argmax
happen
in the
must be the case
it
-
[R(i)
i],
so
we have
these two inequalities yields
)
R(i
strictly increasing in
takeovers affect the manager's investment?
will
(16)
argmax W(i),
SB €
)
R(i)
>
l
the other hand
)
How do
(15)
z[e J ( J,f,i^ 1 )-e,(yf,z-^ 2 )]-l
SB
i
fi)e t
- i FB Combining
—
FB maximizes
fi(yl-e(yli,e 2 ))-i.
SB W(i FB - R(i FB + i FB > W(i
However, by (16) W(i)
i
we have
=
W(i FB ) > W(i SB ).
R(i SB )
j/f ,
= (l-n)(yf-e(yf,i,e
^[W{i)\
0.
,
production phase.
13
If
i,
SB
)
+
SB
i
a contradiction.
The manager
there
(17)
.
is
rationally antic-
no takeover, he keeps
his
.
job and receives an information rent in the good state of the world whereas in case of
a takeover he gets only his outside option
function of his investment
is
Therefore his expected utility as a
utility.
given by
U = (l-q)R(i)-i
Suppose that the takeover probability
is
q
(18)
.
exogenously given. Proposition 2 summarizes
the impact of a potential takeover on the manager's investment decision.
Proposition 2 An optimal investment
for
all q
€
Furthermore
[0, 1].
Proof: Note that R(i)
because by Assumption
2.1
and that
i
e(yf,i,#i)
is
€ argmax,- U(i,q') and
By
the definitions of
%'
(1
Vs
any given yf R(i)
.
is
bounded above
e(j/f,i,#2)
>
>
Vi
0.
Take any
€
q',q"
[0,1],
q'
<
q"
Let
.
G argmax, U(i,q").
i"
and
for
is
prove the second part of the proposition
using a simple revealed preference argument.
i'
and so
bounded above and
We
maximum.
Therefore (18) must have a
which maximizes (18) exists
i"(q) is decreasing in q
continuous in
is
level i*(q)
i"
we know
that
- q')R(i') -
i'
>
(1
-
q')R(i")
-
i"
(19)
-
i"
>
(1
-
q")R(i')
-
i'
(20)
[(1
-
q')
-
-
and
(1
-
q")R(i")
Thus
[(1
Since
V
q'
(i),
<
-
q" and R(i)
which
is
q')
is
-
(1
-
q")]R(i')
>
strictly increasing in
strictly increasing in
i
i
(1
q")]R(i")
by Assumption
by Proposition
1, is
2.4
it
(21)
.
follows that
i'
>
i".
therefore decreasing in q as
Q.E.D.
well.
Proposition 2 shows that anticipated takeovers
lationship specific assets as
compared
may
cause underinvestment in
to the second best characterized
re-
above and thus
aggravate the efficiency distortion of the manager's investment decision. This underinvest-
ment has a negative impact on the company's
we have shown
potential value
that a raider can increase the value of the
14
Vs
.
In the previous section
company ex
post, for a given
level of investment,
through rent shifting and enhancing the efficiency of production. But
Proposition 2 makes clear that the costs and benefits of takeovers cannot be judged only
on the basis of
ex ante.
If
this
ex post value increase since
the underinvestment effect
is
it
does not reflect the reduction in value
strong enough
it
may
offset
the shareholders'
potential gains from selling their company.
The Tender Procedure
4
we model the
In this section
raider's tender offer
and the shareholder's strategies with
respect to this offer explicitly in order to determine takeover price and takeover probability
endogenously.
We
whom
is
assume that the target company
holds one single share.
tendered. This
The
means that the
raider
is
common
small shareholders each of
makes a conditional
offer
ir
each share that
for
raider buys the tendered shares only
at least the majority of all shares to gain control of the
offers are
N
owned by
if
he can acquire
company. Such conditional tender
practice on the market for corporate control.
For his takeover attempt, the raider has to incur transaction costs
is
a random variable, with support
distribution function F(t) which
is
[0,i],
and
assumed
is
t.
Ex ante
t
distributed according to the cumulative
to have a continuous
and
strictly positive
density f(t).
Given the tender
We
to tender.
offer
ir,
each shareholder has to decide individually whether or not
assume that the number
of shareholders
is
so large that each individual
11
shareholder neglects the influence of his tender decision on the takeover success.
What
are the payoffs of this tender
and suppose the takeover succeeds.
his payoff is the
because
of the
if
expected value of
If
his
game? Consider
for his
own
an individual shareholder
he tendered his share, he gets
minority share. This value
a raider succeeds in acquiring the majority of
company
first
all
may
7r.
If
differ
he did not,
from
shares he might dilute the value
benefit but at the expense of minority shareholders.
n This
V T /N
12
The
assumption is made in most tender models, starting with Grossman and Hart (1980).
For instance he can pay himself a wage that exceeds the market wage or he can sell assets of the
company under value to another company he owns.
12
15
extent to which he can do this
Let d denote the
is
V
Jd
,
maximum
because the raider
is
determined by corporate law and the corporate charter.
possible level of dilution.
always dilute as
will
Then the value
much
decision, since the tender offer
The
raider's payoff
is
Suppose the takeover
as possible.
does not succeed. In this case the shareholder's payoff
is
of a minority share
^-, irrespective of his tender
conditional.
depends on the number of tendered shares.
tendered, the takeover succeeds and his payoff
n
If
> —
shares are
is
n^—jr--^j+d-t.
If
n <
y
shares are tendered, the takeover attempt
(22)
fails
and the
raider's payoff
is
-t.
Note that an
offer
problem which was
first
7r
<
v
d
^
is
due
to the free rider
analyzed by Grossman and Hart (1980). Since each individual
own tender
does not consider his
has no chance to succeed. This
decision to be decisive,
it is
a weakly dominant strategy
to hold out.
To exclude
body weakly
so called "no trade"-equilibria in
prefers to tender,
we
which the takeover
restrict attention to equilibria in
fails
although every-
undominated
strate-
gies.
Proposition 3
If
d
>
t,
then there
outcome of the tender game
in
is
a unique
subgame perfect equilibrium
which the raider offers
takeover succeeds. Equilibrium payoffs are d
—
t
7r*
~
=
for the raider and n*
and
=
the
~d
for each shareholder.
Proof: See Appendix.
The aggregate
make a
is
q
=
payoff for
all
shareholders
is
II
=
Nir*.
If
d
<
t,
the raider cannot
profitable tender offer. Hence, the ex ante probability of a takeover in equilibrium
prob(d
-t>
Note that
after a takeover.
0)
this
=
F(d).
ex ante probability does not depend on the value of the company
This implies that the manager does not influence the likelihood of a
16
takeover with his investment decision. So Proposition
investment decision assuming an exogenously given
dogenously. However, this
dilution d as
ally
a.
lump sum
assume that d
is
is
2,
q,
which characterized the optimal
holds also
q
if
due to our particular specification of the
transfer
from the company to the
We
raider.
is
determined en-
maximum
level of
could more gener-
an increasing function of the potential ex post value of the company.
In this case the manager's investment would increase the possible level of dilution
thus the likelihood of a takeover.
This would give an additional disincentive to invest
and thus reinforce our underinvestment
we
tractable
Now we
However, to keep the model
result of Section 3.
will stick to the simpler specification
5
and
where d
is
sum
a lump
dilution.
The Optimal Corporate Charter
We
can turn to the optimal choice of the corporate charter.
consider two in-
struments with which the shareholders can influence the takeover probability and the
manager's investment incentives.
(1)
The shareholders choose
the parameter
d, i.e.
the
maximum
level of dilution.
By
choosing d the shareholders can determine the takeover probability and affect the
takeover price. 13
(2)
They decide whether
or not the
manager
is
allowed to use takeover defenses.
We
distinguish two cases
(i)
the manager
is
(ii)
the manager
is
not allowed to use any poison
free to use poison pills
for the raider to acquire the
How
is
will the shareholders use these
pills at all (referred
and thus can make
company
(referred to as p
instruments?
The
first
=
0),
virtually impossible
it
=
to as p
1).
step to answer this question
to determine the optimal level of dilution for a given p.
13
Our tender model
of Section 4 with free riding and dilution introduces in a natural
variable d that can be influenced by the shareholders in their corporate charter.
would be Holmstrom and Nalebuff's (1991) tender model where the
specifying in the corporate charter the necessary control majority.
17
A
way the
policy
possible alternative
raider's profits can
be influenced by
Suppose the manager
not allowed to use poison
is
value of the firm as a function of d
V e (d\p=0)
Recall that
E
pills (p
=
Then
0).
the expected
is
=
(l-q)V S (i) +
=
(l-q)V S (i) + q(V T (i)-d)
=
V s (i) +
[E
+
qIT(i)
R(i)
-
(23)
F(d)
d]
.
independent of the manager's investment. Let
is
= argmaxF e (<f|p =
d
(24)
0)
d
The
choice of do
is
driven by three effects:
• a direct effect
on the takeover price 77
• a direct effect
on the raider's
•
an indirect
effect
profit,
increasing functions of
d — t, and thus on the takeover probability F(d);
V s (i)
is
a decreasing function of the
and the takeover price II (i)
= V T (i) — d
are
i.
Suppose next the manager
Certainly not.
holders can "bribe"
d;
on the manager's investment which
takeover probability. Recall that
any takeover?
= VT —
him not
is
allowed to use poison
pills (p
Once the manager has made
to use his poison
pills.
The
=
his
idea
is
1).
Does
this prevent
investment the shareto offer
him a golden
parachute which compensates him for the expected information rent he forgoes in case
of a takeover.
selling the
14
company which
"bribe" the manager.
but poison
pills
G reduces the shareholders' expected profit from
s
II — V — G = E — d. But as long as E > d it pays to
This golden parachute
is
now
As a consequence takeover attempts occur with
positive probability,
are never used in equilibrium.
In contrast to the situation where p
on the manager's investment. The reason
golden parachute
G
=
is
takeovers do not have a negative impact
that the shareholders have to specify this
only after the manager's investment has
become observable. Thus
they can use this observation to evaluate the expected information rent R(i) given the
14
The
implicit assumption here
is
that the shareholders have
all
the bargaining power. This corresponds
to our previous assumption that shareholders as the principal have
a contract offer to the manager. See footnote
7.
18
all
the bargaining power
when making
actual investment
rent. Since the
i,
and choose the golden parachute G(i) to match exactly the forgone
manager
receives R(i) anyway, either as an information rent or as a golden
parachute, he will choose the second best investment level
the firm as a function of d
V
e
is
(d\p=l)
now
i
SB
15
.
The expected
value of
given by
=
(l-q)V S (i SB +
=
V s (i SB + [E-d\-
)
q [Il(i
)
SB
F(d)
)-G(i SB )}
(25)
.
Let
dx
=
a,rg
max V e (d\p =
1)
(26)
In contrast to the situation above the choice of d\ affects only
•
the takeover price as a function of d and
• the raider's profit
but
it
and thus the takeover probability F(d),
does not affect the investment level i SB
.
Proposition 4 summarizes the discussion of poison
Proposition 4 There
pills.
are two candidates for a corporate charter which
max-
imizes shareholders profits:
'
-
A
corporate charter which allows the
tect his
manager
poison
pills to
pro-
information rent induces second best investments.
Given
this
to use
corporate charter there will be a takeover attempt
E>
d.
if
and only
15
G=
R(i)
manager
leads
In this case the shareholders offer a golden parachute
which the manager accepts and he does not use poison
-
ift<d and
pills.
A
corporate charter which gives no control rights to the
to
underinvestment as compared
to the
second
best.
Under
this corporate
Obviously, a golden parachute fixed before the investment has been taken would not do the job,
it cannot condition on the actual investment and thus would not have any effect on the manager's
investment incentives. Note that our combination of poison pills and golden parachutes has an interesting
since
parallel in the renegotiation literature. It
is
well
known
that renegotiation of ex post inefficient incentive
schemes has a negative impact on ex ante efficiency. However, Hermalin and Katz (1991) have shown
that this need not be the case if new information becomes available to the contracting parties after the
contract has been written.
19
charter a takeover occurs
and only
if
<
if t
d.
In case of a takeover
shareholders capture the manager's information rent in form of a higher
takeover price.
Which
two alternatives the shareholders
of these
importance of the underinvestment
effect
will
choose depends on the relative
and the gain from rent
shifting,
on the
i.e.
parameters of the model.
In the remainder of this section
maximizing corporate charter which
Proposition 5 Suppose d can
we
is
be
contrast these two alternatives with the welfare
characterized in Proposition
16
chosen continuously. Then the welfare max-
imizing corporate charter allows the manager to use poison
pills,
p
i
SB
< i FB and
=
1.
Thus
first
i
=
SB
i
.
order condition for the optimal level
dw
e
and since f(dw)
>
E. Since
W
e
is
+ I E_
Jo
{
[E-d w ]f(d w
dw =
is
)
FOC
}
j^ dt
is
(
2? )
is
=
0.
increasing in d \/d
by assumption, the
t
it is
achieved by
,
-SB
This implies
and
Given i SB the expected welfare as a function of d
= yS^SB} + R ( iSBj _
W
The
1,
since distributional concerns are not relevant for welfare
optimal to give the manager the best possible investment incentives. This
choosing p
=
dw = E.
chooses
Proof: Since
5.
< E and
(28)
decreasing in d
uniquely characterizes a global
W>E
maximum.
Q.E.D.
This result has a very intuitive interpretation. The optimal level of dilution should
be specified such that a takeover
cost
t is
is
carried out whenever the realization of the takeover
smaller than the expected efficiency increase E.
Proposition 6 compares the welfare maximizing corporate charter with the two candidates for a privately optimal corporate charter.
16
Welfare maximizing given the constraints of incomplete contracts, asymmetric information and that
the manager cannot be
made
residual claimant.
20
Proposition 6 Suppose
a) If the
manager
allowed to use poison
is
manager
=
pills (p
is
not allowed to use poison
may
optimal level of dilution, do,
privately optimal
I), the
can never exceed the socially optimal
level of dilution
b) If the
the level of dilution can be chosen continuously.
pills (p
=
level, i.e. d\
0),
<
d\\r.
then the privately
exceed the socially optimal
level,
dw-
Proof: a) Consider the function
g(d)
=
W\d) - V e (d\p =
=
V s {i SB + R(i SB -
1)
)
)
SB
i
+f[E- t)f(t)dt - V s
(i
SB
)
-(E- d)F{d)
Jo
rd
=
R(i
SB
)-i
SB
+
[E
-
- (E -
t]f(t)dt
d)F{d)
(29)
.
Jo
Note g(d)
is
increasing since
^- = [ESuppose
d\
> dw-
Since d x 6 argmax
V
Similarly, since
-\E- d]f{d) + F(d) =
d]f(d)
dw = argmax
e
e
(d\p
e
e
(31)
g(d w )
But
b)
is
this is
and (32)
=
W
e
(d
e
b)
"unproductive",
dV e (d\p =
we
i.e.
(d
W
>
)
w -V
must be true that
w \p=l)
e
{d x )
)
e
(d
w \p =
1)
>
=
0.
Then
^=
W
e
is
and
0)
dd
di
=
(31)
.
(32)
.
(d,)
-
V'idrlp
ai
^
=
1)
=
g(d 1 )
(33)
d.
> dw- Suppose
0.
.
the investment
Therefore
^ + [E + R(i) - d]f{d) - F(d)
[E+R(i)-d]f(d)-F(d).
21
=
increasing with
construct an example in which do
e,-
(30)
.
yields
a contradiction to the fact that g(d)
To prove part
1) it
>
we have
W {dw
Combining
=
\p=\)>V
(d 1
W
V
F(d)
(34)
^r
d
0)
= -^f(d) +
=
Suppose that f(d)
the
derivative of
first
dV e {d\p =
want
+
constant.
Then
V
V
=
0) at
+
R(i)
0)
dd
We
is
[E
e
=
to construct the
Thus, by choosing
fj.
\i
p.
close
(d\p
= dw
d
d]f'(d)
=
0)
-
2f(d)
(35)
.
a concave function of
is
d.
Evaluating
shows that
- d w ]f(d w - F(d w
)
example such that
(l-A*){yf
close to
increases, then R(i)
exists a
[E
e
-
=
)
R(i)f(E)
- F(E)
=
1,
s
this
is
B
we can make
E
and F(E)
— e(yf,z,0 2 )]
e(yf
I
i^)}.
(37)
arbitrarily small. Furthermore,
goes up. Since /(•)
constant there
is
enough to one such that
0)
=
R(i)f(E)
- F(E) >
(38)
.
dd
Given concavity
is
(36)
strictly positive. Recall that
-e(yf ,i,»i)-»f +
p[e(yf,i,6i)
dV e (d\p =
than
.
d=d w
E =
if
(d\p
R(i)
+ R(i)-d]f(d)-2f(d)
[E
this implies that the shareholders will
choose a higher level of dilution
Q.E.D.
socially optimal.
Proposition 6 says that in case of p
=
1
the shareholders choose d always inefficiently
low because they do not take into account the positive externality of d on the raider. Thus,
some
of the takeovers that would be ex post efficient (in the sense that
take place. In case of p
=
0,
however, the shareholders
encourage takeovers that are ex post
inefficient.
from rent shifting which add nothing to
It is this last effect
which
critics
The
may
reason
t)
do not
choose d too high and thus
is
that there are private gains
social welfare.
may have in mind when
to appropriate rents leads to socially inefficient takeovers.
not very large with respect to managerial rents
also rents of other stakeholders of the
E >
company
it
While
may be
(e.g.
they claim that the intention
this effect
of considerable
is
presumably
importance
if
workers or suppliers) are on stake.
6 Conclusions
This paper shows that the impact of takeovers on the value of a company cannot be
judged only on the grounds of the value increase a raider might bring about when he
22
actually takes over the company.
The reason
is
that this value increase
may be
by
offset
ex ante reactions of stakeholders to anticipated takeovers, for instance by underinvesting
in relationship specific assets.
We have shown that shareholders can overcome this underinvestment problem caused
by takeovers by choosing the appropriate governance structure
for the
company.
If
the
shareholders give up part of their residual control rights and allow the manager to use poi-
son
pills
they can induce second best investments. Using additionally golden parachutes
the shareholders can
make
sure that poison
post profitable takeovers can take place.
holders approve of poison
pills
and
offer
pills
are never used in equilibrium
This result explains
golden parachutes
is
why
enough the shareholders
vestment. In this case
it
will
Thus,
if
the fact that share-
not so puzzling after
However, our analysis has also made clear that shareholders
shifting through a higher takeover price.
and so ex
may
benefit
all.
from rent
the gains from rent shifting are high
not choose the corporate charter which prevents underin-
may happen
that ex post inefficient takeovers are encouraged
solely for the sake of rent shifting.
23
Appendix
Proof of Proposition
=
Let z
1:
L(z,A)
By
=
{u>i,w 2 ,yi,y 2 } and A
the Kuhn-Tucker
=
{A x A 2 A 3 A 4 } and define the Langrangian function
,
,
- fi)(yi - wi) + n(y 2 - w2
I
(
,
-w
+
+
Aj [wi
+
\2[w2-w +
+
A 3 [t0i-e(yi,i,0i)]
+
A 4 [w 2
Theorem
2
-
e(y 2
Let z*
=
{ttij, i«2, 2/1,2/2}
show that
this
i,
#i)
-
e(y u
i.e.
i,
X )]
(39)
6 2 ))
(see e.g. Intriligator (1971, p. 56))
saddle point of the Lagrangian,
<
i,
,
,
e{yi,i,0 2 )-e(y2,i,0 2 )]
l
the maximization problem of the shareholders
L(z,A*)
e(y 2
)
we know
that z* solves
there exists a A*, such that (z*, A*)
if
is
a
if
Z,(z*,A*)
<
L(z*,A)
Vz>0,A>0.
as characterized in Proposition
1
and A*
=
(40)
{0,/z, 1,0}.
We
will
indeed a saddle point of the Lagrangian and hence, that z* solves the
is
maximization problem.
Note
2/i
< y[
B
first
<
y2
that by (8) e y (y*,i,6i)
B =
IR2 are not. Thus,
Given
y\.
it is
z*,
<
1,
whereas by
=
1.
Thus,
easy to check that A* indeed minimizes Z(x*,A), so the second
is satisfied.
To prove that the
part of (40) holds (given A*)
first
e y (y2,i,8 2 )
conditions IC2 and IR1 are binding whereas ICl and
part of inequality (40)
first
(9)
order conditions for a
maximum
we have
to
show that
and that L(z, A*)
of L(z, A*)
is
z* satisfies the
concave.
The FOCs
are given by:
-Q-
dL
-^—
dy 2
dL
dL
dw 2
=
(^-^) +
=
fi- fi-e y (y 2 ,i,6 2 )
=
-(l-/i)-/z +
=
-fi
+
fi
^-ey(yi,i,0 2 )-e y {y u i,e 1 )
l
<
<
<
<
(41)
(42)
(43)
(44)
24
Obviously they are
with equality at z", so the complementary slackness conditions
satisfied
hold as well.
To
see that L(z, A*)
is
concave,
we have
to
show that the Hessian matrix of
negative semidefinite. Note that by Assumption 2.1 and 2.2
while
all
=
/«i«(yi.*,02)-e w (yi,»,0i)
L V2U2
=
-^yy(y2,i,0 2 ) < 0,
-jj?-
=
Suppose the raider
if
<
-^-
=
offers
the majority of
—t, so
7r
—Tp-
<
Suppose the raider
(46)
.
<
7r
-£f-
<
by Assumptions
2.3
and
2.5
Q.E.D.
^-j^-
The
individual shareholder strictly prefers not to
shareholders tenders, and he
all
is
indifferent
is
the only undominated strategy.
the majority does
if
The
raider's payoff
is
cannot have been optimal.
offers
=
7r
VT -
N
However, for any given positive
v ^~ d
+
e,
t
>
0.
are indifferent no matter
In this case each individual shareholder
is
V ~d
+ c\-t = d-N-e-t.
(
e
there exists a smaller
So the only equilibrium candidate
an equilibrium only
2.3,
2.5.
weakly prefers to tender and the raider's payoff
payoff.
(45)
3:
not tender. Thus, to hold out
II
by Assumption
by Assumption
Proof of Proposition
tender
<
the other second derivatives vanish.
=
-|j-
is
we have
Lyivx
Finally, note that
and
L(-)
offer is
ir*
e'
(47)
which would have yielded a higher
—
^
Given
.
tt'
what the other shareholders are doing. But
if it is
accepted for sure by at least n
> —
all
-k*
shareholders
can be part of
shareholders. Otherwise
the raider would have done better offering slightly more. Hence the only subgame perfect
equilibrium outcome
is
that the raider offers
ir*
=
v
^
d
and
at least
> y
n
shareholders
accept. Note that the raider's payoff
nd
N—n
)..,.__,*-* =
+
-
.
(£-)
n-(
is
independent of
-
n.
-
,
,
d
nd
---t
,
Furthermore, the equilibrium payoff of each shareholder
matter whether he tenders or not.
25
.,„.
= d-t
(48)
is
N no
Q.E.D
References
AUERBACH, A.J
(ED.) (1988), Corporate Takeovers: Causes and Consequences, Chicago,
London.
Grossman,
S.
and O. Hart
(1980), "Takeover Bids, the Free-Rider Problem,
Theory of the Corporation", Bell Journal of Economics,
GROSSMAN,
S.
Vol. 11,
No.
1,
and the
42-64.
"The Costs and Benefits of Ownership: A Theory
and Lateral Integration", Journal of Political Economy, Vol. 94, 691-
AND O. Hart
of Vertical
(1986),
719.
HERMALIN
B.
AND M. KATZ
(1991), "Moral
Renegotiation in Agency"
HOLMSTROM,
AND
,
Hazard and
Econometrica, Vol. 59,
NALEBUFF
Verifiability:
1
The
Effects of
735-1 753.
"To the Raider Goes the Surplus?
Reexamination of the Free- Rider-Problem", mimeo.
B.
INTRILIGATOR, M.
wood
LAFFONT
B.
(1990),
(1971), Mathematical Optimization
A
and Economic Theory, Engle-
Cliffs.
AND
J. TlROLE (1988), "Repeated Auctions of Incentive Contracts, Investand
ment,
Bidding Parity with an Application to Takeovers", Rand Journal, Vol.
J.
19, 516-537.
MANNE, H.
Corporate Control", Journal of Political
(1979), "Incentive Compatibility
and the Bargaining Problem", Econo-
Economy,
MYERSON, R.
and the Market
for
(1965), "Mergers
Vol.
73,
110-120.
metrica, Vol. 47, 61-74.
SAPPINGTON, D.
(1983), "Limited Liability Contracts between Principal
and Agent",
Journal of Economic Theory, Vol. 29, 1-21.
SCHARFSTEIN, D.
(1988),
"The Disciplinary Role of Takeovers", Review of Economic
Studies, Vol. 55, 185-199.
SHLEIFER, A. AND L. SUMMERS (1988), "Breach
Auerbach (1988), 33-67.
WILLIAMSON, O.
(1985),
The Economic
of Trust in Hostile Takeovers", in:
Institutions of Capitalism,
Press.
27
New
York:
The Free
MIT LIBRARIES DUPL
3
TOflO
007MbT3E