HD28
.M414
no.
«<?
|
MAY 9
1991
Shareholder Value Maximization and
Product Market Competition
Julio J. Rotemberg
David S. Scharfstein
Sloan School of Management
Massachusetts Institute of Technology
WP# 2516
was 1916-89
(L
February 23, 1989
L
u
M.I.T.
LIBRARIES
'YOa
Rb
1991
Shareholder Value Maximization
and Product Market Competition
Julio J.
David
*
Rotemberg
S. Scharfstein
Sloan School of Management,
Massachusetts Institute of Technology
current version: February 23, 1989
Abstract
This paper investigates product-market competition under the assumption that
managers maximize shareholder value. This contrasts with the standard assumption in studies of industrial organization that
discounted value of firm profits.
managers maximize their expected
These assumptions are not equivalent when
shareholders are imperfectly informed about firm profitability.
conditions under which shareholder-value maximization leads to
We
demonstrate
more aggressive
product- market strategies than profit maximization. This effect arises because
rival firm's profits
provide information about a firm's value: lower profits of
vals leads investors to believe that the firm
reduce their
rivals' profits in
an
is
more valuable. Thus, firms
effort to increase their
own
stock price.
ri-
try to
We
draw
implications of these effects for corporate financial structure.
•We thank Steven Kaplan, James Poterba, and Jeremy Stein for helpful comment* and the National Science and Sloan
Foundation! for research support.
Introduction
1.
Financial economists typically assume that firms maximize shareholder value.
dustrial economists typically
assume that firms maximize the discounted value of
Sometimes these assumptions amount
per,
we explore
to the
same
thing; often they
do not. In
the product-market effects of shareholder-value maximization
In-
profits.
this pa-
when
these
assumptions are not equivalent.
The equivalence breaks down
when
in
many
realistic
environments.
A
leading example
is
firms have outstanding debt obligations. In this case, shareholders receive the resid-
ual cash flows after
who maximize
payments to
share value care
ers receive all the cash flows
maximize
less
when cash
about cash flows
in those states in
which bondhold-
and shareholders receive none. By contrast, managers who
show that neglecting bondholders can lead
maximization; Myers (1977) demonstrates that
As Miller and Rock
flows are uncertain, managers
do not care how these cash flows are divided. Jensen and Meek-
total firm value
ling (1976)
creditors. Thus,
this
to overinvestment relative to profit
can also lead to underinvestment. 1
(1985) and Stein (1989) show, differences between profit maxi-
mization and share value maximization can also arise for pure equity-financed firms when
information
is
imperfect. Given that stock prices represent the (rational) forecast of the
discounted value of profits this
may seem
surprising.
How
can the maximization of
rational forecast differ from the maximization of the present value itself?
cording to these authors
is
that under imperfect information, firms
enhance investor perceptions (and thus the share
profits.
price) even
this
The answer
may have
ac-
the ability to
without actually increasing
In particular, firms can increase current profits at the expense of future profits
by foregoing valuable investment projects.
If
investment
is
unobservable, and firms vary
according to their inherent profitability, high current profits lead investors to believe that
the firm
is
inherently
more
profitable.
This raises the question of why managers would maximize share value rather than
their expectation of the firm's present discounted value of profits.
1
possibility
is
that
Jensen and Meckling show that the option-like featurei of debt can lead to excessive risk-taking. Myen shows that because
first to bondholders (who have higher priority), shareholders are reluctant to invest when the
the benefits of investment accrue
firm
One
is
in financial distress.
the former objective
is
also that of current shareholders. Indeed, shareholders prefer value
they intend to
maximization
if
takeover bid.
Shareholders will also want a high future stock price
issue
more equity
To
new
shares for liquidity reasons or in response to a
if
the firm plans to
investment.
exhibit the effects of shareholder-value maximization
we
tition,
to finance
sell their
reconsider the standard Cournot oligopoly model.
maximizing firms maximize a weighted average of expected
on product-market compeIn our model, share-value
profits
and stock
We
price.
introduce imperfect information by assuming that firms' costs and the level of market
demand
are stochastic
and not
demand
correlated, while market
The
profits are jointly
it
profits
serially independent.
and those of
determined by the
suggests that industry
and the others have low
As a
specific cost shock.
price.
is
stock market tries to infer a firm's costs (and hence future profitability) based
on the firm's realized
firms
directly observable by investors. Firms' costs are serially
They can do
its rivals.
level of
demand was
profits,
it
result, firms
Rivals' profits are useful because their
market demand.
low. In contrast,
When
if
profits are low for all
one firm has high profits
suggests that the former received a favorable, firm-
wish to outperform each other to increase their stock
so not just by ensuring that their
own
profits are high,
but also by
lowering their rivals' profits.
These
model
it
effects
make
firms
means that firms
profit maximizers.
To
will
more aggressive
in the
product market.
produce more output than they would
In our
if
Cournot
they were pure
see this, suppose that firms' production decisions were the
as those in the standard
Cournot model.
A
same
small increase in a firm's output reduces
expected profits to the second order, but decreases
its rivals'
expected profits to the
its
first
order by lowering the market price. So raising output above the Cournot level increases
shareholder value, by raising current stock price.
The
basic structure of our
model
is
similar to the "signal- jamming''
model of Holm-
strom (1982a), Fudenberg and Tirole (1986) and others. In these models, managers do
not have private information about their characteristics so they are not signaling models.
Instead, managers learn along with the market about their types. Nevertheless, managers
try to manipulate the market's assessments
through unobservable actions. Similarly,
in
our model, firms are not privately informed about their inherent profitability, but try to
influence the market's assessments through their unobservable product-market strategies.
Our
analysis has several implications for the effect of the stock market
performance.
on
industrial
has been argued that U.S. firms care more about short-term results than
It
Japanese firms because U.S. shareholders (typically institutional investors) have shorter
investment horizons than Japanese shareholders (typically banks and corporations with
long-term relationships with firms). In addition, the active takeover market in the U.S.
forces
American firms
to maintain a high stock price,
whereas the concentrated pattern
of
shareholdings in Japan insulates Japanese companies from the threat of takeover. Indeed,
borne out
this observation
is
(1985), Japanese
managers
while American managers
The
result
is
in the data:
list
in
a survey reported by Abegglen and Stalk
list it
is
that American managers are
more reluctant than
investment in market share: Scherer (1988)
Japanese are more willing to incur short-term
losses
is
may
in
which
Our model
this seemingly
myopic project choice
As a
result firms
be valuable. One such
for
Stein (1989) presents a
It is
true that low profits
might avoid valuable projects
if
the market cannot observe
on part of the
story,
may
lower
but they lower the profits of other firms as well. The positive reputational
effect
their actions
by ignoring the other firms
of lowering rivals' profits
own
But
Japanese coun-
exists in equilibrium.
with low short-term payoffs (such as competing aggressively)
and thus the reason
their
by competing more aggressively
points to a potential weakness of this argument.
are a negative signal of a firm's profitability.
profits,
of ten objectives,
one of many to argue that the
current market share and hence greater future market power.
model
list
second.
terparts to take projects with deferred payoffs, though they
project
on a
share-price maximization as last
in the
profits are low.
this only focuses
market. Aggressive product-market strategies
can outweigh the negative reputational
effect of
lowering one's
profits.
The
implication
is
erable takeover activity
that a stock market with short investor horizons or with consid-
may
induce firms to compete more, not
less, aggressively.
Note,
however, that while this improves the efficiency of the product market and increases social
welfare,
lowers overall industry profits.
it
The framework presented here
Firms with ample cash are
ture.
they are
less
also has implications for the choice of financial struc-
less likely to issue
equity in future periods.
concerned with their stock price and thus are
market strategies. Such firms
To overcome
this
lose
less
As a
result,
aggressive in their product-
market share to their leaner
rivals.
problem, firms try to gain competitive advantage by paying out free
cash flows, thereby committing to return to the capital market for further financing.
It
optimal for individual firms to pay high dividends, repurchase stock, or issue debt, to
is
commit
compete more
to
inefficient
aggressively.
from firms' viewpoint.
Although individually optimal,
means that
It
all
firms will
it
is
collectively
compete more aggressively,
reducing industry profits.
The paper has three
present the
main
than
maximizing
profit-
sections following this one. In Section 2
result that share-value
some
describe the model and
maximizing firms may compete more aggressively
firms. Section 3 considers the implications of these
effects for financial policy
the context of
we
along the lines discussed above.
related
product-market
Section 4 puts our results in
work and discusses the empirical implications of the model.
Section 5 contains concluding remarks.
The Model
2.
We
consider an industry with n firms producing a homogeneous product.
demand
function for this industry
and
a normally distributed
£
is
is
p(Q,e)
random
= P[Q) —
variable with
n firms choose simultaneously the quantity of output
model. The equilibrium price
Each firm
particular, the
The use
is
of the
t
=
1, ...,n
is
marginal cost of firm
i,
g±
is
the industry's output
zero and variance a\
to sell as in the
at constant,
denoted
facilitates exposition
particularly simple in this case. This simplicity
good
mean
Q
inverse
The
.
standard Cournot
the one that clears the market given these quantities.
produces output
normal distribution
where
e,
The
comes
0j, is
but stochastic, marginal cost. In
normally distributed 3 with
mean
8
and
(without affecting the main conclusions) because Bayeeian updating
people are willing to pay for the
at the cost of ignoring that the price
negative for really large realirations of t.
distribution again simplifies the presentation. In the care of costs, normality of t has the disadvantage that
costs are sometimes negative. This feature too could be eliminated without affecting the main conclusions but at the cost of
substantial complication.
is
The normal
variance Og.
The random
efficiency of the firm's
demand
uncertainty,
variable
a permanent firm characteristic;
is
6{
managers or the productivity of
e, is
transitory
asymmetry plays an important
and does not
role in
our results.
its
measures the
it
durable assets. In contrast, the
affect firms' future profitability. 4
We
This
give reasons for this assumption in
Section 3 below.
We make
two key informational assumptions. The
They observe neither the random
serve only the level of profits.
equilibrium prices and quantities.
The assumption that outside
informed about the random variables affecting the industry
that these
random
that outside investors ob-
first is
is
variables
-
t
and
e
nor the
investors are imperfectly
Below we argue
plausible.
variables are so difficult to isolate that even the firm's managers lack
complete information about them.
What may be more
controversial
our assumption that investors do not observe
is
the equilibrium prices and quantities. In our model,
single
market price which
unrealistic to
a
more
all
goods are identical so there
consumers can presumably observe. Thus,
all
realistic
model,
in
which firms produce a variety of different products for
little
know nothing about
prices
and
We
is
it
is
that firms cannot observe
is
We make
this informational
reasonable to suppose that
when
(which can also be thought of as their investment
know
the level of
demand when
non-observability of
is
e.
focus instead on
0±
and
e
directly.
not a signaling model; firms do not have private information that
action in equilibrium.
First,
we
quantities.
they signal through their actions. Since firms are uninformed about
same
different
nonetheless retain the assumption that outsiders
Our second informational assumption
Therefore, our model
in
information about the firm's chosen
output. Rather than work with a complicated multi-product model,
a simple homogeneous goods model.
a
would appear
it
assume that investors do not observe the price while consumers do. But,
customers, the observation of a few prices conveys
the
is
the products
assumption
this
all
random
in
for
two reasons.
production capacity) they do not
must actually be
variable as affecting industry
firms in the industry.
firms must take
firms choose the quantity to produce
sold.
Moreover, when the firm chooses quantities
4
While it is more plausible to view
viewed as affecting marginal cost of
0, all
demand, the
This rationalizes the
it is
analysis
also uncertain of
is
formally identical
if c
its
This uncertainty could stem from partial ignorance of actual input
marginal costs.
costs, like the cost of labor or materials.
But, this uncertainty
is
perhaps best thought
of as uncertainty about the firm's underlying technical efficiency. For example, the firm's
managers might learn only slowly about the productivity of the firm's assets or how
effective
they are at controlling costs.
A second
is
that
it
reason for assuming that firms,
simplifies the analysis
like investors,
are uninformed about
6^
and
e
without sacrificing realism. Unlike signaling models where
different "types" take different actions, in our model, all types of firms take the
same
model shares an important feature with signaling models: firms
action. Nonetheless, our
spend resources to try to manipulate investors
beliefs
about their characteristics.
so-called separating equilibria in signaling models, firms are unable to fool the
And
like
market
in
equilibrium. 5
In the
Cournot model, firms simply maximize current
considered here that
of profits.
By
is
contrast
profits.
For the type of model
equivalent to maximizing the managers' expected discounted value
we assume that
The
alent^, shareholder value).
firms maximize the current stock price (or, equiv-
stock price
is
simply the discounted value of profits ex-
pected by outside investors. Therefore, under symmetric information the maximization of
shareholder value
is
equivalent to the maximization of discounted profits.
This equivalence breaks down when the outside investors know
agers of the company.
less
than do the man-
Then managers concerned with value maximization
to take actions that reduce the present value of profits
if
are willing
these actions enhance investor
perceptions of future profits.
Formally,
we assume
and the stock price
at the
that managers maximize a weighted average of current profits
end of the period. As Miller and Rock and Stein point out, and
as discussed in the Introduction, this
realistic
in
is
in the
average shareholder's interest in a number of
environments. First, investors might have to
sell
their shares for liquidity reasons,
which case the weight on the share price would be the fraction of shareholders who must
sell
&
their shares. Second, there might be the threat of a takeover. Finally,
Holmatrom (1982a)
it
perhaps the
first
to take this approach.
jamming.*
6
Fudenberg and Tirole (1986)
call thii
if
the firm
must
type of model •signal-
issue equity,
in
more
it
will
put some weight on
We
develop this
we assume
stock price) V, which
is
that firm
t
is
maximizes current shareholder value
i
a positive constant, pt
is
E[ni
+
api],
(1)
the firm's share price at the end of the period, and the
expectation operator, E, refers to the firm's expectations over
We
assume risk-neutral valuation and a discount rate of
the firm distributes
(or current
given by:
V =
9.
last interpretation
detail in Section 4.
For the moment,
where a
share price.
its
and the n
c
zero.
realizations of
Note that
if
a =
1
and
profits as dividends, (l) simply restates the familiar equality
all its
between the value of the firm and the current dividend plus the discounted end-of-period
stock price.
We now come
make
could
price
We
is
is
to the forces determining the stock price itself.
that the competition
firms
is
As Riordan (1986) shows
it
in
have an incentive to influence their
stock
introduces other effects that are not the focus of our
a model with repeated rounds of competition, firms
rivals'
perceptions of their productivity. Firms want
to convince others that they have low costs to induce
same
The
carried out several times.
then the investors' expectation of the firm's profits in future rounds of competition.
avoid this assumption because
paper.
among
One assumption we
qualitative effects as those
wish to abstract from
we study:
this effect to focus
firms
them
to
compete more
produce
less.
This has the
aggressively. However,
on those which are mediated through
we
financial
markets.
We
thus assume that there
is
only one oligopolistic interaction while the stock price
nonetheless depends on investor perceptions about
firms have available to
them
to their cost parameter, 9±.
0±.
What we have
projects in the second period
Firms with lower values of
-
t
in
mind
whose expected return
are better
managed
productive physical assets, so the return they can hope for on other projects
is
in this
sense that
In this section,
9± is
is
that the
is
related
or have
is
more
higher. It
a permanent shock to costs.
we assume that the stock
the stock market's expectation of
0,.
price
is
a function g(9{), where
Thus, the stock price depends only on
7
-
t
represents
this investor
expectation.
We assume
higher costs are worth
that this function
Given
this
decreasing in
In the next section
less.
features of the model, but
is
now we
we
its
argument so that firms with
derive g(.) from
more fundamental
treat g(.) as fixed.
assumption, firms maximize
£?[»,
Note that since a firm cannot
+ ag{$i)).
affect its payoffs
(2)
from the second period project
in the first
period, the actual project payoffs are omitted from (2).
The next
step
is
to calculate
t
given any realization of n^ for the k
=
l,...,n firms.
While investors cannot observe the outputs of the n firms, they can make rational conjectures about the outputs chosen in equilibrium.
Let qT be the conjectured equilibrium
outputs of the n firms. Of course, a requirement of the equilibrium
is
that these conjec-
by the following
tures are correct. Thus, investors recognize that profits are determined
equation:
Trjfc
where Q*
=
=
Ylk^k- From equation
[D(Q') -
(3), investors
Given the conjectured output of the n
zk
From
each Ok
in
the n realizations of
Bayesian fashion.
arguments then imply that
«
=
'»
where
s
=
o^fo*
is
0,-
=
7r^,
k
n
+
t))ql
can
(3)
infer the realization of 6 k
+
e
=
zk
.
firms,
+
e
= D(Q*) -
investors derive z k
Suppose
is
{6 k
this prior
is
n k /ql
.
(4)
They then update
also
their prior
normal with mean
Q.
about
Standard
given by:
z
d
7T^
s
+n +
s
+n—
„
s. + n
v-^
1
*i
~
1
E
+n
f— 7TT
'
s
/
Z*'
the signal to noise ratio and the weights on the
x
5
n+
1
sum
to
increasing in
z,-
variables
one.
There are two important features of
and thus decreasing
in
7r
t
;
this
updating
rule.
a higher profit realization for firm
8
First, $t
i
is
leads investors to believe
that the cost parameter Q x
is
low. Second, the revision in
6± is
decreasing in
all
other firms'
realizations of z\ high profit realizations for these firms lead investors to believe that
low so that firm
»
is
Note also the
have a high cost parameter
likely to
0,-.
increase in the signal to noise ratio s increases the weight on the firm's
The
the weight on the other z^s.
rationale for this
common component,
Other firms'
e.
performance because firms' costs are
t
An
own
2,
and reduces
increase in s
0, relative to
amounts
the variance
gauging individual
Stated differently, any difference between
less similar.
is
more
be due to an unusual
likely to
.
The updating
rule captures the simple notion that relative
tool in assessing ability
arise in the
simple.
profits are less useful in
a given firm's performance and prior expectations
realization of
is
component,
to an increase in the variance of the idiosyncratic
of the
An
informativeness of the signals on the updating rule.
effect of the
t is
when
there are
common
performance
is
a valuable
shocks to performance. Similar effects
tournaments literature (Lazear and Rosen (1981), Holmstrom (1982b), and
Nalebuff and Stiglitz (1983). There, relative performance provides information about the
actions of individual agents since performance
is
affected
by a
common
shock to
all
agents'
performance.
3.
Equilibrium
This section analyzes product-market equilibrium
cussed above.
We
study Nash equilibria
other firms' output choices.
use.
Of
They
in
in light of the
updating rules
dis-
which firms choose output taking as given
all
also take as given the updating rule (5) that investors
Finally, the updating rules themselves take as given the strategies of the
course, in equilibrium, everyone
is
at
n
firms.
an optimum given the others' strategies and the
conjectured strategies are the chosen ones.
There are two
effects that
directly affects profits; second,
it
a firm considers
when
it
chooses output.
alters the market's assessment of
-
t
First,
and hence determines
the firm's stock price. This latter effect arises because output affects the profits of
and hence
affects z^ for all firms.
The
effect of
the driving force behind our results.
9
output
all
firms
output on the profits of the other firms
is
We now
6t
and
determine the
To do
e.
this
we
effects
on
6 X of
5
=
+ n-l <(gt> 6)
5
<fg.
where
7r '(0
t
respect to
t
,£)
its
=
output
on
output, not the output
The
first
q*
k
and hence
all 7T£
£
-
+n
o?
+ P'{Q*)q
t)
—
This
z^.
because q£
is
term on the right-hand
the market's assessment of 8 X
is
if
and only
effect of g,
We
(6)
+n
X
and
if
i
with
(..
the effect of an increase
the market's conjecture of the chosen
effect of qx
side of (6) reflects the effect of qx
on the n^.
on
2,.
It
lowers
expected profits are increasing in output. The
on the
profits of the (n
-
all
s
f-'
qx unambiguously lowers the assessment of 6 X because
lowers
P'(Q')
y.
So we simply determine the direct
itself.
second term captures the
:
when determining
l,..,n as given
for a given realization of
the derivative of profits of firm
is
x
for a particular realization of
Note that we take
of qx
-
[-P(Q')
output
in
with respect to qx
differentiate (5)
Mj(°i,€)
an increase
it
—
l)
other firms. Increasing
lowers market price and hence
firms' profits.
are
now ready
to derive the first-order condition for the firm's choice of output.
This will give us each firm's reaction function -
optimal choice of output given
its
all
other
firms' output choices. Differentiating (2) with respect to qx yields:
+ n-l)
(/(ejajs
1
+
(s
The standard Cournot
this case to serve as a
a =
Thus,
0, firms
all
n)q
:
condition
is
}
h
,
a special case of equation
benchmark against which
maximize
profit;
to
(7).
We
compare the more general
first
case.
discuss
When
they place no weight on the second-period stock price.
the terms with a drop out and equation (7) simply states that the firm sets
the expected marginal profits of output equal to zero: this
familiar Cournot condition.
informed about
choice so that
all
t
even
if
This condition
a >
0.
is
also optimal
In that case, 9 X
term with g1 drop
is
the stochastic version of the
when
investors are perfectly
cannot be affected by the firm's output
out. Again, equation (7) states that firms set expected
marginal profits equal to zero.
10
Figure
1
below shows the reaction curves of the two firms
lines labelled i2,(0) give the
maximize
profit (a
downward. 6 That
=
is,
0).
As
optimal reaction of firm
is
t
in
The two
a duopoly.
to the output of firm j
when
firms
standard, we assume that the Cournot reaction curves slope
an increase
in rivals'
conjectured output reduces the amount that
an individual firm wishes to produce. The Nash equilibrium occurs at the intersection of
the two curves: at that point neither firm has an incentive to change
its
output given the
other's output choice.
Equation
(7)
shows that the combination of share-value maximization and imperfect
information has two effects on equilibrium.
effect
is
is
The
first
(and, in our context, less interesting)
that the firm weighs marginal profitability of quantity increases by
a positive constant.
If
— kg' where
1
k
g depends on 9, these weights on profits affect the firm's output.
For example, suppose that g
the firm's profits are low.
is
sensitive to changes in 9 only for high values of 9,
Then the
firm
is
most concerned with increasing
its
when
i.e.
profits in
these states. This will lead firms to reduce output below the Cournot point because the
marginal profitability of output in the high 9
+
e
states
is
negative.
If,
in contrast, g
only sensitive to changes in 9 for low values of 9 the firms tend to produce
Cournot
This
is
more than the
level.
effect is similar in
They analyze the
effect of
some
respects to one that Brander and Lewis (1986) consider.
debt financing on product-market equilibrium
when
uncertamly, but symmetric information. Assuming that the value of equity
firms cannot fully meet their debt obligations, share- value maximizing
weight on small profit realizations. Formally, this
sitive to large values of 9.
firms produce
is
is
there
is
when
zero
managers place no
similar to assuming that g(9)
is
insen-
Like our model, this implies that under Cournot competition,
more output than the standard Cournot
the effects analyzed by Brander
We
level.
wish to abstract from
and Lewis because, with equity financing, there
is
no a
priori reason to expect g' to be particularly large for either high or low realizations of
'Provided the »econd-order condition it met, the reaction curves tlop« downward
output* ia negative. This will be the case if
rivals'
P'iQ') + /"'(<? *)«.-<0.
11
if
the derivative of (7) with respect to
We
0.
effect
therefore
would
assume
for the
fundamental determinants of firm
The second - and,
in
to focus
it
is
is
increases in output have
no
level
effect
on
by lowering the market
This effect amounts to a
when
shift
g' is
We
its
which maximizes
its
own
profits.
is
This term (which
competitors' outputs, each firm
profits.
It
does so because small
But, the increase in output lowers
9^.
out in firms' reaction curves; given the output of
downward
in the
One can
rival
this
indeed show
standard Cournot case, they
will slope
as well.
illustrate this graphically in
positive,
(7).
This lowers the market's assessment of
price.
Figure
1 for
the intersection of the two lines labelled Ri(0)
q
from more
produces more output. Provided the reaction curves slope downward,
the reaction curves slope
downward here
g'
ef-
constant, at an optimum, the
translates into an increase in the equilibrium output of all firms.
if
on the competitive
effect of imperfect information
apparent from the second term in
negative. Thus, holding fixed
produces more output than the
that
Also, since this
profitability.
This means that,
positive.
expected value of n'
firms, each firm
a negative constant.
our context, more interesting -
and share- value maximization
rivals' profits
is
maximization. In the next section, however, we derive
fects of share-value
is
that g
monopolist as well, we ignore
arise for a
includes P'{Q))
1
moment
both reaction curves
is
the duopoly case.
As discussed above,
the standard Cournot equilibrium.
shift out to Ri(a).
When
Equilibrium output increases.
Several important features of this equilibrium are worth noting. First, since output
is
greater than the profit-maximizing equilibrium level (and hence further
monopoly
level), firms' profits are
lower
when they maximize shareholder
away from the
value.
Second, investors are not fooled in equilibrium. They correctly predict the equilibrium
outputs and make correct estimates of
7
The
0,-
+
c.
derivative of (7) with respect to the output of other firrru
i
_
g '(*,)a (j
(«
+
Nonetheless, firms try to manipulate the
it:
+ n-l)
<*l'P"{Q')
")«.*
For the reac*ion curves to slope downward, this expression must be negative. In the usual case (which corresponds to a equal
to zero) the corresponding condition is that
[p'(Q')
must be negative. Simple algebra establishes that
+ P"{Q')h]
this condition implies the
12
one stated above.
market's beliefs through their output choices. In the end, they fool no one, but the market's
expectations drive
them towards
this
seemingly sub-optimal strategy. 8
This implies that the second-period stock price
is
the
same
regardless of whether
firms maximize profits or shareholder value. Given that profits are lower this
means that
although firms set out to maximize shareholder value, shareholders are worse
when
firms
maximize
equilibrium than
and
profits.
when
social welfare
is
off
than
But, note that because the firms' outputs are greater in this
firms maximize profits, prices are closer to expected marginal costs
higher.
We conclude this
section with a discussion of
some comparative
static results
and
their
economic implications.
Importance of stock
From
price.
the above discussion
should be clear that an
it
increase in the weight firms place on second-period share price, shifts out firms' reaction
curves and increases equilibrium output, lowers profits and shareholder value, and increases
social welfare.
Uncertainty.
An
In this case, firms are
performance
reluctant to
is
increase in the signal to noise ratio,
more dispersed
more important
compete aggressively
in affecting
the market's beliefs.
to lower their rivals' profits.
interpretation of this result
creases real economic activity.
If
we
is
interpret
g,-
performance
is
shift
inward,
social welfare declines.
demand
variables, re-
in uncertainty that affect all firms
- for
shocks - increase corporate investment because then relative
a better measure of firm profitability.
certainty can reduce investment. Their idea
this sense,
Thus, firms are more
as investment in production capacity,
Bernanke (1983), Pindyck (1988), and others have
*In
and
in firm-specific cost or
duces corporate investment. In contrast, increases
common demand
a firm's individual
Reaction curves
rise-,
effect.
that an increase in idiosyncratic uncertainty de-
our model predicts that greater variability
example,
has the opposite
in their characteristics so that
equilibrium output declines, profits and shareholder value
One
s,
the model
is
is
also
shown that increases
in
un-
that uncertainty increases the option value
similar to the labor-market model of Holmstrom (1982). In his model, worker* exert effort to
Of course, in equilibrium they have no affect on the market's beliefs. But,
influence the market's assessment of their abilities.
given the market's beliefs, a worker who shirked would be viewed a* lets able. Workers might be better off if they could band
together and reduce their effort, but non-cooperative behavior makes this impossible. This equilibrium feature is also found in
other contexts in papers by Riordan (1986), and Fudenberg and Tirole (1986).
13
of investments; the
more uncertainty there
the more can be learned about the value of
is
investing. Thus, firms tend to defer investments
Our
versible.
their investment decisions are irre-
results suggest that different types of uncertainty
investment. Provided the uncertainty
and Pindyck
when
But,
will hold.
if
is
have different impacts on
idiosyncratic, similar results to those of
the uncertainty
Bernanke
systematic, the opposite result will hold.
is
how
This comparative static result also has implications for
the stock market affects real economic activity.
If
the informativeness of
stock prices are determined to
some
extent by noise trading, prices deviate from fundamental values, and firm behavior has less
impact on stock
and compete
is
prices.
Thus, firms have
aggressively.
production capacity
less incentive to invest in
Interestingly, however, the noisier
firm profitability since the product-market outcome
is
the market, the greater
is
closer to the standard
Cournot
equilibrium.
Number
number
the
an increase
output of
of
Q
of firms. The analysis also has implications
of firms
and product-market equilibrium.
n does not
in
In the standard
affect firms' reaction functions;
produced by n —
1
firms
is
same
the
for the relationship
between
Cournot model,
the optimal response to total
as the optimal response to total
output
produce by n firms. However, since there are more firms, industry output increases
Q
in equilibrium.
The
analysis
makes
8{.
it
firms
maximize value rather than
of rivals shifts out firms' reaction functions
number
of rivals, increases the
easier to distinguish the
Hence, firm
common demand
number
profits.
and induces them
An
to
t
+
e
realizations,
profits
when
recognizes that reductions in other firms' profits have a
impact on the market's assessment of
output of Q, firm
of 6
i
will
0,.
The
produce more than
if
equilibrium both because there are
implication
n-
1
more
aggressively.
14
and so
shock from the idiosyncratic cost shock.
market places more weight on other firms'
result, the
rises in
if
aggressively. This can bee seen formally by differentiating (7) with respect
Increasing the
to n.
As a
substantively different
number
increase in the
compete more
is
is
that
if
it
tries to infer
more favorable
n firms produce a
total
firms produce Q. Total industry output
firms
and because firms compete more
The
results of this section are driven by the assumption that the only determinant
of second-period profitability
c,
is
a firm- specific characteristic,
0,.
Suppose instead that
which we regarded as a transitory demand shock actually has a permanent
both demand and firm profitability.
dg/de
is
positive.
Thus
demand was low and
In this case, second-period profits are g(9,
realizations in which
will
effect
all
firms' profits are
low suggests
e)
on
where
part that
in
be low in the subsequent period. This tempers aggressiveness in
goods markets. Firms realize that actions which increase other firms' profits tend to make
markets believe that the current and future
financial
acts in the opposite direction of the
Indeed,
than
sively
in
if
this effect
in the
Second,
will
4.
be
strong enough,
is
First, in
likely to
demand
may even
it
We
lead firms to
believe that this
less
is
compete
less aggres-
an unlikely outcome
a stable, proven industry, uncertainty about individual
be greater than uncertainty about the industry as a whole.
firms are expanding into other businesses the industry
if
are high. This effect
one we have analyzed.
standard Cournot model.
many environments.
firm performance
is
levels of
important while the firm-specific efficiency measure,
demand
9, will
uncertainty,
c,
be more important.
Implications for Financial Policy
Above, we argued that firms care about their stock price because they
to issue equity to finance investment.
This section presents a more detailed analysis of
the effect of equity issues on product-market equilibrium.
makes firms more aggressive
in the
may need
We show
that issuing equity
product market. Therefore, firms wish to commit to
an aggressive product-market strategy by increasing the importance of external, equity
financing.
They can do
so
by distributing cash
flows in the
form of dividends, stock
repurchases, or debt payments.
To explore these
ideas
we consider the
following extension of our basic framework.
Suppose that each firm has an investment project
an
initial outlay.
The
return on the investment
of cost efficiency. In particular,
where
b
>
0.
we assume
is
in the
second period that requires / as
inversely related to 9, the firm's
measure
that the cash flows from this project are e~ bs
,
This particular functional form simplifies the analysis for reasons we discuss
below.
15
C
In the first period, the firm has
choosing
its
output how much of
to shareholders.
C
This distribution
is
publicly observable to
We
period profits as a dividend.
its first-
must decide prior
It
to distribute to shareholders. Define x as the
we assume
For the moment, and for simplicity,
firms.
where I > C.
in cash,
payout
and competing
investors
all
to
that the firm distributes
later discuss the implications of
of
all
dropping
this
assumption.
a firm distributes x
If
The
the investment.
must
it
equity issue gives
second-period cash flows.
The
— (C —
raise I
new
fraction, 7,
from the equity market to finance
x)
on a fraction 7 of the
investors a claim
set so that the
is
appropriate rate of return on their investment of
J.
As
new
before,
firm's
shareholders receive the
we
set this rate of return
equal to zero.
Summarizing, the timing of the model
to shareholders.
Then,
all
profits to shareholders.
firms
compete
the
begin the analysis by
new
in the
First,
managers distribute cash
product market, distributing
first
its
of their
cash flows as a liquidating dividend.
considering the equity issue process. Investors
equity form expectations about
its
value based on the observations of
who buy
firms'
all
Thus, they demand a fraction, 7, of the cash flows such that
profits.
7£(e-*'«)
where the expectation
is
=
/
-C+
1,
(8)
conditioned on the observations of
all
firms' profits.
9
Let N(9i,... ,9 n ,e) represent the multivariate normal distribution of the
As before, investors observe
all
firms' profits
and thus observe the
Their conditional expectation of the firm's cash flows
is
E{e- M <)= [e- b °<N( Zl -e,...,z n -e
where
all
Finally, firms issue equity to finance the second-period project,
undertake the project, and distribute
We
as follows.
is
0,-
t
's
set of variables z\
and
.
.
.
e.
zn
.
therefore:
e)de
=
e-
b§< +
T a\
and o are the conditional mean and the conditional variance of
•There may
t
(9)
-,
t
respectively.
exiat realitations of the B't and t such that £(e~ '-) ii len than / + z. In this cue, the firm would have
to satisfy (8), which is clearly infewible. Taking full account of this possibility introduces technical complications
"i
into the model that we think are economically irrelevant. So we assume this effect is mathematically small and ignore it in
to iet
>
what
follows.
1
1
16
C
Thus, given investors'
beliefs:
I-C + x
1 =
follows
It
from
(10) that
(10)
investors believe that
if
period project are high) the firm must issue
Firm
t's
objective
V,
=
is
+ f
Je
xt
\[P(Q)
l ...e
equity to finance the investment.
maximize shareholder
n ,e
-
dx
-
l
game, each firm
In the first stage of the
Xj.
to
less
low (and the returns on the second-
0+ is
value:
%+
(1
-
-y)«-
M *]jV(*i
...*„,
(14)
e).
J
simultaneously chooses an amount to distribute,
t
Then, given these distributions, firms simultaneously choose outputs.
We
given
solve the
all
model
by determining the equilibrium of the product-market game
first
firms' choices of x.
To do
so,
we
differentiate (11)
with respect to
<ft.
Using
(10),
this gives:
dV
f
N(0
n ,e)=O.
}
-b&i+^a
dqi
.*!»,«
(12)
Changing the variables of integration and using
[P(Q') -
d9;
+ P'iQ-)*} ~ b-±[I
Zi]
/,z\...zn
where
6(s
N m (zi,
.
.
zn )
.
+ n-l)(7-C +
(.
+ „),;
The second term
the marginal distribution of [zi
is
x)'
] [
holding fixed
level of
its
in
equation (14)
we
is
—
+^
,...,zn )
=
0.
(13)
is
8
—
We
n ).
i (/-c+,)P'(C-)
=
o.
(.4)
P'(Q*)qi], equal to a negative number. Thus,
more than the profit-maximizing
just another version of our result in the previous section.
consider the firm's decision of
in the first period.
...
negative. Therefore, firms set expected marginal
competitors' outputs, each firm produces
output. This
Next,
{ Zl
becomes:
p W)- s - F '(«•)*]
from increased output, \P{Q*)
profits
-C + x] N m
dqi
Finally, using equation (5), equation (13)
H
(9) this simplifies to:
show that
how much
cash, x to distribute to shareholders
in equilibrium firms eet
cash to shareholders.
17
x
=
C; they distribute
all
their
To
see this, suppose that firm
small increase in
We
£,-.
did not distribute
*
show that an increase
first
all its
cash:
z,-
<
C. Consider a
out the firm's reaction
in x± shifts
curve and then show that this unilateral shift increases shareholder value.
To show that an increase
in X{
differentiate (14) with respect to
(holding fixed
respect to x,
(s
=
,
[s
+
inequality in (15)
exceeds 6) or the
+
njq^
*
n)q
is
all
-
s
i
+ »- i)(W) -*)-
+n
sP'iQ'h:} >
(is)
o.
J
L
strict as long as either price
is
slopes
downward
exceeds marginal cost (so that
(so that
P'
is
only a second order effect on
increases the equilibrium price
all
The
other firms.
its profitability,
P
negative). Holding the
other firms fixed, this change in x± increases the firm
output, while decreasing the outputs of
firm
met, dqjdx^
i
demand curve
reaction curves of
we
positive provided the derivative of (17) with
L
[(-
.
is
is
t
with respect to z t we have:
positive. Differentiating
'qidxi
The
Given that the second-order condition
.
other firms' outputs)
all
is
zt
indeed shifts out the reaction curve of firm
increase in firm
equilibrium
t's
t's
output has
while the decrease in other firms' outputs
and hence has a
first-order positive effect
on the
profits of
t.
This establishes that given other firms' cash distributions, each firm has an incentive
to distribute all of
its
first-period cash.
Such distributions make the firm more aggressive.
This, in turn, leads competitors to reduce their output,
making the
worthwhile. These reductions are more valuable the larger
and marginal
of one's
cost.
own output expansion on
so.
the difference between price
Moreover, the reductions themselves are larger the larger
While such a policy
that, in equilibrium, firms have
is
is
the effect
price.
Given that each firm wants to distribute
doing
is
distribution of cash
all its
cash, the equilibrium has
individually rational,
no
effect
it
lowers
on second-period stock
all firms' profits.
Note, however, that social welfare
is
this equilibrium as firms
compete more aggressively and prices move
marginal costs.
18
firms
Given
price, shareholder value
(first-period stock price)
also lower.
all
is
higher in
closer to expected
If
the product market
relevant
demand curve
is
perfectly competitive price equals marginal cost while the
for the firm
has zero slope.
In this case, (15) implies that firms
have no incentive to distort their balance sheets. Indeed, the incentive to distribute cash
is
strictly increasing in the excess of price over
Our
marginal cost and the steepness of demand.
basic result in this section has a simple interpretation.
makes a firm more concerned with
its
it
firm sacrifices
more
to act
of
more
its
Formally, our model
aggressively vis a vis
current profits to tarnish
is
distribution of cash
reputation in financial markets because of the in-
creased reliance on the stock market for future financing.
concern leads
The
similar to
many
its
its
This increased reputational
product-market competitors. Each
competitors' image.
that analyze strategic interactions in the
product market. For example, Spence (1977) and Dixit (1980) show that
reasons there
is
an excessive incentive to invest
in
for competitive
By
cost-reducing technologies.
lowering
marginal costs, a firm's reaction curve shifts out, thus lowering the equilibrium output of
rivals.
This
is
analogous to the result presented here: firms distribute cash, shifting out
and lowering
their reaction curves,
Dixit framework,
if all
rivals' outputs.
And,
as in our model, in the Spence-
firms follow this individually rational strategy, industry profits will
be lower.
It is
worth comparing the implications of our model to those of Jensen's (1986) theory
of "free cash flows."
In Jensen's view, managers with control over corporate cash would
rather invest the cash in negative present value investments than distribute
holders. Shareholders can curb this form of managerial slack - emphasis
profitability
- by forcing managers to distribute cash flows to them
it
to share-
on growth over
in the
form of
divi-
dends, stock buybacks, or debt. In this way, investments must be financed externally, and
thus must receive careful scrutiny from the capital market. In support of this view, Jensen
cites
numerous event studies showing that cash
distributions
and debt-equity exchange are
associated with increased share prices. 10
Our model
also predicts profit increases
changes in financial structure.
and positive shaxe price responses
to these
But, these do not stem from a reduction in managerial
10
Thii evidence is a!so consistent with signaling models of financial structure in which dividends, stock buybacks and debtequity exchanges are interpreted as positive signals. See, for example, Myers and Majluf (1984).
19
slack:
our model, there are no agency problems between shareholders and managers.
in
may
Rather, cash distributions induce firms to compete more aggressively, behavior which
look like a reduction in managerial slack.
This implies that while the firm's own share
price should rise after a cash distribution, the share prices of other firms in the industry
should
fall.
There are other important differences between the models.
First, there is
distinction between unilateral actions that increase an individual firm's profits
share price, and industry-wide actions that have the opposite
firms distributed cash there
contrast, there are
share price should
Second,
and
in
would be a reduction
no such interactions
rise
even
if all
In our model,
its
if all
each firm's profits and share price. In
model so that each
firm's profits
and
firms distribute their cash.
our model, cash distributions are associated with increases in investment
prices. In contrast, in Jensen's
and
in Jensen's
and
This unilateral investment and sales increase lead to greater profits and share
sales.
sales
in
effect.
a subtle
framework cash distributions
this reduction raises profits
result in less investment
and
and share price by eliminating unnecessary capital
expenditures and growth.
One
to
of the problems with Jensen's interpretation of the evidence
McConnell and Muscarella (1985), investment
price responses. Their research indicates that
expenditure, share prices
investment
is
rise.
itself is
when
firms
is
that, according
associated with positive share
announce an increase
in capital
This evidence raises some doubts about whether over-
a pervasive part of the U.S. economy,
and hence whether cash distributions
are intended to overcome this problem.
Unfortunately, there
is
no direct evidence on the
effects of
cash distributions on invest-
ment. However, the recent wave of leveraged buyouts (LBOs) provides useful information
about the validity of the ideas presented above. In the typical LBO, the firm
for a large
is
purchased
premium, financed mainly by issuing new debt. Although LBOs typically
volve taking the firm private, successful
LBOs
generally end with a
the company's shares. Indeed, the often stated objective of an
new
LBO
is
in-
public offering of
to reorganize the
one exception ii capital expenditure on oil exploration and development, for which McConnell and Muscarella show
fall upon their announcement. Jensen ha* presented strong evidence that oil companies have over-invested
and that the recent takeover wave in that industry has been a response to these inefficient investments.
"The
that share prices
20
company so
that
In this way,
perception of
that
is
As a
its
can be taken public once again.
it
an
LBO
makes a company's fortunes more dependent on the stock market's
its profitability.
Our theory
thus predicts that one of the benefits of an
leads competitors to retrench (or perhaps to respond with an
it
result,
investment and sales by the newly leveraged firm should
competitors should
issues.
LBO
of their own).
rise,
while those of
fall.
Kaplan's (1988) study of management-led
on these
LBO
LBOs
provides
some evidence which bears
His most relevant results are summarized in Table
1.
The
table reports
changes in sales and capital expenditures from one year before the buyout to two years
after the buyout.
These changes are normalized so that they would equal zero
companies had the same rate of growth
Our theory has mixed
first
to firms in the
same
mind two
First,
fall
of Table
industry.
1
in the
same
in interpreting
industries.
In contrast to the theory's
shows that sales and capital expenditure
However,
buyout
fall relative
these results one should keep in
caveats.
most of these declines occur between the year before the buyout and the year
of the buyout
they
non-buyout companies
success in explaining the results.
column
predictions, the
as
if
itself.
In this year industry adjusted capital expenditures fall by 1.63 while
only by 1.34 from the year before the buyout to two years after the buyout.
the effect of the buyout
is felt
only after the
fiscal
If
year of the buyout, this suggests that
buyouts actually raise capital expenditure by roughly
.31.
12
Indeed, this interpretation of
the evidence gains in attractiveness once one looks at pre-buyout growth in sales and capital
expenditure.
From two
years before the buyout to one year before, buyout companies have
lower growth in sales and capital expenditure than their industry peers. Thus, the poor
performance
A
in the year of the
second caveat
to sell divisions
in interpreting this
than other companies
capital expenditures even
in
buyout seems to be part of a previous downward trend.
if
evidence
is
that
in their industry.
the companies become
LBO
Thit
ii
not exactly true because the
flrrru in
more aggressive
the two sample! differ somewhat.
21
likely
This tends to depress sales and
which they remain.
,3
companies are more
in the lines of business
The
results presented in the
second column of Table
1
provide evidence in favor of our
LBO
model. This column shows the change in sales and capital expenditure for
went back to the capital market either to
industry
rivals
rivals.
These are precisely the
because they are more
likely to
LBO
and capital expenditure.
other than to
A
commit
to a
in sales
firms that
is
to outperform their
LBO
These firms probably underwent an
more aggressive product-market
different interpretation
While
we would expect
their
undergo a subsequent sale exhibit lower growth
for reasons
strategy.
that only successful firms go back to the capital market
and that successful firms also expand capacity and
firms.
and capital expenditure than
be concerned with the capital market's perception of
their value. In contrast, firms that did not
in sales
who
themselves to the public or to another firm.
resell
These firms had substantially higher growth
firms
this alternative interpretation
is
sales
plausible,
more rapidly than unsuccessful
we note that
it
runs counter to
Jensen's free cash flow theory which posits that successful companies are the ones that
reduce their capital expenditure the most.
Table 1
The
Effects of
Management Buyouts on
Sales
and Investment
Industry adjusted growth from one year before buyout
to two years after buyout
IPO,
Overall difference
sale
and releveraged companies
Sales
-8.31
17.36
Capital Expenditures
-1.34
0.33
See Kaplan (1988) for variable definitions
Our theory
of an
LBO
prediction
also predicts that there should be a share price rise at the
and that the share price of other firms
is
obviously true, while there
is
in the industry
no evidence on the
magnitude of stock price increases accompanying LBOs
22
is
should
latter.
announcement
fall.
The former
Of course, the
so large that
it
is
actual
unlikely to
be due exclusively to this phenomenon.
If
we add
LBOs almost
these other benefits to our model,
competitors are weakened. The reason
may convey
is
its
certainly have other advantages. 13
predictions concerning the share prices of
that a successful
information about the ease with which
LBOs
LBO
by a firm
in the industry
(and their attendant
efficiencies)
can be carried out by the competitors. Therefore, even when one of the benefits
retrenchment of competitors, the price of competitors' shares
led to revise
5.
upwards
their estimate of the profitability of
may
LBOs
increase
if
is
the
investors are
in the industry.
Conclusions
Classical studies of financial markets treat the productive part of the firm as an
abstract generator of cash flows; they ignore the product-market structure that gives rise
to these cash flows. Similarly, studies of industrial organization, typically neglect the role
played by financial markets; instead, they start by postulating that firms maximize the
present discounted value of cash flows without asking
to this objective. In this paper,
resulting
we have
tried to bridge
model enables us to explore the
competition and, in turn, the
effect of
what
financial structure gives rise
both of these shortcomings. The
effect of financial
markets on product-market
product-market competition on corporate financial
structure.
This paper
is
not the
and Lewis (1986) and
its
first
to take this approach.
We
have already mentioned Brander
formal connection to our model. Bhattacharya and Ritter(l983),
Gertner, Gibbons, and Scharfstein (1988), Rotemberg (1984,1988), and Bolton and Scharfstein (1989) also analyze the interaction
last
paper
is
most relevant
to our work.
between the product and capital markets. This
In their framework, creditors prevent
from diverting resources to themselves by threatening to cut
formance
is
if
the firm's per-
to aggressive product-market competition. Rivals understand that
predatory actions which reduce the firm's profits are
The optimal
"Amonjst
funding
poor. This optimally designed "shallow pocket" reduces agency problems, but
makes firms vulnerable
exit.
off
managers
financial structure balances the
likely to
be followed by the firm's
agency and product- market
effects.
the advantage! mentioned in the literature are tax reduction!, bondholder expropriation*, and increased efficiency
VUhny (19S8) for a review of the reason* firm* undertake LBO*.
of investment plant. See Kaplan (1988) and Shleifer and
23
This result conflicts in some ways with the one presented here. In our model, firms
commit
to return to the capital
they compete more aggressively
market
in
a
for further financing as
the product market; this
way
of ensuring that
commitment
gives
them a
competitive advantage. In contrast, in the Bolton-Scharfstein model, the commitment to
return to the capital market for further financing makes the firm
product market by inducing
rivals to
compete more
more vulnerable
Our view
aggressively.
whom
for
cash
inducing predatory behavior by
for fear of
is
vulnerable to predatory behavior
is
important.
relatively likely will be unwilling to distribute
Cash-poor firms
exit
that both of
is
these effects can be at work, but that for most firms only one of the effects
may be more
More
rivals.
liquid firms
in the
who
willing to distribute cash as a
needed
are less
way
of
committing to commit more aggressively. More generally, the two theories suggest that
there are product-market costs and benefits of distributing cash.
6.
References
Abegglen,
J.
and Stalk, G. (1985), Kaisha: The Japanese Corporation (New York:
Basic Books)
Bernanke, B.
terly
Uncertainty and Cyclical Investment," Quar-
S. (1983), "Irreversibility,
Journal of Economics, Vol. 98, pp. 85-106
Bhattacharya,
S.
and
Ritter, J. (1983), "Innovation
and Communication: Signalling
with Partial Disclosure," Review of Economic Studies, Vol. 50, pp. 331-346.
Bolton, P. and Scharfstein, D. (1989),
Problems
Brander,
J.
in
"A Theory of Predation Based on Agency
Financial Contracting," American Economic Review, forthcoming.
and Lewis, T. (1986), "Oligopoly and Financial Structure: The Limited
Liability Effect,"
Dixit, A. (1980),
American Economic Review,
"The Role
Vol. 76, pp. 956-970.
of Investment in Entry Deterrence,"
Economic Journal,
Vol. 90, pp. 95-106.
Fudenberg, D. and Tirole,
RAND
J.
(1986),
"A 'Signal-Jamming' Theory of Predation,"
Journal of Economics, Vol. 17, pp. 366- 376.
24
Gertner, R., Gibbons, R., Scharfstein, D. (1988), "Simultaneous Signalling to the
Capital and Product Markets," RAND Journal of Economics, Vol. 19, 173-190.
Holmstrom, B. (1982a), "Managerial Incentive Problems: A Dynamic Perspective,"
in Essays in Economics and Management in Honor of Lars Wahlbeck (Helsinki:
Swedish School of Economics)
(1982b), "Moral Hazard in Teams," Bell Journal of Economics, Vol.13, pp.
324-340.
Jensen,
M.
Jensen,
M. and Meckling, W.
"Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers," American Economic Review, Vol. 76, pp. 323-329.
(1986),
(1976),
"Theory of the Firm: Managerial Behavior,
Agency Costs and Ownership Structure," Journal of Financial Economics,
3,
pp. 305-360.
Kaplan,
S.
(1988),
mimeo
"Management Buyouts:
Efficiency Gains or Value Transfers?,"
(October).
Lazear, E. and Rosen, S. (1981), "Rank-Order Tournaments as
tracts," Journal of Political
Miller,
Vol.
M. and Rock, K.
Optimum Labor Con-
Economy, Vol.89, pp. 841-864.
(1985), "Dividend Policy
under Asymmetric Information,"
Journal of Finance, Vol. 40, pp. 1031-1051.
Myers,
S. (1977),
"The Determinants of Corporate Borrowing," Journal of Financial
Economics,
Myers,
S.
When
and Majluf, N.
(1984), "Corporate Financing
Firms Have Information That Investors
Do Not
and Investment Decisions
Have," Journal of Financial
Economics, Vol. 13, pp. 187-221.
Nalebuff, B. and Stiglitz, J. (1983), "Prizes and Incentives:
of
Toward a General Theory
Compensation and Competition," Bell Journal of Economics, Vol.
14, pp.
21-43.
Pindyck, R. (1988), "Irreversible Investment, Capacity Choice, and the Value of the
Firm," American Economic Review, Vol. 78, pp. 969-975.
25
Riordan, M. (1986), "Imperfect Information and Dynamic Conjectural Variations,"
RAND
Rotemberg,
Journal of Economics, Vol.16, pp. 41-50.
J.J. (1984), "Financial
Transaction Costs and Industrial Performance,"
Sloan School of Management Working Paper 1554-84
(1988), "Short Investor Horizon
M.
Economic
Scherer, F.
and Industrial Performance", mimeo
(1988), "Corporate Takeovers:
Perspectives, Vol.
2,
The
Efficiency
Arguments," Journal of
pp. 69-82.
and Vishny, R. (1988), "Management Buyouts as a Response to Market
Pressure," in A. Auerbach, ed., Mergers and Acquisitions. Chicago: University
of Chicago Press.
Shleifer, A.
Spence, A.
M.
(1977), "Entry, Capacity, Investment, and Oligopolistic Pricing," Bell
Journal of Economics, Vol.
8,
pp. 534-544.
Stein, J. (1989), " Efficient Capital Markets, Inefficient Firms:
A Model
Corporate Behavior," Quarterly Journal of Economics, forthcoming.
26
of Myopic
A
cUrs-^l^
^**<*
*''',
i
J
£<3 v\ )\)}fiu
RiU)
/
/
/
M°> RiW
1
..--.•
Ffq^re
^
^1
39
Date Due
FEB.
5
1998
Lib-26-67
MIT LIBRARIES DUPL
3
TOfiO
DDbbbbOl
D