Sie sind auf Seite 1von 17

American Economic Association

On the Irrelevance of Corporate Financial Policy


Author(s): Joseph E. Stiglitz
Source: The American Economic Review, Vol. 64, No. 6 (Dec., 1974), pp. 851-866
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1815238
Accessed: 02-11-2017 01:23 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

American Economic Association is collaborating with JSTOR to digitize, preserve and extend
access to The American Economic Review

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
On the Irrelevance of Corporate
Financial Policy
By JOSEPH E. STIGLITZ*

This paper extends to a multiperiod the irrelevance of financial policy is of far


model, the argument of Franco Modigliani greater significance.
and Merton Miller and the author (1969) We can divide the decisions of the firm
that the financial policy of the firm is of no into four groups:
consequence. In doing so, we are able to (a) How should the firm finance its
consider a much wider class of financial investment?
policies: not only does the firm choose a (b) How should the firm distribute its
debt-equity ratio, but it also selects a revenue?
dividend-retention ratio and a maturity (c) How much should the firm invest?
structure of debt, and it may even decide (d) Which projects should the firm
on holdings of assets (securities) in other undertake (or what techniques of
firms. I wish to show, in the context of a production should the firm em-
general equilibrium model, that none of ploy)?
these policies has an effect on the valuation The first two decisions of the firm are
of the firm, under certain seemingly weak the financial decisions of the firm, the lat-
circumstances. Whether these assumptions ter two the real decisions. The theory of
are ''realistic" or not is a question of some corporate finance focuses on the financial
debate, about which I shall have a few decisions. The two financial decisions are
words to say later. But by clarifying the closely related (see below), and so are the
assumptions, I hope at least to focus the two real decisions. What is not obvious is
discussion on the relevant issues. the relationship between the real decisions
The question of the effect of firm finan- and the financial decisions. An answer to
cial policy on the valuation of the firm is this question requires an analysis of the
obviously of central concern to students of relationship between corporations and the
corporate finance; if the conditions under household sector of the economy, and to
which the "irrelevance" theorems obtain further our understanding of this rela-
are deemed realistic, it robs them of much tionship is a primary object of this paper.
of their stock in trade. But the question of If the hypothesis that the financial policy
of the firm makes no difference to the
* Stanford University. This is a revised version of a
lecture delivered at a conference at Hakone, June 25-26,firm's market valuation is correct, it also
1970. I am very much indebted to Hirofumi Uzawa, who means that if firms maximize their market
organized the conference, and to the other participants
value, the real decisions are the only de-
for the helpful comments and discussion. The research
described here was supported by the Guggenheim cisions that count, and the financial de-
Foundation, the National Science Foundation, and the cisions have no bearing on them. In par-
Ford Foundation. Like all workers in the area of corpo-
ticular, it means that analyses of the real
rate finance, I owe a great intellectual debt to Merton
Miller and Franco Modigliani; in addition, I have sector based on "flow of funds analysis"-
benefited greatly over the years from helpful discussions and conclusions such as that of Nicholas
with them on the issues discussed here. I am also in-
Kaldor that because the flow of funds from
debted to Robert Merton and Peter Kerbel for their
helpful discussions. the household sector to the corporate sec-

851

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
852 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

tor is very small, the decisions of house- The paper will proceed as follows. In
holds with respect to savings are of rela- Section I, the basic model is set up. In
tively little significance in the determina- Section II, I prove my fundamental
tion of the equilibrium of the economy- theorem on the irrelevance of the firm's
are not likely to give us much insight into financial policy from the point of view of
what is really going on: at best they pro- any individual. Section III will comment
vide us with some spurious correlations.' briefly on the assumptions made and their
Moreover, if the maturity structure of limitations. Section IV will show that
debt is of no consequence, it casts some financial policy need not be of concern to
doubt about the validity of the partial any particular firm, even if it is of concern
equilibrium models attempting to relate to individuals, under much weaker condi-
the maturity structure to the term struc- tions than those used to demonstrate the
ture of interest rates (see, for instance, my earlier proposition.
1970 paper).
I. The Basic Model
In the literature, two different but
closely related propositions have been con- A. Firms
fused: they both assert that the financial The various financial decisions of the
policy of the firm has no affect on its firm are very closely related. One of the
valuation. One asserts, however, that the interests in a multiperiod model is to ex-
individual is indifferent to alternative plore these relationships. A decision to in-
financial policies, in particular to debt- crease the amount to be distributed as
equity ratios, and hence there is no de- dividends means that if the firm were to
terminate debt-equity ratio for the econ- leave its investment decision unchanged,
omy as a whole. That is to say, any change it would have to raise additional revenue
in the financial policy of the firm can be to pay for the planned investment. If it
completely offset by the actions of the raises more by issuing bonds, the amount
stockholders (and indeed will be offset in left over for distribution next period will
the new general equilibrium situation). be decreased, and hence either retained
The second proposition asserts that the earnings or dividends next period must be
individual may not be indifferent to al- reduced. If it raises the revenue by issuing
ternative financial policies, that there may shares, it means the amount distributed to
be for instance a determinate debt-equity each shareholder next period (if retained
ratio for the economy as a whole, but the earnings were unchanged next period)
financial policy of any particular firm would be reduced. Thus, the interrelations
makes no difference. The first asserts, in among all the decisions are complex and
other words, the irrelevance of the financialany decision today may have ramifications
structure for the entire economy, and for many periods into the future.3
therefore of the particular firm; the second
only asserts the irrelevance of the financiallatter type: they show that if there are two or more firms
of the same risk class (the same pattern of returns across
structure of an individual firm. Clearly the
the states of nature), then the debt-equity ratio of any
former proposition is a much stronger one particular firm is indeterminate. The author's theorem
than the latter.2 We are concerned here (1969) was of the former type.
3 The importance of these relationships has often
with both kinds of propositions.
been missed by even as astute students of the theory of
1 E.g., the correlation between retained earnings and corporate finance as William Baumol and Burton
investment does not provide an explanation of the de- Malkiel, and Modigliani and Miller. In discussing the
termination of the level of investment. See, for instance, impact of taxation on the optimal financial policy of the
John Meyer and Edwin Kuh. firm, they observe that increasing the debt reduces the
2 The Modigliani-Miller theorem was really of the tax liability of the firm and hence increases its value.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 853

For expositional simplicity, we shall use retained earnings or by issuing new securi-
a "one-commodity" model;4 each period ties.
there is a single commodity input and a (b) If it issues new securities, it can
single commodity output (dollars or yen). use a number of different financial instru-
We shall look at the consequences of al- ments: common stock, bonds, preferred
ternative financial plans on the firm's stock, convertible bonds, etc. Each of these
market valuation, given a "real plan" of financial instruments carries with it differ-
the firm. A real plan is characterized by a ent contractual rights with respect to the
statement of the investment level and distribution of the gross profits of the firm,
choice of technique in each period con- and the part the owner of those instru-
tingent on the state of nature (the set of ments can play in the decision making of
events that have occurred up to that the firm. For instance, bonds yield a fixed
time). Thus, given the real plan, we know sum in every state of nature except when
the level of profits in each period, depend- the firm goes bankrupt, in which case the
ing of course, on the state of nature. Let proceeds of the firm are divided among the
bondholders. However, except when there
Ij(t, 0(t), k)=the level of investment of
is the distinct possibility of the firm not
the ith firm at time t, if the
being able to meet its debt obligations, the
state of nature at that date
bondholders generally have no voting
is 0(t), under plan k.
rights in the management of the firm.
Xi(t, 0(t), k) = the output or gross profits
The return to a common stock, on the
of the ith firm at time t, if
other hand, is variable except when the
the state of nature at that
firm goes bankrupt, in which case it is
date is 0(t) under plan k.5 6
zero. A shareholder is entitled to receive
There are a number of alternative ways a proportionate share of the dividends of
a firm may finance its investment: the firm. The dividends, of course, depend
(a) It can finance its investment with not only on the real policy of the firm but
on the particular financial policy chosen.
For an ongoing firm with a given investment policy, To know the stream of returns, the share-
increasing the debt-equity ratio implies that the firm holder must know both the real and the
retains less of its earnings and thus the capital gains
will be smaller.
financial decisions of the firm. On the other
4 The analysis for the multicommodity model is hand, if our argument that financial policy
identical, except- now a new set of financial decisions is irrelevant is correct, then although
becomes available to the firm: it can denominate its
changes in dividend policy affect the pat-
bonds in terms of money or in terms of some other com-
modity, or some composite commodity (e.g., the cost of tern of returns received by any single share
living index). Indeed, certain decisions which may ap- of the firm, the individual is indifferent to
pear to be "real" are in fact financial: when relative
these changes. The shareholder (like the
prices are uncertain, firms must decide on whether to
buy futures (or hold inventories) of inputs or sell futures bondholder) can sell his shares at any date
(or hold inventories) of outputs. In short, all such and receive what he can for them. Finally,
"hedging" decisions have (under the assumptions be-
low) no affect on the market value of the firm.
ownership of shares generally gives one a
proportionate vote in the stockholders'
I No loss of generality is had by interpreting I, and
Xi as vectors of inputs and outputs. In the proofs, we meeting (although some firms also issue
then need to replace Ii by v* Ii and Xi by v Xi, where v
is the vector of relative prices (at time t, in state 6(t)).
shares which do not have voting rights).
I Since in our simplification there are no other inputs, In the ensuing analysis, we shall assume
the output of the firm and its gross profit (before paying for simplicity that there are only two
interest on debt, etc.) are identical. The modifications
required when there are other inputs are straightfor-
classes of financial instruments, bonds and
ward. common shares.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
854 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

(c) If it decides to issue bonds, it must Si(t) = the number of shares outstand-
decide on what maturity-one year, two ing at the end of the period.
year, etc.-and what the coupon rate will REi(t) = retained earnings.
be. For simplicity we shall assume that
There is, of course, no natural unit for
bonds carry no coupons. Thus, a t period
shares, so it is just as simple to define
bond is a promise to pay in t periods 1 dol-
lar. When it is issued, it obviously sells at
Et (t) = qi(t)Si(t)
a discount.7 Let
as the value of the shares outstanding at
p(t, r, 0(t)) = the price at time t in state 6(t)
the end of the period, while
of a bond which promises to
pay 1 dollar at time r. Ei (t) = qi(t)S(t - 1)
If there is uncertainty, the individual will as the value of the shares outstanding at
not know what the price of such a bond the beginning of the period, i.e., the value
will be in future periods except that, if at t-th period prices of the shares outstand-
there is no bankruptcy, ing at the end of the previous period.
Thus E+(t) -E-(t) is the value of the
p(t, t, 0) = 1 for all 0, t
change in the number of shares outstand-
In the discussion below all variables are ing resulting from issuing new shares dur-
state-contingent, but for notational sim- ing the t-th period; this should not be con-
plicity we omit the 6(t) except when it fused with E (t+ 1)-E+ (t), which is the
would otherwise be confusing. Similarly, change in the value of those shares out-
all real variables (Xi, Ii) are dependent on standing at the end of the t-th period from
the "plan k," but k too will be suppressed. the t-th to the t+ 1st period. The latter is
The relationships among the various the capital gain (or loss) on existing shares.
financial decisions are expressed by the The second accounting identity states
two accounting identities: Total invest- that total income in state 0 at time t must
ment must be equal to the value of the be equal to the income distributed (to
change in outstanding bonds plus the value bondholders and to shareholders) plus
of the change in outstanding shares plus that retained by the firm.8
retained earnings:
(2) Xi(t) = Bi(t - 1, t) + Di(t) + REi(t)
00

()Ii(t) = E [p(t,-T) (Bi(t,T)-Bi(t-1,T)) where Di(t) are the dividend payments to


T-=t+lI stockholders on record at the beginning of
+qi(t) (Si(t)-Si(t-1)) ]+REi(t) the period; i.e., each share receives
D(t) S(t- 1) or the dividend per dollar in-
where
vested at time t-1 is D(t)/E+(t-1).
Figure 1 illustrates a flow of funds dia-
Bi(t, T) = the number of bonds outstand-
ing at the end of period t with gram for this economy over time. (Be-
maturity at time T. cause the flow of funds occurs over time,
qi(t)= the price of a share of the ith the diagram is not made circular.) It
firm at time t. should be noted that we have drawn the
line for retained earnings through the
7There are of course still other financial decisions,
including ownership claims on other firms, what numer- 8 Recall that we are assuming for notational sim-
aire to denominate bonds in, etc. All of these could be plicity that there are no coupons on bonds; thus bond-
included in our model-at some expense in notational holders only receive income from the firm upon matura-
complexity. tion of bonds.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 855

x(Ct, es(t))
q q(t, 01 ((t)) (E(t, 91(t)
erlp(t,-r, a,(t))

Household Sector

etoined Eorn ngs el (tt+1)


x(t,02_(t))fl
I( I -t,e9 (t -i )) e q(t I e2(t)) (E (t, e O1(t))) Dvidends > New E 1(t,Om(L)) e1(tti)
| 3 (t,r>, t).) e D((tZ() | E) (9(t)) - E e(t, eO(ttl)
\~~~~~~~~~~Rte W) -(s-(t- GZW) e3

x(t,e3(t)) ond Payments New Debt

ew P(t ,t,e5,(t)) Bt1 , 3-))

Investment Stote of Profits and Distribution of New Capitol roised New Investment
noture prices of profits to shore- (issuing shores and
announced securities holders (dividends) debt) which, together
with (shores and bondholders: with retoined eornings
associoted and bonds) remoining profits finance
retained

FIGURE 1

household sector with a dotted line: the ately been substantially narrowed. In
fact that the earnings do not pass through Figure 1 we can, for instance, now com-
their hands directly does not necessarily pletely ignore all but one of the environ-
mean that the household sector does not mental paths passing through t.
include (in some sense) such retained Given the new information embodied in
earnings in their income. As the analysis the announcement of the state 0(t), the
below shows, the retained earnings will be value of the shares and the prices of bonds
directly reflected in the price of outstand- are determined. In particular, the value of
ing shares. the equity of the firm now is Ey(t, 02(t)).
The diagram also serves to clarify the Moreover, at this point, for the particular
timing implicit in our analysis: Let us plan we have denoted by k, we know
break into the diagram at a point where exactly what the firm plans to do this
the firm has just made its "new" invest- period: we know its investment, I(t, 02, k),
ment decision having raised the requisite how much dividends it plans to give out,
capital. The output (profits) next period how much it plans to retain, how many
(which depends not only on investment in bonds and of what maturity it plans to
the period just ended, but on investment issue, how many new shares it plans to
in all preceding periods, as well as the issue, etc. We still don't know, of course,
specification of the environmental path for what its investment will be in the future;
these preceding periods) is unknown; we for this we await further information. But
await the specification of the "environ- we do assume that we know what the firm
ment" for time t, for example, the rainfall, will do in each contingency.9
temperature, etc. The "state of nature" is
I The assumption that the raising of new capital
then announced, i.e., 0(t) is then given. follows (in each period) the distribution of the profits
This means that the set of possible out- and the bond payments is made simply for expositional
convenience. In fact, these two operations may be
comes for t+1 and beyond has immedi- (over)

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
856 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

In the theorem presented in the next min Xi(0)


Bi < -
section, two concepts play a crucial role: 1 + r*
one is the value of the firm, the other is
bankruptcy. while it is positive if
The total value of the ith firm is the mmn Xi(0)
present value of its outstanding bonds plus
1 + r*
the value of its equity: at the beginning of
the period (before maturing bonds are re- See Figure 2. The analogous statement
deemed), this is here would be

(5) Xj(Ij, 0(t)) < Bi(t- 1, t, O(t - 1)


(3a) V,(t) = Et(t) + Z p(t, r)Bi(t-1, - )
T=t
But this will not do. For firms always have
the
while the value of the firm at the end of option, if their returns in a particular
the
period is period in a particular state are low, of
borrowing more or issuing new shares to
meet these debt obligations. Indeed, this is
(3b) VT(t)-=Ej(t) + E p(t, r)Bj(t, r)
T-t.+ 1 exactly what they would normally do,
provided their future prospects of returns
Using (1) to (3), we can solve for the
are sufficiently good. Therefore the condi-
value of dividends in terms of the change
tion stated in (4) is at best a statement
in the value of the firm, its gross profits,
about very short-term liquidity, not about
and its investment:
the solvency of the given firm. In the last
period of a multiperiod model with a
(4) Di(t) = Xi(t) - Ii(t) + V,(t) - V(t)
terminal date, the condition for bank-
Bankruptcy is somewhat more difficult ruptcy is given by (5), since the firm can-
to define. The basic notion is, of course, not (by assumption) issue new shares or
that the firm is unable to meet its debt borrow further. But there is no reason to
obligations. In the two-period model dis- restrict ourselves to a finite period model.
cussed in my 1969 paper, a firm is bank- Clearly, what we mean by bankruptcy
rupt whenever the profits are less than the is that at some date, in some state of
nominal claims of bondholders. nature, the value of the maturing bonds of
a firm is less than the face value
Xi < (1 + ri)Bi

where r, is the nominal rate of interest on p(t, t, 0(t)) < 1


the bond. If r* is the nominal rate of inter- for some t and 0(t). This is equivalent to
est on a perfectly safe bond, and min Xi(O) saying that at that date and in that state
is the minimum profit in any state of na- of nature the value of the equity of the
ture, the probability of bankruptcy is zero firm is zero (or negative if there is not
provided limited liability).1"

thought of as occurring simultaneously. The important (6) Et (t, 0(t)) ? 0 or


assumption is that the same price of a security prevails
at the beginning of the period as at the end. Since once
the state of nature is announced, everything that will
(6') V(t, 6(t)) < E Bi(t
occur during that period is known, this is not an un-
reasonable assumption; alternatively, if we think of the *-(t, 76 0(t))
distribution of returns and the raising of new capital
as occurring simultaneously, this is clearly the appro- 10 Clearly, if the price of a share is zero, the firm can-
priate assumption. not issue more equity to pay off the debt holders.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 857

B. Households

Let wi-(t) denote the jth individual's


X~~~~~~~~~~~~~~~~Xo wealth at the beginning of the period and
wi+(t) be his wealth at the end of the
Poyments to Stockholders period. If ES+(t) is the individual's owner-
ship of equity in the ith firm at the end of
(14.r)B{ Poyrments to Bond Holders
the period and Bj(t, r) his ownership at
FIGURE 2a. DISTRIBUTION OF RETURNS: the end of the period of bonds maturing at
No BANKRUPTCY date -r (because of the no bankruptcy as-
sumption, the bonds issued by all firms are
identical), then
l y ren ts to Stkholders
w'j (t) = E Ei (t) + E p(t r)B (t, r)
i t=t+l
Payments Ito Bond Holders

Hence, using (3), we obtain

(7) 7v j+(t) =Eaj(0)V+ (t)


Return per Dollar

invested lo Secrity
Rettan per
Dollor Invested + E p(t, r)(B (t, r)
_ _ I ~~~~~~~~~~~~in Bond r=t+ 1

FIGURE 2b: DISTRIBUTION OF RETURNS: BANKRUPTCY - a &(t)Bj(t, T))

The fact that the value of the equity of where i(t)-=EI+(t) /E(t) is the fraction
the firm in some state of nature at some of the equity of the ith firm owned by the
date in the future is zero does not mean jth individual at the end of the t-th period.
that the value of the equity will be zero At the beginning of the next period his
today; if there is some chance that the portfolio is worth
firm will not go bankrupt, clearly the
value will be positive. But it does mean (8) w (t + 1) = t a V(t) V (t + 1)
that bonds issued with maturities at the
date of potential bankruptcy or beyond
are risky securities, i.e., their terminal + > p(t + 1, r)(Bj(t, r)
r=3t+ 1
value is uncertain, and clearly the price of
these bonds will not be the same as the
aEoe ()Bj(t, r))
price of a bond whose terminal value is i

certain. A change in the financial policy


Finally, since during the period he will
which results in a chance of bankruptcy,
receive dividends equal to Ei a:DI if he
i.e., in a chance that the firm will not be
consumes ci(t), his end of period wealth
able to meet its debt obligations, thus
will be related to his beginning of period
changes the prices of the bonds the firm
wealth by the equation
issues, and it is the invariance of the price
of the bonds which will be crucial in the
(9) in (t) =-@} (t)-c (t)
argument of the next section.
We shall assume in the subsequent anal-
+ Ea'(t -1) D (t)
ysis that there is no bankruptcy.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
858 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

Thus, substituting (4), (7), and (8) into THEOREM 1: (a) Assume there is no
(9), we obtain bankruptcy of any firm in any state of na-
ture. (b) A ssume that there is a perfect
(10) ck(t)= ZE (t-1) {IXi(t)-Ii(t) market for perfectly safe bonds of all ma-
00
turities. (By perfectly safe, we mean that the
+ V_T (t)-E p (t, r7)Bi(t- 1, T) amount that they pay upon maturity is
7--t known for certain; the price of these differ-
ent maturities at all other dates may be highly
+ E p(t, r)Bj(t-1, T)-w (t),variable.) II (c) A ll firms haze already made
T=-t

their real decisions (i.e., the value of k for


allj, T
each i is given). (d) A ssume there is a gen-
Equations (7) and (10) can be thought of eral equilibrium, with all markets clearing,
as defining the individual's consumption characterized by a given price in each state
opportunity set. Given any set of owner- of nature at cach time for each maturity
ship of shares 4(t- 1) and bonds bond and each firm having a given valuation
Bi(t-1, r) at t-1, (10) defines the value in each state of nature and at each time t,
of consumption plus end of period wealth and a given financial policy (i.e., a specifi-
at t. Given the value of wealth at the end cation of debt-equity ratio, retention ratio,
of period t, (7) defines the possible owner- maturity structure of bonds).
ship of shares and bonds at t, which in turn
Then, there is another general equilibrium
defines the value of consumption plus end
solution whtere any firm (or group of firms)
of period wealth for t+ 1. Note that the
has changed any (or all) of its financial
opportunity locus does not depend on
policies, but in which the value of the firm
either dividends or retained earnings; the
and the price of all maturities of all bonds
intuitive reason for this will be made
are unchanged (for all periods and states
clearer in the discussion following the
of nature), and investors have made of-
theorem in the next section.
setting portfolio adjustments, i.e.,

C. General Equilibrium
(13) AB(t, ) = ai(t)ABi(t, r) all t, r, j
Market equilibrium requires the total
value of ownership claims on the ith firm i.e., each investor alters his holdings of bonds
equal the value of its equity, i.e., by exactly his stockholder's share of the
change in debt of each maturity of all firms
E,i(t) = a &t(t)E,(t) = E,(t) or and

(14) Aac(t) = 0 for all i, j, t


(11) Z ta(t) = 1 for all t, i
or equivalently
Similarly, demand for bonds of each ma-
turity must equal the supply: (14') AE (t) = ai(t)AE' (t)
Each investor changes his equity holdings
(12) , Bi(t, T) = B(t, r) for all t, T
it J
11 From a consutmption point of view, it cannot be
said which maturity is the "safer" bond, which is the
II. The Basic Theorem riskier; i.e., it cannot be said the long-term bonds are
riskier than short-term 'bonds. See my 1970 paper.
We are now prepared to state and prove What we shall show here is that from the point of view
our central theorem. of thefirm, the maturity structure is irrelevant.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 859

of each firm in proportionwith


tothe changed
the financial
firm's policy of the
change
in total equity capital. firm, and conversely.
2. Since the consumption opportu-
The argument of the proof is simple. I nity set is unchanged, if a particular con-
shall show that if the value of the firm and sumption path { ci*(t, 0(t)) } is preferred to
the price of all maturities of all bonds are all other consumption paths in the original
unchanged (for all periods and states of situation, then it is preferred in the new
nature) then the set of consumption possi- situation. Thus, the indicated changes in
bilities available to any individual is un- portfolios are not only feasible, but opti-
changed. Since the set of consumption pos- mal.
sibilities is unchanged, the individual will 3. If in the original situation, all
choose the same consumption path (i.e., markets cleared at each point of time (for
the same plan of consumption over time, each state of nature), they also do in the
which is clearly a contingent plan, de- new situation. Since ao(t) are unchanged,
pending on which states of nature occur). clearly if Ej aj(t) = 1 before, it still does,
To do this, he changes his investment- and all markets for securities clear (equa-
portfolio allocation plan as described by tions (11) above are all satisfied). The
(13) and (14). Finally, I show that if the change in demands for bonds of a given
set of investment-portfolio allocation plans maturity at a given date is given by
originally adopted by the different indi-
viduals in the economy (that is, before the Z E a (t)ABi(t, r) = E ABi(t, T)
j i i
firm changed its financial policy) was an
equilibrium, so that each date and in each i.e., it just equals the change in supply, so
state of nature markets cleared, then the that if demand equaled supply in the
set of new investment-portfolio allocation original situation (equation (12) was satis-
plans also constitutes an equilibrium. fied) it still does.

PROOF
III. Comments on the Theorem
and Its Proof
1. The consumption opportunity set
is unchanged. Consider any feasible con- There are four kinds of comments which
sumption path and its associated portfolio I have to make. In Section lIIA we provide
allocation. From (7), it immediately fol- an intuitive interpretation of the theorem.
lows that if wj+(t), p(t, r), and V+(t) are In Section ItIB we point out how much
unchanged after a given change in financial weaker the assumptions employed in our
policy of a firm, then the changes in port- analysis are than those used in previous
folio allocation described by (13) and (14) proofs. In Section IIIc we discuss briefly
are feasible; and from (10) if these changes the limitations on the proof, and how crit-
are undertaken, then c'(t+?)+ww+(t+1) ical they are for the general validity of the
is unchanged. Thus, if cj(t+ 1) is un- theorem. In Section IIID we discuss the
changed, wj+(t+ 1) is unchanged. Clearly, competitive forces at work to eliminate
if the values of the firm and prices of bonds the "inefficiency" resulting from the re-
are unchanged, the value of wi for dates source allocation to financial management.
before the contemplated change in finan-
A. Intuitive Interpretation
cial policy begins is unchanged.
These statements immediately imply The basic argument of the theorem is
that a consumption stream that was that individuals can exactly "undo" any
feasible in the original situation is feasible financial policy undertaken by the firm.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
860 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

Let us consider verbally what actions of classes, as many states of nature as secur-
the individuals are required to offset vari- ities, or the assumption that returns can
ous actions by the firm. Assume the firm be described by means and variances,
decreases its dividend payout ratio. This assumptions which have been crucial in
means that it has more retained earnings, other proofs of the more limited theorem
so, if the two basic financial accounting on the irrelevance of debt-equity ratios.
identities are to be satisfied, either it must
borrow less (perhaps it even lends) or (b) Competitiveness of capital market. No
issue fewer new shares. To make up for the assumption about the competitiveness of
loss in dividends, i.e., to keep the same the capital market has been made; the
consumption path, individuals buy fewer only assumption is that there be no dis-
new shares in the firm or buy fewer new criminatory pricing, i.e., the price paid by
bonds. Assume the firm simply issued one individual (firm) for a bond (or share)
fewer shares. In one case, the value of the be the same as for all other individuals.
equity grew because of issuing new shares, But the market rate of interest-and
in the other case, the value of the equity hence the interest rate paid by a firm-
grew because of retained earnings. From may be affected by the amount of capital
the point of view of the stockholders, the it raises from the market.
two are perfectly equivalent. This change
in dividend pay-out ratio thus leaves the (c) Rationality of consumers. The only
debt-equity ratio unchanged. On the other restriction on individual behavior is that
hand, if the firm decreases the number of given a set of feasible consumption paths,
bonds issued, it will lead to a lower debt- he always selects the same consumption
equity ratio. Then individuals borrow on path. Thus, the individual may maximize
their own account. One can think of it as his discounted expected utility, but no
if the individual takes the proceeds of the such restrictive assumption is required for
loan to purchase the increased equity in the result to obtain.
the firm (since the two are exactly equal,
this is only a convenient way of looking at (d) "Control" of firm. Even if the indi-
it; since all funds are fungible, there is no vidual does care about his political power
real connection between the two). The (control) within the firm (which he may if
increased borrowing by individuals ex- the real decisions of the firm depend on the
actly offsets the decreased borrowing by stockholders), if the role of each stock-
firms so markets continue to clear. Simi- holder in decision making is simply a func-
larly, if the firm decides to issue more tion of the proportion of the total shares
three-year bonds and fewer five-year he owns, the financial policy makes no dif-
bonds, the individual can undertake ex- ference, since political power of any share-
actly offsetting actions in his own port- holder is identical in the two situations.
folio. One might have argued that a smaller
equity base would make a "take-over"
B. On the Generality of the Theorem more likely; but under the assumption of
(a) Risk classes, Arrow-Debreu securities, no bankruptcy, this would not be true,
mean-variance analysis. It should be em- since the group taking over the firm could
phasized that in the proof, it is not as- borrow on the strength of the equity in the
sumed that there are two or more firms firm as collateral; if in the low equity
which are otherwise identical; the argu- situation, the group taking over could
ment does not require the existence of risk raise the requisite capital for a take-over,

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 861

they would have no problem doing so in (h) Market clearing. The particular path
the high equity situation.12 described in the above analysis is an equi-
librium path where individuals make plans
(e) Source of uncertainty. No assumption all of which are consistent with one another,
about the source of uncertainty is re- i.e., they are market clearing. In fact, the
quired.'3 only thing required for the analysis is
market clearing at time 0. In making his

(f) Multiplicity of equilibrium. Theorem 1 portfolio-consumption decision for time 0,


is a theorem about market equilibria. It the individual must have expectations of
states that there are an infinite number of prices and firm valuations at all future
general equilibrium solutions of the econ- dates and states. These may not, of course,
omy all of which are identical in all re- be realized; at each successive date ex-
spects except for the financial policies of pectations are then revised. It is important
firms and the value of bonds and shares to the analysis that these revisions depend
(separately) held by individuals (although on " real events" not on the financial
the proportions of the shares of each firm structure of the firm (see below).
owned by any individual are the same).
There may of course be more than one gen- C. Limitations on the Theorem
eral equilibrium solution to the economy There are three critical limitations on
at any given set of financial policies. As the theorem.
usual, very strong conditions would be
required to ensure uniqueness. But what
(a) Independence of expectations from fi-
our theorem does assure us is that if there
nancial policy. Our analysis requires that
are two (or three or . . . ) equilibria at a
these expectations be unchanged as the
given set of debt-equity ratios, then there
firm changes its announced financial policy
are two (or three or . . . ) at any other set
for the future.
of debt-equity ratios. The theorem has
If it should turn out that these expecta-
nothing to say about the important ques-
tions are a function of the financial policy
tion as to which one of these will in fact
of the firm, then in fact the financial policy
be chosen.
of the firm will affect its valuation this
period. The expectations that financial
(g) Differing expectations. The argument of
policy will affect market valuation are, at
the proof does not require that individuals least in this very rough sense, fulfilled.
have the same expectations. The only But note that the argument for equi-
agreement in expectations that is required librium paths shows that there is no rea-
is that the firm will not go bankrupt in any son that these expectations ought to
state of nature. (See below for a discussion change.
of this assumption.)

(b) Individual borrowing an imperfect sub-


12 For a more extended discussion of the stitute for firm borrowing. Perhaps the
relationship
between debt-equity ratios, bankruptcy, and take-overs, major objection to the proposition that the
see my 1972 paper.
13 In particular the distinction between technological
firm's financial policy is irrelevant is that
uncertainty and price uncertainty, which played such individual borrowing is not a perfect sub-
an important role in Diamond's analysis, is of no conse- stitute for firm borrowing. There are four
quence here. It should also be noted that Diamond's
assertion that his results do not depend on the no-
principle reasons for this: 1) higher interest
bankruptcy assumption is incorrect. rates for individual borrowing than for

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
862 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

corporate borrowing; 2) limitations on the the debt-equity policy to be pursued by


amount that individuals can borrow from any particular firm, only the set of debt-
the market; 3) transactions costs; and 4) equity policies that groups of firms can
special tax provisions (differential treat- follow; i.e., even if the constraint is bind-
ment of capital gains and deductibility of ing, in general one firm can increase its
interest payments for the corporate income debt-equity ratio when another firm de-
tax). I have discussed these limitations in creases its debt-equity ratio. One cannot
greater detail elsewhere (1969, 1973). Here speak of an optimal debt-equity or optimal
I wish only to make a few observations. retention ratio.
First, the higher nominal interest rates
individuals pay and the quantitative re- (c) Bankruptcy. In my judgment, the most
strictions on their borrowings are pri- restrictive assumption is that of no bank-
marily a reflection of the higher proba- ruptcy.
bility of default on the part of individuals. The careful reader may have wondered
They are, in other words, a particular where the restriction of no bankruptcy
manifestation of the general problems was used in the proof. Because of limited
that the chance of bankruptcy brings to liability laws, it is clear (as I noted before)
the analysis. that Et ?O. If the firm issues a sufficiently
Second, the first three reasons given large number of bonds so that in some
above place restrictions on the set of state of nature at some date E pBt> Vi
financial policies among which the indi- for Vi to be the same as in the original
vidual is indifferent, but there is no reason (reference) situation, Ei would have to be
to believe that these restrictions are very negative. But this is impossible.
severe. They may mean that firms can- The assumption was not only critical to
not have all-equity policies, but indi- the proof, but I would argue, critical to
viduals will still be indifferent among a the general validity of the theorem. To
wide set of debt-equity ratios. If, for put it one way, it is not reasonable to
instance, a firm were to decrease its debt- assume that the price of bonds for which
equity ratio, the analysis does not require there is a positive probability of default at
that individuals borrow from the market maturation would be the same as a per-
to purchase the additional shares issued fectly safe bond. One might argue that
by the firm; it only requires that they de- the decline in the nominal value of bonds
crease their holdings of bonds. Hence so is compensated for by an equivalent in-
long as the total debt of those firms whose crease in the value of equity, and under
shares the individual owns is sufficiently certain circumstances-the existence of
large that the individual is a net lender as many securities as states of nature or
rather than a net borrower, the individual the mean-variance model with homoge-
is indifferent. This places a lower bound on neous expectations-this is true. But in
the "average debt-equity ratio" of the the more general case, bankruptcy changes
firms in the individual's portfolio (although the opportunity set facing a given indi-
not on the debt-equity ratio of any single vidual so that the value of the firm is
firm). This constraint may become an im- changed. Not only is the financial policy
portant constraint if at those debt-equity of importance, but no separation between
ratios there is a finite probability of bank- the financial and real decisions is pos-
ruptcy; that is, it is only in conjunction sible.14 (See my 1972 paper.)
with the bankruptcy constraint that this
constraint becomes significant. 14 These remarks should serve to clarify the difference
Third, it does not place restrictions on between my theorem, both its meaning and its proof,

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 863

D. Competitive Forces to Eliminate "Waste" pectations about profits may not be as


of Resources on "Financial Management" "irrational" as the above analysis sug-
gests. Changes in financial policy may be
One might ask, if financial policy is
an important signal for the "real pros-
really of no importance, why do firms
pects" of the firm. This would not be the
waste resources on "money managers"-
case in my model, because in it there is no
shouldn't competitive forces lead all firms
such thing as a liquidity crisis; but in the
to ignore financial policy? Since worrying
about it costs resources and can't increase
real world bankruptcy may be important,
and the fact that banks and other lend-
the market valuation of the firm, clearly
firms that spend resources on financial
ing institutions are unwilling to lend the
firm money (for instance forcing a reduc-
management have lower profits to dis-
tion in dividends to meet the liquidity re-
tribute to their stockholders than firms
who don't. There are five answers to this:
quirements) may be a signal that those
1) I have ignored some important who know more about the prospects of the
firm than the relatively uninformed share-
considerations, in particular, taxes, which
holder are not sanguine about the pros-
do make it profitable to worry about
financial structures. Does this mean that I pects of the firm.15

believe that in the absence of taxation 4) There is, moreover, no reason that
in the short run the different valuations
financial managements would wither
lead to any inconsistencies or more gen-
away? Not necessarily, or only very
slowly, as the remaining points argue. erally that there are any forces leading
individuals to reformulate their expecta-
2) I have already argued that if indi-
viduals believe that financial policy affects tions so that valuations are independent
firm valuation, then it will, and the firm of financial policies. Even if we have two
firms which are identical in every real
that ignores the popular "prejudices" may
respect (that is, they belong to the same
do worse than one which takes them into
risk class, in the terminology of Modigli-
account. There may have been no rational
reason for the prices of tulip bulbs to rise ani and Miller), there is not necessarily any
method by which individuals can arbi-
as they did in the tulip bulb mania, but
since they were rising, at least in the short trage (over any short- or medium-run16

run, one could make a "profit" by invest- period) .17

ing in them (see below). 15 Indeed, one might argue that this signalling effect
of financial policy is one of its more important functions.
3) Moreover, this relationship be-
If firms never issued dividends, simply retaining earn-
tween the firm's financial policy and ex- ings (even in the form of bond purchases) then it might
be possible for firms to postpone letting shareholders
and that of Modigliani and Miller. They assert that know when they are in "bad straits" even longer than
the financial policy is of no consequence. But Modigliani they do at the present. This may provide part of the
and Miller made use of risk classes in their proof, explanation of why, in spite of strong tax advantages
the use of which seemed to imply objective rather for not issuing dividends, firms continue to do it.
than subjective probability distributions over the pos- 16 This qualification is imposed because, under certain
sible outcornes. The mechanism which ensured that the circumstances, it can be shown that if different financial
debt-equity ratio made no difference for the value of the policies are pursued with the firms having different
firm was individual arbitraging among different firms in valuations and equal returns to the individual, then, in
the same risk class. Such arbitrage does not play any finite time, the relative valuations must become infinite.
role in my analysis. Moreover, their argument was But finite, in this context, may be very long indeed.
based on partial equilibrium analysis rather than gen- Such differences in valuations are (at least mathematic-
eral equilibrium analysis. It was not clear from their allv) very similar to the speculative booms (or depres-
analysis whether the theorem held only for competitive sions) which often seem to characterize price move-
markets, nor how the possibility of firm bankruptcy ments on the stock market. For a general discussion of
affected their results. The basic insight of the M-M these problems in a slightly different context, see Karl
analysis, that individual leverage could substitute for Shell and the author.
firm borrowing, remains the basis of my argument. 17 To see this in the extreme case, we need only con-

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
864 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

5) Finally, we note that the resources acterized by a given market rate of interest
wasted on financial management may be (on safe bonds), by a given nominal rate of
relatively minor (relative, say, to total interest on the risky bonds of each of the
profits of the firm) and hence the "com- firms which faces a chance of bankruptcy,
petitive forces" to eliminate this ineffi- and by each firm having a given market valu-
ciency may operate with relatively little ation and a financial policy (dividend-
strength. retention ratio, maturity structure of debt,
etc.), and in which a given fraction of the
IV. Irrelevance of Financial Policy shares of thefirm are owned by the i-th indi-
of Any Particular Firm
vidual.
The above proposition established the Let any firm (or any group of firms)
irrelevance of the financial structure of change its financial policy. If financial
the economy as a whole. The crucial as- intermediaries may be established costlessly,
sumption employed was that of no bank- then there exists a new general equilibrium
ruptcy. We can remove this assumption solution for the economy with the same mar-
and prove a weaker theorem about the ket rate of interest, in which every firm has
irrelevance of the financial policy of any exactly the same market valuation as before,
particular firm. and in which the proportion of each firm's
shares owned by the i-th individual, either
THEOREM 2: Assume there is a general directly or indirectly through intermedi-
equilibrium for the economy which is char- aries, is exactly the same as before.

sider the situation where profits minus investment are


known for certain and the firm issues no new shares.
Since the argument for changes in debt-
Then dividends for, say, firm 1 may be written (in a equity ratio is perfectly analogous to
continuous time formulation) changes in other financial policies, I shall
Dl(t) = XI(t) -Il(t) -rBl(t) + Al focus my remarks on the debt-equity
where all bonds are assumed to be short-term bonds, ratio. Assume in the initial equilibrium,
earning an instantaneous rate of return of r(t). Assume there were no financial intermediary pur-
there are two firms which are identical except that one
issues no bonds at all, i.e.,
chasing bonds and shares of the given
firm. The firm changes its debt-equity
Dl(t) = Xl(t) - I(t)
ratio. A financial intermediary is created
The total rate of return from owning shares in either
company is simply the sum of dividends and capital
which reconstitutes the firm, i.e., purchases
gains, which we require to be the same for both. Drop- all of its bonds and shares, then issues
ping the subscripts on X and I, we have bonds and shares in exactly the same ratio
E1 D1 V1 X -I rBi 12 X-I as in the original situation. The oppor-
p - + = -- + - = -- +
E1 E1 E1 E1 E1 V2 V2 tunity set facing the individual is com-
or pletely unchanged, and hence the market
i1 V2 B1 (X-I) (X-I) valuations, rates of interest, etc. are com-
V1 V2 V1 V1 V2 pletely unchanged.'8
It is clear that if p = r, and V1 = V2 initially, then the
two firms have the same value forever. But note that the 18 More generally, assume the original debt-equity
second equation can also be satisfied with p = r and ratio of the firm is d, and the kth financial intermediary
V1 F# V2, in which case there will be cumulative changespurchases ak of the bonds and equity of the firm and
in the ratio of the valuiations. Eventually, these will issues a debt-equity ratio of dk. Thus 1- ot ck iS the
probably lead to one of the firms having an unusually proportion of the firm purchased directly by individuals
low or unusually high earnings-valuation ratio, and this (not through intermediaries). Now assume the firm
will probably lead to a revaluation of the firm. But changes its debt-equity ratio to d'. Then all interme-
there is no reason (without perfect futures market), as diaries except the one for which dk=d' are unaffected.
we have argued in detail elsewhere, that this might not It now issues a debt-equity ratio of d, and purchases
go on for a long time. 1- 1 ak of the bonds and equity of the firm.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
VOL. 64 NO. 6 STIGLITZ: CORPORATE FINANCIAL POLICY 865

One might argue, however, that the op-


Income x (e)
portunity set has been changed, because in
principle the individual can buy bonds
I EI
and shares in the firm directly as well as Distribution [ I |
I r( a X-/ '
through the intermediary. Thus, if there
is a probability of bankruptcy, his oppor-
tunity set is larger now than it was before.
This may lead to an increase in the de-
mand for the securities (bonds and stocks)
o 8, ox E
of the given firm, so that the new situation
is not an equilibrium one. There will be a ofk(1 +r) B o+(x- ( +r) 8)
new general equilibrium situation, with ok

the value of bonds and equities of the


firm greater than before. But this would B E

imply that the original situation could not (+r)B' 0c X - (I t-F r) B


have been an equilibrium. For a financial
intermediary could have purchased and FIGURE 3. OPERATION OF FINANCIAL INTERMEDIARY
issued the same fraction of the bonds and (r(f) =interest payments on bonds in state 0)
stocks of the given firm, thus obtaining a
given fraction of the income of the firm in
The fact that we do not see such a pro-
every state of nature, and then issued
liferation of financial intermediaries sug-
bonds and shares in the ratio of the debt-
gests either that they are unnecessary
equity ratio of the "new" situation. The
(that is, Pareto optimality can be ob-
organizers of the intermediary would have
tained with a limited number of such inter-
then made a pure profit for themselves,
mediaries,20 or the conditions of Theorem
equal to the difference between the value
1 are satisfied so that the financial struc-
of the firm in the two situations.
ture is of no consequence), or that there
The point of the argument is the follow-
are significant transactions costs relative
ing: If, corresponding to a given set of real
to the gains to be had by the creation of
and financial decisions of the other firms,
such intermediaries.2'
the financial decision of the firm does
make a difference, free entry of financial
V. Concluding Comments
intermediaries will ensure that a set of
financial securities will be marketed which We have established the irrelevance of
maximizes the value of the firm regard- financial policy under a fairly general set
less of the debt-equity ratio of the firm. of conditions. Three classes of limitations
If one took the assumption of costless were noted to our result :22 1) expectations
creation of intermediaries seriously, there of real returns dependent on firm financial
is no reason to suppose that the process of
20 As in those circumstances in which the portfolio
proliferation of intermediaries would stop
separation theorem is valid. See David Cass and the
short of creating as many securities as author.
states of nature; in which case, not only is 21 Included in "transactions costs" are the cost of
obtaining information about different securities. If there
the financial structure of any individual
is a finite probability of bankruptcy, purchasers of
firm of no consequence, the financial bonds have to evaluate the riskiness of the bonds. Thus
structure of the economy is irrelevant.19 not only does bankruptcy result in transactions costs
when it occurs, but the potentiality of bankruptcy re-
sults in transactions costs at the time the bonds are sold.
19 That is, all financial structures that provide as 22 Besides the obvious distortionary effects of taxa-
many securities as states of nature are equivalent. tion.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms
866 THE AMERICAN ECONOMIC REVIEW DECEMBER 1974

policy; 2) individual borrowing not a per- ing of the implications of these limitations
fect substitute for firm borrowing; 3) bank- and more refined tests to discriminate
ruptcy. Whether these limitations are im- among the alternative hypotheses.
portant in practical applications is a moot
question. But whether they are or are REFERENCES
not, the theoretical importance of the W. Baumol and B. G. Malkiel, "The Firm's
theorem is not diminished: an understand- Optimal Debt-Equity Combination and the
ing of it is to corporate finance as an under- Cost of Capital," Quart. J. Econ., Nov.
standing of the frictionless surface is to 1967, 81, 547-78.
the understanding of the physics of mo- D. Cass and J. E. Stiglitz, "The Structure of
tion. For some practical problems, fric- Investor Preferences and Asset Returns, and

tion can be ignored in a first approxima- Separability in Portfolio Allocation: A Con-


tribution to the Pure Theory of Mutual
tion; in others it cannot. But even when it
Funds," J. Econ. Theory, June 1970, 2,
cannot, an understanding of what would
122-60.
happen in the absence of friction is essen- P. Diamond, "The Role of a Stock Market in
tial. The empirical testing of the model is a General Equilibrium Model with Tech-
another matter: in physics we can attempt nological Uncertainty," Amer. Econ. Rev.,
to approximate a frictionless surface and Sept. 1967, 57, 759-76.
observe motions under those conditions; N. Kaldor, "Marginal Productivity and the
to do the analogous thing here would re- Marco-Economic Theories of Distribution:
quire us to create a world without transac- Comment on Samuelson and Modigliani,"
tions costs, tax distortions, and other fric- Rev. Econ. Stud., Oct. 1966, 33, 309-19.
J. R. Meyer and E. Kuh, The Investment De-
tions, and see if in these circumstances
cision: An Empirical Study, Cambridge
firms ignored their financial structure. It
1957.
is, of course, essentially impossible to do
F. Modigliani and M. H. Miller, "The Cost
this. Fortunately, the issue is not whether
of Capital, Corporation Finance, and the
under those circumstances the financial Theory of Investment," Amer. Econ. Rev.,
policy would be irrelevant-most of us June 1958, 48, 261-97.
would agree that it would be-the issue is: K. Shell and J. E. Stiglitz, "The Allocation of
how significant are the "limitations" and Investment in a Dynamic Economy," Quart.
in what way do they affect corporate J. Econ., Nov. 1967, 81, 592-609.
financial policy. The tests performed so J. E. Stiglitz, "A Re-Examination of the
far-such as examining the value (per Modigliani-Miller Theorem," Amer. Econ.
unit scale) of firms thought to be essen- Rev., Dec. 1969, 59, 784-93.

tially identical except for their debt-equity ' "A Consumption Oriented Theory of
the Demand for Financial Assets and the
ratio-do not discriminate between worlds
Term Structure of Interest Rates," Rev.
in which Theorem 1 is valid, those in
Econ. Stud., July 1970, 37, 321-51.
which Theorem 2 is valid but not Theorem
"On Some Aspects of the Pure The-
1, or worlds in which neither theorem is ory of Corporate Finance, Bankruptcies and
valid (in which financial policy is impor- Takeovers," Bell J. Econ., Autumn 1972, 3,
tant, but in which value-maximizing firms 458-82.
have selected the set of financial policies ,"Taxation, Corporate Financial Policy,
which maximize the firm's valuation). and the Cost of Capital," J. Publ. Econ.,
What is required is a greater understand- Jan. 1973, 2, 1-34.

This content downloaded from 200.128.60.61 on Thu, 02 Nov 2017 01:23:20 UTC
All use subject to http://about.jstor.org/terms

Das könnte Ihnen auch gefallen