WEALTH MAXIMIZATION AND THE COST OF

CAPITAL*
ALAN J. AUERBACH
I. Introduction, 433.Il. The model, 434.Ill. The problem, 435.IV. Wealth
maximization: no personal taxes, 436.V. Wealth maximization with personal taxes,
438.VI. Firm value and the cost of capital, 440.Vu. Financial policy and the cost
of capital, 442.VHI. Conclusion, 445.

I. INTRODUCTION

In a simple world of certainty, with perfect capital markets, no

taxes on capital income, and all investment financed through direct
ownership, a utility-maximizing investor would strive to maximize
the present value of his investment. This wealth maximization would
be achieved by the acceptance of all investment projects having a
positive present value when discounted at the individual's personal
rate of time preference.'
Once the possibility of corporate finance is introduced, complications arise concerning the optimal choice of financing method and
the appropriate discount rate to use in present value calculations. In
the absence of taxation certain principles are generally accepted. First,
it is irrelevant whether equity-source investment funds come from
retained earnings or the sale of new shares. Second, firms should use
a composite "cost of capital" in their discounting decisions, a weighted

average of the interest rate on debt and the rate of time preference
of stockholders. Finally, if the cost of capital varies with the degree
of debt finance, or "leverage," firms should choose the debt-equity
ratio for which it is minimized. While many authors have arrived at
these results starting from the objective of wealth maximization, the
exact relationship remains somewhat unclear. In addition, there has
been considerable debate over the impact of corporate and personal
income taxes on firm policy.2
In this paper we explore the issue of wealth maximization and
the implied behavior of the firm, paying particular attention to the
results discussed above and how they are affected by the existence
* The author would like to thank Martin Feldstein and Jerry Green for many
stimulating discussions relating to this topic.
1. Modigliani and Miller 119581 have argued that, in the absence of taxation, it
cannot.

2. See Farrar and Selw)m [1967], Pye [1972], Stapleton [1972], Stiglitz [1973, 19761,
King 11974], Miller [1977], Auerbach [1979], and Bradford [1978].

1979 by the President and Fellows of Harvard College. Published by John Wiley & Sons, Inc.
00:/:3.553://79/u09;t.04J:/o1 (st

The Quarter/v Journal of Economics, Augut 1979

434

QUARTERLY JOURNAL OF ECONOMICS

of capital income taxes. Our results indicate that a tax structure

similar to that in existence in the United States influences the cost
of capital in a very different way than has been assumed previously
and that the relative advantages of debt over equity as a method of
finance, and capital gains over dividends as a vehicle for personal
realization of corporate profits, may have been greatly overstated.
These findings may help to explain certain aspects of corporate financial behavior that have seemed puzzling.
II. THE MODEL
We consider the behavior of competitive firms in a discrete-time,
infinite-horizon model. These firms finance their investment projects
through sales of common stock, retention of earnings, and sales of
debt, which, for the sake of simplicity, we take to have a term of one
period. At the beginning of each period, firms distribute dividends
to pre-existing shareholders, pay interest, and repay principal on debt
outstanding in the previous period, and sell new shares, ex dividend.
Corporate profits are taxed at rate , but interest payments may be

deducted from taxable profits. Dividends are taxed at rate 0 at the

personal level. We assume that capital gains are taxed upon accrual
at a rate c <0. We thus abstract from the issue of tax deferment that
arises when, as in the United States, capital gains are taxed upon realization, and concentrate on the favorable rate at which capital gains,
relative to dividends, are taxed. All stockholders face the same personal tax rates, (1 and c.
The one-period discount rate of equity owners during period t
is Pt,3 while the interest rate is The ex dividend value of pre-existing

shares at the beginning of period t is denoted V. The value of new

equity sold is V. Thus, the ex dividend value of the firm's equity at
the beginning of period t is
(1)

V=V+V.

The degree to which equity is "diluted" by new equity sales may be

represented by the fraction,
= V/V.
(2)
Debt outstanding in period t is B, arid the firm's "leverage" is defined

by the term,
(3)

= B/(B + Vi).

3. Note that p is the rate at which stockholders discount nomLnal flows from equity. If inflation is present, at rate ir, the real discount rate is p

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

435

If we let x represent the cash flow of the firm at the beginning of

period t, net of corporate tax, but before account is taken of net sales
of debt and equity, interest payments, and the tax savings on such
payments, then dividends paid at the beginning of period t + 1 (i.e.,
the end of period t) on shares held during period t are

D = Xt+i + B+1 + V1 [1 + i(1 T)]Bt.

(4)

By the very definition of dividends, D must always be nonnegative.

III. THE PROBLEM

We assume that firms seek to maximize the wealth of existing

shareholders. This is accomplished by choosing, at the beginning of
period t, an investment policy, represented by the vector i = (Xt,
. .), a debt policy, B = (Be, B+1,. . .), and an equity policy, VN = (Vi",
.), which, among feasible choices, maximize the value of
shares owned as of the beginning of period t, including the concurrent
after-tax distribution.
Before proceeding any further, we must derive an expression for
the valuation of equity. We assume that shares are equal in value to
the present discounted value of the distributions their owners receive.
At the end of period t, the total dividends paid to all stockholders are

D. However, if there are personal taxes, the net distribution is

smaller. The personal tax liability at the end of period t is the tax on
dividends
plus that on accrued capital gains on existing stock,
c (V Vi). Thus, the net distribution received by all shareholders
at the end of period t is

E = (1 O)D1 c(V?+1

(5)

V,).

If new equity is issued between period t and a subsequent period

s, not all of the distribution E. will go to shares held as of period t. A
fraction will go to the shares arising from the new issues. The fraction
of shares held during period s that were in existence before the start
of period t is, from (2),
= (1

(6)

)(1

(1

Thus, the distribution at times t to shares owned as of the begin-

ning of period t is

and the value of such equity is

V=

(7)
4.

fl (1 + pY'

.=1 z=t

The choice of i, , and

determines the dividend policy D =

436

QUARTERLY JOURNAL OF ECONOMIC.

Using (1) and (2), and applying (7) for successive values oft, we obtain
ptVt = E + (V1 Vi).

(8)

(8) states that, for all t, the one-period holding yield on

equity, which includes the net distribution and capital gain, must
equal the rate of time preference associated with equity holding Pt
Equation

The wealth of existing stockholders that the firm seeks to maximize is

(9)

W?

= V + E_1.

Using (1), (4), and (5), we may rewrite this expression as

(10) W =

(1

c)Vt + (1 0)
[1 + i_1(1 r)]B_1 +

Xt}

(0 c)V + Vt_i.

B_1 are predetermined at the beginning of

period t, W is maximized if and only if the firm maximizes
(11) W = (1 c)V + (1 0)B + (1 O)Xt (0 c)V.
Since itI, V1_1, and

We are now ready to examine the characteristics of wealthmaximizing behavior. Because of the complexity of the general
problem, we attack first the simpler case in which 0 =

0.

IV. WEALTH MAXIMIZATION: No PERSONAL TAXES

In this situation the firm's objective function (11) reduces to

(12)

WV+B+x1.

Firms desire to maximize the sum of their securities' market value and

current cash flow. This is a standard result.

To simplify the exposition in this and succeeding sections, we
introduce the term,
(13)

F=

V1,

which is the dividend payable at the end of period t, before account

is taken of the sale of new equity.
Using (1), (5), (8), and (13), we obtain for the case with no personal taxes
(14)

V = (1 + pY1(V+1 + Ft).

437

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

Solving for V, we obtain5

(15)

V=

fl

st 2t

(1

+ p' F

Thus, as first argued by Miller and Modigliani [1961], the particular

source of equity funds used to finance investment has no impact on
the firm valuation. Once x and B are determined, so is V, since F is,
by definition, independent of VN. Because of this result, we may ig-

nore the question of share sales and repurchases when considering

the question of wealth maximization in the absence of personal
taxes.
Using expressions (1) and (13) to substitute into (14), and then
rearranging terms, we obtain an expression for shareholder wealth:
(16)

W = Xt + (1 + r)1W+1,
-

where we define the term,

(17)

rt = bi(1

r) + (1 b)p,

to be the "cost of capital," an average of the returns the firm must earn

on equity holdings and debt, weighted by their respective portions

in the firm financial structure. Because of the deductibility of interest
payments, the net cost of debt is i (1 r), rather than the gross interest rate.
Solving (16) for W, we obtain6
(18)

W=x +

st+1 z=t

(1 + r)1 x.

Shareholder wealth may be expressed as the present value of the

stream of after-tax corporate cash flows, where the financial policy
determines the appropriate discount rate.
It is clear that the cost of capital, defined above, determines the
cutoff point for investment projects. Given the choice of financial
policy and, hence, the value of r = (r1, rt+i,. . .), the firm will increase
W by accepting those projects having positive present value when
the cost of capital in each period is used as the discount rate.
Up until this point we have said very little about the determi5. This result is dependent on the assumption that the term fl., (1 + pY'VT

approaches zero as T approaches infinity; that is, the firm's value grows at a rate less
than p. This restriction is necessary to rule out "chain letters" in which F is continually
less than or equal to zero and dividends on existing shares are supported solely by the
sale of new shares. A similar assumption will be made concerning the value of W in
equation (18) to rule out the continual sale of ever larger amounts of new debt to support
payments to existing holders of debt and equity.
6. See note 5.

438

QUARTERLY JOURNAL OF ECONOMICS

nation of p and i, or about the place of risk in our model. A completely

adequate treatment of risk is beyond the scope of this paper. It is even

questionable whether firms should always seek to maximize shareholder wealth when there is uncertainty as to the particular state of
nature that will occur in each period. One fairly simple, albeit imperfect, approach is to assume that the market evaluation of a firm's
"riskiness" increases with the degree of leverage, and that the required

rates of return p and i must increase;7 that is,

(19)

Pt

= p(b), it = i(bt), p', i' > 0.

If we assume that Pt and t are determined in this way, it follows from

(17) and (18) that, for any given stream x, the firm maximizes W by
choosing b for each period s t to minimize the cost of capital. Thus,
the overall wealth maximization, in this case, may be viewed as a
two-stage procedure in which firms first determine the financial policy
that minimizes the cost of capital faced in each period, and then use
this cost of capital as the discount rate in determining the optimal
investment strategy.
V. WEALTH MAXIMIZATION WITH PERSONAL TAXES

A more interesting, and more difficult, problem arises once we

remove the assumption that fi = c = 0. Many of the familiar results
of the previous case must be altered to take account of the existence
of personal taxes.
The first difference, seen in the definition of W* (in equation
(11)) is that the firm should not strive to maximize the sum of current
cash flow and the market value of its securities, unless 0 = c. This
condition is, in our model, equivalent to the integration of corporate
and personal income taxes, since all corporate source income, whether

retained or distributed, would be taxed at the same rate; T' = r + 0

= T + c.
Following the same procedure used to obtain equation (15), we

get, after various substitutions,

(20) v = st

[+
z=t

Pz

1c)1]

1c

(0_1c
C) v+1J.

Comparing (20) to (15), we note that, unless 0 = c, taxes enter in a

complex way into the valuation formula. However, there is one extremely important observation that can be made. For any given values
7. See Feldstein, Green, and Sheshinski [1979] for a recent example of this approach.

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

439

of x and B, and hence F, a decrease in the net issue of new equity, V,

at any time s > t, increases V and hence W, since 0> c. Thus, it will
never be optimal to issue new shares and pay dividends at the same

time, since V may be increased by an equal decrease in D5_1 and

V.

The same logic would compel firms to continue decreasing VN

below zero, repurchasing equity, as long as dividends were positive,
if such behavior were not otherwise hindered. This dominance of share

repurchases over dividends as a method of distribution has been

recognized in the past by many authors.8 However, share repurchases

have never constituted an important part of firm distributions. Part

of the answer probably lies in Section 302 of the Internal Revenue
Code, which prohibits firms from repurchasing shares in lieu of distributing dividends.
Recent U. S. data confirm that new share issues have constituted
only a small portion of new equity finance, and that repurchases have
been a very unimportant method of distribution, compared to dividends. For example, in 1976, net issues of debt by U. S. nonfinancial
corporations amounted to $47.8 billion, retained earnings were $44.3
billion, and dividends were $32.2 billion. Net issues of equity were only

$10.5 billion,9 and share repurchases were, in turn, only a small percentage of this number. 10
Given the above discussion, we shall henceforth assume that
firms neither issue new shares nor repurchase existing ones. While
not greatly abstracting from reality, this restriction constitutes a very

helpful simplification. Assuming that V 0, we see that (20) becomes

=

(21)

st z=t
[n (i

2
+ 1

cY'l

1
(-_)F

Following the same procedure used in deriving (18), we obtain

(22)

W=

(1

6) x +

sI

fJ
st+1 zt

(1

+ r2)' x,

where the cost of capital is

8. Among them, Bierman and West [1966], Pye [1972], and King [1974].
9. The preceding figures were obtained from the 1976:IV issue of the Flow of

Funds Accounts published by the Board of Governors of the Federal Reserve

System.
10. The breakdown between new issues and repurchases is nOt given for nonfinancial corporations. However, the corresponding figures for all corporate business
in 1976 are listed in the December 1977 issue of the Federal Reserve Bulletin. While
gross sales of new equity amounted to $14.1 billion, repurchases were only $3.1 billion.

440
(23)

QUARTERLY JOURNAL OF ECONOMICS

rt =

[(1

c) (0 c)b]'[bi (1 T)(l

0) + (1 b )Pr].

As in the case without personal taxes, the firm maximizes wealth,

given financial policy, by choosing all projects that have a positive

present value when discounted by the cost of capital, and if i and p
are determined by (19), should choose the degree of leverage that
minimizes the cost of capital.
The formula for the cost of capital given in (23) differs from resuits obtained previously, and there are a number of interesting observations to be made concerning this expression. If the firm is financed solely through debt, then r reduces to the net of tax rate of
interest, i(1 r), which the firm must pay on its debt. At the other
extreme, an equity-financed firm faces a cost of capital of p/(1 c).
There are two rather surprising aspects of this result. First, the cost
of capital does not depend on the dividend payout rate. Second, the
rate of dividend taxation does not enter at all. Equity gains are effectively taxed at rate c, which may be very small compared to the
statutory rate of taxation of dividends. Even for the general case in
which there is both debt and equity in the firm, the payout rate does
not directly influence the cost of capital.

The next section contains an example that should aid in the

understanding of the above results.
VI. FIRM VALUE AND THE COST OF CAPITAL

In this section we consider the particular case of an economy in

which there is no inflation, in which i, p, b, and hence r, are all constant, and assume that firms produce output using homogeneous,

nondepreciating capital and labor, subject to a constant-returnsto-scale technology. These assumptions are made to facilitate the
exposition and do not influence the characteristics of the results.
The investment in an additional unit of capital at any given time
t will decrease concurrent cash flow by the purchase price, and increase the cash flow in each succeeding period by the marginal product
of such capital, less corporate taxes, ['(1 r). Such a project increases
W by f'(1 T)/rI 1. Perfect competition ensures that firms will
invest until the adoption of new projects does not increase wealth; that

is, the after-tax marginal product of capital must equal the cost of
capital:
[(1 r) = r.
(24)
Given that firms equate the wage rate with the marginal product
of labor, the cash flow available to the firm at the beginning of period

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

441

t if no new investment occurs is, by Euler's Theorem, ['(1

where K_1 is the capital on hand during period t 1. The net cash
flow at the beginning of period t is therefore this amount less new

investment:

Using

(22), (24), and (25), we obtain

(26) W = (1 0)

+
I.

= (1 O)[Xt
Substituting

(27)

(K K1_1).

x = ['(1

(25)

st1

(1 + r)(t)[(1 + r)K8_1 K8]

+ Ks].

(26) into (11), we obtain

V =(--)

[K Be],

which is similar to results obtained by Auerbach [1978] for the case

of all equity finance and Bradford f 1978] for the case in which r = c
= 0. Since 0 > c, the market value of equity is lower than the reproduction cost of its capital stock, less the market value of its debt. In
the language of Tobin [1969], "q" is less than unity.

Since this capitalization occurs, the presumed superiority of

capital gains over dividends disappears, and the effective tax on equity

income is lower than a weighted average of the statutory taxes on

dividends and capital gains. For example, imagine an all-equity firm
considering taking one dollar out of current dividends for reinvestment. The net loss to stockholders in the current period is (1 0),
since their dividend tax liability is also reduced. The capital stock is
increased by one unit, and by (27) the stock's market value goes up
by (1 0)1(1 c). Therefore, current capital gains taxes are increased
by c(1 0)1(1 c). If the stockholders then sell off the value of the
stock corresponding to the new investment, they receive a capital gain
of(1 0)/(i c), and the value of their remaining stock is the same
as it would have been if the dividend had been paid out. But the total
distribution is (1 0)/(1 c) c(1 0)1(1 c) = (1 0), which equals
the dividend foregone. Thus, payout policy is irrelevant from the point
of view of the stockholders.
Now, suppose that instead of selling off the increase in stock
value, individuals hold it forever. Also, suppose that all returns from
the new capital in succeeding periods are paid out in dividends. (Since
we have just demonstrated the irrelevance of payout policy, this assumption poses no restrictions.) The initial cost is then the lost dividend (1 0), plus the capital gains tax accrued c(1 0)/(1 ). for

442

QUARTERLY JOURNAL OF ECONOMiCS

a total of (1 0)1(1 c). The net distribution in each period is the

marginal product of capital, less corporate and dividend taxes, ['(1

r)(1 0). Thus, this marginal asset has zero present value, discounted at p, when f'(l r) = p1(1 c), the all-equity cost of capital
derived in the previous section.'1
VII. FINANCIAL POLICY AND THE COST OF CAPITAL

Thus far, we have had little to say about the choice of financial

mix between debt and equity. One important question to ask is

whether, in the absence of any notion of risk or bankruptcy, firms will
finance with both equity and debt or, as has been suggested by Stiglitz
[1973], will finance with only debt at the margin.

If all future streams are certain, then the degree of leverage of

a firm has no impact on the values of p and i that it faces. Shares in

all firms are perfect substitutes, as are bonds, so that firms are
price-takers with respect to p and i.
Differentiating (23) with respect to b, for i and p constant, we
obtain (dropping subscripts)
(28)

1(1

c)

(0

c)b]2(1

0)[i(1

r)(1

- c)

p],

which is either always positive, always negative, or always zero. Thus,

firms will desire all debt if i(1 r)(1 c) <p, all equity if i(1 r)(1

c) > p, and be indifferent if i(1 T)(1 c) = p. The determination

of i and p depends, in turn, on the characteristics of individual investors. We consider two cases, one in which all investors face the same

tax schedule, the second in which they do not. In each case we let
denote the individual tax rate on debt income.
Case 1: One Class of Investors
Without risk, debt and equity should be perfect substitutes, and
investors will hold whichever yields a higher net rate of return, holding
both only if these returns are equal. For an equilibrium to exist with
debt and equity present, both firms and individuals must be indif-

ferent between the alternatives. From our discussion above, the

condition for firm indifference is i(1 r)(1 c) = p. By construction,
individuals receive p from holding equity. After paying a personal
11. There is a question as to whether this undervaluation equilibrium is feasible.
In another paper we explored this problem in the context of an all-equity, overlapping-generations growth model (Auerbach, 1979) and found that such a regime could
occur there only if the after-tax marginal product of capital, ['(1 r), exceeds the
population growth rate.

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

443

income tax, they receive i(1 ) from holding debt. Thus, individuals
will be indifferent if and only if i(1 0) = p. Combining these two

conditions, we have the result that debt and equity will exist together
in equilibrium if arid only if (1 r)(1 c) = (1
Now assume that we are initially in a position in which both debt
and equity are held, so that i(l = p, and suppose that (1 T)(1

c) > (1 ). Then i(1 r)(1 c) > p, and firms will shift entirely
to equity. Similarly, if (1 r)(1 c) < (1 0), they will shift entirely
to debt. Thus, the equilibrium will be characterized by all equity, all
debt, or both according to whether (1 r)(1 c) is greater than, less
than, or equal to (1 4).12
It is interesting that the tax on dividends 0 does not enter directly
into the process at all)3 Since the question of which method of finance

will be used depends on c, it would be helpful to relate this hypothetical tax on accrued capital gains to the currently existing tax on
realized gains. Bailey [1969] has shown that the effective tax rate on
accrued capital gains approximately equals the statutory rate on realizations, which cannot exceed 0.50, times the ratio of realizations
to accruals, which he estimated to be 0.2 for the long run in the United
States. If we let c = 0.10 and assume that 0 = , then the equilibrium

will be characterized by all equity, all debt, or both according to

whether (1 is greater than, less than, or equal to 0, where
(29)

0 = r/(0.9 + 0.1T).

If we take r to be the U. S. statutory rate 0.46, then 0 = 0.486. That

is, an all-debt equilibrium will occur only if 0 <0.506. In comparison,
the maximum statutory rate on personal capital income in the United
States is presently 0.7.
In the model just presented, an equilibrium will have all equity
or all debt according to whether or not "the" tax rate on personal income exceeds 0. In reality, however, the marginal tax rates faced by
individuals vary greatly.
Case 2: Two Classes of Investors

Suppose that there are investors of two types, differing only by

their marginal tax rates. Let (Oi, , c1) be the rates faced by class 1,
and let (09, 02, c2) be those faced by class 2. For simplicity, we again
assume that 0 = and 02 = 02. Our objective is to derive the condi-

tions under which debt and equity coexist in equilibrium, with one

= 0.

12. A similar result has been derived by King ]1974].

13. Bradford [1978] has demonstrated this point for the case in which r = C =

444

QUARTERLY JOURNAL OF ECONOMICS

class holding equity and the other debt. For this to occur, firms must
be indifferent between debt and equity, equity-holders must prefer
equity, and debt-holders must prefer debt. Without any loss of generality, we take class 1 to be the equity-holders.

In equilibrium the rate of return received by equity investors

must equal their rate of time preference. Combining (5) and (8), we
obtain
p1 =

(30)

(10)+ (1 c1),

where p1 is the rate of time preference of class 1 members and V is

the capital gain they receive. Similarly, the net return to holders of
debt must equal their rate of time preference:
p2 = (1

(31)

The potential return to equity holders from holding debt is

p1 = (1

(32)

and the potential return to debt holders from holding equity is

p2 = (1

(33)

09)+ (1 C2).

The conditions for an equilibrium with debt and equity are

(34.1)

p1 p'

(34.2)

p2 p2
i(1 r)(1

(34.3)

c1)

= p1;

that is, debt-holders prefer debt, equity-holders prefer equity, and

firms are indifferent between debt and equity. Combining (30)(34)
we obtain
(and substituting 0 for and 09 for
(35)

((1_cl)/(1_0l))(1_a)+a
1 c1

((1 c2)/(1

02))(1

a) + a

(1

)> 1101

c1

where a is the fraction of gross equity gains received through dividends:

(36)

V+D

It is clear that (35) can hold only if (1 c1)/(1 0) (1 c9)/(1

(19). (Since we may rearrange indices, this is always possible.) What
this tells us is that if equity and debt coexist in equilibrium, equity

WEALTH MAXIMIZATION AND THE COST OF CAPITAL

445

must be held by those who receive a relatively greater tax advantage

in the statutory treatment of capital gains. This result is similar to
that of the simple Ricardian trade model which dictates that countries
will specialize in the production of the good for which they have a
comparative advantage.
If (1 c1)/(1 ()) = (1 c2)/(1 (12), then condition (35) will
bemetifandonlyif(1 T) = (1 0)/(1 c1) = (102)1(1 c9),in
which case both classes will be indifferent between debt and equity.
Thus, the two-class example collapses to the one-class example, and
no segmentation will occur. Suppose, at the other extreme, that class
2 consists of tax-exempt investors, such as nonprofit institutions.
Then (35) becomes
(37)

(1

a) + a (1_

(1

T)

(li)
(11c1
01).
Unless a is near unity,14 the first inequality in (37) will be satisfied,
and the condition reduces to that necessary for the existence of equity
in a one-class economy. Thus, in a two-class world, with one class
tax-exempt, and the other facing a marginal tax rate on regular capital
income between 0.51 and 0.70, debt and equity would coexist, with
the former group holding debt and the latter holding equity.

VIII. CONCLUSION

This paper has reviewed the investment and financial behavior

of corporations seeking to maximize the wealth of their shareholders.
We have focused on the impact of personal income and capital gains
taxes, finding that, in the presence of differential taxation of dividends

and capital gains, wealth maximization does not imply maximization of firm market value and the source of equity financing is not
irrelevant.
The appropriate cost of capital in the presence of personal taxes
does not depend directly on either the dividend payout rate or the tax
on dividends. Equity shares have a market value lower than the difference between the reproduction cost of a firm's assets and the
market value of its debt obligations. Because of this capitalization,
it need not be true that an economy without risk or uncertainty would
have no equity financing.
HARVARD UNIVERSITY AND NATIONAL BUREAU OF EcoNoMic RESEARCH

14. For c = 0.1 .0, 0

unless a 0.68.

0.70. and r = 0.46, the first inequality in (37) must hold

QUARTERLY JOURNAL OF ECONOMICS

446

REFERENCES
Auerbach, A. J., "Share Valuation and Corporate Equity Policy," Journal of Public
Economics, 1979.
Bailey, M. J., "Capital Gains and Income Taxation," in A. C. Harberger and M. J.
Bailey, eds., The Taxation of Income from Capital (Washington, D.C.: Brookings
Institution, 1969).
Bierman, H., and R. West, "The Acquisition of Common Stock by the Corporate Issuer," Journal of Finance, XXI (Dec. 1966), 68796.
Bradford, D., "The Incidence and Allocation Effects of a Tax on Corporate Distributions," mimeographed, 1978.
Farrar, D. F., and L. L. Selwyn, "Taxes, Corporate Financial Policy, and Return to
Investors," National Tax Journal, XX (Dec. 1967), 44454.
Feldatein, M. S., J. Green, and E. Sheshinski, "Corporate Financial Policy and Taxation
in a Growing Economy," this Journal, XCIII (Aug. 1979), 41132.
King, M. A., "Taxation and the Cost of Capital," Review of Economic Studies, XLI
(Jan. 1974), 2135.
Miller, M. H., "Debt and Taxes," Journal of Finance, XXXII (May 1977), 26175.

and F. Modigliani, "Dividend Policy, Growth, and the Valuation of Shares,"

Journal of Business, XXXIV (Oct. 1961), 41 133.

Modigliani, F., and M. H. Miller, "The Cost of Capital, Corporation Finance, and the

Theory of Investment," American Economic Review, XLVIII (June 1958),

26197.

Pye, G., "Preferential Tax Treatment of Capital Gains, Optimal Dividend Policy, and
Capital Budgeting," this Journal, LXXXVI (May 1972), 22642.
Stapleton, R. C., "Taxes, the Cost of Capital, and the Theory of Investment," Econobsic
Journal. LXXXII (Dec. 1972), 127392.
Stiglitz, J. E., "Taxation, Corporate Financial Policy, and the Cost of Capital," Journal
of Public Economics, II (Feb. 1973), 134.
"The Corporation Tax," Journal of Public Economics, V (April 1976), 303
11.

Tobin, J., "A General Equilibrium Approach to Monetary Theory," Journal of Money,
Credit and Banking, I (Feb. 1969), 1529.
U.S. Federal Reserve System, Board of Governors, Flow of Funds Accounts.
, Federal Reserve Bulletin, LXIII (Dec. 1977).