True Man 1988

Accounting Research Center, Booth School of Business, University of Chicago
An Explanation for Accounting Income Smoothing

Author(s): Brett Trueman and Sheridan Titman
Source: Journal of Accounting Research, Vol. 26, Studies on Management's Ability and
Incentives to Affect the Timing and Magnitude of Accounting Accruals (1988), pp. 127-139
Published by: Wiley on behalf of Accounting Research Center, Booth School of Business, University of
Chicago
Stable URL: http://www.jstor.org/stable/2491184 .
Accessed: 21/08/2013 16:17
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
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.
Wiley and Accounting Research Center, Booth School of Business, University of Chicago are collaborating
with JSTOR to digitize, preserve and extend access to Journal of Accounting Research.
http://www.jstor.org
This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

All use subject to JSTOR Terms and Conditions
Journal of Accounting Research
Vol. 26 Supplement 1988
Printedin U.S.A.
An Explanation for Accounting

Income Smoothing
BRETT TRUEMAN* AND SHERIDAN TITMANt
1. Introduction
It is widely believed that corporatemanagersoften engage in income
smoothing,taking actions to dampenfluctuationsin their firms'publicly
reportednet income. One reason given for this is that managersthink
that investors pay more for a firm with a smoother income stream. (See
Ronen and Sadan [1981].) While there have been many empiricalstudies
attempting to document whether or not income smoothing actually
occurs,there has been little explorationof why a managermight ration-
ally want to smooth his firm's income. Explainingwhy smoothing might
be observed, and how it can result in an increased stock price, is the
focus of this paper.
In two recent studies, Lambert [1984] and Dye [1988] demonstrate,in
agency settings, that a risk-aversemanagerwho is precludedfrom bor-
rowing and lending in the capital markets has an incentive to smooth
his firm's reported income. In this paper, in contrast, it is shown that
within a market setting an incentive exists for a manager to smooth
income that is independentof either risk aversion or restrictedaccess to
capital markets. The market setting is also useful in that it allows for
the analysis of the effect of income smoothing on stock prices.
It is shown here that if a managercan choose which of two periods to
recognize certain income, he may prefer the choice that is expected to
* University of California, Berkeley; tUniversity of California, Los Angeles. We would
like to thank Ron Dye, David Hirshleifer, Jack Hughes, Pat Hughes, Paul Newman, Eric
Rasmusen, Chris Tang, Siew Hong Teoh, participants at the 1988 Journal of Accounting
Research Conference, and especially Yoon Suh for their helpful comments. All remaining
errors are, of course, our own. Sheridan Titman gratefully acknowledges financial support
from the Batterymarch Fellowship Program.
127
Copyright (C, Institute of Professional Accounting 1989

128 MANAGEMENT S ABILITY TO AFFECT ACCRUALS: 1988
result in a smoother income stream. (This action is often called temporal

income smoothing). By smoothing income the manager may be able to
reduce the estimate of various claimants of the firm about the volatility
of its underlying earnings process, which, in turn, lowers their assessment
of the probability of bankruptcy. This is valuable for the firm's stock-
holders since it decreases the firm's cost of borrowing and favorably
affects the terms of trade between the firm and its customers, workers,
and suppliers, as described in Titman [1984].
The result that income smoothing can affect the perceived probability
of bankruptcy follows from two related assumptions. The first is that the
amount of discretion any firm has to shift income from one period to
another depends on the nature of its operations and so differs across
firms. This assumption is formalized here by positing that some firms
can shift income across periods while others cannot. The second assump-
tion is that claim holders cannot fully observe each firm's operations and
so are not certain of the flexibility it has to shift income. Hence, when
claim holders observe a smooth income stream, they are uncertain about
whether it came from a firm with high volatility that smoothed its
reported income stream, or if it came from a firm with low volatility that
did not have the flexibility to smooth income. The implication of this is
that firms with the flexibility to shift income across periods can, in
general, project lower volatility than can firms without this flexibility.
If smoothing is costly, a manager may choose not to smooth the firm's
reported income. Such costs could arise, for example, if early recognition
of income for financial reporting purposes also requires early recognition
of the tax liability, resulting in a higher present value of the firm's tax
bill. However, if the smoothing costs are sufficiently small, the manager
still has an incentive to smooth the firm's reported income.
Given differential costs and benefits to smoothing, predictions are
made about the types of firms that are more likely to smooth income.
These predictions can potentially be empirically tested. Such tests would
be based on the observation that smoothing reduces the information
content of earnings announcements; the market reaction to these an-
nouncements is, therefore, expected to differ systematically across firms
based on the market's perception of the likelihood that income is being
smoothed.
In section 2 is a description of the general economic setting. This is
followed in section 3 by an analysis of the manager's motivation to
smooth income. Section 4 presents a discussion of the results of the
analysis. Possible extensions of the analysis are presented in section 5.
Finally, in section 6 is a summary and conclusions.
2. Economic Setting
Consider a firm whose economic earnings each period, before deduction
of any interest payments or bankruptcy costs in the case of insolvency,

EXPLANATION FOR INCOME SMOOTHING 129
are generated according to the following process:

it = , + et (1)
where , is known to all agents. The distribution of et is normal with
mean zero and is stationary over time. The variance of this distribution
is initially unknown to all but the firm's manager. It takes on either a
low value, a,2, or a high value, Ub2. The prior probability that the variance
equals 2b (U,2) is given by Pb (Pa)-
The assumption that the mean of the earnings distribution is known
is made in order to simplify the analysis and to focus on the effect that
an uncertain variance has on the manager's behavior. This is in contrast
to most signaling models which assume that the variance is known in
order to focus on the effect that an unknown mean has on managerial
decisions. Implications for the manager's behavior if the mean and
variance are both unknown are discussed in section 5.
It is assumed that only the manager can observe the firm's economic
earnings for any period.1 Others, however, may be able to infer something
about them from the firm's reported income, as disclosed by the manager.
In preparing this report from his knowledge of the firm's economic
earnings, the manager may be able to shift, between periods one and two,
the time at which some of the economic earnings are recognized as
income on the income statement.2 Such shifting could be accomplished,
for example, through the timing of the recognition of pension costs,
adjustments to the estimates of assets' useful lives, and adjustments to
the write-downs for bad debt. It is assumed that if the firm's manager
does not shift income across periods, then the firm's reported income in
any period equals that period's economic earnings.
The flexibility to shift income across periods is assumed to be available
only in some firms., Furthermore, only the manager knows whether or
not he has this flexibility. Others believe that with probability ps the
manager can shift income across periods and that with complementary
probability 1 - p, he cannot.4
In this setting, income smoothing is defined as the manager shifting
the recognition of some of the firm's income, if there is that flexibility
within the firm, from the second period to the first (the first period to
the second), whenever the first period's economic earnings are less than
(greater than) the expected per period economic earnings, At.For simplic-
' This implies that in examining the firm's records the external auditor cannot perfectly
observe the period's economic earnings.
2
More generally, the ability to shift income could extend beyond the first two periods
without changing the model's implications.
Alternatively, it can be assumed that the ability to shift income is available to all firms
but that the extent to which it can be done differs across firms.
ZIThe manager is assumed to be unable to signal to outsiders his ability to shift income
across periods.

130 B. TRUEMAN AND S. TITMAN
ity, it is assumed that even if the manager is able to smooth income, he

can shift only a fixed proportion, g, of the difference, At - x, from the
second period to the first, where 0 < g < L' (Note that if Auis less than
x1 this means a shift of a positive amount of income from the first period
to the second.) The magnitude of g is assumed to be known to everyone.
Furthermore, it is assumed initially that the timing of income recognition
does not affect the present value of the firm's cash flows. This restriction
is relaxed in the later analysis.
In this setting, if the manager engages in income smoothing, first-
period reported income, xis, is given by:
xiS = (1-g)x1 + gA (2)
and second-period reported income, X2S, is given by:

X2S= X2 - g(A - X1). (3)
3. The Manager's Decision to Smooth Income

The purpose of this section is to show that in this setting a manager
with the goal of maximizing the present value of the cash flows accruing
to the firm's risk-neutral shareholders may have an incentive to smooth
the firm's reported income. This incentive arises if the firm's future
claimants are adversely affected by the firm's volatility. While these
claimants could include future customers, workers, and suppliers, the
focus here is on the firm's future debt holders. Specifically, it is assumed
that the firm plans to sell debt at the end of the second period with a
maturity value of D dollars.6 So as not to introduce changes in the firm's
operating results as a consequence of the sale, it is assumed that the debt
is used to repurchase some of the firm's outstanding equity. While in
this model the decision to change the mix of debt and equity is made
exogenously, in a more general setting it could result, for example, from
a change in the present value of the tax advantages of additional debt
relative to the disadvantages due to the increased likelihood of bank-
ruptcy.
The price that the firm receives for the debt, or alternatively the
required interest rate, is a function of the risk-neutral debt holders'
beliefs about variance of the firm's economic earnings, since the proba-
bility that the firm will go bankrupt is directly related to this variance.
The greater the probability that this variance is high, the greater the
probability of bankruptcy and the lower the selling price of the debt.
5 The upper limit on g of one ensures that if economic earnings for period one are less
(greater) than Az,reported income for period one is also greater (less) than Az.A shift in the
time of income recognition is, therefore, guaranteed to imply that first-period reported
income is closer to the mean level of economic earnings.
('Having the debt issued at the end of the second period more fully highlights the effects
of income smoothing. However, results similar to those presented below would be obtained
if the debt was sold in the first period or in any future period.

Denote by B, (Bb) the price that debt holders would pay if they knew for
certain that the variance of economic earnings is low (high), where Ba >
Bb. Then, with their imperfect information about the variance, they set
the price at:
B(pa') = papBa + PbBb (4)
where Pa' (Pb') is the debt holders' posterior probability that the variance
is low (high) given their observation of the firm's reported income in the
first two periods. (Explicit expressions for these probabilities are derived
below.)
In order to increase the selling price of the debt, and thereby increase
the present value of the cash flows accruing to the shareholders, the
manager has an incentive to raise debt holders' posterior probability that
the variance of economic earnings is low. To see how income smoothing
accomplishes this, assume that debt holders believe that the manager
smooths income whenever he can. Then, if the manager reports an income
pattern of (xir, x2/) over the first two periods, debt holders know that
either the manager was unable to smooth income and that the realized
economic earnings were (Xir, X2r) or that the manager did shift income
across periods and that the economic earnings were (x1', x2') where:
Xlr (gA5)
1-g
and:
= X2 + g(u -
X2 X1'). (6)
(Relations (5) and (6) are derived by inverting (2) and (3), respectively.)
Since the debt holders cannot observe whether the manager was able to
smooth income, they are not certain of the actual economic earnings
pattern. (From observation of the reported income pattern, they do
update their prior probability, pa, that the manager smoothed income.
But there still remains some uncertainty over whether or not the manager
engaged in income smoothing.) Given the reported income pattern, and
using Bayes' rule, debt holders assess the probability that the variance
of the firm's economic earnings is 0-a2, denoted by Pa' (Xr, X2r), as:
Pa~(Xr, X~ = [(1--
(1 Ps)fa(Xi,
psjfa(Xir, X2 )Pa ++ Psfa
Xr)P pfaXr,(xir,
f X2 )Pa
Xjr)Pa
+ (1 - pS)fb(Xl, X2 )Pb + pSfb (X1, X2 )Pb]
where fi(x1r, X2) [fi'(Xlr X2)] is the value of the density for reported
earnings, evaluated at (Xir, X2), if the manager cannot (can) smooth
income and if the variance of the economic earnings is ,i2 i = a, b.
In order to show that it is optimal for a manager to smooth income if
he has that flexibility, note, first, that when the manager makes his
smoothing decision at the end of the first period, he does not yet know
the value of second-period economic earnings, x2. He is able to calculate

only the expected value of debt holders' posterior probability,

Pa/(Xir, over all x2. Let Ei[pa/(Xis,
X2r), x25)] (Ei[Pa'(Xi, X2)]) denote this
expectation if he does (does not) smooth income. (The subscript i denotes
that the manager is in a firm whose underlying economic earnings have
variance a,2.) That Ei[Pa'(XIs, xi2)] is greater than Ei[Pa'((x, x2)] is shown
in the following lemma.
LEMMA. If the firm's economic earnings are normally distributed, then:
Ei[Pa'(Xls, xi2)] > Ei[Pa'(XI, X2)].
Proof. To begin, note that fi(X1r, X2r) is equal to hi(xlr)hi(X2r) where

hi( Xr) [hi(X2r)] is the value of a normal density with mean Atand variance
Ui2, evaluated at Xir (X2r). Further, fi(X1r, X2r) is equal to hi(Xr)hi'(X2 rIXr)
where hi' (Xir) is the value of a normal density with mean Atand variance
(1l- g)2 ai2, evaluated at Xlr, and hi' (x2rI Xr) is the value of a normal
denritywith mean u- g(U - (Xir-gu)/(l - g)) and variance a,2, evaluated
at X2r. (These properties of fi' (Xir, X2r) are derived from relations (2) and
(3).) Examination of hi' (Xlr)hi'(x2rI Xir) reveals that it can be rewritten as
hi(Xlr)hi(X2r)/(l - g).
Consider, next, the expressions y(x1, x2) and z(x1, x2) where:
x2) =~ ha(Xis)ha(X2s)Pa + hb(xls)hb(x2s)pb

ha(xIs)ha(X2s)Pa
AXI,
y(x1, X2) (8)
and:
ha(xl)ha(X2)Pa 9
Z(XI,X2)
z(x1 x2) = ha(Xi)ha(X2)Pa + hb(Xl)hb(X2)Pb(
Given the preceding discussion, in order to show that Ei [Pa'(Xls, x2s)] >
Ei[pa'(XI, 2)], it is sufficient to show that Ei[y(x,, x2)] > Ei[z(x1, X2)]
Ei[y(x1, x2)] is given by:
f [y(x1, -e2)]hi( -e2)de2
+ [y(x1, A + e2)]hi (A + e2)de2. (10)
Further, since the normal distribution is symmetric, z(x1, A - e2) = Z(XI,

,u + e2). This implies that to prove the lemma it is sufficient to show
that:
y(x, Au+ e2) + y(x1, u - e2) > 2z(x1, AU+ e2) V e2 > 0. (11)
With normally distributed random variables, it is straightforward to
show (after some algebraic manipulation) that (11) is satisfied for all e2
> 0. Q.E.D.
The implication of this lemma is that if the debt holders believe that
the manager smooths the firm's reported income, the manager has an
incentive to do so. By reporting (X1I, X2s) rather than (x1, x2), the manager

increases, on average, debt holders' assessment of the probability that

the firm's underlying economic earnings have low variance.
Along similar lines, it can be shown that even if the debt holders
believe that the manager does not smooth the firm's reported income,
the manager still has an incentive to do so. By smoothing income, he
again increases, on average, debt holders' assessment of the probability
that the firm's economic earnings have low variance. An equilibrium,
therefore, cannot exist in which the manager does not smooth income.
The main result of this section follows directly from the lemma.
PROPOSITION 1. If income smoothing is costless, then in equilibrium
it is optimal for the manager to smooth reported income.
To understand the intuition behind this proposition, consider first a
simple case where xl < u and where the manager knows that the second-
period economic earnings will be greater than A. If the manager shifts
income from period two to period one, the debt holders observe the
smoother reported income stream (x],`x2s). Since there is some possibility
that this income pattern resulted from a firm whose manager does not
have the flexibility to smooth income, debt holders assess a positive
probability that the underlying economic earnings stream was actually
(xiS,x2S). This, in turn, increases their assessment of the probability that
the variance of the firm's economic earnings is low and increases the
price they pay for the debt.
However, the manager does not know the value of x2 when he makes
his income-smoothing decision at the end of the first period. If x1 is less
than ,u, then smoothing reported income toward the mean results in a
smoother reported income stream over the two periods when the reali-
zation of x2 is greater than At (and also in some cases when x2 is less than
A); but it results in a less smooth income pattern if the realization of x2
is sufficiently less than At (and, in particular, if it is less than xl). That
is, smoothing income in the first period might actually cause debt holders
to decrease, rather than increase, their assessment of the probability that
the variance of economic earnings is low. On average, though, with x2
centered around ju,a shift of income from the second period to the first
results in a smoother reported income stream over the two periods.
(Similar reasoning applies if xi is greater than At; smoothing first-period
income toward the mean, on average, results in a smoother reported
income pattern.) Even if he does not know the value of x2, the manager
still expects the firm's shareholders to benefit if he smooths the firm's
reported income at the end of the first period.
4. Discussion
The preceding analysis provided conditions under which a manager
desires to smooth first-period reported income toward its mean value.
Such action, on average, increases debt holders' assessment of the prob-
ability that the variance of the firm's economic earnings is low and

consequently increases the selling price of the debt. This increases the
present value of the portion of the firm's cash flows accruing to its
shareholders.
The shareholders of firms that are able to smooth income gain at the
expense of shareholders of those firms who do not have this flexibility.
To understand this, recall that debt holders do not know whether or not
the firm's manager has the ability to smooth income. Shareholders of a
firm that has smoothed income gain because debt holders, in setting the
price of the debt, assume that there is a possibility that the reported
income was in fact not smoothed (and that the variance of the economic
earnings was the same as that of the reported income). This causes them,
on average, to underestimate the true variance of the firm's economic
earnings, resulting in a higher selling price for the debt. Conversely,
shareholders of a firm that is unable to smooth income lose because debt
holders make their calculations assuming that there is a possibility that
the reported income was actually smoothed (and that the actual economic
earnings were more volatile than the reported income). On average, they
overestimate the variance of the firm's economic earnings, resulting in a
lower selling price for the debt.
Since the gains to the shareholders of firms that can smooth income
are exactly offset by the losses to the shareholders of firms that cannot,
the following proposition holds.
PROPOSITION 2. At the beginning of the first period, shareholders, not
knowing whether or not the manager has the flexibility to smooth income,
are indifferent between an accounting system in which income smoothing
is allowed and one in which it is precluded, as long as income smoothing
is costless.
Proof. In an accounting system in which income smoothing is pre-
cluded, shareholders in a firm with a variance of economic earnings of
a.2 know that, on average, debt holders set the price of the debt at:
Ei[B(p.')] = Ei[Pat'(Xi1 i2)]Ba + (1 - Ei[Pa'(Xii 52)])Bb (12)
where the expectation is now taken over all x1 and x2. (With income
smoothing precluded, p, must be set equal to zero in the expression for
Pa'(X1, x2).) By definition E[pa'(il, x2)] is just equal to the prior proba-
bility, Pa.
In an accounting system in which income smoothing is allowed, these
shareholders know that debt holders, on average, set the price of debt at:
Ei[B(pa')] -Ei[Pa'(xi% i2S)]Ba + (1 - Ei[Pa'(xi15 i2S)])Bb (13)
if the manager is able to smooth income and:
EJB(pa')] = Ei[Pa'(Xi1 i2)]Ba + (1 - Ei[Pa'(Xi1 i2)])Bb (14)
if the manager is unable to smooth income. But since shareholders do

not know whether or not the manager is able to smooth income, their

overall expectation for the price of the debt is:
{pSEl[paa(i1S, x25)] + (1 - ps)Ej[pa' (xl, X2)] }Ba

+ [1 - lpsEi[paI(xiS i2s)] + (1 -ps)Ei[paj'(x, i2)] R]Bb. (15)
But, again, by definitionpE [Pa(i x95)] + (1 -ps)Ei[Pa'(xi X2)] is equal
to Pa
Therefore, at the beginning of the first period, shareholders are indif-
ferent between an accounting system in which income smoothing is
allowed and one in which it is precluded, as long as smoothing is costless.
The expected selling price of the debt is the same in either case. Q.E.D.
The previous results have focused on the effect income smoothing has
on the cash flows accruing to the firm's shareholders. Consider, now, the
effect of income smoothing on the value of the firm as a whole. Without
costs to bankruptcy, it is clear that, given the firm's economic earnings,
the total value of the firm after the debt is issued is unaffected by income
smoothing; shareholders gain only at the expense of debt holders. How-
ever, as the following proposition demonstrates, if bankruptcy is costly,
income smoothing can increase the firm's total market value by decreas-
ing investors' assessment of the probability that the firm goes bankrupt.
PROPOSITION 3. If bankruptcy is costly, the market value of the firm
at the end of period two is, on average, higher if the manager smooths
income than if he does not.
Proof. From the investors' perspective, at the end of the second period,
the value of the firm is equal to the present value of the firm's cash flows,
not including bankruptcy costs, less the present value of expected bank-
ruptcy costs. The investors believe that the firm is less likely to go
bankrupt and incur these bankruptcy costs the greater is Pa'. Since
income smoothing, on average, increases Pa't the market value of the
firm at the end of the second period, on average, increases as a conse-
quence. Q.E.D.
Although the market value of the firm is affected by the manager's
action to smooth income, the true value of the firm (calculated with the
manager's superior information about the economic earnings) is unaf-
fected if bankruptcy imposes costs only on the firm's debt and equity
holders. This follows because the firm's total economic earnings and
total cash flows are, in this case, the same whether or not the manager
smooths income. However, if bankruptcy is costly to other claim holders
of the firm, such as its customers, workers, and suppliers, then income
smoothing can affect the firm's total cash flows and, therefore, its true
value. This is because income smoothing, by affecting these claim holders'
perception of the probability of bankruptcy, affects the firm's production
costs and output prices.
The analysis to this point has assumed that income smoothing is
costless. Consider now the case where there are costs to smoothing. Such

costs might arise if smoothing implies an earlier recognition of income

and if this early recognition is accompanied by an acceleration of the
payment of income taxes. Costs could also result if management must
expend more effort in convincing the firm's auditor that the actions
taken to smooth income are legitimate.
To examine the implications of costly income smoothing, assume that
if the manager does smooth income, the firm incurs a cost of K. Given
the manager's goal of maximizing the present value of cash flows to
shareholders, it is apparent that for sufficiently high costs of smoothing,
the manager prefers not to engage in income smoothing for some or all
realizations of x1;the expected gain from increasing debt holders' assess-
ment of the probability that the variance of the firm's economic earnings
is low is outweighed by the costs incurred. Income smoothing, therefore,
remains optimal for all values of x1 only if the gain to smoothing, after
deduction of the smoothing cost, remains positive. For a firm whose
economic earnings have a variance of a,2, this gain, denoted by G, is given
by:7
G = Ej[p.'(x1S, i2S)]Ba + (1 - Ei[pa'(xiS, x2)])Bb
- Ei[pa'(X1, 2)]Ba - (1 - E[pa'(Xl, x2)])Bb - K. (16)
The first two terms in (16) represent the expected selling price of the
debt when there is smoothing (and a higher price is paid, on average),
while the second two terms represent the expected selling price of the
debt when there is no smoothing. Equation (16) can be simplified to
yield:
G = (Ei[pA'(x1S, xii)] - Ei[pa'(Xi, 2)])(Ba - Bb) - K (17)
The following proposition results from this relation.

PROPOSITION 4. In order for costly income smoothing to be optimal,
the prior probability that the firm's economic earnings are of low vari-
ance, Pa, must be sufficiently different from both zero and one.
Proof. Examination of Ej[pa'(Xis, XS)] and Ei[pa'(x1, x2)] reveals that
at either Pa = 0 or Pa = 1, these two expressions are the same. Therefore,
by smoothing income the manager does not affect debt holders' posterior
probability that the variance of the firm's economic earnings is low.
Then, if smoothing is costly, the manager prefers not to smooth the
firm's reported income. By continuity, if Pa is close enough to either zero
or one, the expected gain from smoothing is again outweighed by the
cost. Q.E.D.
If the prior probability, Pas is close to either zero or one, debt holders
have little uncertainty over the variance of the firm's economic earnings.
This implies that the prior probability is not substantially revised by
'Consistent with the modeling here, (16) does not explicitly consider the effect of income
smoothing on any claim holders of the firm other than its debt holders. Including the
impact on other claim holders would have the effect of increasing the gain from smoothing.

their observation of the firm's reported income, so that the benefit of

income smoothing may not justify its cost. For such firms, smoothing is
less likely to be observed.
Given (17) the following also can be shown.
PROPOSITION 5. The smaller is g, the less profitable is income smooth-
ing.
Proof. Differentiationof Ei[pa' (Xis, X28)] and Ei[pa' (x1, X2)] with respect
to g reveals that the former decreases as g decreases while the latter
increases. Therefore, the gain from smoothing, (17), decreases as g
decreases. Q.E.D.
This proposition states that the less income the manager is able to
shift between periods, the less he gains by income smoothing. Another
way to interpret this result is that smoothing is less profitable for firms
in which debt holders can to a large extent infer the level of economic
earnings each period. Income smoothing is less likely to be observed for
such firms.
From equation (17) it is also apparent that in order for costly income
smoothing to be optimal, the selling price of the debt in a firm whose
economic earnings have high variance must be sufficiently less than the
selling price in a firm with a low variance of economic earnings. Other-
wise, with Ba - Bb small, the cost of smoothing outweighs the benefit.
This means that firms with a low probability of going bankrupt, whether
the firms' economic earnings have high or low variance, would be less
likely to smooth income. Conversely, those firms with only a small
probability of going bankrupt when the variance of economic earnings is
low but a much higher probability when the variance is high would be
more likely to smooth income. Similarly, firms which impose high costs
on claim holders if they go bankrupt would be more likely to smooth
income. Since both the probability and costs of bankruptcy differ less
among firms within industry groupings than across industries, the tend-
ency to smooth income should, therefore, be related to industry classifi-
cations.
Since income smoothing is assumed to be unobservable, these conjec-
tures about the types of firms more likely to smooth income cannot be
directly tested. However, indirect tests may be possible, based on the
observation that income smoothing reduces the information content of
earnings announcements. The stock price reaction to any earnings an-
nouncement should, then, vary systematically with the above-mentioned
factors which relate to the probability that the firm smoothed its income.
5. Possible Extensions
The analysis here has shown that the sale of debt provides an incentive
for a firm to smooth its reported income. However, it is easy to see that
this result generalizes to the case where the firm does not sell any
securities. As long as the firm's customers, workers, and suppliers impose

higher costs on the firm if they perceive the variance of its economic
earnings to be higher, the firm's manager is still motivated to smooth
income.
In deriving the results of this model, it was assumed that the variance
of the firm's economic earnings is unknown to investors but that their
mean is known. The model can also be applied to the case where the
variance is known but the mean is unknown. In this case, it is easy to
show that the manager would have an incentive to increase reported
income just prior to the sale of securities, regardless of whether or not it
results in a smoother income stream. This would increase investors'
perceptions of the mean of the firm's future economic earnings and raise
the price at which the securities could be sold.8 More generally, if both
the mean and the variance of the firm's future economic earnings are
unknown to investors, there would be incentives both to smooth reported
income (so as to project a lower variance) and to increase reported income
prior to the sale of securities (so as to project a higher mean). However,
without adding more structure to this setting, it is difficult to predict
how these incentives would interact to determine the manager's optimal
method of managing earnings.9
The problem would become more complex if the manager were allowed
to obtain private information abut the firm's future economic earnings
before making his smoothing decision. In this case, whether the manager
smooths income depends on his private information. Because of this, the
income reported in the current period may also provide a signal to
investors about the firm's future economic earnings, affecting the price
at which the firm can sell its securities.
It has also been assumed that in making his smoothing decision the
manager's objective is to maximize the present value of the cash flows
accruing to shareholders, regardless of the effect on the firm's current
stock price. A further extension of this analysis would be to examine how
the manager's decision changes if he also weighs the effect that income
smoothing has on this price. This is a meaningful issue if the firm's
initial capital structure includes debt since a decrease in perceived risk,
brought about by a smoother income stream, lowers the market value of
equity at the end of the first period. Depending on the magnitude of this
decline, the manager could be deterred from smoothing income.
6. Summary and Conclusions

This paper has provided a reason that a corporate manager may
rationally want to smooth reported income-namely, to lower claim
8 It is necessary that the action to increase reported income take place just prior to the
sale so that the required reversing transaction, decreasing reported income, occurs after the
sale. Otherwise, investors' expectation for the firm's economic earnings would remain
unchanged.
9 Similar difficulties in the context of signaling models explain why virtually all of the
models allow for only one source of uncertainty.

holders' perception of the variance of the firm's underlying economic

earnings. In turn, it was shown that such action could have a positive
effect on the firm's market value. This paper is related to others which
have attempted to explain seemingly irrational behavior on the part of
managers (see, for example, Trueman [1986] and Kanodia, Bushman,
and Dickhaut [forthcoming]) and may be useful to shareholders and
regulators in deciding the extent to which they should monitor and
control managers' actions.
REFERENCES
DYE, R. "Earnings Management in an Overlapping Generations Model." Journal of Ac-
counting Research (Autumn 1988): 195-235.
KANODIA, C., R. BUSHMAN, AND J. DICKHAUT. "Escalation Errors and the Sunk Cost
Effect: An Explanation Based on Reputation and Information Asymmetries." Journal of
Accounting Research (forthcoming).
LAMBERT, R. "Income Smoothing as Rational Equilibrium Behavior." The Accounting
Review (October 1984): 604-18.
RONEN, J., AND S. SADAN. Smoothing Income Numbers: Objectives,Means, and Implications.
Reading, Mass.: Addison-Wesley, 1981.
TITMAN, S. "The Effect of Capital Structure on a Firm's Liquidation Decision." Journal of
Financial Economics (March 1984): 137-51.
TRUEMAN, B. "Why Do Managers Voluntarily Release Earnings Forecasts?" Journal of
Accounting and Economics (March 1986): 53-71.


True Man 1988

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

True Man 1988

Uploaded by

Copyright:

Available Formats

Accounting Research Center, Booth School of Business, University of Chicago

An Explanation for Accounting Income Smoothing

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

An Explanation for Accounting

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

result in a smoother income stream. (This action is often called temporal

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

are generated according to the following process:

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

ity, it is assumed that even if the manager is able to smooth income, he

and second-period reported income, X2S, is given by:

3. The Manager's Decision to Smooth Income

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

only the expected value of debt holders' posterior probability,

Ei[Pa'(Xls, xi2)] > Ei[Pa'(XI, X2)].

Proof. To begin, note that fi(X1r, X2r) is equal to hi(xlr)hi(X2r) where

x2) =~ ha(Xis)ha(X2s)Pa + hb(xls)hb(x2s)pb

f [y(x1, -e2)]hi( -e2)de2

+ [y(x1, A + e2)]hi (A + e2)de2. (10)

Further, since the normal distribution is symmetric, z(x1, A - e2) = Z(XI,

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

increases, on average, debt holders' assessment of the probability that

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

if the manager is unable to smooth income. But since shareholders do

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

{pSEl[paa(i1S, x25)] + (1 - ps)Ej[pa' (xl, X2)] }Ba

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

costs might arise if smoothing implies an earlier recognition of income

- Ei[pa'(X1, 2)]Ba - (1 - E[pa'(Xl, x2)])Bb - K. (16)

The following proposition results from this relation.

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

their observation of the firm's reported income, so that the benefit of

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

6. Summary and Conclusions

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

holders' perception of the variance of the firm's underlying economic

This content downloaded from 205.133.226.104 on Wed, 21 Aug 2013 16:17:45 PM

You might also like