Académique Documents
Professionnel Documents
Culture Documents
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.
American Accounting Association is collaborating with JSTOR to digitize, preserve and extend access to The
Accounting Review.
http://www.jstor.org
(e.g., current cash equivalents) are feasi- lobbies on accounting standards based
ble. on its own self-interest. For simplicity,
Watts [1974] and [1977] has started to (since this is an early attempt to provide
develop such a theory. This paper ex- a positive theory) it could be argued that
pands on this initial work by focusing on we should assume that management's
the costs and benefits generated by ac- self-interest on accounting standards is
counting standards which accrue to congruent with that of the shareholders.
managements, thereby contributing to After all, that assumption has provided
our understanding of the incentives of hypotheses consistent with the evidence
management to oppose or support vari- in finance (e.g., the risk/return relation-
ous standards. Management, we believe, ship of the various capital asset pricing
plays a central role in the determination models). However, one function of finan-
of standards. Moonitz supports this cial reporting is to constrain manage-
view: ment to act in the shareholders' interest.
(For example, see Benston [1975], Watts
Management is central to any discussion of
financial reporting, whether at the statutory [1974], and Jensen and Meckling
or regulatory level, or at the level of offi- [1976a].) Consequently, assuming con-
cial pronouncements of accounting bodies. gruence of management and shareholder
[Moonitz, 1974, p. 64] interests without further investigation
Hence, it seems appropriate that a pre- may cause us to omit from our lobbying
condition of a positive theory of stan- model important predictive variables. To
dard-setting is understanding manage- reduce this possibility, we will examine
ment's incentives. next the effects of accounting standards
The next section introduces those fac- on management's self-interest without
tors (e.g., tax, regulatory, political con- the congruence assumption. The purpose
siderations) which economic theory leads of the examination is to identify factors
us to believe are the underlying determin- which are likely to be important predic-
ants affecting managements' welfare and, tors of lobbying behavior so that we can
thereby, their decision to consume re- include them in our formal model.
sources trying to affect the standard- The assumption that management se-
setting process. Next, a model is pre- lects accounting procedures to maximize
sented incorporating these factors. The its own utility is used by Gordon [1964,
predictions of this model are then tested p. 261] in an early attempt to derive a
using the positions taken by corporations positive theory of accounting. There have
regarding the FASB's Discussion Mem- been several attempts to test empirically
orandum on General Price Level Adjust- Gordon's model, or variants of it, which
ments (GPLA). The last section contains we call the "smoothing" literature.3
the conclusions of the study. Problems in the specification of the em-
FACTORS INFLUENCING MANAGEMENT 2 Many economic models assume a rather limited
version of economic man. In particular, they assume that
ATTITUDES TOWARDS FINANCIAL man maximizes his own welfare when he is constrained
ACCOUNTING STANDARDS to play by certain rules and in certain institutional set-
tings, ignoring his incentives to avoid or change the
In this paper, we assume that individ- rules. setting. etc. Meckling [1976] analyzes this issue.
uals act to maximize their own utility. In 3 Ball and Watts [1972]; Barefield and Comiskey
doing so, they are resourceful and in- [1972]; Barnea, Ronen and Sadan [1975]; Beidleman
t1973]; Copeland [1968]; Cushing [1969]; Dasher and
novative.2 The obvious implication of Malcom [1970]; Gordon [1964]; Gordon, Horwitz and
this assumption is that management Meyers [1966].
pirical tests in the smoothing literature through i) taxes, ii) regulatory procedures
leave the Gordon model essentially un- if the firm is regulated, iii) political costs,
confirmed.4 Also, certain aspects of the iv) information production costs, and
Gordon model contribute to the model's directly via v) management compensa-
lack of confirmation. Essentially, Gordon tion plans. The first four factors increase
[1964] assumed that shareholder satis- managerial wealth by increasing the
faction (and, presumably, wealth) is cashflows and, hence, share price. The
solely a positive function of accounting last factor can increase managerial wealth
income. This assumption avoids the con- by altering the terms of the incentive
flict between shareholders and manage- compensation. Each of these five factors
ment by implying that increases in stock are discussed in turn.
prices always accompany increases in
Factors Affecting Management Wealth'
accounting income. However, recent re-
search casts serious doubt on the ability Taxes. Tax laws are not directly tied to
of management to manipulate directly financial accounting standards except in
share prices via changes in accounting a few cases (e.g., the last-in-first-out in-
procedures.5 ventory valuation method). However, the
We assume that management's utility indirect relationship is well documented
is a positive function of the expected Zeff [1972] and Moonitz [1974]. The
compensation in future periods (or adoption of a given procedure for finan-
wealth) and a negative function of the cial accounting does not decrease the
dispersion of future compensation (or likelihood of that procedure's being
wealth). The question is how do account-
ing standards affect management's 4 For these defects see Ball and Watts [1972], Gonedes
[1972] and Gonedes and Dopuch [19741.
wealth?6 Management's total compensa- 5 Fama [1970] and Goedes and Dopuch [19741.
tion from the firm consists of wages, in- Further, the results of studies by Kaplan and Roll [19721,
centive compensation (cash bonuses and Ball [1972] and Sunder [1975] which address the specific
issue support the hypothesis that the stock market can
stock or stock options), and nonpecuni- discriminate between real events and changes in account-
ary income, including perquisites (dis- ing procedures. Given that the market can on average
cussed in Jensen-Meckling, 1976a). Since discriminate, then it must be concluded that managers
(on average) expect the market to discriminate. Obvious-
it is unclear what role accounting stan- ly, managers do and will attempt to influence their share
dards play in the level of nonpecuniary price by direct accounting manipulation, but if these
income, we exclude it and focus on the attempts consume resources, then incentives exist to
eliminate these inefficient allocations.
first two forms of compensation. To the 6 For earlier discussions of this question see Watts
extent that management can increase [1974] and Gonedes [1976].
either the level of incentive compensa- We have purposefully excluded from the set of fac-
tors being examined the information content effect of an
tion or the firm's share price via its choice accounting standard on stock prices. We have done this
of accounting standards, they are made because at present the economic theories of information
better off. and capital market equilibrium are not sufficiently de-
veloped to allow predictions to be made regarding the
This analysis distinguishes between influence an accounting standard on the capital market's
mechanisms which increase manage- assessment of the distributions of returns (see Gonedes
ment's wealth: 1) via increases in share and Dopuch, 1974). We believe that a theory of the
determination of account ng standards can be developed
price (i.e., stock and stock options are and tested ignoring the information content factor. If at
more valuable) and 2) via increases in in- some future date, the information content factor can be
centive cash bonuses. The choice of ac- specified and included in the theory, then the predictions
and our understanding of the process will be improved.
counting standards can affect both of But we see no reason to delay the development of a
these forms of compensation indirectly theory until information content is specified.
Political Costs. The politicalsectorhas size and anti-trust [Siegfried 1975]. In proposed anti-
trust legislation, size per se has been mentioned specifical-
the power to effect wealth transfersbe- ly as a criterion for action against corporations. See the
tween various groups. The corporate "Curse of Bigness," Barron's, June 30, 1969. pp. 1 and 8.
sector is especially vulnerable to these Also see a bill introduced into the Senate by Senator
Bayh (U.S. Congress, Senate, Subcommittee on Anti-
wealth redistributions.Certaingroupsof trust and Monopoly (1975), pp. 5-13) would require
voters have an incentiveto lobby for the divesture for oil firms with annual production and/or
nationalization,expropriation,break-up sales above certain absolute numbers. In the hearings on
that bill, Professor Mencke of Tufts University argued
or regulationof an industryor corpora- that absolute and not relative accounting profits are the
tion.'0 This in turn providesan incentive relevant variable for explaining political action against
for elected officials to propose such ac- corporations.
Menke said, "Nevertheless, precisely because the
tions. To counterthese potentialgovern- actions of large firms are so visible, the American public
ment intrusions,corporationsemploy a has always equated absolute size with monopoly power.
number of devices, such as social re- The major oil companies are among the very largest and
most visible companies doing business in the United
sponsibility campaigns in the media, States.
governmentlobbyingand selectionof ac- Huge accounting profits, but not high profit rates, are
counting procedures to minimize re- an inevitable corollary of large absolute firm size. This
makes these companies obvious targets for public
portedearnings." By avoidingthe atten- criticism." (U.S. Congress, Senate, Subcommittee on
tion that "high"profitsdraw becauseof Anti-trust and Monopoly (1976), p. 1893).
cism suggests that Superior Oil Com- ers and nonmanaging directors have
pany (1974 sales of $333 million) incurs more incentive to adjust for and control
considerably less costs from anti-trust, increases in reported earnings due to
''corporate responsibility,'' affirmative changes in accounting standards than do
action, etc., than Exxon with sales of $42 politicians and bureaucrats.
billion.
Incentives for Various Groups to Adjust
InformationProduction(i.e., bookkeep-
for a Change in Accounting Standards
ing) Costs. Changes in accounting pro-
cedures are not costless to firms. Ac- An individual (whether a shareholder,
counting standard changes which either nonmanaging director, or politician) will
increase disclosure or require corpora- adjust a firm's accounting numbers for a
tions to change accounting methods in- change in accounting standards up to the
crease the firms' bookkeeping costs (in- point that the marginal cost of making
cluding any necessary increases in the adjustment equals the marginal bene-
accountants' salaries to compensate for fits. Consider the incentives of the outside
additional training).13 directors to adjust bonus compensation
Management Compensation Plans. A plans due to a change in accounting
major component of management com- standards. If these directors do not adjust
pensation is incentive (bonus) plan in- the plans, management compensation
come (Conference Board [1974]), and rises and share price falls by the full dis-
these plans are based on accounting in- counted present value of the additional
come. Our survey of 52 firms in our compensation."7 Each outside director's
sample indicates that the majority of the wealth declines to the extent of his owner-
companies formally incorporate account- ship in the firm and there is a greater
ing income into the compensation plan. 4 chance of his removal from the board. 18
Hence, a change in accounting standards
13 We are assuming that any change in accounting
which increase the firm's reported earn-
standards does not reduce the firm's information produc-
ings would, ceteris paribus, lead to greater tion costs. Although there may be cases where a firm is
incentive income. But this would reduce using a costly procedure which is eliminated by a simpler,
the firm's cashflows and share prices cheaper procedure, information production costs in this
case may decline, but we expect these situations to be
would fall. As long as the per manager rare.
present value of the after tax incentive 14 The frequency is 69 percent.
'5 At this early stage in the development of the theory,
income is greater than the decline in each we assume that management of the firm is composed of
manager's portfolio, we would expect homogeneous (i.e., identical) individuals to simplify the
management to favor such an accounting problem.
16 Our examination of the description of 16 manage-
change. 5 But this assumes that the share- ment compensation plans indicated that all the plans
holders and nonmanager directors do not were administered by the nonmanaging directors.
oppose such an accounting change or do 17 Likewise, we would expect the outside directors to
not adjust the compensation plans for the adjust the incentive compensation targets in those cir-
cumstances when it is in the shareholders' interest to
change in earnings. 6 In fact, the in- report lower earnings (e.g., LIFO), thereby not reducing
creased cashflows resulting from the the managers' incentive via bonus earnings to adopt
political costs, regulatory process and LIFO.
18 Our analysis indicates that outside (nonmanaging)
tax effects of an accounting change as- directors are "efficient" monitors of management, Watts
sumes that various politicians/bureau- 11977].If this were not the case, the capital market would
crats (i.e., the electorate) do not fully quickly discount the presence of outside directors. As far
as we can determine, firms are not required by the New
adjust for the change. A crucial assump- York Stock Exchange listing requirements or Federal
tion of our analysis is that the sharehold- regulations to have outside directors. Paragraph 2495G
If nonmanaging directors did not con- For small rate increases, the per capita
trol management (including adjusting coalition costs each consumer (or some
the compensation plans for changes in group of consumers) would bear lobby-
accounting standards), the decline in ing for the regulator's removal would
firm value offers incentives for an out- vastly outweigh the small per capita bene-
sider or group to tender for control of the fits they would receive via lower regulated
firm and install outside directors who rates. Hence, rational consumers would
will eliminate those managerial activities not incur large monitoring costs of their
which are not in the best interest of the regulators and other politicians (Downs
shareholders.19 This group would then [1957]; Alchian [1969]; and Alchian
gain a proportionate share of the full and Demsetz [1972]). Knowing this, it is
capitalized value of the eliminated abuses not in the regulators' and politicians'
(e.g., the present value of the incremental interests to adjust changes in accounting
compensation resulting from the change standards as fully as if they were con-
in accounting standards). Therefore, the fronted with the same change in account-
benefits for shareholders and nonman- ing standards in the role of outside di-
aging directors to adjust compensation rectors or shareholders in the firm. The
plans for changes in accounting stan- benefits of adjusting for changes in ac-
dards are immediate and direct, if there is counting standards are lower in the politi-
an efficient capital market for equity cal sector than in the private sector.2'
claims. Hence, there is a greater likelihood that a
However, for the politicians and bu- given accounting standard change will
reaucrats, our analysis suggests that the result in increased tax, regulatory, and
lack of a capital market which capitalizes political benefits than will the same
the effects on the voters' future cashflows change result in increased management
reduces the benefits accruing to the compensation. For a given accounting
politicians of monitoring accounting standard change, managers should expect
standards, and the result is that they will their own shareholders and outside di-
perform less adjustments for changes in
accounting standards.20 For example,
of Commerce Clearing House, Volume 2, New York
what are the benefits accruing to a utility Stock Exchange encourages listed firms to appoint out-
regulator for adjusting a utility's account- side directors. "Full disclosure of corporate affairs for the
ing numbers for a change in standards? information of the investing public is, of course, normal
and usual procedure for listed companies. Many com-
In the previous case of an outside direc- panies have found this procedure has been greatly aided
tor, the share price will fall by the dis- by having at least two outside directors whose functions
counted presented value of the increased on the board would include particular attention to such
matters.- This listing statement is consistent with our
compensation resulting for an incom- observation that outside directors provide monitoring
plete (or inaccurate) adjustment of the benefits.
compensation plan. But if the regulator 19 This assumes, of course, that such takeovers earn
a fair rate of return net of transactions costs.
does not completely adjust for a change 20 See Zimmerman [ 1977] and Watts [ 19771for further
in accounting standards and allows the discussion of this issue.
utility's rates to increase (resulting in a 21 It could also be argued that politicians and regu-
wealth transfer from consumers to the lators have a higher marginal cost of adjusting than do
shareholders, nonmanaging directors, and other capital
utility's owners), then the only cost the market participants since the former group does not
regulator is likely to incur is removal necessarily have a comparative advantage of adjusting
from office due to his incomplete adjust- financial statements, whereas, existing capital market
participants probably have a comparative advantage at
ment. He incurs no direct wealth change. such activities.
rectors to make a more complete adjust- have an incentive to oppose the standard
ment than politicians. since their bonus compensation plans will
Given this analysis, we predict that have to be adjusted (a costly process), if
managers have greater incentives to their incomes are to remain unchanged
choose accounting standards which re- by the new standard. Above size E, the
port lower earnings (thereby increasing political, regulatory, and tax benefits of
cashflows, firm value, and their welfare) reporting lower earnings due to the new
due to tax, political, and regulatory con- standard are assumed to dominate the
siderations than to choose accounting incentive compensation factor.
standards which report higher earnings The benefits (costs) of a proposed ac-
and, thereby, increase their incentive counting standard are expected to vary
compensation. However, this prediction with the firm's size. This relationship can
is conditional upon the firm being regul- exist for two reasons: (1) the magnitude
ated or subject to political pressure. In of the reported income change may be
small, (i.e., low political costs) unregul- larger for larger firms and (2) for an in-
ated firms, we would expect that man- come change of a given magnitude, the
agers do have incentives to select ac- benefits (costs) vary with firm size.23
counting standards which report higher Hence, the present value of the stream of
earnings, if the expected gain in incentive benefits (or costs) to the firm, GB, are an
compensation is greater than the fore- increasing function of firm size.24
gone expected tax consequences. Finally, Information production costs, curve
we expect management also to consider IC, are also expected to vary to some ex-
the accounting standard's impact on the tent with firm size due to the increased
firm's bookkeeping costs (and hence complexity and volume of the larger
their own welfare).
The next section combines these five 22 The expected effect of an accounting standard could
factors into a model of corporate lobby- vary over time (i.e., it could increase current reported
income and decrease some future reported income). In
ing standards. that case, the analysis is slightly more complex, but the
criterion is still the same (i.e., the effect on the manager's
A POSITIVE THEORY OF MANAGEMENT wealth). However, for simplicity, the remainder of the
paper refers to standards increasing or decreasing re-
LOBBYING ON ACCOUNTING STANDARDS ported income as though the whole time series of future
Given a proposed accounting stan- income shifts uLp or down.
23 Whether the magnitude of the income change does
dard, management's position depends on vary with firm size depends on the particular accounting
the size of the firm (which affects the standard in question. For certain accounting standards
magnitude of the political costs) and (e.g., requiring all firms to report depreciation based on
current replacement costs) it is apparent a priori that
whether the proposed standard increases there will be a correlation between the income change
or decreases the firm's reported earn- and firm size. For other standards (e.g., general price level
ings.22 Figure I separates the standard's accounting) a priori, it is not obvious that a relationship
will exist (e.g., net monetary gains may offset depreciation
impact on earnings into decreases (1A) in larger firms). However, since political costs depend on
and increases (IB). The curve GB in firm size then we expect the benefits (costs) of standard
Figure IA (earnings decrease) denotes changes to vary with firm size. For example, if all firms'
earnings decline by $1 million (due to a standards
the proposed accounting standard's pres- change) then we would expect larger firms to incur larger
ent value to management including the benefits since the likelihood of anti-trust actions are
tax, regulatory, political, and compensa- expected to be associated with firm size.
24 We would expect firms in different industries to be
tion effects as a function of firm size. For subject to different political pressures, tax structures, and
small firms (below size E), not subject to regulation. Hence, Figure 1 is developed for firms in the
much political pressure, these managers same industry that only differ by size.
FIGURE I
~~~~~NB
_ - C IC
_~~~~~~~~~~wo-0
/
I A
t0, I
Size
Unfavorable INo I Favorable
Submission Submission | Submission
No
Submission
Firm
IB. Accounting Earnings Increase
ENB
ExpectedENCS
Present EBC
Value
Ah ~~~~~~IC
D'
0
AM-
~~~~~~~~~wo-A Size
firm's accounting system. The difference the amount CS, ENB-CS. A firm will
between the gross benefits, GB, and the make a submission if ENB-CS is posi-
additional information costs, IC, yields tive. This occurs in the regions DA,
the net benefits curve, NB. where opposing submissions occur, and
If the firm size is in the region OB, the beyond C, where favorable submissions
net benefits curve, NB, is negative, and are made. Between 0 and D and between
the firm will consider making an un- A and C no submissions are made.
favorable submission to the FASB. Be- In Figure 1B, the proposed standard
fore the firm makes a submission, man- increases reported income. This case is
agement holds beliefs regarding the like- similar to the previous one except the
lihood the FASB will adopt the standard gross benefits are only positive for small
and the likelihood the FASB will adopt firms where the management compensa-
the standard if the firm makes an oppos- tion plans are expected to dominate the
ing submission.25 The difference between tax, political, and regulatory factors.
these beliefs is the change in the adoption Beyond size E' gross benefits are nega-
likelihood if management makes a nega- tive since, for those firms, the income
tive submission. The product of this increases are expected to increase gov-
difference and the negative net benefits, ernmental interference (political costs),
NB, (i.e., the present value26 of the cash- raise future tax payments, and lead the
flows arising from the five factors) is the public utility commission to reduce the
expected present value of the net benefits firm's revenues (if the firm is regulated).
curve, ENB. For example, a firm will The net benefits curve is again the
incur negative net present value benefits algebraic sum of GB (gross benefits) and
of $100,000 if the standard is adopted. IC (information costs) and the submis-
They believe the likelihood of adoption sion's expected net benefits less sub-
is .60. By making a negative submission mission costs, ENB-CS, cuts the axis
to the FASB the likelihood falls to .59. at A'. Accordingly, firms with asset sizes
The expected net present value of the in the interval OA' make no submissions
benefits of the submission is then and firms of sizes beyond A' make un-
+ $1000. favorable submissions.
Firms larger than size B face positive
net benefits if the standard is adopted. 25 In this situation, it is possible that management will
They will consider supporting the stan- lobby on an accounting standard because of secondary
(or gaming) effects (i.e., vote trading thereby influencing
dard to the FASB, thereby increasing subsequent FASB pronouncements). We chose not to
the standard's likelihood of adoption.27 introduce gaming because it complicates the model and
Hence, the expected net benefits curve is such complication is only justified if it improves or is
likely to improve the empirical results. We are able to
also positive beyond point B since it is predict corporate behavior without considering gaming.
the product of a positive net benefit and and we do not consider it likely to improve these results.
26 The firm is discounting the future cashflows with
a positive change in the FASB's likeli- the appropriate, risk-adjusted discount rate. Further-
hood of adoption given a favorable sub- more. we are assuming that this discount rate is not in-
mission. creasing in firm size which is consistent with the avail-
able evidence.
If the cost of the submission is $CS, 27 We are assuming that the likelihood of the FASB
consisting primarily of the opportunity adopting the standard, if the firm makes a submission,
cost of the manager's time, then the total is independent of firm size. This is unrealistic since large
firms, we expect, would have more influence with the
expected net benefits of a submission Board. However, inclusion of this additional dependency
given the submission cost is a vertical does not change the results; in fact, it strengthens the
downward shift in the ENB curve by predictions.
porting the Effects of General Price- tained from the COMPUSTAT tape and the
Level Changes in Financial Statements" 1974 Moody Manuals. In addition, data
and scheduled a public hearing on the on the existence of management incen-
topic for April 25, 1974. Public com- tive compensation plans was obtained by
ments and position papers were solicited. a questionnaire mailed to the chief
One hundred thirty-three accounting
firms, public corporations, industry or- TABLEI
available for our sample. in their net income and ranked by their
Davidson and Weil [1975c] also esti- asset size (Table 2). Of the 26 firms with
mate the effect of GPLA on income for income decreases, eight voted yes and
1974 (which was in the future at the time 18 no.34 The eight yes votes came from
of the submissions). Even though the the larger firms, thus supporting our pre-
adjustment procedure was slightly dif- diction. To test the null hypothesis that
ferent, only two of our 19 firms in the the eight firms which voted yes are drawn
combined samples reverse the direction from the same population of firms (with
of the income effect between 1973 and respect to size) as the 18 that voted no,
1974. Similarly, all of the utilities, (24), we performed a Mann-Whitney U test.
and 35 of the 50 other companies in their Our tables indicate that we can reject the
sample have income effects of the same null hypothesis at the .001 level.35
sign in both years. Since the effects of Of the eight firms with income in-
income changes in the immediate future creases or no changes in net income,
are less heavily discounted, these results seven voted no. Thus, the general ten-
suggest that the error introduced by our
assumption of stationary income changes 3 In this case, firm size is measured by the firm's
is not likely to be severe. Fortune 500 rank in assets. The results are identical when
rank in sales is used. Furthermore, the intent of govern-
ment intervention depends on the metric used by the
Tests of the Theory courts, legislators, and regulators. Market share, con-
In the reported tests, we use asset size centration and size are among the commonly used indi-
cators. Absolute size is important in explaining govern-
as the surrogate for firm size.33 Based on ment regulation for both theoretical and empirical rea-
our model, we can make predictions sons. An implication of Peltzman's (1975, p. 30) theory
about the relationship between asset size of regulation is that the amount of wealth redistributed
from firms by government intervention is a positive func-
and firm submissions. We predict that tion of economies of scale. Since we expect large firm size
firms whose earnings are increased by to indicate the presence of economies of scale, implica-
GPLA will oppose GPLA regardless of tion of Peltzman's theory is that government interven-
tion will be greater for larger firms. Empirically, we ob-
their size (i.e., there will be no association serve numerous cases of politicians and regulators echo-
between size and submission). However, ing the conventional wisdom of certain segments in
for firms whose earnings are decreased by society, that big business is inherently bad. (See, "Curse
of Big Business," Barron's June 16, 1969 and footnote
GPLA, we predict that they will either 12).
support GPLA or will not make a sub- 3 We use the term "vote" to mean responding to a
mission depending on where asset size discussion memorandum by issuing a corporate opinion.
5 Siegel [1956], p. 274. Even after any reasonable
C (Figure 1) occurs in their industry. adjustment for the degrees of freedom lost due to pre-
Since we cannot determine the asset size vious statistical analysis, this result is still significant.
corresponding to point C, we are in An intuitive idea of the strength of the relationship
between management's attitude and firm size can be ob-
a position analogous to being able to tained by considering an analogy. Suppose we put 26
predict the sign of a regression coefficient balls in an urn representing the firms with earnings de-
but not its magnitude. Consequently, our creases; eight red balls representing the firms that voted
yes; and 18 black balls, representing the firms that voted
test of the model does not include asset no. Now, we randomly draw 13 balls out of the urn with-
size C (analogous to the magnitude of the out replacement representing the largest 13 firms (out of
coefficient). The test is only of the pre- the 26). The probability that we draw eight red balls
(analogous to the probability of the eight firms voting
diction that there is a positive relation- yes being the "large" firms if the null hypothesis of no
ship between asset size and submission association between votes and size is correct) is .001.
for firms with income decreases. If the votes of firms are not independent, as in the case
of gaming, this analogy is inappropriate. But we do not
Firms making submissions were classi- have any evidence of vote dependence (via gaming or
fied according to the direction of change otherwise).
TABLE 2
ASSET SIZE, DIRECTIONOF EARNINGSEFFECTAND CORPORATEPOSITIONON GPLA
I Exxon 1 Yes
2 General Motors 2 Yes
3 Texaco 3 No
4 Ford 4 Yes
5 Sears Roebuck (Rank I in retail sales) 7 No
6 IT&T 8 No
7 Gulf Oil 9 Yes
8 Standard Oil of California 10 Yes
9 General Electric 11 No
10 Standard Oil of Indiana 12 No
11 Shell Oil 16 Yes
12 Dupont E.I Nemours 18 Yes
Point C*
13 Union Carbide 22 No
14 Continental Oil 26 No
15 Marcor (Rank 2 in retail firms) 33 Yes
16 International Harvester 34 No
17 Caterpillar Tractor 47 Yes
18 Rockwell International 54 No
19 W. R. Grace 55 No
20 Owens-Illinois 80 No
21 Inland Steel 85 No
22 American Cyanamid 92 No
23 United Aircraft 107 No
24 Seagrams Sons Inc. 108 No
25 Eli Lilly & Co. 135 No
26 Merck 143 No
27 General Mills 156 No
28 Texas Instruments 164 No
29 Gillette 167 No
30 Reliance Electric 332 No
31 Harsco 368 No
32 Masonite 386 No
33 Automated Building Components Not Ranked No
34 Copeland Corporation Not Ranked No
* Point C in Figure I is determined by minimizing the number of misclassifications.
t Yes = Favored GPLA
No = Opposed GPLA
frequency of firms with income increases tively larger regulated firms favor GPLA.
that support price-level adjustments is If we assume our model is correct and
one out of eight (12.5 percent). that asset size C is the same for all indus-
The above results support the relation- tries, we can estimate C by minimizing
ship between management's attitudes on the number of prediction errors (analog-
GPLA and firm size for the 23 unregu- ous to estimating a regression coefficient
lated firms. However, if we assume that by minimizing the sum of squared errors).
firm size and the direction of the income This estimate provides information on
change are independent (Table 2 supports the relative importance of political and/or
this assumption), then (if there is no size tax costs for different size firms. Given the
effect) the average size of firms support- data, C is between the 18th and 22nd
ing GPLA should be the same as the largest firms in the Fortune 500 in 1973
average size of firms opposing. Thus we (see Table 2). This suggests that reduced
can use the voting behavior of all 52 political and/or tax costs outweigh infor-
firms in our sample to test the size rela- mation production and/or management
tionship. compensation factors in determining
Table 3 presents the median rank on management's position on GPLA only
asset size for both regulated and unregu- for very large firms. For most other firms,
lated firms favoring and opposing GPLA. information production costs dominate.
The median rank in the Fortune 500 of Are the major benefits of reporting
the nine unregulated firms supporting lower adjusted incomes derived from tax
GPLA is 10. The median rank of the 25 or political considerations? It is very
unregulated firms opposing GPLA is 92. difficult to differentiate between these
two factors, but one possible way is the
TABLE 3 following. Is the change in adjusted in-
MEDIAN RANKS OF FIRM SIZE BY REGULATION AND come proportional to firm size? If it is,
POSITION ON GPLA*
then both the tax and political factors
may be operating. But if there is no asso-
Regulated (N = 18) Unrequlated(N = 34)
ciation between firm size and the magni-
In Favor Against In Favor Against tude of the income change, then the tax
(9) (9) (9) (25) effect cannot explain why larger firms
Median fdvor GPLA. Therefore, this result could
Rank 13 38 10 92 only be due to political costs. We can
* Fortune [Mayand July, 1974]. obtain estimates of the income effect of
GPLA for 11 of the firms whose incomes
For regulated firms, there also appears would be reduced by GPLA (six sup-
to be a relationship between size and porting, five opposing).36 The average
management attitudes. The net incomes reduction in income for the six firms
for all the utilities investigated by David- which supported GPLA is $177.7 mil-
son and Weil [1975b] are increased by lion, while the average reduction for the
GPLA suggesting none of the utilities five which opposed GPLA is $38.5
should favor GPLA. However, as noted million. Thus, it appears that the income
in the preceding section, operating earn- change does vary with size and the pre-
ings are relevant to rate determination. 36 This test was performed on 11 firms with income de-
Those earnings fall for all the utilities creases which Davidson and Weil reported 1973 ad-
justed earnings. Firms which were manually adjusted by
investigated by Davidson and Weil us for Table 2 were excluded from this test since only the
[1975b] and this could explain why rela- sign of the earnings change was calculated.
ceding results are consistent with both agement's attitudes on accounting stan-
the tax and political costs affecting man- dards. Although we are not able to ex-
agement's attitudes. plain some of the notable nonsubmitting
The preceding results test only whether firms' decisions, we would point out that
the size effect exists for firms which did most of the firms submitting are large,
submit to the FASB. It is interesting to and the likelihood of submission in-
examine the effect of GPLA on firms creases with asset size (12 of the 18 firms
which did not submit. In particular, the ranked 1-1 8 in the Fortune500 submitted,
firms of asset size above our estimated C four of the 18 firms ranked 19-36 sub-
which did not submit are of interest since mitted, two of the 18 firms ranked 37-54
our model predicts they would submit on submitted, one of the 18 firms ranked
the basis of the income effect. Dupont is 55-72 submitted, etc.).
the last firm above asset size C in Table 2
to vote. It is ranked 18th in the Fortune Discriminant Analysis
500 in 1973. There are seven firms ranked The preceding tests were based on the
higher than 18th which did not make a direction of the earnings change, not the
submission to the FASB. They are IBM magnitude of the change. A discriminant
(ranked 5th), General Telephone (6th), analysis is conducted including manage-
Mobil Oil (7th), U.S. Steel (13th), Chrys- ment compensation, depreciation, and
ler (14th), Tenneco (15th), and Atlantic net monetary assets as independent vari-
Richfield (17th). ables, and using data on 49 of the 53
The size of the income change is crucial firms making submissions to ensure con-
to determining why these seven firms did sistency of the Davidson and Weil pro-
not submit. If changes are not associated cedures.
with firm size, the expected benefits of a The change in price-adjusted income is
submission could be very small and may correlated with the magnitudes of depre-
not exceed the submission costs. Unfor- ciation and net monetary assets. The
tunately, Davidson and Weil only esti- larger both of these variables in unad-
mated the change in earnings in 1973 for justed terms, the larger will be the decline
three of these seven firms: IBM, U. S. (in absolute dollars) in adjusted net in-
Steel, and Chrysler. All three have in- come. We do not perform an actual price-
come reductions with GPLA and their level adjustment, but rely on the unad-
average reduction is $88 million. This is justed magnitudes of depreciation and net
less than the average reduction for the six monetary assets.
firms with income reductions which did 3 A more likely explanation of U.S. Steel's failure to
submit ($177 million), but it is not trivial. submit is the fact that the steel industry was not as politi-
cally sensitive as the oil industry (for example) at the
Further, the reductions for two of the time. In other words, a given earnings effect has less politi-
three nonsubmissions (IBM and General cal cost or benefit. This possibility is not included in our
Telephone) exceed the reductions for four model. This could also explain Chrysler's failure to sub-
mit. As number three after General Motors and Ford they
of the six submissions. Consequently, it is may be subject to less political pressure (and hence cost).
difficult to attribute the fact that the In addition, the "free rider" effect may explain some of
three firms did not submit to the lack of these nonsubmissions.
While we can only expect a positive theory to hold on
an income effect.37 average, the failure of IBM to submit is puzzling. That
In summary, these tests confirm the firm has anti-trust suits outstanding and some economists
relationship between size and manage- allege that it earns monopoly profits. For a discussion of
one of these suits and statements by economists that IBM
ment attitudes on GPLA. Political costs earns monopoly profits, see "The Breakup of IBM"
and, perhaps, tax effects influence man- Datamation, October 1975, pp. 95-99.
TABLE4
DISCRIMINANTANALYSIS
Coefficients (t-statistics)
Yates
Model SALES Adjusted
Num- DEP/ NMA/ SALES TSALES Chli
her N Sample Constant MKTVL MKTVL x CHG CHG MCOMP REG R2 Square*
1 49 total -.0241 122.6 -38.9 .000044 -.4131 -.2355 -.3443 .358 9.25
sample (-.12) (.60) (-1.62) (3.67) (-1.11) (-1.42) (-1.29)
2 49 total -.0855 160.4 - 14.2 .000043 - .4381 - .1619 .332 9.25
sample (-.44) (.79) (-.98) (3.53) (-1.17) (-1.03)
3 49 total -.0973 143.0 -15.6 .000034 -.1601 .311 9.25
sample (-.50) (.70) (-1.07) (3.58) (- 1.02)
4 34 unregulated .0431 74.0 -36.5 .000044 - .3271 -.2186 .366 19.96
firms (.19) (.27) (-1.06) (3.58) - .89) (-.89)
5 34 unregulated .0412 86.2 -35.3 .000038 -.2335 .347 13.16
firms (.18) (.32) (-1.03) (3.73) (-.96)
6 49 total -.0079 215.3 .000033 -.2365 .0077 .293 11.74
sample (-.04) (1.09) (3.44) V 1.39) (.05)
7 49 total - .0662 .000033 .201 5.98
sample (- 1.03) (3.44)
* The Yates correction for continuity is useful in establishing a lower bound on the X2 statistic.
capture the tax effects, are predicted to are proxies for political costs. These two
be positive under that hypothesis (the variables, assume that political costs are
larger the depreciation and net monetary symmetric for both earnings increases
assets the greater the decline in adjusted and decreases. The multiplicative dum-
income and the greater the tax benefits). my, CHG, is positive if earnings decline
The sign on normalized depreciation (based on the Davidson-Weil [1975a]
is as predicted, but normalized net mone- results) or if the firm is regulated.42
tary assets is of the wrong sign. One of The sign on SALES x CHG is as pre-
the following three hypotheses explain dicted, positive, and in addition has the
this result: the tax effect is only operating highest t-statistic of all the independent
via depreciation;41 depreciation and net variables. In addition, the coefficient on
monetary assets, being inversely related SALES x CHG is the most stable co-
(correlation coefficient ranging from efficient across various realizations and
-.41 to -.55), are entering the regres- subsamples which leads us to conclude
sion with opposite signs; or the tax effect that firm size is the most important vari-
is not an explanatory factor. Since our able. The sign of
sample is very small, it is not possible to
use a holdout subset to distinguish be- 4' That is, this sample of firms does not expect the
tween these hypotheses. tax laws to be changed to include in taxable income gains/
The next two variables, losses on net monetary assets.
42 Since the regulatory commission bases rates on
depreciation, net monetary assets are not expected to be
(SALES) CHG and ALES) CHG, an important consideration, hence operating earnings
decline for regulated firms.
MCOMP and MKTVL plans indicates that although the minimum and maxi-
mum amounts transferred to the bonus pool depend on
the final net income number, we find that the actual
(correlation coefficients of -.60 and bonus paid is most highly associated with operating or
current income (depreciation is included, but extraordi-
-.86 respectively). When nary gains and losses are excluded). We correlated the
NMA change in management incentive compensation expense
for 271 COMPUSTAT firms with changes in operating
MKTVL income and changes in net income after extraordinary
items. The correlation coefficient for changes in operating
is deleted from the model (Model 6), the income exceeded that for changes in net income after
extraordinary items for over two-thirds of the firms.
sign of REG reverses, the importance of Gains or losses on monetary assets are not included in
operating income. Consequently, only adjusted depreci-
DEP ation (ignoring inventory adjustments) are expected to
affect management compensation and the effect is to
MKTVL reduce management pay.
all models the constant is close to zero. rectly through possible governmental in-
The estimated discriminant functions tervention (antitrust, price controls, etc.),
are consistent with the tests of the theory. can affect their future cashflows by dis-
All of the discriminant functions are couraging government action through
statistically significant and the inter- the reporting of lower net incomes. The
vening variable driving these findings is empirical evidence with respect to the
firm size. In fact, firm size explains over position 52 firms took before the FASB
half the explained variance in voting on price level restatements is consistent
behavior (Model 7). with respect to this hypothesis.
These results are consistent with those The single most important factor ex-
using the Davidson and Weil findings. plaining managerial voting behavior on
The discriminant functions indicate that General Price Level Accounting is firm
the political cost factor is more important size (after controlling for the direction
than the tax factor in affecting manage- of change in earnings). The larger firms,
ment's attitudes. ceteris paribus, are more likely to favor
The major empirical problem in the GPLA (if earnings decline). This finding
discriminant analysis is the rather small is consistent with our government inter-
sample size which precludes using a hold- vention argument since the larger firms
out sample and, furthermore, does not are more likely to be subjected to govern-
allow more sophisticated econometric mental interference and, hence, have
techniques to control for the multi- more to lose than smaller corporations.
colinearity. Hence, it is difficult to con- The existence of costs generated by
trol for the interaction between the under- government intervention may have more
lying factors. However, these preliminary fundamental and important effects on
results are encouraging and suggest that the firm's decisions than just its lobbying
additional research in this area is war- behavior on financial accounting stan-
ranted. dards. Not only would we expect the firm
to manage its reported earnings, but also
SUMMARY AND CONCLUSIONS to alter its investment-production deci-
We have focused in this paper on the sions if the potential costs of government
question of why firms would expend interference become large. For example,
resources trying to influence the determi- government intervention costs may lead
nation of accounting standards. The his- the firm to select less risky investments in
tories of the Committee on Accounting order to eliminate the chance of high
Procedures, the Accounting Principles returns which then increase the likeli-
Board, and FASB are replete with exam- hood of government intervention. If the
ples of managements and industries exert- total risk of these less risky investments
ing political pressure on the standard- tends to be positively correlated with the
setting bodies. systematic risk of the firm, then we would
A possible answer to this question is expect the beta (the estimate of the co-
provided by the government intervention variance between the return on the stock
argument, namely, that firms having and the market return normalized by the
contact (actual or potential) with govern- variance of the market) on the common
ments, directly through regulation (pub- stock to be significantly below one (aver-
lic utility commissions, Interstate Com- age risk) for those firms facing large
merce Commission, Civil Aeronautics government intervention costs. The evi-
Board, etc.) or procurement, or indi- dence from the sample of firms making
REFERENCES
Alchian, A. A., "Corporate Management and Property Rights," in Economic Policy and th/e Rcgulation
of Corporate Securities (H. Manne, ed.), (American Enterprise Institute, 1969).
and H. Demsetz, "Production, Information Costs and Economic Organization," American
Economic Review (December 1972), pp. 777-795.
Armstrong, Marshall S., "The Politics of Establishing Accounting Standards," A speech before the Third
Annual Securities Regulation Institute in San Diego, California, January 16, 1976, as reported in
Arthur Andersen & Co., Executive News Briefs, (February 1976), p. 1.
Ball, R., "Changes in Accounting Techniques and Stock Prices," Empirical Research in Accounting:
Selected Studies, 1972. Supplement to Journal of Accounting Research (1972), pp. 1-38.
and Ross Watts, "Some Time Series Properties of Accounting Income," Journal of Finance
(June 1972), pp. 663-82.
Barefield, R. M., and E. E. Comiskey, "The Smoothing Hypothesis: An Alternative Test," THE ACCOUNT-
ING REVIEW (April 1972), pp. 291-298.
Barnea, A., J. Ronen, and S. Sadan, "The Implementation of Accounting Objectives-An Application to
Extraordinary Itens," THE ACCOUNTING REVIEW (January 1975), pp. 58-68.
Beidleman, C. R., "Income Smoothing: The Role of Management," THE ACCOUNTING REVIEW (October
1973), pp. 653-667.
Benston,. George J., "Accountants Integrity and Financial Reporting,' Financial Executive (August
1975), pp. 10-14.
The Conference Board, Top Executive Compensation (Conference Board, 1974).
Copeland, Ronald M., "Income Smoothing," Empirical Research in Accounting: Selected Studies. Sup-
plement to Journal ofA counting Research 1968, pp. 101- I 16.
Cushing, B. E., "An Empirical Study of Changes in Accounting Policy," Journal of Accounting Research
(Autumn 1969), pp. 196-203.
Dasher, B. E. and R. E. Malcom, "A Note on Income Smoothing in the Chemical Industry," Journal of
Accounting Research (Autumn 1970), pp. 253-259.
Davidson, Sidney, Clyde P. Stickney and Roman L. Weil, Inflation Accounting (McGraw-Hill, 1976).
and Roman L. Weil, "Inflation Accounting: What Will General Price Level Adjusted Income
Statements Show?" Financial Analysts Journal, (January-February 1975a) pp. 27-3 1; 70-8 1.
and Roman L. Weil, "Inflation Accounting: Public Utilities," Financial Analysts Journal, (May-
June 1975b), pp. 30-34; 62.
and Roman L. Weil, "Inflation Accounting: Some 1974 Income Measures," Financial Analysts
Journal, (September-October 1975c), pp. 42-54.
Downs, A., An Economic Theoryof Democracy," (Harper and Row, 1957).
Fama, Eugene F., "Efficient Capital Markets: A Review of Theory and Empirical Work,' Journal of
Finance, (May 1970), pp. 383-417.
Gonedes, N., "Income-smoothing Behavior Under Selected Stochastic Processes," Journal of Business
(October 1972), pp. 570-584.
, "Class Discussion Notes: Section 8," unpublished manuscript, University of Chicago (January
1976).
and N. Dopuch, "Capital Market Equilibrium, Information Production, and Selecting Account-
ing Techniques: Theoretical Framework and Review of Empirical Work," Studies on Financial Ac-
counting Objectives. 1974. Supplement to Journal of Accounting Research (1974).
Gordon, M. J., "Postulates, Principles and Research in Accounting," THE ACCOUNTING REVIEW (April
1964), pp. 251-263.
, B. N. Horwitz, and P. T. Meyers, "Accounting Measurements and Normal Growth of the Firm,"
Research in Accounting Measurement, eds. Jaedicke, Ijiri and Nielsen (American Accounting Associa-
tion, 1966), pp. 221-231.
Haskins and Sells. Public Utilities Manual (New York, 1974).
Horngren, Charles T., "The Marketing of Accounting Standards," Journal of Accountancy (October
1973), pp. 61-66.
, "Setting Accounting Standards in 1980," unpublished speech before the Arthur Young Professors
Roundtable (March 30-31, 1976).
Ijiri, Yuji, "The Price-Level Restatement and its Dual Interpretation," THE ACCOUNTING REVIEW,
(April 1976), pp. 227-243.
Jensen, Michael C., "Reflections on the State of Accounting Research and the Regulation of Account-
ing," Presented at the Stanford Lectures in Accounting, May 21, 1976.
, and William H. Meckling, "Theory of the Firm: Managerial Behavior, Agency Costs and
Ownership Structure," Journal of Financial Economics (October 1976a), pp. 305-360.
, and William H. Meckling, "Can the Corporation Survive?" Public Policy Working Paper Series,
PPS76-4, Graduate School of Management, University of RocIester, (April, 1976b).
Kaplan, R. S., and R. Roll, "Investor Evaluation of Accounting Information: Some Empirical Evi-
dence," Journal of Business (April 1972), pp. 225-57.
Meckling, William H., "Values and the Choice of the Model of the Individual in Social Sciences," Revue
Suisse d' Economic Politique et de Statistique (December 1976).
Moonitz, Maurice, Obtaining Agreement on Standards. Studies in Accounting Research No. 8 (Sarasota,
Florida: American Accounting Association, 1974).
Peltzman, S., "Toward a More General Theory of Regulation," Journal of Law and Economics, (August
1976), pp. 221-240.
Posner, Richard A., "Theories of Economic Regulation," The Bell Journal of Economics and Management
Science, (Autumn 1974), pp. 335-358.
Siegel, Sidney, Nonparametric Statistics (McGraw-Hill, 1956).
Siegfried, John, "Determinants of Antitrust Activity," Journal of Law and Economics (October 1975),
pp. 559-581.
Stigler, G. J., "The Theory of Economic Regulation," The Bell Journal of Economics and Management
Science, (Spring 1971), 3-21.
Sunder, S., "Empirical Analysis of Stock Price and Risk as They Relate to Accounting Changes in In-
ventory Valuation," THE ACCOUNTING REVIEW (April 1975), pp.305-315.
U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly of the Committee on the Judiciary,
Hearings, The Petroleum Industry, Part I, 9th Congress, 1st Session, 1975.
U.S. Congress, Senate, Subcommittee on Antitrust and Monopoly of the Committee on the Judiciary,
Hearings, The Petroleum Industry, Part III, 9th Congress, 1st Session, 1976.
Watts, Ross, "Accounting Objectives," Working Paper Series No. 7408, Graduate School of Manage-
ment, University of Rochester, (April 1974).
Watts, Ross, "Corporate Financial Statements: Product of the Market and Political Processes," Austral-
ian Journal of Management(April 1977), pp. 53-75.
Zeff- Stephen, Forging Accounting Principles in Five Countries: A Historv and an Analysis of Trends,
Arthur Andersen Lecture Series (Stipes Publishing Company, 1972), pp. 110-268.
Zimmerman, Jerold, "The Municipal Accounting Maze: An Analysis of Political Incentives," Supple-
ment to the Journal of Accounting Research (1977).