THE ACCOUNTING REVIEW

Vol. 65, No. 1

January 1990 pp. 131-156

I
Positive Accounting Theory: A Ten Year
Perspective
Ross L. Watts and Jerold L. Zimmerman University of Rochester

ABSTRACT: This paper reviews and critiques the positive accounting liter- ature following
publication of Watts and Zimmerman (1978, 1979). The 1978 paper helped generate the
positive accounting literature which offers an explanation of accounting practice, suggests
the importance of con- tracting costs, and has led to the discovery of some previously unknown
empirical regularities. The 1979 paper produced a methodological debate that has not
been very productive. This paper attempts to remove some common misconceptions about
methodology that surfaced in the debate. It also suggests ways to improve positive
research in accounting choice. The most important of these improvements is tighter links
between the theory and the empirical tests. A second suggested improvement is the
develop- ment of models that recognize the endogeneity among the variables in the regressions.
 A third improvement is reduction in measurement errors in both the dependent and
independent variables in the regressions.

T is more than a decade since our two papers, "Towards a Positive Theory of the Determination of
Accounting Standards" and "The Demand for and Supply of Accounting Theories: The Market for
Excuses" were published in The Accounting Review. The intervening time allows us to look back on these
papers and the ensuing literature with some perspective.
The two papers were controversial ten years ago and remain so today. The papers (primarily
Watts and Zimmerman 1978) contributed to a literature that has uncovered empirical regularities in accounting
practice (Christie forthcom- ing; Holthausen and Leftwich 1983; Leftwich forthcoming; Watts and
Zimmer- man 1986). The empirical regularities have been replicated in different settings
Financial support was provided by the John M. Olin Foundation and the Bradley Policy Research Center at
the University of Rochester. The comments of Ray Ball, James Brickley, Andrew Christie, Linda DeAngelo, Robert Hagerman, S.
P. Kothari. Richard Leftwich, Tom Lys, Clifford Smith, Jerold Warner, and Greg Whittred are gratefully acknowledged.
We thank William Kinney for encouraging us to pursue this project. An earlier version of this paper was presented at
the Accounting Association of Australia and New Zealand, July 4, 1989, Melbourne, Australia.

131
Manuscript received May 1989. Revision received
September 1989. Accepted September 1989.
132
The Accounting Review, January 1990

(Christie forthcoming) and it is clear there is a relation between firms' accounting choice and other firm variables,
such as leverage and size and the signs of the relations are mostly consistent across studies. Positive accounting research guided
the search for the empirical regularities and provided explanations for them. To date, there are no systematic alternative sets of
explanations for those regularities articulated and tested in the literature. Further, the literature has moved beyond the first
simple exposition of the theory in the 1978 paper. The explanation for accounting choice is now richer and more sophisticated.
Our first objective in this paper is to convey our perspective on the evolution and current state of positive
accounting theory and to summarize the evidence on systematic empirical regularities in accounting (Section I). The second
objec- tive is to evaluate the research methods and the methodology used to document the empirical regularities. We discuss
criticisms of the original papers and of the subsequent positive accounting literature in Secuon II. While the positive ac-
counting literature has explained some accounting practice, much remains un- explained. Our third objective is to provide our
views about future directions for positive accounting literature (Section III).

Evolution
I. Evolution and State of Positive Accounting Theory

Modern positive accounting research began flourishing in the 1960s when Ball and Brown (1968), Beaver
(1968), and others introduced empirical finance methods to financial accounting. The subsequent literature adopted the
assump- tion that accounting numbers supply information for security market invest- ment decisions and used this "information
perspective" to investigate the relation between accounting numbers and stock prices.' The "information per- spective" has
taught us much about the market's use of accounting numbers. But, except for the choice of inventory methods, the
“information perspective” has not provided hypotheses to predict and explain accounting choices. The “in- formation
perspective" has not provided hypotheses to explain why entire in- dustries switch from accelerated to straight-line
depreciation without changing their tax depreciation methods.
An important reason that the information perspective failed to generate hypotheses explaining and
predicting accounting choice is that in the finance theory underlying the empirical studies, accounting choice per se
could not affect firm value. Information is costless and there are no transaction costs in the Modigliani and Miller (1958) and capital asset
pricing model frameworks. Hence,

'The "information perspective" views accounting data (usually earnings, dividends, and cash flows) as providing information on
inputs to valuation models (e.g., discounted cash flows) and tests for associations between accounting disclosures and stock prices or
returns. In the contracting approach adopted in the literature and discussed in this paper, accounting methods are primarily determined
by the use of accounting numbers in contracts between parties to the firm. Under this approach account- ing disclosures directly affect parties'
(including stockholders') contractual claims and, hence, the values of those claims (including stock prices). To the extent accounting
disclosures are correlated with attributes investors use in valuing securities, these disclosures contain information and affect stock prices. Thus, under
both an "information perspective" and a "contracting perspective," accounting disclosures have the potential to alter securities prices
(Holthausen forthcoming).
Watts and Zimmerman-Positive Accounting Theory
133

if accounting methods do not affect taxes they do not affect firm value. In that
134
The Accounting Review, January 1990

market prices; within the firm alternative mechanisms such as standard costs are used (Ball 1989). Which productive
activities are carried out by markets and which by firms depends on which arrangement is cost effective.3 In competition
among firms, those that organize themselves to minimize contracting costs are more likely to survive (Fama and Jensen
1983a, 1983b). It was a short step to suggest that accounting methods affect the firm's organizational costs and so the
accounting methods that survive are the result of a similar economic equilibrium (Watts 1974, 1977).* Accounting researchers
have recently returned to using that notion of an efficient set of accounting methods to explain accounting choice (Zimmer 1986).

As noted above, the agency costs associated with debt and management compensation contracts and the agency,
information, and other contracting costs associated with the political process provided the hypotheses tested in the early
empirical accounting choice studies (bonus plan, debt/equity, and political cost hypotheses). However, the more general
approach suggested agency and other costs associated with other contracts (e.g., sales contracts) could also affect accounting
choice. This potential for many contracts to play a role in explaining organizational choice (including accounting choice) and the
fact that agency costs used to explain the contracts often arise in contractual scenarios that differ from those of the standard agency
problem led researchers to start to use the term "contracting costs" instead of agency costs (Klein 1983; Smith 1980). The concept
of contracting costs and the notion of accounting methods as part of effi- cient organizational technology play key roles in
contemporaneous positive accounting theory.

Contemporaneous Positive Accounting Theory

Contracting costs arise in (1) market transactions (e.g., selling new debt or equity requires legal fees and
underwriting costs), (2) transactions internal to the firm (e.g., a cost-based transfer price scheme is costly to maintain and can
pro- duce dysfunctional decisions), and (3) transactions in the political process (e.g., securing government contracts or
avoiding government regulation requires lobbying costs). Contracting costs consist of transaction costs (e.g., brokerage

3 Coase (1937) suggests that economies of scale in long-term contracting are what cause activity to be organized in firms. Alchian and
Demsetz (1972) point out that those economies are not sufficient since market arrangements could achieve the same economies (e.g., contracting consultants).
What is necessary is some unique advantage of firm organization over market arrangements. Alchian and Demsetz suggest it is the advantage firms have
in metering inputs to team production that generates firms. Monitors meter individual inputs and the monitors' incentive problem is solved by
giving them the residual claim to the firm (hence, the firm structure), Klein et al. (1978) suggest firms emerge to solve post contractual opportunism
associated with specialized assets. Meckling and Jensen (1986) suggest that firms have an advantage in generating information by aggregating
data and using that in- formation. Difficulties in capturing the information's benefits in the market result in the firm being the optimal form of organization.
* Watts adopted such a view in "Accounting Objectives" which he presented to the Annual Con- gress of the N.S.W. branch of the Institute
of Chartered Accountants in Australia in 1974. The paper was later substantially revised given Jensen and Meckling (1976) and joint work with Zimmerman
and published in Watts (1977).
* The influence of sales contracts on accounting choice is considered by Watts and Zimmerman (1986, 207) and by Zimmer (1986) and
joint venture contracts by Zimmer (1986). Further, Ball (1989) suggests intrafirm transactions affect internal accounting choice (e.g., the basis for transfer
prices).
Watts and Zimmerman-Positive Accounting Theory
135

fees), agency costs (e.g., monitoring costs, bonding costs, and the residual loss from dysfunctional decisions),
information costs (e.g., the costs of becoming in- formed), renegotiation costs (e.g., the costs of rewriting
existing contracts be- cause the extant contract is made obsolete by some unforeseen event), and bank- ruptcy
costs (e.g., the legal costs of bankruptcy and the costs of dysfunctional decisions). Throughout this paper,
we use the term "contracting costs" to incor- porate this wide variety of costs. The term "contracting parties"
is meant to include all parties to the firm including “internal" employees and managers and "external" parties, such
as suppliers, claim holders, and customers."
The existence of contracting costs is crucial to models of both the organiza- tion of the firm and accounting choice.
Meckling and Jensen (1986) suggest that within the firm the lack of a market price is replaced by systems for
allocating decisions among managers, and measuring, rewarding, and punishing manage- rial performance.
Accounting plays a role in these systems and so appears to be part of the firm's efficient contracting technology.
Trying to predict and explain the organization of the firm with zero contracting costs is pointless (Coase 1937; Ball
1989). How the firm is organized, its financial policy, and its accounting methods, are as much a part of the
technology used to produce the firm's product as are its production methods. Hence, modelling accounting choice
while assuming zero contracting costs is not productive.
The extent to which accounting choice affects the contracting parties' wealth depends on the relative
magnitudes of the contracting costs. For example, assume accounting-based debt agreements have higher
renegotiation costs than accounting-based bonus plans. Then, mandatory changes in accounting proce dures by
the FASB impose greater relative costs on firms with debt agreements than on firms with bonus plans,
ceteris paribus. And, firms with debt agree- ments will conduct more lobbying and undertake more (costly)
accounting, financing, and production changes to undo the effects of the mandatory change than firms with only
bonus plans. Thus, developing a positive theory of account- ing choice requires an understanding of the relative
magnitudes of the various types of contracting costs.
Contracts that use accounting numbers are not effective in aligning man- agers' and contracting parties' interests if
managers have complete discretion over the reported accounting numbers. If managers know (or can
determine) which accounting methods best motivate subordinates, then the contracting parties want managers to
have some discretion over the accounting numbers. Hence, we expect some restrictions on managers' discretion over
accounting numbers, but some discretion will remain. When managers exercise this discretion it can be because (1) the
exercised discretion increases the wealth of all contracting parties, or (2) the exercised discretion makes the
manager better off at the expense of some other contracting party or parties. If managers elect to exercise
discretion to their advantage ex post, and the discretion has wealth re- distributive effects among the
 contracting parties, then we say the managers acted "opportunistically."

* See Watts (1974) for an earlier and Ball (1989) for a later discussion of contracting parties other than capital suppliers and managers.
136
The Accounting Review, January 1990

Ex ante, the set of accounting choices restricted by the contracting parties is determined by "efficiency" reasons (to
maximize firm value). One cost of allow- ing managers more rather than less discretion is the increased likelihood of some ex post
managerial "opportunism" (i.e., wealth transfers to managers) via ac- counting procedures. However, ex ante the contracting parties
expect some redistributive effects and reduce the price they pay for their claims. Ex post, wealth is redistributed by managerial
opportunism, but ex ante some redistribu- tion was expected and the parties price protected themselves. Price protection does
not eliminate the incentive to act opportunistically nor does price protection eliminate the dead weight costs of managers taking
opportunistic actions. The extent to which contracts can be written ex ante to preclude such ex post behavior that causes dead
weight costs increases the chance the firm will survive in a competitive environment (Klein 1983, fn. 2).
The set of accounting procedures within which managers have discretion is called the "accepted set." It is
voluntarily determined by the contracting parties. Managerial discretion over accounting method choice (i.e., the “accepted set") is
predicted to vary across firms with the variation in the costs and benefits of re- strictions. These restrictions produce the "best" or "accepted”
accounting prin- ciples even without mandated accounting standards by government. The restric- tions are enforced by external
auditors. Reacting to the incentive of managers to exercise accounting discretion opportunistically, the accepted set includes
“conservative” (e.g., lower of cost or market) and “objective” (e.g., verifiable) accounting procedures (Watts and Zimmerman
1986, 205-206).
Figure 1 represents the concept of the “accepted set" of accounting methods as a Venn diagram. Al denotes the
accepted set of methods for firm 1. Ex ante, the accepted set is determined jointly by the contracting parties to maximize the value
of the firm (e.g., set A1 vs. A2 in Fig. 1). Managers have discretion to choose any method within the accepted set (e.g., X1). Also,
managers in firm 2 are con- strained ex ante to the set A2 and choose X2 ex post. For example, within the accepted set of
procedures used for bonus plans managers might select the method that maximizes their utility, even if it comes at another
contracting party's expense. Managers' ex post choice can either increase the wealth of all contracting parties or redistribute
wealth among the parties. Empirically, it is difficult to separate ex ante from ex post. Contracts are continually being written,
rewritten, and revised.
Variations across sets of accepted accounting procedures (e.g., A1 and A2 in Fig. 1) explain some cross-sectional
variation in accounting choice (e.g., man- agers in firm 2 cannot choose method X1). For example, Zimmer (1986) argues
Australian real estate development firms are restricted by accepted practice from capitalizing interest except for cost plus
contracts that allow interest as a cost. His evidence is consistent with that hypothesis.
Most accounting choice studies assume managers choose accounting methods to transfer wealth to themselves at the
expense of another party to the firm because they can take the firm's observed contracts as given and then deter- mine managers'
incentives for accounting choice. Some research studies assume accounting methods are chosen for efficiency reasons (i.e., they
increase the pie available being shared among all parties to the firm (Watts 1974, 1977; Leftwich
Watts and Zimmerman-Positive Accounting Theory

Figure 1
Relation Between the Accepted Set of Accounting Methods and the Choice of Method from within the
Accepted Set
 XI
All Feasible Accounting Methods

Al

X2
A2

Al denotes the set of accepted methods for firm 1 A2 denotes the set of accepted methods for firm 2
X1 denotes the choice of method from within the accepted set by firm 1 X2 denotes the choice of method from within the
accepted set by firm 2
137
 et al. 1981; Zimmer 1986; Whittred 1987; Ball 1989; Malmquist forthcoming: Mian and Smith forthcoming). However, no
study to date has explained both the ex ante choice of the accepted set and the ex post choice of accounting method from within
the accepted set. Most studies that assume opportunistic choice of accounting methods do not control for the fact that managers
in different firms likely are choosing accounting methods from different constrained accepted sets.
The accepted set of accounting methods is one part of the firm's implicit and explicit contracts including the firm's
capital structure, compensation plans, and ownership structure. All the contracting provisions (including the accounting policies)
are endogenous. Capital structure choice is related to compensation policy and to accounting policy. But, the relation is not
necessarily causal. Capital structure changes do not cause changes in the accepted set of accounting methods. Rather,
some exogenous event, such as a new invention or government deregulation occurs and this causes changes in the
contracting variables includ- ing accounting methods (Ball 1972; Smith and Watts 1986).
138

Evidence on the Theory

The Accounting Review, January 1990

Two types of tests of the theory have been conducted: stock price tests and accounting choice tests.
The stock price tests have been reviewed extensively elsewhere (Foster 1980; Ricks 1982; Holthausen and Leftwich 1983; Lev
and Ohlson 1982; Watts and Zimmerman 1986; Bernard 1989). Stock price tests of the theory reveal some price reactions to
mandatory accounting changes, espe- cially involving oil and gas accounting (Lys 1984).' Stock price studies are prob- ably
relatively weak tests of the theory (Watts and Zimmerman 1986). The more promising ones are accounting choice studies.
Most accounting choice studies attempt to explain the choice of a single ac- counting method (e.g., the choice of
depreciation) instead of the choice of combinations of accounting methods. Focusing on a single accounting method reduces
the power of the tests since managers are concerned with how the com- bination of methods affects earnings instead of the effect
on just one particular accounting method (Zmijewski and Hagerman 1981). Some studies seek to ex- plain accounting accruals
(the difference between operating cash flows and earnings). Accounting accruals aggregate into a single measure the net effect of
all accounting choices (Healy 1985; DeAngelo 1986, 1988a; Liberty and Zimmerman 1986). But use of accruals as a
summary measure of accounting choice suffers from a lack of control of what accruals would be without manage- rial
accounting discretion.
Most accounting choice studies use combinations of three sets of variables: variables representing the manager's incentives to
choose accounting methods under bonus plans, debt contracts, and the political process. Bonus plan and debt contract variables
are used because they're observable. The three particular hypotheses most frequently tested are the bonus plan hypothesis,
the debt/ equity hypothesis, and the political cost hypothesis. The literature has tended to state each of these hypotheses as
managers behaving opportunistically. The bonus plan hypothesis is that managers of firms with bonus plans are more
likely to use accounting methods that increase current period reported income. Such selection will presumably increase the
present value of bonuses if the compensa- tion committee of the board of directors does not adjust for the method chosen. The
choice studies to date find results generally consistent with the bonus plan hypothesis (Watts and Zimmerman 1986, chap. 11;
Christie forthcoming).

7 Using Lys' own calculations, Frost and Bernard (1989, 20) and Bernard (1989, 14) conclude Lys' evidence is inconsistent with a link
between stock price reactions to mandated oil and gas accounting and the violation of debt covenants. However, that conclusion is unwarranted. Lys estimates
the aver- age cost of violations as 2.5 percent of the stock value, the same order of magnitude as the stock price reactions observed. Frost and
Bernard argue that given an average cost of violation of 2.5 percent, the average stock price reaction should be much less since according to Foster (1980)
very few firms have a debt covenant violation as a result of the mandated accounting change. There are at least three prob- lems with the Frost and Bernard
argument. First, the Lys point estimates are likely to have large stan. dard errors. Second, to obtain an estimate of the stock price reaction, the estimated cost
of a violation has to be weighted not by the relative frequency of violation but by the change in the likelihood of vio- lation. While few firms
violated covenants, many firms' probability of violation likely increased sub- stantially. Third, Malmquist (forthcoming) suggests Foster's description of oil and
gas firms' covenants is incorrect. Frost and Bernard (1989) also use their own empirical study's results to argue that there is no link between the stock price
reaction and debt covenants. Because of selection biases, however, their study provides little evidence on the Issue (Begley forthcoming).
 Watts and Zimmerman-Positive Accounting Theory
139

The early tests of the bonus hypothesis are not very powerful tests of the theory because they rely on
simplifications of the theory that are not appropriate in many cases. For example, a bonus plan does not
always give managers incen- tives to increase earnings. If, in the absence of accounting changes,
earnings are below the minimum level required for payment of a bonus, managers have incen-
tive to reduce earnings this year because no bonuses are likely paid. Taking such an "earnings bath"
increases expected profits and bonuses in future years. By using bonus plan details to identify
situations where managers are expected to reduce earnings, Healy's (1985) tests encompass more kinds
of manipulation. His results are consistent with managers manipulating net accruals to affect their
bonuses.

The debt/equity hypothesis predicts the higher the firm's debt/equity ratio, the more likely
managers use accounting methods that increase income. The higher the debt/equity ratio, the closer
(i.e., "tighter") the firm is to the con- straints in the debt covenants (Kalay 1982). The tighter the covenant
constraint, the greater the probability of a covenant violation and of incurring costs from technical default.
Managers exercising discretion by choosing income increasing accounting methods relax debt constraints
and reduce the costs of technical default.
8

The evidence is generally consistent with the debt/equity hypothesis. The higher firms' debt/equity
ratios, the more likely managers choose income increasing methods. Press and Weintrop
(forthcoming) and Duke and Hunt (forthcoming) find that debt/equity ratios are correlated with
closeness to bond covenants as assumed in the debt/equity hypothesis. Some studies, however, have
avoided using the debt/equity ratio as a proxy variable for closeness to the covenant constraint by using
more direct tests. For example, Bowen et al. (1981) examine whether accounting choice varies with
tightness of the dividend con- straint as specified in the debt covenant and measured by "unrestricted
retained earnings." The association between leverage and accounting method choice is an empirical
regularity unknown prior to the positive accounting studies.
The political cost hypothesis predicts that large firms rather than small firms are more likely to use
accounting choices that reduce reported profits. Size is a proxy variable for political attention.
Underlying this hypothesis is the assump tion that it is costly for individuals to become informed about
whether account- ing profits really represent monopoly profits and to "contract" with others in the
political process to enact laws and regulations that enhance their welfare. Thus, rational individuals are
less than fully informed. The political process is no differ- ent from the market process in that respect. Given
the cost of information and monitoring, managers have incentive to exercise discretion over accounting
profits and the parties in the political process settle for a rational amount of ex post opportunism.

* Holthausen (1981) and Healy (1985) fail to reject the null hypothesis of no association between leverage and accounting
method choice (see Christie forthcoming, table 1).
Researchers are beginning to distinguish between how close the firm is to a given covenant con- straint versus the
existence of the covenant. For example, Press and Weintrop (forthcoming) find the existence of a covenant has additional
explanatory power in a model predicting accounting choice after including a leverage variable.
140
The Accounting Review, January 1990

The evidence is consistent with the political cost hypothesis. However, the result only appears to hold for the
largest firms (Zmijewski and Hagerman 1981) and is driven by the oil and gas industry (Zimmerman 1983). Difficulties with
using firm size to proxy for political costs, including the likelihood that it can proxy for many other effects, such as industry
membership, are discussed in Ball and Foster (1982). The interesting finding is the consistency of the sign of the relation
 between size and accounting choice across a variety of studies. The largest firms tend to use income decreasing
accounting methods. Presently, there is no alternative theory for the empirical regularity between firm size and
accounting choice other than the political cost hypothesis.
Bonus plan, debt contract, and political process variables other than bonus plan existence, leverage, and size
have also been found to be associated with accounting choice. Christie (forthcoming) aggregates test statistics across the various
studies and concludes “... six variables common to more than one study have explanatory power. These variables are managerial
compensation, leverage, size, risk, and interest coverage and dividend constraints. Another con- clusion is that the
posterior probability that the theory taken as a whole has explanatory power is close to one."
While bonus, debt, and political process variables tend to be statistically significant (p-values smaller than .10), in
many studies the explanatory power (R2) of the models is low. In Zmijewski and Hagerman (1981), the model of cross-
sectional choice of accounting methods is not significantly better than picking the most common combination, although
Press and Weintrop (forthcoming) achieve slightly improved explanatory power. The alternative predictive model is that
each firm uses the most common combination of accounting methods, a model with little explanatory appeal. The
alternative model begs the ques- tion of what determines the majority accounting choice. Many accounting teachers
would be uncomfortable with the explanation that managers choose their accounting procedures based on what most
other firms are doing. The real issue is the lack of an alternative model with greater explanatory power, not the low
explanatory power of the extant theory. Several problems with the existing research methods contribute to the low
explanatory power. These are discussed next.

II. Criticisms of Positive Accounting Research

Table 1 lists most of the published papers with critical comments on our 1978 and 1979 papers. The second
and third columns list the number of explicit references made by the authors to our 1978 and 1979 papers. These
columns in- dicate which of the two papers is the primary focus of the article. The fourth col- umn lists the general topic
of the paper and the fifth column lists the major criti- cisms raised in the paper.
The criticisms in Table 1 can be dichotomized into two mutually exclusive sets: those concerning research methods
(including the inferences drawn) and those concerning methodology (including the philosophy of science). For example, Ball
and Foster (1982), Holthausen and Leftwich (1983), and McKee et
Table 1
Summary of Papers Reviewing Watts and Zimmerman (1978 and 1979)
Number of References
Authors
WZ
(1978)
WZ
(1979)
Ball and Foster (1982)
13
1
Tinker et al. (1982)
1
Watts and Zimmerman-Positive Accounting Theory
Topic
Review of
Empirical
Accounting Research
Positive versus normative
theories
Methodology of Positive Accounting
Review of
"Economic
Consequences Literature"
WZ (1979)
Major Criticisms
• Firm size and bonus plans can proxy for omitted
 variables
• Weak theoretical underpinning for size-political cost
construct
• Holdout sample not used
• Positive theories are value-laden and mask a conservative blas
• Ignores underlying class struggles
• Logical Positivism is an obsolete methodolog- ical
approach
• Approach is a "sociology of accounting" in- stead of
accounting theory
Tests introduce ad hoc arguments to excuse the
exceptions to the theory
• Inappropriate methods are used for construct- ing
explanatory theories
Interpretation of results limited because:
• Incomplete political and contracting theories
• Specification problems in left-hand-side and right-
hand-side variables
• Economic framework is unjustified Positive approach open to
dispute Nature of proof is unscientific Contrary evidence
presented
Christenson (1983)
6
9
Holthausen and Leftwich (1983)
Lowe et al. (1983)
O
12
141
Authors
McKee et al. (1984)
Number of References
TABLE 1-Continued
WZ
(1978)
WZ
(1979)
0
Whittington (1987)
0
7
Hines (1988)
4
0
142

12
Topic
Replication of WZ (1978)
Review of WZ
(1986)
Christenson
(1983) and Methodology
Major Criticisms
• Results do not hold in a new sample
• Holdout sample not used
• Foreknowledge of sample proportions biases parameter
estimates
Presentation of arguments and evidence is unbalanced
Extreme methodological stance
Positive theories are value-laden
• Approach is a "sociology of accounting" in- stead of
accounting theory
 • Popper is not a practical evaluative guideline for
empirical accounting research
The Accounting Review, January
Watts and Zimmerman-Positive Accounting Theory
143

al. (1984)1o discuss research methods problems and not philosophy of science issues. The remaining authors
concentrate on philosophy of science issues to the near exclusion of problems with research methods. Except for
Holthausen and Leftwich (1983), all the reviews of positive accounting ignore the accumulating body of
evidence consistent with the theory. For example, Hines (1988) cites McKee et al. (1984) as contradictory evidence to
Watts and Zimmerman (1978). Yet, she ignores 21 studies reviewed in Watts and Zimmerman (1986, chaps. 11, 12)
and Christie (forthcoming) that present evidence generally consistent with the theory.
The research method issues are important and future research must attempt to address them. However,
it is unlikely that the positive accounting literature or any other empirical literature will ever totally eliminate such
issues. We do not agree with many of the philosophy of science issues raised and seek to eliminate the
common misconceptions they reflect. Research method issues, some raised by others and some by us, are discussed
first in this section and philosophy of science issues are discussed second.

Research Method Issues

The first research method issue involves the tests' lack of power. The second issue involves the
possibility that the results obtained in the positive accounting literature are due to unrecognized alternative hypotheses,
not the stated hy- potheses.
Reductions in the tests' power. Tests of the theory lack power for several reasons: problems with model
specification, problems specifying the left-hand- side and right-hand-side variables, and omitted variables. Each of these are
dis- cussed next.
Model specification. All the studies to date have assumed accounting choice results either from efficiency
reasons or managerial opportunism. This produces two model specification errors. First, in probit type regressions where
the choice of accounting method depends on the effect of the choice on the manager's wealth, the right-hand-side or
explanatory variables reflect the wealth effects of the choice via compensation plans, debt agreements, and the
political process. Implicitly researchers are holding constant the firm's investment opportunity set and contracts and
interpret the compensation plan variable as managerial oppor- tunism. But, the debt and political variables can represent
both efficiency and opportunism. Thus, the model is misspecified. The second specification error results from ignoring the
interaction effects among the right-hand-side variables. Higher earnings impose political costs and so reduce the
size of the pie for the contracting parties and at the same time increase the manager's bonus compen- sation. The
manager's increased share of the smaller pie might be larger than a smaller share of the larger pie. The bonus plan
and political process effects inter-

10 McKee et al. (1984) discuss problems of the tests in Watts and Zimmerman (1978), extend the tests to another sample of firms,
and offer some statistical refinements. The only satistically significant explanatory variable in our 1978 paper was firm size. McKee et al. find that
their refined measure of firm size, (SALES/MAXSALES)DTREND, is statistically significant in both our sample and their sam- ple and remains
statistically significant after various refinements are made. They do not discuss the importance of this finding.
144
The Accounting Review, January 1990

act. However, in the empirical models the right-hand-side variables are treated as additive and interaction effects
are ignored. Solving these two specification prob- lems requires researchers to specify the intertemporal interaction
between opportunism (including managerial reputation incentives) and efficiency effects (see Christie 1987).
Left-hand-side variable. Problems specifying the accounting choice variable reduce the power of
the tests. One such problem mentioned earlier is the use of single method choices as the left-hand-side variable.
Zmijewski and Hagerman (1981) and Press and Weintrop (forthcoming) use sets of accounting methods and still
achieve relatively low explanatory power. However, ranking the effects of various portfolios of accounting methods on
earnings requires assumptions about the relative effects on earnings of the various accounting choices (e.g.,
the effect of depreciation choice vs. inventory choice). These assumptions induce error in the left-hand-side variable. Healy
(1985) tries to overcome this problem by using net accruals as his left-hand-side variable. But, the variable "net accruals"
is a noisy measure of the net accruals manipulated by managers. Some accounting decisions that affect accruals
have been made earlier and are prob- ably beyond the manager's discretion at the time of the measurement. Ideally, net
accruals should be measured relative to what they would be without manipu- lation, so these variations are excluded from
the left-hand-side variable. This re- quires a model of accruals that currently does not exist (Moyer 1988; McNichols and
Wilson 1988; DeAngelo 1988b).
Right-hand-side variables. Some variables in accounting choice studies are mismeasured. For
example, both the closeness to the covenant (i.e., the differ- ence between the number specified in the covenant and
the actual number) and the existence of the covenant are likely important determinants of accounting choice. But the
debt/equity ratio by itself is an imprecise measure of both close- ness to the constraint and the existence of a constraint. Also,
the use of a zero-one variable to measure a bonus plan effect is simplistic. Ball and Foster (1982, 184) point out that other
components of pay, such as salary, can depend on account- ing earnings without a formal compensation plan and
that even with a formal accounting-based plan the outside directors can adjust the incentive pay for accounting changes.
However, finding an association between an indicator vari- able representing a bonus plan and choice of accounting
methods is informative and suggests that further research with more refined measures based on the bonus plans'
details will yield stronger results than the zero-one variable. Also, more direct measures of political sensitivity than
firm size (Wong 1988; Jones 1988; Sutton 1988) provide more powerful tests of the political cost hypothesis.
Omitted variables. There are three different omitted variable problems in the current literature: omitting
standard accounting-based contracts, omitting less standard contracts, and omitting variables representing the
accepted set. First, contracting cost variables for standard contracts, such as bonus plans occasionally are omitted
because such variables are costly to collect. For exam- ple, Daley and Vigeland (1983) omit a variable representing
accounting-based management compensation plans from their regression. Because leverage, com- pensation
contracts, and accounting policy are part of the firm's efficient con- tracting technology, these variables covary and
also vary with firm size. Omitting
Watts and Zimmerman-Positive Accounting Theory
145

a right-hand-side variable correlated with included variables causes the existing right-hand-side variables to become
surrogates for the omitted variables. This produces biased coefficients of the estimated right-hand-side variables and hampers
their interpretation.
A second omitted variables problem is that to a large degree, the literature to date focuses only on debt and
compensation contracts. Other contracts influence management's choice of accounting methods, but these are omitted
in most tests. For example, the existence of a bonus plan is likely correlated with other organizational devices such as stock option
plans. These other organizational structures might be driving the accounting choice rather than bonus plans (Ball and Foster
1982, 185). And, it is incorrect to ascribe all the explanatory effect of the bonus plan indicator variable results to the bonus plan.
Corporate control issues also are often omitted as explanatory variables in seeking to explain ac- counting choices.
DeAngelo (1988a) finds that net accruals are more positive (i.e., higher reported earnings) during proxy fights.
Zimmerman (1979) and Ball (1989) argue that accounting numbers are part of the internal control process and, thus, affect
manager's choice of accounting methods (e.g., cost allocations). Ignoring these, other less frequently researched
informal contracts can produce biased coefficients.
Third, as discussed under specification problems above, the left-hand-side variable in most studies is the manager's
choice of accounting methods. Even without a government regulatory defined set of accounting methods, this choice is
made from within the "accepted set of methods" (see Fig. 1)." Yet, most studies do not control for differences across firms'
accepted sets. Such control re- quires a theory of how the sets of accepted accounting methods vary and such a theory does not
exist. Failure to control for differences in accepted sets induces another correlated omitted variables problem in the tests. The severity of
this correlated omitted variables (and model specification) problem is likely to be larger in studies in which the sampled firms are
drawn from several industries than in studies where the sampled firms are drawn from the same industry.
Alternative hypotheses. Alternative hypotheses can explain the bonus, debt/equity, and size results found
in the positive accounting literature. Several scenarios illustrate how this problem might arise:
1. If the accounting system is part of the firm's efficient set of implicit and explicit contracts, accounting choice is
endogenous. Contracting, invest- ment, and production decisions are determined jointly. The type of con- tracts
used (including the accounting methods) depends on the firm's investment opportunity set. Hence, the firm's investment
opportunity set (e.g., whether it includes growth options or not) is correlated with the firm's financial, dividend,
compensation, and accounting policies. Smith and Watts (1986) find significant cross-sectional correlations among

"Mian and Smith (forthcoming) find that accounting policy decisions regarding consolidations vary by type of organization structure.
Consumer finance subsidiaries are more prevalent where the parent is in the financial services industry and choice of consolidation is more
homogeneous within like organization structures than in dissimilar structures. Also, operating interdependencies between the parent and subsidiary drive
some accounting choices.
146
The Accounting Review, January 1990

firms' investment opportunity sets, financial policies, dividend policies, and compensation policies. The
documented correlations between debt/ equity and accounting choice and between bonus plans and
accounting choice could be due to the correlation between financial and compensa- tion policies and the
optimal set of accounting procedures for contracting. Most researchers, however, interpret these associations
as resulting from opportunistic actions by managers and have not considered efficiency- based hypotheses.
2. Accounting choice also is endogenous in the political process. The poten- tial costs of a proposed
accounting standard affect the standard before it is released. The correlation between financial and compensation
policies and accounting policy is likely affected by the firm's tax accounting policies. While some financial
accounting method choices do not affect taxes, reducing bookkeeping costs by keeping one set of books and the
possibility that tax audits or future taxes might be levied using reported income induce a relation between financial
accounting and tax account- ing methods. 12
One cannot test claims that variables like debt/equity and size are surrogates for alternative explanations until
those alternatives are identified and the relation specified. Given the investment opportunity set and taxes are identified as
possi- ble explanatory variables, future research can investigate their implications as alternative hypotheses to those
currently advanced. For example, changing accounting methods can result from a change in the firm's investment oppor-
tunity set causing the efficient contracts and accounting methods to change. Or, some exogenous event occurs
(such as reduced demand for the firm's products) and managers take opportunistic actions to undo the adverse
compensation and debt contract effects of the exogenous event. Accounting changes likely are due to both efficiency
reasons and managerial opportunism. Probing the relative im- portance of efficiency and opportunism for accounting
method changes requires more refined theories and more linkage between the theory and the tests. Philosophy of
Science Issues
Positive theories are value-laden. Tinker et al. (1982, 167) argue that all re- search is value-laden and not socially
neutral. Specifically, "Realism, operating in the clothes of positive theory, claims theoretical supremacy because it is
born of fact, not values" (p. 172). We concede the importance of values in determining research; both the researcher's
and user's preferences affect the process.
Competition among theories to meet users' demands constrains the extent to which researcher values influence
 research design. Positive theories are “If ..., then..." propositions that are both predictive and explanatory. Researchers
choose the topics to investigate, the methods to use, and the assumptions to make. Researchers' preferences and
expected payoffs (publications and citations) affect their choice of topics, methods, and assumptions. In this sense,
all re-

12 The Corporate Alternative Minimum Tax under the Tax Reform Act of 1986 requires a portion of reported income be in
the tax base. This act increases the tax incentives on financial reporting. Re- search to date has not documented the effect of 1986
tax reform on financial reporting incentives.
Watts and Zimmerman-Positive Accounting Theory
147

search, including positive research is "value-laden." The usefulness of positive theories depends on their predictive and
explanatory power and on the user's preferences or objective function. To the extent that the researcher's values inter- fere
with the theory's ability to predict and explain, the theory's usefulness is reduced.
Approach is a “sociology of accounting" instead of accounting theory. Christenson (1983, 5) writes, "The program of the Rochester
School is concerned with describing, predicting, and explaining the behavior of accountants and managers, not that of
accounting entities." His definition of an "accounting en- tity" is "A business enterprise or other economic unit, or any subdivision
thereof for which a system of accounts is maintained" (Kohler 1975, 14). Christenson (1983, 6) supports his criticism with an analogy from
the physical sciences, “Chemical theory consists of propositions about the behavior of chemical entities (molecules and atoms) not
about the behavior of chemists." In chemistry, chemical reactions exist without chemists and one can study reactions without studying
chemists. But, there would be no accounting without accountants, managers, or preparers of the numbers; there would be no
numbers or systems to investigate because people “maintain” the system (Lavoie 1989). Analogously, there would be no study
of political science if politicians and voters were ignored. The study of accounting (or political science) is a social science
(Christenson 1983, fn. 5). An accounting theory that seeks to explain and predict accounting cannot divorce accounting research
from the study of people. The contracting approach to studying accounting requires researchers to understand the incen-
tives of the contracting parties.
Inappropriate methods are used for constructing explanatory theories. We apply traditional, generally accepted
research methods and methodology from accounting, finance, and economics. Christenson (1983, 6) states, "The Roches- ter
School has drawn its concept of 'positive theory' from that guru of the Chicago School of Economics, Milton Friedman.""
Whittington (1987, 331) states, "... Watts and Zimmerman are not unique in owing intellectual alle- giance to the Chicago view. . . .
The majority of North American empirical ac- counting researchers would fall into this category, and their collective achieve-
ments are formidable.”
The economic approach we and many others use applies a simple proposi- tion: To predict and explain
individual behavior, people (including accountants, regulators, and researchers) consider the private costs and benefits
(broadly de- fined) of an action and choose the action if the benefits exceed the costs. This economics-based research methodology may be
fundamentally flawed in ways we do not now understand. But, accounting research using this methodology has produced
useful predictions of how the world works (e.g., association between earnings and stock prices, random walk model of earnings,
contracting and size variables associated with accounting choice). A methodology that yields useful results should not be
abandoned purely because it may not predict all human

13 Christenson is referring to Milton Friedman's views on scientific methodology as expounded in Friedman (1953). In our opinion,
Friedman places too much emphasis on prediction vis-à-vis explana- tion.
148
The Accounting Review, January 1990

behavior. Do we discard something that works in some situations because it may not work in every circumstance? Despite what
the critics think methodology should be, the methodologies that survive are the ones that produce useful theories.
Competition in the marketplace of ideas will produce future research that uncover the errors of our present ways.
Time will tell whether our approach is inappropriate.
Choice of the term "Positive Accounting Theory." Positive accounting re- search existed long before the
publication of our 1978 and 1979 papers. Early examples include Gordon (1964), Gordon et al. (1966), and Gagnon
(1967). We applied the label "positive" to a set of existing research studies. The prime rea- son we attached this
adjective in "Towards a Positive Theory of the Determina- tion of Accounting Standards" was to emphasize
that accounting theory's role is to provide explanations and predictions for accounting practice.
In Watts and Zimmerman (1986, 2) we state the objective of an accounting theory is to explain and predict
accounting practice. Neither prediction nor ex- planation is preeminent. We adopted the label "positive" from economics
where it was used to distinguish research aimed at explanation and prediction from research whose objective was
prescription. Given the connotation already attached to the term in economics we thought it would be useful in distinguishing
accounting research aimed at understanding accounting from research directed at generating prescriptions. In the
1960s researchers were still debating various normative theories of accounting (Chambers 1966; Sprouse and Moonitz
1962).
Our use of the term "positive" differentiated our and other people's (positive) research from traditional
normative theories by emphasizing the importance of prediction and explanation. It helped place normative theories and their
role in a clearer perspective. Our work was not directly related to the debate over alterna- tive normative theories and we wanted to
differentiate our work from that debate. The phrase "positive” created a trademark and like all trademarks it conveys information.
“Coke,” “Kodak,” “Levis” convey information. A positive theory differs from a normative theory, though a positive theory
can have normative implications once an objective function is specified (Jensen 1983).
In retrospect, the term "positive" generated more confusion than we antici- pated. For example, some
thought we meant logical positivism (Christenson 1983). We merely intended to distinguish positive propositions from the
extant normative propositions in the literature. While the term "positive" avoided de- bates over normative uses of the
work, the term "positive" generated consider- able debate over philosophical issues.
Despite its problems, we prefer "positive accounting literature" to alterna- tive terms that have arisen,
particularly the term "economic consequences liter- ature." This latter term suggests accounting standards are
decided on some higher basis and that economic consequences are a secondary factor only consid- ered after the initial
decision is made on the higher basis.14

11 Some have suggested the term "contracting theory." While descriptive of most of the elements in the existing
theory, it seems to preclude noncontractual variables that might be discovered later (e.g.. taxes or information for the capital markets,
Holthausen forthcoming).
Watts and Zimmerman-Positive Accounting Theory
149

Debate over methodology. Several papers listed in Table 1 involve a debate over what constitutes
"proper" methodology (Tinker et al. 1982; Christenson 1983; Lowe et al. 1983; Whittington 1987; Hines
1988). For example, Christen- son (1983, 1) concludes, ". . . [T]he standards advocated by the Rochester
School for appraisal of their own theories are so weak that those theories fail to satisfy Popper's (1959)
proposal for demarking science from metaphysics." Hines (1988) then criticizes Christenson for relying on
Popper (1959) which later philosophers of science have questioned. Hines (1988, 658) argues these
methodology issues are important and if ignored will "harmfully limit the nature and domain of ac- counting
research."
The methodology criticisms have failed the market test because they have had little influence on
accounting research. Researchers have not changed their approach. Referees and editors of journals
have not asked researchers to alter their methodology based on these published critiques. There are at least
three reasons these criticisms have had little effect on published research. First, the criticisms are written in an
abstract fashion. Instead of just criticizing extant papers, if the critics would repeat studies without making the
 alleged errors, then users of the corrected research would demand such procedures be followed in the future. If
the alleged errors are important to users, then other researchers, edi- tors, and referees would adopt the
suggestions. Second, critics who place unrea- sonable demands on studies cause other researchers to
disregard their com- plaints. For example, Hines (1988, 661) argues that Watts and Zimmerman (1978) should
have: (1) avoided crude proxies, (2) avoided unrealistic assump- tions, (3) investigated the anomalies, (4)
clarified their theories, and (5) rigorously tested their theories against competing hypotheses. All these
standards are rele- vant, but if all were applied rigorously to individual papers (especially early papers
in an area of thought), no research would be published. Third, to most researchers, debating methodology is a
"no win" situation because each side argues from a different paradigm with different rules and no common
ground. Our reason for replying here is that some have mistaken our lack of response as tacit acceptance of
the criticisms.

III. Summary and Conclusions

Our prime objective in this paper is to provide a perspective on our 1978 and 1979 Accounting
Review papers. The 1978 paper has proven more important than the "Excuses" paper. Based on citations,
the 1978 paper has received over three times as many citations as the 1979 paper (Brown and Gardner 1985,
97). The 1978 paper was a catalyst for research into the choice of accounting meth- ods. Except for
generating debates over methodology, the 1979 “Excuses" paper has remained outside the mainstream
of accounting research probably because of the more subjective type of evidence necessary to test theories
of the effect of accounting research on policy.
The debate over methodology has been less useful than the discovery and explanation of empirical
regularities. The positive accounting literature has dis- covered several empirical regularities in accounting
choice and provided an explanation for them. Critics of the 1978 and 1979 papers raise issues involving
150
The Accounting Review, January 1990

research methods and philosophy of science. The methodology we and the sub- sequent literature use is the methodology of
economics, finance, and science generally. This methodology has been successful in accounting and we feel no
necessity to apologize for it. Under this methodology, a theory is not discarded merely because of some inconsistent
observations. The best theory is determined in a competition to meet the demand from students and practitioners for theories
that explain and predict accounting choice. It is unlikely an accounting or a social science theory with perfect predictions
will ever exist. Researchers are influenced by their values. But, to the extent those researchers are competing to meet
student and practitioners' demand for theories, they have incentives to reduce that influence. Further, the careful dichotomy between
theory and prescription helps reduce that influence. Lastly, accounting is an activity carried out by people and one
cannot generate a theory that predicts and explains accounting phenomena by ignoring the incentives of the individuals
who account. In this final section we summarize the contributions made by this literature, our views on promising research
directions, and some conclusions. Positive Accounting Literature Contributions
Discovering systematic patterns in accounting choice outlined in the preced- ing sections and providing specific
explanations for the patterns are the litera- ture's major contributions. However, we believe the literature has made other contributions: it
provides an intuitively plausible framework for understanding accounting. A plausible framework is a useful pedagogy
for teaching accounting. The literature also encourages researchers to address accounting issues and emphasizes the central
role of contracting costs in accounting theory.
The literature explains why accounting is used and provides a framework for predicting accounting choices.
Choices are not made in terms of "better mea- surement" of some accounting construct, such as earnings. Choices are made in
terms of individual objectives and the effects of accounting methods on the achievement of those objectives. For example, some
accounting instructors teach that certain accounting methods (e.g., current cost) are better than others (e.g., historical
 cost). But, no explanation is offered why these "better" measures are not adopted. The positive accounting literature takes as given the
proposition that the accepted set maximizes the wealth of the contracting parties and then seeks to understand how wealth is
affected by specific accounting methods.
The literature's emphasis on predicting and explaining accounting phenom- ena encourages research that is
relevant to accounting. One of the first questions one pursuing this approach asks of a new model is whether it has any
relevance to predicting and explaining accounting practice.
Another contribution of the literature is to highlight the importance of con- tracting costs (including
information, agency, bankruptcy, and lobbying costs). Contracting costs have long been important in economics and
date to Coase (1937). Positive accounting research has more recently recognized the impor- tance of contracting costs to
explain accounting. In the late 1960s and 1970s, financial economists derived pricing models (capital asset pricing models,
option pricing models, arbitrage pricing models). These models were developed under assumptions of costless information
and such models explain why different
Watts and Zimmerman-Positive Accounting Theory
151

securities sell for different relative prices. Such models do not explain institu- tional differences,
such as open- and closed-end mutual funds. To explain such institutional differences requires assumptions of
costly information and con- tracting. Likewise, accounting would not exist without contracting costs and so it
is difficult to produce a theory that predicts and explains accounting without making assumptions about the
relative magnitudes of these costs. The central role of contracting costs highlighted by positive accounting
research makes it difficult to ignore these costs in accounting theories. It directs researchers' atten- tion to
the appropriate issues.

Future Research Directions

Section II discussed two major research methods issues: the lack of power of the tests and
alternative economic explanations for the empirical regularities. The following research suggestions focus on
these two issues. We believe these suggestions will be more fruitful in advancing the understanding of
accounting choice than "merely conducting more studies using existing formulations of the theory and
existing ways of measuring variables" (Christie forthcoming) (also see Holthausen and Leftwich 1983,
109-114).
First, the single most important task facing positive accounting researchers is improving the
linkage between the theory and empirical tests. The theory pre- dicts that the magnitude of debt
renegotiation costs will affect managers' choice of accounting methods and will set an upper bound on
the magnitude of the de- fault costs. To date, researchers have been unable to document the magnitude of
the costs imposed by a technical violation of a debt covenant or the magnitude of renegotiation costs
(Holthausen 1981; Leftwich 1981; Lys 1984; Leftwich forth- coming). Greater attention has to be placed on
developing a unified theory that incorporates both the ex ante efficient restrictions on the managers' accepted
set of accounting methods and the ex post exercise by managers of their discretion to choose accounting
methods from within the accepted set. The empirical tests can no longer assume accounting choice is made for
either efficiency or oppor- tunistic reasons. Both must be incorporated into the tests. Also, estimates of the
relative magnitudes of the various components of contracting costs can help to further refine the linkage
between the theory and tests by identifying those costs most influential in driving accounting choice.
Developing and testing alternative hypotheses for the existing empirical reg- ularities also will
enhance the linkage between the theory and the tests. Hypotheses can be developed to predict new empirical
regularities. Under the contracting approach, debt and compensation contracts are only some of the contracts that
affect firms' cash flows. Other (explicit and implicit) contracts can be used to develop new predictions (DeAngelo
1988a). Particularly promising is the effect of accounting procedures for internal control on external
reporting (Ball 1989). For example, Mian and Smith (forthcoming) find that the prevalence of consolidated
reporting of financing subsidiaries depends on the extent to which the subsidiary is interdependent with the
parent's main business. How the firm is organized internally (e.g., functionally or by product line), the type of
internal compensation systems, and the investment opportunity set are likely associated with the type of
internal accounting performance measurement systems. Inter-
152
The Accounting Review, January 1990

nal contracting parties may well turn out to be as important a determinant of external financial reporting as the
external contracting parties.
Finally, the political process can affect firms' cash flows other than via the simple political cost
hypotheses. More detailed specification of government regu- latory processes that rely on accounting
numbers can be used to develop new hypotheses and a tighter linkage between the theory and tests by suggesting
more precise proxy variables other than firm size (Sutton 1988; Wong 1988; Jones 1988).
Second, when accounting choice is cast as part of the efficient contracting technology, variables often
used to explain and predict accounting choice are endogenous. For example, changes in accounting procedures occur
simulta- neously with changes in the firm's investment opportunity set, its financial and compensation contracts,
its organizational structure, and even in its political en- vironment. Managers choose packages of accounting
policy, financial policy, and organizational structure (including performance evaluation and reward sys- tems).
Theoretical and empirical models have to be developed to sort out the endogeneity problems among the variables
and, thereby, increase the power of the tests. While this is no easy task, it seems essential to significant advances in
both the theories of the firm and of accounting.
Accounting numbers are used in different ways across industries. Besides the obvious regulatory uses
of accounting numbers in financial institutions and public utilities, differences in industries' opportunity sets are
likely to affect the accepted set of accounting methods. Two types of studies are likely to prove use- ful and
again increase the tests' power. First, studies investigating differences in investment opportunity sets (e.g., the relative
amount growth opportunities to assets in place, Myers 1977), accounting policies, organizational structures, and
financial policies across industries are likely to produce information useful for the modelling suggested in the
preceding paragraph. Second, intra-industry studies of accounting choice while requiring significant amounts
of industry-specific knowledge by the researcher, have the potential of generating useful insights about the
magnitude of contracting costs.
Third, measurement errors in net accruals can be reduced to increase the tests' power. This requires a model
of net accruals not subject to managerial ac- counting discretion (Kaplan 1985; McNichols and Wilson 1988;
DeAngelo 1988b; Moyer 1988). Also, replacing the simple indicator variables used to represent a bonus plan or an
accounting-based debt covenant with continuous variables that better measure the relative magnitudes of
various contracting costs will probably increase the theory's predictive power.
Conclusions

While the positive accounting literature has yielded empirical regularities and explanations for these
regularities, it is clear there are many research oppor- tunities available beyond those currently exploited. The tests of the
debt, bonus, and political cost hypotheses represent very limited exploration. Incorporating both ex ante contracting
efficiency incentives with ex post redistributive effects is likely to prove useful. Likewise, investigating the
implications of internal con- tracts and external contracts other than debt and bonus contracts is likely to be
Watts and Zimmerman-Positive Accounting Theory
153

productive. The major breakthroughs are likely to come from viewing account- ing as a choice that is
endogenous with the choice of organization, contracting, and financial structures. Such a breakthrough will be difficult to
 achieve, but important foundations can be laid by stressing the linkage between the theory and the empirical tests and by
investigating inter- and intra-industry variations in accounting methods and other organizational choices.

References
Alchian, A. A. 1950. Uncertainty, evolution and economic theory. Journal of Political
Economy. (June): 211-221.
•

and H. Demsetz. 1972. Production, Information costs and economic organization. American Economic Review. (December): 777-795.
Ball, R. 1972. Changes in accounting techniques and stock prices. Journal of Accounting
Research. (Supplement): 1-38.
. 1980. Discussion of accounting for research and development costs: The impact of research and development
expenditures. Journal of Accounting Research. (Supple- ment): 27-37.
•

1989. Accounting, auditing and the nature of the firm. Working paper. William E. Simon Graduate School of Business
Administration, University of Rochester.
>

and P. Brown. 1968. An empirical evaluation of accounting income numbers. Jour- nal of Accounting Research. (Autumn): 159-
178.
4

and G. Foster. 1982. Corporate financial reporting: A methodological review of em- pirical research. Journal of Accounting Research.
(Supplement): 161-234.
Beaver, W. 1968. The information content of annual earnings announcements. Journal of
Accounting Research. (Supplement): 67-92.
Begley, J. 1990. Debt covenants and accounting choice. Journal of Accounting & Eco-
nomics. (Forthcoming).
Bernard, V. 1989. Capital markets research in accounting during the 1980s: A critical re-
view. Manuscript, University of Michigan.
Bowen, R. M., E. W. Noreen, and J. M. Lacey, 1981. Determinants of the corporate decision
to capitalize interest. Journal of Accounting & Economics. (August): 151-179. Brown, L. D., and J. C. Gardner. 1985.
Using citation analysis to assess the impact of jour- nals and articles on contemporary accounting research (CAR). Journal of Accounting
Research. (Spring): 84-109.
Chambers, R. J. 1966. Accounting, evaluation, and economic behavior. Prentice-Hall. Christie, A. A. 1987. On cross-sectional
analysis in accounting research. Journal of Ac-
counting & Economics. (December): 231-258.
1990. Aggregation of test statistics: An evaluation of the evidence on contracting and size hypotheses. Journal of
Accounting & Economics. (Forthcoming). Christenson, C. 1983. The methodology of positive accounting. The Accounting
Review.
(January): 1-22.
Coase, R. H. 1937. The nature of the firm. Economica. (November): 386-405.
Daley, L. A., and R. L. Vigeland. 1983. The effects of debt covenants and political costs on the choice of accounting methods: The case of
accounting for R&D costs. Journal of Accounting & Economics. (December): 195-211.
DeAngelo, L. E. 1986. Accounting numbers as market valuation substitutes: A study of
management buyouts of public stockholders. The Accounting Review. (July): 400-420. 1988a. Managerial competition, information
costs, and corporate governance: The use of accounting performance measures in proxy contests. Journal of Accounting &
Economics. (January): 3-36.
•

1988b. Discussion of evidence of earnings management from the provision for bad debts. Journal of Accounting Research.
(Supplement): 32–40.
Duke, J., and H. Hunt. 1990. An empirical examination of debt covenant restrictions and accounting-related debt proxies. Journal of
Accounting & Economics. (Forthcoming).
154
 The Accounting Review, January 1990

Fama, E. F., and M. C. Jensen. 1983a. Separation of ownership and control. Journal of Law
and Economics. (June): 301–325.
and
1983b. Agency problems and residual claims. Journal of Law and Eco- nomics. (June): 327-349. Foster,
G. 1980. Accounting policy decisions and capital market research. Journal of Ac-
counting & Economics. (March): 29-62.
Friedman, M. [1953] 1966. The methodology of positive economics, essays in positive eco-
nomics. Reprint, Phoenix Books.
Frost, C. and V. Bernard. 1989. The role of debt covenants in assessing economic conse- quences of limiting capitalization of
exploration costs. The Accounting Review. (October): 788-808.
Gagnon, J. M. 1967. Purchase versus pooling of interest: The search for a predictor. Journal
of Accounting Research. (Supplement): 187-204.
Gordon, M. J. 1964. Postulates, principles and research in accounting. The Accounting
Review. (April): 251-263.
+

B. N. Horwitz, and P. T. Meyers. 1966. Accounting measurements and normal growth of the firm. Research in
accounting measurement. Eds. R. K. Jaedicke, Y. Ijiri, and O. Nielsen, 221–231. American Accounting Association.
Healy, P. M. 1985. The effect of bonus schemes on accounting decisions. Journal of Account-
ing & Economics. (April): 85-107.
Hines, R. D. 1988. Popper's methodology of falsificationism and accounting research. The
Accounting Review. (October): 657-662.
Holthausen, R. W. 1981. Evidence on the effect of bond covenants and management com- pensation contracts on the choice of
accounting techniques: The case of the depreciation switch-back. Journal of Accounting & Economics. (March): 73-109.
1990. Accounting method choice: Opportunistic behavior, efficient contracting and information perspectives. Journal of
Accounting & Economics. (Forthcoming).
•

and R. W. Leftwich. 1983. The economic consequences of accounting choice: Impli- cations of costly contracting and
monitoring. Journal of Accounting & Economics. (August): 77-117.
Jensen, M. C. 1983. Organization theory and methodology. The Accounting Review.
(April): 319-339.
-, and W. H. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial
Economics. (October): 305–360. Jones, J. 1988. The effect of foreign trade regulation on accounting choices, and produc-
tion and investment decisions. Working paper, University of Michigan.
Kalay, A. 1982. Stockholder-bondholder conflict and dividend constraints. Journal of
Financial Economics. (July): 211-233.
Kaplan, R. S. 1985. Comments on Paul Healy: Evidence on the effect of bonus schemes on accounting procedure and accrual decisions.
Journal of Accounting & Economics. (April): 109-113.
Klein, B. 1983. Contracting costs and residual claims: The separation of ownership and con-
trol. Journal of Law & Economics. (June): 367–374.
R. Crawford, and A. Alchian. 1978. Vertical integration, appropriable rents, and the competitive contracting process.
Journal of Law & Economics. (October): 297-326. Kohler, E. L. 1975. A dictionary for accountants. 5th ed. Prentice-Hall.
Lavoie, D. 1989. The accounting of interpretations and the interpretation of accounts: The communicative function of "the
language of business." Methodology and accounting research: Does the past have a future. Ed. O. Johnson, 107–149. Orace
Johnson. Leftwich, R. 1981. Evidence of the impact of mandatory changes in accounting principles on
corporate loan agreements. Journal of Accounting & Economics. (March): 3–36.
1983. Accounting information in private markets: Evidence from private lending agreements. The Accounting Review.
(January): 23-42.
1990. Aggregation of test statistics: Statistics vs economics. Journal of Accounting & Economics.
(Forthcoming).
R. L. Watts, and J. L. Zimmerman. 1981. Voluntary corporate disclosure: The case of interim reporting. Journal of
Accounting Research. (Supplement): 50-77.
Watts and Zimmerman-Positive Accounting Theory
155

Lev, B., and J. A. Ohlson. 1982. Market-based empirical research in accounting: A review, interpretation, and extension. Journal of Accounting
Research. (Supplement): 249-
322.
Liberty, S. E., and J. L. Zimmerman. 1986. Labor union contract negotiations and account-
ing choices. The Accounting Review. (October): 692-712.
Lowe, E. A., A. G. Puxty, and R. C. Laughlin. 1983. Simple theories for complex processes: Accounting policy and the market
for myopia. Journal of Accounting and Public Policy. (Spring): 19-42.
Lys, T. 1984. Mandated accounting changes and debt covenants: The case of oil and gas
accounting. Journal of Accounting & Economics. (April): 39–65.
Malmquist, D. 1990. Efficient contracting and the choice of accounting method in the oil and
gas industry. Journal of Accounting & Economics. (Forthcoming).
McKee, A. J., Jr., T. B. Bell, and J. R. Boatsman. 1984. Management preferences over ac- counting standards: A replication and additional
tests. The Accounting Review. (Octo- ber): 647-659.
McNichols, M., and G. Wilson. 1988. Evidence of earnings management from the provision
for bad debts. Journal of Accounting Research. (Supplement): 1-31.
Meckling, W., and M. Jensen. 1986. Knowledge, control and organizational structure.
Working paper, University of Rochester.
Mian, S., and C. Smith. 1990. Incentives for unconsolidated financial reporting. Journal of
Accounting & Economics. (Forthcoming),
Modigliani, F., and M. H. Miller. 1958. The cost of capital, corporation finance and the theory
of investment. American Economic Review. (June): 261–297.
Moyer, S. 1988. Accounting choices in commercial banks. Dissertation, University of
Rochester.

Myers, S. 1977. Determinants of corporate borrowing. Journal of Financial Economics.

(November): 147-175.
Peltzman, S. 1976. Toward a more general theory of regulation. Journal of Law and Eco-
nomics. (August): 211-240.
Popper, K. R. [1959] 1965. The logic of scientific discovery. Reprint, Harper & Row. Press, E., and J. Weintrop. 1990. Accounting-
based constraints in public and private debt agreements: Their association with leverage and impact on accounting choice. Journal
of Accounting & Economics. (Forthcoming).
Ricks, W. 1982. Market assessment of alternative accounting methods: A review of the em-
pirical evidence. Journal of Accounting Literature. (59-102).
Smith, C. W. 1980. On the theory of financial contracting: The personal loan market. Jour-
nal of Monetary Economics. (July): 333-357.
-, and J. B. Warner. 1979. On financial contracting: An analysis of bond covenants. Journal of Financial Economics. (June): 117–
161.
7

and R. Watts. 1982. Incentive and tax effects of U.S. executive compensation plans. Australian Journal of Management.
(December): 139-157.
—, and ———. 1986. Investment opportunity set and corporate policy choices. Working paper, University of Rochester.
Sorter, G. H., S. W. Becker, T. R. Archibald, and W. H. Beaver. 1966. Accounting and finan- cial measures as indicators of corporate
personality-some empirical findings. Re- search in accounting measurement. Eds. R. K. Jaedicke, Y. Ijiri, and O. Nielsen, 200-210.
American Accounting Association.
Sprouse, R., and M. Moonitz, 1962. A tentative set of broad accounting principles for busi- ness enterprises. Accounting
Research Study No. 3. American Institute of Certified Public Accountants.
Stigler, G. J. 1971. The theory of economic regulation. Bell Journal of Economics and Man-
agement Science. (Spring): 3-21.
Sutton, T. G. 1988. The proposed introduction of current cost accounting in the U.K.: Deter- minants of corporate preference, Journal of
Accounting & Economics. (April): 127- 149.
Tinker, T. A. M., B. D. Merino, and M. D. Neimark. 1982. The normative origins of positive
156
The Accounting Review, January 1990

theories: Ideology and accounting thought. Accounting, Organizations and Society 2: 167-200.
Watts, R. L. 1974. Accounting objectives. Working paper, University of Rochester.
1977. Corporate financial statements, a product of the market and political processes. Australian Journal of Management.
(April): 53–75.
and J. L. Zimmerman. 1978. Towards a positive theory of the determination of ac- counting standards. The Accounting Review.
(January): 112-134.
and
1979. The demand for and supply of accounting theories: The market for
excuses. The Accounting Review. (April): 273-305.
and
1986. Positive accounting theory. Prentice-Hall. Whittington, G. 1987. Positive accounting: A review article. Accounting and Business
Research. (Autumn): 327-336.
Whittred, G. 1987. The derived demand for consolidated financial reporting. Journal of
Accounting & Economics. (December): 259-285.
Wong, J. 1988. Economic incentives for the voluntary disclosure of current cost financial
statements. Journal of Accounting & Economics. (April): 151-167.
Zimmer, I. 1986. Accounting for interest by real estate developers. Journal of Accounting
& Economics. (March): 37-51.
Zimmerman, J. L. 1979. The costs and benefits of cost allocations. The Accounting Review.
(July): 504-521.
-. 1983. Taxes and firm size. Journal of Accounting & Economics. (August): 119-149. Zmijewski, M., and R. Hagerman. 1981. An income
strategy approach to the positive theory of accounting standard setting/choice. Journal of Accounting & Economics. (August):
129-149.