An examination of share price behavior during the 1973-1976 lessee accounting standard-setting process

Material Information

An examination of share price behavior during the 1973-1976 lessee accounting standard-setting process
Thompson, Robert B., 1949- ( Dissertant )
Ajinkya, Bipin B. ( Thesis advisor )
Abdel-khalik, A. Rasha ( Thesis advisor )
Smith, E. Daniel ( Reviewer )
Radcliffe, Robert ( Reviewer )
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Copyright Date:
Physical Description:
viii, 223 leaves : ill. ; 28 cm.


Subjects / Keywords:
Accounting interpretations ( jstor )
Accounting standards ( jstor )
Assets ( jstor )
Debt contracts ( jstor )
FASB standards ( jstor )
Financial accounting ( jstor )
Financial statements ( jstor )
Leases ( jstor )
Prices ( jstor )
Shareholders ( jstor )
Accounting thesis Ph. D
Dissertations, Academic -- Accounting -- UF
Leases -- Accounting ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )
Spatial Coverage:
United States


Generally accepted accounting principles, the financial reporting rules followed by U.S. business entities, are continuously modified by agencies such as the U.S. Securities and Exchange Commission and the Financial Accounting Standards Board. This process occasionally flares into controversy when it is felt that the financial statements of a particular class of firms will be adversely affected by a contemplated rule change. It is sometimes claimed that the accounting effects of the rule change will result in a downward revision of the equity values of affected firms. One argument offered in support of such claims is that a mandated accounting change may effectively alter the degree to which shareholder activities are restricted by accounting-based negative loan covenants. An alteration of this type may result in a transfer of wealth from shareholders to creditors, or in deadweight losses born primarily by shareholders, as resources are consumed in coping with the contract-related effects of the change. In this dissertation, the behavior of share prices during the 1973-1976 "lessee accounting" standard-setting process is examined for consistency with the above arguments. This policy process culminated in 1976 with the issuance, by the Financial Accounting Standards Board, of Statement of Financial Accounting Standards No. 13 - Accounting for Leases . In some cases this accounting standard, which mandated the capitalization of a class of "financing leases," caused non-trivial changes in financial statement relationships frequently used as a basis for loan covenants. Subject to certain caveats , the data examined are consistent with the hypothesis that lessee shares were valued downward (relative to non-lessee shares) as a result of the process. The observed cross-sectional pattern of returns is consistent with two predictions which are deduced from the "contracting hypothesis" described above. Firms which are highly leveraged, and for which the potential financial statement impact of the standard was large, exhibited significant negative abnormal price behavior relative to other groups of firms. And within this class of firms, the data suggest a significant relationship between the magnitude of abnormal performance and the proportion of long-term debt actually subject to accounting-based restrictions.
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: leaves 217-222.
General Note:
General Note:
Statement of Responsibility:
by Robert B. Thompson II.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000473767 ( AlephBibNum )
11665271 ( OCLC )
ACN8976 ( NOTIS )


This item has the following downloads:

Full Text







I would like to thank the members of my committee for their

advice, and above all for their patience: Bipin Ajinkya (Chairman),

A. Rashad Abdel-khalik (Cochairman), Robert Radcliffe, and E. Dan


P. C. Venkatesh helped me to unravel the mysteries of computing,

and Sarah Terry typed each manuscript with unflagging cooperation.

For purely personal reasons, special thanks are also due to Joan

Cartier, Eugene T. Martin,and Alan G. Mayper.

Financial support for this dissertation was provided by the Ernst

and Whinney Foundation in the form of an Ernst and Whinney Doctoral

Dissertation Award. This support was invaluable, and I am grateful to

have received it.



ACKNOWLEDGEMENTS................................................ ii
LIST OF TABLES....................................... ............
ABSTRACT ........................ ............................... vii


INTRODUCTION ..... .. .................................. 1
1.1 Notes....................................... 10

2.1 Introduction ................................ 11
2.2 Recent Evolution of Lessee Financial
Accounting and Reporting Standards......... 12
2.3 The Financial Statement Impact of
Lease Capitalization...................... 25
2.4 Review and Analysis of Previous Studies..... 27
2.5 Summary ............... ...................... 51
2.6 Notes........................................ 53

3 THE CONTRACTING HYPOTHESIS............................ 56
3.1 Introduction................ ................ 56
3.2 Leftwich Study........................... ... 57
3.3 Accounting-Based Loan Covenants and
Lease Capitalization..................... 62
3.4 Factors Influencing the Cross-Sectional
Distribution of "Contracting" Share
Price Effects............................ 70
3.5 Hypotheses................................... 90
3.6 Notes........................... ............ 93

4.1 Introduction............................... 95
4.2 Sample Selection ............................ 96
4.3 Data...................................-- .. 98
4.4 Research Design for First-Stage Tests....... 105
4.5 Empirical Results.......................... 112
4.6 Interpretation and Discussion............... 155
4.7 Notes........................................ 164

5.1 Introduction.......... ..................... 165
5.2 Restatement of Hypotheess................... 166
5.3 General Description of Test Procedure........ 168


5.4 Description of Sample and Data
Collection Procedures..................... 174
5.5 Measurement of Independent Variables........ 177
5.6 Results................................... 179
5.7 Robustness of Results....................... 192
5.8 Notes................................ ...... 205

6 SUMMARY AND CONCLUSIONS............................... 206

APPENDIX DERIVATION OF BIAS EQUATIONS.......................... 215

REFERENCES..... .................................................. 217

BIOGRAPHICAL SKETCH............................................. 223



4-1 Summary of Sample Selection Procedure............

4-2 Distribution of TA and (NCL/TA)..................

4-3 Index of Events Connected with the Evolution
of Lessee Accounting Standards.................

4-4 Event Weeks and Event Periods....................

4-5 Distribution of "Market Model" Regression

4-6 Estimated Coefficients of Cross-Sectional
Regressions Event Weeks 1-16.................

4-7 Estimated Coefficients of Cross-Sectional
Regressions Event Periods 1-11...............

4-8 Summary of Estimates a3t from Cross-
Sectional Regressions..........................

4-9 Summary of Weekly Return Observations
Used in Estimating "Local" 8...................

4-10 Distribution of "Local" 3-Estimates..............

4-11 Estimated Standard Errors of "Prior" and
"Local" B-Estimates............................

4-12 Variance of Market Model Disturbance
Term "Prior" and "Local" Estimates...........

4-13 Revised Regression Results Event Weeks.........

4-14 Revised Regression Results Event Period
Intervals ......................................

4-15 t-Statistics and Estimated Standard Errors -
Event Period Intervals.........................

4-16 Confidence Intervals and t-Statistics
for Ztnta3t .. ............. ... ...... .... ...

4-17 Comparison of Results............................


5-1 Partitioning Scheme...............................

5-2 Independent Variable Definitions ..................

5-3 Regression Variables Descriptive Statistics.....

5-4 Regression Variables Product Moment

5-5 Cross-Sectional Regression Results................

5-6 Alternative Specification of Dependent

5-7 Alternative Specification of X' ...................

5-8 Alternative Specification of D.....................

5-9 Industry Distribution in Total Sample
Regressions........................ ....... ......

5-10 Alternative Specification with Industry
Dummy Variables ...................................

5-11 Lessee Subsample Only .............................

5-12 Lessee-Only Regressions Implied Cumulative
Abnormal Returns.................................














Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Robert B. Thompson II

April 1984

Chairman: Bipin B. Ajinkya
Cochairman: R. Rashad Abdel-khalik
Major Department: Accounting

Generally accepted accounting principles, the financial re-

porting rules followed by U.S. business entities, are continuously

modified by agencies such as the U.S. Securities and Exchange Commis-

sion and the Financial Accounting Standards Board. This process

occasionally flares into controversy when it is felt that the

financial statements of a particular class of firms will be adversely

affected by a contemplated rule change. It is sometimes claimed that

the accounting effects of the rule change will result in a downward

revision of the equity values of affected firms. One argument offered

in support of such claims is that a mandated accounting change may

effectively alter the degree to which shareholder activities are

restricted by accounting-based negative loan covenants. An alteration

of this type may result in a transfer of wealth from shareholders to

creditors, or in deadweight losses born primarily by shareholders, as

resources are consumed in coping with the contract-related effects of

the change.

In this dissertation, the behavior of share prices during the

1973-1976 "lessee accounting" standard-setting process is examined for

consistency with the above arguments. This policy process culminated

in 1976 with the issuance, by the Financial Accounting Standards

Board, of Statement of Financial Accounting Standards No. 13 -

Accounting for Leases. In some cases this accounting standard, which

mandated the capitalization of a class of "financing leases," caused

non-trivial changes in financial statement relationships frequently

used as a basis for loan covenants.

Subject to certain caveats, the data examined are consistent with

the hypothesis that lessee shares were valued downward (relative to

non-lessee shares) as a result of the process. The observed

cross-sectional pattern of returns is consistent with two predictions

which are deduced from the "contracting hypothesis" described above.

Firms which are highly leveraged, and for which the potential finan-

cial statement impact of the standard was large, exhibited significant

negative abnormal price behavior relative to other groups of firms.

And within this class of firms, the data suggest a significant

relationship between the magnitude of abnormal performance and the

proportion of long-term debt actually subject to accounting-based





Beginning with the burgeoning popularity of the "sale-and-

leaseback" in the late 1940s and early 1950s, and continuing through

to the present day, the long-term noncancellable lease has grown

steadily in importance as a financing vehicle. During most of this

same period, the question of appropriate financial reporting for these

"financing leases" has been a matter of considerable debate. One

faction in this controversy -- lessee firms and their representatives

-- has generally favored treating leases for financial reporting

purposes as simple period-by-period rental arrangements, on the theory

that lease contracts, like wage contracts and purchase commitments,

are strictly executory in nature. Another faction, peopled largely by

academics, financial analysts, and (to a limited extent) financial

accounting policy makers, has maintained that, regardless of their

outward form, financing leases are in substance asset acquisitions

financed by borrowing, and should be accounted for as such. And in

the region separating these extremes, many shades of accounting

argument have been voiced over the years.

In terms of actual financial reporting practice, the "rental"

approach to lease accounting was used almost exclusively by lessees

until the late 1970s. A slow shift in policy orientation toward

mandatory capitalization began in April 1973, when the newly organized

-- and soon to be official -- Financial Accounting Standards Board

(FASB) announced a comprehensive review of lease accounting as one of

seven items on its initial agenda. In its own last official

pronouncement, Accounting Principles Board Opinion No. 31, the

Accounting Principles Board moved in June 1973 to standardize the form

and content of disclosure by lessees. In October of the same year,

the U. S. Securities and Exchange Commission (SEC) issued Accounting

Series Release No. 147, which in essence amended Regulation S-X tc

include the disclosure requirements of APB 31, plus additional

requirements for the disclosure of pro forma income statement and

balance sheet effects based on an assumed set of capitalization rules.

Following these actions, the FASB entered upon its own standard

setting process, which culminated almost three years later, in

November 1976, with the issuance of Statement of Financial Accounting

Standards No. 13 -- Accounting for Leases (SFAS 13). From the

lessee's perspective, the general thrust of this new standard was to

extend the subset of leases (some would say to create a subset of

leases) for which capitalization is the required accounting treatment.

Moreover, in an almost unprecedented fashion, the Board held the new

capitalization provisions to be retroactively applicable to existing

leases. Thus, for many lessees, the financial statement impact of

SFAS 13 was dramatic.

The standard setting process just outlined was heatedly contested

at every step along the way, and, as it became increasingly evident

that the final outcome of the process would be an incrementally

stringent capitalization requirement, the policy debate centered more

and more on welfare-type arguments than on the technical accounting

arguments mentioned earlier. Proponents of capitalization (including

the SEC) argued for the welfare of the investor, citing his need for

comparable financial statements and maintaining that similar trans-

actions (i.e., asset acquisitions financed using conventional loans

and leased acquisitions) should receive similar accounting treatment

in order to minimize confusion. Opponents argued that mandatory lease

capitalization would, for a large number of lessees, so adversely

affect key financial statement relationships that access to capital

would be impaired. The costs to shareholders of such firms were held

to be so great as to outweigh any potential benefits of capitalization

to investors, particularly in light of the fact that full and uniform

disclosure of leasing activities had been achieved with the promulga-

tion of APB 31 and ASR 147. Any investor who felt that capitalization

was the appropriate accounting treatment for financing leases could,

it was argued, easily adjust a firm's financial statements using the

data disclosed in compliance with these standards.

The broad purpose of the present dissertation is to examine

whether the behavior of lessee share prices during the 1973 1976

"policy change" period was consistent with the predictions of those

who opposed mandatory lease capitalization. This question should be

of particular interest to the accounting community, since the anti-

capitalization arguments referred to above imply that it is the manner

in which information is disclosed, rather than the information per se,

which is responsible for the predicted adverse effects.

At the time, this notion was greeted with a degree of skepticism

by some accountants, and in particular by many academics. The

perception of the present author is that, throughout most of the

1970s, a preponderance of accountants tended to view the financial

accounting process as being primarily "informational" in nature. In

the academic community, this informational perspective was coupled

with an almost rigid belief that securities markets are extremely

efficient in the semi-strong form defined by Fama (1970). A joint

implication of the informational perspective and semi-strong form

efficiency is that security prices should be invariant, under a

reasonable range of conditions, to the form in which data are publicly

disclosed. According to a well-known exposition of the implications

of market efficiency for accounting written during the period in


Since information is publicly available even if it
is only reported in a footnote to the published
financial statements, or in a 10-K annual report
to the Securities and Exchange Commission, market
efficiency would predict that such information
would be properly impounded in security prices.
Further, any subsequent disclosure of such infor-
mation in a more visible manner . would not
cause the security's price to change. (Dyckman,
Downes and Magee, 1975, p. 86)

Under this view, and given the existing disclosure regulations, lease

capitalization would add no new information to the public domain, and

hence would have no impact on prices.

The counterargument, presented vigorously by lessees at every

opportunity, can be paraphrased as follows. The financial accounting

process actually serves a dual purpose. On the one hand, it provides

information which is useful in assessing the risk and return charac-

teristics of individual securities or other financial claims, and

viewed solely from this perspective, the value of such claims should

be invariant to the form in which data are publicly disclosed. On the

other hand, its output (periodic financial statements) is the product

of reasonably stable measurement rules (i.e., generally accepted

accounting principles), and is subject to verification by independent

third parties. Thus, financial statement numbers and relationships

have also provided a basis on which these same claimants can form

contracts, and, where such accounting-based contracts are actually in

force, the invariance of price to financial statement content can no

longer be asserted. In the specific instance of lease accounting, one

impact of mandated capitalization on firms which are significant


lessees is an adverse, and sometimes radical, transformation of

balance sheet numbers (e.g., the level of retained earnings) and

relationships (working capital, the debt-equity ratio, the ratio of

net tangible assets to funded debt) which are commonly employed as

control devices in loan contracts. In the case of a lessee firm which

is bound to comply with contract terms of this type, capitalization

can lead to an immediate technical contract violation or, at a

minimum, to an increase in the probability that such a violation will

occur in the future. If the penalty for violation is severe, the

firm's shareholders must bear the cost of renegotiating the affected

contract or, alternatively, of revising investment and financing

programs to compensate for what amounts to an unanticipated increase

in the restrictiveness of the contract.

The intuitive appeal of this counterargument leads directly to a

secondary purpose of the present study. If lessee share price

behavior during the policy change period is consistent with the

hypothesis that lessee equities were adversely affected by the change,

it then becomes a matter of interest to inquire whether the

cross-sectional pattern of price behavior is reflective of the

"contracting" notion on which the counterargument is based.

Some results on this question are already available. For

example, Lev (1979) and Collins and Dent (1979) provide evidence that

the equities of firms using the "full cost" method of accounting for

oil and gas exploration costs were valued downward (at least in

relative terms) during the standard setting process leading to the

issuance of Statement of Financial Accounting Standards No. 19 --

Financial Accounting and Reporting by Oil and Gas Producing Companies

(FASB, 1977).3 This standard (later withdrawn), which mandated use of

the "successful efforts" method, would have resulted in balance sheet

and income statement effects similar to those of SFAS 13. Moreover, a

cross-sectional analysis by Collins, Rozeff and Dhaliwal (1981) indi-

cates that the adverse price behavior was more severe on average for

full cost firms having accounting-based loan contracts outstanding.

In a different vein, results presented by Leftwich (1981) indicate

that the equities of "merger-prone" firms were valued downward during

the deliberations leading to the adoption of Accounting Principles

Board Opinion No. 16 -- Business Combinations (1970), and a

cross-sectional analysis is again suggestive of a "contracting"-type

explanation. In both of these studies, however, the evidence in

support of the contracting hypothesis is relatively weak, and it is

hoped that the present study will shed additional light on the


The whole question of whether non-informative (or not very infor-

mative) changes in the form and content of financial statements can

have a material impact on security values is one of more than purely

academic interest. Agencies charged with the responsibility for

setting financial reporting standards must mediate among the interests

of all parties who may be affected by their decisions regarding

specific accounting policy issues. The argument that shareholders of

class x would be adversely affected by policy decision y as a result

of y's impact on accounting numbers used in loan contracts has become

quite a common one, and the FASB in particular has shown a tendency to

be responsive. In connection with the retroactive application

provision of SFAS 13, for example, the Board established a four year

"transition period," an express purpose of which was to allow firms to

renegotiate any loan contracts which would be seriously affected by

the capitalization of existing leases. Moreover, a glance at the

Board's current agenda reveals at least two incipient issues --

pension accounting and accounting for income taxes -- whose resolution

could conceivably have as dramatic a financial statement impact as did

SFAS 13, and in which contracting arguments are likely to play a

critical role. As indicated earlier, however, empirical evidence in

support of such arguments is rather slender, and it is felt that, by

contributing to the present body of evidence, this study should be of

at least indirect value to accounting policy makers.

The remainder of this dissertation is organized as follows.

Chapter 2 provides a more detailed description of the 1973-1976 lessee

accounting standard setting process, a brief discussion of the

financial statement impact of lease capitalization, and a critical

analysis of some existing studies of the share price impact of APB 31,

ASR 147, and SFAS 13. The "contracting hypothesis" is explicated in

Chapter 3. Included therein is a discussion of the use of accounting

constraints in loan contracts, the effect of lease capitalization on

the firm's accounting position with respect to such constraints, and

the cross-sectional pattern of share price behavior which might be

expected as a consequence. The results of empirical tests are

reported in Chapters 4 and 5. Chapter 4 addresses the question of

whether lessee share prices were valued downward (relative to those of

non-lessees) during brief intervals surrounding a number of events

which occurred during the course of the 1973-1976 standard setting

process, and Chapter 5 examines whether the observed "abnormal

performance" corresponds to the cross-sectional pattern predicted in

Chapter 3. Finally, in Chapter 6 the study as a whole is briefly

summarized, some limited conclusions are stated, and some suggestions

for future research are made.


1.1 Notes

1. The history of the controversy over appropriate lessee accounting
is described in greater detail in Chapter 2. For an excellent
review of the technical accounting issues involved, see Financial
Accounting Standards Board (1974).

2. Under an "informational" perspective on the financial accounting
process, the primary purpose of periodic financial statements is
to provide external users with data which can be used in
assessing the risk and return characteristics of the various
financial claims against the reporting firm. This perspective is
perhaps best exemplified by the spirit of the Report of the Study
Group on Objectives of Financial Statements (American Institute
of Certified Public Accountants, 1973), more commonly known as
the Trueblood Report.

3. For a different interpretation of the evidence, however, see
Dyckman and Smith (1979).



2.1 Introduction

As indicated in Chapter 1, a considerable controversy surrounded

the adoption, during the mid-1970's, of APB 31, ASR 147, and SFAS 13.

The first and second of these standards, promulgated respectively by

the Accounting Principles Board and the Securities and Exchange

Commission, had the effect of standardizing and expanding the set of

disclosures required of lessee firms. The third standard, issued by

the Financial Accounting Standards Board, required in addition that

lessees capitalize leases meeting certain criteria.

Opponents of these standards argued that increases in the

required level of disclosure or in the stringency of capitalization

criteria could have a serious impact on the economic position of

lessee shareholders. Four existing studies of lessee share price

behavior during the APB 31, ASR 147, and SFAS 13 standard setting

periods suggest that this may indeed have been the case, although they

shed little light on the question of why such effects might have

occurred. The primary purpose of the present chapter is to examine

these studies in order to see what conclusions can reasonably be drawn

from their results. Before undertaking this task, however, some

background information is in order. Accordingly, section 2.2 de-

scribes the evolution of standards for financial accounting and

reporting by lessees during recent years, ending with the adoption of

SFAS 13. Section 2.3 provides a brief discussion of the financial

statement impact of lease capitalization. Section 2.4 provides a

review of the studies mentioned above, together with a critical

analysis of the results reported therein. The chapter as a whole is

briefly summarized in Section 2.5

2.2 Recent Evolution of Lessee Financial Accounting and
Reporting Standards

while the use of the noncancellable long term lease as a

financing device has increased steadily and dramatically since the

late 1940's,1 the financial reporting rules applicable to leases show

very little change prior to 1973. However, the question of whether

financing leases should be capitalized (i.e., recorded as assets and

corresponding obligations) or treated as rental arrangements has been

controversial almost from the beginning of this period. An early

position, expressed in Accounting Research Bulletin No. 38 by the

AICPA's Committee on Accounting Procedure, was that a lease should be

capitalized only if it were "clearly in substance an installment

purchase of property" (Committee on Accounting Procedure, 1949, p. 7).

Despite subsequent arguments by some writers (e.g., Myers, 1962) that

financing leases should be capitalized because they represent property

rights, the "installment purchase" criterion was reaffirmed by the

Accounting Principles Board in APB Opinion No. 5 (1964).

ARB 38 was never actually binding on reporting firms, and APB 5,

while binding, did not require capitalization of lease contracts

written prior to 1965. In addition, companies apparently had little

difficulty in writing new lease contracts which avoided the capital-

ization criteria set forth in APB 5. Therefore, although the practice

of leasing grew at a rapid rate throughout the 1950's and 1960's, few

leases were actually capitalized.4 Moreover, because neither of these

pronouncements explicitly required a detailed supplemental disclosure

of the significant aspects of noncapitalized lease contracts,

financial statements included relatively little information about the

leasing activities of reporting firms prior to the mid-1970's. For

example, May, Harkins and Rice (1978) found that during the fiscal

year preceding the year in which APB 31 (see below) became effective,

the lease disclosures of over half of a large random sample of lessees

consisted at most of a footnote detailing rental expense for the

current and/or past year.

The ultimate resolution of these disclosure and accounting issues

began to take shape in 1973, during a period of some confusion in the

accounting policy making environment. In July of that year, private

sector standard setting authority was formally transferred from the

AICPA's Accounting Principles Board to the newly created Financial

Accounting Standards Board. And throughout this period, private

sector accounting policy formation appears to have been heavily

influenced by an increasingly "activist" Securities and Exchange

Commission, under the leadership of Chief Accountant John C. Burton.

The disclosure issue was addressed in 1973 by both the APB

(Opinion No. 31) and the SEC (Accounting Series Release No. 147).

During 1971 and 1972 the APB had already taken some tentative steps

toward dealing with lease accounting problems. A public hearing on

the subject was held in October 1971, and various draft proposals were

privately circulated both before and after the hearing. In April

1973, the FASB prepared an initial agenda which included (among other

things) appropriate financial accounting and reporting procedures for

both lessees and lessors. Although the Board clearly indicated that

it did not intend to consider the issues involved on a piecemeal

basis, and intimated that the final lease accounting standard would

not be quickly forthcoming, the APB more or less simultaneously

announced that its own proposals for lessee reporting rules would be


In May 1973, an SEC official charged that the accounting profes-

sion "has probably failed" [to keep up] "with the phenomenal growth

and complexity of lease arrangements," ("SEC Plans to Require," 1973,

p. 9) and announced that the SEC was likely to propose a new set of

disclosure rules for lessees. Such a proposal was in fact published

for comment on June 6, 1973, under the title Securities Act Release

No. 5401. SAR 5401 suggested amendments to Regulation S-X which would

require not only an increase in the general level of disclosures

concerning noncapitalized leases, but also the presentation of pro

forma income statement and balance sheet figures calculated under the

assumption that certain leased assets had been purchased using

borrowed funds.

One week later, and two weeks prior to its own demise, the APB

issued Opinion No. 31, which might be regarded as its own "comment" on

SAR 5401. Passed with only one dissenting vote, APB 31 required foot-

note presentation of a detailed schedule of future minimum rental

commitments arising from noncapitalized, noncancellable leases,

together with a disclosure of rental expense for the current year and

a brief description of sundry aspects of existing lease contracts. In

framing the opinion, the Board emphasized the interim nature of the

disclosure requirements established therein, and pointedly declined to

adopt the SEC's suggestion that pro forma capitalization disclosures

be mandated:

. the Board is refraining from establishing
any disclosure requirements which may prejudge or
imply any bias with respect to the outcome of the
FASB's undertaking, particularly in relation to
the questions of which leases, if any, should be
capitalized and how such capitalization may
influence the income statement. Nevertheless, in
the meantime the Board recognizes the need to
improve the disclosure of lease commitments in
order that users of financial statements may be
better informed.(Accounting Principles Board,
1973, p. 5)

The SEC, apparently unimpressed either by the APB's position or

by the 134 letters of comment fall negative) received in response to

SAR 5401, issued ASR 147 in October 1973. ASR 147 formally amended

Regulation S-X to require 10-K presentation of (a) the lease data

required to be disclosed under APB 31, (b) the discounted value of a

subset of lease commitments referred to as "noncapitalized financing

leases," and, provided that certain materiality criteria were met; (c)

the impact on net income if all noncarcellable financing leases were

capitalized, the related assets amortized on a straight line basis,

and interest recognized on the outstanding lease liability.6 ASR 147

defined a financing lease as one which,

during the noncancellable lease period, either (i)
covers 75 percent or more of the economic life of
the property or (ii) has terms which assure the
lessor a full recovery of the fair market value
. of the property at the inception of the lease
plus a reasonable return on the assets invested.
(U.S. Securities and Exchange Commission, 1973b,
p. 547)

In its introduction to ASR 147, the Commission provided a very

clear statement of its own attitude regarding the adequacy of the APB

31 disclosure requirements and the importance of pro forma

capitalization disclosures to investors:

The Commission has carefully considered the con-
tents of Opinion No. 31 to determine whether it
provided for sufficient disclosure to meet the
needs of investors and has concluded that it does
not. . Specifically, the Commission believes
that disclosures of the present value of financing
leases and of the impact on net income of capi-
talization of such leases . are essential to
investors . and are] . necessary to
enable investors to compare meaningfully the
capital and asset structures and operating results
of companies making use of different methods of
acquiring and financing assets.(U.S. Securities
and Exchange Commission, 1973b, p. 546)

Having thus defined what, in its own eyes, constituted "adequate

disclosure," the Commission concluded its introductory remarks by

stating that its intention was not "to prejudge the issues of lease

accounting now being considered by the Financial Accounting Standards

Board" (U.S. Securities and Exchange Commission, 1973b, p. 546).

SEC spokesmen pointed out that an FASB pronouncement on lease

accounting was not expected before 1975, and that "two years was just

too long to wait in terms of disclosure" ("SEC Adopts Rule," 1973, p.


By disposing of the disclosure issue, the APB and the SEC cleared

the ground for a consideration by the FASB of lease-related accounting

issues. During the three year period following the SEC's adoption of

ASR 147, the FASB engaged in its by now familiar routine of discussion

memorandum, public hearing, and exposure draft. The culmination of

this process was Statement of Financial Accounting Standards No. 13 --

Accounting for Leases, issued by the Board in November 1976. The

process itself, which is of major significance from the perspective of

the present study, is reviewed here.

As indicated previously, the FASB placed lease accounting on its

initial agenda in April 1973. In consultation with a broad based

independent task force which included representatives from industry,

government, public accounting, the financial community, and academe,

the FASB staff prepared a discussion memorandum which outlined and

analyzed the lease accounting issues to be considered during the

standard setting process. These included accounting by lessees and

lessors, accounting for sale-and-leaseback and leveraged leasing

transactions, and matters of implementation. With regard to lessee

accounting, the major questions to be resolved were simple, albeit

highly controversial. First: for what subset of the universe of

leasing transactions is "capitalization" the appropriate accounting

treatment? And second: given a capitalization policy, should that

policy be implemented on a retroactive basis or on a prospective

basis? Basically a very neutral document, the discussion memorandum

proposed no answers to these questions, but merely outlined the

theoretical and popular support for existing schools of thought.

Issued in July 1974 under the title An Analysis of Issues Related to

Accounting for Leases, and widely disseminated, the discussion

memorandum called for letters of comment and position papers, and

invited interested parties to present their points of view at a four

day public hearing to be held during the following November.

The public hearing was apparently well attended and was the

subject of fairly detailed press coverage by a Wall Street Journal

staff reporter (Andrews, 1974). Some generalizations can be made

concerning the attitudes of the "interested parties" in attendance.

Large public accounting firms, professional and academic accounting

organizations, financial analysts, and the New York Stock Exchange

tended to favor capitalization rules more stringent than those of APB

5, generally on the grounds that the resulting financial statements

would be more "useful" or more "comparable" across firms. However,

these organizations exhibited some disagreement on the question of

retroactive v. prospective application. Lessees and organizations

representing lessees tended to oppose capitalization, and, given

capitalization, to oppose retroactive application. Although this

opposition was sometimes justified using accounting arguments (e.g.,

it would be inappropriate to impose a lease capitalization requirement

without first considering appropriate accounting procedures for the

entire spectrum of executory contracts), lessees more often than not

cited economic arguments against capitalization. For example, Robert

O. Whitman, representing the Financial Executives Institute, suggested

that mandatory capitalization would cause some lessees to be in

default under existing loan agreements, and would generally make it

more difficult for lessees to raise new capital through borrowing.

At the macroeconomic level, such a requirement would "conflict with

the nation's goals of combating inflation, fighting unemployment,

protecting the environment, and seeking energy self-sufficiency"

(Andrews, 1974, p. 18).

Formally, the public hearing was no more than a forum organized

to allow interested parties to express their views on the subject of

lease accounting. The FASB itself adopted no position either during

the hearing or at its close. However, Andrews, the Wall Street

Journal reporter, clearly departed under the impression that "the

weight of accounting opinion favors capitalizing most long term

leases" (Andrews, 1974, p. 18), and that opponents of capitalization

had presented arguments which were less than convincing. This was

also the assessment of John C. Burton, the SEC's Chief Accountant, who

noted that "those opposed [to lease capitalization] sense that they

have lost the battle in the accounting profession, and they are now

looking to Congress." Mr. Burton warned that the FASB should "be

prepared to make a case to Congress" (Andrews, 1974, p. 18).

However clear the "weight of accounting opinion" may have been at

the time of the public hearing, a draft version of the proposed final

standard was not issued for public comment until late August 1975. In

the matter of accounting by lessees, this document (hereafter referred

to as the "first exposure draft") proposed five conditions individu-

ally sufficient for capitalization. These conditions, which

corresponded very roughly to the ASR 147 "financing lease" definition,

were designed to operationalize the fundamental position adopted by

the FASB: that a lease should be capitalized if it "transfers

substantially all of the benefits and risks incident to the ownership

of property" (FASB, 1976, p. 54) to the lessee. Basically, the

proposed standard required capitalization of any lease which confers

ownership of property on the lessee, either through an outright

transfer of title or through a bargain purchase option. Even in the

absence of an effective transfer of ownership, however, the proposal

required capitalization if the noncancellable lease term equals or

exceeds 75% of the economic life of the leased property, if the

residual value of the leased property is anticipated to be less than

25% of its fair value at the inception of the lease, or if the leased

property is special purpose property to the lessee and not readily

saleable to another user.

The apparent intent of the Beard in proposing these conditions

was to require a level of lease capitalization beyond that implied by

the "installment purchase" criterion of APB 5. On the question of

implementation, however, the Board proposed that the new capitali-

zation conditions be applied only to lease transactions effected on or

after January 1, 1976. In some 240 letters of comment received in

response to the first exposure draft, the Board was criticized both by

lessees (for proposing an increased level of capitalization) and by

proponents of capitalization (for proposing that the capitalization

conditions be applied on a prospective basis). The latter group

included the SEC, which argued that it was nonsensical to adopt stiff

capitalization rules and at the same time to exempt billions of

dollars in existing leases.

In November 1975, Marshall S. Armstrong, the FASB chairman,

announced that the Board needed more time to consider the comments

received in response to the first exposure draft, and that a final

pronouncement on lease accounting rules would not be issued until

early in 1976. This projection proved to be somewhat optimistic: the

Board's "consideration," which extended throughout the first half of

1976, resulted in alterations of the proposed standard which were felt

to be sufficiently significant to warrant yet another opportunity for

public comment. In order to allow for this, the Board issued a second

exposure draft in July 1976.

In the matter of lessee accounting, the terms of the second

exposure draft differed from those of the first in two respects.

First, the technical conditions for capitalization were altered. The

"special purpose property" condition was dropped entirely, on the

argument that such a condition was not itself sufficient to indicate

that the risks and benefits of ownership had passed from lessor to

lessee. In addition, the "nominal residual value" condition was

dropped in favor of an "investment recovery" condition: capitaliza-

tion would be required if the discounted value of non-cancellable

minimum rentals equals or exceeds 90% of the fair value of the leased

property (less ary investment tax credit retained by the lessor) at
the inception of the lease. Second, and perhaps of primary

importance to lessees, the Board revised its position regarding

implementation, proposing retroactive capitalization of existing

leases at the close of a four year "transition period." During the

transition period, the financial statement effects of retroactive

capitalization were to be given footnote disclosure, but early appli-

cation of the new capitalization rules to all leases was encouraged by

the Board.

Many respondents to the discussion memorandum had indicated that

retroactive application of stringent capitalization rules would cause

some lessee firms to be in technical violation of the terms of

existing loan agreements, and the Board seemed to have this problem in

mind when it proposed "prospective application only" in the first

exposure draft. In the second exposure draft the Board argued that

the four year transition period would give such firms ample oppor-

tunity to renegotiate affected contracts prior to the time at which

retroactive capitalization became mandatory, and thus to avoid most if

not all of the potential costs of compliance. An analysis of letters

of comment received by the Board following the release of the second

exposure draft suggests that many lessee firms were unimpressed by

this argument. Of 116 lessees responding negatively to the second

exposure draft, 63% opposed retroactive capitalization for some

reason, and of the latter group, slightly more than half cited "loan

covenant problems" as a primary reason for their opposition. A number

of respondents argued that a renegotiation of existing loan agreements

would entail two types of costs. The out-of-pocket costs of

renegotiating some loans -- those publicly held -- would be large,

and, moreover, lenders would be likely to demand interest rate

concessions in exchange for a favorable adjustment of accounting-based

loan provisions.

Despite these and other more technical arguments related to the

form of the capitalization conditions, the Board had apparently

decided by this time that a final standard could be issued. In late

November 1976, the Board issued Statement of Financial Accounting

Standards No. 13 -- Accounting for Leases. The provisions of SFAS 13

-- virtually identical to those of the second exposure draft -- became

effective immediately for lease contracts negotiated on or after

January 1, 1977, and retroactively for reporting years beginning after

December 31, 1980. Although one Board member voted against SFAS 13 in

its final form, it is apparent from a discussion of the reasons for

his dissent that the Board as a whole was reasonably satisfied with

those portions of the standard which dealt with lessee accounting and


The adoption of SFAS 13 brought the long controversy over lessee

accounting to a tentative close.1 It is evident from the history

recited above that the period 1973-1976 was one in which the "rules"

for lease-related financial accounting and reporting changed dra-

matically and, at least in the eyes of many lessees, detrimentally.

With regard to APE 31 and ASR 147, some lessees maintained that the

mere disclosure of new data concerning their leasing activities could

be economically harmful. It was felt that, as a result of such

disclosure, investors and creditors would perceive lessees as being

more risky, therefore less credit-worthy and, more generally, less

valuable. However, the sharpest criticism by lessees during the APB

31/ASR 147 period was directed at the SEC's requirement for pro forma

balance sheet and income statement disclosures calculated under the

assumption that "financing leases" had been capitalized. At issue was

whether, by mandating these particular disclosures, the SEC was

prejudging or predetermining the outcome of the yet-to-begin FASB

deliberations on the matter of capitalization by lessees. And it is

clear from the history of those deliberations that, with respect to

lessee accounting, the thorniest questions faced by the Board were

whether leases should be capitalized, which leases should be capital-

ized, and whether capitalization should be instituted on a retro-

active or on a prospective basis. In response to the Board's

proposals, lessees were quick to suggest that the financial statement

effects of more stringent capitalization rules could impair a firm's

ability to raise capital and perhaps result in technical violations of

loan covenants. And these effects would be magnified, they claimed,

should retroactive application be mandated. A number of lessee

respondents argued rather heatedly that it would be inappropriate to

apply the Board's capitalization conditions ex post to transactions

which would have been structured differently had the new rules then

been in effect.

Most of the remainder of this chapter is devoted to a review and

discussion of existing studies which have sought to determine whether

the adverse effects predicted by lessees during the APB 31/ASR 147/

SFAS 13 period did in fact materialize. However, because these

effects were alleged to flow mainly from the financial statement im-

pact of lease capitalization, that impact is briefly described below.

2.3 The Financial Statement Impact of Lease Capitalization

As indicated above, ARB 38 prescribed that certain leases --

those which were, in substance, installment purchases of property --

should be capitalized, and APB 5 set forth criteria to be used in

determining whether a particular lease satisfied this condition.

However, the APB 5 criteria were sufficiently vague that, prior to

SFAS 13, financing-type leases were typically accounted for simply by

charging periodic rental payments against income, with no asset or

obligation recognized in the lessee's accounts.

Under SFAS 13, a lease which satisfies any one of the standard's

four capitalization conditions is given accounting treatment as a

"capital lease." The lessee records both an asset and a liability

equal to the present value of the minimum rental payments to be made

during the noncancellable lease term. While the asset is noncurrent,

the liability will generally include both current and noncurrent

components. Expenses recognized under this scheme include depreci-

ation (where applicable) and interest on the unamortized lease

obligation. In the very common case of a lease which calls for level

rentals, annual expense will decline (in total) year by year,

exceeding the rental amount during the early years of the lease and

exceeded by the rental amount during later years. Thus, a change from

"rental" accounting to capitalization involves both additions to the

balance sheet and an accelerated pattern of expense recognition.

For firms with a prior history of leasing, the effects of retro-

active capitalization on key balance sheet relationships are largely

unambiguous. Indicators of leverage, such as the debt-equity ratio,

increase. Indicators of liquidity (working capital, current ratio)

and indicators of debt coverage (e.g., the ratio of net tangible

assets to funded debt) decline. In addition, because the total

earnings accumulated since the inception of the firm's leasing

activities will be lower on a post-capitalization basis, both total

retained earnings and retained earnings "unrestricted as to dividend

payments" should decrease.

The effects of retroactive capitalization on financial statement

"flow" relationships are somewhat less predictable, because the impact

of capitalization on reported income for a given period depends on the

relative age of lease contracts considered in the aggregate. As an

empirical matter, however, studies by Ro (1978) and Pfeiffer (1980)

suggest that for virtually all firms affected by ASR 147 and SFAS 13,

the effect of capitalization on net income is either negligible or

negative. This in turn implies a deterioration in financial ratios

which measure fixed charge coverage and book rates of return.

One final point regarding the ultimate financial statement impact

of SFAS 13 is worth noting. It is apparent from post-SFAS 13 finan-

cial reports that many long term leases written during the 1950's,

1960's, and early 1970's, while avoiding the capitalization provisions

of APB 5, fell well within the scope of the SFAS 13 capitalization

conditions and were therefore subject to retroactive capitalization.

On the other hand, because the SFAS 13 capitalization conditions are

highly arbitrary, they appear almost to invite circumvention, and it

has been suggested that many post-SFAS 13 lease contracts are

written close to the "capitalization borderline," thereby achieving

accounting treatment as rental arrangements. The latter observation

is based on purely anecdotal evidence, but, taken jointly with the

former, at lease tentatively implies that the standard's hypothetical

"adverse potential" results largely from its provision for retroactive


2.4 Review and Analysis of Previous Studies

As indicated above, the new financial accounting and reporting

requirements imposed by the APB, SEC, and FASB elicited a response

from lessees which can only be described as hostile. Lessees who

opposed the new standards tended to offer equity and welfare

arguments, rather than accounting arguments, against further

disclosure and increased capitalization. In particular, it was argued

that if the disclosure standards were adopted, those firms which

leased extensively would be placed at a disadvantage in the capital

markets relative to those which did not. If retroactive

capitalization were required, as in SFAS 13, some lessees would find

themselves in technical violation of loan indentures. And, at a

second order level, progress toward a number of economy-wide goals

would be hindered.

To the extent that these fears were warranted, one would expect,

other things equal, to see a downward adjustment in lessee share

prices during the 1973-1976 period, reflecting the present value of

new costs associated with the accounting standards in question. The

relative behavior of lessee share prices during this period has been

examined in four major studies. Two of the studies analyze the share

price impact of the APB 31/ASR 147 disclosure rules; the others focus

almost exclusively on SFAS 13. These studies, which differ

substantially as to hypotheses tested, research design, and reported

results, are reviewed below. An assessment is then made of the

conclusions that can reasonably be drawn from the four studies taken

jointly, and a discussion of unresolved issues is provided.

The first systematic study of the effect of a lessee accounting

pronouncement on share prices was undertaken by Ro (1978), who

hypothesized that the lease disclosures mandated by the SEC in ASR 147

would influence investors' perceptions of the risk-return character-

istics of, and therefore the share prices of, affected firms. This

proposition was tested using a sample of 99 pairs of lessee and

nonlessee firms, matched on the basis of systematic risk and (where

possible) SIC industry code. Ro examined mean monthly return dif-

ferences from January 1973 (six months prior to the SEC's initial

proposal of ASR 147) to September 1974 (six months after the first ASR

147 disclosures). Several overlapping subperiods within the primary

study period were also investigated. Finally, mean return differences

for calendar 1972 were examined in order to determine the adequacy of

the matching procedure.

Ro performed statistical tests on mean return differences for the

entire matched sample and for various subsamples, reporting the

following results:

a. During the 1972 (nonstudy) period, mean return differences
were essentially zero.

b. During the primary test period and during all subperiods but
one, lessee returns were significantly lower than those of
the matched non-lessee firms.

c. The pattern indicated in (b) was apparently attributable to
the subsample of firms ("PV-IE firms") which responded to ASP
147 by reporting both the "present value of noncapitalized
financing leases" and a material pro forma income effect.
Mean return differences were generally insignificant for the
group of firms ("PV" firms) which did not report a material
income effect.

d. Return differences of high-S firms were on average more
pronounced than those of low-B firms.12

Based on these results, Ro concluded that "the SEC's extended lease

disclosure decision of 1973 had an impact on the pricing of securities

for the firms which were affected by the decision" (Ro, 1978, p. 336).

To explain the markedly different share price behavior of the "PV" and

"PV-IE" firms, he reasoned that either "the balance sheet effect

itself of capitalized lease disclosure . may not be as great as is

generally believed," or alternatively, that "the balance sheet effect

(if any) was already impounded in security prices prior to the SEC

lease decision" (p. 336).

Ro's study was extended by May, Harkins, and Rice (MHR, 1978),

who suggested that if the disclosures mandated by the APB and SEC

affected the market's assessment of the risk-return characteristics of

lessee firms, then at least one of two possible types of price

behavior should be observed. First, if estimates of the extent of

firms' leasing activities were for some reason systematically biased

-- perhaps due to the costliness of obtaining leasing information from

non-accounting sources -- then one should observe a change in the

level of lessee share prices relative to those of non-lessees.

Second, even in the absence of such a systematic bias, lessee shares

should exhibit greater than normal price variation as new information

about specific firms is "impounded" by the market. MHR also argued

that, at the firm level, the degree of price response should vary

directly with the degree of change in the level of the firm's

disclosures and with the materiality of the firm's lease obligations.

Finally, the authors pointed out that the timing of the observed price

response would depend on the extent to which speculative information

acquisition and trading occurred during the period preceding the first

disclosures under the new standards.

MHR investigated these hypotheses by examining 1972-74 10K and

annual reports for a random sample of 314 firms. "Degree of change in

the level of lease disclosure" was measured qualitatively by reference

to pre- and post-APB 31/ASR 147 disclosures. In addition, a rental

coverage ratio was computed for each firm as an indicator of the

materiality of lease commitments. The total sample was partitioned

into subsamples representing combinations of materiality level and

level of change in disclosure practice, and an average cumulated

residual return (as per Fama and MacBeth (1973)) was computed for each

subsample over the period February 1970-December 1974.

MHP's extensive analysis of average cumulated residuals produced,

among others, the following results:

a. Over the period studied, lessee share prices declined
significantly relative to those of non-lessees.

b. The magnitude of relative decline was found to be strongly
positively correlated with the materiality measure.
Moreover, the materiality of lease commitments appeared to
account for both "PV/PV-IE" effect and, to a lesser extent,
the risk dependence of mean return differences observed by

c. The most dramatic relative price declines were exhibited by
the upper 40%, materiality-wise, of those firms which had
previously disclosed some quantitative information about
lease commitments. In February or March 1972, these shares
entered a lengthy period of steady decline which persisted on
average until June 1973, the month in which APE 31 was

d. Little evidence of significant abnormal price behavior was
found around the date of the first APB 31/ASR 147

Subject to numerous caveats, the authors appear to interpret their

evidence as being consistent with a systematic "market under-

assessment," prior to the beginning of lessee accounting deliberations

by the APB, of the extent of the leasing activities of those firms for

which leasing was "very material." They suggest that speculative

information production and trading during the deliberation period had

the effect of "unbiasing" estimates of leasing activity, and that this

"unbiasing" was largely accomplished prior to the time at which the

new disclosures first became available.

The Ro and MHR studies of share price response to the 1973

APB/SEC disclosure rules assume that the imposition of these rules

represented a pure "information event." In the two studies of SFAS 13

share price response to be reviewed below, such information effects

are de-emphasized. The general argument made by the authors of these

latter studies is that any abnormal price response to SFAS 13 must be

attributed to the standard's indirect effect on the production,

investment, or financing decisions of the firm.

Abdel-khalik, Ajinkya and McKeown (AAM, 1981) analyzed common

stock returns for groups of lessee and non-lessee firms over periods

of 13-22 weeks surrounding each of five standard setting events: the

issuance of the first and second exposure drafts, the adoption of SFAS

13, the issuance of ASR 225,13 and the period during which most firms

applied SFAS 13 retroactively. The lessee sample included all NYSE

and AMEX airline, fast food, and retail firms; and from other

industries, all firms for which the present value of non-capitalized

financing leases (per ASR 147) exceeded both 25% of book long term

debt and 5% of book long term capitalization as of the end of fiscal

1975. Non-lessee control firms were randomly selected from the

remainder of the NYSE/AMEX population.

AAM argued that a firm's production-investment-financing

decisions would be affected by SFAS 13 only if (a) non-capitalized

leases represented a large component of capital structure, and (b) the

firm lacked the necessary flexibility to mitigate the financial

statement effects of the standard. The ratio of the present value of

noncapitalized financing leases to book value of long term debt, and

firm size as measured by the book value of total assets, were adopted

as proxies for these two characteristics. The lessee sample was thus

cross-partitioned by industry type, size, and the materiality of SFAS

13's effect on financial statistics. Control samples were partitioned

on the basis of firm size only. The expectation implicitly expressed

by the authors was that if lessee share price behavior differed during

the SFAS 13 deliberation period from that of similarly sized non-

lessees, the observed differences would be most pronounced in the case

of small firms with substantial lease financing.

AAM examined these arguments empirically by computing average

cumulative standardized market model residuals for each lessee and

non-lessee subsample over each of the "event periods" studied. The

residuals were tested for statistical significance with the following


a. During the first exposure draft period, lessee firms
generally exhibited a strong pattern of significantly
positive residuals.

b. During the second exposure draft period, lessee firms
generally exhibited a strong pattern of significantly
negative residuals.

c. The behavior of lessee residuals during the other periods
studied was either non-significant or industry specific.

d. During the first and second exposure draft periods, lessee
and non-lessee residuals exhibited qualitatively similar
behavior. Moreover, in the case of lessee firms the
anticipated pattern of residuals according to firm size and
materiality of lease commitments was not observed.

Based on these results, and most specifically on result (d), AAM con-

cluded that while significant shifts of relative share prices occurred

during the periods studied, these shifts did not appear to be related

to the SFAS 13 standard setting process.

The opposite general conclusion is drawn by Pfeiffer (1980) from

the results of a study which is in many ways similar to that of AAM.

Pfeiffer suggested that SFAS 13 could have been expected to have a

negative impact on lessee share prices due either to its effect on the

firm's position with respect to loan covenants, or to its effect on

the accounting numbers which, in many cases, are used in determining

the level of management incentive compensation. In this study, the

price behavior of 260 lessee firms was examined for periods of 11-18

weeks surrounding each of the following events: the adoption of APE

31, the adoption of ASR 147, the issuance of the discussion memorandum

on lease accounting by the FASB, the FASB public hearing period, the

issuance of the first and second exposure drafts, and the final

adoption of SFAS 13. "Lessee firms" were those for which capital

leases exceeded 5% of total assets, apparently at the end of fiscal

1976. A control sample of non-lessee firms was also constructed for

use in some parts of the study.

Pfeiffer's analysis of returns took on two forms. The first was

a series of portfolio return comparisons, in which portfolios were

formed to reflect either firm characteristics (e.g., lessee v.

non-lessee, management incentive plans present or absent) or aspects

of SFAS 13 financial statement impact (e.g., large v. small effect on

income, retained earnings, book leverage, working capital). Under the

assumption that returns are generated by a multifactor process, and

the further assumption that a firm's sensitivity to various factors

can be indexed by a set of market- and accounting-based descriptorss,"

individual securities were weighted to form portfolios which were

equivalent in terms of systematic risk, sensitivity to short term

interest rate fluctuations, dividend yield, book leverage, size, and

current position.14 To the extent that this weighting scheme controls

for all factors affecting returns in general, the likelihood is

greatly increased that an observed difference in portfolio returns is

in some way related to the dimension on which the portfolios are

differentiated. Pfeiffer's portfolio return comparisons produced the

following results:

a. Returns to the shares of lessee firms were significantly
lower (dramatically so, considering the short time periods
involved) than returns to non-lessee shares during the
"public hearing" and "second exposure draft" periods.

b. During the same periods, lessee firms for which the financial
statement impact of SFAS 13 was "large" exhibited
significantly lower returns than did lessee firms for which
the financial statement impact was "small."

c. A small number of firms which were known to have renegotiated
loan contracts as a result of SFAS 13 exhibited relatively
low returns during the "public hearing" period.

d. During other periods, and for other dimensions used as a
basis for portfolio formation, portfolio returns did not in
general differ significantly.

In the second phase of Pfeiffer's analysis, which focused only on the

public hearing period, lessee mean returns were regressed cross-

sectionally, first on a market-based estimate of systematic risk, and

then on an extended set of variables which also included four measures

of the impact of SFAS 13 on financial statement numbers and

relationships. As might be expected from the portfolio comparison

results, the extended regression model "explained" a significantly

greater proportion of mean return variation than did the model based

solely on systematic risk.

Based on the various empirical results described above, Pfeiffer

concluded that "SFAS 13 had a significant negative impact on stock

returns of lessee firms," and observed that "the magnitude of the

effect appears to be related to the effect on lending agreements"

(Pfeiffer, 1980, pp. 41-42).

Ignoring for the moment the various interpretations and

conclusions expressed by the authors, the broad empirical results of

these studies can now be summarized. Three of the studies reviewed --

Ro, MHR, and Pfeiffer -- indicate that after some attempt has been

made to control for the effects of factors which influence security

returns in general,

(1) the shares of lessee firms appear to earn lower returns than
those of non-lessee firms during a lengthy period surrounding
the adoption of APB 31 and ASR 147, and also during shorter
periods surrounding at least some of the events which led to
the adoption of SFAS 13; and

(2) the magnitude of the observed difference appears to be posi-
tively associated with the "materiality" of a firm's lease
commitments (as measured, for instance, by Ro's distinction
between "PV" and "PV-IE" firms, by MHR's rental coverage
ratio, or by the financial statement impact indices used by
Pfeiffer as a basis for portfolio formation).

The results of the fourth study are difficult to compare to those of

the first three for several reasons. First, AAM present only

qualitative information concerning the return behavior of their

various portfolios. The results of statistical tests on portfolio

return differences are not reported. Second, while the natural

comparison to make is AAM v. Pfeiffer, the most consistently negative

lessee share price behavior observed in the latter study occurred

during the "public hearing" period. This period was not examined by

PAM. Third, it is possible that the "materiality" measure employed by

AAM as a basis for portfolio formation -- the ratio of noncapitalized

financing leases per ASR 147 to long term debt -- is not, whatever its

other merits, very highly correlated with those used in the other

studies. For example, it is easy to envision situations in which a

low level of leasing activity could be quite large in relation to long

term debt and yet immaterial in terms of its impact on other financial

statistics. Because of this lack of comparability, the discussion to

follow will be based mainly on the results of the first three studies.

Given these results, two questions of interest can now be

addressed. First, can it be concluded with any large measure of

confidence that lessee share prices were actually affected by the

standard setting activities of the APB, SEC, or FASB? And second, if

one assumes that such a conclusion is warranted, to what extent does

the available evidence allow one to distinguish between possible

reasons for such an effect?

The most general answer to the first question, as the authors of

all studies reviewed here have pointed out, is that ex post empirical

studies do not prove causal relationships, since it is technically

impossible to control for all competing explanations of a particular

phenomenon. Speaking in more practical terms, however, if all

competing explanations which are more than "remotely probable" can be

eliminated via the use of appropriate controls, cause and effect

interpretations can be stated with greater confidence than would

otherwise have been the case. Thus, faced with evidence which is

consistent with the hypothesis that lessee share prices were affected

by the imposition of lessee accounting standards, what needs to be

assessed is the extent to which more than remotely probable competing

explanations have been satisfactorily excluded by the control pro-

cedures employed in these studies.

The set of more than remotely probable competing explanations for

such evidence includes any simultaneous event, or series of events,

whose share price impact would probably have been systematically

different for lessee and non-lessee firms, or for "material" and "non-

material" lessees. One competing explanation, of potential importance

in any study which examines contemporaneous security returns, is that

if lessee and non-lessee firms tend to respond differently to general

economic events, the use of an incorrectly specified model of

conditional expected returns could conceivably be responsible for the

observed pattern of empirical results. This alternative deserves

serious consideration, particularly in light of the fact that the

period examined in these studies was characterized both by dramatic

changes in the general level of share prices and by greater than usual

return volatility.

Each study reviewed here deals similarly with the problem of
simultaneous events. It is first assumed that the j security's

realized return over a given period can be decomposed into three

components, as follows:

Rt = Ej + f (lt' ... kt) + jt (1)

In (1), E. is an unconditional expected return, f.(.) is a function of

k random economic factors which affect all security returns (the

intensity of the effect depending on a particular security's

characteristics), and uj reflects the net impact of all

security-specific events, including the effect (if any) of the lessee

accounting event of interest. Because the lessee/non-lessee

distinction could conceivably proxy for inter-security differences in

any or all of these components, a specific empirical model is selected

to represent (1), and model characteristics (i.e., E. and f.) are
3 3

estimated for each security to be studied. These estimates are used

to control for systematic lessee/non-lessee return differences which

could have been expected, given the general pattern of economic events

during the period studied. In effect, such a procedure allows

attention to be focused directly on the behavior of the ujt

Obviously, this procedure can produce spurious results if the

empirical model chosen to represent (1) is incorrectly specified. One

plausible source of such misspecification might be referred to as the

"missing factor" problem. Suppose that the correct model for Rt

depends on k random economic factors, of which only (k-l) are con-

trolled for successfully by the assumed model. In addition, suppose

that returns to lessee and non-lessee shares tend -- for any reason --

to differ in their sensitivities to the behavior of the kth factor.

Under these conditions, it would hardly be surprising to observe a

significant difference between lessee and non-lessee returns for a

given period, even after controlling for the effects of the (k-l)

factors recognized in the assumed model.

In the Ro, MHR, and AAM studies, conclusions are based on condi-

tional prediction errors which were calculated using either a one

factor "market" model or on a time series analogue of the traditional

capital asset pricing model. As an empirical matter, it is well kncwn

that prediction errors associated with these versions of (1) are not

independent across securities, which indicates that an additional

factor or factors must be taken into account in order to explain the

systematic behavior of returns.5 Thus, one can interpret the results

of these studies as being consistent with an "accounting standard

effect," a "missing factor effect," or both.

Pfeiffer recognized and nominally dealt with this problem by

representing (1) as a k factor linear model:

Rjt = E + bjtt + ... + bjk kt + ujt (2)

This is the returns generating process on which the Ross (1977)

arbitrage pricing model is based. Actually, Pfeiffer left k

unspecified and assumed that each of the coefficients bj, ... bjk

could in turn be represented as a linear combination of certain

measureable or estimable characteristics of security j. As pointed

out previously, these characteristics included j's "market B," the

sensitivity of its return to short term interest rate changes, its

dividend yield, and the size, leverage, and current position of the

underlying firm. To the extent that this assumption is correct,

controlling for these characteristics would also control for any

systematic differences between lessee and non-lessee returns.

However, the appropriateness of this control procedure was never

directly tested, and the degree to which it was successful in

eliminating systematic return differences is unknown.

It should be pointed out in passing that even if Pfeiffer's pro-

cedure were capable in principle of perfectly controlling for

systematic lessee/non-lessee return differences, it is still possible

that his use of accounting numbers to proxy for certain firm

characteristics resulted in "matched" lessee and control samples which

were actually fundamentally different. One example can be provided to

illustrate this point. Pfeiffer in essence computed the following

measure for each lessee and non-lessee firm:

LEV. = (D./A.)/M(D/A) ; (3)

D and A. are the book values of firm j's total debt and total assets,
3 3
respectively, and M(D/A) is the mean value of this ratio for all

Compustat firms. In order to control for financial leverage when

comparing returns to portfolios of lessee and non-lessee firms, each

portfolio was weighted to satisfy the requirements

Zw.LEV. = 1 (4a)
3 3

Zw. = 1 (4b)

and w. > 0 for all j, (4c)

where [w] is a vector of portfolio proportions. That is, the

"leverage characteristic" of each portfolio was set to equal a market

wide average. However, because end of fiscal 1972 book values were

used as a basis for these measures, the leasing component of financial

leverage was effectively ignored. If the leverage measure is

redefined to include the present value of -non-capitalized leases, i.e.

D. + NCL.
LEV = + NC M D + NCL (5)
j A. + NCL. A + NCL

it can easily be shown that portfolio weights satisfying (4) will

produce a lessee portfolio whose financial leverage is strictly

greater than that of the related non-lessee portfolio. If, as

Pfeiffer assumed, leverage is an important argument in the

"systematic" portion of (2), the fact that his lessee and non-lessee

portfolios are likely to have differed on this dimension provides an

alternative explanation for the observed return differences. A

similar argument can be made with respect to the measurement of firm

size, for which the book value of total assets was taken as a proxy.

To summarize, because of a potential "missing factor" problem,

lessee/non-lessee return differences which have tentatively been

attributed (in the studies reviewed above) to the impact of lessee

accounting standards may simply reflect -- to some extent --

systematic differences in the return behaviors of lessee and

non-lessee shares.

In addition to the "missing factor" problem, a second plausible

source of model misspecification, which will be referred to here as

the "nonstationarity" problem, could also account for some or all of

the lessee/non-lessee return differences observed in these studies.

This problem arises if the parameters (E. and f.) of (1) are

nonstationary, if their direction of change and rate of change are in

part the result of an interaction between security-specific charac-

teristics and aggregate economic phenomena, and if the lessee/

non-lessee distinction proxies for one or more of the relevant

security-specific characteristics.

Many studies of stock price behavior indicate that the market

model slope parameter (8) is nonstationary.6 And in the theoretical

realm, a number of studies suggest that changes in leverage occasioned

by drifts in firm value may be an important source of such

nonstationarity. For example, Galai and Masulis (1976) show that, if

the equity of a levered firm which pays no dividends is regarded as a

European call option on the firm's assets, then

= (LS- )v =sv (6)

where B and 5 are instantaneous values (for the firm's equity and
s v
assets) of the slope parameter of the market model version of (1), and

n represents the instantaneous elasticity of equity value (S) with

respect to firm value (V). For a given level of debt, n is shown to

be a convex and monotonically decreasing function of firm value,
approaching unity as V becomes large. Thus, if B is presumed to be

stationary, B should increase as V falls, and vice versa. Moreover,

the convexity of n indicates that if firms i and j differ only by a

scale factor and by degree of leverage -- i.e., if Vi/Si > Vj/Sj, then

6 should be more volatile than B with respect to changes in V.
Sl s]
These results suggest (but do not absolutely imply) that during a

period of steady decline in market values, the market model slope

parameter will increase for all levered equities, and that the rate of
increase will be faster for highly levered firms.8 Likewise, during

a period in which market values steadily increase, the equity slope

parameter will generally decrease, and again, the rate of decrease

will be faster for highly levered firms.

The results of the studies reviewed above can be interpreted in

the context of this scenario. Ro, MHR, and Pfeiffer estimated market

model B's over a period preceding APB 31 and ASR 147. These estimates

were used in turn to compute conditional prediction errors (i.e.,

estimates cf uj in (1)) during subsequent periods of three months or

more surrounding various standard-setting events. The fact that

average prediction errors for lessee firms were more negative than

those for non-lessees over some of the periods studied is the primary

evidence supporting the "adverse share price effect" hypothesis.

Evidence will be presented in Chapter 5, however, to indicate

that on average, lessee firms are more highly levered than non-lessee

firms. That is, the lessee/non-lessee distinction proxies for "degree

of leverage." In addition, an examination of various stock market

indices shows that equity values (and, presumably, firm values)

declined steadily and dramatically throughout 1973 and 1974, the

period during which most of the results of the above studies were

observed.19 These facts, together with the nonstationarity scenario

just outlined, suggest that some proportion of those results can be

explained in terms of differential biases in the measurement of ujt.

In particular, if "true" equity B's are increasing relative to their

estimated values, the rate of increase being faster for lessees than

for non-lessees, and if market returns are on average negative, then a

downward bias will be imparted to the difference between lessee and

non-lessee average prediction errors.

Figure 2.1, a plot of cumulative average lessee/non-lessee pre-

diction error differences over the period 1973-1978 (inclusive), lends

some credibility to the notion that the Ro, MHR, and Pfeiffer results

may be due in part to "nonstationarity" problems. As was done in

those studies, the values of ujt used to construct this plot were

computed using pre-1973 B estimates. The negative drift of the

curve over the 1973-1974 declining market period is consistent with

earlier findings. However, the direction of the drift is reversed

during the generally rising market of 1975-1978, and such a reversal

is precisely what one would expect under the nonstationarity scenario

described above. That is, if "true" equity B's are larger (relative

to their estimated values) for lessees than for non-lesses, and if

market returns are on average positive, then an upward bias will be

imparted to the difference between lessee and non-lessee average

prediction errors.

The preceding discussion can now be summarized. Three of the

four studies reviewed above report empirical results which are

consistent with the hypothesis that lessee share prices were adversely

affected by the various lessee accounting standards adopted during the

mid-1970's. However, because these results are derived from an

examination of contemporaneous security returns, it is argued that

failure to adequately control for systematic return differences, due

either to the "missing factor" problem, to the "nonstationarity"

problem, or to biases in the measurement of firm characteristics,

constitutes a "more than remotely probable" alternative explanation of


) C

4- U

., -H
C W "L

LI -H 04
---3--- ----------'J-- __ ___ 9---_- Si

> C


-- --- -------------_-- T

'.,* DL/L z C

-~ ~ --- -------- _- -__ _-------__ T i

0 c


the evidence. It is further argued that until this competing expla-

nation can be satisfactorily eliminated, only limited confidence can

be placed in the conclusions that a lessee accounting standard/lessee

share price effect actually occurred.

In order to accomplish this, one must have in hand a fully

developed (and fully validated) dynamic model of the "returns

generating process" -- something that is not likely to be available in

the foreseeable future. Thus, the discussion above is offered as an

explication of a problem, rather than as a criticism of the various

studies reviewed. The importance of the problem itself, however,

cannot be overemphasized, particularly in an era in which policy

making groups appear to be using studies of the type reviewed either

as inputs to the policy decision process or as ex post validations of

decisions already taken. While the measurement approach taker in the

present study (to be described in Chapter 4) will not solve the

problem, it is hoped that it will provide additional insights

concerning the magnitudes of lessee accounting standard share price


Although it is important to know whether lessee accounting

standards -- or for that matter any accounting standards -- have had a

significant impact on share prices, the value of this knowledge

depends on the degree to which the impact can be explained. Thus, if

it is temporarily assumed that the empirical results reviewed above

are indicative of a share price impact, the second of the two

questions posed earlier must still be addressed: to what extent does

the available evidence allow one to distinguish between possible

reasons for such an effect?

Two broad hypotheses can be discerned in the studies described

above, both of which have frequently been reflected in the comments

submitted by respondents to various lessee accounting proposals.

These hypotheses can briefly be stated as follows:

H1: Information whose production cost previously exceeded
any related benefit becomes available to the market as a
result of the adoption of a lessee accounting standard.
This information affects share prices because it alters
investors' perceptions of the risk/return
characteristics of lessee securities.

H2: Certain contract terms, which are expressed with
reference to financial accounting numbers, and whose
negotiation was conditioned on established financial
accounting practice for lessees, are effectively altered
due to the adoption of a new lessee accounting standard.
Share prices are affected because the alteration in
contract terms narrows the set of wealth maximizing
opportunities available to shareholders.

Throughout the remainder of this study, H and H2 will be referred to

as the "information" and "contracting" hypotheses, respectively. As

indicated in Chapter 1, a primary goal of the present study is to

determine whether the cross-sectional pattern of (apparent) lessee

accounting standard-related share price effects is consistent with a

particular aspect of the contracting hypothesis: namely, that the

price impact is "more adverse" in the case of firms having substantial

loan contracts whose terms are stated with reference to financial

accounting numbers and relationships.

While the footnote disclosures required by APB 31 and ASR 147 may

have resulted in the provision of new information, it is unlikely that

these standards had any direct effect on contract terms. At first

glance, then, it would seem that the information hypothesis, and only

the information hypothesis, is relevant in the context of the Ro and

MHR studies. On the other hand, SFAS 13 can be viewed as a standard

whose principal effect was to require that these disclosures be

incorporated in the firm's financial accounts, thus potentially

affecting contract terms but providing little in the way of new

information. This characterization of SFAS 13 suggests that the

contracting hypothesis, and only the contracting hypothesis, is

relevant in the context of the Pfeiffer and AAM studies. It is

possible, however, to interpret all the evidence which links lessee

accounting pronouncements with negative share price behavior as being

consistent with either hypothesis, or with some combination of the

two, depending on what assumptions one makes about the mechanism by

which information is impounded in prices.

For example, Ohlson (1979) assumes a market composed of rational,

wealth maximizing investors having homogenous beliefs based on

identical access to information. In this setting, he considers the

effect of a change in the disclosure environment (e.g., a shift to a

finer set of disclosures regarding leasing activities). A basic

result of his analysis is that, if the change has no impact on firms'

production-investment-financing programs, the average price effect of

the change (over a large number of securities) should be essentially

zero. This is because pre-change security prices will have been based

on unbiased estimates of all price-relevant security character-

istics.21 If it is believed that "real" market behavior approximates

that which would obtain under these idealized conditions, it becomes

difficult to reconcile the Ro and MHR results (and virtually

impossible to reconcile the Pfeiffer results) with the information

hypothesis. However, if one interprets the various events leading to

the adoption of APB 31, ASR 147, and SFAS 13 as signalling increases

in the probability that lease capitalization would ultimately be

mandated, then the contracting hypothesis -- in which an impact on

production-investment-financing programs is explicit -- provides an

appealing explanation of the observed results. In particular, the

association between the magnitude of share price decline and the

materiality of leasing activities makes sense under the contracting

hypothesis since, other things equal, the more material a firm's

leasing activities, the greater the potential impact on contract


The characteristics of Ohlson's market represent one extreme of a

spectrum of possible characteristics. At the other end of this

spectrum, one might envision a market composed of investors who differ

widely in access to information, in information processing capabil-

ities, and therefore in beliefs about the price relevant attributes of

each security. If these differences were sufficiently pronounced,

security prices could depend on the disclosure environment in a

systematic way. A change in disclosure policy might then create

conditions under which, for example, investor class A could profit at

the expense of investor class B by adopting trading strategies which
anticipate the effects of the change on B's beliefs.22 In such a

market, the information hypothesis could conceivably explain the

results obtained in the various studies described here, even if

contracting effects were totally absent.

If the market is indeed "segmented" in the manner just described,

however, it is difficult to make other than a non-directional

hypothesis concerning share price impact. In order to go much further

than this, one would require a fairly detailed specification of the

actual structure of the market. Since the focus of the present study

is upon the extent to which the observed pattern of share price

movements is consistent with that which could be expected under the

contracting hypothesis, a market of the "Ohlson" type is assumed as a

maintained hypothesis throughout the study. The immediate logical

impact of this assumption is to rule out the information hypothesis as

an explanation of observed share price movements. In practical terms,

the ability to disregard the information hypothesis greatly simplifies

matters of research design. Moreover, the assumption is not

unreasonable, in that very few studies in the accounting and finance

literatures have produced results which actually contradict the notion

of an "Ohlson"-type market. In the final analysis, however, the

making of such an assumption means that any conclusions regarding the

aptness of the contracting hypothesis as an explanation of share price

movements must be tempered by recognition of the possibility of

simultaneous "pure information" effects.

2.5 Summary

The present chapter provides a description of the 1973-1976

lessee accounting standard-setting process, a discussion of the

financial statement impact of SFAS 13, the eventual output of that

process, and a critical review of four empirical studies which have

attempted to determine whether lessee equity prices were significantly

influenced by the process. It is concluded that the results of the

four studies, taken jointly, are at least consistent with the

hypothesis of significant influence. It is also pointed out that the

major focus of these studies is on whether, rather than why, such

influence was exhibited. A primary intention of the present study is

to examine the consistency of the data with one possible explanation

of why share prices should be influenced by the adoption of accounting

standards which appear, to all intents and purposes, to have been

largely cosmetic. This explanation, referred to as the "contracting

hypothesis," will be described in greater detail in the chapter which


2.6 Notes

1. Using data from a source quoted in Bowman (1978) it can be
calculated the annual revenues to the leasing industry
(approximately $10 billion in 1977) grew at a compound annual
rate of about 25% over the period 1950-1977.

2. See Cary (1949), Gant (1959), Vancil and Anthony (1959), Dieter

3. The 1962 study by Myers was sponsored by the Research Division of
the American Institute of Certified Public Accountants, and
published as Accounting Research Study No. 4. Myers' broad
conclusion is stated succinctly: "To the extent . that
leases give rise to . property rights, those rights and
related liabilities should be measured and incorporated in the
balance sheet (Myers, 1962, p. 4). This conclusion was rejected
by the APB in Opinion No. 5, apparently on the theory that a
"property rights" approach to lease accounting would be
inconsistent with accepted modes of accounting for other types of
executory contracts.

4. According to the task force which prepared the 1974 FASB
discussion memorandum on lease accounting, capitalization under
Opinion No. 5 was in practice largely limited to those leases
securing Industrial Development Bonds.

5. MHR presents the results of a detailed comparison of pre- and
post-APB 31 lease disclosures for a random sample of 314 firms.
An interesting aspect of these results is that, while the level
of disclosure prior to APB 31 was low for the average lessee
firm, those firms for which leasing activities were "most
material" tended to disclose the most detailed supplementary data
concerning leasing activities. Thus, it is not at all clear that
such disclosures were generally inadequate prior to APB 31.

6. The "financing lease" distinction was required to be made if
gross annual rental expense exceeded one percent of consolidated
revenues. The pro forma income disclosure was required if either
(a) the present value of noncapitalized financing leases equalled
or exceeded five percent of the sum of long term debt,
stockholders' equity, and the said present value, or if the
impact of the capitalization on average net income for the most
recent three years would have equalled or exceeded three percent.

7. Consider a lease in which the lessor retains title to the leased
property upon termination. Holding constant (a) the ratio of
residual value to the fair value of the property at the inception
of the lease, and (b) the yield rate implicit in the lease, the
substitution of an "investment recovery" condition for a "nominal

residual value" condition would make it more (less) likely that a
relatively long term (relatively short term) lease would be

8. A firm which makes all payments of principal and interest
according to the schedule established in the loan agreement, but
which violates one or more other covenants, is said to have
committed a "technical violation." However, many such violations
-- including violations of negative covenants involving
accounting numbers -- are "events of default" in the legal sense.
Thus, "technical violations" should not be regarded as mere

9. Of the 282 letters of comment received by the Board in response
to the second exposure draft, 121 were direct submissions by
lessee firms. The remaining 161 letters were submitted by
accounting firms and organizations, various trade and
professional associations, law firms, and lessors. Five lessee
firms supported the second exposure draft generally ard the
retroactive capitalization provision in particular.

10. SFAS 13 was not the Board's final word on the subject of lease
accounting. As of this writing, the Board has issued 7
amendments to and 6 "interpretations" of the standard. The SEC
also (effectively) amended SEAS 13 by moving forward the required
date for retroactive application (see ASR 225 and ASR 235, both
issued during 1977).

11. See Dieter (1979), Dieter and Wyatt (1980), or any intermediate
financial accounting text of current vintage.

12. In this context, "B" is the market model slope parameter which,
in the capital asset pricing model of Sharpe (1964), is a measure
of the riskiness of the related security or portfolio.

13. In ASR 225, the SEC required that SFAS 13 be implemented with
full retroactive application in filings with the Commission
subsequent to December 24, 1977.

14. This second assumption seems to be based, at least in part, on
previous work by Rosenberg and Marathe (1975), who showed that
estimates of systematic risk conditioned on a large number of
such descriptors reflect subsequent return behavior more clearly
than estimates based solely on the behavior of past returns.
However, Pfeiffer provides no rationale for the particular set of
descriptors used in his study.

15. See Roll and Ross (1980) for a review of the literature on this

16. See Sunder (1980) for a review and for empirical estimates of the
step variance of B, calculated under the assumption that
follows a random walk.

17. Black and Cox (1976) show that if the firm pays continuous
dividends, and if the instantaneous dividend is determined as a
constant proportion of firm value, then n will be decreasing
"almost everywhere." s

18. In order for this suggested behavior to unambiguously follow from
the Galai and Masulis results, 6 must be stationary. Moreover,
other implicit arguments of (6) -- for example, the riskless
interest rate, the variance of asset returns, and the required
level of payments to creditors -- must remain constant.

19. For example, by December 1974, the NYSE Composite Index had
fallen to approximately half of its January 1973 level.

20. At the firm level,

ut = R. a. 6.R ,
]t jt ] ] mt
where R and R are the realized returns, in week t, on
t j mt
security j and on the CRSP equal-weighted portfolio,
respectively. The ordinary least squares parameters a. and 8.
were estimated using 50 alternate weeks of returns during the 100
weeks preceding the first trading week of 1973.

21. The prices of individual securities would of course be expected
to fluctuate if the "disclosure environment" change truly alters
the available information set. However, these fluctuations
should be independent across securities.

22. The behavior of prices in markets of this type have been modelled
by Goldman and Sosin (1979) and by May and Noreen (1981).



3.1 Introduction

The ownership structure of a corporate firm can be viewed as a

configuration of contracts which allocates the risks and benefits of

the firm's investments among various claimants. The value of a given

claim depends not only on the characteristics of those investments,

but also on the particular contract terms which determine the

(state-contingent) payoff stream to which it is entitled. Each class

of claimants, motivated by economic self-interest, can be presumed to

have adopted wealth maximizing strategies with respect to the contract

terms by which it is affected. Thus, a potential for wealth transfer

among claimholders is created when existing contract terms are

perturbed through the action of some external agency. Moreover, the

value of the firm as a whole should decline if transactions costs are

incurred either in renegotiating affected contracts to eliminate, or

in modifying production-investment-financing programs in order to

adapt to, the effects of the perturbation.

Given this view of the firm, a mandated change in accounting

method can affect security values by operating on contract terms which

are stated with reference to accounting numbers measured in accordance

with generally accepted accounting principles. This is the essence of

the contracting hypothesis. As was seen in Chapter 2, lessee

opposition to lease capitalization and to retroactive application was

routinely supported by arguments which were simply variations on this

theme. Partly for this reason, but more generally because it offers a

plausible explanation for an adverse share price reaction to an

apparently "cosmetic" change in financial accounting policy, the

contracting hypothesis as it relates to loan contracts is a major

focus of the present study. Its various ramifications are discussed

in detail in the remainder of this chapter. Section 3.2 is devoted

to a review of the basic existing empirical study of loan contract-

related share price effects. Section 3.3 provides some general

information concerning the use of accounting constraints in loan

contracts, together with an example of the effect of retroactive lease

capitalization on the firm's position with respect to such

constraints. In section 3.4, a model is developed of the cross-

sectional distribution of returns arising from the adoption of an

"adverse" accounting change. This model is summarized as a series of

testable hypotheses in section 3.5.

3.2 The Leftwich Study

The first empirical investigation of contracting share price

effects is provided by Leftwich (1981), who noted that an accounting

standard which reduces the permissible set of financial accounting

techniques also narrows the opportunity set faced by shareholders who

are party to "accounting-based" contracts. The adoption of such a

standard should therefore decrease the aggregate value of share-

holders' claims by an amount which, as discussed above, includes both

wealth transfer and transactions cost components.

Leftwich's study focused on loan contracts which include

provisions stated in terms of accounting numbers. In addition to

making the general argument that share prices are adversely affected

by an "option reducing" accounting standard when such contracts are

outstanding, he also reasoned that the adverse effect should be

mitigated to the extent that the firm's debt is directly placed,

callable, or convertible. If debt is directly (i.e., privately)

placed, the number of creditors involved is small, and the direct

costs of renegotiating the loan contract to eliminate the effects of

the accounting rule change should be relatively low. In the case of

debt which is publicly offered but callable, the adverse price impact

of an accounting rule change is limited to the cost of refunding the

debt. And finally, the wealth transfer potential associated with such

a rule change is more limited when the value of a loan depends in part

on the presence of a conversion privilege.

These ideas were explored through an examination of the behavior

of stock returns surrounding "significant events" during the period

leading to the adoption of APB Opinion No. 16 -- Accounting for

Business Combinations (Accounting Principles Board, 1970). The intent

of APB 16 was to limit the use of pooling of interests accounting, and

to require the use of purchase accounting, in many merger situations.

Leftwich argued that, relative to purchase accounting, pooling of

interests accounting has a beneficial effect (from the shareholders'

perspective) on those financial statement numbers and relationships

typically used in restrictive loan covenants. Thus, the adoption of

an accounting standard which narrows the range of conditions under

which pooling is permissible should have been economically harmful to

the shareholders of "merger-prone" firms.

When averaged over a large sample of firms having some history of

merger activity, market model prediction errors were found to differ

significantly from zero during (two week) periods surrounding nine of

twenty two APB 16-related "events." For each of these nine events,

prediction errors were then regressed cross-sectionally on measures of

publicly offered, privately placed, callable and convertible debt, and

on a measure of firm size. Generally, the regression results suggest

that the magnitude of the adverse price response associated with APB

16 varied directly with the extent to which firms were debt financed

and inversely with firm size. Beyond this, it appears that the

direct placement, callability, and convertibility hypotheses were not

particularly supported by the data, leading Leftwich to conclude that

"contracting cost theory provides, at best, an incomplete explanation

of the observed abnormal performance of firms during the APB

deliberations on business combinations" (Leftwich, 1981, p. 33).

Whether this conclusion is warranted or even meaningful is open

to question. If a complete explanation is taken to be one which

perfectly models the non-systematic component of security returns at a

given point in time, or even one which accounts for a large proportion

of the cross-sectional variation observed, then such a conclusion

could have been stated with a high degree of confidence even in the

absence of the accompanying research. On the other hand, if

completeness is defined to exist only when a particular set of

predictions is fully supported by a particular set of results, then

"incompleteness" may be inferred either because the theory itself is

false, or because the experimental design is not sufficiently fine to

allow detection of the predicted effects. The fact that the Leftwich

results do not support the direct placement, callability, and

convertibility predictions may be due to the latter problem in

particular. For example, suppose that a number of firms are equally

"bound" by accounting-based loan covenants, and that their financial

statements would be identically affected by the provisions of APB 16.

Also suppose that the firms are identically levered, where leverage is

measured as the ratio of the book value of debt to the market value of

equity. If contracting costs are indeed significant, one would then

expect the data to support second-order predictions such as those

concerning direct placement, callability, and convertibility. These

rather strong conditions are clearly absent from the Leftwich study,

however, and hence it is difficult to interpret the regression

coefficients estimated therein as straightforward tests of the

predictions in question. For this reason, it seems more useful to

suggest that Leftwich's results provide tentative support for the

hypothesis that contracting effects of detectable magnitude do occur,

than to conclude that, as an explanation of abnormal performance, the

contracting hypothesis is incomplete.

Having reached this assessment, it becomes appropriate to search

for a context other than APB 16 in which the contracting hypothesis

may be further explored. It can be argued that, for a number of

reasons, the lessee accounting standard developments of the 1973-1976

period provide such a context. First, it seems clear that for many

firms, the impact of lease capitalization on financial statement

numbers and relationships -- and therefore the potential impact on

contract terms -- was far from trivial. Second, under the final

version of SFAS 13, the financial statement impact of lease capi-

talization was intensified due to the retroactive application

provision, a feature that was completely absent in the case of APB 16.

Third, the standard setting process leading to SFAS 13 included a

major policy reversal on the matter of prospective v. retroactive

application, which suggests a similar reversal in share price

behavior. And finally, because the financial statement impact of

lease capitalization is more easily proxied than is the impact of a

standard which precludes (to some uncertain degree) the future use of

pooling of interests accounting, the "lessee accounting context"

provides a better opportunity for predicting the cross-sectional

incidence of contracting share price effects.

The aim of the remainder of this chapter, then, is to develop

testable hypotheses concerning the impact of APB 31, ASR 147, and SFAS

13 on share prices, given the contracting hypothesis. As in the

Leftwich study, the discussion will focus on loan contracts, and in

fact it will be assumed that the effect of these standards on cther

classes of contracts having accounting-based terms -- e.g., certain

incentive compensation plans for top management -- was negligible. A

general description of the use of accounting restrictions in loan

agreements is presented in the section which follows.

3.3 Accounting-Based Loan Covenants and Lease Capitalization

Jensen and Meckling (1976) have suggested that restrictive loan

covenants are negotiated by shareholders in order to control the

natural conflict of their own interests with those of creditors,

thereby reducing the agency costs associated with debt financing. If

penalties for violation are sufficiently large, such covenants place

bounds on the set of actions shareholders may take in transferring

wealth from creditors to themselves. These actions could include (a)

the payment of dividends which are financed either through a liquida-

tion of existing investments or a reduction in planned investment; (b)

issuance of additional debt whose seniority equals or exceeds that of

existing obligations; or (c) alteration of the returns distribution of

the firm's assets, either through asset substitution, manipulation of

new investment, or merger. If the shareholder-creditor conflict of

interests is imperfectly controlled through contract provisions, then

rational creditors will rely on "price protection" to insure an

expected return which is consistent with the level of risk accepted.

And, once a loan contract has been negotiated, rational shareholders

can be expected to adopt a set of investment and financing programs

which maximize their own wealth, subject only to the constraints

imposed by the contract.

A number of common loan covenants restrict shareholder activities

without reference to accounting numbers -- for example, virtually all

debenture contracts include clauses prohibiting the mortgage or sale

and leaseback of those assets owned by the firm at the time the

debentures are sold. On the other hand, many standard types of loan

covenants are explicitly based on financial accounting numbers. Some

of these place absolute floor or ceiling limits on accounting numbers

or on relationships among accounting numbers. Others permit share-

holders to undertake actions potentially detrimental to creditors only

if conditions are met which depend directly on such numbers or

relationships. Although by no means an exhaustive list, three of the

most frequently encountered types of accounting-based loan covenants

are described here to illustrate ways in which accounting numbers are

used to limit the activities of shareholders.

Dividend pool. Many loan indentures limit the sum of dividends

and net amounts expended for share repurchase to a fraction of

accounting earnings accumulated since the inception of the contract.

Thus, the ability of the firm to distribute assets to shareholders

depends on the level of its accounting earnings and on the acceptable

range of accounting techniques for computing earnings.

Restrictions on merger activity and the sale of additional debt.

Many indentures permit these activities only if, on a pro forma basis,

the firm's "assets in place" provide some acceptable level of coverage

for funded debt. Accounting numbers are normally used as proxies for

this relationship. For example, mergers might be permitted only if,

on a post-merger basis, the ratio of net tangible assets to funded

debt exceeds some constant. When such restrictions are present, the

firm's ability to alter its asset or capital structures depends on

accounting definitions of assets and liabilities.

Working capital maintenance. Frequently, an indenture will

define a minimum level of working capital which must be maintained

while the loan is outstanding. The firm's ability to meet this

restriction depends on, among other things, accounting definitions of

assets and liabilities and on acceptable methods of segregating these

items into their current and noncurrent components.

A very detailed discussion of the use of accounting numbers in

loan agreements may be found in Commentaries on Indentures (American

Bar Foundation, 1971, hereafter Commentaries). Evidence from this

source, from a number of private placement agreements reviewed in

Leftwich (1980), and from a selection of "public" indentures reviewed

by the present author, suggests that "generally accepted accounting

principles" constitute the basic measurement framework for loan

covenants which involve accounting numbers.6 That is, GAAP is binding

on the parties except where modified through specific contract

provisions. Moreover, the same sources of evidence suggest that the

firm's position with respect to accounting-based loan conditions is

generally determined on the basis of GAAP in existence as of the

moment of calculation. It is for these reasons that a policy-making

body such as the APB or the FASB may inadvertently alter the

provisions of loan agreements by expanding or contracting the set of

accounting techniques which are permissible for financial reporting


The lessee accounting standards of 1973-1976 reduced the set of

accounting techniques available to lessees, and the impact of the

reduction on accounting numbers and relationships frequently used in

loan contracts was (from the shareholders' perspective) clearly

adverse. For example, with respect to the typical loan covenants

described above, retroactive capitalization of existing leases has the

immediate effect of reducing working capital, the dividend pool, and

the ratio of net tangible assets to funded debt. And, to the extent

that tightened capitalization conditions are binding on future leases,

a firm's rate of growth with respect to typical accounting-based loan

restrictions will also be reduced.

The effect of retroactive capitalization can be clarified with a

simple example. Assume that a lease capitalization requirement has

been imposed, effective as of the end of Year 5, and consider a

hypothetical firm having the pre-capitalization balance sheet shown


Pre-Capitalization Balance Sheet

End of Year 5 -

Current assets $ 200,000 Current liabilities S 75,000
Non-current Debentures (non-
assets (net) 800,000 current portion) 300,000
Capital stock 200,000
Total assets $1,000,000 Retained earnings
restricted as to
dividends 200,000
unrestricted 225,000

Total equities $1,000,000

This firm pays dividends which increase moderately from year to year;

the Year 5 dividend was $45,000. The contract between debenture-

holders and shareholders includes four accounting-based restrictions

which are binding so long as the bonds are outstanding. First,

working capital of at least $100,000 is to be maintained at all times.

Second, the current ratio must be maintained at a level which equals

or exceeds 2.0. Third, dividends are limited in total to $50,000 plus

earnings accumulated subsequent to the contract date (resulting, by

assumption, in "unrestricted" retained earnings of $225,000). And

fourth, both mergers and additional borrowing are prohibited if, as

the result of such activities, the ratio of net tangible assets to

funded debt would be less than 2.0. And finally, two non-accounting

restrictions have been included in the agreement. The first of these

places limits on "off-balance-sheet" financing: aggregate annual

rentals under leasing arrangements are limited to $50,000. Second,

proceeds from the mortgage or premature disposal of existing fixed

assets must be used to liquidate outstanding debentures.

At the beginning of Year 1, the firm in question leased equipment

with a fair market value of $200,000. On that date, the equipment had

an expected life of ten years and negligible expected residual value.

The noncancellable lease term is ten years, and a payment of $32,549

is due to the lessor at the end of each year. The interest rate

implicit in the lease is 10%. And at the end of the lease term, the

firm may purchase the equipment from the lessor for an amount equal to

its appraisal value on that date.

Until the present time (i.e., the end of Year 5), this lease has

been accounted for by charging $32,549 annually against revenues.

Some simple calculations will show that, if the lease is capitalized

retroactively at the end of Year 5, $100,000 will be added to

non-current assets (assuming straight line depreciation) and $123,387

will be added to liabilities. Of the latter, $20,210 and $103,177

will be classified as current and non-current, respectively. And

unrestricted retained earnings will decline by $23,387, the amount by

which the increase in liabilities exceeds the increase in assets.

Using these numbers, the firm's pre- and post-capitalization position

with respect to the loan restrictions described above can be

constructed, as indicated below.

Pre Post

Current ratio 2.67 2.10

Working capital $125,000 $104,790

Unrestricted retained earnings (URE) $225,000 $201,613

URE/Current dividend 5.00 4.48

Net tangible assets/funded debt 3.08 2.49

It should be noted that this example, while hypothetical, is in

no way extreme or "pathological." The set of accounting-based loan

restrictions assumed is representative of those faced by a broad range

of firms, and the assumed scope of leasing activities is not atypical.

The example illustrates the point alleged earlier, that the effect of

capitalization is, from the shareholders' point of view, unambiguously

adverse along every accounting dimension of the loan agreement. If

the policy change from "rental" accounting to capitalization was

largely unanticipated at the moment of contract negotiation -- that

is, if the levels of the various accounting restrictions were set on

the assumption that rental accounting would be used over the life of

the contract -- then capitalization has the effect of "tightening"

those restrictions.

Under the rationality assumptions which are characteristic of all

economic analysis, one must presume that the shareholders' precapi-

talization position with respect to loan restrictions was optimal. In

this context, an optimal position would consist of a contemplated set

of future investment and financing activities which maximizes current

shareholder wealth, given existing loan restrictions, penalties for

violation, and the current state of the firm. Since a policy change

to capitalization has the effect of tightening loan restrictions, it

seems clear that an optimal pre-capitalization position must, ceteris

paribus, be sub-optimal on a post-capitalization basis.

It was pointed out earlier that a number of Leftwich's

hypotheses, while themselves technically consistent with the general

notion of contracting share price effects, were apparently not

supported by the APB 16 data. It was also argued that these results

should not be interpreted as grounds for rejecting the contracting

hypothesis, and that further investigation in a different context

would be appropriate. Actually, it seems highly probable that such

effects do in fact occur. The real questions at issue are those of

magnitude and generality. Do circumstances exist in which the

contracting costs imposed by a mandated accounting change amount to a

non-trivial proportion of equity value? And are these circumstances

sufficiently pervasive to allow the related share price effects to be

detected? What will be attempted here, in order to begin to answer

these questions, is the development of a model of the share price

impact of a mandated change in lessee accounting rules when

accounting-based loan restrictions are in force. First, some recent

results from the finance literature are used to provide a very basic

characterization of a loan contract, and the role of accounting

restrictions in such a contract is considered. This characterization,

together with some severely simplifying assumptions, is used to

propose a model of the cross-sectional distribution of share price

effects resulting from the adoption of an adverse accounting change.

Attention is then devoted to the effect of relaxing the initial

simplifying assumptions. The end result of this process consists of a

set of hypotheses which relate a small number of observable variables

to the distribution of returns at the time of an accounting policy


In the discussion which follows, the phrases "share price

impact," "share price effect," etc. are meant to be understood in

relative terms. That is, the contracting share price effect due to a

mandated change in accounting practice is defined as the ratio of any

contract-related costs incurred by shareholders to the pre-change

market value of equity. It is important that share price effects be

defined in this way, first to permit comparisons across firms, and

second because the empirical analysis in this study will focus on

returns rather than on absolute price changes.

3.4 Factors Influencing the Cross-Sectional Distribution of
"Contracting" Share Price Effects

Consider the following characterization of a loan contract.

Shareholders borrow B dollars from creditors to be invested, along

with S dollars of personal funds, in specific assets having a current

market value V = S + B. Over the life of the contract, shareholders

are permitted to receive continuous dividends amounting to 6V per unit

time. When the contract terminates at time T, creditors are to re-

ceive min(C, VT) dollars, and shareholders are to receive max(O, V -C)

dollars. C is a fixed amount. The (ex-dividend) terminal value of

the firm's assets, V is assumed to have a lognormal distribution,

the parameters of which are known both to shareholders and to


If asset and financial markets are frictionless, and if it can be

assumed that shareholders will not deviate from the investment/

financing program specified in the contract, then the exact value of B

-- the amount that creditors are willing to lend -- can be determined

(see Black and Scholes (1973), Merton (1974), Galai and Masulis

(1976)). Generally speaking, B is a function of V, C, 6, and o2, the

last of which is the variance of dV/V, the instantaneous rate of

return on the firm's assets. Tn somewhat different terms, if one

defines B = CD, so that D is the current value per dollar of promised

payment to creditors, then

D = f (V, C, 6, o2), (1)


DV > 0; D C D6' D 2 O,

where subscripted values of D represent partial derivatives. Once C,

6, and o2 have been fixed, creditors and shareholders will "share" all

subsequent changes in V, with dB = CD dV and dS = (1-CDV)dV = dV.

This means that a unique level of B (and therefore, of S) is

associated with each possible level of V. A representative plot of B

and S against V is shown in Figure 3-1. For very large values of V, B

-1 -1
will approach R1 C and S will approach V-R C, where R1 is the
f f f

current price of a riskless dollar to be delivered at time T. In this

region, Dv and SV will be very close to zero and one, respectively.

For very small values of V, both B and S will approach zero, and D

and S will tend respectively toward (1/C) and zero.

This model of a loan contract, which has frequently been

presented in the finance literature, differs from a typical loan

agreement in two important respects. First, and most fundamentally,

the partial derivatives of D indicate that shareholders will have an

incentive to deviate from 9*, the promised investment/financing

program, once the loan is in place. In particular, any action which

preserves V and simultaneously increases, C, 6, or G2 will increase

the wealth of shareholders at the expense of creditors. Realizing

this, creditors will be unwilling to lend unless the promises of

shareholders can be adequately enforced. Enforcement is a simple

matter if the value and variance rate of the firm's assets, the amount

promised to creditors, and the amount disbursed as dividends can be

observed at each point in time. With this information, creditors can

determine whether shareholders have complied with contract terms, and

the contract can be made self-enforcing if a sufficiently large


I [
S \t


\ \ \ \

\ \
\ \ \
\ \ \
\ \ \N
\ \ \ N

penalty is imposed for non-compliance. In point of fact, however,

neither V nor o2 -- nor, by extension, 6 -- is directly observable.

They are subjective quantities which must be appraised or assessed,

and about which opinions are likely to differ where a conflict of

interests exists. Thus, in a "real" contract, both shareholders'

promises and the method of their enforcement must depend on something

other than the theoretical variables which determine the value of the


The second of the differences mentioned above arises because

"fixing" the investment/financing program may not be efficient. If

the firm's investment opportunity set changes over time -- for

example, if new projects with positive net present value become

available -- then it may be possible for shareholders to increase the

value of equity without impairing the value of creditors' claims. In

principle, this problem is easily remedied: shareholders simply

promise not to depart from a given investment/financing program 8* if,

by so doing, creditors' interests would be adversely affected. In

practice, of course, one is still faced with the question of how such

a promise is to be enforced.

In order to make a real debt contract palatable to lenders, then,

shareholders must include a system of promises (covenants) which can

be monitored (preferably at low cost) by reference to observable

variables, and which provide a reasonable degree of control over the

wealth transfer problems associated with the partial derivatives of D:

asset substitution, claim dilution, and excessive disinvestment. In

addition, from the standpoint of efficiency, one would also expect the

system of covenants to allow shareholders some flexibility in the

choice of investment/financing programs.

In many cases, the system of covenants chosen to achieve these

purposes is based largely on financial accounting numbers computed in

accordance with (but sometimes allowing clearly specified departures

from) generally accepted accounting principles. These numbers are

observable and, due to the existence of a well organized auditing

profession, can be monitored by reasonably independent outsiders at

reasonably low cost. Moreover, it is apparent that relationships

among accounting numbers are used to control or eliminate opportuni-

ties for "creditor to shareholder" wealth transfers. In the system of

covenants described above in connection with the numerical leasing

example, for instance, the fact that proceeds from the disposal of

fixed assets must be used to liquidate existing loans means, in

effect, that "asset substitutions" must be financed using new capital.

The "claim dilution" problem is controlled by disallowing the sale of

new debt if, as a consequence, total debt would exceed a specified

proportion of tangible assets. The "excessive disinvestment" problem

is controlled by establishing dividend limits which depend on the

growth in net assets. Finally, it is equally apparent that systems of

covenants can be designed which protect creditors' interests and

which, at the same time, allow shareholders considerable latitude in

the choice of investment/financing programs. Looking once more at the

example just cited, it can be seen that on the investment side, no

restriction is placed on the composition of the firm's asset

portfolio. And on the financing side, only an overall constraint on

the relationship between tangible assets and funded debt need be met.

Subject to this constraint, the financing arrangements for particular

projects are unrestricted (except, as noted above, in the case of

asset substitutions).

This discussion suggests the following characterization for a

typical loan contract. Shareholders, contemplating a particular

investment/financing program 6*, promise to pay creditors min(C,V ) at

time T. In addition to this promise, the contract also includes a

system of covenants based on observable variables, and a system of

penalties which are of sufficient magnitude to guarantee shareholder

compliance. Now, define AI as net new funds invested in the firm

after the inception of the loan agreement, and DV* as the partial

derivative associated with 0* at the point V* = V-AI. The system of

covenants is designed to have the approximate effect of insuring that

dD dv dl
d > DV* (d --) (3)

This contract has two desirable properties. First, the

constraint established by the covenants means that, regardless of the

manner in which 6 is altered subsequent to the loan date, changes in

firm value (other than those due to new investment) will be "shared"

by creditors exactly as if 6 has been fixed at @*. And second, the

contract provides appropriate investment incentives for shareholders.

Since S = V-CD,

dS dV dD dC
C D--

dV dl dD dl dC
= (L c + ( D-) (4)
dO dr dO d4 9- (8)

Substituting the constraint on -* and rearranging and collecting terms


dS dI dC dV dI
( D - -- d d-) ( CDv) (5)

Wealth maximizing shareholders will require strict equality between

the left and right hand sides of (5). The LHS can be interpreted as

the change in the value of existing equity due to a small change d6 in

the firm's investment/financing program. Shareholders will find de

attractive only if the RHS of (5) is positive. If de is viewed as a

"project" to be accepted or rejected, then the RHS is simply the

product of the project's net present value and a positive number less

than one. Thus, it will be in the best interests of shareholders to

undertake projects which increase firm value and to reject those which

do not.

The "bottom line" effect of this contract can be expressed in the

following way. Regardless of the manner in which shareholders changeS

subsequent to the contract date, the system of covenants guarantees

(at least approximately) a minimum value per dollar of C at each

possible level of V-AI. If the actual value per dollar of C is

allowed to fall below the required minimum, which can be denoted as

D*, severe penalties are imposed, On the other hand, if the actual

value per dollar of C is allowed to rise above D*, shareholder wealth

can be increased via further modifications of 8. Thus, if share-

holders consistently adopt programs which maximize equity value, one

would always expect to observe

S = V CD*, (6)

where D* is determined jointly by the level of V-AI and by the col-

lection of covenants included in the contract.

One assumption explicit in this characterization of a loan

contract is that both real assets and financial claims can be traded

at zero cost. The addition of three further assumptions will make it

possible to develop a very simple and unambiguous model of share price

reaction in the face of an adverse accounting change. First, both

within and across firms, all loan contracts outstanding are assumed to

include equally restrictive accounting constraints. Second, it is

assumed that no outstanding contract includes a call provision. And

third, maximization of the value of a given loan is assumed to be

consistent with maximization of creditor wealth.

Within this scenario, a mandated accounting change which tightens

loan restrictions can be thought of as increasing D in (3) or

equivalently, as increasing D* at each level of V-AI. Denoting the

post-change required minimum for D as D**, the "contracting" effect of

the change on equity value can be written (in returns form) as

S- (D** D*) (7)

In absolute terms, an adverse accounting change has the effect of

transferring C(D**-D*) dollars from shareholders to creditors. The

specific method by which this wealth transfer is accomplished will

depend on the circumstances faced by a particular firm, but the set of

available methods is indicated by the partial derivatives of D.

Ceteris paribus, D can be increased by reducing the rate at which

dividends are paid, by redeeming some portion of outstanding debt, by

substituting "low variance" projects for existing or contemplated

"high variance" projects, or by new infusions of equity capital.

Another possibility is that the contract itself may be renegotiated to

eliminate the effect of the accounting change, with compensation to

creditors in the form of an increase in promised payment or an

acceleration of loan maturity.

Expression (7) suggests a qualitative description of the cross-

sectional distribution of AS/S. First, if the impact of the change on

those accounting numbers and relationships used to frame loan

covenants is slight, then (D**-D*) should be close to zero and AS/S

should be small. Likewise, if amounts promised to creditors are small

in relation to pre-change equity value, then C/S is close to zero and

AS/S should again be small. If these conditions are individually

sufficient for a "small" share price effect, their complements must be

jointly necessary for a "large" share price effect. And since by

assumption creditor wealth is maximized when the value of a given loan

is maximized, their complements must be jointly sufficient for a

"large" share price effect. More generally, at a given level of C/S

(which can be interpreted as an index of leverage), the magnitude of

AS/S should vary directly with (D**-D*), and therefore with the

magnitude of the impact of the accounting change on financial

statement numbers and relationships.

An implication of this description is that the contracting

hypothesis can be tested using an experimental (or quasi-experimental)

design which allows for a "leverage by financial statement impact"

interaction. Before this idea is addressed, however, it must be

admitted that the four assumptions involved above are hardly realis-

tic, and that the proposed cross-sectional distribution of share price

effects may not be appropriate once they are relaxed. Thus, some

consideration should be given to the role of, and the effect of

relaxing, each assumption.

The "zero transactions costs" assumption insures that none of the

firm's resources need be expended in adapting to a mandated accounting

change, and thus limits the shareholders' loss to an amount w =

-C(D**-D*). As indicated in Figure 3-2, an adverse accounting change

can be thought of as increasing D* (and therefore the value of debt,

B = CD*) at each level of V, in effect shifting the curves which

relate B and S to V upward and downward, respectively, to B' and S',


B'(V) B(V) = -(S'(V) S(V)) = C(D**(V) D*(V)) = w(V). (8)

That is, w(V) is the vertical distance between the pre- and post-

change curves. Thus, if transactions costs are zero and V = Vo, the

value of creditors' claims should rise to B' = B + w, and the value
o o

of equity should fall to S' = S w.
o o

When transactions costs are present, the situation must be viewed

somewhat differently. In this case, tightened loan restrictions




I: \ o

\ v" .
\ -

I I \ 2"

\ I I 4
S\ I
I I o 4
S 1 \ ,s

1 1
\I j \ .> \ f

hI .f'
\ I i*

I I I iu

I I i I .
SI 'a>
\ I *

\ I \ I\ \
i I\ I
\ -> X ^' ^ H ^
v J^ V\3 9 ii

0 0

0 0
0n c0i

require that the post-change shareholder-creditor relationship be

"located" along B' and S', but the process of "getting there" may

require an expenditure of the firm's resources. When creditors'

claims are risky, changes in firm value are shared by creditors and

shareholders. The relevant question is this: does the cross-

sectional distribution of AS/S have the same properties in the

presence of transactions costs as that hypothesized in the case where

such costs are absent? In particular, should one still expect to

observe the proposed "leverage by financial statement impact"


An answer to these questions can be approached by noting that

shareholders, in order to "move" from B to B', may adjust investment/

financing programs using methods which are strictly unilateral (e.g.,

issue new equity, alter asset composition, decrease dividend rate) or,

alternatively, using bilateral methods which require creditor cooper-

ation (e.g., early redemption of loans, renegotiation of contract

terms). Let the minimum transactions costs associated with unilateral

and bilateral actions be represented as t and tb, respectively. If

shareholders choose to adjust 8 using strictly unilateral methods,

firm value will decline by AV = tu, equity value will decline by As =

(w + at ), and the value of creditors' claims will increase by AB =
(w (l-a)tu). The proportion of tu absorbed by shareholders, a, is

related to the slope of the curve S. Thus, when S is very large (very

small) relative to B, shareholders will bear virtually all (none) of

these transactions costs. In any event, since the adjustments

involved can be pursued without creditor cooperation, (w + t u) is the

maximum loss that can be imposed on shareholders as a result of the

accounting change in question.

Since only the "zero transactions costs" assumption has been re-

laxed at this point, it can also be argued that (w + at ) is the

minimum loss imposed on shareholders by a given adverse accounting

change. Shareholders will be unwilling to adjust using bilateral

methods unless the resulting equity value exceeds S-w-at Creditors,

on the other hand, will not participate in a bilateral solution unless

the resulting loan value exceeds B + w (1 a)t For example, sup-
pose that loans are to be redeemed at a price P with accompanying

transactions costs tb. Shareholders will require

V P -tb S w at Creditors will require P >B + w (l-a)tu.
b u u
These conditions can be written jointly as

0 < P-[B + w (1 a)t ] < t t. (9)
u u b

Basically, (9) says first that redemption is a possibility only if tb

is lower than the minimum transactions cost associated with unilateral

shareholder action, second that the cooperation of creditors nust be

secured by a positive "premium", and third that this premium must be

smaller than the achievable savings in transactions costs. Wealth

maximizing creditors will simply set a premium which differs from

(tu-tb) by a small number. The resulting wealth effects will be

(approximately) S = -(w + atu) and B = w (-a)tu + (tu tb). A

similar argument can be made in the case of loan renegotiation, so

that in general the wealth effects of an adverse accounting change can

be written gainn approximately) as

AS = -(w + at ) (o1a)

AB = w (1 a)t + max[O, t u t] (10b)

Equations (10) indicate that AS will be approximately the same whether

the response to the accounting change is "unilateral" or "bilateral",

since wealth maximizing creditors will be able to absorb any cost

savings achievable through bilateral action.

Thus, when the zero transactions costs assumption is relaxed, the

appropriate characterization (in returns form) of the "contracting"

share price impact of a mandated accounting change becomes

AS u C
= (D**-D*) (lla)

a 1
= t w (llb)
S u S

This means that the distribution of AS/S will have the same general

shape in the presence of transactions costs as in their absence if the

magnitude of (atu/S) is a constant or increasing function of the

magnitude of (w/S). This condition will obtain if one is willing to

make the plausible assumption that the dollar magnitudes t and w are

directly related. In this case, comparing the terms of (llb), t

increases as w increases. It can also be shown that, while a itself

approaches zero as S becomes very small, the ratio (a/S) becomes

infinitely large. That is, (a/S) increases with (1/S). Therefore,

the product (a/S)tu must increase as the product (1/S)w increases.

In summary, it is argued that when an adverse accounting policy

change is imposed, the cross-sectional distribution of AS/S should

have the same qualitative empirical properties whether transactions

costs are present or absent.

The second assumption included in the analysis above is that all

loan contracts, both within and across firms, contain equally

restrictive accounting constraints. Actually, a given firm may have

numerous loan agreements outstanding, which vary widely in the extent

to which they place restrictions on shareholder activities subsequent

to the contract date. And the general restrictiveness of contract

terms may vary from one firm to the next. This suggests that a factor

other than those already proposed (i.e., leverage and financial

statement impact) must be taken into account in describing the

distribution of share price effects following a mandated accounting

change. In particular, it is argued here that if the financial

statements of two firms are identically affected by an adverse

accounting change, shareholders of that firm whose accounting-based

loan covenants are most restrictive at the margin will experience the

largest wealth effect as a result of the change.

Unfortunately, measurement of the restrictiveness of loan

covenants at the margin is far from being a straightforward task. It

is virtually impossible to learn from publicly available sources the

precise terms of the possibly many loan contracts to which a given

firm is party. Moreover, even if these data could be obtained, it is

not clear that the various types of accounting constraints faced by an

individual firm can be aggregated to provide a quantitative index of

"restrictiveness" that would apply across firms. On the other hand,if

one is willing to assume a direct relationship between the

restrictiveness of a firm's "average" loan contract and its "most

restrictive" loan contract, a crude proxy for the latter can be

developed. While public data concerning the terms of particular

contracts is sparse, it is fairly easy to determine that proportion of

C represented by loans subject to no accounting-based constraints

whatsoever. Denote this proportion as X, with 0 < X < 1. One can

argue that when A is large (small), the average and marginal

restrictiveness of loan covenants is low (high), and the related share

price effect should be small (large).

It can also be argued that when the third assumption -- that of

"noncallability" -- is relaxed, the proportion A is important from

another perspective as well. Noncallability was initially assumed in

order to guarantee that creditors would be able to take full advantage

of the wealth transfer implied by an adverse accounting change.

However, most real loans appear to include some provision for pre-

mature redemption at the option of the borrower within a reasonably

short period after the contract date.0 In the context of the current

scenario, the "striking price" of such a call option would typically

be the discounted value of C, at the rate of yield in force on the

contract date, plus a penalty for early termination of the contract.

Thus, the effect of relaxing the noncallability assumption is to

increase the range of unilateral actions that shareholders may pursue

in responding to the accounting change. If it is momentarily assumed

that all outstanding loan contracts include accounting constraints

(i.e., A=0), that these loans can be called at an aggregate price P ,

and that the transaction itself entails costs tc, then shareholders

will prefer "calling" over other unilateral options whenever

(P B) + tc < w + at (12)

where other notation is as previously defined. If this condition is

met, (P B + t ) becomes the maximum loss that shareholders may
c c

sustain as the result of an adverse accounting change.

The magnitude of (P B) in (12) depends heavily on the spread

between current interest rates and those which prevailed on the

contract date. If current rates are relatively low, (P B) should

be close to zero. If current rates are relatively high, (P B) may

become quite large. Since interest rates were historically high

during much of the period to be examined in this study, it can be

argued that, at least under the conditions just stated, exercising the

option to call existing loans would tend to be a very costly response

from the shareholders' point of view, and therefore that relaxing the

noncallability assumption should have little real effect on the

analysis presented thus far. On the other hand, if X is large -- that

is, if most of a firm's loans are not subject to accounting

restrictions -- then calling those loans which are subject to

accounting restrictions may be the minimum cost shareholder response

to an adverse accounting change even during a period in which loans

are "trading" at a substantial discount.

The final assumption which requires further comment -- and

perhaps the most important of the four -- is that maximizing the value

of a given loan is, from the creditor's point of view, consistent with

wealth maximization. One might think of an adverse accounting change

as creating a wealth transfer potential which creditors may or may not

choose to realize. Wealth maximizing creditors would always choose to

realize this potential if the act of so choosing had no external

effects on other (present or future) contracts. Where external

effects are present, it is conceivable that the gain to be achieved by

realizing such a wealth transfer could be partially or completely

offset by losses connected with other contracts. In the latter case,

a rational creditor would be willing to renegotiate the terms of a

loan contract in order to neutralize the effect of the accounting

change, and the costs imposed on shareholders could, in the final

analysis, be no greater than the direct costs associated with the

renegotiation process. In short, the pattern of share price behavior

explicated in the preceding pages can be viewed as one which is

possible, but which will be guaranteed to materialize only if external

effects associated with the realization of "shareholder to creditor"

wealth transfers are generally negligible in magnitude.

One distinction that may be useful in this context is that be-

tween privately placed loans (which for present purposes can be

defined to include unregistered securities held by insurance companies

and other institutional lenders, and term loans provided by commercial

banks) and public debt issues (those registered under the Trust

Indenture Act of 1939). Loans in the latter category are more likely

to be widely held by creditors whose relationship with the firm is

essentially anonymous. In this situation, any external effects on

other contracts to which these creditors are (or will be) a party

should be minimal. Thus, "public" creditors should be willing to

fully exploit the wealth transfer potential inherent in an adverse

accounting change.

There is, on the other hand, a distinct possibility that private

creditors may not be so willing. For example, Zinbarg (1975),

apparently speaking in his capacity as an executive of the Prudential

Insurance Co. of America's Bond and Commercial Loan Department, has

this to say regarding the modification of contract terms:

. some [private placement] lenders will make
modifications only in exchange for a higher
interest rate or a tightening of other parts of
the agreement. However, private placement lenders
generally view the Loan Agreement as a living
document destined to be modified periodically to
take account of changing circumstances.
Accordingly, these lenders make most modifications
routinely, with no quid pro quo exacted from the
borrower unless the proposed corporate action will
compromise the lender's margin of safety.(Zinbarg,
1975, p. 35)

The reason for this Marquess of Queensbury-like behavior is also

clarified in Zinbarg's article:

. .the bulk of each year's new loans are
additional credits extended to existing borrowers.
Indeed, the very fact that continuous communica-
tion is maintained through the loan modification
process makes both parties more comfortable with
each other, and facilitates the repeated financing
of borrower's long term growth. (p. 52)

These remarks suggest, in fact, that the direct placement creditor/

borrower relationship is an enduring one which is costly to replicate,

and therefore that the creditor's incentive to capitalize on the

wealth transfer potential of an adverse accounting change may be

virtually nil.

Hayes (1977), in a lengthy summary of the term lending policies

of eight large mid-western banks, indicates that bankers take a

similar view of the creditor/borrower relationship:

Because loan officers as a rule maintain close
contacts with their customers, often acting almost
as regular financial advisors, they generally seek
every conceivable means of bringing a borrower
through periods of difficulty without permanent
damage to his earning capacities. Thus flexi-
bility in adjusting terms of loan agreements to
changed conditions is regarded as an axiom of
policy.(Hayes, 1977, p. 124)

On the other hand, the banker's view of the creditor/borrower

relationship in the case of other institutional creditors is curiously

at odds with that expressed by Zinbarg:

. life insurance companies and pension funds
use basically the same approach to long-term
direct loans as to purchases of bonds in the
market. There is seldom any expectation of a
continuing relationship or of ancillary business.
As the relation by the nature of the case is
impersonal, these institutional investors have an
understandably strong preference for avoiding
frequent negotiations for changes in the terms of
loan agreements. (p. 124)

Each of these sets of remarks may contain a certain element of

"window dressing." However, they at least suggest that, other things

equal, one may expect to observe a larger (smaller) share price effect

in response to an adverse accounting change when the greatest part of

a firm's debt is publicly (privately) placed. In passing, it should

be mentioned that this same distinction was proposed by Leftwich, but

for a different reason. Leftwich argued that the transactions costs

involved in renegotiating and amending a public indenture would

greatly exceed those necessary to amend a private loan agreement

It has been suggested here that the incentives inherent in the

"public" and "private" situations may be substantially different. The

effects of these arguments are, however, complementary.

3.5 Hypotheses

A number of arguments have been made in preceding sections of

this chapter. First, because the investment/financing opportunity set

available to shareholders is frequently restricted by the terms of

existing loan contracts, and because these restrictions often take the

form of constraints on financial statement numbers and relationships,

it was suggested that the effect on equity values of the "lessee"

accounting policy events of the 1973-1976 period -- culminating in

more stringent capitalization conditions and the necessity for retro-

active application -- was potentially adverse. Second, it was

proposed that, if one is willing to make certain simplifying assump-

tions about the characteristics of loans and the attitudes of

creditors, a reasonably uncomplicated model of the cross-sectional

distribution of such "share price effects" would result. In this

model, the magnitude of the adverse share price effect for a given

firm depends jointly on the extent to which the firm is leveraged and

on the impact of the policy change on financial statement numbers and

relationships. Third, it was acknowledged that the assumptions

underlying this model could well be excessively unrealistic, and two

further variables were identified which might be expected to have a

moderating influence on the distribution of share price effects.

These arguments can be distilled into a number of testable

hypotheses. First, suppose that a group of firms has been partitioned

according to "degree of leverage" and "degree of financial statement

impact," as indicated in Table 3-1. Define ut as the non-systematic

component of the realized return on firm j's equity in period t, and


Vk = E(ujt )j E cell k.

Table 3-1

Partitioning Scheme

Degree of Leverage

Low High

Degree of

can be identified as ti

ee accounting change is

tests the following hypol

Null Hypothesis

H o: II =P2=3 =4

he period in which the effect of an adverse

impounded in prices, then equation (7)


U1 F2
(Cell 1) (Cell 2)

"3 14

(Cell 3) (Cell 4)

If t



Alternative Hypotheses

Hal : 2 P1 < 0
H a2: 3 < 0

H : (u1 U3) (W2 Ul)
32) (3 ) < 0

= (i4 U2) (.13 111) < 0

Beyond the simple cell mean differences indicated by equation (7), the

analysis of this chapter also suggests that the magnitude of ut


H : vary directly with A', the proportion of the firm's debt
which is subject to accounting-based restrictions; and

H : vary directly with P, the "publicly traded" proportion of
a such debt.