‘5’

1...: wry-wu- ’7 6

 

 

‘1.

’v-

 

 

ABSTRACT

A STUDY OF TEE PREDICTIVE SIGNIFICANCE OF
SEVERAL INCOME MEASURES RELATIVE TO THE ACCOUNTING
FOR EXTRAORDINARY ITEMS AND PRIOR PERIOD ADJUSTMENTS

by James E. Parker

r~

Summary of purpose and research design.

The problem addressed in this study concerns the

 

method of reporting extraordinary items and prior period ad-
justments in annual financial reports of business corporations.
The criterion of predictive assistance was selected as a means
of providing a possible solution to the diversity of views
surrounding this area of financial reporting. Finally, his-
torical empirical evidence was examined with respect to the
predictive assistance criterion in comparing several income
reporting alternatives. Comparisons were made on both a con-
ceptual level, i.e. between reporting concepts, and on a prag-
matic level, i.e. between reporting practices. The "concepts”

and related "practices" were as follows:

 

Concept - Related practice
Current Operating performance As reported
Modified all—inclusive Modified all-inclusive

All-inclusive --

The basic technique of this study was to examine each

of several income reporting alternatives with respect to

 

 

 

<-— “,-

H.-

 

 

James E. Parker

resulting forecasts where earnings per share figures based on
each reporting alternative were used as inputs to two differ-
ent prediction models. The resulting forecasts were then com—
pared to actual results (measured (a) by the same income re-
porting alternative and (b) by the all-inclusive alternative),
with the dollar difference between the forecast value and
actual value being expressed as a percentage of the actual.
These ratios constitute the forecast error measures that were
compared to determine which of the associated reporting alter—
natives would have been (for the test period) of greater aid

in forecasting earnings potential.

Findings, conclusions, and implications

The findings of any predictive study depend in part
upon the particular model(s) employed. However, from the
various findings of this study, it would appear that in gen-
eral there was a greater effect upon both forecast magnitude
and associated error resulting from the choice of historical
time periods than from the choice of either prediction model,
forecast objective, or income reporting alternative.

Arguments for both the current operating performance
concept (as a theoretical ideal) and the "as reported" prac-
tice (as a realized resultant of this ideal) are based solely

on the generation of data more useful in making future

 

 

James E. Parker

predictions. Given the research design of this study, histor-
ical empirical evidence failed to substantiate this argument.
Thus the writer considers this study to be empirical support
for the position advanced in Accounting Principles Board
Opinion No. 9 with respect to the single amount to be desig-
nated as net income for the period.

In addition to comparing the various concepts and
practices with respect to goodness of forecasts (i.e. magni-
tude of absolute percentage forecast error), they were also
examined from the viewpoint of conservatism, i.e. as to which
alternative, if any, tended to consistently result in lesser
forecasts. Examination of the findings led the writer to the
following conclusions. In general, the current operating
performance concept resulted in significantly lesser forecasts
than did either of the other two concepts. However, there was
no significant consistent difference with respect to forecast
amounts under each of the two practices. This suggests the
notion that in practice management varies its interpretation
of accounts and/or method of accounting in such a way as to
maximize the firm's apparent earnings potential. Accordingly
a comparison was made between the current operating perform—
ance concept and the "as reported" practice. The findings

are consistent with this notion. A similar comparison was

 

 

 

 

 

James E. Parker
also made with respect to goodness of forecasts. However, the

findings showed no significant differences.

 

 

 

 

 

A STUDY OF THE PREDICTIVE SIGNIFICANCE OF
SEVERAL INCOME MEASURES RELATIVE TO THE ACCOUNTING
FOR EXTRAORDINARY ITEMS AND PRIOR PERIOD ADJUSTMENTS

BY

James E. Parker

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Accounting and Financial Administration

1969

 

 

 

 

 

Copyright

by
James E. Parker

1969

 

 

 

 

ACKNOWLEDGMENTS

Each of my committee members has contributed toward
the successful completion of this dissertation in his own
way. However, a special note of thanks is due to Dr. George
C. Mead, chairman, and Dr. Richard J. Lewis for extraordinary
contributions.

A note of appreciation is also due Dr. Roland R.
Salmonson and Dr. Herbert E. Miller. While not members of
the dissertation committee, they made substantial contributions
toward its eventual completion. The dissertation is a descend-
ant of a research pr0posal originally presented in a Doctoral
Seminar conducted by Dr. Salmonson. Dr. Miller, through in-
formal discussions, provided guidance and inspiration during
the earlier stages of refining the original prOposal. During
this process a number of fellow doctoral candidates and faculty
at other schools offered helpful criticisms and suggestions.
Such colleagues included Barry E. Cushing, Gary M. Crooch,
Hugh A. Hoyt, Eugene H. Sauls, and Elba F. Baskin, Robert G.
May, Robert E. Jensen, and Howard J. Snavely.

Several of the statistical computer programs employed
during the course of this study were drawn from the library
of the Computer Institute for Social Science Research. These

programs and others were run on the CDC 3600 and CDC 6500

ii

 

 

computers of the Michigan State University Computing Labora-
tory. Without these facilities this study would not have
been feasible.

Financial support received during the course of my
doctoral studies is gratefully acknowledged. The following
organizations and individuals made it possible for me to
hasten the completion of my doctoral studies: American
Accounting Association, Haskins & Sells, and the many con—
tributors to the fellowship fund of the Department of
Accounting and Financial Administration.

To Professors James Don Edwards (Chairman) and
Gardner M. Jones (Acting Chairman 1967—68) of the Department
of Accounting and Financial Administration, I wish to express
a special note of thanks.

Thanks are also due to Mrs. Jo McKenzie for a most
skillful typing job, and to my wife, Jean, for help in the

sorting and tabulation of data.

iii

 

 

 

computers of the Michigan State University Computing Labora-

tory. Without these facilities this study would not have
been feasible.

Financial support received during the course of my
doctoral studies is gratefully acknowledged. The following
organizations and individuals made it possible for me to
hasten the completion of my doctoral studies: American
Accounting Association, Haskins & Sells, and the many con-
tributors to the fellowship fund of the Department of
Accounting and Financial Administration.

To Professors James Don Edwards (Chairman) and
Gardner M. Jones (Acting Chairman 1967-68) of the Department
of Accounting and Financial Administration, I wish to express
a special note of thanks.

Thanks are also due to Mrs. Jo McKenzie for a most
skillful typing job, and to my wife, Jean, for help in the

sorting and tabulation of data.

iii

 

A-

ACKNOWLEDGMENTS.

LIST OF TABLES

LIST OF
Chapter

I.

II.

III.

IV.

FIGURES.

F‘F‘F‘F‘H

U'lthNI-J

o
GUI-FUN?“

NNNNNN

wwww
btulord

ufibb

 

 

TABLE OF CONTENTS

INTRODUCTION. . . . . . . . . . . . . . .

Purpose of study . . . . . . . . .
Statement of problem . . . . . . .
Authoritative Opinions . . ... . .
The modified all-inclusive concept
Other doctoral research. . . .,. .

RESEARCH APPROACH . . . . . . . . .

Fundamental nature of the problem.
The function of accounting . . . .
The predictive ability criterion .
Concepts versus practices. . ... .
Summary of the approach. . . . . .
Hypotheses . . . . . . . . . . . .

ME THODOL O GY O O O O O O O O O O O O O O 0

Overview . . . . . . . . . . . . .
Sample selection . . . . . . . . .
Data sources 0 O O O O O O O O O 0

Procedures for determining earnings

per Share 0 O O O O O , O O O O O 0
Procedures for determination of
forecast errors. . ... . . . . .

EMPIRICAL FINDINGS. . . . . . . . . . . .

Overview . . . . . . . . . . . . .
Descriptive statistics . . . . . .

Procedures and findings relative
to each of three income concepts

iv

Page

ii

vi

viii

l3
l7
19

21
21
23
27
36
38
41
43
43
44
48
49
59
64

64
65

73

 

 

 

V.

APPENDIX

A.

B.

 

4.4 Procedures and findings relative
to each of two income practices. .

SUMMARY AND CONCLUSIONS . . . . . . . . . .
5.1 Overview . . . . . . . . . . . . . .
5.2 Brief summary of purpose and

research design. . . . . . . . . .

5.3 Findings, conclusions, and

implications 0 O O O O O O O O O O
5.4 Suggestions for future research. . .

Companies included in the study . . . . . .

Form used to collect data on each extra-~
ordinary income determinant, prior period
adjustment-new, or prior period adjust—
ment-old. . . . . . . . . . . . . . . . .

Form used to collect data and make
computations by company, by year. . . . .

Form used to summarize, by company, certain
information contained on form shown as
Appendix C O O O O O O O O O O O O O O O O

BI BLI OGRAPIIY O O O C O O O O O O O O O I O O O O O O

Page

90

100

100

100

101
105

109

111

112

113

114

 

 

 

 

Table

II.

III.

IV.

VI.

VII.

VIII.

IX.

 

 

LIST OF TABLES

Page
Number and Per Cent of 600 Companies
Reporting in Accordance with Each
of Three Alternatives . . . . . . . . . . . . 10

Number and Per Cent of Items Treated in
Accordance with Each of Three Reporting
Alternatives by the 600 Sample Companies. . . 11

Percentage of Materiality of Individual
Items Reported by 600 Sample Companies. . . . 12

Means of Absolute Percentage Forecast
Error Measures for All Companies for
Each of Four Income Measures. . . . . . . . . 69

Average Deviations of Absolute Percentage
Forecast Error Measures for Each of
Four Income Measures. . . . . . . . . . . . . 70

Ordinal Ranking of Means and Average
Deviations as Shown in Tables IV and V. . . . 71

Percentage of Companies for Which Each
of Four Income Measures Resulted in
Over forecaSts O O O O O C O O O O O O O O O O 72

Chi-square One-sample Probabilities of
the Observed Number of Companies for
Which Income Measures Under Each of
Three Income Concepts Resulted in
Least Absolute Forecast Error . . . . . . . . 76

Comparisons (with respect to absolute
amounts) of Forecast Errors Associated
with Income Measures Under the Modified
All-inclusive (MOD) and Current Operating
Performance (CUR) Concepts. . . . . . . . . . 81

Comparisons (with respect to absolute
amounts) of Forecast Errors Associated
with Income Measures Under the All-
inclusive (ALL) and Current Operating
Performance (CUR) Concepts. . . . . . . . . . 82

vi

 

 

 

Page

XI. Chi-square One-Sample Probabilities of
the Observed Number of Companies for
Which Income Measures Under Each of
Three Income Concepts Resulted in
the Least Forecast. . . . . . . . . . . . . . 85

XII. Comparisons (with respect to relative
amounts) Between Forecasts Based Upon
Income Measures under the Modified All—
inclusive (MOD) and Current Operating
Performance (CUR) Concepts. . . . . . . . . . 88

XIII. Comparisons (with respect to relative
amounts) Between Forecasts Based Upon
Income Measures Under the All-inclusive
(ALL) and Current Operating Performance
(CUR) Concepts. . . . . . . . . . . . . . . . 89

XIV. Comparisons (with respect to absolute
amounts) of Forecast Errors Associated
with Income Measures Under the As
Reported and Modified All-inclusive
Practices . . . . . . . . . . . . . . . . . . 93

XV. Wilcoxon Matched-Pairs Signed-Ranks Test
With Respect to Lesser Absolute Forecast
Error of As Reported (REP) and Modified
All-inclusive (MOD) Income Practices. . . . . 95

XVI. Comparisons of Forecasts Associated with
Income Measures Under the As Reported
and Modified All-inclusive Practices. . . . . 97

XVII. Wilcoxon Matched—Pairs Signed-Ranks Test
on Forecasts Based Upon Income Measures
Under As Reported (REP) and Modified
All—inclusive (MOD) Practices . . . . . . . . 99

XVIII. Comparisons of Forecasts Associated with
Income Measures Under the Current Operating
Performance Concept and the As Reported
Practice. . . . . . . . . . . . . . . . . . . 106

 

 

 

LIST OF FIGURES

Figure Page

1. Comparisons of Forecast Errors Generated
From Each Data Set. . . . . . . . . . . . . . 40

2. Comparisons of Forecast Amounts . . . . . . . . 4O
3. Determination of Percentage Forecast

Error Measures. . . . . . . . . . . . . . . . 6O
4. Outline of Comparisons. . . . . . . . . . . . . 65

viii

H

 

 

 

 

 

 

 

CHAPTER I

INTRODUCTION

1.1 Purpose of study

This research is an attempt to find a reasoned (based
upon the criterion of predictive ability) and empirically
supported resolution of conflicting viewpoints as to the in-
clusion or exclusion of extraordinary items and prior period
adjustments in the determination and reporting of the periodic
net income of business corporations for both the current and
prior reporting periods.

The measurement in the "real world" of business income
is a complex and loosely defined process. This has been in no
small part due to a lack of agreement among businessmen, ac-
counting practitioners, and accounting theorists as to the
basic concept of income itself. As a result, many controver-
sies have arisen among and between those engaged in the prac—
tide of preparing financial statements and those using said
statements which purport to "present fairly". Noting that
“There is a considerable diversity of views as to whether
extraordinary items and prior period adjustments should enter
into the determination of net income of the period in which

1

 

 

 

 

they are recognized,"1 the Accounting Principles Board of the

American Institute of Certified Public Accountants acted
recently in the controversy to which this study is addressed.
The fact that Accounting Principles Board Opinion Num-
ber 9 deals specifically with this problem area does not make
it resolved. "At stake is the usefulness of the income state-
ment and the public's confidence in it."2 The following ex-
amples illustrate the continuing controversy over this subject:

The recommendations for reporting corrections
and extraordinary items listed in Accounting Prin—
ciples Board Opinion No. 9 (December 1966) appear
deficient. As will be shown below, they do not
always result in full disclosure, and they permit
inaccurate reporting of periodic income. In effect,
they represent authoritative pronouncements rather
than consistent applications of reasoned principles.

Possibly because of the absence of a research
study issued by the AICPA and available for discus-
sion by all concerned prior to the drafting of an
.opinion and possibly for other reasons, APB Opinion
No. 9 falls short, in the author's view, of meeting
the need for change in some areas while going too
far in others.

 

lAccountingPrinciples Board, "Reporting the Results
of Operations," Opinion No. 9 (New York: American Institute
of Certified Public Accountants, December, 1966), p. 109.

2Le0pold A. Bernstein, Accounting for Extraordinary
Gains and Losses (New Ybrk: The Ronald Press Company, 1967).

p. 9.

 

3Donald A. Corbin, "Reporting unexpected Items:

.A Dissent to APB Opinion No. 9," paper presented to the
accounting faculty and graduate students, University of

'Washington, November 8, 1968, p. 1.

4Bernstein, Accountinggfor Extragrdinary Gains and
Losses, p. 310.

 

 

 

 

A contribution toward a reasoned and empirically sup—

ported solution to this "diversity of views" under present

day conditions was the objective of this research effort.

1.2 Statement of problem

This section includes only a brief discussion of the
problem and two opposed views of income. For a more detailed
description of the problem including the historical evolution
of both the problem and authoritative opinion related thereto,
the reader is referred to Part I of Bernstein's Accounting for
Extraordinary Gains and Losses.

Under currently acceptable accounting practices, cer-
tain gains and losses are often realized during a period or
periods for which they are deemed not to be resulting from the
regular operations (for that period) of the firm. Examples of
such items include the following:

1. Gains or losses on sales or other disposals

of fixed assets, investments, and divisions.

2. Gains or losses on changes in valuation bases

of inventories, investments, and fixed assets.

3. Foreign exchange adjustments.

4. Plant expenses deemed to be nonrecurring.

5. Prior year adjustments.

6. Catastrophe losses.

 

A.
v

“a

E?
a;

 

 

 

The above examples do not constitute an exhaustive listing

but rather should only be considered as illustrative of the
nature of the items under consideration. Adjustments and/or
transactions of a capital nature are not included since it is
widely agreed that such items are not income. The signifi—
cance of those items considered "extraordinary“ lies in the
nature and timing of the recognition of their effect on the
earned portion of owners' equity and on the likelihood of
their reoccurrence in future periods. The distinction between
income and capital adjustments and/or transactions is not the
concern of this research, i.e. distinctions as reflected in
published annual reports will be taken as given for purposes
of this study.

The question of singular importance becomes "How
should extraordinary gains or losses and/or prior period in-
come adjustments be reflected in the reporting of the firm's
income (and earnings per share)?" This question is phrased
from a reporting viewpoint because it is an answer to a
reporting problem that is sought as the objective of this
study. It may be argued that two matters are involved:

(1) determining financial information and (2) reporting finan-
cial information. This may be, but the distinction is irrel-

evant for the external user of the financial statements. For

 

 

 

 

him the most important (and often the only) source, either
directly or indirectly, of financial information is the firm's
published financial statements. To this user, there is only
the reported information. Hence, this study is directed at
the reporting level.

Basically, two opposed reporting concepts of income
have appeared both in the literature and practice. These are

the all—inclusive concept and the current operating perform-

 

ance concept. Under the all-inclusive concept, all items
affecting the net change in owners' equity between two points
in time, excepting capital transactions and adjustments, are
included in the determination of that amount labeled "net
income for the period" in the firm's annual financial state-
ments. Under the current operating performance concept,
extraordinary items and prior period adjustments are not
included in the amount labeled "net income for the period",
but rather are shown as direct increases or decreases to

the earned portion of owners' equity.

1.2.1 Argpments pro and con

The arguments for and against the all—inclusive and
current operating performance reporting concepts of income

have long appeared in the literature. Meigs, Johnson, and

 

 

Keller have summarized these arguments as follows:5

Arguments for all—inclusive income reporting

1. Income for any period is the realized increase
in net assets during the period as a result of
the entire set of conditions facing the business
during that period. An item is no less a part
of this picture because it is unusual in nature
or amount. Gains or losses relating to activi-
ties of prior years but arising because evidence
is now available that was not clear in the past
are likewise a part of that picture.

 

2. Periodic income is difficult if not impossible
to determine precisely; many elements of revenue
and expense are a product of events relating to
several accounting periods. The reader of finan-
cial statements is best served if he can accept
the income report with confidence that it fully
reflects all events that either occurred or came
to light during the current period. At the same
time, careful labeling and disclosure will en-
sure that he can distinguish unusual items and
corrections of past errors from the other ele-
ments of the net income figure.

Arguments against the current operating performance concept

A feeling that the alternative is less desirable is
a form of argument on the side of any proposition,
and many who support the all-inclusive view are
persuaded of its merit by the dangers they see in
alternatives: The user of accounting statements
who is not trained in accounting may not be aware
that the income statement is incomplete and thus
may not look to the retained earnings statement
for extraordinary items. This opens the way for
possible manipulation of reported income data,
since there are many borderline cases between
what is recurring and nonrecurring, current and

 

Keller,
Company,

5Walter B. Meigs, Charles E. Johnson, and Thomas F.
Intermpdiate Accounting (New York: McGraw-Hill Book
1963). pp. 104—5.

 

 

 

 

.111- far?!) 1 - ‘

 
 

 

 

noncurrent, ordinary and extraordinary. Manage-
ment may thus be tempted to 'normalize' reported
income by varying its interpretation of what
falls in these categories at different times.

Arguments for current operating performance income reporting

1.

Readers of financial statements attach great
importance to the figure labeled 'net income,‘
both in comparing the performance of different
companies and as a guide in estimating probable
earnings in the near future. A net income
figure that reflects the current operating
performance of a company best serves these
objectives since unusual items by their very
nature are peculiar to a given firm and are

not likely to recur in future years.

The accountant is familiar with the underlying
events behind income statement figures. He is
thus better equipped than the reader of a pub-
lished statement to make a distinction between
the factors that are a part of a company's cur-
rent operating performance, and those that re-
flect past errors or unusual events.

Arguments against the all-inclusive concept

1.

If errors of prior periods are included as a
part of current reported net income a double
distortion results: Past income was misstated;
and the income of the current year is misstated
in the opposite direction.

If unusual or extraordinary charges or credits
are included in reported net income, the casual
reader of financial statements may be misled;
the reported 'earnings per share' of companies
is a figure widely reported in the financial
news, and often the fact that corrections or
unusual items are included in the figure goes
unmentioned.

..rvu—w- v—Dt quart ‘1 - , n - .'."'¢: 4"}..‘43 17.?

-.

 

\— u

1,.
.“

 

.r'~ 1.

n ._ . ...__—.._.__—. _ J
— _ .7 .—-—-. .... ...-u.- —_-_ - .

   

1.2.2 .Empirical significance of the_problem

The practical importance of any research study in
accounting should be considered. To obtain an indication of
the empirical significance of extraordinary items and prior
year adjustments in published annual reports of industrial
and commercial corporations in the United States, a prelimi- ,5
nary review was made covering the most recent eleven year

period (1956-66) for which data were readily available. This

 

review was based solely on information reported in Accounting

6 an annual survey of the accounting as-

Trends & Techniques,
_pects of the annual reports of 600 industrial and commercial
corporations, cOnducted by the staff of the American Institute
of Certified Public Accountants.

In order to be able to make definitive statements con-
cerning intuitively perceived tendencies observed from the
data presented in Tables I, II, and III, best fitting straight
lines~were determined for certain of the time series data by
employing the method of least squares. .The lepes of these
lines are parenthetically referred to in the following remarks

,concerning trends.

Table I below reveals the number of companies identifying

 

.6American Institute of Certified Public Accountants,

Accountinqurends & Techniques (New York: AICPA, 1957-67).

 

 

 

 

 

 

 

 

 

 

one or more items as extraordinary in their annual reports

and their reporting treatment of such items.7

Over the entire period an average (mean) of 203 com-
panies or 34 per cent reported one or more extraordinary items
in each year. Taking the sample of 600 companies to be repre-
sentative, it would appear that for approximatelyone—third
of all United States industrial and commercial corporations
in the average year, there is the problem of determining the
prOper disposition of one or more extraordinary items in their
financial statements.

As can be seen from Table II, the total number of
extraordinary items reported in each year by the 600 companies
tended to decrease slightly (lepe of -7.09) over the entire
period, noting in particular that for those companies report-

ing.one or more extraordinary items, the following tabulation

 

7The reader is warned of the following inconsist—
ency between years in the information reported by Accounting
Trends & Techniques. For 1963 and prior years, extraordinary
items shown after net income for the year on the income state-
ment were designated "Income". For 1964 and 1965, such items
were denoted "Retained Earnings". This inconsistency was ob-
served from a detailed description of the accounting treatment
given extraordinary items which has been summarized into the
designations of "Income" and "Retained Earnings" as shown in
Tables I and II. It was not possible to eliminate this in-
consistency in summarizing the information due to the fact
that Accountinngrends & Techniques reports detail only by
number of items - not by number of companies.

 

 

 

 

 

 

 

 

 

 

 

10

TABLE I

NUMBER AND PER CENT OF 600 COMPANIES
REPORTING IN ACCORDANCE WITH EACH OF THREE ALTERNATIVES

 

 

 

 

No Adjusting Accounts for
Extra- Extraordinary Itgmg
Ordinary Retained Capital
Items Total Income Earnings Surplus
Year No. % No. % No. % No. % No. %
1956 360 60 240 40 211 35 27* 5 2 -
1957 416 69 184 31 158 26 24** 4 2 - “
1958 368 61 232 39 191 32 37+ 6 4 1 I
1959 387 65 213 35 194 32 32 3 2 -
1960 370 62 230 38 193 32 36 6 1 -
1961 378 63 222 37 183 31 37 6 2 -
1962 366 61 234 39 186 31 45 7 3 1
1963 397 66 203 34 156 26 45 8 2 —
1964 413 69 187 31 119 20 68 13 0 0
1965 426 71 174 29 96 16 77++ 13 1 -
1966 481 80 119 20 60 10 59+++ 10 O 0

 

*Includes 6 companies adjusting both income and retained
earnings.
**Includes 4 companies adjusting both income and retained
earnings.
+Inc1udes 12 companies adjusting both income and retained
earnings.
++Includes 18 companies adjusting both income and retained
earnings.
+++Includes 3 companies adjusting both income and retained
earnings.

shows a rather stable pattern in the average number of such

items per company per year:

1956 = 1.20 1960 - 1.41 1964 - 1.35
1957 - 1.40 1961 - 1.41 1965 - 1.44
1958 - 1.39 1962 - 1.58 1966 - 1.36
1959 - 1.31 1963 - 1.30

 

NUMBER AND PER CENT OF ITEMS

TABLE II

11

 

TREATED IN ACCORDANCE WITH EACH OF THREE

REPORTING ALTERNATIVES BY THE 600 SAMPLE COMPANIES

 

 

Accounts Adjusted for
Extraordinary Items

 

 

 

 

Total Retained Capital

Number of Income Earnings Surplus
Year Items No. ‘% No. ‘% No. %
1956 289 254 88 33 11 2 1
1957 257 226 88 29 11 2 1
1958 322 277 86 41 13 4 1
1959 280 246 88 32 11 2 1
1960 324 263 81 60 19 l -
1961 312 246 79 64 21 2 1
1962 369 273 74 93 25 3 l
1963 264 201 76 61 23 2 1
1964 252 152 60 100 40 0 0
1965 250 143 57 106 42 l -
1966 162 79 49 83 51 0 0

 

Although there was a slight decrease in the number of

extraordinary items, an examination of Table III reveals an

increase in the materiality of all such items.

This state-

ment is based upon slopes of 1.78 and 1.26 respectively for

the best fit lines associated with the number of items and

percentage re the highest percentage materiality group (over

50%) of Table III,

and lepes of -5.99 and -1.53 for those

lines associated with the number of items and percentage re

 

-L—s___—.——-—.—-I— _ kw 4- ‘

 

 

 

.wanmcfifiumuop no: muaamanmume mo mmmucmoaom z

 

.mucmEumsnpm Mm»
mEoocH paw meua mumsapaomauxm How popmsnpm mmsflcamm Hmmm usoaaso on Emufl mo oaumms

 

 

s a «m mm ea mm 6a mm mm as ea am. moa coma
m ma om om 0a SN ma as ea as cm on omm mama
a ma om am 0a «m ma mm ea mm mm em «mm «Goa
6 Na am mm Ga ms ma 04 ea mm mm an «em moma
m“ «a as ea ow ea om Na ms Na ms mm oma mom moma
m an ea mm «a ms ma 64 «a me am so mam aoma
m ea ma om ma as ma we ma as «m aaa «mm coma
ma em aa om ma as ma as ma me am mm omm mmma
aa om ma a6 «a me ma ms ma mm hm mm «mm mmma
aa em ea om ma mm Na mm ma mm mm am emu smma
m oa m ma ea as ma «4 ma mm as mva mmm omma
“R. .02 X. .02 X .02 . “X. .02 X .02 X. .02 HmuOB HMO?

 

 

 

 

 

 

Z nRom H0>O wmomlam NONIHH A\oOHIw wwmlo

 

 

mmHZ¢mEOU mflmz<m 000 Mm QmBmOmmm
WSHBH AQDQH>HQZH m0 *MBHQ¢HmmB€S m0 mwflfizmvmmm

HHH mqmdfi

 

. - . hf‘ )5.

 

 

 

 

 

"w rm; ,1 z.
.

 

 

13

the lowest percentage materiality group (0-5%) of Table III.
Thus there has been both a relative and absolute increase in
those extraordinary items of relatively greater amount. As

a summary of the practical significance of the problem, we
may say that over the period examined 34 per cent of the com—
panies reported an average of 1.38 extraordinary items per
year with the average materiality of the items increasing

slightly.

1.3 Authoritative Opinions

The diversity of views as to whether extraordinary
items and prior period adjustments should enter into the
determination of net income of the period in which they are

recognized, is not of recent origin. Writers on the subject

8 10

include W. G. Rowe, A. C. Littleton,9 Thomas W. Leland,

11

Maurice H. Stans, and Arthur Andersen & Co.12 among others.

 

8W. G. Rowe, "Surplus Adjustments," Accounting Review,
VIII (October, 1933) 293.

 

9A. C. Littleton, "The Integration of Income and Sur-
plus Statements," JOurnal of Accountancy, LXIX (January,
1940), 40.

10Thomas W. Leland, "Revenue, Expense, and Income,"
Accounting Review, XXIII (January, 1948), 22.

11Maurice H. Stans, "Modernizing the Income State-
ment," Accounting Review,XXIV (March, 1949), 7.

12Arthur Andersen & Co., Accounting and Reportipg
Problems of the Accounting Profession (Chicago: Arthur
_Andersen & Co., 1962), p. 39.

 

 

 

__—a:___ ....

 

 

14

The arguments for and against the all-inclusive and current
Operating performance concepts were summarized above and will
not be repeated. At this point only the conclusions reached
by three authoritative bodies will be presented.

The following quotations reflect the American Account—
ing Association's committees' preference for the all—inclusive
approach toward the reporting of period income:

The income statement for any given period should
reflect all revenues properly given accounting recog-
nition and all costs written off during the period,
regardless of whether or not they are the results of
operations in that period.

For any one year the income statement should
reflect all realized revenues, and all costs and
losses written off during that year, whether or not
they have resulted from ordinary operations.

The income of an accounting period should be re—
ported in a statement providing an exhibit of all
revenue and expense (including losses) given ac-
counting recognition during that period.

The realized net income of an enterprise measures
its effectiveness as an operating unit and is the
change in its net assets arising out of (a) the ex-
cess or deficiency of revenue compared with related

 

l3American Accounting Association, Executive Commit—
tee, "A Tentative Statement Of Accounting Principles Under—
lying Corporate Financial Statements — 1936," Accounting and
Reporting Standards for Corporate Financial;Statements and
Supplements (Madison, Wisconsin: American Accounting Asso-
ciation, 1957), p. 67.

14Ibid., 1941 Statement, p. 55.

15Ibid., 1948 Revision, p. 17.

 

a b h. .\~....§.

 

 

 

 

 

 

 

15

expired cost and (b) other gains or losses to the
enterprise from sales, exchanges, or other conver-
sions of assets.16

However, in its 1957 revision, the Association's committee
stated an exception (with respect to prior period adjustments)
in its preference for the all—inclusive approach:17

The reports for the period will encompass not
only those transactions which arise from Operations
of the period, but also some transactions completed
during the period and related to activities of prior
periods. Transactions relating to current operations
should be reported as components of realized net in-
come of the period. Income-determining transactions
recognized in the current period but primarily re-
lating to prior activities should not affect the
determination or reporting of realized net income of
the period.

Until APB Opinion Number 9, the American Institute of

Certified Public Accountants had expressed its preference for

the current Operating performance approach as follows:18

The committee has indicated elsewhere that in
its opinion it is plainly desirable that over the
years all profits and losses of a business be re-
flected in net income, but at the same time has
recognized that, under appropriate circumstances,
it is proper to exclude certain material charges
and credits from the determination of the net in-
come of a single year, even though they clearly

 

16Ibid., 1957 Revision, p. 5.
171bid., p. 8.

18American Institute of Certified Public Accountants,
Committee on Accounting Procedure, Accounting Research and

Terminology Bulletins, Final Edition (New York: AICPA,

1961), p. 63.

   

16

affect the cumulative total of income for a series
of years. In harmony with this view, it is the
Opinion of the committee that there should be a
generalppresumption that all items of profit and
loss recognized during the period are to be used
in determining the figure reported as net income.
The only possible exception to this presumption
relates to items which in the aggregate are mate-
rial in relation to the compapy's net income and
are clearly not identifiable with or do not result
from the usual or pypical business Operations of
the period.

 

In December, 1966, however, the official stance of
the institute was changed with the issuance of Opinion No. 9

of the Accounting Principles Board. As pointed out by Savage

and Snavely,19

Part I of Opinion No. 9 makes the following
changes: (1) It changes the definition of extra-
ordinary items. The criteria to be used in deter-
mining whether or not an item qualifies as extra-
ordinary differ significantly from those in use
prior to the effective date of the Opinion. (2)

It changes the definition of prior ppriod adjust-
justments. Here again, the prescribed criteria
differ significantly from the ones formerly used.
(3) It requires that the amount shown as net income
on income statements include all items of profit
and loss recognized during the period except for
prior period adjustments as newly defined. ,(4) As
a result of the changes in definitions of extra-
ordinary items and prior period adjustments, the
Opinion causes the amount shown as net income be-
fore extraordinary items to include items of profit
and loss that heretofore were considered to be
extraordinary and/or corrections of prior years'

 

 

19Allan H. Savage and Howard J. Snavely, "The Account-
ing Principles Board and Opinion No. 9/' Unpublished paper
prepared at the University of Texas at Arlington, p. 1.

 

 

 

 

[1.1:

 

17

income figures. (5) It requires that all extra—
ordinary items (as newly defined) be shown in a
Separate section of the income statement. (6) It
requires that all prior period adjustments (as

newly defined) be shown in the statement of retained
earnings.

As Bernstein points out, the Securities and Exchange
Commission has generally favored the all—inclusive approach,
. . . , . V "20
Its p051tion 'materially' unchanged over the years. How-
ever, Rule 5-03 of Regulation S—X does provide for the addi-
tion or deduction from that amount labeled net income or loss,
at the bottom of income statements filed with the Commission,
of items equivalent to direct credits or charges to retained

21

earnings. Thus, certain items may be excluded from the

amount specifically designated as net income for the period.

1.4 The modified all-inclusive concppt

As discussed above there have been basically two Op-
posed concepts Of reporting income appearing both in the
literature and in practice over the years, the all-inclusive
concept and the current Operating performance concept. How—
ever, with the advent of APB Opinion No. 9 a new concept was

advanced. This concept will be referred to as the "modified

 

20Bernstein, Accounting for Extraordinary Gains and
Losses, p. 25.

21Securities and Exchange Commission, Regulation
S-X; Washington, D.C., Revised 1958.

 

 

 

 

   

18

 

all—inelusive concept." Under the modified all-inclusive
concept the only exclusions in the determination and report-
ing of net income for the current period are those items
deemed to be "prior period adjustments." According to
Opinion NO. 9, prior period adjustments are "limited to those
material adjustments which (a) can be specifically identified
with and directly related to the business activities of par-
ticular prior periods, and (b) are not attributable to eco-
nomic events occurring subsequent to the date of the financial
statements for the prior period, and (c) depend primarily on
determinations by persons other than management and (d) were
not susceptible of reasonable estimation prior to such deter-

mination."22

The Board expects such items to be rare in
modern accounting and that, in most cases where there is a
prior period adjustment, "the Opinion of the reporting inde-
pendent auditor On such prior period would have contained a
qualification because of the uncertainty (then existing)."23
The concern of this dissertation research is an eval-
nation (in terms of the criteria of predictive ability) of

the relative merits of the three reporting concepts of income

with regard to the inclusion or exclusion from reported net

 

22Accounting Principles Board, Opinion No. 9, p. 115.

23Ibid.

 

 

 

 

   

19

 

income 013 extraordinary items and/or prior period adjustments.
To smmmarize, those favoring the all-inclusive approach would
include in current income all changes in owner's equity not
due to capital transactions; those favoring the modified all-
inclusive concept would do likewise except that prior period
adjustments (as newly defined) would be excluded; and those
favoring the current Operating performance concept would ex-
clude all extraordinary items and prior period adjustments

from the current period's net income.

1.5 Other doctoral research
‘A canvass of the literature revealed two recent doc-
toral dissertations in the area of accounting for extra-

24 Both studies

ordinary items and prior yearadjustments.
approached the solution to "the diversity of views" on the
basis of.g priori reasoning, an approach that has refused to
yield a practical solution for many years. Wright's empir—

ical findings, covering a five year period and expressed mostly

in industry aggregates, simply support the contention that a

 

24Albert Welter wright, Jr., Accounting for Extra-
ordinary Charges and Credits and Their Implipations For Net
Income, Unpublished Doctoral Dissertation, University of
California, Los Angeles, 1965, and LeOpOld A. Bernstein,
Extraordinary Items of Gain or Loss: A Research Study of
Theory and Practice, Unpublished Doctoral Dissertation,
New'YOrk University, 1966.

 

20

pro‘blem exists. The empirical portion of Bernstein's work,
mostly concerned with a two year period, is also a cross-

sectional analysis as Opposed to a study of time series data.

 

 

 

 

2.1 km

mount re
3n tWO gr<
éefinitior
Of the ent
by the gen
'Jiew is thl
it is not d
fflmatiOn g
TEE/Ofted 0n
tathe reads
ZCted that P
all'infi‘lusiv

mas and/0r

 

HOWe

 

 

CHAPTER II

BASIC RESEARCH APPROACH

2.1 Fundamental nature of the problem
The idea of including extraordinary items in the
amount reported as net income for the period can be supported
on two grounds. The first is that reported income should, by
definition, reflect all changes in the economic well—being
of the entity to the extent that such changes are recognized
by the generally acceptable accounting practices in use. This
view is thus based upon an absolute concept of income, i.e.,
it is not dependent upon a specified use of the resulting in-
formation generated under the concept. Secondly, income
reported on this basis may be considered as the more useful
to the reader in obtaining a given objective. It should be
noted that prOponents of both the all—inclusive and modified
all—inclusive concepts advocate the inclusion of extraordinary
gains and/or losses and thus argue on each of these two grounds.
However, prOponents of the current Operating perform-
ance concept base their arguments largely, if not solely,
on grounds of providing more useful infOrmation in financial
:reports. The prOper amount to be labeled net income for the

21

‘7'

 

——
‘ ’flcwm-

 

arm

 

notion C
The com

of reali
its utilj
$315 in I
be pproac
ing eXtI'dO
rent periO<
changes th
ever. 1310130
that such e;
apparent ea:

period ad jus

1i
1 reported

 

 

22

period is that amount that will prove most useful to the
users of the financial statements. Under this approach, the
notion of a "true" or "actual" income concept is meaningless.
The concept of income is an abstraction, i.e., an abridgment
of reality. Its justification for existence lies solely in
its utility.

The question of whether to include prior period adjust-
ments in the amount reported as net income for the period can
be approached in a parallel fashion to that employed regard-
ing extraordinary gains and losses. Net income for the cur-
rent period should not, by definition, reflect economic
changes that did not occur during the current period. How-
ever, proponents of the all-inclusive concept have argued
that such exclusions Open the door for presenting misleading
apparent earnings potential, i.e., that the exclusion of prior
period adjustments could actually reduce predictive assistance
of reported income figures.

It should be noted that this study is not directly
concerned with the desirability of eliminating what some1
have called an "undue" emphasis on the needs of those who

will not or cannot use the financial statements as a whole,

 

1For example, see Bernstein, Accounting for Extra-
ordinary Gains and Losses, p. 198. ". . .anyone who chooses
to use single figures out of its (pig) context (should not)
blame the accountant for the possible adverse results of his
poor judgment."

78-;
- ‘-:p—-v~— in

 

 

 

adjustme

teristic

the modi f
Whereas t}

income cou

CQRCGpts j".

more useful

 

 

 

23

i.e.. the focus on a single net income figure as Opposed to
all items making up the total change in owners' equity over
a period. This study takes this "undue" emphasis on the net
income figure as a fact of life and attempts to resolve the
single question of how extraordinary items and prior period
adjustments should be reported given this behavioral charac-
teristic on the part of some users.

To summarize, the idea of "true" income supports only
the modified all-inclusive concept toward income reporting,
whereas the notion of employing the "most useful" concept of
income could be used in support of whichever one of the three
concepts is deemed to result in the generation of information

more useful in obtaining a given objective.

2.2 The function of accounting

Accounting has been defined as fthe process of identi—
fying, measuring, and communicating econOmic information to
permit informed judgments and decisions by users of the infor-
mation."2 From this definition, it follows that the function

of accounting is to "permit informed judgments and decisions."

 

2American Accounting Association, Committee to Pres
pare a Statement of Basic Accounting Theory, A Statement of
Basic Accounting_Theory, (Evanston, Illinois: American
.Accounting Association, 1966), p. l.

MM

 

 

 

Sabre E
COmuniC
bot not

jecision
Shuttle:

decis

h
In )
0‘

annual f1
and Canno'
a 90th M
affected b

It
be limited
abeut the P
knawledge s

ei’ents and l

tion dOes nc

 

 

24

Before Economic information can be identified, measured, and
communicated to serve this function, it is first necessary,
but not sufficient, to know the nature of the judgments and
decisions that are to be based upon this information. It is
submitted here that the fundamental character of all judgments
and decisions, based upon past data reported in published
annual financial reports, is anticipatory. The past is gone

and cannot be recalled. The present is only a point in time,

 

a point without dimension. Only events yet to occur can be
affected by man. The future is the domain of concern.

It has been suggested that the domain of accounting
be limited to present knowledge - knowledge in the present,
about the past and the present - on the grounds that such
knowledge serves as the basis of both judgments about past
events and of plans to procure future events.3 This sugges—
tion does not deny the anticipatory nature of judgments.
Even judgments about past events have as their purpose
assistance in determining future action; otherwise such
judgments would not be made. Likewise, any effort toward
determining one's present position is impelled by antici-

patory motives.

 

3Raymond J. Chambers, Accounting, Evaluation and
Economic Behavior (Englewood Cliffs, New Jersey: Prentice-
Hall, Inc., 1966), p. 97.

 

 

AS

approach to
notion 0f u
"far what P
p~rposes f0
dominant Pu
coostance i
serial per
current per
present man .
be used solr
sent, there
determined T
approach ass.
ordinary" a}!

35 action ta
Of c
co“1:01 anc

mlortantly

 

E?51
.uate mar

4
f

35p?
. .0
d8: CUT

H
r:

   

A.
1
‘1
1

5.
F

 

As noted above, the current operating performance
approach to the reporting of periodic income relies on the
notion Of usefulness. The term "useful" raises the question:
" For what purpose?" To the extent there are opposed possible

purposes for which the item labeled net income may be used, a

 

dominant purpose must be selected. For example, in a cir-
cumstance in which there has been a recent turnover of man—-
agerial personnel, certain gains and losses recognized in the
current period may clearly not be the result of action of the
present management. If the amount labeled net income is to
be used solely as an index for evaluating the present manage—
ment, there can be little doubt that the "proper" income is
determined by following the current operating performance
approach assuming that all items considered to be "extra—
ordinary" are both beyond the control of and not the result

of action taken by the present management.

 

 

Of course, there is the problem of determining the

" control and responsibility" of present management. But more

 

 

importantly, we should recall our purpose in attempting to
evaluate management in the first place: to serve as a guide
to future) action. If the gains or losses from, say, a long-
term non-cancellable contract continue to occur in future

periods, current Operating performance income may be both a

 

 

 

good evalua

of the firm
item labele
In this exau
rent and (2
W‘ni«
financial 5:
part upon t‘:
influence f;
aP’PIOBCh ma}
“filament a
eitheI the n
Pirfomance
“We: i.e.
Fillies [’11in
part 05 user
stated Previ
.llke undLJe
301p there
Home in o
ordinary ite
tht the Co“

 

    
   
 

N

5.

26

good-evaluation of the present management and a poor predictor
of the firm's earnings potential at the same time. Thus the

item labeled net income may be used for more than one purpose.
In this example, the two purposes are (l) to evaluate manage-

ment and (2) to predict the future earnings of the firm.

 

Which purpose is dominant for "the user" of published
financial statements in the above example? This depends in
part upon the courses of action open to the user. If he can

influence future management action, the current operating

 

approach may be preferable. On the other hand, if future
management action is given, even though unknown, to the user,
either the modified all-inclusive, or the current operating
performance concept may result in the more useful measure of
income, i.e., the better ‘prediifcrtor- For United States com-
panies publishing financial statements, a passive role on the
part of users is assumed for purposes of this study. As
stated previously, the user group of concern are those who
"place undue emphasis on a single measure." Except for this
group there would be no controversy since all three concepts
provide in one way or another for the disclosure of extra-
ordinary items and prior year adjustments. But the fact is
that the controversy does exist and the purpose Of this study

is to aid in finding a solution.

£1

 

 

surexer
extraor
not all.
over the
another.
’zrsefulne
criterion
torically
ThiCh in 1
decision In
556 to be '
ables will
Of some fut

Elie Withou-

 

 

 

   

27

2,3 The redictive abilit criterion

 

The problem of evaluating alternative accounting mea-
surements is not unique with respect to the accounting for
extraordinary items and prior period adjustments. "Most, if
not all, accounting controversies can be viewed as disputes
over the relative merits of one measurement alternative versus

another . " 4

The American Accounting Association has viewed
"usefulness of the information" as "the all-inclusive
criterion." Beaver, Kennelly, and Voss5 point out that his-
torically, usefulness has been related to decision making
which in turn raises two difficulties: (1) to define the
decision model and (2) to specify how the decision variables
are to be measured. Often one or more of the decision vari-
ables will be a probabilistic expectation of the occurrence
of some future event. Thus an important relationship between
predictions and decisions is.) drawn: "A prediction can be

made without making a decision, but a decision cannot be made

without, at least implicitly, making a prediction."6

 

4William H. Beaver, John W. Kennelly, and William M.
Voss, "Predictive Ability as a Criterion for the Evaluation
of Accounting Data," Accounting Review, XLIII (October 1968) ,
675.

5Ibid.

61bid., p. 680.

 

 

 

 

 

 

 

not at pr
cone conc
mama
concepts

requires (
a parametq
little ab:
their deci
be employe

This concl‘

7
$110915 ;

Bec
d601sic
ability
to the
the sam
re93rdi;
Embseqm
eVentuaJ
ability
COnsiste
tion, 1
bring ac
tems of

The

mociatmn

7

 

I

”7

1

/

 

 

Sinice financial statement users' decision models can

not at present be fully specified, evaluating the three in-
<:ome concepts in terms of their relative predictive ability
is a potentially significant approach. To evaluate the three
<:oncepts in terms of their ability to predict future income
:requires only that we assume that expected future income is
a parameter of the decision process, even though we know very
little about how the users use this parameter in reaching
‘their decisions. Thus the predictive ability criterion can
13s employed prior to further specification of decision models.
'This conclusion was reached by Beaver, Kennelly, and Voss as
7
follows:
Because prediction is an inherent part of the
decision process, knowledge Of the predictive
ability of alternative measures is a prerequisite
to the user of the decision-making criterion. At
the same time, it permits tentative conclusions
regarding alternative measurements, subject to
subsequent confirmation when the decision models
eventually become specified. The use of predictive
ability as a purposive criterion is more than merely
consistent with accounting's decision-making orienta—
tion° It can provide a body of research that will
bring accounting closer to its goal of evaluation in
terms of a decision-making criterion.
The usefulness of an accurate estimate of future earn-

:ings is held to be self—evident. The American Accounting

lassociation Committee to Prepare a Statement of Basic

 

7Ibid.

 

 

 

Accour
3:1 e)
:rofit

earnin

earnings
tents of
tion as p

the purpo:

”fir prime

C3319 rEPOr

31g future

 

 

\

 

29

Accountirug Theory expressed this notion as follows: "Almost
all external users of financial information reported by a
Iprofit—0riented firm are involved in efforts to predict the
earnings of the firm for some future period."8 The committee
also Observed that "The past earnings of the firm are considered
'to be the most important single item of information relevant
'to the prediction of future earnings,"9 and that "A person
lising financial statements as an aid in predicting future
eearnings has a right to demand from the accountant measure—
mnents of past earnings that supply as much relevant informa-
tion as possible."10 An even stronger statement concerning
the purpose of income reporting was made by Robert Sprouse:
"The primary purpose of the measurement of last year's in-
come reported to investors is to provide a basis for predict-

' l ' IIll
ing future years 1ncome.

 

8American Accounting Association, Committee to Pre-
pnare a Statement of Basic Accounting Theory, A Statement of
IBasic Accounting Theory, p. 24.

91bid.
10Ibid.

1Robert T. Sprouse, "The Measurement of Financial
Position and Income: Purpose and Procedure," Research in
chcounting Measurement, R. K. Jaedicke, Y. Ijiri, and O.
hiielsen (eds.) (Evanston, Illinois: American Accounting
Itssociation, 1966), p. 106.

 

 

 

 

 

 

 

 

 

2.3.1 The
A c
"predicting
isthe writ
future conc
economic in
in; measure
Sup
advanced by
l’earnings p:
This notion
P059. One 5
BanY's Comm:
two prir
prepare:
the earr
Second 5
fot the
0f CdSeS
Constitt

p oSsible

(0f any
fife Past

 

 

 

2.3.1 The Question of forecast objective

"predicting earnings" or "estimating earnings potential.“

A question may be raised as to what is meant by

It

is the writer's position that the desired knowledge of the

future concerns the total change in the shareholder's future

(economic interest in the entity. Thus, the relevant account-

:ing measure of this objective is future all-inclusive income.

Support for this View of "earnings potential" can be

aadvanced by examining the need for having an indication of

"earnings potential." The ultimate criterion is usefulness.

(Phis notion raises the questions for whom and for what pur-

Iaose. One such purpose relates to the valuation of a com—

;Dany's common stock:

12

. . .the valuation of most common stocks involves
two principal steps or procedures. The first is the
preparation of some estimate of the probable range of
the earnings potential for the future. . . . The
second step . . . is to establish a reasonable price
for the estimated earning power. . . In the majority
of cases the statistical record of past earnings re-
flected by the income (profit and loss) statements
constitutes the starting point for the calculation of
possible future earning power. . . . The objective
(of any adjustments made by the investor) is to make
the past record indicative to the greatest extent pos-
sible of the economic activities which seem most
likely to prevail in the future.

 

12Douglas A. Hayes, Appraisal and Management of

5;ecurities (New York: The Macmillan Company, 1956), pp. 284-

£35.

 

 

 

In

there is n:
estimates c
may be labe
It may be t
Also such "
different t
are not igr‘

A s
earning pow
PWEI, as m
after Capit‘
into at lea

30w'er. th<

 

°Pinion tha1
the Sim3103

As Minted c
grcuP of use
‘ltree comeI
“naming”
“Wes. give,
certain “Sex

period best

_ =
“‘n— p.

‘a
h

    

I.

31

In the above discussion of future economic activities
there is no reason to assume that the investor desires to omit
estimates of future economic activities simply because they
may be labeled "special" or "extraordinary" by the company.

It may be that such items are more difficult to estimate.

 

 

Also such "earnings“ may be valued at different rates and
different techniques employed in their estimation, but they
are not ignored.

A sensible approach toward the prediction of future

 

earning power (defined to include all changes in the economic
power, as measured by accounting, of the stockholders equity
after capital adjustments) would be to divide the projection
into at least two elements -- the normal and the extraordinary.
However, there is ample evidence to suggest a widespread
opinion that many users of financial statements focus only on
the single figure designated as net income for the period.
As pointed out previously, it is the very existence of this
group of users that gives rise to the controversy since all
three concepts recognize the need for full disclosure of the
extraordinary nature of such items. The question then be-
comes, given this single figure fixation on the part of
certain users, which single income figure for the current

period best serves as an aid in predicting future changes in

 

 

 

.:-n:' . $.3‘r’3‘5’“

:‘te accou:
:jjustmen
A:
:eant by I
cancern M
1,9,, to f
3fthat sa
sive incom‘
income, pas
future modi
azing incorr
income.

Sup
that income
determined
Barket valu
it:

".w i.e

flaunt repo

 

«L'

“‘18 regard
a: this pro
5&(et Valli;

e: .
"rungs in

Ulrrent OPE‘

i

 

 

 

the aCCOUJTting measure of stockholders equity after capital

adjustments?

An alternative view toward the question of what is
nneant by predicting earnings potential is to give primary
(zoncern to the predictability pg; gs, of an income concept,
i.e., to forecast future values of an income from past values
caf that same income series. Under this view, past all-inclu-
sive income would be used to predict future all-inclusive
income, past modified all-inclusive would be used to predict
future modified all-inclusive income, and past current oper-
ating income would be used to predict future current operating
income.

Support for this alternative view lies in the notion
that income can not exist, much less be reported, except as
determined by accountants. Likewise it may be argued that
nnarket values are determined on the basis of reported earn-
ings, i.e. that to the investing public, income is that
anmunt reported by the company as "earnings per share." In
this regard an interesting question lying beyond the scope
(of this proposed study is raised: to what extent does the
Inarket value all-inclusive earnings or modified all-inclusive
earnings in a different manner from that in which it values

current operating earnings?

 

 

Fox
be regardec
i.e. each c
for its ter
of (1) futt
in the accc

effustments

2.3.2 E
The
in this app
atternpt to
ing to any
that the li
”we! to av
izatiOng.
The
the partiCu
M linear 1
Mela, Whi
iZed Vers i O:
army a ttrj
l:fomuation;

given to 1";

 

 

For purposes of this study, "earnings potential" will
be regarded from each of the two viewpoints discussed above,
i.e. each of the three concepts of income will be examined
for its tendency to generate data useful in making predictions
of (1) future income of the same series and (2) future changes
in the accounting measure of stockholders' equity after capital

adjustments, i.e. all-:inclusive income.

2.3.2 Limitations of the predictive ability criterion

The following limitations are not uniquely inherent
in this approach but rather are apt to be encountered in any
attempt to evaluate alternative accounting procedures accord-
ing to any purposive criterion. However, it is desirable
that the limitation of any research method be kept in mind in
order to avoid drawing unwarranted inferences and/or general-
izations.

The finding of predictive ability depend in part upon
the particular prediction model(s) employed. In this study,
two linear time series regression models were used. The
models, which in the writer's Opinion represent Operational-
ized versions of intuitively perceived characteristics gen-
erally attributed to lesser sephisticated users of accounting
information, varied only with respect to relative weights

given to historical data and are fully specified in the

 

hm- .

- awm

 

 

following
sible to
from the
sent a 3'0
treasure wi
either or

Ar
positive (
ing error '
assume inc
C€pt PIQdi1
inclusive 1
”Eights ar(
“"9 incomi
Power of t:
the modifie
S'Jlts are C

possibilit.

 

iredictive

CO

Study.

ml

 

fie .

 

 

34

following chapter. It is important to note that it is impos-
sible to test the income measure under examination separately
from the prediction model employed. Positive results repre-
sent a joint confirmation of both the model and the income
measure while negative results may be due to an "error" in
either or both elements.

Another potential problem arises when results are
positive (i.e. there is a significant difference in forecast-
ing error between alternative income measures.) For example,
assume income measures under the modified all-inclusive con-
cept predict better than say, income measures under the all-
inclusive concept when using a given model. If different
weights are introduced into the forecast model the all-inclu—
sive income measure might contribute more to the predictive
power of the newly weighted model than income measures under
the modified all-inclusive concept. Even if consistent re-
sults are observed for all models tested there is always the
possibility that some untested model possesses still greater
predictive power and at the same time suggests the Opposite
conclusion concerning the alternative income measures under
study.

The evaluation of the relative predictive ability of

the income measures may require an assumption concerning the

 

loss functi
it may be t
damaging to
forecast er
study was t
under forec
functions 1
Aft
limitations
contribute
is addresse ]
Kennelly, a
Two

sion:
may app
diCtiVe
Possibl
alterna
Within
conside

The
not an 1
difficu
emPil'ic
impolite
icy-088
kn0wled
We d' ‘

 

 

 

35

loss function associated with prediction errors. For example,
it may be that overforecast errors of future earnings are more
damaging to the achieving of the users objectives than under-
forecast errors of the same magnitude. The approach of this
study was to present the distribution (regarding over and
under forecasts) so that the reader might apply his own loss
functions in choosing among.the income measures.

.After considering both the potential advantages and
limitations of using the predictive criterion in an effort to
contribute toward a solution to the problem to which this study

is addressed, the writer is in complete agreement with Beaver,

Kennelly, and Voss's concluding remarks.13

Two implications emerge from the previous discus-
sion: (1) The preference for an accounting measure
may apply only within the context of a specific pre—
dictive purpose or prediction model. It may be im-
possible to generalize about the 'best' measurement
alternative across different contexts. (2) Even
within a specific context, the conclusions must be
considered as tentative.

The inability to generalize is a possibility, but
not an inevitability. We have cited only potential
difficulties, whose relevance can only be assessed
empirically, not byig priori speculation. What is
important is to know to what extent we can generalize
across purposes, and the only hOpe of acquiring this
knowledge is to conduct the predictive studies. .1;
we discover that different measures are best for

 

13Beaver, Kennelly, and Voss, "Predictive Ability as
a Criterion for the Evaluation of Accounting Data," pp. 682-
83.

 

 

 

 

 

36

different purposes, it would be erroneous to believe
that theppgedictive studies are any less important
because Of that discovery. The inabilipy to gener—
glizeLif it does exist, is not a flaw Of the predic-
tive ability methodology. It merely reflects the
state of accounting theory, but in neither case is

it an indictment Of the methodology that exposes

that fact. (Emphasis supplied.)

Even within a specific context, the preference
for one measure over another is tentative. A measure
that performed poorly may not be permanently rejected
in the sense that the researcher may refine the mea-
sure (and its theory) or redesign the study in the
hope that future research will demonstrate that the
measure is really better. Also there is always the
possibility Of an unknown or untested measure that
performs even better than the best measure tested.
Theory construction in other disciplines is an evolu-
tionary process, where the hypotheses are continuously
being revised, redefined, or overturned in the light
Of new theory and new evidence. There is no reason
to believe that accounting theory will be different.

Although it is important that a general awareness
Of these factors exists, neither the potential inability
to generalize nor the tentative nature of the conclu-
sions should be regarded as a deterrent to conducting
the predictive studies. Extension Of research efforts
into the predictive ability Of accounting data is nec-
essary for the fulfillment Of accounting's decision-
making orientation and for the meaningful evaluation
Of alternative accounting measures.

2.4 Concepts versus practices

Thus far, three concepts Of reporting income have been
discussed. These have been referred tO as the all-inclusive
concept, the modified all—inclusive concept, and the current

Operating performance concept. As noted in Chapter I, the

American Institute Of Certified Public Accountants favored

 

37

the current Operating performance concept prior to Opinion
NO. 9. However, the preliminary review based on information
reported in Accounting Trends & Techniques showed that prac-
tice was not consistent with any single concept during the
period prior to Opinion NO. 9. This Observation is also con-

sistent with the empirical findings Of Arnett,l4 Bernstein,15

and Wright.16

Thus it becomes important tO distinguish be-
tween a given concept and actual practice under the AICPA's
advocacy Of that concept. Therefore, this study was directed
at evaluating the relative predictive ability Of income mea-
sures under both two practices, as well as the three theo-
retical concepts.

The two practices under which income measures were
examined are referred to as, the "as reported" practice and
the "modified all-inclusive" practice. The "as reported"

practice refers to income as actually reported by the sample

fimms over the periods Of the test, whereas the determination

 

l4Harold E. Arnett, "Application Of the Capital Gains
and Losses Concept in Practice," Accounting Review, XL
(January 1965).

5Bernstein, Accounting for Extraordinary Gains and
Losses.

16 . . .

' Wright, Accounting for Extraordinary Charges and
Credits andpgheirigmplications for Net Income.

 

...—.....—

 

 

 

‘ I

 

38

Of income measures under the modified all—inclusive practice
Often required the recasting Of published income statements
to show how they would have appeared if APB NO. 9 had been
in effect. Because Of the necessary exercise Of judgment by
the writer, the latter must therefore be viewed as simulated
measures.

In fact, income measures under the modified all-inclu-
sive concept are identical with those under the modified all-
inclusive practice. The purpose in evaluating the relative
predictive ability Of income measures under each Of these two
practices was tO attempt to measure the overall potential
desirability Of the AICPA position as reflected in Part I of
Opinion NO. 9, and thus contribute toward a solution to the
seemingly endless stream Of apparent paradoxes such as the
following:

The APB has chosen disclosure. . . rather than
accuracy. . . . Their approach may tend to prevent
attempted manipulation (but) it may also force. . .
innocent distortion.

2.5 Summary Of the approach

The test Of usefulness in evaluating a) concepts --

current Operating performance, modified all-inclusive, and

 

17Corbin, "Reporting Unexpected Items: A Dissent
to APB Opinion NO. 9," p. 5.

 

I

 

 

 

 

 

 

 

 

39
all-inclusive -- and b) practices -- "as reported" and
modified all-inclusive —- reduces to a test Of their relative

ability to generate data more useful in making future predic-
tions Of "income". Unless at least one Of these concepts or
practices can be shown tO be of greater use in this respect,
the criterion Of predictive assistance is irrelevant to the
controversy. This would destroy the arguments grounded in
"usefulness," for both the current Operating performance (and
its realized resultant -- the "as reported" practice) and all-
inclusive concepts. The burden Of proof is upon these con-
cepts (and practice) because they are dependent upon a speci-
fied use of the resulting information. In contrast, the
modified all-inclusive alternative seems clearly superior on
the basis Of.§ priori accounting theory.

As mentioned previously, two prediction models were
employed during the process Of comparing the relative predic-
tive ability Of (l) the three income concepts and (2) the
two income practices. Also two forecast Objectives were used
for each Of these comparisons. Thus for each set Of data
there were four comparisons Of resulting forecast errors due
to the four possible combinations Of prediction models and
forecast Objectives as indicated by the following illustra-

tion.

 

 

 

 

 

 

COMPARISONS OF FORECAST ERRORS
GENERATED FROM EACH DATA SET

40

FIGURE 1

 

 

Forecast
Objective A
(all—inclusive income)

Forecast
Objective B
(income under same
concept or practice)

Prediction Model I
(unweighted linear

regression)

Prediction Model II
(weighted linear

regression)

 

Comparison Of
3 concepts

Comparison Of
2 practices

Comparison Of
3 concepts

Comparison Of
2 practices

 

 

Comparison Of
3 concepts

Comparison Of
2 practices

 

Comparison Of
3 concepts

Comparison Of
2 practices

 

 

 

 

In addition comparisons were made with respect to forecast

amounts as indicated by Figure 2.

COMPARISON-OF FORECAST AMOUNTS

FIGURE 2

 

 

Prediction Model I

Comparison Of 3 concepts
Comparison Of 2 practices

Prediction Model II

Comparison Of 3 concepts
Comparison Of 2 practices

 

The purpose Of these illustrations is only to summarize the

basic approach used during the course Of this study. The

following chapter describes the methodology in detail.

 

 

 

41

2.6 Hypotheses

The following hypotheses were tested by the method-
ology set forth in the subsequent chapter. The hypotheses
can be divided into two groups (1) those (C.l - C.2) con—
cerned with comparisons between income reporting concepts
and (2) those (P.3 - P.6) concerned with income reporting
practices. Hypothesis C.l is concerned with whether, among

the three income concepts, there is any consistent signifi-

 

cant difference as tO which income measure resulted in better
forecasts with the p priori expectation being that the current
Operating performance concept would generate better forecasts
than either Of the other two concepts since all items deemed
extraordinary are excluded therefrom. Hypothesis C.2, also
concerned with the three concepts, involves comparisons of
forecast amounts in order to ascertain whether any concept
can be considered more conservative than the others. Hypoth-
eses P-3-and P.4, responsive to both frequencies and relative
amounts. parallel hypothesis C.l with respect to the two
practices while hypotheses P.5 and P.6 parallel hypothesis 0.2.
The hypotheses were as follows:
Hypothesis C.1 H g the probability of a better prediction
(i.e. a lesser absolute forecast error) is
the same for predictions generated from in—

come measures under each of the three/in=

come concepts. Has the probabilities of a

  ___]

 

 

Hypothesis

Hypothesis

Hypothesis

Hypothesis

Hypothesis

C.2

P.3

P.4

P.5

P.6

 

42

better prediction are greater with the cur-
rent Operating performance concept.

Ho: the probability of a lesser prediction
is the same for predictions generated from
income measures under each of the three in-
come concepts. Ha: the probabilities Of a
lesser prediction differ according to the

income concept.

Ho: the probability of a better prediction
is the same for predictions generated from

income measures under each of the two income
reporting practices. Ha: the probabilities
Of a better prediction are greater with the

"as reported" practice.

Ho: the aggregate predictive ability of the
"as reported" practice and the modified all-
inclusive practice do not differ. Ha: the
aggregate predictive ability Of the "as
reported" practice is greater.

Ho: the probability Of a lesser prediction
is the same for predictions generated from
income measures under each Of the two in-
come reporting practices. Ha: the proba—
bilities Of a lesser prediction are greater

with the "as reported" practice.

Ho: the probability of lesser aggregate
predictions is the same for predictions
generated from income measures under the
"as reported" and modified all-inclusive
practices. Ha: the probabilities differ
according to the income practice.

‘9' “P's 4.“ 1‘.

 

 

 

CHAPTER III

METHODOLOGY

3.1 Overview
The basic means by which the relative predictive
ability of the two income practices and three income concepts

was evaluated was by comparing the associated relative fore—

 

cast error measures. All comparisons were between forecast
error measures associated with various income measures for
the same company for the same accounting period, i.e., the
forecast error associated with income measure A for Company
XYZ in the year 1960 was compared with the forecast error
associated with income measure B for Company XYZ in the year
1960; there was no comparison with any measure associated
with, say, Company ZZZ, nor was there any comparison with
any measure associated with Company XYZ's 1962 Operations.
Thus all extraneous variables which might have influenced the
outcome Of any Of the individual comparisons were eliminated.
Two groups Of comparisons were made: (a) those be-
tween forecast error measures associated with income measures
Obtained under each of three income concepts and (b) those

between forecast error measures associated with income

43

 

games 0

each groupi

for each

 

of forecas
are repres

On
the method
be divided
share on e
and evalua
a11d foreca
0f procedt
chailter w}

forth in (

 

 

measures <5btained under each Of two income practices. For

each group» a comparison was made for each Of 761 companies,
for each Of 16 time periods, and under each Of 4 combinations
of forecast model and forecast Objective. These combinations
are represented graphically in Figure 1.

Once the sample was selected and the raw data obtained,
the methodology followed during the course of this study can
be divided into three processes: determining earnings per
share on each of four bases; determining forecast errors;
and evaluating the significance Of the resulting forecasts
and forecast error measures. The first two processes (groups
of procedures) are described in the subsequent pages Of this
chapter while those procedures related to evaluation are set

forth in Chapter IV along with the findings generated thereby.

3.2 Sample selection

A random sample of 76 companies2 was selected from

 

lException: for the single most recent Of the 16
time periods, data wensnot available for 8 companies.

2Sample size was determined as follows. From Table
I (of Chapter I) it was noted that the percentage of com—
panies reporting one or more extraordinary items ranged from
20 per cent in 1966 to 40 per cent in 1956 with the average
for the period 1956-66 being 34 per cent. Assuming the com—
panies reported upon in Table I to be representative Of the
universe Of this study, it appeared that a sample size of
approximately 76 would be adequate in order to Obtain (for
each year) a minimum of 15 companies reporting one or more
extraordinary items.

 

 

_ 5,241: _. .

 

among thos
New York 5
requiremer
pany still
cannon stc
change as
engaged in
services,
of such co
(4) that t
undergOne
1967. A l
Seated as
Th
a8 £0110wS

l‘drket Dea

 

 
 
 

45

among thCJse companies whose common stock was traded upon the

New York Stock Exchange during the calendar year 1946. Further

requirements for inclusion in the sample were (1) that the com-

pany still be in existence as Of January 1, 1968, (2) that its
common stock continue to be listed on the New York Stock Ex-
change as Of that date, (3) that the company not be primarily
engaged in the areas Of transportation, communication, utility
services, finance, or insurance since the accounting practices
of Such companies are governedrbyfregulatory’agencies,:and

(4) that the nature Of the company's Operations must not have

undergone a total change3

during the time period 1947 through
1967. A list Of the companies included in the sample is pre-
sented as Appendix A.

The procedure employed in selecting the sample was

as follows. A OOpy was Obtained Of the listing "1946 Stock

Market Dealings" as published in The New York Times,4

 

 

3In order to clarify what is meant by a total change
in a company's Operations, the only two examples encountered
during the course Of this research were International Mining
Corporation (formerly National Department Stores Corp.)
whose Operations changed from retailing to mining, and
Madison Square Garden Corporation (formerly Graham-Page
Corporation) whose Operations changed from the manufacture
Of automobile parts to the Operation Of an amphitheater.

4"1946 Stock Market Dealings," The New jgrk Times.
January 2, 1947, sec. L, p. 36.

 

.. "—é-JJI'
|

 

 

, 51w,

 

h

 

January 2:
secotively
ing withir.
tyreferen
totline of
ifthe ite
described
essary to

76. A bre~

Pr
Me
De
Li
No
Ch

 

 

 

 

January 2. 1947. Each item on this list was numbered con-

secutively from 1 to 1,340. A random sample Of numbers fall-
ing within the range of 1 to 1,340 (inclusive) was selected
by reference to "Appendix VII Random Numbers" Of Schaum's
Outlingpof Thpory and Problgpg Of Statistic_s_.5 If and only
if the item corresponding to this number met the requirements
described above, was it included in the sample. It was nec-
essary to examine 219 listings in order tO Obtain a sample of

76. A breakdown Of the 143 rejections is as follows:

Preferred stocks and class B issues 66
Mergers and acquisitions 32
Delistings 15
Liquidations 6
Non-qualifying industries 22
Changes in operations __2

143

It should be noted that companies were not omitted from the
sample as a result of name changes. Distinctions between
name changes and the creations Of new entities were based on
information reported in various editions (covering the 1946

to 1968 time span) of Mpody's Industrial Manual,6 Moody's

 

 

5Spiegel, Murray R., "Appendix VII Random Numbers,"
Schaum's Outling Of Thppry and Problems of Statisticp (New
York: Schaum Publishing Co., 1961), p. 349.

égpody's Industrial Manual (New York: Moody's
Investors Service, Inc., 1946 through 1968.)

 

 

public Uti
body's Ba
Tl

the purpos
the sampl
company 1.
Poor's Re
10

Nfl sh
ber (3.1. ‘
major sen
first two
which of 1

S'I'C- nur

115%
a Short w

or busine

\
7

Investors‘
8

1
InVe5t°rs

9!
InvastOrs‘ ‘

10

l
hEcutive‘
1968) .

 

47

Public Utilipy Manual,7 Moody's Transportation Manual,8 and
Moody's Bank & Finance Manual.9

I The procedure employed in classifying companies (for
the purpose Of deciding which ones were to be included in
the sample) owith respect to industry was as follows. Each
company listing in the "Corporation Directory" section of
Poor's Register Of Corporations, Directors and Executives-

10 shows the Standard Industrial Classification code num-

lggg
ber (S.I.C.) for the principal products manufactured or the
major services furnished by the company in question. The
first two digits Of this four digit S.I.C. number show in
which Of ten major divisions the company is classified.
S.I.C. numbers were not used in pre 1960 editions Of Poor's
Register Of Corporations, Directors and Executives; however

a short verbal description Of the company's principal product

or business is given. The S.I.C. number reported in 1968

 

7Moody's Public Utility Manual (New YOrk: Moody's
Investors Service, Inc., 1946 through 1968).

8Moody's TranSportation Manual (New York: Moody's

Investors Service, Inc., 1946 through 1968)-

SMOOdX'S Bank and Finance Manual (New York: Moody's

Investors Service, Inc., 1946 through 1968)-

10Poor's Re ister Of Cor orations. Directors and .
Egecutives - 1968 (New York: Standard & Poor's Corporation,
1968).

 

 

along wit:
classify e

used by t

3.3 Data
T2
in the sar

and subse'

 

 

procedure
nual repo.
“1967" we.
cellected
not be 10
has based
Published
docum(“—nts
R
1947 to N
those cov
t0 Novemb
the above
order to
”2‘53 wt

Covered t]

48

along with verbal descriptions Of 1959 and 1949 were used to
classify each company into one Of the ten major divisions

used by the Standard Industrial Classification System.

3.3 Data source

The published annual reports Of each company included
in the sample for fiscal periods ending December 31, 1947
and subsequent provided the source data for this study. This
procedure called for the examination Of a total Of 1,596 an-
nual reports. Reports for eight companies covering the year
"1967" were not yet available at the time this data was being
collected. Also the 1947 annual report for one company could
not be located. With these exceptions, all data collection
was based upon examination Of original source documents, i.e.
Published annual reports of the companies. A total Of 1,587
documents was examined.

Reports for fiscal periods ending from December 31,
1947 to November 30,‘1948 (inclusive) were labeled "1947",
those covering fiscal periods ending from December 31, 1948
to November 30, 1949 were labeled "1948", etc. (In applying

the above criterion a latitude of seven days was allowed in

order to accommodate those firms reporting on the baSis Of

a 52-53 week fiscal period.) As a result, this study

covered the 20 year time span immediately preceding the

 

 

 

 

effective
50. 9 whi
after Dec
included
the time
including
and after

I
this stud
annual re
3 detaile
Year. a 5
ing which

1

3,4 Proc .

relates t

Stated in

 

49

effective date (for annual fiscal periods) Of APB Opinion

NO. 9 which became effective for fiscal periods beginning
after December 31, 1966. Annual reports for "1967" were also
included to the extent available (68 Of the 76 companies) at
the time these data were being collected. The purpose of
including this period was to allow for comparisons before

and after APB Opinion NO. 9 became effective.

The financial statements examined for purposes Of
this study were those included in the company's published
annual report to stockholders. In those cases in which both
a detailed and condensed report was prepared for a single
year, a single report was examined. The criteria for select-
ing which report to examine were as follows:

1. The financial statements must have been

audited as evidenced by the inclusion of
the auditor's Opinion in the annual report

to stockholders.

2. The report must have been sent to stockholders
by the company without the stockholder making

a specific request to receive it.

3.4 Procedures for determininqearninqs per share

Since the individual investor's forecast Objective

relates tO his individual share in the company's future earn-

ings potential, all income measures used in this study were

stated in terms of earnings per share. Earnings per common

 

 

 

share’ af
were d€te
period (1

23515:

It will b
the three
income E
accounts
tent of r
transacti
determina
adeStmen
to determ

C

“!

 

 

Urns llSel

 

 

snare: after adjustments for stock splits and stock dividends,

vmre determined for each fiscal period in the 21 year test
period (1947-1967) for each company on each Of the following
basis:

(A) As reported in the annual report

(B) Current Operating performance

(C) Modified all—inclusive (per APB Opinion No. 9)

(D) All-inclusive
It will be recalled that income measures B, C and D represent
the three income concepts whereas A and C represent the two
income practices. The procedure followed was to classify all
accounts reported on either the income statement or the state-
ment of retained earnings as either capital adjustment or
transaction, normal income determinant, extraordinary income
determinant, prior period adjustment—Old, or prior period
adjustment-new. The following relationships were then used

to determine the various income measures:
CUR - (XDET + PPAO) = MOD - PPAn = ALL

where CUR = Current Operating performance income
XDET Extraordinary income determinants
PPAO = Prior period adjustments—Old*
MOD Modified all-inclusive income
PPAn = Prior period adjustments-new*
ALL = All-inclusive income

*These terms are defined in Section 3.4.2
Forms used for the purposes Of collecting data and determining

income measures have been reproduced as Appendices B, C, and D.

 

 

 

 

 

camper is c
amount O
mined fc

use of ‘

period E
the £01]

mites 3.1",

“‘93 0r

fell int

 

51

3.4.1 Puijustment for stock dividends and splits_

Adjustments were made for stock dividends and/or
stock splits. The procedure employed that amount reported
as earnings per share for the year 1967 as a standard of
comparison. An adjustment factor which was applied to that
amount originally reported as earnings per share was deter-
mined for each company for each year Of the test period by

use Of the following formula:

 

, l l 1
Adj1967—n ' 1 ' 1+P1 - 1+P2 ' 1+Pn
where P1, P2,. . .Pn = Number Of new shares

received as a stock divi-
dend or in a stock split
expressed as a percentage
of the number of shares
previously held, for splits
and stock dividends occur—
ring during the year 1967 - n.
3.4.2 Identifiqgtion Of “items"

The term "item" was used to refer to an extraordinary
income determinant, prior period adjustment-Old, or prior
period adjustment-new. When examining financial statements,
the following criteria were used in determining what consti—
tutes an "item".

Accounts appearing on the statement Of retained earn—

ings or earned surplus were regarded as an "item" unless they

fell into one of the following classifications: earnings

 

 

 

distribut
A

evaluate

....

“triordi

01d. and
I

trite“0n ‘

 

 

52

distrllnxtions or capital adjustments.
Accounts appearing on the income statement were
evaluated as follows in deciding whether an "item" exists:

1. An account was considered as being an "item" if
one or more Of the following terms were used in
either the title Of the account or in the title
of the caption under which the account is report-
ed: extraordinary, special, nonrecurring,
abnormal, prior period, or prior year. It should
be emphasized that the terms "other income or
expense", "non-Operating", and "from other
sources" are not included herein.

2. An account was considered as being an "item" if
it falls within any Of the following classifica-
tions with respect to its nature:

(A) Disposal or sale of:
Property, plant, and/or equipment
Investments or securities
Subsidiary, affiliate, or division
(B) Change in valuation bases:
Inventory write—down to market
Change in investment valuation
Other property, plant, and equip-
ment adjustments
Lifo liquidation or replacement
(C) Expenses, losses, gains, etc.:
Foreign exchange adjustments
Nonrecurring plant expense
Discontinued operations
(D) Extraordinary depreciation

3. An account was not considered as an "item" if
it meets neither of the two criteria described
above.

Having identified an "item", its classification as

extraordinary income determinant, prior period adjustment-

Old, and prior period adjustment-new was based upon the

criterion that for an "item“ to be considered as a prior

 

 

period adjus
in APB Opini
prior period
for prior ye
annual repor
could be bot
adjustment-n

ordinary inc |

3-4-3 Ad 'us
the d

It we
CoupaniesI ar
than one Gale ‘
either the us
in the Compar
adjustments w
Year Since an
Cant magnitud
latter Case,
parability of
Other periOdS

jescrib'ed as

It wi

 

 

53

Period adj\1l8tlttent—new, it must meet the criteria set forth

in APB Opinion No. 9. For an "item" to be considered as a

prior period adjustment-old, restatement must have been made
for prior years' reported income in the company's published
annual report. It should be noted that an individual "item"
could be both a prior period adjustment-old and a prior period ?!
adjustment-new. All other "items" were regarded as extra-

ordinary income determines. ' u

 

3.4.3 Adjustments forpperiods not of
the duration of one calendargyear

It was observed that in a small minority of cases,
companies' annual reports cover periods of duration other
than one calendar year. This situation was the result of
either the use of a 52—53 week reporting period or a change
in the company's annual closing date. In the former case no
adjustments were made for a slight deviation from a calendar
year since any resulting distortion would be of insignifi-
cant magnitude for purposes of this study. However, for the
latter case. an adjustment was made in order to obtain com-
parability of data relating to thisperiod with data of
other periods. The procedure for making this adjustment is
described as follows.

It will be recalled that a report for a fiscal year

 

 

 

 

 

 

ending du
seven day
with the
fiscal pe:
ated with
poses. As
pany's anrv
Zine, or twc
Year. Howe
this proced
536d with ea
instances in
513919 caler
IEports were
as111916? per
Of the two P‘
The

Years in whi

Eon

 

ths 01‘ mo

were adjusts.

ifinths COVE}:

Cilen
dar Yea

 

 

54

ending during the calendar year on a date other than within
seven days Of December 31, was designated as being associated
with the preceding calendar year, i.e. a report covering a
fiscal period extending into two calendar years was associ-
ated with the earlier calendar year for identification pur-
poses. As a result it was possible that a change in a com-
pany's annual closing date could have resulted in either zero,
one. or two reports being identified with a given calendar
year. However, for all such changes relevant to this study,
this procedure resulted in at least one report being associ-
ated with each calendar year for each company. In those
instances in which two reports were so identified with a
single calendar year, all flow measures contained in the two
reports were combined and regarded as being associated with
a single period equal in duration to the sum of the durations
of the two periods covered by the two separate reports.

The adjustment procedure followed for all calendar
years in which the associated report or reports covered six
months or more was one of annualization, i.e. flow measures
were adjusted by multiplying by a fraction in which the
numerator is 12 and the denominator is the number of calendar
months covered in the report or reports identified with the

calendar year in question.

 

 

 

3.4.4 PEI

Ti
and compar
reported,"
operating ;
each "item'
as either a
come for th
that amount

measures the

(A)

(B)

 

(C)

 

 

55

3.404 Determination Of income tax effects Of "items"'

 

The purpose of this study was to determine, examine,
and compare four measures of income, termed herein: "as
reported," all-inclusive, modified all-inclusive, and current
Operating performance. For a given company in’a given year,
each "item" disclosed in the financial statements was reported
as either a determinant of that amount designated as "net in-
come for the period" or subsequent to the determination of
that amount. Thus, in order to determine the four income
measures the following conversions had to be made:

(A) From current Operating performance basis to
all-inclusive basis. This involved adding
the reported amount of each "item" to that
amount originally reported as net income.

(B) From current Operating performance basis to
modified all-inclusive basis. This involved
the same procedure as used in conversion A
above except that it was limited to those
"items" which did not meet the criteria for
prior period adjustments as set forth in
‘APB Opinion No. 9.

(C) From all-inclusive basis to current Operating
performance basis. This involved deducting
the amount Of each "item" on a net of tax
basis from that amount originally reported
as net income.

(D) From all-inclusive basis to modified all-
inclusive basis. This involved the same
procedure as used in conversion C above
except that it was limited to those "items"
which meet the criteria for prior period
adjustments as set forth in APB Opinion No. 9.

 

 

 

- Au!—

 

 

In
required 5.
taked as r
required a
"item" in c
effect on 1
durell was
D: and E a}:

Ste

 

 

 

56

(E3 From modified all—inclusive basis to cur-
rent Operating basis. This involved the
same procedure as used in conversion C
above except that it was limited to those
"items“ which did not meet the criteria
for prior period adjustments as set forth
in APB Opinion NO. 9.

(F) From modified all-inclusive basis to all-
inclusive basis. This involved the same
procedure as used in conversion A above
except that it was limited to those "items"
which meet the criteria for prior period
adjustments as set forth in APB Opinion NO. 9.

It should be noted that conversions A, B, and F above
required simply that the amount of the item in question be
taked as reported. However, conversions C, D and E Often
required an estimation of the income tax consequence of the
"item" in question in order to approximate the net of tax
effect on reported net income. Hence, the following proce—

ll

dure was employed in and only in performing conversions C,

D, and E above:

Step 1. If the tax effect of the "item" was dis-
closed in the annual report, that amount
as reported was used. Otherwise, the
procedure employed was as described below.

 

llReference was made to the following sources in
develOping this procedure: Bruton, Paul W. and Bradley,
Raymond J., Federal Taxation (St. Paul, Minn: west Pub-
lishing Co., 1955); Commerce Clearing House, Inc., 1957 U.S.
Master Tax Guide and 1967 U.S. Master Tax Guide, (Chicago,
Ill: Commerce Clearing House, Inc., 1957 and 1967); and
Montgomery, Robert H., Montgomery's Federal Taxes - Corpora-
tions and Partnerships - 1947-48 Vol. I - Gross Income and
Deductions (New York: Ronald Press Company, 1948).

 

 

 

 

1i

 

 

S:

Ste

 

 

 

 

Step 2.

Step 3.

Step 4.

Step 5.

57

If specific information disclosed in the
annual report could be used to determine
the tax effect, the tax effect of the
"item" in question was determined on this
basis. Otherwise, the procedure employed
was as described below.

A limit equal to the amount Of the income
tax expense reported for the year, was
Observed when employing the following
estimation procedures. Thus if there were
no income tax expense for the year in which
the "item" was reported, the tax effect

of the "item" was regarded as zero, i.e.
loss carrybacks and/or carryforwards were
not considered. If and only if a positive
tax expense was reported, was the procedure
described below employed.

From the information contained in the annual
report, the "item" in question was classi-
fied as one of the following:
(a) Ordinary gain or loss
(b) Capital gain or loss
(c) Gain or loss on prOperty used
in trade or business

(A) For ordinary gains and losses, the tax
effect was estimated by multiplying the
amount of the "item" times the average
tax rate, i.e. that ratio existing be-
tween the reported income tax expense
for the year and reported net income
before tax.

(B) For capital gains, the tax effect was
estimated by multiplying the amount of
the item times 25 per cent, the alter-
native capital gains tax rate existing
for corporations over the entire time
period under consideration in this
study, i.e. from January 1, 1947 to
date. For net capital losses, the tax
effect was considered to be zero.

 

 

 

 

 

89!
above Proce
tax effects
were Origin
“as the Gas
ailing the
were used w

Played only

$59 for 3C

 

 

58

(C) For net losses on property used in
trade or business, the income tax
effect was estimated in the same
manner as for ordinary gains and
losses. For net gains on property
used in trade or business, the pro—
cedure for estimating tax effects gave
recognition to the depreciation re—
capture provisions of United States
Federal Income Tax Law which generally
became effective for taxable years
beginning after December 31, 1962.

The procedure was to estimate the tax
effect of these gains reported in
"1947" through "1962" in the same man-
ner as that used for capital gains.
However, for such gains occurring
subsequent to "1962" the rate applied
to the gain was selected from the
following schedule (arbitrarily
determined) according to the period
covered by the financial statements
in which the gain is reported:

 

1963 .34 1966 .42
1964 .37 1967 .44
1965 .40

Several points should be emphasized concerning the
above procedures for estimating income tax effects. First,
tax effects were relevant only where the "items" in question
were originally reported as a determinant of income. This
was the case for 503 out of a total Of 1,033 "items" examined
during the course of this study. Secondly, reported amounts
were used when reported. The estimation procedures were em-
ployed only where there was no alternative. This was the

case for 303 out of the 503 items originally reported as a

 

 

 

determ
proced:
s:ch es
closer
ported 1
income)

which no

35 m

to Figure
detail bej
cance of t
were used

MOdels. E
and Splits
tents) for
determine

 

C‘Jldted by I.

f0110wing

 

l’ear. Als

 

59

Qeteminant of income. Thirdly, it is not claimed these

procedures produced precise approximations but rather that
such estimates did yield converted income measures that are
closer to actual (i.e. that amount that would have been re-
ported had the company employed the alternative concept of
income) than would income measures based on conversions in

which no attempt is made to adjust for tax consequences.

3.5 ‘grocedures for determination Of forecast errors

Before beginning this section the reader is directed
to Figure 3 for an overview of the procedures described in
detail below. To evaluate the relative predictive signifi-
cance Of the various income measures, series Of such measures
were used as inputs to two least-squares linear regression
models. Earnings per share (as adjusted for stock dividends
and splits and, where approPriate, for prior period adjust-
ments) for the firm's most recent five years, were used to
determine the parameters for the model from which a forecast
was made for the following year. After making the forecast
for each year, the linear trend line of best fit was recal—
culated by incorporating actual earningSJxr share for the
following year (based on the income concept under examination)
and omitting. the earningsper share for the sixth most recent

year. Also, for certain income measures, any prior period

GECUCH

 

 

 

 

1!!
11"!

 

 

 

 

 

 

 

wmm3m<mz mommm BWflUNNOh WUQHZWUMNQ ho ZOHEQZHZMWEWQ

MU HHHNNDU Hoar!

 

 

 

 

- m>flmsaocanaa< ,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hops:
. -n , x
.5 mmm HMSHUG
jvm uouum .x. A 1 ”flame Honum X _ _ , . r ,
- n r 05Hm> Hopes Huxmmm
1 unmomuom counmfloz
mm .Al‘ Vfim< Honnm R .- qixil , .
Houum X “W, «,0>HumcH0#H¢ mlxmmm
, x Hops:
mmm Hm5#04 N
- w mu mmm
m>sm5a0gnuaa<
. - Hmong wax
mm Hounm $.1T m<.HOHHm x.fi -- xmmm Hmsuod mam
. a r: OSHm> . Hope: mix
. - - mMUOHOh . o mam3c n. mmm
. Ha . u u up . o ,
Hm Hounm .x. Houum w» -

 

 

 

- w m>aumcwouafi

 

 

 

 

 

 

 

0.
6 x Hausa
mmm Hmsuud
. £ua3 comm sx ximdww HOW}, .maopofi munmcw
o>HumcuouHm IEOO son» use» map How monam> Hmsuom coauoflcoum mm poms
coma omnmn who noflag monummoe Hon uaonMMflo m ucmHOMMHO mum whom»
monommofi Honuw tum unmoouom QuooumHmu mum 039 09 m mchoo
unmomHOm 0mm ommusouuom x now» How loam on» How
lucooumm ©o>flH. UCOHOMMHO w mummowuom ms“ mousmflm mam
IOU haumawewm as assuaomwm IuHSmou N one
ImluwwmmmmwMAd. mmmmmmmamu. < m>aumcuouad mcfluuomom oEoosH
mcfluhomom
OEOUGH
nuuuuuuli

:mmmwmdfl2:m0mmm BmflUflMOb mwflfizmommm ho ZOHBdZHZMflBmQ

m mmpth

 

 

61

adjustments (Old or new, depending upon the income measure
under examination) reported in the most recent year‘were used a:
adjust earnings per share for the appropriate priOr year(s)
with the recomputed trend being based on the preceding five
years as adjusted. In short, each forecast was based on in-
come measures associated with the immediate five years in—
cluding the effect of, where appropriate, prior period ad-
justments, if any, to the extent that such items were reported
prior to the year for which a forecast was being made.

The two forecasting models were as follows. One was
a least-squares linear regression model in which earnings
per share figures for each of the five previous years was
given equal weight. The second was similar to the first
except that in making the regressions, earnings per share
for the most recent year was given a weight Of 5: earnings
per share for the second most recent year was given a weight
Of 4, for the third most recent year a weight of 3, for the
fourth most recent year a weight of 2, and for the fifth most
recent year a weight Of 1.

Forecast values, as derived from each of the four
income reporting alternatives were then compared to each of
the following actual values: (1) earnings per share of the

same income series and (2) earnings per share based on the

 

 

(h

 

 

62

all—inelusive concept of income. The dollar difference between
the forecasted value and the "actual" value was then expressed
as a signed (positive or negative) percentage of the absolute
value of the relevant "actual" value in order to (l) avoid
distortions due to differences in per share income levels and
(2) maintain an indication Of whether the forecast value was
greater or less than the actual value. These signed ratios
constitute the "forecast error measures" which were then eval-

uated as described in Chapter IV.

3.5.1 Restatements for_priorjperiod adjustments

Earlier it was noted that where apprOpriate, prior
period adjustments were used to restate the earnings per
share Of prior years, upon which the forecasts are based.

The purpose Of this procedure was to give effect to the prac-
tice (recommended in APB Opinion No. 9) of including historical
summaries in published annual reports which provide investors
with a convenient source of data for making earnings predic-
tions.

It will be recalled that the objective of this study
was to determine which of several income measures would have
been (for the test period) of greatest aid in forecasting
some measure Of earnings potential. Also, it will be re-

called that under the modified all-inclusive concept of

 

 

63

income as set forth by APB Opinion No. 9, the criteria for

an item constituting a prior period adjustment differ signifi-
cantly from those used formerly. Thus restatements of prior
years' earnings per share due to prior period adjustments

were varied depending upon the measure of income being eval-
uated. To be specific, no restatements were made for prior
period adjustment when generating forecasts from earnings per

share figures based upon the all-inclusive or current Operat—

 

ing performance measures of income. Restatements for prior
period adjustments-new as defined by APB Opinion No. 9 were
made for generating forecasts with the modified all—inclusive
income measures. Restatements were made for prior period
adjustments-old, i.e. those so treated in published annual
reports, when generating forecasts from earnings per share
figures as originally reported after adjustment for stock

splits and dividends.

...—, V_.
_ . «....

 

“ rm -.7.’ “If."

 

 

CHAPTER IV

EMPIRICAL FINDINGS

4.1 Overview
The empirical findings of this study are set forth in

the three major sections of this chapter. Specifically, this

 

chapter contains only a statement of the empirical findings 3;
of the study. A discussion of their significance and the re- '
sulting implications for accounting practice is deferred to
Chapter V.
Section 4.2 includes descriptive statistics concern-
ing the forecast error measures associated with each of the
four income measures. Section 4.3 includes a description of
the procedures and findings relative to evaluating the sig-
nificance of the forecasts and resulting forecast error mea—
sures under each of the three income concepts. Section 4.4
includes a description of the procedures and findings rela-
tive to evaluating the significance Of the forecasts and
resulting forecast error measures under each of the Egg
income practices.

Figure 3 (in conjunction with Figure l of Chapter II)

gives a graphic outline of the comparisons reported upon in

64

 

 

A

A

 

 

 

   

 

 

Sections 4.3 and 4.4. Figure 4 can be viewed as single cells

of Figure l and Figure 2.

FIGURE 4

OUTLINE OF COMPARISONS

 

 

Criteria of Comparison

 

Best, i.e. lesser Most conserva—
absolute percentage tive, i.e.,
Comparison of forecast error lesser forecast
3 Concepts:
CUR-MOD-ALL Section 4.3.1 Section 4.3.2
CUR-MOD 4.3.la 4.3.2a
CUR—ALL 4.3.lb 4.3.2b
2 Practices:

 

Codes used: ALL
CUR
MOD
REP

4.2 Descriptive

Table IV

All—inclusive

Current Operating performance
Modified all-inclusive

= "As reported"

statistics

shows for each year, for each four year

period, and for the total period studied, the mean absolute

percentage forecast error measure associated with each of the

four income measures for the 76 companies sampled, where each

income measure is examined under all possible combinations of

the two prediction models and the two forecast Objectives.

Each annual mean

figure (M5) was determined according to the

 

 

" 1 .37.. ._ v. ..-
. '3.
uh

 

 

 

 

 

 

66

following formula :

\l

6
M. = Fji Aji

J i l A

[‘1

 

ji

 

76

where j = year

i number of the company. fa

Fji = forecast amount for year j for company i. g
actual amount for year j for company i.

Aji

Table V shows for each year, for each four year period,

 

and for the total period studied, the average deviation of fl
absolute percentage forecast error measures associated with

each of the four income measures for the 76 sample companies,

where each income measure is examined under all possible com-

binations of the two prediction models and the two forecast

objectives. The annual average deviation figures (MDj) were

determined according to the following formula:

 

__ 76 ._ .
MD. = Z 31' 3
=1

where j = year
i = number of company

ji

El
ll

 

In order to assist the reader in analyzing the data
of Tables IV and V, Table VI was prepared. Table VI shows by

year and within each of the four possible combinations of

 

 

 

 

 

 

\\
“I

 
     
 

"I! A...

67

predietdnle model and forecast objective, the relative ordinal
rank of both the means and average deviations of the forecast
error measures associated with each income measure. The rank
of 1 is assigned to the mean or average deviation of least
magnitude (i.e. “best"), the rank of 2 is assigned to the
next "best", and so forth. Sums of these ranks are also
shown and when ranked give a measure of the overall rank of
the various income measures with respect to specified charac—

teristics of the resulting forecast error measures.

 

The following Observations are based upon the data of
Tables IV, V, and VI:

(1) There is a lack Of consistent support for
either of the two prediction models, i.e.
neither model consistently resulted in better
forecasts.

(2) The choice of forecast Objective resulted in
no consistent effect upon forecast error.

(3) It would appear there is in general a greater
effect upon forecast error resulting from
changes in time periods than from changes in
either prediction models, forecast objectives,
or income measures.

(4) Over the entire period, the "as reported" and
"current operating performance" measures re-
sulted in "better” forecasts than did the
"modified all—inclusive" and "all—inclusive"
income measures. This statement is based
solely on the sums of the annual ordinal
ranks of the means and mean deviations of the
forecast errors associated with each of the
income measures.

 

 

(5)

 

68

There was a significant positive correlation
between the ordinal ranks of the means and
the ordinal ranks of the average deviations.
In fact, for 194 out of the 256 cells in
Table III, or 75.8%, the rank of the mean
was identical with the rank of the average
deviation.

Table VII shows for each year, for each four year period,

and for the tOtal period studied, the percentage of companies
for which each of the four income measures resulted in a fore-
cast in excess Of the related "actual", where each income mea—
sure is examined under all possible combinations of the two

prediction models and the two forecast objectives. The follow-

 

ing observations are set forth for emphasis:

(l)

(2)

(3)

(4)

(5)

Neither of the two prediction models seems to
result in overforecasts more frequently than
the other.

The choice Of forecast Objective results in
no consistent effect upon forecast error.

None of the four income measures seems to
result in overforecasts more frequently than
the others.

It would appear there is a greater effect
resulting in changes in time period than from
changes in either prediction models, forecast
objectives, or income measures.

For the test period as a whole there was an
overall tendency for each of the four income
measures to result in more underforecasts
than overforecasts. This situation prevailed
under all possible combinations of prediction
models and forecast Objectives.

 

 

 

 

69

 

 

 

 

 

wwa won an OHH OOH th NHN am and avH wNH mmH mmH mfia HHN om helema
«MN NMN mma me vMN onN mmm hw HNN mHN NNH ¢®H HNN MNN me mm holvwma
we me me we we #v mHoN Ne om m¢ am we Om mv NNON 0v hwma
sh mm 00 mo eh Ho Ho H0 H5 Or oo 00 Ah om 00 mm mead
mm Om mm Nm mm Om vN MN mv hm mm mm me hm Hm 0N momd
awn mph vmw mam Hmh has FDA mfia man mar mmm cam man hNh NnH hHH vomd
NON OHN Hm mm NON moN an NmH hmH NmH omH NON bmH mfia NvH eva molowmd
Nm vm he he Nm 0m we mfi hm «m be m¢ hm mm me ov nomd
oHH NOH Fm wm OHH HOH Nnv hw¢ MHH «OH hm mm MHH voH hmv Nnv Nome
mm mm me me mm am an an cm 00 Nm vm cm on nv me Homa
mmm 0N0 @ma NmH mmm MNo mm mm mNm Nmm mNm new mNm mhm 0v Nm oomd
me moH mw mNH mwa mmH mh am mmH me Nm NNH mmH 0mm on Am mmlwmoa
«ma mma mm mm va mmH wm mm ewa non vm am ¢®H #wa hm Nw mmma
hwa vma mm hha hmH mmH Ho so mmH boa om cma mod hoH mw oh mmma
mHA waa mm NHH mHH mm ow no mad ONH No mHH mHH Nm mm 00 hmma
ASH cud Nm me Aha mma mma hwa boa Nha HMH mma Fwd mmH QMH me wmma
He no mm ow He no No hm mm mm mm vw mm mm H@ mm nmlemH
Hm 0m av Nm Hm Nm mv mv Nm am am mm Nm Nm am as mmma
mm om on we mm om moH vb mm mm mm em mm mm mm mm vmmd
ev Vm mm mm we mm mm mm VG #0 am No vv ow N¢ no mmmd
hm mm mm hm hm mm mm mm mm mm ow mm mm Mm mm mm Nmma
w wm um .... w mm mm w H mm um w H mm mm w .233:
I I? 11 I YLP I I I .lP 11 I st? 11
. .1 31 H _ .1. 31 H _ .1 31 H _ .1 §J 8
1 T3 03 a 1 T3 03 a 1. T3 03 a T T3 03 a
u ur. mu d u u? mu d u UT. mu d u ur. mu 5
a as 3 o a a a 3 o O a a 3 o a a a 1 o
I 1:? e J I v.9 E I T 1.9 B 1 I T.P E J
n n u 0 3 n n u 0 3. n n u 0 3 n n u 0 3
S S Dd a S B D d a S 8 3d a S 3 ad a
t. I as P I. I. as P I r. as P I t. as P
A A 1 A A J A A J A A 1
a e e a a e a a e a a E
3 3 2 3
m. m. m. m.
5 6 5 5
w>aw3HocfilaH< muwhmm flEmm 0>wn5HOENIHH< wwwuum wfimm uo>M000nAO
nub—3.83 wounmwwS pounmwozcb couzuwuzzs " H000:

 

mmmbmtfll MEOUZH Mbom ho IU<H mom mmHZ4AZOU Rafi mom
mmmbm<m= mommm Bw<om¢0m mwtfizmummm NBDAOmfld m0 dem2

>H WAQEB

 

 

 

 

7O

 

 

 

 

 

 

th VON v2” mmfl th NON New mNH N¢N VNN me nMN NVN NNN hmm NNH helemH
oov wmm mmN mwN 00* Nos mmoa mm mnm Ohm ohN mwN mum mhm m2: 5. hwlvwma
Nv N». we .3. Nv Ne comm He me hv om wv me we mmmm wv hwma
Hoa HOH om mm HOH OHM mm mm mm mm E. .8 mm 3:” Nu am wwma
NM ON on 5N Nm mN ma Z” vc mm mm Nm Ev mm mN NN mwma
vaH 0N: Ohm HNm vaH mN¢H av.“ HAN VOMH .vOMH mvm mam VOMH non.“ mmH mmH vwmd
NNN omm moa Q: th m¢m omN NmN gm VMN th Hmv Sum OMN mNN NNN melooma
mm pm we we mm mm mm on mm mm me wv mm ow mm .3 moma
va 0: end oma VmH mvH Hmw emm em.“ 9: NMH mma wma Y: #2. w: Noma
moa mm Ne 3. mod em mm ov moH Vb mm mm moa mm vv me mea
wow Em: mmH NNN vmm 3:: er. mo mm¢ mmo mmm mwma mmm wmw X” Vm owoa
mmN oeN m.: era mmN HMN Nm HHH mHN HNN ma." mo." wHN HHN em MHH amlwmma
owN omN Nm hm omN omN mm hm NmN HmN mm mm NmN NmN oo 0h mmma
th mwN m: mmN NON hoN mm mm BUN ¢NN hm mON SN NNN .3 cm wmma
we." #3 OOH mma Y: Hm mm No o: Nma NHH 0.3 0v." hm vm v0 hmmd
NmN CNN .34 NmN NON VON ooN HON omN mmN so." NNN omN mwN HON mmN mmmfi
mo ow mm Ho mw no aw co #0 Nb pm so «.0 NF Nc mo mmINmma
mm ov mm Ne On Me av oe mm mm mm mm On an an FM mmma
n.3,. was Nm Nm m2” W: 1: Na ONH mNH NOH mma OMH 0N” 9: new. vmma
on em mN «m an hm mm mm mm mm Nm mm mm .2. mm 2. mmmH
em Vm mm mm vm Nm hm mv Nm Nm hm vm Nm we Nm he Nmma
V V N do V V V H d V V V d D V V V H d V
n nm mm .. n um ”m .. u um .....m .. n um um _. .32....
. . T. 1.1 8 . _ .... 31 8 . .1. 31 H _ . t. 1.1 H
.... 1...: o a a .... I. 1. 06 a I. II. o a .- .... .... I. 00 a
u u 1. mu .0 u u t. mu d u u.... u d u u .... u d
a o a 3 o a o s .4 o a as w .4 o a a s m? o
I I D. e 1 I I D. a J I ID. 1 I I D. J
n n u 0 3 n n u.” 3 n n u .m .4 n n u“ 3.
s s ad 3 s s a s s I a a I s o a
I. .... s s P t. I. as D. .... .... s a D. .... .... as D.
A A J A A 1 A A ... A A 3
s s e a s e s s e s s I
... .4 .4 .4
.... t. .... ....
u u u u
6 6 .b .b
0>wm5H0CNIHH€ DOHMQW gum QDvIF—FLEVIHH< nfldkflm gflm ufl>Hu00flQO
03:3 03 vuucmwoz pousmigcb pounmwozcb "Hog:

 

magma NZOUZH gorm mo :04” mom
mam/NH: moan BméummOh Hughzwvmmm NEH—How: ho WZOHB¢H>NQ N0§m>¢

> Sud“.

 

we av VN NM m¢ mv NM OM hv Mv MN mM mv be HM ON axe—mm

 

 

 

D: no can
em Ov MN HM om Hm HM ON HO 3. ON vM Hm me HM MN 32mm
emu: mo Sam
NM NN vv HH NM NN 3 HH MM NN .3. HH MM NN vv HH MOMH
Mv MM HH NN MM vv HH NH Mv MM HH NH MM 3‘ NN HH OOMH
3. HH MM NN 3. MM NN HH 3 NH NM HN .vv MM NN HH MOMH
Mv VM HH NN MM Mv HH NN MM MM NN HH vM Mv NN HH VOOH
MM vv HH HH NM 3 MN HH .3. MM NH HN Nv MM NN HH MOOH
3. MM HN HH NN HH Mv VM vs MM HH NN NN HH Mv eM NOMH
3» MM HH NN we MM HH NH vs MM HH NN vv MM HH NN HOmH
MM vM HH NN MM 3‘ HH NN NM HH MN 3 g MM HH NN OOMH
MM Mv HH NN MM Mv HH NN vM Me HH NN MM MM HN NH mmmH
Mv QM HH NN Mv MM HH NN MM .3. HH NN MM 3. HH NN mmaH
ve MM HH NN 3.. MM HH NN NM vv HH MN vw MM HH NN hmmH
MM .vv HH NN MM vv HH NN MM 3 HH NN NM 3V HH MN OmmH
NM MN HH 3 HM 3 MN NH HM HH HH ve NM MM MN HH mmmH
1 av MM HH HH MM NN 3. HH vM MM HH NN vM Mv NN HH vmoH
7 NN Mv HH MM NN 3 HH MM NN Me HH MM HN Mv NH MM MmmH
HN HH 3 MN MM NN iv HH HN HH 3 MN MM .NH MM HH NmmH
.1. m. a... W. m mm a... w w w... «m. w w mm s... .... .23....
I I? J: I I? J: I I9 11 I IP 1:
_ _I. 31 a _ _I. 31 H _ .I J: H _ .I. 3: u
I. I3 es s I. I; 09 s I I; 03 s I I; 03 s
u UI. Ju d u uI. mu .0 u UI. an d u ur. Ju d
a as m: o a as 1 o a as m3 0 a as m? o
I In. E 1 I ID. 2 1 I ID. a J I ID. a. 1
n n U0 .4 n n no 3. n n u.“ .4 n n uO .4
s 3 ad a s 3 ad a s s a s s 5 ad a
I. I. as P .I. I. 33 D. I. .... 3 3 D. .... .... a a D.
A A J A A J A A 1 A A 1
s s e s s E s s e s s s
3 3 3 1
I I. I. I
u u u u
.b .b .b 5
GD mus—FL: wldfifl DUHHOm 05mm ¢>w¢=~UCMIHH¢ mflfihflw OENm u0>MUUUflQO
vauonoz pouanoz pousoHozcb pousmHosca "Homo:

 

 

nosHm> ann 0» 5H 503 5:11:31 coo... mo scan a .n . uHuHO vacuum
«93.3 £3": 0» 30H Ecuw 59: no xcmu moumoHucH unHu unuHh

> g >H man—28 2H 232m 2
mZOHB<H>uQ HU§N>¢ 924 mg m0 02H; EHQMO

H> NHNEB

 
 

 

 

 

 

 

 

72

 

 

 

 

 

 

 

3. 3. 2. ma 3 3 2. ma 3 ma 3 $-33
ma ov as an an an mm. mm am am mm souvomH
ON ON ON OO OO OO OO OO OO OO OO NOOH
HO HO OM OM NM OM OM OM NM OM OM OOOH
OM ON ON NN HN ON OH NN HN ON HN mOOH
HN OH HN HN NN OH NH HN HN NH ON OOOH
om OO mO OO om OO OO OO OO OO mO NO OO NO NO NO MOIOOOH
NM NM MM MM NM NM OM OM NM MM OM OM NM MM OM OM MOOH
OO NO NO OO OO NO MO OO Om OO OO Hm Om Hm Hm mm NOOH
Om Om Om Om Om Om Om Om Om Mm Om Mm Om Om Mm Mm HOOH
mm mm Om Mm mm Om Mm Om om om NO NO Om OO NO mO OOOH
mm mm Om Hm Mm Mm Om mm mm Hm Hm Om Hm Mm Om Om OmnOmOH
NN HN ON HN NN ON ON ON ON mN ON ON ON ON MM ON OmOH
OO OO OO NO OO OO ON ON NO NO OO OO NO NO ON HN OmOH
HN ON OO OO HN ON ON ON OO HN ON ON OO mN mN ON NmOH
Hm Hm NO OO Hm Hm Hm Hm HO HO NM NM HO NO NO NO OmOH
HO 0O OM OM HO OM OM OM NO HO OM OM NO CO CO OM mmINmOH
ON ON NH ON ON mN HN ON NN HN OH HN NN ON OH HN mmOH
NM NM OM OM NM MM OM OM mO NO HO HO MO OM NO HO OmOH
OM OM OM OM OM OM OM OM HO OM OM NM H9 OM OM OM MmOH
HO Om Om Nm HO HO HO Om HO HO Nm Om HO HO mO Om NmOH
V Vw d3 V we VW d3 V V Vw .gu V V VW d3 V
I I0 an S I I0 an S I I0 an 8 I I0 8n 3 "OHDOMOZ
I TIP .41 I TIP 4‘1 I TIP 11 I ID. 31
_ .I. 3: H _ .I. 31 H _ .I. 31 H . .I. 31 8
I I49 09 s I. I4: 08 a I. 151. as a I. I4: 08 s
u UT. 1U d u UT. lu d u UT. Ju .d u UT. mu d
D 3 a m 1. O D D a m 3. 0 D D a m .4 O D 38 3 O
I TIP 9 1 I TIP 9 1 I TIP 9 1 I ID. 9 1
n n Unv 3 n n uxu 3 n n "OO 1 n n "uO 3
s 8 ad a s 3 ad a s 5 ad a s 8 ad a
I. I as D. I. I- axe P I. .5 as D. I. I. as P
A A J A A 1 A A J A A 1
a 8 E a a p. a a e a a D.
14 ...O 3. .4
I. I. I. I.
u u u u
6 5 5 5
O>HOSHUCHIHH¢ moHuwm OEMO O>HODHUCHIHH< OOHHOO 06mm uo>HuuwnnO
poucmez OOuLOHo3 pouanozcs couanozco ”H000:

 

 

mbom m0 m0<m SUHm3 “Oh OMHZ<OEOU ho mdeZMUKWA

mbmdommommm>0 2H QWBHDOQO mmfibmdmz mZOUZH

HH> mHm<E

 

73

A‘°3 Procedures and findin 5 relative
I—_0 each of three income concepts

 

This section is divided into two sub—sections. Sub-
section 4.3.1 concerns hypothesis C.l, i.e. it deals with
comparisons relating to absolute amounts Of forecast errors.
Sub-section 4.3.2 concerns hypothesis C.2, i.e. it deals with
comparisons relating to signed (positive or negative) amounts
of forecasts.

4.3.1 Procedures and findings relative to evaluatipg
theppigpificance of absolute forecast errors

Null hypothesis: When comparing the results Of em-
ploying different income concepts to a company for a time
period, there is no difference in the expected number of
lowest absolute forecast error measures associated with each
of the income concepts, and any observed differences are merely
chance variations to be expected in a random sample from the
rectangular pOpulation where fl = f2 = f3. This is a more
rigorous restatement Of hypothesis C.l.

Alternative hypothesis: The frequencies f1' f2, and
f3 are not all equal.

Statistical test: Since this procedure involved com-
paring sample data with a presumed pOpulation, the goodness-
of—fit type of statistical test is apprOpriate. The chi-square

one-sample test was chosen because the hypothesis under test

 

A

4|.—

 

 

 

74

Concerned a comparison of Observed and expected frequencies
in discrete categories., The categories were the three income
concepts.

In order to acquaint the reader with the method of
the chi-square one sample test, the following description is
1
presented:

The null hypothesis may be tested by

 

2
2 .. (Oi " Ei)
X = E E-
i=1 1
where Oi = Observed number Of cases categorized
in ith category
Ei = expected number Of cases in ith
category under Ho
k
E directs one to sum over all (k) categories
i=1

(This) formula directs one to sum over k categories
the squared differences between each Observed and
expected frequency divided by the corresponding ex-
pected frequency.

If the agreement between the Observed and expected
frequencies is close, the differences (0i — Ei) will
be small and consequently X2 will be small. If the
diVergence is large, however, the value of X2 as
computed from (the) formula will also be large.
Roughly speaking, the larger X2 is, the more likely
it is that the Observed frequencies did not come
from the pOpulation on which the null hypothesis is
based.

 

1Sidneysiegel, Nopparametric Statistics for the Be-

havioral Sciences (New YOrk: McGraw-Hill BOOk Company, 1956),

p. 43.

 

 

 

 

—2;- ] |

 

 

75

“In contemporary statistical decision theory, the
procedure of adhering rigidly to an arbitrary level of sig-
nificance, say .05 or .01, has been rejected in favor of the
procedure of making decisions in terms of loss functions,
utilizing such principles as the minimax principle (the prin-
ciple of minimizing the maximum loss)."2 However, for this
study, as is the case with most research in the behavioral
sciences, there is a lack of information concerning loss
functions. Thus in order to allow the reader to reflect his
own views regarding the "losses" or “gains" involved in
achieving information of greater aid in making forecasts,
probability levels associated with the chi-square statistics
calculated from the observed data are presented in Table VIII.
These figures indicate the probability of a statistic equal
to or greater than the given chi-square statistic. Once the
reader has determined his own significance level, he can turn
to Table VIII to determine whether the null hypothesis is to
be rejected for any given combination Of time period, fore-
cast model, and forecast Objective.

Based upon an examination of Table VIII it is the
writer's opinion that there is an "overall significance" in

favor of the current operating performance concept of income

 

21bid., p. 8.

 

 

 

 

OmHo. ODD Mooo. ODD mNNo. ODD Oooo. ODD NOIOOOH

 

 

 

 

 

NOON. ODD mHOH. ODD OOOO. ODD MNON. ODD NOOH
MNOO. ODD mNmo. ODD OONO. ODD OMoo. ODD OOOH
OONN. ODD OOMO. ODD ONHO. HH< OMmM. ODD mOOH
HOOO. ODD HONN. ODD OONm. ODD OMmM. ODD OOOH
OOOO. ODD OOOO. ODD OOOO. ODD oooo. ODD MOIOOOH
HOOO. ODD NHOO. ODD MNOO. ODD NOOO. ODD MOOH
OMNH. ODD HMMo. ODD MNOO. ODD HMMO. ODD NOOH
OONO. ODD MOOO. ODD OMHO. ODD OHOO. ODD HOOH
OOOH. ODD OMOO. ODD mNmO. ODD HOOO. ODD OOOH
OOHO. ODD OOOO. ODD OMHO. ODD NMOO. ODD OmIOmOH
ONOM. ODD HMMo. ODD ONOM. ODD HONN. ODD OmOH
HOOO. ODD mOON. ODD mNmo. ODD ONOM. ODD OmOH
mNmO. ODD OMmM. ODD OONO. ODD HMMO. ODD NmOH
OONm. ODD OOOH. ODD NOHO. HHO HOHO. ODD OmOH
HMoO. ODD OOHO. ODD mooo. ODD ONOO. ODD mmINmOH
ONHO. ODD OOHO. ODD OOHO. ODD ONOM. ODD mmOH
OmNO. ODD HOOO. ODD OOOH. ODD ONOO. HHO OmOH
NOOH. ODD OMHO. ODD OmNO. ODD Omoo. ODD mmOH
HOOO. ODD OONm. ODD ONHO. ODD mOom. ODD NmOH
. . . d.a .3 .
m www.mm m www.mm m .mNmsmw m armamo 60.3..
0 un... u 0 un... u 0 un... u 0 un... ..u
q IlsIa q IJsIa q TlsIa q IJaIa
p. 38308 9 92303 E 89008 E 39003
q s.Le Md q s.le Md q S.IE Md q s.lp.Md
I. 7:581. T. T.Sa_.+ I. Isa? T. Teen.
I 1.45 I 3.48 I 14+s I 3.48
I. A q+1 I. A .41 I A .41 I A .41
1. a a 3 a a .4 a 3 3. 8 8
A 01.9 s A O..e s A o 1 Be. A 07.9 s
316th JJQn 1:10..“ valqnu
OSI OSI OSI OSI
3103. 3103 310.4 3101.
0 III. 0 III. 0 .LI. 0 TIT.
m3nu m3nu mJnu NJnu
.035 .035 .035 .035
la la la 13
O>HmaHDcHIHH¢ mmHumm wEmm w>HmsHucHIHH< mmHuOO OEOO “waHuUOMnO
823...... H529. a: .5233. 3233.5 Honor.

 

HnmecmmEoo O0 NuHHmuDHQ.uoO uouum ummuouoO ouaHOmHm
umm3oH :H OOuHammu um>mc udwucoo meu OUCHm 30Hmn umommm uoc moon. ammocou O>HmaHucHIHHm OOHOHOOZ u DOS
umoocou oucmEuoOuOO OcHumquo ucouuao u ODD
uOmocou O>HmaHucHIHH< HHO "OunD mOOOD

OOOOm BOODDOOO HBDHOmmd Bmdmd ZH QmBHDmmO
OBOODZOD OZODZH mmOIB O0 dem OmDZD mmODmdmz NZODZH DDHI3 OOO
mmHZCOZOD OO OmmZDz Qm>OmmmO ODE m0 OWHEHHHOODOOO mHO2<mINZO MOQDOOIHDD

HHH> HHO<E

 

 

 

 

   

77

 

IEPOItiJug relative to the all-inclusive and modified all-
inclusive concepts, with respect to associated absolute fore-
cast errors. HOwever, there is a possible explanation of
this "overall significance" apart from a tendency of the cur-
rent Operating performance income measure to produce better
forecasts than those generated with respect to either of the
other two concepts. This "overall significance" could have
been the result of relative extreme similarity between income
measures under the all—inclusive and modified all-inclusive
concepts, a situation which one could support on ania priori
basis since. for current year figures, the only differences
are prior period adjustments-new and only 118 such adjustments

were involved in the sample data of this study.3 In order to

 

3For example, assume the following Situation:
Percentage Forecast Error

 

 

 

Current
Company Operating Modified
No. Performance All-inclusive All-inclusive
1 20 9 8
2 20 8 9
3 8 20 22
4 9 22 20

Frmm this data current Operating performance would appear
"best" when comparing all three income concepts, i.e. it is
"best" for 2/4 of the companies as compared to a "no differ-
ence" expectation of 1/3. However, if only the current
Operating performance and modified all-inclusive concepts are
compared, the observed number of companies for which each con—
cept is better = the "no difference" expectation = 1/2. This
is also true when comparing only the current Operating per-
formance and all-inclusive concepts.

 

 

eliminate

sons were
ance and tr
the currer
These proc
two sub-se
4.3.13 PJ
c
a
N
Ploying t,
inclusive
no differ
"ith each
the direc
terms of
the Sum O
A

ranks < t

associate

 

will be 1'
TI
I

and inv01

     

78

eliminate this possible bias two additional sets of compari-
sons were made: one involving the current Operating perform-
ance and modified all—inclusive concepts, and one involving
the current Operating performance and all—inclusive concepts.
These procedures and findings are presented in the following

two sub-sections.

 

4.3.la Procedures and findings relative to E
comparing the current Operating_performance a _
and modified all-inclusive concepts 9 ‘-

 

 

Null hypothesis: When comparing the results of em-
ploying the current Operating performance and modified all-
inclusive concepts to a company for a time period, there is
no difference in expected absolute forecast error associated
with each of the two concepts. Relative magnitude as well as
the direction of observed differences is considered, i.e. in
terms of the Wilcoxon test, the sum of the positive ranks =
the sum of the negative ranks.

Alternative hypothesis: The sum of the positive
ranks (:the sum of the negative ranks, i.e. forecast error
associated with the current Operating performance concept
will be lesser.

The Wilcoxon matched-pairs signed-ranks test was
chosen because this hypothesis employs two related samples

and involves two scores which can be ranked in order of

 

 

 

 

absolute

possess Ir
ordered I:
I
sethod of
following
L
r:
s.
d
w
s:
e
a
T
T
d
i
r
‘I
\

K

W

 

 

79

absolute magnitude, i.e. the absolute forecast error measures
possess measurement characteristics in the strength of an
ordered metric scale.

In order to acquaint the reader with the rationale and
method of the Wilcoxon matched-pairs signed-ranks test the
following description is presented.4

Let di = the difference score for any matched pair,
representing the difference between the pair's
scores under the two treatments. Each pair has one
di‘ To use the Wilcoxon test, rank all the di's
without regard to sign: give the rank of 1 to the
smallest di' the rank of 2 to the next smallest,
etc. When one ranks scores without respect to sign,
a d- of -l is given a lower rank than a di of either

1
—2 or +2.

Then to each rank affix the sign of the difference.
That is, indicate which ranks arose from negative
di's and which ranks arose from positive di's.

Now if treatments A and B are equivalent, that is,
if H0 is true, we should expect to find some of the
larger di's favoring treatment A and some favoring
B. That is, some of the larger ranks would come
from positive di's while others would come from
negative di's. Thus, if we summed the ranks having
a plus sign and summed the ranks having a minus sign,
we would expect the two sums to be about equal under
Ho° But if the sum of the positive ranks is very
much different from the sum of the negative ranks,
we would infer that treatment A differs from treat-
ment B, and thus we would reject Ho. That is, we
reject Ho if either the sum of the ranks for the
negative d.'s or the sum of the ranks for the posi-
tive di's 1S too small.

 

4Sidney Siegel, Nonparametric Statistics For The
Behavioral Sciences, p. 76.

 

 

 

Th

tive data.

4.3.11) g;
5:_c
E?
T?
cal test,
of sub-sec

here. The

Presented

4.3. 1c 9}
T‘
Tables VI

(

 

 

80

The findings of this test, along with certain descrip-

tive data, are presented in Table IX.

4.3.lb Procedures and findings.relative to
comparing the current operating
performance and all-inclusive concepts

The null hypothesis, alternative hypothesis, statisti-
cal test, and presentation of findings exactly parallel those
of sub-section-4.3.la and consequently will not be discussed
here. The findings, along with certain descriptive data, are

presented in Table X.

4.3.lc Observations concerning Tables VIII, IX, and X
The following observations are drawn from the data of
Tables VIII, IX, and X.

(l) The choice of prediction model apparently has
no consistent effect as to which income con-
cept yields the smaller forecast error.

(2) The choice of forecast objective apparently
has no consistent effect as to which income
concept yields the smaller forecast error.

(3) Forecasts associated with the current operating
performance concept involved smaller relative
absolute error during the four year period 1960
through 1963 than during any other four year
period covered by this study. It should be
noted that for this four year period only,
the null hypotheses were rejected at the .10
level for all three comparisons given any
combination of forecast objective and predic-
tion model.

 

lili|||

tr .....z...L :: C Di R.,: l; .L... f i :1 ....ZDOOZU QZ<

FCC—s: ....)OMZJUZOI44< sum—~1— ~QCZ ”OZ? 1:023 WHEDE<EZ KZCUZH :Lfn3 QEE<~UOWW<
IOszm .rm<.x:.CL ...: 33.25:: HEDJOmnax O... FUHLmBz ILL}: m20m~O<LZOU

 

Xi ”Jr—(LL

 

 

 

 

81

 

.onuu on: oucoueuuwo ozu shut: ncoHum>uomno Has

no msflmmouv v5» 0» $50 uoaaa>

: unauOuuwc co comma cum noaufiafinnnoum amaze.

 

 

 

 

Hm ODD Hm ODD mm ODD Vm ODD nonmmma
mama. Do: am no: mmmm. no: Hm DOE hmuv. ODD mm ODD NHow. ODD Nm ODD hwlvomd
ovum. 90: Nm ODD moNv. ODD hm ODD HHho. no: on II H~ON. no: mm ODD hwmn
vaw. ODD mm ODD Hhmm. ODD em ODD mwoo. ODD hm ODD mmho. ODD mm ODD ooma
mmmo. 002 mm no: ammo. no: am 902 wvoH. ODD mm ODD mmmv. ODD om II mwwn
HDOM. DOE mm Do: oohm. D0: mm 90: @mmm. not an no: Nmnv. 002 vm no: vwma
mmoo. ODD mm ODD mmwo. ODD hm ODD hHoo. ODD ow ODD Hooo. ODD mo ODD mwlooma
mmod. ODD wm ODD Hmmv. ODD mm ODD OVHO. ODD mm ODD mmHo. ODD mo ODD nwwa
mnnm. ODD Fm ODD mva. ODD mm ODD mmNm. ODD om ODD 00mm. ODD mm ODD NDOH
movM. ODD vm ODD ammo. ODD oo ODD mmmo. ODD mm ODD vmoo. ODD mo ODD mea
mmma. ODD wm ODD 05mm. ODD hm ODD mmmo. ODD mm ODD omoo. ODD mo ODD coma
hamo. ODD Hm ODD mmco. ODD am no: mOmH. ODD Hm ODD bane. ODD Hm ODD mnIOmmH
monm. ODD mm ODD mmmv. ODD am no: moHH. ODD hm ODD Hmmv. no: am no: mmma
nvvm. DOE pm Do: Noma. ODD mm ODD vmmv. ODD mm no: awe“. ODD Hm ODD mmmH
ammo. ODD mm ODD NNHO. ODD mm ODD ¢©mm. ODD am no: ovoH. ODD mm ODD hmmH
Doom. ODD mm ODD ommm. 00: mm DOS ovmm. ODD Hm ODD hamw. ODD mm 902 wmma
mohm. Do: 0m ODD wmvs. ODD Hm ODD Oman. Do: an D02 mmmv. no: on II mmINmmH
mvvo. Do: 00 D02 mawo. no: Vm DOE mDmN. DOS mm D02 mmmv. no: Hm ODD mmmH
VHmm. ODD Vm ODD nvmm. ODD mm ODD Omen. 002 an no: HNHN. no! em Do: cmMH
mmom. ODD cm ODD Hmmm. ODD om II mmvo. ODD mm ODD «wao. ODD ow ODD mmmd
some. ODD mm ODD mvmm. ODD mm ODD mmmo. no: cm DOS ovho. 902 mm 902 ~mm~
o O maID 33d 1.313 0 O maID 33d IDIO o O mDID 35d 1.3.13 0 0 maID 33d 1.3.1.3 "DOAHDO
3U 1.900 003 3200 Ju 1.900 009 3900 3U 1.200 003 .4900 ..su 1.200 009 1.900 . .

a USMu uzz ASMu a USMu uxJ Asmu a usnu UJJ ASMu a USMu UJJ Asnu
M 3393 O a .493 M 9.43: o o 333 M 3.43: o a 133 M 9333 a D 393
1.1. D. 13 One o :3 1.1 P 13 SM; o 19 1.3 P 13 ans 0 :9 1.3 D. 13 ans 0 13
In. a d du.u 39 d [B e d dqu 1.9 d I? 3 d duu 3.9 d In. a d du.u 3.3 d
o 1. 0.193 31.3 193 o 1. 6.12.4 31.3. 12.4 a 1. file: 3.1.3. 12.4 a 1. 0.19:. 3.1.3 123
0... A10. as 310. CI Ala. 32 310. OI A10. 32 31a. OI Ala. as DIG.
x a 0 s 1 M udu o o s z x a o s 1 Mnflb o o s J x a o s 1 M UPD five a 1 x a o s J Mnub o o S 1
on. M103 9 a mxoa OP M103 2 a wzoa on. M109 9 3 mice OD. M108 9 a mzoa
u 1. 1.5 s m d 1.8 u 1. T.s s m d 1.5 u I. T.s s w d Is u I. 1.5 csm d 1.8

d I)nn so eJnn d I)nn so EJnn d I)nn eo EJnn d I)nn 90 ann
l J 3 e 3 I Ire} u 0 3 I L J D E 3.I rn[,3 u 0 3.l L 1 3 E 3 I 11E.; u 0 3.l L 1 0 E 3 I 1)L,3 u 0 3 I

o 0593 00 3193 o 0593 00 1133 o 0333 00 3133 o 0333 00 1193

q x 1. M 13 a 1. q x 1. M 13 e 1. q x 1. M J: a 1. q x 1. M1 3 a I

s o P.3u a IO 5 m1au e o P~2u 3 10 s m.:u e o P.9u 3 I0 sm 5U e o P.5u 3.90 s m.su

q uaob Jnm 205 q uaob 11m sob q uaofi 11m 906 q uaofi 13m 206

1. 31 Ad Isa T 1 1 .Ad rs) I 1.1 .Ad rrJ I .4: A.a [.1

I 3 a a 1. E 09 1. I 1.3 a 1. E o a I I 1.3 a 1. e o a I I 1.9 a 1. e o a T

1. a 1 .u u 1. u 1. a J .u u 1 .u 1. a J .u u I .u 1. a J _u u 1 .u

1 s . 1. _ 3 s _ 1. _ 3 s . 1. _ 3 s _ I .

A 3 a A 1 a A 3 a A 3 a

a ( s t ( s u ( s a ( s

"0
uwozmfloz 1 44 . : uuuzmwoz 1 A: : uu>wuwowmu

 

a>iu:rocm-ss<

mowuom uEmw

 

mBOmDZOD AODDV ODZOSOOOODO UZHEOOHOO BZOOODD Dz<
ODOZV O>HquDzquq< DmHOHDOZ HOE OOQZD meDm<mE NSODZH OFHZ DOBOHDOmmO
mOOOOm EmODOOOO O0 AmFZDOZO OBDAOmQO OB BDOOmNO OFHSV WZOmHO<OZOD

XH NAO<B

 

.L .....ALZCD :DDU. E..2<ZO0.0ZEL ".2~.O.<OH._C EZHZKDU Q2<
D _ is» L./h.nu: OLZHI.O.—< .27? ZEEO VACICIKOHZ ..OZCUZO :bnz COL.<HUADV.V.<
9.1221. y~.n..<r..n-._:.— L: ”thuxuré L._L_‘—CU.:< CLI Fume—V131 57—2—3» HMZGWHK<rOO¢FvU

X :.:.<L.

 

.ouun mm: oucmuowuflv wcu ouozz mcofiumsnwmno .72 mo chmnouv 0:... cu 030 mwsam.» c ucwumuwfiu co uwmmn mum mowujwnmnoum www.19-

 

 

 

 

 

 

vm ODD «m ODD mm ODD 0m ODD nonmmmA
wAnm. ODD Nm ODD NamV. ODD Hm ODD OmOH. ODD mm ODD ONDH. ODD mm ODD Owlvwma
flown. AAO Am ODD DOMV. ODD 0m ODD vmam. AAO Om II Ommm. AA< mm ODD homH
Amwo. ODD mm ODD ammo. ODD mm ODD ammo. ODD H0 ODD HOmO. ODD aw ODD ooma
Ofinv. QAO mm AA< DHVN. AAO ow AAO ONHO. ODD m0 ODD hmmm. ODD Hm ODD mwma
Hoov. AA< Hm AAO thv. AAO mm AA< NNOm. AA< Vm AAO ommv. ODD Hm AA< #mma
Amoo. ODD mm ODD mooo. ODD Ho ODD oooo. ODD no ODD OOOO. ODD we ODD molooma
OOON. ODD mm ODD MNMH. ODD mm ODD MONO. ODD no ODD HAHO. ODD Ow ODD moma
NHOH. ODD mm ODD vmvo. ODD no ODD mmNH. ODD oo ODD OmOH. ODD Aw ODD NomH
mooo. ODD mo ODD mmoo. ODD m0 ODD nmoo. ODD mm ODD @Hoo. ODD Nb ODD HomH
NNAA. ODD mm ODD CONN. ODD mm ODD Ahoo. ODD v0 ODD mHoo. ODD mo ODD ome
OVNM. ODD mm ODD Demo. ODD mm ODD OVNN. ODD mm ODD NmmH. ODD Hm ODD mmuomma
mmmfl. AAO Am AA< mmvm. AAO Nm AA< hmmH. ODD mm ODD HVNV. AAO mm ODD OmOH
mHmN. ODD mm ODD ammo. ODD Om ODD OOP¢. ODD mm ODD Amav. ODD mm AAO mmmd
vmoo. ODD Om ODD VOOO. ODD m0 ODD hmow. ODD Am AAO Ahmo. ODD Hm ODD nmma
Oman. ODD Am AAO Hamm. AAO ow AA< thw. ODD Hm AAO mhmv. AAO mm AAO Omma

0‘ mwmm. ODD mm ODD nmmw. ODD mm ODD mawm. ODD mm ODD Omen. ODD mm ODD mmlmmma
no NovA. AAO mm AAO hnmA. AAO Hm AAO omhv. AA< Om II mamm. ODD mm ODD mmmH
vgwm. ODD mm ODD mmmm. ODD mm ODD Hmmv. ODD Am AA< bmmm. AAO mm AAO meH
meow. ODD Om II ONNN. ODD hm ODD Como. ODD mm ODD ammo. ODD No ODD mmmH
mmOA. ODD Ho ODD mama. ODD om ODD meow. AAO mm ODD mvvm. AAO Hm AA< Nmmfi
O 0 m 3 1.3 a I.d 1.3 Isa O O m 3 I.J a 1.d 1.: 1.3 O O m Dylo 3 1.d 1.0.13 0 O m D I.J 0.9d 1.3.l3
3U Y..P.00 009 3200 3U 1.900 003 3900 3U 1.900 009 .4300 ..su [.200 003 3'00 nflOfihwm
a USMU U11 .ASMU a USMU U11 .ASMU 9 USMU UJJ VAIMU a USMU UIJ ’AHMU
M 6330 0 D 360 M 8330 D D 1.33 M 3390 D D 3.80 M 8390 0 D .433
13. 0. 19 3MB 0 33 1.3 0.13 3M8 0 13 I.) P13 9MB 0 13 1.3 p13 9MB 0 J3
ID. 3 .u dUlu 1:3 d In. a d du..u 1.9 d I? 3 d du.u 1.9 d I? 3 d dU.U 1:3 .0
3 1. O.J 2.4 31.1 J E 3 a 1 ~w1 E 3 3 1.3 1 E 3 3 1. QJ 9.4 3 1.3 1 9.4 D 1. Q: 9 3 3 1;4 1.21
OI. AJQ. DE 310. CI (.10. 0a. 010. OT. Ala. DD. 310. OI A10. 32 010.
X9 05) MUSD 0051 X9 OS} MU.b 0051. X9 051 MUSD 0052 X8 051 qu OOBJ
OD. M103 D. 3 mice OD. M103 P 9 NJOB OD. M103 9 9 NJOQ OD. M103 9 a NICO
u ... .15 SN .0 IS u I, IS SM d IS u I. IS 5m .0 IS U I. IS 5m .0 IS
.0 I)nn PO 9300 d I)nn EO BJnn d Illnn BO EJHH d T.)nn 9.0 23““
i1 DC3I 1L1: UO3I L1 DE3I If} u03I L1 393.1 Ir...) UO3I L1 DE3I IL} UO3I
O 0593 00 7:183 O 0333 00 1:133. 0 0833 00 1.133 O 0333 00 1.131.
q x I. M J a a 1, q x 1. M J D a 1. q x 1. M J a a 1. q x 1. M 1 a a 1
E OPIU 61.0 SWJU E OPJU 31:0 SWJU v. OPJU 31.0 BW3U E OPJU 3.1.0 SWJU
u. UBOD 13m 905 q U905 Ilm 905 O. U305 13m 90b 0. U306 13m 05
1. 3 1 A d [.1 1. 31 A.d rs: 1. 3 1 A.d [.1 1. 3 J Ind rs:
I 3901. E 091. I 3931. D. 031. I 3.331. E 031. I 389.... v. 03....
I, 31_U U 1.U I. 31.U U J_U I. 31.U U 1.U T. aJ_U U J_U
3 S. 1. _ 3 S. I. _ 3 S. I. _ .3 S. 1. _
A 3 a .A 3 3 A 1. a ..A .4 3
a I s 1 ( s a I S a ( s
amigos IEFEac: “SEES. . S . : Jones
OanDHUCAIOD/s, won—how 02mm uw>duoomno

 

mHOODZOD AODDO MDZOSOOOOHO UZHEOOOOO FZOOODD 924
”AAOV m>HmDADZH|AAO NOB OOQZD WOODm<m2 HZODZH OEHS DOE<HDOmm<
wOOOOO EmODmOOO O0 AWEZDOZO mkDAOmm< OP FDmOmmO DEHSV mZOmHOOOEOD

X OAOOE

 

 

the leas

Moi/ing

 

 

4.3.2.

(i-e- th

 

 

83

(4) In general the significance levels at which
the null hypotheses could be rejected in favor
of the current operating performance concept
were higher for those comparisons between
only two concepts than for those comparisons
involving all three concepts. Frequencies
regarding majorities and pluralities support
this observation also.

(5) Based upon the data of Tables IX and X, it
is the writer's opinion that the modified all—
inclusive concept fared "better" against the
current operating performance concept than did
the all-inclusive concept.

(6) Based upon the data of Table X, it is the
writer's opiniOn that the current operating
performance concept fared "better" but not
significantly better than did the modified
all-inclusive concept. The difference as
reflected by the Wilcoxon test was not great
enough to allow rejection at the .05 level
for any four year period, excepting 1960
through 1963, for any combination of forecast
objective and prediction model.

4.3.2. Procedures and findinggxelative to evaluating
the significance of relative forecast amounts

 

The purpose of comparing relative forecast amounts
(i.e. the projected amount — n23 the resulting error) under
each of the three concepts of income reporting, was to allow
for evaluation of the relative conservatism of each of the
three concepts. By definition, the concept that results in
the least forecast is the most conservative.

Null hypothesis: when comparing the results of em-

ploying different income concepts to a company for a given

 

 

84

time PBrixxi, there is no difference in the expected number of
least forecast measures associated with each of the income
concepts, and any observed differences are merely chance vari-
ations to be expected in a random sample from the rectangular
population where fl = f2 = f3. This is a more rigorous restate—
ment of hypothesis C.2.

Alternative hypothesis: The frequencies f1, f2, and
f3 are not all equal.

Statistical test: Since this procedure involved com-
paring sample data with a presumed population, the goodness-
of-fit type of statistical test is appropriate. The chi-
square one sample test was chosen because the hypothesis under
test concerned a comparison of observed and expected frequencies
in discrete categories. The findings of this test are shown
in Table XI.

Based upon an examination of Table XI it is the
writer's opinion that there is an "overall significance"
regarding the current operating performance concept of income
reporting relative to the all-inclusive and modified all—
inclusive concepts, with respect to associated forecast
amounts. As shown in Table XI, in every case, it was the
current operating performance concept which resulted in the

most conservative forecasts. However, there is a possible

 

 

 

031-90th
COMPAE

 

 

Codes Use

 

Model:

 

 

85

TABLE XI

 

CHI‘SQUARE ONE-SAMPLE PROBABILITIES OF THE OBSERVED NUMBER OF
COMPANIES FOR WHICH INCOME MEASURES UNDER EACH OF THREE

INCOME CONCEPTS RESULTED IN THE LEAST FORECAST

 

 

 

 

 

Codes Used: ALL = All-inclusive concept
CUR = Current operating performance concept
MOD = Modified all-inclusive concgpt
Model: Unweighted Weighted
0‘“ exp
S m w c m m
-H m o -a m o
u u u o
H m O H o O
5 u u s u u
m 0-H O m 0-H O
m m H u w m H u
u m m -H u m m -a
u u w H u H m H
u m D-H -H u m D-H -H
o.m H c n o.m H s a
32mg 2 82mg 3
Period: 8 c 3 g 8 S c 3 E 3
U-H m o O U-HlH O O
1952 CUR .0001 CUR .0015
1953 CUR .0000 CUR .0015
1954 CUR .0039 CUR .0575
1955 CUR .3536 CUR .1426
1952-55 CUR .0000 CUR .0000
1956 CUR .0015 CUR .0074
1957 CUR .0491 CUR .1371
1958 CUR .1040 CUR .3679
1959 CUR .3827 CUR .5045
1956—59 CUR .0000 CUR .0010
1960 CUR .1082 CUR .1738
1961 CUR .1738 CUR .2638
1962 CUR .1738 CUR .5248
1963 CUR .5045 CUR .3827
1960-63 CUR .0023 CUR .0111
1964 CUR .0344 CUR .1606
1965 CUR .0012 CUR .0241
1966 CUR .0241 CUR .1738
1967 CUR .1414 CUR .1289
1964-67 CUR .0000 CUR .0000

 

 

 

86

explanatixn1 of this ”overall significance“ apart from a tend-
ency 0f the current operating performance income measure to
produce lesser forecasts than those generated with respect to
either of the other two concepts. This "overall significance"
could have been the result of relative extreme similarity
between income measures under the all-inclusive and modified
all-inclusive concepts. In order to eliminate this possible
bias two additional sets of comparisons were made: one in-
volving the current operating performance and modified all—
inclusive concepts, and one involving the current operating
performance and all-inclusive concepts. These procedures and

findings are presented in the following two sub-sections.

4.3.2a Procedures and findings relative to comparing

the current operating performance and
modified all-inclugive concepts

Null hypothesis: When comparing the results of em-
ploying the current operating performance and modified all-
inclusive concepts to a company for a time period, there is
no difference in expected ratios between signed (positive or
negative) forecasts (based on each of two concepts) and asso—
ciated absolute all-inclusive income measures. Relative
magnitude (of the ratios) as well as the direction of observed
differences is considered, i.e. in terms of the Wilcoxon test,

the sum of the positive ranks = the sum of the negative ranks.

 

 

lute mag

istics i
ings of

presente

4.3.2b

 

87

Alternative hypothesis: The sum of the positive ranks
# the sum of the negative ranks.

The Wilcoxon matched-pairs signed-ranks test was
chosen because this hypothesis employs two related samples
and involves two scores which can be ranked in order of abso-
lute magnitude, i.e. the ratios possess measurement character—
istics in the strength of an ordered metric scale. The find-
ings of this test, along with certain descriptive data, are
presented in Table XII.

4.3.2b Procedures and findings relative to thg current
operating performance and all-inclusive concepts

 

The null hypothesis, alternative hypothesis, statis-
tical test, and presentation of findings exactly parallel
those of sub-section 4.3.2a and consequently will not be
repeated here. The findings, along with certain descriptive

data, are presented in Table XIII.

4.3.2c Observations concerning Tables XI, XII, and XIII
The following observations are based on the data of
Tables XI, XII, and XIII.
(1) The choice of prediction model apparently
has no consistent effect as to which income
concept yields the smaller forecast.
(2) Based upon the data of Tables XI, XII, and

XIII, it is the writer's opinion that in
general the current operating performance

 

 

 

.onmN mm3 oucwuwwwflu on» whoa: mcoflum>uom£o :m mo mcfloaouv 23 ca 9.6 moi?» c ucuuwwuflu no woman mum nowadaflnmuoum among...

 

 

 

 

mm mDU hm «DU hmlmmma
OOHO. mDo om mDU N000. mDU #0 who >o|¢oma
«man. mDU A0 “DU mmho. KDU oo mDU mama
wnmw. KDU Om KDU mva. «:0 mm Mao coma
VVNO. mDU mo mDU wooo. mbU we mbo moma
vaH. mDU om mDU Nvmm. mDU mm mDU vomH
omfifi. mbo Hm Q02 @mVH. mbo om II Molowma
mfiwm. mDU mm DOE Vann. no: hm no: momfi
memn. who on in ¢m¢m. mDU Nm mDU mood
mvmm mDU am no: mhma. KDU mm mbo aema
oooa. KDU Nm mDU wmam. mbU mm mDU ooaa
vaM. mDU mm 530 mvma. “DU mm mDU mmIOmmH
vomw. Q0: 0m 90: vam. mDU mm DOE mmma
maom. no: mm no: vmmm. Q0: mm mDU mmma
Nmmfl. moo om mDU wnma. mDU Vm MDU hmaa
Onoa. mDU no mDU woma. mDU no :50 omma
8 @000. ~50 mm “DU Nooo. ~50 No ~50 mmlmmma
no obmm. mDU mm mDU oHOM. MDU vm “DU mmma
OmmH. mDU om mbu momo. KDU mo mDU mea
mmao. mDU mm mbo ommo. :30 00 «DO mmma
mafia. “DU mm «DU meNO. KDO No KDU Nmau
). . ). a a .D .
Wm Memm www mem nu mémm om,“ ono 60:91
I. l. u w.“ m1; .0 ....L w.” "...;
U) WPMW 3M9 Emoa 1.3 00.08 9M9 “NOB
ID. Xemd dqu UEMd IE XBMd dutu HEMd
o 1 o 1 a 3 3 r.) t§$e 3 3 I. o 1 a J 1 I.) T:L.9 3
01 US... 3 901 0.1 Hal 3 301
x a 1 1 M H o s x 1 x a m J M H o s 1 x
CD. QNJB p. x: 1.1.9 0D. 3 3.9 v. 1: t.1:3
U 3105 EN 105 u 31.05 sm 105
d s u 1 n e o A J n d s u 1 n e a A 1 n
.11 14391 IfO 8.1 LI 339T IC.O at
O (0.31 00m 033 O (0.31 00m 033
H e I. M 1 d J? t. q 9 T. M I d 1.E F
e u.s u a 1.2 s u e a.fl u a t.E s u
W KID zlu 36 q £36 zlu 3b
1. s A F s I. s K I s
.l a T a
l s I s
1 1
,A 'A
u a
coucmflwz n 4; . :
"awuoz

 

mkmmUZOU AEDUV MUZdZdOhmmm 02Hb¢mum0 EZNmKDU 924 AQOZv
mpHmDAUZHIAAd DMHLHQOZ HEP KMQZD mumbm<m2 NZOUZH ZOAD Qmm<m
mhmduwmom zumZFum flmEZDOZ< H>H8<qmm OE Fummmmm ZEHSV mZOmHmanOU

HHX mumdh

 

 

ill. I
....LLZ. V. 7%; c ....2..7;.,_; T. u.2_._(:,f:. ...2L:..:_L ZZC »_:.\

na>uw:._UZOl._.h( “Err EECZD “Elam-3‘2: BICUZH ZCAD 33mg
"......m‘UnLZCL ZEEENE AME-230$ E>HL~<¢EK CL. POWER“! 15H»: EZOMHfil—‘OU

Hhhx wing?

 

 

.ouon mmz oucoumwwwp on» ouoz3 ucofium>uomno Ham mo mcfimmoup any 0» 090 nuoan> : acouowmwn so comma own uofiufiawnmnoum amonaa

 

 

 

 

mm moo mm moo monNmmH
mooH. moo mm moo ammo. moo mm moo mouvoms
woes. moo om moo mama. moo mm moo moms
muss. moo mm moo ocmv. moo «m moo moms
Oman. moo mm moo comm. moo vo moo moms
moms. moo mm moo ~mom. moo mm moo oomm
«moo. moo Hm moo ooHo. moo cm moo mouooma
aoov. moo Hm mom mmom. moo Hm mom memo
mmmm. moo Hm mom mNHv. moo mm moo moms
ovoo. moo cm moo mace. moo om moo Homo
cmmo. moo mm moo cmoo. moo mm moo comm
News. moo cm moo ovoa. moo mm moo mmuomma
coho. moo Hm mom mvem. moo Nm moo ammo
mmma. mom mm mom «mam. moo om moo ammo
moon. moo om moo ooom. moo mm moo ammo
emmm. moo mo moo some. moo mo moo ommm
oooo. moo mm moo oooo. moo No moo mmummmm

0/ mmom. moo on In ammo. moo mm moo mama
no mooo. moo cm moo oamo. moo No moo vmmm

«moo. moo mo moo omoo. moo no moo mmmH
cmoo. moo mo moo mmoo. moo mo moo «mas
o ) .D o ). u
..m. memo a: 3.3. mm. ml” 3...... mo...” 8......

0 IS U U11 m1 U 0 .l8 U U11 NJ U
M 3 ID 3 0 .6 IO M D ID 3 D .0 ID
1.3 00.08 3M8 EWOB 1.3 00.09 9M3 Emoa
To. xamd du.U UQMd In. xaMd dU.U UEMd
a? 0383 3T3 IC33 a? 0393 3T3 TCa3
OT. U31 3 801 OT. U81 0 901
x a 1 x M u.o 8 J J x a 1 J M u.o B J 1
o P 3 m J a e 3 T.1.a o D. 3 m 1.9 e 1. 1.1.8
u a 1.0 s s w 3 o s u a 1.0 s 8 m «40 8

d SUI“ 23 £1“ d SUI“ 90 K1“
L1 333T. T...L.0 BI L1 389T IC.O 8....

0 (p33 00m 033. 0 (0.31. 00m 00.4

q 9 I. M J.d 3 e I a. e I. M 1 d .32 I.

2 q s u a t.2 S u 9 0.9 U a I.E s u

0.. K35 JQU 35 0.. 53.5 JQU 35

I. S K I. S I. 8 K I. B

T? S T. s

3 3

A. ....

noocmwoz ooucmmmzco
"mono:

 

 

mfimmuzoo AKDUV moz<2m0mmmm UZHFédmmO BZNMZDU 92¢ AAA¢V
N>HmDAUZHIQA¢ NIB meZD mmmbmdmz HZOUZH 2095 Qfimdm
mbm<0mm0m zmm39mm AWEZDOZ< m>HB<ANm OF BOflQWWd EBHZV mZOmHmdmzoo

HHHX MAQ<B

 

 

 

 

 

9O

concept resulted in significantly lesser fore-
casts than did either the modified all-inclusive
or all-inclusive concepts. It should be observed
that in all but two combinations of four year
periods and forecast models the current Operating
performance concept resulted in a majority of
lower forecasts with respect to both the modified
all-inclusive concepts. There was no exceptions
with regard to the Wilcoxon test. The average
(mean) Wilcoxon probabilities were .1974 and .1
.1138 respectively. 1

4.4 Procedures and findings relative to
each of two income practices

 

This section is divided into two sub-sections. Sub- g
section 4.4.1 concerns hypotheses P.3 and P.4, i.e. it deals
with comparisons relating to absolute amounts of forecast
errors. Sub-section 4.4.2 concerns hypotheses P.5 and P.6,
i.e. it deals with comparisons relating to signed amounts of
forecasts.

4.4.1 Procedures and findipgs relative to evaluating
the significance of absolute forecast errors
4.4.1a Test of hypothesis P.3

Null hypothesis: When comparing the results of em-
ploying different income practices to a company for a time
jperiod, there is no difference in the expected number of
.lesser absolute forecast error measures associated with each

(of the income practices, and any observed differences are

Jtmerely chance variations.

 

 

 

“‘~ ”Wm

 

91

Alternative hypothesis: The probability of a lesser
absolute forecast error is greater with respect to the "as
reported" practice.

Statistical test: Since this procedure involved com-
paring the results of two different treatments (i.e. income
practices) to each company for a given time period, the two
related sample type of statistical test is appropriate. The
sign test was chosen because the hypothesis under test con-

cerns only the relative size of absolute forecast errors.

 

This test requires no assumptions regarding the distribution
of differences, nor is it necessary to assume that the com-
panies are similar in any way.

In order to acquaint the reader with the rationale
and method of the sign test the following description is
5

presented.

The null hypothesis tested by the sign test is that
P(XA>XB) = P(XA<XB) = 1’5

where XA is the judgment or score under one of the
conditions (or after the treatment) and X8 is the
judgment or score under the other condition (or
before the treatment). That is, XA and XB are the
two "scores" for a matched pair. Another way of
stating H0 is: the median difference is zero.

 

51bid., p. 68.

 

92

In applying the sign test, we focus on the direc-
tion of the differences between every XAi and XBiI
noting whether the sign of the difference is plus
or minus. Under Ho' we would expect the number
of pairs which have XA2>XB to equal the number of
pairs which have XA< XB. That is, if the null
hypothesis were true we would expect about half
of the differences to be negative and half to be
positive. H0 is rejected if too few differences
of one sign occur.

. . . The probability associated with the occur-

rence of a particular number of +'s and -'s can

be determined by reference to the binomial dis-
tribution. . . . If a matched pair shows no
difference (i.e., the difference, being zero,

has no sign) it is drOpped from the analysis. . . .
The findings of this test, along with certain descrip-

tive data, are presented in Table XIV.

4.4.1b Test of hypothesis P.4

Null hypothesis: When comparing the results of em-
ploying different income practices to a company for a time
period, there is no difference in expected absolute forecast
errors. Relative magnitude as well as the direction of ob-
served differences is considered. i.e. in terms of the
Wilcoxon test, the sum of the positive ranks (i.e. excesses
of "as reported" forecast errors over modified all-inclusive‘
forecast errors) = the sum of the negative ranks (i.e. ex-
cesses of modified all-inclusive forecast errors over "as

reported" forecast errors).

 

 

 

 

 

 

 

 

93

.OMON

wm3 mucoumwwfio oru muozz mconm>uwmno :m we mafimmouc or... no mop mooamo : acouowwwo co momma mum moflgwnmpoum omocm;

 

 

 

 

 

 

mm mm mm mm nolmmma
Mann. mm wwma. mm Nmoo. oo «#00. Ho holvoma
mmmH. mm nowo. mo Omwfi. mm wave. mo noma
flown. mm 050m. om N000. V0 momo. N0 woma
mono. mv moms. 0v Howe. >0 mhma. H0 mwma
whom. om mmma. mm 000m. Hm mmwm. Om vwmfl
Mano. mm nomm. mm H000. or «000. m0 mwlowma
mvmo. m0 OOnm. om mooo. mo wane. no mwoa
CONN. mm nvnm. 0m woma. 00 00mm. mm mme
OOmV. mm mmwfl. mm >000. m0 NNHO. Hm amma
moma. ow worm. 0m mmmo. mo cmvo. mo owmfl
ommv. Hm Nvom. mm wmom. Vm 0mv~. mm mmlomma
OmOH. ~o homo. m0 homo. no mONH. H0 mmma
000m. Om mmOH. Ho hmmm. hm ommo. m0 mmma
vmmo. we vomm. mm mans. 0v OOHO. aw hmma
mmmn. vv mmom. 0m nmwm. om omom. av 0mma
momm. av mONN. mm owNn. Nm OHmH. vm mmlmmma
mmmw. NV wmam. vv mnmv. mm OMmM. mm mmma
oocm. Hm mnmv. mm hmmm. hm hmhw. mm mea
mHom. om ammo. No Have. we smoo. ow mmma
mvno. me mmmm. mm whom. hm mumm. mm mmma
n08 Ide n08 Id3d n08 Ide n08 Tde
U}? 8109 uJI. 9103 UJI. 3209 u}.... 9103 "COAL—Om
P u s e 1 x P u s e 1 1 P u s e I 1 P u s 2 J 1
a o o s o o a o o s o a a o o s o o a o o s a a
11m 93MB JQm 33Ma ZQN 93MB JQN 83MB
5 I. 1 I.H u s I. 1 I.u.u s I. x I.u.u s I. J I.“ u
u 9 E 3 I 3 u a e 3 I 3 u a E D I.3 u a E D 1.3
n1I Jess nJI Jase n11 Jase nJI 3932
[A 0 Nb IA 0 Nb TA 0 H6 TA 0 Nb
1.9 d J J a I a d 1 1 a T.e d 1 J a I a d J J a
P: 399 P: 932 Pl 939 91 see
H o o s s o u. o a s s o q o a s s o q o o s s o
A o q s n J & o q s n 3 M o q e n 3 fl 0 q s n 3
do? 511 DE 5.11 Do. 811 DE 8T1
o 3 3 3 3 a o o o q 3 3 a a o a q 3 3 e o o o q 3 3 a D
3 n I. a d 0 3 n T a d 0 3 n I. a d O 3 n I. a d 0
H11 aPom QJI apom H11 ePom H11 spam
9 1 I. 1 1 d a J I. J 1 d a J I. 1 1 d a 1 I. J 1 d
s a 3 J I.3 s s a 3 J I.3 E s a 3 J I.3 e s a 3 1 I.3 e
t,u A 0 u a u I u A o u a u I u K o u a u I.u K o u a u
SD... J 9.! SD; 1 0.1. SD... 1 0.7.. SD! 1 0.1.
a a a a a a a a
s s s s
poozoooz toucmmozcn nwucmfloz uwuzmwo3cb "Hone:
o>mmsaocH|~H¢ mownwm oEmm uw>muuonno

 

mmuHFU<Mm m>HmDQUZHIQA< QmHmHQOZ
924 Dmfimommm m4 HIP ZNDZD mm¢bm<m2 HZOUZH EPH3 QHP‘HOOmm1
mmommm Fm<omm0h ho AmEZDOS< mFDAOmm< OE BUNQWHM EFHZH mZOmHm<AZOU

>Hx mumde

 

 

 

94

Alternative hypothesis: The sum of the negative
ranks'>-the sum of the positive ranks.

The findings of this test are presented in Table XV.

4.4.lc Observations concerning Tables XIV and XV

(l) The choice of prediction model apparently has
no consistent effect as to which income report-
ing practice yields the lesser absolute fore-
cast error.

(2) For the earlier years of the test period, the
choice of forecast objective had no consistent
effect as to which income reporting practice
resulted in the lesser absolute forecast error.
However, for the latter years, the "as reported"
practice resulted in significantly better fore-
casts under the same series forecast objective
than did the modified all-inclusive practice.
For such combinations of four-year periods and
prediction models, both the sign and Wilcoxon
tests rejected the respective null hypotheses
at the .01 level.

(3) Based upon the data of Tables XIV and XV, it
' is the writer's Opinion that in general the
"as reported" practice resulted in "better"

but not significantly better forecasts than
did the modified all-inclusive practice.

4.4.2 Procedures and findings relative to evaluating

the significance of relative forecast amounts
4.4.2a Test of hypothesis P.5

Null hypothesis: When comparing the results of em-
ploying different income practices to a company for a time
period, there is no difference in the expected number of

lesser forecasts associated with each of the income practices,

 

 

 

 

‘w.

! tiny-SI... .r

 

Nb"

 

.ooom mm: oocwuowwmo min 0.3:... mc0mu~>uomno :5 Mo 9.5.30.5 mm.» Cu won mooam> : oswuowwflp :0 woman won mofiugfinmnbum mam—E...

 

 

 

 

OHVM. mum mmmN. mum omoo. mum Omoo. mmm nonvwma
Ovmfi mum ommo. mmm nvma. mum mNNO. mum homa
mmwfi. mum wmmv. no: mHHo. mum onwo. mum 000a
Omom. no: 0mmm. no: NNNO. mum VVHH. mum moma
Vmfiw DOE hmVN. mum movv. mum tha. mum vowa
vao. mmm H¢0m. mum H000. mum 0000. max mwlooma
Nova. mum Homm. DOE mmoo. mum ooao. mum mood
Onom. mum moan. mmm mmmo. mum wvmo. mum mama
Nmmv. mum ooom. mum ammo. mum moao. mum Homa
mmwo. mum 00mm. mum OOHO. mum Omao. mum ome
meo. mmm move. mum mwao. mm& meo. mum mmlomma
nvma. mmm NOmH. mum wnmo. mum «Cam. mmm mmmH
hmVN. mum HOHO. mum hobo. mmm HOHO. mam mmma
wnad. mum Homo. mum vhwm. mum mmwm. mum hmma
mmmv. no: mmnfi. DOS mmofl. mum mmov. mmm 0mmH
EJ mem. 00: mflom. no: omen. mum Nmno. mum mmlmmma
OJ NOAH. DOE oNvH. no: HONV. mam muom. mum mmmd
mvwv. Go: owma. DOS «mam. mum vao. mum vmmH
mNOM. mum mmmo. no: HOMO. mum hmoo. mum mmmH
mmmv. mum VVON. mmm memo. no: meH. Q02 NmmH
boO )Jsd 600 )Jsd 500 )Jsd boo )JSd
.d O n 1 .
m3w wmmm MJW wmflu wa wmmu “1w axIs peduom
e 0 1 a 3 a e o 1 a 3 D E o I a 3 0 E o 1 a 3 D
3 q 3 o e 3 3 q 3 0 a 3 3 q 3 D a 3 3 q 3 D a 3
a s e 1 e P I a s E J 2 P I a s E z e P I. a s e 1 s P I
J a I. B S 3 J a I. E s D 1 9 I. E s 3 1 9 I. E S D
(.1 [ u 3 I.3 (.1 I u 3 I.a (.1 I u 3 I.a (.1 I u 3 I.3
A X u A X u A X u A X u
M a d e M M a d a M M e d a M M a d a M
I P J s 1 I q I.P 1 s 1 I q I.P 1 s 1 I H I.P 1 s 1 I q
I 0 n J a I. I 0 n J a I. I O n J a I. I O n I a I
o)q moss o)q moss a)q moss a)q moso
o o o s 1 5 H o o e s J s H o o E s 1 s q o o E s 1 s H
x 1 q ( s a x 1 q (.s a x 1 q (.s a x 1 q (.s a
o I 1 1 o I. 1 J o I. 1 J o I. 1 J
u I a u I a u I a u I a
T. _ .... _ I. _ .... _
.Im 3 (m 3 Tu .3 L 3
A K K A
a 3 4 4
Uwonmmoz pounmaw3ca pwucmHmS Dvozmfio3cb H0002
33305;? magnum 9.5m 533930

 

mmUHBU<¢m mZOUZH AQOZV W>HmDAUZHIAA< QmHmHQOZ 92¢
Ammmv DmHmOmmm m4 LO mommm Bm<UmmOm MBDQOmmd mmmmmg
OE Bowmmmm :EHZ HmHE mxz<MIDm20Hm mdemIQmmohdi ZOXOUAH3

>X Mdmflh

 

 

 

 

 

 

96

and any observed differences are merely chance variations.

Alternative hypothesis: The probabilities of obtain-
ing lesser forecasts are not equal with respect to the two
practices.

Statistical test: Since this procedure involved com-
paring the results of two different treatments (i.e. income
practices) to each company for a given time period, the two
related sample type of statistical test is apprOpriate. The
sign test was chosen because the hypothesis under test con-
cerns only the relative size of forecasts. Findings of this
test, along with certain descriptive data, are presented in

Table XVI.

4.4.2b Test of hypothesis P.6
Null hypothesis: When comparing the results of em-

ploying the "as reported" and modified all-inclusive practices
to a company for a time period, there is no difference in ex—
pected ratios between signed (positive or negative) forecasts
(based on each of two practices) and associated absolute all-
inclusive income measures. Relative magnitude (of the ratios)
as well as the direction of observed differences is considered,
i.e. in terms of the Wilcoxon test, the sum of positive ranks

= the sum of the negative ranks.

 

 

 

 

 

 

Model:

Period:
1952
1953
1954
1955

1952-5

195E
195‘
195
195
1956-

TABLE XVI

97

 

COMPARISONS OF FORECASTS ASSOCIATED WITH
INCOME MEASURES UNDER THE AS REPORTED
AND MODIFIED ALL-INCLUSIVE PRACTICES

 

 

 

 

 

 

M063 1 3 Unweighted Weighted
I I
I o I U
E m s m
0.: H s r4 0.: H a H
U o mw4+I ram Ira o o Qw4+3 .344. H
-H w m o u o -a m m o H 5
¢I£:='U m -H o c WI;:='U m ml 5 :
030a”) Emu o3pmtI Eso
mum OOOH mum OOUH
OHUHH :uomw muuHu cuomm
o\o H o o -H-H 'c-a bno H o o wIwI 'U-H
MMOmm ancm mmomm nchm
u 0.0 -H m o m +I 0.0 -H m o m
c:m o H H H.Q > A: c m $3484 H.Q > J:
m m H w m m H OJJ m m H (D m m H mIJ
0-H m w m.n m o o 0-H Inca m.Q m o o
. 33333 Iaflsa. ségza $835§
Period: mm: 41H ...—Imoux: mau=+IH 4040*“:
1952 55 .3258 53 .4402
1953 48 .4388 54 .3746
1954 55 .3179 51 .5000
1955 49 .5000 50 .5660
1952-55 51 .3792 52 .3454
1956 55 .3220 56 .3089
1957 63 .0631 62 .0998
1958 66 .0298 53 .4340
1959 37 .0877 38 .1077
1956-59 56 .0780 52 .3079
1960 44 .2983 40 .1808
1961 38 .1077 34 .0551
1962 46 .4253 48 .5000
1963 38 .1077 35 .0843
1960-63 41 .0297 39 .0126
1964 47 .4321 53 .4321
1965 69 .0205 69 .0307
1966 65 .0607 71 .0147
1967 65 .0607 65 ’0607.
1964-67 61 .0052 64 .0010
1952-67 53 52

 

*These probabilities are based on different n values due to
the dr0pping of all observations where the difference was

zero.

 

98

Alternative hypothesis: The sum of the positive ranks
# the sum of the negative ranks.

The Wilcoxon matched-pairs signed—ranks test was chosen
because this hypothesis employs two related samples and in-
volves two scores which can be ranked in order of absolute mag-
nitude, i.e. the ratios possess measurement characteristics in
the strength of an ordered metric scale. The findings of this
test, along with certain descriptive data, are presented in

Table XVII.

4.4.2c observations concerning Tables XVI and XVII

(l) The choice of prediction model apparently
has no consistent effect as to which income
practice yields the lesser forecast.

(2) Based upon the data of Tables XVI and XVII,
it is the writer's opinion that in general
there was no significant consistent differ-
ence with respect to forecast amounts under
each of the two practices.

 

 

 

WI?

\\ I,

Model:

Period:

‘

1952
1953
1954
1955
1952-55

1956
1957
1958
1959
1956-5§

 

1960
1961
1962
1963
1960.6

1964
1965
1966
1967
1964—6

“\“I

*These
the d
Zero.

 

 

 

99

TABLE XVII

 

WILCOXON MATCHED-PAIRS SIGNED-RANKS TEST ON FORECASTS
BASED UPON INCOME MEASURES UNDER AS REPORTED (REP)

AND MODIFIED ALL-INCLUSIVE (MOD) PRACTICES

 

 

 

Model: Unweighted Weighted
s «
>1 >I
u u
c -H c -H
-r-| H -.-I H
m '3 m :3
c w m c m m
H p A H u n
u m 0 u m 0
H m H a H m H a
50 04 :10 04
m o c m o c
m H o o w H o o
H o o x u o w x
m H o m H o
u -H o u H U
m H m H m H 0'4
OCD 4J-v-l 00 4.)...I
.. 3 23 .3
Period: 8 g g m o o 3IH
0 PI as o o H a o
1952 REP .6067 REP .9070
1953 REP .9054 REP .6452
1954 REP .8088 MOD .8670
1955 MOD .7444 MOD .9124
1952-55 REP .6826 REP .7200
1956 MOD .9352 MOD .9624
1957 REP .1064 REP .0810
1958 REP .1928 REP .5716
1959 MOD .6116 MOD .6268
1956—59 REP .1976 REP .2778
1960 REP .8224 REP .9754
1961 MOD .9256 MOD .6536
1962 MOD .8554 MOD .7732
1963 MOD .1970 MOD .2584
1960-63 MOD .5176 MOD .4256
1964 REP .8576 REP .3514
1965 REP .0360 REP .0390
1966 REP .0956 REP .0812
1967 REP .1558 REP .3090
1964-67 REP .0112 REP .0064

 

*These probabilities are based on different n values due to
the dropping of all observations where the difference was

zero a

 

CHAPTER V

SUMMARY AND CONCLUSIONS

5.1 Overview
The primary purpose of this chapter is to present the
conclusions reached as a result of this study and to briefly

examine their implications for accounting practice. The

 

chapter begins with a brief summary of the purpose of the
research and research design. The next section reviews the
various hypotheses and findings. This section also includes
the general conclusions and implications for accounting prac-
tice. The final section suggests some directions for future

research.

5.2 Brief summary of purpose and research design

The problem addressed in this study concerns the
method of reporting extraordinary items and prior period
adjustments in annual financial reports of business corpo—
rations. The criterion of predictive assistance was selected
as a means of providing a possible solution to the diversity
of views surrounding this area of financial reporting.

Finally, historical empirical evidence was examined with

100

 

 

 

 

‘1
I

l

respect L

several
on both It.

pragmat

of seve:
sulting
each re
ent prq
pared ‘
differ
expres
StituI
deter]
WCmld

5.3

Fred
Emm
it V
“Do:

frm

 

101

respect to the predictive assistance criterion in comparing
several income reporting alternatives. Comparisons were made
on both a conceptual level, i.e. between concepts, and on a
pragmatic level, i.e. between practices.

The basic technique of this study was to examine each
of several income reporting alternatives with respect to re-
sulting forecasts where earnings per share figures based on
each reporting alternative were used as inputs to two differ—
ent prediction models. The resulting forecasts were then com-
pared to each of two different actual values with the dollar
difference between the forecast value and actual value being
expressed as a percentage of the actual. These ratios con—
stitute the forecast error measures that were compared to
determine which of the associated reporting alternatives
would have been (for the test period) of greater aid in

forecasting earnings potential.

5.3 FindingsL conclusions, and implications

As was pointed out in Chapter II, the findings of any
predictive study depend in part upon the particular model(s)
employed. However, from the various findings of this study.
it would appear that in general there was a greater effect
upon both forecast magnitude and associated error resulting

from the choice of historical time periods than from the

 

 

 

 

choice of ei

income repor
Hypo
three income
difference a
casts. The
ating perforz
either of W

testing of H

 

 

102

choice of either prediction model, forecast objective, or
income reporting alternative.

Hypothesis C.1 was concerned with whether, among the
three income concepts, there was any consistent significant
difference as to which income measure resulted in better fore—
casts. The alternative hypothesis was that the current Oper-
ating performance concept would generate better forecasts than
either of the other two concepts. The results of the formal
testing of Hypothesis C.1 are presented in Tables VIII, IX,
and X.

After examining the findings reported in the above
named tables, it is the writer's conclusion that for the test
period as a whole there was a slight tendency for the current
Operating performance concept to generate better forecasts
than either of the other two concepts. However, in view of
divergency of results between various sub-periods within the
total period covered by the study, it is the writerLgopinion

that this observed overall tendency favoring the current

Operating performance concept was so slight as to render it

totally insignificant with respect to the establishment of

 

theoretical ideals which might serve as gpidelines for future
accounting practice. Thus for Hypothesis C.1, Ho was not

rejected.

 

Hikgg;

 

 

 

 

 

 

 

 

.

 

103

Even though comparisons between concepts showed no
significant differences, in order to evaluate the modified
all-inclusive alternative (per APB Opinion No. 9) on a
practical basis it was still necessary to compare the two
practices for two reasons: (a) possible biases inherent to
this study's method of Operationalizing the various concepts
and (b) the possibility that past (pre APB Opinion No. 9)
practice was better (i.e. resulted in greater predictability)
than any consistently applied recognized theoretical concept.
Hypotheses P.3 and P.4 were concerned with whether, between
the two practices, there was any consistent significant dif-
ference as to which income measure resulted in better fore-
casts. The alternative hypotheses were that the "as reported"
practice would generate better forecasts. The results of the
formal testing are presented in Tables XIV and XV.

Examination of the findings shown in Tables XIV and
XV led the writer to the conclusion that in general the "as
reported" practice resulted in "better" (in 15 out of 16
cases per Table XIV and in 14 out of l6_cases per Table XV),
but not significantly better (significant at .01 for only 2
of the 15 cases per Table XIV and for only 4 of the 12 cases
per Table XV), forecasts than did the modified all-inclusive
practice. Thus for Hypotheses P.4 and P.5, H6 was again not

rejected.

 

   

 

 

 

 

104

As was pointed out in Chapter II, arguments for both
the current Operating performance concept (as a theoretical
ideal) and the "as reported" practice (as a realized resultant
Of this ideal) are based solely on the generation of data more
useful in making future predictions. Given the research design
Of this study, historical empirical evidence failed to sub-
stantiate this argument. Thus the writer considers this study
to be empirical support for the position advanced in Accounting
Principles Boardggpinion NO. 9 with respect to the single
amount to be designated as net income for the period.

In addition to comparing the various concepts and
practices with respect to goodness Of forecasts (i.e. magni-
tude of absolute forecast error), they were also examined
from the secondary viewpoint Of conservatism, i.e. as to which
alternative, if any, tended to consistently result in lesser
forecasts. Hypothesis C.2 was concerned with whether, among
the three income concepts, there was any significant differ-
ence as to which measure resulted in lesser forecasts. The
results of the formal testing of Hypothesis C.2 are presented
in Tables XI, XII, and XIII. Hypotheses P.5 and P.6 involved
comparisons between the two practices. The results of test-
ing Hypotheses P.5 and P.6 are presented in Tables XVI and XVII.

Examination of the findings led the writer to the

following conclusions. In generalL the current Operating

   

 

105

erformance conce t resulted in si nificantl lesser forecasts

 

than did either of the other two concepts. However, there was
no significant consistent difference with respect to forecast

amounts under each of the two practices. This suggests the

 

notion that in practice management varies its interpretatiOn
Of accounts and/or method of accounting in such a way as to
maximize the firm's apparent earnings potential. Accordingly
a comparison was made between the current Operating perform-
ance concept and the "as reported" practice. The findings
(pee Table XVIII)_are consistent with this notion since under
the Wilcoxon test the current Operating performance income
measure resulted in lower forecasts for all cases with average
probabilities of .0923 and .1565 relative to the unweighted
and weighted prediction models respectively. A similar com-
parison was also made with reSpect to goodness of forecasts.
HOweverithe findings showed no significant differences and

are not reported herein.

5.4 Suggestions for future research

This study concerned the designation of a single
figure as net income for the period. In this respect its
findings support the modified all-inclusive approach as set
forth in APB Opinion NO. 9. However, as pointed out in

Chapter I this Opinion also requires the amount shown as net

 

 

1‘1

 

 

 

 

 

COMPARISONS OF FORECASTS ASSOCIATED WITH INCOME
MEASURES UNDER THE CURRENT OPERATING PERFORMANCE
CONCEPT AND THE AS REPORTED PRACTICE

106

TABLE XVIII

 

 

 

 

 

E.

U‘Hm U'I u

s o O c -H

w-l‘H-I-l -H H

u c I u m :H

3‘68. 8.2.1: 3".” f:

m m E U»U G m m U n

o O 0 W1“ u o O-u o
H O O IH.C O m H o H a

H O 3.u 3 H c a.
o O‘H PIC o o o c
>mo OHMH >mx no
-H UIO -H O o x
p H >. aI¢I>Im .p H U FIO
see 2.3: 33: :2
H O H o o H I4c>3 u-H
m-H o 0-H o o ouH 3

II *5 H c-ra> u H o
.H G m o m m-H FIG o G‘H
d-H E m QIEIJ d-H m 0 0

Model: Unweighted
1952-55 CUR 60 CUR .0001
1956-59 CUR 51 CUR .3299
1960-63 CUR 60 CUR .0001
1964-67 CUR 55 CUR .0429
1952-67 CUR 57
Model: Weighted
1952-55 CUR 55 CUR .0006
1956-59 REP 50 CUR .4679
1960-63 CUR 57 CUR .0012
1964-67 REP 51 CUR .1564
1952-67 CUR 53

 

*These probabilities are based on different n values due to
the drOpping of all observations where the difference was
zero.

 

 

 

 

   

 

107

income before extraordinary items to include certain amounts
that heretofore were considered to be extraordinary. It is
the writer's suggestion that the Board's new standards for
ascertaining net income before extraordinary items be sub-
jected to the test of empirical verification.

The subject of this study has been the source of many
controversies among and between those people engaged in pre—
paring financial statements and those using said statements.
Even though the very existence of this controversy was accepted
as justification for this study, the writer recognizes that
whether or not there is a "real" problem at all rests upon
the truth of the assumption of functional fixation, i.e. the
notion that many users do focus upon the single amount desig-
nated as net income for the period and do not adequately use
the financial statements as/a whole. It is the writer's sug-
gestion that much additional research be done in this be-
havioral aspect of accounting. If alternative methods of
presentation are found not to affect certain decisions, then
the entire discussion over which alternative is desirable is
irrelevant with respect to those decisions. Such behavioral
studies could take the form of questionnaire surveys, labora-
tory simulations, and real world Observations. The latter
form of study would be involved in addressing the question

raised in Chapter II: to what extent, if any, is a firm's

 

u].

 

 

 

 

108

market price affected by alternative methods of income
reporting? “The greatest accounting need both at the present

and in the future is the determination Of the nature of infor-

mation needs Of users of accounting communication.“

American Accounting Association, Committee to Pre-

gfre a Statement of Basic Accounting Theory, A Statement of

-——§Esic Accounting Theory, p. 69.

A

 

 

 

 

 

   

 

m1 4,,“ -m. _

 

Iu—l
O

N
o

H'H
F'OIO(D~JO\UIAID
O O O O

F‘H
can:

14.
15.

16.
17.
18.
£19.
.20.
21.
22.
23.
24.

25.
26“

2F7.
213.
£259.
.3().
.3LL.
13:2-
£353.

 

 

APPENDIX A

Companies Included In The Study

Allied Chemical Corporation
& Dye Corp.)

(formerly Allied Chemical

American Consumer Industries, Inc. (formerly American

Ice Co.)
The American News Company

American Seating Company

American Smelting & Refining Company

Bayuk Cigars Incorporated
Beech Aircraft Corporation

The Bendix Corporation (formerly Bendix Aviation Corp.)

Bethlehem Steel Corporation
Briggs & Stratton Corporation

Brunswick Corporation (formerly Brunswick-Balke-Collender

CO.)
Caterpillar Tractor Co.

Celanese Corporation

Central Aguirre Sugar Company
Chris—Craft Industries, Inc. (formerly NAFI Corp.,

National Automotive Fibres, Inc.)

Cincinnati Milling Machine
City Stores Company

CO.

Columbia Pictures Corporation

Congoleum—Nairn Inc.
Continental Baking Company
Continental Steel Corporat
Corning Glass Works
Culter-Hammer, Inc.

DeSoto, Inc. (formerly DeSoto Chemical & Coating, Inc.,

United Wall Paper. Inc.)
Detroit Steel Corporation

ion

Diamond International Corporation (formerly Diamond

National Corp., Diamond Gardner Corp., Diamond Match Co.)

The Dow Chemical Company
The Duplan Corporation

13- I. du Pont de Nemours & Co.

ESB Incorporated (formerly Electric Storage Battery Co.)

Eastern Stainless Steel Corporation

Eastman Kodak Company
Evans Products company

109

 

   

 

 

 

 

34.
35.
36.
37.
38.
39.
40.

41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.

57.
58.
£59.
(50.
(51.
62.
(53.

64.
65.
66.
67.
68.
69.
7o.
71.
72.
73-
74-
75-
76-

 

 

110

Federated Department Stores

General American Transportation Corporation

General Cigar Company, Inc.
General Electric Company

General Foods Corporation

Getty Oil Company (formerly Pacific West Oil Corp.)
Gould-National Batteries, Inc. (formerly National

Battery CO.)
The Great Western Sugar CO.

Gulf Oil Corp.
Hotel Corporation of America (formerly Childs Company)

Hudson Bay Mining and Smelting Co., Ltd.

Interco Incorporated (formerly International Shoe Co.)
JOy Manufacturing Company

Keebler Company (formerly United Biscuit Co. of America)
Kimberly Clark Corporation

Magnavox Company

The Manhattan Shirt Co.

The Maytag Company

McGraw-Hill, Inc.

McQuay-Norris Manufacturing Company

Mercantile Stores Company, Inc.

Metro-Goldwyn—Mayer, Inc. (formerly Loew's Inc.)
Midland-Ross Corporation (formerly Midland Steel Products

CO.)

Minnesota Mining & Manufacturing Company

Mission Corporation

Munsingwear, Inc.

G. C. Murphy Company

.NVF Company (formerly National Vulcanized Fibre Co.)

JNational Can Corporation
:National Service Industries, Inc. (formerly National
:Linen Service Corp.)

National Steel Corporation

Royal Crown Cola CO. (formerly Nehi Corp.)
St. Joseph Lead Co.

Sheraton Corporation of America

Simmons Company

Sinclair Oil Corporation

The L. S. Starrett Company

Sun Chemical Corp.

Sm} Oil Company

United States Tobacco CO.

F- W- Woolworth CO.

weatinghouse Electric Corp.

The Youngstown Steel Door CO.

   

 

 

ENDIX B

     

 

 

 

 

111

APPENDIX B

Form used to collect data on each extraordinary income determinant, prior
period adjustment—new, or prior period adjustment—old.

Company Number

Year ending

Number of items

*********‘k‘k**************************

Charge
PRESENTATION IN REPORT

Credit
BEFORE net income:

Amount before tax I I . Among ordinary items

Amount after tax , , . Separate section

$ effect on EPS Notes, letter, etc.

Adjustment Factor AFTER net income:
$ effect on Adj EPS On income statement
96 effect on Adj EPS Combined statement

Tax effect: Reported , Estimated R E statement

Caption item reported under
Item title

Item description

 

 

 

 

7 1 .- ‘x‘lrw-

 

 

 

 

APPENDIX C

 

   

 

 

 

Form used to collect data and make

Company

 

Year ending

(A)

(B)
(c)
(D)
(E)
(F)
(G)
(H)
(I)
(J)
(K)
(L)

(M)

(N)

(O)
(P)

(Q)

112

APPENDIX C

 

Net income before federal
income tax (B+D). . . . . . . .

Federal income tax expense. . .
Average tax rate (BéA). . . . .
Net income as reported. . . . .
Total preferred dividends . . .
Earnings to common stockholders
EPS as reported . . . . . . . .
Average number of common shares
Common shares 0/5 at year end .
Adjustment Factor . . . . . . .

Adjusted reported BPS (GxJ) . .

Number

computations by company, by

 

Number

of items

 

 

 

 

 

Items included in (K) above: Credit (Charge)

No $ No $

 

 

No $ NO $

 

 

Current Operating income (K—L).

No

 

No

 

All items except PPAne - Credit (Charge)

w.

NO 3 No $

 

 

NO $ No $

 

 

No

 

No

 

Modified all-inclusive income (M+N)

PPAnew: No S No

 

All-Inclusive income (0+P). . .

 

$

$

S

 

 

 

 

 

 

 

   

 

 

 

APPENDIX D

 

 

 

 

 

...... Tb. ’

 

 

113

APPENDIX D

Form used to summarize, by company, certain information contained on
form shown as Appendix C.

Company Number

 

 

EPS under the following income concept:

As Current Modified All-
Year Reported Operating All-inclusive Inclusive

 

1947

1948

1949

1950

1951

1952

1953

1954

1955

1956

1957

1958

1959

1960

1961

1962

1963

1964

1965

1966

1967

 

SELECTED BIBLIOGRAPHY

Reference WOrks

Accounting Principles Board. Opinion No. 9, Reporting the
-Results of Operations. New YOrk: American Institute
of Certified Public Accountants, December 1966.

 

American Accounting Association, Committee to Prepare a
Statement Of Basic Accounting Theory. A Statement Of
Basic Accounting Theory. Evanston, Illinois: American
Accounting Association, 1966.

American Accounting Association, Executive Committee.
Accounting and Reporting Standards for Corporate
Financial Statements and Sppplements. Madison,
Wisconsin: American Accounting Association, 1957.

American Institute of Certified Public Accountants. Accounting
Trends & Techniques. New York: American Institute of
Certified Public Accountants, 1957 through 1967.

American Institute of Certified Public Accountants, Committee
on Accounting Procedure. Accounting Research and
Terminology Bulletins, Final Edition. New York:
American Institute of Certified Public Accountants,
1961.

Arthur Andersen & CO. Accountingpand Reporting Problems of
the Accounting Profession. Chicago: Arthur Andersen
& Co., 1962.

Bernstein, LeOpOld A. Accounting_for Extraordinary Gains and
Losses. New York: The Ronald Press Company, 1967.

. Extraordinary Items of Gain or ppss:.rA Research
Study of Theory and Practice. Unpublished Doctoral
Dissertation, New YOrk University, 1966.

 

Bruton, Paul W. and Bradley, Raymond J. Federal Taxation.
St. Paul; Minn.:~ west Publishing Co., 1955.

114

 

 

   

115

Chambers, Raymond J.. Accountirg, Evaluation and Economic
Behavior. Englewood Cliffs, New Jersey: Prentice-
Hall, Inc., 1966.

Commerce Clearing House. 1957 U. S. Master Tax Guide.
Chicago: Commerce Clearing House, 1957.

. 1967 U. S. Master Tax Guide. Chicago:
Commerce Clearing House, 1967.

 

Friedman, Milton. "The Methodology Of Positive Economics,"
Essays in Positive Economics. Chicago: University
of Chicago Press, 1953.

Hayes, Douglas A. Appraisal and Management of Securities.
New York: The Macmillan Company, 1956.

Meigs, Walter B., Johnson, Charles E., and Keller, Thomas F.
Intermediate Accounting. New York: McGraw-Hill BOOk
Company, 1963.

Montgomery, Robert H. Montgomery's Federal Taxes --
Corporations and Partnerships -- 1947-48 Vol. I --
Gross Income and Deductions. New YOrk: The Ronald
Press Company, 1948.

Moody's Bank and Finance Manual. New York: Moody's Investors
Service, Inc., l946.through 1968.

Moody's Industrial Manual. New YOrk: Moody's Investors
Service, Inc., 1946 through 1968.

Moody's Transportation Manual. New York: Moody's Investors
Service, Inc., 1946 through 1968.

Moody's Public Utilipy Manual. New YOrk: Moody's Investors
Service, Inc., 1946 through 1968.

The New York Times. "1946 Stock Market Dealings," The
New York Times, sec. L, p. 36, January2, 1947.

Poor's Register of Corporations, Directors and Executives -
1968. New York: Standard & Poor's Corporation, 1968.

Securities and Exchange Commission. Regulation S-X,
Washington, D.C.: Government Printing Office, 1958.

   

 

 

 

 

 

116

85-9981. Sidney. Nonparametric Statistics For The Behavioral
Sciences. New York: McGraw-Hill Book Company, 1956.

 

Spiegel, Murray R. Schaum's Outline Of Theory and Problems
of Statistics. New YOrk: Schaum Publishing Co.,
1961.

wright, Albert Walter, Jr. Accounting for Extraordinary
Charges and Credits and Their Implications For Net
Income. unpublished Doctoral Dissertation, University
of California, Los Angeles, 1965.

 

Various corporate annual reports.

 

Periodical Literature £3

Arnett, Harold E. "Application of the Capital Gains and
Losses Concept in Practice," Accounting Review, XL
(January 1965).

Beaver, William H., Kennelly, John W., and Voss, William M.
"Predictive Ability as a Criterion for the Evaluation
of Accounting Data," Accounting Review, XLIII
(October 1968).

Corbin, Donald A. "Reported Unexpected Items: A Dissent to
APB Opinion NO. 9," Unpublished paper presented to
the accounting faculty and graduate students,
University of Washington, November 8, 1968.

Frank, Warner. "A Study Of The Predictive Significance Of
Two Income Measures," unpublished paper written at
the University of Wisconsin, 1968.

Leland, Thomas W. "Revenue, Expense, and Income,“ Accounting
Review, XXIII (January 1948).

Littleton, A. C. "The Integration of Income and Surplus
Statements," The JOurnal Of Accountancy, LXIX
(January 1940).

Rowe, W. G. "Surplus Adjustments," Accounting Review, VIII
(October 1933).

 

   

   

   

117

Savage, Allan H. and Snavely, Howard J. "The Accounting
Principles Board and Opinion No. 9," Unpublished paper

prepared at the University of Texas at Arlington,
1967.

Stans, Maurice H. "Modernizing the Income Statement,"
Accounting Review, XXIV (March 1949).

fr '.
E.,

 

   

 

   

 

   

 

«tis-