Citation

Material Information

Title:
A critique of income reporting for shareholders
Added title page title:
Income reporting for shareholders
Creator:
Whitehurst, Frederick Dozier, 1938-
Publication Date:
Language:
English
Physical Description:
iv, 161 leaves : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Accountancy ( jstor )
Business structures ( jstor )
Cash ( jstor )
Corporations ( jstor )
Correlation coefficients ( jstor )
Earnings per share ( jstor )
Financial accounting ( jstor )
Financial investments ( jstor )
Investment decisions ( jstor )
Net income ( jstor )
Accounting thesis Ph. D ( lcsh )
Dissertations, Academic -- Accounting -- UF ( lcsh )
Income accounting ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis - University of Florida.
Bibliography:
Bibliography: leaves 154-161.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Resource Identifier:
021718866 ( ALEPH )
13299806 ( OCLC )

Full Text

A CRITIQUE OF INCOME REPORTING
FOR SHAREHOLDERS

FREDERICK DOZIEy WHITEHURST
FREDERICK DOZIER WHITEHURST

A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA
1968

Frederick Dozier Whitehurst

1968

To Delmas D. Ray

Chapter Page

I. INTRODUCTION 1

II. THE ACCOUNTANT' S APPROXIMATION OF TRUE INCE 14 III. STATEMENT OF STATISTICAL PROCEDURE 43

IV. PREDICTION OF EARNINGS 84 V. PREDICTION OF CASH RETURN 101 VI. CONCLUSIONS 118 Appendix

I. STOCK IDENTIFICATION 128 II. ADJUSTED EARNINGS PER SHARE INPUT DATA 134

III. MARKET VALUATIONS INPUT DATA 141

IV. OUTPUT AND RANK CORRELATIONS INPUT 146

BIBLIOGRAPHY 154

LIST OF TABLES

Table Page

I. Projected Earnings Per Share Compared With
Actual Earnings Per Share 87

II. Projected PE Ratio Compared With Actual PE Ratio 89

III. Actual Rank Compared With Predicted Rank Earnings Per Share 90

IV. Actual Rank Compared With Predicted Rank
Price-Earnings Ratio 93

V. Mean and Standard Deviation Actual Earnings Per
Share as a Percent of Projected Earnings Per
Share 96

VI. Rank Correlation Coefficients Actual Compared With
Projected Earnings Per Share Stocks Grouped by
Industry 98

VII. Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation
Earnings Per Share 99

VIII. Projected Price-Earnings Ratio Compared With
Actual Cash Return 103

IX. Actual Rank of Cash Return Compared With
Predicted Rank of Price-Earnings Ratio 106

X. Projected Price-Earnings Ratio Compared With
Actual Cash Return Utilizing Time Flexible
Investment Decisions 109

XI. Actual Rank of Cash Return Compared With Predicted Rank of Price-Earnings Ratio Utilizing Time
Flexible Investment Decisions 110

XII. Rank Correlation Coefficients Predicted
PE Ratio Compared With Actual Cash Return
Stocks Grouped by Industry 112

Table

Page

XIII. Interindustry Rank Correlations Compared With
Intraindustry Average Rank Correlation
Predicted Price-Earnings Ratio Rank Correlated
With Cash Return 113

XIV. Rank Correlation Coefficients Predicted PE Ratio
Compared With Actual Cash Return Utilizing Time
Flexible Decisions Stocks Grouped by Industry 116

XV. Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation Predicted
Price-Earnings Ratio Rank Correlated With Cash
Return Utilizing Time Flexible Decisions 116

C-PTER I

INTRODUCTION

The contemporary certified public accountant is involved with many functions. Typically, he is concerned with taxation matters, systems analysis, various management services, and auditing. Paramount a.ong these diversified functions is that of lending disinterested or independent attestation to the conformity of management's financial representations to that which has come to be termed "generally accepted principles of accounting." More precisely still, the certified public accountant is said to be vitally interested in the "fairness" of the representations of management to third parties including stockholders and creditors. He is viewed as the guardian of integrity in those financial representations which are inevitably associated with a conflict of interest. From the first public accounting statute to the present day, his function in society has been associated with terms such as "equity," "justice," and "fairness." Integrity in financial reporting to third parties has come to be interpreted as the correspondence of procedures used with accepted principles of accounting which, in turn, imply a "fair" presentation. Financial presentations are, therefore, considered fair only so long as they are consistent with procedures which are in conformity with generally accepted treatment rules. These treatment rules have as their major core of emphasis the income determination process. Hence, the principal raison d'etre of the certified public

accountant is that of lenrding his attestation to the "fairness" of reported "income."

Perhaps the most astonishing feature of modern accounting is the fact that a whole profession has devoted itself, for the most part, to the task of determining income for an enterprise for a period of time and yet is at a total loss to arrive at a consensus as to the meaning or significance of that aggregate. Few would disagree that an

important function of the modern C. P. A. is that of "income determination." Fewer still would be able to relate a meaning for that notion on other than a procedural basis. Perhaps this is a failure of research effort. Perhaps it is due to the lack of applicability of scientific method to such "assertions." More fundamentally, perhaps

the asserted phenomena of "income" in a real or true sense simply does not exist. If this is the case, there is little wonder that there is such confusion in economic and accounting realms as to the significance, definition, computation, meaning, verification, and utility of computed monetary income.

A. recent outcrop of a relatively large research effort in terms of both expenditure and effort by the Amnerican Institute of Certified Public

Accountants suggests as an appropriate definition of net income the
1
"...increase (decrease) in owner's equity...." This is substantially a duplication of the definition offered by Canning which reduced the significance of net income to that which uas left over after: the

1
Robert T. Sprouse and Maurice oonitz, A Tentative Set of Broad
Accounting Principles for aBusiness Enterprise (American Institute of Certified Public Accountants, 1962), p. 9. This meaning of income is a recurrent theme throughout most of te official pronouncements of both major professional associations in accounting.

2
accountant completed his procedures. If one understands "change in owner's equity" to mean the change brought about through the procedures of the accountant in closing the nominal accounts, then cet. par., the two statements do indeed agree. Unfortunately, such definitions not only fail to indicate the meaning or significance of this computational aggregate, but worse still, they appear to be entirely appropriate. To our knowledge, the only meanings attributed to income in accounting literature are statements of definition on procedural bases. Why then, has income enjoyed a historical prominence in financial affairs? Why has enterprise net income been the building block of substantially an entire profession?

There is evidence that profit calculations started in the Middle Ages as a tool for the computation of appropriate cash distributions which would be made before the termination of parternership voyages under3
taken as single ventures. In modern times, income calculations have been propounded in a variety of ways, not the least of which was the federal income tax. If tax is based on "income," there would naturally exist the tendency to explicitly formulate procedures for the determination of exactly what would constitute one's income. In an effort to minimize individual federal tax payments, the treasury department has been forced to formulate procedures and guidelines for the determination of procedural meanings for "income." Transfer this emphasis

2John B. Canning, The Economics of Accountancy (The Ronald Press Company, 1929), p. 98.
3Raymond de Roover, "The Development of Accounting Prior to Luca
Pacioli According to the Account-books of Medieval Merchants," in Studies in the History of Accounting, edited by A. C. Littleton and B. S. Yamey Iichard D. Irwin, Inc., 1956), pp. 115-ll6.

to the corporate taxation area and the importance of not only the calculations of net income but also the procedural meanings of net income become greatly expanded. Perhaps the greatest single factor in the rise of income determination was the interpretation which the courts made with respect to corporate fiduciary relationships.

Corporate officers have been included in the category of the

fiduciary. They have the responsibility of the fiduciary with respect to those on whose behalf they hold a position of trust. Courts have

interpreted this relationship as having the responsibility for publishing annual statements of corporate affairs. This includes the balance sheet and income statement. Some state corporation statutes require financial reporting by the corporate fiduciary.

The national stock exchanges have had some influence in developing

financial procedures which have as a result the statement of "net income." They have the authority to prescribe' reporting standards as a prerequisite for listing. Traditionally, such reporting standards include statements of financial position and income statements.

This listing requirement is reinforced by the S.E.C. which, through its regulations, requires the annual submission of balance sheets and income statements which have been audited by independent accountants reflecting "condition" and "earnings." The standard of acceptability of procedures used in such reports has been outlined by the S.E.C. under the term of "substantial authoritative support." This rule was first
4
established in 1934, at the very start of securities legislation.

4United States Securities and Exchange Commission, Accounting
Series Releases (United States Government Printing Office, 1956), p. 5. This particular release (Release Number Four of April 25, 1938) presumes financial statements to be misleading unless there is "substantial authoritative support" for the procedures employed.

Reports filed with the Securities and Exchange Commission become public record and they are available for both inspection as well as duplication.

A heavy reliance on income determination is evident with respect to the rate regulatory powers of many federal and state agencies, particularly those whose powers extend to the prescription of accounting procedures whereby the "rate of return" is policed.

The Accounting profession itself has brought noticable attention to

the concept of income in recent years. The American Accounting Association has published many works devoted to the analysis of income and related valuation topics. The American Institute of Certified Public Accountants has been active in the promulgation of "sound" accounting treatment rules ever since the formation of the Committee on Accounting Procedure and the issuance of its opinions which started in 1938. Currently, the Accounting Principles Board of the American Institute of Certified Public Accountants carries on the work started by the Accounting Procedure Committee. There is little doubt that the prescriptions coming forth from such organizations have as their primary aim the more accurate determination of financial position and operating results.

Evidence of the primary position which "income determination" holds in accounting is found in the literature of accounting. Paton is perhaps the foremost advocate of the central position of income. He states that income determination is "...the most significant and crutial phase
5
of the accountant's task." Professor Reed K. Storey has stated that

William A. Paton, Advanced Accounting (Macmillan and Company, 1947), p. 438. See also William A. Paton, Accounting Theory (Accounting Studies
Press, Ltd., 1962), p. 142. Most of the writings of this "father" of modern accounting have centered around the development of a coordinated and consistent "theory" of accounting with income determination and asset valuation its central functions.

the determination of periodic net income "...is the most important
6
function of financial accounting."

Notwithstanding such a heavy reliance on the conceptual : ticn of income, there is little evidence that the net income aggregate, to the fairness of which independent auditors lend their attestations, has an understandable meaning which would facilitate (or even permit) a determination oft relative fairness. The independent accountant attests to the fairness of such income calculations without any reference to a standard to which a given net income figure might be compared for an objective determination of its degree of approximation,
7
nearness, or fairness. The fairness of management's representations

wuich have been determined by procedures which are consistent "ith generally accepted accounting principles appears to be a conclusion

rather than a criterion. Financial position and operating results are "fair" only so long as they are determined by use of rules of procedure which have as the sole criterion of warrant their general acceptance. It would constitute more efficient conmunication if references to the "fairness" of position and operating results were omitted. The certifying auditor himself generally sidesteps this fairness issue by indicating in the same sentence that the balance sheet and the income statement present fairly the financial position and the results of

operations in conformity with generally accepted principles of accounting

6Reed K. Storey, "Cash Movements and Periodic Income Determination," he . ounting Review, July, 1960, p. 449. See also A. C. Littleton, "Concepts of Income Underlying Accounting," The Accounting Revie-, March, 1937, p. 22.
7
R. J. Chambers, "A Matter of Principle," The Account*n Review, July, 1966, p. 455. Professor Chambers insists that withcu-t such a standard of comparison allusions to Hfairness" are both unwarranted as well as meaningless.

applied on a basis consistent with that of the preceding year. If there is no agreement as to the significance of income which is the direct product of the use of accepted accounting principles, and if the result of the use of such procedures is a "fair" presentation, then the inevitable conclusion is that "fairness" has no generally agreed upon meaning either. This is particularly evident when one considers the quality of such generally accepted accounting principles. They appear so ambiguous to one author that he concludes that they
8
have no definitive meaning.

The accountant, then, attests to the balance sheet and income

statement presentations as being procedurally consistent with these accepted rules. He concludes that net income has been determined in a "fair" manner. Yet, as Professor Chambers indicates, through the wide choice of equally acceptable alternatives in procedure, any independent auditor could certify the "fairness" of any one of something like thirty-million possible and equally acceptable net income figures for a large corporation with a sufficiently large diversifi9
cation of transactions. On this basis it would appear that the conclusion of Professor Canning in 1929 was not dated. After a careful study of the methodology of the accountant he concluded that "...account10
ants have no complete philosophical system of thought about income."

George O. May, "Generally Accepted Principles of Accounting," The Journal of Accountancy, January, 1959, p. 24.
9R. J. Chambers, p. 455.
lOJohn B. Canning, p. 160. It would furthermore seem that to revert to statements of "fairness" which are implicit in the conformity of financial presentations with generally accepted accounting principles constitutes a negation of the presumed professional attitude. It would seem that "fairness" should be more explicitly defined as constituting the conformity of presentations with that which has become accepted.

The discussion above supports the observation that all is not

well with respect to the clarity of the accountants conceptual rigor. Yet, many accounting writers rely on such conceptual apparatus to support the actions of practicing accountants. Income does exist. It

simply falls to the accountant the task of arriving at "income" as nearly as possible. Many assert that the income calculation of the accountant is clearly not exact. It is, however, thought to be close enough to the true income to be valid. This type argument represents through the observation of the author one of the two most popular

approaches to asserting the propriety of determining accounting net income and reporting it to third parties in the annual report. In summary form this argument simply states that income of an enterprise for a period of time does exist and the accountant goes about measuring it as closely as possible. The conceptual framework of "accounting

theory" is said to be the guide for the operations of the practicing accountant. We must trace the meaning of the computed net income, therefore, backwards to the concepts that are said to be involved.

In addition to this approach, many accounting writers offer the outcomes of the net income calculation as support for its propriety. Net income is said to be of use. Whatever methodology is used for its computation, net income does represent a valid basis for the orientation

of decision and action. It is said to have functional utility. Among the suggested applications of accounting net income is the popular assertion that the computed accounting net income represents relevant information in the decision environment of the prospective shareholder. There are then, two basic assertions about accounting income. The

first asserts the existence of enterprise net income and suggests that

it is the accountant's task to determine the monetary expression of that net income as accurately as possible. The second asserts that

computed accounting net income (as an outcome) is useful for decision. The decision of the prospective sharcholier is thought to be the primary utility application of reported net income.

In a popular article written in 1943, Professor George O. May 11
suggested ten uses of financial statements:

1. As a report of stewardship
2. As a basis for fiscal policy
3. As a criterion for the legality of dividends
4. As a guide to wise dividend action
5. As a basis for the granting of credit
6. As information for prospective investors in an enterprise
7. As a guide to the value of investments already made
S. As an aid to government supervision
9. As a basis for price or rate regulation
10. As a basis for taxation

Item six above seems to have emerged as the most popular justification from the standpoint of outcomes for the currently reported net income figures derived from the application of "financial accounting theory." Decision relates to the future and financial net income is thought to be a good basis on which to make investment decisions. Specifically, past income trends are asserted to have relevance in estimating the future trends or levels of income. This being so, the prospective investor is thought to benefit a great deal from the

information disclosed in the past incomane statements of enterprises. Past income calculations are thought to be predictive to the extent of rendering to the investor a basis for an intelligent guess.

11
George O. May, "The Nature of the Financial Accounting Process," The Accounting Review, July, 1943, pp. 189-190. Italics added.

In this connection Professor M'orrison states:

As we have stated, the intelligent investor is interested
fundamentally in the earning power of the company.... This involves
forecasting or budgeting because earning power involves the future.
Unless the intelligent investor has quite complete and accurate
information about the ranner in which earnings were created in the
past, his forecast will be a very poor guess. On the other hand, if he has a good historical record he has a good starting point
from which to forecast the future.12

Professor Robert T. Sprouse has summarized this popular viewpoint

by saying in a recent paper:

My paper is concerned with specifically the accounting
measurements reported in the financial statements provided by
publically held corporations to current and prospective investors
financial statements about which certified public accountants
render their professional opinion concerning fair presentation in
conformity with generally accepted accounting principles. These
are the financial statements which the Securities and Exchange Commission, the New York Stock Exchange, the American Institute
of Certified Public Accountants, the Financial Analysts Federation,
and a significant portion of the membership of the American
Accounting Association and other professional groups are most
vitally concerned. It seems clear from the nature of the
intended audience and the nature of the interested professional
groups that such financial statements are intended to provide
information that is helpful in making rational investment
decisions.13

The following statement by Professor Sprouse is presented as a

representation of the consensus of accounting writers, in the opinion

of the author, with respect to the primary utility of financial net

income reported in the annual report to stockholders.

The primary purpose of the measurement of last year's income
reDorted to investors is to provide a basis for predicting future
years' income.1A

12Paul L. Morrison, "The Interest of the Investor in Accounting Principles," The Accounting Review, March, 1937, p. 39.
13Robert T. Sprouse, "The Measurement of Financial Position and Income: Purpose and Procedure," Research in Accounting Measurement, edited by R. K. Jaedicke, Y. Ijiri, and 0. Nielsen (American Accounting Association, 1966), p. 103.
14Ibid., p. 106. Italics added.

In -whole, or in part, the investor is thought to be able to utilize

income determination by looking at past trends in order to gain some insight into the probable future status of the enterprise under consideration. It is important to note that this claim is made of income figures and not simply of financial data in general. Income is an interpretative aggregate which is thought to add something to unprocessed financial data which would otherwise be lacking. 1nphasis is placed not on the reporting of financial facts but on the interpretation of those facts in the form of net income. Historical financial facts could easily be presented in the amual report without requiring that they adhere to the income determination model. Factual statements of historical transactions together with selected current values for assets (if this information is factually determined to be useful) could be presented without any references to the net income of an enterprise for a period. As it is, judgements as to which alternative accounting treatment rule franm among the many "principles" of accounting is the correct one under the circumstances is made by reference to its anticipated consequences in the income determination process. It would appear, therefore, that the primary concern is not on the presentation of data but on the interpretation of these data. Net income represents an opinion as to how well an entity performed in the past (whether this opinion is a "generally accepted"

opinion consistent with the dictates of collective authority or not is irrelevant). Hence, when the certifying auditor renders his opinion on financial statements, he is, in effect, saying that it is his opinion that the opinion expressed in the form of net income has been determined on a basis consistent with generally accepted accounting principles (an additional opinion). Essentially, the auditor is saying that it is my

12

opinion that the opinion of past performance expressed as net income is "fair" in that my basis for a clean certificate rests on generally accepted opinion with which the reported net income figure is consistent. It appears to be unpopular to hold that net income is interpretative. Many writers claim what might be termed an existential status (non conceptual status) for income. Income is. It falls to the accountant the task of determining as closely as possible this reala" income amount. Here then, are the two basic beliefs in the propriety of income reporting. One holds that income enjoys an existential status, perhaps much like gravity, or the behavior of oxygen. The other holds that income amounts in the annual report provide valuable information for shareholder investment decisions. One is thought to be able to predict the earnings of the future from the earnings trends of the past and thereby gain some measure of insight as to what to expect in the future. These two claims are, in the opinion of the author, the most frequently encountered assertions as to the value of income figures appearing over the opinion of certified public accountants. We interpret the function of a critique as that of determining the qualitative aspects of the correspondence

between the assertions made of income utility and the data base which ostensibly supports these assertions. Perhaps one or both such beliefs are justified in the light of extended logical argument and in the face of relevant data. Perhaps also one or both such beliefs require too great a leap of faith to be held in the light of evidence. We shall offer a critique of income determination based on the warranted assertability of

these two beliefs in income.

Chapter II will deal with the first of these two beliefs. That is, it will examine the efficiency of tracing the significance of a reported

net income figure backwards to the concepts which are said to direct the practicing accountants operations. Is there a meaning of income in "accounting theory" which directs operations in such a way as to insure the establishment of an income figure which has significance on other than procedural grounds. Instead of reiterating the claim, "income is,"' we ask the question "is there income" in the claimed "real," "ultimate," or "fundamental" sense.

The remainder of the study (Chapters III through V) will deal with the second isolated belief. We will be interested in developing data of a relevant nature which will indicate whether and to what extent past income trends factually do indicate things to come. Yet, the susceptibility of future income amounts to projection from the extrapolation of past trends is not the central issue in this section. We have documented the belief that past income trends establish a decision base which is thought to be valid for the prospective shareholder. The real issue in this assertion of .utility for income is the extent to which the relatively better stock investment decisions will be made if the basis for selection is the relatively better projected income amounts as determined by the extrapolation of past trends. Chapter III will set forth the methodology by which relevant data might be used to establish the assertability of this utility claim for income determination. Chapters IV and V will present the findings of this statistical study. Chapter VI will present the overall conclusions with respect to both claims.

CHAPTER II

THE ACCOUNTANT' S APPROXIKATION OF TRUE INCOME

If, as some suggest, the role of "accounting theory" is to provide a conceptual "base" which might be used to "direct" operations of the practicing accountant, then the meaning and significance of the accountant's computations must be found in such conceptualizations. In genaral, no reference is made to "outcomes" or "consequences" when it is asserted that the support for accountant's computations is to be found in "theory." Instead of tracing meaning forward to consequences, those who assert an exclusively directional function of theory suggest (most of the time unavoidably) that the significance of the accountant's statement of operating results (net income) is to be found in the concepts that are said to direct his actions. This argument was identified in Chapter I as one of the two most popular paths to belief in the propriety of the accountant's income figure. The phenomenon termed "income" is and the accountant measures it as closely as it can be measured. It is important to note that in this argument there is no burden placed on income to be a utility device the quality of which must be settled upon by a determination of the consequences of its use in association with the asserted functions it has. The emphasis here is not on the utility of the income computation but on the correspondence between "guiding# concepts and the accountant's operations in practice. Practice is justified on the basis of its conformity with the concepts and not on the basis of the utility potential of the net income aggregate.

These two paths to belief in income stem from what might be termed a

polar separation of emphasis between the two materials of inquiry,

conception and perception. Note this divergence in the statement by

Professors Paton and Littleton and the statement of Professor May.

Is net income what management chooses to say it is or is
it the result of objective conditions? ...net income of an
enterprise for a period is not made large or small by a method
of calculation; its real amount is determined by operating
activities and attendant economic conditions. An accounting reckoning should be an attempt to capture objective realities
which exist whether the reckoning is made or not.1

Professor May does not hold much promise for objective realities.

...it is essential to determine what are the fruits of the
years activities and this can be done only on the basis of estimates and on fundamental conventions which must find their warrant not in "truth" but in usefulness and practicality.2

1W. A. Paton and A. C. Littleton An Introduction to Corporate Accounting Standards (American Accounting Association, 1964), p. 86. Italics added. Professor Charles T. Horngren implies that economic
2George O. May, "The Choice Before Us," The Journal of Accountancy, March, 1950, p. 207.

Professors Paton and Littleton suggest the existence of a real, true, objectively determinable, and ongoing income while George 0. May holds no hope for a referential base of income in "truth." Professors Paton and Littleton are associated with the argument that a meaning for income might be found in concepts whereas George 0. May is associated -ith the contention that income meanings are to be found in the practical application of the income aggregate, the consequences stemming from its use, or, in a word, percepts.

The task at hand is the examination of the asserted concepts which are said to underlie the accountant's operations. The matter of "outcomes" and the asserted functional utility of financial income will be taken up in Chapter III.

With allusions to reality, truth, and objective conditions, there is naturally some part to be played by scientific method in both the search for as well as the evaluation of income. All too often logical is presumed to be"a synonym for scientific.

It would seem that in the field of accounting the word
"scientific" as applied to the determination of the amount of corporate income has the meaning of a method of determination which is to the greatest degree possible, objective, unbiased,
logical, and capable of being verified and tested by other
observers or accountants.3

Yet, without any indication that there is a standard against which computed net income might be compared for a determination of its degree of approximation, this same author goes on to say that "...although such a determination can never be exact or precise, it is an approximation

3Arthur C. Kelly, "Can Corporate Incomes be Scientifically
Ascertained," The Accounting Review, July, 1951, p. 290. Italics added. Note should be made of the naive use of the term "tested" in the statement. It is in the use of this word that the others (objective, unbiased, and logical) take on the meaning intended by the author.

4 17 to the truth." Hence, the financial net income is an approximation to the truth which is represented in the concepts directing the actions of the accountant. Truth is conceived and the accountant aims as closely as he can to state the truth, the "objective realities which exist," in a manner consistent with the directive forces of the concept. The situation is this: A concept is settled upon and operations are undertaken to put the concept to use. When the directing forces of the concept have shaped operations in such a way as to make them as consistent as possible with the concept, the process stops. The result is termed an approximation to truth, an approximation to reality, or an approximation to the objective conditions which exist. Since the operational result of the implementation of the concept is termed an approximation to truth, and since the operational result is also termed an approximation to the ideality of the concept, the ostensible reference to truth must mean the concept. The concept and truth are synonymous and to the extent that operations fall somewhat away from the ideality of the concept, they are only approximations to truth. Here lies one of the two basic beliefs in income, that it is true, that it is real, and that the only reason the accountant cannot state it exactly is due to the dilutions of the "pure theory" encountered in practice. Are these contentions sound?

The first task is to state what these concepts are. A thoughtful

forewarning to those who would take to deliberations on income the hope of arriving at a concrete irreducible meaning was issued many years ago

4Ibid. The popular notions on the relationship of absolute truth
to approximations to truth as far as income determination is concerned are well summarized by Kelly. He says, "So far as accounting is concerned, it is well recognized that'there is no such thing as the absolute truth in a presentation of corporate net income. Every statement of income is merely an approximation to the truth, or to reality..." Ibid.

by Professor Frank n. Knight. knight suggested that no other concept in economics was used vith a more bewildering variety of well-established
5
meanings than profit or income. The wisdom of this caution is evident to those who are familiar with the variety of computational counterparts to conceptualized notions of income which exists in accounting theory.

We shall have to concern ourselves initially with those conceptualized notions of income which are said to be economic (as opposed to accounting or computational) in nature. These statements of conceived interaction (truth, in the argument of the proposers) generally represent what might be termed the more popular searches for "central" or "ultimate" meaning. Among these concepts, those of Professor John R. Hicks are perhaps the most frequently appropriated for use in accounting income determination discussions. Why Professor Hicks' definitions of income seem so attractive to accountants is, of course, conjecture. Perhaps quite a substantial case might be developed in favor of the contention that what accountants actually do in practice appears to be

less painful a dilution of the strict definition than the concepts of other writers. Somehow one cannot possibly imagine an accountant coming to grips with the Fisherian "psychic benefit" crystallization of meaning of income. Because of their popularity, we shall consider Hicksian definitions rather closely at first. We shall want to determine, if possible, the qualitative aspects of this "truth" even if the accountant could follow the dictates of the concept exactly. Might income be "pinned down" by tracing the significance and meaning of the operational aggregate "backwards" to the bedrock of "truth" as it is claimed?

5Frank H. Knight, Risk, Uncertainty, and Profit (Harper and Row, 1965), p. 22. See also the author's article on "Profit," Encyclooedia of the Social Sciences, 1933, p. 480.

Economic Income - The Accountant's Guide

Economic concepts of income are often attributed the quality of being a referential base to which operational counterparts in accounting might be related, however diluted that operational expression might be. Many prominent writers in accounting suggest that these concepts of income provide an "ideal" toward which accounting income might tend to approximate. Professor Sidney S. Alexander, while stipulating the limited appeal of any one economic income concept, goes on to build a theoretical framework utilizing the Hicksian central definition as
6
a clear starting point. Professor Willard Graham noted that, "Although the practical applicability of the economist's theory of income has been seriously questioned, it has value as an 'ideal' or
7
'yardstick' by which to evaluate other theories of business income." Professor Graham went on to indicate that the economic theory of income
8
has pointed to certain shortcomings in the accountant's procedure. Such prescriptions suggest that economic conceptualizations are formulated in statement form. For the most part they are attempts to deal with the meaning of income on a constitutive basis. As a guide, these concepts are settled upon as the narrative description of that which their counterparts in accounting set out to make operational. Concepts are thereby used in the sense of guiding operations in what might be termed "blind concordance." No burden is placed on such operations for establishing the assertability of the conceptualized indication of "truth."

6Sidney S. Alexander, "Income Measurement in a Dynamic Economy," Studies in Accounting Theory, edited by W. T. Baxter and S. Davidson (Irwin, 1962), pp. 127-146.
7Willard J. Graham, "Some Observations on the Nature of Income,
Generally Accepted Accounting Principles, and Financial Reporting," Law and Contemporary Problems (School of Law, Duke University, 1965), p. 654.
SIbid.

Once the operational counterpart has performed the indicated procedure, the process stoos. If the processes of "theory formulation" in both economics as well as in accounting are properly viewed as knowledgeacquisiticn processes, then this interpretation of the function of concepts would seem to be somewhat inefficient. Note that there is no place in this notion of the idealness of economic income for operations designed to establish the warranted assertability of hypotheses. Note that in the conceptualization processes there is no place for indicated if-then relationships toward which operations might be undertaken on an existential (non-conceptual) plane. If, as Professor Graham noted, the shortcomings of operational expressions are disclosed in theoretical conceptualizations, then we would seem to be placing the cart in front of the horse. Utilizing the method of science, it is found that existential reconstructions are generally made to point to any possible shortcomings which the conceptualized structure might have had. Notwithstanding this seemingly fundamental difficulty, we shall now turn to a discussion of the Hicksian concepts in an effort to understand what is claimed in the way of a relationship between operations and the concepts which are said to direct them.

Hicksian Definitions

Many discussions of income implicitly assume that there is but one concept of income. This is simply not the case. In both economics as well as in accounting there is a plurality of notions of income no one

of which receives a clear consensus of meaning or definition. Historically, accountants have settled on the Hicksian definitions. His time-honored irreducible statement of the meaning of income crystallized from his own observations might be associated with Littleton's definition of "principle."

Littleton states that a principle "...is a crystallization of ideas into
9
a clear verbal statement of a significant relationship." If this is what Hicks and others have done, we might be able to gather some indications as to the effectiveness and reliability of this principle formation approach as we trace the meaning of the Hicksian definitions. If we find merit to the Hicksian mode of analysis, perhaps knowledge acquisition processes (presuming that Littleton excluded from his meaning of principle such matters as "principles" of good manners) do envision the crystallization of ideas.

We shall need to examine closely not only the central concept of

Professor Hicks, but also the three approximations to this core of meaning which he offers. Each such approximation is said to be a more accurate definition of the central concept than the former. By taking each definition separately, we should gather a better appreciation of the utterly subjective nature of the central concept. Moreover, a more accurate characterization of the accounting adaptations of this definition might be established.

The definitions of Hicks will be related verbatim. We shall wish to bring into isolation one recurring characteristic of Hicksian def10
initions, their total lack of extrapersonal significance. It will be of benefit to note this recurring characteristic as each approximation is considered.

Without attempting to deal with the meaning of the term "as well

9A. C. Littleton, Structure of Accounting Theory (American Account-ing Association, 1953), p. 23.
10Extrapersonal is used here to represent that quality of propositions whose related existential statuses presumably have no dependence on the perceiver.

off," Hicks defines a man's income as that which he might "...consume during the week and still expect to be as well off at the end of the week as he was at the beginning." This is the asserted "central concept." The first of the three prior approximations follows:

Income No. 1 is thus the maximum amount which can be spent
during a period if there is to be an expectation of maintaining
intact the capital value of prospective receipts in money terms.11

Income calculations which are consistent with the aforementioned definition are necessarily associated with future expectations on a personal basis. This first approximation I...makes everything depend 12
on the capitalized money value of the individual's prospective receipts." This definition is said to have an identity with those calculations of individuals who are anticipating a constant stream of receipts of a uniform amount. For, "...any one who expects a constant stream of receipts (and does not expect any change in interest rates) will reckon that
13
constant amount as his income, on that definition." This constant stream of receipts is perfectly consistent with the formal definition in terms of maintaining intact a uniform capitalized value of prospective receipts in money terms. An annuity, for example, of a constant amount indefinitely into the future must necessarily have a capitalized value which remains intact period-to-period. A principal amount, if left intact, will yield interest indefinitely into the future. But in order to maintain capital value, one must come to grips with the status of the future interest rates. The use of the limitation "in money terms" in the formalized definition eliminates considerations of changes in the purchasing power of money, but there is nothing in the definition which

llJohn R. Hicks, Value and Capital (Clarendon Press, 1946), p. 172.
12Ibid.
13Ibid.

eliminates the requirement of dealing with future interest rates. How else is it to be known that the capitalized value of prospective receipts has been maintained (ex post), or is to be maintained (ex ante)? One may assume that the interest rates will not vary. One migt even assume that they will vary but that he has an exact knowledge of when and how. It is possible to have a constant principal or

a constant income stream so long as the variation in future interest rates is known, estimated, or by default, presumed to remain constant. If it is assumed to remain constant, there will be no difficulty in maintaining capital. The present value of an annuity of $100 per year indefinitely into the future is$ 2,000 assuming a permanent interest rate of 5 percent without regard to when this computation is made. Present value in this case will always be the same regardless of when the calculation is made. If the rate is expected to vary, then it becomes imperative that an estimation of variation both with respect to amount as well as time be made. The future status of interest rates in one way or another must be dealt with if income is to be stated in terms consistent with this definition. The problem may not be sidestepped by viewing the capital amount as a simple principal amount. To maintain a money amount on a perpetual basis is one thing. If, however, the principal amount is to represent the monetary expression of the present value of anticipated income streams, then not only will there have to be a reference (by assumption, estimate, or knowledge) to future interest rates, but also to the level of anticipated income streams. In order to maintain the capitalized value of prospective receipts under Hicks' first definition, there must be a comprehension of both the level of future interest rates as well as the level of future income streams. There is no alternative.

Under the definition, therefore, the return from the principal amount is termed "income" only on the condition that there has been either an assumption as to future events or a more-or-less careful estimation thereof. A statement of monetary income consistent with

definition "No. 1," therefore, is conditional upon the income receiver's personal expectations of future events. The interest rate stability

assumption does not alter the argument in any manner. The point is, that to come to grips with a monetary statement of income consistent with the definition, there is required an inevitable reference to anticipated future events on a personal basis on the part of the income receiver. Under these conditions, the future expectations of one income receiver might well differ from those of another notwithstanding an identical investment environment. Income, as a statement of a monetary magnitude in accordance with the first formal definition, is dependent upon the estimation (or whim) of what the future holds by the personal income receiver in correspondence with that degree of sophistication which his analysis of the probable status of future events contained.

At issue is not the question of whether this can be performed or not. The issue is the warranted assertability or extrapersonal significance of this utterly personal statement. Common agreement (market determined or otherwise) with respect to anticipations only postpones the inevitable. This subjective statement of income would vary dependent upon varying notions as to what is thought will transpire in the future. If such a concept is used as the guiding directive for the operations of the accountant in his search for "true income" (identified as the first of two major paths to belief in income), then an

asserted statement of existential interaction (income is) rests on intuitive estimate. The status of statements of such existential interaction depends not on the perceptive faculty of the income receiver in the structure of if-then propositions but on conceptualized conclusions drawn from experience. If "truth" is said to be approximated through the use of such a concept of income as this as a directing force, then "truth'. would seem to depend on personal estimates of future events. The argument utilizes an extremely poor sense of scientific objectivity. If anything, the assertion that income under this definition was so-and-so for a.particular period would have to be considered as a statement of a rather crude hypothesis the merits of which would as yet be unverified by conceptualized if-then propositions coupled with perceptualized reconstructed materials. Future events might well support the subjective statement of income under this first definition. Scientific explanation requires the testing of hypotheses with events ex cost to the statement of the hypothesis. But if we are to await the final termination of all events affecting the unit under consideration in order that the asserted hypothesis of monetary income under the definition might be tested, a question arises as to the practical utility of the whole system. Hypotheses always contain an "I think" element or a rather personalized notion of what existential conditions are. There is, however, a vast difference between an "I think" which is capable of existential verification and an "I think" which is not. Indeed, this appears to be the substantive difference between factual quality capable of reliance and mere opinion.

Consider the following summarized characteristics of Hicks' first approximation:

1. To state a capital value in monetary terms requires a reference

to future events. To say that a principal of $2,000 will yield an annual income stream of$ 100 indefinitely into the future under the assumption that the interest rate is to remain constant at 5 percent requires a reference not only to future interest rates but also a reference to the anticipated income streams of the property. The very statement of the principal at $2,000 is a reflection, by definition if nothing else, of the present worth of whatever the future income streams are thought to be. Likewise, if there is thought to ensue a constant stream of constant return on whatever capital value there is, then statistically that capital value is bound to remain intact, the proviso being that the return is spent or otherwise disposed of and termed "income." 2. If income may be extracted without impairing the present worth of prospective receipts, then ipso facto, its statement is conditional on the capitalization procedures which themselves are subject to future expectations. It makes little difference whether you estimate future interest rates or, by default, assume them to remain constant. 3. To the extent that there is common agreement with respect to such future events, the statement of income under the definition might be said to become interpersonal. It is, however, not extrapersonal by virtue of simple agreement. Simple agreement is not a condition for knowledge. To be sure, scientific verification of hypothesized statements requires a sensory perceptional faculty. This is consistent with the observation that all relationships with the existential world come initially to inquirors by way of human senses. Note the difference, however, between intuitive common agreement on the one hand, and the quality of the data perceived by the senses when if-then propositions are formulated or conceptualized and their indicated interaction existentially (non-conceptually) prceptualized. Whatever mode of scientific explanation one holds to be the most efficient, and indeed there are many, those which have been the most successful in the past have all had one indisputable condition for knowledge or warranted assertability. It will clearly include the provision that before statements are held to constitute "knowns," if-then evidence must be developed extrapersonally supporting the correspondence of perception with conception. If data indicate a lack of correspondence, it falls to conception the task of rearranging possibilities. The assertability of conception depends on the quality of the correspondence between conception and perception. Income No. 1 does not appear to have any such extrapersonal qualities. It appears that income is as income is thought to be. 4. There rmigt arise the question of the validity of using a market-derived figure to represent a consensus of discounted future expectations. This is the popular technique of those who appropriate the Hicksian definitions to accounting operations. Since our inquiry at this stage is evaluating the possibility of grasping the significance of income by reverting to the asserted concepts which are said to "lie behind the operational expression," it would have to be deduced that income calculated on the basis of market-determined estimates of discounted future income streams does not represent a figure which has significance extrapersonally. A significance cannot be ascribed to it by tracing the concept which is said to direct the operational actions undertaken to measure it. Income is, as income is thought to be whether individualized notions as to future events are involved or market consensuses. It matters little how many there are who are doing the guessing. What do the above four characteristics of this first formal definition demonstrate with respect to our discussion? We are interested in the possibility of coming to terms with the meaning of income on other than procedural grounds. Hicks' first definition does not offer any extrapersonal significance for income calculations. Under the formal definition, there is no conceptual or perceptual methodology by which income calculations could have any substantive extrapersonal meaning. That income exists existentially (in an ultimate, true, or real form), or that factual or assertable income is susceptible to measurement or statement in a monetary term in a manner consistent with both the formal definition as well as the mode of scientific inquiry appears to be a categorical absurdity. The second definition supposedly comes close to the "central concept" of income; better than the first yet worse than the third. We now define income as the maximum amount the individual can spend this week, and still expect to be able to spend the same amount in each ensuing week.4 Hicks adds that if the rate of interest is thought to remain constant, the second definition and the first definition will "...come 15 to the same thing." An interpretation of the first definition in the light of the second might conclude that the only difference between them is the relaxed assumption with respect to the interest rate in the second; the interest rate being held constant by assumption in the first. There is, however, nothing in the first formal definiti on which requires lIbid. p. 174. 15I'bid. or suggests the assumption of a constant interest rate indefinitely into the future. It might be noted that an assumption might equally be made with respect to the second definition. Either way, the future interest rate is dealt with on a personal expectation basis. Moreover, if, as Hicks suggests, that definition No. 1 is the definition which most people use in their private affairs, then the formal definition would seem to need rephrasing. The formal definition of Income No. 1 (or that of definition No. 2 assuming a relaxed interest rate constancy assumption) is not consistent with reckoning as income the weekly salary of an individual who expects no change in his weekly salary in the future. A capitalization of prospective receipts would require the estimation of life-expectancy. There is a difference between the present value of an annuity which has a limited life and one which is expected to continue indefinitely into the future. If an individual earns$ 100 under either definition No. 1 or No. 2, this might well be reckoned as his income, but the $100 is not the amount which might be spent with the anticipation of maintaining intact the capital value of prospective receipts. Statistically, there is no difference between the first definition and the second. They are computationally identical. If it was Hicks' intention to assume away the complexity of changes in the anticipated interest rate in the first definition, there were undoubtedly better ways of doing so. The characteristics of the second definition are much the same as the first. What is required in order to state a monetary figure which might be consumed leaving all future periods with the expectation that this saxme amount might be consumed then? We must know the income streams of the valued assets in each of the future periods as well as the interest rates prevailing in those periods. Once these are settled upon, we can compute an amount which might be extracted in such a manner as to have this consumption amount identical in all periods. Indeed, an exact knowledge of the future would permit utter accuracy in establishing capital values and such a situation defies a variation 16 in income period-to-period. Once again, common agreement with respect to expectations does not establish a greater weight of assertability. Subjective expectations cannot be criteria of "truth." Again, it seems to be categorically absurd to hold that income computed on the basis of the second definition represents a reflection of established modes of existential interaction. The existential attribute assigned to income under this definition simply would not exist without the perceiver of expectations. Income is, as income is thought to be. Up to this point we have been considering monetary expressions of income without regard for possible variations in price-levels. We have been concerned with monetary capital maintenance or monetary income continuance. Both were dependent on a statement of subjective opinion as to past events (income) with an anticipatory referential base in the future. The quality of past events was determinable only by reference to the pending future. We must now add a third and yet more bewildering complexity to this already cluttered "path" to central meaning. How do we redefine income if the perpetually anticipated constant monetary stream is expected to vary in purchasing power from period to period? 16 Sidney S. Alexander, pp. 154-156. Professor Alexander suggests that income variations as representations of errors of measurement might come as a shock to some. Under conditions of certainty variations in income'might be considered as normal by an accountant but must be termed "errors in measurements" by the economist. Income No. 3 must be defined as the maximum amount of money which the individual can spend this week, and still expect to be able to spend the same amount in real terms in each ensuing week.17 To those elements of the future which were required to be known or estimated with respect to the first two definitions, we add another. Now the anticipatory horizon is expanded to include not only future interest rates and income streams, but also the future changes in pricelevels. To be precise, we would have to arrive at a statement of the future changes in the price-level which are expected to affect the "market basket" of the individual unit in question. Of course, this estimate would have to be made net of any change in technology affecting the quality of the market basket items so that comparable goods would be considered. In order to determine this, some indications would have to be gathered which would establish the probability of product uniformity through time. These are but a few of the imponderable complexities which are introduced with this innocent clarification and closer approximation to the central meaning of income. Yet there is still another. We now come to the final complexity which, when added to the analysis and after having made compensations therefor, will bring us to the Hicksian assertion of central meaning. This final adjustment is for consumption goods which are durable to the extent of lasting through more than one income period. Income is "...not the maximum amount the individual can spend while expecting to be as well off as before at the 18 end of the week; it is the maximum amount he can consume." Money expenditures on durable consumption goods will cause spending to exceed 17John R. Hicks, p. 174. 18Ibid. consumption. The use of previously acquired consumption goods would, to that extent, cause consumption to exceed spending. Hicks suggests that only when the use of previously acquired consumption goods exactly matches current expenditures thereon will consumption be equated with 19 spending. How are we to know this? Without a perfect market for each item of durable consumption goods for each particular degree of wear possible, there is no way of estimating the extent to which spending in one period on durables compensates for consumption in another. If a person uses durable consumption goods in one period, he must compensate for this by a reduction in his spending. Otherwise, he would be "worse off" to the extent of the wear involved in the use of these durables. Hicks contends that the extent to which a greater stream of income is necessary in the future to compensate for current use of dur20 ables cannot be known without a perfect "durables" market. Hence, the conclusion that "...we are forced back on the central criterion, that a person's income is what he can consume during a week and still expect to 21 be as well off at the end of the week as he was at the beginning." The above four definitions of income are said to be ex ante in nature. That is, the calculations involved are to take place before the start of the period in question. Income ex ante for a year is the amount which might be consumed during the year with the expectation of being as well'off at the end as at the beginning all calculations having been made at the start of the year. There corresponds to each of the ex ante definitions an ex 1st counterpart. This ex post definition simply makes the calculations in point of time after the period under 19Ibid. 2011bid. 21Ibid. consideration. If the expectations were not attained, then the value of one's prospect at the end of the period will fall short of what they were expected to have been. Thus, changes in expectations during the period will reflect themselves in ex nost income calculations whereas they would not under the ex ante definitions. Consequently, ex ante income will equal ex Dost income only if expectations were exactly attained and no changes therein took place. Hence, income ex ante, plus or minus the effect of changes in expectations which took place during the period, would equal income ex post. Are the characteristics of ex post income calculations significantly different from those of ex ante calculations? Professor Hicks maintains that ex post income calculations are objective in nature. Income "No. 1" ex post "...is not a subjective affair, like other kinds of income; it is almost completely objec22 tive." It will become apparent that the significance of this statement depends on the interpretation one gives to the term "objective." If, by objective, Professor Hicks means that the computational aggregates as "inputs" to the income calculation were derived from market place transactions totally apart from personal evaluations of worth by the individual computing his incomane, then we must proceed to an analysis of how such market prices were determined. Supply and demand is said to determine price. This price, however, represents a coincidence of agreement and' not a consensus of value (discounted future expectations). For example, if the market value of capital at two different dates is known, and the degree of consumption of durables is known, then an income amount might be computed in accordance with the interpretation offered by Hicks of Income No. 1, ex post. It is altogether another matter if 22Ibid., p. 179. the term "objective" is to represent extrapersonal verification of the assertability of a monetary income. Even if market derived input values represented consensuses, any change in such opinions as to future events would change the monetary expression of income. Thus, the degree to which the monetary expression approximated the immutable (true or real income as a term representing an existential status) would depend on conglomerate opinion. In either case (Income No. 1, ex ante or Income No. 1, ex post), the suggested meaning for "objectivity" must refer to the coincidental interaction of supply and demand which itself is the subject of expectations. It will be recalled that Income No. 1, ex ante was defined as the maximum amount that could be spent with the expectation of being as well 23 off at the end of a period as at the beginning. Hicks now asserts that this capital value of prospective receipts is susceptible to objective measurement. It is held to be an objective magnitude capable of 24 direct calculation. The meaning of objectivity must be determined by reference to the nature of the capitalizations involved. The formal definition requires the capitalization of expected cash receipts at two dates in order to determine income for the period between the dates. Income would equal the difference between the present value of prospective receipts at the two dates. Ex ante, this computation would be made at the beginning of the period. Ex post, the computation would be made at two different times, the first at the beginning of the period and the second at the end. By shifting the estimator from the individual to the market, we might extract from the computation any overtones of 23Ibid., p. 172. 24Ibid., p. 179. bias or purposeful deceit. But the fact remains that they are estimates. Reference to future events is an essential part of the search for "knowns" or reliable statements of existential statuses. Future events are central in scientific inquiry. Perceived future events are clearly the determinants of the assertability of conceived hypothesized existential statuses. They must "bear out" the hypothesis before the hypothesized statement of existential status is accepted as a tentative known. It is interesting that "income" is asserted before such future events transpire. Income ex ante or ex post is stated with dependence upon future events (present value of anticipated income streams) for both its magnitude as well as for its very existence. Even in the face of a perfect market, income calculations could not be held to be objective in the sense of meaning extrapersonal. The existence of income under any of the definitions requires and is dependent upon a perceiver. Ex post definitions draw us no closer to the assertability of an existential status than do ex ante definitions. They both might be said to have an inevitable character of intuitiveness. In summary, Professor Hicks does not offer any evidence of an intelligible existential state of affairs corresponding to what he terms "income." Unfortunately nothing more might be said of the bed-rock meaning or significance of incone now than before. Essentially income is as income is thought to be. The accountant has appropriated the Hicksian definition at his peril. The role of concepts has received a great deal of attention from philosophers of science. It has received somewhat less attention among accountants. Hicks himself was keenly aware of the shortcomings of his concepts as well as the function attributed to income as an economic tool. He says that it is a tool which breaks in our hands when we endeavor to employ it. He finally allowed the doubt to escape as to whether income "...in the last resort stands up to analysis at all, or whether we have not been chas25 ing a twill-o'-the-wisp.'" The Fisherian Conceot Aside from the Hicksian concepts there are other concepts perhaps less computational in nature but nonetheless efforts toward a statement of irreducible meaning. Professor Irving Fisher compiled an extensive discussion of the meaning of income emphasizing the transitional nature of any income computation. According to Fisher, any computation of monetary income or any compilation of "objective services" (money income transformed into physical units of want-satisfying commodities or 26 services) is only an approximation to the central meaning of income. Subjective satisfactions from physical means are the motivational force behind all economic activity. VWhat we really mean when we speak of income are the psychic satisfactions which are derived from the added command of material goods and services brought about by concerted economic activity. In simple terms, a man's efforts in his business or occupation are undergone with the hope of thereby being able to satisfy wants. His labor, if it is socially desired, is rewarded with a greater command over physical goods and services from which he might derive a 27 net addition to his satisfaction level. Fisher pictured three successive stages or aspects of income. Money income was considered the first stage. It was defined in a rather 25Ibid., p. 176. 26Irving Fisher, The Nature of Capital and Income (Macmillan and Co., 1912), pp. 165-168. 27Ibid. general way as that money aggregate which a man received which could be used to secure those material goods and services vital to his existence. Real income, then, was money income adjusted for price-level fluctuations. The last stage, the real meaning of the term "income," was said to be this "psychic income." Psychic income is represented by all those "agreeable sensations" and "experiences" which come to an 28 individual through the consumption or use of material goods and services. This is asserted to be the real concept or the true meaning of income. Mattessich notes when juxtaposing the Fisherian and Hicksian concepts of income that the one (Fisher's) is based on flow variables while the other (Hicks') is based on stock variables. This seems to be consistent with the interpretation of the Hicksian concepts as 29 capital maintenance concepts. Thus, we are led back to the duality exploited by accountants when measuring income, on the one hand, in the balance sheet through the difference in stocks and, on the other, in the income statement through the accumulation of flows. The logico-mathematical principle behind this phenomenon is easily stated: Any change can be quantified in two ys: (1) by measuring the sum total of all contributing increments and decrements (flows) or (2) y measuring the difference between the two totals (stocks) connected by this change. This intuitively self-evident proposition...yields the two basic forms from which every definition of income must be chosen.30 There might seem to be some cause-effect confusion here. Hicks' notion of income would appear to make capital a derivative of income whereas the Fisherian notion would seem to make income a derivative of capital. The intuitive self-evidence statement of Mattessich appears to be self-evident only to the extent that income and capital are antecedently thought of in the Fisherian and Hicksian senses. If this 281bid. 29Richard Mattessich, Accounting and Analytical Methods (Irwin, 1964), pp. 22-23. 30Ibid., p. 23. is the case, the intuitively self-evident nature of this proposition has no more weight than any other logically correct statement determined on the basis of the implied interaction of "givens" or "assumptions." The proposition is intuitively self-evident only as it depends on what is already understood about the meaning to be attributed to capital and income. What are the known elements of existential interaction settled upon by either (stock or flow) notion? Fisher asserts that income constitutes the psychic benefits derived from the consumption of material goods and services. Hicks asserts that income has significance in that it is the amount which might be consumed with the expectation that the individual will be as well off at the end of a period as he was before. Where is the indicated known? Where is the criterion of usefulness for either definition? How would one go about reconciling such varying points of view? Both are, in the final analysis, totally dependent on a personalized notion of well-off-ness. Why might it not also be held that factual current well-off-ness can never be known in the present for it inevitably requires a reference to anticipations? Definitive statements of irreducible crystallized experiences do not change the nature of the factual extant world. Nor do we gain any greater insight into this world or its interacting components through the exclusive use of 31 concepts alone. The definitions of Professors Fisher and Hicks do not lead one to the formulation of if-then propositions, indeed, none are even indicated. Income is, as income is thought to be. Those in accounting who endeavor to app opriate the concepts of "economic income" to 31This distinction between concepts and percepts and the pitfalls which await those who seek to rely exclusively on one or the other in inquiry is set forth in detail in Harvey T. Deinzer, Methodological Presuppositions in Financial Accounting Models (Department of Accounting, College of Business Administration, The University of Florida, 1968), pp. 7-8 and 13-14. accounting operations in the form of making them a referential base for guiding the accountant's operations do not appear to have a very good "case" when it comes to stating an approximation to true, real, or fundamental income. In sumary, many writers in accounting have insisted that even if income is such a slippery notion that it cannot be stated exactly, accounting income is still a good approximation to the real thing. In these cases it is seldom known just exactly what the writers mean by income, but more seriously is the question of how we are to come to terms with assertions of approximation without any knowledge of absolutes. What justification is there for statements attributing a quality characteristic to income computations when no knowledge of such an "ideal" exists? If income computations are said to be good approximations, the question must be asked with respect to what. If we are to knowledgeably assert the approximate nature of accounting income calculations with respect to the "irreducible" statement of income or this ideal, then perhaps we would also be in a position to go ahead and state the real thing since we would have to know the relationship between it and the asserted approximate figure. Some engaging insight might be gathered from a consideration of the following irreducible notions of income. This list of constitutive essences of income is by no means exhaustive. It might, however, be indicative of the efficiency of evaluating the meaning and significance of computed income by tracing the operation "backwards" to the concepts which are said to be the directing force. Not all of the following statements have been suggested as having the quality of a "guide" for directing the operations of the accountant. They all do, however, represent what the individual author thinks the meaning of income to be. They are attempts at stating a conceived notion of an existential status. John B. Canning agrees for the most part with Fisher, although his "services" might be interpreted as being once removed fran the psychic benefit notion, "For income in essence is services - the desired 32 element in economic events." David Solomons isolates income as an unqualified growth in present value for "...growth in present value must be what we had better under33 stand income to mean." Richard Mattessich defines income as the "...flow of goods and services within a well-defined period, between the production side and 34 the consumption side of an entity." The difference between the flow of services and the psychic benefits or satisfactions which are derived from their use might be made note of at this point. A. C. Littleton says, "The really basic idea goes deeper. Income is the 'service charge' needed to induce the production of goods and 35 the employment of labor." Professor Norton M. Bedford selects what he terms a constitutive definition for income. It amounts to the "...extant intuitive notion 36 that income is a constitutive concept." Bedford later suggests that 32John B. Canning, "Some Divergencies of Accounting Theory from Economic Theory," The Accounting Review, March, 1929, p. 8. 33David Solomons, "Economic and Accounting Concepts of Income," The Accounting Review, July, 1961, p. 375. 3Mattessich, p. 21. 35A. C. Littleton, Structure of counting Theory (American Accounting Association, 1953), p. 22. Italics added. 36Norton M. Bedford, Income Determination Theory (Addison-Wesley Public ing Company, 1965), p. 72. an overpreoccupation with the preciseness of a concept might "...violate our commonsense intuitive feeling that there is something real about a 37 concept, as opposed to just a set of measurements." If one were to transfer his "constitutive" notion about income to the above statement, the result would be "income is bound to be, for we all feel that it exists intuitively." Still another shade of income meaning is expressed by Professors Backer and Bell. "Ideally, it would seem that income should represent the increment in entity value that occurred during the period, exclu38 sive of equity transactions." John M. Keynes suggests as a meaning for income of an individual "...the excess of the value of his finished product sold during the period 39 over his prime cost." When all else fails, we are "...forced back on the central criterion, that a person's income is what he can consume during a week and still expect to be as well off at the end of the week as he was at the begin40 ning." To summarize in one sentence, income is as income is thought to be. There is clearly no evidence to support the statement that there is an intelligible existential state of affairs which corresponds to any of the definitions examined. All notions as to meanings within the topic of "income theory" are definitive in character and their merits must be evaluated by virtue of recourse to the effectiveness with which they 37Ibid., p. 70. 38Morton Backer and Philip W. Bell, "The Measurement of Business Income, Part 1 - The M tching Concept," Modern Accounting Theory (PrenticeHall, 1966), p. 69. 39John Maynard Keynes, The General Theory of Employment Interest and Moy (acmillan and Company, 1936), p. 53. 40Hicks, p. 172. carry out their asserted functions. Income is not a term which represents an assertable existential status. It is not even representative of simple agreement as to definition. Tracing income computations "backwards" to their conceptual meaning offers nothing more than the blind hope of concept agreement, but then what . Proponents of this "economic income ideal" adaptation hold that income exists in a real or fundamental sense and that it falls to the accountant the task of approximating this real amount as closely as possible (we have seen also that there are some who claim that modern accounting methods factually do approximate this ideal figure). Belief in the propriety of income reporting supported by the argument which was examined in this chapter would appear to be just that, belief. If income determination is to receive a favorable critique, the search for its warrant will have to be continued elsewhere. The second of the two major beliefs in income was isolated to be the contention that income reporting is vital to sound investor decisions. Future levels of income are said to be determinable from the indications of past trends of income. This predictive potential is thought to provide prospective shareholders with information on which to base good investment decisions. The remainder of this study will deal with this belief. CHAPTER III STATEMENT OF STATISTICAL PROCEDURE We have, then, disposed of the possibility of dealing effectively with an evaluation of income reporting for shareholders on the basis of its asserted correspondence with conceptualized statements of interaction (alone) or conceived ideal definitions. Aside from having a lack of consistency with modern knowledge-acquisition processes, and aside from the sub-question of the lack of general agreement on a definition for this conceptualized ideal, it was found that there existed two insurmountable difficulties when an effort was made to apply this "correspondence-to-the-economic-income-ideal theory." The first was the impossibility of stating this income figure in terms which would be extrapersonally significant. The second major difficulty stems from the first. Without the possibility of such a statement, there can be no "ideal" against which the operational counterpart in accounting might be contrasted. There is, therefore, no possible way of determining whether and to what extent the income calculation of the accountant approximates the "true income" (as represented by such "ideal" economic income definitions) of the enterprise for a period. Any figure settled upon as the net income of an enterprise for a period of time cannot be said to be "efficient," "approximate," "true," or "correct," by reference to any asserted correspondence to conceptualized "essences" of income. Those who would propose this approach evidently think of the accountant as proceeding with a measure of income knowing that the specific attribute which he seeks to measure is there somewhere. There is simply no evidence of the existence of a measurable attribute in the economic realm with which extrapersonal significance might be associated by recourse to a supposed correspondence between the measured amount and some "ideal" amount, or economic income. Intellectual discussions pertaining to income and "income theory preoccupy many accounting writers. As a profession, public accountancy has gone to no small length to refine and delineate what is considered "proper" accounting procedures for the "realistic" reporting of net 1 income. In general, this approach has been an appeal to the concepts which are said to be involved in the traditional historical cost allocation and revenue recognition notions associated with income determination. In broad terms, it is fair to say that the accountant defines income as the excess of that which is properly classified as revenue over that aggregate of items which is properly classified as expense. This is a traditional procedural definition which, under somewhat more detailed consideration, appears to be a statement of how to "get to" income without any indications of the meaning of what you have "gotten to" once you have "gotten to it." It is a matter of fact that the computational result of the accountant's operations has simply been termed "income" through tradition, and that there does not exist any known method of relating it to what has been called "economic income," "real income," or "fundamental" income. The writings of Churchman may be IThe contention that income will be more "realistically" stated with the adoption of a particular procedural alternative instead of another is nowhere more evident then in an article title written by a former high-ranking individual in a major accounting firm. Herbert T. McAnly, "The Case for LIFO: It Realistically States Income and is Applicable to Any Industry," The Journal of Accountancy, June, 1953, pp. 691-700. Italics added. interpreted as indicating the confusion which arises from the exclusive 2 reliance on concepts in the study of economic behavior. The other possible approach to an evaluation of income reporting for shareholders appears to rest in the area of the asserted functional utility of reported income figures. Income reporting has many uses aside from those associated with the annual report. It is a useful device for taxation purposes, for example. Regulated monopolies are supervised by a form of income determination to insure that rate charges meet socially acceptable criteria. All of these uses of various forms of income reported are associated with a certain group of individuals. The income determination technique is designed with a particular use in mind. Yet, financial income reported in the annual report is designed by appeal to such notions as "realistic income," or income figures which are neither "overstated" nor "understated." Paramount among the asserted uses of the financial net income figure is that of being a component of rational investor decisions. Ostensibly, a rational investor acting within the purview of the asserted functional utility of income reported for shareholders would tend to maximize his financial gain by examining the past income trends of all investment alternatives. This assertion carries with it many if-then implications which would need to be investigated before any merit might be placed on the statement. The place of market values would, for example, need a careful look. Since a part of any financial return on any common stock (particularly those which are 2Churchman came upon this insight when some ambiguities in the meaning of "cost" occurred to him. Cost, as a representation of the value of the alternative not taken, should logically be extended to represent the optimal alternative not taken, which implies a knowledge of all alternatives, which implies a knowledge of the future. C. West Churchman, Prediction, and Optimal Decision (Prentice-Hall, 1961), pp. 59-62. listed on national exchanges) will be determined on the basis of the change in stock price, such an assertion would seem to have predictive implications as to the behavior of market prices in response to income trends. If the implications of this assertion can be established, then the statement would be supported by evidence which would indicate that it might be relied upon and acted upon with confidence in the consequences. The hypothesized implications of the assertion, the then element in implied if-then propositions, is the central focus of attention in the remainder of the study. Hypotheses are tested by developing the implications of the thought-to-exist existential state of affairs by reference to selectively determined experimentation. IF past income figures constitute a component of rational investor decisions by virtue of their ability to indicate income amounts to be realized in the future, THEN: 1. The extrapolation of past income trends must be an efficient indicator of future income trends and, 2. Decisions based on the relative attractiveness of projected earnings (qualified, of course, by investment cost) must lead to the selection of the factually higher cash return stocks determined on an ex post basis. Although the statistical implications of the tests indicated above are rather complex matters, the two implications of this second belief in income which are identified as the then elements above, constitute the central tests of the hypothesis which will be made. It is believed that the quality of a contribution to the literature of any discipline is not determined on the basis of positive suggestions only. If the hypothesis (and consequently the belief) is supported by the evidence, we will have more to base our actions on than pure assertion. If it is not supported, there will be a need for a redirection of effort. Either outcome would constitute a partial determination of a currently unsettled and untested situation. In order to convey a broad overview of technique, a brief outline of the entire study will be made initially. There are two basic questions involved. The first is the extent to which earnings of past years are efficient indicators of the earnings of future years. The second is the question of decision efficiency when stocks are chosen on the basis of the relatively more attractive projected price-earnings ratios. Since changes in market price constitute a part of investment return, it is possible that even if earnings are susceptible to prediction the relatively more favorable stocks might not have been chosen (ex post determined). Predictions may well exactly match factual performance but unless the market price of the stocks is also consistent with the predictions, the relatively more favorable stocks might not be chosen. Hence, there is more to the question than the mere prediction of earnings from past trends. There are two central procedures involved. The first is designed to yield a measure of the efficiency obtained in predicting future years' incomes from past trends. Basically, this part of the study will use an extrapolation of past earnings per share figures from a least squares trend line and thereby estimate the anticipated earnings per share figures for a future period. Stocks will then be arranged in the order of highest anticipated earnings per share figures to lowest. This array will then be rank correlated with a similar array with factual earnings per share figures as the value ranked from highest to lowest. A high rank correlation coefficient would tend to indicate that past earnings per share are good estimators of future earnings per share. But even if the rank correlation coefficient is determined to be 1.0, this still is no indication that prospective investors are any better off with income reporting than without it. If they are said to be able to make decisions based, in part at least, on the information supplied in income statements, we must determine whether it makes sense to choose investment stocks on the basis of the relative attractiveness of projected income figures. The decision ostensibly implied in this contention is made on the basis of the higher projected price-earnings ratios of some stocks in relation to all others. The second major analysis, then, will be a rank correlation study between an array of projected price-earnings ratios on the one hand, and an array of actual cash return on the other. If the same stocks with the higher priceearnings ratios also turn up in the upper sections of the array of stocks listed in terms of cash return from highest to lowest, then the hypothesis would tend to be supported and the second belief in income valid. The cash return will be determined on the basis of cash dividends received during the investment period plus or minus the net change in stock value 3 from the start of the investment period until the end. 3The indicated inputs for the rank correlation follow in formula form: Projected Price- - Projected EPS 1963-1965 From Trend Extrapolated Earnings Ratio Least Squares Regression of EPS of Prior Period Investment Cost Factual Price- Factual EPS 1963-1965 Earnings Ratio Investment Cost Rate of Cash Return = Dividends 1963-1965 plus Stock Value Change Investment Cost It is imperative to recall that a stipulation was made in Chapter I to the effect that casual observation would convince one that there were other factors which would enter into any particular investment choice. Of interest, however, is the efficiency lent to the decision process to the extent that past income trends influence the decision. It is fully clear that the majority of investors make their decisions on the basis of many factors, not the least of which is a good honest guess. We are interested in the extent to which the decision process is helped or hampered by full or partial reliance on extrapolated past trends. The question of real interest is whether the decision base is fully or partially valid. Does income determination do what it is asserted to do? The Sample The sample consists of one-hundred-ninety-eight common stocks selected primarily on the basis of the availability of relevant data for the study. The sample was not selected purely at random in strict accordance with random sampling techniques. More will be said later on the consequences of this limitation. The sample was selected from data published in the 1963 and 1966 editions of Moody's Industrial Manual. The data were originally derived from either the content of the annual report to stockholders or the related statements which were filed with the Securities and Exchange Commission. There were comprehensive data available for some three-hundred-fifty stocks which the Moody Manual goes into with somewhat significant detail. Of these three-hundred-fifty stocks, only one-hundred-ninety-eight of the listings contained sufficient data with which appropriate adjustments could be made to translate reported data into relevant data for this analysis. .50 In many cases stocks were rejected on the basis of inconsistencies in the raw data reported in the financial statements. In many other cases, insufficient data were available with which to adjust earnings per share figures necessay for the establishment of valid trend lines. In other cases the time sequence when both cash dividends as well as stock dividends were paid was not reported. This precluded a determination of the actual share equivalents held at the end of an investment period which were attributable to one full share held on the investment decision date. This constraint would have precluded an accurate determination of the cash return for the stock. If, for example, a cash dividend were paid immediately before a stock dividend, the cash dividend per share (or part thereof) would be more than if the same cash dividend (in total amount) were paid after a stock dividend. In the absence of an explicit indication in those cases in which the figures were so close as to indicate doubt as to which occurred first, the stock was rejected. If the timing sequence was indicated or if there was other information indicating which occurred first, the share was included and an accurate share equivalent figure was determined. The three-hundred-fifty comprehensively treated corporations represent the most important American corporations in terms of total resources, total income, or both. The sample, therefore, will represent a fairly good cross section of American industry. It does not encompass all major stocks. A data constraint prevented a consideration of all important stocks. To include all such stocks would have required a tracing of data from every stock for every year in every set of financial statements. This would have entailed a mammoth undertaking. Yet, these limitations are not as serious as they might at first appear. The large sample size (n value) suggests that little more would be expected to be uncovered by including all stocks or by expanding the "n" value very significantly. The sample size should compensate to some degree for the lack of complete randomness in sampling technique. When considering the conclusions drawn from the study, it would be helpful to keep in mind not only the limitation as to the lack of randomness, but also the large "n" value for they tend to compensate for one another. Since the hypothesis suggests a predictive quality for past income calculations, the true universe size is quite large in that it comprises all common stocks. There is a vast difference, however, between the importance of this assertion as it applies to all stocks and its importance as it applies to major corporations. The sample also represents some fifteen broadly defined industries. The data will be analyzed both among all stocks as well as among stocks grouped by industry. In the way indications as to the predictive capacity of the data intraindustry might be determined. The selection of the industry groups is, at best, a subjective affair. Rather than use the industry groupings which Moody's uses for ratio analyses, which would, in this case, result in a "thin spreading" of some forty-five industries, (many of which would have only one common stock in the classification), some of the industry classifications were combined yielding a total of fifteen industries. These represent what the author considers adequate classifications for the purposes described above. The classifications are summarized below: 1. Aerospace - 3 Corporations 2. Steel - 10 Corporations 3. Variety Stores - 4 Corporations 4. General Retailers - 8 Corporations 5. Chemicals - 7 Corporations 6. Cement - 4 Corporations 7. Pharmaceuticals - 11 Corporations 8. Machinery, Equipment, and Tools - 11 Corporations 9. Building Materials - 5 Corporations 10. Oil Refining, Integrated, and Producing - 11 Corporations 11. Railroad Equipment - 4 Corporations 12. Tobacco - 4 Corporations 13. Food - 14 Corporations 14. Paper, paperboard, and packaging - 6 Corporations 15. Electrical Equipment, Controls, and Instruments - 6 Corporations It will be noted that the total number of corporations represented in the above classifications does not equal the total sample. This is explained by the fact that there were some corporations whose products did not fit appropriately into any of the above classifications. Many of the largest corporations are so diversified that to include them in any single classification would be to exclude them from other groups into which they might be equally placed. Also, consistent with the original Moody classifications, there were thirteen corporations whose stocks would have comprised one separate industry each. In ces such as these, the common stocks of those corporations were excluded from the industry grouped analyses but included in the all stocks analysis. Relevant Data Techniques designed to transform raw data into relevant data need to be considered before procedural aspects of the statistical technique are entertained. The earnings figures of prior years in the form in which 53 they are normally reported in successive annual reports are not valid bases from which to strike a trend line representing the "direction" which earnings are taking. Data need to be gathered for appropriate "rates of return" figures. These inputs to the rank correlations need to be computed from the raw data published in the annual reports. Earnings per share The reported operating income or net income is not the decisiondetermining factor to a prospective common stockholder. There are two reasons for this. Corporate earnings are irrelevant first of all because normally the cost which the shareholder pays for his share is stated in per share terms. Hence, corporate earnings would have to be reduced into the appropriate per share earnings. Cost is one of the most important elements of any investment decision regardless of what the decision criterion might be. On whatever basis an investment is selected, that basis is generally reduned to a common association with whatever is known about investment gost. The second reason, and perhaps the more important, stems from the possibility of there being a preferential issue which stands, as it were, in front of the common shares as to earnings. As a consequence, all that is reported as corporate earnings or financial net income cannot in any way be necessarily associated en masse with the common shares alone. Corporate earnings are, therefore, irrelevant for both the decision environment of the prospective investor as well as for the current purpose. If the trend of earnings is to be extrapolated, the relevant element of corporate earnings is the per share earnings which are attributable to one common share. Yet, this is still not enough. If earnings per share figures are to represent the average number of shares outstanding during any given fiscal year divided into the opera- ting or net income of a corporation, then stock dividends or stock splits will significantly affect the average number of shares outstanding. This will, in turn, affect the per share earnings reported in financial statements. For example, if corporate income is constant over a two-year period, and there is a two-for-one split at the end of the first year, the second year's earnings per share figure will be onehalf that of the first year's. If there were reason to establish a trend of earnings attributable to one share held or purchased at the start of the first year, the reported earnings per share figure would have to be doubled for the second year in order for the trend line to 4 be valid. Corporate earnings as reflected in the income statement are, therefore, twice removed from relevance as far as establishing earnings behavior of a single share. In the first place, any earnings attributable to any preference stock must be extracted from corporate earnings. In the second place, earnings attributable to the common stockholder's equity (earnings attributable to all common shares) need adjustment for any stock dividend or stock split which might have taken place during the trend determining period. As in the case of the sample, there were limitations on the number of years for which adjusted earnings per share figures could be calculated. For all one-hundred-ninety-eight stocks there will be sufficient data available for two separate trend periods, 1956 through 1962 and 5 1960 through 1962. For forty-one stocks in the sample, an additional 4Although there are different accounting entries for stock splits and stock dividends, the effect on relevant data in the study is the same. A 100 percent stock dividend is identical to a 2 for 1 split. 5Trend period refers to the years in which earnings are regressed. Investment period refers to the years in which the investment is held. trend period can be established for the period 1951 through 1962. Thus, for forty-one stocks there will be data available for three separate periods in which valid trends might be established, a twelve-year period, a seven-year period, and a three-year period. For one-hundred-fiftyseven stocks, there will be data available for only two trend periods, the seven-year period (1956 through 1962) and the three-year period (1960 through 1962). It is hoped that through the use of such varying trend periods the study might cover possible criticisms over the timing of economic causes and effects. For example, if only the shorter trend period was used, it might be asserted that stocks must be evaluated by reference to their income performances over long periods of time. The year-to-year fluctuations could be held to be inevitable with no significant effect on.the general long run trends. On the other hand, if only the long run trend was used, it could be held that long run trends are not predictive due to the constant changes taking place in technology and marketing. It could be held that the shorter periods alone constitute the best basis for predicting future years' incomes because if trend periods are stretched out over too long a period of time, the economic and technological "causes" of past performance are no longer operative. Through the use of varying trend periods, it is hoped that questions such as these might be avoided. If the longer periods are, in fact, more predictive, we shall find out, and vice versa. The primary question, it should be recalled, is not whether the one is more efficient than the other, but whether either is efficient in predicting which of all possible investment alternatives (in common stocks) are the better prospects. In all components of the study, the central question must be kept in mind. Every element is designed to establish the qualitative aspects of the implications of the assertion that currently reported financial income computations constitute relevant data for shareholder and prospective shareholder decisions in that they form a basis for predicting future years' incomes. With the above background in mind, a detailed exposition of the earnings per share adjustments can be made. The nature of the problems encountered in adjusting reported earnings per share figures is best described by recourse to an illustration. While it is the central purpose of the study to examine whether the data implications support or refute the principal assertion, there is the equally important matter of determining whether earnings may be efficiently predicted in the first place. To settle this question the three trend periods were established to see whether future incomes are, in fact, susceptible to projection and prediction from the indications of past trends. This trend line will be drawn in accordance with a least squares regression formula representing a mini=m total distance (positive and negative) of the combined distances of all adjusted figures from the trend line. No mention has been made, however, about the base year to which all earnings per share figures must be adjusted for a valid trend expression. To adjust the published earnings per share figures in the context used here means to state them all in equivalent dollars of earnings (not adjusted for price-level change) consistent with a common base year. One might establish adjusted earnings per share figures in terms of the equivalent earnings of a single share held at the start of the ten year period, at the end of the period, or any year of the ten year period. Assume a hypothetical corporation has constant earnings for a ten-year period (1951 through 1960). The only change that takes place in the financial affairs of this corporation is a two-for-one common stock split at the end of the fifth year. In this case, the reported earnings per share data might look like this: Earnings Per Share - Unadjusted 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 2.00 2.00 2.00 2.00 2.00 1.00 1.00 1.00 1.00 1.00 In order to establish a valid trend representative of the average historical experience of a common share in earnings, some adjustment must be made. One has a choice. A trend might be struck which would represent the earnings in each of the ten years which are attributable to a common share held at the start of the trend period (1951). In this case, the adjusted earnings per share data would be set forth as follows: Earnings Per Share - Base 19251 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 The major point is that adjustments are made to reflect earnings attributable to a single common share held at the start of 1951. In 1956, there is held two shares instead of the one which was held from 1951 through 1955. The two shares in 1956 are directly attributable to the share held in 1951 and, accordingly, the earnings attributable to that single share of 1951 are represented by the earnings of both shares held in 1956. If, on the other hand, the trend is established in terms of the historical record of the equivalent of a single share held at the end of the trend period, the following data are appropriate: Earnings Per Share - Base 1960 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Either of the two trends is valid, but the choice of which to use must be made on the basis of the function of the trend. The selection of a particular base year for adjustments carries with it implications as to the consistency of the investment cost figure with that year. For example, if the choice is made to use 1951 as the base year, one would have to double the investment cost prevailing in 1960 for a valid projected price-earnings ratio. At 1960, the prevailing stock cost would represent only one-half the equity to which all trend earnings have been made consistent. If it were deemed necessary to use the prevailing market price in 1960 for an investment decision, then it would seem more likely that the earnings adjustments would be made using 1960 as the base year. In order to avoid the extra computations required of two separate base year adjustments, the base year to Ahich all (trend period as well as investment period) earnings per share figures were adjusted in this study was the year 1965. As a consequence, investment cost figures used in this study need not necessarily be the simple cost of a share on or near the hypothetical investment decision date. If a stock were split subsequent to the investment decision date, the actual earnings per share for the investment period would reflect the equivalent earnings which were attributable to a single share held as of the end of the investment period, not the earnings attributable to a share held as of the investment decision date. In the event of a two-for-one split in the investment period, for example, the value for the investment cost would have to be 59 the published price for a share on (or near) the investment decision date divided by two. All factors must bc made consistent once a decision is made as to a base year. The earnings per share data, the dividends received data, the investment cost data, and the sale value of the stock all must be mutually consistent for the analysis to be valid. Investment and disinvestment decision dates Because of the lack of information as to when within any year dividends were paid to common stockholders, the investment decision and the disinvestment decision were timed to insure accurate cash return data. The investment decision date for all stocks will be the start of the fiscal year (for the particular stock) 1963 and the disinvestment decision date will be the end of the corresponding fiscal year in 1965. A corporation whose fiscal year coincides with the calander year will have decision dates of January, 1963, and December, 1965. A corporation with a fiscal year ending in August will have investment decision dates of September, 1963, and a disinvestment decision date of August, 1965. More is said below on the exact determination of investment cost. Investment cost Hypothetical investment decisions will be made on the basis of the relative attractiveness of predicted price-earnings ratios which will have been determined on the basis of predicted adjusted earnings per share as a percent of investment cost. It will be noted that projected earnings is not the sole criterion of advisable investment. Per share anticipated earnings must be related to investment cost before the relative attractiveness of each alternative can be determined. A stock with highest projected earnings per share need not place among the "best bets" since the related cost of acquiring that share might be so high as to make the cost of those anticipated earnings prohibitive. The technical decision criterion (as implied in the statements supporting the second belief in income outlined in Chapter I), therefore, will be the ratio of average projected earnings per share over investment cost. The settlement on a single figure for investment cost is a difficult matter and the process is clearly not without subjective overtones. To simply seize upon one single day as representative of the cost of the particular investment under consideration appears to be completely inadequate for the purposes here. It was considered more reasonable to go through an averaging process for the stock prices of a particular share which surrounded the decision date. In this way, it is hoped that a more representative investment cost might be established. To pick a single trading day to establish stock prices might well catch many stocks at a peak or trough thereby lending a severe bias to the results. In both the projection as well as the evaluation segments of the study, an averaging process was used. The mid range of the stock price in the month preceeding the investment decision date as well as the mid range of the stock price in the month following the investment decision date were averaged to produce what is termed herein "investment cost." For example, if January 31 were the hypothetical investment decision date, the investment cost would be the average of the mid ranges of the stock price for the months of January and February. Or, what amounts to the sa&me thing, it would be the high price for the months of January and February plus the low prices for January and February, divided by four. Investment cost is, therefore, the average stock price for the twomonth period surrounding the investment decision date as determined by the stock price range. Since a part of'the return on any stock con- sists of price fluctuations, to carry the averaging process too far would result in averaging out the very changes which constitute a part of the implied predictive potential of past earnings. The two month averages noted above are considered the minimum averaging process consistent with averaging out only day-to-day market price fluctuations. To take the process beyond the two month period would be as unrealistic 6 as not to utilize any averaging at all. This procedure mitt be objected to on the basis that observations may indicate that investors do not have access to such a representative stock price and that, in practice, investors seek to await the low price, or the "good time to buy," and the high price or "the good time to sell." This observation is readily conceeded as a possibility and appropriate adjustments in investment cost and sale price figures will be made. In order to accomodate this question, a sub analysis will be undertaken with appropriate rank correlations just as with the major study, which utilize a time-flexible investment cost and sale price. Investment cost will, in this case, be interpreted as the lowest possible price prevailing within the two month period surrounding the investment decision date. This will correspond to a sale price which will represent the highest possible price prevailing within the two month period surrounding the sale date. In this way the study will encompass the possibility of this behavior and stipulate its probability, but it will not leave it at that. If the contentions of the would-be critics are warranted (and this question will not occupy us in this study), we will have an answer. 6 These data were gathered from Standard and Poore's Monthly Stock Guide for the years 1962 through 1966. Since all stocks are listed, the prices are consistent with the New York and American Stock ExchangeAssociated Press price quotings. Share equivalents It will be recalled that stock dividends and stock splits caused a rather bothersome problem in gathering relevant data for earnings per share figures for trend periods. These transactions also cause problems in determining an appropriate investment cost figure. Since the adjustments for all earnings per share figures are to be made on the basis of the year 1965, projections as well as evaluations will have to be made consistent with that same base year. Hence, any computation for investment cost will have to be divided by the share equivalent at the end of the investment period for a given share held at the start of the investment period. The equivalent of a single share purchased at the start of an investment period would be two shares if there had been a two-for-one split during the investment period. In the event of a 10 percent dividend, for example, the share equivalent would be "1.1" if this stock dividend took place during the investment period. In the former case, a holder of a single share would hold two shares at the end of the investment period, and in the latter case, he would hold 1 1/10 shares. Since all earnings per share figures are adjusted to be consistent with the past equivalent earnings of a single share held at the end of the investment period, whatever computations are made to arrive at an investment cost figure will have to be divided by the share equivalents if the original unadjusted share price figure is gathered from published prices of the share at the investment decision date, whether an averaging process is used or not. For example, if actual price quotings yield an averaged investment cost of$ 50.00 on an investment decision date, the actual cost of the investment for the earnings attributable to a single share held at the end of the investment period

would be $25.00 if there had been a two-for-one split. Hence, a share equivalent value in this case would be "2," the value to be divided into published investment cost at the start of the investment period. No such division will be required at the end of the investment period for the determination of the sale price since all data are made consistent with the 1965 termination date. 'Whatever is settled upon as the averaged investment sale value will not need share equivalent adjustment. The use of the averaging process does not negate the requirement of dividing investment cost by share equivalents for an appropriate investment cost figure. The individual components of that averaging process are unadjusted so the final average is also unadjusted. The adjustment would be required whether a single trading day was siezed upon as an appropriate index of share value or whether the average of many such days is used. The investment cost figure (not the sale price) must be divided by the share equivalents whether it is averaged or not. Dividends received There is a subtle difference between the constraints of adjustment in establishing the decision criterion .nd the actual cost of the investment decision itself. For the projection of earnings and for the establishment of the anticipated price-earnings ratios on the basis of which the investment decision is made, adjustments are required to make all data mutually consistent. 1When the actual investment is made, it will be made on the basis of the then current unadjusted cost of a share. This share cost will have been adjusted in the sense of having been subjected to the averaging process, but it will be unadjusted in the sense of not having been subjected to share equivalent adjustments. The investment cost used in the projection segment might bear no relationship with the investment cost used in the evaluation segment. The evaluation segment will use a stock price which is the average of the actual stock prices current at the time the investment is made. This accounts for the difference between the cost of investment used for projection and the cost of investment used for the decision in those cases in which the stocks have a share equivalent value greater than 1. This is not an inconsistency. The only reason that projected investment cost varies from the decision cost is due to the requirement that cost and adjusted earnings per share figures must be consistent in order to get valid projected price-earnings ratios. This insight is important when a consideration of just exactly what is meant by the term "share" in computing dividends received per share. Dividends are obviously a component of cash return, but does "share" refer to the dividends received on the equivalent of a share held at the end of the investment period? Is it also subjected to the share equivalent adjustment? Does it refer to the dividends received per share in the sense of what is meant by that term in the annual reports? Or, does it refer to the total amount of dividends which were paid to an investor by virtue of the single share purchased at the start of the investment decision period? For accurate determination of cash return, dividends per share must encompass all dividends (cash) which were received by virtue of that single share purchased at the investment decision date. If there had been a two-for-one split, then, the reported dividends per share would have to be doubled for those years in the investment period after the occurrance of the split. If the split took place at the start of 1964 (the second year of the three-year investment period 1963-1965), the dividends received would consist of the reported dividends per share for 1963, plus twice the reported dividends per share for the years 1964 and 1965. These adjustments are implicit in the dividends per share figures. the adjustments having been mrde at the time the data were gathered. Capital gain or loss Dividends attributable to a single share held on the investment decision date constitute only one part of the investor return. The other part of the return would be reflected in any increase or decrease in an investment price (market value of the stock) from the time of investment to the time of sale. These two sources constitute the only two ways in which an investor can get purchasing power from an investment in common stock. Consequently, whatever device is used for the prediction of the better stocks, the decision criterion must be predictive with respect to either dividends or capital gain, or both. Those who assert belief in the propriety of income reporting on these premises assume a predictive power of past earnings trends (although this is generally not made explicit). Again, it would appear to be rather unrealistic to simply take one trading day and establish as the sale price for the stock the high or low price at which the stock was traded on that day. Therefore, the same averaging process will be used in establishing a sale value as was used is establishing the purchase price or investment cost. The mid ranges of the two month period surrounding the sale date (the last day of the corporate fiscal year) will be averaged. In addition to the selection of this sale value f.o: the stock, an entirely different set of rank correlations will be used when time flexibility is introduced. This will correspond to the time flexible sub-analysis which was described above with respect to investment cost. The investment cost was determined both on the basis of the average of mid ranges for the two month period surrounding the investment decision date as well as on the basis of the lowest stock price prevailing during that two month period. This procedure was allowed so that tests could be run which would give indications of the reliability of past income trends in settling on the relatively better common stocks if this "good time to buy" and "good time to sell" notion were, in fact, operative. The "good time to sell" aspect of this sub analysis is interpreted as the highest possible stock price prevailing during the two month period surrounding the investment decision sale date. As with dividends, it is important to compute capital gain or loss associated with or attributable to the originally purchased share. Thus, if there had been a two-for-one split during the investment period, it will be necessary to double the share price prevailing at the time of sale in order to get a valid cash return figure. In this case the prevailing cash value of the stock which was purchased on the "buy" date is represented by two shares at the sale date. In summary then, the decision criterion will be the relatively more attractive stocks in terms of the projected price-earnings ratios which are determined on the basis of extrapolated (from three past periods) earnings adjusted to be consistent with an investment cost. The evaluation criterion will be actual cash return in the form of dividends attributable to the single share held on the investment purchase date plus or minus the net change in investment value (not share value), divided by the averaged cost of a share on the investment decision date. The values in these two arrays will be ranked in the order of highest value to lowest value, and then the two arrays will be rank correlated. The rank correlation coefficients will indicate the degree to which this belief in income under test is to be supported. Implicit interest Since there are generally two basic reasons why common stocks are chosen as investments (growth and income), it might appear that a study of this complexity should adjust the data further to give explicit effect to implicit interest involved in investment costs and negative interest adjustments for dividends when they are received. This lack of adjustment might be considered by some as a serious limitation of this study. The relevant data for such adjustments were not available and, thus, such adjustments could not have been made even if it were desired. The limitation, however, is not as serious as it might at first appear. The extent to which this limitation constitutes a serious weakness should be determined, in patt at least, on the basis of judgements as to whether prospective shareholders actually make such implicit interest adjustments in practice. This appears doubtful from casual exposure to investor behavior. Moreover, in order to cause a serious defect in the findings of the study, the lack of interest adjustments would have to result in a deviation in the rankings of the stocks and not simply deviations in the projected price-earnings ratios. It was for this reason, among others, that rank correlation analysis was chosen as the statistical tool. Since, for the most part, the greater amount of any implicit interest adjustment would work in the same direction for all stocks, this weakness does not appear to be quite as serious as it first seems. Moreover, to make this adjustment would require a knowledge of when within 68 a given year the dividends were received. These data were not available. To establish dividend receipt dates, therefore, would have been conjecture in the first place. Of more importance, however. the arbitrary receipt dates might well have resulted in a serious distortion instead of a slight refinement. Ideally, the study should have made such adjustments in accordance with exact dividend receipt dates. We stipulate the validity of the implicit interest adjustment. The results, however, are not expected to be materially affected by this constraint. The hope of establishing conclusive proof was abandoned in Chapter I. We seek the indications of the data as they shed light on the implications of the belief under review. We now proceed to a detailed illustration of statistical technique. The adjustments which follow below for the single hypothetical share represent in detail the exact same adjustments and procedures followed for each stock in the sample. Projection Technioue Assume the follow-ing hypothetical data: Unadjusted reported earnings oer share (in dollars) 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 .60 .60 .80 1.00 1.20 1.40 .80 .90 .90 1.00 1961 1962 1963 1964 1965 1.20 1.30 1.20 .70 .75 Unadjusted dividends per share (in dollars) 1963 1964 1965 1.00 .50 .50 The hypothetical decision dates as determined by the close of the corporate fiscal year are March 31, 1963 (investment decision date), and March 31, 1965 (sale date). Associated unadjusted market values for the two months surrounding these decision dates follow: Market prices from exchange transactions: Range for the month of March, 1963$ 55.00 - $40.00 Range for the month of April, 1963 45.00 - 40.00 Range for the month of March, 1965 25.00 - 23.00 Range for the month of April, 1965 26.00 - 22.00 Time flexibility data: Lowest price in March and April, 1963$ 40.00 Highest price in March and April, 1965 26.00

The foregoing data are typical unadjusted data which were gathered for each share in the study. The only exception is the availability of earnings per share figures for the full twelve year period 1951 through 1962. This information was available for only forty-one of the stocks. The above data represent, therefore, the maximum that was processed for each share.

In addition to the data above, assume the following facts:

1. There were two separate two-for-one splits in the common

stock. The first took place at the start of 1957 and the second took place in the investment period at the start of 1964.

2. There were no other changes in capital.

3. There were no transactions in Treasury Stock during the entire fifteen year period.

4. There were no material changes in accounting procedure during the entire period.

Earnings per share

Because of the two stock splits, adjustments are required so that

the earnings trends that are to be established will be valid. The base date to which all data are made consistent is -965. The two-forone split in 1964 will require all previous earnings per share figures to be divided by two. The split in 1957 also requires . :-ings per share figures prior to 1957 to be divided by two. Hence, all earnings per share figures prior to 1957 must be divided by four, divided by two once to give effect to the 1964 split and again to give effect to the

1957 split. The adjusted earnings per share figures are presented below:

Adjusted earnings per share - Base date, 1965:

1951 1952 1953 1954 1955 1956 1957 1958 1959 1960

.15 .15 .20 .25 .30 .35 .40 .45 .45 .50 1961 1962 1963 1964 1965

.60 .65 .60 .70 .75

The above earnings per share figures represent the earnings which were attributable to the equivalent of a single share held at the end of the base year, 1965 (the termination of the investment period). Investment cost

The raw data have indicated the price ranges surrounding both the investment decision date as well as the sale date. In accordance with the averaging process described above, the mid ranges of these monthly

price ranges are averaged. The investment cost figure for this hypothetical share (before any adjustments required for share equivalent

values greater than 1) is determined to be $45.00 by the following process. Mid range, March, 1963$ 47.50
$90.00 = 8 4,5.00 Mid range, April, 1963 42.50 2 3 90.00 Share equivalents Since there was a stock split within the investment period, there will have to be a further adjustment to reduce investment cost so as to be consistent with the adjusted projected earnings. If the cost figure were to remain at the averaged raw figure of$ 45.00, it would reflect the cost of anticipated earnings attributable to a share held on March, 1963, whereas anticipated earnings have been adjusted to be consistent with earnings attributable to a share held in March, 1965. Averaged investment cost must be divided by the share equivalent value. Since two shares will be held at the end of the investment period instead of the single share purchased, the share equivalent value in this case will be 2. Investment cost used for projection of price-earnings ratios is $22.50. Twelve year projection The trend extrapolated earnings projection will make use of a linear least squares trend line with a projection of past trends into a future three year period (1963 through 1965). The first trend equation 7 is determined from earnings data from the period 1951 through 1962. 7The following is a derivation of the values for a and b in the trend equation used in the extrapolation. Year X EPS XY X2 Year X EPS XY X2 1951 -ll .15 -165 121 1957 1 .40 40 1 1952 - 9 .15 -135 81 1958 3 .45 135 9 1953 - 7 .20 -140 49 .959 5 .45 225 25 1954 - 5 .25 -125 25 1960 7 .50 350 49 1955 - 3 .30 - 90 9 1961 9 .60 540 81 1956 - 1 .35 - 35 1 1962 11 .65 121 4.45 1,130 572 a = Sum. Y L.45 = .370833 b = Sum. (XY) .019878 n 12 Sum. X 572 Stephen P. Shao, Statistics for Business and economics (Charles E. Merrill Books, Inc., 1967), pp. 521-530. Y = a / b (x) For projected 1963 earnings: Yc : .370833 / .019S78 (13) Y - .629242$ .629242

For projected 1964 earnings:

Y = .370833 / .019878 (15)
C
YVc = .668997 .668997 $2.006992 =$ .668997

For projected 1965 earnings: 3

Yc = .370833 / .019878 (17)

Yc = .708753 .708753

A 2.006992

Hence, $.668997 is the projected average yearly earnings per share for the period 1963 through 1965 as determined from an extrapolation of adjusted earnings per share data from 1951 through 1962. The number of stocks involved in this particular extrapolation in the study itself is forty-one. The other trend periods (1956-1962 and 1960-1962) will have the total one-hundred-ninety-eight stocks. It will be noted that the average projected earnings per share figure is the same as that for the second year. This is due to the use of a straight line trend for the establishment of both the average as well as each individually. A straight line is used to set values for three equally spaced time periods in the future. Naturally the second value will equal the average of the first three. Therefore, we need compute only the second year projected earnings in order to arrive at the desired average of the three future years. The average anticipated earnings per share will be placed over the adjusted investment cost and the result of this procedure will be the indication of the projected price-earnings ratio or the decision criterion. Seven year projection The seven year period refers to 1956 through 1962. We shall generate projected price-earnings ratios from the extrapolation of the trend of earnings in these years for the entire sample of one-hundredninety-eight stocks. The derivation of the projectedarnings in 8 the hypothetical case follows. For projected 1964 earnings: Yc a / b (x) Y = .485714 / .048214 (5) Yc .726784 Since we need compute only the projected earnings per share for the second year, the average of the three future years being the same, the above calculation is all that is needed. Again, this projected earnings figure will be placed over the.adjusted investment cost to determine the projected price-earnings ratios on the basis of which the stocks will be rank correlated. 8The following is a derivation Year X EPS 1956 -3 .35 1957 -2 .40 1958 -1 .45 1959 0 .45 1960 1 .50 1961 2 .60 1962 3 .65 3.40 a = w. Y 3.40 .485714 n 7 Stephen P. Shao, pp. 521-530. of the XY -1.05 - .80 - .45 0 .50 1.20 1.95 1.35 values for a and b above: X2 9 b = .048214 sum.X2 1:35= .048214 28 Three year projection All stocks of the sample will also have a three year trend period as the basis for a short run projection of recent earnings into the immediate future. The projected earnings per share amount from, the three year trend period in the hypothetical case is determined as 9 follows. For projected 1964 earnings: Yc a b (x) Yc .583333 / .0750 (3) Yc = .808333 The average annual projected earnings per share for the three year period 1963 through 1965 as determined from the extrapolation of the earnings trend of the three years immediately preceeding is$ .808333. Actual average earnings per share

The actual average earnings per share is determined simply by placing the sum of the factual earnings per share (adjusted) over 3. In the hypothetical case, this average actual earnings per share amounts to $.683333, the average of$ .60, $.70, and$ .75.

In summary then, we have projected the average earnings per share

figures anticipated in the period 1963 through 1965 from three different

9The following is a derivation of the values for a and b above: Year X EPS XY X

1960 -1 .50 -.50 1
1961 0 .60 0 0
1962 1 ,5 .65 1
1.75 .15 2
a = Sum. Y = 1 = .583333 b = Sum. (XY) .15 = 750
n 3 SumX2 -2 " Stephen P. Shao, pp. 521-530.

past periods of time, each period progressively shorter. The three projections are summarized below.

Trend period Projectcd EPS Actual EPS

1951 - 1962 $.668997$ .683333 1956 - 1962 .726784 .683333 1960 - 1962 .808333 .683333

It will be of interest to know the extent to which actual earnings approximate the projected earnings. In the main body of the study percentage figures will be calculated which are designed to determine the degree of accuracy obtained with such projections. Actual earnings will be compared with projected earnings without any other procedure involved. The percentage figure will represent the extent to which actual earnings conform to the projection. The data for the hypothetical case follow.

Trend period Projected EPS Actual EPS Percentage Actual to Projected

1951 - 1962 $.668997$ .683333 102.09 1956 - 1962 .736784 .683333 93.95 1960 - 1962 .808333 .683333 84.53

The important percentage will, however, not be the individual percentage for each stock but the mean of percentages for all stocks for each of the three trend periods. It will be of great interest to know what the mean percentage actual earnings per share is to projected earnings per share from each trend period. The determination of the percentages for each stock is determined as illustrated above. Projected price-earnings ratios

The projected price-earnings ratios are found by placing the

projected average earnings per share over the adjusted investment cost.

For each stock, then, there will be three separate price-earnings ratios, corresponding to each of the three trend periods. There will be a priceearnings ratio determined on the basis of the extrapolation of the earnings trend established in the period 1951 throUgh 1962 and for the other two trend periods also. In most cases, however, this will mean only two projected price-earnings ratios because the first trend period was established for only forty-one of the stocks of the sample. A data

constraint caused this limitation. In the hypothetical case, the price10
earnings ratios are as follows.

Trend period Projected EPS Investment cost PE Ratio

1951 - 1962 $.668997$ 22.50 .0292 1956 - 1962 .726784 22.50 .0324 1960 - 1962 .808333 22.50 .0360 Actual price-earnings ratio

The actual price-earnings ratio, which will be rank correlated with the projected price-earnings ratios, is determined in a similar manner by using actual earnings per share instead of projected earnings per share. For the hypothetical case, the actual PE Ratio is .0304 determined as follows.

Actual earnings .683333 304
Investment cost $22.50 Evaluation Technique The data above have been illustrated in hypothetical form for the purpose of conveying an understanding of the basis of the asserted decision criterion of shareholders. Investors are thought to make use 1Unadjusted average investment cost is$ 45.00. Placing this
over the share equivalent value of 2 yields $22.50 for average adjusted investment cost. 77 of the past earnings of a corporation in the form of making these data the basis for an anticipation of the earnings to be realized in future years. The extrapolated trends have been established in order to provide a basis for an array or listing in the order of highest to lowest projected earnings as a percent of investment cost. From such an array, the asserted utility of income computations might be tested. The question is can prospective investors make rational and good investment decisions if that decision base is the relative attractiveness of projected earnings as established from the extrapolation of the trend of past earnings. An array of projected price-earnings ratios rank correlated with a similar array of actual price-earnings ratios will indicate the extent to which the investment criterion factually turns out to be what was thought of it. Yet, this is still not the whole picture.. Even if actual price-earnings ratios were susceptible to projection, it will not mean that the prime assertion under investigation is supported. One-must be able to predict the relatively better cash return stocks from this basis, not simply those stocks which will turn out with the better earnings performance. Earnings is one thing, cash return is another. The technique for determining cash return is illustrated below. It will be recalled that dividend figures for the three years of the investment period as reported in-the financial statements (unadjusted) were$ 1.00, $.50, and$ .50 per share for 1963, 1964, and 1965 respectively. To determine cash return it will be necessary to collect all dividends which were attributable to a share held on the investment decision date, not the dividends attributable to a share held at the end of the investment period. This same problem will cause an apparent

lack of consistency between the investment cost used for projection and the investment cost used for determining cash return. It will be noted that it is of no consequence that the figure used for investment cost for projection differs from that used for evaluation. The

data do need to be internally consistent within the ranked arrays. The separation between the rankings, however, is complete. Once the data have been refined and adjusted to provide a basis for the ranking in the order of relative attractiveness of projected earnings as

qualified by investment cost, the investment decision has been made. So long as the evaluation criterion actually ranks those same stocks in the order of the relative attractiveness of factual perfortance, there is no compelling reason for the data to be in agreement as to a single investment cost. The problem arises because of the necessity of carrying one single share through several processes of trend determination and price-earnings projections. Had there been no stock dividends or splits, there would be only one investment cost for all computations.

The determination of actual cash return (the evaluation criterion) is illustrated below.

Cash Return

Capital Gain
11
Sale value of the investment $48.00 Cost of the investment 45.00 Capital Gain$ 3.00

llSale value of the investment constitutes the proceeds received from the sale of all interests arising from the single share purchased on the investment decision date. The average market price must be doubled at the sale date because there are two shares (due to the 1964 split) ht the sale date for the one share which was purchased.

79

Before the stock split

Received in 1963 $1.00 After the stock solit Received in 1964$ .50

Received in 1965 .50

Share eoaivalents X $1.00 2.00 Dividends Received$ 3.00

Cash Return 6.00 Actual rate of return on investment

The determination of the actual rate of return on investment is accomplished by placing cash return over averaged (unadjusted) investment cost. In the evaluation criterion section, there is no necessity of requiring all data to be consistent with a share held at the end of the investment period.

Cash return $6.00 Cost$ 45.00 .1333 or 13.33 percent Cash return with time-flexible decisions

To accomodate possible questions as to the arbitrary selection of decision dates, it was decided ea..lier to develop data with respect to the validity of the prime assertion when the "good time to buy" and "good time to sell" behavior observations were operative. In this case, the projection of future earnings is the same as before. The only change is in the evaluation criterion. Now, the determinate of cash return will allow the purchase price to be the lowest possible prevailing market price during the two month period surrounding the investment decision date. The sale value will be the highest possible

S80

market price prevailing in the two month period surrounding the sale date. The originally assumed hypothetical data indicated a low price in the two month period surrounding the puchase date at $40.00 and a high price in the two month period surrounding the sale date at an unadjusted$ 26.00. Since two shares are held at the end of the investment period for the one originally purchased at the investment decision date, the $26.00 will have to be doubled, which yields a sale value for the investment of$ 52.00. The computation of the rate of return

allowing for time flexible decisions follows.

Cash Return

Capital Gain

Sale value of the investment $52.00 Cost of the investment 40.00 Capital Gain$ 12.00

Same as before 3.00

Cash Return $15.00 Rate of cash return Cash Return$ 15.00
Cost 40,00 = .375 or 37.5 percent
Cost 4 0.00

Summary

The above illustrations reflect the procedures undertaken to

gather relevant data for the rank correlations which will follow later in the study. A summarized statement of those rank correlations follows. All elements represent arrays stated in the order of highest to lowest values. It will be recalled that the 1951 - 1962 trend period has a sample size of forty-one whereas all others have a value of one-hundred-ninety-eight.

Projected earnings per share:

From 1951 - 1962 trend rith Actual EPS 1963 - 1965 From 1956 - 1962 trend with Actual EPS 1963 - 1965 From 1960 - 1962 trend with Actual EPS 1963 - 1965

Projected price-earnings ratio:

From 1951 - 1962 trend with Actual PE ratio

From 1956 - 1962 trend with Actual PE ratio From 1960 - 1963 tron- with Actual PE ratio

Projected price-earnings ratio:

From 1951 - 1962 trend with Rate of cash return From 1956 - 1962 trend with Rate of cash return

From 1960 - 1962 trend with Rate of cash return

Projected price-earnings ratio:

From 1951 - 1962 trend with Rate of cash return
Time flexible

From 1956 - 1962 trend with Rate of cash return
Time flexible

From 1960 - 1962 trend with Rate of cash return Time flexible

In addition to calculating the coefficients of rank correlation for all stocks in accordance with the above procedure, there will be an additional full rank correlation for each array, but instead of including all stocks, the stocks will be industry grouped. In this manner, evidence will be forthcoming relevant to the validity of the second belief in income on an intraindustry basis.

If earnings of future years are susceptible to projection from the established trends of the past, then the rank correlation coefficients for the projected earnings per share-actual earnings per share

correlation should be quite high. If the future earnings do not follow the established trends of the past, a low or negative rank correlation coefficient should be forthcoming. If decisions based on the relative attractiveness of projected earnings as a percent of investment cost are valid, a high rank correlation coefficient should be indicated for the

analysis which rank correlates projected price-earnings ratios with actual cash return. The details of the rank correlation procedure are outlined below. Only one rank correlation will be demonstrated since

all are procedurally identical.

For the illustration, projected price-earnings ratios will be used

and rank correlated with actual cash return. To have the data needed for the rank correlation analysis, the data obtained in the hypothetical stock above will be supplemented with assumed data for nine additional stocks. The assumed data follow.

Projected price-earnings ratios Actual rate of cash
From trend 1951-1962 return

Stock Number 1 .0501 .1795 2 .0312 .1206 3 .0292 .1333 4 .0280 .1222 5 .0215 .1262 6 .0210 .1200 7 .0205 .1010 8 .0200 .0937 9 .0198 .0998 10 .0176 .0930 Rank correlation analysis does not use a relationship between the values above for each share. The relationship indicated in a rank correlation coefficient is a relationship between the assigned numbers for the stocks. The two arrays for the rankings follow. The hypothetical stock in the illustrations in this chapter is stock number .3.

83

Projection Cash Return
Rallnking* Ranking
(Y)() (Y-X) (Y-X)2

1 1 0 0 2 3 -1 1 3 5 -2 4 4 4 0 0 5 2 3 9 6 6 0 0 7 7 0 0 8 9 -1 1
9 8 1 1
10 10 0 0

The formula for rank correlation is

rk = -n(n2 2 ) 6(162) = 1 - 0.146 = 0.90301

The coefficient of rank correlation in the above case is .903, which would substantially uphold the validity of the prime assertion. An answer of 1.0 would have indicated perfect positive rank correlation and an answer of -1.0 would have indicated perfect negative rank corre12
lation. The rank correlation above indicates that if stocks were chosen on the basis of projected price-earnings ratios which were determined from an extrapolation of the trend of past earnings, the investment decision would be a good one. The predictive potential of past earnings has been established for the hypothetical case. This, however, is a dry run. Chapters IV and V will deal with the outcomes when actual data are treated.

12The rank correlation formula is derived from Stephen P. Shao, pp. 694-697.

CHAPTER IV

PREDICTION OF EARNINGS

By systematically dismissing the possibility of dealing effectively

with a critique of income reporting through recourse to the degree of

correspondence between conceptualized "essences" of income and operational

computations of the accountant, the burden of justification for the

reporting of financial net income to third parties was placed on what

may be termed "outcomes." That is, if net income computations are

to "make sense" in the context of the annual report in a manner consistent with the assertions which are made of it, such justification must come from the demonstrated utility of net income computations in shareholder decisions. It is not possible to defend income reporting by recourse to an assertion of correspondence between the accountant's computation and the co-called concept of "economic income." Confusion in this area was hypothesized to have stemmed from an exclusive reliance on "concepts" in income theory. Operations were directed by concepts in blind concordance and the quality of such operations was determined to rest entirely on the closeness of the operations to those Vhich were suggested by the formulated concept. Concepts were assigned the

task of determining the qualitative status of operations, and not vice versa. Ostensibly the only quality requirement for the concept came by

way of general agreement that the concept was indeed a "true" and reliable statement of observation.

With the burden of proof, as it were, of the advisability of

reporting income figures to shareholders shifted from purely conceptal argument to correspondence betwoon asserted use and factual indications as to whether the assertion is valid, a concrete set of tests mirat be undertaken. When the indications are analyzed, there will be a data-based indication of the warranted assertability of the claim that is made of income reporting. The basic question now is, does

income reporting factually do what is claimed of it.

There are many such claims of the utility of income reporting.

The one thought to be the most frequently encountered and the one which appears to represent a consensus of importance was isolated for study

in this analysis. Essentially, this claim asserts the utility of past income trends in predicting future income trends thereby lending to prospective investors valuable information on which to form an opinion on the investment potential of alternative investment possibilities in common stocks. This is the single claim under investigation here.

The extent to which income reporting assists credit grantors in their function is altogether another matter. Likewise, the utility of income reporting in determining the legality of dividends is not under review. We are interested in a critique of income reporting for shareholders.

Chapter III set forth the details of how this claim was to be

tested. There were two questions involved. The first was whether income amounts to be realized in future periods could, in fact, be predicted on the basis of the indications of trends in income from past periods. This chapter will deal exclusively with this first question. Chapter V will take up the question of whether the relatively better stocks

in factual performance will be chosen if this decision base is used.

We must first know, however, whether earnings can be projected. Without an affirmative answer to this question, the primary question of decision efficiency will have a rather dismal prospect. Yet, an affirmative answer to this question at hand need not indicate any relevance of past earnings to the choice of good investments in common stocks. Earnings may be precisely predictable without having any relevance to the decision base for investment. There is reasonable ground to suspect that this is indeed the case. If earnings were susceptible to projection, and every investor had access to this same information, would not the

market price of the attractive stocks be expected to be bid up or down consistent with an acutal cash return equating itself to common interest in the long run? This would appear to be a reasonable expectation. Moreover, many investors readily conceed that the prospects of a certain stock are determined more on the basis of what the average investor's opinion is expected to be rather than on the basis of what the operating efficiency of the related corporation is expected to be. Also, if an investor is in the market for short term capital gains, there is serious doubt that financial statements will be of any use whatsoever. If he is in the market for longer term growth and sustained dividends, there would be some question as to the relevance of past income trends in that changes in technology are taking place so rapidly that the technological advantage (which in many cases yields an economic advantage) is not held for very long. Future income performance in these cases would be thought to be determined more on the basis of a constant research effort than on the basis of the continuance of economic advantages held in the past. These are but a few of the reasons why the prime assertion is thought to be generally somewhat overstated.

We will initially set forth what the data indications are with respect to the predictability of earnings fcr all stocks. Secondly, we will make identical analyses with stocks grouped in accordance with the established industry classifications. All Stocks Analysis

Rank correlation coefficients obtained from two arrays on the basis of the relationship between projected earnings per share and actual earnings per share are presented below. The tests of significance were run via an estimation of the population rank correlation coefficient at a confidence limit of 95 percent. On the assumption that the basis of sample selection did not contribute a material bias to the analysis, the indicated population rank correlation coefficients are stated with 95 percent confidence. That is, there are 95 chances out of 100 that the population rank correlation coefficient will fall between the upper and lower indicated interval limits.

TABLE I
Projected Earnings Per Share Compared With Actual Earnings Per Share

Trend Sample Rank Correlation Population Population
Period Size Coefficient Upper Limit Lower Limit

51 - 62 41 .556625 .655625 .457625 56 - 62 198 .327217 .466769 .187665 60 - 62 198 .432938 .572490 .293386

The rank correlation coefficients do not indicate a very good

degree of relationship between projected earnings per share and actual earnings per share. The best coefficient as well as the best population interval estimate is obtained from the trend data from 1951 -

1962. The general observation sometimes encountered to the effect that the longer the length of past earnings periods, the greater the weight to the projection appears to hold little merit. The longer of the two remaining trend periods, however, has the poorest record of all. The best is the 1951 - 1962 trend period and the worst is the 1956 - 1962 trend period. With maximum population rank correlation estimate, the coefficient is .655625, which appears to be indicative that the statements of assertion heretofore made with respect to the predictive potentials of income trends are, perhaps at best, only vaguely correct. There does not appear any reason to believe that such trends have the so-called obvious relevance to shareholder decisions. Moreover, since we faced a data constraint when trying to gather sufficient data to determine these rank correlation coefficients for the 1951 - 1962 trend period, it would appear quite reasonable to believe that prospective shareholders in general are also constrained y this lack of data. At best, they would have been able to gather sufficient data for only forty-one of the major American corporations. The importance of this highest coefficient is diluted because of the lack of ability to put the relationship to practical use. Yet, all rank correlation coefficients are positive; indicating some degree of correlation. The question now becomes, are we to presume that that degree of correlation which is evidenced is sufficient to make rational and good investment selections from among common stocks. With the positive correlation coefficients demonstrated to exist, can an investor make good decisions? Is the relationship positive enough?

Clearly, an investor would not invest soley on the basis of the anticipated earnings. A share of stock generally costs something.

Consequently, investment decisions, to the extent that they are made on the basis of predicted earnings from past trends, would seem to be made on the basis of projected price-earnings ratios. In other words, qualify the future projection of earnings by current investment cost and you have the hypothesized decision base. The attractiveness of an investment alternative is not a simple matter of determining the projected earnings of that investment. An out-of-line investment cost would surely diminish the attractiveness of even the highest projected earnings.

Table II will demonstrate that when the added complexity of investment cost is included in the analysis, the rank correlation coefficients deteriorate still more.

TABLE II
Projected PE Ratio Compared With Actual PE Ratio
Trend Sample Rank Correlation Population Population Period Size Coefficient Upper Limit Lower Limit

51 - 62 41 .420693 .519693 .321693 56 - 62 198 -.007269 .132283 -.146821 60 - 62 198 .184763 .324315 .045211

With confidence limits again set at 95 percent, the indicated

maximum population rank correlation coefficient is found to be .519693. The lower limit is -.146821. The projection of price-earnings ratios would appear to be the rational basis for decision if past earnings

trends are thought to indicate things to come.. With a sample maximum correlation of .420693, the least which might be said is that for the one-hundred-ninety-eight stocks of the sample, there is very little

evidence that investment decisions made on the basis of extrapolated past trends (qualified by investment cost) are as valid as they are thought to be. Moreover, in the shorter period (1956 - 1962), the sample rank correlation coefficient is negative. This is evidence that the prospective investor is better off without this so-called obviously relevant information.

Since the ostensible purpose of arranging projected price-earnings ratios in the order of highest to lowest is so that the better "looking" stocks will be grouped toward the top of the array, it might be interesting to see what the relationship between the five highest and five lowest factual stocks is in comparison with how they were predicted to turn out. The following table will assign numbers one through five to the highest factual stocks in predicted earnings per share and the last five digits for the lowest in projected earnings per share (37 through 41 when "n" 41, 194 through 198 when "n" = 198). Table III will indicate these relationships for earnings per share figures and Table IV will indicate them for price-earnings ratios.

TABLE III
Actual Rank Compared With Predicted Rank - Earnings Per Share

Actual Top Five Actual Bottom Five
In Rank Predicted In Rank Predicted
Highest to Lowest Rank Lowest to Highest Rank

1951 - 1962 (n = 41)

1 31 41 40 2 3 40 41 3 15 39 38 4 8 38 39 5 6 37 4

TABLE III Continued
Actual Rank Compared With Predicted Rank - Earnings Per Share
Actual Top Five Actual Bottom Five
In Rank Predicted In Rank Predicted
Highest to Lowest Rank Lowest to Highest Rank 1956 - 1962 (n: 198)

1 7 198 17 2 196 197 27 3 41 196 193 4 9 195 188 5 50 194 76

1960 - 1962 (n : 198)

1 56 198 39 2 1 197 12 3 114 196 196 4 179 195 169 5 120 194 167

In the 1951 - 1962 analysis, the best that might be said is that the five factually highest stocks were, with one exception (stock number one), all within the top half of the predicted rankings. Likewise, the five least favorable stocks in factual earnings performance, with one exception (stock number thirty-seven), were all in the bottom half of the predicted rankings. In the 1956 - 1962 analysis, this record deteriorates to one exception in the top five and three exceptions in the bottom five. In the 1960 - 1962 analysis, it is worse still. In this time period, there are three exceptions in the top five

and two exceptions in the bottom five. These data indicate that earnings of past periods, particularly when projected and used in interfirm projections of relative attractiveness of anticipated earnings, clearly do

not have the predictive potential that they are commonly presumed to have.

Another, and perhaps an even more interesting analysis, is to see

how well the upper 25 percent in predicted earnings factually turned out. Data indicate that if one were to choose from among the top 25 percent of stocks as determined by an array in the order of highest to lowest projected earnings per share, the 1951 - 1962 trend period would have rendered to such an investor a 50 percent chance that the chosen stock would turn up in the factually highest 25 percent. Of the ten stocks (25 percent of 41 = 10) with the highest predicted earnings per share, five were among the factually highest ten.

In the 1956 - 1962 trend period, this percentage drops to 40 percent. Only 40 percent of those stocks predicted to be in the upper 25 percent actually were in the upper 25 percent. Again, as the time period shortens, the relationship becomes worse. In the 1960 - 1962 trend period, this percentage drops to 36 percent. Of the top 50 stocks (25 percent of 198 50) predicted to have the highest earnings per share, only 18 of them turned up in the factually highest 25 percent of stocks.

The above indications of relationships represent a measure of the effectiveness of predicting future income amounts based on the extrapolation of past trends. When the relevant factor of investment cost is added, the presumed basis of investment decision is determined. Table

II indicated the rank correlation coefficients and population estimates based on intervals. It would be interesting to know how the top and

bottom five stocks performed in this category. Table IV contains this information. It is organized the same as Table III except that the relationships are between projected and actual price-earnings ratios instead of earnings per share.

TABLE IV
Actual Rank Compared With Predicted Rank - Price-Ernings Ratio
Actual Top Five Actual Bottom Five
In Rank Predicted In Rank Predicted
Highest to Lowest Rank Lowest to Highest Rank
1951 - 1962 (n = 41)
1 32 41 40 2 6 40 37 3 9 39 41 4 3 38 1 5 10 37 8 1956 - 1962 (n = 198)
1 1 198 198 2 196 197 197 3 29 196 22 4 64 195 186 5 94 194 193 1960 - 1962 (n = 198)
1 162 198 193 2 5 197 176 3 42 196 4 4 82 195 195 5 76 194 74

For the most part the projection of the relatively poorer stocks appears to be quite reliable. It is unfortunate that investment decisions cannot be made on this basis. The performance of the predicted top five stocks in the 1951 - 1962 trend period appears to be the best of the three periods. In each period, there is only one exception to

the rule that all predicted top five stocks factually turned up in the top half of the factual performance array. This is the most favorable comment which might be made. It is doubtful that this would offer any great reassurance to a prospective shareholder considering an investment. If there are other indicators of good investment stocks, and this extrapolation of past income trends appears to support those indicators, then perhaps this methodology (and hence, a qualified version of the prime assertion) is useful in such cases. The indications of this part of the study, however, reflect evidence that to the extent that such a process as this is considered, the consequence should not be expected to be too favorable. For the most part, it would seem reasonable to suspect that there would be other factors quite a bit more important in investment decisions than the relative quality of past income trends. The evidence above is interpreted as indicating that however small a weight is assigned to the quality of past income performance, that segment of the investment decision can be expected to be at best only vaguely sound.

It would be interesting once again to see what the chance would be for having a stock in the upper 25 percent of the predicted array turn up in the upper 25 percent of the factual array. Of the ten stocks with the highest predicted price-earnings ratio, five turned up with the factually highest 25 percent, a percentage of 50. For the trend period 1956 -

Full Text

PAGE 1

A CRITIQUE OF INCOME REPORTING FOR SHAREHOLDERS Â•'Â•,V. " . ~ -V FREDERICK DOZIEil WHITEHURST A DISSERTATION PRESENTED TO THE GRADUATE CX)UNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1968

PAGE 2

Copyright byFrederick Dozier Whitehurst 1968

PAGE 3

To Delmas D. Ray i

PAGE 4

TABLE OF CONTENTS Chapter Page I. INTRODUCTION 1 U. THE ACCOUNTANT'S APPROXIMATION OF TRUE INCOffi 14 in. STATEMENT OF STATISTICAL PROCEDURE 43 IV. PREDICTION OF EARNINGS 84 V. PREDICTION OF CASH RETURN 101 VI. CONCLUSIONS 118 Appendix I. STOCK IDENTIFICATION 128 II. ADJUSTED EARNINGS PER SHARE INPUT DATA 134 in. K\RKET VALUATIONS INPUT DATA 141 IV. OUTPUT AND RANK CORRELATIONS INPUT 146 BIELIOGPuAPHI 154 iv

PAGE 5

LIST OF TABLES Table Page I. Projected Earnings Per Share Coinpared With Actual Earnings Per Share S7 II. Projected PE Ratio Compared mth Actual PE Ratio 89 III. Actual Rank Compared With Predicted Rank Earnings Per Share 90 IV, Actual Rank Compared With Predicted Rank Price-Earnings Ratio 93 V. Mean and Standard Deviation Actual Earnings Per Share as a Percent of Projected Earnings Per Share 96 VI. Rank Correlation Coefficients Actual Compared With Projected Earnings Per Share Stocks Grouped by Industiy 98 VII. Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation Earnings Per Share 99 VIII. Projected Price-Earnings Ratio Compared With Actual Cash Return 103 IX. Actual Rank of Cash Return Compared With Predicted Rank of Price-Earnings Ratio 106 X. Projected Price-Earnings Ratio Compared With Actual Cash Return Utilizang Time Flexible Investment Decisions 109 XI. Actual Rank of Cash Return Compared With Predicted Rank of Price-Earnings Ratio Utilizing Time Flexible Investment Decisions 110 XII. Rank Correlation Coefficients Predicted PE Ratio Compared VJith Actxial Cash Return Stocks Grouped by Industry 112 V

PAGE 6

Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation Predicted Price-Earnings Ratio Rank Correlated With Cash Retvim Rank Correlation Coefficients Predicted PE Ratio Compared VJith Actual Cash Return Utilizing Time Flexible Decisions Stocks Grouped by Industry Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation Predicted Price-Earnings Ratio Rank Correlated With Cash Return Utilizing Time Flexible Decisions vi

PAGE 7

TABLE OF CONTENTS Chapter Page I. INTRODUCTION 1 n. THE ACCOUNTANT'S APPROXIMATION OF TRUE INOCME 14 in. STATEMENT OF STATISTICAL PROCEDURE k3 IV. PREDICTION OF EARNINGS d4 V. PREDICTION OF CASH RETURN 101 71. CONCLUSIONS US Appendix I, STOCK IDENTIFICATION 128 II. ADJUSTED EARNINGS PER SHARE INPUT DATA 134 in. MARKET VALUATIONS INPUT DATA 141 IV. OUTPUT AND RANK CORRELATIONS INPUT 146 BIBLIOGRAPHY 154 iv

PAGE 8

CR^iPTER I INTRODUCTION The contemporary certified public accountant is involved vath many functions. Typically, he is concerned vath taxation matters, systems analysis, various management services, ani auditing. Paramount among these diversified functions is that of lending disinterested or independent attestation to the conformity of management's financial representations to that which has come to be termed "generally accepted principles of accounting." More precisely still, the certified public accountant is said to be vitally interested in the "fairness" of the representations of management to third parties including stockholders and creditors. He is vievred as the guardian of integrity in those financial representations which are inevitably associated va.th a conflict of interest. From the first public accounting statute to the present day, his function in society has been associated with terras such as "eq-uity," "justice," and. "fairness. " Integrity in financial reporting to third parties has come to be interpreted as the correspondence of procedures used with accepted principles of accounting which, in turn, im.ply a "fair" presentation. Financial presentations are, therefore, considered fair only so long as they are consistent with procedures which are in conformity with generally accepted treatment rules. These treatTCtBTit rules have as their major core of emphasis the income determination process. Kence, the principal raison d_|.etre of the certified public

PAGE 9

2 accountant is that of lending his attestation to the "fairness" of reported "income," Perhaps the most astonishLng feature of modern accoxmting is the fact that a whole profession has devoted itself, for the most part, to the task of deternaining income for an enterprise for a period of time and yet is at a total loss to arrive at a consensus as to the mearong or significance of that aggregate. Few would disagree that an LTiportant fxinction of the modem C. P. A. is that of "income determination." Fewer still would be able to relate a meaning for that notion on other than a procedural basis. Perhaps this is a failure of research effort. Perhaps it is due to the lack of applicability of scientific method to such "assertions." More fundamentally, perhaps the asserted phenom.ena of "income" in a real or true sense simply does not exist. If this is the case, there is little wonder that there is such confusion in economic and accounting realms as to the significance, definition, computation, meaning, verification, and utility of computed monetary income. .. _. A recent outcrop of a relatively large research effort in terms of both expenditure and effort by the American Institute of Certified Public Accountants suggests as an appropriate definition of net income the 1 ",,, increase (decrease) in owner's eqxaity, ..." This is substantially a duplication of the definition offered by Canning which reduced the significance of net income to that which vas left over after the Robert T, Sprouse and I^urice Xoonitz, A Tentative Set of Broad Accounting Principles for Business Snteirprise (American Institute of Certified Public Accountants, 1962), p. 9. This meaning of income is a recurrent theme throughout most of the official pronouncements of both major professional associations in accounting.

PAGE 10

2 accountant completed his procedures. If one understands "change in owner's equity" to mean the change brought about through the procedures of the accountant in closing the nominal accounts, then cet. par ., the two statements do indeed agree. Unfortunately, such definitions not only fail to indicate the meaning or significance of this computational aggregate, but worse still, they appear to be entirely appropriate. To our knowledge, the only meanings attributed to income in accounting literature are statements of definition on procedural bases. Why then, has income enjoyed a historical prominence in financial affairs? Why has enterprise net income been the building block of substantially an entire profession? ' There is evidence that profit_calculations started in the Middle Ages as a tool for the computation of appropriate cash distributions which would be made before the termination of partemership voyages mder3 taken as single ventvires. In modem times, income calculations have been propounded in a variety of ways, not the least of vAiich was the federal income tax. If tax is based on "income," there woxild naturally exist the tendency to explicitly formulate procedures for the determination of exactly what would constitute one's income. In an effort to minimize individual federal tax payments, the treasury department has been forced to formulate procedures and guidelines for the determination of procedural meanings for "income." Transfer this emphasis John B. Canning, The Economics of Accountancy (The Ronald Press Company, 1929), p. 98. ^Raymond de Hoover, "The Developraent of Accounting Prior to Luca Pacioli According to the Account-books of Medieval Merchants," in Studies in the History of Accounting , edited by A. C. Littleton and B. S, Yamey Taichard D. Irwin, Inc., 1956), pp. 115-116.

PAGE 11

4 to the corporate taxation area and the importance of not only the calculations of net income but also the procedural meanings of net income become greatly expanded. Perhaps the greatest single factor in the rise of income determination was the interpretation vAiich the co\irts made -with respect to corporate fiduciary relationships. Corporate officers have been included in the category of the fiduciary. They have the responsibility of the fiduciary vdth respect to those on whose behalf they hold a position of trust. Courts have interpreted this relationship as having the responsibility for publishing annual statements of corporate affairs. This includes the balance sheet and income statement. Some state corporation statutes require financial reporting by the corporate fiduciary. The national stock exchanges have had some influence in developing financial procedures which have as a result the statement of "net income," They have the authority to prescribe' reporting standards as a prerequisite for listing. Traditionally, such reporting standards include statements of financial position and income statements. This listing requirement is reinforced by the S.E.C. v^ich, through its regvLIations, requires the annual submission of balance sheets and income statements which have been audited by independent accountants reflecting "condition" and "earnings," The standard of acceptability of procedures used in such reports has been outlined by the S.E.C. under the term of "substantial authoritative support," This rule iiras first 4 established in 1934, at the very start of securities legislation. 4united States Securities and Exchange Commission, Accounting Seriss Releases (United States Government Printing Office, 1956), p. 5. This particular release (Release Nimiber Four of April 25, 193S) presumes financial statements to be misleading unless there is "substantial authoritative support" for the procedures employed.

PAGE 12

5 Reports filed \^dth the Securities and Kxchange Commission become public record and they are available for both inspection as well as duplication. A heavy reliance on income determination is evident with respect to the rate regulatory powers of many federal and state agencies, particularly those vdiose powers extend to the prescription of accounting procedures whereby the "rate of return" is policed. The Accounting profession itself has brought noticable attention to the concept of income in recent years. The American Accounting Association has published many works devoted to the analysis of income and related valuation topics. The American Institute of Certified Public Accountants has been active in the promulgation of "sound" accounting treatment rules ever since the formation of the Committee on Accounting Procedure and the issuance of its opinions vtoLch started in 193S. Currently, the Accounting Principles Board of the American Institute of Certified Public Accountants carries on the work started by the Accounting Procedure Committee. There is little doubt that the prescriptions coming forth from such organizations have as their primary aim the more accurate determination of financial position and operating results. Evidence of the pidmary position Â•vrfiich "income determination" holds in accounting is found in the literature of accounting. Paton is perhaps the foremost advocate of the central position of income. He states that income determination is "...the most significant and crutial phase 5 of the accountant's task," Professor Reed K. Storey has stated that ^'jilliam A, Paton, Advanced Accounting (Macmillan and Company, 1947), p. A.3S. See also William A. Paton, AccountiAg; Theory (Accovmting Studies Press, Ltd., 1962), p. 1^2. Most of the vn-itings of this "father" of modem accoxmting have centered around the development of a coordinated and consistent "theory" of accounting with income determination axd asset valuation its central functions.

PAGE 13

the determination of periodic net income "...is the most important 6 function of financial accounting." NotTsdthstanding such a heavy reliance on the conceptual notion of income, there is little evidence that the net income aggregate, to the fairness of which independent auditors lend their attestations, has an understandable meaning vdiich would facilitate (or even permit) a determination of -its relative fairness. The independent accoimtant attests to the fairness of such income calculations without any reference to a standard to i^ich a given net inccme figure might be compared for an objective determination of its degree of approximation, 7 nearness, or fairness. The fairness of managemait's representations Vj-hich have been determined by procedures which are consistent with generally accepted accounting principles appears to be a conclusion rather than a criterion. Financial position and operating results are "fair" only so long as they are determined by use of rules of procedure which have as the sole criterion of warrant their general acceptance. It would constitute more efficient communication if references to the "fairness" of position and operating results were omitted. The certifying auditor himself generally sidesteps this fairness issue by indicating in the same sentence that the balance sheet and the income statement present fairly the financial position and the results of operations in conformity v/ith generally accepted principles of account ins: ^Reed K, Storey, "Cash Movements and Periodic Income Determination," The Accomting Review , July, I960, p. 449. See also A. C. Littleton, "Concepts of Incons Underlying Accounting," The Accounting Review, I-Iarch, 1937, p. 22. 7 R. J. Chambers, "A Matter of Principle," The Accoxmting Review , July, 1966, p. 455. Professor Chambers insists that mthout such a staridard of comparison allusions to "fairness" are both univ'arranted as well as meaningless.

PAGE 14

7 applied on a basis consistent with that of the preceding year. If there is no agreenient as to the significance of income which is the direct product of the use of accepted accounting principles, and if the result of the use of such procedures is a "fair" presentation, then the inevitable conclusion is that "fairness" has no generallyagreed upon meaning either. This is particularly evident -v.-hen one considers the quality of such generally accepted accounting principles. They aopear so ambiguous to one author that he concludes that they have no definitive meaning. The accountant, then, attests to the balance sheet and income statement presentations as being procedurally consistent with these accepted rules. He concludes that net income has been determined in a "fair" manner. Yet, as Professor Chambers indicates, through the wide choice of equally acceptable alternatives in procedure, any independent auditor could certify the "fairness" of any one of something like thirty-million possible and equally acceptable net income figures for a large corporation with a sufficiently large diversifi9 cation of transactions. On this basis it would appear that the conclusion of Professor Canning in 1929 was not dated. After a careful study of the methodology of the accomtant he concluded that "...account 10 ants have no conrolete philosophical system of thought about income." George 0. Kay, "C-enerally Accepted Principles of Accounting," The Journal of Accomtancy , January, 1959, p. 24.. J. Chajnbers, p. 455. ^'^John B. Canning, p, 160. It i\ro\ild furthermore seem that to revert to statements of "fairness" which are implicit in the confonnity of financial presentations with generally accepted accounting principles constitutes a negation of the presumed professional attitude. It would seem that "fairness" should be more explicitly defined as constituting the conformity of presentations with that which has become accepted.

PAGE 15

The discussion above supports the observation that all is not well vd.th respect to the clarity of the accountants conceptual rigor. Yet, many accounting witers rely on such conceptual apparatus to support the actions of practicing accountants. Income does exist. It slTiply falls to the accountant the task of arriving at "income" as nearly as possible. Many assert that the income calculation of the accountant is clearly not exact. It is, however, thought to be close enough to the true income to be valid. This type argument represents through the observation of the author one of the two most popular approaches to asserting the propriety of determining accounting net income and reporting it to third parties in the annual report. In summary form this argument simply states that incone of an enterprise for a period of time does exist and the accountant goes about measuring it as closely as possible. The conceptual framework of "accounting theory" is said to be the guide for the operations of the practicing accountant, VJe must trace the meaning of the computed net income, therefore, backyards to the concepts that are said to be involved. In addition to this approach, many accounting writers offer the outcomes of the net income calculation as support for its propriety. Net income is said to be of use . VJhatever methodology is used for its computation, net income does represent a valid basis for the orientation of decision and action. It is said to have functional utility. Among the suggested applications of accounting net income is the popular assertion that the computed accovmting net income represents relevant information in the decision environment of the prospective shareholder. There are then, tvro basic assertions about accounting income. The first asserts the existence of enterprise net income and suggests that

PAGE 16

it is the accountant's task to determine the monetary expression of that net income as accurately as possible. The second asserts that computed accounting net income (as an outcane) is useful for decision. The decision of the prospective shareholder is thought to be the pidmary utiQjLty application of reported net income. In a poDular article vnritten in 1943, Professor George 0. lÂ«lay 11 suggested ten uses of financial statements: 1. As a report of stewardship 2. As a basisfor fiscal policy 3. As a criterion for the legality of dividends 4. As a guide to vdse dividend action 5. As a basis for the granting of credit 6. As information for prospective investors in an enterprise 7. As a guide to the value of investments already made 8. As an aid to government supervision 9. As a basis for price or rate regulation 10. As a basis for taxation Â•Â• ' . Item six above seem.s to ha.ve emerged as the most popular justification from the standpoint of outcomes for the currently reported net income figures derived from the application of "financial accounting theory." Decision relates to the future and financial net incane is thought to be a good basis on v;hich to m.ake investment decisions. Specifically, past income trends are asserted to have relevance in estimating the future trends or levels of income. This being so, the prospective investor is thought to benefit a great deal from the information disclosed in the past income statements of enterprises. Past income calculations are thought to be predictive to the extent of rendering to the investor a basis for an intelligent guess. George 0. May, "The Nature of tte Financial Accounting Process," The Accounting Review, July, 1943, pp. 189-190. Italics added.

PAGE 17

10 In this connection Professor Morrison states: As we have stated, the intelligent investor is interested fimdamentally in the earning power of the company.,,. This involves forecasting or budgeting because earning power involves the future. Unless the intelligent investor has quite complete and accurate information about the manner in which earnings were created in the past, his forecast vri.ll be a very poor guess. On the other hand, if he has a good historical record, he has a good starting point from vrtiich to forecast the future. Professor Robert T. Sprouse has summarized this popxilar viewpoint by saying in a recent paper: My paper is concerned with specifically the accounting measurements reported in the financial statements provided by publically held corporations to current and prospective investors financial statements about which certified public accountants render their professional opinion concerning fair presentation in confonuity with generally accepted accounting principles. These are the financial statenents v/hich the Securities and Exchange Commission, the New York Stock Exchange, the American Institute of Certified Public Accountants, the Financial Analysts Federation, and a significant portion of the membership of the American Accounting Association and other professional groups are most vitally concerned. It seems clear from the nature of the intended audience ard the nature of the interested professional groups that such financial statements are intended to provide information that is helpful in making rational investment decisions . ^3 The following statement by Professor Sprouse is presented as a representation of the consensus of accounting writers, in the opinion of the author, with respect to the primary utility of financial net income reported in the annual report to stockholders, . . The primary purpose of the msasurement of last year' s income reported to investors is to provide a basis for predicting future years ' incomeT l^ ^aul L. Morrison, "The Interest of the Investor in Accounting Principles," The Accounting Review, March, 1937, PÂ» 39. 13 Robert T, Sprouse, "The Measurement of Financial Position and Income: Purpose and Procedure," Research in Accounting Measurement , edited by R. K. Jaedicke, Y. Ijiri, and 0. Nielsen (American Accounting Association, 1966), p. 103. Â• ^Ibid . , p. 106. Italics added.

PAGE 18

11 In \hole, or in part, the investor is thought to be able to utilize inccane determination by looking at past trends in order to gain some insight into the probable future status of the enterprise under consideration. It is important to note that this claim is made of income fig:ures and not simply of financial data in general. Income is an interpretative aggregate vMch is thought to add something to unprocessed financial data Â•which woxild otherwise be lacking. Emphasis is placed not on the reporting of financial facts but on the interpretation of those facts in the form of net income. Historical financial facts could easily be presented in the annual report vathout requiring that they adhere to the income determination model. Factual statements of historical transactions together vdth selected current values for assets (if this ijnformation is fact'ually determined to be useful) could be presented vathout any references to the net income of an enterprise for a period. As it is, judgements as to -which alternative accounting treatjuent rule frcm among the many "principles" of accounting is the correct one under the circumrstances is made by reference to its anticipated consequences in the income determination process . It would appear, therefore, that the primary concern is not on the presentation of data but on the interpretation of these data. Net income represents an opinion as to how well an entity performed in the past (whether this opinion is a "generally accepted" opinion consistent with the dictates of collective authority or not is irrelevant). Hence, when the certifying auditor renders his opinion on financial statements, he is, in effect, saying that it is his opinion that the opinion expressed in the form of net income has been determined on a basis consistent with generally accepted accomting principles (an additional opinion). Essentially, the auditor is saying that it is my

PAGE 19

opinion that the opinion of past performance expressed as net income is "fair" in that my basis for a clean certificate rests on generally accepted opinion with which the reported net income figure is consistent. It appears to be unpopular to hold that net income is interpretative. Fiany writers claim what might be termed an existential status (non conceptual status) for income. Income is. It falls to the accountant the task of determining as closely as possible this "real" income amount. Here then, are the tv/o basic beliefs in the propriety of income reporting. One holds that income enjoys an existential status, perhaps much like gravity, or the behavior of oxygen. The other holds that income amounts in the annual report provide valuable information for shareholder investment decisions. One is thought to be able to predict the earnings of the future from the earnings trends of the past and thereby gain some measure of insight as to what to expect in the future. These two claims are, in the opinion of the author, the most frequently encountered assertions as to the value of income figures appearing over the opinion of certified public accountants. We interpret the function of a critique as that of determining the qualitative aspects of the correspondence between the assertions made of income utility and the data base which ostensibly supports these assertions. Perhaps one or both such beliefs are justified in the light of extended logical argument and in the face of relevant data. Perhaps also one or both such beliefs require too great a leap of faith to be held in the light of evidence. We shall offer a critique of income determination based on the warranted assertability of these two beliefs in income. Chapter II will deal with the first of these two beliefs. That is, it will examine the efficiency of tracing the significance of a reported

PAGE 20

13 net income figure back\%ards to the concepts ifjhich are said to direct the practicing accoiantants operations. Is there a meaning of income in "accounting theory" which directs operations in such a way as to insure the establishment of an income figure which has significance on other than procedural grounds. Instead of reiterating the claim, "income is," we ask the question "is there income" in the claimed "real," "ultimate," or "fundamental" sense. The remainder of the study (Chapters III through V) will deal with the second isolated belief. â€¢e will be interested in developing data of a relevant nature v/-hich will indicate whether and to what extent past income trends factually do indicate things to come. Yet, the susceptibility of future income amounts to projection from the extrapolation of past trends is not the central issue in this section. Vie have documented the belief that past income trends establish a decision base which is thought to be valid for the prospective shareholder. The real issue in this assertion of -utility for income is the extent to which the relatively better stock investment decisions will be made if the basis for selection is the relatively better projected income amounts as determined by the extrapolation of past trends. Chapter IH will set forth the methodology by which relevant data might be used to establish the assertability of this utility claim for income determination. Chapters IV and V will present the findings of this statistical study. Chapter VI will present the overall conclusions with respect to both claims.

PAGE 21

CHAPTER II THE ACCX)UNTANTÂ» 3 APPROXIMATION OF TRUE INCOME If, as some suggest, the role of "accounting theory" is to provide a conceptual "base" which might be used to "direct" operations of the practicing accountant, then the meaning and significance of the accountant's computations must be found in such conceptualizations. In genaral, no reference is made to "outcomes" or "consequences" vrtien it is assez*ted that the support for accountant's computations is to be found in "theory. Instead of tracing meaning forward to consequences, those who assert an exclusively directional function of theory suggest (most of the time unavoidably) that the significance of the accountant's statement of operating results (net income) is to be found in the concepts that are said to direct his actions. This argument was identified in Chapter I as one of the two most popular paths to belief in the propriety of the accountant's income figure. The phenomenon termed "income" is and the accountant meas\ires it as closely as it can be measured. It is important to note that in this argument there is no bvirden placed on income to be a utility device the quality of which must be settled upon by a determination of the consequences of its use in association with the asserted functions it has. The emphasis here is not on the utility of the income ccxnputation but on the correspondence between "guiding" concepts and the accomtant's operations in practice. Practice is justified on the basis of its conformity with the concepts and not on the basis of the utility potential of the net income aggregate.

PAGE 22

PAGE 23

16 Professors Paton and Littleton suggest the existence of a real, true, objectively detenninable, and ongoing income while George 0. May holds no hope for a referential base of income in "truth." Professors Paton and Littleton are associated vfith the argument that a meaning for income might be found in concepts whereas George 0. May is associated with the contention that income meanings are to be found in the practical application of the income aggregate, the consequences stemming from its use, or, in a word, percepts . The task at hand is the examination of the asserted concepts which are said to underlie the accountant's operations. The matter of "outcomes" and the asserted functional utility of financial income will be taken up in Chapter III. ^^Lth allusions to reality , truth, and ob.jective conditions, there is naturally some part to be played by scientific method in both the search for as well as the evaluation of income. All too often logical is presimed to be 'a synonym for scientific. It woTild seem that in the field of accoimting the word "scientific" as applied to the determination of the amount of corporate income has the meaning of a method of determination which is to the greatest degree possible, objective, unbiased, logical, and capable of being verified and tested by other observers or accountant 3.3 Yet, without any indication that there is a standard against which computed net income might be compared for a determination of its degree of approximation, this same author goes on to say that "...although such a determination can never be exact or precise, it is an approximation ^Arthur C. Kelly, "Can Corporate Incomes be Scientifically Ascertained," The Accounting; Review, July, 1951, p. 290. Italics added. Note should be made of the naive use of the term "tested" in the statement. It is in the use of this word that the others (objective, unbiased, and logical) take on the meaning intended by the author.

PAGE 24

to the truth," Hence, the financial net income is an approximation to the truth which is represented in the concepts directing the actions of the accountant. Truth is conceived and the accovmtant aims as closely as he can to state the truth, the "objective realities "which exist," in a manner consistent with the directive forces of the concept. The situation is this: A concept is settled upon and operations are undertaken to put the concept to use. ^Vhen the directing forces of the concept have shaped operations in such a way as to make them as consistent as possible vdth the concept, the process stops . The result is termed an approximation to truth , an approximation to reality , or an approximation to the objective conditions which exist . Since the operational result of the implementation of the concept is termed an approximation to truth, and since the operational result is also termed an approximation to the ideality of the concept, the ostensible reference to truth must mean the concept. The concept and truth are synonymous and to the extent that operations fall somewhat away from the ideality of the concept, they are only approximations to truth. Here lies one of the tvro basic beliefs in income, that it is true, that it is real, and that the only reason the accountant cannot state it exactly is due to the dilutions of the "pure theory" encountered in practice. Are these contentions sound? The first task is to state what these concepts are. A thoughtf\il forewarning to those who would take to deliberations on income the hope of arriving at a concrete irreducible meaning was issued many years ago 4 ibid . The popular notions on the relationship of absolute truth to approximations to truth as far as income determination is concerned are well summarized by Kelly. He says, "So far as accounting is concerned, it is well recognized that" there is no such thing as the absolute truth in a presentation of corporate net income. Every statement of income is merely an approximation to the truth, or to reality,,," Ibid .

PAGE 25

18 by Professor Frank ti. Knight, Knight suggested that no other concept in economics was used vdth a more bevdldering variety of well-established 5 meanings than profit or income. The wisdom of this caution is evident to those who are familiar with the variety of computational counterparts to conceptualized notions of income which exists in accounting theory. ]fle shall have to concern ourselves initially with those conceptualized notions of income which are said to be economic (as opposed to accounting or computational) in nature. These statements of conceived interaction (truth, in the argument of the proposers) generally represent vihat might be termed the more popular searches for "central" or "ultimate" meaning. Among these concepts, those of Professor John R, Hicks are perhaps the most frequently appropriated for use in accoianting income determination discussions. IVhy Professor Hicks' definitions of income seem so attractive to accountants is, of course, conject\ire. Perhaps quite a substantial case might be developed in favor of the contention that what accovintants actually do in practice appears to be less painful a dilution of the strict definition than the concepts of other writers. Somehow one cannot possibly imagine an accountant coming to grips with the Fisherian "psychic benefit" crystallization of meaning of income. Because of their popularity, we shall consider Hicksian definitions rather closely at first. We shall want to determine, if possible, the qualitative aspects of this "truth" even if the accountant could follow the dictates of the concept exactly. Might income be "pinned down" by tracing the significance and meaning of the operational aggregate "backvrards" to the bedrock of "truth" as it is claimed? 5 ^ ^^^^ ^* Kni^t, Risk , Uncertainty , and Profit (Harper and Row, 1965;> p. 22. See also the author's article on "Profit," Encyclopedia of the Social Sciences , 1933, p. 480.

PAGE 26

19 Economic Income The AccQ-untant ' s Guide Economic concepts of income are often attributed the quality of being a referential base to vAiich operational counterparts in accounting might be related, however diluted that operational expression might be. Many prominent writers in accounting suggest that these concepts of income provide an "ideal" toward -vAiich accounting income might tend to approximate. Professor Sidney S, Alexander, vdiile stipulating the limited appeal of any one economic income concept, goes on to bviild a theoretical framework utilizing the Hicksian central definition as 6 a clear starting point. Professor Willard Graham noted that, "Although the practical applicability of the economist's theory of income has been seriously questioned, it has value as an 'ideal' or 7 'yardstick' by vAuch to evaluate other theories of business income." Professor Graham went on to indicate that the economic theory of income 8 has pointed to certain shortcomings in the accountant's procedure. Such prescriptions suggest that economic conceptualizations are formulated in statement form. For the most part they are attempts to deal with the meaning of income on a constitutive basis. As a gviide, these concepts are settled upon as the narrative description of that v^aich their comterparts in accounting set out to make operational. Concepts are thereby used in the sense of guiding operations in what might be termed "blind concordance." No burden is placed on such operations for establishing the assertability of the conceptualized indication of "truth." Sidney S. Alexander, "Income Measurement in a Dynamic Economy," Studies in Accounting Theory, edited by V/. T. Baxter and S. Davidson (Irvdn, 1962), pp. 127-146. '''vfi.llard J. Graham, "Some Observations on the Natxire of Income, Generally Accepted Accounting Principles, and Financial Reporting," Law and Contemporary Problems (School of Law, Duke University, 1965), p. 654. %id.

PAGE 27

. .> . 20 Once the operational comterpart has performed the indicated procedure, the process stops . If the processes of "theory formulation" in both economics as well as in accounting are properly viewed as knowledgeacquisition processes, then this interpretation of the fxmction of concepts would seem to be somevrhat inefficient. Note that there is no place in this notion of the idealness of economic income for operations designed to establish the \irarranted assertability of hypotheses. Note that in the conceptualization processes there is no place for indicated if-then relationships toward which operations might be undertaken on an existential (nonconceptual) plane. If, as Professor Graham noted, the shortcomings of operational expressions are disclosed in theoretical conceptualizations, then we vrould seem to be placing the cart in front of the horse. Utilizing the method of science, it is found that existential reconstructions are generally made to point to any possible shortcomings which the conceptualized structure might have had. Notwithstanding this seemingly fundamental difficulty, we shall now turn to a discussion of the Hicksian concepts in an effort to understand what is claimed in the way of a relationship between operations and the concepts which are said to direct them, Hicksian Definitions Â• Many discussions of incccie implicitly assume that there is but one concept of income. This is simply not the case. In both economics as well as in accomting there is a plurality of notions of income no one of which receives a clear consensus of meaning or definition. Historically, accountants have settled on the Hicksian definitions. His time-honored irreducible statement of the meaning of income crystallized from his own observa^tions might be associated with littleton's definition of "principle."

PAGE 28

21 Littleton states that a princiole "...is a crystallization of ideas into 9 a clear verbal statement of a sigiificant relationship." If this is vrfiat Hicks and others have done, we might be able to gather some indications as to the effectiveness and reliability of this principle formation approach as we trace the meaning of the Hicksian definitions. If we find merit to the Hicksian mode of analysis, perhaps knowledge acquisition processes (presuming that Littleton excluded from his meaning of principle such matters as "principles" of good manners) do envision the crystallization of ideas. We shall need to examine closely not only the central concept of Professor Hicks, but also the three approximations to this core of meaning which he offers. Each such approxii-ation is said to be a more accurate definition of the central concept than the former. By taking each definition separately, we should gather a better appreciation of the utterly subjective nature of the central concept. Moreover, a more accurate characterization of the accounting adaptations of this definition might be established, , The definitions of Kicks will be related verbatim. We shcill wish to bring into isolation one recurring characteristic of Hicksian def10 initions, their total lack of extrapersonal significance. It will be of benefit to note this recurring characteristic as each approximation is considered. V/ithout attempting to deal with the neaning of the term "as well _ A. C. Littleton, Structure of Accounting Theory (American AccountÂ•ing Association, 1953), p. 23. ^^Extrapersonal is used here to represent that quality of propositions v^ose related existential statuses presumably have no dependence on the perceiver.

PAGE 29

22 off," Hicks defines a man's income as that which he might "... consume during the week and still expect to be as well off at the end of the week as he was at the beginning." This is the asserted "central concept." The first of the three prior approximations follows: Income Mo. 1 is thus the maximum amoiant which can be spent during a period if there is to be an expectation of maintaining intact the capital value of prospective receipts in money terms. ^ Income calculations >Mch are consistent with the aforementioned definition are necessarily associated with future expectations on a personal basis. This first approximation "... makes everything depend 12 on the capitalized money value of the individual's prospective receipts." This definition is said to have an identity with those calculations of individuals who are anticipating a constant stream of receipts of a uniform amount. For, "...any one who expects a constant stream of receipts (and does not expect any change in interest rates) will reckon that 13 constant amount as his income, on that definition." This constant stream of receipts is perfectly consistent with the formal definition in terms of maintaining intact a uniform capitalized value of prospective receipts in money terms. An annuity, for example, of a constant amount indefinitely into the future must necessarily have a capitalized value which remains intact period-to-period. A principal amoxmt, if left intact, will yield interest indefinitely into the future. But in order to maintain capital value, one must come to grips with the status of the future interest rates. The use of the limitation "in money terms" in the formalized definition eliminates considerations of changes in the purchasing power of money, but there is nothing in the definition which 'Â•'Â•John R. Hicks, Value and Capital (Clarendon Press, 1946), p. 172. l^Ibid. 13ibid. . .

PAGE 30

?3 eliminates the requirement of dealing with future interest rates. How else is it to be known that the capitalized value of prospective receipts has been maintained (ex gost), or is to be maintained (ex ante )? One may assume that the interest rates will not vary. One mi^t even assume that they \vill vary but that he has an exact knowledge of when and how. It is possible to have a constant principal or a constant income stream so long as the variation in future interest rates is known, estimated, or by default, presumed to remain constant. If it is assumed to remain constant, there will be no difficulty in maintaining capital. The present value of an annuity of $100 per year indefinitely into the future is$ 2,000 assuming a permanent interest rate of 5 percent without regard to vdien this computation is made. Present value in this case will always be the same regardless of when the calculation is made. If the rate is expected to vary, then it becomes in^Derative that an estimation of variation both with respect to amomt as well as time be made. The future status of interest rates in one way or another must be dealt with if income is to be stated in terms consistent with this definition. The problem may not be sidestepped by viewing the capital amount as a simple pidncipal amount. To maintain a money amoimt on a perpetual basis is one thing. If, however, the principal amount is to represent the monetary ejqDression of the present value of anticipated income streams, then not only will there have to be a reference (by assumption, estimate, or knowledge) to futtire interest rates, but also to the level of anticipated income streams. In order to maintain the capitalized value of prospective receipts under Hicks' first definition, there must be a cOiT^jrehension of both the level of future interest rates as well as the level of future income streams. There is no alternative, .

PAGE 31

Under the definition, therefore, the return from the principal amount is termed "income" only on the condition that there has been either an assumption as to future events or a more-o]>-less careful estimation thereof. A statement of monetary income consistent vdth definition "No, 1," therefore, is conditional upon the income receiver's personal expectations of future events. The interest rate stability assxmption does not alter the argument in any manner. The point is, that to come to grips vath a monetary statement of income consistent with the definition, there is reqiiired an inevitable reference to anticipated future events on a personal basis on the part of the income receiver. Under these conditions, the future expectations of one income receiver might well differ from those of another notvfithstanding an identical investment environment. Income, as a statement of a monetary magnitude in accordance >d.th the first formal definition, is dependent upon the estimation (or whim) of -what the future holds by the personal income receiver in correspondence vdth that degree of sophistication vjhich his analysis of the probable status of future events contained. At issue is not the question of whether this can be performed or not. The issue is the warranted assertability or extrapersonal significance of this utterly personal statement. Common agreement (market determined or otherwise) with respect to anticipations only postpones the inevitable. This subjective statement of income would vary dependent upon varying notions as to v/hat is thought will transpire in the future. If such a concept is used as the guiding directive for the operations of the accountant in his search for "true income" (identified as^the first of two major paths to belief in income), then an

PAGE 32

asserted statement of existential interaction (income is) rests on intviitive estimate. The status of statements of such existential interaction depends not on the perceptive faculty of the income receiver in the structure of if-then propositions but on conceptualized conclusions drawn from experience. If "truth" is said to be approximated throu^ the use of such a concept of income as this as a directing force, then "truth'.' would seem to depend on personal estimates of future events. The argument utilizes an extremely poor sense of scientific objectivity. If anything, the assertion that income under this definition was so-and-so for a . particixlar period vrould have to be considered as a statement of a rather crude hypothesis the merits of irrfiich vrould as yet be unverified by conceptualized if-then propositions coupled with percept ualized reconstructed materials. Futxire events might vrell support the subjective statement of income mider this first definition. Scientific ejqalanation requires the testing of hypotheses with events ex post to the statement of the hypothesis. But if we are to await the final termination of all events affecting the unit under consideration in order that the asserted hypothesis of monetary income under the definition might be tested, a question arises as to the practical utility of the whole system. Hypotheses always contain an "I think" element or a rather personalized notion of what e:d.stential conditions are. There is, hovrever, a vast difference between an "I think" which is capable of existential verification and an "I think" which is not. Indeed, this appears to be the substantive difference between factual quality capable of reliance and mere opinion. Consider the following summarized characteristics of Hicks' first approximation; Â• .

PAGE 33

1. To state a capital value in monetary terms reqviires a reference to future events. To say that a principal of $2,000 will yield an annual inconE stream of$ 100 indefinitely into the future imder the ass\imption that the interest rate is to remain constant at 5 percent requires a reference not only to future interest rates but also a reference to the anticipated income streams of the property. The very statement of the principal at $2,000 is a reflection, by definition if nothing else, of the present worth of whatever the future income streams are thought to be, Iikevd.se, if there is thought to ensue a constant stream of constant return on whatever capital value there is, then statistically that capital value is bound to remain intact, the proviso being that the return is spent or othervase disposed of and temed "income." 2. If income may be extracted without impairing the present worth of prospective receipts, then ipso facto , its statement is conditional on the capitalization procedures which themselves are subject to future expectations. It makes little difference whether you estimate future interest rates or, by default, assume them to remain constant. 3. To the extent that there is common agreement with respect to such future events, the statement of income under the definition might be said to become interpersonal . It is, however, not extrapersonal by virtue of simple agreement. Simple agreement is not a condition for knowledge. To be sure, scientific verification of hypothesized statements requires a sensory perceptional faculty. This is consistent with the observation that all relationships with the existential world come initially to inquirers by way of human senses. Note the difference, however, betv.'een intuitive common agreement on the one hand, and the quality of the data perceived by the senses when if-then propositions PAGE 34 ^ 27 are formvilated or conc eptualized and their indicated interaction exLstentially (nonconceptually) per ceptualizedÂ» VJhatever mode of scientific explanation one holds to be the most efficient, and indeed there are many, those viiich have been the most successful in the past have all had one indisputable condition for knowledge or warranted assertability. It will clearly include the provision that before statements are held to constitute "knowns," if-then evidence must be developed extrapersonally supporting the correspondence of perception with conception. If data indicate a lack of correspondence, it falls to conception the task of rearranging possibilities. The assertability of conception depends on the quality of the correspondence between conception and perception. Income No. 1 does not appear to have any such extrapersonal qualities. It appears that income is as income is thought to be. .' : Â•', . . . 4. There mi^t arise the question of the validity of using a market-derived figure to represent a consensus of discounted future expectations. This is tire popular technique of those who appropriate the Hicksian definitions to accounting operations. Since oxir inquiry at this stage is evaluating the possibility of grasping the significance of income by reverting to the asserted concepts which are said to "lie behind the operational expression," it would have to be deduced that income calculated on the basis of market-determined estimates of discounted future income streams does not represent a figure vdiich has sigiificance extrapersonally. A significance cannot be ascribed to it by tracing the concept which is said to direct the operational actions undertaken to measure it. Income is, as incone is thou^t to be whether individualized notions as to future events are involved or PAGE 35 market consensuses. It matters little how many there are who are doing the guessing. Â• What do the above four characteristics of this first formal definition demonstrate with respect to our discussion? V/e are interested in the possibility of coming to terms with the meaning of income on other than procedural groimds. Hicks' first definition does not offer amy extrapersonal significance for income calculations. Under the formal definition, there is no conceptual or perceptual methodology by which income calcxilations could have any substantive extrapersonal meaning. That income exists existentially (in an ultimate, true, or real form), or that factual or assertable income is susceptible to measurement or statement in a monetary term in a manner consistent with both the formal definition as well as the mode of scientific inqiairy appears to be a categorical absurdity. The second definition supposedly comes close to the "central concept" of incomej better than the first yet worse than the third. V/e now define income as the maximum amomt the individual can spend this v/eek, and still expect to be able to spend the Hicks adds that if the rate of interest is thought to remain constant, the second definition and the first definition will "...come to the same thing." An interpretation of the first definition in the light of the second might conclude that the only difference between them is the relaxed assumption with respect to the interest rate in the second; the interest rate being held constant by assumption in the first. There is, however, nothing in the first formal definition which requires same amount 15 -^Ibid., p. 174. ^^ibid. PAGE 36 29 or suggests the assumption of a constant interest rate indefinitely into the future. It might be noted that an assumption mi^t equally be made vdth respect to the second definition. Either way, the future interest rate is dealt vdth on a personal expectation basis, Koreover, if, as Hicks suggests, that definition No. 1 is the definition which most people use in their private affairs, then the formal definition vrould seem to need rephrasing. The formal definition of Income No, 1 (or that of definition No. 2 assuming a relaxed interest rate constancy assumption) is not consistent with reckoning as income the weekly salary of an individual who expects no change in his weekly salary in the future. A capitalization of prospective receipts woxild require the estimation of life-expectancy. There is a difference between the present value of an annuity which has a limited life and one -vAiich is expected to continue indefinitely into the future. If an individual earns$ 100 \inder either definition No. 1 or No, 2, this might well be reckoned as his income, but the $100 is not the amount which might be spent with the anticipation of maintaining intact the capital value of prospective receipts. Statistically, there is no difference between the first definition and the second. They are computationally identical. If it was Hicks' intention to assume away the complexity of changes in the anticipated interest rate in the first definition, there were \mdoubtedly better ways of doing so. The characteristics of the second definitiDn are much the same as the first. ViTiat is required in order to state a monetary figure vriuch might be consumed leaving aH future periods with the e3q)ectation that this sa,me amount might be consumed then? Vfe must know the income streams PAGE 37 of the valued assets in each of the future periods as vrell as the interest rates prevailing in those periods. Once these are settled upon, we can compute an amo\int vdiich might be extracted in such a manner as to have this consimtption amount identical in all periods. Indeed, an exact knowledge of the future would permit utter accuracy in establishing capital values and such a situation defies a variation 16 in incoins period-to-period. Once again, common agreement with respect to expectations does not establish a greater weight of assertability. Subjective expectations cannot be criteria of "truth," Again, it seems to be categorically absxird to hold that income computed on the basis of the second definition represents a reflection of established modes of existential interaction. The existential attribute assigned to income under this definition simply would not exist without the perceiver of expectations. Income is, as income is thought to be. Up to this point we have been considering monetary esq^ressions of income without regard for possible variations in price-levels. We have been concemed with monetsiry capital maintenance or monetary income continuance. Both were dependent on a statement of subjective opinion as to past events (income) with an anticipatory referential base in the future. The quality of past events was determinable only by reference to the pending future. Vfe must now add a third and yet more bewildering complexity to' this already cluttered "path" to central meaning. How do we redefine income if the perpetually anticipated constant monetary stream is expected to vary in purchasing power from period to period? 16 Sidney S. Alexander, pp. 154-156. Professor Alexander suggests that income variations as representations of errors of measurement might come as a shock to some. Under conditions of certainty variations in income "might be considered as normal by an accountant but must be termed "errors in meas\irements" by the economist. PAGE 38 31 Income No. 3 must be defined as the maximum amovmt of money v^ch the individual can spend this week, and still expect to be able to spend the same araoimt in real terms in each ensuing vreek.-'-7 To those elements of the futxire v;hich were required to be known or estimated with respect to the first two definitions, we add another. Now the anticipatory horizon is expanded to include not only future interest rates and income streams, but also the future changes in pricelevels. To be precise, v;e would have to arrive at a statement of the future changes in the price-level which are expected to affect the "market basket" of the individual xmLt in question. Of course, this estimate would have to be made net of any change in technology affecting the quality of the market basket items so that comparable goods would be considered. In order to determine this, some indications vrould have to be gathered which would establish the probability of product uniformity through time. These are but a few of the imponderable complexities which are introduced with this innocent clarification and closer approximation to the central meaning of income. Yet there is still another. V/e now come to the final complexity which, when added to the analysis and after having made condensations therefor, will bring us to the Hicksian assertion of central meaning. This final adjustnent is for consumption goods vAiich are durable to the extent of lasting through more than one incom.e period. Income is "...not the maximiim amount the individual can spend while expecting to be as well off as before at the 18 end of the week; it is the maximiam amount he can consvune ." Money expenditures on durable consumption goods will cause spending to exceed 17 John R. Hicks, p. 174. . Ibid . , r> PAGE 39 consumption. The use of previously acquired consumption goods vrould, to that extent, cause consumption to exceed spending. Hicks suggests that only when the use of previously acquired consumption goods exactly matches current expenditures thereon vd.ll consumption be equated vdth 19 spending. How are we to know this? I^ithout a perfect market for each item of durable consumption goods for each particular degree of wear possible, there is no way of estimating the extent to which spending in one period on durables compensates for consumption in another. If a person uses durable consumption goods in one period, he must compensate for this by a reduction in his spending. Otherwise, he would be "worse off" to the extent of the wear involved in the use of these durables. Hicks contends that the extent to which a greater stream of income is necessary in the future to compensate for current use of dur20 ables carjiot be known without a perfect "durables" market. Hence, the conclusion that "...v/e are forced back on the central criterion, that a person's income is what he can consme during a week and still expect to 21 be as well off at the end of the week as he was at the beginning." The above four definitions of income are said to be ex ante in nature. That is, the calculations involved are to take place before the start of the period in question. Income ex ante for a year is the amount which might be consumed during the year with the expectation of being as well 'off at the end as at the beginning all calculations having been made at the start of the year. There corresponds to each of the ex ante definitions an ex post counterpart. This ex post definition simply makes the calculations in point of time after the period under Ibid . . ^Qlbid. _ V ^^Ibid. PAGE 40 33 consideration. If the expectations were not attained, then the value of one's prospect at the end of the period vdll fall short of what they were expected to have been. Thus, changes in expectations during the period vdll reflect themselves in ex post income calculations whereas they would not under the ex ante definitions. Consequently, ex ante income v/ill equal ex post income only if expectations were exactly attained and no changes therein took place. Hence, income ex ante, plus or minus the effect of changes in expectations which took place during the period, would equal income ex post . Are the characteristics of ex post income calculations significantly different from those of ex ante calculations? Professor Kicks maintains that ex post income calculations are objective in nature. Income "No, 1" ex post ",,.is not a subjective affair, like other kinds of income; it is almost completely objec22 tive." It will become apparent that the significance of this statement depends on the interpretation one gives to the term "objective." If, by objective. Professor Hicks means that the computational aggregates as "inputs" to the income calculation were derived from market place transactions totally apart from personal evaluations of worth by the individual computing his income, then we must proceed to an analysis of how such market prices were determined. Supply and demand is said to determine price. This price, however, represents a coincidence of agreement and" not a consensus of value (discounted future expectations). For exam.ple, if the market value of capital at two different dates is known, and the degree of consumption of dvirables is known, then an income amount might be computed in accordance with the interpretation offered by Kicks of Income No. 1, ex post . It is altogether another matter if ^ ^Ibid . . p. 179. PAGE 41 34 the tern "objective" is to represent extrapersonal verification of the assertability of a monetary income. Even if market derived input values represented consensuses, any change in such opinions as to future events would change the monetary expression of income. Thus, the degree to which the monetary expression approximated the immutable (true or real income as a term representing an existential status) would depend on conglomerate opinion. In either case (Income No, 1, ex ante or Income No. 1, ex post), the suggested meaning for "objectivity" must refer to the coincidental interaction of supply and demand v^ich itself is the subject of expectations. It will be recalled that Income No, 1, ex ante was defined as the maximum amount that could be spent with the expectation of being as well 23 off at the end of a period as at the beginning. Hicks now asserts that this capital value of prospective receipts is susceptible to objective measurement. It is held to be an objective magnitude capable of 24 direct calculation. The meaning of objectivity must be determined by reference to the nature of the capitalizations involved. The formal definition requires the capitalization of expected cash receipts at two dates in order to determine income for the period between the dates. Income would equal the difference betv/een the present value of prospective receipts at the two dates. Ec ante , this conputation would be made at the beginning of the period, ac post , the computation would be made at two different times, the first at the beginning of the period and the second at the end. By shifting the estimator from the individual to the market, we might extract from the computation any overtones of ^^ Ibid .. p. 172. ' Â• ^^'ibid.. p. 179. ''; : . PAGE 42 bias or purposeful deceit. But the fact remains that they are estiinates. Reference to future events is an essential part of the search for "knowns or reliable statements of existential statuses. Future events are central in scientific inquiry. Perceived future events are clearly the determinants of the assertability of conceived hypothesized existential statuses. They must "bear out" the hypothesis before the hypothesized statement of e^dstential status is accepted as a tentative knovm. It is interesting that "income" is asserted before such future events transpire. Income ex ante or ex post is stated vdth dependence upon future events (present value of anticipated income streams) for both its magnitude as well as for its very existence. Even in the face of a perfect market, income calculations could not be held to be objective in the sense of meaning extrapersonal. The existence of income under any of the definitions requires and is dependent upon a perceiver. Qc post definitions draw us no closer to the assertability of an existential status than do ex ante definitions. They both mi^t be said to have an inevitable character of intuitiveness. . In summary. Professor Hicks does not offer any evidence of an intelligible existential state of affairs corresponding to what he terms "income," Unfortunately nothing more might be said of the bed-rock meaning or significance of income now than before. Essentially income is as income is thought to be. The accomtant has appropriated the Hicksian definition at his peril. The role of concepts has received a great deal of attention from philosophers of science. It has received somewhat less attention among accountants. Hicks himself was keenly aware of the shortcomings of his concepts as well as the function attributed to income as an economic tool. He says that it is a tool PAGE 43 36 which breaks in our hands when we endeavor to employ it. He finally allowed the doubt to escape as to whether income "...in the last resort stands up to analysis at all, or vAiether we have not been chas25 ing a 'will-o'-the-wisp.'" The Fisherian Concept Aside from the Hicksian concepts there are other concepts perhaps less computational in nature but nonetheless efforts toward a statement of irreducible meaning. Professor Irving Fisher compiled an extensive discussion of the meaning of income emphasizing the transitional nature of ajrjr income computation. According to Fisher, any computation of monetary income or any compilation of "objective services" (money income transforaed into physical units of wantsatisfying commodities or 26 services) is only an approximation to the central meaning of income. Subjective satisfactions from physical means are the motivational force behind all economic activity. IJhat we really mean when we speak of income are the psychic satisfactions which are derived from the added command of material goods and services brought about by concerted economic activity. In simple terms, a man's efforts in his business or occupation are \;indergone with the hope of thereby being able to satisfy vnnts. Kis labor, if it is socially desired, is rewarded with a greater command over physical goods and services from viiich he might derive a 27 net addition to his satisfaction level. Fisher pictured three successive stages or aspects of income. Money income was considered the first stage. It was defined in a rather 25ibid,, p. 176. ^^Irving Fisher, The Nature of Capital and Income (Macmillan and Co., 1912), pp. 165-168. 27lbid. PAGE 44 37 general way as that money aggregate which a man received which could be used to secure those material goods and services vital to his existence. Real income, then, was money income adjusted for price-level fluctuations. The last stage, the real meaning of the term "income," was said to be this "psychic income," Psychic income is represented by all those "agreeable sensations" and "experiences" which come to an 28 individual through the consujiption or use of material goods and services. This is asserted to be the real concept or the true meaning of income. Kattessich notes when jioxtaposing the Fisherian and Hicksian concepts of income that the one (Fisher's) is based on flow variables v/hile the other (Hicks') is based on stock variables. This seems to be consistent with the interpretation of the Piicksian concepts as 29 capital maintenance concepts. ' Thus, we are led back to the duality exploited by accountants when measuring income, on the one hand, in the balance sheet through the difference in stocks and, on the other, in the income statement through the accumulation of flows. The logico-mathematical principle behind this phenomenon is easily stated: Any chan;;;e can be quantified in tv;o vfays ; (l) b^ measuring the sum total of all contributing increments and decrements ( flows ) or (2) by measuring the difference betvreen the two totals ( stocks ) connected by this change . This intuitively self-evident proposition. . .yields the tvro basic forms from which every definition of income must be chosen.-^^ There might seem to be some cause-effect confusion here. Hicks' notion of income would appear to mai PAGE 45 38 is the case, the intuitively self-evident nature of this proposition has no more vreight than any other logically correct statement determined on the basis of the implied interaction of "givens" or "assumptions." The proposition is intxiitively self-evident only as it depends on what is already understood about the meaning to be attributed to capital and income. Miat are the known elements of existential interaction settled upon by either (stock or flow) notion? Fisher asserts that income constitutes the psychic benefits derived from the consumption of material goods and services. Hicks asserts that income has significance in that it is the amount which might be consumed with the expectation that the individual ;^all be as well off at the end of a period as he was before. VJhere is the indicated known? IVhere is the criterion of usefulness for either definition? How would one gp about reconciling such varying points of view? Both are, in the final analysis, totally dependent on a personalized notion of well-off-ness. Why might it not also be held that factual cxirrent well-off-ness can never be known in the present for it inevitably requires a reference to anticipations? Definitive statements of irreducible crystallized experiences do not change the nature of the factual extant world. Nor do we gain any greater insight into this vrorld or its interacting components through the exclusive use of 31 concepts alone. The definitions of Professors Fisher and Hicks do not lead one to the fomulation of if-then propositions, indeed, none are even indicated. Income is, as income is thought to be. Those in accounting v;ho endeavor to appropriate the concepts of "economic income" to -^^This distinction between concepts and percepts and the pitfalls which av/ait those who seek to rely exclusively on one or the other in inquiry is set forth in detail in Harvey T. Deinzer, Methodological Presuppositions in Financial Accounting Models (Department of Accounting, College of Business Administration, The University of Florida, 196S), pp. 7-8 and 13-14. PAGE 46 39 accounting operations in the form of making them a referential base for guiding the accountant's operations do not appear to have a verygood "case" when it comes to stating an approximation to true , real , or fundamental income. In summary, many vn?iters in accounting have insisted that even if income is such a slippery notion that it cannot be stated exactly, accounting income is still a good approximation to the real thing. In these cases it is seldom knovna just exactly what the writers mean by income, but more seriously is the question of how we are to come to terms with assertions of approximation without any knowledge of absolutes. What justification is there for statements attributing a quality characteristic to income con^jutations \ihen no knowledge of such an "ideal" exists? If income ccmputations are said to be good approximations, the question must be asked with respect to v^at . If we are to knowledgeably assert the approximate natxare of accounting income calculations with respect to the "irreducible" statement of income or this ideal, then perhaps we would also be in a position to go ahead and state the real thing since vre would have to know the relationship between it and the asserted approximate figure. Some engaging insight might be gathered from a consideration of the following irreducible notions of income. This list of constitutive essences of\income is by no means exhaustive. It mi^t, however, be indicative of the efficiency of evaluating the meaning and significance of computed income by tracing the operation "backwards" to the concepts which are said to be the directing force. Not all of the following statements have been suggested as having the quality of a "guide" for directing the operations of the accomtant. They all do, however. PAGE 47 represent what the individual author thinks the meaning of income to be. They are atten^ts at stating a conceived notion of an existential status. John B. Canning agrees for the most part vd.th Fisher, although his "services" might be interpreted as being once removed fron the psychic benefit notion, "For income in essence is services the desired 32 element in economic events." David Solomons isolates income as an unqualified grovrt-h in present value for "...groxrth in present value must be what we had better under^ 33 stand income to mean." Richard lÂ»Iattessich defines income as the "...flow of goods and services vdthin a well-defined period, between the production side and 34 the consimption side of an entity." The difference between the flow of services and the psychic benefits or satisfactions which are derived from their use might be made note of at this point. A. C. Littleton says, "The really basic idea goes deeper . Income is the 'service charge' needed to induce the production of goods and the employment of labor." Professor Norton M, Bedford selects v;hat he terms a constitutive definition for income. It amounts to the "...extant intuitive notion 36 that income is a constitutive concept." Bedford later suggests that John B. Canning, "Some Divergencies of Accounting Theory from Economic Theory," The Accountine; Review , March, 1929, p. 8. Â•^^David Solomons, "Economic and Accomting Concepts of Income," The Accounting; Review , July, 1961, p. 375. ^Slattessich, p. 21. ^^A. C, Littleton, Structure of Accounting^ Theory (American Accounting Association, 1953), p. 22. Italics added. Norton M. Bedford, Income Determination Theory (Addi son-Wesley PublicliLng Company, 1965), p. 72. PAGE 48 an overpreoccupation with the preciseness of a concept might "...violate our commonsense intxiitive feeling that there is something real about a 37 concept, as opposed to just a set of measurements." If one were to transfer his "constitutive" notion about income to the above statement, the result would be "income is bound to be, for we all feel that it exists intuitively." Still another shade of income meaning is esqsressed by Professors Backer and Bell. "Ideally, it would seem that income should represent the increment in entity value that occurred during the period, exclu38 sive of equity transactions." John M. Keynes suggests as a meaning for income of an individual "...the excess of the value of his finished product sold during the period 39 over his prime cost." When all else fails, we are "...forced back on the central criterion, that a person's income is what he can consume during a week and still expect to be as well off at the end of the week as he was at the begin40 ning." To summarize in one sentence, income is as income is thou.ght to be . There is clearly no evidence to support the statement that there is an intelligible existential state of affairs which corresponds to any of the definitions examined. All notions as to meanings within the topic of "income theory" are definitive in character and their merits must be evaluated by virtue of recoxirse to the effectiveness with which they 37 ibid . . p. 70. ^^Morton Backer and Philip V/. Bell, "The Measurement of Business Income, Part 1 The Matching Concept, " Modern Accounting Theory (PrenticeHall, 1966), p. 69. 39john feynard Keynes, The General Theory of Employment Interest and Money (l^acmillan and Company, 1936), p. 53. ^Hicks, p, 172. PAGE 49 A2 carry out their asserted functions. Income is not a term which represents sin assertable existential status. It is not even representative of simple agreement as to definition. Tracing income computations "backmrds" to their conceptual meaning offers nothing more than the blind hope of concept agreement, but then what I Proponents of this "economic income ideal" adaptation hold that income exists in a real or fundamental sense and that it falls to the accountant the task of approximating this real amount as closely as possible (we have seen also that there are some vrfio claim that modern accounting methods factually do approximate this ideal figure). Belief in the propriety of income reporting supported by the argument which was examined in this chapter vrould appear to be just that, belief . If income determination is to receive a favorable critique, the search for its warrant will have to be continued elsevrfiere. The second of the two major beliefs in income was isolated to be the contention that income reporting is vital to sound investor decisions. Future levels of income are said to be determinable from the indications of past trends of income. This predictive potential is thovight to provide prospective shareholders with information on v;hich to base good investment decisions. The remainder of this study will deal with this belief. PAGE 50 CHAPTER III STATEMENT OF STATISTICAL PROCEDURE We have, then, disposed of the possibility of dealing effectively vath an evalxiation of income reporting for shareholders on the basis of its asserted correspondence with con c eptuali zed statements of interaction (alone) or conceived ideal definitions. Aside from having a lack of consistency vdth modem knowledge-acquisition processes, and aside from the sub-question of the lack of general agreement on a definition for this conceptualized ideal, it was found that there existed two insurmountable difficiilties when an effort was made to apply this "correspondence-to-the-economic-income-ideal theory." The first was the impossibility of stating this income figure in terms vdiich vrould be extrapersonally significant. The second major difficulty stems from the first. Without the possibility of such a statement, there can be no "ideal" against which the operational counterpart in accounting might be contrasted. There is, therefore, no possible way of determining whether and to what extent the income calculation of the accountant approximates the "true income" (as represented by such "ideal" economic income defi. * nitions) of the enterprise for a period. Any figure settled upon as the net income of an enterprise for a period of time cannot be said to be "efficient," "approximate," "true," or "correct," by reference to any asserted correspondence to conceptualized "essences" of income. Those who would propose this approach evidently think of the accountant as proceeding with a measure of income knowing PAGE 51 , 44 that the specific attribute which he seeks to measvire is there somewhere. There is sinply no evidence of the existence of a measurable attribute in the economic realm with which extrapersonal significance might be associated by recourse to a supposed correspondence between the measured amount and some "ideal" amount, or economic income. Intellectual discussions pertaining to income and "income theory" preoccupy many accounting writers. As a profession, public accountancy has gone to no small length to refine and delineate what is considered "proper" accounting procedxires for the "realistic" reporting of net 1 income. In general, this approach has been an appeal to the concepts which are said to be involved in the traditional historical cost allocation and revenue recognition notions associated with income deteiv mination. In broad terms, it is fair to say that the accountant defines income as the excess of that which is properly classified as revenue over that aggregate of items which is properly classified as expense. This is a traditional procedural definition vdiich, under somewhat more detailed consideration, appears to be a statement of how to "get to" income without any indications of the meaning of what you have "gotten to" once you have "gotten to it," It is a matter of fact that the computational result of the accountant's operations has simply been termed "income" through tradition, and that there does not exist any known method of relating it to what has been called "economic income," "real income," or "fundamental" income. The writings of Churchman may be e contention that income will be more "realistically" stated with the adoption of a particular procedural alternative instead of another is novrfiere more evident then in an article title written by a former high-ranking individual in a major accounting firm. Herbert T. McAnly, "The Case for LIFO: It Realistically States Income and is Applicable to Any Industry," The Journal of Accountancy ^ June, 1953, pp. 691-700. Italics added. PAGE 52 45 interpreted as indicating the confusion which arises from the exclusive 2 reliance on concepts in the study of economic behavior. The other possible approach to an evaluation of income reporting for shareholders appears to rest in the area of the asserted functional utility of reported income figures. Income reporting has many uses aside from those associated with the annual report. It is a useful device for taxation purposes, for example. Regulated monopolies are supervised by a form of income deteroination to insure that rate charges meet socially acceptable criteria. All of these uses of various forms of income reported are associated vrith a certain group of individuals. The income determination technique is designed with a particular use in mind. Yet, financial income reported in the annual report is designed by appeal to such notions as "realistic income," or income figiires which are neither "overstated" nor "understated." Paramount among the asserted uses of the financial net income figvire is that of being a component of rational investor decisions. Ostensibly, a rational investor acting within the purview of the asserted functional utility of income reported for shareholders would tend to maximize his financial gain by examining the past income trends of all investment alternatives. This assertion carries with it many if-then implications which would need to be investigated before any merit might be placed on the statement. The place of market values vjould, for example, need a careful look. Since a part of any financial return on any common stock (particularly those which are Churchman came upon this insight when some ambiguities in the meaning of "cost" occurred to him. Cost, as a representation of the value of the alternative not taken, should logically be extended to represent the optimal alternative not taken, which implies a knowledge of all alternatives, which implies a knowledge of the future. C. West Churchman, Prediction and Optimal Decision (PrenticeHall, 1961), pp. 59-62. PAGE 53 listed on national exchanges) vdH be determined on the basis of the change in stock price, such an assertion vrovild seem to have predictive implications as to the behavior of market prices in response to income trends. If the implications of this assertion can be established, then the statement would be supported by evidence which would indicate that it might be relied upon and acted upon with confidence in the consequences. The hypothesized implications of the assertion, the then element in impHed if-then propositions, is the central focus of attention in the remainder of the study. I^otheses are tested by developing the ingjlications of the thought-to-exist existential state of affairs by reference to selectively determined experimentation. IF past income figures constitute a con^jonent of rational investor decisions by virtue of their ability to indicate income amounts to be realized in the future, THEN : 1, The extrapolation of past income trends must be an efficient indicator of future income trends and, 2, Decisions based on the relative attractiveness of projected earnings (q;ialified, of course, by investment cost) must lead to the selection of the factually higher cash return stocks determined on an ex post basis. Although the statistical implications of the tests indicated above are rather complex matters, the two implications of this second belief in inconB which are identified as the then elements above, constitute the central tests of the hypothesis viiich will be made. It is believed that the quality of a contribution to the literature of any discipline is not determined on the basis of positive PAGE 54 47 suggestions only. If the hypothesis (and consequently the belief) is supported by the evidence, we will have more to base our actions on than pure assertion. If it is not supported, there will be a need for a redirection of effort. Either outcome would constitute a partial determination of a currently unsettled and untested situation. In order to convey a broad overview of technique, a brief outline of the entire study vdll be made initially. There are two basic questions involved. The first is the extent to which earnings of past years are efficient indicators of the earnings of future years. The second is the question of decision efficiency when stocks are chosen on the basis of the relatively more attractive projected price-earnings ratios. Since changes in market price constitute a part of investment return, it is possible that even if earnings are susceptible to prediction the relatively more favorable stocks might not have been chosen (ex post determined). Predictions may well exactly match fact\xal performance but unless the market price of the stocks is also consistent with the predictions, the relatively more favorable stocks might not be chosen. Hence, there is more to the question than the mere prediction of earnings from past trends. There are two central procedures involved. The first is designed to yield a measure of the efficiency obtained in predicting future years' incomes from past trends. Basically, this part of the study will use an extrapolation of past earnings per share figures from a least squares trend line and thereby estimate the anticipated earnings per share figures for a future period. Stocks will then be arranged in the order of highest anticipated earnings per share figures to lowest. This array will then be rank correlated with a similar array with factual PAGE 55 48 earnings per share figures as the value ranked from highest to lowest. A high rank correlation coefficient vrauld tend to indicate that past earnings per share are good estimators of future earnings per share. But even if the rank correlation coefficient is determined to be 1.0, this still is no indication that prospective investors are any better off with income reporting than without it. If they are said to be able to make decisions based, in part at least, on th3 information supplied in income statements, we must determine whether it makes sense to choose investment stocks on the basis of the relative attractiveness of projected income figures. The decision ostensibly implied in this contention is made on the basis of the higher projected price-earnings ratios of some stocks in relation to all others. The second major analysis, then, will be a rank correlation study between an array of projected price-earnings ratios on the one hand, and an array of actual cash return on the other. If the same stocks with the higher priceearnings ratios also txim up in the upper sections of the array of stocks listed in terms of cash ret\im from highest to lov/est, then the hypothesis would tend to be supported and the second belief in income valid. The cash return will be determined on the basis of cash dividends received during the investment period plus or minus the net change in stock value 3 from the start of the investment period until the end. . . '^The indicated inputs for the rank correlation follow in formula form: Projected Price_ Projected EPS 1963-1965 From Trend Extrapolated Earnings Ratio ~ Least Squares Regression of EPS of Prior Period Investment Cost Factual Price_ Factual EPS 1963-196 5 Earnings Ratio ~ Investment Cost Rate of Cash Return = Dividends 1963-1965 plus Stock Value Change Investment Cost PAGE 56 49 ' It is imperative to recall that a stipulation was made in Chapter I to the effect that casual observation would convince one that there were other factors which would enter into any particular investment choice. Of interest, however, is the efficiency lent to the decision process to the extent that past income trends influence the decision. It is fully clear that the majority of investors make their decisions on the basis of many factors, not the least of which is a good honest guess. We are interested in the extent to which the decision process is helped or hampered by full or partial reliance on extrapolated past trends. The question of real interest is whether the decision base is fully or partially valid. Does income determination do what it is asserted to do? The Sample The saciple consists of one-hundred-ninety-eight common stocks selected primarily on the basis of the availability of relevant data for the study. The sample was not selected purely at random in strict accordance with random sampling techniques. More will be said later on the consequences of this limitation. The sample was selected from data published in the 1963 and 1966 editions of Moody' s Industrial Manual. The data were originally derived from either the content of the annual report to stockholders or the related statements which were filed with the Securities and Exchange Commission. There were comprehensive data available for some three-hundred-fifty stocks which the Moody Manual goes into with somewhat significant detail. Of these three-hundred-fifty stocks, only one-himdred-ninety-eight of the listings contained sufficient data with which appropriate adjustments could be made to translate reported data into relevant data for this analysis. PAGE 57 .50 In many cases stocks were rejected on the basis of inconsistencies in the raw data reported in the financial statements. In many other cases, insufficient data were available with which to adjust earnings per share figures necessay for the establishment of valid trend lines. In other cases the time sequence when both cash dividends as well as stock dividends vrere paid was not reported. This precluded a determination of the actual share equivalents held at the end of an investment period which were attributable to one full share held on the investment decision date. This constraint vrould have precluded an accxirate determination of the cash return for the stock. If, for exasiple, a cash dividend were paid immediately before a stock dividend, the cash dividend per share (or part thereof) would be more th^ if the same cash dividend (in total amount) were paid after a stock dividend. In the absence of an explicit indication in those cases in which the figures were so close as to indicate doubt as to which occurred first, the stock v;as rejected. If the timing sequence was indicated or if there was other information indicating vfhich occurred first, the share was included and an accurate share equivalent figure was determined. The three-hundred-fifty comprehensively treated corporations represent the most important American corporations in terms of total resources, total income, or both. The sample, therefore, will represent a fairly good cross section of American industry. It does not encompass all major stocks. A data constraint prevented a consideration of all important stocks. To include all such stocks would have required a tracing of data from every stock for every year in every set of financial statements. This would have entailed a mammoth undertaking. Yet, these limitations are not as serious as they might at first appear. PAGE 58 51 The large sample size (n value) suggests that little more would be expect ( to be uncovered by incliiding all stocks or by expanding the "n" value very significantly. The sample size should compensate to some degree for the lack of complete randomness in sampling technique. When considering the conclusions dravci from the study, it would be helpful to keep in mind not only the limitation as to the lack of randomness, but also the large "n" value for they tend to compensate for one another. Since the hypothesis suggests a predictive qxiality for past income calculations, the true universe size is quite large in that it comprises all common stocks. There is a vast difference, however, between the importance of this assertion as it applies to all stocks and its importance as it applies to major corporations. The sample also represents some fifteen broadly defined industries. The data will be analyzed both among all stocks as well as among stocks grouped by industry. In the way indications as to the predictive capacity of the data intrai ndustry might be determined. The selection of the industry groups is, at best, a subjective affair. Rather than use the industry groupings which Moody's uses for ratio analyses, which would, in this case, result in a "thin spreading" of some forty-five industries, (many of which would have only one common stock in the classification), some of the industry classifications were combined yielding a total of fifteen industries. These represent viiat the author considers adequate classifications for the pxirposes described above. The classifications are summarized below: 1. Aerospace 3 Corporations 2. Steel 10 Corporations 3. Variety Stores 4 Corporations PAGE 59 52 4. General Retailers 8 Corporations . ' ' ' 5. Chemicals 7 Corporations 6. Cement 4 Corporations 7. Pharmaceuticals 11 Corporations 8. Mchinery, Eqiiipment, and Tools 11 Corporations 9. Building Materials 5 Corporations 10. Oil Refining, Integrated, and Producing 11 Corporations 11. Railroad Equipment 4 Corporations 12. Tobacco 4 Corporations 13. Food 14 Corporations 14. Paper, paperboard, and packaging 6 Corporations 15. Electrical Eq\iipment, Controls, and Instruments 6 Corporations It vail be noted that the total number of corporations represented in the above classifications does not equal the total sample. This is e:x53lained by the fact that there were some corporations whose products did not fit appropriately into any of the above classifications. Many of the largest corporations are so diversified that to include them in any single classification would be to exclude them from other groups into v;hich they might be equally placed. Also, consistent with the original Moody classifications, there were thirteen corporations whose stocks would have comprised one separate industry each. In cises ouch as these, the common stocks of those corporations were excluded from the industry grouped analyses but included in the all stocks analysis. Relevant Data Techniques designed to transform raw data into relevant data need to be considered before procedural aspects of the statistical technique are entertained. The earnings figures of prior years in the form in which PAGE 60 53 they are normally reported in successive annual reports are not valid i bases from which to strike a trend line representing the "direction" which earnings are taking. Data need to be gathered for appropriate "rates of return" figures. These inputs to the rank correlations need to be computed from the raw data published in the annual reports. Eamin,s:s per share The reported operating income or net income is not the decisiondetermining factor to a prospective common stockholder. There are two reasons for this. Corporate earnings are irrelevant first of all because normally the cost which the shareholder pays for his share is stated in per share terms. Hence, corporate earnings irould have to be reduced into the appropriate per share earnings. Cost is one of the most important elements of any investment decision regardless of vdiat the decision criterion might be. On whatever basis an investment is selected, that basis is generally reduned to a common association with whatever is knovm about investment cost. The second reason, and perhaps the more important, stems from the possibility of there being a preferential issue vtoch stands, as it were, in front of the common shares as to earnings. As a consequence, all that is reported as corporate eanoings or financial net income cannot in any way be necessarily associated en masse with the common shares alone. Corporate earnings are, therefore, irrelevant for both the decision envirorjaent of the prospective investor as well as for i the current purpose. If the trend of earnings is to be extrapolated, the relevant element of corporate earnings is the per share earnings -vdiich are attributable to one common share. Yet, this is still not enough. If earnings per share figures are to represent the average number of shares outstanding during any given fiscal year divided into the opera- PAGE 61 54 ting or net income of a corporation, then stock dividends or stock splits vail significantly affect the average n-umber of shares outstanding. This will, in turn, affect the per share earnings reported in financial statements. For example, if corporate income is constant over a two-year period, and there is a two-for-one split at the end of the first year, the second year's earnings per share figure will be onehalf that of the first year's. If there were reason to establish a trend of earnings attributable to one share held or purchased at the start of the first year, the reported earnings per share figure would have to be doubled for the second year in order for the trend line to 4 be valid. Corporate earnings as reflected in the income statement are, therefore, twice removed from relevance as far as establishing earnings behavior of a single share. In the first place, any earnings attributable to any preference stock must be extracted from corporate earnings. In the second place, earnings attributable to the common stockholder's equity (earnings attributable to all common shares) need adjustment for any stock dividend or stock split vAiich might have taken place during the trend determining period. As in the case of the sample, there were limitations on the nximber of years for which adjusted earnings per share figures could be calculated. For all onehundred-ninety-eight stocks there will be sufficient data available for two separate trend periods, 1956 through 1962 and 5 I960 through 1962. For forty-one stocks in the sample, an additional ^Although there are different accounting entries for stock splits and stock dividends, the effect on relevant data in the study is the same. A 100 percent stock dividend is identical to a 2 for 1 split. ^Trend period refers to the 3rears in which earnings are regressed. Investment period refers to the years in which the investment is held. PAGE 62 55 trend period can be established for the period 1951 through 1962. Thus, for forty-one stocks there vail be data available for three separate periods in which valid trends might be established, a twelve-year period, a seven-year period, and a three-year period. For one-hundred-fiftyseven stocks, there will be data available for only two trend periods, the seven-year period (1956 through 1962) and the three-year period (I960 through 1962). It is hoped that through the use of such varying trend periods the study might cover possible criticisms over the timing of economic causes and effects. For example, if only the shorter trend period was used, it might be asserted that stocks must be evaluated by reference to their income performances over long periods of time. The year-to-year fluctuations could be held to be inevitable with no significant effect on. the general long run trends. On the other hand, if only the long run trend was used, it could be held that long run trends are not predictive due to the constant changes taking place in technology and marketing. It could be held that the shorter periods alone constitute the best basis for predicting future years' incomes because if trend periods are stretched out over too long a period of time, the economic and technological "causes" of past performance are no longer operative. Through the use of varying trend periods, it is hoped that questions such as these might be avoided. If the longer periods are, in fact, more predictive, we shall find out, and vice versa . The primary question, it should be recalled, is not whether the one is more efficient than the other, but whether either is efficient in predicting vMch of all possible investment alternatives (in common stocks) are the better prospects. In all components of the study, the central question must be kept in mind. Every element is designed to establish the qualitative PAGE 63 56 aspects of the implications of the assertion that currently reported financial income computations constitute relevant data for shareholder and prospective shareholder decisions in that they form a basis for predicting future years' incomes. With the above backgroxind in mind, a detailed exposition of the earnings per share adjustments can be made. The nature of the problems encomtered in adjusting reported earnings per share figures is best described by recourse to an illustration. While it is the central purpose of the study to examine whether the data implications support or refute the principal assertion, there is the eqtially important matter of determining whether earnings may be efficiently predicted in the first place. To settle this question the three trend periods were established to see whether future incomes are, in fact, susceptible to projection and prediction from the indications of past trends. This trend line will be drawn in accordance with a least squares regression formula representing a minimum total distance (positive and negative) of the combined distances of all adjusted figures from the trend line. No mention has been made, however, about the base year to which all earnings per share figures must be adjusted for a valid trend expression. To adjust the published earnings per share figures in the context used here means to state them all in eqviivalent dollars of earnings (not adjusted for price-level change) consistent with a common base year. One might establish adjusted earnings per share figures in terms of the equivalent earnings of a single share held at the start of the ten year period, at the end of the period, or any year of the ten year period. PAGE 64 57 Assume a hypothetical corporation has constant earnings for a ten-year period (1951 through I960). The only change that takes place in the financial affairs of this corporation is a two-for-one common stock split at the end of the fifth year. In this case, the reported earnings per share data might look like this: Earnings Per Share Unad.justed 1951 1952 1953 1954 1955 1956 1957 195S 1959 I960 2.00 2.00 2.00 2.00 2.00 1.00 1.00 1.00 1.00 1.00 in order to establish a valid trend representative of the average historical experience of a common share in earnings, some adjustment must be made. One has a choice. A trend might be struck which would represent the earnings in each of the ten years which are attributable to a conmon share held at the start of the trend period (1951). In this case, the adjusted earnings per share data vrould be set forth as follows: . Earnings Per Share Base 1951 1951 1952 1953 1954 1955 1956 1957 1958 1959 I960 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 The major point is that adjustments are made to reflect earnings attributable to a single coimion share held at the start of 1951. In 1956, there is held two shares instead of the one which vas held from 1951 through 1955. The two shares in 1956 are directly attributable to the share held in 1951 and, accordingly, the earnings attributable to that single share of 1951 are represented by the earnings of both shares held in 1956. If, on the other hand, the trend is established in terms of the historical record of the equivalent of a single share held at the end PAGE 65 58 of the trend period, the follovdng data are appropriate; Earnings Per Share Base I960 1951 1952 1953 195h 1955 1956 1957 1958 1959 I960 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Either of the two trends is valid, but the choice of which to use niast be made on the basis of the function of the trend. The selection of a particular base year for adjustments carries with it implications as to the consistency of the investment cost figure with that year. For example, if the choice is made to use 1951 as the base year, one would have to double the investment cost prevailing in I960 for a valid projected price-earnings ratio. At I960, the prevailing stock cost vrould represent only one-half the equity to which all trend earnings have been made consistent. If it were deemed necessary to use the prevailing market price in I960 for an investment decision, then it would seem more likely that the earnings adjustments would be made using I960 as the base year. In order to' avoid the extra computations required of two separate base y^ar adjustments, the base year to which all (trend period as well as investment period) earnings per share figures were adjusted in this study was the year 1965. As a consequence, investment cost figures used in this study need not necessarily be the simple cost of a share on or near the hypothetical investment decision date. If a stock were split subsequent to the investment decision date, the actual earnings per share for the investment period would reflect the equivalent earnings which were attributable to a single share held as of the end of the investment period, not the earnings attributable to a share held as of the investment decision date. In the event of a two-for-one split in the investment period, for exanple, the value for the investment cost would have to be PAGE 66 Â• 59 the published price for a share on (or near) the investment decision date divided b^; tvro. All factors must be made consistent once a decision is made as to a base year. The earnings per share data, the dividends received data, the investment cost data, and the sale value of the stock all must be mutually consistent for the analysis to be valid. Investment and disinvestment decision dates . Because of the lack of information as to when vdthin any year dividends were paid to common stockholders, the investment decision and the disinvestment decision were timed to insure accurate cash return data. The investment decision date for all stocks will be the start of the fiscal year (for the particular stock) 1963 and the disinvestment decision date will be the end of the corresponding fiscal year in 1965. A corporation whose fiscal year coincides with the calander year will have decision dates of January, 1963, and December, 1965. A corporation with a fiscal year ending in August will have investment decision dates of September, 1963, and a disinvestment decision date of August, 1965. More is said below on the exact determination of investment cost. Investment cost Hypothetical investment decisions will be made on the basis of the relative attractiveness of predicted price-earnings ratios which vdll have been determined on the basis of predicted adjusted earnings per share as a percent of investment cost. It will be noted that projected earnings is not the sole criterion of advisable investment. Per share anticipated earnings must be related to investment cost before the relative attractiveness of each alternative can be determined. A stock with highest projected earnings per share need not place among the "best bets" since the related cost of acquiring that share might be so high PAGE 67 60 as to malte the cost of those anticipated earnings prohibitive. The technical decision criterion (as implied in the statements supporting the second belief in income outlined in Chapter I), therefore, vail be the ratio of average projected earnings per share over investment cost. The settlement on a angle figure for investment cost is a difficult matter and the process is clearly not vrLthout subjective overtones. To simply seize upon one single day as representative of the cost of the particular investment under consideration appears to be completely inadequate for the purposes here. It was considered more reasonable to go through an averaging process for the stock prices of a particular share which surrounded the decision date. In this way, it is hoped that a more representative investment cost might be established. To pick a single trading day to establish stock prices might well catch many stocks at a peak or trough thereby lending a severe bias to the results. In both the projection as well as the evaliiation segments of the study, an averaging process was used. The mid range of the stock price in the month preceeding the investment decision date as well as the mid range of the stock price in the month following the investment decision date were averaged to produce what is termed herein "investment cost." For example, if January 31 were the hypothetical investment decision date, the investment cost would be the average of the mid ranges of the stock price for the months of January and February. Or, what amounts to the same thing, it would be the high price for the months of January and February plus the low prices for January and February, divided by four. Investment cost is, therefore, the average stock price for the twomonth period surrounding the investment decision date as determined by the stock price range. Since a part of the return on any stock con- PAGE 68 61 sists of price fluctuations, to carry the averaging process too far would result in averaging out the very changes which constitute a part of the implied predictive potential of past earnings. The tvro month averages noted above are considered the miniiraim. averaging process consistent vdth averaging out only day-to-day market price fluctuations. To take the process beyond the two month period woiold be as \mrealistic 6 as not to utilize any averaging at all. This procedure migit be objected to on the basis that observations may indicate that investors do not have access to such a representative stock price and that, in practice, investors seek to av^ait the low price, or the "good time to buy," and the high price or "the good time to sell." This observation is readily conceeded as a possibility and appropriate adjustments in investment cost and sale price figures will be made. In order to accomodate this qi;iestion, a sub aaalysis will be undertaken with appropriate rank correlations just as with the major study, which utilize a time-flexible investment cost and sale price. Investment cost will, in this case, be interpreted as the lowest possible price prevailing within the two month period surrounding the investment decision date. This will correspond to a sale price which will represent the highest possible price prevailing within the two month period surrounding the sale date. In this v/ay the study will encompass the possibility of this behavior and stipulate its probability, but it will not leave it at that. If the contentions of the would-be critics are v^arranted (and this question will not occupy us in this study), vre will have an answer. Â— Â— Â— These data were gathered from Standard and Poore's Monthly Stock Guide for the years 1962 through 1966. Since all stocks are listed, the prices are consistent vdth the New York and American Stock ExchangeAssociated Press price quotings. , PAGE 69 62 Share equivalents It will be recalled that stock dividends and stock splits caused a rather bothersome problem in gathering relevant data for earnings per share figures for trend periods. These transactions also cause problems in determining an appropriate investment cost figure. Since the adjustments for all earnings per share figures are to be made on the basis of the year 1965, projections as vjell as evaluations vill have to be made consistent with that same base yesir. Hence, any computation for investment cost vdll have to be divided by the share equivalent at the end of the investment period for a given share held at the start of the investment period. The equivalent of a single share purchased at the start of an investment period would be tvro shares if there had been a two-for-one split during the investment period. In the event of a 10 percent dividend, for example, the share equivalent would be "1.1" if this stock dividend took place during the investment period. In the former case, a holder of a single share would hold two shares at the end of the investment period, and in the latter case, he would hdd 1 l/lO shares. Since all earnings per share figures are adjusted to be consistent with the past equivalent earnings of a single share held at the end of the investment period, whatever computations are made to arrive at an investment cost figure will havB to be divided by the share equivalents if the original unadjusted share price figure is gathered from published prices of the share at the investment decision date, \d:iether an averaging process is used or not. For exarple, if actual price quotings yield an averaged investment cost of$ 50.00 on an investment decision date, the actual cost of the investment for the earnings attributable to a single share held at the end of the investment period

PAGE 70

be $25.00 if there had been a two-for-one split. Hence, a share equivalent value in this case would be "2," the value to be divided into published investment cost at the start of the investment period. No such division vdll be required at the end of the investment period for the determination of the sale price since all data are made consistent vdth the 1965 termination date. Miatever is settled upon as the averaged investment sale value vdll not need share equivalent adjustment. The use of the averaging process does not negate the requirement of dividing investment cost by share equivalents for an appropriate investment cost figure. The individual components of that averaging process are unadjusted so the final average is also unadjusted. The adjustment would be required vxhether a single trading day was siezed upon as an appropriate index of share value or whether the average of many such days is used. The investment cost figvire (not the sale price) must be divided by the share equivalents whether it is averaged or not. Dividends received There is a subtle difference between the constraints of adjustment in establishing the decision criterion and the actual cost of the investment decision itself. For the projection of earnings and for the establishment of the anticipated price-earnings ratios on the basis of which the investment decision is made, adjustments are required to make all data mutually consistent. VJhen the actual investment is made, it will be made on the basis of the then current unadjusted cost of a sh^re. This share cost will have been adjusted in the sense of having been subjected to the averaging process, but it will be unadjusted in the sense of not having been subjected to share equivalent adjustments. PAGE 71 The investment cost used in the projection segment might bear no relationship vd.th the investment cost used in the evaluation segment. The evaluation segment will use a stock price which is the average of the actual stock prices current at the time the investment is made. This accounts for the difference betvreen the cost of investment used for projection and the cost of investment used for the decision in those cases in which the stocks have a share equivalent value greater than 1. This is not an inconsistency. The only reason that projected investment cost varies from the decision cost is due to the requirement that cost and adjusted earnings per share figures must be consistent in order to get valid projected price-earnings ratios. This insight is important when a consideration of just exactly what is meant by the term "share" in computing dividends received per share . Dividends are obviously a conponent of cash return, but c^oes "share" refer to the dividends received on the equivalent of a share held at the end of the investment period? Is it also subjected to the share eqiaivalent adjustment? Does it refer to the dividends received per share in the sense of what is meant by that term in the annual reports? Or, does it refer to the total amount of dividends v^ch were paid to an investor b^ virtue of the single share purchased at the start of the investment decision period? For accurate determination of cash return, dividends per share must encompass all dividends (cash) which were received by virtue of that single share pvirchased at the investment decision date. If there had been a two-for-one split, then, the reported dividends per share would have to be doubled for those years in the investment period after the occurrance of the split. If the split took place at the start of 1964 (the second year of the three-year X;. PAGE 72 investment pericxi I963-I965), the dividends received vroiild consist of the reported dividends per share for 1963 ^ plus twice the reported dividends per share for the years I964 and 1965. These adjustments are implicit in the dividends per share figures, the adjustments having been made at the time the data were gathered. Capital .gjain or loss Dividends attributable to a single share held on the investment decision date constitute only one part of the investor return. The other part of the return would be reflected in any increase or decrease in an investment price (market value of the stock) from the time of investment to the time of sale. These two sources constitute the only t\-TO ways in which an investor can get purchasing power from an investnent in common stock. Consequently, vrhatever device is used for the prediction of the better stocks, the decision criterion must be predictive with respect to either dividends or capital gain, or both. Those who assert belief in the propriety of income reporting on these premises assume a predictive power of past earnings trends (although this is generally not made explicit). Again, it would appear to be rather mrealistic to sLmply take one trading day and establish as the sale price for the stock the high or low price at which the stock was traded on that day. Therefore, the same averaging process will be used in establishing a sale value as was used is establishing the purchase price or investment cost. The mid ranges of the two month period surrounding the sale date (the last day of the corporate fiscal year) mil be averaged. In addition to the selection of this sale value fc;? the stock, an entirely different set of rank correlations win be used when time flexibility is introduced. PAGE 73 66 This will correspond to the time flexible sub-analysis which was described above with respect to investment cost. The investment cost was determined both on the basis of the average of mid ranges for the two month period svirrounding the investment decision date as well as on the basis of the lowest stock price prevailing during that tv:o month period. This procedure was allovred so that tests could be run which would give indications of the reliability of past income trends in settling on the relatively better common stocks if this "good time to buy" and "good time to sell" notion v/ere, in fact, operative. The "good time to sell" aspect of this sub analysis is interpreted as the highest possible stock price prevailing during the two month period surroundirig the investment decision sale date. ' ' ^ As with dividends, it is important to compute capital gain or loss associated vdth or attributable to the originally purchased share. Thus, if there had been a two-for-one split during the investment period, it will be necessary to double the share price prevailing at the time of sale in order to get a valid cash return figure. In this case the prevailing cash value of the stock which v^as purchased on the "buy" date is represented by two shares at the sale date. . In summary then, the decision criterion will be the relatively more attractive stocks in terms of the projected price-earnings ratios which are determined on the basis of extrapolated (from three past periods) earnings adjusted to be consistent with an investment cost. The evaluation criterion vdll be actual cash return in the foim of dividends attributable to the single share held on the investment purchase date plus or minus the net change in investment value (not share value), divided by the averaged cost of a share on the investment decision date. PAGE 74 67 The values in these tvro arrays -will be ranked in the order of highest value to lowest value, and then the two arrays will be rank correlated. The rank correlation coefficients will indicate the degree to which this belief in income under test is to be supported. Implicit interest Since there are generally two basic reasons why conmon stocks are chosen as investments (growth and income), it might appear that a study of this complexity should adjust the data f\urther to give explicit effect to implicit interest involved in investment costs and negative interest adjustments for dividends when they are received. This lack of adjustment might be considered by some as a serious limitation of this study. The relevant data for such adjustments were not available and, thus, such adjustments could not have been made even if it were desired. The limitation, hovrever, is not as serious as it might at first appear. The extent to which this limitation constitutes a serious weakness should be determined, in pal^t at least, on the basis of judgements as to whether prospective shareholders actually make such implicit interest adjustments in practice. This appears doubtful from casual exposure to investor behavior. Moreover, in order to cause a serious defect inthe findings of the study, the lack of interest adjustments would have to result in a deviation in the rankings of the stocks and not simply deviations in the projected price-earnings ratios. It was for this reason, among others, that rank correlation analysis was chosen as the statistical tool. Since, for the most part, the greater amount of any implicit interest adjustment would work in the same direction for all stocks, this weakness does not appear to be qxiite as serious as it first seems. Moreover, to make this adjustment would require a knowledge of when within PAGE 75 a given year the dividends vrere received. These data were not available. To establish dividend receipt dates, therefore, would have been conjecture in the first place. Of more importance, however, the arbitrary receipt dates nd-ght well have resulted in a serious distortion instead of a slight refinement. Ideally, the study should have made such adjustments in accordance with exact dividend receipt dates. We stipulate the validity of the impHcit interest adjustment. The results, however, are not expected to be materially affected by this constraint. The hope of establishing conclusive proof v-ras abandoned in Chapter I. We seek the indications of the data as they shed light on the implications of the belief under review. Vfe now proceed to a detailed illustration of statistical technique. The adjustments wiich follow below for the single hypothetical share represent in detail the exact same adjustments and procedures followed for each stock in the sample. Projection Technique Assume the following hypothetical data: Unadjusted reported earnings per share (in dollars ) 1951 1952 1953 1954 1955 1956 1957 195S 1959 I960 .60 .60 Â«80 1.00 1.20 1.40 .80 . 90 . 90 1.00 ,1961 1962 1963 1964 1965 1.20 1.30 1.20 .70 .75 Unadjusted dividends per share (in dollars) Â• ' 1963 1964 1965 1.00 .50 .50 The hypothetical decision dates as determined by the close of the corporate fiscal year are March 31, 1963 (investment decision date), and PAGE 76 March 31, 1965 (sale date). Associated unadjusted market values for the tvro months siirrounding these decision dates follow: Market prices from exchangje transactions ; Range for the month of March, 1963$ 55.00 $40.00 Range for the month of April, 1963 45.00 40.00 Range for the month of Fiarch, 1965 25.00 23.00 Range for the month of April, 1965 26.00 22.00 Time flexibility data ; ' Â•' Lowest price in March and April, 1963$ 40.00 Highest price in March and April, 1965 26.00 The foregoing data are typical unadjusted data which were gathered for each share in the study. The only exception is the availability of earnings per share figures for the full twelve year period 1951 through 1962. This information was available for only forty-one of the stocks. The above data represent, therefore, the maximum that was processed for Jeach share. In addition to the data above, assume the following facts: 1. There were two separate tvra-for-one splits in the common stock. The first took place at the start of 1957 and the second took place in the investment period at the start of 1964. 2. There were no other changes in coital. 3. There were no transactions in Treasviry Stock during the entire fifteen year period. . 4. There were no material changes in accounting procedure during the entire period. Earnings per share .-.-^ Because of the two stock splits, adjustments are required so that

PAGE 77

70 the earnings trends that are to be established will be valid. The base date to which all data are made consistent is 1^965 . The tv/o-forone split in 1964 vail require all previous earnings per share figures to be divided by two. The split in 1957 also requires a' / carrjjigs per share figures prior to 1957 to be divided by tvra. Hence, all earnings per share figiires prior to 1957 Kiust be divided by four, divided by tvra once to give effect to the 1964 split and again to give effect to the 1957 split. The adjusted earnings per share figures are presented below: Adjusted earning; s per share Base date , 1965 1 1951 1952 1953 1954 1955 1956 1957 1958 1959 I960 .15 .15 .20 .25 .30 .35 .40 .45 .45 .50 1961 1962 1963 1964 1965 Â• .60 .65 .60 .70 o75 The above earnings per share figures represent the earnings which were attributable to the eqxiivalent of a single share held at the end of the base year, 1965 (the termination of the investment period). Investment cost ' The raw data have indicated the price ranges surrounding both the investment decision date as vrall as the sale date. In accordance with the averaging process described above, the mid ranges of these monthly price ranges are averaged. The investment cost figure for this hypothetical share (before any adjustments required for share equivalent values greater than l) is determined to be $45.00 by the following process. ' Fiid range. Inarch, 1963$47.50 ^ 90.00 _ . Fiid range, April, 1963 42.50 2 ^ ^^'^^ $90.00 PAGE 78 71 Share equivalents Since there was a stock split vathin the investment period, there will have to be a further adjustment to reduce investment cost so as to be consistent with the adjusted projected earnings. If the cost figure were to remain at the averaged rav; figure of if 45.00, it vrould reflect the cost of anticipated earnings attributable to a share held on March, 1963 , vrfiereas anticipated eaimings have been adjusted to be consistent vdth earnings attributable to a share held in March, I965 . Averaged investment cost must be divided by the share equivalent value. Since two shares will be held at the end of the investment period instead of the single share purchased, the share equivalent value in this case vdll be 2. Investment cost used for projection of price-earnings ratios is S 22.50. , Â• ' Twelve year pro;' action The trend extrapolated earrAngs projection will make use of a linear least squares trend line with a projection of past trends into a future three year period (1963 through 1965). The first trend equation 7 is determined from earnings data from the period 1951 through 1962. 7 Tne f ollovdng is a derivation of the values for a and b in the trend equation used in the extrapolation. Year X EPS XY X2 Year X EPS XY X2 1951 -11 .15 -165 121 1957 1 .40 40 1 1952 9 .15 -135 81 1958 3 .45 135 9 1953 7 .20 -140 49 1959 5 .45 225 25 1954 5 .25 -125 25 I960 7 .50 350 49 1955 3 .30 90 9 1961 9 .60 540 81 1956 1 .35 35 1 1962 11 .65 715 121 4.45 1,130 572 a = Sum. Y = 4.45 = .370833 b = Sum. = 11.30 n 12 Sum. X'i 572 " Â• Stephen P. Shao, Statistics for Business and Economics (Charles E. Merrill' Books, Inc., I967), pp. 521-530. PAGE 79 72 Y = a / b (x) c ' For pro.jected 1963 earnings : = .370833 / .019Â£7a (13) Y^ I 0629242 For projected 1964 earnings: Y = .370833 / .019878 (15) 0 = .668997 For pro.jected I965 eamingis ; Y^. = .370833 / .019878 (17) Y^ = .708753$ .629242 .668997 I 2.006992 , ^ ^^68997 .708753 $2.006992 Hence,$ ,668997 is the projected average jearly earnings per share for the period 1963 through 1965 as determined from an extrapolation of adjusted earnings per share data from 1951 through 1962, The number of stocks involved in this particular extrapolation in the study itself is forty-one. The other trend periods (1956-1962 and 1960-1962 ) -will have the total one-hundred-ninety-eight stocks. It will be noted that the average projected earnings per share figure is the same as that for the second year. This is due to the use of a straight line trend for the establishment of both the average as Â•well ^s each individually. A straight line is used to set values for three equally spaced time periods in the future. Naturally the second value vd.ll equal the average of the first three. Therefore, we need compute only the second year projected earnings in order to arrive at the desired average of the three future years. The average anticipated eamingp per share will be placed over the adjusted investment cost

PAGE 80

73 and the result of this procedure vdll be the indication of the projected price-earnings ratio, or the decision criterion. Seven year projection The seven year period refers to 1956 through 1962. We shall generate projected price-earnings ratios from the extrapolation of the trend of earnings in these years for the entire sanple of one-hundredninety-eight stocks. The derivation of the projected ''earnings in Since we need .compute only the projected earnings per share for the second year, the average of the three future years being the same, the above calculation is all that is needed. Again, this projected earnings figure will be placed o-ver the .adjusted investment cost to determine the projected price-earnings ratios on the basis of which the stocks will be rank correlated. 8 The following is a derivation of the values for a and b above: 8 the hypothetical case follows. For projected 1964 earnings ; = a / b (x) = / .048214 (5) : .726784 Year 1956 1957 1958 1959 I960 1961 1962 X -3 -2 -1 0 1 2 3 EPS XY .35 -1.05 .40 .80 .45 .45 .45 0 .50 .50 .60 1.20 .65 1.95 3.40 1.35 28 ^ zMa= .485714 b =^ " ^ = .048214 Stephen P. Shao, pp. 521530.

PAGE 81

74 Three ;^'Bar projection All stocks of the sanK)le will also have a three year trend period as the basis for a short run projection of recent earnings into the imediate future. The projected earnings per share amount frc::n the three year trend period in the hypothetical case is determined as The average annxial projected earnings per share for the three year period 1963 through 1965 as determined from the extrapolation of the earnings trend of the three years imrrjediately preceeding is $,808333. Actual averagje earning:s per share The actual average earnings per share is determined simply by placing the sum of the factual earnings per share (adjusted) over 2Â» In the. hypothetical case, this average actual earnings per share amounts to$ .683333, the average of $,60,$ .70, and $.75. ' In summary then, ve have projected the average earnings per share figures anticipated in the period 1963 through 1965 from three different 9The following is a derivation of the values for a and b above: 9 f ollovj-s . For projected 1964 earnings ; a / b (x) Y^ = .583333 / .0750 (3) Yc = .808333 Year X EPS XY X^ I960 196i 1962 Â•1 0 1 .50 -.50 1 .60 0 0 1.75 .15 2 a = Sum. Y = 1.75 ^ . 3 .583333 b = Sum. (XY) Sum X2 iii .C750 Stephen P. Shao, pp 521-530. PAGE 82 75 past periods of time, each period progressively shorter. The three projections are summarized belov;. Trend period . Pro.jected EPS , Actual 2PS 1951 1962$ .668997 $.683333 1956-1962 , .726784 ' .683333 I960 1962 .808333 ' .683333 It vdll be o? interest to know the extent to which actual earnings approximate the projected earnings. In the main body of the study percentage figures i\ri.ll be calcxilated which are designed to determine the degree of accuracy obtained with such projections. Actual earnings will be compared with projected earnings without any other procedure involved. The percentage figure will represent the extent to which actual earnings conform to the projection. The data for the hypothetical case follow. Trend period Projected EPS Actual EPS Percenta.^e Actual to Projected 1951 1962$ .668997 $.683333 102.09 1956-1962 .736784 .683333 .. ^ 93.95 1960 1962 . 808333 ' .683333 ' 84-53 . The important percentage will, however, not be the individual percentage for each stock but the mean of percentages for all stocks for each of the three trend periods. It vdll be of great interest to know what the mean percentage actual earnings per share is to projected earnings per share from each trend period. The determination of the percentages for each stock is determined as illustrated above. Projected price-earnings ratios The projected price-earnings ratios are found by placing the projected average earnings per share over the adjusted investment cost. PAGE 83 76 For each stock, then, there will be three separate price-earnings ratios, corresponding to each of the three trend periods. There vri.ll be a priceearnings ratio deterniined on the basis of the extrapolation of the earnings trend established in the period 1951 throtigh 1962 and for the other two trend periods also. In most cases, however, this will mean only two projected price-earnings ratios because the first trend period was established for only forty-one of the stocks of the sample. A data constraint caused this limitation. In the hypothetical case, the price10 earnings ratios are as follows. Trend period Projected EPS Investment cost PE Ratio 1951 1962$ .668997 ^ 22.50 .0292 1956-1962 .726784 . 22.50 ' .O324 1960-1962 .808333 22.50 .0360 Actual price-earnings ratio The actual price-earnings ratio, which will be rank correlated with the projected price-earnings ratios, is determined in a similar manner by using actual earnings per share instead' of projected earnings per share. For the hypothetical case, the actual PE Ratio is .O3O4 determined as follows. , . Actual earnings $.683333 = .03O4 " ' Â• Investment cost ^ 22.50 *" Evaluation Technique The data above have been illustrated in hypothetical form for the purpose of conveying an understanding of the basis of the asserted decision criterion of shareholders. Investors are thought to make use "'^Unadjusted average investment cost is$ 45.00. Placing this over the share equivalent value of 2 yields $22.50 for average adjusted investment cost. PAGE 84 of the past earnings of a corporation in the form of making these data the basis for an anticipation of the earnings to be realized in future years. The extrapolated trends have been established in order to provide a basis for an array or listing in the order of highest to lowest projected earnings as a percent of investment cost. From such an array, the asserted utility of income computations might be tested. The question is can prospective investors make rational and good investment decisions if that decision base is the relative attractiveness of projected earnings as established from the extrapolation of the trend of past earnings. An array of projected price-earnings ratios rank correlated with a similar array of actual price-earnings ratios will indicate the extent to vdiich the investment criterion factually turns out to be what was thought of it. Yet, this is still not the whole picture. . Even if actual price-earnings ratios were susceptible to projection, it will not mean that the prime assertion binder investigation is supported. Onemust be able to. predict the relatively better cash return stocks from this basis, not simply those stocks which will turn out with the better earnings performance. Earnings is one thing, cash ret\im is another. The technique for determining cash return is illustrated below. . Â•It will be recalled that dividend figures for the three years of the investment period as reported in the financial statements (unadjusted) were$ 1.00, $.50, and$ .50 per share for 1963, 1964, and 1965 respectively. To determine cash return it will be necessary to collect all dividends vMch were attributable to a share held on the investment decision date, not the dividends attributable to a share held at the end of the investment period. This same problem will cause an apparent

PAGE 85

78 lack of consistency between the inv3strr.3nt cost used for projection and the investment cost used for determining cash return. It will be noted that it is of no consequence that the figure used for investment cost for projection differs from that .used for evaluation. The data do need to be internally consistent vithin the ranked arrays. The separation between the rankings, however, is complete. Once the data have been refined and adjusted to provide a basis for the ranking in the order of relative attractiveness of projected earnings as qualified by investment cost, the investment decision has been made. So long as the evaluation criterion actually ranks those same stocks in the order of the relative attractiveness of factual perfonr.ance . there is no compelling reason for the data to be in agreement as to a single investment cost. The problem arises because of the necessity of carrying one single share through several processes of trend determination and price-earnings projections. Had there been no stock dividends or splits, there \irould be only one investment cost for all computations. ' ,.^ Â• The determination of actual cash return (the evaluation criterion) is illustrated below. Cash . Return Capital Gain . -'Sale value of the inveG-r.ent $48.00 _ Cost of the investment 45. 00 Capital Gain ^ 3,00 ^Sale value of the investment constitutes the proceeds received from the sale of all interests arising from the single share purchased on the investment decision date. The average market price must be doubled at the sale date because there are two shares (due to the I964 split; kt the sale date for the one share which was purchased. PAGE 86 ; 79 Â• : '. Dividends Received Before the stock split Received in 1963$ 1.00 After the stock split Received in 1964^ .50 Received in 1965 .50 Share equivalents X ^1,00 2.00 Dividends Received $3.00 Cash Return 6.00 Actual rate of return on investK.ent The determination of the actual rate of return on investment is accomplished by placing cash return over averaged (unadjusted) investment cost. In the evaluation criterion section, there is no necessity of requiring all data to be consistent vdth a share held at the end of the investment period. Cash return$ 6.00 Cost $45.00 ~ .1333 or 13.33 percent Cash return vath time-flexible decisions To accomodate possible questions as to the arbitrary selection of decision dates, it was decided earlier to develop data vdth respect to the validity of the prime assertion v.iien the "good time to buy" and "good time to sell" behavu.or observations were operative. In this case, the projection of future earnings is the sam.e as before. The only change is in the evaluation criterion. Now, the determination of cash return will allow the purchase price to be the lowest possible prevailing market price during the two month period surrounding the investment decision date. The sale value will be the hi^est possible PAGE 87 Â• 80 market price prevailing in the t\fo month period surrounding the sale date. The originally assumed hypothetical data indicated a low price in the two month period surrounding the puchase date at$ 40.00 and a high price in the two month period surrounding the sale date at an unadjusted $26.00. Since two shares are held at the end of the investment period for theone originally purchased at the investment decision date, the$ 26.00 will have to be doubled, viiich yields a sale value for the investment of $52.00. The computation of the rate of returri allowing for time flexible decisionsfollows. Cash Return Capital Gain Sale value of the investment$ 52.00 Cost of the investment 40.00 Capital Gain $12.00 Dividends Received ^ Same as before 3.00 Cash Return$ 15,00 Rate of cash return ~~~~~~~~ ; * . Cash Return _ $15.00 Cost "$ 40.00 ~ '^^^ ^7.5 percent Summary , The above illustrations reflect the procedures undertaken to gather relevant data for the rank correlations which will follow later in the study. A summarized statement of those rank correlations follovre. All elements represent arrays stated in the order of highest to lowest values. It will be recalled that the 1951 1962 trend period has a sample size of forty-one whereas all others have a value of one-hundred-ninety-eight, .

PAGE 88

Projected earnin.E^s per share: From 1951 1962 trend with Actual EPS 1963 1965 From 19^6 1962 trend vdth Actual EPS 1963 1965 From XyOU i.yo
PAGE 89

82 correlation should be quite high. If the future earnings do not follow the established trends of the past, a low or negative rank correlation coefficient should be forthcoming. If decisions based on the relative attractiveness of projected earnings as a percent of investment cost are valid J a high rank correlation coefficient shoxxld be indicated for the analysis which rank correlates projected price-earnings ratios with actual cash retiu:'n. The details of the rank correlation procedure are outlined below. Only one rank correlation will be demonstrated since all are procedurally identical. For the illustration, projected price-earnings ratios will be used and rank correlated with actual cash return. To have the data needed for the rank correlation analysis, the data obtained in the hypothetical stock above will be supplemented with assumed data for nine additional stocks. The assumed data follov:. Projected price-earnings ratios Actual rate of cash From trend 1951-1962 return Stock Number 1 ,0501 . .1795 Rank correlation analysis does not use a relationship between the values above for each share. The relationship indicated in a rank correlation coefficient is a relationship between the assigned numbers_ for the stocks. The two arrays for the rankings follow. The hypothetical stock in the illustrations in this chapter is stock number .2. 2 3 4 5 6 7 8 9 10 .0312 ,0292 .0280 .0215 .0210 .0205 .0200 .0198 ,0176 ,1206 .1333 .1222 .1262 .1200 .1010 .0937 .0998 .0930

PAGE 90

83 Projection Cash Return RanlcLnff RaTikins 1 1 2 3 3 . 5 4 U 5 2 6 6 7 7 8 9 9 8 10 10 The formula for rank correlation is ^ = 1 6(Y-X)^ 6(16) n(n2 1) 10(102-1) (Y-X) (i-x)2 0 0 -1 1 -2 4 0 0 3 9 0 0 0 0 -1 1 1 1 0 0 _0 2 1 0.146 = 0.90301 The coefficient of rank correlation in the above case is .903, which would substantially uphold the validity of the prime assertion. An answer of 1.0 would have indicated perfect positive rank correlation and an answer of -1.0 would have indicated perfect negative rank corre12 lation. The rank correlation above indicates that if stocks were chosen on the basis of projected price-earnings ratios which were determined from an extrapolation of the trend of past earnings, the investnent decision would be a good one. The predictive potential of past earnings has been established for the hypothetical case. This, however, is a dry run. Chapters IV and V will deal with the outcomes when actual data are treated. 12 The rank correlation formiOa is derived from Stephen P. Shao. pp. 69V697.

PAGE 91

CH/VPTER IV PREDICTION OF EARNINGS By systematically dismissing the possibility of dealing effectively vath a critique of income reporting through recourse to the degree of correspondence between cone eptxiali zed "essences" of income aiid operational con^^utations of the accountant, the burden of justification for the reporting of financial net income to third parties was placed on viiat may be termed "outcomes." That is, if net income computations are to "make sense" in the context of the annual report in a manner consistent with the assertions which are made of it, such justification must come from the demonstrated utility of net income computations in shareholder decisions. It is not possible to defend income reporting by recourse to an assertion of correspondence between the accountant's computation and the co-called concept of "economic income." Confusion in this area was hypothesized to have stemmed from an exclusive reliance on "concepts" in income theory. Operations were directed by concepts in blind concordance and the quality of such operations was determined to rest entirely on the closeness of the operations to those viiich were suggested by the formiilated concept. Concepts were assigned the task of determining the qualitative status of operations, and not vice versa. Ostensibly the only quality requirement for the concept came by v^ay of general agreement that the concept was indeed a "true" and reliable statement of observation. With the burden of proof, as it were, of the advisability of 64

PAGE 92

85 reporting income figures to shareholders shifted from purely conceptal argument to correspondence between asserted use and factual indications as to v/hether the assertion is valid, a concrete set of tests mi^t be undertaken. Ivhen the indications are analyzed, there will be a data-based indication of the warranted assertability of the claim that is made of income reporting. The basic question now is, does income reporting factually do what is claimed of it. There are many such claims of the utility of income reporting. The one thought to be the most frequently encountered and the one which appears to represent a consensus of importance was isolated for study in this analysis. Essentially, this claim asserts the utility of past income trends in predicting future income trends thereby lending to prospective investors valuable information on which to form an opinion on the investment potential of alternative investment possibilities in common stocks. This is the single claim under investigation here. The extent to v/hich income reporting assists credit grantors in their function is altogether another matter. Likewise, the utility of income reporting in determining the legality of dividends is not xonder review. We are interested in a critique of income reporting for shareholders. ' Chapter III set forth the details of how this claim was to be tested. There were two questions involved. The first v/as whether income amounts to be realized in future periods could, in fact, be predicted on the basis of the indications of trends in income from past periods. This chapter will deal exclusively vri.th this first question. Chapter V will take up the question of v;hether the relatively better stocks in factual performance v.d.11 be chosen if this decision base is used.

PAGE 93

&6 We mst first know, hovrever, v^iether earnings can be projected. Vfi-thout an affirmative answer to this question, the primary question of decisLon efficiency vail have a rather dismal prospect. Yet, an affirmative answer to this question at hand need not indicate any relevance of past earnings to the choice of good investments in common stocks. Earnings may be precisely predictable without having any relevance to the decision base for investment. There is reasonable ground to suspect that this is indeed the case. If earnings were susceptible to projection, and every investor had access to this same information, would not the market price of the attractive stocks be expected to be bid up or down consistent with an acutal cash retvmi equating itself to common interest in the long run? This would appear to be a reasonable expectation. Moreover, many investors readily conceed that the prospects of a certain stock are determined more on the basis of what the average investor's opinion is expected to be rather than on the basis of viiat the operating efficiency of the related corporation is expected to be. Also, if an investor is in the market for short term capital g^ins, there is serious doubt that financial statements will be of any use viiat soever. If he is in the market for longer term growth and sustained dividends, there would be some question as to the relevance of past income trends in that changes in technology are taking place so rapidly that the technological advantage (vrhich in many cases yields an economic advantage) is not held for very long. Future income perfoi^nance in these cases would be thought to be determined more on the basis of a constant research effort than on the basis of the continuance of economic advantages held in the past. These are but a few of the reasons why the prime assertion is thought to be generally someviiat overstated.

PAGE 94

. S7 V/e vri.ll initially set forth what the data indications are with respect to the predictability of earnings for all stocks . Secondly, we will make identical anatlyses with stocks grouped in accordance with the established industry classifications. All Stocks Analysis Rank correlation coefficients obtained from two arrays on the basis of the relationship between projected earnings per share and actual earnings per share are presented below. The tests of significance were run via an estijnation of the population rank correlation coefficient at a confidence limit of 95 percent. On the assun?)tion that the basis of sangjle selection did not contribute a material bias to the analysis, the indicated population rank correlation coefficients are stated with 95 percent confidence. That is, there are 95 chances out of 100 that the population rank correlation coefficient will fall between the upper and lower indicated interval limits. TABLE I Projected Earnings Per Share Compared With Actual Earnings Per Share Trend Period Sample Size Rank Correlation Coefficient Population Upper Limit Population Lower Limit 51 62 41 .556625 .655625 .457625 56 62 198 .327217 .466769 .187665 60 62 198 .432938 .572490 .293386 The rank correlation coefficients do not indicate a very good degree of relationship between projected earnings per share and actual earnings per share. The best coefficient as well as the best population interval estimate is obtained from the trend data from 1951 -

PAGE 95

88 1962. The general observation sometimes encountered to the effect that the longer the length of past earnings periods, the greater the weight to the projection appears to hold little merit. The longer of the two remaining trend periods, however, has the poorest record of all. The best is the 1951 1962 trend period and the worst is the 1956 1962 trend period. With maodLmum population rank correlation estimate, the Â• coefficient is .655625, vrfiich appears to be indicative that the statements of assertion heretofore made with respect to the predictive potentials of income trends are, perhaps at best, only vaguely correct. There does not appear any reason to believe that such trends have the so-called obvious relevance to shareholder decisions. Moreover, since we faced a data constraint when trying to gather sufficient data to determine these rank correlation coefficients for the 1951 1962 trend period, it would appear qxiite reasonable to believe that prospective shareholders in general are also constrained J^y this lack of data. At best, they would have been able to gather sufficient data for only forty-one of the major American corporations. The importance of this highest coefficient is diluted because of the lack of ability to put the relationship to practical use. Yet, all rank correlation coefficients are positive; indicating some degree of correlation. The question now becomes, are we to presvime that that degree of correlation which is evidenced is sufficient to make rational and good investment selections from among common stocks, \iith. the positive correlation coefficients demonstrated to exist, can an investor make good decisions? Is the relationship positive enough? Clearly, an investor would not invest soley on the basis of the anticipated earnings. A share of stock generally costs something.

PAGE 96

89 Consequently, investment decisions, to the extent that they are made on the basis of predicted earnings from past trends, would seem to be made on the basis of projected price-earnings ratios. In other words, qualify the futvire projection of earnings by ciu-rent investment cost and you have the hypothesized decision base. The attractiveness of an investment alternative is not a single matter of determining the projected earnings of that investment. An out-of-line investment cost woxild surely diminish the attractiveness of even the highest projected earnings. Table II will demonstrate that when the added con^jlexLty of investment cost is included in the analysis, the rank correlation coefficients deteriorate still more. ^ TABLE II Projected PE Ratio Compared Vdth Actual PE Ratio Trend Sajxrale Rank Correlation Population PoDulation Period Size Coefficient Upper Limit Lower Limit 51 62 56 62 60 62 41 198 198 .420693 -.007269 .184763 .519693 .132283 .324315 .321693 -.146821 .045211 With confidence limits again set at 95 percent, the indicated maximom population rank correlation coefficient is found to be , 519693 Â• The lower limit is -.146821. The projection of price-earnings ratios would appear to be the rational basis for decision if past earnings trends are thought to indicate things to come.. With a sanple nm y inn^ ra correlation of .420693, the least which might be said is that for the one-hundred-ninety-eight stocks of the san?3le, there is very little

PAGE 97

90 evidence that investment decisions made on the basis of extrapolated past trends (qualified by investment cost) are as valid as they are thought to be. Moreover, in the shorter period (1956 1962), the sample rank correlation coefficient is negative. This is evidence that the prospective investor is better off vdthout this so-called obviously relevant information. , ; Since the ostensible purpose of cirranging projected price-earnings ratios in the order of highest to lowest is so that the better "looking" stocks vdll be grouped tovrard the top of the array, it mi^t be interesting to see what the relationship between the five highest and five lowest factual stocks is in comparison with how they were predicted to turn out. The following table will assign numbers one through five to the highest factual stocks in predicted earnings per share and the last five digits for the lowest in projected earnings per share (37 through 41 when "n" r 41, 194 through 198 when "n" = 198). Table III will indicate these relationships for earnings per share figures and Table IV will indicate them for price-earnings ratios. TABLE III Actual Rank Compared With Predicted Rank Earnings Per Share Actual Top Five In Rank Highest to Lowest Predicted Rank Actual Bottom Five In Rank Lowest to Highest Predicted Rank 1 2 3 4 5 1951 1962 (n = 41) 31 41 3 40 15 39 8 38 6 37 40 a 38 39 4

PAGE 98

91 T.'iBLS III Continued Actual Rank Compared With Predicted Rank Earnings Per Share Actual Top Five In Rank Predicted Actual Bottom Five In Rank Predicted Highest to Lovrest Rank Lowest to Highest Rank 1956 1962 (n = 198) 1 7 198 17 2 196 197 27 Al 196 193 Â•f 9 195 188 c 7 50 194 76 I960 196 2 (n = 198) 1 56 198 39 2 . 1 197 12 : 3 llA 196 196 179 195 169 5 120 194 167 In the 1951 1962 analysis, the best that might be said is that the five factually highest stocks were, vdth one exception (stock number one), all vdthin the top half of the predicted rankings. Likevri.se, the five least favorable stocks in factual earnings performance, with one exception (stock number thirty-seven), were all in the bottom half of the predicted rankings. In the 1956 1962 analysis, this record deteidorates to one exception in the top five and three exceptions in the bottom five. In the I960 1962 analysis, it is worse still. In this time period, there are three exceptions in the top five

PAGE 99

/ .-92 and tv.^i exceptions in the bottom five. These data indicate that earnings of past periods, particularly when projected and used in interfirm projections of relative attractiveness of anticipated earnings, clearly do not have the predictive potential that they are commonly presumed to have. Another, and perhaps an even more interesting analysis, is to see ho"w well the upper 25 percent in predicted earnings factually turned out. Data indicate that if one were to choose from among the top 25 percent of stocks as determined by an array in the order of highest to lowest projected earnings per share, the 1951 1962 trend period would have rendered to such an investor a 50 percent chance that the chosen stock would turn up in the factusilly highest 25 percent. Of the ten stocks (25 percent of 41 = 10) with the highest predicted earnings per share, five were among the factually highest ten. In the 1956 1962 trend period, this percentage drops to 40 percent. Only 40 percent of those stocks predicted to be in the upper 25 percent actually vrere in the upper 25 percent. Again, as the time period shortens, the relationship becomes worse. In the I960 1962 trend period, this percentage drops to 36 percent. Of the top 50 stocks (25 percent of 198 z 50) predicted to have the highest earnings per share, only 18 of them turned up in the factually highest 25 percent of stocks. The above indications of relationships represent a measure of the effectiveness of predicting future income amounts based on the extrapolation of past trends. When the relevant factor of investment cost is added, the presimed basis of investment decision is determined. Table H indicated the rank correlation coefficients and population estimates based on intervals. It would be interesting to know how the top and

PAGE 100

93 bottom five stocks performed in this category. Table IV contains this information. It is organized the same as Table III except that the relationships are between projected and actual price-earnings ratios instead of earnings per share. TABLE IV Actual Rank Compared Vfi.th Predicted Rank Price-Earnings Ratio Actual Top Five Actual Bottom Five Try Pnnlr Predicted In Rank Predicted Hi/3;hest to Lowest Rank Lowest to Highest Rank 1951 1962 (n = U) 1 32 Al 40 2 6 40 37 3 9 39 4L 4 3 38 1 5 10 37 8 1956 1962 (n = 198) 1 1 198 198 2 196 * 197 197 3 29 196 22 64 ' 195 186 5 94 194 193 I960 1962 (n = 198) 1 162 198 193 2 5 197 176 3 42 196 4 4 82 195 195 5 76 194 74

PAGE 101

For the most part the projection of the relatively poorer stocks appears to be quite reliable. It is unfortunate that investment decisions cannot be made on this basis. The performance of the predicted top five stocks in the 1951 1962 trend period appears to be the best of the three periods. In each period^ there is only one exception to the rule that all predicted top five stocks factually turned up in the top half of the factual performance array. This is the most favorable comment which might be made. It is doubtfvil that this vrould offer any great reassurance to a prospective shaireholder considering an investment. If there are other indicators of good investment stocks, and this extrapolation of past income trends appears to support those indicators, then perhaps this methodology (and hence, a qualified version of the prime assertion) is useful in such cases. The indications of this part of the study, however, reflect evidence that to the extent that such a process as this is considered, the consequence shoxild not be expected to be too favorable. For the most part, it would seem reasonable to suspect that there vovHd be other factors quite a bit more important in investment decisions than the relative quality of past income trends. The evidence above is interpreted as indicating that however small a weight is assigned to the quality of past income performance, that segment of the investment decision can be expected to be at best , only vaguely sound. ; . Â• It would be interesting once again to see what the chance would be for having a stock in the upper 25 percent of the predicted array txim up in the upper 25 percent of the factual array. Of the ten stocks with the highest predicted price-earnings ratio, five turned up with the factually highest 25 percent, a percentage of 50. For the trend period 1956 -

PAGE 102

95 1962, only 14 percent of the predicted highest 25 percent actually appeared in the factually top 50 stocks. The trend improves slightly for the i960 1962 trend period. In this case the relationship improves to 25 percent. Once again, these data are somewhat less than encouraging to those who would assert an obvious relevance of past earnings to investment decisions, particxilarly when such past earnings are thought to yield indications of the relative attractiveness of alternatives. Perhaps the most damaging evidence of all is the data with respect to the mean actual earnings per share achieved as a percent of projected earnings per share. For each stock the actual earnings per share (the average of the three years in the projection period) was con^juted as a percent of the projected earnings per share. The mean of all such percentages is indicated in Table V. If there is a tendency for actual earnings per share to follow the established trends of the past, we woiild expect a mean percentage figure near 100 percent with a rather small standard deviation. Since we are always interested in the relative performance of all stocks, even a mean percentage significantly different from the 100 percent figure would not be damaging to the prime assertion. For example, if one were to make investment choices on the basis of the relative attractiveness of past earnings, then even if past earnings are not attained the decision criterion wo\xld be valid if the economic conditions causing the deviation from the iieal 100 percent figure acted in the same direction for all stocks. If an unusiial period of prosperity were to cause all stocks to lift to^ say, 130 percent of the anticipated earnings as derived from the extrapolation of past trends, the decision criterion wouli still be valid. In such an event, the relatively better stocks will still have been chosen. Therefore, the mean which

PAGE 103

96 is indicated in the table below is not nearly as important as the stardard deviation. Regardless of the mean performance of alL stocks, so long as the standard deviation is small, the decision criterion is upheld. This will indicate to us some measure of the performance of all stocks in attaining the predicted earnings per share amounts. Table V follows. TABLE V Mean and Standard Deviation Actual ii^amings Per Share as a Percent of Projected Earnings Per Share Trend Period Mean of Actual Earnings as a Percent of Projected Standard Deviation 51 62 170.655 393.268 56 62 139.789 1,234.183 60 62 69.682 A28.772 The standard deviations indicate that the mean in each case is aH but meaningless. Not only is there no evidence that past inccme trends are continuing in the future, but more importantly, there appears to be no percentage which would appropriately be thought of as representing the performance level about vhich factual performance tends to gravitate. The individual stocks deviate from the established means so widely that the established mean, in fact, is not a representative percentage describing average performance. With standard deviations running at best twice the mean and at worst almost nine times the mean, it appears to be a fair conclusion to state that past earnings trends extrapolated into the future is not an adequate decision criterion, in whole or in part, for choosing the relatively more attractive stocks in future income performance. To the extent that a prospective investor relies on the past earnings performance of various stocks

PAGE 104

97 as a basis for indicating to him which of the stocks will have the relatively better future earnings performance, it would appear that his decision criterion is extremely poor. In fact, past earnings do not extend themselves into the future in a predictable way. If a prospective investor is thought to rely on such a procedure to the extent of about 10 percent of his total decision weight distribution, then we may conclude that 10 percent of his decision will be quite poor. It is clear that the earnings trends of the past have no such obvious relevance to shareholder decisions. It should be noted that this entire analysis presumes the continuance of financial net income computations in their same approximate manner. Any material change in accountLiig procedvire would require another such analysis to determine if the change in accounting procedure tended to yield past trends a better indicator of things to come. Under the presumption that the accountant will not materially change his procedure, the conclusion is that there is no basis for believing that past income trends are efficient indicators of future income trends. . ' The above analyses have been concerned with all stocks without regard for whether the asserted usefulness of past income trends is any more true on an intrai ndustry basis. We now move to an analysis of industry grouped stocks. . " 'Â• Industry Group Analysis The major question to be asked in this section is whether the asserted utility of past income trends in predicting futxire income trends is any more va l id when considering stocks in the same general industiy than it is when considering all common stocks as investment alternatives. We will deal, initially with the projection of earnings per share and .

PAGE 105

then vdth the projection of price-earnings ratios. Table VI vail present the derived rank correlation coefficients between industryarrays of projected earnings per share and actual earnings per share. Since the sample size for the 1951 1962 trend period is only fortyone, there vdll be times when a rank correlation coefficient would be meaningless because of the small number of stocks in a given industry. This accounts for the notation , "Not Available" in the table below. TABLE VI Rank Correlation Coefficients Actual Compared V/ith Pro.iected Earnings Per Share Stocks Grouped by Industry 1 Indus tr:y Group Trend 1951 1962 Trend 1956 1962 Trend I960 1962 1 Not Available -.5000 .5000 2 Not Available .4909 -.3091 3 Not Available -.8000 .4000 .4 .9000 .9701 .5988 5 .5000 -.1081 .7500 6 Not Avai lable .2000 1.0000 7 .5000 .1546 .5636 8 1.0000 .4788 .8303 9 Not Available .8000 .6000 10 -1.0000 .1455 .1818 11 Not Available .4000 -.8000 12 1.0000 .6000 .8000 13 1.0000 .3168 .3388 lU Not Available .4857 .6571 15 Not Available .4286 -.6000

PAGE 106

It may be observed that approximately one-fifth of all industry grouped rank correlations are negative and that approxinately one-half of all rank correlations are under .50. This is consistent with the poor record past earnings trends had in projecting future earnings in the all stocks analysis. This does not answer the primary question being asked in this section, however. We need a measure of the extent to vAiich this procedure yields more valid data for decision purposes on an intraindustry basis than for all stocks. Table VII will show the differences foimd between the average rank correlation of all industry grouped stocks and the rank correlation of all stocks industry grouped, the latter figure representing a measxare of the rank correlation of all stocks together. If this process yields better data on an intraindustry basis, we should expect that the average rank correlation of individual industry rank correlations vrould be significantly higher than the full rank correlation of all of those same stocks taken as a whole. In other words, the predictive ability of past earnings should be better on an industry basis than on an all stocks basis. Table VII follows. TABLE VII Interindustry Rank Correlations Compared '.'ith Intraindustry Average Rank Correlation Earnings Per Share Tr^nd Interindustrv Intraindustry Ax'erage Period Rank Correlation Rank Correlation 51 62 .8035 .8428 56 62 .2670 .2709 60 62 .3524 .3674

PAGE 107

100 It was concluded earlier that past income trends are very poor predictors of the relatively more attractive stocks in terms of future income performance. The data in Table VII indicate that while the assertion of predictability is more valid (very slightly more valid) for intraindustry analyses, that improvement which is indicated is clearly not of the magnitude suggested by those proponents of this procedure. To answer the question asked in this segment of the study, yes, the prime assertion is more valid for intraindustry analyses. The improvement, hovrever, is so very slight that there can be placed no significant weight on the applicability of the prime assertion as it applies to intraindustry analyses. It should be noted that the validity of the second belief in income does not stand or fall on the basis of findings in this chapter. The prospect for finding support for it in data is extremely diÂ£>mal in view of the findings here. We have concluded that earnings of the past are not efficient predictors of the relatively more attractive stocks as far as future earnings are concerned. But the precise question of whether this second belief in income is warranted or not does not depend on the ability of past trends to predict future incomes. It depends on the ability of projected price-earnings ratios to predict the better stocks in cash return performance. We simply conclude here that one of the tvro most important inputs to this main procedure is not predictive. This clearly does not speak well for the prospect of a favorable conclusion in Chapter V, but we must develop the data to be satisfied. One might not be able to predict future earnings from the trends of the past, but he might be able to choose the relatively better stocks from the earnings trends of the past. Chapter V will handle this possibility.

PAGE 108

CHAPTER V _ Â• PREDICTION OF CASH RETURN With the unfavorable evidence unfolded in Chapter IV, we come to this, the heart of the second belief in income, vdth a rather poor prognosis. If the rank correlations appear to be worse in the cash return analyses, we laight well attribute the increasingly diluted validity of past earnings as efficient indicators of the relative attractiveness of investments to the disturbing effect that variations in market prices have on the prediction of the better stocks. If the rank correlations are better, we will have to tentatively presume that this methodology is better than expected (how much better will depend on how much better the rank correlations are) in spite of the inability to accurately predict future earnings from past trends. One possible solution to this dilemma would be that past earnings are related to future market prices regardless of the lack of susceptibility of past earnings to projection. This is only a possibility. On the facv^ of it, one would not place very much hope for favorable evidence to support this alternative. We stipulate at the outset that if the rank correlations in this segment of the stiody are better' than those associated with the predictions of earnings, we will have encountered a serious question which we will not only be unable to answer, but also one which will tend to jeopardize the prospect of arriving at any conclusion. If the rank correlations in this segment are better than those before, we will, in

PAGE 109

102. effect, be indicating the means lack a susceptibility to projection but the outcome is valid. This opens up the next logical area for analysis which is the susceptibility of predicting future market prices throu^ the extrapolation of past earnings trends. There was already evidence on hand that pointed to the futility of taking this approach. Stock prices rose at an annual average of 10.2 percent from 1949 to 1965, while per share earnings have been rising at an annual average of only 3.2 percent in that same time span. This would suggest that there was something significantly more important than prior earnings 1 , performance in the prediction of future market prices. For this reason, among others, we chose not to provide for this analysis in the study. Therefore, if the rank correlations in this segment are bditer tha-i those associated with the prediction of future earnings, we will have come face-to-face with evidence contrary to the indications. This situation will have to be interpreted as a fxirther indication of some degree of methodological error or statistical bias in this study. If, on the other hand, the rank correlations in this segment are somewhat worse, the data indications will be both consistent with the prior analyses of this stiidy as well as consistent with other data. The rate of cash return, it will be recalled, was computed on two different bases. The first was dividends plus capital gain over averaged investment cost. The second, allowing for the "good time to buy and good time to sell" behavior observation, will be computed on the basis of dividends plus the highest possible capital gain over the ^J. Fred Weston and David K. Eiteman, "Economic Trends and Security Values A Bleak or Bountiful Future for Investors? " Financial Analysts Journal. February, 1965, p. 2?.

PAGE 110

103 lowest possible investment cost for the two month period surrounding the investment decision date. Rank correlation coefficients were determined between an array of predicted price-earnings ratios and an array of actual cash return (cash return being interpreted initially in the conventional manner ard, secondly, utilizing time flexible decision data). The rank correlation coefficients, together with population interval estimates, will afford us indications as to whether factually better investments can be made if the basis for decision is the relative attractiveness of projected price-earnings ratios. This will represent the first attempt, to our knowledge, that evidence will h^ve been gathered which was directly intended to test this asserted utility of financial net income reported to shareholders. The first analysis is the rank correlation coefficients obtained when price-earnings ratios (decision criterion) were rank correlated with actual cash return (evaluation criterion). TA3LE Vlll Projected Price-Samings Ratio Compared With Actual Cash Return Trend Period Sample Size Rank Correlation Coefficient Population Upper Limit Population Lower Limit 51 62 41 .391060 .490060 .292060 56 62 19S -.097256 .042296 -.236808 60 62 198 .146064 .285616 .006512 Assuming that the limitations in sampling already mentioned have not contributed a material bias to the data, the population rank correlation coefficient will be between the upper and lower confidence limits with a confidence level of 95 percent. Since these coefficients

PAGE 111

are substantially lower than those obtained in ranking projected and actual earnings per share, there is some evidence that market prices play an important part in invalidating the asserted utility of past income trends in investment decisions. There appear to be other sets of factors more important than the earnings history of respective corporations in setting market valuations. These observations, of course, are purely possibilities. The data indications point in their general direction -without conclusively demonstrating their accuracy. The main point is the stedily worsening rank correlation coefficients. We were able to predict the relatively better stocks as far as future earnings are concerned only vaguely. We were able to predict factual price-earnings ratios only poorly. It appears now that we are unable to predict the relatively better stocks in actual cash return. For the majority of corporations, data will be available to prospective investors only for relatively short periods of time (from 1956 to present). The maximum number of stocks which we were able to accumulate relevant data for when the time period was stretched longer was forty-one. From 1956 onward, we were able to gather relevant data for one-hundred-ninety-eight. The rank correlation coefficients for these periods (subsequent to 1956) gives clear indications that past earnings trends are of no assistance in determining the relatively better cash return stocks. The maxinivim rank correlation coefficient escpected would be the upper limit of .285616 in the I960 through 1962 trend period. The upper limit of .490060 for the 1951 through 1962 trend period is not good to start with, and when the added problem of data constraint is considered, its significance decreases still more. There is evidence that if prospective shareholders had access to relevant information with vMch to make

PAGE 112

105 longer trend axialyses, perhaps the assertion of utility of past incomes might take on greater validity. This is currently an iinknovni. Even the slight evidence of this possibility should be carefully vreighed in light of the rather small sample available for the 1951 through' 1962 trend period. It might be instructive to note the positioning of the factually five highest and loxvest stocks in cash return in contrast with the predicted positioning these shares were given. Table IX vd.ll present this data, it appears on the following page. ' It will be noted that for all five of the highest stocks in cash return in each of the three trend periods, the predicted price-earnings ratio ranking is in the upper half of that ranking, with one exception in each case. The relationship for the lowest five stocks is not that good. For the trend period 1951 through 1962, there are no exceptions, all predicted rankings are in the bottom half. In the 1956 through 1962 trend period' there are two exceptions (numbers 196 and 195). In the I960 through 1962 trend period there are three exceptions. In this case, the exception becomes the rule. While the quality of the rank correlations degenerates in moving from the prediction of earnings to the prediction of cash return, the quality of the relationships among the five highest factual cash return stocks and their predicted rankings improves. We shall not place any significance on this development because the major purpose of illustrating these relative positionings was to indicate where one would have had to go to find the better cash return stocks in his array of projected price-earnings ratios. It appears that one would have a wide choice from among the top half of such an array. It is highly doubtful that this prospect adds greatly

PAGE 113

106 TABLE Actual Rank oÂ£ Cash Return Compared V/ith Predicted Rank of Price-Earnings Ratio Cash Return Actual Top Five PE Ratio Cash Return Actual Bottom Five PE Ratio In RarJc Â• Richest to Lowest Predicted Rank In Rank Lovrest to Highest Predicted Â• Rank 1951 1962 (n = 41) 1 9 41 J7 2 17 40 37 38 39 28 14 38 24 5 21 37 36 1956 1962 (n = 198) 93 198 190 2 80 197 151 3 71 196 49 4 61 195 ' ^ 22 5 156 194 176 I960 1962 (n = 198) 76 193 I 3 12 197 6 3 21 196 100 4 15 195 4 5 107 194 121 to the ease of investment decision. In short, we have demonstrated that evidence does not support the argument of those who assert an "obvious relevance" of past earnings trends to investment decisions. To be ^ore than fair, and to be cognizant of the stipulated limitations

PAGE 114

. 107 of this study, particularly with regard to sampling technique, vre hold that the evidence vdll support an assertion of vague relevance of past earnings performance in the selection of the better investment stocks. This appears to be justified in view of the fact that most of the rank correlation coefficients thus far in the study have been at least positive . That past earnings have the obvious relevance ascribed to them by most writers in financial accounting appears to be a gross overstatement of the case for income determination for annual report purposes in view of the findings of this study. One is even tempted to say that there is boxond to be a set of factors more reliable than past earnings in investment decisions in the light of the rank correlation coefficients determined here. If one were to give 100 percent weight to the extrapolation of past income trends in choosing his investments, his prospects would be dismal indeed. It would likewise seem that one's investment decision would have a poor prospect to the extent that past earnings were relied on to yield indications of the better stocks. If one were to select his investments from among the top 25 percent of stocks with predicted price-earnings ratios determined from an extrapolation of the earnings trend of the period 1951 through 1962, there would be a 20 percent chance that the selected stock would turn up in the highest 25 percent of cash return stocks. Of the top 10 stocks (25 percent of Al) in predicted price-earnings ratios, only 2 turned up in the highest 25 percent of stocks in cash return performance. For price-earnings ratios determined for the earnings trend of the period 1956 through 1962, the percentage is 5For the 19601962 trend period, 32 percent of the predicted highest were among the 50 factually highest stocks. Chapter IV indicated that the related

PAGE 115

108 percentages for relationships between projected earnings per share and actual earnings per share were 50, 40, and 36 for the trend periods beginning in 1951, 1956 and I960 respectively. These percentages, when compared with the 20, 5, and 32 for the current analysis, indicate a serious deterioration in relationship for the 1951 1962 and 1956 1962 trend periods. The I960 1962 trend period data did not change significantly. For the most part, these data reflect a progressively worsening predictive capacity of the data when additional relevant variables are introduced. The factor of investment cost was added and the rank correlation coefficients and percentages went down. The above analysis interpreted cash return as the summation of dividends and capital gain (determined consistent with the averaging procedure used for purchase and sale values) divided by investment cost. VJe now interpret capital gain as the difference between the lowest price a particular share sells for within the two months srirrounding the investment decision date and the highest price in that period. This will accomodate those who observe that viiatever decision criterion is used for the selection process, stocks are generally bought and sold consistent vdth good market prices. Table X will present the rank correlation coefficients when cash return (as interpreted above using time flexible decisions) is rank correlated with the same array of predicted price-earnings ratios. Table X appears on the following page. The rank correlation coefficients tmder this new interpretation of capital gain do not appear to be materially different from prior ones. It might be noted that in this case, as in all other rank correlation analyses of this study, the 1951 through 1962 trend period enjoys the best correlation coefficient, the I960 1962 and 1956 1962 trend

PAGE 116

109 TABIÂ£ X Projected Price-Earnings Ratio Compared With Actual Cash metum Utilizing Time Flexible investment Decisions Trend Period Sample Size Rank Correlation Coefficient Population Upper Limit Population Lower Limit 51 62 41 .388507 .487507 .289507 56 62 198 -.112201 .027351 -.251753 60 62 198 .155070 .294622 .015518 periods follovdng in that order. This is a phenomenon for which we cannot offer an explanation. It is conceivable that cyclical fluctuations of the late 1950 's caused this pattern. Yet, that trend period is clearly long enough for such economic fluctuations to reflect their effects throughout the gamut of all stocks. It vri.ll be recalled that any external variable mast affect rankings , not simply the performance of all stocks. . The observed rank correlation coefficients have not materially changed \mder this new interpretation of capital gain. Hence, the same comments would apply as before. The assertion under test does not appear to be any more valid under this interpretation of capital gain than before. Table XI on the following page will present the positioning of the highest five stocks in cash return under the nev/ capital gain interpretation with respect to how they were predicted to rank. In general, the upper five stocks all came from the top half of predicted rankings (there is one exception in each trend period). The lower five stocks in factual cash return performance were all within the lower half of the predicted rankings for the 1951 1962 trend period. For the-

PAGE 117

110 TABLE XI Actual Rank of Cash Return Compared V.'ith Predicted Rank of Price-Earnings Ratio Utilizing Tinie Flexible Investment Decisions Cash Return Actual Top Five In Rank Highest to Lovrest PE Ratio Predicted Rank Cash Return Actual Bottom Five In F^ank Lowest to Highest PE Ratio Predicted Rank 1951 1962 (n = kl) 1 : 9 41 39 ' 2 17 40 37 : 3 38 39 28 14 38 24 21 37 : 36 1956 1962 (n = 198) 94 198 190 2 80 197 151 3 71 196 *Â• 49 4 60 195 22 5 156 ' 194 176 1960 1962 (n = 198) 1 76 198 6 2 21 197 ' " 3 3 12 196 4 15 195 98 107 194 158 1956 1962 trend period there are tvro exceptions and for the I960 1962 trend period there are four exceptions.

PAGE 118

Ill Of the 25 percent of the highest stocks in predicted cash retiim, ^ only 20 percent tiimed up in the factually highest 25 percent in the 1951 1962 trend period. Only 6 percent of the predicted highest 25 percent turned up in the factually highest 25 percent in the 1956 1962 trend period. In the I960 1962 trend period, 24 percent turned up in the highest 25 percent of factual cash return. These indications are somewhat less than consistent vdth assertions which are made on the utility of income reporting in interfirm and interindustry investment decisions. The following discussion will present data on how well thisasserted utility stands up on an intraindustry basis. Basically, a grouping of stocks on the basis of the industry classifications presented in Chapter III does not yield data which would support the claim of the utility of past income figures to a degree ^ any greater than that already established on an all stocks basis. In , most cases, the rank correlation coefficients obtained from a conglom1 erate of all stocks vdaich were previously industry grouped appesirs to j be better than the average rank correlation coefficient of those same stocks when grouped by industry. If past income trends factually \ indicated the better cash return stocks, and this indication potential were more valid on an intraindustry basis (as it is often claimed) than on an interindustry basis, we would expect that the average of the rank correlation coefficients from industry grouped stocks would be significantly better than a rank correlation coefficient obtained from all such stocks taken as a whole. This is interpreted as the implication of the assertion of the greater relevance of the prime assertion on ^ an intraindustry basis. If this is a valid interpretation of these implications, the data appear to refute the prime assertion as it applies

PAGE 119

Â•; ! :; . .^ Â• 112 to intraindustry analysis. Table XII will indicate the industry grouped rank correlation coefficients obtained vdien predicted price-earnings ratios were rank correlated with actual cash retxim. Table XIII will then indicate how the average of these rank correlation coefficients compared with a rank correlation coefficient of all industry grouped stocks taken as a whole. TABLE XII Ranlc Correlation Coefficients Predicted PE Ratio Compared VJith Actual Cash Return Stocks Grouped by Industry Industry Grouo Trend 19^1 19^2 Trend Trend I960 1962 1 Â— . 2 Anon 4 .9000 .4524 .6667 5 1.0000 .1071 -.3571 6 Not Available .2000 1.0000 7 1.0000 -.0364 .0636 8 -1.0000 -.4134 .1030 9 Not Available .5000 .1000 10 1.0000 -.1909 .3105 11 Not Available -.6325 .8000 12 -1.0000 .2000 .8000 23 -1.0000 .0440 -.13a U Not Available -.9429 -.4286 15 Not Available -.3714 -.0286

PAGE 120

113 It will be noted that of the total of 37 rank correlation coeffi cients, 43 percent are negative and 73 percent are less than ,5000. The data in Table XIII, hovrever, are more important. TABLE XIII Interindustry Rank Correlations Compared V/ith Intraindustry Average Rank Correlation Predicted Price-Earnings Ratio Rank Correlated With Cash Retiim Trend Period Interindustry Rank Correlation Intraindustry Average Rank Correlation 51 '(^2 .4513 .1286 56 62 -.1470 -.1416 60 62 .2567 .2156 Two of the three rank correlation conqjarisons are, in fact, worse on an intraindustry basis than on an interindustry basis. The one comparison v;hich is better constitutes an insignificant improvement in the first place, and in the second place, both rank correlation coefficients are negative (-.1470 and -.1416). The data indications do not appear to support the argument of those who assert a greater validity of the prime assertion in intraindustry investment comparisons. If anything, the data point to the probability that intraindustry stock comparisons are worse when the basis of comparison is the projected earnings per share qxialified by investment cost as the criterion of investment. Although Table VII in Chapter IV indicated that one is better able to predict a ranking of stocks as to future earnings per share on an intraindustry basis than on an interindustry basis, the improvement was not as significant as it was perhaps thought to have been. In this analysis, we find that sample data indicate that two of the three rank correlation comparisons are opposite from what they were

PAGE 121

114 were presimed to have been. Hence, it is likely that the method by which vase investment decisions are thought to be assisted (projection of the earnings of the past) is not invalid so much because of the inability to predict future earnings from past earnings trends as it is because market valuations do not appear to respond in a logical or predictable manner to reported earnings. We could reasonably conclude that the earnings of the future are vaguely predictable from the established trends of the past. It appears, however, that we are unable to conclude that good investment decisions can be made in whole or in part on such a basis. Those who assert the utility of past income trends in investment decisions lack data support more because of the "irrational" fluctxiations of the market than because of the absolute inability to vaguely predict the esirnings of the future, from the trends of the past. It should be noted, however, that even this prediction of earnings was far less accurate than it was presumed to have been. There is even a question as to whether it was accurate enough to support the conclusion that such predictions were vagiuely accurate . Perhaps this inability to predict would have destroyed the prime assertion without recourse to these additional cash return prediction analyses. We presume this vague prediction potential because all rank correlation coefficients pertaining to the prediction of earnings were positive; indicating some degree of relatedness. When one comes upon deteriorating rank correlation coefficients in the cash return prediction analyses, it is to be presumed that the additional variables of dividends and market valuations significantly impair the predictive potential of this technique. This is not altogether unexpected. Average investor attitudes responding to a vd.de range of events, clearly form a partial determinant of cash return. This range

PAGE 122

115 includes everything from the physical condition of the President to (and perhaps most important in this era) the expected opinion of nutual funds executives and other institutional investors. With factors such as these determining a part of the cash return of listed stocks, it is little vronder that the claims of accountants as to the relevance of past income performance in investment decisions are overstated. Table XIV vdll present the industry rank correlation coefficients obtained when industry grouped stocks were rank correlated as between predicted price-earnings ratios and cash return with time flexible decisions. This analysis will determine the extent to which the better investment decisions will be made if the decision criterion is the relative attractiveness of projected earnings per share qualified by investment cost, and the evaluation criterion includes cash return with capital gain interpreted as the difference between the highest prevailing market value and the lovrest prevailing market value s\irrounding the decision dates. Table XV will compare the average rank correlation coefficient of the industry grouped stocks with a rank correlation coefficient of all such stocks taken as a whole. Again, if the prime assertion xinder review is more relevant in intraindustr/ decisions, then we should expect a higher average rank correlation coefficient from the industry groupings than from the single rank correlation coefficient of all stocks taken as a whole. Table XIV appears on the following page. Once again, we find that, of the total of 37 rank correlations, 43 percent are negative and 73 percent are belov/ .5000. The data in Table XIV were organized so that we might determine the relevance of the prime assertion as it applies to intraindustry analysis under the new interpretation of capital gain. Table XV appears below Table XIV.

PAGE 123

T/iBLB XIV Rank Correlation Coefficients Predicted PE Ratio Compared vJith Actual Cash Return Utilizing Time Flexible Decisions Stocks Grouped By IndustryIndustry Group Trend lypl 1962 Trend 1956 1962 Trend I960 1962 Not Available -.5000 '-.5000 Not Available .1394 .5350 o J Not Available -.8000 .4000 4 .9000 .4524 06667 c t> .0901 -.4505 L O Not Available -.4000 .8000 ( x.uuuo -.0091 .1273 d o Â— l.UUUU .35S7 .1515 Q 7 Not Available .5000 .1000 10 1.0000 -.2273 .3607 11 Not Available -.6325 .8000 12 -1.0000 .2000 .8000 13 -1.0000 .0396 -.1253 14 Not Available -.9429 -.4286 15 Not Available -.3714 -.0286 Â— T.^3LE XV Interindustry Rank Correlations Compared With Intraindustry Average Rank Correlation Predicted Price-Earnings Ratio Rank Correlated With Cash Return Utili::.ing Time Flexible Decisions Trend Period Interindustrv Rank Correlation Intraindustry Average Rank Correlation 51 62 56 62 60 62 ..4558 -.1682 .2730 -.1286 -.1402 .2134

PAGE 124

117 From the data in Table X7, it is apparent that the prime assertion is no more valid \inder the new interpretation of capital gain in intraindustry analyses than otherwise. Two of the three rank correlation, comparisons are opposite from what the case was presumed to be iinder the assertion. The one which did show an improvement, as before, is negative to start with and reflects a very small improvement (-.1682 to -.1402). The sample data indicate clearly that the suggestions of some as to the greater relevance of this methodology in intraindustry investment considerations are without vrarrant. In suxmaxy-y the data indications presented in this chapter tend to refute the prime assertion of the relevance and utility of past earnings performance to investment decisions on interfirm bases. There is seme positive rank correlation between predicted price-earnings ratios and actual cash return (under both interpretations of cash return), but that degree of association demonstrated herein does not warrant the statement that past earnings have an "obvious relevatnce" to shareholder decisions. For one-hundred-ninety-eight major stocks, the rank correlation between the presumed decision criterion and the evaluation criterion ranged from a high of .155070 (the I960 1962 trend period under time flexible decisions. Table X) to a low of -.112201 (the 1956 1962.trend period under time flexible decisions. Table X). This is presumed to be a clear refutation of the asserted validity of income reporting under the second belief in income vhxch was isolated in Chapter I.

PAGE 125

. ' CHAPTER VI Â• ' CONCnJSIOKS To a great extent, the value of this study will stand or fall on the basis of the accuracy of the observation that the two most popular paths to "belief" in income determination for annxial report purposes are the belief that there is an existential status associated vdth notions of income, and secondly, that currently reported income figures constitute a factor of obvious relevance to prospective investors in their deliberations over investment alternatives on an interfirm comparison basis. These two great beliefs in income were isolated in Chapter I as the two claims that are most frequently encountered in the literature. â€¢e have interpreted the function of a critique as that of examining the factual and logical bases for such claims. Income determination is the central occupation of substantially an entire American profession. The existence of that profession is, to some extent, inexorably associated with the belief its clients hold in its services. If one believes that income detennination is essential in financial affairs, he will naturally support, in financial terms as well as dji rhetoric, the profession which has devoted itself to the task of "realistically" stating enterprise income. In essence, we are dealing with the susceptibility of an asserted existential status to verification "in principle" on the one hand, and the susceptibility of a defined income to measurement on the other. 118

PAGE 126

119 In the conclusion phase of this study we shall treat each of the two "paths to faith in income" separately. True Income and ExistentialStatuses The conclusions drawn from the analysis of the so-called "true," "real," or "fundamental" income are rather definite and quite precise. We have confidence in these conclusions because the proponents of the assertion have been precise in their claims. This belief in what may be termed something "ultimate" in income stems from a polar separation of the two materials of inqviiry. Appeal to concepts as an exclusive mode of analysis on the one hand, and an exclusive reliance on percepts or outcomes on the other has been identified by Deinzer (althoxigh in a somewhat different setting) as this polarity of methodo1 logical commitment. On the one hand there is appeal to warrant focused exclusively on the so-called conceptualized essences of income (economic income definitions, particularly those of Hicks), and on the other, an exclusive appeal to asserted utility functions of currently reported amounts labeled "operating income" or "net income." It appears that the tvro paths to warrant (as assertions but not demonstrations) in income "theory" (and perhaps in all of accounting theory) take on an exclusive dependence on concepts of income or asserted practical utility functions of income vihatever process directs the computations. Historically, most of the attention in the literature of accounting has been directed toward establishing whether and to what extent operations conform to a priori notions of propriety. Littleton, for example, would presumably discard any alternative ^Harvey T. Deinzer, pp. 7-3.

PAGE 127

120 proposal v;hich would not "fit in" \d.th other already agreed upon 2 "principles" of accounting. This selectivity function relies exclusively on internal consistency among the agreed upon operations vdiich are a priori presumed to have warrant. Professor Graham, as another example, appealed to concepts of economic income as a guide for the accountant and as a referential base which vcis said to point 3 to the shortcomings of the accountant's operations. Here again, that which is accepted in practice depends upon the extent to which the suggested practice conforms to or is implicit in the conceptual reference base. It is suggested that operations confoim to the conceptualized notions of proper action in blind concordance and id-thout any methodology by which one might establish tte warrant of the concept in the first place. Generally, even in day-to-day affairs, one confirms ideas or "I think" elements by reference to data indications. Scientific inquirers presume that data indications determine the qualitative content of the "I think" element, the idea, or, in this case, the assertion. Certainly concepts always contain directives. The process of inquiry, however, is always tentative in that before the concept is held to be warrantedly assertable, data indications poinl^ ing to the accuracy of the concept must be forthcoming. General agreement on the validity of the concept obtained through simple observation without formulated if-then propositions constitutes a hazardous reliance. Until the time of Columbus it was a generally agreed upon observation that the world was flat. Until the time of the Copemican ^A. C. Littleton, Structure of Accounting Theory , pp. 11-14. Willard J. Graham, "Some Observations on the Nature of Income, Generally Accepted Accounting Principles, ani Financial Reporting," p. 654.

PAGE 128

121 Revolution it was generally agreed that the sun rotated about the earth. This yielded the conclusion that the earth was the center of the \ini verse. Conceptual structures devoid of if-then data indications are clearly unreliable in knowledge acquisition processes. Accountants are thought to be involved in practical affairs when it cones to closing the accounts and preparing an income statement, l^any economists exhibit a sense of "pity" toward accountants who have the practical problems of determining an income amount. Economists are generally (and superficially) thought to be involved with purely theoretical matters while accountants are thought to be exclusively practical. This is an utterly meaningless distinction particularly when the accountant takes it upon himself to put into practical operation elements of the conceptual framework of the economist as in the assertion vmder review. "Knowns" are disclosed through the interaction of the conceptual and perceptual elements of inqxiiry. If the^ accoxmtant finds it impossible to put into practice the concepts of the economist, then there is one of two things wrong. Either the concepts are invalid or incapable of verification "in principle" in the first place, or there is something wrong with the manner in which the accountant goes about applying them. If the economist is thought to engage in knowledge acquisition processes, then that knowledge which he disseminates must have a data base, intelligible on an extrapersonal basis in inquiry, consistent vdth logical rigor, and devoid of an exclusive reliance on intuitive notions of existential statuses derived from sin^jle observation. Accountants have appropriated income as a conceptual tool of the economist and have interpreted it as an end. ^It is flatly impossible to state a monetary expression of income

PAGE 129

122 as an indication of an existential status (the objective reality which exists, fundamental income, true income, income vÂ±iich is neither understated or overstated) when the concept which is said to be directing that computation is based on the subjective notions of individuals reflecting their personal opinion of their well-offness. If the socalled economic concepts of income are to direct the operations of the accountant (presuming one can picture an accoimtant stating in a monetary aggregate the combined "agreeable sensations" of Fisher as well as the difference between the Hicksian statements of well-offness at two points in time), then accomtants cannot appeal to statements attributing to income a characteristic of "iiltimateness" intelligible on an extrapersonal plane. Statements indicating that this or that particular procedure will cause the income amount to be more "realistic" are also ill-advised, for such claims implicitly presume the existence of this ideal against which an income amo\mt might be compared for a factual determination of its realism or nearness. When this ideal is thought to be economic income, the claijn deteriorates further because economic income cannot be stated in monetary terms on an extrapersonal basis. The claimant is left with nothing more than an assertion that one particular alternative more closely aligns the reported income amount to an utterly subjective income definition. In short, there is no data basis for an assertion that a particular income figure is or is not an approximation to some ideal. The extent to which an income amount does approximate this ideal cannot be known because the ideal itself cannot be stated in a monetary amount which is significant to anyone other than the income measurer himself. The efforts toward refining the operations of the accountant to make them

PAGE 130

more closely consistent with economic concepts of income are futile because no such economic concept even claims extrapersonal significance. Market determined statements of value (marginal utility or marginal productivity) represent expectations. To the extent that market mechanisms are relied upon to determine the current valuations for the determination of well-offness, the statement of monetary income is no closer to a representation of an existential status than before. It matters little how many there are who are doing the guessing, ' It should be noted that we are not critical of the economist's claim as to the behavior tendencies of economic actors, ;rnether these claims are reliable or not is of no concern here. The issue is the advisability of the accountant's appropriation of such notions of income and the claim that they represent a referential guide for coinputing "real" income (not in the sense of data adjusted for price-level change or a stock of goods and services) or claims pertaining to the existence of ",,,the objective realities which exist whether the reckoning is 4 made or not, " Therefore, with respect to the first of the two beliefs in income, we conclude that definitions of income which are generally referred to as economic income definitions are not useful referential bases tov/ard which the operations of the accomtant should be made to gravitate. We conclude that the authors of those economic income concepts attempted to formulate a narrative description of the tendencies of human economic behavior and never did they suppose that income v/as capable of repre senWilliam A, Paton and A. C. Littleton, An Introduction to Corporate Accounting Standards, p. 84. '

PAGE 131

L24 tation in a monetary amount giving effect to an existential status of extrapersonal significance. Vfe conclude that, upon examination, there is no basis v^hatever for the belief of many that income simply is and that it falls to the accountant the task of determining that income as closely as possible. Past Income Trends sind Investor Decisions ;> , , The conclusions drawn from the analysis of the statistical study of the second half of this work are less precise than those drawn from the first. It is asserted that income figures constitute a relevant segment (some say indispensable) of prospective shareholder decisions in that past income trends yield indications of the future income amounts to be realized. We have interpreted the implications of assertions such as this as the ability of a prospective investor to make the relatively better investment decisions on the basis of the relatively better extrapolated future income amounts qualified only by investment cost. Hence, the predicted price-earnings ratio was interpreted as the decision base. Cash return, interpreted as cash dividends plus or mijius the net change in investment value, was interpreted as the evaluation criterion. Rank correlation analyses were undertaken to determine the extent to ^^felch the relatively better stocks in factual cash return performance vrould be correlated with the relatively better stocks in predicted price-earnings ratios. Rank correlation analysis was chosen as the statistical device because it appeared to be more in line with the way in ^diich stocks are thought to be chosen under the assertion. In addition, any external variable which would tend to influence the data would thereby be required to disturb the rankings and not merely the values in the analysis.

PAGE 132

125 In general, it was found that the prediction of future earnings from the established trends of the past represented a wholly unreliable methodology. In the all stocks analysis it was found that the standard deviations associated with the mean of actual earnings per share as a percent of projected earnings per share ranged from more than twice the mean to more than nine times the mean (Table V Chapter IV). Under these conditions, the indicated mean appears to be little better than meaningless. The rank correlations between projected earnings per share and actual earnings per share ranged from approximately ,56 to ,33 with population interval estimates at a 95 percent confidence level of from .66 to ol9. VJe might conclude on the basis of these data that the relatively better stocks in earnings performance in the future may be vaguely determined from the extrapolation of past earnings trends on an interfirm comparison basis. A3J. related rank correlation coefficients were positive. But in view of the extremely large standard deviations obtained, one must conclude that the selection of stocks on the basis of predicted future earnings performance from past trends represents a very hazardous undertaking. It is apparent that the prediction of future earnings amounts from past trends is not as accurate as it is generally presumed to be As to whether the factually better cash return stocks will be chosen if the basis of choice is the relatively better stocks in predicted price-earnings ratios, we can be somewhat more definite. The rank correlations in this case ranged from approximately .39 to -.10 with poptilation interval estimates at a 95 percent confidence level of from .49 to -.24. The measiure of relationship in this case is extremely poor. , In some cases, particularly in the 1956 1962 trend period.

PAGE 133

126 the indicated rank correlation coefficients vrere negative vrfiich indicates that one vroxild be, in fact, better off vAthout the past earnings record of the investment alternatives, Vfe presume that the generally lower rank correlation coefficients in the cash return analysis are due, in part, to the added effect on cash return of market behavior. This tentative conclusion is also consistent vrith a hmorous observation which John Maynard Keynes made of the securities markets. The actual, private object of the most skilled investment to-day is "to beat the gun", as the Americans so well express it, to outwit the crowd, and to pass the bad, or depreciating, halfcrown to the other fellow. This battle of vdts to anticipate the basis of conventional valuation a few months hence, rather than the prospective yield of an investment over a long term of years, does not even require gulls amongst the public to feed the mavrs of the professional; it can be played by the professionals amongst themselves. Nor is it necessary that anyone should keep his simple faith in the conventional basis of valuation having any genuine long-term validity. For it is, so to speak, a game of Snap, of Old Maid, of I-iusical Chairs a pastime in which he is victor who says Snap neither too soon nor too late, who passes the Old Maid to his neighbor before the game is over, vAao secures for himself a chair when the music stops. Thesd games can be played with zest and enjoyment, though all the players know that it is the Old Maid which is circulating, or that when the music stops some of the players will find them:selves unseated. Or, to change the metaphor slightly, professional investment may be likened to those newspaper competitions in vAiich the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those which he thinks likeliest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one's judgement, are really the prettiest, nor even those vriiich average opinion genuinely thinks the prettiest. We ha-ve reached the third degree where we devote our intilligences to anticipating vdiat average opinion expects the average opinion to be. And there are some, I believe, v^o practice the forth, fifth and higher degrees.^ 5john I4aynard Keynes, The General Theory of Employment Interest and Money (Macmillan and Company, 1936), pp. I55-I56.

PAGE 134

127 The data indications also point to the lack of significantly greater validity of the assertion in intraindustry ana]o^ses. The comparison between average ranlc correlation coefficients of industry grouped stocks and a rank correlation coefficient of all such stocks taken as a whole failed to demonstrate a material difference. In suirjnary, then, there is a total lack of support for the first belief in income and a substantial lack of basis for the second. To assert that past income trends constitute an obviously relevant element in investor decisions requires quite a substantial lea^ of faith which data indications fail to support. Perhaps Hicks was right after all. Perhaps income is nothing more than a rough and crude measure used by the businessman to "... steer himself through the bewildering changes of situation which 6 confront him." Perhaps also, income calculations have nothing more^ to their credit than their "...obedience to the laws of arithmetic." It does appear to be a reflection of a grasp for certainty in an utterly uncertain situation, particularly in view of the effort to synthesize in one meaningful figure all the complex transactions of large modem enterprises. We share the faith of Hicks and the outlook of Solomons. Income determination in annual reports appears to have more support in tradition than in warrant. It is more than just possible that "...so far as the history of accounting is concerned, the next twenty-five years may subsequently be seen to have been the twilight of income 8 measurement." Â°John R. Hicks, p. 171. 7lbid., p. 178 ^David Solomons, "Economic and Accounting Concepts of Income," The Accounting Review, July, 1961, p. 383.

PAGE 135

APPENDIX I

PAGE 136

APPENDIX I STOCK IDENTIFICATION STOCK CORPORATE NAIffi NUI^IBER as of Jirne , 19^6 1 Allied Stores Corporation 2 Federated Department Stores, Incorporated 3 Gimbel Brothers, Incorporated 4 W. T, Grant Company 5 Montgomery Ward and Company Incorporated 6 Sears, Roebuck and Company 7 Great Western Sugar Company 8 Beaxinit Corporation 9 First National Stores, Incorporated 10 McKesson and Robbins, Incorporated 11 Todd Shipyards, Incorporated 12 H. J. Heinz Company 13 Kimberly-Clark Corporation 14 United States Plywood Corporation 15 General Mills, Incorporated 16 The Pillsbury Company 17 Archer-Daniels-Midland IS International Minerals and Chemical Corporation 19 Procter and Gamble Company 20 Quaker Oats Company , 21 The Torrington Company 22 United Merchants ard Manufacturers, Incorporated 23 lfi.nn-Dixie Stores, Incorporated 24 Addressograph 25 Anderson, Clayton and Company 26 Campbell Soup Company 27 R. H. Macy and Company, Incorporated 28 Northrop Corporation 29 Kelsey-Kayes Company 30 Metro-Goldyn-]yiayer, Incorporated 31 Ansted Industries, Incorporated 32 Bendix Corporation 33 The Black and Decker Manufacturing Company 34 Burlington Industries 35 North American Aviation, Incorporated 36 Outboard Marine 37 Ralston Purina Company 3B Walgreen Company 39 Vfin. Wrigley Jr., Incorporated 40 Continental Motors 129

PAGE 137

130 STOCK CORPORATE NAI^IE NUMBSR as of June , 19^6 41 Dresser Industries, Incorporated 42 International Harvester Company 43 Swift and Company 44 West Virginia Pulp and Paper Company 45 Wilson and Company, Incorporated 46 Avco Corporation 47 Eagle-Pitcher Industries, Incorporated 48 Youngstown Sheet and Tube Company 49 Ex-Cello Corporation 50 General Tire and Rubber Company 51 Interco, Incorporated 52 Admiral Corporation 53 Air Reduction Company, Incorporated 54 Allegheny Ludlum Steel Corporation 55 Allis-Chaljners Manufacturing Company 56 Alpha Portland Cement Company 57 Aluminum Company of America 58 American Cyanamid Company 59 American Can Company 60 Anchor Hocking Glass Corporation 61 American Home Products Corporation 62 American Machine and Foundry Company 63 American Radiator and Standard Sanitary Corp. 64 American Smelting and Refining 65 Armstrong Cork Company 66 Atlas Chemical Industries, Incorpora.tion 67 Bethlehem Steel Corporation 68 Blaw-Knox Company 69 Boeing Company 70 Borg-Warner Corporation 71 Bristol-Meyers Company 72 Bucyrus-Erie Company 73 . The Budd Company 74 Burroughs Corporation 75 The Carborundum Company 76 Caterpillar Tractor Company 77 Chicago Pneumatic Tool Company 78 Chrysler Corporation 79 Cincinnati Milling Machine Company 80 Cities Service Company 81 Clark Equipment Company 82 The Cleveland-Cliffs Iron Company 83 Cluett, Peabody and Company, Incorporated 84 The Coca-Cola Company 85 Container Corporation of America 86 Continental Baking Company S7 Continental Can Company, Incorporated 88 Continental Oil Company 89 Corn Products Company 90 Coming Glass Works 91 Crane Company

PAGE 138

131 STOCK COKPOHATE NAME MJl^IBER as of June, 1966 92 Crucible Steel Company of America 93 Cutler-Hammer, Incorporated 94 Diamond International Corporation 95 The Electric Storage Battery Company 96 Flintkote Company 97 WZ Corporation 9a Ford Motor Company 99 Freeport Sulpher Company 100 Fruehauf Corporation 101 General American Transportation Corporation 102 Zenith Radio Corporation 103 General Electric 104 General Motors 105 Getty Oil Company 106 Gilette Company 107 B. F. Goodrich Company 108 Â• Halliburton Company 109 Hercules, Incorporated 110 Hershey Chocolate Corporation 111 Homestake I-Iining Conpany 112 Honeywell, Incorporated 113 Ingersoll-Rand Company 1:14 Inland Steel 115 Johns-ManviUe Corporation 116 Johnson and Johnson 117 Jones and Laughlin Steel Corporation 118 Keebler Com.pany 119 Kennecott Copper Corporation 120 Koppers Company, Incorporated 121 S. S. Kresge Company 122 The Kroger Company 123 Leihigh Portland Cement Company 124 Libbey-Oifens-Ford Glass Company 125 Liggett and Myers Tobacco Company 126 Lone Star Cement Corporation 127 P. LoriUard Company 128 M. Lowenstein and Sons, Incorporated 129 Marathon Oil Company 130 I'larshall Field and Company 131 The Maytag Company 132 The Mead Corporation 133 Mead Johnson and Company 134 Melville Shoe Corporation 135 Merck and Company, Incorporated 136 Mnnesota Mining and l-Ianuf acturing Con^jany 137 Mobil Oil Corporation 138 Motorola, Incorporated 139 G. C. Murphy Company 140 National Biscuit Company 141 National Distillers and Chemical Corporation 142 National Lead Company

PAGE 139

132 STOCK CORPORATS NAME NmSER as _of June , 196^ 143 National Steel Corporation 144 National Tea Company 145 Newport News Shipbuilding and Drydock Company 146 Olin Mathieson Chemical Corporation 147 Otis Elevator Company 14s Oi-i'ens-Illinois, Incorporated 149 Parke, Davis and Company 150 Penn-Dixie Cement Corporation 151 Pennscilt Chemicals Corporation 152 Pepsico, Incorporated 153 Chas. Pfizer and Company, Incorporated 154 Phillips Petroleum Company' 155 Pullman, Incorporated 156 Rayonier, Incorporated 157 Republic Steel Corporation 15s R. J. Reynolds Tobacco Company 159 The Ruberoid Company 160 Safeway Stores, Incorporated 161 Schering Corporation 162 Scott Paper Company 163 Scovill Manufacturing Company 164 Sharon Steel Corporation 165 Simmons Company 166 Sinclair Oil Corporation 167 The Singer Company I6S Skelly Oil Company 169 Smith, Kline and French Laboratories 170 Square D Company 171 Standard Brands, Incorporated 172 The Standard Oil Company (N.J.) 173 Sterling Drug, Incorporated 174 Sunray DX Oil Company 175 Sunshine Biscuits, Incorporated 176 Texas Gulf Sulphur Company 177 Tidewater Oil Company 178 Time, Incorporated 179 Timkin Roller Bearing Company ISO Union Camp Corporation 181 Union Carbide Corporation 182 Uniroyal, Incorporated 183 United Engineering and Foundry Company . 184 United Fruit Company 185 United States Gypsum Company 186 United States Pipe and Foundry Company 187 United States Steel 188 United States Tobacco Company 189 Vanadium Corporation of America 190 Walworth Company 191 Warner-Lambert Pharmaceutical Company 192 Westinghouse Air Brake Company

PAGE 140

133 STOCK CORPORATE NAME NIE-IBSR as of June, 1966 193 Westinghouse Electric Corporation 194 Weyerhaeuser Company 195 Wheeling Steel Corporation 196 Vlhirlpool Corporation 197 F. W. Woolworth and Company 198 Worthington Corporation

PAGE 141

APPENDIX H ADJUSTED EARNINGS PER SHARE INPUT DATA

PAGE 142

135 UN CO O H to vO O C^^vC O H rn H -:} H ^'NsO LAOCOrH-^C^OrH'JA C^vO C tO -ir\C^C^NO C^nOnO W^"^-i'Â•f-: O O --L^^0NLf^Ou^-dÂ•Or^ o^md u-\ on On os -;t OH-<^-2Â•"^'-'^ONo:^lAONc^^^r^ON^--i-c^c^Nocvto<^o<^o^c^HOC^"^ou^NO c^^c^Jc^iHHHc^Jc^J-^c^^HC^Jc^oJHHc^iHc^ic^"^HHc^20>^^^c^^c^i-^NOC^i-4^^r^ I oc^H<^c^NO^--ONOu^c^u^oNou^<}Of^ONOONM^^oOr^-^NONONOHHNOtx)<:^^ Hr-!c^c^H-^tX)-^ONc^iNONOcvONc^i^^c^'A^.oc^i{Â»tX)cnOONc<^c\>Hu^O'^Hc^ C\2CNJC\iHrSHCN2CNi-^C\JOCVcnHHHCN!H c^NocouNc^irNcvcVNOCOoN^i-uAtX) c<^no r^tSONOc^CNJHNOOpH-J-CviON CsJHC\iHHHcÂ»\H.^C^i^C^-^NO-c^Â•-;^NONOlAtx)-^tooNONtÂ»HOc^c^NOH^oH-dÂ•Hc^ tOtÂ»ONtOOI-|HC^CM^O^^^^tX)r^CVtOC'^CVO^C^i(r^OHU^c^OC\iOtX)ON<^HHC^ to UN cA O r^UN-^-l-ONr~-U\H:>-^HONO O^tO c^^AI-AmD CxJnOO !>-C^On-C0nO mOUNUNtXiONH NONOtOtcHOONsO CÂ«^nO COC^C^CVONCV-:^OJ>-CVONCVtXl-^-d-ONC-H-Ct-HC^ON C^nO HHHr-|C\ii-i0^rHi/\C\ic^HC\2iHHHC\'HH0^C\JH H C\! CV cA C\J(r^. OONUNOOOiHCVUN-^UNCVC^CN2 CÂ—uMKinOnO i-i-^O ONtX) r4 O C^nQnO O C^^NOtXl C\J C-OnC^CO CM O H iH C^tX) HHHHC\iH-d-c^-;i-c\iNOC\2C\iHHHcf\HHcr\c^H H-;}H H nO O VTN H NO c^c^tONOH!>-ocv o^ta Noc^NoooNOo^-c^o^NO-^c^ON02H^--:^toto^-oo^f*^ ir\ UM^OtOvO OCN-^0NC^cr^ONOCÂ«^C0HN0O>^^^O-dC^r^t0ONC^v0C^2t0C^JtXl(^it0 cv O NO H NO cv Â• Â• Â• r-i H cA I c/i o o NO O O -I. H to o o Â• CM -4m H H o to ON Â• Â• H Â• O cv NO a Â• Â• r-i to to ON cv CTn to -4to to CNi o On ci CM ON NO UN Â• Â• Â• i-l r-i UN o -4C3N cv CM r-i UN Â• Â• Â• Â• H to H CO HCMO^-;?-UNsOC^tOONOr-jCMCÂ«N-J-UNvOC--tOC;NOiHCM cn UNvO f-tOONOHCVcA-ct rHiHrSHHHHHHi-|C\2CMCMC\lCVCNiOJCMCMCMO^C*NC*N<Â»NC*N

PAGE 143

136 UN c^u^toc^<)M^c^>o^oc^^^cÂ»^cHH-d-c^rH(rNH voc^2HHH^o^^cr^^/^c\ir^c^ou^u^so^X)C\^c><^cv^\^oc^^-oc^ONHc^f^^ tX)OvOcnONvOC^C^OOiHOHOÂ«)Hc^vOvOOo^C\JC>t>Â£>CNJvCOOC^OC\2HCO v^HHc^iu^c^JHcf^c^^cÂ»^w^c^ic^JVAcÂ«^<^^c^JHc^-:^HH<^2^^c^2-^cvHc^2c^ c^^tx)"^C7^-4OHc^JC^JC^coc^^-c^^oOvoHo^^-^--<^icvolAÂ•' '~ ' ' " HHi^CXJHOJiHHCNirHHC^OJCMHHc^CN vO H O o cA k CO cA C^vo C---0 OO H C^vO c^tX) H O O O >^CV LP>Oc^tO tX> C^C^H ir\0 OcTnOJ Cn-J-O C~Hto^oo^c^^<^oo^^/^a)0^toc^M^c^w^Hc^^c^ic^i^oc^iCNJOJCN2HHHHHC\iC\J H02C\iHCNJHc^C\iHCNJHOJC\2HC\JHHHH CV CNi tX)^or-tx)c^tx)tx)l^c^HtxÂ»'^o^r^-dHHHHoJ^-1Â»(^J^oc\^HcvM^oc^w^c^ c^^-^^(>Hc^o^u^<^c^^coc^.c^^c^.p^so^-^-tooN-rHCVOOtOvÂ£)CVlX)-d-^ O C^sO COtOtXlOv.' C^OnOUN c^HHHwNHHHCNiHc^HHCviOi CNi H H C*N CV CNi cÂ»N C\! C\i -4H HHHH-d-c*N vO UN cA UN U-N cA C^.C^CVtÂ»^-CV^-tX)UNCy^C^voo^^^-voc^c^^-dÂ•c*Nvo C^HHHun~*c*NH CN CNi H CM -4to I -4C\i Cf\ -;f UN CM -jH Hc^HiHUNcvN H C!N CNi Â• Â• UN r-i CM CNi UN cA o to UN UN CNi UN UN UN cA r^ CM UN CNi to NO 3i CNi UN I cn H UN cA CO H -4H H CNi H O CM UN CM C
PAGE 144

137 ^oH<^HN01X)^-OHvo<^c^cv<^o^Â£)cvtx)Ho>^^o^\^voc^c^vo'^o^HojHc^ UMX)vO i^O^vO C^tX)0 C^tX) Â»J^(^i^O OvO -j--d-tO CC^^OvO C^OOtX) Oi tX) O Or^r^-dÂ•tX)u^c^0^t0Ov0C^i C^H c^c^HHHHc^HcviCNl-4-C\iCNio^C^HHcnc'^C\JCvJ-d-HHHHOiHHCVHCviH^ H ^-c^c^c^H1X)^HOc^Hu^c^^[^-^^J-J-tÂ»cv^-Hc^^o^c^JOc^^^c^c^ic^sooc^^cf^o^ -^'-^^c^il^^l^u^lÂ•'^otocÂ»^D-H^O(^^0Â«^^oc^i0^otÂ»cvc^^OcÂ»^o^c^^u^t^ vOc^tX>0^^<^OC^.C^ -t cv r-ivo^-lAtoHc^ic*^c>o^^H UN O o CO cv Â• Â• Â• Â• Â• Â• CNi UN UN H CV UN CV H CV o UN O OfN O CV ON Â• Â• Â• Â• Â• Â• Â• H -4CV H C--^-c^c^c>-^-c^cotÂ»tX)tococototÂ»cocooNc^o^c^.cJ^c^o
PAGE 145

13^ 1^ Hr^-^o^soc*^^^o^JHHw^'^^-'^H^\2Hcvtx)H>J^ou^c\^l^HH'^!>-C^!>HC^C^C^HvOCN2 lAHCH r4 H H r4 oj H CNl CÂ»^ C<^ -t ITS C*^ H H H -dH a r-\ r-i r-i Â•-{ H H (MHt>-cvooH -d-vo <^<^^<^H!^-oc\.'oc^-4oÂ«<^o^^~c^^c*^sq O vOi^toc^I>-CÂ«AO-d--d-C^O(OC\J c^-*OOvOH'^C^-MDC^-eOC<\C\JsOtO vO cVc^vOC^tOCV'^ir\C^lX)ONOHvOOOsOONOOCOHC^'^OtXlvOCVvO'^Os<>COC^ ojcA *H *c*^H^^c^iHc^J(^^c*^o^i-^c^^^c<^r^c^iH^^^JH-dHc^jr^H<^i H H o ONvo^^H'^^^^-oo>^to-^Htxlc\^o^oc^2o^tÂ»c^c^ooc>c^toMDc<^toc^icÂ»^s^ HO^OHHC^C^C^JC^iHtX>^OC~-M^u^^^HHOOtOHHC^iCVvO"^r^C^.vOCOOC^C^^ c^c^ HOcÂ»^HHC^^c\ic\^C^Jc^C^^c^-4l^c^iC^2(^JC^^l^<^C^l-4Hc^iHr^c^! CM H to tX)cvo^c^Â£^^ocv'^l^^s^H^-cntx)ON^o-dÂ•cvoc^>H^<^^^H^-o>^^oc^HOto^^-cv^-c^a^c^ooovo-dÂ•c^tÂ»c^c^o-ct^oc^tx)HOtooo'^c^c^iCVHC\J<\JCv2CM^C\2^-C^C\iHCV,CMC\JC\iHHC\i C\J H U\ -ctOOvJi-l'^OOOH C^vO H Â£> -*vO 0^C\2OMDCM-d-C0v0CVv0OO0NC\JC^C^ OvO C>cA H cA C\2 CV CM CM CM 0^ UNC^O Â»rvÂ«o ^ HCMH vOvOC^ tt)-* CM CM C\2 1^ CViHO Ht-ltX) to J>CMCM CJN Os\0 OtO O CMH cm CMc<\ rH-3CM CM CM H CM H ^ CO::?^'55::-?9^0'HCMo^-4U\vOC^1Â»<>OHCMCCNOrHC\}C^-
PAGE 146

139 UN so r-t o OS to 0^ lA 0^ C\2 UN cA cA sOUNHOj CV UMAH-ctC^C^CAsOHCV C^CV (r\CV(r\-4-CV C<\ H o r-c^osONUNv\c^vo c^ixi D-ONvotÂ»vO H -d-MDsoJH-4-CVC\2C\iUNOiHC\JHC^CVHcA^O>20iNLf\OsO^-C^C~-tXlOs U\sO !>cv -4(r\ lA c\! cvi c^no c^^totX)C^voos<^oHHOu^c^i-:iÂ•!^--^-^tx>o-Â«^ cnCÂ«^C\JH02-4-C<^H-2CNJOJOc^C\iCNJHC^HCv2C^CVC^HC\JHCNiOJC\JC\lC\^ I -d-O u\OssO u-\0 C^OsOO HO [>-vOtÂ»tÂ» UMAOsOsOvO O UACNJ "AO C^cnH H C^vO cvos^-uA^-^o<^o^H^-H-d-o^oc^Â«)-^<^tÂ»^oc^u^toc^oc^c^^c^^oc^2c^^tX)<^tx) <^r-ir-{Hr-{C<\Oir-{-:t CVCNJOC^HHHCNJH (TNCviCSiHO^ O C\iC^H-4-H I i vOOvoc^.u^-;fr^-4Â•ooc^.to^-^o^oOsoAHH OS lA o o vO H C^sO o CV Â• Â• Â• m Â• Â• m H H a cv UN UN CM OS H O -4to to C*N Â• Â• * Â• Â• H H H H H O UN sO UN o vO H OS OS -4H Â• Â• Â• Â• Â• Â• Â• H H H cv UN to OS Cv2 UN vO H H SO H CÂ»N SO O Â• Â• Â• Â• Â• Â• Â« H H OssO UN vO to sO H -dO o Â• Â« Â• Â• Â• * Â• H H to cv CO ^XS22^Q.^i<^^^"i^'^r'^<>Â•PHC\2c^<^uNsOC-tOOsOHCVc^^^ C^C^C^<^4-C^-d-JÂ•-4-4Â•<^ rHHrHHHHrHrHr^r-ir-ir-iHHt-ii-irHr-ir-iHr-ir-iHHHi-iHHr-{HHt-iHr-i

PAGE 147

140 I CTj so 0^ vO u^o^^oc^Â«)Â•M^^itX)voHc*^c\'r^cv^oc^tc^'tx)c\iO^-(;^w ix>cvj>cv u^_dc^CN<^:^u^-^OOvo^>Â•l-^Â»^Â«^o^o c^^-rHcv c^^o -*OtX)C>OC^cnOsC^U\HHtX)OCVC^ONC\' u^ vAvo c\( H o o^so O md tc -dw^ p i H CnCn-CO CAC\ C^iH 'AC^-jfCV C^'^c'^'ACV i^-CO to CV CV OH CV iAJ>-vOC0 C^1C^JC*^HC^ (Vr-iC\JCViHHHrH--*r-lCVHCj>C3rHH-OHCVHH >^HOCVCNHMDC\'OtOvO(>!>C^Or--f'^a)vOvOC>C^r-CVC\.>CVC>0 ^-Olr\r-iI-ic^i^--^Â»-'^HcvtoHOÂ«^'^o^-otoocN-tiÂ•c^>^^0"^c^ CVCVc^CNic^HC^CVtX) HOC^H-i-C0O<-OHr^UN C--vO OO^r-ltOOC^CNrnvOsOvOr^Hc^ CV I-i CV CV O H i CV ^H H CV H -4-tO CN-^-UNCTNUNUNCNUNo^CV OtOO^O^C^UN-ctHO UNvO CV O -C CV OUNOOCVC;sHOC^-d-i>-r-iO>J^r~-UNO^i>-HrHONtOtX)C^C^tOiHi-l CV iH -4Cvj CV OvJHCVCVC^-4-CVC^Hc^CV CV iH UN <^txlHOOO(>i:^ocv.Â— ^cvv^ocvc^c<^ c>--c>r^-^r-icocvcv^u^oto-<'Â•cv-^ocvo-c'-^^-c^-c^too^ocVlHHC;s iHH-JCVC*N CV CV -j-C^CV CV H C^i^C^vO H -4-H CV CV tO H r-i CV UN UN vO CV UN NO O to o Â• Â• Â• Â• Â• Â• H -4UN to O UN to O UN o o O UN Â• Â• Â• Â• 9 cA H C\i UN to UN !>. O UN -4 O UN Â• Â• Â• Â• Â• Â• to C\2 UN O? CVmD to o H CV iH UN i-i UN C. Â£>-!>-C^;>-C^i>-ir-C^(>-tCtOtOtOtOtOtOt)0-COtO CTnOGnOnOnOOOnCn Hr-ir-iHrHHr-iHHHHrHHr-ii-irHr-ii-ii-iHr-iHHr-it-ir-irHr-{

PAGE 148

APPENDIX III MARKET VALUATIONS INPUT DATA

PAGE 149

142 STK SEQ MVl-RAV/ KV2-RAW 1 1.0 64.500 73.750 2 1.0 50.219 73.188 3 1.0 S5.S13 48.625 4 1.0 27.531 40.875 5 1.0 33.563 38.063 6 2.0 79.934 127. 188 7 1.0 34.375 40.344 8 1.0 26.594 40.688 9 1.0 61.313 46.875 10 1.0 41.813 41.063 11 1.0 23.531 26.906 12 1.0 51.844 45.750 13 1.0 65.625 54.250 14 1.0 46.031 43.625 15 1.0 25.656 59.000 16 3.0 49.625 42.313 17 1.0 35.313 38.813 18 2.0 38.375 67.531 19 1.0 65.281 64.656 20 1.0 66.625 55.625 21 2.0 51.750 36.438 22 1.0 19.344 38.313 23 1.0 25.500 32.156 24 1.0 55.031 46.563 25 1.0 38.063 27.625 26 3.0 92.8I3 35.188 27 2.0 55.250 55.313 28 1.0 23.688 25.188 29 1.0 31.500 40.094 30 1.0 33.875 37.125 31 1.0 26.906 47.031 32 1.0 50.781 59.906 33 2.0 47.813 45.594 34 2.0 21.3A4 38.500 35 1.0 63.938 55.000 36 1.0 12.906 16.000 37 1.0 31.000 40.906 38 2.0 42.250 30.875 39 1.0 23.187 66.031 40 1.0 10.000 13.469 41 2.0 20.438 26.922 42 2.0 46.656 38.938 43 1.0 34.813 49.156 44 1.0 29.125 46.281 45 1.0 36.563 47.250 46 1.0 24.906 25.625 47 1.0 20.437 30.875 48 3.0 86.281 42.469 49 1.0 40.906 58.438 [Vl-ADJ MYl-LOW MV2-HI 64.500 61.625 76.500 50.219 46.750 73.500 85.813 79.000 49.375 27.531 26.000 43.000 33.563 31.875 38.125 39.967 72.000 128.750 34.750 32.750 42.000 26.594 24.500 43.500 61.313 57.875 49.250 a. 813 40.250 42.375 23.531 21.875 27.875 51.844 40.500 48.500 65.625 50.750 56.125 46.031 39.000 45.125 25.656 21.500 61.500 16.542 41.250 46.625 35.313 33.000 42.750 19.187 34.000 71.750 65.281 56.375 66.875 66.625 62.650 60.250 25.875 50.000 38.500 19.344 17.875 29.750 25.500 22.000 33.375 55.031 49.750 46.875 38.063 36.750 28.250 30.938 87.250 35.625 27.630 50.375 56.000 23.688 22.000 27.000 31.500 29.375 42.750 33.875 29.250 40.625 26.906 24.750 51.750 50.781 48.000 65.375 23.906 42.250 48.875 10.672 20.000 41.875 63.938 60.375 59.500 12.906 10.500 18.500 31.000 28.125 41.500 21.123 39.250 31.875 23.187 21.125 75.250 10.000 9.125 14.750 10.219 17.500 31.000 23.328 43.500 41.500 34.813 31.500 51.875 29.125 26.500 48.500 36.563 33.750 50.125 24.906 31.750 27.750 20.437 19.500 37.000 28.760 81.000 45.000 40.906 36.500 62.250 DIV-63 DIV-64 Diy-65 3.0000 3.0000 3.0000 1.4500 1.4500 1.5500 1.2500 1.3625 1.6000 1.2000 1.2000 1.2000 1.0000 1,0000 1.0000 1.7500 2.0000 2.2500 1.6500 1.9500 1.9500 1.2000 1.2000 1.4000 2.5000 2,5000 2.5000 1.5000 1.5000 1.5500 1.4000 1.4000 1.4000 1.0000 1.0000 1.0000 1.8500 2,0000 2.0000 2.0000 1.7000 1,2500 1.2000 1.2000 1.3000 1.5250 1.6000 1.7750 2.0000 2.0000 2,0000 1.6000 2.0000 2,0000 1.5500 1,6750 1.8000 2.2000 2.2000 2.2000 2.5000 2,5000 3.0000 1,0000 1.1000 1.1500 0.9600 1.0800 1,2000 1.0000 1.1000 1.3000 1,5000 1.0000 1.0000 2.2000 2.3500 2.6250 2.2000 2.2000 2.7500 1.0000 1,0000 1,0000 1,7000 2.0000 2.0000 1.7500 1.5000 1.5000 1.6000 1.7000 1.9000 2,4000 2,4000 2.4000 1,5400 1,8500 2,2000 1,2000 1.4000 1.3000 2.7000 2.4000 2.0000 0.6000 0.4500 0.6500 0.8500 1.0000 1.0000 1.6000 2.0000 2.4000 1.2000 1.20-00 1.6000 0.4000 0.4000 0.4000 1.2000 1.2100 1.4000 2.4000 2.7000 2.9500 1,6000 1,7500 2,0000 1,2000 1.2250 1.3500 1.6000 1.6000 2.0000 0.8000 1.0000 1.0000 1.2000 1.2000 1.2000 5.0000 5.4000 5.4000 1.6000 1.6000 1.7000

PAGE 150

143 SIK SEQ m-RAW MV2-RAW 50 1.0 21.219 27.281 51 1.0 24.406 38.625 52 1.0 13.031 75.843 53 1.0 56.281 73.531 54 1.0 34.375 51.875 55 1.0 15.375 33.625 56 1.0 17.750 12.219 57 1.0 55.469 75.906 58 1.0 49.625 87.875 59 1.0 45.500 57.000 60 1.0 29.656 42.344 61 1.0 54.094 87.219 62 1.0 21.531 20.344 63 1.0 13.219 20.875 64 1.0 57.500 73.469 65 2.0 67.688 58.875 66 1.0 15.000 22.562 67 1.0 30.281 39.687 68 1.0 27.562 31.375 69 1.0 38.312 142.062 70 1.0 Z^.OOO 48.500 71 2.0 86.531 99.781 72 1.0 15.593 63.031 73 1.0 12.531 20.718 74 1.0 29.594 49.812 75 1.0 46.438 91.313 76 2.0 37.312 50.406 77 1.0 27.031 43.781 73 2.0 75.406 54.84A 79 1.0 39.687 74.875 80 2.0 56.093 42.438 81 1.0 31.218 64.375 82 1.0 38.906 86.375 83 1.0 45.906 79.843 84 2.0 86.781 85.961 85 1.0 24.250 35.188 86 1.0 44.562 47.500 87 1.0 45.062 63.437 88 1,0 54.250 68.875 89 1.0 50.906 52.156 90 1.0 162.750 236.250 91 2.0 42.812 40.000 92 1.0 16.781 28.906 93 2.0 58.781 59.281 94 2.0 48.281 40.000 95 1.5 51.062 55.500 96 1.0 18.813 20.718 97 1.0 37.188 80.750 98 1.0 45.312 55.156 99 1.0 23.187 66.031 m-ADJ KVl-LOW MV2-KI 21.219 18. 250 30.000 24.406 23.000 40.250 13.031 11.500 107.875 56.281 54.125 77.750 34.375 31.000 55.250 15.375 14.000 36.250 17.750 16.125 13.000 55.469 51.250 84.875 49.625 46.125 96.000 45.500 44.250 59.000 29.656 28.000 44.875 54.094 51.000 90.975 21.531 19.250 21.875 13.219 12.500 23.500 57.500 53.625 82.375 33.844 63.000 62.000 15.000 13.875 25.375 30.281 28.625 /^2.375 27.562 25.000 33.875 38.312 36.500 172.000 42.000 40.250 50.000 43.265 79.250 109.500 15.593 14.875 70.000 12.531 n.500 23.250 29.594 27.375 55.250 46.438 44.125 99.500 18.656 35.375 53.000 27.031 25.625 46.000 37.703 68.875 60.375 39.687 82.500 82.500 28.046 51.750 46.875 31.218 29.000 66.250 38.906 36.750 89.500 45.906 43.750 85.000 43.390 83.125 90.750 24.250 23.250 37.875 44.562 42.750 49.500 45.062 43.750 67.500 54.250 53.250 72.875 50.906 49.375 55.500 162.750 155.500 251.000 21.406 41.250 45.000 16.781 15.125 37.500 29.390 56.000 62.250 24.140 45.750 42.125 51.062 48.750 57.500 18.813 17.250 22.625 37.188 35.000 84.OOO 45.312 42.750 56.875 23.187 21.125 75.250 DIV-63 DIV-64 DIV-65 0.4250 0.5000 0.6000 1.2000 1.2000 1.2000 0.0 0.0 0.0 2.5000 2.5000 2.5000 2.0000 2.0000 2.0000 0.5000 0.5000 0.5025 0.8750 0.5000 0.5000 1.2000 1.2000 1.4000 1.8000 2.0000 2.1500 2,0000 2.0000 2.0500 1.4000 1.4000 1.4000 1.6000 1.7000 1.9000 0.9000 0.9000 0.9000 0.8000 1.0000 1.0000 2.7000 2.9000 2.8000 2.1000 2.5000 2.7500 0.6000 0.6000 0.7000 1.5000 1.5000 1.5000 1.4000 1.4000 1.4000 2.0000 2.0000 2.5000 2.0000 2.0500 2.2000 1.8000 2.1000 2.6000 0.6000 1.7000 3.0000 0.5250 0.6000 0.6500 1.0000 1.0000 1.0000 1.8000 1.9000 2.2000 1.1500 1.5000 2.0000 1.4000 1.5500 1.6500 1.7500 2.0000 2.5000 1.6000 1.6000 2.0000 2.8000 2.6500 2.8500 1.3500 1.5500 I.75OQ 2.0000 2.5000 3.0000 1.5500 1.7000 2.1000 2.7000 3.0000 3.4000 0.9250 1.0000 1.1500 2.2000 2.2000 2.2000 1.9500 2.0000 2.2000 1.9000 2.1000 2.4000 1.4000 1.5000 1.5000 2.5000 2.5000 2.5000 2.0000 2.0000 2.8000 0.8000 0.8000 1.2000 2.0000 2.0000 2.2000 1.9500 2.2500 2.7000 1.9300 2.4700 2.7700 0.8000 0.8500 1.0000 0.8000 1.0000 1.2000 1.8000 2,0000 2.1000 1.2000 1.2000 1.6000

PAGE 151

144 STK SSQ MVl-RAW IN2-?A\'J MVl-ADJ MVl-LOW MV2-HI DIV-63 DIV-64 DIV-65 100 1.0 25.562 34.125 25.562 24.000 36.750 1.5000 1.7250 1.5000 101 2.0 69.312 44.531 34.656 66.000 48.250 2.425O 2.5500 2.7500 102 1.0 55.750 119.438 55.750 52.500 138.000 1.3500 1.5500 2.0000 103 1.0 76.688 116.468 76.688 74.125 120.000 2.0000 2.2000 2.3OOO 104 1.0 58.938 103.906 58.938 55.000 IO8.25O 4.OOOO 4.4500 5.2500 105 1.0 17.875 36.938 17.875 16.375 42.125 0.1000 0.1000 0.1000 106 1.0 30.938 39.156 30.938 28.000 41.500 1.2000 1.2000 1.2000 107 1.0 44.719 56.938 44.719 41.375 6O.625 2.2000 2.2000 2.2000 108 2.0 52.469 49.281 26.234 50.000 56.500 2.4000 2.6500 3.0000 109 1.0 42.125 43.781 42.125 39.000 50.000 0.7500 1.0000 1.0000 110 1.0 33.500 33.690 33.500 31.750 36,125 0.9000 0.9250 1.0000 111 1.0 45.250 46.125 45.250 42.000 49.375 1.6000 1.6000 1.6000 112 2.0 86.375 72.625 43.185 81.000 76.500 2.0000 2.0500 2.2000 113 2.0 67.938 54.188 33.969 62.5OO 57.000 4.OOOO 4.OOOO 4.OOOO 114 1.0 38.125 43.625 38.125 35.000 43.375 I.65OO I.85OO 2.0000 115 1.0 44.122 54.781 44.122 4I.625 59.250 2.0000 2.0000 2.0500 116 1.0 76.000 175.313 76.000 73.250 188.000 1.1000 1.2000 1.4500 117 1.0 49.062 69.249 49.062 45.625 74.250 2.5000 2.5000 2.5500 118 1.0 38.875 30.719 38.875 36.500 34.250 1.5000 I.5OOO 1.2000 119 1.0 68.188 126.094 68.188 65.OOO 135.250 4.0000 4.OOOO 4.7500 120 1.0 40.469 63.875 40.469 37.750 68.375 2.0000 2.4000 2.4000 121 1.0 22.813 80.688 22.813 21.250 85.500 1.2000 1.2000 I.4OOO 122 1.0 24.688 35.750 24.698 23.759 24.688 1.1000 1.1250 1.2250 123 1.0 17.688 16.375 17.688 I6.25O 17.375 1.0000 1.0000 1.0000 124 1.0 52.125 55.750 52.125 49.250 58.250 2.6000 2.9500 2.9500 125 1.0 68.875 73.250 68.875 66.000 76.875 5.0000 5.0000 5.0000 126 1.0 18.938 18.094 18.938 17.875 19.375 1.0000 1.0000 1.0000 127 1.0 43.656 44.750 43.656 41.250 47.750 2.4750 2.5000 2.5000 128 1.0 9.781 25.219 9.781 9.000 27.000 0.0 0.0 0.4500 129 1.0 45.156 55.438 45.156 42.875 58.250 1.7000 2.0000 2.1500 130 1.0 35.438 58.344 35.438 33.250 6I.OOO 1.4500 1.6000 1.8000 131 1.0 30.000 36.188 30.000 28.000 39.375 I.65OO 1.9000 1.9000 132 1.0 39.906 47.406 39.906 38.125 53.000 1.7000 1.7000 1.7000 133 1.0 21.250 29.313 21.250 19.000 32.125 0.6800 0.6300 0.4800 134 1.0 26.750 28.719 26.750 24.125 31.750 1.3000 0.8000 1.0000 135 3.0 81.469 71.563 27.156 77.500 75.750 2.0000 2.S3OO 3.6OOO 136 1.0 56.375 68.500 56.375 52.125 71.625 0.9000 1.0000 1.1000 137 1.0 58.781 94.031 58.781 55.625 97.500 2.6000 2.8000 3.O5OO 138 1.5 64.625 166.531 43.083 59.750 191.625 1.0000 1.0000 0.7500 139 1.0 25.031 22.844 25.031 24.000 24.500 1.2000 1.2000 1.2000 140 1.0 43.438 53.063 43.438 40.25O 54.875 1.5750 1.6750 1.7750 141 1.0 24.656 34.000 24.656 23.125 36.875 1.2000 1.2000 I.4OOO 142 1.0 70.250 70.781 70.250 67.125 74.500 3.2500 3.2500 3.2500 143 1.0 37.063 60.688 37.063 33.750 65.750 I.65OO 1.9500 2.1250 144 1.0 15.563 15.688 15.563 14.500 16.125 0.8000 0.8000 0.8000 145 1.0 48.000 37.844 48.000 43.500 53.125 2.4000 2.5000 2,5000 146 1.0 33.125 59.406 33.125 30.500 67.875 1.0000 1.3000 I.4OOO 147 1.0 56.625 58.063 56.625 53.500 61.875 1.8000 I.825O 2.0000 14s 2.0 74.656 62.094 37.328 72.000 64.500 2.5000 2.5000 2.9000 149 1.0 25.844 33.219 25.844 23.750 37.250 1.0000 1.0000 I.I5OO

PAGE 152

U5 STK SEQ MVl-RAW MV2-RAW 150 1.0 17.969 12.438 151 1.0 36.906 49.813 152 1.0 46.531 78.438 153 1.0 47.406 72.375 154 1.0 49.781 56.750 155 1.0 25.313 59.688 156 1.0 22.219 37.750 157 1.0 36.188 43.156 158 1.0 41.938 44.281 159 1.0 29.125 35.344 160 2.0 45.469 30.688 161 1.0 41.219 88.375 162 1.0 32.281 37.781 163 1.0 21.969 64.625 164 1.0 20.125 30.531 165 2.0 34.219 35.469 166 1.0 36.344 62.406 167 2.0 127.438 60.594 168 1.0 56.813 108.000 169 1.0 63.375 78.688 .170 1.0 37.375 80.281 171 1.0 66.469 73.094 172 1.0 58.531 81.063 173 3.0 72.969 38.688 174 1.0 25.469 32.125 175 2.0 102.125 45.563 176 1.0 14.406 88. 3A4 177 1.0 21,281 51.188 178 3.0 67.500 98.813 179 2.0 54.218 47.688 180 1.0 35.625 48.906 181 2.0 104.313 68.844 182 1.0 42.063 75.156 183 1.0 18.563 19.781 184 1.0 23.031 30.688 185 1.0 76.563 65.219 186 1.0 15.969 21.750 187 1.0 45.813 52.313 188 1.0 27.000 32.250 189 1.0 13.781 28.281 190 1.0 6.813 9.719 191 1.0 24.344 41.594 192 1.0 26.000 41.969 193 1.0 33.281 63.125 194 1.0 26.063 40.625 195 1.0 29.656 29.594 196 2.0 36.625 38.688 197 3.0 65.375 30.344 198 2.0 28.219 41.344 [Vl-ADJ MVl-LOW M72-HI 17.969 16.625 13.375 36.906 34.000 52.750 46.531 42.000 82.125 47.406 43.750 75.250 49.781 47.125 58.125 25.313 23.375 69.875 22.219 20.000 43.000 36.188 33.500 45.250 41.938 39.625 45.875 29.125 26.875 37.250 22.735 42.125 31.750 41.219 38.250 97.000 32.281 30.000 39.875 21.969 20.500 73.875 20.125 17.500 33.500 17.109 31.875 36.750 36.344 33.375 65.875 63.719 121.000 64.750 56. 813 55.000 116.000 63.375 61.250 86.250 37.375 34.500 85.625 66.469 63.625 75.500 58.531 55.875 84.125 24.323 67.250 41.250 25.469 23.375 34.000 51.062 95.500 52.250 14.406 13.375 104.625 21.281 19.125 53.875 22.500 66.500 103.750 27.109 50.000 49.875 35.625 34.125 55.250 52.156 99.500 72.750 42.063 40.000 79.750 IS. 563 17.625 21.250 23.031 21.000 34.625 76.563 72.750 67.760 15.969 15.000 22.875 45.813 43.250 55.750 27.000 25.500 33.250 13.781 12.375 29.250 5.625 11.375 11.375 23.344 22.875 43.750 26.000 25.000 46.375 33.281 31.125 66.000 26.063 25.250 42.875 20.656 26.625 36.000 18.312 34.500 41.250 21.791 63.000 32.375 lif.109 25.625 44.000 DIV-63 DIV-64 DIV-65 1.0000 1,0000 0.9000 0.9000 1.0000 1.1000 1.4000 1.4000 1.6000 1.0500 1.1500 1.3000 1.9750 2.0000 2.0500 1.4000 1.9000 2,2000 1,0000 1.1500 1,4000 2.0000 2.0000 2.0000 1,6500 1.8000 1.8500 1.6000 1.6000 1.6000 1.6500 1.8500 2.0000 1.5000 1.5500 1.7000 0.8250 0.9000 0.9250 1.2000 1.6250 2.0000 0.4000 0.4000 0.5000 2.0000 2.5000 2.8000 2,0000 2.0000 2,0500 3.3900 4.0000 4.0000 1,8000 2.0000 2.0000 1,5000 1.5500 1.8500 0.9000 1.7500 2.2500 2.0500 2.2500 2.4000 2.7500 3.0000 3.1500 1.9200 2.0625 2.2875 1.4000 1.4000 1.4000 2.5200 2.2800 2.2800 0,4000 0.4000 0.4000 0.0 0.0 0,0 3,5000 4.9500 6.0000 3.0000 3.6000 4.0000 1.5000 1.5000 1.5000 3.6000 3.6000 4.0000 2.2000 2.2000 2.2000 1.0000 1.0000 1.0000 0,6000 0.6000 0.1500 3.1000 3.2000 3.2000 1.2000 1.2000 1.2000 2.0000 2.0000 2.0000 1.4000 1.4000 1.5000 0.2000 0.2500 1.5000 0.0 0.0 0.0 0.7000 0.7500 0.9000 1.4000 1.5000 1.6500 1.2000 1.2000 1.2500 1.2000 1.2000 1.2000 3.1800 1.0000 0.0 1.6000 1.9000 1.8000 2.7250 2.9000 3.0000 1.5000 1.5000 3.8750

PAGE 153

APPENDIX IV OUTPUT AND RANK CORRELATIONS INPUT

PAGE 154

CO w " Q CO CO H O K tH in O CO CO C\2 147 Uocvcoc^^c^tÂ»-^c\ivpc^^c'^cvcvl^c^tÂ» HOC^-4o^C^'Xc^OOc^2C^H'^c^i<^HOu^OC^JOHM3C^OOcÂ«^ Â• ^ ', 1 ' vO OvO i^CNOvO-d-H-d-OO i^vO C^toc^C><^HC^lC^J^--4c^^iÂ£^C7^.H-5C^i I I to H iH H I H I CNi o WN CO MD -4Â• Â• Â• iH H H H H -4H to O H crs O Â• Â• . H rH i H H . -d-sOCNOstOC^-^O^CNJOH^ OvOCVCNOHOOOCviC^C--C\JtOOH'AHCO-4-ONOi Â•tosoo-4-Â»^Nc^r-c\ja)-4-i>-iACM H O O -dlA CNJ OHOÂ«)rHcnCOC^HtOt50-4-OC^C--vOC^OvO to Cr^^O O CNi C\>'CjNCSvdtOr-lr-lr-tO-4'^C~-tOCX CV -4H HI HCM Hi-t I I H H Â«-j I CNJ !>-C^CVC\J OnÂ£)-4HCJnnOvO CNiH-4-C^OC^iAU\-d-QvO-ci-f-lvO f^-^-4-OsHH COvOCvJ OOO lAc^HvOvO C^C\!CO c^H I>-tO ^-^-0 -4^0 H C^OvOvO -4-Â»AO c<^^ocÂ«^^-c^iÂ•^o<^ooH-4c^^c^^oJO^-4Â•r^c^iOÂ£>tovoOI^o^c*^ cavo "a OOOCDOCNj'OOCSOCDOOOCMCVOc^OOOOOOOOHOOOH I I I cAvo <^u^(r\c^Hc^c^r^cvO-4Â•u^-4^-c>u^c^^oc^c\Jtocvc^^u^c^JO-dvOr^ to -^tO H C\J lACM C~-rH OOjvOtO ir\-3-^vOvOsOvOvO O^tO OtO H C^CT^lAC^-ch CViAcAMDC^ICNJc^vOHOc^OOO-TMDCVvOOOUNvOC^OHCNiHH-d-CNiCN o o o o o I CNJ OOOOOOOOHHOCVOOOOOOOOHOOOO III I II O CM c^ ^-d-vo totoÂ»At-loc\;uM>-iAOir\u%vOr^iA-4'cy^tOvOCNc^H^O <^>AcAto<^-^OC>^ovo>Au^l^\u^tocv'Avo<^lAc\'c^^voc^votOC\J-4-^J>-CVvOC7sOOtOC3NCVCN)OHCMCVOtOO C^UMACJ^vO lAHC^!>-^MDtO C^O^OtO O^CVvO-d-OsOtO -^^OvO C^OvO O "A OOOOHOOOOOHOOOOHHHOOC^OOOOOOHCVJOO O O O o I o H O o o o o o i i UN o o o o o o o o I o o o o o I oooooooo to to o 1 1 C\2 lA 0\2 UN ON O O o -4Â• Â• Â• o o O C\2 U\r-OHOvOH'ACNtOC\2vOvOI>-C^HsOHiHO _*IA1A-^C\J -4C\J O C--lAr-vOvO C^C^C^cnOvO "AvO tO tO lAc*NvO H H rH OWN OOOOOOOOOOCVOOOOHOHOOCNJOOOOOHHOOO M CDOOOC3 0 00CDCD(3CDOCDf^C\lC\JCNJC\JC\JCVCA>AC\lHCNJCNJHCAHCAOCA I C-OCA^-C^CAC-C^C^Ol>CAJ>-OC~-OC-CAC^OCACAOOOOC^CACACA CAHvOtOOC>CA-4-NOÂ£^-4iAOHtOH<>tO-:tiAtOCAO<>ClNO-4-C\JtOCA^ -4-C3NCVVO -4H-4-ONU\cfNONC\2 CsJvOHO Â£>-vO OvO CrvCAHCMCV OnnO-4--4-sOCA C\2C\J-4CViACMOH-;t-C\JcA-*C\JiHOC'N-4HcAC\2 C^HC\JC\2C\JHCvJC\2tOCAH till I II rH I rH -4lA to C\! CA CA CA CO CA CA O lA o Â• Â• Â• Â• Â• Â• Â• Â• CM CM O CA H H CM CA -4C0CNrHrHC^CMrHC^lAHC0 CAMD tOCAtOCMtOvO-4-CVvOCfNONCAC^CAU\CM-4 tOtOCM rHvOtOI>-HtO-4-CA-4-HC^tOtOrHO OCS I>-vO CV O . _ vO 0^ -4-vO O vOsO to nO . CVvOCfNONCAC^CAU\CM-4-^CACV t>-tO >A i >A CAvO rH CM O -4m3 iA CA CD CA CMCM-4THOrHOO-4CMvOC^-4rHOCMrHCM -4-CAvOrHCMC^HrHcACMOir-rH ill I H CM CA-4lAvO C^tO O O rH H H C\2C^-;flAvOI>-tO CNO rHCMC^-4U\MDC--tOOOrH HHrHrHrHpHHHCMCMCMCMCMCMCMCMCMCMCACA

PAGE 155

1A8 o PL, Eh Â§ O CO CO cd O S -p a, g < en a: CO o CNJ O Â•H -P cd M PL, o -P Q U M w H a< EH el CO Cf\ ot3 I OHO U cd\0 P-. CO H Q vO O W H CL, EH CO C\2 Â•o S I O W vO Si W P^ EH CO ,. ,^\J(>ocv^-vocoou^cvo\ic^^^tx)<^(>c^^cv-4Â•c^c>tX)^^^"^c^u^vq -4-AO(^ir\iAvOCNJHOHC\JC\J-4-C7sH O^-d-C^-d-tO C ^tOvOtX) i^C^vO -*>o OH H CM I u^ H I Â» c\j I NO irv o u^c^-4Â•ONOr^oc^o-JHÂ«DC^of^o^c^Hcn"^c^--^tOvoc^^c^.-dtt)co-dÂ•^OU^^^O^-HO^^C!NH-dÂ•HC^JC0C^^C-O^^^C^i OC--Crs-d-OC^WN-C:N~tsOCViA-CO-J'CÂ«AvO'AC\i C^HC^vO-*OOsO'COOOOC~-C^iArHC^ rH C^O^^tO r-tX) HtX)vO 0^J>Lf\iA-cfvO "A-dT^tO t>-C^C^Â«5 CV r-i O C^CVtO-S' .... ^ . _4-c0i-) H H I H I H O H H H I -4H C\J I tÂ» NO CNJ I cvjONc\jcn<}-i>-!>-^tx)H ' o CNJ t> NO CAC^C^CTNirycVCNJOC^C-OlANOC^-tOO ONHr~-ci^oÂ«)NO-4-c\joto!>-CN-cyNc^r^pH, . _ _ -j-O-tX) u^^?^Hb^u^C^HC^C^OÂ«)ONQ^-^OtX)tÂ»C^C^ CJnnO o " u\ -4OJ C\J OHc^OOOOOOCNJrHOOOOHOOOOtX)OOH o o I H O O O O H vO O O H iH CNJ CV O CNJ I> H O O -4tX)C-HCNjN0CrNUNHI>-O-C0t0OOO-4-'AN0C^rHN0tXiNOOC^ NOO-4tX)tX)iAC :fHNONO'Ac^r~tX)-4-NOCNJCNJ-ci-tOCÂ«AW^C^O OHCN!OOOOOOHOOOOOOOOOO-d-OOHOOOOOOO I I tOHCM0-4-'AC>HH to to o to to Lf\o u^Â£>-irNcM/\to o t>-o lAOo CJNc^^-p^ o-^mtv^-vo cÂ»^u^u^ OOCVOOOrHOOHHOOHOHHOHOHOi-IOOOOOHOO OOOCDCDOCDCDCDOCDCDOCDOOCDOOOOOOOOOOOOOO (^H-:t^\Jtou^totoOc^c^vocf^c^JNO^c^-dÂ•Â«^u^c<^cNJCJNto4ONCO-dÂ•u^<^_^ ooto^--4Â•ONOlAC^>cvo^^oHNO'ACNJr^c^icN^^^OHNO^^ow^^^'^'-^NOc^J OHHl^CVHOOCNir^r^OO-dCN^OHCNJ^^OCf^OOHr^H^-lr^CNJOO I i i i I I o NO r-J o ^ONtOC^-J-CNiCNJ . _ . _ r-uNo i.rNH!>-c^Lf\o^-j-ot>-ONOto oocvooo ----HONOtOtOtOCJNHo^OCVHOC^OO r::;cN!QNO-o^O c^O O to-*NOCrNOOtOrHcÂ»^t>--*NOHr-u^to -dÂ•CJNHc^^c^CNJl-i!^-lAOOu^toc^r^ S -^f H lA O' O NO CN-cf O UM.'^ CV' 0^ O "AC-CNjc^tOOtO-tOtONOCTNHC^H -C:-Â£>-C^Oo^OOQOC^OC-C^C-OOC~-OO^C^cnc^OC^C^C^O^OOc^ O NO Q O NO C^nO OnO CrNON!>-ONNO CAnO UMr\Lr\_4-tO !>-CNJ >Ac^-d-H CN--4-CN! H-d-ONOCNJ CV -d-OCjNCNJ O CNJ O^tO C^.4-C^HnOC0 CJnnO c^totototo O C^-4-'A OCNIrHHC^C^HLTNCM H I o CV O NO NO o I H to (JN c\2 O i O CNJ I H(HÂ«0WM>-nOCÂ«AO I I H On NOCV-^CNJi-l-4-ONONOO^-d-C)NtO-4--4-CVlAo^(>OCNJ(NJC>ONOC\'NO>AtOCVC^ NOOCN^c^tocÂ»^tOf-ltou^tooc^cv^oc^-d'-'^tooNOC3NHu^tocA"^r^c^JVf^u^ ^--4-CNjr^HCNJC^o^c^-d-CVC^CNJiAir\-j-CNJiAC7N-4-ONOONOi>-OOCNJtOo^Lr\ rH CNJ C^O CNJ I CNJ u^ o CNJ I CNJCNJOCM-JOOCNJCNJOHiHH H O O O I I CNJ H I C^CNJ OJ CA Oi^-:fi/NsO>-tOOOrHCNJ C^-d-VAvOC^tOONOHCNJo^
PAGE 156

149 CO CO O S O tH CO CO n3 O W ^ Eh -tJ Oh <; o =H CO CO cv " O S in o -p o <5 f:^ n: Â« Eh 1^ CO 3 c^c? iH cv iXoN H c<\to cv ^:-o^c---J^-l^c^:oc-toc^o^^-toc^^^--4oc^.cf^^^r-t^-cÂ«^ ,.Â•.Â»Â•Â•Â•Â•..Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•Â•*Â•Â•Â• to o u^vo Â«)^ovO^OH^-oÂ«l<^Ovo^oc^^^ou^of^^JOOs-4;TOÂ« ^^c^^oHo^\-^o^o'^-c^c^p^^c>-a^r^o^ I H H H H H I I o CNi iXN J>H vOC>i^vOvOtOC^C\ivOr~-C^vOC3^tOoc^c^^oc^tx)--dH^-cv |c^-*H vfNr-ic\JHHlHC\iC\JH-3-H HHC\ii-l c^HCMH CV i I I -d-CÂ«>WNvOtO O^O to ir\CNJvOtO OtO "^-^H CV UNC^vi/Nc^H C\i C^CNC^CM C^-*-d-CvJvO O 0-CV CO CV C 3T^C--Â±C<^vO OvO HCO tO HsO vOUAtOC^tOOH^^C<^HC^tOcnO^-4Â•^OOCVC^C^^O-4Â•0^2vOC*^sOC*^0^ ^^C>r-^OOOc*^CDl-^-^^^^^I^C^IoO^^OHHHHOOOOOOHHHOOO HHtOHt>HtO-d-HC\; '^'^c^vOOtOtOHHCÂ«^vOvOtO-4-<}-tOr-1tO!>-OC\i QH osc\;-^c^>^CNr--ototooNtoc>cN2C>c<^HvoHiH'^tor^Hc\jcoc\jOt>t>-4tovo-dÂ•c\^toc^^^c'^c^c^otoc^lAOMDcv--tcoolA^^^l^c^r^^O'^Hs0CN!C^i'AU^C^OW^CNi^O-4C0O<0C<^ cÂ«^^ococ^c^MDu^o^c3^0vocN2HOtocvHcvc^o^'^*^c^to^o-^C^^-H?^cy^^oH HOOOOOHHOHHOHHrHOHi-jHrHOOOOOOOOHHOOOH OCDOOOCDOOCDCDOOCDCDCDOOOC^OOOOOOOOOOOOOOO O H o c>H O d-t~-C\' C^OOnO -d-vO f^<^-*C^J lAvO f-VNc^cV UAC^H I>tO u^vO OvO C^vO r^M3-d--d-iAcyscNj u^w^cÂ»^vO i^o-o HOcv c^c^o HO o c^iao^cq h hco O-^-HOOHOOo^CNiCNJOHOHOCNiOOOO-d-H H O O o o o o 1 CNJ o ooooooooooo I i r-i oooooooooo ooooooooo I o o CO CJN to o o o o o o * Â• Â• Â• Â• o o o o o so H CO Â— ~ ' * O O O O ( lACACVtO i>-r\J O OsvOCO O H C^OcAOstOtOvOvO OtO CNJ O^LA -^-'ITNirNCAOCNi ONvO cntO >AirN!>-cnOÂ»AO-UNcA'-A>Ao^CACV C\' C^MD OOOOOOOOHOOOOOOOOOOOOOOHOOOO oooooooooo I I o o o o I oooooooooooooooooooo I I l>-Ot>OCÂ«-OOC^Or^c^Ot>-Oc^^-Oc^Ol>-cnOc^C^C^c^OC^!>-00000 cv^-o^ic^Jor--<^I-fc^voc^Â•c^.^olAtou^HtotocNJ^-vOHH-dÂ•HCNJOHOcotocNl tO^OC^r-^tOC^ON-3CNJOc^^-^OHtOH:^CVC^CN2C^JCVOCVOC^JC^Ht^--dÂ•CAH-d^^ HcAC^2HCN2H^-^CNJcAHHlACNiCNicA-^cAc<^LfN-dCN^CNJ-4Â•-dcACN2u^cNiCN^CN2c^^ o j>-o c^o t~-c^cAcAcAC^c^o r~-o cAO c^o c^o c^c^o o c^cao c^tvc^^-oo^ 0\JsOvOH<^LACN^^OM^-:^CMAC^^-O^^CN2^^CÂ»^ C^OC^^-ilA^-OC^CJ^C3^t^-C^i-^tO-4Â•CO -*^o I o CM C^ r-i WNnO -d-O H H H 1 H 1 H to CNJ vO H CNJ o -4" to O lA CO vO t-i Â• Â• Â• Â• Â• Â• H 0^ CNJ O i to CNJ -do 'AO-dÂ•u^^co^Osc\ivO O O^iHtO C^vO UNCÂ—vO i>-0 tO -ctH O -4--:}-^t0C0 CAOCOCOvO CO U\tO OnO !>-lAH WNlTNlAvO WNUNOH -4-vO -*C0 U^tO CNJ ^-f-vO CAO C^vO CNJ CNJ CAH COvOtOHCVOCNIr^-^iAHvO'^CNJ'AOc^'A^HHCNJC>c^^-HC>C>r--HCCC^CNJ-iO O CNJ I CNJCNJOrHO-4-HC^CtO CJnO HOJ C^-d-VNvO vOsovOvovovovo ^-r>-c~-^-c^i>-c^i>-c*-c^totoco tococo to COtO to OC^C^nCT^CJ^CJ^CTn

PAGE 157

150 CO to fv-^ H -o W 1 ctJ O K !h Eh -P a. , U u <; o tH CO CO H o S f-i Eh &! O Â«H n3 -P O <: CO CO C\! Q . . S O ;h Eh Â£| O Â«H -P CO 1-:! -P a; nj Â•< cu o a. Eh O H Eh O Â•H a. CNJ Q CO o CO o w o fL, Eh CO H I vO ; 1 I H Â•'"5 ; o t U i . Eh CO Oj O W sO Ph Â£h to _ , , .HOvO r^vO r~-otx)txioo^c\ji-ioocM iH H -4I I I H H I H 0^ -4UN H o H to vO CNJ H CO UN CNJ * o H H -4vO o UN ON Â• o CNJ H NO HCVONCNJC\!vOC^C\JOCOC^CÂ«Ar-HH^<^tONOtONOt>-tOC\J UNO^H-^CV HUNOUNCfNC^-J-CNJ UNrHUN-tCNJ CNJHt0H00sHCN-4-t0^HC-HHt0ON0-4OCVÂ£>-OC0-4-Hv0UNOOt>-t>-HrvOC^CV-4Â•vOOUNUNCÂ«^C^J-;tvO'OHtOOC^U^U^HC^HCN2C^^CNJ^O^^-4UN<Â»^^OHO-4Â• cotoc*SS-i-(c>-oo -rfvo cSr-c^r^usc\j o'-Â±uncnjso t>-oco unh f-co -4-tototo -4vO J> H H I I HOCVOOOOOOHOO ^0Â«)HtOC^C\J^-C^HHU^HC^^-0^0^0-4C)NtOC^C^^-r^CN2t^--4Â•U^C^JCNJ<>OUNO taOUNCNJ t>-CVOHOHC\JCV UNCVOvOHUNvOHC^>UN^-CT\0 [>-UNUNCrNUNcrNUNOstO -4--4-C^Â£>-UNÂ£>-J>CV UNvOvO -dT^CNJ CV 00 -Cj-UNvO C^O CÂ»J OCN! r-CNJ C^C^CVJ C^O -J-CN HOCVCDCDr4c3iHHCDOHOOCNJUNtO CNJ r^O UNc^CNJvOCJNOc^-d-OiHOHr-C-CNJ C^c^UNHO C^C-unOCNJCNJ HHvOC^C^ I CVON-H 0!>-v0 O^O ON-4-UNI: j-COON-4C^OC<^OvO O.VO -cf<}-HCNivOUNC^QCr\-4-CVUNOvOUNr~0ofNO UNtO C^HvO C^O -4-i-!C--tO C\2H^-C^UNC^OlAOcr^C^r^COUNtXlÂ£^-UNr-^^OUN-ctOtOCVHr-^^OUNOtOUNCJ^O-4Â•^0 OOOHOOOHOOOOOOOOHOOOOHOOOOOHOOiHOOO OOOOOOOCDOOOOOC>OOCDCDC3CDOCDOOC3CDCDCDCDOC:)C>C:^ I I I c^o o o^!>o^f-o c*NO r-c--cfNO c^o o o or-c^o^c^p^ir-D-o !>-c^o^c^c^o cÂ— -d-tO CÂ— iH !>-CNi i>--c}-|-| C^C^-d-HOtO -4-C^CjNtO C^C^OtOtO O^H I>-UNOCV UNO^H CNJ OvOHvOOtOHc^tO-J-OOCJNOOvOH UNvQ cV -4-HC^OOtOCV UNCN' C^Oc^tOO ^^^^JCVCVC^JCV0^v0OH-^C^HHCVCV-C^C*^C^-4^OU\v0-4C^HHC<^-4HC^^^Cf^CV c^r^Oi>-OOOo^t>c^OOJ>-Oc^c^OOc^c^r^crNOOcfNOC~-c^I>-0<*NOÂ£>c^ vOCOCrsc^vOHHUNtxDOOONOCV C^vOr^NvDOtOvO OtOtOOHHUNOCV UNUNUNCV UNOCVHU^^-C7^^^0Hr^tOC^CVC>C^O-^Â•UNO-ClÂ•U^HtOC^JOOO^OOc^C7^C^^UN C*N -J' 0^ I C^CVHC*-H-4-nO-4-C^C^MDC^C\JC^CV HUN-d-HUNCVJvOOC^C^r^HO I I I i CV to H o -4 CV UN H Â• C^ ON OvJ vO CV CV UN UN Â• C*N H -4TO CO -4" O to H CV r^vO vOrHHOH UN^^U^cVO^^^OC^^C -cfsO H Â£>H O CV UN t0-4c^CVC0C>0,'C0v0J>-O-4-0NÂ£>-t0 CO C^vO ^-O-d-tHt-ONOvOONr-JcV-4-C^HH Â£>UN^OUNC^CJ^OC^^UNCV
PAGE 158

151 o cn CO n5 O K cd -P p., Oj O tH CO CO H EH i 2: H Â•-ja -4Â«u^HHC^<^Ovoc^Â•-^vocvo^OMDu^^OMDU^^oto^-HC^HO^ c^JO^O"^^^coOOO^^^^^w^cvc^totX) OtOOOOOOOHOvOtXjvOHO^O^ <^Ou^c^2HC^OO^-w^u^o-d^CO^vOtX)cÂ»^Oc;-\vp vo ir\c--vo c'^vO "Ncv cvtx) -4-cr\c>-c-to H C^C^H o^o O H O tX) O I H -4CM i H C\ H I OS tX) vO vO H o H -4!>-4O . H 1 -4tX) o o vO {> CvJ ON -4 I> !>CO Â• Â• Â• Â• rH to -4o o o o o o H H -4-i^tM r^-4I On -4U>Â£>-tX)Lf\UM>-CNJrH-4-C\i-40HtOOHC\JiHo^CVH iAÂ£>-CNi-CO WNC\JtOOSvO-I>u^^^tÂ»ON-4H^oc^2toc^^oooscY^-4o^-4HOtx)^-HlAtx)Hr^o^a^ ^-oscou\c\i c^^i^o^I H I cv H H H H -4 H-4-UNU-\HCM i/\l>-H-C0ONU>tX)iX)Ht>-CNJ HHHOOOHUM>-Oi-(i^Â«)Ot>-C3N OC^itX)-4-4Â•C'^Ou^^~C^^'^H^-0s-4v0OOCM-4^0O^0C^^C^JC^0^v0C^^0H00 volA^--3o-4Â•ONtx)H<^^~cvHcvc^of^c^^o^^-o^ootx)c<^c^^c<^'^cvl^^o^o-4tx> OOOOCMOOc^OOOOHOO HOOOOOHOO I CNiHOOOOHOCXJ COvO-4CrsOt)0 r^tOvOC-C^vOCV CM ^-mUNOr^vOi-lOrHH'^OstXJCVtaHCJNCVH _u^ocyvoo-4Â•c^tx)OsOt>>^^co^-c^w^u^vo C^cA-4HC>C\' C^OOr>SuNr-iC^rHOCJNr-iC^-4H-4-C^vOCM Lf\t)OC^Ho^-4CNJOJ H OOOOHOOCNJ OOOOOOOOOOOO t I OOOOHOOOOOHOCM ^OvOO^Â•-4C^<^C^U^^-HC\J^OvOr^Osr^-4vOOOCVC^i^'^sOvO I>-I>-'^-00 i^C^OOvOO'COvO (rNI>-ONONU-\;>-vO r-i^O -4^-sO cv CVvO oco C~ 4C~OOOOOOOrHOOOOHOOOOOHOOOOOOHHOOOOOH OOC^CDOCDOCDC>CDOOOOOOOOOOOOOOOOOOO ^^-HHHCT^^-voHc^Htx)H^-l^c^Jvoc>^^^^c^^r^c<^cr^o^tQ"^c^^Â•^vou^^ i^^ I ii-N tv^ I LTV <^ -ivi oj -CO C Â— O tSTX) -S" Uf\ _4 CM ~ " CO O ...... ^ _ . . ^ H^^~-HÂ•1^ C'S OOrHHOOOHOOCM OOOOOOOOOOOOOOOOOOH I I ooooooooooooo I I -4 o o-\ H H UA CO vO O o o o o o Â• Â• Â• Â• Â• Â« o o o o o o CM OvOMDCMC^tXiO OCMvO^-4--4-vOO-4-!>-i^ Q i-IOOOOOOHOOOO ^-C^r-l-4-H'>--4-0Nv0Â£:^t0lX)Ov0-4-Â»^O-4c^tXl .4-cMONp-l-O O r-i O vO -4H -4tX) CM _ ~ OOCMOOOOOOOOO O c^vO H o o o oooooooooooooooo o o o o o I o o oooooooooo I I OC>-c^C^t>-o^Â£>Oo^r^OO[:^OOOC^c^c^OC^c^cnc~--Oc^o^c^C>-0!>-C^O C^HJ^^OtÂ»0-4tX) CMri^HCMHfHLa-4rHCMCM-4-UAH-4-C^C^CM-4iHcsJCMCMcf\HCM-4CMCMHc^Hc^ c'^H<^c;^o^ocovoONOc\2C^tx)^>o^o^oc^c^-4Hc^-^u^sJD<^-^^n^O^-C^HtX)W^C^lAr^C0CMO-40^Hc^tX)HtX)"^^--4Â•v0-4HÂ£^-^^ C^CM H o I CM I lACMHCMCOHOo^C^CM cf^ H H I CM H >A to H O CM I l/N O -4-vO O to Â• Â• H CM to CVJ ^9 rf\OONONiHtONO-4-C>tOC>.COOsO-4-4-CMCi^C>vOC^-4-r\J Â£>CMCMO-d-OiH lAvO ir\ Â• oo^oc>^oc\2-4^^to^-c^c^Ovoc^Â•tou^oc<^-<^co-4oc>oc^Hr^^-4Â•^-^^co OO-4C^HC^iAv0rHOc^t0O-4-v0rH0i^t0v0Â»Av0-4v0C3Ns0v0OHCJNH^-v0C^ CÂ«^HHi-IHCMC^-4-iH0^pHCM rHH^CM-4-HiAOHC^CMC^O H CM 1 O O H H H I . -~. tOCJ^Ot-^CMC*^-4Â•W^^OI^-tOONO^^CMC^-4lAvO^-tOC^Or^CMc^ c^c<^c^c^^^c<^c^c^c^-4Â•-4Â•-4-4-3Â•-4<^<^-4-4Â•'J^u^vAu^ H CM UNvO ' 1 cf^ c^ ^ _ H rH rH i-\r-{r-\r-it-{r-{r^r^r-{

PAGE 159

152 CO to Â§ -o i O p( >H Eh Softs. a, o tH U\vO O C\J vO vO -4-HvOvOC>tOr-'^CNi-ltX)Hi^OOV\ONC<^Lr\HOO i^vO O CnvQ C^COU^(VO^-dCVHC^^V^1Â»/^M^O<Â»OHc^-4c^i~^tno-4^ou^c^v0^to-Â±vOH OS C) ^ I-l I I H U^C^C\iC^i>^Os^OOsc^JOH-^COO<^-4sOÂ•-^C^r^'^^-C^^r-lQtO C<^OC^20svOCOH^-^O^OOtÂ»0^Â•-40svOU^C^-C^COC^Jr^CVOO^^ c^r^ -4H HCV HC^r-HCMH IHOJCVHI HI -dI II I I Osi/MrN-d-^AtO-d-CvJ tr\c^-4-tOCVC^tOC^H'^OssOC^CV lAtO C^r^C^JW^^-tX)HO^"^ ^-Â•-Â•Hc^c^-tO-40 sO -4 C\2 01 OOsvOHOstOCOC^J^-'AHc^^>-0^-4-4Â•'^OOOC^JCOW^w^ OHHOHOHOOOOCDsOHcnHOOHOOOOOOHHHHHOOH C\2sO~tOO>OsCrsOÂ£^ C^sQ CO sO C!dCOCO-4-OHC^Oo^Â£>-CH(^i-l Â»/NC\JC0OO0^C\iCviir\vO-4-O iAOsC^H-4-C~C^^-4-Â£>-l^Os> _ mOsCN2-4C\!OsCNJCOC^W\OscVOCO CV -4sO Os-:t-TC:scV c^O "^cnc^H -TtO Â£> OS Â£>OS cr^ lA sO r-l C\J OiHOOrHOHOOOOO"AHc^OOOOOOOOOO I I C\2 HOOOHOOH r^OHOC\2-Hvr\tO voO-0^sn-d-Ci^HHOsp OS H 'A sO Â£> OS J> o r-so -4 OOOHOOHOOcnO O -4O O H O O C\! CV r-l !>O cv H : , o o c^ o H . H ri -co t>to ^ 4 O sO to lA lA OS rH CNi Oi CV O O H O -4H r\ H so H O o O H H o O Â• Â• Â• Â• Â• Â• * o o o o O o o I o o I u^o^-coc^ivotx)sOvoc^H^-osHHc^^-os~^^^^Hc^Hcr^r^o^-o-4Â•c^Jso^-Â«^ -4 H c^vO CV-4^OsOt0CVA-4Â•t0U^^^-^-C^2C0^^s0H Q OiOOOOOOOOHOOHHOOOOOHHOOOOi^sOOOHOHH 0 ooooooooooooooooooooo Fill I I I I I I I I cnOC^QJ>-OOi>-OC^OÂ£>-00!>-C^Ot>-C^cr\Cr\( vOOtOOsC?sCNitOsO>^OOOC\JHOs_:tiHc\iH O -4 H 0^ C\J H O I CM ir\ OJ c\J c\i C\2C^Os-^UNOsHCO-4CV-4-sOtOsOOstX)HCVC\JtOsOCsiOHC\JtO-4CVC^Oc^O^O ~ " k CNJ -4Â£>to to sO r-l Â»A OS U\ H so sO O CM O H to so -4H O -4 r-l C^ -4 C'N < HHOsc^^-4-4w^lAc^Oc^^soHHÂ«^c^^OOO c^cr\HCMc\iOiHCNic\JHOiI I I I I iH I II I o CN o I so so vO -4. -4 CM o CV so OS OS OS CÂ«^sO to Â£> Â• Â• Â• Â• Â• Â• Â• Â• CM 0^ H H O o 1 CO CM Â•OS I O W so u oi ^ CO S3 O sO H -4 Â£>sO OS H I I I-) '^-4CM -4(nOC^-4-C^O UNCJssO CMsO Os C-sOsQ CJsCTsCM C^WNvO l^C<^>^^ H CM H lAvO 0it0-4-iH-4U^v0^-t0C^^C^01t0H'AOt0iHCM rH-4-CM "^0 cnCMC:sLf\c^O0sO-4HiH-4Os0-4--4-O OOtOOSrHiAvOsOsOO-4-OsH -4HC\JCM-4UACMHcr\C\CMOO I O I HCM 0^-4OHCMC~--4HOsOO-4C\i I t I I I I <^u^M^c--cocysOHc^ic^-4lAsoc^cocJso^-^c^2c^-4Â»Avoc^toc^^oHc\^c^-4Â•u^vo sqsOsOsqsOvOC^C^C^C^D-Â£>-C>-i>-C^C--tOtOtOtOtOtOtOtOtOtOC7sOsCJ\C7sCysCrsOs HrHHHHr-ii-^r-iHrHHHHHHr-iHr-ii-ii-irHi-ii-ir-iHrHHrHr-it-irHr-iH

PAGE 160

to CO c\ W -P Pi d, "< o
PAGE 161

BIBLIOGRAPHY Alexander, Sidney S. "Income Measurement in a Dynamic Economy," revised by David Solomons, in Studies in Accounting Theory, Second Edition. Edited by W. T. Baxter and S. Davidson. Hcmewood, Illinois: Richard D, Irwin, Inc., 1962, pp. 126-21?. American Accounting Association. Accounting and Reporting Standards for Corporate Financial Statements and Preceeding Statements and Supplements. Iowa City, Iowa: American Accounting Association, 1957. . A Statement of Basic Accounting Theory . Evanston, Illinois: American Accounting Association, 1966. American Accomting Association, Committee on Concepts and Standards Underlying Corporate Financial Statements. "Price-Level Changes and Financial Statements," The Accounting Review , October, 1951, pp. /f68-474. American Accounting Association, Executive Committee. "Accounting Concepts and Standards Underlying Corporate Financial Statements," The Accounting Review . October, 194S, pp. 339-344. American Institute of Accountants. Changing Concepts of Business Income. New York: Macraillan and Company, 1952. -Anreder, Stephen S. "Pitfalls for the Unwary," Barrens , December 24, 1962, pp. 3, 8, 10, and 13. Amett, Harold E. "The Concept of Fairness," The Accounting Review . April, 1967, pp. 291-297. . "VJhat Does 'Objectivity' Mean to Accountants," The Accounting Review . May, 1961, pp. 63-68. Arthur Andersen and Company. "Establishing Accoiinting Principles A Crisis in Decision Making," Arthur Andersen and Company, 1965. Backer, Morton, and Bell, Philip W. "The Measurement of Business Incom.e, Part 1 The Matching Concept, " Modern Accounting Theory . New York: Macmillan and Company, 1966, pp. 68-74. Baxter, VJ. T. "Recommendations on Accounting Theory," reprinted in Financial Accounting Theory ; Issues and Controversies . Edited, by S. A. Zeff and T. F. Keller. New York: McGraw-Hill Book Company, 1964, pp. 427-A37. 154

PAGE 162

155 Bedford, Norton M. "A Critical Analysis of Accounting Concepts of Income," The Accounting Review , October, 1951> PPÂ» 526-537. . Income Determination Theory ; An Accounting Framevrork . Reading, Massachusetts: Addis on-â€¢esley Publishing Company, 1965. Bernstein, Peter L. "Profit Theoiy Where Do â€¢e Go From Here?" T^iS. Quarterly Jovimal of Economics , August, 1953> PPÂ» 407-422, Bimberg, Jacob G., and Raghu, Nath. "Laboratory Experimentation in Accounting Research," The Accounting Review, January, 1958, PP. 38-45. Blough, Carmen G. "Changing Accounting and Economic Concepts Affect Methods of Inventory Pricing, " Tte Jo^^^al of Accountancy . September, 1948, pp. 204-212. Boxilding, K. E. "Economics and Accounting: The Uncongenial Tvans," in Studies in Accounting Theory . Edited by W. T. Baxter and S, Davidson. Homewood, Illinois: Richard D. Irwin, Inc., 1962, pp. 44-55. Bradish, Richard D. "Corporate Reporting and the Financial Analyst," The Accounting Review, October, 1965, pp. 757-766. Buckley, John V7., Paul Kircher and Russell L. Mathews. "Methodology in Accounting Theory," The Accounting Review, April, 1968, pp. 274-283. Canning, John B. "A Certain Erratic Tendency in Accomtants' Income Procedure," Econometrica. I (1933), 52-62. . The Economics of Accountancy . New York: The Ronald Press Company, 1929. . "Some Divergencies of Accounting Theory from Economic Theory," The Accounting Review. March, 1929, pp. 1-8. Cannon, Arthur M. "Tax Pressures on Accounting Principles and Accountants' Independence," The Accounting Review. October, 1952, pp. 419-426. Larson, A. B. "Cash Movements: The Heart of Income Measurement," The Accounting Review. April, 1965, pp. 334-337. Chambers, Raymond J. Accounting , Evaluation, and Economic Behavior . Englewood Cliffs, New Jersey: PrenticeHall, Inc., 19661 . "Edvjards and Bell on Business Income," The Accounting Review , October, 1965, pp. 731-741. . "A Matter of Principle, " The Accounting Review . July, 1966, pp. Z^43-457.

PAGE 163

156 Ch^irchman, C. West. Prediction and Optimal Decision . Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1961. Cotter, Arundel. Fool's Profit . New York: Barron's Publishing Company, 1940. Cowan, Tom K. "Are Truth and Fairness Generally Acceptable," ThÂ£ Accounting Review . October, 1965, PP. 78S-794. ^ox, Garfield V. "The Relation of Stock Prices to Earnings," The Journal of Business . October, 1929, pp. 3S3-395. Deinzer, Harvey T. Tte American Accoiinting Association Sponsored Statements of Standards for Corporate Financial Reports ; A Perspective . Gainesville, Florida: Department of Accounting, College of Business Administration, The University of Florida, 1964. . Development of Accounting Thought . New York: Holt, Rinehart and Vfijiston, Inc., 19o5^ . "Explanation Strains in Financial Accounting, " ^The Accounting Review, Janviary, 1966, pp. 21-31. . Methodological Presuppositions in Financial Accounting Models . Gainesville, Florida: Department of Accounting, College of Business Administration, The University of Florida, 1968. de Roover, Raymond. "The Development of Accounting Prior to Luca Pacioli According to the Account-books of Medieval Merchants, " in Studies in the History of Accounting . Edited by A. C. Littleton and B. S. Yamey. Homewood, Illinois: Richard D. Irwin, Inc., 1956, pp. 114174. Devine, Carl T. "Depreciation and Income Measurement," The Accounting Review. January, 1944, pp. 39-47. , "Research Methodology and Accounting Theory Formation," The Accounting Review . July, I960, pp. 387-399. . "Some Conceptual Problems in Accounting Measurement," in Research in Accounting Measurement . Edited by R. K. Jaedicke, Y. Ijiri, and 0. Nielsen for the American Accounting Association. Evanston, Illinois: The American Accounting Association, 1966, pp. 13-27. Dewey, John. Logic ; The Theory of Inqviiry . New York; Holt, Rinehart and VJinston, Inc., 1938. Dickerson, Peter J. Business Income A Critical Analysis . Berkeley: University of California, 1965.

PAGE 164

157 Edwards, Edgar 0. and Bell, Philip W. The Theory and Keasurament of Business Income . Berkeley and Los Angeles: University of California Press, 19^4!^ Edwards, Ronald S. "The Nature and Measurement of Income," in Studies in Accounting Theoiy , Second Edition. Edited by W. T. Baxter and S. Davidson. Homewood, Illinois; Richard D, Irwin, Inc., 1962, pp. 70-125. "An Essay in Etymology: Revenue, Income, Profit, ani Earnings, " T^ Accountant , April 9, 1955, p. 391. Fisher, Irving. The Nature of Capital and Income . New York: Macmillan and Company, 1912. Grady, Paxil. Inventory of Generally Accepted Accounting: Principles for Business Enterprises . New York: Am.erican Institute of Certified Public Accountants, Inc., 1965. Graham, Willard J. "The Effect of Chan^ng Price Levels upon the Determination, Reporting, and Interpretation of Income," Kie Accounting Review , January, 1949, pp. 15-26. . "Depreciation and Capital in an Inflationary Economy," The Accoimting Review, July, 1959, pp. 367-375. . "Some Observations on the Nature of Income, Generally Accepted Accounting Principles, and Financial Reporting," Law and Contemporary Problems . School of Law, Duke University, 1965, pp. 652-673. Green, David Jr. and Joel Segall. "The Predictive Power of First-Quarter Earnings Reports: A Replication," in Empirical Research in Accounting Selected Studies , 1966. Edited by S. Davidson. Chicago: Institute of Professional Accounting, Graduate School of Business, University of Chicago, 1967, pp. 21-43. Hackney, V/illiam P. "Accounting Problems in Corporation Law," and Contemporary Problems . School of Law, Duke University, 19657 PP. 791-823. Halvorson, Newman T. "The Search for Accounting Principles," in Symposium for Educators . New York: Ernst aM Ernst, 1967, pp. 23-50. -Harkavy, Oscar. "The Relation Between Retained Earnings and Common Stock Prices for Large, listed Corporations," The Journal of Finance , September, 1953, pp. 283-297. Heilman, E. A. "Realized Income," The Accounting Review , June, 1929, pp. SO-87. Hepvrorth, Samual R. "Smoothing Periodic Income," ^^e Accounting Review , January, 1953, pp. 32-39.

PAGE 165

153 HLcks, John R. Value and Capital , Second Edition. Oxford: At the Clarendon Press, 1946. (First published in 1939.) Hills, George S. The Law of Accounting and Financial Statements. Boston: Little, Bro;iffi and Company, 1957. Homgren, Charles T. "Security Analysts and the Price Level," The Accounting Review , October, 1955, PP. 575-581. . "Hovf Should We Interpret the Realization Concept," The Accounting Review, April, 1965, pp. 323-333. Iselin, Errol R. "Chambers on Accounting Theory," The Accounting Review , April, 1968, pp. 231-238. Johnson, Charles E. "Inventory Valuation The Accountant's Achilles Heel," Tte Accounting Review, January, 1954, PP. 15-26. Jones, Ralph C. "Effects of Inflation on Capital and Profits: The Record of Nine Steel Coii?3anies," The Journal of Accountancy , January, 1949, pp. 9-27. Kelly, Arthur C. "Can Corporate Incomes be Scientifically Ascertained?" The Accounting Review , Jxily, 1951, pp. 289-298. Kendall, Maurice G. Rank Correlation Methods , Second Edition. London: Charles Griffin and Company, Ltd., 1965. Kerr, Jean St. G. "Three Concepts of Business Income," Tte Australian Accountant , April, 1956, pp. 139-146. Kester, Roy B. "Sources of Accounting Principles," The Journal of Accountancy , December, 1942, pp. 531-535. Keynes, John Maynard. The General Theory of Employment Interest and Money . New York: Ifecmillan and Coiipany, 1936. Knight, Frank H. Risk , Uncertainty and Profit . New York: Harper and Row, 1965. (First published in 192T7) . "Profit," Encyclopedia of the Social Sciences , Vol. XII, 1933. ^^rson, Kermit and R. W. Schattke. "Current Cash Equivalent, Additivity, and Financial Action," The Accounting Review, October, 1966, pp. 634-641. Lemke, Kenneth W. "Asset Valuation ard Income Theory," ThÂ®. Accounting Review . January, 1966, pp. 32-41. Levy, Saul. Accountant's Legal Responsibility . New York: American Institute of Certified Public Accountants, 1954.

PAGE 166

159 Littleton, A. C. "concepts of Income Underlying Accounting, "_The Accountin.^ Review ^ l-Iarch, 1937, PP. 13-22. . structure of Accounting Theory . Madison, Wisconsin: American Accounting Association, 1953. . iiVftiat is Profit," The Accounting Review . September, 1928, ^ 278-2S8. Lorig, Arthur N. "Some Basic Concepts of Accounting and their Implications, The Accounting Review, July, 1964, PP. 563-573. McAnly, Herbert T. "The Case for LIFO: It Realistically States Income and is Applicable to Any Industry," The Journal of Accountancy, June, 1953, pp. 691-700. l-Iackenzie, D. H. "Contemporary Theories of Corporate Profits Reporting," The Accounting Review , October, 1949, PP. 360-368. Mattes si ch, Richard. Accounting and Analytical Methods . Homewood, Illinois: Richard D. Irwin, Inc., 1964. l^Iautz, R. K. "Btnphasis on Reporting not Calculation, Could Settle Income Statement Controversy, " The Joarnal of Accountancy . August, 1953, pp. 212-216. lÂ«Iay, George 0. "Generally Accepted Accounting Principles," The Journal of Accountancy . Janviary, 1959, pp. 23-27. . iiThe Choice Before Us," The Journal of Accountancy. March, 1950, pp. 206-212. . "The Nature of the Financial Accoxmting Process," The Accounting Review. July, 1943, PP. 189-190. Moody's. Moody's Industrial llanual. New York: Moody's Investors Service, Inc., 1963 and 1966. Morrison, Paul L. "The Interest of the Investor in Accounting Principles," The Accounting Review, March, 1937, PP. 37-42. -^Jerlove, S. H. "Recent VJri tings on Profits," The Journal of Business . October, 1929, pp. 361-382. O'Donnell, John L. "Relationships Betvreen Reported Earnings and Stock Prices in the Electric Utility Industry," The Accounting Review . January, 1968, pp. 135-143. Paton, V/illiam A. Accounting Theory . Chicago: A. S. P. Accounting Studies Press, Ltd., 1962" (First Published in 1922.) . Advanced Accounting . New York: Macmillan and Company, 1947.

PAGE 167

160 Paton, William A. "Measuring Profits Under Inflation Conditions, " The Journal of Accountancy^ January, 1950, pp. 16-27. . Paton on Accounting; . Ann Arbor, Michigan: Bureau of Business Research, Graduate School of Business Administration, The University of Michigan, 1964. and A. C. Littleton. An Introduction to Corporate Accounting Standards . Iowa City, Iowa: American Accounting Association, 19o4. (First published in 1940.) Peterson, miliam A. "Significance of Prospective Income Data," The Accounting Review ^ April, 1966, pp. 275-282. Pigou, A. G. "Net Income and Capital Depletion," T^ie Economic Journal. June, 1935, PP. 235-241. Rappaport, Alfred. "Establishing Objectives for Published Corporate Reports," The Accoxinting Review . October, 1964, pp. 951-962. Sanders, Thomas Henry, Henry Rand Hatfield and Under hill Moore. A Statement of Accounting Principles . Columbus, Ohio: American Accounting Association, 1959. (First published in 1938.) Schmidt, Fritz. "Is Appreciation Profit?" The Accounting Review , December, 1931, PP. 289-293. Shao, Stephen P. Statistics for Business and Economics . Colvmbus, Ohio: Charles E, Merrill Books, Inc., 1967. Solomons, David, "Economic and Accounting Concepts of Income," The Accounting Review . July, 1961, pp. 374-383. . "Income True and False, " in Studies in Accounting Theory . Second Edition. Edited by W. T. Baxter and S. Davidson. Homewood, Illinois: Richard D. Irwin, Inc., 1962, pp. 299-306. Spacek, Leonard. "Burdens Placed on the Investor by Questionable Accounting Practices," an address before the Financial Analysts Federation, May 25, 1966. New York: Arthur Andersen and Company, Subject File AD 7220, Item 63. . "Can â€¢e Define Generally Accepted Accounting Principles?" The Journal of Accountancy . December, 1958, pp. 40-47. Sprouse, Robert T. "The Measurement of Financial Position and Income: Purpose and Procedure," in Research in Accountin,g Measurement . Edited by R. K. Jaedicke, I. Ijiri, and 0. Nielsen for the American Accounting Association. Evanston, Illinois: American Accounting Association, 1966, pp. 101-1L5. . "Observations Concerning the Realization Concept," The ' Accounting Review . July, 1965, pp. 522-526.

PAGE 168

161 SproTise, Robert T. and Maurice Koonitz, A Tentative Set of Broad Accounting Principle s for Business .hiiterprises . New York: The American Institute of Certified Public Accountants, 1962. Standard and Poore's. Standard and Poore' s Monthly Stock Guide . New York: Standard and Poore, 1962 through 1966. Stone, Donald E. "The Objective of Financial Reporting in the Annual Report," The Accounting; Review, April, 1967, pp. 331-337. Storey, Reed K. "Accounting Principles: AAA and AICPA, " The Journal of Accountancy , June, 1964, pp. 47-55. \/_ . "Cash Movements and Periodic Income Determination," The Accounting; Review , July, I960, pp. 449-454. The Search for Accounting; Principles . New York: American Institute of Certified Public Accountants, Inc., 1964. A Study Group at the University of Illinois, Center for International Education and Research in Accounting. A Statement of Basic Accounting; Postulates and Principles . Urbana, Illinois: The Board of Trustees of the University of Illinois, I964. United States Securities and Exchange Commission. Regulation SX . Washington, D. C: United States Government Printing Office, 1962. Accounting Series Releases Release Number Four . Dated April 25, 1938. Washington, D. C: United States Government Printing Office, 1938. Republished in 1956. Vatter, X^illiam J. "Income Models, Book Yields, and the Rate of Return," The Accoxmting Review , October, 1966, pp. 681-698. "Obsticles to the Specification of Accounting Principles," in Research in Accounting Measurement . Edited by R. K. Jaedicke, Y. Ijiri, and 0. Nielsen for the American Accounting Association. Evanston, Illinois: American Accoimting Association, 1966, pp. 71-87. Vfemtz, William W. "The Impact of Federal Legislation Upon Accounting," The Accounting Review , April, 1953, pp. 159-169. Weston, J. Fred and David K. Eiteman. "Economic Trends and Security Values A Belak or Bountifvil Future for Investors?" Financial Analysts Journal , January-February, I965, pp. 24-31. --"VJhat are Eamijigs? The Growing Credability Gap." Forbes . May 15, 1967, pp. 28-44. Wilcox, Edv/ard B. and Howard C. Greer. "The Case Against Price-Level Adjustments in Income Determination," The Journal of Accountancy. December, 1950, pp. 492-504.

PAGE 169

BIOGRAPHICAL SKETCH Frederick Dozier Whitehurst was bom September 13> 193S, at Portsmouth, Virginia. In June, 1956, he was graduated from Saint Paul's High School in Portsmouth, Virginia. In June, 1964, he received the degree of Bachelor of Business Administration with a major in Accounting from Old Dominion College in Norfolk, Virginia. In September, 1964, he commenced his graduate studies at the University of Florida vAiere he held a three-year National Defense Education Act Title IV Fellowship until September, 1967. He received the degree of Master of Arts at the University of Florida in August, 1965. From September, 1965, until September, 1967, he pursued his work toward the degree of Doctor of Philosophy. In September, 1967, he accepted the position of Assistant Professor of Accounting at Old Dominion College. He is currently on the faculty of that institution.

PAGE 170

This dissertation was prepared under the direction of the ^ chainaan of the candidate's supervisory committee and has been approved by all members of that conmiittee. It was submitted to the Dean of the College of Business Administration and to the graduate council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. , ; . : Â• ; August, 1968 ..^-^^All Administration Dean, Graduate School Supervisory Committee: Chairman ^0/r/^ \ \