David Solomons
UNIVERSITY OF PENNSYLVANIA

THE TWILIGHT OF INCOME MEASUREMENT: TWENTY-FIVE YEARS ON

Abstract: The paper reviews events and trends since 1961,when the author incautiously forecast a possible decline in the importance of income measurement. He finds that little has changed in the intervening 25 years, and the forecast has not been borne out by events. Historical cost accounting has survived a period of serious inflation with hardly a dent. Earnings seem to be as important to financial analysts and to academic researchers as they ever were, and recent tax changes bring taxable income somewhat closer to accounting income than previously, thereby increasing the importance of the income concept rather than diminishing it.

“Each of us sees the future differently, no doubt. But my own guess is 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 measurement.”
The Accounting Review, July 1961, p. 383.

If I had realized in 1961 that I might be called to account 25 years later for that incautious statement, I would probably have been more circumspect than I was. I did, it should be noted, say “may be seen,” not “will be seen,” and I could hide behind that. But one should know better than to make anything but vague prognostications. Kierkegaard wrote that life must be lived forwards, but it can only be understood backwards. That seems to be as good an excuse as any for these reflections on my 1961 paper.

REASONS FOR PESSIMISM

Let me remind the reader why I took such a pessimistic view of income measurement. The statement quoted above came at the end of a paper in which I analyzed the differences between economic and accounting concepts of income. Accounting income, I argued, was incomplete because, aiming only to measure realized income, it did not take into account unrealized changes in the value of net assets accruing during a period, while it did include value changes, if realized, that had accrued in an earlier period. On the other hand, economic income (increase in well-offness) took into account all changes in the value of net assets that occurred during a period, both those due to changes in the underlying circumstances and those due to mere changes in expectations about those circumstances. (The distinction can be illustrated by comparing the changes in the value of oil reserves resulting from new discoveries with changes in their value resulting from improved expectations about the percentage of existing reserves that can be recovered.) In my paper I discussed Sidney Alexander’s1 concept of “variable income,” from which he attempts to exclude changes in expectations, and I concluded that such exclusion could not be made operational. This led me to the gloomy conclusion that neither accounting income, as it was then understood, nor economic income was a satisfactory measure of enterprise or management performance.

If by accounting income is meant income as accountants now measure it, i.e., historical accounting income, I see no reason to change that opinion. Yet the remarkable thing is that, despite the ferment of the inflation accounting debate since 1961, nothing fundamental has changed in regards to the practice of accounting for income. Accounting income, as a concept, is now what it was then, and it seems to receive as much attention now as it did then. Earnings per share is still the summary indicator that attracts more attention than any other. There is no support for my twilight prediction in that direction. It must be regarded more as a statement of what I thought should happen rather than what, in the light of hindsight, did happen.

To be sure, there have been some changes of emphasis. Though importance is still attached to “the bottom line,” analysts pay more attention than they did to the components of income. The growth in the diversification of business units has led to segmented income statements, showing separate results for major lines of business and for the geographical areas in which a company does business. The results of discontinued businesses must be disclosed separately. Accounting standards

1Sidney Alexander’s monograph, Income Measurement in a Dynamic Economy, in which variable income is discussed, was originally written for the Study Group on Business Income, organized by the American Institute of Accountants (now the American Institute of Certified Public Accountants) in 1948. It was given a limited circulation as one of the Group’s Five Monographs on Business Income. It was revised by David Solomons, and the revised version is reprinted in Studies in Accounting, eds. W. T. Baxter and Sidney Davidson (London, The Institute of Chartered Accountants in England and Wales, 1977), pp. 35-85.

now require a number of other special items to be disclosed, e.g., research and development expenditures. But these are changes of detail; they do not alter the fundamental nature of accounting income.

THE INFLATION ACCOUNTING DEBATE

The year 1961, which saw the publication of my paper, also saw the publication of The Theory and Measurement of Business Income, by Edgar O. Edwards and Philip W. Bell (Berkeley: University of California Press, 1961). For most English-speaking accountants unfamiliar with the German and Dutch literature on accounting and changing prices, this book opened up new possibilities of bridging the gap between economic and accounting income. (That is not to say that any but a few academics paid any attention to it.) By showing in detail the valuation and bookkeeping procedures necessary to implement a system of current cost accounting to reflect changes in specific prices, combined with price level adjustments to correct for changes in the general price level, the book brought within reach a method for giving an approximation to at least one variant of economic income.

What the Edwards and Bell approach did, as is by now well known, was to compute current operating profit (the excess of sales revenues over the current cost of goods sold) and then to add or subtract real holding gains or losses (i.e., holding gains and losses net of inflation), including real holding gains or losses on monetary assets and liabilities. It could not claim to measure economic income in the fullest sense, for it did not take into account changes in the value of intangibles. But this was hardly a defect, for what was lost in theoretical purity was more than made up in practicability. Concentrating on tangible assets, the system used current costs, which could be derived, in most cases, from market information. The net income that it measured did approximate to the change in the real value of net tangible assets. Anyone interested in attaching a value to intangibles could do it for himself, being better equipped to do so than if only historical cost accounting results were available

If, in 1961,1 had foreseen the serious inflation of the late 60s and the 70s — in 1960, the consumer price index rose by 1.6%, in 1980 by 14.4% — and if I had foreseen the adamant refusal of businesses everywhere to adapt their financial reporting methods to inflationary conditions, even when a feasible method of doing so was at hand, I might have been even more pessimistic about the future of income measurement than I was; and again I would have been wrong. After flirting with constant purchasing power accounting (CPPA) for a while, the standard-setting bodies in both the USA and the UK developed current cost accounting standards. In both cases, this was in response to government initiatives. Here, the SEC’s ASR190 (1976) mandated disclosures of replacement cost data, and this led to the FASB’s Statement of Financial Accounting Standards (SFAS) No. 33 in 1979. In the UK the government-appointed Sandilands Committee reported in favor of current cost accounting in 1975, and this led to the issue by the Accounting Standards Committee of Statement of Standard Accounting Practice (SSAP) 16 in 1980. But the flowering of current cost accounting has been short-lived. With the substantial diminution in the rate of inflation in both countries since 1980, both standard-setting bodies are in retreat. SSAP 16 has been withdrawn, with no likely successor in view, and SFAS 33 has been made voluntary (which is equivalent to its abandonment). Once more, historical cost accounting has demonstrated its remarkable hold on life in the unlikeliest of circumstances.

THE RECOGNITION AND MEASUREMENT FIASCO

The FASB had unusual opportunity to pave the way for improvement in income measurement, without committing itself to taking action in the immediate future, when it published Statement of Financial Accounting Concepts (SFAC) No. 5, Recognition and Measurement in Financial Statements of Business Enterprises, in 1985. This was to have been the crowning achievement of the Board’s conceptual framework, on which work had been proceeding almost since the Board’s inception in 1973. By definition, a conceptual framework is a statement of philosophy, not a standard or a set of standards, and it does not bind the Board to take any particular action. It does, however, impose a constraint on what the Board can do if its actions are to be seen to be consistent with its words. SFAC No 5 was expected, at least by some academics, to examine the historical cost accounting model in a fundamental way, to evaluate the information it provides, to analyze its strengths and weaknesses, and to point the way towards possible improvements, notably in the matter of income measurement. In the event, it did none of these things.

In regards to recognition, SFAC No. 5 formulated what it called fundamental recognition criteria that an item should satisfy in order to be recognized in financial statements. The Concepts Statement listed four criteria, but there were in reality only two. The first was that an item must satisfy the definition of an element of financial statements (as defined in SFAC No. 3). The second was that the item has a relevant attribute that can be measured with sufficient reliability. The terms “relevance” and “reliability” were defined in SFAC No. 2.

When it came to measurement, SFAC No. 5 had virtually nothing new to say. In various places in the statement, it described the existing historical cost accounting model with a complacency that showed little desire for change. If changes are to come, they are to be the result of a process of evolution. With this statement, coupled with the abandonment of SFAS No. 33, the Board virtually abdicated its role as a change agent, so far as income determination is concerned.

One of the major gaps in SFAC No. 5 was in the treatment of “earnings.” The word was constantly used but never defined. Clarification of the meaning of ” earnings” would have been a real contribution. Is it possible that this was not done because the result would have disclosed a philosophical mess that would have been hard to defend?

As further evidence that little of a fundamental nature has changed during the last 25 years, one may cite the ongoing debate between those who give primacy to the balance sheet in the accounting model and those who think that the balance sheet is little more than a list of balances left over after income has been determined by matching costs and revenues in the income statement. The debate, of course, is really over the nature of income. The “balance sheet school” views income as the increase in net worth that has occurred during a period — the economic view, one may say — while the “income statement” school sees income as the result of certain activities that have been completed during a period. One cannot but wonder why this argument generates so much heat, especially among the enthusiasts for “matching.” Generally, the most passionate advocates of that view are financial executives, the preparers of financial statements. It seems likely that they take that position because matching gives them more control over the bottom line; matching offers a great range of options in deciding how to measure income.

There are still many possible alternatives among the different ways of allocating costs, choosing among inventory cost-flow assumptions, selecting among depreciation methods and” managing” the timing of asset realizations. Even though the proliferation of accounting standards has closed off a number of alternatives, a goodly number remain. On the other hand, a consistent “value to the business” model, substantially tied to current cost, would offer fewer alternatives, contrary to the usual charge that it would be too subjective, and that is probably why it is so fiercely resisted by preparers.

ACCOUNTING INCOME VS. TAXABLE INCOME

One reason for my pessimism about the usefulness of the income concept in 1961 was the gap, which seemed to be getting wider, between accounting income and taxable income. The more the tax system relied on its own definition of the taxation base, and the less it relied on income as accountants defined it, the less the importance that would attach to accounting income. Now the Tax Reform Act of 1986 to a limited extent reverses that trend. Examples are the equating of corporate capital gains with ordinary income, and the requirement to capitalize certain interest and other costs of producing inventory that formerly could be expensed. Each step that brings taxable income more into conformity with accounting income increases accounting income’s importance.

CONCLUSION

Many sophisticated accounting academics have lost patience with the income concept and have turned their backs on it. One of the best-known reactions of this kind is the paper by Beaver and Demski, “The Nature of Income Measurement” (The Accounting Review, January 1979), which concludes with a “challenge to accounting theorists … to address the primitive question of the propriety of the accrual concept of income.” But, proper or not, “earnings” (the practical embodiment of the accrual concept of income) has been the focus of a vast amount of academic empirical research in the last two decades, and though there are signs that the volume of this kind of work is beginning to abate somewhat, it does not look as though the concept is about to drop out of the literature any more than it seems to be about to lose interest for financial analysts.

But let me draw back from further predictions. One bad call every 25 years is quite enough.