Tax Analysts®Tax Analysts®

Article Archive

February 27, 1996
The Birth Pangs Of The Modern Income Tax -- An Early Treasury Study (Part 1).
Carl Shoup and Roy Blough

====== SUMMARY ======

Many of the policies being discussed during the 1996 debate on tax reform have been considered before. In 1937, two Treasury economists, Carl Shoup and Roy Blough, prepared a study that considered a number of tax reform proposals.

The late 1930s and early 1940s was a time of great change in the federal tax system. It was the period when the income tax was transformed from a class tax to a mass tax; it was the time when the modern income tax was born. The debate going on at the time about the structure of a federal tax system bears striking similarities with the current arguments over tax reform. Therefore, this 1937 study illuminates current discussions about replacing the income tax system.

====== FULL TEXT ======


Depression, War, and the Creation of the Modern
Personal Income Tax

ROSWELL MAGILL (September 20, 1937)


Chapter 1
Standards for a Federal Tax System

Chapter 2
Sources of Federal Revenue

Chapter 3
Income Taxation

Chapter 4
The Taxation of Business

Chapter 5
Manufacturers Sales Tax

Chapter 6
Death and Gift Taxes

Chapter 7
Social Security Payroll Taxes

Chapter 8

Chapter 9
Liquor Taxes

Chapter 10
Tobacco Taxes

Chapter 11
Automotive Excises

Chapter 12
Manufacturers' Excise Taxes, Other Than
Automotive Taxes

Chapter 13
Documentary, etc., Stamp Taxes

Chapter 14
Miscellaneous Taxes

Chapter 15
Recapitulation of Recommendations for Change


Between 1939 and 1944, the number of Americans filing individual income tax returns jumped from 7.6 million to more than 51 million. At first, this expansion of the tax base was driven by the chronically insufficient revenue yields of a depression-wracked economy. Later, the staggering demands of World War II made the need for money even more acute. Casting about for solutions, policymakers turned naturally and repeatedly to the income tax. As one former Treasury official recalls, "It's where the money was."

The mass income tax, in other words, was less the product of careful planning than it was the child of necessity. And, it never disappeared, outlasting both war and depression to become the centerpiece of America's postwar revenue system.

Historian John Witte has argued persuasively that tax policy is usually made incrementally, proceeding by small steps rather than long leaps. Over time, small steps can cover a lot of ground. Only rarely does a big change happen quickly and deliberately. The transformation of the individual income tax into a broad-based tax was just such a piecemeal process; revolutionary in effect, it was evolutionary in development. Indeed, the lower exemptions that transformed the income tax were enacted gradually over a 15-year period, stretching from the early 1930s to the mid-1940s.

To appreciate the incremental process of policy formulation, it is important to keep in mind what didn't happen. Federal tax policymakers might have struck out in new directions during these critical years. Change, after all, was afoot in Depression-era Washington. Officials might have opted for some form of a broad-based consumption tax. More than a few voices were raised in favor of such proposals.

Ultimately, however, taxwriters avoided that route -- at least until the revenue demands and inflation concerns of World War II made their consideration unavoidable. They preferred to stick with the devil they knew: an income tax that seemed more predictable than its alternatives. Consumption taxes, moreover, were unpopular with key members of the New Deal coalition, especially organized labor. Sales taxes, in particular, drew heavy fire from this important political constituency.

But political concerns don't fully explain the emergence of the broad-based income tax. For while the sales tax certainly had foes, the income tax also had friends. Treasury officials, for instance, recommended several times in the late 1930s that the exemption thresholds for the income tax be lowered substantially. As Treasury Secretary Henry Morgenthau Jr. told an audience in 1937, "We would be applying the principle of capacity to pay more justly if we were to reduce the number of consumer taxes and increase the number of income taxpayers."

Morgenthau's comment echoed the conclusions of economists and lawyers inside Treasury who by the early 1940s had developed a consistent argument for revamping the federal revenue structure. Drawing on a series of major studies, they called for a broadening of the income tax base and a simultaneous reduction in federal excise taxes and tariff duties. Taxing income, they argued, was generally fairer than taxing consumption.

Ultimately, of course, policymakers chose to tax both; faced with wartime strictures, they proved unwilling to part with many tried and true excise levies. Arguments for a broader income tax, however, enjoyed growing support among tax specialists, and many of Treasury's arguments carried the day in postwar tax debates. These same arguments remain important today, for they bear directly on current debates about federal taxation. As today's policymakers consider scrapping or severely restructuring the federal income tax, it is important to know how it was supposed to work.

In this Special Supplement, Tax Notes reprints a key 1937 Treasury study that helps explain how policymakers viewed the income tax during the critical years of its dramatic growth. Prepared by economists Carl Shoup and Roy Blough, both of whom were to figure prominently in Treasury Department tax policymaking during the late 1930s and early 1940s, the study considers a variety of tax reform proposals. It gives special attention to the undistributed profits tax and the treatment of capital gains and losses -- two of the hottest tax issues of the late 1930s. The report, however, also considers a wide range of other issues, including tax rates, exemption levels, income averaging, excise taxes, and integration of the corporate and individual income taxes. Before summarizing the report's conclusions, it may help to explain the political circumstances surrounding its inception and reception.

Lost But Not Forgotten

Buoyed by his landslide reelection in 1936, Franklin Roosevelt began his second term in the White House with an ambitious tax agenda. To begin with, he championed a vigorous investigation of tax evasion and avoidance. Hearkening back to his election rhetoric chastising America's "economic royalists," he raged against the efforts of wealthy Americans to minimize their tax payments. Treasury and congressional officials cooperated on a prominent anti-loophole initiative.

Roosevelt also asked Morgenthau to begin a thorough reevaluation of the federal tax structure with an eye toward improving fairness and efficiency. In response, Treasury officials recruited two prominent tax economists, Shoup and Blough, to serve as consultants. Shoup and Blough, both of whom joined Treasury's tax staff later that year, finished their report late in the summer of 1937. Undersecretary of the Treasury Roswell Magill, in turn, used it as the basis for his recommendations to Congress, forwarded in November of the same year. Roosevelt's anti-loophole campaign was a political success, sailing through Congress on unanimous votes. Treasury's tax reform study, however, was lost amid the political fallout from 1937's resurgent economic woes.

The political fate of Treasury's 1937 study, however, should not obscure its contribution to the Depression-era debate over tax reform -- or its relevance to today's version of the same debate. In particular, the report provides a clear set of standards for evaluating possible tax reforms. While some of the standards were clearly rooted in the problems and intellectual climate of 1937, others remain central in today's debates over tax reform.


"Discussion of standards should be encouraged," the report begins, "otherwise the arguments tend to become entangled with purely technical questions and the basic reasons for disagreement never become clear." While acknowledging that their particular set of standards are neither immutable nor universally accepted, the authors offer them as a starting point for debate. "Once approximate agreement has been reached on standards," they add optimistically, "it should be possible for technicians to agree rather closely on the particular kinds of taxes that will accomplish the ends in view."

The standards fall into six main categories: revenue needs, economic effects, social control, justice, administration, and tax consciousness. Some standards are clear, sharply delineating good taxes from bad. Others, however, serve as gauges, yielding measurements rather than judgments.

Revenue standards set the boundaries of debate. The most basic requirement is that federal taxes yield an adequate amount of total revenue. Individual taxes, however, must also be compared with other revenue standards. How, for instance, will the revenue yield from a particular tax change when the rates are changed? How will the revenue yield respond to variations in business activity or to rapid inflation?

Economic standards cover a variety of issues, most relating to the overall functioning of the economy. How will a given tax affect shifts of capital among industries? What will be its impact on saving, risk-taking, or the incentive to work? Will it encourage or discourage consumer spending?

Social control standards concern the effect of various taxes on specific industries, whether through protective tariffs, excise taxes, or other discriminatory levies. This category also includes the impact of different taxes on forms of business organization, the control of speculation, and the consumption of socially undesirable commodities, such as liquor or tobacco.

Justice standards, according to the report, must play a key role in shaping the federal tax system. "A tax system, to be workable in a democratic country," they assert, "must in general appeal to the sense of fairness of the people." Perhaps the most important justice standard is that a tax take account of personal differences in economic power. Justice standards also touch, however, on benefit charges, taxes on unmerited gain, and equal taxation of competitors.

Administration standards are fairly straightforward, favoring taxes with smaller administrative costs and reduced possibility of evasion. At the same time, however, the authors are quick to point out that such standards must be weighed against other tests, such as justice: "It is better to have an income tax with a certain amount of evasion, for example, than a salt tax with none."

Finally, the authors establish tax consciousness as an important criterion for evaluating a tax, looking kindly on taxes that make their presence known to taxpayers. Indeed, the visibility of the income tax is one of the main reasons the authors ultimately recommend its expansion into lower-income groups.

The body of the report uses these standards to evaluate the federal tax system. In the process, the authors develop a set of proposed reforms. Perhaps the most significant is a reduction in personal exemptions. While arguing that more Americans should pay income taxes, however, the authors are careful to warn that rates for these new taxpayers should be kept low. Under then-current law, they point out, lower-income Americans already labor under a regressive collection of local property taxes and federal tariff duties.

The report also recommends lower rates for upper-income taxpayers. Reductions at the top of the rate scale, they argue, would diminish incentives to avoid and evade. This, in turn, would reduce the administrative burden of the income tax.

The report evaluates but does not recommend the introduction of a general manufacturers sales tax. The burden of such a tax is regressive, the authors point out, exacerbating unequal wealth distribution. Such a levy would also do little to promote tax consciousness. On a more positive note, however, a sales tax would raise considerable revenue and could probably be administered at a reasonable cost. They also note that the tax is less likely than other taxes to decrease the level of saving.

- Joe Thorndike


The authors refer to a report on tax reform titled "Facing the Tax Problem." That report is available from the various Tax Analysts databases as Doc 96-4769 (309 pages).


September 20, 1937

by Roy Blough and Carl Shoup


This report attempts to set standards for a sound federal tax system and to analyze the existing taxes in the light of these standards. Where the existing system seems unsatisfactory, changes -- in the form of either entirely new taxes or merely modifications of existing taxes -- are suggested.

In the course of this attempt, two main approaches were kept in mind:

(1) The general aim was to keep the entire problem of the federal tax system in view, so that nothing of marked importance would be omitted and a proper sense of proportion with respect to the various details would be obtained. Thus, the report opens with a discussion of what standards should be set for a sound federal tax system and indicates how the existing federal system appears, judged by these standards. The second chapter shows what proportions of the total revenue required come from each major source and how these proportions might reasonably be changed in view of the recommendations in this report. The succeeding chapters discuss each tax in turn, with respect to the standards set up in the first chapter. The last chapter recapitulates the recommendations for change.

(2) Another aim was to devote particular attention to the problems raised by the undistributed profits tax and the provisions concerning capital gains and losses. The amount of space given to these problems will not, it is hoped, lead the reader astray with respect to their importance compared with other problems of the tax system. Concentration on these two problems seemed desirable because there is at the moment great pressure from many sources for change. There is a general feeling that the existing provisions are so unsatisfactory that an improvement is possible. The difficulties caused by these provisions are so pressing that action of some sort should not be indefinitely delayed. However, the analysis makes clear, it is hoped, that, for some time to come, intensive study of these issues will yield important discoveries to guide policy.

In order to economize space and time, many of the other problems in the tax system are mentioned only briefly if other sources readily available do as much as or more than the present writers could do here in description and analysis. The sources chiefly referred to are the series of memoranda submitted by the tax consultants engaged by the Treasury in the summer of 1934, and the recently published volume, Facing the Tax Problem.* Specific reference would also have been made often to the series of studies by the Division of Research and Statistics in the summer of 1937, but for the fact that it was necessary for the present writers, who started their work about August 1, to end it by September 15, whereas the division's reports were scheduled for completion September 1. However, most of the material in the present report has been written with the benefit of much of what appears in the division's reports, since the division cooperated extensively in making available preliminary drafts and in invitations to attend discussions by the division's staff members in the course of their work.

The problem of coordinating the federal, state, and local tax systems has not been analyzed in this report, except incidentally in the description and analysis of each tax, and in two paragraphs at the end of the first chapter. This restricted treatment was deliberately planned at the time this study was first outlined, for reasons given in Facing the Tax Problem: paragraph 97, pages 426-27; paragraph C, section XI, page 436; and pages 450-451. Briefly, it appears best to leave existing federal-state tax relations unchanged and at the same time press for the appointment of a commission to study the problem of federal-state-local fiscal (not merely tax) problems. Any piecemeal repair now is likely to make less possible the formulation of a general revision program that will be acceptable to all parties.

Of course the federal government should refrain, so far as possible, from adding to existing conflict. This aim has been kept in mind in the present report (for instance, with respect to the kind of income surtax scale recommended).

An attempt has been made to reach definite conclusions (see Chapter XV), but on the whole the study was undertaken with the expectation that the most important thing was to list, explain, and analyze the major points at issue in each case, and thus to prepare the way for an intelligent decision. Therefore, some details have been included which would have been excluded if one of the requirements for inclusion had been a definite recommendation with respect to each detail. For instance, although an averaging system of some kind seems both practicable and probably essential to a solution of the capital gains and undistributed profits problems (see Chapter III), it is not yet quite evident which of the two or three varieties of averaging explained in Chapter III, is to be preferred. Further immediate study by others is needed, and with such study it should be possible, perhaps within two or three weeks, to reach a satisfactory conclusion.

The three-cent postage charge on letters is not usually considered a tax problem, and has not been discussed in Chapters I to XIV, but in reality it is in part a tax on those who pay the charge, to the extent, apparently, of about one cent, since there is some evidence that the three-cent charge yields about $100 million revenue above costs. As a tax, it is a very poor one except that the revenue is moderately stable and the extra administrative costs are practically nothing at all. It has no support from the principles of tax justice; it is imposed on something that is in part a personal necessity (with a luxury element only occasionally) and in part a business expense. Moreover, it represses tax consciousness in an inadvisable manner -- by deceiving the payer, who surely does not know that it is a tax. With respect to everything except administration, a salt tax would not be much worse, which indicates the low level at which it must be placed. The logical step seems to be to reduce the charge to two cents and raise the rates on other classes of mail (and perhaps charge for some mail services now rendered free of charge), except to the extent that other classes of mail must be subsidized -- which, it is to be hoped, is not extensive. Insofar as a subsidy must be granted, additional revenue can be found in some of the additional sources suggested in Chapter II for the $8 billion total. If this total must be obtained, however, and also some of the mail services subsidized, it might be necessary to delay lowering the three-cent rate.

Another matter to which attention must be called, although time has not permitted an adequate study, is the tax of approximately cent a pound on manufactured (refined) sugar manufactured in the United States, imposed by the Act of September 1, 1937. If there can be assurance that this tax is indeed a tax on monopoly profits of refiners, and is not a burden on consumers, there seems to be no reason for opposing it; but this conclusion appears not to be prima facie evident. Insofar as the tax is a burden on consumers, it is a very poor one, as reference to the standards in Chapter I will readily show. If it burdens the sugar growers or their landlords, it is scarcely a logical tax, to the extent that the purpose of the Act was to improve the profits of sugar growers.

Although virtually all of this report is a work of collaboration, it may be useful, in case questions of detail arise later, to list the general way in which the work was divided and carried out:

I. Standards for a federal tax system (Shoup)

II. Present and proposed sources of federal revenue
(Blough and Shoup)

III. Income taxation

A. Problems restricted to personal
taxation (Shoup)

B. Problems restricted to corporation
taxation (Blough)

C. Problems common to both personal and
corporation taxation (Shoup)

D. Problems of integrating personal and
corporation taxation (Blough)

IV. Business taxation

A. General principles (Blough)

B. Proposals for changing present business
taxes (Blough)

C. Excess profits taxation (Shoup)

V. Manufacturers sales tax (Shoup)

VI. Estate and gift taxes (Blough)

VII. Social security payroll taxes (Blough)

VIII. Customs duties (Blough)

IX. Liquor taxes (Shoup)

X. Tobacco taxes (Shoup)

XI. Automotive excises (Blough)

XII. Other manufacturers excises (Shoup)

XIII. Documentary stamp taxes (Shoup)

XIV. Miscellaneous taxes (Shoup)

XV. Summary of recommendations (Blough and Shoup)


Standards for a Federal Tax System

Necessity for Standards

To construct a sound federal tax system, fairly definite standards must be set up, and the kind of taxes selected that will most nearly meet those standards.

The present chapter sets up standards; the succeeding chapters examine the available taxes in the light of these standards, and make the appropriate selection.

The standards to be set up are not immutable; neither are they universally accepted. As given here, they represent what the present writers believe to be in the public interest at the present time, and, in most cases, would probably receive the general approval of students of taxation. They are, however, the basic source of conflict in tax policy. Once approximate agreement has been reached on standards, it should be possible for technicians to agree rather closely on the particular kinds of taxes that will accomplish the ends in view.

Discussion of standards should be encouraged; otherwise the arguments tend to become entangled with purely technical questions and the basic reasons for disagreement never become clear. Attention becomes unduly centered on one or a few taxes, to the neglect of promising alternatives. Some important standards tend to be overlooked entirely, and the relative weight to be given to different and possibly conflicting standards is scarcely thought of. For example, a discussion of the proper treatment of capital gains and losses is apt to degenerate into an involved but futile argument over technical details before there has been a meeting of minds on such matters of standard as (1) What tests should be used for relative ability to pay? (2) What effect or lack of effect on security markets is desired, and how important is this consideration compared with standard 1? (3) What limits must be set to administrative complexity, and how important is this compared with other standards? -- and so on.

The standards set up in this memorandum are drawn for the most part from the recently published volume, Facing the Tax Problem, but they are restated and a selection is made among them in view of the fact that the federal system alone -- not the federal, state, and local system as a whole -- is the subject of the recommendations herein. Also, the effects of a tax system on savings, risk-taking, and willingness to work are set up somewhat more distinctly as standards than in that volume. Other sets of standards may be found in other books -- for instance, Lutz's "requisites of a sound tax system" -- but in general the various lists cover much the same ground.


1. General Volume of Revenue

The first set of standards to be discussed here are those concerning revenue, and the first standard in this set is the volume of revenue needed over a long period. It appears that the federal government is spending at the rate of somewhere between $75 billion and $100 billion over a decade, though the figure may conceivably fall as low as $50 billion if a period of prosperity is long- sustained. How much of the total should be raised in any given year is a matter discussed below under the standard, degree of fluctuation desirable in annual revenue, but the writers have been requested to put it, for the purposes of the present report, at a minimum of $7 billion or $7.5 billion for the fiscal year 1939 if business recovery continues. No attempt is made here to pass on the wisdom of spending $50 billion, or $75 billion, or $10 billion, or on the debt policy implied by the annual total just stated -- the figures are simply regarded as given data. The existing revenue system will probably yield about $7.3 billion in 1939 (see Chapter II). Hence the immediate problem, in terms of these assumptions, is a readjustment of the tax burden rather than an increase in it.

2. Response of Revenue to Change in Rates

Another revenue test is whether the yield of the tax will change proportionately to a change in rate. Thus, if the cigarette tax rate of six cents a package were increased to seven cents, would the yield increase by one-sixth, or by something less? And if the rate were reduced by one cent, would the yield decrease by one-sixth or by something less? If a tax is so high that it is restricting demand appreciably, the yield will show a change less than proportionate whether the rate is increased or decreased. Is it desirable to have a tax at this stage? Obviously the answer depends on whether the Treasury contemplates an increase in the rate or a decrease. It is pleasing to know that the yield of a tax will not drop proportionately with a cut in the rate -- but, as already implied, the usual accompaniment of this characteristic is that the yield will not rise pari passu with a rate increase. On the whole, since the Treasury is likely to be faced for some time with the problem of tax increases -- not decreases -- any member of the existing tax system that will not increase in yield almost proportionate to a moderate rate increase may be considered undesirable in that respect. The fact that, by this standard, a satisfactory tax is, incidentally, likely to show a decrease in yield almost fully proportionate to a decrease in rate may be ignored, as long as revenue needs continue pressing. The existing system, on the whole, would not respond, in revenue, proportionately to rate changes over the whole system. This fact reflects the high levels to which some of the taxes have been raised. It also reflects the magnitude of the total revenue.

3. Variations in Revenue With Variations in Business

Some taxes show a fairly stable yield through a period of business boom and depression -- through a business cycle -- after eliminating fluctuations due to such things as changes in rates, definition of the base, etc. Others show a highly fluctuating yield. It is probably desirable that the federal revenue system as a whole have a fairly marked degree of cyclical instability. If the tax proceeds decline in depression, the government is forced to issue new money (in effect) by borrowing from the banks. If they rise sharply in prosperity they offer the government a strong inducement to retire the debt accumulated in depression. This statement is much too simplified to do justice to the great complexity of the situation, and in any case the advice of money and banking authorities is essential on this point, but at least there seems little reason for holding rigidly to the doctrine that stability of revenue through boom and depression is an essential characteristic of a sound tax system. The present federal system is highly unstable, but not too much so, especially since state and local revenues have a marked tendency toward stability. However, if the experience of the next three or four years (assuming business recovery continues) shows that Congress will not reduce the debt to somewhat the same extent it did in 192029, then the conclusion above may have to be modified, and ways sought to make the federal system more stable. To do this and yet retain highly progressive taxes probably means that some sort of averaging device would have to be used, whereby the taxpayer would have to pay, in depression, taxes based in part on past prosperity income. The history of such averaging devices is not encouraging. /1/ Perhaps any marked movement toward stability would therefore have to be at the expense of tax justice, by discarding some of the progressive-rate features of the tax system for flat-rate or regressive taxes, largely indirect taxes. Some data on the degree of instability of the present federal system are given in Facing the Tax Problem, pp. 32229, but in general this branch of tax research has been neglected by all tax students.

In view of the marked instability of the income tax and death taxes, this report will favor other taxes only insofar as their instability is moderate, and indeed, it will be well to have some taxes in the system with a very stable yield.

4. Response of Revenue to Rapid Rise in Prices

If there is a likelihood of a sharp increase in the general price level during the next few years (and such appears to be the case), it is desirable to have a revenue system that is particularly responsive to the rise, so that the dollar income of the Government will increase sharply enough so that the real income (command over goods and services which the Government needs) will at least not fall. However, if the taxes are so responsive that the Government's dollar income increases at a greater rate than the general price level, the tax system may prove to be of at least some help in ultimately checking the price rise. The present federal system, chiefly owing to the personal income tax, is moderately well-equipped for such a situation, but some further study is needed at once to prepare the system against loss of real income through taxpayers' delay in payment. Moreover, the system may prove too effective, in that a large increase in revenue will be obtained at the expense of severe encroachments on tax justice. Any given progressive rate scale becomes (for a while, at least) more severe in terms of real income when prices rise, and similarly the relief granted by given personal exemptions shrinks. /2/ Research into methods of correcting quickly any great injustices of this kind could well be instituted now.

Economic Effects

5. Shifts of Capital and Labor Among Industries

Some taxes are heavy enough, considering their general type, to cause a shift of capital and labor from the taxed industry into some other industry -- or, what amounts to much the same thing, less capital and labor is put into the taxed industry than if the industry were not taxed. If this effect is not intentional, and is therefore an incidental, by-product sort of effect, it can almost always be marked down as an undesirable characteristic. Men's ideas (ex-tax) of how to allocate capital and labor may be far removed sometimes from whatever standard for allocation the reader may adhere to, but it is not likely that a random, incidental effect of a tax will better the situation. On the whole, the federal revenue system is only fair, judged by this standard, since the customs duties and some of the heavier excise taxes undoubtedly cause or have caused shifts of capital and labor.

6. Saving, Risk-Taking, and Work

Some taxes probably have a much greater restrictive effect than others on such factors of production as volume of saving, willingness to take risks, and willingness to work. It is often possible to say, with a reasonable degree of assurance, that tax A exerts a greater restrictive effect than tax B. However, the degree of difference seems impossible to ascertain within a wide range. Moreover, it is often difficult, perhaps impossible, to decide whether a given degree of restriction is desirable (at least as to the first two factors noted above). In consequence, the present writers are inclined to give no opinion in many cases, either as to what the situation is with respect to a particular tax, or (assuming that the situation is known) whether the tax is desirable in this characteristic. It is recognized, however, that others may reasonably feel that the attitude above represents an excess of caution and an understatement of the amount of knowledge available, and it is chiefly for their convenience that these possible economic effects have been listed.

SAVING. The volume of savings in the country -- or, in other words, the accumulation of capital -- is affected by the tax system. For example, high income tax rates in the high brackets probably result in a smaller total of savings being accumulated than if the revenue were instead raised by a general sales tax. It is usually possible to say whether a given tax will result in a larger or smaller volume of savings than another given tax -- but, as indicated in the preceding paragraph, it seems impossible, in the present state of knowledge, to say how much larger or how much smaller the volume will be. The existing federal system probably on balance tends to lower rather than to increase domestic savings, compared with the situation if the revenue were coming, say, from some foreign source, but this comparison is so unreal as to make the conclusion of very restricted value. Moreover, there is no general agreement among economists on how large the volume should be. Indeed, there is not even agreement on whether the volume should be (indeterminately) larger or (indeterminately) smaller than it now is. In view of these uncertainties, the effect on the volume of savings is not much more than a hypothetical standard for judging any given tax. However, the present writers are inclined to lean toward a tax that at least does not restrict the volume of savings to the present level.

RISK-TAKING. The effect on the willingness to risk is another standard, much like the one discussed in the immediately preceding paragraph -- that is, although it is usually possible to say whether one tax encourages or represses this willingness compared with another tax, it is not possible to say how great the difference is, and there is no agreement on whether -- and how much -- that willingness should be repressed or encouraged. If it is repressed, the volume of savings may decline somewhat (to this extent this question is the same as the one in the paragraph above), and, more important, the savings that are accumulated and invested will go instead into safer lines, lowering interest and profit margins there, and thus tending to lower prices of consumers' goods resulting from those lines. Correspondingly, the development of entirely new products or of known products that involve risk-taking is checked. The present writers suspect that risk-taking has in many cases been carried much too far, as concerns the welfare of society as a whole, and are inclined to see no cause for the alarm in measures that reduce risk-taking compared with what it was in the decade of the 1920s. Admittedly, however, views on this subject must be vague. The present federal system definitely exerts a repressive effect on risk- taking, particularly through some of the details of the income tax.

WORK. The effect on the willingness to work is another standard by which to test a tax. Here there is general agreement that the tax system should encourage and not repress. The actual offset in quantitative terms is, however, unknown. Even the tendency of one tax compared with another is perhaps more difficult to ascertain than in the case of savings or willingness to risk.

Regrettably, then, three of the most important standards for a tax -- those described immediately above -- must be used sparingly if at all in deciding what taxes to include in a system, simply because of lack of knowledge. Possibly some time from now there will be somewhat more foundation to build upon, but it would only be deceiving oneself to base judgments, now, primarily on the presumed effects of taxes on these three economic factors.

7. Consumers' Spending

Some taxes tend to restrict the immediate volume of consumers' spending (the long-run tendency may be in the other direction). Subject to some qualifications, it may be said that this characteristic is the obverse of the restriction-on-volume-of-savings noted above. Again, differences between taxes may be found, but the quantitative aspects and the desirability of these aspects are usually unknown.

8. Social Control

Some taxes are much better suited for economic and social control than other taxes. Whether any given control -- for example, control of spirits consumption, or control of beer consumption -- is desirable, is of course a problem that far transcends tax matters. Nevertheless, some answer has to be given to such problems, if one is to say whether the existing spirits tax or beer tax is a good tax. The paragraphs below will note the chief kinds of control relevant in a discussion of the federal tax system, and the writers' attitude toward these controls will be stated or implied. For a general statement of the problem of control through taxation, see Facing the Tax Problem, pages 129140.

PROTECTIVE TARIFF. The protective tariff encourages certain domestic business firms at the expense of (a) certain other domestic business firms, (b) domestic consumers in general, and (c) certain foreign business firms. Its proponents have often argued that it guarantees the workingman a higher standard of living (on the grounds that he is protected from cheap labor abroad) and increases the productive capacity and the welfare of the country as a whole. Practically all economists are agreed, however, that the results of the tariff are the opposite of these claims. For a resume of the reasoning that leads to this conclusion, See Facing the Tax Problem, pp. 14243. In its essence, the reasoning is that any measure that checks the specialization in those lines of production for which the country is naturally best suited tends to lower the country's productive power and hence its welfare. Such a lowering may be necessary in some countries that need to develop certain industries useful in case of war, but this reasoning appears not to apply to the United States, or in but a minor degree. Hence this method of control, so intensively developed in the existing federal system, is decidedly undesirable. Some difficult problems of tax justice arise, however, in a transition period of lowering tariff duties; these are discussed in paragraphs below.

PROMOTION BY INTERNAL TAXES. The promotion of one domestic business almost wholly at the expense of another domestic business (that is, without directly affecting foreign business firms), is a form of tax control exemplified in the federal system by the 10 percent tax on state bank notes and the tax on oleomargarine. Such promotion and repression is about as poor a standard for a tax system as that discussed in the immediately preceding paragraph, and for the same reason. A strong case on special grounds must be made out if any such control measure is to be approved. A strong case does exist with respect to state bank notes, but certainly not with respect to oleomargarine.

SIZE OF BUSINESS UNITS. The encouragement or repression (usually the latter) of large business units is another standard of current interest. No doubt the tax system can be framed so as to discourage materially the growth of large business units. If such units are to be discouraged, however, the tax system should be a supplementary method of control rather than the main instrument. Moreover, despite the undesirable features of bigness in business, the evidence for broad restrictions on size of business units is not yet convincing -- and it is only broad restrictions that are likely to be feasible if the tax system is the instrument of control. Similarly, no reason is evident why the federal tax system should be weighted against small units. Fortunately, the existing federal tax system is not seriously weighted in either direction.

CORPORATION VS. unincorporated concerns. The encouragement or repression of the corporate form of doing business, compared with the unincorporated forms, is another possible control standard for the tax system, but there is no decisive evidence that any more power of control is needed than is already available in the laws governing incorporation, use of the partnership device, etc. The tax system is an even cruder method of control here than in size of business. Of course it is possible to argue for a special tax on corporations to recoup some of the benefits the corporate owners get from the privilege of incorporation, but this is a somewhat different matter, and is treated in the paragraphs below under benefit taxation. The existing federal system, including the effect of the undistributed profits tax, is, on balance, definitely weighted against the corporate form to a much greater extent than seems justifiable on the privilege basis (if indeed this basis has any justification at all). The measures to be taken to eliminate this discrimination are discussed in detail in Chapters III and IV below. Special taxes on special forms of corporation -- for instance, the surtax on personal holding companies -- do not raise the general question of corporations versus unincorporated enterprises.

DIRECTORS VS. stockholders. A shift of control over corporate affairs away from directors and toward stockholders can be accomplished by the federal tax system -- at least with respect to control over the use of the corporation's current earnings. The present tax on undistributed profits causes a shift of this nature. The shift could be made greater by certain changes in the tax -- e.g., allowing only cash dividends to be used in the dividends-paid credit. Again, however, there is not yet sufficient evidence to say that the tax system must be called upon to supplement the laws governing corporation organization and operation with respect to allocation of control between directors and stockholders. This standard, therefore, although not definitely rejected, is not accepted in the present report as a basis for formulating a federal tax program.

ASSET DISTRIBUTION AND BUSINESS CYCLE. The conclusion in the preceding sentence applies likewise to the use of the tax system to force earnings out to stockholders (in such a way that corporate assets are decreased -- that is, not merely by means of a taxable dividend in stock) in order to increase the amount spent on consumption goods and decrease both the amount of (money) saving and the amount spent on producers goods. This standard differs from that in the preceding paragraph in that it does not rest on the question of who should decide what should be done with the corporation's earnings (if stockholders generally showed an inclination to reinvest all the extra dividends, the standard in the preceding paragraph would still be satisfied, but that in this paragraph would not). However attractive this objective may seem superficially, as a means of dampening business cycles, the fact remains that economists generally do not support the kind of business cycle explanation on which this plan rests. Conversely, there seems no adequate reason for shaping a tax system deliberately to decrease the percentage of the earnings paid out in dividends. The federal income tax system prior to the Revenue Act of 1936 apparently had some influence in this direction -- largely unintended by the law makers, it may be presumed. /3/ Note that the problem in this paragraph is the same one touched on above when discussing taxes that affect saving or consumers' spending.

HOLDING COMPANIES. Holding companies can be largely put out of existence by certain adjustments in the income tax, chiefly by including dividends received in corporate net taxable income. If a less broad, more refined attack on the holding company problem is desired, however, the tax system will probably not be found very helpful as an instrument of control. The existing federal system has a slight tendency toward repressing holding companies in that it allows only 85 percent of intercorporate dividends to be deducted.

SPECULATION. Control of speculation, particularly on security and commodity exchanges, can be achieved by transfer taxes. The repression of ill-informed, reckless speculation without discouragement of carefully planned capital undertakings is a highly desirable aim, and upon further study the transfer tax may be found to have more potential usefulness in this connection than has been supposed. The only federal tax that shows a distinct regulatory purpose of this kind is the prohibitive cotton futures tax levied on transfers not made under certain kinds of contract. The taxes on the transfer of stocks, bonds (and possibly on real estate, though the rate is very low) doubtless have some slight repressive effect, and the treatment of capital gains and losses also bears on this point.

LABOR REGULATIONS. In the control of labor regulations the federal tax system has been of little use, and does not give much promise of development (unless the crediting provisions of the unemployment compensation tax, designed to insure state action in this field, is included here). The only federal tax of importance in this way is the prohibitive tax on the manufacture of white phosphorus matches.

SUMPTUARY TAXES. In order to restrict the consumption of harmful commodities, a tax may sometimes be useful, notably with liquor, especially spirits. Indeed, it is probably very important that the federal spirits tax be regarded primarily as an instrument of control rather than a source of revenue, though it can of course still serve the latter purpose to some extent. At its present level, however, the federal spirits tax is probably not a very effective measure of control, even when state spirits taxes are considered; and aside from the spirits tax, the federal system does not attempt to repress the consumption of harmful commodities. The narcotics tax might be considered an example, but here the tax is only incidental to much more stringent methods of control.


`A tax system, to be workable in a democratic country, must in
general appeal to the sense of fairness of the people. Of course no
tax system can appeal to everyone as being eminently fair, and it is
certain that everyone will have some complaint against any system.
However, ideas of fairness in taxation are usually nebulous.
Nevertheless, it is unthinkable that a tax system could be designed
and operated in this country at the present time without regard to
the tests of justice.

9. Taking Account of Personal Differences in Economic Power

'To raise the revenue that cannot be supplied by benefit taxes and
taxes on "excess" (to be explained below), use should be made, so far
as possible, of taxes that can take account in many ways of personal
differences in economic power to support the government. The personal
income tax, for instance, can do this, but a cigarette tax can take
account of these differences only in a very crude and incomplete
manner. The outstanding taxes from this point of view are those on
personal income (and corporate income, so far as the tax is
integrated with the personal income tax -- see Chapter III below),
estates, and gifts. The federal government probably does not even now
make as much use of these three taxes, compared with other, cruder
taxes, as it might. In 1937 it obtained $2,455 million from these
four taxes -- 48 percent of total federal tax revenue. In 1938 the
percentage will be somewhat higher. But since the revenue from
benefit taxes and "excess" taxes is small (gasoline tax, $197
million; excess profits tax, $25 million; unjust enrichment tax, $6
million; old-age tax $207 million -- total $435 million in 1937),
these figures mean that well over one-third the federal tax revenue
is being raised by crude taxes not readily adjustable to differences
in personal economic status. Moreover, as will be indicated in later
chapters, many of the details of the more refined taxes are in
themselves crude, and thus fail to take account of relevant
differences in personal economic status. Notable examples are the
restrictions on deductions of capital losses and the isolation of one
annual accounting period from preceding or succeeding ones, in the
income tax.

`PROGRESSION AND PRESENT BURDEN. One reason for desiring taxes that
can be adjusted in complex ways to personal differences is that
progressive graduation of a personal tax is generally accepted as
just. However, no generally accepted measure for the severity of the
progression that should be imposed can be found. The federal tax
system as a whole is probably progressive in its burden at all points
along the income scale above $2,000. It is certainly steeply
progressive at income levels above $5,000.

Professor Newcomer's figures in Chapter 16 of Facing the Tax Problem (see especially table 24) indicate that a New York farmer with an income of $500 pays about $7 in federal taxes, direct and indirect, while one with $1,000 income pays about $14, and one with $2,000 income, about $25. Thus the federal tax burden as a percentage of income is probably either proportional or slightly regressive (the opposite of progressive) in these low levels. The burden is so light -- less than 2 percent of income -- that the matter would not be one of grave concern, were it not that the far heavier state and local taxes at these levels also fail to show a progressive burden.

The $500 farmer apparently pays between $60 and $100 a year in total federal, state, and local taxes in New York and Illinois. The $1,000 farmer apparently pays between $100 and $160, while the $1,000 wage-earner apparently pays between $150 and $190.

The lesson seems to be that the federal government should aim to make its own system markedly progressive at these levels, and at least avoid any new tax measures that would accentuate the existing tendency. A word of caution must be added, however; Professor Newcomer's figures were developed chiefly as a means of furthering the technique of investigation in this difficult and little-explored field of burden distribution, and a high degree of accuracy is not claimed for them.

The $5,000 man in New York apparently pays about one-fifth of his income in total federal, state, and local taxes, if he is a salaried worker and about one-fourth if he is a merchant. At the $20,000 level the fraction becomes nearly one-third (salaried worker). For a corporation official with $100,000 income, it is not far from one-half, and at the $1 million level it is over four- fifths. /4/ Thus, even with the proportional or regressive effect of state plus local taxes, the net result is a decidedly progressive burden, owing to the sharply progressive nature of the federal tax system at levels above $5,000. These figures are no proof that the federal system is progressive enough -- or too progressive -- but they give some indication of the degree of progression. /5/

TAXES THAT TAKE LITTLE OR NO ACCOUNT OF PERSONAL DIFFERENCES. The rest of the taxes in the federal system -- the customs duties, the liquor and tobacco taxes, the manufacturers' excises (except the gasoline tax), the documentary stamp taxes, and the "miscellaneous" group, and possibly also the capital stock tax -- can be supported on the basis of tax justice only as undesirable but possibly necessary alternatives. As a matter of fact a large part of the revenue from these taxes represents a burden on people who either are or could be reached by an income tax. This is especially true of the excises, stamp taxes, and miscellaneous taxes. The distribution of the burden represented by the revenue received from the tobacco and liquor taxes seems pretty capricious. (Have the abstemious and the nonsmokers some special claim to partial relief from carrying the burden of government?) The burden of the customs duties is more than the revenue collected: in large part it is higher prices charged by protected domestic concerns. From the point of view of tax justice alone, little, if any, reason can be found for utilizing the taxes discussed in this paragraph. It may be necessary to use them because of administrative considerations, or regulatory considerations, but here it becomes a matter of weighing different and somewhat conflicting standards. The remarks in this paragraph apply also to a general manufacturers sales tax.

10. Benefit Charges

"Persons who directly use and benefit from certain governmental services should pay a benefit charge, according to a generally accepted standard of tax justice. A benefit charge directly burdens only the beneficiary:.... Some [benefit charges] require the direct beneficiaries as a group to cover the entire cost of the service; others require them to cover only part. The cost can be divided among the individual beneficiaries in any one of several ways." /6/

The federal government has relatively little chance to levy benefit charges; most of its services are too generalized, not localized enough. The present postal rates are benefit charges, except for that part of the first-class rate that is above the cost of handling first-class mail, which really represents a crude form of tax. The federal gasoline tax can be supported in large part on a rough benefit basis, in view of federal aid to state highways. The other federal automobile taxes, with the partial exception of the tax on tires and tubes, are technically deficient as benefit taxes. Possibly the old-age payroll taxes can be regarded as benefit taxes in part. All the other federal taxes have to be justified on some basis other than direct benefit. Incidentally, a tax system is not necessarily sounder the more use it makes of benefit charges. Nobody advocates, for example, putting public education on a benefit-charge basis. The extent to which benefit charging should be used -- granted it is technically feasible up to a certain extent -- is one of the most controversial problems of tax justice. In general, it is well not to push very hard for an extension of benefit taxation; such taxation implies that existing differences in wealth, income, and opportunity should not be disturbed by the tax-expenditure system. The present federal system probably has about all the benefit charges that it should have in it.

11. Taxes on Unmerited Gain

Gains that "violate prevailing standards of economic morality" /7/ are, almost by definition, reasonably subject to special taxes, regardless of who obtains the gains -- the only qualification being the matter of innocent vested interests, which is discussed a few pages below. Minor examples of this kind of tax are found in the federal system -- the unjust enrichment tax, and the 100 percent "tax" in the Vinson Act on excessive profits of builders of naval vessels and aircraft. There are strong reasons, from this aspect of tax justice, for introducing, gradually, a true excess profits tax.

12. Equal Taxation of Competitors

Another tax standard from the point of view of justice is the avoidance of taxing business competitors unequally. Of course no tax system is likely to treat all competitors alike, especially if "competitor" is taken in its broadest sense, so that cigar manufacturers may be said to compete with sporting goods manufacturers, for instance, for the consumer's dollar. Reference here is instead to unequal treatment of businessmen in the same or closely allied lines of business, or to business in general if the tax is that broad. For example, the payroll tax for unemployment insurance gives unequal treatment to competitors in that it requires those with eight employees to pay a tax while those with seven employees pay no tax. On the whole, the federal system does not tax nearby competitors unequally, but some improvement may be possible.

13. Unrecorded Loss of Consumers' Satisfaction

When the price of a commodity is raised substantially by a tax, some persons will no longer buy it. They suffer an unrecorded burden in the loss of satisfaction they would otherwise get from the use of the article, compared with the lesser satisfaction they have to accept by turning to some other article. Similarly, some consumers, while not ceasing entirely their purchases of the taxed article, buy less of it than before. If they buy so much less that they have money left over even after paying the price plus tax of the portion they still consume, the inference must be that they too are being forced by the high price to turn to untaxed substitutes and to accept a diminution in satisfaction -- a diminution that is not represented by dollars of tax collected. Any substantial diminution of this kind is to be counted against a tax. A heavy tax on cigarettes, for example, may have this fault, whereas an income tax has not. Income taxpayers must restrict their consumption to the extent that they turn money over to the government -- but the same is true of cigarette smokers, and in addition the smokers bear the burden of loss of satisfaction of the cigarettes they do not smoke (or, it might be said, the tax they do not pay). The federal tax system is not grievously at fault with respect to unrecorded burdens of this kind, but there may be an appreciable amount under some of the taxes, notably the customs duties and the cigarette and spirits taxes.

14. Vested Interests

One standard for a good tax system is that it does not, through too frequent changes or the wrong kind of changes, greatly disturb many innocent vested interests. For the existing federal system, this standard is of chief importance with respect to the following possible changes: (1) the introduction of a genuine excess profits tax, (2) the reduction or removal of customs duties, and (3) the imposition of heavy excise taxes on special commodities where the demand might decline so much that the producer would have to shoulder much of the burden of the tax himself. It is of some importance, in special instances, when changes in the income tax burden are contemplated. /8/ Sometimes the damage is lessened by making the change gradually.

Any change in a tax system is apt to upset established economic relations and result in considerable economic loss to people who have made commitments on the assumption (implicit, at least) that no such change would occur. For instance, an investor may buy into a company that is making excessive profits, but buy in on the basis of a normal yield of, say, 5 percent. Thus, he might pay $300 for a share, with a book value of $100, earning $15 a year. If a heavy excess profits tax is now unexpectedly levied on the concern, the remaining profits may give him no more than 1 percent or 2 percent yield on his purchase price. He has had a vested interest -- an "innocent" vested interest -- that has been harmed by the change in the system.

15. Bonus by Change

The obverse of the vested interest problem is that of the bonus that may be given to a taxpayer by the lowering of the tax. The bonus is an increase in value of the taxpayer's property, over and above the immediate yearly relief (if any) intended for him by the tax change. For example, if the corporation normal income tax is suddenly abolished, the existing owners of corporate common stock may get, in the current year, a relief equal, not merely to the current year's tax, but to a capitalized series of the tax for several years into the future -- which, it may be assumed, would be more than the legislation would intend them to get, and so to that extent would be a bonus.

Neither the vested interest nor the bonus problem arises unless some change is contemplated in the tax system.


16. Cost of Administration and Compliance, and Degree of Evasion

Any tax system will cost something to administer and to comply with and will at the same time permit some evasion. The problem then is to hold both phenomena within reasonable limits. One criterion of reasonableness is degree of compatibility with other standards discussed in this chapter. It is better to have an income tax with a certain amount of evasion, for example, than a salt tax with none. Considering all things, the present federal system is working reasonably well from an administrative point of view, with a possible exception in the spirits tax. The system can, however, and should be made still better, at least with respect to amount of evasion, and cost and inconvenience to the taxpayer, especially in the income tax. /9/ Also, it is desirable to try to find out what the cost of administration of each of the major internal-revenue taxes is. The only figure available at present is an over-all one.

It is also desirable to have a system that does not cause the taxpayer too much expense (apart from the payment itself) -- that does not, in other words, have a very high cost of compliance.

Tax Consciousness

17. Reasons and Opportunities for Tax-Consciousness

To instill a considerable amount of tax-consciousness into those who really bear the tax burden is desirable. It gives those persons a better chance to allocate their total spending power between governmental and non-governmental activities in the proportion that they desire. Taxes like the income and death taxes can do this more than shifted taxes like manufacturers' excises can. Even some of the latter can be made to help toward this objective, however. Thus the amount of the cigarette tax might be noted on the tax stamp that is affixed to each package. The big opportunity at present, however, is the possibility of lowering exemptions under the federal personal income tax. It appears that if the exemptions were lowered from $1,000, $2,500, and $400 to $800, $1,500 and $300, respectively, the total number of returns would apparently increase from 3.7 million to 7 million or 8 million even in a year like 1933, and taxable returns, from 1.3 million to 3.7 million. /10/

Federal-State-Local Coordination

18. Tax Problem as Part of a Fiscal Problem

A good tax system in a federalized state must have the federal taxes so coordinated with the tax systems of the underlying jurisdictions that the desired degree of decentralization and local independence is maintained at the minimum of administrative and compliance cost, and also without too much departure from the other standards discussed in this chapter. The present federal system does not have much to boast of in this connection, though it does take account of the state and local situations by the crediting devices under the estate and unemployment compensation taxes, by the allowance of state income taxes and state and local property taxes as deductions in computing net income subject to federal tax, and by using a gradual rather than a steep progression in income tax rates for the first $10,000 or $15,000 of income. Unfortunately, the problem is much more than a tax problem. It is also an expenditure and debt problem. For reasons given elsewhere, /11/ it seems unwise to push farther with plans for federal-state-local tax coordination alone. Rather, an immediate attack should be made on the whole problem of fiscal coordination.

Meanwhile, each tax may be examined as to whether the present use of it by the federal government (a) avoids conflicts of governments, and (b) avoids repressing decentralization -- that is, tends to allow the states and localities some degree of fiscal independence and responsibility. By these standards, the customs duties are satisfactory, while the gasoline tax is not, for example.

Special Note for Mr. Magill

The revenue estimates for 1939 are based on the confidential figures contained in Vol. I of the Tax Revision Studies, 1937, of the Division of Research and Statistics, except that for the smaller taxes not shown in detail in those figures, the writers have made rough estimates of their own.


Sources of Federal Revenue

I. For a Hypothetical Full Year 1939:
Total, $7.5 Billion

1. Recommended Amounts of Revenue

If $7.5 billion in revenue must be raised by the federal government in a year such as 1939 is estimated /1/ to be, it is recommended that the following amount be obtained from the following sources:

                               Table 1
                   Recommended Amounts of Revenue,
                     Hypothetical Full Year 1939
                                                  (In Billions
          Tax                                      of Dollars)
Personal income                                        2.93

Corporations: normal tax      .60
              privilege tax   .15                       .75
Estate                                                 1.08

Gift                                                    .01

Stamp                                                   .05

Manufacturers' excise (wholly gasoline)                 .21

Liquor                                                  .67

Tobacco                                                 .32

Other internal revenue: payroll taxes /a/               .80

Customs                                                 .55
Total taxes                                            7.37

Miscellaneous receipts                                  .21
Total revenue                                          7.58

     /a/  The source from which this estimate is taken (Tax Revision
Studies, Vol. I, table on p. 17) does not specify that all of this is
assumed to be from payroll taxes, but there seems little doubt that
all or practically all of this $800,000,000 is from these taxes.

For the actual year 1939 the pattern will have to be somewhat different, because the changes in the estate tax and the income taxes cannot be made soon enough to be fully effective in 1939. Hence the table above may be taken to represent a "hypothetical 1939" /2/ in which it is assumed all recommended tax changes will be fully effective.

Before considering what the actual 1939 pattern might be, it is best to explain briefly what the figures above on the hypothetical 1939 imply as to changes in the existing system. No detailed arguments for the changes will be given at this point; they are contained in later chapters of this report.

2. Comparison With Existing System

First, the figures above may be compared with what the existing system would yield in a "hypothetical 1939," as estimated in Vol. I (p. 17) of Tax Revision Studies, 1937. The two systems, as the following table shows, would produce about the same amount of revenue, but the proposed system relies much more heavily on the personal income tax (divided by $1.04 billion) and the estate tax (divided by $0.68 billion) and much less on corporation taxes ($0.79 billion) and the stamp, sundry, and excise taxes, including the tobacco tax ($0.66 billion).

                               Table 2
   Amounts of Revenue Under Recommended System and Existing System
                     Hypothetical Full Year 1939
                                         (In Billions of Dollars)
           Tax                          System     Existing System
Personal income                           2.93          1.89 /a/
Corporation normal                         .60          1.28 /a/
Corporation privilege                      .15          0
Corporation undistributed profits         0              .08
Excess profits                            0              .03
Capital stock                             0              .15
Estate                                    1.08           .40
Gift                                       .01           .03
Stamp                                      .05           .08
Manufacturers' excise                      .21           .47
Sundry                                    0              .13 /b/
Liquor                                     .67           .67
Tobacco                                    .32           .56
Other internal revenue:
  payroll taxes /c/                        .80           .80
Customs /d/                                .55           .55
                                          -----         -----
Total taxes                               7.37          7.12
Miscellaneous receipts                     .21           .21
                                          -----         -----
Total revenue                             7.58          7.33
     /a/ Back taxes split $160 million to corporation normal tax and
$100 million to personal income tax.

     /b/ Does not include tax on coconut, etc., oils processed.

     /c/ See footnote a to Table 1.

     /d/ Includes tax on coconut, etc., oils processed.

3. Comparison With Division's Recommendations

Finally, the system for a hypothetical 1939 recommended by the present writers may be compared with the system for the same year recommended by the Division of Research and Statistics. The comparison has limited value, because the Division's system produces a total of $9.05 billion compared with $7.58 billion. The present writers' system would depend less on corporation taxes ($0.66 billion), and stamp, sundry, and excise taxes, including tobacco taxes ($0.73 billion), and slightly less on personal income taxes ($.08 billion). However, the Division says (p. 12, first section of Vol. I) that if less revenue than the $9.05 billions is needed, it recommends (a) eliminating more of the stamp, manufacturers' excise (not including tobacco), and sundry taxes, (b) a lowering of the rates of the privilege tax, (c) certain adjustments in the income tax, and (d) reduction of the rates of the gasoline tax and some other excises. Except for the reduction of the gasoline tax rate, these steps are in the same direction taken by the present writers, and it is therefore likely that there is not much disagreement as to how much each tax should be called upon to raise in a hypothetical 1939 full fiscal year if the total were to be $7.5 billion.

4. Implications of Recommendations

The implications of the present writers' program, in Table 1 above, with a few brief supporting arguments, will now be presented, tax by tax.

                               Table 3
             Amounts of Revenue Under System Recommended
             in Present Report and System Recommended by
           Research Division, Hypothetical Full Year 1939
                                         (In Billions of Dollars)
                                        System          System
                                     Recommended      Recommended
                                      in Present      by Research
          Tax                           Report          Division
Personal income                           2.93           3.01 /a/
Corporation normal                         .60           0
Corporation privilege                      .15           1.36
Corporation undistributed profits         0               .05
Estate                                    1.08           1.08
Gift                                       .01            .01
Stamp                                      .05            .15
Manufacturers' excise                      .21            .47
Sundry                                    0               .13 /b/
Liquor                                     .67            .67
Tobacco                                    .32            .56
Other internal revenue:
  payroll taxes                            .80            .80
Customs /c/                                .55            .55
                                         ------          ------
Total taxes                               7.37           8.84
Miscellaneous receipts                     .21            .21
Total revenue                             7.58           9.05
     /a/ The present writers have assigned all back taxes to personal
income tax, none to undistributed profits tax.

     /b/ Excluding tax on coconut, etc., oils processed.

     /c/ Including tax on coconut, etc., oils processed.

5. Personal Income Tax

The estimate of $2.93 billion in Table 1 was reached by taking the estimate made in the table on p. 17 of Vol. I of Tax Revision Studies, 1937, and subtracting an estimated $80,000,000 loss to come from setting the maximum surtax plus normal tax bracket rate at 67 percent instead of 79 percent (to lessen administrative difficulties and undesirable economic effects). This method of computation may seem to imply approval of all the personal income tax features, including the rate scale (aside from the cut in the top rates) proposed in the Studies. Actually, it does not. Instead, the use of the figure $2.93 billion implies merely this: That the present writers favor, in general, the kind of surtax rate scale proposed in the Studies (except for the top rates), and agree that the personal income tax can be called upon to produce, in a hypothetical full year 1939, about $3 billion. In matters of detail, however, the personal income tax that the writers favor would differ appreciably from the tax that forms the basis of the Studies estimate. The treatment of capital gains and losses, the use of an averaging device, the procedure with respect to undistributed profits, the size of the earned income credit, and the lumping of spouses' incomes for tax determination are all examples of matters where the Studies and the present report do not agree entirely. In some of these cases, adoption of the present writers' proposals would call for a higher rate scale to raise the same amount of revenue as the Studies' plan would; in other cases, a lower scale. Time and resources have not permitted the writers to make estimates of the revenue effect of each of these differences, but it appears that the net loss would not be so great that it could not in fact be made up by increasing rates. To the present writers, it seems most important to make the income tax a tax that will appeal to the community's sense of justice, even at the cost of having a higher rate scale. General observation suggests that it is reasonable to expect to be able to raise $3 billion out of a total $7 billion to $9 billion by a personal income tax that is equitable -- especially when it is considered that the total federal, state, and local tax revenue will be $15 billion to $18 billion, /3/ with less than half a billion from state personal income taxes.

6. Corporation Income Tax

The privilege tax estimate is based on a rate of between 1 and 2 percent, with the same kind of base as recommended in the Studies, except that (a) interest and dividends received from other corporations are not included in taxable income, since deductions for those items are not allowed to the paying corporation, and (b) rents paid and, subject to some reservation, royalties paid, are deducted, largely because of the injustice of not allowing the depreciation element in rents as a deduction. The Studies' volumes apparently do not give the rate of privilege tax necessary to raise the $1.4 billion noted on p. 17 of the first section of Vol. I, but it would probably be above 10 percent, even on a base that includes dividends, interest, rents, and royalties from other corporations. This, it is submitted, is much too high even for a permanent tax, since the tax is impersonal and no relief is granted to the stockholder through the personal income tax when he gets his dividends. A rate of 5 percent should be the maximum, and for the transition period it should be much lower, for fear of the sudden new burden on corporations heavily in debt and (as concerns the plan submitted in the Studies), renting much of their property (e.g., certain railroads). Meanwhile, part of the normal tax may be kept -- say at a 7 percent rate to yield about $0.60 billion. Year by year the privilege tax rate would be increased and the normal rate decreased until the former got to 5 percent as a maximum (lower, if revenue considerations permit) and the latter disappeared, the total revenue from corporation taxation then being about $0.50 billion, instead of the $0.75 billion shown in Table 1 -- but if at that time the extra $0.25 billion was still needed, it could be obtained by retaining some of the taxes scheduled below for abolition. It should be emphasized that the estimates of yield in this paragraph to come from given rates are highly uncertain, and it is desirable to get some data on this point as soon as possible.

7. Estate Tax and Gift Tax

Here the estimate is simply a copy of the Studies' estimate. It is here assumed, however, that the amount given is net, after any payments to be made to states according to the Division's recommendation. In any case, the rate schedule and in general the other features on which this estimate is based, are favored by the present writers.

8. Excise and Sundry Stamp Taxes

The figure of $0.05 billion for the stamp taxes represents the abolition of all existing stamp taxes except those on stock transfer and bond transfer and the very light tax on real estate transfer -- and does not, as the Studies' estimate apparently does, represent any increase in the existing stock transfer tax. The manufactures' excise figure of $0.21 billion represents the abolition of all the manufacturers' excises except the one cent gasoline tax. The sundry figure of zero represents the abolition of all the sundry taxes, notably the tax on telephone, etc., messages, and the admissions tax.

9. Liquor and Tobacco Taxes

The liquor tax figure of $0.67 billion represents the existing law, unchanged.

The tobacco tax figure of $0.32 billion represents a lowering of the cigarette tax rate from six cents a package to three cents a package, which still leaves this tax at a higher rate, compared to retail value, than the taxes on other tobacco products (unchanged).

The payroll tax figure is unchanged because the rate of the chief tax, the old-age pensions tax, will still be only 1 percent on employees and 1 percent on employers in 1939. Whether it should ever go higher, especially to 3 percent and 3 percent, respectively, is a complex problem that does not need to be settled this year.

As to customs, the present writers favor reducing the existing rates sharply, and, ultimately, coming as close to free trade as national defense conditions permit. This program would probably, for several years, increase rather than decrease revenues from customs, but to be conservative the Studies' estimate (which assumes no change) is adopted.

II. For the Actual Year 1939: Total, $7.5 Billion

The changes recommended by the present writers for the income tax and the estate and gift taxes cannot be fully effective for the fiscal year 1939. These recommendations call for a constitutional amendment -- but the delay caused by this would not affect revenues very adversely if at all, since the purpose of the amendment would be not to raise more revenue, but to get a system that is more equitable and that has fewer undesirable economic repercussions (particularly as to capital gains and losses and undistributed profits). Instead, the delay comes chiefly from the delayed payment systems under the income tax and the estate tax. The results of the actual 1939 would be perhaps about $0.60 billion short of the hypothetical 1939 for the personal income tax and perhaps $0.50 billion short for the estate tax -- judging from the tables on pp. 15 and 17 of Vol. I of the Studies (the corporation tax yield would not be much affected by the difference between the actual 1939 and a hypothetical 1939, under the system proposed in the present report). To obtain this $1.1 billion until the transition has been accomplished, the program implied in Table 1 above should be modified as shown in Table 4.

                               Table 4
            Sources of Additional Revenue for Actual 1939
                     (Add to Hypothetical 1939),
                 Total Approximately $7,500,000,000
                                                        in billions
                                                         of dollars
Delay the decrease of 3 cents in the cigarette tax           .24

Delay repeal of the taxes on passenger
automobiles and tires and tubes                              .10

Delay repeal of the tax on toilet preparations               .02

Delay repeal of the tax on admissions                        .02

Delay reduction in corporation normal
tax rates                                                    .60

Count on gift tax payments as under present system           .02

Payments still due on repealed capital stock and
excess profits tax                                           .10
Total                                                       1.10

III. For a Full Year 1939: Total, $8 Billion

If a total revenue of $8 billion is needed in a hypothetical full year 1939 -- that is, $0.42 billion more than is provided in Table 1 at the beginning of this chapter -- it is recommended that the decrease of three cents in the cigarette tax be delayed, that the repeal of the taxes on passenger automobiles and tires and tubes be delayed, and that the normal tax on corporations be reduced only to a rate of about 8 percent instead of 7 percent.

IV. For the Actual Year 1939: Total, $8 Billion

If a total revenue of $8 billion is needed in the actual year 1939 -- that is, before the changes recommended in the estate and income taxes can be fully effective -- the measures shown in Table 5 are recommended. These are in addition to those in both Table 1 and Table 4 above (except that the personal income tax figure in Table 1 is reduced to $2.33 billion and the estate tax figure to $0.58 billion).

                               Table 5
        Sources of Additional Revenue, Total $8,000,000,000,
                  Actual 1939 (Add to Actual 1939,
                Figures for Total of $7,580,000,000)
                                                       Billions of
Delay repeal of other stamp taxes                           .03

Delay repeal of other manufacturers' excises                .14

Delay repeal of other sundry taxes                          .11

Increase income tax rates enough to gain, in two
quarterly installments                                      .14

In effect, the total federal system would here be practically the same as it now is except for the increases (which would not yet be fully effective) in the income tax and the estate tax, and the substitution of the privilege tax for the capital stock and excess profits taxes.

V. For Depression Years

The recommended system, compared with the existing system (see Table 2 above) would on its face be somewhat more unstable. The personal income tax and the estate taxes, both unstable, play a larger part, and the capital stock tax, the stamp taxes, the manufacturers' excise taxes, the sundry taxes, and the cigarette tax -- all relatively stable -- play a smaller part. On the other hand the proposed system places less reliance than the existing system on the corporation normal tax, the undistributed profits tax, the excess profits tax, and the gift tax -- all unstable taxes -- and relies slightly on the corporation privilege tax with the moderately broad base, probably a moderately stable tax.

Adjustments within the income tax structure will also make appreciable differences in stability. The net effect of the changes can be ascertained only after considerable statistical investigation, if then. If on the whole, it is decided that the system would be too unstable -- which the writers are not, on the whole, inclined to believe -- the averaging device suggested in Chapter III below can be adjusted so as to get almost any rough degree of stability desired. If the adjustment is pushed so far that taxpayers complain too much about having to pay a depression tax on prosperity income, /4/ the only alternative seems to be to impose heavy or numerous excise taxes that are undesirable from almost every other point of view. There is a fundamental conflict between equity in taxation and cyclical stability in revenue, unless taxpayers are willing to average, over the years, their income tax payments (as well as their income tax liability, which is about all that is recommended in the averaging system in Chapter III).

Income Taxation

The present chapter will be concerned with income taxation conceived of as an instrument based on relative personal ability to support government. Insofar as the taxation of corporate incomes is treated here, it is as a collection-at-source device that must be integrated with the personal income tax. The taxation of corporate incomes for impersonal reasons is discussed in Chapter IV.

This chapter is divided into four sections:

a) Problems restricted to the personal income tax.

b) Problems restricted to the corporation income tax.

c) Problems common to both the personal and corporation taxes.

d) Problems of integrating the two taxes.

It will be assumed in this chapter that the income tax needs no defense as a permanent and major source of revenue to the federal government. The discussion will therefore be concerned with ways in which the tax can be improved.

Section A

Problems Restricted to the Personal Income Tax

The federal personal income tax can be improved, the writers believe, by changes that concern the following matters: /1/

1) Rate scale.

2) The amount and type of personal exemptions.

3) Returns of husband and wife.

4) Earned income.

5) Adjustments to maintain justice in case of a rapid rise in

6) Tax-exempt salaries.

7) Bad debts, interest, etc.

8) Imputed income.

Each of these will now be briefly discussed.

1. Rate Scale

THE NORMAL TAX. The chief reasons for having both normal tax and surtax rates have been:

a) It was a convenient way of taking account of the fact that dividend income had already been taxed by the corporation income tax. /2/ Since the Revenue Act of 1936 removed the exemption of dividends from the normal tax, this reason has no force at the present. It is to be hoped, however, that this exemption will be restored or that some other way will be found to integrate the corporation income tax with the personal income tax. /3/ If some satisfactory way of handling the undistributed profits question were found, the corporation income tax could be abolished and dividends treated in the hands of individuals just like other taxable income. In this case, the dividend question would not call for a normal tax separate from the surtax.

b) The interest on all outstanding federal notes, bonds, etc., is exempt from the normal tax, and the interest on obligations of instrumentalities of the federal government is exempt likewise if the act authorizing the issue so specifies. /4/ Whatever the wisdom of this provision in the first place, it would probably be a breach of faith to circumvent it by abolishing completely the normal rate or lowering it to a nominal figure and replacing it with higher surtax rates. On the other hand, the existence of this exemption is a strong reason for raising, not the normal rate, but the surtax rates, if any rates at all are to be raised.

c) It offers a convenient way to handle the earned income allowance, but is not essential for this purpose. /5/

d) It offers a convenient way to restrict the effect of the personal exemptions and the credit for dependents in reducing the tax of the well-to-do taxpayer. Again, however, there are other ways of achieving this restriction.

In view of these considerations, it is not desirable to abolish the normal tax or reduce it to a nominal amount at this time, but it is also not desirable to increase it, unless an increase is deemed absolutely necessary in the process of integrating the corporation and individual taxes.

Moreover, any substantial increase in the normal tax -- assuming it is levied at a flat rate -- is likely to add too little tax burden to the middle and upper brackets compared with what it adds to the lower brackets. Certainly a flat rate increment is a much less flexible instrument than an adjustment of the graduated surtax schedule.

COMBINED SURTAX AND NORMAL TAX RATES. The federal personal income tax rates are very low at the bottom and very high at the top. They are probably higher at the top than they should be, in view of the incentives to avoidance and evasion that they create, and the consequent administrative troubles, and in view, also, of the danger of stifling the willingness to take risks. Granted that it is desirable to make the richest taxpayers pay the total amount that the present rate schedule calls for, it is advisable to lower the top bracket rates somewhat and increase the middle bracket rates. Of course this procedure implies a net addition to the federal revenue since the middle bracket increase would also hit taxpayers who do not get into the top brackets. However, as Chapter II shows, there is little doubt that income tax rates should be increased in the course of improving the tax system. Consequently it is recommended that all bracket rates that now are higher than 63 percent be reduced to that figure.

The overall, or average rate, of federal personal income tax at certain levels is as follows, for a single man with no dependents: /6/

        Net Income                 Tax
          $1,500             9/10 of 1 percent

           2,000             1.6 percent

           3,000             2.3 percent

          10,000             5.6 percent /7/

For higher income levels it is advisable to include the burden of the corporation income tax. The following figures assume that half of the taxpayer's income is from salaries and half from corporate earnings. They include the federal corporation normal tax but assume that all earnings are paid out so that there is no undistributed profits tax:

        Net Income                 Tax
         $30,000             18.3 percent

          50,000             23.4 percent

          70,000             28.3 percent

         100,000             36.2 percent

For a married taxpayer the percentages are of course smaller, because of the personal exemption. State income taxes add to the burden in two-thirds of the states, but it must be recalled that the combined burden cannot be obtained merely by adding the state rates to the federal rates, since the federal law allows deductions of state income taxes paid, in computing net income, and some states allow a similar deduction for federal income tax paid. /8/

Since time and resources did not allow the present writers to make estimates of amounts to come from specified surtax schedules, it is here recommended simply that the general type of surtax schedule proposed on p. 11 of Vol. II of Tax Revision Studies, 1937 be adopted. This means that substantial increases will be confined to taxpayers with $20,000 income or more -- a procedure that is advisable because (a) there is much evidence that in the next few years the states will be in considerable need of raising substantial sums from income taxes, and, owing partly to the migratory tendencies of large incomes, it is these lower incomes that the states will want to depend on; (b) the federal and state payroll taxes will before long be appreciable burdens on the low incomes.

2. Personal Exemptions.

The personal exemptions /9/ raise two problems that are sometimes confused with one another:

a) What individuals should be exempted from
paying any personal income tax at all?

b) Among those who are required to pay something,
what differences should be allowed on account of
marital status and dependents?

With respect to the first question, the present writer has nothing to add to the analysis and suggestions made in Facing the Tax Problem, pages 354358 and 451 (showing number of returns under existing and hypothetical exemptions), 413414 (paragraph number 59), 42132 (tax-consciousness), 430 and 454456 (recommending a lowering to $1,000, $500, and $200). See also the report on the federal income tax submitted to the Treasury by Dr. Roy G. Blakey in 1934. Possibly the Blakey recommendations ($1,600, $800, and $400), milder than that in Facing the Tax Problem, should be adopted -- this depends chiefly on the administrative task involved. Moreover this important proviso must be recalled: in view of the regressiveness of the heavy local taxes on property, the federal government should impose only a light rate -- say, not more than 4 percent -- on the new taxpayers, and should reduce other taxes that are now a burden on them. These new taxpayers should not, in the writers' opinion, be called upon yet to bear an addition to their total tax burden. Reduction and repeal of some of the miscellaneous excises, etc., offer some opportunity in this direction, but not much. The sole important opportunity (aside from the social security tax program) that the federal government has of relieving all of the low income taxpayers of a hidden tax burden is by reduction of the tariff duties. This may or may not be politically impossible, but there can be no escape from the conclusion that a reform of the tariff duties is the major step that the federal government must take if it is really to offset, for these new taxpayers, the burden of the income tax.

With respect to the second question, more time is usually spent discussing the various forms in which the exemptions can be stated than the subject is worth. Whether the exemptions should or should not be applicable to the surtax schedule; whether they should come out of the lowest brackets or the highest brackets; whether they should be in the form of credits against tax; whether they should or should not "vanish" above a certain level; the answers to all these questions depend on the degree of spread in tax liability that one wishes to maintain between the single person and the married person (and those with and those without dependents) at various income levels. For the rest, anything that can be done by one or the other of these exemption devices can also be done practically as well by adjusting the surtax rates. Without feeling strongly one way or the other, the present writers are inclined to favor leaving the mechanism (but not the absolute amounts) of the exemptions as at present, in order to maintain a substantial difference, in terms of percent of income, between the tax liability of the single and married (and those with and without dependents) at the higher income levels. See also the remarks in the following section, "Joint vs. Separate Returns."

3. Joint vs. Separate Returns

This analysis embraces the problem of community property, but goes further, to examine in a general way the relation of spouses with respect to the income tax. No details of the community property problem will be given here, since it is assumed that the readers are already familiar with the facts.

Ultimately, the question seems to be: what should be the taxpaying unit for purposes of an income tax? The present writers have had considerable difficulty in isolating the issues in this problem and in forming a judgment, but at the moment the relevant points seem to be as follows:

The personal income tax rates are progressive because, as the number of dollars a person has increases, any one dollar means much less to him. Roughly, as his income grows, he gets to receiving "semi-luxury" income and then "luxury" income, instead of simply "necessity" income, and it is judged reasonable to take a larger share of his luxury income in taxes than of his necessity income. Suppose, now, a man and woman, not married, each earn a "necessity" income (say $4,000 each), but not a luxury income. If they marry, and pool their incomes for their joint use, does any of it become a semi- luxury or luxury income? Yes, insofar as two can live more cheaply together than they can apart, and thus can have something to spare towards luxuries that wasn't to spare before. Otherwise, no. At least, this much seems certain: neither the man nor woman is in the position he would have been if he had remained single and his salary had been raised to $8,000, or if he had come into possession of investment income of $4,000 a year. Yet, this seems to be in part the assumption of an income tax law that would force this man and woman to file a joint return, so that $4,000 of the family income would be subject to higher rate brackets than any part of either's income had been subject to before marriage.

Suppose that before marriage the woman had been receiving no income, having been living with her parents, and that the man has been earning -- and continues to earn, after marriage -- $8,000 a year. May it not be said on his behalf that at least some of what was luxury or semi-luxury income becomes now necessity income, entitling him (in the name of the family) to a lower average rate on his $8,000? -- and correspondingly, is not some part of the woman's parents' income released from the necessity class and floated up into the semi-luxury or luxury class?

This approach suggests that the family group should be considered the taxpaying unit, but that the family income should be divided into more or less equal parts -- as many parts as there are members of the family -- and the progressive rate scale applied separately to each part. Each part, incidentally, would be treated as the single taxpayer's income is now treated -- a low exemption. There would be no married exemption. Likewise, there would be no credit for dependents (probably the share of the family income allotted to very small children, for tax purposes, would be smaller than the part allotted to the adults in the family group: thus, each child under 10 might be allotted half the amount of income allotted to each parent, but in no case more than $5,000).

A second plan is to (a) total the family's income, (b) divide by the number of members in the family, but counting anyone other than the spouse as only a fractional part of a member, (for example, each child might be a member), (c) compute the tax on the resulting income, (d) multiply the tax by the number of members. This plan does not, perhaps, lend itself so readily to setting a limit to the tax benefit that the wealthy may derive from children, etc.

The significance of this line of reasoning is not that some such plan should be at once enacted -- for it will need severe scrutiny before it can safely be pushed -- but rather that it indicates the direction in which future reforms with respect to this matter are likely to occur; and that, therefore, any step that would force a family group to file a joint return would be a step in the wrong direction. Such a step has been suggested as a means of insuring uniformity of treatment, community property states with noncommunity property states. To some students of the problem, uniformity in this matter is more important than any one particular type of treatment, and if a joint return method was easy to secure, while another method was not, they would not have much regret about forcing joint returns. In view of the considerations adduced above, the present writers do not share this attitude. The path of equity seems to lead in the direction of the community property states, not away from them. It would necessitate steeper rate scales to get the same revenue (a real political disadvantage, of course), and would have the effect of lightening the burden of married taxpayers at the expense of single taxpayers. The fact that the married exemption ($2,500) is more than twice the single exemption ($1,000) may be unconscious recognition of the injustice done in subjecting necessity income to luxury rates, but in any case it of course results in inadequate compensation to those with incomes of more than $7,000 or $8,000, while perhaps somewhat over-compensating those with smaller incomes.

If the plans noted above for dividing the income were not feasible, a rough measure of justice might be done by substantially increasing, as soon as the incomes get into the progressive part of the rate scale, (a) the spread between the married exemptions and single exemptions and (b) the credit for dependents. Measure (a) would tend to put married taxpayers in non-community-property states on a par with those in community-property states, for those with moderate incomes. Perhaps the well-to-do families in non-community- property states can take care of themselves anyway, since they often have a substantial investment income that can be judiciously divided among members of the family (with some penalty in the way of a gift tax, it is true) in order to minimize income tax. At the other end, it may be noted that as long as the rate is flat, not progressive, the problem of luxury vs. necessity income does not arise (likewise, the families in community property states get no advantages). /10/ But as incomes get larger and get into the graduated rate scale, the spread between single and married exemption needs to widen. Thus the alternative to separate returns would be a system along this pattern (in practice, approximations could be used for each income bracket). The figures were obtained by finding an exemption for which the tax is equal to what it would be if income were divided as suggested in the second plan (see below).


                Single personal  Married personal   Credit for first
If income is:     exemption is:    exemption is:      dependent is:
$ 5,600.00 /1/      $ 800.00         $ 1,600.00         $ 200.00

  6,000.00            800.00           1,800.00           100.00

  7,000.00            800.00           2,300.00           100.00

  8,000.00            800.00           2,800.00           100.00

  9,000.00            800.00           3,300.00           100.00

 10,000.00            800.00           3,644.44           266.66

 15,000.00 /2/        800.00           4,400.00           640.00

 20,000.00            800.00           5,384.61           653.84

     /1/ Supposing graduation to begin at $4,000 ($5,600 $1,600).

     /2/ The exemptions might stabilize at about this level.

For other points in connection with joint and separate returns, the reader may be referred to Chapter VII of Blakey's report to the Treasury in 1934 on income tax matters. The present writer has nothing to add to the analysis on the community property states in this memorandum and in the comments made on the matter in congressional committee reports and statements by the Secretary of the Treasury.

The optional plan that Blakey proposes points in the same direction as the plans suggested in the present chapter, but it does not go quite so far. Moreover, it is not apparent that the rate scale's progression should necessarily be steepened in the manner suggested by the Blakey memorandum.

4. Earned Income

The existing earned income credit is not large enough to warrant serious criticism -- and by the same token, not large enough to be of very much importance. Some criticisms of the structure of the existing credit, and of the general idea itself, are, however, in order, in case a move develops to enlarge the credit (as suggested, for instance, in Vol. II of Tax Revision Studies, 1937).

As to the structure of the existing credit, one effect of the credit is to raise the personal exemptions (without affecting the requirement for filing returns) from $1,000 to $1,111 (single) and from $2,500 to $2,778 (married). This results from the assumption that all income up to $3,000 is earned net income, and therefore reducible by 10 percent. Thus a single person with a net income of $1,111 gets a personal exemption of $1,000 and an earned income credit of $111, and pays no tax. It would seem advisable to reshape the earned income credit so that those with net incomes between $1,000 and $1,111 (or $2,500 and $2,778, married) should pay at least some tax, especially if tax-consciousness is important enough to justify the much more difficult step of lowering exemptions. The present system simply adds to the number of non-taxpaying returns. One suggestion is to make the 10 percent apply, not to the net income, but the net income less the personal exemption, credits for dependents, and (if desired) credit for exempt bond interest. The percentage might be raised somewhat, in this case, without loss of revenue. This net income less credits for normal tax (before earned income credit, of course) would then, for simplicity's sake, be assumed to be earned net income up to, say, $3,000.

As to the earned income credit idea in general this warning may be given: it is such a crude device that it is likely to result in considerable injustice if pushed very far. The chief argument for it seems to be that income other than "wages, salaries, professional fees, and other amounts received as compensation for personal services actually rendered" (section 25 (a)(4)(A)) -- or in general, investment income -- represents a more lasting source of revenue and one that comes in with less effort. But compare, with respect to both of these points, the following persons:

A: Wage earner or salary receiver.

B: Receives part of his income in coupons from the bonds
issued by some concern that has gotten into
difficulties so that there is considerable doubt
whether the company can continue service on its bond.
This investor has purchased the bonds with savings from
a salary equal to that of A.

C: Same as B, except that the bonds have passed to him by

D: Receives part of his income in coupons from the bonds
issued by a concern of first-rate credit standing. This
investor has purchased the bonds with savings from a
salary equal to that of A.

E: Same as D, except that the bonds have passed to him by

As far as permanence of source of revenue is concerned, B and C may well be worse off than A, yet the earned income credit favors A. As concerns the effort or sacrifice necessary to obtain the income, there may be a case for disallowing to B and D some relief granted to A. But it is possible that in some cases the process of saving (which is of course being carried on continually by B and D simply by refraining from consumption of the capital) takes as much effort and sacrifice as the process of earning. Compared with [illegible], A does indeed have a sound case for differential treatment, but here it may be argued that the "unearned" aspect should be taken care of by taxes that strike at the root of the "unearnedness" -- death and gift taxes, land increment taxes, excess profit taxes, etc.

There may be enough to the earned income idea to justify a failure to repeal the present degree of discrimination, but it seems inadvisable to extend the discrimination.

5. Adjustments in Case of a Rapid Rise in Prices

This matter is not one that requires immediate attention, but it may require study within the next year or two.

If prices and money incomes rise rapidly, the rate scales of graduated taxes like the income, death, and gift taxes become distorted in terms of real purchasing power. Likewise, the real relief afforded by the personal exemptions decreases. The real pressure of the tax grows much heavier on all income receivers, but especially so on those with low and moderate incomes. If a married taxpayer with two dependents has been receiving a salary of $8,000, and if this salary doubles at the same time that prices in general double, his real purchasing power has not increased. Therefore, his dollar tax total should no more than double. Actually, under the existing federal law, it would increase to more than five times the former amount (from $184 to $948). The increase is even more severe in the lower income ranges. If the taxpayer has been receiving a salary of $4,000 and it doubles (general prices doubling also), his tax, instead of doubling, increases to 15 times the former amount.

Moreover, as money incomes rise, many persons hitherto exempt entirely become taxable.

If the price rise reaches astronomical proportions, the real burden finally becomes constant instead of growing heavier, as almost everyone finds most of his income in the upper reaches of the schedule, when progression tapers off and finally ceases. Meanwhile, however, the relative tax burdens have been much distorted.

Insofar as taxpayers delay payment while prices are rising, they tend to benefit, since they pay finally in cheap money as obligation incurred in terms of dear money. But some taxpayers will delay more than others, and the net result will be an arbitrary reshuffling of the real burden of the income tax among taxpayers, with probably an increase in the net total burden, for a while at least.

The factors described above may be, on balance, helpful in checking inflation, but the cost in terms of tax justice is heavy.

How to forestall this arbitrary reshuffling of the burden, and possibly also the increase in its net total, is not immediately apparent, but at least the matter seems worth serious study, in view of (1) the possibility of a rapid rise in prices, and (2) the difficulty of getting any reform measure enacted in the midst of the confusion incident to such a rise. Possibly some form of index number treatment is indicated. /11/

6. Tax-Exempt Salaries

This brief section is merely to recall that in any program for readjusting federal-state relations, particularly in connection with tax-exempt securities, the matter of tax-exempt salaries should be included. From the point of view of revenue, it is by no means as clear as with tax-exempt securities that the chief loser under the existing arrangement is the federal government. Those who receive government salaries usually have small total incomes, and many if not most of the income-taxing states tax those small incomes more heavily than the federal government does. On the other hand, about one-third of the states do not levy personal income taxes. It might be worthwhile to explore this problem of where the balance of advantage might lie if all government salaries were subject to income tax. The federal government might find that it had a bargaining point with the states in connection with any plan to tax government bond interest.

Probably the federal government -- and more likely the states, since the federal tax is nationwide in application -- pay smaller salaries than they would be doing if all salaries had been subject to tax. Therefore, in time, salaries might be expected to rise if they were made taxable. However, even if the result was no net fiscal gain to either layer of government -- and this seems highly doubtful -- the procedure would be justified on the basis of getting a closer adjustment of relative real burden among various taxpayers and on the basis of tax-consciousness. Also, it would create a better feeling toward the income tax on the part of other taxpayers.

7. Bad Debts, Interest, etc.

The general philosophy of the income tax disallows deductions of any expense not incurred with the intent of gaining taxable gross receipts. In other words, personal expenses -- expenses that contribute to one's personal satisfaction -- in contrast with business expenses, are not deductible. The existing federal law departs from this concept in allowing deduction of all bad debts, all interest (except on debts incurred to purchase or carry certain wholly exempt bonds) and losses from casualty or theft. (There is, however, a general restriction refusing deduction if the deductible item is allocable to wholly exempt income). The casualty and theft deduction may be supported on the ground that it is not the kind of using-up of income that would indicate the taxpayer had power to support government, /12/ but it seems clear that the bad debt and interest deductions now granted are too broad. Bad debts might be assimilated to casualty (or theft!), but there is an element of volition here that changes the situation. Incidentally, if uninsured casualty or theft losses are deductible, should not deduction be allowed for premiums on casualty and theft insurance? In both cases the taxpayer suffers a loss -- it is chiefly a question whether he chooses to undergo a number of small but certain losses or to risk a single large but uncertain loss.

Any restriction on interest or bad debt deductions should probably be made applicable only with respect to transactions entered into after the change in the law. Otherwise extreme hardships might result in some case, where loans had been contracted, or made to others, on the assumption that the item would be deductible.

8. Expenses That Are Both Personal and Business

This item is inserted, not with any suggestion that action will be taken in the near future, but because it is a fundamental problem that will probably have to be faced sooner or later. Some parts of the expenses for food, clothing, shelter, and medical attention are clearly connected with particular kinds of income-earning efforts. It is not clear whether practically all of them should, as they now are, be disallowed as deductions. For a development of this point, see the forthcoming "Income Conference Volume," National Bureau of Economic Research, Part VI (Shoup), "The Distinction Between 'Net' and 'Gross' In Income Taxation."

9. Imputed Income

Imputed income, or nonmonetary income, arises from the possession of a relatively durable good or from the creation, by the taxpayer, of goods or services for his own use. Home owners, compared with dwellers in rented premises, enjoy imputed income. A farmer who gets his own food from his land by his own efforts enjoys imputed income, and may be contrasted with the urban dweller who buys his food with his taxable money wage.

The existing federal income tax law -- and apparently all the state income tax laws -- do not tax imputed income. A considerable injustice is thereby done the taxpayers who get none or little of this kind of income.

For a brief analysis of the subject, see pages 29-40 of Memorandum G submitted to the Treasury in 1934 by Shoup.

The taxation of imputed income should be introduced slowly, to prevent undue hardships, and it should not even be started until administrative authorities in the Treasury have made a detailed study of the problems involved. The task might prove impossible to achieve with more than rough justice, but at least even that would be an improvement over the present situation.

Section B

Issues of Corporation Income Taxation

This section concerns four issues of policy in corporation income taxation. They are treated from the point of view that the corporation income tax is a tax at source on the individual stockholder and should be adjusted to approach most nearly the standard of taxing persons with reference to their total economic positions. Some of the issues will be discussed again in the next chapter in the analysis of the impersonal business taxation of corporations.

1. Progressive Corporation Tax Rates /1/

"Prior to 1932, the federal corporation income tax was given a slight degree of progression by the exemption of $3,000 of net income. Since 1935 the federal tax has been imposed at progressive rates. Those of the 1935 act ranged from 12 percent on the first $2,000 of income to 15 percent of income above $40,000. The 1936 rates ranged from 8 percent on the first $2,000 to 15 percent on income above $40,000." /2/ The significance of these rates rests not so much in their present mild progressiveness as in the general issue that they raise with regard to future policy.

The burden for any corporation tax falls proportionally on all the holders of any one class of stock. Accordingly, the relative burden on wealthy stockholders and poor stockholders is not affected by the proportionality or progressiveness of the corporate tax rate structure. Only if wealthy persons in general owned the stock of large corporations and low-income persons in general owned the stock of small corporations would progressive tax rates on corporation income have the effect of progressive rates on individuals. Common observation indicates that aside from extremely small corporations, this situation probably does not exist generally. Because of their marketability the stocks of large corporations are in general more suitable for investors with small incomes than are the stocks of smaller corporations. Investors with large incomes can diversify their risks among the less marketable and generally more speculative stocks of small corporations. Even if the wealth of the stockholder were positively correlated to the size of the corporation whose stock he owns, a great deal of inequity among stockholders in corporations of different sizes would result because of the numerous cases that did not follow the general rule. Tax burdens would vary widely among persons with the same incomes.

For these reasons, the conclusion is reached that if the corporation income tax is intended to be a personal tax collected at source, the corporation rate should be flat rather than progressive.

2. Consolidated Returns

The use of the corporation income tax as a method of personal taxation implies that insofar as possible the economic position of the individual will be treated as a unit with gains and losses from all sources merged. Taxing each corporation separately, although in general necessary, violates this principle, since the loss incurred by a person in one corporation cannot be offset against the income gained in another. Where the stockholders of two or more corporations are identical or nearly so, it is not necessary to tax each corporation separately; the incomes and losses of all the corporations could be consolidated as was formerly permitted in certain circumstances by the income tax law. This procedure comes closer to treating the individual stockholder as a person than when consolidated returns are not recognized. In the absence of consolidation, the individual is treated as if none of the corporations were suffering losses -- although adequate provisions for offsetting losses against the income of future years reduces the significance of this effect.

The use of consolidated returns reduces revenue somewhat, which may be a compelling factor in some fiscal situations.

The consistency of allowing consolidated returns in the impersonal taxation of business as well as the objection to them on the ground that holding company structures are socially undesirable will be discussed in the next chapter. These questions have no bearing on personal taxation as such. From the point of view of personal taxation consolidated returns should be freely allowed.

3. Taxation of Inter-Corporate Dividends

Taxation of inter-corporate dividends is not consistent with the use of the corporation income tax as a personal tax. It is a matter of little moment to the final individual stockholder whether the income passes through the hands of several holding companies on its way to him from the earning corporation. It may be that dividends will reach him sooner if they are received directly from the operating corporation, than if they pass first through several holding companies. This may be significant in integrating the burden, which is discussed in a later section. However, so far as the taxation of stockholders with reference to their total economic position -- the ideal of personal taxation -- is concerned, taxing inter- corporate dividends is not logical, since it places heavier burdens on persons with investments in holding companies than on persons with investments in operating companies.

Taxation of inter-corporate dividends may be considered also in connection with the impersonal taxation of business or the control of holding company structures. These matters will be discussed in the next chapter.

4. Tax Base

The following chapter analyzes the proposal that a business tax on corporations should be based not on corporate net profit but on the operating profit of the business regardless of its capital structure. Where, as in the present chapter, the corporation tax is viewed as a personal tax collected at source, the use as a tax base of corporate net after payment of interest and other contractual charges is more logical than the use of operating net without deducting such payments. The income received by the stockholders comes from the corporate net, so that if the corporate tax is substituted for the personal tax it should apply to corporate net. /3/

Section C

Problems Common to Both Corporation and
Personal Income Taxes

Both the federal personal income tax and (in case it is retained) the federal corporation income tax can be improved by changes that concern the following matters:

1) Carryforward of losses.

2) Averaging of income over a period of years.

3) Capital gains and losses.

4) Depletion allowances.

5) Tax-exempt securities.

Each of these will now be discussed in turn.

Carryforward of Loss

The failure of the existing law to allow a loss -- whether from business operations or otherwise -- to be carried over and absorbed in future years' income penalizes the taxpayer with irregular income. It particularly penalizes the businessman who takes risks, and who must count on large profits from some of his ventures because of what is almost a certainty that he will lose on some of them (which case, he of course does not know). Similarly, it penalizes the taxpayer who operates a business that is particularly sensitive to business cycle influences. These injustices are bad enough in the ordinary income tax, but they are particularly bad under the undistributed profits tax. Over a number of loss and gain years the concern may have no net profit, but the undistributed profits tax will have put pressure on it to pay dividends -- in effect, to pay dividends out of capital, or at least prior accumulated surplus. The loss years eat into the common stockholders' equity, and in the good years they have no chance to rebuild it, so that if dividends are declared in taxable preferred stock, the common stock equity gets very thin.

Logically, no limit should be set to the number of years the carryover should be permitted. Administrative expediency might require some limitation, but at least a five-year carryover seems advisable.

The recommendation holds whether or not an averaging plan (see next section) is adopted.

Averaging Income Tax Over Period of Years /1/ /2/

1. Problems Caused by Fluctuating Incomes

The income of many taxpayers, notably those that have capital gains and losses, fluctuates from year to year.

This fluctuation in net income may be disadvantageous for the government because it may result in too great a fluctuation in its annual receipts from all sources.

It is usually disadvantageous to the taxpayer in question. Of two taxpayers, each having a total income of, say, $100,000 over a period of 10 years, the one whose income is the less regular will pay a greater total tax over the 10-year period, if rates remain unchanged. The extra tax that he pays by being thrown into the higher surtax brackets in his good years is greater than the tax he saves by being thrown into the lower surtax brackets in his poor years. This result necessarily follows as long as the higher brackets carry higher rates than the lower brackets.

Moreover, if the taxpayer's income is so erratic that it is a minus quantity in some years, he suffers still more (compared with the regular-income taxpayer) under the present federal law, which allows no carry-over of losses or credits against normal tax net income and surtax net income.

The burden put on the taxpayer with fluctuating income also has important economic influences, at least with respect to capital gains and losses. The taxpayer, in his endeavor to even his income over the years, may buy or sell at different times than he would if there were no tax. Sometimes he will change his buying and selling policies for reasons that are irrational, but that are developed by the pressure of the high tax that fluctuating incomes pay. And if the tax law tries to relieve this pressure by some measure such as the existing step-scale plan, it is apt to interfere with investment and speculative procedure even more. /3/

2. Three main devices are available for decreasing the irregularity
of annual net income

The simplest, and least effective, is to allow a minus net income of one year to be carried over to succeeding years until it is absorbed in being offset against positive net income.

Another method is to average the incomes of two or more years and compute the tax each year on such an average income. This method may readily be used in conjunction with the carryover-of-losses method, and it is this combination that is analyzed in the present chapter.

A third method is to lengthen the accounting period, so that a return would be made out, for example, only once every two years, and would show a net income figure for two years. This method, too, may be used in conjunction with the carryover-of-losses method.

3. Chief Objection to Averaging Systems

Averaging systems have been used, but have in several cases -- perhaps in most cases -- been abandoned, as in Great Britain and Wisconsin. The chief reason for abandoning them in these two cases seems to have been dissatisfaction on the part of the taxpayers who discovered that they had a tax to pay even in a year when they had no income. Perhaps they also objected simply to paying more, in a year when they had some income, than they would pay under a straight annual system. They had to do this under the ordinary averaging system, in the relatively poor years. Apparently they did not consider the relief from taxation in the good years an adequate offset.

To the extent that tax policy is guided by the objection just stated, no averaging device or any other plan can be used as a means of decreasing the year-to-year fluctuation in the government's revenues. The only way to decrease that fluctuation is to assume that taxpayers will save more in good years in order to pay taxes in bad years. The local property tax, for example, operates on this assumption. But if taxpayers were willing to do this, they would not raise the objection just stated.

Supposing, however, that tax policy must be guided by that objection, an averaging device may still be feasible as a means of decreasing the discrimination between recipients of irregular income and recipients of regular income. What is necessary in this case is that the fixing of the tax liability be on an averaging basis and the payment of the tax be on the usual annual basis. Thus, large payments would, as usual, be made in good years and small payments in poor years.

This disassociation of liability and payment may also eliminate some part of the economic influences on the speculative and investment markets mentioned above.

4. Subject of the Present Section

The present section describes a device for disassociating liability and payment. A section below lists the main characteristics of the plan. An illustration shows in some detail how the plan would operate. If this plan merits further analysis, the analysis should be pushed at once, since the averaging device may make it unnecessary to give capital gains and losses any special treatment at all. They could be simply thrown in with other income and deductions (for further details, see the section below in this chapter on capital gains and losses). In other words, before deciding whether to replace the existing step-scale plan by some other special gains and losses plan that may be equally complicated (though more just), it would be better to explore first the possibilities of the averaging device, since such a device, if feasible, not only may be the best way to handle capital gains and losses, but may improve the income tax system in other ways as well.

5. Major Results of Proposed Devices

The chief characteristic of the averaging plan proposed here is that, as far as payment is concerned, the taxpayer with irregular income pays slightly more (owing to the increase in rates to be noted below) during his poor years, than he pays under the existing system, but in his good years pays decidedly less than he pays under the existing system, -- and, on balance, over a period of years, pays less. To raise the same amount of revenue from all taxpayers as is raised by the present system, the schedule of rates would have to be somewhat higher. In other words, over an entire cycle of fluctuations, the regular-income taxpayer would be taxed more in order that the irregular-income taxpayer might be taxed less. (Actually, it might not be necessary to raise the present level of rates, since there would probably be a revenue gain in taxing gains and losses under the averaging device instead of under the existing step- scale device).

The net effect on the fluctuation in government revenues would be to decrease that fluctuation slightly. In the poor years, all taxpayers would pay, at slightly higher rates, on the same base as under the existing system. Since the same revenue would be obtained, over an entire period of fluctuation, as under the present system, it follows that in good years the total payments would be somewhat less than under the existing system.

6. Chief Technical Points

The chief points in the averaging device described in this chapter are:

(1) The averaging feature is cast in the form of a relief, or credit against tax, to taxpayers with irregular incomes.

(2) Use of this relief provision is optional with the taxpayer, but it is intended to be simply enough constructed so that almost every taxpayer with irregular income will take the trouble to use it (but it may be advisable to restrict the relief to taxpayers filing on Form 1040).

(3) The average is an average of annual income over a fairly long period, say six years.

(4) The plan can be introduced gradually, using a two-year average the first year, a three-year average the next, and so on, until a six-year average is reached. The present analysis will assume that this method is used. It is possible, however, to formulate the plan so that it will not take effect until six years (or whatever the length of the averaging period may be) after it has been enacted, but so that it will, at that sixth year, be retroactive, in that it will reduce taxes due for the preceding five years.

(5) A new average is struck each year -- that is, the average is a moving average.

(6) In accordance with the implication in (a) above, no taxpayer is, in any given year, asked to pay more, because of the averaging device, than he would otherwise. The only qualification to this is that, since the rate schedule might be slightly higher, all taxpayers might be on a higher level of payment.

7. Two Options

The details noted in (1) to (6) above will apply in any case, but there are some other matters where a choice must be made. Briefly, there are two extremes toward which the averaging system may tend.

1) The taxpayer's real tax liability is, in any case, going to be computed, not on his annual income, but on his average income, whereas he will be paying on the basis of his annual income. Subtracting the former from the latter, one obtains either a positive excess, or credit, in favor of the taxpayer, or a minus quantity representing money due the government. When a positive excess, or credit, tends to pile up, the taxpayer is allowed to use it -- with certain limitations, to keep him from using it up too fast -- to reduce his annual taxes in good years. When a minus quantity in favor of the government appears, the taxpayer is simply prohibited from reducing his annual taxes in good years until a positive excess replaces the minus quantity. The question now becomes: shall the positive or negative balance be carried forward from year to year, modified up and down, of course, each year, but with no break in this accounting and no termination until some such logical stopping point as the taxpayer's death, or bankruptcy, or change in citizenship? This method would in effect try to average one's income over one's entire lifetime, in most cases.

2) Or, shall the positive or negative balance resulting from one year -- say 1942 -- be discarded (or, what is left of it be discarded) after a few years, say by 1947? Sometimes the government would lose, sometimes the taxpayer would lose, by this discarding process. In effect, this plan gives a taxpayer from 1943 to 1947, inclusive, to use, in tax reduction, any excess of annual tax over average tax that was created in 1942 by, for instance, the realization of exceptional capital gains in that year.

It is not prima facie evident which plan would be the more desirable, and therefore each is studied in the pages below. The first plan is called the "continuous balance plan" and the second the "shortened balance plan."

8. The Continuous Balance Plan

a. THE TRANSITION PERIOD. Under the continuous balance plan the taxpayer would, for the first year of operation -- say the income year 1938 -- compute the tax on his annual income, just as he does now, and would pay that tax. In the second year, 1939, he would compute his tax on annual income and would also compute a tax on his average yearly income for the two years, 1938 and 1939 (that is, add the two incomes, divide by two, and compute the tax on the result). If the annual tax were larger than the average tax he would note the difference, or balance. Then he would reduce his annual tax by 20 percent, but in no case reduce it by more than the difference, or balance, just noted. If some balance remained, he would carry it over to the next year. The 20 percent is a rather arbitrary figure, inserted to prevent, in general, too rapid using-up of the credit balance, for this would in turn mean a correspondingly longer time to get back to a credit balance position after hard times. Also, it might result in somewhat erratic revenue fluctuations.

If the average tax were larger than the annual tax, a minus item would be entered in the credit balance line, and carried forward to next year. The taxpayer would pay the tax on the annual income, with no relief.

In the third year -- 1940 -- the same process would be followed, except that the tax on 1940 income would be compared with a tax on an average income of three years -- 1938, 1939, and 1940 -- instead of two years. The credit balance (positive or negative) would be increased or decreased, as the case might be, and, if the result was still a positive balance, it could be used as credit against the 1940 tax, subject to the 20 percent limitation.

Each year, the number of years included in computing the average would grow until six years were included. Thereafter, the average would drop the oldest year and include the current year. The six-year figure is somewhat arbitrary; the longer the period, the better the averaging, but six years is perhaps as long as is practicable.

b. CREDIT BALANCES. No payments by the taxpayer on account of a minus credit balance would ever be demanded. This may seem at first sight like giving something away, but in reality it is probably not so in most cases. Over a taxpayer's lifetime, the total of the tax on his average incomes will always be less than the total of the tax on his annual incomes. The appearance of a minus credit balance may mean that he has been using up some credit that in the long run will be seen to have been only tentative; but with the 20 percent restriction, or something like it, in force, a minus balance in many, if not most cases, will mean that the averaging period is not long enough to do real averaging with the taxpayer in question, or that its introduction came at an unfavorable, unfair time for him. However, there is probably no need to argue the point at length; such part of such minus balances as does remain, when the taxpayer finally drops out of the picture, that is traceable to too vigorous absorption of prior positive credits, may simply be put down as one of the imperfections of the plan, one of the costs that it may be desirable to pay to get the benefits of an averaging plan.

In view of these remarks, one may question whether a taxpayer should ever be given a refund on account of a positive credit balance existing at, say, death, bankruptcy, or change in citizenship. It might be advisable to let the credit lapse -- or, in other words, provide that the credit is a personal item good only for reducing personal income tax otherwise payable. On the other hand, further study might show that the chances of the Government's paying much in refunds would be small, and that it would be, again, a cheap price to pay for the averaging system.

c. ADMINISTRATIVE REQUIREMENTS. From the administrative point of view, this continuous balance plan would make it necessary for the Bureau of Internal Revenue to file the returns of each taxpayer for at least two successive years in the same folder. The taxpayer, also, would have to look at his last year's return in making out the current year's. He would not have to go back farther. His last year's return would have on it a schedule showing the income of each of the preceding six years (except of course during the transition period).

d. NET LOSSES. A net loss (whether from business or otherwise) in any one year would be carried over into succeeding years until absorbed by net income. If a person entered bankruptcy or died with an unabsorbed net loss it might or might not be feasible to refund to him or his estate a tax on the amount of the net loss. If this could be done it would be advisable, perhaps, as a matter of tax justice, if his taxable lifetime showed a net income after allowing for all losses. In any case, however, it represents a greater problem of injustice under the existing system than under an averaging system.

e. CORPORATIONS. The plan would be just as feasible for the corporation normal tax and undistributed profits tax as for the personal tax. When the rate is flat, or when the graduation is mild, as under the corporation normal tax, the irregularity of income causes relatively little injustice as between taxpayers with regular incomes and taxpayers with irregular incomes. However, the undistributed profits tax places a heavy penalty on the fluctuating- income corporation, and the averaging device would be of real benefit here, if the tax is retained.

f. CHANGES IN RATES AND BASE. In some years the tax rates, and the provisions governing the normal-tax and surtax net incomes and the credits against tax, will be different from those of some of the preceding years. Should the averaging device take account of these changes? The simplest kind of averaging device takes the normal-tax and surtax net incomes as actually reported for the past years, and, after averaging these incomes, applies the rate scales in force in the current year (treatment of credits against tax would be difficult in any case, but it is a minor point, usually). Thus the averaging would compromise by using old tax-base provisions and new rate scales. Whether a more logical but more difficult procedure is justified is at least doubtful.

g. REDETERMINATION OF INCOME. In some cases the taxpayer's net income as reported in prior years will not be the same as the net income finally computed after conference or litigation. If such a final computation has been arrived at before the taxpayer makes out his return in the current year, he might be required to use the corrected tax base, not the one he first reported. /4/ If the settlement has been reached by simply compromising on the amount of tax due, some arbitrary method of translating this compromise into terms of changes in the tax base might be required of the tax officials, the compromise being pretty arbitrary anyway (the writer understands such compromises are rare). If the settlement is not reached until after the return is filed, there are two practical methods: (1) allow the change to affect only those returns filed after the settlement is reached, on the assumption that over a number of years the pluses and minuses will in general cancel each other, especially if the period taken for averaging is fairly long; (2) enter the change as if it were a change in the current year's income (thus it would affect income over the full averaging period). Probably the latter would be the better.

h. TAX FORM. The checking of the taxpayer's computation of average income, etc., would increase the administrative task, but there might be nearly counter-balancing savings. For instance, this device would, in the writers' opinion, make it possible to treat capital gains and losses exactly like other income and other deductions, thus allowing omission of a complicated schedule for gains and losses. Insofar as a net increase in complexity was the result, the question would be one of weighing the administrative disadvantages against the net economic and equity advantages. As the present writer visualizes it, the tax return form would have on it a box something like this (the tax return form for 1944, when the system is well under way; the forms for the earlier years would be somewhat simpler):

[This to replace at least part of the capital gains and losses schedule]

            Computation of Tax Relief for Taxpayers With
                 Fluctuating Income: Return for 1944
(a) Return filed shows       normal            surtax
                             tax net           net
                             income            income
1939*                        ....              ....

1940*                        ....              ....

1941*                        ....              ....

1942*                        ....              ....

1943*                        ....              ....

1944*                        ....              ....
                             ----              ----
(b) Total                    ....              ....

(c) Divide by 6 to get
average income               ....

(d) Normal tax on
average income               ....

(e) Surtax on average
income                                        ....

(f) Total tax on
average income
(d) + (e)                    ....

(g) Tax payable
(before relief) on
this year's income
(Item 33, present
return form)                 ....              If Item (g) is not
                                               larger than Item (f)
If item (g) is larger                          make the following
than Item (f) make the                         computation:
following computation:
                                               (h) Plus or
                                               minus credit balance
(h) Difference between                         from last year's
Item (g) and Item (f)        ....              return ...

                                               (i) Difference betwe
                                               Item (g) and
(i) Plus or minus                              Item (f). . .
credit balance from
last year's return           ....
                             -----             (j) Item (h) minus
                                               Item (i) -- this is
(j) Present balance                            plus or minus credit
(Item (h) + Item (i))        ....              balance to carry
                                               over to next year
                                               (no tax relief this
                                               year) ...
(k) Twenty percent of
Item (g)                     ....

(l) Tax relief: Item
(j) or (k), whichever
is smaller (if Item
(j) is minus, no
relief is granted)           ....

(m) Item (j) minus
Item (l): plus or
minus credit balance
to carry over to next
year                         ....

     *Data for these years would be copied from 1943 return

9. The Shortened Balance Plan

The shortened balance plan has all the features of the continuous balance plan, except that:

1) The return form does not use the words "credit balance," and in general avoids the implication that the government owes the taxpayer anything or that the taxpayer owes the government anything.

2) Instead of a credit balance item, carried forward from the previous year's return, the form has space for entering the average tax and the annual tax of each of the six years. (These data for the five preceding years will be found on the return the taxpayer made last year). The taxpayer adds each set, gets the difference, and sees whether he has a positive difference in his favor, and how much. If the difference is negative he stops his computation there (no relief).

3) Each year, of course, after the plan is in full force, the plus or minus balance of the seventh year alone drops out of the picture, being replaced by the plus or minus balance of the current year alone (the plus or minus balances of the intermediate five years remaining in effect, of course).

The return form would have a box in it something like this (replacing at least part of the capital gains and losses schedule):

                           Return for 1944
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income

A. In 1939 return /a/                 _____            _____

B. In 1940 return /a/                 _____           _____

C. In 1941 return /a/                 _____           _____

D. In 1942 return /a/                 _____           _____

E. In 1943 return /a/                 _____           _____

F. In 1944 (present) return           _____           _____

G. Total, six years                   _____           _____

H. Divide Item C by 6                 _____           _____

I. Tax on Item H (average income)     _____           _____

      II. Comparison of (a) Tax Paid, and Tentatively Payable
            for 1944, With (b) Tax Due on Average Income

                    1939    1940   1941  1942   1943   1944   Total

J. Tax on annual    ___a    ___a    ___a  ___a   ___b   ___c    ___

K. Tax on average
income              ___a    ___a    ___a  ___a   ___a   ___d    ___

L. Excess of tax on annual income (Item J minus Item K)

        III. Computation of Tax Relief -- To Be Made Only if
        Tax on Annual Income (1944 Item in Line J) Is Greater
          than Tax on Average Income (1944 Item in Line K)

M. Twenty percent of 1944 tax (in line J)              _____

N. Tax relief for 1944: Item L or Item M,
   whichever is smaller                                _____

                           IV. Tax Payable

O. Tax payable (Item J for 1944, minus
   Item N)                                             _____

     /a/ To be copied from 1943 return.

     /b/ Item O, 1943 return.

     /c/ Item 33 on present return form.

     /d/ Total of items in line I.

The following illustration carries the shortened balance plan through the transition years and into the first full year (it starts with 1937, but it might instead, of course, start with 1938 or 1939). Also, it uses a 10 percent tax relief instead of a 20 percent relief. The 10 percent figure was the one the writer had in mind when he worked the illustration out, but he now believes that a 20 percent relief would be better, to avoid making the taxpayer wait long for his relief.

Time did not permit working the illustration beyond the year 1942.

            Income Fluctuates Periodically with Averaging
                        Period: No Loss Years
A. Two men, A and B, each single, no dependents, have the
following income from salary and professional fees:
                            A                            B
1937                          0                        50,000

1938                     50,000                        50,000

1939                    100,000                        50,000

1940                    100,000                        50,000

1941                     50,000                        50,000

1942                          0                        50,000

1943                          0                        50,000

1944                     50,000                        50,000

1945                    100,000                        50,000

1946                    100,000                        50,000

1947                     50,000                        50,000

1948                          0                        50,000
                        -------                        -------
Total, 12 years         600,000                       600,000
B. Their tax payable under the Revenue Act of 1936: /1/
1937                          0                         9,334

1938                      9,334                         9,334

1939                     33,354                         9,334

1940                     33,354                         9,334

1941                      9,334                         9,334

1942                          0                         9,334

1943                          0                         9,334

1944                      9,334                         9,334

1945                     33,354                         9,334

1946                     33,354                         9,334

1947                      9,334                         9,334

1948                          0                         9,334
                        -------                        ------
Total, 12 years         170,752                       112,008

     /1/ All tax computations in this illustration have been made
only once and, though it is believed no significant errors are
present, a check should be made when time permits.

                           Return for 1938
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income
A. In 1937 return /a/                     0                0

B. In 1938 (present) return          50,000           50,000 /b/

C. Total, two years                  50,000           50,000

D. Divide Item C by 2                25,000           25,000

E. Tax on Item D (average income)     1,000            2,070

       II. Comparison of (a) Tax Paid, and Tentatively Payable
            for 1938, with (b) Tax Due on Average Income

                                  1937        1938        Total

F. Tax on annual income           0 /a/     9,700 /c/     9,700

G. Tax on average income          0 /d/     3,070         3,070
H. Excess of tax on annual income
   (Item F minus Item G)                                  6,630

        III. Computation of Tax Relief -- To Be Made Only if
        Tax on Annual Income (1938 Item in Line F) Is Greater
          Than Tax on Average Income (1938 Item in Line G)

I. Ten percent of 1938 tax (in line F)                      970

J. Tax Relief for 1938: Item H or Item I,
   whichever is smaller                                     970

                           IV. Tax Payable

K. Tax payable (Item F for 1938, minus Item J)            8,730

     /a/ To be copied from 1937 return.

     /b/ In this and all subsequent years, it will be assumed, for
simplicity in computation in this illustration, that the normal tax
income and surtax income are the same, though in fact this would not
be the case.

     /c/ Item 33 on present return.

     /d/ For this year only, same as Tax on Annual Income.

                           Return for 1939
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income

A. In 1937 return /a/                      0                0

B. In 1938 return /a/                 50,000           50,000 /b/

C. In 1939 (present) return          100,000           50,000

D. Total, three years                150,000           25,000

E. Divide Item D by 3                 50,000            2,070

F. Tax on Item E (average income)      2,000            7,700

       II. Comparison of (a) Tax Paid, and Tentatively Payable
            for 1939, with (b) Tax Due on Average Income

                             1937      1938        1939        Total

G. Tax on annual income      0 /a/   8,730 /b/   34,000 /c/    42,730

H. Tax on average income     0 /a/   3,070 /a/    9,700 /d/    12,770
I. Excess of tax on annual
income (Item G minus Item H)                                   29,960

        III. Computation of Tax Relief -- To Be Made Only if
        Tax on Annual Income (1939 Item in Line G) Is Greater
          than Tax on Average Income (1939 Item in Line H)

J. Ten percent of 1939 tax (in line G)                          3,400

K. Tax Relief for 1939: Item I or Item J,
   whichever is smaller                                         3,400

                           IV. Tax Payable

L. Tax payable (Item G for 1939, minus Item K)                 30,600


     /a/ To be copied from 1938 return.

     /b/ Item K, 1938 return.

     /c/ Item 33 on present return form.

     /d/ Total of items in line F.

                           Return for 1940
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income

A. In 1937 return /a/                     0                0

B. In 1938 return /a/                50,000           50,000

C. In 1939 return /a/               100,000          100,000

D. In 1940 (present) return         100,000          100,000

E. Total, four years                250,000          250,000

F. Divide Item E by 4                62,500           62,500

G. Tax on Item F (average income)     2,500           11,850

       II. Comparison of (a) Tax Paid, and Tentatively Payable
            for 1940, with (b) Tax Due on Average Income

                  1937     1938        1939         1940      Total

H. Tax on annual
   income         0 /a/   8,730 /a/   30,600 /b/   34,000 /c/  73,330

I. Tax on average
   income         0 /a/    3,070 /a/   9,700 /a/   14,350 /d/  27,120
J. Excess of tax on annual income
   (Item H minus Item I)                                       46,210

        III. Computation of Tax Relief -- To Be Made Only if
            Tax on Annual Income (1940 Item in Line H) Is
      Greater Than Tax on Average Income (1940 Item in
                               Line I)

K. Ten percent of 1940 tax (in line H)                          3,400

L. Tax Relief for 1940: Item J or Item K,
   whichever is smaller                                         3,400

                           IV. Tax Payable

M. Tax payable (Item H for 1940, minus
   Item L)                                                      30,600

     /a/ To be copied from 1939 return.

     /b/ Item L, 1939 return.

     /c/ Item 33 on present return form.

     /d/ Total of items in line G.

                           Return for 1941
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income

A. In 1937 return /a/                      0                0

B. In 1938 return /a/                 50,000           50,000

C. In 1939 return /a/                100,000          100,000

D. In 1940 return /a/                100,000          100 000

E. In 1941 (present) return           50,000           50,000
                                     -------          -------

F. Total, five years                 300,000          300,000

G. Divide Item F by 5                 60,000           60,000

H. Tax on Item G (average income)      2,400           10,960

       II. Comparison of (a) Tax Paid, and Tentatively Payable
            for 1940, With (b) Tax Due on Average Income

               1937    1938      1939       1940       1941    Total
I. Tax on annual
   income      0/a/  8,730/a/  30,600/a/  30,600/b/  9,700/c/  79,730

J. Tax on average
income         0/a/  3,070/a/   9,700/a/  14,350/a/  13,360/d/ 40,480
K. Excess of tax on annual income
(Item I minus Item J)                                          39,250

        III. Computation of Tax Relief -- To Be Made Only if
        Tax on Annual Income (1941 Item in Line I) Is Greater
          Than Tax on Average Income (1941 Item in Line J)

L. Ten percent of 1941 tax (in line I)                            970

M. Tax relief for 1941: Item K or Item L,
   whichever is smaller                                           970

                           IV. Tax Payable

N. Tax payable (Item I for 1940, minus Item M)                  8,730


     /a/ To be copied from 1940 return.

     /b/ Item M, 1940 return.

     /c/ Item 33 on present return form.

     /d/ Total of items in line H.

                           Return for 1942
               I. Computation of Tax on Average Income

                                    Normal tax
                                      Income           Surtax
                  Item                                 Income

A. In 1937 return /a/                      0                0

B. In 1938 return /a/                 50,000           50,000

C. In 1939 return /a/                100,000          100,000

D. In 1940 return /a/                100,000          100 000

E. In 1941 return /a/                 50,000           50,000

F. In 1942 (present) return                0                0
                                      ------          -------
G. Total, six years                  300,000          300,000

H. Divide Item G by 6                 50,000           50,000

I. Tax on Item H (average income)      2,000            7,700

    II. Comparison of (a) Tax Paid, and Tentatively Payable for
              1942, With (b) Tax Due on Average Income

        1937    1938     1939       1940      1941     1942    Total

J. Tax on annual
income  0/a/  8,730/a/  30,600/a/  30,600/a/  8,730/b/   0/c/  78,660

K. Tax on average
income  0/a/  3,070/a/  9,700/a/ 14,350/a/ 13,360/a/ 9,700/d/  50,180
L. Excess of tax on annual in-
   come (Item J minus Item K)                                  28,480

        III. Computation of Tax Relief -- To Be Made Only if
        Tax on Annual Income (1942 Item in Line J) Is Greater
          Than Tax on Average Income (1942 Item in Line K)

L. Ten percent of 1942 tax (in line J)                           ____

N. Tax relief for 1942: Item L or Item M,
   whichever is smaller                                          ____

                           IV. Tax Payable

O. Tax payable (Item J for 1942, minus Item N)                   ____


     /a/ To be copied from 1941 return.

     /b/ Item N, 1941 return.

     /c/ Item 33 on present return form.

     /d/ Total of items in line I.

In concluding this section on averaging, there is submitted herewith a memorandum from Mr. William Vickrey describing an averaging device which he developed as a result of being requested to analyze the plans described above and to attempt to formulate one that would accomplish the task more adequately than those do. Time limitations have prevented the present writers from commenting in detail on Mr. Vickrey's plan (Mr. Vickrey had available for this work only parts of the last few days of the study) but it is evidently an important contribution that deserves serious study at once. Its chief virtues seem to be that, by a recomputation, each year, of the taxes, past and present, on the present and all previous years' incomes, with allowance for interest on both tax and income, the taxpayer is practically prevented from avoiding or minimizing tax liability in any way by shifting income or loss from one year to another, and a far more complete averaging is obtained than by the other systems. If tax avoidance is really checked, the need for an undistributed profits tax or any special treatment of capital gains and losses (e.g., inventory plan) disappears. The behaviour of the annual tax liability under this plan is also such as to remove most of the necessity for reliefs and credits. The chief disadvantages seem to be that the Bureau of Internal Revenue will have to reorganize its procedure somewhat, that the taxpayer will have to use a rate schedule that he may not understand because it will be a synthesis of the rate schedules of the present and past years, and that some troublesome problems of inventorying assets at certain dividing lines in the taxpayer's life (e.g., marriage) may have to be solved.

Averaging of Income for Income Tax Purposes

William Vickrey

It has long been considered one of the principal defects of the graduated individual income tax that fluctuating incomes are on the whole subjected to much heavier rates of tax than incomes of comparable average magnitude that are relatively steady from year to year. Thus for example A, whose income is alternately $25,000 and $75,000, will pay a much heavier tax over a period of years than B who enjoys a steady $50,000 income. /1/ It is generally considered that irregularity of income is an inconvenience rather than otherwise, and that A has if anything a smaller ability to pay than B, rather than the greater ability to pay that might justify this heavier tax. /2/

Several attempts have been made to remedy this situation. In the United States, the only steps towards this goal have been the provisions for the carryover of net business losses and the various special privileges granted to taxpayers having income in the form of capital gains, such as the special 12 percent rate in effect from 1922 to 1933, the percentage discounts allowed from 1933 to 1937 according to the length of time for which the asset has been held, and the combination of discount and special rate now in effect. Although other reasons have been advanced, and possibly were chiefly responsible for the adoption of these special treatments, they were based, in part at least, on the theory that capital gains constituted a type of income that was much more subject to fluctuations than others. While this is true on the average, it does not necessarily follow for the individual taxpayer, nor is the amount of relief given under these provisions at all accurately proportioned to the hardship to the taxpayer involved. Thus one result of such provisions has been to open a considerable loophole for tax avoidance; taxpayers will tend to throw as much of their income as possible into the form of capital gains, and thus obtain the relief without necessarily having suffered from fluctuation of their total income. Other income may fluctuate, of course, as well as that arising from capital transactions. No relief is provided by this method for such cases. The relief is of course to a considerable extent a slight concession to those who maintain that capital gains are not income and should not be subject to the income tax.

A more straightforward method of dealing with the problem is by the use of some sort of average income as a basis for the assessment of the tax. In Great Britain, such averages have had a long history. A three-year average basis of assessment for certain types of income was introduced with the first income tax during the Napoleonic wars, retained in the income tax of 1843, and continued with various modifications from then until 1926, when the average basis of assessment was abolished. The chief reasons for abolishing the average basis of assessment seem to have been its complexity (superimposed on what was already a fairly complex method of administration) and the objections of the taxpayers in years of reduced income to paying a heavy tax based on the larger incomes of previous years. The principal obstacles to the abolition of the average basis of assessment seem to have been the difficulties of transition, and the fear of the Inland Revenue that evasion might become easier and revenues less stable and predictable, rather than any advantage that the average basis might have had in providing equity between taxpayers with steady and with fluctuating incomes. The amount of relief provided by the average basis to taxpayers with fluctuating incomes was almost negligible; /3/ the British experience can thus hardly be cited as an example of the failure of averaging as a method of alleviating the hardships of taxpayers with fluctuating incomes. /4/

The state of Wisconsin enacted in 1927 an averaging provision that was specifically directed towards dealing with the hardships imposed by the previous law on taxpayers with fluctuating incomes. It provided that the tax should be equal to the tax on an average of the income of the three years previous to the year in which the tax was paid. /5/ The law was repealed, however, in 1931. The repeal of the averaging provision may be laid chiefly to the fact that with the depression setting in and incomes thereby greatly reduced, taxpayers were finding that it was a considerable hardship to be compelled to pay income tax based in part on the high incomes of previous years. /6/ Another serious drawback was the fact that the state had encountered constitutional difficulties in trying to collect tax on the untaxed portions of income in the case of estates of persons who died with substantial amounts of income left "untaxed," and similarly in the case of corporations dissolving, persons leaving the state, and so on. (Fitch vs. Wisconsin Tax Commission, 201 Wis. 383) The serious loopholes thus left open, and the popular discontent with the lag in payments compelled the legislature to abandon this promising experiment. /7/

In 1921, the Commonwealth of Australia enacted an averaging provision that apparently is the only comprehensive averaging provision to have survived for any length of time. /8/ The Act is unique in that while the rate of tax is determined with reference to an average of five years' income, this rate of tax is applied to the income of the current year rather than to the average income. This type of provision seems roughly to satisfy over short periods of time the requirements for equalizing taxation as between the taxpayer with a fluctuating income and the one whose income is relatively steady. The problem of the taxpayer with reduced income is much less troublesome here than in the Wisconsin scheme, for while he may have to pay at a high rate because of the incomes of previous years, this rate applies only to the relatively low current income. The transition problem is much easier also, as there is no question of taxing any year's income more than once, or of letting any income go untaxed, either as the scheme is abandoned, or as the taxpayer leaves the jurisdiction. On the whole, the plan seems fairly satisfactory, except for some capricious cases that may arise when large losses are followed by large gains. If the rates of tax are not constant, but change from year to year according to the needs of the government for revenue, there is a further capricious influence in that the individual whose income happens to be especially large in the years of heavy taxation will pay more taxes than the individual who, with average income of the same magnitude, manages to have his income lower than usual in the years in which the rates are high. /9/

None of the foregoing averaging methods achieves more than a very rough equality of treatment as between the taxpayer with a steady income and the taxpayer whose income fluctuates in various ways. The ideal averaging plan would provide that no taxpayer would bear a heavier or lighter burden merely because his income happened to be earned or realized in one year rather than another, whether by chance or by design of the taxpayer, and this without regard to the fluctuations in the needs of the government for revenue or to the rates of tax in effect at various times. If a practicable method of averaging that satisfies this criterion can be developed and put into effect, then conversely it will be impossible for the taxpayer to avoid tax in the long run by changing the shape of his income stream. /10/ As a consequence, many of the arbitrary, unpopular, and complicated provisions designed chiefly to prevent the manipulation by the taxpayer in his own favor of the shape of his income stream will no longer be necessary and can be discarded. /11/

The general requirements for the ideal averaging device may then be summarized as follows (in approximate order of importance for the present discussion):

1. The discounted value of the series of tax payments made by any taxpayer should be independent of the way in which his income is allocated to the various income years.

2. The revenue for any given year should be capable of being raised or lowered by suitable modification of the rates without too long notice.

3. If the taxpayer leaves the jurisdiction at any time, there should be no accumulation of untaxed income left behind, and no tax due except possibly the regular income tax for the last year. Conversely, there should be no necessity for tax refunds by the government.

4. Any given tax payment should not be too large in relation to the income of the last year.

5. Transition to and from other methods of levying income tax should be simple.

6. The method of computing the tax should not be beyond the ordinary taxpayer's capacity.

The first step is to obtain a basis for the tax that will satisfy the first criterion. Consider two taxpayers A and B starting with equal capital, having the same earned income during a given period, and spending the same amount, the only difference being that A pays taxes on this income during the period whereas B manages by one method or another to postpone his taxes until the end of the period. A's total income for the period (before taxes) will be less than B's by an amount equal to what the taxes that A paid would have been able to earn if invested at compound interest up to the end of the period. If then to A's total income we add the compound interest on the taxes A paid during the period from the time they were paid to the end of the period, an "adjusted total income" is obtained which is the income A would have had had he paid no taxes during the period. It can be shown that this adjusted total income will remain the same for any given taxpayer regardless of any changes that may occur either by chance or by design in the allocation of the realization of the income to this various income years. /12/ To satisfy the first criterion it is only necessary to make sure that all persons whose adjusted total income is the same (over the same period of years) make tax payments with respect to this income of the same aggregate present value. A standard for the magnitude of these payments may be set up by providing that an individual with a steady income shall pay the same present value of tax as he would have done without the application of the averaging provisions. A simple method of satisfying criterion 2 is to treat the taxpayer each year as though that year were the last of an averaging period. This treatment will avoid any questions of unpaid and uncollectible taxes, will cause the taxes to keep step fairly well with the income of the taxpayer, and will permit the revenue in any one year to be varied by suitable changes in the basic surtax rates.

The principles involved in the computation of the tax for each year will then be as follows: The adjusted total income of the taxpayer will be calculated by adding to his total income for the period the compound interest on the taxes which he has paid with respect to this income. The size of the constant income that would have yielded the same adjusted total income over this period is then calculated. The next step is to calculate the present value of the taxes that would be payable on such an income according to the present methods of assessment, including the tax for the current year. The present value of the taxes already paid by the original taxpayer would then be deducted from this sum and the remainder will be the tax currently payable.

At first sight, this method of determining the annual payments to be made by the taxpayer seems hopelessly complex; it is possible, however, to arrange the computation so that the actual work required to be performed by the taxpayer will be considerably less than that involved in the computations at present required in the case of taxpayers having capital gains and losses, computations the necessity for which would be eliminated by the use of the averaging method. In this arrangement, the Treasury would prepare special surtax tables, comparable in every respect to those now in use, which would give the total tax payable on given amounts of adjusted total income, with marginal rates to be applied to income between the bracket limits given in the table. There would be one such table prepared for each number of years during which individual taxpayers had been subject to the averaging method. From a previous return or certified transcript of some sort the taxpayer would copy the total adjusted income and the total value of payments as of last year, compute the interest on the tax, add this interest and the income of the current year to the old adjusted total income to get a new adjusted total income, and add interest to the old value of taxes paid to get a new value of taxes paid. He then enters the table for the appropriate number of years with the total adjusted income and obtains a total tax in exactly the same manner that the present surtax is obtained from the table for a given surtax net income; from this total tax the taxpayer then deducts the present value of past tax payments to obtain the payment due for the current year. /13/ As a computation this should compare very favorably with many of the more recent proposals for special treatment for capital gains and losses (most of which are far from completely satisfactory), and even with the computations required under the formulas of the Australian system, with separate treatment of earned and unearned income.

The chief drawback seems to be that there must be a separate table made up for each number of years for which taxpayers are permitted to average their income; thus after 15 years of operation, 15 separate tables would need to be drawn up, and while each individual taxpayer would need to consult only one table, the burden of selecting the proper table will either have to be placed upon the taxpayer, or the Treasury will have to undertake to mail each taxpayer the proper table on the basis of the records of previous returns. This should not prove an insuperable obstacle.

Now it is possible to show mathematically that the amounts of tax successively due under such a method of computation and payment will under very general conditions not bear too high a relationship to the income of the last year. Several limits may be derived under various conditions. If the current income is larger than the average income, the residue left after taxation will be greater than the residue from an income equal to the average income under a straight annual assessment. If the current income is less than the average income, and the rates are constant through time, then the tax payable will be less than the tax payable on an annual basis of assessment. If rates are changed, but not so drastically that the average rate in the current basic schedule is greater than the mean marginal rate has been in the past, then the average effective rate of tax will be less than this mean marginal rate for past years. If current income is greater than average income and the rates have been raised so that the average current basic rate is greater than the mean marginal rate, then the average rate of tax on the entire income will be less than the average rate in the current schedule on an income equal to the average income. /14/ It may be seen from the above conditions that the payment due in any given year will bear an unreasonable relationship to the income of that year only in the case of a drastic rise in the general scale of rates accompanied by a sharp decrease in the income of the individual taxpayer. Thus no provision for relief of hard cases need be made except in years in which a sharp increase in rates has been made. Even in the years in which such relief is necessary, the form of the relief may be fairly simple without opening any very serious loopholes, since the number of taxpayers who will require the relief will be very small. The relief might take the form of a provision that the tax payable in any one year shall not exceed an amount determined by applying a supplementary rate schedule, somewhat higher than the regular one, to the income of the taxpayer for the previous year. The reduced tax payment may be carried over as a basis for calculating subsequent payments, so that in most cases the government will not in the long run lose any revenue through the application of the relief except in very extreme cases.

It is apparent that such an averaging device can prevent the avoidance of tax by the shifting of income only with reference to shifts of income within the averaging period. It if is permitted to shift income between years included in an averaging period and previous or subsequent years whether or not they happen to be included in separate averaging periods, the possibility of avoidance will reemerge. It will therefore be necessary to reintroduce at the close or commencement of each averaging period such safeguards as may be available to prevent such shifting of income between one averaging period and another.

The simplest method of preventing such shifting of income between averaging periods is to require that an inventory be made at the end of each averaging period and that the capital gains so revealed be included in the income of the last year of the averaging period. (This is equivalent to requiring that G(0) = 0 in the analysis of footnote 12.) While this procedure might involve a prohibitive amount of administrative work if valuation each year were required, as is the case when assessment on a strict accrual basis is proposed, only a small fraction of this work would be required under an averaging scheme, since the valuations would be made only at relatively long intervals.

There are many reasons why the average period should be made as long as possible. Obviously, the longer the averaging period the smaller will be the administrative task of valuation and checking valuations. If the averaging periods are arranged so that their ends are staggered, then the effect of the valuation date on the markets may be reduced by lengthening the averaging period since the averaging period amount to be valued at any one time will be smaller. The incentive for the taxpayer to attempt to shift his income from one period to another will be smaller the longer the averaging period, since it will be more difficult for him to forecast for the longer period what the size of his income will be and to what rates he will be subject; moreover, the actual variations in average income and average rates as between one period and another are likely to be smaller. In general, the increase in equity afforded by the operation of the averaging method will become greater as the period is made longer.

The logical limit would seem to be to extend the averaging period from the majority of the taxpayer until his death (although it would be possible to start the averaging period at birth, the difficulties involved seem to outweigh any possible advantages especially as such a procedure would tend to favor those who had taxable incomes during their minority). If this plan is adopted, then only two valuations become necessary; one of these valuations, that at death, is in most cases already required for estate tax purposes, and the other valuation, a valuation at the taxpayer's majority, would usually involve only a very small amount of property, and would be a valuation where in general it would be to the taxpayer's advantage to report as much property as possible, so that enforcement would be relatively simple. That capital gains accrued upon a taxpayer's property at the time of his death should be taken into account in the return for the last year has already been proposed as an independent reform, although doubts have been expressed as to the constitutionality of such an assessment. If the direct imposition of such a method of assessment does prove unconstitutional, there is the possible alternative device of offering the averaging plan coupled with the voluntary acceptance by the taxpayer of such an assessment as an alternative to being taxed on an annual basis; since in most cases the averaging method would under identical rate schedules result in a reduction of the tax burden, most taxpayers might accept such a proposal. The rates for the annual basis of assessment might be made higher than the other rates, or particular attention might be paid to the auditing of the returns of taxpayers not electing the average basis. Another method of inducing the taxpayer to accept such an assessment voluntarily might be to impose a special estate tax on the transfer of assets containing such unrealized gains. The tax on such unrealized gains might indeed avoid the constitutional issue by being formulated as such an estate tax graduated by reference to the taxpayer's average income.

Extending the averaging period from the taxpayer's majority to his death automatically provides for a staggering of the averaging periods of different taxpayers so as to reduce the influence of the necessary valuations on markets to a minimum. This procedure however raises the problem of how to treat taxpayers whose family status changes. One solution is to arbitrarily out the averaging period at the time of marriage, divorce or death of spouse. Such a procedure, however, imposes a fairly heavy tax burden on marriage, since the individual who marries will not be able to average his previous low income, with his subsequent presumably higher income (or in less frequent cases vice versa), and will therefore have to pay heavier taxes than the man who remains single and is able to average over the longer period. This may altogether outweigh the concessions given the family man in increased personal exemptions, and if it does so, will run directly counter to most accepted notions of ability to pay. Another method of dealing with the problem would require separate returns to be filed and the tax computed and paid separately by each member of the family. A more radical but probably more satisfactory method in the long run would be to go one step farther than the community property states and consolidate the total family income in one return, apportion this family income among the various members of the family according to proportions fixed by statute, and compute a tax separately for each member of the family using for each person the appropriate previous total adjusted income. This method had the advantage that it is likely to prevent to a very large extent the avoidance of tax by various methods of redistributing income between members of the same family. Moreover it would eliminate the arbitrary and unjustified advantage now enjoyed by residents of community property states. On the whole, it seems a rather more equitable method of taxation than the one at present in effect, even if not used in conjunction with any averaging method. In order that it should not be thought of as merely a method of increasing the relief given the wealthy on account of marriage, this change in method of assessing family incomes should be accompanied by a decrease of about 50 percent in the income levels at which the various existing rates of tax become effective, so that the actual change in the tax burden of the married will be relatively slight, with a substantial increase in the tax burdens of those individuals with large incomes who have no family with which they share this income. /15/

In the case of corporations, the opportunities for avoiding taxation by shifting income from year to year are more limited than in the case of individuals, since the corporation income tax rates are but slightly graduated and the rates themselves seem to have been, at least in the immediate past, rather more stable from year to year than the rates of the individual income tax. Nevertheless the application of the present averaging method, or a modification of it, may be advantageous even here, since by so doing the treasury would be freed from the necessity of checking inventory, depreciation, obsolescence, and the like, except to make sure of arithmetical accuracy and the absence of double counting. The question is not so important as with the individual income tax, however, as the need for the undistributed profits tax and the special capital gains provisions would not be affected.

A more important question is how far the averaging device should be extended to individuals with lower incomes. A rigorously thoroughgoing application of the first criterion would result in every adult filing a return no matter how small his income, and permitting individuals to cumulate as a deficit, to be offset against future income, any excess of exemptions and allowable deductions over gross income. This procedure is open to two very serious objections. The administrative job of auditing this vast number of returns would be staggering, especially as it would be necessary to check even those returns that were obviously not taxable, since if the taxpayer later has a large income, the amount of deficit reported in previous returns will affect his tax in such a year. The statistics might be interesting, but probably rather expensive. Another serious drawback is that Congress might some day find that having in the past been somewhat more generous with exemptions and deductions than they will want to be at that time, the taxpayers will nevertheless have accumulated a backlog of deficits against which to offset income so that no matter how far the exemptions are then reduced, very little income from the lower levels can be brought into the taxable brackets. Thus the result could conceivably be considerable financial embarrassment on the part of the government, or a breach of faith with the taxpayer through the abrogation of the right to set off these accumulated deficits against current income. A simple way of getting around these difficulties is to allow personal exemptions to be deducted only to the extent of net income. This would restrict the carrying forward of negative income to cases where business losses, capital losses, and other deductions exceed net income; in such cases a slight penalty would attach in that the benefit of the exemption for that year would be lost. Thus in the bulk of the cases there would be no incentive, under such limitations, for the filing of nontaxable returns in the expectation of future increased income.

In connection with a flat rate tax at a fairly low rate, such as the normal tax, the application of the averaging method under the above restrictions would make a slight but on the whole insignificant difference, provided only that full carryover of losses is permitted. It may therefore be quite sufficient, at least at first, to apply the averaging method to the computation of surtax only, continuing to calculate the normal tax on an annual basis as heretofore. This plan would cut the initial administrative load down very sharply, and after experience has been gained, the plan might be extended to cover all taxpayers if this seems desirable on the basis of such experience.

It is to be noted that the method of assessment here described is in no sense a moving average or mere smoothing device, as were the other plans reviewed at the outset. It is rather a method of making it possible to consider the income of several years as a unit in assessing income tax, and of making due provision for suitable installments of tax as the income accumulates. It would be possible to modify the method of assessment to work as a moving average, but if this were done the computations would become more complicated while the property of making the tax burden independent of the manner in which the income is shifted from year to year would be lost, although the profitability of such shifting might be greatly curtailed. (Unlike such moving average plans as that of Wisconsin discussed above, no particular problems arise at the time of inauguration or abandonment of an averaging plan such as outlined above or at the beginning or end of the averaging period of any individual taxpayer, other than those problems of allocation of income with respect to accounting period that arise with even greater magnitude in any case under a straight annual basis of taxation.)

Any averaging device will of course require a certain amount of recordkeeping. This is the price that must be paid for the increase in equity of the income tax generally and for ease of administration in other directions. It is quite possible that the decrease in other administrative work may more than make up for the keeping of more complete records. There are also other minor points, such as the problem of dealing with part- year returns, and with aliens who in some years draw their income from foreign sources and may so obtain unfair advantages under an averaging system that considers only a part of their income; the problem of changes in residence status would be particularly important if states wish to adopt some such averaging device. These difficulties and the many possible methods of dealing with them cannot be discussed at length here.

The adoption of such a method of assessment by states and other local subdivisions is somewhat less advantageous than in the case of the central government, as the rates are in general lower and the graduation less steep. At present most taxpayers govern their actions chiefly with reference to the federal income tax so that the decrease in avoidance is likely to be slight. However, if the federal government should employ such an average basis of assessment, taxpayers might be more influenced in their actions by state taxes, and under such circumstances the effect of the subsequent adoption of the average basis by the state in checking avoidance will be correspondingly greater. The chief gain, in the case of a state tax, is likely to be in equity, at least at first, and while this gain is not to be despised, it seems usually less sought after than the checking of avoidance. The method would probably be inapplicable to nonresidents, at least not without considerable complication.

Such a method of assessing income tax is thus not without limitations and disadvantages. The keeping of records, the slightly more complicated method of computing the tax, the existence of several surtax tables among which the correct one must be selected, the required final valuation at death, the special treatment required for part year and other special types of returns, the occasional need for payment of refunds, (in addition to refunds of overpayments resulting from error), and the more detailed treatment of family returns are the chief points at which objections may be raised. Against these minor drawbacks are to be set substantial gains in equity as between taxpayers with steady and those with fluctuating incomes, taxpayers with fixed and with manipulatable incomes, and taxpayers with capital gains and other forms of income, and between single and family taxpayers.

The fact that similarly circumstanced taxpayers are treated similarly produces in turn still further desirable results, such as a substantial decrease in the worry, expense, and economic waste that now results when taxpayers seek to minimize their tax burden, reductions in the amount of litigation, and a decrease in the influence of the income tax on business transactions and the economic life of the community. For example, the securities market should be freer from the extraneous influence of arbitrary rules concerning capital gains, while the influence of the income tax in reducing the amount of capital available for risky enterprises should be diminished, as abnormal profits in one or two years will no longer push income up to such high brackets.

The undistributed profits tax, as well as the surtaxes on personal holding companies and corporations improperly accumulating surplus, could be repealed completely without fear of reopening loopholes for avoidance. The special provisions for the taxation of capital gains could also be repealed, as such gains may be included in net income without discount without imposing any special hardship on the recipients of such gains, while the limitations on the deduction of capital losses could be removed without thereby opening any loophole for avoidance. /16/ The removal of the incentive to shift income from depression years when rates are high to prosperous years when rates are low should increase the cyclical stability of yield and in return reduce the pressure to limit the deductibility of losses in time of depression, as was done in 1932.

The method of assessment outlined above has been developed on the assumption that the base to be adopted for income taxation is "accrued income" as opposed to the expenditure base proposed by Dr. Irving Fisher. It is possible to adapt the method of assessment, with a few slight changes, to operate on the latter base; however, the advantages of an average basis of assessment in the case of an expenditure tax are much less striking than in the case of the accrued income tax, chiefly because the expenditure tax is inherently less difficult of administration. The possibility of the simplification of administration in the one case, and the equalization of burden in both cases resulting from such an average basis of assessment should make it easier to discuss the relative merits of these two bases for taxation on the basis of their economic and social effects, without having the issue confused by the consideration of relative ease of administration and relative degree of discrimination against fluctuating income.

Problems Common to the Personal Income Tax
and the Corporation Income Tax

Certain problems are common to the personal income tax and the corporation income tax. What appear to be the most pressing of these are discussed in this section. The problems covered are those of:

1) Capital gains and losses

2) Depletion allowance

3) Tax-exempt securities

Actually, this section consists almost entirely of an analysis of the first problem -- capital gains and losses. This is by far the most urgent, and the time spent on it allowed only a few brief paragraphs for the other two. Another problem, carryover of losses, might logically be treated here, but it has instead been mentioned in the previous section on personal income tax problems in connection with the averaging methods, which are peculiarly applicable to the personal tax problem, though not without significance for corporation taxation.

Capital Gains
Summary of Chief Plans

The major kinds of treatment that may be given capital gains and losses, with their chief advantages and disadvantages, are as follows:


This method lightens the administrative burden. It probably results in a loss of revenue compared to most of the other treatments. More serious, it is not a neutral policy as concerns economic effects on the markets -- it merely substitutes one set of effects for another, since taxpayers will try to throw as much of their income as possible into the form of capital gains, in order to lessen tax liability. This would of course also be a serious loophole from the point of view of equity. Still more serious, perhaps, it results in an unjust distribution of the tax burden in that it ignores important factors influencing one's ability to support government.


This method is relatively simple, administratively, and probably results in a net gain in revenue, but it subjects taxpayers to unfair burdens when they realize a net gain that is large because it has accrued over a number of years. In some cases the added tax would be very heavy indeed, and would have an appreciable effect on decisions to hold or sell capital assets. Moreover, full benefit is not given to a net loss accrued over a long period, since the loss is taken out of lower surtax rates, on the average, when it is lumped in one year than it would be if spread over the years of accrual. This fault might be remedied somewhat for some taxpayers by limiting gross loss deductions to gross gains but allowing an indefinite carryover of unabsorbed losses.


Under this plan the taxpayer could report any accrued gain that he wanted to, the gain on realization being correspondingly lessened; but he would not be allowed to report any accrued loss, and in other respects likewise the gain and loss would be treated as other income. This method has been in use in a canton in Switzerland. /1/ It differs from No. 2 above in giving the taxpayer a chance to even out his gains, if he can forecast well enough, and thus escape unjustly high surtax rates. Whether taxpayers would in fact use the device to any marked extent seems very doubtful. The chief value of this plan is as a supplement to some other plan, for such an asset as real estate.


This method is the basis of the existing federal system of handling gains and losses. The full possibilities of that method are not, however, shown by this system, which restricts capital loss deductions to capital gains of the same year (plus $2,000). Moreover, it runs the loss percentage down, like the gain percentages, as the period held increases, whereas the loss percentages should go up when there is a net realized loss. /2/ The best possible single step-scale plan would have administrative advantages over most other plans. It would probably increase revenue appreciably compared to the method that would ignore gains and losses altogether. But it, too, is not a neutral policy as concerns economic effects on the markets -- the taxpayer has the option to realize his gains and losses as he will, and he will in many cases, presumably, buy and sell at times and prices different from those he otherwise would select, in order to throw gains into low-income years or into the long-period low- percentage sections of the step-scale. A converse, but probably much less forceful pressure, is exerted with respect to losses. Finally, a single-step scale plan can never be even approximately equitable for all taxpayers. If the gain percentages go down rapidly, in order not to overburden the taxpayer whose top income falls in a steeply progressive part of the rate scale, /3/ unjustifiable tax relief is given to the taxpayer whose top income is taxed in a mildly progressive part of the rate scale. Moreover, as a practical matter, it is not likely that Congress will allow a loss on an asset held over five years, for instance, to be deducted at, say, 140 percent, while a loss on an asset held only one year is deducted at 100 percent -- yet when the net results of the years' realization is a loss, some such adjustment would be necessary if patent discrimination among taxpayers was to be avoided.


Administratively, this plan is somewhat more difficult than a single step-scale plan, particularly in the lengthy schedule that the taxpayer must fill in if he realizes gains and losses in the same year. This schedule is not in reality very difficult if the taxpayer follows instructions closely, but it is formidable at first sight, is particularly open to misrepresentation, and is apt to leave the taxpayer in a state of confusion even after he has followed instructions and obtained the correct answer. The plan probably increases revenue considerably compared to the method that would ignore gains and losses altogether. Again, however, there is no neutrality as concerns economic effects on the markets. Presumably the inducement to "time" realization is not so great in many instances as it is under a step-scale plan, but the shrewd taxpayer will still try to throw his gains into low-income years. Indeed, in certain hypothetical cases it will actually pay a taxpayer to fail to obtain some "other income" -- he will have more money left, after tax, if he gets a cut in salary, for instance, in a year when he realizes a large gain than if he does not get the cut (cases where the taxpayer did reduce his other income in order to lessen tax might be rare in practice, but the mere possibility makes a forceful argument in the hands of opponents of capital gains and loss taxation. /5/ This defect is not found in the step-scale plan. As to avoiding evident discrimination among taxpayers, the apportionment plan is in its general structure better than any single step-scale plan could be (particularly in giving adequate allowance to losses, if no time limit is set to the carryover of net capital losses) but it permits of odd results in many cases, chiefly because it assumes that "other income" is stable over a period of years and that capital gains (or capital losses) realized in different years do not overlap in their accruing. /6/


Under this multi-step-scale method, the tax form would contain a table giving one step-scale plan for taxpayers whose "other income" was less than $5,000, for example; another for taxpayers in the $5,000 - $6,000 class, etc. Several variants of this plan are feasible, but this brief description gives the general idea. The administration would be somewhat more difficult than under the single step-scale plan, and the taxpayer, particularly, might be somewhat puzzled at first by the table, although perhaps not as much as by the Division's apportionment-plan schedule. The policy, like all the others above, is not a neutral one as concerns effects on the market, though the taxpayer would have more difficulty than under the single step-scale plan in ascertaining how much tax he would save by delaying or hastening realization. There is a nearer approach to equity among taxpayers, probably, compared with the single step-scale plan, but the vexing question of rising percentages for net losses remains. Moreover, unless the table is to be so detailed as to be unwieldy, there will be sudden jumps in tax from one income level to another (as in the illustration below, from just under the $20,000 other-income level to just over) so large as to cause some criticism. Consider, for example, a taxpayer with ordinary income of $21,990 and capital gains on assets, held two years, of $2,000. These gains would, under the percentages shown in Table 1, be taxed at 94.12 percent of their value, and the total tax would be $1,878.31, leaving $22,111.69 net income after taxes. If income increased to $22,005, the gain would be assessed at 99.04 percent and the tax increased to $1,897.59, leaving only $22,107.41. Thus a loss of $4.28 results from the $15 increase of income. The percentages actually decrease with increasing income in some portions of the table, so that the opposite anomaly may occur: i.e., a slight increase in net income may result in a decrease in the tax paid. Finally, no one set of answers will give the same degree of justice for different sized gains and losses. Table 1 [see next page] shows what a table such as this would look like. /7/

This particular table is built up on the assumption that the average gain accrued over each year equals 10 percent of other income. If instead it equals 30 percent, the percentages in the table should be as shown by the figures in parentheses (not computed for all the cases). The difference, it can be seen, is enough to make any one table of this kind capricious in its effects.


Under this method, the one-table plan would apply only to gains or losses where the asset has been held for more than two years (other gains and losses to be treated as ordinary income): otherwise the length of time the asset was held would be ignored; in place of length time, the size of the capital gain or loss would be considered. The chief question in choosing this plan rather than No. 6 above, is whether the length of time held or the size of the gain or loss (in its relation to the degree of progression of the rate scale) is the more important factor from the point of view of equity, in practice. Otherwise, the two plans have much the same effects. The answer to the equity question seems to be a close one, with much to be said on both sides. The kind of table used in this plan is shown in Table 2. The answers are here given in terms of tax rate applicable to gains (or tax relief for losses), but they could instead be like Table 1 and be in terms proportion of gain or loss to include in other income. This table suffers from a defect analogous to that explained for Table 1 -- as the third, unaccounted-for variable (in this case, length of time held) changes, the percentage answers in the table should change. Also, this table suffers from the defect of sudden jumps in the tax, as explained in the footnote to the table -- a defect also noted in the discussion of Table 1, but more serious here. This defect may be overcome by the kind of computation shown in Table 3, but the added complexity is rather formidable.

                               Table 1
Other Income     Series of Step-Scales for Gains, According to Size
(in thousands    of Other Income, /1/ Showing Percent of Gain or Loss
of dollars)         to Be Taken Into Account. (Sample Computations)
                          Time Held (years)
             1        2       3       4       5        7         10

  0 - 5     100

  5 - 6     100

  6 - 7     100

  7 - 8     100

  - - -    - - -

 20 - 22    100     94.12   92.16   89-47   87.37    84.96     80.95
                   (92.98)                 (80.99)            (67.14)

 22 - 26    100     99.04   95.85                              88.23

 26 - 32    100

 50 - 56    100     95.45   91.79   88.72   85.49    80.55     74.90
                   (89.79)                 (74.37)            (66.67)

200 - 250   100      100    99.48           98.44              96.21
                   (98.23)                 (95.10)            (92.13)

     /1/ The figures not in parentheses assume a gain per year equal
to 10 percent of other income. The figures in parentheses (not as
many computed, owing to lack of time) assume the gain per year is 50
percent of other income. In practice, some one assumption as to size
of gains would have to be made; the differences between the figures
in and out of parentheses give some idea how important that
assumption would be. The figures are here computed for the lower
limit of the bracket, but would ordinarily be computed to apply to
the entire bracket. Some of the figures are capricious, especially
for 2- and 3-year periods and small gains. These might be smoothed.

                               Table 2
         Table for Determining Rate to be Applied to Capital
                     Gains (Sample Computations)

        (Assuming all gains over two years to be four years)
Income other than                         Capital Gains
capital gains                      (in thousands of dollars)
(in thousands of
dollars)                  0 - 4     4 - 8      50 - 60       60 - 70
     Up to 0

      0 - 5

      5 - 6

      6 - 7

      - - -

      - - -

      - - -

     ?? - 20                                   16.76%        17.27%

     ?? - 22                                   17.80%        18.33%

Note that if ordinary income is $19,900 and capital gains $59,900,
tax is $11,188.24, leaving $68,611.76, but if gains increase to
$60,100, tax becomes $11,628.27, leaving only $68,371.76, or less
than if taxpayer had smaller gains. A similar anomaly occurs when
ordinary income passes from one bracket to the next. (The figures in
the above table are based upon the lower limit of the brackets).

                               Table 3
          Segment of Table 2, Refined To Avoid Jumps in Tax
Ordinary in-           Capital Gains (In thousands of dollars)
come (in
thousands of
                         50             50-60               60
     18                $8,380           19.80%           $10,360

  18 - 20               26.0%            0.6%             32.0%

    20                 $8,900           21.00%           $11,000

For an ordinary income of $18,500 and capital gains of $54,000, tax
is computed as follows:

(1)  Block of tax on $50,000 of gains with
     $18,000 other income. . . . . . . . . . . . . . . . . .$8,380.00

(2)  Excess income over $18,000 = $500

(3)  (2) X 26.0 percent. . . . . . . . . . . . . . . . . . .$  130.00

(4)  Excess gains over $50,000 = $4,000

(5)  (4) X 19.80%. . . . . . . . . . . . . . . . . . . . . .$  792.00

(6)  Excess in (2) X excess in (4) 1,000 = $2,000

(7)  (6) X 0.6%. . . . . . . . . . . . . . . . . . . . . . .$   12.00
Total tax on gains. . . . . . . . . . . . . . . . . . . . . $9,314.00

     This table may be modified to give the total tax, inclusive of
the tax on ordinary income, directly without requiring a separate
surtax table computation. The table then becomes:

       18           50          50-60            60

     18-20       $ 9,380        19.80%        $11,360

       20          39.0%         0.6%           45.0%

       20        $10,160        21.00%        $12,260

It will be noted that in some parts of the table percentages are high and the taxpayer may object to the apparent heavy rates more than if they were concealed.


The existing plan uses one variable -- the length of time the asset is held. /8/ That is, the portion of the gain or loss taken into account depends solely on that one variable. The plans in the immediately preceding paragraphs (Nos. 6 and 7) use two variables -- in No. 6, length of time and size of other income (or, what amounts to the same thing, range of rate brackets within which other income falls), and in No. 7, size of other income and (since this also helps determine what range of rates the gain or loss falls in) absolute size of gain or loss. The plan in this paragraph uses three variables -- length of time held, size of other income, and the absolute size of the gain or loss. Here, two tables are necessary. One table has, as column and row headings, two of the variables, and the figure located in the table by reading across and down from these variables forms the column heading in a second table which has as row heading the third variable -- and reading across and down again gives the answer, that is, the proportion of gain or loss to be taken into account (or the answer might instead be in terms of rate of tax or tax relief for gains or losses). /9/ Administratively, this plan is somewhat worse than either of the one-table plans (Nos. 6 and 7 above) but perhaps still not so troublesome as the schedule of the Division's apportionment plan (No. 5 above), though the two tables would take up considerable room on the return form. The economic effects remain, diminished a little more, perhaps, by the added difficulty of forecasting the tax results of any plan of action. As to equity, the plan comes fairly close to the apportionment plan, except that in practice the loss percentages might not be set at reasonable levels (see Plan No. 4 above), and the problem of sudden jumps (see Plans No. 6 and 7 above) remains.


This plan was submitted by Dr. Robert H. Haig in September 1937, upon the writer's request for a plan that would follow the general philosophy of the Division's apportionment plan, but that would be very simple for the taxpayer to compute. It adds each gain or loss realized, to get a net grand total. It then divides each gain or loss by the number of years held and adds the results, getting a total net gain or loss attributable to the current year. This is added to ordinary income and taxed as such. This net gain or loss attributable to the current year is then subtracted from the net grand total, obtained above, and the balance is attributable to previous years. To this balance is applied the top rate (normal plus surtax). This plan is much simpler than the Division's apportionment plan (No. 5 above) -- almost as simple as the existing step-scale plan, if not more so - - and might have about the same revenue results as the Division's apportionment plan, except that it gives much more importance to short-term losses in the fixing of the total tax liability. Under the Division's plan, a short-period loss affects the tax rate on only the short-period part of a long-period gain, whereas the modified plan allows a short-period loss to affect the rate on all of the long- period gain. Thus if a loss is on a two-year asset and a gain is on a ten-year asset, the Division's plan allows the loss to drag down the rate of the tax only on one-fifth of the gain, while the modified plan allows the loss to drag down the rate of tax on the whole gain. In somewhat extreme cases, the taxpayer may, under this modified plan, have more left over, after tax, if he can somehow create a short-term capital loss than if he cannot. For example, a man with a $50,000 ordinary income and a $300,000 gain over a ten-year period would pay a tax of $92,260 in the year of realization. If he could create a loss of $30,000 on a two-year asset, and realize it in the same year as the gain was realized, the tax would be only $60,400 -- a gain in tax of $31,860 compared with the capital loss of $30,000. (The difference is of course much greater in more extreme cases.) This kind of case may be of slight practical importance, since losses -- even short-term losses -- cannot be created at will. Rather, the importance of this extreme case is that it illustrates strikingly the tendency of this plan to give great importance to short-term losses - - a feature not present in the Division's apportionment plan. Of course one answer is that on the average short-term losses are bound to be small losses, and long-term losses, large losses -- indeed, this is the basic assumption underlying all the apportionment plans. If this were always so, the defect just noted (heavy weighting of short-term losses) would not be important -- but it will not be always so, especially because the plan itself will create an inducement for it not to be.

This modified plan has the same defect as the Division's plan with respect to the size of other income, and the possibility of a net saving by a decrease in such income.

In other respects, this modified plan seems to have about the same results in justice as the apportionment plan. That is, it is somewhat of an improvement over a single step-scale plan, possibly also over a multi-step-scale plan (one-table plan in No. 6 above), but, like all apportionment plans, still containing many imperfections.


This plan calls for apportionment by dividing by number of years held, as in the Division's apportionment plan (No. 5 above), but reopening previous years' returns and adding the yearly segments of gain or loss to the actual other income of the years concerned, instead of relying wholly on the other income of year of realization. Administratively, this plan has almost insuperable objections from both taxpayer's and official's points of view because of the necessity of reopening returns. This is unfortunate, since otherwise the plan is fairly good. It is not entirely neutral economically but much more nearly so than the other apportionment plans above. In general, it might be said to be a satisfactory solution of the economic-influence problem. From the point of view of equity, too, it is much superior to those other apportionment plans. It takes account of gains that are overlapping as they accrue, and of changes in tax rates and other income from year to year. A modification of the plan that might make it administratively acceptable without sacrificing too much equity is to reopen returns for, say, five years, but for any longer period use one of the other apportionment plans for the excess over five years. This still seems too difficult to the writers, however, from an administrative point of view.


If an asset has been held ten years, for instance, the gain will, under this plan, be apportioned in equal installments over the next ten years. Administratively, this plan has simplicity except in the necessity of running an account forward, which, however, is by itself perhaps too complex in view of the limited advantages obtained. This plan is more nearly neutral as to economic effects than any of the plans so far discussed except No. 10 above, and is about on a par with that one. As to equity, it is perhaps as reasonable to assume that the taxpayer's future "other income" will be stable as it is to assume that his past income has been stable. Moreover, this plan does avoid the inequities involved, in all the apportionment plans except No. 10 above, when accruing gains (or losses) overlap. However, it might be advisable to add an interest charge to the delayed tax. In effect, this plan is the obverse of No. 10 -- in place of reopening past returns, future actual returns are used as a basis for computing tax on realized gains or losses. If this future-years plan is used, it might be offered as an option -- a kind of exceptional relief -- the taxpayer's alternative being to include gains and losses, less some arbitrary discount -- say, 10 percent -- as other income.


The taxpayer's income, including gains and losses (all of which would be reported simply as ordinary income) would be averaged over a period of years -- say, six years -- in one of the ways outlined in an earlier section of this chapter. This plan would probably produce more revenue than ignoring gains and losses altogether, and probably no less than the existing step-scale plan, though it is almost impossible to give even a tentative answer to this question without considerable study. Administratively, the averaging plan would be somewhat less simple than the existing step-scale plan, but probably simpler than the apportionment plan (No. 5 above) or even a one-table multi-step-scale plan (No. 6 or No. 7 above), at least from the taxpayer's point of view. For the price involved, the plan gives extra service, since it handles some troublesome problems in other kinds of fluctuating income (e.g., lawyers' fees). Again, however, this plan is not neutral with respect to market action, /10/ although the averaging, like the two-table plan, makes it difficult for the taxpayer to calculate what his best course of action may be. At least there would probably be few, if any, instances where taxpayers could save money by reducing other income, and, more important, there would be no way for them to forecast with certainty whether this saving would really exist after the averaging device had done its work. As to equity among taxpayers, the averaging plan seems to come much closer than any of the other plans above. It does not need to assume that there is no overlapping of accruing gains (or losses) or that "other income" has been stable -- in fact, its reason for existence is that the contrary is assumed to be the normal case.


This method, the true "accrual method" is here called the "inventory method" to avoid confusion (see the footnote to No. 6 above), but it does not mean inventory at cost or market, whichever is lower. Rather, the asset is valued at the end of each year and the plus or minus difference from the value at the end of the preceding year -- or cost or other basis, if the asset was acquired within the year -- is brought to account as ordinary income or loss. To avoid too severe fluctuations, an averaging system should probably be used. Administratively, the plan would be feasible for listed securities and actively traded commodities and unlisted securities, but not for other commodities and unlisted securities and not for other assets (e.g., real estate). It would, incidentally, get rid of much existing complexity by making unnecessary (a) the undistributed profits tax, (b) the personal holding company and section 102 provisions, (c) probably also most of the tax-free exchange provisions in section 112 and 113. Inconvenience to the taxpayer in paying before realization might be avoided by granting him a few years' delay (at [illegible] percent interest) in payment (somewhat like the estate tax provisions). This plan is the only one that removes practically all influence on the securities and commodities exchanges. The only effect left would be a mild pressure on the market now, rather than later, in so far as taxpayers sold part of their securities now, rather than later, to pay the tax, and possibly a pressure to lower stock prices at the year's end for inventory purposes -- a pressure probably of slight practical significance in view of (a) the many other and powerful factors influencing stock prices and (b) the tendency toward evening up the account upon realization. As to equity, it would, over those assets to which it applied, treat all income recipients alike. And as to revenue, there seem to be no a priori grounds for assuming that there would be a great difference one way or the other from the apportionment plans, especially if an averaging device is used. There are no difficulties as to changes in other income or tax rates, and no difficulties over assumptions that accruing gains do not overlap. But since this inventory plan cannot apply to all assets, it follows that (a) there would be some difference in treatment of taxpayers owning different kinds of assets, and (b) the plan would have to be coupled with some other plan that would cover the other assets -- say, the general averaging plan in No. 12 above. It would be desirable to have the averaging plan cover all transactions, including the inventoried assets. As to procedure in enactment, this method almost surely requires a constitutional amendment, such as adding to the end of the Sixteenth Amendment the qualifying phrase, "including income accrued, whether or not realized, and ownership of or beneficial interests in incomes of legal persons." A more specific amendment might instead read something like this: "and the term 'incomes' shall not be interpreted to exclude (a) accrued gains or losses determined on the basis of valuations taken at the close of the taxable years or (b) shares in undistributed income of corporations or trusts." /11/ If the states are to be given the power to tax income on the same basis -- and this would seem desirable if only from the viewpoint of possible future tax coordination -- the amendment should be correspondingly broadened.

Conclusion of Present Writers

For reasons that are in part given above and in part in the more detailed section below, the writers favor the restricted inventory plan in No. 13, coupled with the averaging device in No. 12. If this combination plan is not feasible, the averaging device alone (No. 12) will probably furnish a satisfactory permanent plan. Otherwise, there is not much to choose from among the other plans, /12/ except that in any case the existing step-scale plan should not be allowed to stand. A single step-scale plan may be a necessary makeshift, but if it is used, its rates for proportion of gains taken into account should be much higher and the loss proportions should also be much higher than those in the present law. If this were done, it would probably be as well to keep the single step-scale plan as to use any of the plans in Nos. 5 to 11, chiefly because the other plans seem not to promise enough to counterbalance the unsettling effect of a change in method, if the change is to be tentative pending a solution of the constitutional problem.

Transfers of Property by Gift or at Death

In any case a determined effort should be made to take gains and losses into account upon transfer of property by gift inter vivos or upon death. The writers doubt that these loopholes can be satisfactorily plugged without a constitutional amendment. They are very serious, especially when assets with large unrealized gains are given to tax-exempt persons such as charitable organizations. In some cases the taxpayer has more money left for his own use if he transfers such an asset than if he sells it and keeps the proceeds. This peculiar result comes from the operation of this loophole coupled with the 15 percent deduction for gifts. /13

Detailed Study of Four Plans

The foregoing pages of this section have summarized the chief plans that may be considered, giving their major advantages and defects. In order to weigh the various factors further, it is desirable to study these features in more detail in the light of four of the above plans. Not all the relevant points of all the plans will be encountered, of course, but most of the important issues will be treated. The four plans chosen are those to which the writer's attention happened to be directed early in the study, this latter part of the present section having been written before the part just concluded.


Capital gains and losses in the broadest sense are changes in money value of any piece of property, after subtracting any outlay made on that property. In practice, however, the income tax laws usually exclude from capital transactions all property that, according to generally accepted accounting methods, should be inventoried each year. /14/

It may be assumed that no plan to be considered at the present for taxing capital gains and losses should be extended to such inventoriable property. However, there remain some troublesome questions as to which of all the other kinds of gains and losses should be taken into account. The objective should be to take account only of those kinds of gains and losses that are deemed to affect one's ability to pay tax or one's duty to support the government. From this point of view, it is useful to divide all gains and losses (aside from those arising from inventoriable property) into the following groups. /15/

Gains and losses arising from:

(a) A change that is proportionate to, and associated with, a change in the general price level.

In this case the asset's real purchasing power is not changed, and there seems to be universal agreement that no such gain or loss should be taken into account. Unfortunately, there is also almost universal agreement that no acceptable method of segregating such gains and losses has yet been devised. The present writers are not willing to assume that the task of segregation is hopeless, but it is clearly one that needs much analysis. Nothing can be done in time for the coming session of Congress, probably, but it is recommended that the Treasury start an intensive study of the subject.

(b) Using for personal consumption (e.g., the usual decrease in the value of an owner-occupied dwelling).

Losses only, not gains, are considered under this heading. Such gains as do occur while using for personal consumption may better be put in groups (a) and (c) to (f) -- the process of using for personal consumption normally results in a decrease in value. These losses cannot be deductible, any more than personal expenditures generally. If, however, during the period of using there occurs some loss that is unusual, in the sense that it is in no wise counterbalanced by consumption satisfactions to the taxpayer (e.g., the house burns down), it may be reasonable to allow a deduction. Such a decrease is not a decrease caused by consumption for personal satisfaction. It is treated in group (d) below. Similarly, decreases in the value of the land on which a dwelling rests may be considered deductible (see group (e) below).

(c) A change in the rate of interest.

This problem has not been discussed much, and the present writer has some doubts on both sides. Again, however, no method of segregation has been devised, nor is it likely to be devised without considerable study.

(d) A discovery (e.g., gold mine) or a loss (e.g., theft) of an asset.

Probably most observers would agree that the gain should be taxed, if any capital gains are to be taxed. Possibly less agreement would be found as to deductibility of the loss.

(e) A change in the prospective flow of net income to come from the asset -- e.g., a change in the prospective earning power of a business, or in a piece of real estate.

Some troublesome problems of analysis arise here, especially as to whether these gains usually denote a gain for the community as a whole (a "social gain"), and, if they do not, whether there must be somewhere in the community offsetting losses that should be allowed as deductions under the income tax. This analysis has not, however, been developed yet to a point where it can be useful in guiding practice, and meanwhile there is probably almost universal agreement among those who would tax capital gains at all that this type of gain should be taxed (and this kind of loss allowed as a deduction).

(f) A change in prospective purchasers' belief that they can resell at a higher price.

This category is intended as a catch-all for cases not classifiable under (a) to (e) above. It includes such a phenomenon as a Florida land boom. Generally, in such instances, one member of the community makes money at the expense of another and it can be argued that these transactions connote changes in relative ability to support government that should find expression in the income tax return through capital gains and losses.


Once it has been settled what kinds of capital gain and loss should be taken into account, various methods of taking them into account must be compared. The most useful way to compare them is by measuring them against a set of reasonable standards. Some of the standards used in other parts of this report for other taxes will not be used here because all the methods of treating gain and loss measure up in the same way with respect to those standards. For instance, none of the methods is suitable at all for taxing special benefits. Still other standards are omitted in this section because, although the various methods conceivably differ with respect to them, not enough is known about the difference to be useful in making a choice, and "no opinion" would have to be noted in each case. This is true with respect to the method's effect on the volume of saving, for instance.

On the other hand, a greater number of standards altogether, and much more specific standards, will be used here than in those sections dealing with whole taxes. Here we are concerned with a particular part of a particular tax, and it becomes both feasible and necessary to state standards in some detail. In this section of the chapter, each standard used will be taken up in turn, and then a recapitulation made at the end.


`The present section tests four methods of taking capital gains and
losses into account. It would be highly desirable to test more, if
time limitations permit. The three plans tested are Plan A (taking
certain unrealized gains and losses into account, and using an
averaging system -- hereafter referred to as the "accrual-averaging
method"), Plan B (taking only realized gains and losses into account,
and using an averaging system -- hereafter referred to as the
"realization-averaging method"), Plan C (the existing step-scale
plan), and Plan D (the plan recently submitted by the Division of
Research and Statistics -- hereafter referred to as the
"apportionment plan"). The alphabetical order carries no implication
whatever of relative merit.

A question mark in the standards table indicates that the writer does not wish to hazard an opinion on the result.

In each standards table a "YES" answer indicates an advantage that the method has; a "NO" answer indicates a disadvantage.

i) Plan A:     Accrual-Averaging Method (Taking
               Certain Unrealized Gains and
               Losses into Account and Using
               Averaging System)

a. Securities and commodities traded in at any time during the year on any organized exchange are valued at beginning of year (this value being the price shown by the first sale on any exchange) and at the end of the year (this value being the price shown by the last sale on any exchange), and the plus or minus difference is entered on the return in the same way as other income or loss. The Treasury compiles and publishes, within a short time after the end of the taxable calendar year, an official list of prices of every security and commodity, on the above basis. /16/ Realization, by sale or any exchange whatever (including what are now tax-free exchanges) is made an occasion for computing gain or loss from year-beginning value. Year-end value is compared with cost, or value at time of gift or time of donor's death, if asset was not in taxpayer's (donee's) hands at beginning of year. Passage of property by gift inter vivos or at death is deemed realization, and an accounting is made in that year's return. For the transition from the existing method of treatment, the first year's-end-valuation could be treated as realization, but the gain or loss taken into account in accordance with the existing step- scale plan, with probably some upward shift in the percentages, which at present go down much too sharply.

b. Securities not traded in on organized exchanges are taxed to owner on basis of increase in book value from year-beginning to year- end (see discussion of undistributed profits tax in Chapter III) or, if security was not in hands of taxpayer at year-beginning, the basis is as explained for similar circumstances in 1 above. The corporations in question must, within one month of close of their taxable year, inform taxpayer of change in book value (if he is a registered stockholder or bondholder, by direct mail; otherwise, by publication in a financial journal). If later definitive settlement with Treasury shows change in book value more than 5 percent greater or less than the statement to security owner, he shall be notified and shall include the plus or minus difference in his return for the current year. Commissioner may make special allowance in case of change in book value due to transfer of amounts among items on liability side of balance sheet. In case of colorable sale to shift tax liability (e.g., sale near the year-end with agreement to repurchase within a given time at a fixed price), both buyer and seller shall be held liable for a gain, and neither shall be allowed a loss. Upon transfer by gift or at death, realization shall be deemed to occur, and an accounting made in that year's return. Upon sale, and possibly also upon what are now tax-free exchanges, gain or loss is brought to account, the basis being increased by the amount of book value changes already taken into account in taxpayer's prior returns.

c. For all other assets (notably real estate) taxpayer may write up (but not down) the value each year as much as he wishes; otherwise, no account is taken until realization by sale or possibly also upon what are now tax-free exchanges. Aside from these provisions, no special treatment is given to the gains or losses (but see averaging device in No. 5 below). Upon transfer by gift or at death, realization shall be deemed to occur, and an accounting made in that year's return.

d. If the taxpayer's net income is a minus quantity -- whether because of capital losses or for other reasons -- he is allowed to carry over this loss until it is absorbed by future years' income (no restriction on number of years).

e. An overall averaging device, applicable to all income, is adopted. Possible averaging devices are explained in detail in an earlier section in this chapter. /17/

ii) Plan B:    Realization-Averaging Method
               (Taking Only Realized Gains and
               Losses into Account, and Using an
               Averaging System)

a. No gains and losses are taken into account until realization as defined in the present law (with respect to what are now tax-free exchanges, the averaging system might soften the tax blow enough to warrant repeal of these complicated measures, but this is doubtful).

b. No special treatment at all is given to realized gains or losses. The gains go in with other income and the losses go in with other deductions. There are no restrictions on the deductibility of losses.

c. If the taxpayer's net income is a minus quantity -- whether because of capital losses or for other reasons -- he is allowed to carry over this loss until it is absorbed by future years' income (no restrictions on number of years).

d. All income is averaged by some one of the methods described in an earlier section of this chapter.

e. If possible without amending the constitution, gift inter vivos or transfer at death is deemed realization, and the gain or loss taken into account in that year's tax return (it is extremely doubtful, however, whether this can be done without amending the constitution).

iii) Plan C:   Step-Scale Plan (Method Now in

a. Gains taken into account at a fraction of full amount that depends on length of time asset has been held. Losses treated same way, but limited to gain plus $2,000, with no carryover. Wash sales prohibition and other relevant items as in existing law. No gain or loss accounted for upon gifts inter vivos or transfers at death.

iv) Plan D:    Apportionment Plan /18/ (Prorating
               Gain over Past Years, Taxing on
               Present Year Basis)

a. This plan is the one submitted by the Division of Research and Statistics in its recent Studies. It divides each gain and loss by the number of years the asset was held, calculates the tax or tax relief due, on the basis of the present year's income, on the result, and multiplies by the number of years. This description is too simplified to give an exact impression of the plan; by using a weighted average annual investment period, some potential anomalies are obviated. Capital losses can be offset only against capital gains of the same year, except a carryover of a tax credit on account of loss is allowed for two years, to be offset against tax liability on capital gains. Gains and losses to be taken into account upon gifts inter vivos and transfers at death, without amending the constitution (but there are serious doubts whether this is possible, and some less desirable alternatives are therefore suggested).


Several standards concerning revenue are listed in the following table of comparisons. All of the statements of probable or possible results are based simply on the writer's general impressions. It is of course advisable to analyze such data as exist, if time permits, but in most cases it is unlikely that these data will give the answers. As the answers stand now, it is difficult if not impossible to base a choice among the methods on revenue considerations because of the uncertainties involved.


The most important economic effect that is susceptible of fruitful analysis in this case is the effect on security and commodity markets. This matter has been the subject of so much controversy that it seems desirable here to try to state the essential issues and the appropriate analysis in brief but comprehensive form, in terms of the effects of the existing step- scale method. The analysis covers all the major points that are likely to arise under any of the four plans discussed in this chapter. The plans are compared, with respect to each point, in the table of standards at the end of this analysis.

The existing method, or step-scale plan, of taking capital gains and losses into account affects the decisions of holders of assets when to sell, for several reasons. These may be treated under two general heads; cases where the asset shows an unrealized gain, and cases where it shows an unrealized loss:


a. Assuming that the taxpayer can forecast the price of the asset with a fair degree of confidence, the step scale may make it advantageous for the taxpayer to hold, even if he thinks the price will decline somewhat. For example: if the taxpayer's top bracket rate is 50 percent, and he has held the asset for slightly more than two years, the effective tax rate on his gain (assuming it is all in the top bracket) is 30 percent (60 percent of 50 percent) if he sells at once, but only 20 percent (40 percent of 50 percent) if he holds for three years and sells. As far as this factor alone is concerned, therefore, it pays him to hold three years so long as the gain does not decrease in the interim by more than 12 percent. /19/

Conceivably, these considerations would not influence the level of market prices if the taxpayer were strictly logical and fully aware of his opportunities, since the indicated procedure would be to sell short the same security, or, if the tax regulations did not consider it a short sale, some security with similar prospects for a decline. Thus the same pressure would be on the market as if there were no tax. However, it may be doubted if most taxpayers are rational and astute enough and have the opportunity to follow this procedure. Moreover, it may be argued that if there were no step- scale tax on gains, the taxpayer would both sell the security he owned and sell short, thus resulting in greater pressure on the market than results when the tax is in force. After all, however, the chief consideration is what taxpayers do do, not what they might do, and general observation leads to the conclusion that most of them do not sell short in the case in question.

b. The earning asset represented by the amount of the potential tax may make it advisable for the taxpayer to hold rather than sell, although he anticipates a decline in price. This pressure is operative, however, only when (a) he thinks that the asset, after declining in value, will rise considerably above its present price or will give him substantial dividends over a long period of holding, and (b) if, were he to sell and repurchase, he could buy no more shares than can be purchased by the proceeds of the sale minus the tax paid. A more general statement is: by selling and paying the tax he loses a potential earning asset (equal, that is, to the size of the tax), so that he may find it desirable to hold even if the asset is going down before it yields returns in dividends or later capital gain, or is not going up as fast as another asset (thus the potential tax may prevent switching that would otherwise occur). As in case 1 above, the possibility of selling short must be considered.

c. If the taxpayer can forecast his other income (including other capital gains and losses) with a fair degree of confidence, he may decide to:

1) Withhold gains until a year when his income is

2) Withhold parts of his gains until years when his
income is the same (thus the gain may be kept within a
single rate bracket instead of lumping up into other,
higher brackets).

3) Realize gains earlier than he would otherwise
because he foresees years of large income ahead (little
effect on market, since immediate repurchase is

d. If the taxpayer can forecast rate changes or other changes in the law with a fair degree of confidence, he may withhold or hasten to realize gains, as the case may be. In case of hastening, little effect on the market may be expected, since repurchase is advisable.

e. Since capital losses can be set off only against the same year's gains (plus $2,000) and no carryover is allowed, it may be advisable to delay realizing a gain until a year when a loss is realized, even though the gain decreases meanwhile. For example: if the taxpayer's top bracket rate is 50 percent, and all the gains and losses fall in that bracket, and if he has no capital loss this year, but will realize one of $5,000 next year, a gain of $10,000 realized this year leaves the taxpayer no better off than a gain of $7,000 realized next year. /20/ Again the possibility of selling short must be considered.

For a similar reason, it may be advisable to realize on a gain earlier than would otherwise be the case, but presumably the asset would be repurchased at once (assuming he would never have sold but for the tax consideration), with, consequently, little effect on the market.

f. The fact that gains accrued at death escape the income tax may well influence one's decisions as to which securities to convert into cash and which to pass on to one's heirs. Indeed, permanent retention of an asset, through several generations, because of this loophole, seems not out of the question. A similar influence -- but one less powerful, because of the provisions for carrying over the low basis to the donee, except when the donee is exempt from tax -- is exercised when a gift is made inter vivos.


a. The step scale may make it advantageous for the taxpayer to sell, even though he thinks the asset will appreciate somewhat in value. For example, if the taxpayer's top bracket rate is 50 percent, and if he has gains enough in either year to make the losses fully effective, and if all his gains and losses fall within that bracket, a realized loss of $1,000, four years old, leaves him no worse off than a realized loss of only $875 a year later. /21/ The net effect on the market price level may not be great, however. The taxpayer would probably be strongly inclined to buy, as soon as he had sold, a similar security with like prospects for a rise (or the same security, after the delay necessitated by the usual wash sale provision). There would not be the same hesitation here to embark on the compensating transaction as in the case of gains (see example No. 1 under gains), where the compensating transaction has to be a short sale.

b. The earning asset representing the amount of the potential tax saving may make it advisable for the taxpayer to sell rather than hold, although he anticipates an increase in price. As in the preceding paragraph, however, he would be strongly inclined to buy back into the market.

c. If the taxpayer can forecast his other income (including other capital gains and losses) with a fair degree of confidence, he may (assuming there are enough gains to allow the losses to take full effect) decide to:

1) Refrain from realizing losses until a year when
his income is higher. /22/

2) Refrain from realizing losses until a year when
his income is the same (thus the loss may be kept within a
single rate bracket instead of lumping down into other,
lower brackets). /23/

3) Realize losses earlier than he would, because he
sees years of small income ahead (but he may then buy back
into the market).

d. If the taxpayer can forecast rate changes or other changes in the law with a fair degree of confidence, he may withhold /24/ or hasten to realize losses, as the case may be. If he hastens the realization, he may buy back into the market.

e. Since capital losses can be set off only against the same year's gains (plus $2,000), and no carryover is allowed, it may be advisable to hasten the realization of a loss to get it into a year of gain, even though the asset would increase in value later. Again, however, the taxpayer would be inclined to buy back into the market at once. If there were no gains this year, but some were expected in future years, it might pay to delay realization of the loss (here the logical counter-action is a short sale).

f. The fact that he may not be able to pass on a loss basis to his donee or heirs may cause the taxpayer to choose a loss-bearing asset if he has to sell some assets anyway.

In the two parts immediately above ("When asset shows unrealized gain" and "When asset shows unrealized loss") paragraphs a are relevant only to the step-scale plan, not to the other three plans. Paragraphs e are relevant only to a plan that has a limitation of this nature on losses. Paragraphs f are relevant only to plans that fail to take account of gains and losses at time of gift inter vivos or transfer at death. Paragraphs b, c, and d, however, are applicable, with varying degrees of strength, to any plan of taking gains and losses into account except a plan that removes from the taxpayer all chance to decide at what time the account shall be taken (the only practicable plan for doing this seems to be an annual accrual plan -- i.e., Plan A, the accrual-averaging plan -- although in Plan B, the realization-averaging plan, the influence on market policies is lessened considerably by the averaging device). The reaction described in Paragraph e-1 under gains -- withholding gains until a year when income is lower -- reveals itself in an extreme form under the apportionment plan where, in certain cases, it actually pays a taxpayer to receive less "other income" (assuming he has not the choice of shifting it to other years), as explained in the first part of this capital gains section above. (See chapter III, note 5.)

In general, there appears to be much less likelihood of loss sales affecting the market than gain sales. The "unnatural" effect exercised by taking losses into account is usually one of hastening realization, not delaying it; and a hastening of realization implies that the owner is going to sell even though he thinks the asset is not going any lower (if he did think it was going lower, he would presumably want to sell at once anyway, and the loss provisions would be exercising no distorting influence). But if the asset is not going any lower, the owner can retain his position (which he would presumably want to do if there were no gain-and-loss provisions) by simply buying back into the market, at once if he buys a similar asset, or after the wash sale period if he buys back the very same securities. This buying-back tends to counteract the depressive influence of the sale. In so far as there is a depressive effect on the market, it seems to be due, not to the loss provisions, but to the wash-sales provisions (Section 118, Revenue Act of 1936).

With respect to gains, on the other hand, the taxpayer who holds on to an asset that he would sell if there were no tax can maintain the position he would be in if there were no tax, namely, the position of not holding the asset, only by selling the asset short -- a procedure not readily open to everyone, and unattractive to many. Hence, such influence as the gains-and-losses provisions should exercise over rational investors and speculators is likely to be much stronger with respect to assets showing accrued gains than with respect to assets showing accrued losses. /25/

There remains the real problem of the irrational investor. Frightened or displeased by the prospect of paying a tax if he sells -- or eager for the tax saving that he may get by realizing a loss -- he delays or hastens realization even when it is directly contrary to his own financial interests to do so, at least as far as he could tell, if he took the trouble to reason the matter out. These cases exist; the writer has known of them among investors of large means, intelligent people in their own regular business. How widespread this irrationality is cannot be known, but it is the writer's impression that it is of considerable importance. After all, the factors involved in a really intelligent judgment as to how much and in what direction one ought to allow oneself to be influenced by the gain- and-loss provisions are so complex that it is no cause for wonder that many investors and speculators look to the thing closest at hand -- the tax payment or the tax saving -- and stop there.

As the following table [Table B, see next page] shows, the accrual-averaging method has a decided advantage over all the other three so far as this matter of market influence is concerned. The realization-averaging device probably removes a large part of the influence if the taxpayer really understands what the device does for him over the long run. The existing step-scale plan seems to be the worst, but the apportionment plan may really, on balance, be the worst because of the pressure it may exert to decrease the "other income."

Another economic effect, however, is the tendency, which the accrual-averaging plan would probably show, to delist securities. None of the other plans has this effect. Whether the effect would be very noticeable in fact is, however, doubtful. The taxpayers who pay under the accrual-averaging plan are likely to be better off in the long run, in many cases, than the others -- and there are many factors involved in deciding whether to list a security.


The comparative equity of the several methods seems to depend largely on:

(i) treatment given to taxpayers with same total
income over a period of years, but different accruals
and/or realizations within the period;

(ii) taking into account gains and losses accrued to
death or when gift is made inter vivos;

(iii) extent to which capital losses are deductible;

(iv) similarity of treatment of all kinds of capital

Each of these will now be discussed in turn.


Taxpayers who have the same total income over a period of years are likely to accrue and/or realize this income in different-sized segments at different times within the period. It may not be necessary that all such taxpayers should be treated exactly alike, but there is general agreement that some of the most obvious cases of differential treatment should be avoided. For instance, few would say that it would be equitable simply to throw into one year's return, with no adjustments at all, a large gain accumulated over several years. The questions involve not only one capital-gains taxpayer compared with another, but also a capital-gains taxpayer compared with, e.g., a salary recipient. The comparisons can be made:

(a) between taxpayers with the same accruals, but
different realizations;

(b) between taxpayers with different accruals, but
the same realizations;

(c) between taxpayers with both different accruals,
and different realizations;

-- in every case the taxpayers having the same total
income over the period of years in question.

This analysis is carried out in the following paragraphs and in Table C, Section A.

a) EFFECT OF PROGRESSIVITY OF RATE SCALE. If we compare a number of taxpayers who get the same net income over a period of years, but who accrue it and/or realize it at different times within the period, we find that the progressivity of the rate scale of the tax levies a greater total tax on some of these taxpayers than on others. The following examples show the kinds of discrimination when the taxpayer is not accruing more than one gain (or one loss) at a time (the taxpayers in these examples have no other income, for simplicity in exposition, but the general conclusions reached are valid in any case):

1. A accrues and realizes a gain of $100,000 the
first year and accrues and realizes a gain of $100,000 the
second year. He may be designated a "regular-realization

2. B accrues a gain of $100,000 the first year and
$100,000 the second year, and realizes the $200,000 at the
end of the second year. He may be designated a "delayed-
realization taxpayer -- regular accrual."

3. C accrues and realizes a gain of $150,000 the
first year and accrues and realizes a gain of $50,000 the
second year. He may be designated an "irregular-
realization taxpayer."

4. D accrues a gain of $150,000 the first year and
$50,000 the second year, and realizes the $200,000 at the
end of the second year. He may be designated a "delayed-
realization taxpayer -- irregular accrual."

If the income tax is levied annually, but only upon realized gains, B is taxed more heavily than either the regular- or irregular- realization taxpayers, A and C (since part of B's gain gets into higher surtax brackets than A's or C's). Likewise, D, the other delayed-realization taxpayer, is taxed more heavily than A or C. Also, C is taxed more heavily than A. The only two taxpayers whose total taxes are equal are the two delayed-realization taxpayers, B and D. (See chart below.) All the other five possible comparisons show inequalities. Taxpayers B and D (except with respect to each other) are the unfortunate ones, and C also loses, in one instance.

If the income tax is levied annually, but upon accruals, A and B are taxed alike (the regular-realization and regular-accrual taxpayers). Also, C and D are taxed alike (the irregular-realization and irregular-accrual taxpayers), but heavier than A or B. In only four cases, rather than five, do the taxes differ. Taxpayer D is again unfortunate, being discriminated against twice, but has as his companion in misfortune, not B, but C.

                               TABLE B
     (YES indicates an advantage) (NO indicates a disadvantage)
                                  PLAN A             PLAN B
                                (Accrual-         (Realization-
                                Averaging)          Averaging)
1.  Sales to realize gains are:
(a)  Made just as soon as     YES -- for          NO --
they would be if there        listed              although
were no income tax            securities and      compared with
                              commodities         Plan C, items
                              NO -- for other     (1) and (5), and
                              assets -- but       possibly (6),
                              not quite as        would not apply,
                              much delay as       and the force of
                              under other         items (2), (3),
                              plans               and (4) would be
                                                  dampened by the
                                                  averaging device

(b)  Made no sooner than      YES -- except       NO -- although
they would be if there        for pressure to     compared with
were no income tax            sell now,           Plan C, loss
                              rather than         offset would
                              later, to get       not matter, and
                              cash to pay the     effect of other
                              tax, or             items would be
                              (probably           dampened by
                              rarely) and         averaging
                              sales to            device
                              decrease year-
                              end price on
                              which value is

(c)  Made at least            YES                 NO -- (?)
eventually, assuming they                         depends on
would be made eventually                          effort to plug
if there were no income                           loopholes
tax                                               without
                                                  amendment --
                                                  success highly
                         [TABLE B CONTINUED]
                                   PLAN C                  PLAN B
                                (Step-Scale)          (Apportionment)
1.  Sales to realize gains are:
(a)  Made just as soon as        NO -- several        NO -- although
they would be if there           influences at        compared with
were no income tax               work to delay        Plan C, item
                                 sales to             (1) and
                                 realize gain:        possibly item
                                 (1) Step-scale       (6) would not
                                 itself (2)           apply. Items
                                 Progressive          (2), (3), (4),
                                 rate scale --        and (5) under
                                 better to save       Plan C would
                                 part or all of       apply, and
                                 gain for years       there may be an
                                 when other           incentive to
                                 income is the        decrease other
                                 same or lower        income (this
                                 (3) Possible         would not occur
                                 earnings on          under Plan C)
                                 represented by
                                 potential tax
                                 (4) Possible
                                 rate changes,
                                 etc., in future
                                 years (5) May
                                 want to save
                                 gains for year
                                 when losses are
                                 realized (6)
                                 Loophole upon
                                 gifts inter
                                 vivos and
                                 transfer at

(b)  Made no sooner than         NO -- may be         NO -- same
they would be if there           hastened in          comment as
were no income tax               order to have        under Plan C
                                 something to
                                 offset losses
                                 of same year
                                 against, or to
                                 avoid future
                                 years of large
                                 income or
                                 higher rates.
                                 Likely to
                                 result in
                                 without much
                                 effect on

(c)  Made at least               NO -- failure        NO -- unless
eventually, assuming they        of present           this plan can
would be made eventually         system to plug       include some
if there were no income          loopholes upon       features that
tax                              gifts and at         will plug this
                                 death may cause      loophole
                                 indefinite           without
                                 retention            requiring a
                                                      amendment --
                                                      which seems
                                                      highly doubtful

(d)  Avoided forever, in      YES                 YES
cases (probably rare)
where they would likewise
be avoided forever if
there were no income tax
2.  Sales to realize losses are:
(a)  Made just as soon as     YES -- for          NO -- although
they would be if there        listed              compared to
were no income tax            securities and      Plan C, gain
                              commodities         offset would
                              NO -- for other     not matter, and
                              assets like         effect of other
                              Plan B              items would be
                                                  dampened by

(b)  Made no sooner than      YES -- for          NO -- although
they would be if there        listed              compared to
were no income tax            securities and      Plan C, no
                              commodities NO      step-scale or
                              -- for other        gain-offset
                              assets, like        influence, and
                              Plan B              effect of other
                                                  items (except
                                                  loophole) would
                                                  be dampened by

(c)  Made at least            YES                 YES
eventually, assuming they
would be made eventually
if there were no income

(d)  Avoided forever, in      YES -- for          NO
cases (probably rare)         listed
where they would likewise     securities and
be avoided forever if         commodities NO
there were no income tax      -- for other
                              assets, but not
                              as much as
                              other plans
It avoids exerting            NO -- but must      YES
pressure to delist            consider
securities                    pressure to
                              list that would
                              be exerted by
                              the requirement
                              that unlisted
                              take the
                              trouble to
                              inform each
                              security holder
                              of annual
                              change in book

                         [TABLE B CONTINUED]

(d)  Avoided forever, in         YES                  YES
cases (probably rare)
where they would likewise
be avoided forever if
there were no income tax
2.  Sales to realize losses are:
(a)  Made just as soon as        NO -- may be         NO -- same
they would be if there           delayed until        comment as
were no income tax               there are gains      under Plan C
                                 that can be
                                 offset against
                                 them, or in
                                 order to make
                                 as much of the
                                 loss as
                                 against as high
                                 surtax rates as
                                 factors are
                                 size of other
                                 income, and
                                 rate scales in

(b)  Made no sooner than         NO -- may be         NO -- although
they would be if there           hastened             compared to
were no income tax               because of           Plan C, no
                                 step-scale,          step-scale
                                 interest on          influence, and
                                 earning asset,       possibly (see
                                 prospect of          above) no
                                 low-income           loophole
                                 years or low-
                                 rate years, or
                                 years with not
                                 enough gains to
                                 offset; but in
                                 all these
                                 cases, without
                                 such market
                                 effect, since
                                 repurchase is
                                 Loophole upon
                                 gifts inter
                                 vivos and
                                 transfers at
                                 death may have
                                 some effect

(c)  Made at least               YES                  YES
eventually, assuming they
would be made eventually
if there were no income

(d)  Avoided forever, in         NO                   NO
cases (probably rare)
where they would likewise
be avoided forever if
there were no income tax
It avoids exerting               YES                  YES
pressure to delist
                     B, C, AND D, OVER TWO YEARS
               (Break in line represents realization)

- 50,000

It will be noted that:

The A-B comparison                   are between taxpayers with the
and the C-D comparison               same accruals but different

The B-D comparison                   is between taxpayers with
                                     different accruals but the same

The A-C comparison                   are between taxpayers with
and the A-D comparison and           both different accruals and
the B-C comparison                   different realizations

To generalize: The realization method and the accrual method each discriminates between taxpayers with both different accruals and different realizations; the realization method also discriminates between the taxpayers with the same accruals but different realizations, while the accrual method discriminates between the taxpayers with different accruals but the same realizations. Of the four plans above, it will be recalled, the first one is an accrual plan and the other three are realization plans -- but the discrimination of the realization-averaging method is dampened by the averaging device.

b) INFLUENCE OF THE EARNING ASSET, ETC. The foregoing paragraphs have been concerned with differences in burden that arise because the rate scale is progressive. But differences may also arise because the taxpayer has to pay sooner than another and so loses an earning asset represented by the amount of the tax. Furthermore, changes may occur, from one year to the next, in the legal definition of taxable income, in the rate schedules, and in the taxpayers' other income. All these factors tend to make for difference in tax payable by taxpayers, who have the same total income over a period of years. As Table C (Section A) shows, the differences appear among different kinds of taxpayers, depending on the particular plan of taking account of gains and losses that is in force.

As the table clearly shows, none of the four plans will give the same treatment to all taxpayers who have the same total income over a period of years. A choice must therefore be made as to whether identical-accrual taxpayers should be treated identically or, instead, identical-realization taxpayers treated identically. The present writers prefer the former choice -- or perhaps it reflects their meaning more precisely to say that they dislike the implications of the latter choice. On the grounds of relative ability to support government, they dislike to see taxpayer A in the chart above taxed differently from taxpayer B, and taxpayer D taxed differently from taxpayer C, more than they do to see taxpayer B taxed differently from taxpayer D.


The degree of success in checking avoidance of the tax by gifts inter vivos and transfers at death, and in allowing for losses accrued to those dates is not likely to be great without a constitutional amendment, which is called for in Plan A but not in the other three. Section B of Table C compares the plans in this respect [Table C, see next page].


The extent to which capital losses are deductible has, in part, been covered by implication in Sections A and B of Table C. The chief questions are: Should capital losses be allowed at all? If so, should they be deductible only against capital gains, or against income from all sources? In either case, if an unabsorbed loss remains, should it be carried over to future years, and if so, for how long? Answers to these questions may be given on the assumption that the matters referred to in Sections A and B have been tentatively settled.

Since capital gains are assimilated to other income and put right on top of that other income, no matter how great the gain is (after adjustments as in Section A and B), there seems to be a prima facie case to the effect that capital losses should be assimilated to other deductions and allowed as offsets against both capital gains and other income. If capital losses are limited to capital gains, with no carryover of losses, a taxpayer who has a capital gain of $1,000 this year and a capital loss of $1,000 next year will pay a tax on $1,000, while a taxpayer who has a gain of $500 and a loss of $500 this year and a gain of $500 and a loss of $500 next year will pay no tax on the gains. This result seems to put altogether too much dependence on the twelve-month period as such, and to give a decidedly unjust distribution of the tax burden. If capital losses are limited to capital gains and a carryover is allowed against future capital gains, much of the injustice is remedied, and, indeed, the result may be better for the loss-incurring taxpayer than if he were permitted to deduct the loss against other income -- he may find that his loss, on the whole, is distributed among higher surtax brackets. But it would seem preferable to devise a method of adjustment (under Sections A and B of Table C) that would make offsetting of the capital loss against other income the equitable procedure. Whether the result would give too great a fluctuation in the government's revenue deserves further study -- much depends on the other features of the particular method used. In any event, if all gains are to be taxed eventually, it appears that all losses should be allowed eventually.


Since some taxpayers will have more of some kinds of assets than other taxpayers, discrimination will result if all assets are not treated alike. Plan A is not satisfactory in this respect, as Table C shows. However, the discrimination tends to disappear upon realization, and, with an averaging system in force, there would probably be little discrimination of importance.

                               TABLE C
                        STANDARDS OF JUSTICE
     (YES indicates an advantage) (NO indicates a disadvantage)
                              PLAN A              PLAN B
                              (Accrual-           (Realization
                              Averaging)          -Averaging)
SECTION "A": Time of
accrual and realization
among taxpayers with the
same total income over a
period of years:
1. It gives the same
   treatment to taxpayers
   with the same accruals
   but different
   realizations, as concerns
   the effect of:
   (a) Progressivity of       YES -- as to        NO -- the
       rate scale             listed              remarks under
                              securities and      Plan D apply
                              commodities         here, except
                              both as to gain     that this plan
                              and loss            has no offset
                              NO -- as to all     limitations.
                              other assets;       The averaging
                              but there is a      system would
                              tendency in the     help
                              right               appreciably in
                              direction, in       many instances
                              the averaging
                              device, and, as
                              concerns gains,
                              in the
                              taxpayer to
                              write up his
                              asset, and in
                              the taxing of
                              book value
(b) Amount of earning assets  YES -- as to        NO -- any
    left to taxpayer          listed              "realization"
                              securities and      plan encourages
                              commodities         holding as far
                              NO as to all        as this factor
                              other assets;       is concerned
                              but see remarks
                              under (a)
(c) Changes from one year to  YES -- as to        NO -- the rate
    another in:               listed              schedule and
    (I) Legal definition of   securities and      other income of
    income                    commodities NO      the year of
    (II) Rate schedules       -- as to all        realization are
    (III)Taxpayer's other     other assets;       used, although
    income                    but see remarks     the definition
                              under (a) above     of previous
                                                  years is given
                                                  some effect
2. It gives the same
   treatment to taxpayers
   with different accruals
   but the same
   realizations (some
   doubt whether equality
   of treatment in this
   case is desirable) as
   concerns the effects

                         [TABLE C CONTINUED]

                                   PLAN C                  PLAN D
                                (Step-Scale)          (Apportionment)
SECTION "A": Time of
accrual and realization
among taxpayers with the
same total income over a
period of years:
1. It gives the same
   treatment to taxpayers
   with the same accruals
   but different
   realizations, as concerns
   the effect of:
   (a) Progressivity of       NO -- only in        NO -- only in
       rate scale             chance cases,        chance cases,
                              and rarely,          and rarely,
                              will the step-       will this plan
                              scale give           give equal
                              equal treatment      treatment,
                              as to gains;         either as to
                              the scale            gains or
                              discriminates        losses, but it
                              against those        would probably
                              whose gains          come closer to
                              fall in a            equality, in
                              steeply              most cases,
                              progressive          than Plan C.
                              part of the          The matter
                              rate scale, and      depends on the
                              in favor of          size of the
                              those whose          other income of
                              gains fall in a      the years in
                              moderately           question, and,
                              progressive or       as to losses,
                              flat-rate part       the amount of
                              of the rate          capital gains
                              scale. The           available as
                              present scale        offsets. This
                              goes down so         offset
                              sharply that         limitation,
                              this                 coupled with a
                              discrimination       loss carry-over
                              -- in practice,      provision, may,
                              in favor of the      however,
                              wealthy -- is        actually work
                              very marked. As      to redress some
                              to losses, the       of the wrong
                              scale tends to       done to the
                              discriminate         taxpayer who
                              against the          realizes his
                              taxpayer who         losses in a
                              realizes less        lump, since it
                              frequently; if       may distribute
                              the year shows       the "lumpy"
                              a net capital        loss among
                              loss, the scale      higher surtax
                              goes in the          brackets, on
                              wrong direction      the whole, than
                              (the                 if the loss
                              percentages go       were deductible
                              down with time;      all in one year
                              they should,         from other
                              for net losses,      income.
                              generally go
                              up, if they are
                              to approximate
                              what would
                              happen under

(b) Amount of earning assets  NO -- same          NO -- same
    left to taxpayer          remarks as          remarks as
                              under Plan B        under Plan B

(c) Changes from one year to  NO -- the           NO -- same
    another in:               definition,         remarks as
    (I) Legal definition of   rate schedule,      under Plan C
    income                    and other
    (II) Rate schedules       income of the
    (III)Taxpayer's other     year of
    income                    realization are
2. It gives the same
   treatment to taxpayers
   with different accruals
   but the same
   realizations (some
   doubt whether equality
   of treatment in this
   case is desirable) as
   concerns the effects

(a) Progressivity of rate     NO                  YES

(b) Amount of earning         NO                  YES
    assets left to taxpayer

(c) Changes from one year     NO -- as to         YES -- by not
    to another in:            listed              taking account
    (I) Legal definition of   securities or       of the earlier
    income                    commodities;        definitions,
    (II) Rate schedules       one taxpayer        etc., for
    (III) Taxpayer's other    will pay on         either taxpayer
    income                    more income
                              under the early
                              law, and less
                              under the later
                              YES -- as to
                              all other
3. It gives the same
   treatment to taxpayers
   when both different
   accruals and different
   realizations (some
   doubt whether equality
   of treatment in this
   case is desirable) as
   concerns the effect of:

(a) Progressivity of rate     NO                  NO

(b) Amount of earning         NO                  NO
    assets left to taxpayer

(c) Changes from one year     NO                  NO
    to another in: (I)
    Legal definition of
    (II) Rate schedules
    (III) Taxpayer's other
1.  It eliminates tax         YES -- for          NO -- see
    avoidance by gift inter   listed              remarks under
    vivos                     securities and      Plan C
                              Even aside from
                              the fact that
                              this plan
                              calls for an
                              accounting upon
                              transfer by
                              gift or at
                              death, the fact
                              that valuations
                              are taken every
                              year makes it
                              impossible for
                              much gain to
                              YES -- for all
                              other assets,
                              chiefly because
                              this plan
                              calls for an
                              accounting upon
                              transfer by
                              gift or at
2.  It eliminates tax         YES -- for          NO
    avoidance by transfer     listed
    at death                  securities and
                              (see remarks
                              YES -- for all
                              other assets
                              (see remarks
1.  It allows a capital       YES                 YES
    loss to be used
    completely, either as
    an offset against other
    income or against
    capital gains -- using
    a carry-over when
1.  It treats all kinds of    NO -- listed        YES
    assets alike              securities and
                              securities, and
                              all other
                              assets are
                              three classes,
                              each of which
                              is treated
                              from the other
                         [TABLE C CONTINUED]

                                   PLAN C                  PLAN D
                                (Step-Scale)          (Apportionment)
(a) Progressivity of rate     YES                 YES

(b) Amount of earning         YES                 YES
    assets left to taxpayer

(c) Changes from one year     YES -- same         YES -- same
    to another in:            remarks as          remarks as
    (I) Legal definition of   under Plan B        under Plan B
    (II) Rate schedules
    (III) Taxpayer's other
3. It gives the same
   treatment to taxpayers
   when both different
   accruals and different
   realizations (some
   doubt whether equality
   of treatment in this
   case is desirable) as
   concerns the effect of:

(a) Progressivity of rate     NO                  NO

(b) Amount of earning         NO                  NO
    assets left to taxpayer

(c) Changes from one year     NO                  NO
    to another in: (I)
    Legal definition of
    (II) Rate schedules
    (III) Taxpayer's other
1.  It eliminates tax         NO -- although      ? -- doubtful
    avoidance by gift inter   some tendency       whether
    vivos                     to lessen           proposals will
                              avoidance is        be
                              exerted by the      constitutional

2.  It eliminates tax         NO                  ? -- see
    avoidance by transfer                         remarks above
    at death
1.  It allows a capital       NO                  NO -- but
    loss to be used                               restrictions
    completely, either as                         are not as
    an offset against other                       severe as in
    income or against                             Plan C
    capital gains -- using
    a carry-over when
1.  It treats all kinds of    YES                 YES
    assets alike

     In comparing the various methods from the standpoint of
administration, it is necessary to ask several specific, narrow
questions, and this is done in Table D [see below].

                               TABLE D
     (YES indicates an advantage) (NO indicates a disadvantage)
                                 PLAN A               PLAN B
                                 (Accrual-         (Realization-
                                 Averaging)         Averaging)
1. It avoids requiring a              NO               YES
   valuation of listed
   securities and
   commodities at the end
   of each year
2. It avoids requiring a              YES              YES
   valuation of unlisted
   securities at the end
   of each year
3. It avoids requiring                NO               YES
   corporations to notify
   holders of their shares
   in the year's earnings

4. It avoids requiring a              YES              YES
   valuation of assets
   other than securities
   and listed commodities
   at the end of each year

5. The return form,
   compared with what it
   would be if gains or
   losses were treated
   just like other income

   (a) is just as simple         NO -- because       NO -- see
                                 the averaging       remarks under
                                 device would        Plan A
                                 add some
                                 complexity, and
                                 gains and
                                 losses would
                                 have to be
   (b) is less simple, but at    NO -- the           NO -- see
       least as simple as        averaging           remarks under
       existing step-scale       calculation         Plan A; but
       plan                      would be            there would be
                                 somewhat more       fewer items per
                                 complex than        year to list
                                 the computation     than under Plan
                                 of percentages      A
                                 under the
                                 present capital
                                 gains schedule
6. It makes unnecessary
   the provisions of:

   (a) Sections 112(b) to        YES -- as to             NO
       112(i) inclusive,         listed
       dealing with              securities and
       recognition of gain or    commodities. In
       loss, as to transfers     few cases would
       made after the new plan   enough gain
       went into effect          have accrued
                                 from the end of
                                 the preceding
                                 year to make
                                 the removal of
                                 provisions much
                                 if any hardship
                                 YES --
                                 probably, as to
                                 other assets,
                                 since the
                                 hardships would
                                 be greatly
                                 mitigated by
                                 the provisions
                                 allowing the
                                 taxpayer to
                                 write up his
                                 asset, taxing
                                 book value
                                 changes, and
                                 income over
                                 several years

   (b) Most of Section 113,      YES as to             NO
       dealing with adjusted     listed
       basis, as to transfers    securities and
       made after the new plan   commodities
       went into effect          (see remarks
                                 YES --
                                 probably, as to
                                 other assets
                                 (see remarks

   (c) 118, dealing with         YES -- as to          NO
       losses from wash sales    listed
       (but in any case, it      securities
       might be well to allow    YES --
       wash sales, in view of    probably, as to
       the possible economic     unlisted
       effects of repressing     securities. The
       them)                     taking into
                                 account of book
                                 value changes
                                 leaves a
                                 realizable loss
                                 so reduced that
                                 it might
                                 perhaps safely
                                 be left to the
                                 discretion of
                                 the taxpayer,
                                 since the
                                 security might
                                 not have a
                                 ready market.
                                 (See also
                                 remarks in

   (d) 24(a)(6), dealing with    YES --                NO
       losses from sales not     probably, as to
       at arm's length           all assets.
                                 After all, the
                                 philosophy of
                                 this plan calls
                                 for frequent,
                                 rather than
                                 accounting of
                                 gains and
1. It avoids requiring           NO -- as to           YES
   payment of the tax            listed
   until the taxpayer has        securities and
   cash on hand from sale        commodities;
   of asset                      but the
                                 particular kind
                                 of averaging
                                 suggested would
                                 mitigate this
                                 difficulty for
                                 the taxpayer.
                                 provision for
                                 delay in
                                 payment, with
                                 interest, could
                                 be made.
                                 NO -- as to
                                 with respect to
                                 the year's
                                 change in book
                                 value; but the
                                 particular kind
                                 of averaging
                                 suggested would
                                 mitigate this
                                 difficulty for
                                 the taxpayer.
                                 Also, provision
                                 could be made
                                 for delay in
                                 YES -- as to
                                 all other
                                 assets --
                                 although the
                                 taxpayer has
                                 the option of
                                 paying before
1. It avoids requiring a              NO               YES
2. It avoids requiring new            NO               NO

                         [TABLE D CONTINUED]

                                 PLAN C                  PLAN D
                              (Step-Scale)           (Apportionment)
1. It avoids requiring a              YES              YES
   valuation of listed
   securities and
   commodities at the end
   of each year
2. It avoids requiring a              YES              YES
   valuation of unlisted
   securities at the end
   of each year

3. It avoids requiring                YES              YES
   corporations to notify
   holders of their shares
   in the year's earnings

4. It avoids requiring a              YES              YES
   valuation of assets
   other than securities
   and listed commodities
   at the end of each year

5. The return form,
   compared with what it
   would be if gains or
   losses were treated
   just like other income

   (a) is just as simple         NO -- gains and     NO -- if a
                                 losses must be      taxpayer has
                                 listed and the      both gains and
                                 percentage          losses, this
                                 computed            form is
                                                     probably more
                                                     complex than
                                                     any of the
                                                     other three
   (b) is less simple, but at                             NO
       least as simple as
       existing step-scale
6. It makes unnecessary
   the provisions of:

   (a) Sections 112(b) to          NO                     NO
       112(i) inclusive,
       dealing with
       recognition of gain or
       loss, as to transfers
       made after the new plan
       went into effect

   (b) Most of Section 113,           NO               NO
       dealing with adjusted
       basis, as to transfers
       made after the new plan
       went into effect

   (c) 118, dealing with              NO               NO
       losses from wash sales
       (but in any case, it
       might be well to allow
       wash sales, in view of
       the possible economic
       effects of repressing

   (d) 24(a)(6), dealing with         NO               NO
       losses from sales not
       at arm's length
1. It avoids requiring                YES              YES
   payment of the tax
   until the taxpayer has
   cash on hand from sale
   of asset
1. It avoids requiring a              YES              YES

2. It avoids requiring new            YES              YES

The chief disadvantage of the accrual-averaging plan, compared to any of the other three, is the requirement for valuing listed securities and commodities at the year's end, the computation by corporations of the book value changes per share, and the requirement that payment be made before realization -- but this last objection could be largely overcome by granting a delay in payment for a maximum of, say, six years, with interest (somewhat as at present under the estate tax). On the other hand, this plan has the advantage, lacking in the others, of making it possible to dispense with most of sections 112, and 113 as to assets transferred after the plan went into effect, and with sections 118 and 24(a)(6).


Finally, the adoption of a particular method may make possible the modification or abolition of some other tax, or in some other way seriously affect other parts of the tax system. Thus Plan A, in the writers' opinion, makes it possible to dispense with the undistributed profits tax. The annual accounting of value in the case of listed securities should take care of undistributed profits, as concerns tax justice, and the method of taxing unlisted securities is specially designed to reach undistributed profits. Of course, in neither case does the method satisfy those who wish to see the earnings actually forced out to stockholders in the form of corporate assets (cash or otherwise). Plan A also makes possible the abolition of the personal holding company and improper accumulation provisions.

Summary Table

Table E [see next page] concludes this capital gains and losses section by showing in one table the answers -- yes or no, with practically no qualifications -- for all four plans for all the standards.

The accrual-averaging plan is seen to be about on a par with the others as to revenue, much the best as far as economic effects go, appreciably better than any of the others as concerns justice, somewhat worse as concerns administration, and outstandingly better as concerns relation to other taxes.

                               TABLE E
                 Standard of Relation to Other Taxes
indicates an
("NO"            Plan A       Plan B        Plan C         Plan D
indicates a    (Accrual-  (Realization-  (Step-Scale) (Apportionment)
disadvantage)  Averaging)   Averaging)
1. It makes        YES         NO             NO            NO
   unnecessary a
   tax designed
   to force
   (whether or
   not in the
   form of
   assets) to
   as concerns
   treatment of
   owners of
   concerns vs.
   owners of

2. It makes        YES         NO             NO            NO
   any special
   tax provisions
   for personal
   (section 351)

3. It makes        YES         NO             NO            NO
   any special
   tax provisions
   (section 102)

The realization-averaging plan, compared with the step-scale and apportionment plans, seems to have practically no advantages, judged by this summary table. Actually, as reference to the detailed Table C shows (Section A), it comes closer to justice than the existing step- scale plan. Incidentally, these remarks indicate the danger of relying too much on this summary table alone, since important shadings in the answers are hidden by it.


Depletion Allowances

The matter of depletion allowances has already been thoroughly studied by the Treasury, the present writer understands, so no detailed exposition of the matter will be given here. It will suffice to refer to a paper on this subject, available in mimeograph form, /26/ which describes the recent history of this problem in the federal law, and to note that the percentage provisions are probably costing the government an appreciable amount of revenue compared with the more equitable method of depleting on a March 1, 1913, or cost basis. Moreover the saving in administration supposedly obtained by the percentage provisions is, as explained in detail in that paper, somewhat limited.

Tax-Exempt Securities

To advance the state of knowledge appreciably on this complex subject would require much time and effort. The present section must therefore content itself with referring to the material already available elsewhere, especially the brief statement of the problem in Facing the Tax Problem, pages 303-306, and a somewhat more detailed treatment in the "Income Tax Memorandum" by Blakey, submitted to the Treasury in 1934.

The present writer agrees, on the whole, with the conclusions reached by Blake, except that the importance of the social injustice of the exemption (through its effect on the surtax rate) may be greater than is there indicated. Moreover, the Oliver plan probably involves too many difficulties in the distribution of the money to the states. The straight-forward method of a constitutional amendment permitting federal and state governments to levy non-discriminatory income taxes on each other's future issues of securities seems decidedly the best attempt to make first. Possibly, if it were made a part of a general program for rearranging federal and state fiscal relations, it would be accepted by the states.

Meanwhile, the federal government can and should refuse to issue any more securities exempt either from the surtax or normal tax. The higher interest rate it would have to pay would probably be but a slight increase over the present average rate (since so much of the interest is already subject to surtax though none of it is subject to normal tax), and it would be a low price to pay for the increased justice of the rate schedules and the example it would set for the states, to say nothing of a certain amount of increased revenue.




The use of the corporate form of business raises two disturbing problems in personal taxation, (1) the double taxation of earnings received by stockholders from corporations, and (2) the temporary or permanent escape from taxation of earnings withheld from stockholders by corporations.


The first problem arises when income taxes are imposed on corporations that are not imposed on other forms of business. Removal of the taxes would eliminate the problem. These taxes are in final analysis burdens on the stockholders, especially the common stockholders. /1/ To achieve the ideal of personal taxation, which in this case is to impose the same tax burdens on persons with respect to incomes received through corporations and otherwise, double taxation of corporation income must be avoided.

The method employed for achieving approximate equality is to exempt dividends paid by corporations from part or all of the personal income tax. Best conditions for achieving equality are a proportional tax rate on corporation income and an equally high proportional tax rate on individuals. Income received through the corporation is thus subject to the corporation tax but escapes the equally high personal tax. Equality of taxation is not complete, since (1) corporate losses cannot be offset against individual incomes, (2) capital gains and possibly other items are differently treated, and (3) income that would be non-taxable to individuals because of personal exemptions is subject to the corporate tax.

The exemption of corporate dividends from personal normal tax, provided in the Federal law prior to 1936, gave greater equality of taxation than if no exemption had been allowed, but the maximum equality was not attained since the rate of corporation normal tax was much higher than the rate of the personal normal tax. This discrimination against corporation income was in part offset by the fact that the personal surtax applied only to the amount of the dividend and not to the original corporation income, thus giving savings in taxes equal to the personal surtax on the amount of tax paid by the corporation. The amounts of tax saving for any person depended on the surtax brackets and were greatest for persons with incomes in the highest brackets.

Since 1936 no exemption of dividends to individuals is permitted with the result that when (as in this chapter) the corporation normal tax is viewed as a personal tax deducted at source, there is double taxation of income from corporations.


Corporation earnings are not taxable, under present constitutional interpretations, until distributed by the corporation to its stockholders in cash or property, in obligations, or in stocks that alter the distribution of interests among the stockholders. Earnings may be "realized" and so become taxable through the sale of stocks at enhanced prices to other persons. In this case the earnings themselves have not become taxable but have resulted in a taxable capital gain.

The personal surtax may thus be postponed from the time the income was earned by the corporation to the time taxable dividends are paid or the stock is sold. Although merely postponing payment of the surtax may appear not to disturb eventual tax equality, it does in fact do so, because the individual stockholder, through the corporation, has the use of the amount of the tax as an earning asset during the period of postponement. The government in the meantime may be paying interest on borrowed money.

Aside from the interest factor, which is involved in any tax postponement, part or all of the personal surtax may be permanently avoided by one or more of four methods. First, the dividends may be subject to lower surtax rates or may escape surtax entirely because received in a year when the stockholder's total income is smaller than when the dividends were earned, in a year when his tax return otherwise shows a loss, or in a year when tax rates are lower. Unless the stockholder is in control of the corporation, the year in which the dividend is received is largely a matter of chance.

Second, the earnings may be permanently withheld and reinvested by the corporation, after which the investor may sell his stock to realize the capital gain resulting from reinvestment. Through choosing the year in which he will take his capital gain, he can minimize the rate of taxation applied to it. Furthermore, the tax on the capital gain is affected by the provision for taxing only from 80 percent to 30 percent depending on how long the securities have been held.

Third, the earnings may be withheld and reinvested by the corporation and then lost through business reverses, in which case no tax is collected from the individual. Had he received a dividend, he would have had to pay a tax even if the dividend was subsequently reinvested and lost. If he eventually sells the security, the basis of the stock purchased with the dividend (plus original holdings) will be higher than that of the stock on which earnings were not distributed. Under some circumstances this will ultimately eliminate the avoidance from this factor. Furthermore, if the corporation were permitted to carry over its losses and offset them against incomes in future years this method of avoidance would occur only when the corporation did not thereafter earn income equal to such losses.

Fourth, earnings may be withheld permanently and the stockholder may not sell the stock before his death. The tax is then avoided entirely since neither he nor his heirs pay an income tax on the increase in the value of the stock accruing prior to his death. In case the stock is given away during the life of the owner, he avoids income tax on the capital gain although the donee, if taxable, (charitable foundations, etc., are not) is subject to tax in case of subsequent sale.

No such postponement or avoidance is possible in the cases of proprietorship or partnership businesses since for these the business gains or losses are combined with the other incomes of the proprietors or partners.


The policy that should be adopted to prevent personal tax avoidance through the corporation and to achieve equal taxation of income from investments in corporate and non-corporate business depends on the stage or level of control over assets at which it is believed income taxes should be imposed.

Three levels of control over current business earnings may be distinguished. At the first, or "realization," level, the owner has directed possession of, and (so far as the business is concerned) complete freedom to dispose of, some property that was but recently part of the assets of the business. This occurs when the earnings are distributed to a corporation shareholder in the form of cash, securities of other corporations, or other assets, or when a partner or proprietor draws cash, goods, or other assets out of the business.

At the next, or "management and control," level, the owner has the practical power to determine, or to influence strongly the determination of, the amount of the physical separation noted above. No separation may result but the owner has the power to bring it about if he desires.

At the third, or "beneficial interest," level, the owner neither gets a part of the assets nor has the power to influence the decision as to whether he shall get them. He has merely an increased equity or beneficial interest in the assets to the extent of the share of the earnings to which he would be entitled if they were distributed.

At all levels the owner is better off in terms of economic power or exchange value than if he had no connection with the business at all. At the first level the economic power can be exercised by spending the cash or assets in any manner desired. At the second level it can be exercised to further the owner's interest in the business or to bring about realization. At the third level the economic power lies in the expectation by this and other investors that future realization (or future management and control that can produce realization when desired) will result. This expectation has a value on the market, but that value may be much less in amount than the earning giving rise to it since the expected benefit may be uncertain or far removed.

The federal income tax law requires that the economic power of other than business owners be at the first level before taxable income is recognized, that is, the general rule is that income must be realized before becoming taxable. In the case of the business proprietorship and partnership, however, the law taxes business earnings as personal income whether or not a separation of assets from the business unit has taken place. If not separated, such earnings are usually at the second or "management and control" level, although in the case of partnerships some income is probably taxed at the third, or beneficial interest, level, since control by any person begins to disappear as soon as two or more persons share ownership. In large partnerships minority partners may have very little power to force the separation of assets from the business.

In regard to corporation stocks, the law in general recognizes personal income only at the first or realization level. There are exceptions to the rule. The second level (management and control) is recognized in the case of personal holding companies (section 351) and corporations accumulating surplus for tax evasion purposes (section 102), but only as a voluntary alternative to penalties on the corporation. The third level (beneficial interest) is recognized in the personal income taxation of those stock dividends that alter the distribution of interests in the corporation.

The difficulty in knowing at what level to tax corporation stocks grows out of the fact that some stockholders have the power to control the distribution of earnings while others have little or no such power and still others have some influence but not definite power. The separation of the corporation from the persons of its owners is made complete through a separate legal entity with continuous life. The corporation is not subject to be broken up (as a partnership would be) by the disaffection of some member of the group, and is controlled by the majority of the stock, and usually of the common stock only if more than one kind of stock is outstanding. With many closely held family or even other corporations probably all or almost all of the stockholders exercise sufficient control to bring about a separation of cash from the corporation when desired. In other corporations, however, control is centered in one or a few stockholders. In the very large public corporation, it may be that no stockholder or cooperating group of stockholders holds enough stock to control. Instead, the officers of the corporation may retain control though proxies, in which case very few stockholders may have any effectual voice in determining corporate policies, for example, the payment of dividends.

These common stockholders who cannot exert control, together with preferred stockholders and owners of nonvoting stock, probably constitute a large majority of all corporate stockholders. Their relation to the business is not one of management and control but of beneficial interest or equity. Their relative rights to share in case of distribution of earnings or of capital are defined but they rarely have any power to force distribution. Some improvement in the power of minority stockholders could undoubtedly be made through proportional representation or otherwise but the extension of management and control to all stockholders is a logical impossibility.

Partial exemption of undistributed corporation earnings from taxation to stockholders who lack control to force the separation of assets is in harmony with a definition of income as an increase in economic power, for the reason that the increase in beneficial but, for at least the time being, unrealizable interest probably does not usually result in an equivalent increase in economic power. However, the many stockholders who have the power to cause the separation of corporation assets are in a position of economic power comparable to that of most partners and of proprietors. Unfortunately, no adequate criteria exist for determining when a stockholder has power of management and control or how complete his power is.


The varying and uncertain powers of stockholders, just discussed, are largely responsible for the difficulty of removing discriminations between incomes earned through corporations and otherwise. Some sacrifice of justice, as well as possible undesirable economic effects, is present in the case of all of the proposals that have come to the attention of the writers.

A first group of proposals would follow the major tenet of the present definitions of income with regard to the stage at which corporation income is taxed to the stockholders, that is, taxation at the realization stage. The first proposal in the first group is to do nothing to remove the inequalities that exist when earnings are not distributed. This might be upheld on the ground that while injustice is present it is no more serious than many tax injustices and in view of the difficulties and possible bad economic effects of a real solution no change should be made. Unfortunately, the matter is not only one of justice between incorporated and unincorporated businesses, but one of justice between persons with income of different sizes. If nothing is done, the high personal surtax rates constitute a strong inducement to use the corporation for tax avoidance by persons with large incomes, thus destroying the progressive character of the surtax brackets. An incidental result would be to deprive the minority stockholder of earnings that he would have received if the majority stockholder had not been induced to withhold earnings.

The second proposal in this first group would make accrued capital gains taxable at the death of the owner of the stock or at the time it was transferred as a gift. One of four methods might be employed at death. First, the capital gain might be accrued and taxed as part of the last year's income of the decedent. Second, the capital gain might be included in the income tax return of the beneficiary when the inheritance was received. The improvement from using this method would be slight since the tax rate on the gain in the hands of the beneficiary would bear no relation to the rate that would have been imposed on the decedent. Third, a special tax might be imposed on the capital gain as part of the estate tax at time of death. Presumably the capital gain would be taxed at the highest tax rates applicable to the estate. The resulting improvement might be slight since the estate tax rates imposed on the capital gain would bear no necessary relationship to the rates that would have been imposed had the assets been sold by the decedent. Fourth, a special increment tax might be imposed at death on the increase in value of the capital asset during the life of the decedent, at a rate in no way connected with the amount of the estate or of the income of the decedent. This tax would bear little if any relation to anyone's personal situation and the improvement resulting from its use would be small.

The capital gains loophole in the case of gifts might also be filled by adapting to the purpose one of the four methods discussed for use at death. In the case of transfer either at death or by gift, the first method, that of adding the accrued capital gain to the income of the decedent or donor, would most nearly equalize taxes on capital assets sold before death or gift and those not sold. Unfortunately, the constitutionality of the procedure is extremely dubious as is that of most of the others, although an excise tax on increments is probably unconstitutional.

Some of the methods mentioned for taxing capital gains at death or gift makes any adjustment for interest earned on the tax during the postponement period. Interest might be taken into account by increasing the tax rates applicable to capital gains at sale, gift, or death in proportion to the period of time that the capital asset had been held. The effect of methods of taxing capital gains at the time of death or gift would be to diminish the permanent tax avoidance possible and therefore to diminish somewhat the pressure to retain earnings. It would, however, leave most of the problem untouched since hope for tax relief springs eternal and since postponement takes place for reason other than tax avoidance. The writers believe that these loopholes should be shut but do not consider shutting them as an adequate solution of the problem of tax equality.

A third proposal in this group is to require either by law or by punitive tax the physical severance and distribution in cash to the stockholders of those assets of the corporation that represent current income. This policy takes away management and control of business assets from the corporation and places the assets in the complete personal control of the stockholders. Such a separation of earnings from the business is not required of partnerships and proprietorships. It would be more of a shock to corporations than it would be to partnerships and proprietorships, if required of them, because of the close relations between personal and business life and the greater degree of control by owners that exists in partnerships and proprietorships.

A second major type of proposal would be to secure equality by not taxing partnership (or even proprietorship) income until it is separated from the business. /2/ Due to the closeness of control and the magnitude of the possible tax savings, use of this proposal would probably result in earnings not being taxed until withdrawn by owners to be spent for consumers' goods. Administrative necessity and justice to other investors would practically force a similar exemption of all saved income. This would be in accordance with the recommendations of some economists, /3/ but would be unacceptable (1) from the viewpoint of justice to those who consider income as accretion of economic power and (2) from the viewpoint of economic control to those who wish by income taxation to discourage increasing concentration of wealth and income.

A third major type of proposal is to tax the equities of stockholders in undistributed income regardless of whether or not the stockholders are in position to realize the income either by controlling its distribution or by selling the stocks at a higher price. Three methods are proposals of this type: (1) to tax the distributive share of corporate profits as a part of the income of stockholders while allowing the deduction of the distributive share of losses from their incomes, (2) similarly to tax to stockholders the distributive share of profits with no loss deduction, and (3) to require by law or punitive taxation that all earnings be distributed in the form of taxable stock dividends (either with the present definition of taxability or with a definition making all stock dividends taxable). /4/ The injustice of these methods lies in the probable unsoundness of the assumption that earnings by a corporation increase to an equal extent the economic power of stockholders who have only a beneficial interest in them and are completely lacking in power to force out assets. To the extent that the stockholders exercised control over the distribution of assets or could realize the income by selling the stock at higher prices, the proposals would not violate the theory of income as the increase of economic power.

Whether the distribution of stock dividends would make the beneficial interest of stockholders in corporation earnings more of an accretion of economic power to the stockholder than would merely an increase in the corporation's surplus account is debatable. The answer appears to rest on whether the market value of the prior holdings plus the formal evidence represented by the stock dividends would be greater than the value of the holdings had no stock dividends been issued.

A fourth major type of proposal is to inventory the stocks annually at their market values, taxing increases in value as personal income and allowing decreases as loss. This method attempts to reach only so much of the increased beneficial interest as is reflected in increased economic power to realize cash through sale of the securities. The increase in security prices reflects, among other things, the present value placed by investors generally on the probable future realization to stockholders on account of the current year's earnings. The market thus translates an uncontrollable beneficial interest into economic power to the extent that other investors have faith that the beneficial interest will be realized and are willing to pay for the opportunity of realization. Also, of course, the inventory method takes many other factors besides earnings into consideration since a security price is a resultant of many forces. The method does not result in perfect equality because the stockholder in control of the corporation will be taxed no more than the stockholder holding a mere beneficial interest, although the economic power of the former may have increased much more than the security prices were affected by the corporate earnings. Despite its defects the inventory-of-securities method, if it could be applied ideally, would probably come nearest to the application of a uniform taxing principle and to result in greater tax equality than any other method. The difficulties are not so much in justice theory as in absence of ideal conditions for applying the method.

Certain of the above proposals may be disposed of without further analysis. The plugging of the loopholes in capital gains taxation at death or gift does not conflict with other methods and can be used with any. As previously stated the writers consider it desirable but do not view it as even approximately a complete solution. The proposal to exempt undistributed partnership and proprietorship income seems to the writers to involve changes so contrary to the spirit of present income taxation as not to be seriously considered.

The principal competing proposals that seem to merit more careful analysis are the following:

(1) TO MAKE NO ATTEMPT TO ACHIEVE EQUALITY. This would call for repeal of the undistributed profits tax without the adoption of a substitute. It is based on the idea that while the resulting situation would be unsatisfactory and undesirable, any attempt to achieve equality would have even more undesirable results.

(2) TO INTEGRATE CORPORATION AND INDIVIDUAL INCOME COMPLETELY BY TAXING THE STOCKHOLDERS ON THEIR DISTRIBUTIVE SHARE OF CORPORATION EARNINGS, ALLOWING THE DEDUCTION OF THEIR DISTRIBUTIVE SHARE OF CORPORATION LOSSES. If carried out fully this would contemplate ignoring the existence of the separate corporate entity. Corporation capital gains and losses would be attributed to stockholders distinct from other income, in order to make effective the step scale provisions of capital gains taxation.

(3) TO TAX STOCKHOLDERS ON THEIR DISTRIBUTIVE SHARES OF CORPO- RATION EARNINGS. This method is similar to the previous one, except (1) that the corporate entity is recognized, only the final net results being attributed to the stockholders, and (2) that only earnings are recognized (losses being left to become effective ultimately as best they may through the capital gains and losses provisions).

(4) TO CONTINUE THE UNDISTRIBUTED PROFITS TAX ON CORPORATIONS SUBSTANTIALLY AS AT PRESENT. The undistributed profits tax is a combination of the third proposal in the first group and the third proposal in the third group. Continuation of the undistributed profits tax would not prevent adjustments to remove inequities in the law or to change the rates of the tax. However, it would mean that no changes would be made that would alter the general nature or effects of the tax.

(5) TO INVENTORY THE MARKET VALUE OF SECURITIES ANNUALLY AND IMPOSE A TAX ON INCREASES, ALLOWING DECREASES AS DEDUCTIONS FROM OTHER INCOME. An inventory on this basis could be applied only to securities with an ascertainable market value which would in general limit it to listed securities. Unlisted securities might be omitted but would probably be valued on the basis of book value. Changes in book value would represent the distributive share of corporation earnings and losses.


Many factors must be considered in judging the relative merits of the five integration proposals just mentioned. Following is a list of what are perhaps the most important -- although by no means all -- characteristics of these proposals: /5/


(1) At what level of control over earnings is equality
imposed and to what extent is the equality complete?

(2) Has the taxpayer received funds from which to pay the

(3) Are all corporations treated similarly?

(4) What other justice characteristics are present or


(5) In what direction and to what extent is the payment of
dividends influenced?

(6) How is control over the reinvestment of earnings

(7) What are the effects on corporation capital?

(8) Is capitalization of corporate assets encouraged?

(9) What are the effects, if any, on security markets?


(10) Is the method constitutional under present

(11) Does it result in any conflict of federal and state


(12) Does the method increase the problems of valuation?

(13) Does it present problems of tracing corporate
equities to stockholders?

(14) Does it increase cost of compliance?

(15) Does it give rise to increased evasion?

These characteristics are analyzed in the following paragraphs. /6/


(1) NATURE AND EXTENT OF EQUALITY. The proposal to impose the corporation income tax and personal income tax side by side with no attempt to achieve equality would give rise to inequalities of possibly very great intensity although of undetermined total amount. From postponement alone the taxpayer may through the corporation earn, on his share of the corporation income, for the period during which the tax is postponed, at amounts up to five times as much as would be earned on the share if the tax were paid immediately. /7/ Realization through capital gain 10 years or more later may save up to 70 percent of the tax. Avoidance at death saves 100 percent of the tax. /8/ The actual total savings to taxpayers have not been determined and perhaps cannot be determined. Also the extent to which the avoidance is intentional, although perhaps not significant in measuring tax equality, is important in determining the effects of those methods that relieve the pressure, but is immeasurable. The pressure to postpone is strong in the case of wealthy stockholders.

Taxing stockholders on corporate gains and allowing corporate losses as though the corporation did not exist would merge the two incomes entirely and eliminate the inequalities between the corporation and other forms of business. However, it would do so on the beneficial interest level, and as pointed out above, would operate as an injustice to those stockholders who cannot exercise control over corporate assets.

Taxing stockholders on their distributive share of the undistributed profits of the corporation without allowing any deduction for their share of corporate losses results in somewhat more harsh treatment of stockholders than of proprietors and partners, who may deduct business losses from personal incomes.

However, this is compensated at least to some extent in the deduction of capital losses at the time the stock is sold. Furthermore, if corporate losses were allowed to be carried forward to future years, the difference in tax burdens on incomes received through corporations and otherwise would probably be slight. The charge that injustice results from taxing "incomes" at the beneficial interest level applies to this method also.

The imposition of an undistributed profits tax like the existing one has three different results on equality depending on which of the three options is chosen by the corporation management. If the tax is paid and no taxable dividend distribution is made, inequalities increase since the tax burdens imposed only on the corporate form of business are increased. Stockholders who control the corporation and have large incomes may still gain from postponement or avoidance of individual surtaxes while minority stockholders having small incomes are charged a tax penalty as a result of the decision of the controlling stockholders. If the tax is avoided and earnings are distributed in the form of cash dividends, equality of tax burden is produced except insofar as losses are not deductible. However, the corporation is in this case subjected to a control over its assets not exercised by government in the case of the partnership, which constitutes an inequality in treatment if not in tax burden proper. If the tax is avoided by issuing taxable stock dividends /9/ equality is achieved, again except insofar as losses are not deductible, but on the beneficial interest level. As previously indicated, this beneficial interest may, through the formal issuance of the dividend, have added more to the economic power represented by the stock than if no dividend had been issued and a tax had been imposed on the share of undistributed earnings, but whether this will occur is a matter of particular circumstances. In general, the same remarks apply to the justice of taxing the stockholder on beneficial interests by this method as by that of taxing the distributive share of undistributed earnings.

Inventorying securities at market value and taxing the accrued gain as income, while allowing accrued loss as a deduction, would come nearest of any of the methods to taxing beneficial interests on the basis of the economic power that they represented, and thus to placing their taxation on a basis equivalent to that of partnerships and proprietorships. However, the stockholder who was in a position to control the disposition of corporate assets would not be taxed as heavily as he should be on the basis of his gain in economic power arising from the earnings of the corporation.

(2) PROVISION OF FUNDS FOR TAX PAYMENT. In some integration methods, taxes are imposed regardless of whether the taxpayer has received any funds from which to pay them while in other methods he receives funds from which to pay taxes before the taxes are imposed. Since, aside from the taxation of business income, the income tax law in general taxes only fully controlled income, the taxation of corporation stockholders on increases in economic power /10/ resulting from the increased market price of stocks would place stockholders on a substantial equality with other business owners while placing all business on a basis different from other income; the taxpayer would have the power to realize although he was not given realized income before the tax was imposed. Taxing beneficial interest, however, assures the taxpayer neither of fully realized income nor of the economic power to realize.

(3) SIMILARITY OF TREATMENT FOR ALL CORPORATIONS. Most methods of integrating corporation and individual income taxes would treat all corporations similarly. The annual inventory of corporate securities and the undistributed profits tax do not do so. Similarity of treatment in the annual inventory method would probably be impossible because of the lack of knowable market values in the case of many stocks, especially those not listed on exchanges. Either those stocks would have to be omitted entirely or inventoried at book value. The latter plan would result in the taxation of the distributive share of earnings for such corporations. A substantial difference in the amount of gain subject to taxation might occur under the two methods. In periods of rising security prices, listed securities would probably show greater gains than unlisted securities, while in periods of falling security prices listed securities might show losses when unlisted securities were showing gains. The use of some device for averaging gains and losses over a period of years would reduce the discrimination although it would hardly eliminate it entirely. When the securities were sold, differences in capital gains taxes would also tend to offset the annual differences.

The undistributed profits tax discriminates among corporations in several ways. Liability to the tax is not universal. Banks and corporations in bankruptcy are exempt. A credit is allowed corporations that are prohibited from paying dividends by certain narrow contractual provisions, while others equally restricted receive no relief.

Furthermore, possibilities of avoidance are not equally available to all corporations liable to tax. As pointed out below, /11/ corporations with a capital deficit may not legally pay dividends either in cash or stock. Other corporations may be so loaded with debt obligations that tax avoidance is an inaccessible option.

Again, it is apparently contemplated that under the undistributed profits tax a corporation subject to the tax may pay cash dividends and then secure needed capital for expansion by the sale of stocks or bonds to the public. A somewhat similar result can be achieved through an agreement of the stockholders to reinvest their dividends. Many corporations, however, cannot use these methods successfully. The medium-sized corporation especially faces difficult problems. It may have too many stockholders to make possible a general agreement to reinvest dividends, while at the same time it may be too small to dispose of its bonds or stocks to the general public. Corporations may have so difficult a time securing needed capital that they will retain their cash and pay the tax.

Likewise corporations may not be able to use the option of paying taxable stock dividends. Statutory or charter provisions may prohibit the issuance of taxable stock dividends. To issue preferred stock as taxable dividends on common stock, some preferred stock must already be outstanding. The charter may not provide for the issuance of preferred stock and an amendment to the charter may require unanimous consent of the stockholders. All the authorized preferred stock may have been issued and no provision made for issuing more. Stock issued as a dividend must sell at par or above to give the corporation full benefit in avoiding the tax. Weak corporations may not be able to sell their preferred stock at par. Corporation charters commonly require dividends on preferred stock to be paid in cash, yet the corporation may need its capital so badly that it cannot afford to pay cash. /12/

In all such cases, although the law appears on its face to treat all corporations equally, it does not in fact do so because through no fault of their own some corporations are not in a position to take advantage of the law. In time, corporations will adjust their practices with the imposition of the tax in mind and thereafter little unjust discrimination may exist, but in the meantime considerable injustice may result.

So far as the writers are aware no quantitative study has been made to indicate to what extent corporations are not able to take advantage of the avoidance provisions of the undistributed profits tax.

Although injustices result from applying the undistributed profits tax to corporations, the allowance of exemptions would not necessarily improve the situation. Exemptions would be hard to administer. It would be practically impossible to distinguish between the deserving cases and the undeserving. The allowance of exemptions might encourage corporations to put themselves in position to receive the exemption rather than pay or avoid the tax. Furthermore, if corporations heavily in debt or otherwise incapable of paying dividends were to receive a special exemption from the undistributed profits tax on this account, an injustice would be done to other corporations since the repayment of debt constitutes an expansion of the capital of the corporation in a fashion very similar to that resulting from the building up of surplus accounts.

(4) OTHER POSSIBLE INJUSTICES. In addition to the discriminations previously discussed, other charges of injustices have been brought against the undistributed profits tax. One of these is that a corporation may have a taxable income and be subject to the undistributed profits tax although no actual profit was earned to distribute. For example, in determining the adjusted net income upon which the undistributed profits tax is based, no deduction is allowed for capital losses unless such losses have already been deducted in computing taxable income for the normal tax. Tax justice would be promoted by allowing capital losses as deductions in determining income subject to the undistributed profits tax.

A second complaint is that while some corporations have stable income from year to year many others normally have incomes in some years but suffer losses in others. If all annual earnings are paid out in dividends, the latter corporations suffer a gradual depletion of capital and must eventually fail. Annual earnings are not necessarily true earnings in the long run.

This problem is merely a variation of the general one of averaging incomes and losses or of carrying losses forward from one year to the next. It is probably more serious in the case of the undistributed profits tax than in the case of the normal income tax. Justice would be improved by permitting either the averaging of incomes and losses for several years or the carrying forward of losses in computing undistributed profits subject to tax.

It is complained further that if the corporation taxable income for a particular year is increased by a deficiency assessment in some later year, the corporation is obliged to pay the undistributed profits tax regardless of its desire to distribute all profits in the form of dividends. The Division of Research and Statistics has proposed a plan whereby a special tax of 5 percent is imposed on such deficiency assessment income, with the provision that the corporation shall then have the option of paying out such income in the form of dividends or of combining it with the current year's income or loss. The purpose of the special 5 percent tax is to discourage the deliberate misstatement of income in order to postpone the tax or have the income show in a loss year. If the penalty tax is sufficient to prevent this, the plan should solve the problem satisfactorily.

It is also complained that the provision that all earnings must be paid as dividends within the fiscal year in order to count in the dividends paid credit is a hardship, since it is impossible to forecast the income before the year has closed. One reason for requiring payment within the year apparently was the desire not to lose the equivalent of one year's taxes by postponing the application of the undistributed profits tax. This is a purely revenue consideration. The other reason was the fear that if a period after the close of the fiscal year were allowed for the payment of dividends, holding companies might effect a long delay in the final taxation of dividends to individual stockholders. Thus, if earnings of an operating corporation for 1936 could be paid as dividends in 1937 to the first holding company, if in turn this holding company could pay dividends on these earnings in 1938 to the second holding company, and if the second holding company could delay paying dividends to the final individual stockholders until 1939, a two-year delay in taxes on dividends would result. The more holding companies in the structure the longer the delay.

Some device could probably be designed to permit payment of earnings as dividends after the close of the fiscal year without resulting in a long postponement of taxes on the final stockholders. However, the problem is perhaps not very serious. So long as the corporation has some surplus at the time the plan is put into operation, an overpayment of earnings in the form of dividends in one year may be used as an offset for the following two years. Since, however, the corporation may have two or more loss years in succession a carry-over should perhaps be allowed against succeeding years in which income was earned until the loss was offset rather than simply the next two succeeding years.


(5) INFLUENCE ON THE PAYMENT OF DIVIDENDS. The imposition of an individual income tax with no method of integrating corporation and individual income exerts a powerful inducement to retain earnings in the corporation which otherwise would be distributed.

The integration of corporation and individual incomes through taxing stockholders on their distributive shares of corporation profits removes the pressure to retain earnings. The stockholder has nothing to gain by having the earnings retained since he must pay the tax whether or not the dividends are paid. A mild pressure to pay cash dividends is probably exerted by such methods of integration since the stockholder will wish at least enough cash to pay the tax.

Integration through the inventorying of securities would also remove the pressure to retain earnings in the company. Little if any tax postponement or avoidance would result from retaining earnings since the market value of the securities would be increased thereby and taxable income would appear -- although perhaps not in the same amount as if dividends were paid. Some inducement might be exerted to pay sufficient cash dividends for the stockholder to pay his tax. /13/

The undistributed profits tax constitutes a positive pressure for the payment of dividends in order to avoid the tax. The pressure is relatively light on the first part of the earnings and relatively heavy on the last part. Except in cases where stockholders with very large incomes are in control of the corporation, the pressure exerted by the tax to cause the payment of dividends is more than sufficient to offset the pressure of the personal income tax for the retention of earnings.

The pressure of the undistributed profits tax to cause the payment of dividends does not discriminate between cash dividends and dividends in the form of taxable stock. However, the fact that individual stockholders will be taxed on the dividends received would probably lead them to demand payment of at least enough cash to pay the tax.

The desirability of securing the payment of more cash dividends has been argued from several points of view. Effects on power over reinvestment and on corporation capital are discussed below. At this point may be noted the arguments that the payment of cash dividends tends to dampen the swings of the business cycle. Three arguments have been put forward. (1) "Oversaving" resulting from the reinvestment of corporation earnings is responsible for an excess of production over consumption, which results in crises and depressions. (2) The reinvestment of corporation earnings by directors rather than by stockholders results in a misdirection of capital and consequent overproduction in some lines of business, producing maladjustments that result in depression. (3) Withheld earnings may be kept in liquid form during depression and recovery years (when increased spending is needed) and invested in new plant, during boom years (when less spending is needed), thus accentuating both booms and depressions. The payment of cash dividends to stockholders, it is claimed, reduces business swings from all three sources. Total savings are decreased since stockholders are likely to spend at least part of the dividends for current consumption goods. The allocation of capital investment among concerns and industries is improved since to the extent that the stockholder reinvests his earnings he will choose concerns and industries that have the prospect of the largest profits. Accumulation of idle funds in depression and recovery years and overinvestment of such funds in boom years is prevented since stockholders will presumably either spend their dividends or invest them immediately. The writers do not feel themselves in position to pass judgment on the correctness of either the facts or the conclusions involved in these arguments.

The point of view that payment of cash dividends weakens corporations and thus accentuates depressions is discussed in the section on "effects on corporation capital."

(6) DETERMINATION OF THE REINVESTMENT OF EARNINGS. In the absence of any special pressure, the determination of whether earnings shall be reinvested or not is ordinarily made by the corporation officers and directors subject to whatever control the stockholders may exercise over them. The question has been raised whether the officers and directors or the individual stockholders should make the decision as to reinvestment. In favor of having officers and directors exercise the power of reinvestment, it is urged that only the management really knows the needs of the business. The management is likely to be farsighted, seeing the possibility of future expansion and of opportunities that will require capital, while stockholders may be interested in their immediate dividends at the expense of their best interests in the long run. Furthermore, much apparent reinvestment is necessary to conserve the previous investment and is not a real expansion. In opposition, it is urged that since corporate ownership and risks inhere in the stockholders, it is properly their function to determine individually whether the incomes from their investments shall be reinvested or used in some other manner. The management of the corporation has very strong inducements, other than those benefiting the stockholders, to reinvest earnings. Reinvestment adds to the size and prestige of the company. It reduces the pressure from the stockholders for efficient corporation management since it is easier to earn a return on a given par value of stock with a large volume of capital than with a small one. If the management has interests in other business concerns doing business with the one in question they may personally benefit from reinvestment. Placing the decision for reinvestment in the hands of the stockholders puts the particular corporation in competition for capital with all other corporations and makes more likely the investment of capital in those industries and concerns that promise the best return, a result that is usually in the public interest. The minority stockholder is also released somewhat from the absolute domination of the majority.

The only way in which the integration of corporation and individual income shifts the power to decide whether the earnings shall be reinvested is through its effect on cash dividend policy. Accordingly, the remarks in the previous subhead apply to this one also.

(7) EFFECT ON CORPORATION CAPITAL. Results that are claimed to follow from the loss of cash by corporations are that they are deprived of capital needed for expansion and that they are weakened for meeting the strains of a period of depression, both of which results are claimed to be undesirable. It has been especially urged that new corporations, corporations recently emerged from bankruptcy, and all corporations during periods of rapidly rising inventory costs need to retain their assets. However, apparently no one knows how many or what kinds of corporations in need of expansion have been unable to secure capital. How much capital is needed in the public interest and at what points it is needed are questions that also have a bearing on the importance of the problem. The claims of corporate management that new funds are needed cannot always be accepted at face value.

Weakening of corporations to meet depressions is said to result from the payment of cash dividends because surplus needed to pay wages, fixed charges, and other costs during depression cannot be stored up and the corporation is therefore more likely to fail. While this argument has an element of truth, it is subject to considerable limitation. The strength of a corporation during a period of depression does not depend on the size of its surplus account, which is purely an accounting matter. More important is the ratio of assets to liabilities, the ratio of quick assets to current liabilities, and other financial ratios. If accumulated earnings have been devoted to building up those types of assets that will help the corporation in the depression, the surplus is a source of strength. If, however, earnings have been used to build new plants or otherwise to expand, the surplus may be rather a source of weakness.

Furthermore, it has been urged by some that corporate financial strength in a depression is not necessarily a socially desirable condition. The argument is that a financially strong corporation will not make the adjustments in prices or in products and operations that are necessary if the depression is to be overcome. Quicker liquidation, quicker adjustment to depression situations, and therefore quicker recovery, are claimed to follow when corporations are financially weak. The writers do not know of any study demonstrating the truth or importance of this argument.

Most of the methods of integration will have an adverse effect on the accumulation of corporation capital since they remove the inducement to retain earnings for the purpose of tax avoidance. However, only in the case of the undistributed profits tax is a definite pressure imposed that may result in more loss of needed capital than would also be the case with a partnership. In case taxable obligations or taxable stock are issued as dividends, no loss of corporation capital results, a fact that has not been recognized by some critics of the tax. In case the corporation is unable to use one of these methods of tax avoidance, a loss of cash equal to the tax paid results when earnings are retained. In case the earnings are distributed in cash, the loss will be the amount of the dividend distributed.

Whether or not the distribution of capital in the form of cash dividends deprives the corporation of needed capital depends on the ease with which capital can be raised through security issues to stockholders or the public. Large corporations commonly have access to capital markets and can secure the needed capital by the issuance of bonds or stock to the public. However, corporations that need the capital the worst may be in such financial condition that they cannot readily float their securities on the market.

As previously mentioned, small corporations having only a few stockholders may secure capital through an agreement of the stockholders to reinvest their earnings as soon as received. The option of cash or stock rights may also be used by corporations of all sizes although some problems are involved that would make their use impracticable in some cases.

The problem of retaining capital seems greatest in the case of the medium-sized corporation. It may have too many stockholders to make possible a general agreement to reinvest dividends when received. At the same time it may be too small to dispose of its bonds or stock to the general public through the regular capital market. As a result, the effect of the undistributed profits tax might be to drive out the medium-sized corporation by preventing corporate growth, weakening it financially to meet competition, or encouraging the merger of several corporations into a large one. In any of these cases competition would be discouraged and monopoly encouraged.

(8) CAPITALIZATION OF CORPORATE ASSETS. In absence of any method of integrating corporation and personal income taxes there is no positive pressure on the corporation either to "capitalize" its corporate assets, that is to have outstanding stocks and bonds with a total par value substantially equal to the value of corporate assets, or not to do so. However the pressure not to pay dividends tends to result in the building up of capital assets represented by the surplus account rather than by the par value of securities, since it is easier for the corporation to do nothing and thus to build up a surplus account than to take the steps necessary to issue stock dividends.

Capitalization is influenced by integration methods largely according to their effect on dividend payment. Taxation of the distributive shares of stockholders in undistributed profits and taxation of accrued capital gains would accordingly have little effect. The undistributed profits tax however exerts an influence to bring about the capitalization of corporate assets. If cash dividends are paid the expansion of capital by the corporation must be brought about by the sale of securities either to stockholders or to the public. If over a period of years all earnings are paid out in cash dividends and if all expansion is financed by the sale of stocks and bonds at par, all corporate assets should be approximately reflected in the par value of the securities, -- although securities might be sold above or below par value.

If the undistributed profits tax is avoided by payment of taxable stock dividends or obligations the capital assets of the corporation all tend to be capitalized. The exact identity between the value of the assets and the par value of securities outstanding will depend on whether the stock dividends were issued at par or above or below.

The capitalization of corporate assets has been urged on the grounds that a book value of stock near par is less misleading to investors than one far from par, and is otherwise preferable to a large surplus account. On the other hand, two objections have been raised to the complete capitalization of corporate assets. (1) It is contended that the financial condition of the corporation is weakened by the issuance of prior claims to earnings. To the extent that financing is by bond issue or that dividends are paid in the form of obligations, the ability of a corporation to pay its debts is reduced. However, to the extent that dividends are paid in the form of taxable preferred or common stock or that financing on the market is through the issuance of stock, the ability of the corporation to pay its debts is no less than if the same amount of capital were represented on the books by a surplus account.

(2) It is also contended that the issuance of preferred stock in payment of dividends as an option under the undistributed profits tax gradually reduces (a) the share of the common stockholder in the assets, (b) his claim to dividends, and (c) his control of the corporation. Corporate control is not reduced unless the preferred stock carries voting power, in which case although common stock may lose control, common stockholders need not since the dividend stocks being complained of were issued to them. Only when dividend stocks are sold are the proportionate voting shares of the holders of common stocks reduced. Likewise, the earnings, although less likely to be paid as dividends on common stock because of the larger preferred dividends to be paid, would go to the same persons, unless sale of shares took place.

Another objection is that the preferred stock of the company would cease to be desired since with its increase year after year, dividends on it would not be paid regularly. However, the receipt of preferred stock would be at least as desirable as no dividend at all.

(9) EFFECTS ON SECURITY MARKETS. Methods of integration that encourage the payment of cash dividends, and especially the undistributed profits tax when the cash dividend method of avoidance is employed, make it necessary for a corporation to secure needed capital by the sale of securities. To some extent this might be done privately to the stockholders of the corporation rather than through the public markets. Undoubtedly, however, the use of the public markets for selling securities would be increased. The desirability of securing capital through the public capital markets has been alleged, on two grounds. The first is that thereby the importance of the banker's function in determining the allocation of capital among industries is increased. The second is that the increase in capital is brought under the jurisdiction of the Securities and Exchange Commission and made subject to legal provisions for protecting stockholders.

Another effect on the security markets would grow out of the issuance of preferred stock as taxable dividends. The sale of such stock in order to secure cash with which to pay the personal surtaxes might flood the capital market with a large volume of low grade corporate securities and result in a general lowering of security prices. /14/


(10) CONSTITUTIONALITY. The constitutionality of different integration methods is important since if a method is constitutional it might be introduced immediately while otherwise a constitutional amendment with its uncertainties and delays would be necessary before the method could be introduced.

Most of the integration methods described above are probably unconstitutional, as the courts are rather strict in recognizing the distinct entity of each corporate unit. Complete integration of corporate income with the income of the stockholders, by disregarding corporate entities and following a plan similar to that used in the case of partnerships, would unquestionably be unconstitutional. Likewise, the taxation of the distributive shares of undistributed profits would probably be unconstitutional. The annual inventory of corporate securities would also probably be unconstitutional on the theory that income does not accrue to the stockholder until realized through sale of the security or through receipt of dividends.

No decision has yet been made by the Supreme Court on the constitutionality of the undistributed profits tax, but there appears to be little question that such a tax is constitutional. The constitutionality of taxing different kinds of stock dividends is still uncertain.

(11) CONFLICT OF FEDERAL AND STATE LAWS. Almost all corporations receive their charters from state governments and are regulated by state laws. Requirements of federal tax laws may conceivably involve a conflict with the state corporation laws. Of the integration methods under discussion, only the undistributed profits tax is likely to cause such conflict. If the corporation pays the undistributed profits tax, no conflict arises, since liability for the tax is legally the same as for all other types. If, however, the corporation endeavors to avoid the tax (as is contemplated in the law) conflict with state corporation laws may result. A corporation with a capital deficit is commonly prohibited by law from paying dividends so long as the deficit continues, so that avoidance of the federal undistributed profits tax by paying dividends in any form would violate state law. Elimination of capital deficits by reducing the par value of outstanding stock or revaluing upward the value of the assets may be difficult. Additional legal provisions frequently restrict the distribution of non-cash dividends. For example, several states set limits on the amount of debts that may be incurred by a corporation. In some states constitutional or statutory provisions prohibit or limit issuance of stock dividends.


(12) PROBLEMS OF VALUATION. Most of the integration methods raise problems of valuation which are, however, much more difficult in the cases of some than of others. Complete integration by disregarding corporate entities would not raise any new problems of valuation, but would make more important than at present such valuation problems as determination of depreciation and depletion for corporate income tax purposes since the high level of progressive individual surtax would be substituted for or added to the moderately progressive rates of the corporation normal income tax. The year in which such deductions were taken would become more important. The same comments apply to the taxation of undistributed profits to stockholders.

Making an annual inventory of the value of corporate securities would raise very difficult problems of valuation. In the case of securities valued at their end-of-year market prices, such prices would take on an importance which they do not now have. At the present time, market prices of securities are important for income tax purposes only when securities are actually sold. If corporate securities were inventoried annually, however, prices around the close of a year would be of great importance in the case of all securities, sold or unsold. The temptation to reduce imputed values on a large block of stock by selling a few shares in order to reduce the taxable income of individual stockholders might be very strong. Appraisal methods by the tax administrators might be tried in order to correct market values. Any such correction, however, would lead to great difficulties since the value of securities would become a matter for administrative judgment rather than of objective market value. Many of the evils of property assessment would undoubtedly arise. However, they would not be so serious as in the case of the property tax since errors in annual valuations would be compensated at realization of capital gain or loss. The use of a suitable averaging system would further reduce the incentive to error and the seriousness of the errors.

In case of unlisted securities the determination of book value raises all the valuation problems involved in the taxation of undistributed profits to stockholders and also the determination of other items that might affect the book value of assets without affecting current earnings.

The undistributed profits tax does not raise any valuation question if the earnings are retained and the tax paid or if the earnings are distributed as cash dividends. If, however, the earnings are distributed in the form of taxable stock, it is necessary to place a value on such stock in order to determine the value of the dividends taxable to the individual and allowable as dividend credit to the corporation. Under some circumstances this valuation problem might be a very difficult one even in the case of large corporations if the market for these securities was quite thin. If might be almost impossible in the case of small corporations.

(13) PROBLEMS OF TRACING CORPORATE EQUITY TO STOCKHOLDERS. Several of the integration methods involve the administrative difficulty of tracing corporation income or loss from the earning corporation to the final individual stockholders holding the equity therein. The problem would be involved in the greatest degree in the case of a complete integration of corporation and individual income disregarding corporate entity. It is also involved, although to a lesser extent, in the taxation to stockholders of undistributed profits and in applying the inventory method to the value of unlisted corporate securities by the determination of book value.

Tracing corporate equities involves two main problems, one the legal question of ownership of corporate equities in case non- cumulative preferred stock is outstanding, the other the complications and expense of tracing equities to the final individual stockholders perhaps through an intricate holding company system.

The determination of the legal rights to distributive shares would perhaps be impossible if non-cumulative preferred stock were outstanding since the rights of holders of these stocks and all junior securities would be affected by the dividend policy adopted by the board of directors. To meet this uncertainty in ownership the tax law might proceed on the assumption that unless non-cumulative dividends are not paid within the year they will not be paid. This would place the burden of any inaccuracy of this method on the common stockholders.

After the equities of various kinds of stock in the earnings were determined (assuming that they could be), many complications would be met in tracing undistributed income and loss to final individual stockholders, especially if intricate holding company arrangements were involved. Although possible it might be very expensive. The determination of book values of securities would require a similar process. In all of these methods any reopening of the corporation's statement of income by tax authorities might necessitate a reopening of the individual returns of all stockholders, although approximately the same results in most cases might be secured by permitting the difference to be entered on the stockholder's return for the current year -- permission that involves the risk of deliberate misstatement of corporation income.

This difficulty of tracing corporate equity to stockholders does not arise in the cases of the undistributed profits tax.

(14) COST OF COMPLIANCE. Cost of compliance would not necessarily increase under any of the integration methods. However, in those cases where other administrative difficulties are met, the government might be expected to place on the corporations involved an additional burden of furnishing information, thus transferring part of the administrative cost to the corporations themselves.

(15) EVASION. The value imputation methods of integration (namely, disregarding corporate entities, taxing undistributed profits to stockholders and making an annual inventory of the values of corporate securities) involve the question of what amount of income should be imputed to each stockholder in case the stock was held for only a fraction of a year. For perhaps a majority of stockholders this problem would not arise or at any rate would not be difficult because the stock would ordinarily be held by the same person throughout the year or would be sold only once during the year. However, for a considerable minority of stock, turnover is frequent; in some cases the stock may be held only a day or a fraction of a day. The attempt to allocate a portion of the annual income to individuals according to the length of time they held the stock would become an almost impossible administrative task in the cases of such stocks.

To meet this difficulty, provision might be made for the whole year's income to be allocated to the individual holding the stock at the end of the year. This might lead to some sale of securities at the end of the year by very wealthy individuals to less wealthy ones in order that the latter, who would be subject to less tax, might bear the tax instead of the former. The tendency would be somewhat to decrease security prices just before the close to the year and somewhat to increase prices after the opening of the year. It would appear unlikely that the avoidance of the tax by this method would be serious over the long run especially since upon sale of the securities a tax would no doubt be applied to the amount of capital gain not previously taxed.

Somewhat more difficult might be the case of the sale of unlisted securities with an agreement to repurchase at the same price after the beginning of the year. In the absence of a public market for the securities, such sales might be made year after year with an almost complete evasion of the tax. However, the law might provide that repurchase of a security within a given time limit would subject the repurchaser to tax on the distributive share as if he had held it over the year-end -- a provision somewhat like the existing wash-sale provision. Considerable cost might be involved in administering such a provision.


The evaluation of policies proposed to solve the problem arising because corporation earnings are withheld from the stockholders is a difficult one because none of the proposed methods is satisfactory. The fundamental difficulty is that modern corporate business is a combination of individualism and collectivism. /15/ None of the proposals faces this difficulty satisfactorily, undoubtedly for the reason that the two elements are inseparably mixed at the present time. It may be suggested a priori that the proposal that comes nearest to separating the two elements and treating them differently is likely to be the best solution of the problem.

In the following paragraphs, the different methods are discussed in reverse order of preference. Some other observers will [n]o doubt place the proposals in a different order.


(1) SUMMARY OF CHARACTERISTICS. When no attempt to integrate is made, incorporated and unincorporated businesses are treated unequally in that stockholders are temporarily or permanently free from taxation on undistributed corporation earnings, while partners and proprietors are subject to immediate taxation of partnership or proprietorship income. The opportunity to postpone or avoid personal surtaxes presents a strong inducement for wealthy stockholders to have corporate earnings withheld from the stockholders. Withholding earnings destroys the progressive character of personal income taxation; encourages the accumulation of corporation capital beyond the amounts needed in business; and diminishes the use of banks and security markets as guiding conduits for capital investment, with the result that their functions of allocating capital among industries and concerns are made less important. In most other respects, however, the management of the corporation is not subject to tax pressures tending to regulate business decisions.

(2) CONCLUSION AS TO DESIRABILITY. This policy must be considered the least desirable of the group. During the last period of business expansion, observers frequently called attention to the large volume of reinvested corporation earnings. The personal surtax rates during that period were low compared with their level today. With today's high level of surtax rates, it appears almost certain that the fraction of corporate earnings withheld from stockholders would be greatly increased wherever wealthy stockholders could dominate corporation policy. This withholding and reinvesting, it should be noted, would not be for business needs but to save stockholders from taxation. The desirability of the economic results is highly dubious, although the writers have less knowledge than is necessary to arrive at firm convictions with regard to them. From the viewpoint of tax justice, however, the destruction of progressive personal taxation by tax-free withholding and reinvesting is clear. It is the belief of the writers that so long as personal surtax levels are at anything like the present height -- and they do not desire a return to the low levels preceding 1932 -- the pressure to use corporations to avoid the personal income tax is so great that some action to relieve that pressure is necessary.


(1) SUMMARY OF CHARACTERISTICS. This method is almost certainly unconstitutional. It does not cause any conflict of federal and state laws. The method treats the corporation as a purely individualistic (as contrasted to the collectivistic) institution. It would completely equalize treatment of corporations and other businesses by taxing at the beneficial interest level. It imposes a tax on stockholders although no funds are provided for its payment and although perhaps most stockholders could not control dividend payment. It treats all corporations similarly (although they are not all similar), removes the inducement to retain earnings, probably exerts a mild pressure to pay enough cash dividends so that the stockholder may meet the tax, leaves with the officers and directors the power to decide whether earnings should be reinvested, removes the inducement to accumulate corporation capital but gives no inducement not to accumulate capital, does not encourage the capitalization of corporate assets, and has no particular effects on security markets. Its use would increase the importance of problems of valuation, and would raise an almost prohibitively difficult problem of tracing corporation incomes and losses (with capital gains and losses kept separate) to individual stockholders. It would probably raise the cost of compliance and is subject to the danger of evasion through the sale of securities at the end of the year with an agreement to repurchase.

(2) CONCLUSION AS TO DESIRABILITY. This method is not desirable from the viewpoint of tax justice because under our business system the income and losses of the corporation are both legally and economically separated from those of the stockholders, so that the attempt to disregard corporate entities conflicts with reality. Beneficial interest would be taxed even if corresponding increase in economic power had been experienced. The method is not desirable from the administrative viewpoint because of the labor that would be involved in tracing income and losses and capital gains and losses separately through holding company structures to the final individual stockholders.


(1) SUMMARY OF CHARACTERISTICS. The undistributed profits tax is probably constitutional, raises points of conflict between federal and state laws since avoidance of the tax through the payment of dividends may be a violation of the state law, strongly encourages the payment of dividends in some taxable form, tends to increase the capitalization of corporate assets, makes corporation valuation methods more important, does not raise any serious problem of tracing corporate equities to stockholders, probably does not materially increase the cost of compliance or avoidance, and probably will not result in serious evasion.

TAX PAYMENT OPTION. If the tax payment option is chosen, the undistributed profits tax does not improve integration of corporation and individual income taxes with reference to stockholders having large incomes and those having small incomes. It reduces the advantages (or increases the disadvantages as the case may be) of the corporate form as compared to the partnership or proprietorship form. It does not involve conflict of federal and state laws. In general, no tax is imposed where funds for payment have not been received although the failure of the present law to recognize capital losses is an exception. Not all corporations are similarly treated since banks are exempt from the tax as are corporations in bankruptcy and corporations with certain contracts limiting dividends payments. A corporation may suffer an actual loss but show a taxable income and thus be obliged to pay an undistributed profits tax although no profit is available for distribution. Carrying forward losses from one year to the next is not permitted. The power over reinvestment of earnings is retained by the corporate directors although less earnings are left to reinvest due to the loss of each in paying the tax; accordingly the increase in corporation capital may be somewhat diminished.

CASH DIVIDEND OPTION. If cash dividends are paid, there is great gain in integration of corporation and individual income. This integration is not complete when corporations lose money or as long as a corporation tax is imposed for which no offset is provided against personal tax. Funds are made available to the individual stockholder for payment of his individual surtaxes on the dividends. The power over reinvestment of earnings is taken from corporate directors and given to individual stockholders. Expansion of corporation capital is more difficult then with no undistributed profits tax and in the case of medium-sized corporations may be very difficult indeed. Capitalization of approximately all corporate assets is almost inevitable. The use of the security markets is encouraged because of the necessity of raising capital through them. The control of banks and of government over the expansion of corporate capital is increased. No important problems of valuation or of tracing corporate equity to stockholders arise except as valuations in determining corporation income become more important. Costs of compliance are increased insofar as necessity arises for sale of securities to secure capital. No great amount of evasion of the tax by individual stockholders is to be expected.

TAXABLE OBLIGATIONS AND STOCK OPTIONS. The use of this option gives the same degree of integration as the use of cash dividends. The stockholder is not provided with funds from which to pay the tax but is given a separate legal right which he may sell without decreasing the apparent volume of legal rights previously held. Whether receipt of this separate legal right increases his economic power is open to question and probably depends on the circumstances. The determination of the reinvestment of earnings is retained by the directors. Corporation capital is not impaired. Corporation assets are almost completely capitalized. Security markets may be disturbed by the sale of obligations or securities to secure cash with which to pay the personal surtaxes. The problems of valuing the dividends as income to the individual and as dividends paid credit to the corporation are considerable. No new problem of tracing corporate equity to stockholders is raised except as may be involved in valuing dividends. Costs of compliance (or rather costs of intended avoidance) are increased by the cost of issuing the obligation or security. No particular problem of evasion is presented.

(2) CONCLUSIONS AS TO DESIRABILITY. The writers confess a considerable degree of uncertainty in judging the relative desirability of the undistributed profits tax. The tax is difficult to judge, partly because it combines three essentially different approaches. In the case where the tax is paid by the corporation in order that earnings need not be distributed, the effects on tax equality are worse than if no tax were imposed. In the cases where the tax is paid in taxable stock, the effect is essentially one of taxing the distributive shares of undistributed profits, and is subject, although possibly to lesser extent, to all objections to such taxation. In the cases where the tax is paid in cash, the corporation is subject to an economic wrench not imposed on other forms of business and one having uncertain results.

The difficulty of judging the undistributed profits tax arises also out of a lack of information on certain aspects of the tax and its effects, notably the following:

(a) To what extent are corporations unable for legal or emergency economic reasons to issue taxable dividends at all, and how long a transition period will be necessary before this will cease to be a serious problem?

(b) To what extent will corporations find it impossible to issue taxable stock dividends as a method of tax avoidance?

(c) To what extent will corporations, especially those of middle sizes, find difficulties in raising capital in case they are unable to use taxable stock dividends?

(d) Does the forced payment of cash dividends to stockholders have on balance desirable or undesirable economic effects?

The importance of information on these points is as follows: If practically all corporations can now pay dividends either in cash or in stock and if the period of adjustment for the remainder will be short, the rate of the tax might be made very high, which would eliminate the evils of the option to retain earnings and pay the tax. If the forced payment of cash dividends is on balance desirable, the taxable stock option could be removed and actual assets forced out to stockholders, eliminating the evils of taxing undistributed profits to owners of the equity. In such case the resulting injustice in comparison with partnerships and proprietorships might be considered sufficiently unimportant to waive.

In the absence of such information, the undistributed profits tax must be viewed as one containing three options, none very desirable, and subject to choice by the directors, which choice may or may not be in the public interest. If the undistributed profits tax is retained, it should at all events be amended to permit the deduction of capital losses and the carrying forward of net losses to future years.


(1) SUMMARY OF CHARACTERISTICS. This method of integration is in effect a partial application of the method of disregarding of corporate entities. However, it does not disregard them entirely, but taxes to the stockholder the net increase in beneficial interest or equity due to the corporate earnings. While it has many of the characteristics of complete integration, including probable unconstitutionality, there are important differences, most important of which are (1) the tax equality produced by its adoption is not as great, and (2) the tracing of corporate equity is not nearly as difficult since (a) the separate identity of receipts such as capital gains does not have to be maintained, and (b) the amounts need to be traced to final stockholders only in years when the corporation shows an income.

(2) CONCLUSIONS AS TO DESIRABILITY. This method is subject to the same objection of taxing increases in beneficial interest regardless of economic power as is the complete integration of corporation and individual incomes. However, the administrative difficulties are much reduced. For this reason the method, although not satisfactory, is preferred to that of complete integration.


(1) SUMMARY OF CHARACTERISTICS. This method of integration is probably unconstitutional. It does not raise any problem of conflict between federal and state laws. It does not insure that the taxpayer will have received funds before paying the tax, but (in the case of listed securities) taxes only gains in economic power. It treats differently corporations with listed securities and those with unlisted securities, removes the incentive not to pay dividends, and may exert a mild pressure for the payment of enough cash dividends to pay the tax. It has no particular effect on the reinvestment of corporate earnings or on the accumulation of corporation capital, and has no effect on the capitalization of corporate assets. It might affect the security market by encouraging manipulation to lower prices at the end of the year, raises a difficult problem of valuation for listed securities where the market is thin, raises the problems of tracing changes in book values to stockholders in the case of unlisted securities, would probably result in an increase in the cost of compliance, and in the case of unlisted securities involves the possibility of evasion through sale of securities at the end of the year with agreement to repurchase.

(2) CONCLUSIONS AS TO DESIRABILITY. In principle, the annual inventory of securities is a method of reaching increases in economic power and not merely in beneficial interests. It is a possible method of meeting the difficult problems of capital gains taxation as well as those of securing equality in taxing incomes from corporate and other sources. This approach is fully valid for the latter purpose only insofar as changes in beneficial interest are reflected in security prices. It has an advantage over the taxation of pure beneficial interests in that the tax is applied (when conditions are ideal for the use of this method) only to an actual net accretion of economic power, although the plus and minus quantities that make up the net accretion come from a variety of sources in addition to the change in beneficial interests due to current earnings. The method comes perhaps nearer than any other to harmonizing the taxation of beneficial interests and the taxation of economic power. It is not perfect because the market valuation of the increase in beneficial interest probably will be less than the increase in economic power accruing to stockholders in control of the corporation, since only when the controlling stockholders' self interests are the same as those of the other stockholders is the value of earnings to people who cannot control the corporation likely to be as great as to people who control it.

Despite this defect, the inventory-of-securities method appears to come nearest to the application of a uniform taxing principle and nearest to tax equality in terms of economic power than any other proposal. The difficulty lies not in its characteristics under ideal market conditions but in the lack of those conditions. Given a continuous body of savers (with changing membership, of course), given a situation where no mass movements were sufficiently large to upset the equilibrium of the body of savers to the individual members, given sufficiently wide knowledge of the condition of every corporation to insure marketability of its securities, and the results, while not perfect, as previously indicated, would perhaps be superior to those capable of achievement by any other method. Unfortunately, the assumed facts are commonly absent. Many corporation securities are unmarketable at prices that they would command if the corporation were widely known. The multiplicity of corporations and the security of investors' time and energy make it inevitable that numerous corporations will not be widely known. Somewhat more widely known securities will have a "thin," highly erratic and easily manipulated market. Only widely held securities will have a broad market. Furthermore, the assumption of a continuous body of investors in a state of equilibrium is fallacious. Great and often irrational mass movements sweep over the security markets resulting in wide price changes. Subject to rapid and eccentric shifts, the price movements make illusory the economic power apparently existing in the ownership of securities.

Some improvements in the situation can doubtless be achieved so that a combination of the inventory method and the taxation to stockholders of undistributed profits might be fairly satisfactory. Corporations may be classified roughly into two groups in accordance with certain outstanding general characteristics. The exercise of at least partial control by stockholders in general is a characteristic of small, closely held corporations, while it is lacking in large publicly owned corporations. The wide marketability of securities is a characteristic of large publicly owned corporations while it is lacking in small corporations. In the case of small, closely held corporations, the taxation of undistributed profits to stockholders (through adjustment of book values or otherwise) would cause relatively little unjust taxation of non-control-carrying beneficial interests. In the case of large, publicly held corporations, the inventorying of securities would cause relatively little taxation of unrealizable (through sale to others) beneficial interests of stockholders. The two sets of corporations would, of course, not be treated alike. In periods of rising security prices stockholders in large corporations would be discriminated against while in periods of falling security prices stockholders in small corporations would be discriminated against. Yet if there were a fairly rigid division of corporations into the two types, probably largely noncompetitive and serving different purposes in the economic system and probably catering to different types of investors, the injustices would perhaps not be too great. They would tend to compensate in the long run in any event, and could be ameliorated by adoption of an averaging system.

Of course the situation is not as simple and satisfactory as the assumptions make it seem. As previously stated the security markets are full of imperfections and the investors full of irrationalities. Furthermore, corporations are not divided, even approximately, into the two groups assumed above. They are of bewildering variety in size, closeness of control, and marketability of securities. To follow the combination plan of taxing changes of inventory values and taxing distributive shares (or increases in book values) would mean injustice and hardship. However, such devices as the averaging of several years gains, and provision for delay of tax payment in taxing distributive shares could undoubtedly be devised so that even with present security markets and kinds of corporations a fairly satisfactory result would be achieved.

Furthermore, looking at other countries, and to the future here, there is much reason for believing that security markets will be more rational in the future so that market values will carry an assurance of economic power that they do not have today. Likewise, there is reason to believe that corporations may -- especially with a little help from tax laws and changes in corporation laws -- develop into the two groups suggested. Accordingly, it may very well be that plans for present and future tax policy should be designed to move towards a combination of inventory and distributive share methods of taxation as the best available solution of integrating corporation and personal incomes.


Since the more desirable methods are very probably unconstitutional, the writers believe that an effort should be made to amend the constitution. In the meantime, they believe that for the time being the undistributed profits tax should be continued on substantially its present basis and at its present rates (but with amendments to take care of capital losses and carry-over of losses). Special efforts should be made to ascertain the facts as to the operation and effects of the tax, so that the entire problem may be reviewed in another year or two. Integration, involving as it does the problem of capital gains and losses among other things, is a matter so complex and yet so little explored thus far that intensive study will for some time to come result in discovery of facts of importance to policy. Particularly should study be made of the points described as needing exploration.

Policy With Respect to the
Undistributed Profits Tax

Since the passage of a constitutional amendment will take time, the immediate problem relates to the policy that should be adopted with respect to the undistributed profits tax. Following is a brief statement on the writers' position.

The undistributed profits tax should not be repealed at this time, chiefly for two reasons. (1) The pressure of present personal surtax rates to cause withholding of corporation earnings is extremely strong. Some method to reduce that pressure, especially during periods of business prosperity, continues to be urgently needed. (2) The tax has not been in operation long enough for anyone to learn whether the undesirable results claimed will materialize on a significant scale. Practically nothing is yet known of the actual working of the tax or of the reaction of corporations to it. The country is not going to be ruined by the application of the tax for the few years necessary to find out how it is really going to work out, or until a constitutional amendment can be passed to make it unnecessary.

However, certain amendments should be made. Several are to bring the adjusted net income nearer to true corporation income, since a corporation should not be required to distribute nonexistent earnings.

(1) Capital losses should be made fully deductible in computing adjusted net income. In the present situation a corporation may suffer a heavy net loss because of capital losses and still be penalized for not distributing earnings in dividends.

(2) A carry-over of losses should be allowed in computing adjusted net income. Under the present law, a corporation may be losing heavily over a period of years and still be obliged to pay either the tax or dividends. The carry-over should be for as long as there are any losses left to carry; for administrative reasons, however, it may be desirable to limit the carry-over to perhaps five years.

(3) The dividend carry-over should be made indefinite instead of only two years. The requirement that dividends be paid in the year earned is a hardship in that forecasting earnings may be impossible. If any excess dividends can be carried forward to a later year the injustice is minimized. However, a two-year carry-over is not sufficient since a corporation may have two consecutive loss years. Administrative reasons may make it desirable to set a limit of perhaps five years to the carry-over.

(4) Many of the objections to the undistributed profits tax might be met, in large part, by allowing a small portion of the income to go free of tax. It may not be unwise to allow the corporation a little elbow room in the form of an exemption of 5 or 10 percent of earnings. Allowance of such an exemption would not necessarily mean less of a tax or less distribution, as the rates on earnings retained above the 5 or 10 percent might be raised to a higher level than at present. It is recommended that 5 or 10 percent of the earnings be exempt only at a nominal rate.

Several changes that have been proposed do not seem desirable to the writers. One is the proposal that additional allowance should be made for corporations having heavy debts. Although the undistributed profits tax sometimes causes hardship to corporations that are heavily in debt and obliged to repay it on schedule, it would be dangerous to give special exemptions for debt. If debts incurred in the future were made exempt, the result would be a wave of borrowing, since by borrowing and then paying back the corporation could accumulate capital just as truly as through accumulating surplus, and the personal surtaxes could be avoided just the same. If only debts incurred in the past were granted exemption, the distinction between deserving and undeserving corporations would be a very difficult one to make. The exemptions granted by the law in the case of contracts prohibiting dividend payment thus give such corporations an advantage in accumulating capital. The proposal of the Division of Research and Statistics to impose a 10 percent tax on the amount of such credit seems a reasonable one.

Another proposal that should not be followed is one that would exempt earnings invested in new plants. To allow such an exemption would not only mean the death of the tax for all useful purposes, but would give a stronger encouragement than before to uneconomic investment. New plants should stand on their own feet and not be bolstered by a tax exemption.

A third proposal that has been made is to increase the tax rates substantially. The writers believe that to make the tax fully effective in producing tax equality it must be heavy enough to force taxable dividend payments. Eventually the tax rates would have to be very heavy, for this purpose, if the tax were retained as a permanent part of the system. However, the present is a period of transition and experiment. The tax causes much hardship, how much or how little perhaps no one has any clear idea. Certainly some corporations are not able to pay dividends and thus avoid the tax. To increase the rates at once to fully effective levels is a dangerous procedure, at least until more information is gained on the extent of the hardships. Therefore, the writers think that any upward adjustment should await further knowledge, especially since any adjustment at all may prove unnecessary, depending on what is done toward amending the constitution. They would of course favor an upward adjustment of rates to compensate for any exemption or decrease in rates granted for a small part of the earnings, as suggested in (4) above.