Date 25 January 1940
Author Carl Shoup
Title Effects of Taxation on National Income
Description Internal staff memo, Division of Tax Research, U.S. Treasury Department
Location Box 34; Economy in General; Records of the Office of Tax Analysis/Division of Tax Research; General Records of the Department of the Treasury, Record Group 56; National Archives, College Park, MD.


Division of Tax Research
U.S. Treasury Department

Carl Shoup

Draft of

January 25, 1940

                          Table of Contents

     I. General Considerations

        The Meaning of the Question

        A. Effect on National Income Through Increasing Consumption

           1. Possibilities of Increasing Consumption by Changes in 
              the Tax System

              (a) Taxes That Vary Positively With the Amount of 

              (b) Taxes That Have No Connection With Volume of                

              (c) Taxes That Vary Inversely With the Amount of                

              (d) Tax Elements Yielding No Revenue

              (e) Summary of Consumption Analysis

           2. Effect of Change in Consumption on National Income

        B. Effect on National Income Through Increasing Investment

           1. Possibilities of Increasing Investment by Changes in 
              the Tax System

              (a) Taxes That Vary Positively With the Amount of 

              (b) Taxes That Vary Positively With the Results of the 

                  (i) Safe Investments
                        Effect on Willingness to Invest
                        Effect on Ability to Invest

                  (ii) Risky Investments

              (c) Taxes Having No Direct Connection With the Amount 
                  of Investment

              (d) Tax Elements Affecting the Mechanism of the Market 
                  for Investment Funds

              (e) Tax Elements Yielding. No Revenue

              (f) Summary of Investment Analysis

           2. Effect of Change in Investment on National Income

        C. Net Effect of Changes in Consumption and Investment

     II. Specific Tax Features

              (a) Upper bracket Personal Surtaxes

              (b) Loss Carryovers

              (c) Averaging of Income

              (d) Capital Gains Taxation

              (e) Consolidating Corporation Returns

              (f) Taxation of Dividends

              (g) Corporation Undistributed Profits, Capital Stock or 
                  Excess Profits Taxation

              (h) Depreciation Allowance

              (i) Payroll Taxes

              (j) Sales and Other Excise Taxed

              (k) Estate and Gift Taxes


Current interest in promoting the growth of national income has led to the question whether any particular elements of the tax system serve to retard that growth substantially. A common approach to the question has been through the effect of taxes on consumption or new investment, with special reference to such like (a) upper bracket personal surtaxes, (b) loss carryovers, (c) averaging of income, (d) capital gains taxation, (e) consolidated corporation returns, (f) taxation of dividends, (g) corporation undistributed profits, capital stock or excess profit taxation, (h) depreciation allowances, (i) payroll taxes, (j) sales and other excise taxes. /1/

It is impossible to give a fully satisfactory answer to this question. Present knowledge is insufficient to demonstrate conclusively just which of the possible obstacles to the growth of national income are the ones which, at any particular time, must be removed before national income can expand. Furthermore, much more research and study will be necessary to determine beyond doubt the relative effects of various kinds of taxation.

The following discussion endeavors to open up the problems raised by the question and to suggest the answers that appear to be most tenable.

/*/ Prepared by the Division of Tax Research, Treasury Department. [Draft of January 25, 1940 -- C.S.C.]

/1/ These were the items mentioned in Question No. 8 of a confidential memorandum of February 17, 1939, containing questions then being studied by the Fiscal and Monetary Advisory Committee.


A. STATEMENT OF THE PROBLEM. -- In a general way, the problem at hand may be stated as follows:

The removal of what elements of the tax structure would increase the national income compared to what that income would be if those elements remained?

The removal of some of the retarding elements is likely, however, to decrease the total revenue produced by the tax system, even though it increases the national income. The question, to be meaningful, must therefore specify whether any such decrease is to be made up from non-tax sources such as borrowing, or instead from tax sources.

In this analysis the latter assumption will be made; the inquiry is thus confined to tax matters, leaving the question of taxes versus borrowing and other similar inquiries for other analyses. If it were assumed that revenue losses were to be covered by borrowing, it might be found that the influence of substituting borrowing for taxation would overshadow the influence of differences among the several tax elements. The assumption that revenue losses are to be made up from taxation is thus the more fruitful for present purposes, although it is also the more detailed, since it requires that the answer state not only what element is being eliminated, but what element is being used as a revenue substitute. An exception to this requirement will occur in the few cases where the total revenue of the system will not decrease even if there is no replacement. This may occur either (a) because the element being eliminated yields no revenue (e.g., a prohibitive import duty), or (b) because the yield of the rest of the system is increased sufficiently as a result of the increase in national income.

The national dollar income may be increased while the national real income is increasing in a different proportion (or possibly even decreasing). Presumably, it is the real income, not the dollar income, that the question refers to; hence, the conclusions made with respect to dollar income should be adjusted, wherever necessary, to make them applicable to real income. However, so long as tax decreases are replaced by tax increases, not by borrowing, and so long as the present condition of underutilization of labor and equipment exists, the conclusions concerning dollar income will probably need not significant adjustment to real income.

The tax system may affect the growth of the national income through its effect on consumption or investment. It is assumed, in the present discussion, that consumption resulting from investment (as when a bricklayer consumes his wages that he gets because the investment is made) is covered by the analysis of investment, so that the "effect on consumption" refers here to direct effect on personal consumption, not indirect effect working through the effect on investment. Likewise, investment may be increased because consumption is increased. Generally, in the analysis below, this kind of increase in investment will be assumed to be covered by the remarks on consumption.

"Consumption" is taken in the sense of production, purchase, or rental for use on personal pleasure instead of either for resale or for use in producing goods or services for sale. Thus, the purchase of coal to be used up in producing electricity is not consumption but the renting of a dwelling or the building of it by the user is, for purpose of the present discussion, consumption. "Investment" is taken to mean "the current addition to the value of the capital equipment which has resulted from the productive activity of the period," /1/ including inventories in the term "capital equipment." "National income" can be produced by either consumption or investment. Thus if consumers increase purchases, and entrepreneurs thereupon increase production by just enough to keep their inventories and other capital equipment at the hitherto prevailing level, consumption has increased, but there has been to investment. If the increased purchases are met simply by letting inventories run down there has been an increase in consumption offset by a decrease in investment, and the increment to national income is zero. If the increased purchases are met by producing enough on INCREASE inventories and/or other capital equipment, both consumption and investment are increased.

Alternatively, "investment" might be defined in a broader way to include not only additions to capital equipment but also replacements -- e..g., amounts spent out of depreciation reserves, or to replenish stocks. This definition is less useful for present purposes, however, since (a) "investment" and "consumption" could not be added, to get "national income", without danger of double counting, /2/ and (b) the question at issue is growth of national income, not simply the maintaining of the existing national income level.


An attempt to ascertain what elements of our tax structure substantially retard the growth of the national income through their effect on consumption involves two steps:

1. Testing each reasonably possible tax element to ascertain whether any of the existing elements could be removed, and replaced by tax elements that would produce the same amount of revenue, with a net result of an increase in consumption.

2. Ascertaining whether such increase would cause a growth of the national income.

1/ J.M. Keynes, The General Theory of Employment Interest and Money, p. 62.

/2/ We ignore here the refinement whereby, in Keynes's treatment, "investment" (but not his "not investment") does include expenditures to make good the depreciation that is bound to occur whether or not the machines, etc., are used.

/3/ After allowing for the effect on tax revenue of any change in national income caused by the substitution of the latter element.


It must be emphasized that the present discussion, to have any meaning for policy-forming purposes, must, at some point before it is concluded, compare one tax element with another -- never leaving the analysis in terms of one tax element alone, unless it is the rare kind that yields no revenue. The technique of abstracting from the revenue aspect, common in economic discussion, is permissible so long as the conclusions are recognized to be only partial conclusions, not fully developed for policy formation. In a memorandum of the present kind, however, the effects of the removal of an existing tax element must be taken to include the effects of whatever substitution is made, and only tic net effect can be considered as the answer. Since borrowing has been ruled out as a substitute (in order to confine the discussion as much as possible to taxation matters), the answer must be in terms of the net effect of substituting one tax element for another. This requirement makes the topic more cumbersome to handle and involves awkwardness in exposition but my avoid leaving the reader in mid-air.

However, in the initial stages of analysis below, only the existing tax element will be considered; at the conclusion of this section the effects of the possible substitute measures will be considered and an attempt made to reach answers in terms of net results.

The discussion will be largely in terms of the removal of part, not all, of the tax in question, since this seems to be the more realistic approach. Thus in discussing the cigarette tax the question will be, what effects might be expected from a reduction, rather than the elimination, of the tax. The size of the reduction assumed is of course important; for purposes of the first part of this discussion it may be somewhat arbitrarily put at 25 per cent of the existing tax rate. (If, upon further consideration, it seems desirable to expand the analysis to include instances of larger decrement or complete elimination, that can be done, though the answers probably grow even less trustworthy as the size of the decrement increases).

Another reason for analyzing in terms of decrements rather than total repeal is that tic effects are likely to change, gradually but substantially, as the decrements increase. For example, it might be possible to remove two billions of revenue from the chief consumption taxes and get substitute revenue from income taxes and death taxes, with a substantial net increase in consumption; but for the next two billions any possible shift might result in very little change in consumption. The consumption-tax rates would be falling so low that tax relief would not increase consumption much, and the rates on incomes and estates would be reaching such heights that tic increments in tax would be checking consumption appreciably. Hence marginal approach to the problem is essential.

From the point of view of consumption, tax elements that yield revenue may be divided into those where the amount of tax payable by the consumer:

(a) varies positively with the amount of consumption in general - e.g., a general retail sales tax - or with some specific the of consumption - e.g., a tax on alcoholic spirits.

(b) does not vary directly with the amount of consumption -- e.g.., a tax on wages or salary received, and the personal income tax except so much of it as is a tax on investment income.

(c) varies inversely with the amount of consumption -- (e.g., an income tax insofar as it taxes investment income, and a death tax.

Another group of tax elements yield no revenue (at least during the period in question) but may nevertheless affect the amount of consumption. Such for instance is a fear that taxes will be higher in the future.


Preliminary to getting a well proportioned view of the influence exercised by existing Federal, State, and local taxes of the kind that vary (in revenue) positively with the amount of consumption, it will be useful to indicate how complicated their effects on consumption may conceivably be. It will be observed that no general rule may be laid down with complete assurance. However, as will be shown subsequently, the solution of the problem is not in fact so hopelessly indeterminate as this skeletal analysis might seem to threaten.

1. A decrease in the rate of a heavy tax, the tax on alcoholic spirits for example, will probably induce most persons who have been drinking to consume a larger volume of spirits.

A. The increase in consumption of the taxed spirits may be small enough so that the consumers' money outlay on spirits (including tax in money outlay) is smaller than it would be without the decrease in tax. Common observation suggests that this is the most likely result (though where existing rates are very high, as with spirits and cigarettes, the price may possibly have reached an area wire consumption is highly sensitive to changes in price). The consumers thereby have money available to increase their consumption in other directions also and to increase savings. If the article in question is one of mass consumption it seems reasonable to suppose that most of this available money will be spent on consumption.

B. The increase in consumption of the taxed spirits may be large enough so that the consumers' money outlay on spirits is about the same as it would be without the decrease in tax. If this is so, there is no need to examine the consumers' expenditures in other directions or their savings except for the factors noted in (D) and (E) below.

C. The increase in the consumption of the taxed spirits may be so great that the consumers' money outlay on spirits is greater than it would be without the decrease in tax. In this case, there is less money available than otherwise for other consumption or for savings.

D. Whether A, B, or C above is the case, the fact that a greater volume of spirits is being consumed will inevitably change certain other forms of consumption, namely, those that are correlated in demand. The consumption of carbonated water, for instance, will presumably increase, since the demand for this is complementary to the demand for spirits, while the consumption of beer, which is to some extent rival to the demand for spirits, will, to that extent, tend to decrease.

E. The necessity, in the present brief discussion, of dealing with consumers in the aggregate runs the danger of overlooking significant changes within the group. For instance, in D above, where consumers as a whole spend the same amount on spirits as without the tax decrease, it might be that those particular consumers who spent more would correspondingly cut down their other consumption while those who spent less could increase their savings, the net result being a decrease in total consumption even though money outlay on liquor was unchanged.

2. Some persons -- not many -- will probably consume about the same volume of spirits as without the decrease in tax. For this group tide effect on consumption exercised by the tax must be sought in the rearrangement of the pattern of consumption of other things and of savings.

3. It is conceivable that a very few persons will consume a greater volume of spirits without the decrease in tax than with it -- possibly because of the satisfaction derived from displaying the ability to consume such a costly article. In practice, however, this case can be ignored as unimportant.

In transforming the outline above into a usable description and analysis of the United States tax system as it exists today, we may start by noting that the only taxes of importance in this group -- taxes that vary directly with consumption -- are those on spirits, beer, cigarettes, gasoline, motor vehicles (state license taxes), housing accommodation (state and local property taxes on buildings as distinguished from land), imports in general, and (in about half the states) retail sales in general. Any one of the numerous other taxes on consumption directly represses consumption so little and supplies so little total tax revenue that it may be disregarded for purposes of the present study.

Probably in none of these taxes does a decrement in tax fall in sub-groups 2 (consumption of taxed article unchanged by tax reduction) or 3 (consumption of taxed article decreased by tax reduction) to a significant extent.

Under sub-group 1 (consumption of taxed article increased), category D (affecting correlated demand) is probably of significance only with respect to decrements in taxes on motor vehicles and housing accommodation, and, as concerns the former, the effect is almost insignificant because the tax is, on the average, so light.

Whether the decrement in any given tax falls under category A (money outlay on taxed commodity de creased by tax reduction), category B (money outlay unchanged) or category C (money outlay increased) is almost wholly a subject for guesswork based on general observation. The few statistical studies that have been made of demand schedules for specified articles are for the most part not concerned with the heavily taxed commodities; moreover, the techniques employed in deriving the schedules appear to be open to some question.

The remarks in the following paragraphs are therefore to be taken merely as conjectures of doubtful validity, based simply on general observation.

The existing rate is so high (relative to the retail price) in the present taxes on spirits, cigarettes, imports subject to duty, and so much of the property tax as may rest on the dweller, that we may assume that in those cases a 25 percent cut in tax rate would decrease the money outlay on the taxed article little if at all.

For example, it would seem that where a package of cigarettes sells at retail for 18 cents, a cut of 1-1/2 cents in the Federal 6- cent tax might, if fully reflected in the retail price (which would then be 16-1/2 cents), lead to an increase of 9 percent in amount consumed, which would be enough to maintain the money expenditure on cigarettes at its former level. Of course the fact that the rate is high makes correspondingly high the decrement in expenditure represented by the reduction of the selling price; but this tendency is probably counterbalanced almost in full by the degree to which physical consumption is ready to expand under this reduction in price. No high degree of confidence can be placed in this conclusion, but in default of anything bettor, it will be employed in the analysis following. Using it, we can accordingly dispense with a discussion of changes in other consumption or saving (with the qualification noted in E above), on the additional assumption that money outlay on the taxed articles would not be significantly increased either. The property tax on dwellings does have special features of importance, however. The lowering of rentals of dwellings and the increase in construction of owner-occupied homes that would be expected to follow the tax reduction might take some time to materialize; even under the strongest impetus, plans and decisions take time, and the impetus itself would not be very strong unless the taxpayer could be convinced that the reduction in tax was to hold good over a period of ten or twenty years. A number of years of actual tax reduction might be needed to convince him of this.

Toward the other end of the scale, where the money outlay on the taxed commodity would be decreased considerably, we may put the 25 per cent decrements in rate of tax on retail sales in general, gasoline, and motor vehicles. Possibly this statement attributes too great an inelasticity of demand (within the range considered) to gasoline.

The beer tax might be placed with this group, but there is some question whether a reduction of 25 per cent in the tax rate would be fully enough reflected in the retail price to cause an appreciable decrease in money outlay.

Where would the money go that would be released by the tax reductions affecting retail sales, gasoline, and motor vehicles (and, over the short period, dwellings)? Some of it would be in the hands of entrepreneurs, since in all three taxes much of the revenue comes from taxes on sales to entrepreneurs, but the direct effects in that case are on investment, not consumption, and will be considered in a section below. As to the rest of the tax reduction, it seems likely that much the greater part of the money released would go for increased expenditures on consumption, since it would be in the hands of persons of fairly low incomes. The release would be, for any one consumer, distributed in small amounts almost daily over the year (excepting the automobile license charge, and the dwellings tax) and would not be noticeable enough to encourage a revised program of savings -- e.g., the taking out of more insurance or the revision of more or less well defined goals in savings bank deposits. These observations imply that for persons of low incomes the savings element in their total outlay is more rigid than the consumption outlay. This implication admittedly rests on no very secure basis. Where savings are negative, as they may be over long periods for a large section of the very low income groups (the difference being largely covered by a somewhat chronic default on obligations toward the grocer, the landlord, etc.), they can scarcely be considered the more rigid element in total outlay; and in those households where consumption outlay is carefully budgeted, small driblets of tax decrements might easily accrue as windfall savings, especially for the first year or two.


The amount of revenue collected from some taxes varies in no direct manner, either positively or negatively, with the amount of expenditure on consumption. In this group fall the taxes levied on the things entrepreneurs buy, including labor; in the United States the important cases are the customs duties on certain articles, the motor vehicle license taxes on commercial trucks and buses, the employers' part of the payroll tax, the retail sales taxes insofar as sales of equipment and supplies are considered to be sales at retail, and the gasoline tax on gasoline purchased for business use. Taxes imposed on the volume of assets, either all assets or some specified type, likewise belong here (the property tax on business property is by far the most important case but the capital stock tax is also important).

Of course all those taxes bear some resemblance to the taxes levied on sales or production; as sales or production increase, there is a tendency for the tax to increase, since the business must buy more labor, gasoline, etc., and may have to enlarge its plant or its capital stock. But because there may be an appreciable time lag and because in any case the relation of increased cost to price policy is not so close as with taxes on production or sales, it seems better to treat these taxes, in a section below, as charges on investment rather than consumption.

Among the non-business taxes the only important ones in the United States to fall in the present group (taxes that do not vary directly with volume of consumption) are (1) the employees' tax on payrolls; (2) that part of the property tax on dwellings that falls on land rather than buildings (this point is admittedly open to considerable debate); and (3) the personal income tax (with a reservation as to the taxation of investment income -- this point will be explained in the next section). The poll tax is an unimportant illustration. For reasons to be given below the death taxes do not belong in this group.

A decrease in the employees' payroll tax would affect consumption simply by releasing money to the workers. In accordance with some of the points brought out above, it may be considered, tentatively, that almost the whole of any such reduction would go to increased expenditures on consumption.

A decrease of the tax on land used for dwelling purposes may be assumed (subject to correction on further study), in accordance with the generally accepted theory, to result in no decrease in the rental price of dwellings or increase in the amount of construction of dwellings. Thus such a decrease in taxation would affect consumption simply by releasing money to the landowner. In view of the large proportion of dwellings in the United States owned by their occupiers, it is likely that a substantial part, if not the greater part, of this tax relief would go to increase consumption, but here our knowledge is so imperfect that this kind of a guess can scarcely be worth much.

A reduction in so much of the income tax as does not fall on investment income would likewise increase consumption somewhat, simply by releasing money to the taxpayer. (So too would a reduction in the taxation of investment income; but it would do something else also -- it would increase the reward for saving and investing to spend later as compared with the reward for spending now, and hence is analyzed separately in the following section). In principle, the release would be especially significant for consumption if the tax reduction were concentrated in the lower-income and middle-income groups, but in practice it might not be very significant, since the existing tax, especially Federal tax, takes only a small proportion of the low incomes and lower-middle incomes.

A reduction in the death tax is comparable to a reduction in the tax on investment income. It increases the reward for saving and investing (a point covered in the next section) and also releases money for either consumption or saving. Probably the amount of this released money that would in fact be directed to consumption is not significantly large.


A decrease in a tax that varies inversely with the amount of consumption will of course tend to DISCOURAGE consumption. The income tax, if it taxes income saved as well as income spent, and also taxes the fruits of saved and invested income, cuts the reward for saving and investing compared with the reward for spending. In so far as saving and investing varies positively, not inversely, with the rate of reward, an increase in such an income tax tends to increase consumption and decrease saving and investing. Conversely, a decrease in the tax tends to decrease consumption and increase saving and investing. Thus the present income tax may be considered to some extent a "tax on saving and investing" as the spirits tax is a "tax on expenditure on spirits." The reaction of the taxpayer is however less predictable in its direction and much slower to take effect. A decrease in the rate of tax on spirits would surely be followed, at once, by some increase in consumption of spirits, but a decrease in the income tax rate might or might not, so far as this present point is concerned, tend to cause an increase in saving and hence a decrease in consumption, and the tendency might not become significantly strong for a year or two or even more.

Death taxes and gift taxes are taxes on savings in the sane sense as an income tax, though the timing of the tax is, for the individual, more irregular. The direction and timing of the effects, on consumption, of a reduction in those taxes are even less predictable than with the income tax, so far as concerns the reward for saving.


The only important tax measures to be put in this class are the marginal elements in a few customs duties. A more significant element is not a tax measure at all but simply a fear that taxes will be higher within a few years. A taxpayer who holds this belief may restrict consumption in order to build up a liquid reserve.


Evidently it would be possible to alter the tax system, without any change in the tax revenue, so that over the short period, -- say a year or two -- and probably over the long period as well, a substantial increase in consumption expenditures would be a direct result. Whether the increase in consumption would tend to increase the national income will be considered in the following section; and whether these alterations in the tax system would adversely affect investment and, through it, national income, will be considered in a later part of this memorandum.

A substantial decrement in the tax was defined above in terms of a percentage reduction -- say 25 per cent -- of the rate. But in order to point out the specific alterations whereby consumption could be significantly increased, the decrement may better be measured in absolute number of dollars. A 25 per cent decrease in the property tax on dwellings, even if limited to the buildings as distinct from the land, would pose a much larger revenue-replacement problem than would a 25 per cent decrease in the cigarette tax.

It will not be possible here, however, to carry forward an analysis so concrete. About all that pay be said is that a substantial increase in consumption could be achieved within a short time by transferring about one billion dollars in total from the taxes on spirits, cigarettes, gasoline, imports of consumers goods, retail sales, payrolls (employees' tax), and lower ranges of personal incomes to the taxes on middle and higher ranges of personal incomes and the taxes on estates and gifts. Such a transfer would necessitate a reduction of between 25 per cent and 50 per cent (nearer the latter, probably) in the rates of the former taxes, and an increase in revenue of perhaps 75 per cent from the latter taxes. Over a longer period, consumption might be further increased by transferring a half billion or so from the property tax on dwellings, and from the automobile license tax, to the property tax on land, the income taxes and the taxes on gifts and estates, though there is doubt whether this further increment to the income tax would check consumption less than the tax it supplemented. Whether a net direct increase in consumption could be achieved by further decrements and increments in these or any other taxes seems, on the basis of present indications, highly doubtful.


Granted that consumption could be increased in the manner outlined above, would national income be increased thereby? (It will be recalled that for the moment we are abstracting from the repercussions of the changes in the tax system on investment and the consequent effect on consumption). This question is chiefly one of general economic theory, with special reference to business cycle theory and money and banking theory. In view of the highly unsettled state of doctrine in these fields, one's answer will depend largely upon which "school" he finds most attractive --whether, for instance, he attaches the importance that Mr. Keynes does to liquidity preference as a force holding the long-term rate of interest at a level too high to allow full employment, or on the other hand is, with Mr. Hawtrey, more concerned over the rates of change in inventories of entrepreneurs as affected by changes in the short-term rate of interest, or is, in contrast to the "monetary" theorists, more occupied, as is Mr. J. M. Clark, with the greater rate of change in producing capital equipment that is caused by a given rate of change in consumer demand. It seems best in this memorandum, which must be considered preliminary in many respects, to indicate no decided preference for one view or the other, but to confine comments to points that are probably valid and important in any case.

1. The increase in consumption would almost surely not be net by entrepreneurs' simply drawing on stocks. A permanently higher level of consumption would probably come to be expected by them, and they might even increase production enough to keep their stocks at a higher level than before, in view of the higher level of consumption. Thus the increase in consumption would surely increase national income somewhat. Special studies in some trades, particularly those of the heavily taxed articles (spirits, cigarettes, gasoline, etc.) would be useful on this point. A general reservation must be made with respect to that part of the increase in consumption that shove up in an increase of imports. Over the short period, at least, this kind of consumption probably does not increase the national income of the importing country (aside from the value added by importing merchants and manufacturers and subsequent distributors), since the money income goes abroad.

2. It may be doubted that the increase in consumption would be substantial enough to lead to enough investment in DURABLE capital equipment to make an appreciable showing for the economy as a whole and thus bring into play the acceleration principle upon which Clark, Harrod and others lay so much emphasis; but this statement is scarcely more than a shot in the dark, and here, too, special studies of certain industries might reveal much of interest.

3. There seems to be a distinct possibility that over the long period -- say a decade or more -- the increase in consumption will hamper the "deepening" of capital equipment (an "increase in the amount of capital employed for each unit of output" /1/), as contrasted with the "widening" of it ("the extension of productive capacity by the flotation of new enterprises, or the expansion of existing enterprises, without any change in the amount of capital employed for each unit of output"). /2/ Thus there may be a tendency to prevent a rise in the standard of living -- a rise perhaps of a kind that cannot be reflected in national income totals, under existing methods of computing national income, since it is a rise that is reflected primarily in lower dollar selling price per unit, and hence may be accompanied by an unchanged, or even lower, aggregate dollar volume of profits, interest, wages, etc. (the components of national income).

/1/ R.G. Hawtrey, Capital and Employment, p. 36.

/2/ Idem.



The discussion of the effect on investment will parallel in some respects the one above on consumption. The elements will be tested to ascertain whether replacements that would result in an increase in investment are possible. Then it will be considered whether an increase thus brought about would increase the national income.

Moreover, in considering the first of these questions, the discussion can be divided in much the same way as that on consumption. Certain tax elements vary positively with the amount of investment --e.g., taxes on the purchase or holding of capital equipment. Other taxes do not vary directly with the amount of investment -- the consumption taxes and wages taxes discussed above. But from this point onwards the investment discussion diverges from that of consumption.

In the first place there is no category of taxes that "vary directly but inversely" with the amount of investment, in the sense that this phrase has been used above. It might be thought that a tax on consumption would be such a tax; but a tax on consumption is like no tax so far as its effect on the relative rewards of spending now, and saving and investing TO SPEND LATER. If indeed the saving and investing is undertaken with no idea of ever spending the saved amount and the fruits thereof, a consumption tax does alter the relative rewards of spending now, and saving and investing. In the present discussion, it is assumed that practically all saving and investing is done with a view to later spending, either by the investor or his successors.

Secondly, a new category must be introduced: taxes that vary directly and positively with the RESULTS that are the aim of the investment -- e .g., taxes on profits. An analogy in consumption taxation would be a tax that varied positively, not with the amount spent on consumption but with the satisfaction from the consumption, so that a consumer buying an automobile, for instance, would not pay any tax if he were disappointed but would pay if he were pleased (or pleased more than he had expected to be). Finally, taxes may affect the functioning of the investment market for bringing savings and opportunities for investment together; there seem no very closely analogous effects in the consumption market.


These taxes are the ones noted on pp. 14-15 above, levied on the things entrepreneurs buy, or on the possession by them of instruments of production (or on the corresponding financial element, such as capital stock). A reduction in such taxes directly stimulates investment, since it makes investment cheaper.

Most of these taxes concern chiefly the short-lived items of capital equipment -- customs duties, "retail" sales taxes, gasoline taxes, personal property tax. A reduction in these taxes could stimulate investment even though there were no assurance that the reduction would be a lasting one. On the other hand, a reduction of the property tax on long-lived capital (excluding land, discussed below) and a reduction of motor vehicle licenses would probably not have much effect for a few years unless in some way the taxpayers were at once convinced there would be no change in policy.

A reduction of the tax on land might possibly result in some increase in investment, but on the whole it seems safer not to count on this result.

As to the payroll tax paid by employers, the general assumption has been that it encourages the use of labor-saving machinery, and from this it might be inferred that the tax encouraged investment. It probably does encourage the indirect or roundabout use of labor; whether this means an increase or a decrease in the total of investment for the period under consideration is a complex question that needs special study.

A reduction in such taxes as those on imports, "retail" purchases of office equipment, machinery, etc., and gasoline confronts a special difficulty, so far as increasing investment is concerned: the few items, out of the total capital equipment of a firm, that would be affected may not be readily adjustable in amount without changes in the rest of the equipment of so substantial a nature that it would not pay to expand the total investment in the firm, at least over the short period. It is, for instance, conceivable that some firms would not increase their total investment if the gasoline they used were given to them entirely free, to say nothing of a mere reduction or repeal of the gasoline tax. The cost of the trucks, drivers' hire, perhaps an inability to expand output at all except by large increments - such things as these may make the demand for the taxed commodity very inelastic. Of course the tax reduction would then release money that could be used for investment or consumption, but in practice there seems a strong probability that for some time it would be used for neither.



The chief tax in this group is the income tax so far as it reaches interest, dividends, profits, rentals, and royalties.

EFFECT ON WILLINGNESS TO INVEST. -- Even in the safest enterprises, or the safest securities, the income tax tends to repress investment insofar as the amount of investment depends on the absolute level of the reward, for the tax tends to loner that level. A decrease in the tax would amount to a rise in the rate of interest (interest paid to the investor), at least over the short period. Presumably the tendency would then be for savings to increase, and for investment to increase. The increase in investment could come not only out of the increased savings induced by the prospect of a higher reward and created by reducing consumption, but also out of hitherto idle funds -- the rise in the rate of interest paid to the investor might well induce him, not only to consume less of his income, but also to decrease the amount of his wealth that he would prefer to hold in cash.

In the lower ranges of income even the marginal rate of tax is so low that a decrement of, say, 25 per cent would probably have no appreciable effect. With the investment yield being considered subject simply to a marginal (bracket) rate of l0 per cent, for example, a 25 per cent decrease in tax rate means that the investment return has been increased only by slightly more than 2.5 per cent of the former return -- for example, from 3 per cent to 3.083 per cent. Even where the marginal rate is close to 20 per cent, as under the present corporation tax, a one-quarter decrement would look small when cast in terms of an increase in the rate of interest. True, the corporate profits are taxed again when they are paid out as dividends. But in many large corporations the men who make the decision whether to reinvest the earnings or to let them lie idle or to pay them out are men who have most of their wealth placed elsewhere than in the company's stock. Under such conditions they may have in mind chiefly the corporation rate and may be considering little or not at all the individual income tax rates, in deciding whether reinvestment of the company's profits promises enough net return after taxes to be worth making.

Individuals subject to marginal rates of 30 per cent or 40 per cent or more, and even individuals on lower levels if they have a substantial part of their wealth invested in corporations that they control, would probably show an appreciable response to a decrement of one-quarter or thereabouts in the tax rate, viewed as an increase in the rate of interest on safe investments. However, as with changes in the property tax on durable assets, the investor would have to be convinced that the decrease would not be repealed within a short time. Under present circumstances it is difficult to see how he could be convinced before reduction had been maintained for some years.

The possible reduction can, as we have seen, be appreciable (in terms of an increase in the net rate of interest) only for large- income taxpayers; but the relation between the amount of saving and the rate of reward for investing is probably pretty loose for those persons. The reward rate may have a closer connection with the strength of the desire to hold wealth in the form of cash, but even there the relation seems uncertain, since the strength of this desire depends, as Keynes points out, partly upon the difference between the existing rate of interest and the expected future rate, and the decrement of tax would do little or nothing to affect that difference.

Moreover, for a certain type of investment return, namely that realized over a term of two years or more in the form of capital gains, the maximum marginal rate is now only 15 per cent. The prospect of this kind of investment return (tax considerations aside) absorbs a fairly large proportion of high-income savings. For that block of savings there is not much more that can be done in the way of reducing income tax rates as an inducement to invest.

The exemption of state and local government bonds from Federal income tax is a tax element that, if removed, would undoubtedly tend to cause some increase of safe investment in private business. The matter is of little practical importance from the short term, so long as already outstanding bonds are not to be taxed. Over the long term its chief importance is in connection with risky investments, to be discussed below. Moreover, insofar as removal of the exemption tended to decrease state and local borrowing and hence state and local investment, there would be a negative entry in the national income account.

On the whole, so far as safe investments are concerned, it seems unlikely that even a reduction of, say, one-fourth in marginal rates of Federal and state individual and corporation taxes would have a very substantial effect in inducing increased investment as a result of an increase in the net reward offered. At least, this conclusion seems valid over the short term.

EFFECT ON ABILITY TO INVEST. -- A substantial reduction in income tax rates would of course leave more money available either for consumption, hoarding, or investment. But it seems unlikely that under present conditions the appearance of this newly released money (or rather that part of it not destined to be consumed in any case) would result in increased investment. The immense amount of money idle at present indicates that investors are waiting either for a raising of the rate of interest (the "liquidity preference" idea) /1/ or for prospective investments to become ire safe. As explained above, the former goal is not likely to be achieved by decreasing the tax on the reward of safe investments; the latter goal is discussed in the following section. Neither is likely to be brought nearer by simply releasing money for investment through tax reduction -- indeed, the former goal tends to recede under such circumstances, since the pressure of this newly available money against the supply of existing safe investments would tend to lower the interest rate still further. Here the problem becomes extremely complex, however. The fact that the interest rate starts to drop below 3 per cent as the newly released money comes on the market may so change some investors' attitudes about the possibility of its going up to 4 per cent in the future that they will now be willing to invest at some point below 3 per cent when they were not before, and the result may be that the newly released money may be divided between an increase in investment (assuming that there were less-than-3-per cent safe prospects that had gone begging before) and some increase in the total hoarded, with some drop in tile rate of interest. But on the whole it seems unsafe to count on much investment, at least over the short term, as a result of making more money available through income tax reductions; and the same conclusion probably applies to estate tax reductions.

/1/ So far as this point is concerned, they are waiting not because they would not be content with 3 per cent if they felt sure that nobody would ever get any more, but because they fear that, some time after they have invested, the rate of interest will rise to its "normal" level and the capital value of their investment thereby fall.

Over a longer period the results may be very different; we should then expect that the "deepening" of capital equipment, which depends in a very direct manner on the rate of interest, would be increased by the release of money for investment and the consequent pressure on the rate of interest. This deepening is, as indicated above, of fundamental importance with respect to the standard of living.


Where there is a considerable degree of uncertainty what the enterprise will show for investors, an income tax on the return will increase the degree of risk. The willingness to invest will be decreased in much the same way as the willingness to purchase a lottery ticket win there is an increase in the percentage of the government's "take."

A substantial decrease in the bracket rates, middle and upper, of individual income tax and of the corporation tax rate would therefore probably cause an appreciable increase in investment -- provided, again, that the decrease had been in effect long enough to give investors confidence that it would stay for some time. This statement, however, is necessarily a highly tentative one. We do not know how much of an increase in safety the idle money is waiting for, nor do we know how numerous are the hazardous investment opportunities that are waiting to be made safer.

No doubt much of the money now in state and local bonds would be invested in risky enterprises or would be responsive to a lessening of the risk, if tax exemption were removed from those bonds.

In addition to the rates of tax, certain features of income taxation tend to increase the investor's risk. Such are the refusal to allow the loss of one year to be fully offset against the gains of other years (the Federal tax allows this offset in a restricted form) and the refusal to allow capital losses to be offset in full against other income.

A decrease in death taxes would probably have some effect in lessening the risk of investments, but so little is known about this problem that it seems best not to count on any such effect.


A reduction in the consumption taxes noted on pp. 11-12 would affect investment to some extent by releasing money for that purpose, but in the aggregate the result would probably not be significant. Most of the money thus released would be spent for consumption (see pp. 11-14), and of the remainder, some would doubtless be hoarded (see pp. 28-29).


The extent of investment depends largely on the success with which the investment market brings together the owners of funds and the opportunities for investment.

As a general rule, anything that tends to block transfers on the investment market tends to block investment. The tax elements in the United States that tend to block transfers are the security transfer taxes and the fact that capital gains are taxed only upon realization instead of on accrual.

The security transfer taxes are so low that it seems doubtful that they block investment to an important extent.

The realization technique for taxing capital gains may, on the contrary, have some such effect. Generally, of course, if the holder of a security thinks it is going down he had better sell, so long as the capital gains tax is less than 100 per cent, since his thinking that it is going down implies that he can repurchase at a lower price than he sells at, and he will have a new, and higher, basis on which to compute gain or loss on any subsequent sale. But he loses a part of his investment in paying tax, and thus cannot buy back as much as he sold; he may have only the vaguest ideas about a second sale and the tax to pay thereupon; and his thought that the stock is going down is an estimate in odds rather than an assured feeling. Finally, the fact of having to pay the tax nay induce an irrational reaction.

If the investor is considering selling and immediately buying another security (switching), there is no doubt that the realization technique tends to prevent the sale, if only because of the loss (rather, loss earlier than otherwise) of part of his capital in tax payment.

This particular problem offers many complex possibilities that cannot be explored here, /1/ but it is probable that the realization technique both delays some transfers that would otherwise occur and, when losses have accrued, hastens some transfers.

/1/ See the chapter on capital gains in the Blough-Shoup report of 1937 to Under Secretary Magill.

It is still less clear just how a tax element that checks transfers only when profits have accrued will check investment; at least it is difficult to distinguish this checking effect from the effect of the tax rates themselves as reduces of reward and increases of risk. One possible way seems to be that the mere checking or delaying of transfers will make it more difficult to pass a security along from the original investor through a chain of more and more conservative investors as the security ripens and gains in stability.


A fear of higher tax rates in the future may induce a hoarding of cash that would otherwise be available for investment. Probably this influence is stronger than in the parallel case discussed above under consumption, since the taxes that are large enough to endanger one's cash position apply only to wealthy persons (with the exception of the property tax, which, however, gives no indication of rising sharply over the next five years or so); and it is among these people that hoarding (because of tax uncertainty) is likely to come out of usual savings rather than through restricting consumption.

A special case of this kind exists when the taxpayer knows, not that the tax rates are going up, but that in at some future year he will have to pay much more tax than currently, owing to the occurrence of some event that is the basis for tax liability. The estate and inheritance taxes levied at death are the best examples. Where the date of the occurrence is largely within the taxpayer's power to determine -- as with the gift tax -- this point is of much less importance, since it is then not necessary to get into a liquid position until very shortly before the date the tax is due.


There are a number of tax elements whose removal or diminution would encourage investment, but the encouragement has several points of contrast with the stimulus that tax changes could give to consumption.

First, the changes in consumption would occur almost at once, while the encouragement to investment would need a long time -- several years -- to become very effective.

Second, while the decrements of one-quarter might well have an appreciable effect on consumption, the investment reaction to a measure of such size is much less uncertain and might be negligible even over several years, so powerful are the other, non-tax factors that control investment. Thus it is conceivable, for instance, that a 25 per cent cut in income tax rates would not have a significant effect on investment, while a 50 per cent or 75 per cent cut would have a very great effect.

Third, while investment may be more difficult to stimulate than consumption by changes in the tax system, it does not face the same narrow limits of expansion that consumption does. The increase in consumption varies (directly) within limits set by the amount of money released to the consumer; the increase in investments has no such limit.

Specifically, a substantial increase in investment after, say, three or four years, could probably be induced by a drastic decrease in the upper brackets of the income tax (except capital gains, where the top rate is already down to 15 per cent). A further increase would probably result if the loss provisions were liberalized (carry- over, and deduction of capital losses from other income). Over a still longer period of time, the elimination of future issues of state and local bonds would have a further encouraging effect, and somewhere along this time schedule a reduction in estate taxes might make itself felt. The revenue loss from these measures would probably have to total half a billion dollars a year or more for several years before any substantial counterbalancing effect appeared in the form of an increased national income. (This figure rests on no detailed computation; it is simply an impression with respect to the general order of magnitude). But when the increase did finally come it might well be very substantial indeed -- conceivably much more so than could be achieved through encouragement of consumption.

The revenue loss would meanwhile have to be made up, if the original assumption above is kept, as to maintaining the revenue total, by increasing some of the commodity taxes or levying new ones, or by increasing the lower bracket rates (and lowering exemptions in such a way as not to increase the burden on wealthy taxpayers) of the income tax and death and gift taxes.

A still more drastic revision to encourage investment would be a substantial cut in the property tax, particularly as it rests on long-lived assets. Here, replacement revenue would involve a large increase in state aid to localities, or state assumption of functions, and similar adjustments in Federal-state relations.

The stimulation of investment by chaises in the tax system requires much bolder moves and more patience than the stimulation of consumption, but the rewards are probably much greater.


By definition, the increase in investment induced by the changes in the tax system would, of itself, increase national income (a discussion of the effects of the replacement tax measures is postponed to the next section). The important questions are (1) whether it would initiate, or at least be an essential condition to, a cumulative upward movement and (2) whether it would over the long term raise the average level of income above and below which the economy oscillates. At this point any statement is of extremely questionable validity. Pending further study it is enough to say that (1) although a substantial upward movement undoubtedly could start even though no change were made in the tax system, and would probably not be initiated merely by such changes, the chances of its being initiated could be markedly enhanced by such changes; and (2) while there seems prima facie a probability that the long term average level of income could be increased, the many other forces controlling this level are so powerful and so obscure in their working that we face a formidable problem of research in coordinating the study of these other forces with the study of changes in taxation designed to encourage investment.

The increase in income associated with the increase in investment would almost surely, at least over a period of a year or more, be appreciably greater than the amount of increase in investment. Without going extensively into the "multiplier" concept, we may note that some of the income received by the factors of production would shortly be respect on consumption goods. Thus it seems likely that the chances of the national income's being increased by more than the DIRECT effect of the tax change is greater where the direct effect is to increase investment than where it is to increase consumption.


Up to this point, the effect of tax changes designed to increase consumption has been analyzed without considering possible repercussions that the substitute measures might have on investment; and the corresponding analysis has also been omitted in the section on encouragement of investment. Yet it is obvious that in each case the replacement revenues are apt to affect national income adversely through discouragement of investment or consumption, as the case may be. In other words, there is scarcely any change in the tax system, of a kind that maintains the total revenue, that will not, in some one or more of its features considered separately, tend to check the growth of national income.

It certainly does not follow that there is no change the net effect of which would be to increase the national income. But there does seem to be a substantial conflict between short-term and long- term objectives. The net tendency of the measures that would increase consumption would be to increase national income over the short term but possibly -- though not necessarily -- to fail to increase it, and perhaps even to decrease it, over the long run as the dampening effect of the replacement measures on investment gathered strength. The net tendency of the measures designed to increase investment is perhaps more obscure, but an initial effect, lasting perhaps over a year or two, might well be to decrease the national income as the replacement measures checked consumption. There is even some danger that the discouragement to investment resulting from such a drop in consumption would seriously impair the ability of the tax measures to stimulate investment even over the long term, though in this connection the deepening of capital equipment must not be forgotten in our concern over the widening of it.

In this memorandum, which is only a tentative exploration of the subject, it is not feasible to carry this analysis of the "net effect" further. But in conclusion it may be noted that part of the task of subsequent research on this problem is to study possible taxes of a kind that have not been used in the United States, or indeed elsewhere. For example, there is some reason to think that taxes on net personal fortunes -- net personal wealth -- are less restrictive of investment than taxes on net personal incomes. Taxation of "idle money," tax-stamped money, and other similar ideas need to be carefully studied, for even though they prove to be as fantastic as some of them sound, it is conceivable that useful ideas can be developed from them. The present memorandum, however, has been limited to a discussion of taxes that are familiar to us from actual use.