INCOME TAX RATES AND INCENTIVES TO WORK AND TO INVEST George F. Break, University of California, Berkeley That the level of income tax rates can have an important influence upon an individual's incentives to work and to invest is a proposition which it would be rash indeed to dispute, but that this influence is a simple or straightforward one to determine is a great, though widely accepted, fallacy. Upon examination, the effects of taxes on incentives are found not only to be highly complex but also to operate, on occasion, in unanticipated directions. To a country desirous of maintaining a high rate of economic growth the importance of determining these effects as accurately as possible is obvious. For this goal to be achieved measures proposed for tax reform must be closely analyzed in terms of their incentive-creating and incentive-destroying potentialities. By considering briefly the various effects of individual income tax rates on incentives, this paper attempts to provide the basic tools needed for such a task. The principal conclusions may be summarized as follows: (1) Changes in average and marginal tax rates affect work incentives in opposite directions, a decrease in the marginal rate, for example, improving incentives but a decrease in the average rate reducing them. Since changes in the tax law typically move both rates in the same direction, at least for some taxpayers, they are theoretically capable of either increasing or decreasing the total amount of work done by the labor force. In order to predict the net effect with a reasonable degree of assurance, one needs empirical evidence covering the actual behavior of different groups of workers. (2) While the available evidence of this sort is not so complete as one would wish, it does indicate, with impressive consistency, that even rather high income tax rates have very little net influence on taxpayer work incentives. At the same time there are suggestions that either very high marginal tax rates (70 percent and above, for example) or sudden increases in marginal rates at lower levels can have significant disincentive effects. (3) In the case of individual incentives to undertake risky investments it is not only the level of tax rates but also the extent of the allowances for loss offsets that is important. Under a proportional income tax with full loss offsets an increase in the tax rate can be expected, surprisingly enough, to increase the total amount of risk taken in the economy. (4) The incentive-creating or destroying powers of a given income tax reform, therefore, are a function of the changes which that measure brings about in marginal and average tax rates and in the allowances made for business and investment losses. The extent to which these considerations can be used to implement a general policy of broaden ing the income tax base and simultaneously reducing marginal tax rates is considered in the last section of the paper with reference to two specific tax reforms. 1. INCOME TAX RATES AND WORK INCENTIVES: THEORETICAL CONSIDERATIONS The effort-stimulating effects of a general reduction in individual income tax rates are likely to be clearly perceived by most taxpayers. Under the lower rates an additional hour's work yields more takehome income and consequently becomes a more attractive possibility. Regardless of whether the worker reacts to this stimulus or not, the tax change nevertheless provides him with a higher disposable income (i.e., average tax rates have fallen), and as he becomes accustomed to his more prosperous situation he is likely to desire more leisure time in which to enjoy it. Lower income taxes, in short, subject work incentives to opposing influences, the one being immediate and obvious and the other more subtle and slower to take effect. While in the short run, therefore, a reduction of marginal tax rates may well increase the total amount of work done, the accompanying fall in average tax rates is quite capable of subsequently reversing this initial movement. Even when the effect of a given tax change on the average taxpayer is theoretically predicable, some margin of uncertainty is likely to remain because of the potential behavior of one or more untypical groups. This point may be illustrated with reference to one of the tax reforms now before this committee-namely, a tightening of the rules governing allowable deductions under the personal income tax and a simultaneous reduction in marginal tax rates so as to maintain a constant level of total revenues. At first glance it might be thought that, since for the typical taxpayer marginal rates would be lowered but average rates kept constant, such a policy would definitely improve work incentives. There are, however, three other groups of workers whose potential reactions must be considered. The first, which is composed of taxpayers with above-average deductions who would find their average tax rates increased but their marginal rates reduced, presents no difficulty since both of these rate changes tend to increase the amount of work done. In addition, however, some taxpayers, having relatively few deductions to claim, would find both their marginal and average tax rates reduced, while still others, being pushed by the increase in their taxable income into a higher tax bracket, might find both marginal and average tax rates higher than before. For both of these groups of workers the net effect on their work incentives is theoretically unpredictable. We have, then, two groups of workers who can be expected to work harder as a result of the tax change and two groups whose reactions will be mixed, some working more and some less. Although under such circumstances a net disincentive effect appears unlikely, it remains a possibility which cannot be rejected on the basis of theoretical analysis alone. 1 Included in this last group would be those workers who had sufficient deductions to reduce their taxable income to zero (or below) under the old rules but not under the new. Taxpayers moving from one taxable income class to a higher one would also face higher marginal rates unless the reform measure had lowered the rate applicable to that bracket down to, or below, the level to which they had formerly been subject. To answer this and other important questions about incentives empirical evidence must be called upon. In recent years interest in this subject has resulted in the production of a number of such studies which are briefly reviewed in the following section. 2. INCOME TAX RATES AND WORK INCENTIVES: EMPIRICAL EVIDENCE Postwar studies of the effects of income tax rates on incentives to work are notable not only for the extent to which they complement one another (covering a wide variety of worker groups, geographical areas, and research methods) but also for the consistency of their findings. Without exception income taxes were found to have little or no influence, in the aggregate, on the amount of work done. At the same time, however, several significant tax reactions of more limited scope were discovered, and the implications of these with respect to tax rate reform will be noted below. The study of widest scope, that made by Clarence D. Long on the basis of aggregate census data for the United States (1890-1950), Great Britain (1911-51), Canada (1911-51), New Zealand (18961951), and Germany (1895-1950), concentrated on the relationship between labor force participation rates (male, female, and both sexes combined) and the height of income tax rates. No systematic association showed up between these two variables, regardless of whether the analysis was made over time within a single country or between countries in those years when tax rates were highest. These findings, as Long himself notes, apply to a period during which income taxes were not an important burden on the majority of workers-in the United States in 1950, for example, 97 percent of the individual tax returns filed recorded payment of an average tax of 10 percent or less of the total income reported-and this fact alone may account for the absence of relationship observed. On the other hand, a persistent tendency for labor force participation to vary inversely with fluctuations in income was found, and if this relationship continues in the future, higher income taxes can be regarded as a factor tending to increase, rather than to decrease, the size of the labor force. Broad, aggregative studies of the kind just discussed frequently conceal important but offsetting reactions on the part of different segments of the working population, and our second study is of interest because it deals exclusively with a group whose services are of great importance to any dynamic economic system and who, because of the high marginal tax rates which they typically face, should be especially sensitive to disincentive influences. During the 1946–50 period Thomas H. Sanders conducted a sample survey, based on personal interview, of approximately 160 business executives in large and small corporations in New England, New York, Illinois, Ohio, and Texas. His chief conclusion was that in spite of their frequent pro 2 Clarence D. Long, "Impact of the Federal Income Tax on Labor Force Participation," in Federal Tax Policy for Economic Growth and Stability, papers submitted by panelists appearing before the Subcommittee on Tax Policy of the Joint Committee on the Economic Report, 84th Cong., 1st sess. (Nov. 9, 1955), pp. 153-166. For a detailed analysis of the data on which this paper was based see the same author's "The Labor Force Under Changing Income and Employment," a National Bureau of Economic research study (Princeton: Princeton University Press, 1958). 3 Thomas H. Sanders, "Effects of Taxation on Executives" (Boston: Harvard University Graduate School of Business Administration, 1951). 47060-59-pt. 346 testations to the contrary, executives had not in fact abated their efforts because of high tax rates but were "still going full blast." Among the specific tax reactions were a tendency on the part of wives to take jobs in order to augment family income, a postponement of retirement on the part of some executives, and a refusal on the part of others to accept promotions or transfers with greater work burdens but relatively little posttax increase in compensation. The significance of these effects is difficult to appraise because the evidence presented is almost entirely qualitative, and taxes were frequently only one among several important factors influencing the actual decision taken. Finally, it should be noted that the majority of executives interviewed, since they worked directly under the supervision of superior officers, were relatively restricted in the extent to which they could in fact vary the amounts of work done in response to high tax rates. A third postwar study, carried out in 1952 among industrial workers in England and Wales, avoided this last difficulty by concentrating on those who were paid on some piece basis or who had the option of working overtime. Some 1,429 randomly selected workers were interviewed by professional interviewers on the basis of a fixed questionnaire (Sanders' interviews, in contrast, were not based upon a probability sample and took the form of fluid, wide-reaching, and unstructured discussions of the problem of tax incentives and disincentives). Although the general conclusion reached was again that income tax rates had not affected work incentives to any significant degree, two interesting subsidiary effects were brought to light. In the first place, the importance attributed by the workers to tax disincentives varied markedly with the generality and impersonality of the questions asked them. While 73 percent of the men and 60 percent of the women believed that income taxes in general decreased productivity, only 30 percent of the men and 25 percent of the women reported that their own propensities to work were thus affected, and these figures must be offset against the 27 percent of the men and 22 percent of the women who felt that taxes induced them to work harder. At a later point in the interviews when the questions turned to specific instances when work opportunities had been declined, less than 5 percent of these decisions were attributed to the effects of high tax rates. A second effect, which is particularly significant for our present purposes, was the high degree of sensitivity shown by the most knowledgeable of the respondents to the rate at which their marginal rate of tax increased. When the behavior of the male workers who professed to know how much more they would have to earn before going into a higher tax bracket was analyzed, those who faced the sudden increase in the marginal tax rate from just under 22 percent to nearly 37 percent reported 67 percent more disincentive reactions than did those in brackets where the rise in marginal rates was much more gradual." It would appear, therefore, that a sharp jump in tax rates does tend to elicit significantly more tax disincentives. In assessing the significance of the lack of reaction to income taxation shown in this British study, one should keep in mind two char 4 Ibid., p. 17. Ibid., pp. 12, 50, and 63-65. Royal Commission on the Taxation of Profits and Income, "Second Report," Cmd. 9105 (London: Her Majesty's Stationery Office, 1954), app. I, pp. 91-124. 7 Ibid., pp. 117-119. |