6
Progressive Tax Reform and Majority Voting P. M. PESTIEAU*
introduction. The question of why incomes should be taxed at progressive rates rather than at proportional rates is a fundamental one in public economics, tt is usually dealt with in the normative terms of fairness and vertical equity. It remains, however, that tax rates are established not on purely normative grounds but within the institutions of public choice. The purpose of this paper is to examine progressive taxation within such a context; more precisely, it will focus on the distributional implications of using majority voting to select the rates of the income tax schedule. The basic issue at hand centers around the following problem: the government is required to raise a certain amount of revenue by taxing the public. How will the tax burden be distributed, given that the choice is left to majority voting? It is generally admitted that progressive taxation will be preferred over proportional taxation by a majority of voters, since a majority of the voting population has below-average incomes. Progressive taxation is thus viewed as a device for the less wealthy majority coalition to exploit the more wealthy minority *Department of Economics, Cornell University. t wish to thank my colleagues, E. Blackstone, D. Mueller, U. Possen and S. Slutsky, all of whom gave me valuable suggestions for improvements.
70
PUBLIC CHOICE
coalition. 1 Departing from this " t y r a n n y of the majority" approach, some peopIe have suggested that progressive taxation is a form of collectively supplied charity that may be supported by the rich as well. When utility functions are interdependent in that every individual in society is o f a philanthropic nature and believes equality is a good thing, progressive taxation may be viewed as a compact among the more wealthy citizens to tax themselves to subsidize the less wealthy, thereby supplying themselves with a collective good and increasing both their own utility and the utility of the less wealthy citizens. Within this frame of analysis, initially developed by H. Hochman and J. Rodgers (1969), progressive taxation could very well get unanimous support. 2 While the above explanation of progressive taxation might seem to entertaima considerable amount of plausibility, it fails to explain why, in various instances, voters have turned down proposals aimed at replacing a proportional income tax, or even a sales tax, by an equal yield progressive tax. A t this point, one might rightfully wonder why the voters' choice would be restricted to merely two alternative tax schedules. Clearly, progressive taxation is not the only way o f redistributing income and in reality is certainly not the best one. 3 In its defense, two points can be made. First, progressive taxation can be viewed as a formal example of more general policies o f redistribution. Secondly, voters may at times be actually asked to choose between tax schedules o f different progressivity. For example, the state o f Massachusetts raises almost 40 percent of its revenue from a flat-rate personal income tax. 4 An alternative graduated (progressive) income tax was submitted for voter approval. Such a reform would have benefitted about 95% of Massachusetts families: it would have been possible for those who earn no more than $23,000 a year to pay tess taxes. The same amount of revenue would still have been raised. However, each time the proposal was taken to the voters - 1 9 6 2 , 1968 and 1 9 7 2 - i t was defeated by a wide, though decreasing, margin. Another example is the unsuccessful attempt by New Jersey Governor W. T. CahiU to make the state legislature approve a progressive tax reform program. Paradoxically, some of the more stringent opposition to his reform came from blue collar neighbourhoods which stand to benefit form it the most. 5 How can one explain that a large majority of voters turned down a reform which would have benefitted most o f them? The purpose o f this paper is to analyse various factors which explain this apparent paradox. One possible explanation is that the majority coalition is aware o f disincentive effects that arise from progressive taxes and fears that a reduction in the total national income will force some to pay more rather than less taxes. (Section I). A second explanation is that, 1See Buchanan and TulIock (1962) and Downs (1957)..Since the tax..schedules to be. voted on are assumed to be given and perfectly understood, there :s no amb~gu:ty about who as going to receive or pay for the transfers. There is thus no room for the median or any other voters to negotiate in order to improve their own position. 2See also Von Furstenberg and Mueller (1971). In a recent paper,.Tullock 11971b) argues that some people tenct to be charitable when charity costs next to nothing, rarticmarly, they would urge and vote for government redistribution while actually making no sacrifice, Since very few governmental transfers are effectively redistributive. In the Case here presented, however, charity will be costly. The passing of a progressive tax proposal will in fact imply higher taxes for the formerly charitable w14o may therefore be tempted to vote against the proposal. 3See J. Pechman and B. Okner (1974), who show that taxation in fact Nays no redistributive role. 4For more details on the Massachusetts case, see Pechman, et at (1974). 5See Kristol (1972). As another example of such a paradoxical behavior, one might cite the resuIts of a federal sl3onsored survey on Public Ot)inlon and Taxes (1972): both at the federal and the state level, a large majority of respondents favored sales taxation, known to t~e regressive, over income taxation.
TAX REFORM
71
because of avarice and expectation of social mobility, some members of the middle income group are reluctant to help the poor they will never be part of and hurt the rich they hope to become. (Section II). A third explanation as to why progressive taxation can be turned down by a majority of voters is that the less wealthy voters do not effectively outnumber the more wealthy. (Section III). These three explanations will be discussed in turn.
In this section, we show why the introduction of disincentive effects from progressive taxation could affect the proposition that a majority always prefers a progressive over a proportional income tax. For the sake o f simplicity, the tax schedules under consideration are those that are linear in before-tax income and include the possibility of lump-sum grants. 6 It will be convenient to plot before-tax income and tax on a graph as in Figure 1. The 45 o line represents the before-tax income; lines a, b, and c represent three types of linear tax schedules, a is a proportional tax; both b and c are progressive, since they imply increasing average tax, including a level of exemption below which people are paid a subsidy. Clearly, c is the more progressive and indeed, if implemented, would leave everyone with the same after-tax income (as measured by the vertical distance between the 45 ° line and the tax line). The three taxes are supposed to finance a given level of public expenditure; one can easily verify that they all intersect at mean income (y~ and per capita expenditure (g). Since in countries such as the U.S. the income distr~ution is skewed rightforward, the median income (m) lies below the mean income. There is, theref6re, more than a majority who prefer the progressive tax (b or c) over the flat rate tax a. A serious objection to this reasoning is that it abstracts from labor supply considerations. In general, the imposition of an income tax has an impact on individuals' decisions about work and leisure, and it is generally believed that an increase in the marginal tax rate leads to a decrease in the labor supply and thus to a reduction in the total amount of income available for taxation. 7 Consequently, some people in the majority coalition realize that they may end up paying more rather than less taxes and thus vote for a low tax rate, as shown recently by T. Romer (1973). His argument can be illustrated with the aid of Figure 2. Consider two alternative income taxes: a, a proportional tax; b, a progressive tax, both yielding the same revenue. Under b, the mean income (Y-b) is expected to be lower than under a @-a)" In either case, the median income (m a or rob) follows, tf the disincentive effect is very strong, i.e., if the difference between ~.a and ~;~, is ±ouite . . . . substantial, It might happen that the me&an income recipient is better offDwith the flat rate tax than with the progressive tax, as occurs in Figure 2. Then, a majority of voters would indeed prefer the proportional over the progressive tax. This conclusion rests on the assumption o f a very sensitive compensated labor supply, obtained in theoretical models by use of log-linear utility functions. However, the existence of a sensitive labor supply is not verified by available empirical evidence. M. Kosters (1969), for example, has recently shown that the substitution effect of 6This argument draws heavily on Foley (1967). 7One may note that below the break-even point (intersection of the horizontal axis and the tax line) there is an unambiguous drop in work effort, since then both income and substitution effects work in the same direction.
72
PUBLIC CHOICE
FI~EI. t~x~
income.
T
per c==pita expenditure
0
l,z¢~l
/
/
/
/
,, I .... !. m y
.
.
.
.
.
.
before
tax' income. >
73
FIGURE 2. tax, income.
b
5/-
m'b ~b ma
a
before ta~ income.
74
PUBLIC CHOICE
the labor supply is not very important. The labor disincentive effect of income taxation appears, therefore, as insufficient to explain why a proportional tax would be preferred over a progressive tax by a majority of voters. The labor disincentive effect is, however, only one, and not the most important, social cost that the introduction of a more progressive tax involves. First, high tax rates may affect capital formation by altering both the amount of saving that is done and the way it is invested. 8 Both of these can be expected to work very significant reductions in the total real income available for redistribution. Secondly, because of high tax rates some people may be tempted to migrate to another state, or even to another country, as would be the case in Europe. Finally, as pointed out by G. TuUock (1971a), the prospect of income redistribution through progressive taxation may lead individuals or groups to invest resources in attempting either to obtain a transfer or to resist a transfer away from themselves. These resources represent net social waste, which TuUock suggests can be quite large. It is possible for instance, that the beneficiaries o f a progressive tax spent more time and money in the lobbying for reform than they may ultimately receive in net benefits. II Implicit in the previous analysis is the assumption that each individual's utility is a function of his own income and leisure, but not of other people's income. As mentioned above, an important body of literature has arisen arguing that preferences which display positive interdependence might imply that income redistribution benefits all. More explicitly, if everyone dislikes other people having more or less income than he, then predictably any move towards progressive taxation will be supported not only by a majority but by a unanimity of voters. One can nevertheless doubt the validity of such an altruistic assumption, at least as far as the attitudes of voters towards tax reforms are concerned. Especially in the middle income group, many people are indifferent, and are even opposed, to letting government help the low income groups any further. The rationale for such an "avaricious" attitude, as R. Scott (1972) has dubbed it, is that many middle income people find that too much money is already spent on the poor whom they hold responsible for what they are. They often view the government as an ally of the poor, and thus tend to block any measure aimed at alleviating the burden of the low income group. 9 In a recent paper, Irving Kristol (1972) argues that the average American is not rebelling against tax inequities, but against taxes in general and the ways in which the welfare state spends them. "It would not be an exaggeration to say that much of the present discontent with taxation is provoked by the fact that the welfare state, which these taxes support, is too committed to equality-to expenditures that benefit primarily the minority who are poor" (Kristol, 1972, p. l l ) . Another explanation is that many in the lower income group do not expect to remain there permanently. Indeed, in the real life, there is a large mobility of people within the distribution of income. Some who are in the less wealthy segment 8See M. Feldstein for a general analysis of the way in which a tax on labor income affects the accumulation of capital. 9In a recent article, Litwak e t al,(1973) discuss the idea of a collusion between the government and the poor at the expense of the middle and upper income citizens,
TAX REFORM
75
of the population can expect to be in the more wealthy segment at some time in the future. Of course, the converse is true as well, but in general people tend to look only at the bright side of social mobility. 10 By supporting a tax that is execessively progressive, some members o f the less wealthy group realize that they may harm their chances to enjoy greater income in the future, and thus they vote it down even though they would get a larger disposable income if they were to remain in the same income class. 11 For that matter, one can use the Rawls-Lerner concept of original position and ignorance about the income that one will receive the future. Assuming that from the perspective of each individual it is more likely that he will receive a higher income in the future, then it is no longer true, as A. Lerner (1944) and his commentators (see W. Breit and W. Culbertson (1970)) thought, that people would unanimously vote for equality. Indeed many would find their expected utility higher with the proportional tax that leaves income distribution unchanged than with the progressive tax, and they would thus vote for the former. This point can be illustrated by looking at an economy divided into a certain number of successive income classes Yr' r = 1,2 . . . . Each individual, whatever income class he belongs to, has some expectation about his future earnings. These can be expressed by a set of subjective probabilities, rrl,rr2 and 7r3, of moving one Class up, one class down or not moving at all, respectively (lr I + rr2 + rr3 = 1). He will prefer, and thus vote for, a progressive over a proportional income tax if the former brings him an increase in expected utility; i.e., (1)
A y r + 1 u'(Yr+l) rr 1 + AYrU'(Y2) rr3 + AYr_ 1 u'(Yr_l) rr2 > 0
where ur(Yr) denotes the marginal utility of an income of level r and AYr, the change in disposable income due to the substitution of a proportional for an equal yield progressive income tax. For those people, whose income in close to the median and means incomes, a y -~ ( 0 , ay,~ = 0, A y ~ ~ 0 ; moreover, A y ~ -A Yr+l because of the shape of trhe±income di~tributionr~nally, u " ( Y r _ l ) ~ u ' ( ~ ; 1 ) because of assumed decreasing marginal utilities. In these conditions, one can observe that inequality (1) is not likely to hold if nl, the subjective probability of moving up is much larger than 7r2, a condition met for most people.I 2 This holds even more for incomes above the median and mean incomes, whereas for the low income group, inequality (1) is likely to hold. The two factors introduced in this section concern essentially the middle income receivers. For that matter, one may note that in any case they would not gain considerably from the tax reform as can be seen from Figure I, since their incomes are in the vicinity of the median. 13 One can thus understand why a majority might not be inclined to rid themselves of the proportional tax system. l OSee Stone (1970) for a study of high expectations of upward mobility. llAfter having completed the first version of this paper, I found a similar argument in Wagner (1973), and Rodgers (1974). One might note that our model differs from Buchanan and TulIock's insurance moctel in whlcla income redistribution is beneficial for risk averse citizens. 12One may note that these subjective probabilities are not very realistic. In most societies, income distribution persistently exhibits the same characteristics for different periods. This implies that on the average the expected income change cannot be positive. Other~nse, one would v?itness an ever-decreasing dispersion of incomes. 13This would be even more evident with a non-linear progressive tax schedule.
76
PUBLIC CHOICE III
A third type of explanation as to why a majority may turn down progressive taxation is that the less wealthy people do not effectively outnumber the more wealthy, since they tend not to be equally enfranchised. 14 In real societies, the upper income group carries a much greater political weight than its mere number would imply. First, the wealthy tend to vote more often and in a greater proportion than the p o o r ) 5 Second, they can spend more time and more money to campaign forissues dear to them. When the issue is progressive taxation, or any measure detrimental to their interests, the rich can organize well-orchestrated campaign~ to polarize the middle-income group on various fringe issues and distract them from the real issue at stake. Among those fringe issues, one can list the following: aversion to change, dissatisfaction with all public agencies and especially with income maintenance programs, and fear of eventual tax increase. 1. Aversion to change. Current distribution of income is determined by individuals' natural abilities and inherited wealth. For some people, both determinants are the outcome of a natural social order that they do not want to disturb. Such people dislike any measure attempting to affect the natural equilibrium of society, fearing that behind some apparent advantages there is a serious risk of chaos. 2. Dissatisfaction with the public sector. In the last decade, an increasing number of taxpayers have expressed deep dissatisfaction vis-a-vis the public sector; they believe that the money they pay in taxes is lost, buying them very little. By turning down a progressive tax reform or an environmental bond issue, for example, they merely cast a vote of protest. 3. Fear of tax increase. Even ff the tax reform does not involve any increase in tax revenue, but just a reshuffling of the tax burden, many voters can be led to believe that eventually they will have to pay more taxes. By subjecting voters to an intensive campaign of advertisement which focuses their attention on such irrelevant issues, and thus confusing the real issue at hand which is iess taxes for a majority, the wealthier citizens can secure the support of less wealthy citizens and obtain a majority to oppose the progressive tax reform. 16
Conclusion. In this paper, we have tried to explain why voters can turn down a measure which is supposed to increase the disposable income of a large majority of families. The reasons we advanced were the following: 14See Downs (1957). 15See Frey (1971) and various comments on his paper in the subsequent series of Public Choice.
16An example of this is the misleading advertisement used in the Massachusetts campaign against the tax reform: "Don't be fooled--The Graduated Income Tax is just another tax increase-it means more for them, and less for y o u . , . Again!".
TAX REFORM -
77
the increase in disposable income is small for the large middle-income group; some people are reluctant to back a measure aimed at helping low4ncome groups further; - some do not want to burden too much the high income group of which they hope, eventually to be a part; - finally, many voters are subject to an intensive campaign of advertisement which focuses their attention on irrelevant issues. -
One implication of this line of explanation is that public financing of political campaigns, more realistic expectations of upward mobility and more altruism among individuals would probably increase the chances of progressive tax reform. Obviously these three points would merit further research.
REFERENCES Breit, W. and W. P. Culbertson, Jr. Distributional Equality and Aggregate Utility: Comment. American Economic Review, 60 (June, 1970), 435-41. Buchanan, J. and G. TuUock The Calculus of Consent. Ann Arbor; University of Michigan Press, 1962. Downs, A. An Economic Theory of Democracy. New York: Harper & Row. 1957. Feldstein, M. S., Tax Incidence in a Growing Economy with Variable Factory Supply. Quarterly Journal of Economics (forthcoming) 1972. Foley, D. Resource Allocation and the Public Sector. Yale Economic Essays. 7 (1967), 45-38. Frey, Bruno. Why Do High Income People Participate More in Politics? Public Choice, 11, (Fall, 1971), 101-105. Hochman, H. M. and J. D. Rodgers. Pareto Optimal Redistribution. American Economic Review,! 59, (September, 1969), 542-557. Kosters, M. Effects of an Income Tax on Labor Supply in A. Harberger and M. Bailey, eds. The Taxation of Income from Capital. Washington: Brookings Institute. 1969. Kristol, I. o f Populism and Taxes. The Public Interest, 28, (Summer, 1972), 3-12. Lerner, A. The Economics of Control. New York: Macmillan, 1944. Litwak, E., N. Hooyman and D. Warren. Ideological Complexity and Middle American Rationality. The Public Opinion Quarterly, 37, (Fall, 1973), 317-332. Pechman, J. and B. Okner. Who Bears the Tax Burden? Washington: Brookings Institute, 1974. Public Opinion and Taxes. Advisory Commission on Intergovernmental Relations. Washington, D. C. 1972. Rodgers, J. D. Explaining Income Redistribution in Redistribution Through Public Choice. H. M. Hochman and G. E. Paterson, editors. New York: Columbia University Press. 1974. Romer, T. Individual Welfare, Majority Voting and the Properties of a Linear Income Tax. (Mimeo: University of Western Ontario). 1973.
78
PUBLIC CHOICE
Scalar, E., T. Behr, R. Torto, and M. Edid. Tax Taxpayers and Social Changes: The Political Economy of the State Sector. Review of Radical Political Economics, (Spring, 1974), 134-153. Scott, R. H. Avarice, Altruism, and Second Party Preferences, Quarterly Journal of Economics, 86, (February, 1972), 1-18. Stone, J. P. Better Personal Future: A Two Component Model, The Public Opinion Quarterly, 34, (Fall, 1970), 346-359. Tullock, G. The Cost of Transfers. Kyklos, 24, (1971) Fasc. 4, 629-643. Tullock, G. (1971b) The Charity of the Uncharitable. Western Economic Journal, 9, (December, 1971), 373-391. Von Furstenberg, G. and MueUer The Pareto Optimal Approach to Redistribution: A Fiscal Application. American Economic Review, 61, (September, 1971), 628-637. Wagner, R. (1973) The Public Economy. Chicago, Illinois: Markham.