Journal
of Public
Economics
13 (1980) 3955401.
0 North-Holland
TAX CONSTRAINTS
Publishing
ON LEVIATHAN
Some second thoughts on the constitutional E.G. WEST Carleton
Received
October
and
University,
Company
choice calculus
G. CORKE*
Ottawa,
Ontario,
1978, revised version
Canada
received July 1979
Geoffrey Brennan and James M. Buchanan (1977, pp. 255-273) hereafter B&B, have attempted to derive normative tax rules based on the constitutional calculus of a typical voter-taxpayer attempting to insure against future domination by a budget-maximizing, and Leviathan-like, bureaucracy. They derived the rule that the tax base should nor be fully comprehensive, which is in striking contrast to the conventional optimal tax literature. A second major conclusion was that progression in the tax structure was desirable, but for reasons quite unconnected with the usual equity arguments. Our paper attempts to show that a fuller treatment of their seminal model can yield richer insights even though it leads to results that are substantially different. The first requirement is to eliminate what appears to us as ambiguity or incompletion in the B&B analysis. While containing a ‘veil of ignorance’ assumption wherein ‘the individual is presumed to be unable to predict with precision what his own position will be at any particular moment in [the] for the post-constitutional sequence’, the B&B model does not account possibility that the individual’s future position might be inside government; that is, he might eventually enjoy membership of the ‘Leviathan group’. We say Leviathan group because the authors explain that they have in mind a Leviathan that comprises large modern public bureaucracies. Today the chances of being a member of these organisations are quite high. Insofar as the Leviathan expands, as B&B fear, the number of individuals comprising it also expands, and the chances of being a member of it increase. This is the more so if we use ‘bureaucracy’ in its broadest sense to include such persons as civil servants, police, and teachers. *We wish to acknowledge helpful comments colleagues Stan Winer and Michael McKee.
from
referees
and valuable
discussion
with our
396
E.G. West and G. Corke,
Tax constraints
on Leviathan
B&B’s readers will naturally assume (as the authors appear to do) that ‘Leviathan men’ will enjoy substantially higher incomes than ‘nongovernment men’. But if there are two such classes, the ruled and the (rich) rulers, the voter behind the veil of ignorance is confronted with a dual outcome lotterytype choice. For he has a chance of being a member of either group. His fmal selection will depend on the expected relative sizes of the two classes, their differences in income, and individual attitudes to risk. Where the individual is risk-loving, and even where he is risk neutral, he will rtot desire the precise pattern of constitutional restraints that B&B predict. The fact that the individual may himself be part of the Leviathan group will affect his calculus in choosing the tax rate r and the tax base b. Consequently, he will not decide upon B&B’s efficient choice of government revenue R*(b, r). In a sense the individual is at odds with himself. Should he become a member of the Leviathan group he would, on B&B’s argument, favour what they find to be the most exploitative system: a regressive tax structure on a maximum base. If, on the other hand, he were to become simply a member of the taxpaying group only, then on the B&B reasoning he would favour the least exploitative system: an ideally progressive tax structure on a confined base. The above considerations reveal the need to specify more clearly the income or utility function with which the individual behind the veil of ignorance is primarily concerned. At the most elementary level he will wish to maximise the function E(Y)=P(L)Y,
+P(L)YL,
(I*)
where P(L) is the probability that he will be a member of the Leviathan group, P(L) is the probability that he will not be a member, YL is the monetary income of the individual in the Leviathan group, YL represents the monetary income of the non-Leviathan individual, and E(Y) is the expected monetary income of the person behind the veil of ignorance.’ It is necessary next to incorporate into the analysis consumption of government services. In fig. 1 it is assumed that these services provide ‘We assume here that the marginal utility of income is the same for the individual regardless of where that income is earned. We may also abstract from the problem of varying attitudes to risk across individuals, so that P is exclusively a function of the size of the bureaucracy. If we did not make this assumption (l*) would become, assuming a risky state of the world, state m and a nonrisky state, n, E(Y)=p,(L)Y,,+P,(L)Y~,,+p,(L)yt,+p.(L)y,,.
(2*)
This addition could be included since it could well be argued that one of the nonpecuniary benefits that members of the Leviathan group will seek is risk avoidance to the extent that the probability that they will enter the risky state is decreased. (P is a function of (a) the person’s risk behaviour, and (b) the chances of being in or out of government.)
E.G. West and G. Corke,
Tax constraints
on Leviathan
397
positive utility up to the level of government shown by G”. At this point the marginal rates of substitution between monetary income, Y, and government services, G, become zero. To the right of this point further increments of government services are of negative value. The optimum position for the ordinary citizen is at E with G’ government services and OC monetary income. This point maximizes utility derived from a combination of conventional monetary income Y and government services G.
Fig.
1
The individual’s initial monetary income OB is assumed to derive from his given endowments. But this position is one where government services are zero. It is normally assumed, of course, that conventional income is more difficult to earn in the absence of a government-policed system of private contracts. Point E captures this improvement plus the enjoyment of other goods and services that cannot be provided noncollectively.* Since we assume both the Leviathan individual (~5) and the nonLeviathan person (L) to enjoy the ‘consumption’ benefits of government services equally, and assuming a position between 0 and G” in fig. 1, then
To the right of G” in fig. 1 we have aU@G
*B&B’s fig. 1, which has leisure (L) on the ordinate and monetary income (Y) on the abscissa, is not easy to follow because of their exclusion of government services. Their analysis is thus entirely in terms of monetary income benefits instead of the more comprehensive term ‘utility’.
398
E.G. West and G. Corke,
Tax constraints on Leviathan
For the Leviathan individual, however, we need next to examine some special circumstances that separate him from others. Observe first that the effects of the level of government services upon the potential monetary income of the non-Leviathan individual works through the size of the tax base (b) and tax rate (r). Thus dYc/d(b,r)
(2*) where E(U) = expected utility and G, = commonly shared direct consumption benefits from public goods publicly produced. From the previous reasoning we note that the function (b, r) may be positive for YL. We may now return to the decision-making of the individual behind the veil of ignorance. His primary task can now be seen as that of maximizing the expected utility shown in eq. (2*). At least this is the calculus that is rooted in the B&B model. And under its assumptions the individual might clearly choose a G that ‘favors’ Y, over YI;. We shall argue that the more reflective individual will neuer entertain such a choice. We proceed by testing the B&B model with comparative static, and then dynamic considerations. B&B are obviously assuming that YL> YE. This calls for some explicit assumptions concerning entry into the bureaucracy. In the case of a democracy it seems reasonable to suppose that initially everybody can apply for employment in a public office. But with such free entry, rents from bureau jobs will accrue to more and more members of society and Leviathan will grow. Eventually the advantages of succeeding generation Leviathan members will diminish until_ individual earnings are equalised at the margin between individuals L and L. It would be stringent to assume that the Leviathan group is totally unproductive; but according to the Niskanen model one may assume that its efficiency declines with bureau expansion. In his original model, Niskanen (1971) spoke of the enlargement of bureaucracies causing ‘allocative inefficiencies’ and then technical inefficiencies.3 The former occur where, 3Niskanen
describes
the initial position
as the ‘cost constrained’
area.
E.G. West and G. Corke.
Tax constraints
on Leviathan
399
although bureaux operate at minimum cost, they produce above-optimal amounts. Technical inefficiencies begin when, in addition, bureaux begin to operate at above-minimum cost (‘fat’ appears in the bureau).4 In subsequent work Niskanen (1975) has developed a more generalized model showing that both allocative and technical inefficiencies should be observed in practice.5 It is sufficient for our argument that the expansion of the bureau is accompanied by growing technical inefficiency. This point is serious because workers in the public sector approach negative marginal productivities. As more workers transfer from the productive private sector, total national income falls. This will reduce the source of tax revenues so that ‘Leviathan’ individuals lose along with others. Eventually, the individual will be unable to achieve the ‘base’ indifference curve i, in fig. 1 unless he reverts to anarchy or resorts to taxpayer, or other, revolutions. Clearly, a well-organised Leviathan group will have a vested interest in seeing this does not happen. A truly optimizing Leviathan group, such as a ‘nondemocratic’ or ‘natural’ aristocracy, would have an incentive to maintain individuals. B&B’s ‘democratic’ an optimal number of non-Leviathan Leviathan group, in contrast, is likely to be too uncoordinated, unwieldy and inefficient to act in this manner. If it is so deficient there will be a series of crises, if not revolutions; and there can be no long-term stable equilibrium. If, on the other hand, the ‘democratic’ Leviathan group succeeds in limiting its own size, it is likely to do so by making job qualifications at entry more arduous. Expressed in other terms, the costs of rent-seeking will be increased by the incumbent (monopoly) bureaucrats. Candidates will then be involved in the wasteful activity, for instance, of obtaining extra diplomas to meet the stiffer screening. Reduction in the national income will again follow. There will be no incentive from within the bureau to alter this position since the self-interest of existing bureaucrats will prompt them to advise maintenance of the entry restrictions. Only a new constitutional convention agreed upon by all citizens could effect a long-term reform. But since the present analysis is at the ‘veil of ignorance’ stage the problem returns for solution by the would-be constitution maker. Obviously he is obliged to anticipate the above problem in his constitutional planning. At the center of B&B’s argument is an attempt to show whether such constitutional design could be accomplished by fiscal, as distinct from institutional, means. Their conclusion in favour of the constitutional choice of a progressive tax rate could be theoretically correct, but for an entirely 4Bureau expansion brings a development from the ‘cost constrained’ to the constrained’ area in Niskanen’s terms. 5The relative amounts of each inefficiency depend upon the bureaucrat’s preferences reward structure. The bureaucrat can now enjoy a ‘discretionary budget’.
‘demand and the
400
E.G. West and G. Cork
Tux constraints
on Leoiathan
different reason from theirs. Insofar as Leviathan members enjoy higher incomes than others, the difference can be largely removed by appropriately selected rates of progression. For in a democracy, Leviathan members are as subject to taxation as anyone else. And one can speak of a tax base being comprehensive in t&s sense. B&B use the term ‘comprehensive tax base’ with respect to what is taxed, monetary or potential income (including leisure). They favour a reduction of the base because this will curb the exploitative powers of the Leviathan bureaucracy. On our wider use of the term ‘comprehensive’, however, it would appear that the ‘gouging’ of monetary or potential income (including leisure) would recoil on the Leviathan members themselves and this acts as an automatic check. Some may argue that Leviathan members can avoid progressive income taxes better than others. This is because they find more opportunities than usual for nonpecuniary incomes. This seems to be implied in Niskanen’s concept of the bureaucrat’s enjoyment of a ‘discretionary budget’. Members of large private bureaucracies may certainly experience similar advantages, but they are likely to be less in degree. They are correlated, after all, with the extent of monopoly of the respective enterprise; and government bureaucracies constitute monopolies par excellence. But insofar as there is validity in this point, it would call, at least theoretically, for a reliance on a tax base that is nrorr comprehensive in B&B’s sense. For the base would need to be suitably widened so as to subject to tax all the bureaucrats’ nonpecuniary benefits (including leisure on the job). Whether the tax is progressive or not, and however narrow or wide the base, B&B’s tax structure is intended to apply universally. But a tax structure designed so as to capture the rents of one group in society will be inappropriate for others who earn high but rentless incomes. In any case, a constitution that allows bureaucrats to adopt special property rights over others in the form of exclusive rent-earning occupations, and then provides for avenues of recoupment through specially designed penalties (taxes), is a second-best world. Since we are behind the veil of ignorance, the opportunity for first-best design of property rights is clearly available. And since the special rent-earning in question is related positively to the degree of monopoly, the focus should be upon institutional (rather than fiscal) design that attempts to preclude it. Structuring society so as to provide competition in government is one such available strategy. And competition is likely to be a more effective ‘tax’ on the potential exploiter. Nonfiscal means of such constitutional design could include : (a) an ‘antitrust’ monitoring and control of government promote competition among bureaucracies; (b) increased uses of referenda;
departments
to
E.G. West and G. Corke,
(c) (d) (e) (f)
Tax constraints
on Leviathan
401
increases in the minimum vote requirement abow 50 percent; the random assignment of politicians to review committees; shorter periods of government office; and the use of subsidies to promote freer entry into politics and elections.
It is acknowledged by B&B that many different types of constitutionally derived constraints are conceivable. They also observe that ‘these need not be mutually exclusive’. What has been argued here is that well-chosen nontax constitutional constraints to anticipate, and possibly avoid, the problem of monopoly government may well be mutually exclusive of the particular tax reforms in B&B’s helpfully thought-provoking study. The kind of constitution that is likely to emerge clearly depends on degree of sophistication possessed by the man behind the veil. B&B assume that he knows about Niskanen-type bureaucracies. But he will also know that in a democracy all citizens are potential entrants to them. Next, while one citizen in the postconstitutional stage may choose to enter, he cannot control the similar actions of others. Because after some point GNP will fall with continuing expansion of the bureaucracy (as indicated by the propensity to technical inefficiency central to the Niskanen model), the sophisticated man behind the veil has to anticipate a future prisoner’s dilemma situation. To resolve this he must design a constitution that will discourage both an above-optimal size bureaucracy and an above-optimal bureaucratic net income. Progressive taxes that apply ‘comprehensively’ - to bureaucrats as well as to others could provide some such control. But a more ‘fundamental’ solution would be a constitution that contained institutional constraints that prevented the bureaucrat from earning more than his marginal product in the first place. A minimum bureuucrac~~ is recognized as needed, but a Lepiuthan is not. B&B’s model appears to be a constitutional attempt by nonLeviathan citizens to curb the powers of bureaucratic members of the Leviathan. As such it seems to be in front of, rather than behind, the veil of ignorance. Their solution of progressive income taxes focuses on minimizing the exploitation of the weak by a strong body of individuals that seems exogenous to the system. Yet behind the veil everything is endogenous. In the best designed system Leviathan itself can be prevented by well chosen ‘contraception’. And this is more likely to rely on institutional than fiscal means. References Brennan, G. and J.M. Buchanan, 1977, Towards a tax constitution for Leviathan, Journal of Public Economics 8, 255-273. Niskanen, William, 1971, Bureaucracy and representative government (Aldine-Atherton, Chicago). Niskanen, William, 1975, Bureaucrats and politicians, Journal of Law and Economics 18, 617643.