The State

Anthony de Jasay, courtesy of the author
Jasay, Anthony de
(1925- )
Display paragraphs in this book containing:
First Pub. Date
Indianapolis, IN: Liberty Fund, Inc.
Pub. Date
26 of 33

4. Redistribution

Buying Consent


Majorities must be paid for out of minority money; this condition leaves the state little choice about the redistributive pattern to impose.


In competitive electoral politics the winner's reward is profitless power.


A given state-of-nature society unmarked by a state, can be told apart from others by its given set of initial distributions of all the unequal attributes which distinguish its members. These are, as we have seen in another context, virtually countless in number. The various distributions, ceaselessly shifting in historical time, are "initial" only in the sense that logically they precede the activities of the state. A relatively small number of them may yield to attempts at levelling. If a state is superimposed on this society, and if it relies on its subjects' consent to stay in power, it may, and under competitive conditions it will, find it advantageous to offer to change some "initial" distribution in such a way that the redistribution will gain it more support (in terms of clout, or votes, or whatever "mix" of the two it considers relevant to power).


Such a redistributive offer is obviously a function of the initial distribution. For instance, in a society where some people know a lot and others only a little, where knowledge is prized by both and (tall order!) absorbing knowledge is painless, the state might gain support by obliging the knowledgeable to spend their time, not in cultivating and enjoying their knowledge, but in teaching the ignorant. Likewise, if some people own a lot of land and others only a little, the former might advantageously be obliged to give land to the latter. A redistributive offer in the opposite direction, involving transfer of a good from the have-nots to the haves, would presumably prove to be inferior inasmuch as there would be much less to transfer. Poor-to-rich transfers would, in typical democratic circumstances, produce a less favourable, indeed a downright negative balance between support gained and lost.


If there are any number of inequalities (though only a few will really yield to levelling), the state can at least propose or pretend to level a number of them. If so, it is impossible to predict the most efficient redistributive offer from the initial distributions alone. Even the presumption that transfers from the haves to the have-nots (rather than the other way round) are politically superior, may not stand up if clout matters much more than votes and it is the haves who have the clout.*6


In order to make a determinate solution possible, it would help to have a political culture where most inequalities were accepted as untouchable, so that neither the state nor its competitors would include them in a redistributive offer. In such a culture, for example, children would be allowed to be raised by their own (unequal) parents; non-income producing personal property would not have to be shared; people could wear distinctive dress; unpleasant work would be done by those who could not get any other, etc. Obviously, not all societies have this sort of culture, though those we call consent-based by and large do. Culture, then, would severely narrow down the possible variety of political offers. However, to rule out any freak programme and cultural revolution, it will be best to consider first a society where only one inequality is "politically" perceived at all: the amount of money people have.


Money looks the natural object for redistribution because, unlike most other interpersonal differences, it is par excellence measurable, divisible and transferable.*7 But it has a subtler advantage, too. At least conceptually, there are political processes which run their course, achieve their objective and comes to an end. The class struggle between capital and the proletariat is conceived in Marxist thought to be such a process. Once this terminal conflict is resolved and there is no exploited class left for state power to oppress, politics comes to a full stop and the state withers away. Likewise, if politics were about latifundia and landless peasants, or the privileges of the nobility and clergy, or other similar inequalities which, once levelled, stayed level, the state's purchase of consent by redistribution would be an episode, a once-for-all event. At best it could be history made up of a succession of such episodes. However, with money as the object, democratic politics can make sense as a self-perpetuating static equilibrium.


Why this is so is best appreciated by recalling the facile distinction that people so readily draw between equality of opportunity and of end-states. Moderate egalitarians sometimes suggest that it is opportunity that ought to be equal while end-states arising out of equalized opportunities ought to be left alone (which could only be done with mirrors, but that is now beside the point). Peter and Paul should have the same chances of attaining any given level of income or wealth, but if in the end he were to make more, Peter should not be robbed to pay Paul. Inequality of income or wealth is in turn, however, the resultant of a large universe of prior inequalities, some of which can be equalized (but then at least some end-states must be permanently interfered with; compulsory free education must be paid for by somebody), while others cannot. If Peter has in fact made more money, some prior inequalities in his favour must have subsisted.


A little reflection shows that there is no other test of the equality of people's respective opportunities to make money, than the money they do make. For once inheritance of capital is abolished, everybody is made to go to the same school and every girl is given cosmetic surgery at eighteen, there are still ninety-nine well-known reasons why one person may be materially more successful than another. If these known reasons (notably one's parents) were all abolished and it were impossible to inherit more ability than the next fellow, we should be left with the unknown residuals habitually subsumed under "luck."


This need not stop anyone from choosing some stipulative definition of equal opportunity, making it an arbitrary subset (to include, say, equal attendance at school, "careers open to talents" and provision of fixed-sum unsecured loans for starting a business, and to exclude everything else such as happening to be in the right place at the right time) of the set of reasons which make end-states unequal. One might stipulate that all who have danced with the most coveted girl at the ball are deemed to have had an equal opportunity to win her. If she gave her affections to one, rather than equally to all, that was luck.


The point is not only that equality of opportunity is conceptually dubious, nor that as a practical matter serious egalitarians must deal with end-states—for that is how you go about equalizing opportunities—though both points are valid enough. It is, rather, that each time end-states are equalized, sufficient underlying "inequality of opportunity" will subsist rapidly to reproduce unequal end-states. They will not be identically the same ones. Redistribution must, intentionally or otherwise, have some influence on the causes of a distribution, if only through its much-invoked effects on incentives—the idea being that if you keep taking away the golden eggs, the goose will stop laying them. Nevertheless, some new unequal distribution will almost instantaneously come about. It will require redistribution to be recurrent (an annual assessment?) or fully continuous (pay as you earn). In any case, there is no danger that the state, by vanquishing the inequality of money, would unwittingly depreciate its own role and "work itself out of a job."


In looking at the conduct of the state in competitive politics, we will for some of the above reasons make the large simplifying assumption that it rules over a society which is an amorphous collection of people lacking any pattern. It does not coagulate into groups, occupations, strata or classes on the basis of material and moral inequalities. It is the ideal democratic society in Rousseau's sense in that it does not break down into sub-societies, each with a general will of its own, in conflict with the general will proper. There are no intermediaries, historical or functional, personal or institutional, between the individual and the state. Though people are thus homogenous, I will nevertheless take it that they have significantly different amounts of money due to "unequal opportunity" or, less controversially, to luck.


Quite unrealistically but expediently, I will also suppose that everybody's political choices are entirely determined by their material interest, and in a narrow sense at that: there is no altruism, no false consciousness, no envy and no idiosyncrasy. When given the chance, people go for the policy which gives them the most money or takes away the least, and that is all.


The other simplifying assumptions we need are less demanding. The basic democratic rules apply. Tenure of state power is awarded to a contender on the basis of a comparison of open competitive tenders describing redistributive policies. The actual tenant is the state. If another competitor were awarded tenure, he would become it. Tenure is for a specific period. There is some provision for premature termination—"recall"—in case the conduct of the state is in gross breach of the terms of its tender offer. If there were no recall, and the period of assured tenure of power were long enough, the state might promise one thing and do another, inculcating in society the corresponding new tastes, habits and addictions and developing support for what it was doing rather than for what it had said it would do. Though this is obviously happening in real politics, for government would become quite impossible otherwise, our analysis would grow immensely complicated if we did not exclude it by postulating easy recall. Award of state power is to be decided by simple electoral majority, one-man-one-vote and secret ballot. Entry to politics is free, i.e. anyone may tender.


Under these assumptions, towards the expiration of each period of tenure there will be competitive bidding for votes by the state and its opposition. The highest tender will, at the appointed time, earn the award of fresh tenure. Which, however, is the highest tender? Neither the state nor its competitors have any money to offer which does not already belong to somebody in civil society. Neither can, therefore, offer to civil society a total net sum greater than zero. Yet each can offer to give some people some money by taking away at least that much from others. (It makes for ease of exposition if collecting taxes is, at this stage, taken to be a costless operation.) The redistributive policy such an offer represents can be regarded as a tender with discriminatory pricing, some votes being bid positive, and others negative, prices—with the crucial proviso that if the tender in question wins, the people whose votes have been bid negative prices will have to pay them no matter how they voted. (As is perhaps obvious, people offered a negative price for their votes may rationally vote either for or against the tender in question, depending on how much a competing tender, if it prevailed, would make them pay.)


Our argument will lose nothing if we simulate the two-party system and consider only two rival tenders, one submitted by the incumbent state and the other by the opposition (which may of course be a coalition), while assuming sufficient ease of entry of potential competitors to prevent the state and its opposition from reaching collusive agreements to share spoils and underpay votes. (The American political system, for one, has in recent years been showing symptoms of incipient collusion, in the form of the bipartisan commission taking over from the adversary-type legislature, where competition had led to stalemate over such questions as the budget deficit or the lack of control over social security expenditures. Despite the attractions of collusion, ease of entry and many other built-in elements of competitiveness make it in my view unlikely that government by bipartisan commission should get very far in superseding the basic rivalry of "ins" and "outs.")


If society is differentiated by riches only, state and opposition have only two roles to divide between them, that of champion of the rich and champion of the poor. Who takes which role may be decided by historical accident; for our purposes, it may as well be decided by spinning a coin. The winning tender must attract 50.1 per cent of the votes. There are thus always 49.9 per cent of the people whose money can be used to buy the votes of the 50.1. Any greater percentage bought would be wasted. No rational tenderer should under these assumptions bid positive prices for more than 50.1 per cent. If he did that, he would by implication be taking money away from less than 49.9 per cent. He would be proposing to redistribute a lesser total sum among more people. In trying to get too many votes, he would be reduced to offering a lower price for each. He would be outbid by his competitor who (as future generals are taught to do) concentrated his fire to get the necessary and sufficient bare majority. In this streamlined political contest, any election result other than virtual dead heat would be proof that at least one competitor had not got his sums right and had handed victory to the other.


So far, so good; this simplified schema duly reproduces the complicated real world's tendency to make close-run things out of democratic elections in two-party systems where competent professionals on both sides strive to be all things to all men and fine-tune their electoral promises. What, however, seems left unpredicted is the winner. We know that the highest tender wins. But we do not know the terms of the competing tenders.


Let us arbitrarily suppose (the argument will gain no unfair advantage if we do) that you can get, say, ten times as much tax from the rich half of society as from its poor half, and that either competitor for state power can propose to tax the rich, or the poor, but not both at the same time. The latter condition makes redistribution conveniently transparent, though it is of course quite possible to redistribute without respecting it. Let us also suppose that both competitors have the same idea of taxable capacity, more than which they will not attempt to extract from either half of society. "Taxable capacity" is an embarrassingly nebulous concept, to which I shall have to return later in dealing with the causes of "churning." It is usually employed in the sense of some economic capacity, having to do with the effects of varying degrees of taxation on taxable income, output, effort and enterprise,*8 the implicit assumption being that everybody's willing performance of their tasks depends, inter alia, on how hard they are taxed. I am employing the concept in both this sense and also in a parallel one, as a relation between taxation and the subjects' willingness to abide by the rules of a political system under which a given share of their income or wealth is taken away from them, the implicit assumption being that the greater this share, the less the subject feels bound to respect rules under which he is made to surrender so much. "Capacity" suggests that there is some limit beyond which the economic or political tolerance of taxation declines, perhaps quite abruptly. Both the economic and the political senses of the concept are shrouded in fog. No one has yet convincingly depicted the shape of the relation, nor did anyone measure its limits. Discussion of it is apt to degenerate into rhetoric. However, unless we are prepared to take it that for a society at any point in its historical career, there are such limits, and that it takes history, i.e. the long period or large events in the short period, to shift them by a lot, much in social affairs must fail to make sense. In the context of the problems we are pursuing there would, for instance, be no intelligible reason why, spurred on by democratic competition, the state should not subject large sections of society, possibly fully one-half of it, to marginal tax rates of 100 per cent.


(If there is no such thing as a "taxable capacity" which taxation cannot exceed without bringing about a high likelihood of political or economic anomie, turbulence, disobedience and breakdown of some possibly obscure kind, unpredictable as to its specifics but unacceptable in any case, it must be feasible as of tomorrow to tax everybody at 100 per cent—"from each according to his ability"—and to subsidize everybody at the state's discretion—"to each according to his needs"—without first having to put society through the phase of the dictatorship of the proletariat. Despite its apparent convenience, this programme cannot really appeal to socialists who, if they had to choose, would probably rather agree that taxable capacity is limited than give up the requirement of fundamentally changing the "relations of production," i.e. abolishing private capitalist ownership.)


Since the winning tender is one which is "accepted" by not less than 50.1 per cent of the voters, the two competitors will seek to hit upon the winning combination of positive and negative "prices" for the richest 49.9 per cent, the poorest 49.9 percent and the middle 0.2 per cent of the electorate.



(1) The rich party might propose to tax the poor, redistributing the money so collected to its own constituency and (in order to form a majority coalition) to the middle. The poor party might symmetrically propose to tax the rich and transfer the proceeds to its own poor constituency and the middle. Table 1 shows us what we would then have.

Table 1
Rich party
Poor party

To the rich brace bracket +1 -10
To the middle brace bracket +10
To the poor -1

0 0


(2) The rich party, however, would immediately realize that its offer under (1) is bound to be rejected, for there is always more money available for buying the votes of the middle out of the taxes of the richer half than out of those of the poorer half. It must, therefore, steal the poor party's clothes and turn upon its own constituency. (This is, of course, what rich parties do in real-life democracies.) Table 2 shows how the two tenders will then compare.

Table 2
Rich party
Poor party

To the rich -9 -10
To the middle +9 brace bracket +10
To the poor 0

0 0


(3) Under (2) the rich party would win. It would get the acceptance of the rich who would prefer to be taxed 9 instead of 10, and of the middle who would prefer to get all the pay-off rather than having to share it with the poor. However, "going for the middle ground" is a game two can play; to stay in the race, both must. So the outcome is as in table 3.

Table 3
Rich party
Poor party

To the rich -9 -10
To the middle +9 +9
To the poor 0 +1

0 0



Neither competitor can further improve its respective tender. Logically, both are equally apt to secure the consent of the majority. The rich party's tender is voted for by the rich, the poor party's by the poor. The middle is indifferent between the two offers. It is equally rational for it to join the top half or the bottom half of society or to toss a coin.*9


The astute reader will have divined that the simple mechanism laid bare above, through which democracy produces redistribution, would continue to operate, mutatis mutandis, in a setting where a constitution forbade redistribution. (The Fifth and Fourteenth Amendments of the American constitution were, for a time, held to do so.)*10 If there is no way round it, perhaps by taming the guardian of the constitution, it must be amended, brought up to date, adjusted to changing circumstances. Instead of 50 per cent, it is then the qualified majority which the constitution requires for its own amendment, that becomes the dividing line in society between top and bottom, rich and poor. The pay-off out of which to fashion a redistributive offer which will, at least under the assumption of consent being solely a function of alternative offers of public money, secure support for amending the constitution, is the money that can be taken from the blocking minority if it is amended.*11


The artificial mechanics of competitive political tendering, which produce the equally artificial result of finely balanced electoral indeterminacy, must of course be taken with a pinch of salt. Neither the state nor its opposition, no matter how coldly professional and competent at engineering electoral platforms, could possibly formulate patterns of seduction with anything like the precision required for our result. Nor would all voters correctly understand and evaluate the prices that were being bid for their support, i.e. the incidence on their income of complex redistributive policies. Many of these might be presented to look more lucrative to the gainers or less costly to the losers than the probable reality. Ignorance, the unpredictability of true incidence and the opacity of social and economic matter, would handicap not only the electorate but also those seeking to gain its support. Even if both competitors used the same data, the same surveys sold by the same pollsters, they could not risk sailing this close to each other. In reality, the coveted middle ground, too, must be much broader than in our illustration, and its benefits from redistribution more diluted.


Nevertheless, for all their artificiality, observing the workings of our schema of electoral democracy is more useful than looking at the mere spinning of wheels. It confirms in the simplest possible manner an intuitively plausible presumption: that material interest alone is insufficient to determine the award of power to one contender rather than another, for the contenders, even if they carry different flags, end up by appealing to substantially the same interests, which they attract by holding out much the same pay-off. The more familiar corollary of this is the "convergence of programmes," the tendency (which some consider a strength of democracy) to narrow down the range within which policies (as well as the images candidates for high office must project) remain electorally viable. The obverse of this coin, of course, is the complaint of the non-conformists that electoral democracy precludes genuine, distinctive alternatives; the very principle of popular choice leads to there being little to choose from.


Our account of the "pure," rich-to-middle tax-and-transfer kind of redistribution which the state, confronted by rivals in electoral democracy, would adopt under certain simplifying assumptions, is to a general theory of redistribution as, in economics, perfect competition is to a full theory of producers' behaviour. It is a stepping stone or heuristic device without whose help more general propositions might not emerge clearly enough. Though I neither claim, nor require for my arguments, to propose a general theory of redistribution, I do sketch some likely looking components of such a theory in the rest of this chapter. Their intent is to explain some of the dynamics of how civil society, once it grows addicted to redistribution, changes its character and comes to require the state to "feed its habit." From benefactor and seducer, the role of the state changes to that of drudge, clinging to an illusory power and only just able to cope with an inherently thankless task.


We have learnt that consent is, by and large, not bought with acts of once-for-all state help to the majority at the expense of the minority. Help and hindrance must be processes, to maintain a stipulated state of affairs which, without such maintenance, would revert to something rather (though never exactly) like what it was before. The beast must be fed continually. If this must be performed under conditions of open democratic competition, whatever of its subjects' liberty and property the state manages to appropriate, must be redistributed to others. If it does not do so, the redistributive offer of its competitor would beat its own and power would change hands. Tenure of power, then, is contingent upon its not being used at the state's discretion. The resources over which it gives command must be totally devoted to the purchase of power itself. Thus, receipts equal costs, output equals input. The analogy with the firm which, in equilibrium, can by maximizing profit do no more than earn its factor costs (including the entrepreneur's wages), is compelling.


We are nearing the heart of the matter, bumping as we do at this juncture into the theory of the state. If the point in being the state were to have power (that is, if that were the state's maximand, its end), it would mean very little to say that the state has maximized it in the situation whose equilibrium conditions we have deduced above. Social power, as we know from Max Weber, is its holder's capacity to make, by recourse to combinations of physical force and legitimacy, another do what he would not have otherwise done. The quintessential democratic state has the capacity to make given subjects in civil society surrender to it given parts of their good. They would not have done it without its "power." But it has no capacity to make them surrender any more nor any less. It would lose "power" if it tried. It must tax the subset S of society an amount T, and it must distribute T' to another subset U. It cannot alter either S or U, it cannot vary T nor have T' fall short of it. It must not indulge its sympathies, follow its tastes, pursue its hobbies, "make policy" and generally promote the good as it conceives it, on pain of being booted out.*12 Though it can make another do something the latter would not have done, it cannot choose what it will make him do. It lacks the other essential attribute of power: discretion.


If power as an end in itself meant "being in power," it would not matter to the power-holder that he must use it in one unique way, only for this and not for that, as long as he held it. But it would make for shallow theory to put this in the role of maximand. By the same token, we would get only a theory of snobbery if we were to put holding a title of nobility as the purpose of the noble's existence, stripping out estates, privileges, ethos and social and economic functions. The state could not use this residual sort of power, nor seek more of it. It could only have or not have it. If it were satisfied with it, pure electoral democracy would be a sort of terminal stage of political development, and our argument would be substantially at an end.


But while relief from further labours might be a pleasant by-product for the writer and his reader, allowing the state to be motivated by such a shallow, near-empty concept of power would grossly misrepresent historical experience. It would contradict, or at least leave unexplained, the state's evident striving over most of modern history for more autonomy, for discretion in deciding what it will make people do. Only the will to have power as a means can properly explain that. The logic of competition, however, is such that democratic power in the limit becomes the antithesis of power as a means to freely chosen ends.


That the wheel thus comes round full circle is yet another illustration of the distant consequences of actions in and upon society being mostly unintended, unforeseen or both. A state seeking to govern mainly by consent instead of by repression cum legitimacy, may have fallen victim to lack of foresight, weakness of will or inconsistency. But it might equally well have been rational, when seeking greater freedom of manoeuvre, readier obedience, lesser reliance on narrow class support—in short, when seeking more discretionary power—to look for it in democratic reforms, in increasing reliance on consent. At the outset, it positively provoked its subjects to make demands upon it, as a vendor might drum up custom for his wares by passing out samples and testimonials, in order to create a political market in which consent could be earned in exchange for state provision of utility and equality. At the end of the day (most such days lasting about a century), such states found themselves, in a special but quite precise sense, virtually powerless, having their policies decided for them by the need of competitive electoral equilibrium and generally running hard to stay in the same place. It is academic to ask whether they could have foreseen this sort of result. Plainly, they have not. In exoneration, they had less warning than Adam before he ate from the tree of knowledge.

Notes for this chapter

The latter need not be the case. In the winter of 1973-4, the British coal miners proved to have enough clout to break Edward Heath's government; yet with respect to the inequalities which would be liable to figure in a redistributive offer, they would clearly count as have-nots.
I prefer naively to talk of "money" and leave it to others whether it is income or wealth or both that should be redistributed and what difference it makes.
If there were no such effects, taxable capacity would be equal to income, i.e. the very concept would be perfectly redundant. People could be taxed at 100 per cent of their income, for doing so would not adversely affect either their ability or their willingness to go on earning it.
With the same rules and the same players, Robert Nozick, in Anarchy, State and Utopia, 1974, pp. 274-5) reaches the contrary conclusion; he sees the rich party as the sure winner. Nozick's argument is that "a voting coalition from the bottom won't form because it will be less expensive to the top group to buy off the swing middle group than to let it form"; "the top 49 per cent can always save by offering the middle 2 per cent slightly more than the bottom group would." "The top group will be able always to buy the support of the swing middle 2 per cent to combat measures which would more seriously violate its rights."

I cannot find the reason why this should be the case. An identical pay-off is potentially available to either the top or the bottom coalition. It is what the bottom coalition gets, or the top coalition keeps, if it succeeds in forming. (In my example, the pay-off is 10). Rather than become the minority, both the top 49 per cent and the bottom 49 per cent would gain by offering some of the pay-off to the middle 2 per cent to make it join a coalition. The middle group would agree to the higher offer. The potential maximum offer is, of course, the entire pay-off (10 for both parties). But if either half did offer to give the whole pay-off to the middle for the sake of becoming part of the majority coalition, it would end up no better off than by resigning to become the minority. The game would not be worth the candle. The highest offer to the middle which it would be rational for either the top or the bottom half to make, therefore, would be the whole pay-off less the sum needed to make it just worth either half's while to coalesce with the middle rather than passively accept defeat.

This sum may be, for all we know, large or small (in my example, I used 1). Whatever it is, if it is the same for both halves of society, the top and the bottom coalitions are equiprobable and the result is indeterminate. For the contrary conclusion to hold, the poor must require a greater inducement to coalesce with the middle than do the rich. There seems to be no particular reason for supposing that this is more likely to be the case than not—at least, I cannot see one.

Let us note, before passing on, that in Nozick's scheme the top group and the bottom group would have to take the trouble of negotiating a coalition with the middle. In our scheme, the state and its opposition relieve them of this trouble by each presenting a ready-made deal, an electoral platform which they can simply vote for or against.

Cf. F. A. Hayek on the "Curious Story of Due Process," in The Constitution of Liberty, 1960, pp. 188-90.
If 25 per cent can block the amendment, the pay-off is whatever 24.9 per cent can be made to hand over to 75.1 per cent.
Cf. the essay of J. G. March, "The Power of Power," in D. Easton, Varieties of Political Theory, 1966.

End of Notes

26 of 33

Return to top