RTowards a Theory of Taxation 1

Published in Social Philosophy and Policy,

Volume 2 Issue 1, Autumn 1984, pp.161-173.


"Towards a Theory of Taxation" is a proper theme for an Englishman to take when giving a paper in America. After all it was from the absence of such a theory that the United States derived its existence. The Colonists felt strongly that there should be no taxation without representation, and George Ill was unable to explain to them convincingly why they should contribute to the cost of their defense. Since that time, understanding has not advanced much. In Britain we still maintain the fiction that taxes are a voluntary gift to the Crown, and taxing statutes are given the Royal Assent with the special formula, ``La Reine remercie ses bons sujets, accepts leur benevolence, et ainsi le veult," instead of the simple ``La Reine le veult," and in the United States taxes have regularly been levied on residents of the District of Columbia who until recently had no representation in Congress, and by the State of New York on those who worked but did not reside in the State, and so did not have a vote. Taxes are regularly levied, in America as elsewhere, on those who have no say on whether they should be levied or how they should be spent. I am taxed by the Federal Government on my American earnings and by state governments on my American spending, but I should be hard put to it to make out that it was unjust. Florida is wondering whether to follow California in taxing multinational corporations on their world-wide earnings. There is some debate on whether such a move is expedient: but nobody to my knowledge has suggested that, in order to make it just, multinational corporations, or their overseas shareholders, should be given representation in the state legislature. Indeed, were anyone to make such a suggestion, he would be laughed out of court, so dead is the contention that representation is a necessary condition of taxes being justly levied. And yet I do not want to bury the Colonists' argument utterly. They had a point. Although the main issues in an adequate theory of taxation are those of justice, there is also room for a quasi-Lockean approach in which we try to make sense of taxes as a sort of bargain in which the taxpayer surrenders his individual quid in return for a share of a collective quo. [162] If we are to develop a theory of taxation and address ourselves to the Colonists' grievance, we need to purge ourselves of contemporary complacency and take an unfashionable stance. For it is generally accepted among modem thinkers than argument. But it is also due to mistaken views of political morality which are open to rational debate. It is widely taken for granted that it is Wrong to be Rich. It may, indeed, be true that it is wrong to be rich, but it needs to be argued for, not taken for granted, and if it is argued for, the arguments will have an important bearing on how the principle is to be applied in practice. It is quite plausible to maintain that to seek riches has a bad effect on the character. It is arguable that even to possess them imposes an unacceptable cost in worldly worry and concern. Jesus said it was hard for the rich to enter the Kingdom, and advised the young man to give away all he had. It is a matter of dispute whether his advice was directed to just that particular young man or was a general precept addressed to everyone. Many Christians have heard and heeded a call to poverty, although many others have believed that they should live in the world and make the most of its resources, for others as well as themselves. but whatever their decision, the question is one of personal morality, not public morality. It may be right for me to be a slum priest, forgoing money, family and freedom for the sake of the Gospel: but that is a decision that only I can take. It is no good if you decide that I should be poor, celibate and under rigid discipline, or if I wish that vocation on you. There is no merit in the choice unless it is made wholeheartedly by the person concerned, and it is unlikely to work - the unwilling mind will continue to dwell on the goodies it has unwillingly renounced, and its underlying bent will manifest itself in all sorts of covert ways. just as it is better to marry than burn, so it is better to be rich than be always thinking about money, or always being angry at others being rich. Even if we are generally persuaded, as medieval men were generally persuaded, that the vocation to poverty was a better one than that to involvement in the world, we should still not make it a basis of our public morality which we reckon to be incumbent on all men irrespective of their personal inclinations. We may all commend the example of St. Francis, but still should not conclude that the Queen should play the part of Robin Hood.

It is often thought that the State should redistribute from rich to poor not because it is Wrong to be Rich, but because Equality is a Good Thing. I have argued against egalitarianism at length elsewhere and shall not weary you with repetition. In the context of taxation, however, there is one variant of the egalitarian argument which is superficially attractive. It seems natural to regard the relief of poverty as a possible aim of public policy - one does not have to be an egalitarian to regard poverty as a bad thing, and only extreme [163] advocates of laisser faire would say that it should be no concern of the State. But then we seem to have a licence for Robin Hood. Whenever there is a choice we should always devote public funds to the fashionable minority group of the moment rather than reduce the higher rate tax demands on the Duke of Westminster. It is this contention that I want to controvert, and in particular the second half. As regards the first half, it has been increasingly recognized that there are limits to what can be achieved by public expenditure in the way of relieving poverty. No matter how much we spend, the poor will always be with us. But although this is a reason for doubting the efficacy of egalitarian policies, there are always going to be some expenditures which look like they might do some good for some poor people; and unless there is some counter argument for not taxing the rich to the limit, the plea of poverty will win the day.

Some thinkers seek to block the Robin Hood argument by claiming that taxation ought not to be redistributive. This claim is ambiguous. In one sense, taxation cannot help but be redistributive. The tax collector takes money from me, and distributes it to soldiers, sailors, and civil servants. Even if there were no welfare payments as such, much of my money would go on expenditures which benefited others much more than myself. airports and the M25 for the rich, the health service for the ill, public education for the uneducated. We cannot lay down that taxation should not be redistributive in this sense. But even if it is permissible for taxation to have redistributive effects, some people would still argue that it should not be intended to redistribute resources. It is, they argue, contrary to the nature of the State that it should extract from an unwilling individual taxes for any other purpose than the protection of his rights against criminals at home or aggressors from abroad. Such a theory of the State is difficult to refute because it is difficult to formulate coherently. I do not think that any such theory of a Minimum State can be sustained. I have argued the point elsewhere, but only cursorily. 2 For our present purpose, I do not claim that no such theory could be put forward, but only that other theories of the State are put forward and acted on, and that for these, too, a theory of contributive justice is needed. Most people do not believe in the Minimum State. They want to engage in collective action, e.g., in sending a man to the moon, and see no reason why they should not act collectively through the mechanism of the nation-state. And if they may engage in collective action generally, then they can choose as a permissible goal of collective action the relief of poverty, which will be redistributive not only in effect but---at least in one obvious sense---in intention too. Of course, there may still be reasons for being chary of adopting redistributive policies because they are inherently divisive and [164] often give rise to abuse. But keeping pork barrels out of politics is a counsel of prudence rather than the imperative of political morality. Many redistributive policies at present practiced are unwise, and some are plain wrong. But I cannot see that there is, or even could be, a cogent argument, either of justice or of expediency, for holding that the relief of poverty or the amelioration of misfortune is not a proper object of public concern or public expenditure; I can, however, see arguments, both of justice and of expediency, for saying that the tax burden should be fairly borne by the whole community and not put disproportionately upon the shoulders of the rich.

Aristotle considers the allocation of benefits and burdens together under the title of distributive justice. But the cases are not all similar. When benefits are being distributed, I am not likely to shirk my share, and if I do, no great problem arises: if I do not want my piece of cake, there are plenty of others who will take it up for me. With burdens, however, each person is naturally reluctant to bear them, and so is tempted to shirk and may actually do so. Fare-dodging, free-riding, tax-evasion, and generally not pulling one's weight, are endemic problems when in pursuit of a collective goal we have to call on individuals to make unwelcome sacrifices as their contribution towards it. These problems ought to be considered quite separately from those of distributive justice, and I propose the name "contributive justice" instead.

Essentially the problem is that it is in each person's interest, selfishly considered, not to contribute, but if we all fail to contribute, the collective goal will not be achieved, and we shall all be worse off. Although it seems to be a sensible policy for me not to pay my taxes, or not to buy a ticket, relying on everybody else to pay his bit all the same, if it is sensible for me, it is sensible for anybody else, and if it is sensible for anybody else it is sensible for everybody else, and if it is sensible for me and everybody else it is sensible for us all, and yet if we all did it, the results would be disastrous. And so I oppose my seeming self-interest, which I can see to be shortsighted, by an enlightened self-interest, which is longer term and can embrace others as well as myself. But it makes sense for me to contribute only on the condition that everybody else does. If I act on enlightened principles and they act according to their short-sighted, seeming selfinterest, then it is I who am being a fool, not they. I need assurance that if I do my part, they will do theirs. Only if I can rely on them is it individually rational for me to do what it is collectively rational for us to do.

Protagoras was the first to formulate the problem, and Plato saw in it the rationale of the many agreements and tacit understandings which underlie law and conventional morality. Hobbes reckoned covenants without swords to be but empty words, and so sought a sovereign armed with coercive [165] power. In recent years a paradigm example, called the Prisoners' Dilemma, has been much studied with the aid of the Theory of Games. At one level of reasoning Hobbes has been vindicated. Coercion is needed, because some people are moved only by short-term self-interest, and others fear that this is so. Coercion, in the form of tax inspectors and penalties for tax evasion, is essential, because otherwise some will not pay their share, and others, knowing or believing this, will then be reluctant to pay theirs either, for fear of being mugs. But the argument does not go as far as Hobbes thought. The reason is, at the crudest level, that not even Leviathan has enough eyes to see everything that is going on. Nobody knows how much revenue is lost through fiddling and non-declaration, but there is plenty of reason to suppose that it is quite a large sum. Threats by themselves cannot be totally effective because they can only be invoked on the basis of adequate information, and the State is not possessed of perfect information, and can obtain the information it needs only with the co-operation of its subjects. That cooperation will be forthcoming only on the basis of loyalty and trust. Although I shall do what I am told by the Gestapo or the K.G.B., I shall not tell them more than I can help or expose myself to their attentions in the way that I am prepared to co-operate with the reasonable demands of a civil society which does not threaten me and with which I can quite largely identify.

Hence, although the State needs to have some coercive power at its disposal in order to bring sanctions to bear on those who will not otherwise keep their covenants, it cannot rely on the sword alone, but must secure a general measure of agreement to its general aims. Although I want to be assured that those who break the law will be punished, so that I am not being a mug when I myself forbear to break it, my own reason for keeping the law is not primarily that I shall be punished if I break it, but that I see the rationality of my, and everyone else's, keeping it, and am reasonably assured of others' actually doing so.

The fiscal application is clear. Taxes must be not simply what I can make others pay on pain of punishment, but what I can see good reason to pay as a matter of enlightened self-interest and reasonable identification with my fellow members of the State, and in reasonable confidence of their paying too. If these conditions are satisfied, then I will be fairly well disposed to pay my share. And if I can see that it is satisfied for everybody else, I can suppose that they are likewise fairly likely to pay. The temptation not to pay, although present, is well resisted by my recognition of the rationality of the demand and the knowledge that others too will be able to recognize the rationality of what is demanded of them. I am not going to feel myself a mug if most of the others are paying too, and the tax-evaders are a small minority, and in danger of being found out and not only penalized but severely stigmatized as well. Fare-dodging does not seem clever because we all can see why fares should [166] be paid, and know that most other people pay their fares, and do not want to copy Professor Joad.

The case has become very different with taxation, especially income tax. I can still remember the time when it was thought to be a moral duty to pay one's taxes, and tax evasion was a serious social crime. Nobody thinks that now. A generation of high and avowedly punitive taxation has eroded the moral basis of fiscal policy, and has developed in Great Britain an Italian attitude towards the Inland Revenue. The rot set in from the top. Rich men felt---were encouraged to feel by some Chancellors---that they were being unfairly taxed, and ceased to feel any moral obligation to contribute to the Exchequer the very large amounts demanded of them. They became less scrupulous about declaring unconsidered trifles, and understandably eager to rearrange their affairs so as to minimize the tax man's exactions. For forty years successive Finance Acts have sought to block up one loophole after another, but have by and large been defeated by the fact that people who run things have to have discretion, and cannot be stopped, except by moral considerations, from using their discretion to their own advantage. The employer finds a job for his wife, sends the works van to collect his son from school, and gets the milk-maid to lend a hand in the rose-garden. The Inland Revenue seeks to forbid it, but finds that it is forbidding the employer to take on anyone, or is requiring the doctor and his receptionist to live in sin in order that she may be paid for answering the telephone. And the example spreads. If the director or the Cabinet Minister has a company car, why not the executive or Assistant Principal? Top people are inherently influential, and if they try to avoid paying taxes, those lower down follow suit. Even if top people do not, they are thought to. For it is known that the rates are very high, and judging others by myself I reckon that if I were paying 60% on my marginal income, I would see to it that any extra money that came my way came in some untaxable form. And if I believe that Lord Vestey and all the others at the top have sewn up their affairs pretty neatly, I shall soon persuade myself not only that I should be a mug to pay my 30%, but that I have a positive duty to fiddle.

The account I have given is depressingly familiar. But what is to be done? We cannot go back to the pre-Daltonian Garden of Eden. Public morality, although easily eroded, is difficult to restore. Nevertheless, there are measures open to the Government. The crudest, although not the most fundamental, is that tax avoidance is a tedious, inconvenient business. Money is usually much more useful than a perk. I would rather have cash than a company car. In fact, I would settle for quite a lot less cash than the cost of a company car, simply because it was less hassle and enabled me to choose between a new car and a holiday in Dalmatia. Although not all perks would be commutable for cash---some serve a functional purpose---many [167] are. And although once established, a perk is likely to continue, the marginal disutility of paying people by perk instead of in cash could be assessed, and could provide useful pointers to the economically efficient marginal rate of taxation. Better a 3 5% rate which was generally paid than a 65% which was largely avoided.

More important, however, would be a shift in the context of discussion of fiscal policy. At present it is carried on almost entirely in terms of what can be got out of other taxpayers, instead of what can fairly be asked of each. We need to put at the forefront of the debate ``Is it fair to ask this poor man to pay 30% at the margin?'' ``Is it fair to ask this man to pay 25% of his total income in tax?'' ``What is the most that it is fair to demand?'' These are crude questions, but for that very reason widely intelligible. If people are able to understand what is being asked of them and of others they can appreciate the issues of contributive justice involved and, per contra, the questions need to be crude, so that they shall be widely intelligible, and that people can understand not only their own burden but what is being asked of others, and can reckon it fairly tolerable and so likely to be borne. Contributive justice is crucially concerned not only with people's perceptions of their own situation but their perceptions of other people's situation and their perception of other people's perceptions of their own situation. It has to be crude in order to secure second-order intelligibility, and thus provide the general assurance, essential if we are to escape from the Prisoners' Dilemma, that other people too will contribute their share.

For this reason the rationale of all sorts of taxation, and especially of progressive taxation, needs to be articulated and discussed. Although often a per capita contribution seems appropriate, as in a club subscription or public transport fare, we can argue that a proportionate contribution is a more appropriate premium to pay for the protection of one's property and life style from anarchy and external enemies. The argument for progressive taxation is less obvious. Perhaps a pools winner is getting proportionately more from the State than a wage earner, and perhaps he is willing to shoulder a proportionately heavier burden, but the argument needs to be made out. "If you were to win the pools would you think it fair to have to pay 40% of your income in tax?"---if people are asked that question they are likely to have a much clearer idea of the issues at stake than if they think of it merely as what others are going to have to pay. And if they can answer that question hypothetically in the affirmative, they will know why pools winners should, and will have some reason to suppose they would, pay their contribution to the Exchequer. A climate of opinion would begin to grow up of mutual understanding and shared expectations about what different people in different circumstances were to pay, and this would greatly militate against avoidance.


Three underlying issues arise in the debate on fiscal fairness: the justice of selective disincentives; the individual's benefit from the collective good; and the real burden of the sacrifice imposed. The selective imposition of some burdens is justified not simply on grounds of contributive justice, but in order to further some other public policy. We tax alcohol, tobacco, and gasoline, partly in order to discourage people from drinking, smoking, and driving; drinkers, smokers, and drivers put both themselves and others to inconvenience and risk. This is not the only reason for taxing these commodities, but it does give some sort of justification, not itself intrinsic to contributive justice, for picking on some people rather than others on whom to levy taxes. It is because it has been generally reckoned wrong to be rich that the top rate of tax in Britain was set at 98%. But if it is not inherently wrong to be rich, then this extrinsic argument for high income tax fails.

The second justification for imposing heavier burdens on some rather than on others is by reference to the benefits received in return. Some collective goods, although in principle available to all, are in practice enjoyed only by a minority. Only a small proportion of citizens avail themselves of consular services. Many poor people never go abroad at all. Only a minority of people go to university or have children who go to university. Public libraries and municipal museums are open to all, but not patronized by all. If all public expenditure were financed by a poll tax, it would be unfair on the poor, who would be paying for facilities for the rich which the poor either had no opportunity of using or no wish to use.

Collective goods of this type can be individualized by subscriptions: in Britain the National Health Service has always been financed in part by a per capita levy, Mr. Heath tried to introduce museum charges, and there is a growing reluctance on the part of the taxpayer generally to pay for university education for the few. But it is not a clear-cut issue. Many collective goods are difficult to price and cumbersome to provide through market mechanisms. We cannot rule out their being provided by some sort of collective action, even though sometimes the appropriate collective may not be the nation-state. In quite a few cases the collective good will be provided by a public authority, and funded from taxation, but it is not obvious that everyone should pay the same tax. What differentiating principles are appropriate?

The underlying concept, here, is that of the Good Bargain. Each taxpayer is paying a certain amount of taxes and enjoying various collective goods. Nobody is enjoying all the collective goods provided, but those that he is enjoying would cost a great deal more to obtain if they were not provided collectively. My pennies spent on the municipal baths give me no return, but the fire service would cost me much more if I had to pay for it privately. Of course, there is a measure of indeterminacy about the total cost of everyone [169] paying privately for such collective goods. Sometimes only by having actual market alternatives can we agree on a basis of comparison. Often, however, a taxpayer will be broadly content if he perceives himself as receiving some valued services which seem reasonably commensurate with the outlay demanded.

It follows that a rather large margin of efficiency of public over private provision is required before we can justify the compulsory provision of a public service at the expense of the taxpayer. Each taxpayer is going to see a number of the public expenditures he has to help finance as doing no good to him. The remainder have got to carry the cost of the unwanted ones. The elderly couple derives little joy from free schooling but finds the public library a great boon, as, also, clean streets and the availability of the district nurse. They will not grudge half their rates going on education so long as the services they actually use seem a good value for the rates they have to pay, that is so long as those services could not obviously be provided by the market at a cheaper price. This often will be the case, but not always or necessarily: it provides a constraint both on the general political question of what services the public authority should provide and the particular fiscal question of how much the individual can be asked to contribute.

Those with large houses in plush neighbourhoods, or with large incomes, or large cars, can reasonably be asked to contribute more because typically they benefit more. Different people will avail themselves of different goods, but each will avail himself of some, and there will not be any general category of exploited taxpayers who have to pay a lot and do not receive a fair return. Although each would like to opt out of some items of public provision, which do not benefit him personally, he would suffer more by others opting out of the provisions which do benefit him. Granted reasonable efficiency of collective action, he gains on the swings what he loses on the roundabouts, and the package as a whole is one that it is in his enlightened self-interest to accept.

The difficulty with this sort of argument is the packaging. If the State is allowed to determine the package on offer, it can make any level of taxation appear to be fair: if the individual, the package can be unpacked until we have a set of separate subscriptions for communal goods, with individuals opting out whenever it is in their individual interest to do so. There is a certain analogy here with the problem of political obedience that Locke inherited from Hobbes. Hobbes offered a big package. The subject had to obey the sovereign regardless, because the sovereign gave him security. In the same way, if there is just a single tax package, and part of the benefit it confers is security, internal and external, then it can carry a pretty pricey tag. If, however, we can set up a more detailed dialogue between the individual and the community, we can make political obedience conditional rather than [170] absolute, and relate benefits to burdens in detail rather than have them all inextricably lumped together. Although one would pay almost anything to be secure, that benefit is one which everyone enjoys and everyone values, and therefore the burden of providing for defence and the maintenance of law and order should be laid on everyone. In time of war the rich may be called on to sacrifice most of their wealth to pay for defense: but if they are being called on to sacrifice most of their wealth, then an equally burdensome sacrifice should be called for from the poor. The fact that the benefit of security justifies a pricey tag cannot be used to justify loading the burden of taxation on the shoulders of one class of taxpayers rather than another, because we all benefit from security, and contributive justice requires that we all contribute.

The basic benefits of civil society are relatively easy to identify and price, and we can make a stab at apportioning the burden of providing for them. The more optional communal benefits are more difficult to package. We need to distinguish two questions: whether they should be in a package at all, and whether the cost of providing them is being fairly apportioned. The former question is predominantly a political one. Representation was seen as a means of securing that collective policies were in line with the Colonists' wishes. Of course, representation is neither a necessary nor a sufficient condition of that. Many governments have conferred benefits on their populations which were recognized as such although there had been no formal representation of the people's wishes, and in any form of representative government a lot of people get outvoted and saddled with policies they do not want. Nevertheless, representation does help. If there is a system of representation, although we cannot tell each individual taxpayer that he wants the benefits he is being asked to pay for, we can tell him that he is a member of a group which by and large wants them. So the question of whether the benefit should be provided is answered, moderately well. The only question left is whether the burden is being fairly shared. Here, it seems to me, we can make use of a presumption of equality. If the community calls on me to bear an equal share of the cost of carrying out a collective decision, even though it is one I may have opposed, I have no complaint. If the share is heavier than that asked of others, contributive justice requires that we consider all those also being asked to contribute the same amount, and ask whether they as a class are benefiting as much as they are having to pay. If they are, they are getting a good Lockean bargain, and are not being too badly done by. If super-tax payers as a whole are getting a lot of opera, foreign consular services, and national parks, which it would cost them more to provide for themselves by private subscription, then they are not being milked by the tax man. But if they are being asked to pay more without there being any justification in terms of benefits enjoyed or available for [171] enjoyment, then their complaint that they are being unfairly ``taxploited" stands.

The duties levied on alcohol, tobacco, and gasoline can be partly justified by the fair bargain argument. Consumers of these commodities are disproportionately heavy consumers of health services, which in most countries are at least partly funded out of general taxation. Drinkers and drivers also occupy a disproportionate amount of the time of the police and the judiciary. Large sums of money are spent on roads. The road lobby in Britain frequently makes out that all the money raised by road taxes should be spent on roads. The argument is fallacious, because it ignores the other costs imposed on the community by road users. But the fact that it is put forward at all, shows that there is felt to be something in the fair bargain argument.

Besides benefits directly linked to contributions, we can take into consideration, also, partial and contingent correlations in deciding what packages of benefits and burdens may be fairly imputed and apportioned. A propensity to consume dutiable commodities picks out a class of taxpayers who benefit from the provision of certain other minority interests: almost all those who benefit from country parks consume petrol, almost all opera goers drink wine. This, again, is only a partial justification, and depends on the fact that contributive justice is only rough justice, and that classes of contributors need only roughly correspond with classes of beneficiaries.

With regard to direct taxes, the fair bargain argument justifies proportionate taxation, as opposed to a simple poll tax. It might yield a partial justification of progressive taxation, although the main justification, in so far as it is justifiable, must rely on the third underlying principle of fiscal fairness, which is that the burdens should be equal not in monetary but in real terms: The widow's mites cost her much more than the rich man's largesse. And so we conclude that taxes should be not only proportionate but progressive. But there is a limit to this principle. If we take £1M from a man with £1.25M it is not clear that this is costing him less than the 30% levied on the ordinary taxpayer. As a first approximation we can vary the good bargain argument to become the hard bargain argument. The hard bargain argument considers the provision of collective goods where there is no possibility of their being supplied by the market, and then considers how hard a bargain would have to be before someone felt he could not afford to accept it. The classical example is defense, but that has the disadvantage of being so strong as to be unselective---although the rich man will certainly give his all to preserve his life, so will the poor man; so that although we could tax the very rich very heavily for Trident, we could equally tax everybody else, and it would be unfair to demand enormous sacrifices of the one and not the other. Other public goods are less extreme. In times past, [172] public benefactors who were rich would produce public goods at their own expense, and we can see this as giving an approximate indication of what they could afford. We can intelligibly ask, ``How much would those with £10,000 per year be willing to pay for a public good if they alone were having to foot the bill.'' ``How much would those with £15,000 per year?'' "How much would those with £20,000?" "How much would those with £30,000, etc.?" This would give a measure, although only a rough one, of equality of sacrifice. Of course, it would not be the same for each individual, or for each sort of public good. As always in political economy, we cannot be asking everyone his opinion, or have everyone making actual choices. We have to impute values, and say on behalf of people what decisions it would be reasonable for them to take in hypothetical circumstances. I do not pretend that there would be uniform agreement on the answers. But there would be an advantage in adopting an approach which was informed by a concern for fiscal fairness, and which asked rather more precise questions than those usually canvassed when taxation is under discussion.

It would help when we were discussing tax brackets if we considered hypothetical communities in which everyone had the same income, £10,000, £20,000, £30,000, etc., and asked how much each member of such a community would be willing to pay, on condition that every other member of that community did the same, in order to secure a particular public good. This would tell us for each public good how affordable it was at different levels of wealth, from which we could tell what level of taxation for the purpose of securing that good, imposed an equal burden of perceived sacrifice for people of different wealth. There would be a further, but surmountable, difficulty in grossing up the different scales for different public goods into one overall scale. Anglo-American constitutional practice is averse to assigning particular imposts to particular purposes, but there is nothing wrong in principle in doing this as a theoretical exercise, so that if, for example, poor Americans set great store on the United States sending a space probe to Pluto, and a low store on third world aid, and American millionaires did not care much about NASA but were keen on famine relief, then the NASA budget should be funded by a few cents a dollar on everyone's income tax, and foreign aid on the marginal rates paid by millionaires. Such a theoretical exercise is possible, and would, I think, be helpful in securing justice between different classes of taxpayers.

Contributive justice is always rough justice. It cannot take into account all the peculiarities of the individual case, but must categorize members of the community in broad general terms. So long as the contributions are not too onerous, it does not matter---I do not quibble over my National Trust subscription. It is only when we have very high taxation that we have to deal with all sorts of special cases, and make special and complex provision for [173] them. But complexity destroys intelligibility. I don't know what other people are having to pay, have no idea if it is fair, but have a shrewd suspicion that a clever accountant could avoid it. Hence, conversely, if we are to have a system that can be seen to be fair and will command general support, it must be simple, broadly-based, and not too onerous. What exactly that means in practical terms, I leave for future discussion. For myself, I have often been attracted by F. A. Hayek's suggestion that the top rate of marginal tax should not be more than the ration of public expenditure to Gross Domestic Product. But that I leave to discussion. The one point I want to urge is that we should formulate afresh the principles of our taxation policy, and debate them in terms of justice. For too long the taxpayer has been talked of as though he were a milch cow, without any regard to his own rights and legitimate interests. We have had a high taxation rhetoric, with ever increasing complexity and avoidance. But in reality the yields have been much less than the rhetoric suggests. A new rhetoric, of justice rather than dairy management, would release many energies at present devoted to fiddling for more productive purposes; and might actually increase the yield.

1. My thanks are due to Loren E. Lomasky, R. M. Hare, and John Gray for comments and criticisms which I have incorporated into this final version, which was delivered at a conference in Florida shortly after the State Legislature had enacted legislation taxing non-residents.
2. J. R. Lucas, The Principles of Politics (Oxford 1966) (pbk. ed. 1985), Section 67, pp. 287-295.