The case against tax reform

FOREWORD

"Tax Reform" is every populist politician's catch-cry.  It has an initially attractive sound to it because every taxpayer is encouraged to think of it as "tax cuts for me".  Richard Wood shows that tax reform is not all it is cracked up to be.

It is readily apparent that tax reform without change in the total tax liability simply means dealing out the tax liability in some different way:  some people gain, but others inevitably lose.

Change, however, is not simply a zero-sum equation.  The reformers point out that taxes which distort markets least -- neutral taxes -- do least harm to the economy and that efficiency can be gained by changing from discriminatory taxes to neutral taxes.  So far as their argument goes this can hardly be doubted, but Mr Wood explains here that it is not the whole story.  Change itself is harmful:  it disrupts established patterns of economic activity and adjustments that have previously taken place.  As people have adjusted to existing rules, unanticipated change is inequitable.  Like continually changing rules of the road, continually changing tax rules are chaos.

This is not, of course, to say that the total tax burden should not be reduced;  one of the main thrusts of Mr Wood's work is to show how the reduction of government intervention and spending, and thus an eventual reduction of the tax burden, is possible without increasing the deficit.  Nor is it to say that relative tax liabilities should never be changed.  But Mr Wood mounts a persuasive case against change for change's sake or for only modest efficiency gains.

It is important to be wary of any suggestion that tax "reform" will give the community something for nothing.  The word "reform" connotes improvement and a fresh start;  it should not be allowed to cover political redistributions of favours, pork-barreling and general tinkering whose real costs the ordinary taxpayer can never know.

John Hyde

1. INTRODUCTION

So much has been written in recent times about the Australian tax system and possible changes in it that it is difficult to see how anything useful can be added, or anything said that has not been said twenty times before.  And I should at the outset make it clear that I claim little of either utility or originality here.

Nevertheless, in all the recent brouhaha over tax reform, there does seem to be one argument that has not been much aired.  This is the argument against tax reform.  I do not mean the argument against Option A or Option C or indeed any other particular option that is currently fashionable.  Nor do I mean the argument in favour of the particular tax system that prevailed until the recent changes and that everyone seemed bent on reforming.  I mean, rather, the arguments against tax reform in principle.

Any such argument must, by its nature, be radically conservative.  It is an argument for leaving the status quo alone, more or less independently of what that status quo happens to be.  And because this is an a priori conservative argument (something that even conservatives sometimes argue doesn't exist), I should emphasise two things.  First, the argument does not necessarily generalise.  That is, it does not apply -- or at least not with anything like the same force -- to other areas of government policy.  Taxation is "special" in a way I shall seek to spell out:  the tax system is not like just any other area of policy concern -- though much of the "expert" analysis treats it as though it were.  Indeed, my chief complaint about the current tax reform debate is that there seems to be widespread misunderstanding as to what exactly it is that is being reformed.  When that misunderstanding is removed, there may well be somewhat more enthusiasm for leaving well enough alone.  But, for the record, I should state clearly that I am not at all conservative on other matters of public policy:  there is much I would change tomorrow if I had the chance.  It is just that this "much" would not include the tax system.

Second, it should not be thought from my argument that I believe all tax systems are equally good.  I don't.  In my view, some are very much better than others.  But here I intend to suppress my views on that issue in order to focus on more general questions -- such as the question of what function the tax system plays in the political and institutional fabric;  how often the system ought to be changed;  and under what institutional arrangements such changes ought to be determined.

On the face of things, these may seem like rather dull questions compared with the substantive tax issues -- who should pay more and who less, and so on.  But in tax matters, dullness is probably a virtue.  If, in the excitement of the tax reform process, the more general questions are overlooked, as I rather think they have been in the recent past, the results can be disastrous.

2. WHAT IS TAX REFORM, REALLY?

Let me begin with a simple point.  Everything that government provides has to be paid for, somehow, by someone.  If we define those payments to be "taxes" broadly conceived, then total government benefits must equal total tax receipts:  nothing costs nothing.  Precisely who pays, and in what capacity, is determined by the structure of the taxes that government uses.  But whatever else, if one person pays less, others must pay more.  And one of the main preoccupations of orthodox tax theory involves recommendations as to how to distribute the costs of government across the population.  Put another way, tax reform is, among other things, concerned with making some people pay more tax and others less.

Once this is recognised, we immediately confront a puzzle.  How can it be that everyone is in favour of tax reform?  The answer to this puzzle is that pretty well everyone believes that he is paying too much tax and that, once the system is appropriately reformed, he will end up paying a good deal less.  Indeed, this is how most of us define tax "reform".  But this clearly can't be the case for everyone:  not everyone can pay less.  And if it seems to everyone that they are paying less, then we know that some of them have been hoodwinked.  Tax "reform" is necessarily a matter of shuffling the tax cards in the pack:  the same total liability has to be dealt out one way or another.

This is not to say, of course, that the overall size of the public sector could not be reduced, with a consequent lowering of taxes for everyone.  But this is not what is normally meant by "tax reform".  "Tax reform" in the standard parlance is concerned with the manner in which a given amount of public revenue is raised, and not at all with the level of aggregate revenue.  Indeed, questions about aggregate revenue levels are usually explicitly ruled out of court -- though as we shall see, they have a happy knack of creeping in the back door anyway.

Nor do I wish to deny that tax reform may, in principle, make everyone better off.  One of the ways in which people "pay" under the tax system is by expending money and energy in trying to avoid taxes, or by organising their affairs in ways that they would not find desirable were it not for tax considerations.  In the general scramble to avoid taxes, individuals not only endure costs themselves;  they also impose costs on each other.  Taxes which direct economic activity to one activity at the expense of others impose considerable cost on the community.  So do tax measures such as tariffs, which inhibit competition.

Tax reform can, in principle, reduce such costs and in the process make the average taxpayer better off.  But such gains do not necessarily show up in reductions in total tax liabilities:  if the level of public spending is maintained, so will be the average taxpayer's tax bill.  In short, we cannot all reasonably expect that our tax bills will be reduced under tax reform;  what we might reasonably hope for is that we expend a little less in adjusting to the tax system and that by more efficient use of resources the tax base itself may be expanded.

But this raises a further problem for tax reformers.  If all taxpayers are prepared to sustain costs in trying to minimise their tax liabilities in their private capacities, why should we suppose that they would do so any less in their political capacities?  It is hardly to be wondered at if people attempt to use whatever means are available to reduce their taxes.  And one such means -- not necessarily the least important -- is political action, designed to alter the tax law in their own favour.  In any so-called "tax reform" exercise, such political activity is gathered up and focused on a particular exercise.  What emerges from that exercise is necessarily a matter of compromise among the various factions and interest groups, all intent on minimising their own liabilities.  To be sure, the requirements of political debate may constrain the arguments in such a way that they have to be couched in terms of the language of "fairness" and "efficiency".  But a large dose of special interest can be expected from all sides.  And it is notable that most factions define fairness in terms of less tax for themselves.  Given the necessary vagueness of the term, it may even be that individuals genuinely believe that fairness is to be so defined.

There is an important reason why this should be so.  One of the relevant "interest groups" in any tax reform exercise is the government itself, and it is surely not too far-fetched to claim that government's "interests" involve having more revenue, and more fiscal power, rather than less.  Taxpayers seem to regard fairness in taxation as a matter as much of relations between taxpayer and the State as of relations among taxpayers.  Interestingly, orthodox tax analysis defines fairness solely in the latter terms.  Tax "equity" is seen to be a matter of distributing a fixed total burden of taxation across the citizenry in the fairest way:  the question of whether that total burden represents a "fair" claim of the State on individual citizens is explicitly suppressed by the standard "equi-revenue" methodology.  In fact, it is not merely that the conventional methods of tax reform analysis do not allow consideration of the "fairness" of aggregate claims by the State.  It is also a clear implication of Realpolitik that the State's interest in any tax reform exercise is to increase those claims -- if not directly in the context of the tax reform itself, then at least indirectly by making the system capable of generating increased public revenues in the future.  Nor is this objective always implicit.  In the 1974 Treasury Tax Paper No 2, written for the Asprey inquiry, the need to allow for increased revenue was a major element in justifying tax "reform":

[A] major question today is whether government can look to personal income tax for large and continuing increases in revenue.  If the view is taken that it cannot, the need arises to supplement tax revenues from the existing system.  In doing that, means of improving the system may also be found ...

What this statement suggests is that "improving" the tax system is an entirely subsidiary goal:  getting one's hands on more revenue is the chief object.  In the more recent Hawke-Keating exercise, revenue increases were explicitly eschewed, though whether the zero revenue growth promise will be adhered to remains to be seen.

Briefly put, I am saying that tax reform is a political exercise, that it involves substantial conflict between the interests of different groups of individuals, and that one significant "interest" in this conflict is that of government itself in increasing tax revenue.  These simple facts imply that what emerges from any so-called tax "reform" exercise will tend to reflect as much the prevailing distribution of political power as the intentions (benign or otherwise) of the tax-reformer, be he or she politician, bureaucrat or public finance academic.  If the structure of effective political power changes very slowly, then of course this will tend to be a force against change in the tax system.  That is, if particular interest groups have secured special tax treatment for themselves within the prevailing tax system, then it may be politically difficult to alter the state of affairs, however "inefficient" or "inequitable" that special tax treatment may appear to be.  It simply may not be politically feasible to achieve the sorts of "reforms" that are deemed to be ideal.

Quite apart from the political constraints on tax reform, there are significant economic ones.  Suppose one could isolate some group which one felt was currently getting away with paying too little tax, and suppose (more heroically) that group lacked sufficient political clout to defend its privileged position.  Even so, it may well turn out that it is more difficult to make taxes "stick" on this group than might appear.  After all, if the tax system is such that some group (small businesses, taxi-drivers, tradesmen or whatever) systematically pay less tax than others whose real incomes/expenditures are identical, then this may be because the members find it relatively easy to organise their income or expenditure into non-taxed forms.  The actual pattern of tax payments necessarily reflects the relative difficulty of taxing different persons, and it may not be possible to change such relative difficulties merely with a stroke of the legislative pen.

For example, if taxi-drivers under-report their income, or tradesmen do much work on a barter basis, it is by no means clear how tax changes can increase the tax payments of such groups.  It is of course sometimes suggested that a change in the tax mix would solve such problems, because it would involve less reliance on the income tax and more on indirect, consumption taxation.  But it is not obvious why taxi-drivers who avoid income tax could not also avoid paying retail sales tax on the services they sell, and thereby increase the demand for their product and their own incomes.  And tradesmen who work on a barter system seem to have the capacity to avoid almost any tax one might devise.  The actual pattern of tax payments reflects, in short, certain economic realities that may be impervious to changes in the tax system.

This is not to say that changing the tax system in various ways could not, and would not, alter the distribution of tax payments -- and undoubtedly some of those distributions are likely to be widely perceived as "fairer" than others (though as I shall argue in section 4 below, appearances may be extremely deceptive in such matters).  My point here is rather that the changes are unlikely to be anything but small for most taxpayers.  Shuffling the tax cards in the deck will not make much difference to the majority.  Some small groups may do quite well, others quite badly;  but for the most part it will be life more or less as usual.  The net gains, if any, in terms of "fairness" will almost certainly be small.  Moreover, any changes will certainly involve less fair treatment for some taxpayers, as well as fairer treatment for others:  no tax arrangements will be perfect, and even the best will involve a weighing of fairness gains and losses.

Let me summarise the argument in this section in terms of two propositions:

• Tax "reform" is necessarily a political exercise -- one that involves substantial conflict of interests among various groups.  It simply cannot be presumed that, in general, what emerges from this exercise is likely to be "fairer" or "more efficient" than what previously prevailed.  That will depend on the nature of the prevailing political forces, and in particular on whether political power is more evenly distributed than when previous tax decisions were made.  In this sense, the term "tax change" might be both more neutral and more realistic than the immodest term "tax reform" -- though of course political rhetoric will invariably tend towards the more heroic language.
• What is feasible in terms of genuine tax reform may be more modest than is often claimed.  To the extent that perceived unfairness or inefficiency result from the fact that some income and some expenditure are difficult to tax, the unfairness and inefficiency of the tax system will tend to remain.  In any change, some taxpayers will be treated less fairly, while others are treated more fairly.  The net gains in fairness, if any, that are achieved will almost necessarily be modest.

3. WHAT IS SPECIAL ABOUT TAXES?

I have just argued that tax "reform" will at best involve marginal net improvements in "fairness" and "efficiency".  But marginal improvements are, after all, still improvements, and their smallness is no reason not to grab them if we can.  Moreover, everything said about tax reform in section 2 is true of virtually every other area of public policy:  all emerge politically;  all involve major elements of conflict;  in most cases, the net benefits are only small.  Does the argument really amount to a case for no change in anything?

My answer is no.  Taxes, I claim, are different.  The difference at stake is akin to the difference between the rules of a game and particular plays of the game.  No one has any difficulty recognising this distinction in a simple parlour game:  the rules of bridge are easily distinguishable from the description of the play of a particular hand.  In policy contexts, the distinction can be more difficult to discern, because both the rules of the policy formation game (namely politics) and the policies themselves emerge from similar processes, often involving the same set of persons.  All the same, there is a clear distinction between a decision to use majority rule as a way of making a policy decision, and a particular policy decision that emerges from the majoritarian process.  I shall call the former level "constitutional", and the latter "in-period" -- without seeking to imply that the "constitution" must be in writing, or should be in any way restricted to narrowly political rules.  I use the terminology "constitutional" simply to suggest the connection to "rules of the game".

My claim about taxes can now be expressed as follows:  the tax system is, or should be recognised as being, quasi-constitutional.  The tax system is, in other words, part of the institutional context within which political decisions are made, somewhat like majority rule.  This is so because the tax system determines the "prices" that various individuals pay for additional public goods and services, as well as the distribution of the costs of a fixed budget.  If the tax system is not established and in place prior to the point at which individuals have to express their views (electorally or otherwise) on the level of public spending, then individual citizens cannot make relevant political decisions responsibly.  To change the tax system after relevant electoral decisions have been made is equivalent to asking individuals to decide on which house they want without knowing what the prices of various houses are:  no market can work satisfactorily on this basis.  To the extent that politics is like a complex multi-person market, in which various individual voters choose among alternative political parties with their associated policy packages, the proper operation of democratic politics presupposes that individuals know what the costs of a project (or package of projects) to them will be.  If the tax system can be changed in a major way at any time, individual voters simply have to guess what price each will have to pay for what is voted upon:  responsible electoral decision-making is substantially undermined.

Imagine the implications for road traffic if the rules of the road were to be subject to violent and unpredictable change:  if the "give way to the right" rule were to be replaced out of the blue with "give way to the left", or if cars were required to drive on the right rather than the left, or if red meant "go" and green meant "stop" on alternate days.  Over a substantial range, it may not matter too much what the rules of the road are (whether motorists drive on the left or the right, for example) providing those rules are widely known and observed:  the rules will lose their value if they are subject to constant change.  Some of those rules may be inefficient or arguably unfair in some instances:  it is, for example, clearly more "efficient" in the economist's sense to have a driver go through a red light if no one else is in sight.  Yet we do not want a rule stating that it is all right to proceed if you think no other car will crash into you:  this is virtually equivalent to having no rule at all!  Equally, the "drive on the left" rule may be deemed unfair to, say, immigrants from continental Europe or South-East Asia, whose driving skills have to be modified to meet the different requirements.  Yet to seek to accommodate all "special cases" under the traffic law is to commit oneself to the prospect of daunting complexity, and perhaps to an impossible quest.

This makes a general point about the nature of rules, and the desirability of stability in them.  What is true of rules of the road is also true of the rules of political process.  We would certainly wish to leave room for voters to change their minds about desirable policies or desirable political parties.  The rules of politics are precisely designed to make such changes possible -- just as the rules of the road leave open the journeys that individuals will choose to make.  Rules are not designed to restrict choice, but rather to make relevant choices possible.  To secure this end, the rules themselves must be stable.  Once this "quasi-constitutional" nature of the tax system is recognised, therefore, we must also recognise the consequent presumption against changing the tax system, even in a direction that is, in itself, "desirable", too often or too radically.  The onus must always be on those who seek to change the rules of the game.

Several questions naturally arise about this argument.  First, are taxes really unique in this respect?  After all, changes in certain regulations (occupational licensing, for example, or tariff levels) can involve drastic changes in individuals' financial positions -- changes that are in most instances more drastic than any that could be expected under most tax reforms.  Is not the case for policy conservatism even stronger in such cases than in the case of the tax system?  This question somewhat misunderstands the argument, which is that individuals cannot make rational choices between the policies of different governments if they do not know what those policies will cost.  Uncertainty about the citizen's disposable income is not the problem under consideration here, although it should be said that violent fluctuations in the fortunes of different citizens with each spin of the electoral roulette-wheel are not to be encouraged either, and can also ultimately subvert the constitutional order.

Second, it might well be asked why, if the economic constraints on changing the distribution of tax payments are significant and the potential fairness gains likely to be small, whether one should worry too much about tax reform?  If everything turns out to be more or less the same, if tax reform is seen to involve changes that are largely cosmetic, then how can tax reform do much harm?  The answer to this question is twofold.  For one thing, although there may be groups which are strategically placed to avoid taxes, and which it is virtually impossible to tax effectively under any tax regime, tax "reform" can certainly affect relative tax burdens among the rest of society.  For another, if tax changes arise from changes in the distribution of political power, then quite major changes in tax burdens can occur as the new majority coalition attempts to force the minority to bear a disproportionate share of the cost of public programmes.  Even when tax changes do not much affect the overall tax base, there may be substantial changes in the rate structure which affect the relative tax burdens of different taxpayers, and hence the effective prices that citizen-taxpayers have to pay for public services.  It is precisely because tax changes do affect the tax payments people pay, and may do so quite significantly, that changes in basic tax institutions should be made sparingly and with a proper reluctance, if at all.

4. IS TAX REFORM "FAIR"?

In conventional public finance circles, fairness is normally conceived as requiring both that the tax system should serve to redistribute income (or wealth or economic power otherwise defined) and that it should impose identical burdens (or benefits) on those who have identical income in the absence of taxation.  I shall here accept this conception without further debate.  Clearly, to make "fairness" so defined an operational criterion for evaluating alternative tax systems, we must be able to assess the distributional consequences of the various options.  Further, because no available option can, in the nature of things, be totally "fair" in the sense defined, we must have some overall measure of fairness so as to be able to tell which of two tax systems is the fairer.  Let me say a word about this last matter.

One of the problems involved in any measure of fairness is that of comparing smaller numbers of larger "inequities" with larger numbers of smaller ones.  One tax arrangement may involve a few very large inequities (defined, say, as differences in burdens endured by individuals with identical total incomes) while another tax arrangement involves a large number of quite small inequities.  Which is to be preferred on grounds of fairness?  The answer to this question may not be immediately obvious, but it does seem clear on reflection that large inequities ought to be weighted proportionately more heavily than smaller ones.  It seems difficult to object to a tax system in which the burdens imposed on otherwise equally-placed individuals differ by only a few dollars -- one can indeed hardly expect anything else.  But a tax regime which involves some individuals in sustaining virtually no burden at all, while most others sustain a large one, seems to affront the most basic intuitions of fairness.  On this basis, we might, for example, use as a measure of fairness the variance of tax burdens about the average across the set of appropriately narrow income ranges.  The characteristic feature of such a measure is that a doubling of the difference in burdens more than doubles the degree of unfairness measured.

A brief example may help here.  Suppose under one tax regime, individuals A and B, who are judged to have identical "ability to pay", pay taxes that differ by $500.  Suppose under a second tax regime, the difference in taxes paid is $1000.  Then the second tax regime would be more than twice as "unfair" as the first.  Because the variance of the distribution of tax payments involves squaring the tax differences, doubling the difference in tax payments quadruples the variance.  As I have suggested, there seem to be good reasons for believing that some such measure of "unfairness" will accord tolerably well with our ethical intuitions.

There is an important implication of such a measure that merits emphasis.  This is that in cases where the true degree of inequity is unknown, and where a particular tax change is just as likely to move in the wrong direction as the right, to make the change will on average reduce the degree of fairness.  In such cases, although there is a one in two chance of increasing fairness, the gain in fairness if one moves in the right direction is less than the loss in fairness if one moves in the wrong direction.  Similarly, to make tax changes in a setting where the chances of moving in the wrong direction are at all large is likely to reduce fairness.  Again a simple example.  A tax change may equalise A's and B's tax payments;  but it is just as likely to increase the difference from $500 to $1000.  The possible reduction in fairness at stake is then four times as great as the possible increase.  Such a policy gamble is clearly a bad bet.  A gambler who takes on bets where there is a fifty-fifty chance of winning one dollar and losing four can only expect to lose money over the long haul.

Now, I have claimed that this is an important implication.  This is because I reckon that the possibility of moving in the "wrong" direction in making tax changes is always significant.  It is so simply because of the profound uncertainty that surrounds the final incidence of various tax options.  The truth of the matter is that the question of the distributional consequences of alternative tax systems is extremely complex.  And any self-styled expert who claims to have the answer to this question should be treated with great scepticism.

What I seek to do at this point is to indicate briefly some of the dimensions of the complexity at stake here.  To do so, I need to emphasise one aspect of the terminology I used earlier in defining "fairness".  As public finance experts have long recognised (at least in principle), there is no simple one-to-one relationship between the taxes an individual pays over to the government and the burden that individual sustains by virtue of the tax system.  This is a fact readily enough recognised by the ordinary citizen in connection with excise taxes, for example.  The individual -- say the petrol station manager, or the brewer -- who faces the legal liability to pay revenue over to the government does not himself necessarily sustain the full tax burden, and may not sustain any burden at all:  that individual may in effect simply be a collecting agent for the state.  But the point is not restricted to "indirect" taxes at all.  The tax system will, in general, alter the prices of the goods and services that individuals buy and sell (including specifically the wages in different industries);  and different tax systems will change those prices in different ways.  It is the effect of all those price changes on individual incomes that determines the distributional impact of the tax system in each case.  Further, the price change involves incorporating all the responses of all the individuals to the tax rates, and such responses are virtually impossible to predict and extremely difficult to detect even after the event.

To indicate the nature of the problem, let me take two examples.  First, consider the argument, widely circulated within the most august public finance circles during the debate on Option C, that a move towards indirect taxes would moderate the effects of evasion and avoidance.  The reasoning -- apparently plausible -- is that though an individual might be able to evade or avoid income tax, he would when he spent that enlarged income still pay some tax due to the increased indirect taxes in place.  This reasoning depends, however, on an assumption that indirect taxes (excises, sales taxes etc.) are passed forward to consumers while income taxes are borne by the nominal taxpayer.  There is no general presumption in economic logic that this will be so.  Both the income tax and the sales tax drive a tax wedge between the price paid by the buyer and the price received by the seller of a good or service:  it would be somewhat strange if the pattern of shifting were utterly different in each case.  Whether or not a case can be made for a change in the tax mix on the grounds of moderating avoidance and evasion depends largely on whether avoidance and evasion under the income tax is in any way associated with particular industries.  If, for example, avoidance and evasion are more prevalent in industries such as taxi-driving or plumbing where the income derived is difficult for the fisc to monitor, then net-of-tax incomes will be relatively higher in those industries.  Workers will tend to move into those industries and out of others where tax evasion and avoidance are more difficult until net-of-tax wages are equalised.  Such movements do not occur instantaneously.  Moving jobs (and possibly locations) is a costly business.

Over time, however, individuals choosing jobs will tend to be attracted to the employment with the higher net-of-tax returns.  In consequence, when taxes have been in place for a long period, gross-of-tax wages will be higher in the more heavily taxed employment.  The shift of resources to lightly taxed areas will also have effects on the prices of the output of the two industries.  In effect, the income-tax evasion possibilities tend to be shifted forward so that consumers of those products also benefit from the effectively lower taxes.

In the same manner, it might be argued that progression in the tax system has highly arbitrary distributional effects, because the higher income-tax rates imposed on higher income earners are largely shifted forward in terms of the higher prices of the goods and services that those higher income earners provide.  If for example all lawyers have high incomes, and all garbage collectors have low incomes, the imposition of a progressive income tax will tend to increase the price of legal services relative to the price of garbage collection services.  Accordingly, the burden of progression in the rate structure will in part be borne according to the consumption pattern of individuals, and no simple conclusions about the egalitarian impact of these consumer-price changes can be drawn.  I do not wish to argue here that income-tax rate differences are fully shifted forward.  Nor do I claim that there is no case for changing the tax mix on fairness grounds.  My point is rather that the effect of any tax change on the distribution of income is extremely complex, that many of the results that are now part of the standard literature are quite counter-intuitive, and that there must of necessity be considerable uncertainty as to whether any particular change will reduce or increase existing inequities.

Let me finally raise another query.  I have already mentioned the tendency for tax differences to induce resource movements until net-of-tax returns are equalised.  Consider in that connection the issue of fringe benefits.  Suppose that in various industries -- or indeed throughout the corporate sector -- there is a practice of paying employees a certain fraction of their incomes in the form of tax-exempt "fringe benefits".  An individual can, let us suppose, either enter the public service and earn $20,000, or become a junior executive in a private firm and earn $20,000 of which $10,000 is paid in the form of fringe benefits.  In the absence of tax, she will, we suppose, be indifferent between the two jobs.  But introduce an income tax (25%, say) with exemption of fringe benefits, and she will choose the private firm.  As a result, the public service will find it harder to hire personnel, and the corporate sector easier -- and there will be an inducement to move from the one to the other until the net-of-tax incomes are virtually the same in both areas.  In response to the movement of workers, the private firm can cut its salaries to junior executives to, say, $18,000 ($10,000 fringe benefits, $8,000 cash income) and the public service raise its salaries to $21,333.  Net-of-cash income in both cases will then be $16,000.  Suppose now, however, with our junior executive well established in her career, that the government moves to close what it has been advised is an obnoxious loophole.  Given that it takes time for salary levels to re-adjust, those who have joined the private sector in the light of the tax advantages will be unfairly penalised.  Our recently commenced junior executive will now receive net-of-tax income of $13,500 as opposed to her classmate who entered the public service.  Prior to the tax "reform", the two were being treated identically under the tax system:  both received a net-of-tax income of $16,000.  After the tax "reform" the one who entered the private firm will be effectively discriminated against by the tax change.  In this case, the tax reform does not so much remove inequities, as create them.  Individuals will plan their futures with an eye to the prevailing tax system, and once full equilibrium has been reached, the horizontal inequities will have been obliterated.  This is not to say that inequities of a similar kind were not created when the tax was originally introduced -- fringe benefit provisions and all -- but those inequities cannot in general be reversed by altering the tax system some decades later.  Once it is recognised that net-of-tax returns tend to be equalised when a particular tax regime is in place, there seems to be a strong prima facie case for leaving that tax regime in place, unaltered.

This argument is not a knock-down one.  It presumes that individuals are basically very similar, and that none of them take any account of the possibility of tax change.  But there are very few knock-down arguments in economics, and this one stands as a counterpoise to those reformist enthusiasts who naïvely assume that any difference in nominal tax liabilities or actual tax payments across industries or among individuals in differing employments is decisive evidence of unfairness.

5. IS TAX REFORM EFFICIENT?

Public finance orthodoxy advocates that the tax system should be "neutral" -- that is, that it should interfere as little as possible with individuals' decision-making in private markets.  If the tax system induces individuals to choose a different course of action from what they would have chosen in the absence of taxation, it is effectively inducing individuals to substitute something they value less highly for something they value more highly:  the consequent loss in value is referred to as the "excess burden" of taxation.  The object of tax reform is, in large measure, to reduce such excess burdens -- something it seeks to do by making the tax system more "neutral".  This objective generates a strong presumptive case in favour of broad-based rather than narrow-based tax regimes, because the size of any excess burden rises at a faster rate than the tax differential.  That is, if one good is taxed at rate T and another is tax-exempt, the "excess burden" generated will be more than twice as great as the excess burden generated by a tax of half that rate, T/2, on the taxed good.  Broad-based taxes may involve some non-neutralities, but they are in general smaller than those generated by narrow-based taxes.

The general presumption in favour of broad-based taxes and of neutrality as an ideal is, however, based on a highly questionable assumption.  This is that the relevant comparison between alternative tax systems is one that holds the level of aggregate revenue constant.  If, instead, the choice of tax arrangements has implications for the total amount of revenue raised, then the effects of any such change in public spending have to be included in evaluating whether the change is "efficient" in the broad sense.  Suppose, specifically, that the broadening of the tax base makes it easier for the government to raise additional revenue, and that government will in the long run adjust to this greater ease by increasing its total claim on resources.  Suppose one believed, however, that the resultant increase in government spending would involve substantial waste, or would yield to taxpayers lower benefits than they would have received if they had spent the proceeds privately.  Then even if there were gains in terms of more neutral choices among private goods to be obtained from broadening the tax base, these would have to be set against the efficiency losses attributable to the expansion of public spending.  Now, of course, the question of whether expanding public expenditure would benefit taxpayers on balance or would make them worse off is not an uncontroversial one.  But if it is true, as I believe, that broadening the tax base will tend in the long run to greater public spending, there is no way that this question can be ignored.  Orthodox public finance tends to suppress it entirely.

The result is that orthodoxy has an in-built bias towards public sector expansion:  by recommending tax-base-broadening on other grounds, it secures public sector growth by stealth.  Public sector growth is not necessarily something that can be regarded as promoting "efficiency" overall.

A second assumption involved in the argument for more neutral tax arrangements is that the pre-tax arrangements are perfectly efficient.  As far as public finance orthodoxy goes, this assumption creates logical problems, for if it were valid there would be no economic case at all for government spending and hence no need for any taxation (or, at least, only taxation for financing redistributive activities of government).  In fact, public economics is built on the proposition that there is extensive "market failure":  government intervention is taken to be required to provide those goods and services (like defence, law and order, pollution control, and so on) which it would not pay anyone to provide, though the total benefits to citizens outweigh the costs.  However, public economists have long recognised that goods and services do not divide themselves neatly into two mutually exclusive categories, one of which the market handles perfectly efficiently and the other of which the market "fails" utterly.  Rather, goods and services lie in a spectrum running between these two extremes.  As we have seen in recent times, creative economists can make a plausible case for some "market failure" with respect to almost any good at all:  government departments are full of them!

In any event, even a radical anarcho-capitalist like Robert Nozick or David Friedman (Milton and Rose's more extreme son) would have to concede that, in any real world tax reform, there are in fact government activities of all kinds which exercise a profound effect on relative prices in the market sector and which any genuinely "efficient" tax system should seek to offset.  I have in mind here things like tariff policy, the whole panoply of regulations (including zoning laws, labour market regulations and the like) and the set of government provisions relating to the support of monopoly organisations such as labour unions.  Moreover, the government may well be providing some public services in grossly excessive amounts, and grossly under-providing others.  The net effect of all these policies on the prevailing set of prices for private goods is not something that can be taken for granted.  Even if tax policy were somehow insulated from the political pressures that generate such government-induced distortions, so that the makers of tax policy could pursue "efficiency" as an end in itself -- and assuming they were disposed to do so -- there would necessarily be profound uncertainty as to what would be required.

Consider, for example, the vexed question of capital gains taxation.  In the current Australian context, there is a substantial range of effective tax rates applied to assets of various kinds:  these rates vary from zero on owner-occupied houses (that is, the annual stream of services drawn from a highly durable asset, the house, in the form of accommodation is entirely tax-exempt);  through low rates on superannuation savings;  to excessive rates on the return from fixed money value assets (such as debentures and savings accounts) because of the failure to allow the real capital loss due to inflation as a deduction under the income tax;  and to "double" rates of taxation on dividends, due to application of both corporate and personal income tax.  The public finance profession is, moreover, divided on the question of whether it is desirable on efficiency grounds to tax property income (or, more or less equivalently, savings) at all, with the balance of the theoretical argument pretty clearly in favour of not doing so.  Within this total confusion, how can one be sure that the current package of reform proposals in respect of capital gains and corporate dividends actually involves a step in the right direction?  The truth is -- one can't.  What efficiency requires, in practice, is extremely obscure.

Of course, this argument can, like most, be pressed too far.  It is reasonable to expect that where there are large differences in tax rates between easily substitutable activities (e.g. rabbit-net and ring-lock fencing), efficiency will be enhanced by reducing the differences, particularly if it can be done without opening other differences (e.g. by eliminating the tax deduction offered to vermin proof fencing).  There are cases where the tax reformer need not be fazed by his onus of proof, though they are fewer than enthusiastic reformers often make out.

As in the fairness case, there is a general argument in favour of conservatism here.  If, in trying to correct a pre-existing distortion, one moves in the "wrong" direction rather than the "right" -- if, that is, one increases the size of the distortion rather than reduces it -- one stands to do more harm than good.  Suppose, for example, that rates of return net-of-tax on capital in Australia are inefficiently low rather than high, and that the net effect of the new package is further to reduce that net-of-tax rate of return.  Then the losses would necessarily be greater than the benefits that would accrue if the net-of-tax rate of return on capital were actually too high.  In short, changing the tax regime is a poor gamble:  only if one has good reasons for thinking that the tax change will be in the right direction should one proceed.  The onus of proof must, it seems to me, lie on those who seek "reform".  This is not to say that those who seek reform will not sometimes be able to discharge that onus, but it does make it clear that one requires a very strong presumption in favour of the efficiency advantages of the change one is making -- a presumption that is more likely to be met in the case of rather modest, highly specific changes.

To all this, one should add the costs of the change in itself.  These are of two kinds.  First, there are the compliance costs associated with the transition to a new regime.  These costs arise simply because taxpayers have to adjust their behaviour so as to be able to meet the requirements of a new tax arrangement.  Changes can temporarily increase compliance costs even when they will reduce them in the long term.  The fringe benefit tax, once people are accustomed to its operation and have developed habits of maintaining diaries of relevant details, may not perhaps impose the horrendous compliance costs that its opponents claim for it.  But even if this were so, the costs in the meantime, while people are acquiring the habits in question, are formidable indeed.  In the same way, the compliance costs under a value-added tax may not be, in the long run, significantly greater than compliance costs under a retail or wholesale sales tax;  but any change in the administration of the tax system does impose substantial costs in the short run.  The experience with the selective employment tax in the U.K. (or Kaldor's other brainchild, the expenditure tax, in India and Sri Lanka) or the Irish wealth tax are all instructive here:  a tax appears for a year or two and then is abolished.  The compliance costs involved in this sort of "on-again-off-again" caper are very considerable.

Beyond this, there are costs associated with uncertainty about the stability of tax arrangements in the future.  Many decisions are "capital" in nature -- that is, they involve actions the benefits from which accrue over a substantial future.  Such decisions are necessarily subject to considerable uncertainty about future states of the world.  To add extra uncertainty, attributable to changes in tax policy, is itself a cost.  To some extent, that cost will tend to show up in the level of investment:  decision-makers will be more reluctant to trust their fortunes to a future that is more rather than less uncertain.  It will also tend to show up in the composition of investment:  decision-makers will opt for shorter-term rather than longer-term investments (or equivalently for employments where options for exit are more readily available).  To the extent that they do this because of expectations that tax arrangements may change, and to the extent that such expectations are stimulated by more frequent tax "reform" action, the associated changes in the level and composition of investment are properly to be considered a cost of the tax reform as such.  It should be noted here that, if citizens expect tax reform every twenty years or so, the refusal to implement a tax reform exercise after twenty years are up may do relatively little to alter taxpayer expectation of a tax change in the near future.  The object of greater stability in tax arrangements is to reduce taxpayer expectations as to how often tax changes will occur.  Moreover, it may be that changes in the external environment interact with the tax system in such a way as radically to alter tax payments, and hence the effective tax system.  In such a case, tax law changes may be justified to maintain the effective tax regime.  Some would doubtless argue that this is what most tax reforms are really about:  protecting the tax system against the erosion of time.  But, if so, it would need to be shown that the tax reform is effective as maintenance, and does not introduce an entirely distinct source of taxpayer uncertainty.  And further, the language of "equity" and "efficiency" in tax reform circles would have to be replaced by the language of "conservation":  the aim of "reform" would be to establish again what the tax law originally secured, whatever that happened to be.

This form of "constructive conservatism" is, of course, not unknown.  It is precisely what is at stake in indexing the personal income-tax rate structure against inflation, for example.  The object there is to insulate the system from independent changes in the general price/wage level.  And one may argue (though it is not clear how persuasively) that tax "reform" more generally is periodically required to adjust for other independent changes in the economic order.

Of course, any such line of argument presupposes that the general intention of the original tax law was and remains tolerably clear -- something that at least some legal scholars might well dispute.  In any event, I should want to distinguish sharply between "constructive conservatism" as I have defined it here, and the sort of "conservative constructivism" that allegedly characterised the Barwick court on tax matters.  It is one thing to attempt to maintain a prevailing tax regime:  it is another entirely to attempt to re-interpret the tax law along the lines of an ideologically conservative imagination.

6. HOW SHOULD TAX REFORM BE UNDERTAKEN?

My general object here has been to argue the proposition that the onus of proof should, in any tax reform exercise, be with those who seek to alter the prevailing system.  Often enough, this "reallocation" of the onus of proof is enough.  There is an enormous gulf set between the question, "why?" and the question, "why not?".  If one could replace the "why not?" attitude in tax reform circles with the "why?" alternative, particularly among professional tax reform specialists, one would have won a significant victory.

But it would be foolish to pretend that this would scotch tax reform entirely.  And, of course, some tax systems may be so bad that almost any change would be an improvement.  Moreover, my argument -- rendered without rhetorical exaggeration -- is not that tax reform is never desirable but rather that tax reformers must be duly modest about the fragility of their arguments and about what it is reasonable to expect from the tax reform process, and that they must take due cognisance of the costs of any tax change.

When tax reform does take place, there are, it seems to me, better as well as worse ways of going about it.  I want to say a little about this issue here.

To be wise after the event, it seems clear that the "tax summit" method of securing tax reform was certain to be a failure.  To bring representatives of special interests together under public gaze for a brief period to consider any substantive tax reform proposals is to invite the death of those proposals.  The only consensus that would emerge in such a gathering is that the proposals under consideration (including the status quo if that had been on the agenda) were inadequate.  Everyone would want change:  it would simply be that all the desired changes would be in opposite directions.  Each representative is, entirely predictably, more interested in being seen to defend the interests of his group -- in posturing for the benefit of those back home (the "Hey, Mum, look at me" routine) -- than in securing genuine consensus.  In any event, consensus can't usually be pre-packaged:  it has to emerge as a compromise from often painful negotiation.

To see tax reform as the final result of such negotiation is to invite the politicisation of the tax reform exercise -- and, as I have argued, that sort of politicisation is what ought to be suppressed, not encouraged, in tax reform.  For somewhat similar reasons, however, it is not clear that simply implementing tax reform as a matter of floor-of-the-house politics is satisfactory either.  If the reigning party simply proceeds to change the tax system in the light of its electoral interests, then we can only expect that the tax system will fluctuate, possibly violently, with the electoral fortunes of different parties.  Moreover, it may be that floor-of-the-house politics simply reallocates the special interest pleading to the party room or the cabinet table.

The quasi-constitutional nature of the tax system, if accepted, carries with it implications about the processes by which the tax system ought to be changed.  In particular, just as constitutional decisions are supposed to transcend party politics or special interests, so, as far as possible, should tax reform decisions:  some means must be found to insulate tax reform from in-period politics.

The obvious prospect here is hardly a radical one:  it would perhaps be somewhat inconsistent with the general line of argument here if it were.  The proper institutional setting for tax reform, in my view, is the Special Committee of Inquiry or the Royal Commission.  This has been a long-standing method of dealing with issues in which ideological persuasion or narrow sectional interest is supposed to be suppressed.  It has a moderately distinguished history in the tax reform context in particular.  The Asprey and Mathews Committees in Australia, the Carter Commission in Canada, the Meade Committee in the United Kingdom are all examples of the genre.  A small number of individuals -- preferably ones with some popular authority, and with some expertise -- are gathered to deliberate on the question of desirable "tax reform", and to do so under terms of reference that attempt, as far as possible, to suppress special interest interpretation.  The responsibility is, or ought to be, not merely to provide a package of tax reform proposals, but to provide that package on a take-it-or-leave-it basis.  If the recommended package is to be subject to any significant amendment by the prevailing political party, the whole point of the exercise is placed very much at risk:  one might as well have done the exercise of tax reform by ordinary politics from the outset.  Analogously, making the reform package run the gamut of bureaucratic approval after the Committee has ceased to function is effectively to assign the tax reform exercise to the bureaucracy.  The appropriate point for bureaucratic involvement would seem to be to provide advice, "expert testimony" and feed-back as part of the Committee's deliberations.

Of course, it may be that a package so recommended will not survive political approval, in which case the status quo will prevail.  As I have argued here, this may not be the worst possible outcome by a long chalk -- provided there is a clear understanding that appointing a new Committee and proceeding through another "tax reform" exercise is out of the question for a significant period of time.

Much could be said about the right composition of any such Committee -- whether it ought to be headed by a judge, or an academic (who are the most common contenders) or some other.  Such detail, however, doubtless pushes us well beyond the point of relevance here.

One thing should, however, be noted.  While one might for various reasons want any such committee to be broadly "representative" in terms of professional background, ideological persuasion, social class, sex and so on, and while it may be tempting -- as a way of suppressing future opposition -- to seek to have representatives from various broad interests (particularly those with popular influence), it is almost certainly a mistake to have those interests represented formally.  Any committee member who sees himself as "wearing a hat" -- as having the role in decision-making of representing some interest -- undermines the constitutional nature of the process.

The conceptual ideal is the jury where special interests in the outcome are reason for disqualification, rather than the political arena or market place where special interests are intrinsic to the process.  The committee must, above all else, be "independent".

All this may seem to be quite distant from the issue of tax reform itself.  It is, however, quite fundamental.  It is not enough to know what you want -- it is also necessary to know how to get it.  Thinking about how to get genuine tax reform may also serve to alert us to the prospect that we may well get something we don't want.

There is, of course, one type of tax change that is not "reform" in the standard parlance, but which must be accessible if government is to retain any genuine policy discretion at all.  I refer to changes in aggregate revenue levels.  The ruling government must have the power to expand or contract spending -- and this implies changing tax rates, at the very least.  There is therefore some need to distinguish such aggregate revenue changes from what I have here termed "reform".  One procedure for doing this is to use only rate changes for periodic revenue adjustments, and to restrict the nature of such rate changes in some way -- for example, all revenue increments could be derived from flat rate levies or by a surtax in proportion to current taxes.  In this way, the general shape of the tax system can be maintained, with more radical restructuring preserved for institutional arrangements that seem most capable of dealing with them.

7. THE BOTTOM LINE

This essay is an attempted exercise in persuasion.  It is addressed to three audiences -- to the citizen-taxpayer;  to the professional economist-lawyer-accountant-bureaucrat, to whom may fall the responsibility to give advice on tax policy matters to those who make the tax law;  and to those practical politicians who have an interest in genuine tax reform.

To the latter group, there are two messages.  First, there may not be anything like as many votes in tax reform as one might infer from the clamour in its favour.  The reason for this is that no one can agree on what "tax reform" means, except that it should involve himself in paying less tax -- and that is impossible for everyone.  Tax reform is, in short, bound to disappoint.  Second, if one does want genuine tax reform, there are some ways that are more likely to achieve it than others;  and the way most likely to achieve it seems to involve a minimal role for the practical politician, beyond setting up the process in the most promising way.

To the second group, there are again two messages.  First, remember that there is social capital embodied in what it is that is being changed -- and that changes are costly.  If, for example, in the upshot, there were to be negligible improvements in efficiency and equity, would one want to implement change?  My argument is that there are quite substantial costs in doing so, The old adage to the effect that an old tax is a good tax is worth taking seriously.  Second, remember that the efficiency and equity consequences of alternative tax systems are extremely complex.  The tax system is inevitably rather like the amateur golfer's swing -- a mess of compensating errors.  To correct a few of those errors might very well do more harm than good.  Indeed, if the chances of having a better tax system when one's wonderful proposals have finally limped into law are only as good as fifty-fifty, then it is better to do nothing.

My message to the citizen-taxpayer is:  beware!  Beware of the practical politician who will promise you tax "reform" but give you tax hikes.  Beware of him, particularly, if he seems to be promising everyone lower taxes and no expenditure cuts, because he is promising something that it is not possible to deliver.  It is on the cards that he is indulging a little duplicity -- that some less visible tax (like the petrol excise, perhaps?) is being used to substitute for some more visible tax.  Beware of the bureaucrat-economist-public finance specialist, whose whole raison d'etre is tax reform:  the significance of his life is understood by him to be a matter of "having an influence" (a matter he talks about with colleagues in hushed tones in dark corridors), and he will happily sacrifice much in terms of your peace of mind to secure the implanting of some pet scheme into the Australian tax law.  Beware of his general predilection in favour of broad-based taxes:  that may well lead to higher taxes.  Beware of his talk of "fairness":  you may find that he has something in mind quite different from what you do.  Indeed, beware of tax reform generally.  Do not allow what you see to be imperfections in the prevailing system to lead you to accept that change is desirable.  For one thing, change is costly.  For another, "reform" may lead to horrors we yet know not of.

One final word of clarification.  It is a clear implication of what I have argued here that the current Hawke-Keating round of "reforms" is probably a mistake.  However, it seems to be a mistake that we have already committed.  The effective status quo is already now the tax system as currently "reformed" -- capital gains taxation, company tax imputation, fringe benefits taxation, rate adjustments and all.

To undo those changes would be to indulge in a new round of tax "reforms", and it is precisely this that I oppose.  The only actions that can be avoided are future ones.  It may not be an entirely happy thought to everyone that Hawke and Keating have turned out to have a relatively "final" word on tax reform -- but, at this point, that is the way it ought to be.