The Economist September 20, 1997 SURVEY OF WORLD ECONOMY

The future of the state

ALMOST any discussion of public policy nowadays seems to begin and end with the same idea: the state is in retreat. At the turn of the millennium, it is argued, governments are confronted by two old enemies, stronger now than ever before: technology and i deology. The state is proving unequal to the challenge. Its power to rule is fading.

A new industrial revolution is under way. Advances in computing and telecommunications press relentlessly on, shrinking distance, eroding national boundaries and enlarging the domain of the global economy. Increasingly, these changes render government s mere servants of international markets. Reinforcing the technological shift is a transformation in the realm of ideas, starting towards the end of the 1970s and reaching its climax ten years later with the collapse of communism. That destroyed the syste m not only in the form practised in communist countries but, more important for those in the West who never experienced it directly, as a sustaining Utopian myth. Judged as propaganda, 1989 did for big government what 1929 did for laisser-faire.

Today the main points of this story are almost universally accepted: most tellingly, perhaps, by the world's political leaders. Left and right, eagerly or reluctantly, bow down before global capitalism. The debate about ``globalisation''-whether it is a good thing or a bad thing, whether to embrace it or resist it-is vigorous, but nobody seems to doubt the new power of international market forces. Indeed, the two main sides in the debate agree not just about that, but also about the nature of the balan ce that societies must strike as a result.

In one corner are the optimists: an alliance of modernising conservatives and the new post-socialist left. They regard the triumph of international capitalism as largely a good thing. Governments may have lost some of their freedom to direct economies as they wish, but the world is benefiting from faster technological progress, historically unprecedented opportunities for the relief of global poverty, and greater freedom for millions of people across the globe. Few, if any, optimists doubt that a well- functioning society requires a state that is competent in both senses of the word, which makes the state's shrinking economic sovereignty something of a drawback. Nonetheless, they find the balance comfortably positive.

The case for gloom

The pessimists-a coalition of populist conservatives, assorted communitarians and the old left-agree on the need for a balance between an effective state and the economic efficiency that market forces can provide. They also agree that market forces, for r easons of technology and ideology, have lately gained the upper hand. But unlike the optimists, the pessimists find this deeply disturbing. In their view, the gains from globalisation are far smaller than the optimists suppose, and the drawbacks much grea ter. And such benefits as there may be will be divided unfairly within society-a crucial point that the optimists tend to ignore.

The new global capitalism, the pessimists concede, will certainly enrich many-but capitalists rather than workers. Moreover, those who fare worst among the workers will be the unskilled, least able to fend for themselves. Globalisation will widen ineq uality, exacerbate poverty and increasingly lead to social ``exclusion''. These costs will mount even as globalisation succeeds in its own terms, at a time when government's capacity to respond is draining away. Its failure to act will undermine the found ations of the democratic state, challenging its very legitimacy. As the remorseless advance of market forces plunges capitalism into a new crisis, Marx will have the last laugh after all.

If this survey had to take sides in the argument, it would support the optimists. For reasons to be explained shortly, they are much closer to being right than the pessimists. But both they and the pessimists are wrong in accepting so readily, first, th at the state is indeed in retreat; second, that globalisation will play the decisive role in determining the future of the state; and third, that any infringement of the powers of the state must be counted as a cost against corresponding benefits. On the first two points, it seems, both sides share a taste for hyperbole; on the third, they betray similarly collectivist instincts.

A more liberal thinker might ask which powers of government have, in fact, been infringed by international markets, and whether those powers, if any, should have resided with the state in the first place. Perhaps the shrinking state is a good thing in its own right, in which case global market forces deserve a more enthusiastic welcome. Or perhaps, despite the spread of capitalism, the state is shrinking more slowly than it should, or not at all-in which case the main thing wrong with globalisation is that it is not doing its job thoroughly enough. This survey aims to sort through these ideas and come to some conclusions about the future of the state.

One of its main themes will be that the effect of globalisation has been overrated. For the rich West, both its costs and its benefits are less than meets the eye. For poor countries, undeniably, it makes a big difference to the prospects for economic d evelopment. International integration is their fast track out of poverty. With small domestic markets, backward technology and inadequate capital, third-world countries have everything to gain from ending their relative isolation and developing close econ omic ties with the rest of the world. But in the advanced economies, where internal markets are big enough to be fairly competitive, barriers to flows of trade, capital and knowledge have long been lower, and technology is at its cutting edge, there is le ss to be gained. For these countries, globalisation is a good thing, but hardly a matter of life or death.

In the advanced economies, in other words, faster economic growth cannot be imported, it must be built chiefly at home. The issue is not whether governments will hold back the forces of change from abroad, but whether they will resist the more insisten t forces of economic change in general. In the rich countries, most of these forces will be domestic rather than international. Most of them, in one form or another, reflect basic economic liberties. Western nations may have reconciled themselves to freer international movement of trade and capital; but that will not oblige them, this survey will argue, to embrace economic liberty within their own borders. The fundamental question for the advanced industrial countries is not, as many suppose, whether demo cracy is compatible with globalisation, but whether democracy is compatible with liberty.

A funny question

Even to pose that question will strike many as odd. Democracy and freedom seem two sides of the same coin: how can there be one without the other? The connection meets no more scepticism than the idea that global market forces are in the ascendant. Yet bo th are dangerous simplifications. Freedom and democracy are linked, of course-but so are freedom and capitalism. And, unfortunately, the pessimists are right to question whether capitalism can continue to get along happily with democracy.

They are wrong, however, about the form that a conflict between the two would most likely take. In their view, the danger is that capitalism will in the end prove so socially catastrophic that democratic states will be overwhelmed by protest, leaving fa scist or otherwise non-democratic regimes to rise up in their place. Another scenario is more plausible. Democratic states, indulgent of anti-liberal values, may make such demands of capitalism, and place such burdens and restrictions upon it, that it wil l slowly fade away, along with freedom. The emerging polity might still be ``democratic''-but that would make it no less dysfunctional and, at the extreme, hardly any less tyrannical.

Just now, avowedly left-of-centre parties rule in much of the advanced industrial world, including America, Britain and France. Popular anxieties about globalisation, advancing technology and economic insecurity may be partly to blame. Arguably, though , left-of-centre no longer means what it used to. Bill Clinton, New Democrat, took office in the United States proclaiming: ``The age of big government is dead.'' Tony Blair's New Labour is far to the right of the party that lost power in 1979. In France, Lionel Jospin's Socialists seem closest to the old model, but even they are shifting a little.

For the most part, today's left-of-centre governments have made their peace with the world of business. They no longer think of themselves as defenders of organised labour or champions of public spending. Words such as ``enterprise'' and ``opportunity' ' trip lightly off the tongue. These are post-socialist leftists, if they are leftists at all. Perhaps they are no more likely than the conservative parties they defeated to raise taxes and enlarge the role of the state. Many of them won their elections p romising to do no such thing. Does this not suggest that the West has, after all, reconciled itself not just to globalisation but also to a greater measure of economic liberty at home?

It does not. The so-called realignment of the left falls far short of a victory for liberalism. The left, on the whole, may have abandoned its traditional figures of speech and even, to a much smaller extent, its preferred instruments of economic contr ol. But it insists that its traditional values have not changed. From a liberal point of view, the new left's old values-its enduring collectivism and anti-individualism-are precisely the problem. Moreover, many conservative parties, in their own way, are just as keen on these values as the new left. Appeals to collectivism and anti-individualism are clearly popular with voters, especially when detached from any discussion of the policies needed to put them into practice. Democracy is comfortable with the se values. Capitalism is not.

But even if this assessment is correct, why is it more plausible to suppose that democracy will undermine capitalism rather than, as the global-economy pessimists argue, the other way round? The answer is simple: because that is what has been happenin g in the West for the past 50 years.

In all of the advanced economies, government has grown fat

Spend, spend, spend

DID somebody say the age of big government was dead? At the beginning of this century government spending in today's industrial countries accounted for less than one-tenth of national income. Last year, in the same countries, the government's share of out put was roughly half. Decade by decade, the change in the government's share of the economy moved in one direction only: up. During war it went up; during peace it went up. Between 1920 and the mid-1930s, years of greatly diminished trade and internationa l economic contact, it went up. Between 1960 and 1980, as global trade and finance expanded, it went up. Between 1980 and 1990, as this breeze of globalisation became a strong wind, it went up again. Between 1990 and 1996, as the wind became a gale, it we nt up some more (see chart 1).

Among the rich industrial countries, America and Japan have the smallest governments. Last year their public spending was 33% and 36% of GDP respectively. Even so, both have shared in the consistently upward trend of state spending (except that in Japa n, unlike almost every other country, spending was higher just before the second world war than it was in 1960). In America, government spending in 1913 accounted for less than 2% of the economy; by 1937, it was still only 9%.

Since 1960 America's government has grown by about a fifth, a comparatively modest rise. Internationally, the average increase over that period was more than three-fifths, while public spending in Japan more than doubled. The point is that government e verywhere has grown, and kept on growing-even in those countries where, by today's standards, government is small. Big government, far from being dead, is flourishing mightily.

Against the tide

True, there are exceptions. Sweden, where in 1993 the government's share of the economy had been 71%, has since repudiated its social-welfare model and cut public spending savagely, to just 65% of national income last year. Or look at the extraordinary tr ansformation in Britain. In 1980, when Margaret Thatcher began wielding her Conservative axe, public spending accounted for 43% of the economy. After nearly 20 years of ruthless cuts, radical dismantling of the welfare state and hard-faced suppression of public-sector unions, the state's share has shrivelled to just 42%. Sickened in the end by this remorseless brutality, the British electorate earlier this year swept Labour back into power with a landslide majority.

Since 1994, many European governments have been trying to curb their budget deficits in order to satisfy the fiscal-policy criterion laid down in the Maastricht treaty on European monetary union. That may have helped to bring average public spending in Europe down a little last year compared with two years earlier. Even so, last year's average was higher than that in 1990 (which in turn was higher than that in 1980). Public spending in France last year was 55% of the economy-after ``cuts'' that cost th e government an election earlier this year, but actually did no more than hold the total steady. In Belgium, despite supposedly heroic efforts at retrenchment, public spending was still 54% of GDP. In Italy and Austria too the figure was well over 50%.

The trend towards bigger government in the industrial countries has been almost universal (see table 2). One way of finding out why this should be so is to look at the composition of spending. The total falls into four broad categories: (a) government consumption, measured by what the state, as a supplier of services, spends on wages and other inputs; (b) public investment; (c) transfers and subsidies; and (d) interest on the national debt. For the industrial countries as a group, between 1960 and 1990 public spending as a proportion of national income fell in only one of these categories: public investment, down from an average of 3% of GDP to 2%. Of the other categories, debt interest has grown most quickly. Next fastest-growing, and much the biggest category, were transfers and subsidies, followed by public consumption (see chart 3).

The engine room

These numbers, broad averages though they may be, are revealing. Consider first the exceptionally sharp rise in debt interest. This reflects the build-up of government debt, caused by an accumulation of government deficits. But what is behind those defici ts? In the short run, governments may borrow to finance their activities instead of collecting taxes to pay for them all, as a way to stabilise the economy as it moves through the business cycle. If that were the only motive, however, deficits during rece ssions (as governments spend more than they collect in taxes) would be balanced by surpluses during recoveries (when the opposite is true), and there would be no build-up of debt over decades-which is what has actually happened. Deficits have become more or less permanent. This is a sign that governments are persistently spending more than citizens can be persuaded to pay in taxes.

Spending on public consumption (that is, on services such as defence, law and order, education and health) has risen substantially not only in inflation-adjusted terms but also in relation to the size of the economies. Remember that economies have been growing-between 1960 and 1990, by an average of 3.7% a year in the industrial countries-so government has been consuming a rapidly growing slice of a rapidly growing pie. That would seem to suggest an enormous rise in the quantity and quality of the serv ices provided.

But multiplying the change in GDP by the change in the share of GDP devoted to government consumption merely provides a figure for the growth in the cost of delivering public services. It says nothing about the volume. Economists who have studied the p ublic sector have long noted that productivity there rises much more slowly than in private business. (They have even given a name to this oddity: ``Baumol's disease''.) A good part of the enormous increase in resources devoted to public services can be e xplained by that slow-rising productivity. That may be why most voters in most countries appear to believe that, despite the huge growth in spending on public services, improvements in recent years have at best been modest.

The most important cause of the state's expansion since 1960, however, is the growth in transfers and subsidies. These include income support, benefits for the unemployed, disabled and single parents, and above all pensions: altogether, the cash-in-han d sector of the modern welfare state.

In the United States, subsidies and transfers in 1937 accounted for 2% of GDP-a figure that already reflects growth in spending after the Great Depression. By 1960 they had reached 6% of the economy; now they stand at 13%. In Britain, subsidies and tra nsfers in 1937 were already 10% of the economy, far higher than elsewhere. By 1960 they had increased to 14%; today they account for roughly a quarter of the economy. By 1960, however, Britain had lost its commanding lead in the welfare-state race; and by 1970 France, Norway, Sweden, Belgium and the Netherlands had left it far behind.

The spectacular growth of the transfer state explains (in a statistical sense, at least) something that at first sight seems very odd: that the state has grown fastest when the pressure to do so has seemed least acute. Asked which event this century ha s done most to drive up public spending in the advanced economies, most people would probably say the Great Depression, or the world wars, or the development of state-run health and education systems in many countries after 1945. And indeed all those fact ors played a part in enlarging governments, but none of them was the single most important.

Only in America might an event in the 1960s spring to mind as a candidate: the Vietnam war. The timing is right. Much the most powerful impetus for growth in public spending came after 1960, but before the oil shocks of the 1970s. However, Vietnam does not fit the bill even for America: higher defence spending is an increase in government consumption, and nowhere was growth in that category of spending the biggest factor in pushing up the total. Instead, the most powerful change this century was the ex plosive growth in transfers after 1960, a shift that occurred at about the same time in almost all industrial countries. This was not a response to some crisis, but the very opposite: an unforced act of policy at a time of rapid economic growth and relati ve political stability, often regarded fondly as a golden age.

The state grows in bad times, it seems, because it has to. It grows in good times as well, only faster, because governments feel more ambitious.

Cry freedom

Does this growing economic encroachment by the state-underwritten by democracy-pose a threat to liberty? Many will argue that liberty, properly defined, is not at stake as the government's share of the economy grows. Political and civil liberties are not infringed by high taxes and public spending: indeed, they may be advanced by those means. Economic liberty is reduced, of course, but even if it is related to the other kinds, the economic sort is quite inferior. Political liberties-freedom of speech, fre edom of assembly-are natural, inalienable rights. Economic liberty is just a cloak for greed.

The West's modern neglect of economic liberty-or, as it often seems, its conscientious objection to it-goes far to explain why the state has been able to grab an ever bigger share of output. The classical liberals would have regarded it as puzzling, if not insane. Albeit in a moment of hyperbole, Locke famously said that ``government has no other end but the preservation of property.'' His point was that free men occupy a private space; governments which enter without consent are ``masters, or arbitrar y disposers of . . . the lives of the people''-in a word, tyrants. This private space is defined by political, civil and economic liberties. They are all tied up together.

Measure this position against the stance of many of today's political commentators. They write as though a government that confiscated all your property would not be infringing your rights-as long as you were free to complain. Indeed, if the government had a ``mandate'' for its act of confiscation (that is, if it had canvassed it in an election campaign, and won), carrying it out would be a mark of virtue. Thus, when Britain in the 1970s applied a top marginal tax rate of 98% to investment income, the policy was criticised chiefly for its counter-productive effect on revenue, much more than for infringing people's liberty. But you do not have to be a John Locke to be bothered by definitions of freedom that disregard economic issues; or to feel that whe n governments have seized more than half of the entire economy, too little of that private space remains for the people inside it to be called ``free''.

The classical-liberal strictures against confiscation apply most forcefully to punitive taxes applied to a minority of the population, as in aggressively redistributive systems of taxation and spending. The classical liberals accepted the need for comp ulsory taxation to pay for essential goods and services that the market, left to itself, would fail to supply, arguing that free citizens would consent to such arrangements. A good many modern followers of the liberal tradition would also say that even sc hemes for the redistribution of income may be acceptable, so long as they are moderate.

Perhaps, then, the growth of public spending faithfully reflects the wide acceptance of an expanding list of ``essential goods and services''. If that is so, you might argue, liberty has not been infringed-not because economic liberty is excluded from the proper definition of freedom, but because governments are giving citizens what they want. That, after all, is what democracies are for, is it not?

This is a more complicated question than it may look. Before returning to it, and assuming for now that governments are indeed delivering what the voters want, it is worth considering a less philosophical one: have voters been getting good value for mo ney?

Big government, big bill

Democracy at a price

WHEN it comes to providing social goods-better education, say, or better health as measured by greater life expectancy and lower infant mortality-heavy-spending governments seem to be doing little if any better than governments that spend much less. This only deepens the mystery of the ever-expanding state.

To be fair, up until the 1950s increases in public spending did seem to produce worthwhile improvements in health and education. Certainly up until the 1930s, the larger part of expanding government expenditure was taken up by investment in infrastruct ure, the supply of essential public services, and the creation of a low-level safety net to guard against poverty. In all these respects, it could be argued, the state was attending to things that the market, left to its own devices, would neglect. In tho se days, a much larger share of public spending was devoted to dealing with genuine ``market failures'', as they would now be called.

Later, and especially after 1960, priorities shifted. By now, far fewer gaps in the market remained. Increasingly, the state expanded its role by seizing, and monopolising, activities hitherto left to the market; and by transferring resources in ever m ore complicated ways from one part of the economy to another, to serve an ever more ambitious (and opaque) range of goals. As a result, in recent decades the connection between increased public spending and improved social outcomes has become much weaker, and in some cases has broken down altogether.

A quick way to demonstrate this is to divide the advanced economies listed earlier into three groups: ``big government'' countries, where public spending in 1990 was more than 50% of output; ``small government'' ones, where public spending in 1990 was l ess than 35% of output; and the rest, which can be disregarded for this purpose. The big-government group takes in Belgium, Italy, the Netherlands, Norway and Sweden; the small-government group Australia, Japan, Switzerland and the United States. For good measure, add to the small-government group two recent arrivals to the ranks of the rich industrial economies: Singapore and Hong Kong, both of whose governments spend less than 35% of GDP. How do these groups compare on broad measures of welfare?

The best single such measure is output per head, adjusted for international differences in purchasing power. By this test, and on the basis of our small, unscientific sample, the small-government countries are just in front. Their adjusted output per h ead in 1995 was $23,300; the corresponding figure for the big-government countries was $20,400. The small-government economies' growth rates were higher, too: an average rise of nearly 4% a year in income per head between 1960 and 1995, compared with 2.5% for the others (though note that growth in the small-government group comes down to the same lower figure, 2.5%, if Singapore and Hong Kong are left out).

What about health? Spending varies a lot, but outcomes much less. Average life expectancy in big-government and small government countries is very similar, at 78.0 years and 77.8 years respectively. Rates of infant mortality are also much the same in the two groups: 6.0 per 1,000 births in the big-government economies and 5.5 per 1,000 in the small-government ones.

As with health, spending on education varies a good deal; again, however, the results are much closer. The United Nations Development Programme calculates a composite school-enrolment ratio, weighting together the proportion of children attending schoo l at different ages. The higher the ratio, the better a country is doing. In the big-government countries the index is 85%; in the small-government countries it is 78% (or 82% if you exclude Singapore and Hong Kong). The estimated level of adult illiterac y is very low in all the countries in both groups. In a recent international survey of children's skills in mathematics and science, the small-government countries on average did markedly better, particularly in maths, than the big-government countries.

On one social indicator, however, the two groups diverge much more obviously: inequality of incomes. In the big-government countries, the poorest 20% of the population receive 7.4% of the national income, whereas in the small-government economies they g et only 5.6%. Even allowing for the fact that incomes overall are higher in the small-government group, in 1995 the poorest 20% there were worse off in absolute terms than their big-government counterparts (see chart 4). Conversely, the richest 20% were m uch better off in the small-government countries than in the big-government ones, both in absolute terms and relative to the less prosperous in their own economies.

Other things equal

It is possible to argue that the more equal income distribution achieved by transfers in the big-government group is preferable to the greater disparity in the small-government group, even if the relatively small gains at the bottom come at the expense of much larger losses at the top. The question is whether this apparently modest increase in equality justifies surrendering an extra 15-20% of the economy to public control.

The fact that Singapore and Hong Kong qualify for inclusion in the small-government group of advanced economies-they are already richer than Japan and Germany, and still growing fast-suggests a further point. Today's fastest-growing ``emerging-market e conomies'', as they are still called, have much smaller governments than most of their western counterparts did at comparable stages of their development. Even in 1960, before their transfer states began to expand in earnest, the big-government countries of the West (and the middle-sized-government countries, too, for that matter) had dedicated 30% of their economies to public spending. Two of the Asian tiger economies have already far surpassed the incomes the West had achieved by 1960. Others are gettin g there fast, with governments smaller than America's was 40 years ago.

When western analysts reflect on the role of government in East Asia, they concentrate on trade and industrial policies. How much of the rapid growth in East Asia, they ask, is due to clever industrial intervention, and how much to letting markets hav e their say? This debate seems to have blinded people to an equally important fact: in the world's fastest-growing economies, government, measured by the extent of its spending in the economy, has stayed small. In years to come, will the Asian tigers buil d transfer states of their own-and if so, will they be able to maintain their faster rates of growth?

In the meantime, a puzzle presents itself. Government in the West has grown to a point where liberty (at least as liberals understand that term) is being seriously infringed. Yet this enormous expansion of the state appears to have yielded little in re turn. Something seems to have gone wrong. Who, or what, is to blame?

One simple piece of theory sheds a useful light on the question. It shows how taxes, even if they pay for worthwhile things, impose an inescapable economic burden, called a deadweight cost. It also shows that this cost rises more than proportionately a s taxes go up (see box, next page).

This analysis, simple as it is, suggests one way in which big government imposes a cost on an economy: every tax-and-transfer destroys some social welfare. No less important, it shows that big government is likely to produce severely diminishing return s. At low levels of government activity, the deadweight cost of taxes may well be outweighed by the gains that flow from essential public goods. But as the scale of government grows, this trade-off seems likely to move in the wrong direction-and at an acc elerating pace.

On the spending side of the government's activities, further inefficiencies of various kinds arise, often because the government is usually a monopoly provider of the services it supplies. This is either because it forbids competition, or (as in privat e health and education) because its service is ``free'' to users, whereas the private alternatives involve large additional expense. To a much greater extent than a private monopolist, a public monopolist is free from pressures to innovate or become more efficient. Moreover, lack of competition suppresses information that the market would otherwise reveal about the cost of supplying different goods and services and about the value that consumers put on them.

The extreme case of this inefficiency in the allocation of resources was communist central planning. Even those economists who had long doubted the Soviet Union's economic statistics were surprised when the full extent of that failure was laid bare. A ll across the Soviet empire were enterprises that for decades had been engaged, not in adding value to raw materials but in literally subtracting it: the value of the finished products, measured by the unforced willingness of buyers to pay for them, was l ess than the value of steel, plastic, glass and other raw materials that went into them. It is well to remember just how badly bureaucrats can fail if allowed to operate as monopolists over a sufficiently wide range of operations.

But memories are short. Those who favour more government intervention in the economy simply point to some market failure, and rest their case. What they are saying, in effect, is that the superiority of markets in allocating resources is strictly theor etical. The textbooks say that given a great many buyers and suppliers, perfect information, homogeneous goods and a complete set of futures markets, perfect competition is bound to follow, bringing ``efficiency'' in its wake. But if any of these conditio ns is missing, the whole apparatus will collapse. There will be market failure, and the desired results will not materialise. This is all you need to say to justify government intervention.

This argument, repeated with endless variations, is specious every time. The case for the superiority of markets over planners is empirical, not deductive. Market failure in the sense just defined is not merely common but universal: the textbook condit ions are never fully satisfied. But that is not a helpful fact. What matters is to know whether, in practice, imperfect markets work better than imperfect governments. In practice, do competition, incentives to effort and innovation, and survival of the f ittest in the marketplace-however flawed that market may be-work better than bureaucrats supposedly pursuing the public interest?

The lessons of history-in the Soviet empire, in China, in Africa and Latin America, in Europe and the United States-suggest that when it is possible to leave the allocation of resources to the market, the results are nearly always better. People who s till doubt this are surely no longer entitled to argue that any sign of market failure is sufficient justification for yet another enlargement of government. The deadweight cost of taxes, together with the inefficiencies caused by lack of competition a nd proper incentives within the public sector, helps to explain why government begins to fail as it grows. But it cannot explain why voters in democracies keep asking for more. Likewise, analysing the components of public spending takes you only so far to wards an answer.

Government borrowing allows countries to spend beyond their means, or beyond their people's willingness to pay taxes, at least for a while. The fact that transfers have grown so much more quickly than public investment and consumption may also reveal some thing about the demands that the modern state is expected to satisfy.

But the puzzle remains. Given the evident drawbacks of big government, and given that nobody likes paying higher taxes, why is it that public spending and taxes have both risen so dramatically over the past 50 years?

The hidden cost of taxes

THE first diagram in the panel shows a demand curve and a supply curve for some hypothetical good. Usually, as the price of a good comes down, the quantity demanded increases; the demand curve therefore slopes downwards from left to right. Usually, as the price of a good goes up, the supply of it rises too; so the supply curve slopes upwards. With buyers and sellers free to trade, a balance of supply and demand will be established at the point where the two curves cross-point X, where the price is P and w here the same quantity, Q, is both demanded and supplied. That point of equilibrium gives the market's answer to how much of the good will be traded and at what price.

Turning to the second diagram, the shaded area between the demand and supply curves, to the left of the point where they cross, has a special significance, because it represents the net addition to social welfare that is created when the good is bought and sold at the market price.

If you divide the area into two, the upper part, A, represents the so-called consumer surplus. Every unit of the good sold when supply equals demand-the whole of the quantity Q in the diagram-is sold at the market price, P. But smaller quantities of the good could have been sold for more than P. Only for the last (or marginal) unit sold is P the top price the consumer would be willing to pay. In effect, therefore, all but that last unit have been sold for less than they are worth to the consumer. The ar ea A adds up all these surpluses, unit by unit, showing the value of all the transactions to consumers over and above the price they paid.

By the same logic, the lower part of the area between the demand and supply curves in the second diagram, B, represents the producers' surplus. Only the last unit supplied costs its producer exactly P. Other producers would have been willing to supply at a lower price, enough to deliver some smaller quantity of goods to the market. When these not-on-the-margin units are sold at the market price, their producers are paid more than they would have been willing to accept. The area B adds up all the produ cer surpluses.

The third diagram shows what happens when a tax is imposed, raising the price paid by consumers from P to Pc, and lowering the price received by suppliers to Ps. At these new prices, Qt is demanded and supplied. The amount of the tax (the difference bet ween Pc and Ps) multiplied by the number of units sold (Qt) gives the revenue raised for the government (area C in the diagram). Both the consumer surplus, A, and the producer surplus, B, are accordingly smaller than before.

That was to be expected. The point is, though, that the two surpluses, added together, have shrunk by more than the amount taken away in tax. Now that the quantity of goods supplied has fallen to Qt, the triangle D has disappeared: it is not part of th e government's tax yield, and it is no longer part of the economic surplus; it has simply vanished. This part of the reduction in the surplus is a pure loss to the economy, known in the jargon as the deadweight cost of the tax. The implication is that if the government raised the area C in taxes and then handed the money straight back as lump sums to consumers and producers, the economy would still be poorer than before because the area D would still be missing.

In the last diagram the tax is twice as big as before. The price to consumers has increased once more, and the quantity supplied has fallen further. The consumer and producer surpluses are also smaller. The government's tax revenue, C, may quite possib ly be smaller too, despite the higher tax rate, because of the smaller quantity traded. The deadweight cost, however, has increased fourfold.

If the demand and supply curves were indeed curves rather than straight lines, the relationship between tax rise and pure economic loss would not be quite so simple. But the basic point would be the same: in general, the deadweight cost of a tax rises ex ponentially as the tax goes up.

Why do voters settle for such a bad deal?

The enigma of acquiescence

IT IS possible that people have been willing to pay ever higher taxes because they want more and better public services, and believe this is the way to get them. Perhaps, on this view, the willingness to pay higher taxes is only now coming to a halt-eithe r because (as in Sweden) taxes had been pushed to preposterous levels, or because (as in Britain?) people are losing all faith that government will spend their money wisely. The trouble with this explanation is that people have always thought their taxes were too high, complained of official incompetence and, once enfranchised, demanded promises of lower taxes in return for their votes.

Public spending in Europe, at least, has fallen in the past two or three years. Despite the tremendous fuss made over this, the fall is small (from 51% of GDP in 1994 to 50% last year) and partly bogus (books have been cooked as the Maastricht treaty' s deadline approaches). It is too soon to say whether it will prove durable. Perhaps a kind of tax exhaustion is at last setting in. But history inspires little confidence: even though taxes have never been popular, that did not stop them from rising almo st without interruption for decades.

Perhaps voters are just confused: they hate taxes but want more public spending, and expect governments to satisfy them on both counts. There is another possibility, however-a way to solve the puzzle of voters' acquiescence in the ever-proliferating, y et under-achieving welfare state. The trick is to stop thinking that politicians and their bureaucratic helpers are out to serve the public interest. Of all people, the ones who cleave most naively to this idea are those supposedly cynical and dismal fell ows, economists. All the theories of mainstream economics (including the tax theory sketched in the previous section) assume, in effect, that governments are Platonic guardians-selfless servants of the public good. If that were true, the remarkable growth of government during the course of this century would be awkward to explain. But once you assume that governments are made up of people who most of the time are guided by considerations of self-interest, the mystery is solved.

Suppose, for a moment, that politicians in democracies are concerned mainly with winning and exercising power-however outlandish that may seem. Straight away you would expect a pull towards greater public spending, because that is the main way in which a government exercises power. Taking an extreme view, the government's one and only aim might be to maximise its revenue (hence its spending, hence its power). In doing this, it would face the same difficulty that once confronted absolute monarchs: if th e people are taxed too heavily, they become less productive and spend their energies on hiding their money instead. To be a good rancher of citizens, you must look after your livestock. That is equally true for elected and unelected rulers. Unlike an abso lute monarch, however, a modern democratic government also faces a political difficulty: it has to get elected.

This affects the fiscal arithmetic. In the simplest case, a two-party system, both parties will compete to woo the ``median voter'', who occupies the electorally crucial middle ground. Suitably dressed up, the pitch will take the form of promises to re distribute money from the minorities at either extreme in order to provide new transfers or services to the median voter.

But seriously . . .

This will strike many as an absurd caricature, and indeed it should not be taken too literally. But in broad terms, it is a fair account of what has happened in the western democracies since 1945. The huge expansion of the state, especially after 1960, ha s improved the lot of the poor by far less than the increased transfers and higher spending on services might have led you to expect. What all these countries have in common over this period is that their welfare states have been, to a greater or lesser e xtent, ``captured'' by the middle classes.

In most western countries, much the most expensive government programmes are nowadays the middle-class ``entitlement'' programmes-intended not to provide a safety net for the poor but to deliver elaborate and expensive services to all. These include pe nsions, health care, education (up to and including university education), public transport and housing subsidies. The poor get some benefit from these programmes, along with everybody else. Plainly, however, these policies are no longer aimed principally at helping the least well-off, if that was ever the goal.

Studies in many countries show that it is the better-off who benefit disproportionately from these programmes. Julian Le Grand, a British economist with egalitarian sympathies, studied the way British public spending was divided between rich and poor i n the 1980s. In health care, the wealthiest fifth of the population received 40% more public spending than the poorest fifth; in secondary education 80% more; in bus subsidies four times more; in university education five times more; in housing subsidies seven times more; in rail subsidies ten times more. This ignores the subsidies for farming, or for the arts, where the rich-to-poor ratio for ``public services'' such as opera must be approaching infinity.

This idea of middle-class capture should not be pushed too far, or it would beg the question why the rich and poor do not form an electoral alliance to squeeze the middle, or why the middle classes and the rich do not go into permanent coalition to end all transfers to the poor. Clearly politics is too complicated a business to fit such simple theorising. In real life, many facts do not conveniently fit the theory: for example, building any rich-poor coalition would face awesome difficulties, and altru ism plays a part in both the form and the substance of party platforms across the political spectrum. Nonetheless, the idea of the self-interested politician competing for the median vote should not be dismissed altogether. Something that would have seeme d unthinkable 50 years ago-a colossal expansion of the state as a supplier of transfers and services to the middle classes-has undeniably come to pass. And the median-voter theory seems a plausible way to help account for it.

The theory begins to look all the more plausible if you enrich it with the notion of the self-interested bureaucrat. Government officials obviously have an interest in expanding the scope of their departments-the bigger the bureaucracy, the better the promotion prospects for incumbents. They also prefer to monopolise areas of activity rather than work in competition with the private sector, not merely in order to grow, but for the sake of an easier life. When the state is small, the connections between politicians and the people working for them may be personal or financial; later, when the state has grown, public workers often become an influential electoral constituency in their own right. This is one of the factors that makes shrinking the state, or reforming it in any way, so difficult in the western democracies.

One last ingredient is the role of interest groups. Mancur Olsen's classic essay ``The Logic of Collective Action'' explained more than 30 years ago how interest groups come to be formed and why, once formed, they are extremely hard to disperse. The gr oups that establish themselves are likely to be relatively small, with concentrated interests: this makes it more likely that the reward for organised action will be large relative to the cost that each member bears in pursuing the group's cause. Yet they are big enough to demand the attention of politicians and the media, so governments must attend to their demands, as well as to the preferences of the median voter.

It is unlikely to be a coincidence that, along with government, interest groups have proliferated in recent years. Almost everyone seems to be a member of two or three. Many of these groups operate on the same premise: that the fastest way to advance the well-being of their members is to arrange transfers of various kinds from non-members. These may not always be in the form of cash or services, but increasingly these days manifest themselves as regulations. Usually under the banner of ``fairness'', w hich has become an infinitely elastic principle, or perhaps in the name of ``sustainable development'' or some other worthy cause, ever more people devote their best efforts to transfer-seeking rather than to economic production. Historically, their great est triumph is probably agricultural protection, which involves an immense transfer of resources from consumers to farmers, and which continues to this day despite the West's supposedly universal commitment to liberal trade.

A combination of these elements-self-interested politicians, self-interested bureaucrats and self-interested pressure groups-may not be the whole explanation for the remarkable expansion of government this century, but it goes a long way. What it impli es is a kind of democratic failure, akin to the market failures that government intervention is supposed to remedy. Collectively, citizens may understand that big government is producing ever more disappointing results. Individually, they and their politi cal leaders keep demanding it.

But now, it is argued, there is an intriguing new possibility. Technology has greatly strengthened market forces at the expense of government. Perhaps, reflecting this change, ideas have shifted in the same direction. But even if ideas have not moved t hat much, an age-old scepticism about government has now gained an extremely powerful new ally. In either case, this raises a new question. Will globalisation tame the Leviathan state?

Will market forces shrink the state?

Markets go global

WHEN James Carville worked as Bill Clinton's chief spin-doctor during the 1992 election campaign, he changed his mind about his next life. Until then, he had wanted to come back as the pope; but having been told over and over again by colleagues that such -and-such a ``progressive'' policy was unaffordable, and that just discussing it would give Wall Street the jitters, he concluded that he would rather return as the bond market. The financial markets, Mr Carville felt, had finally become more powerful tha n God.

The idea that elected governments have been rendered powerless by market forces has become a well-worn cliche of modern political commentary-although, as the earlier evidence shows, it is still a long way from being true. But perhaps things are startin g to move that way?

Once the automatic assumption that governments are disinterested servants of the public good has been dropped, the much-heralded assault of global markets on the prerogatives of the state takes on a new character. In moderation, it may begin to look li ke a good thing. If you happen to live in Scandinavia, you do not have to be a raving liberal to wish for just a bit less government. No doubt the best solution would be for people to vote for less government, and for governments to be able to oblige. Fai ling that, a bit of shrinkage under duress seems acceptable. However, what might arouse anxiety in Scandinavia and elsewhere is the thought that the power of the state to do anything-including the essential things that only states can do-will eventually b e destroyed. This is the spectre raised by some global-economy pessimists.

At first sight their argument seems persuasive. It goes as follows. Even if governments are to discharge their minimum responsibilities of providing public goods such as defence, and law and order, they must collect taxes. People cannot be expected to volunteer to pay for such things. They will be tempted to become ``free-riders'', calculating that others will pay, and realising that those who refuse to do so cannot be excluded from the benefits. The money needed to pay for public goods, therefore, has to be collected through taxes.

Beyond that narrow core of services that everyone agrees must be provided by the state, the argument continues, there are many others that have some of the characteristics of public goods: basic education, for instance, or primary health care. Left to itself, the market may well supply a lot of these services, but it will still not provide enough, because at least some of the benefits will extend beyond those who pay for them. Once again, though in a more limited way, the difficulty of excluding free-r iders means that too little of the service will be supplied. By stages, this kind of argument can be extended across the entire range of government spending programmes. At the extreme, it has been argued that equality of income is itself a kind of public good-something from which everybody benefits but which individuals, left to themselves, will not pay for. If governments are to supply much from this far longer list of goods, the tax bill is going to be big.

That is all right, the argument goes on, so long as the mobility of capital and labour is limited. But what happens when advancing technology and continuing deregulation allow capital and labour the freedom to roam? Owners of capital, seeking the highe st return, will flee from high-tax regimes to low-tax ones. Skilled labour, too, will start to shop around the globe for a tax haven. Governments, aware that they must compete to attract and retain these scarce factors of production, will cut their taxes and hence their spending. A ``race to the bottom'' gets under way as the provision of public goods drops in line with revenues. As capital and labour become more mobile, dollars begin to count for more than votes-remember these are elected governments-and democracy is defeated.

Certainly, on a variety of measures, the world economy has become far more integrated in the past few decades than it used to be. Trade is one such yardstick. In America the share of exports in national income has nearly doubled, from 4% in the 1950s to 7% in the 1990s. Since the United States is a continental trading zone in its own right, which suggests a relatively low share of international trade, that rise is already impressive. Arguably, though, a better measure for judging how far foreign trade i s impinging on American producers is the share of merchandise exports in the output of manufactured goods. This more revealing ratio, over the same period, rose from 6% to nearly 20%.

A frictionless world?

Europe's producers are even more exposed to foreign competition-not just from outside the European Union, but from elsewhere in Europe as well. A single global market for tradable goods is by no means a reality yet. If it were, the prices of goods in diff erent countries, expressed in a common currency, would be the same. They are not. Even so, the world has moved far and fast in this direction of late (see chart 6, next page).

Integration through flows of capital has increased even faster than integration through trade in goods. The expansion of international finance has been phenomenal, and all the more striking for taking place over a relatively short period. International transactions in bonds and equities and daily turnover on the foreign-exchange market have both increased at an astonishing rate over the past 20 years (see chart 7). It has become a well-known fact that daily turnover on the currency markets now often ex ceeds the global stock of official foreign-exchange reserves-so what chance have central banks of influencing exchange rates by buying or selling currency in the markets? Flows of foreign direct investment have also increased rapidly, though nothing like as rapidly as transactions in currency and securities (see chart 8, next page).

Remarkable as these growth rates are, they need to be put in perspective. Even though vast sums of money are hurtling round the globe, there is still no such thing as a single global market for capital. The most obvious evidence is that, within a margi n of two or three percentage points either way, countries tend to invest about as much as they save. If there were a fully integrated global market for capital, this very strong correlation would not persist the way it has. Poor countries that are short o f physical capital and offer many opportunities for profitable investment would tend to invest far more than they save, and rich ones would tend to do the opposite. Poor countries, therefore, would generally run big current-account deficits and rich ones big current-account surpluses. If there were a single global market for capital, international financial imbalances would be much bigger than they are.

The capital market, it turns out, is integrated only in a narrower sense-in the sense that ``arbitrage opportunities'' are quickly competed away. For instance, the rate of interest on a safe dollar-denominated bond of a given maturity is the same in th e international markets as the rate of interest on a safe sterling-denominated bond of the same maturity after allowing for the cost of ``cover'' (insurance) against the exchange-rate risk. If this were not so-and in the old days of capital controls it wa s not-then there would be a riskless opportunity to make money (in the jargon, an arbitrage opportunity). This level of integration has been achieved mainly thanks to the abolition of capital controls, together with advances in financial technology and t elecommunications which have greatly reduced the costs of such transactions.

But integration at deeper levels has yet to materialise. The markets do not regard assets denominated in different currencies as perfect substitutes for each other: differences in interest rates and differences in expected currency changes do not exact ly cancel out. Most important, there is no single world real rate of interest. This is partly because different currencies are not perfect substitutes and partly because, as noted earlier, the price of goods in a common currency also varies from country t o country. The claim that ``these days, when you want to borrow money, you have to pay the global market rate'' is in fact false. There is no global market rate, even among countries that have fully liberalised their capital flows.

To say that the world, as yet, is far from having truly single markets for goods and capital is not to deny that integration is increasing, or that economic policy has been affected by it. The next section will ask what the change so far means for gover nment. Before that, however, it may be helpful to think a little more about the principles involved.

Slippery Utopia

Consider the limiting case of a fully integrated world economy, with a single currency and a single market for goods. Suppose that capital is completely mobile, and that labour too moves freely across international borders. Imagine, in other words, a degr ee of integration even greater than within a national economy.

Liberals are often said to have a desiccated view of humanity-to see men as rootless, atomistic individuals floating free of any ties to society. For once, let that be true. Imagine not only that all regulatory barriers have been removed (for that would not come close to making capital and labour perfectly mobile), but picture a world without ties of history, language, culture and kinship, in which it is costless for people and objects to move around, and in which capital remains malleable even after it has been applied to production (so that a sewing machine in Bombay one moment becomes a PC in Atlanta the next). In short, imagine a world in which there is no hindrance of any kind to the mobility of people or their assets.

Recall the pessimists' argument: integration undermines the capacity of governments to govern. Under these imaginary circumstances, therefore, it would surely be impossible for the state to collect taxes and supply public goods? Not at all. People, as workers, consumers or owners of capital, do value public goods. That is the only reason why it makes sense to provide such goods in the first place. When the choice is between having them and not having them, people are willing to pay. The difficulty aris es only when there is a third choice: having them without paying for them. In a closed economy, because of this free-rider problem, it is impossible to discover how much people are willing to pay. In a perfectly open economy, this information would be rea dily available.

People would vote with their feet, choosing where to live, or where to invest, according to the value for money that each government provided. That decision in turn would reflect not only each state's technical efficiency (eg, its ability to build a b etter road at any given cost) but also people's individual preferences (eg, whether they preferred roads to open space, extra policemen to extra schoolteachers, or lower taxes to any of the above).

Yet on reflection it seems the pessimists' case is proven. In this peculiar world, there might be pure public goods, but there could be no question of providing generous transfers to the poor. Thanks to the assumption that mobility is costless, any gov ernment that did much to help the poor would attract to itself the destitute of the world. Its taxes would be driven higher and its income-generating taxpayers would be off.

But make this fantasy fractionally more realistic. Assume that governments in countries with lots of rich workers restrict the immigration of poor ones, and/or that poor people cannot afford to move. Further assume, atomistic individuals notwithstandin g, that most people have some altruistic regard for their neighbours. On those assumptions, even if capital and labour were still perfectly free to run away, a moderate degree of redistribution could, and probably would, still take place. More than in the real world, perhaps, some people would presumably flee to enclaves of the rich and selfish; but most people with some altruistic instinct might well continue to prefer living close to others of like mind.

Admittedly, much would depend on whether egalitarians are right that greater equality of income is indeed a kind of public good. If it delivers benefits to the whole society-less crime, a happier and more relaxed society-then, as with other public good s, many taxpayers will be willing to pay for it, so long as the price is not too high.

If equality is not a public good in this sense, then the rationale for public redistribution (as opposed to private, voluntary, benevolence) can only be some variant of the doctrine that the income produced by the members of a society belongs not to the m individually but to society at large, which is then within its rights in sharing it out as it sees fit. In other words, it is the People's Money, not the people's money.

Many egalitarians would indeed argue exactly that, and some modern liberal philosophers are sympathetic. They would find a frictionless world-in which people with the capacity to earn a living were not indentured to their governments-an inconvenient on e. But that is their problem.

Big government will come to terms with the global economy

Still in command

EVEN in that barely imaginable world in which capital and skilled labour could be instantly transported and transformed around the globe, taxes, as well as death, would still be certain. Progressive taxes, too, at a pinch. The real world, of course, is no t remotely like that. Real-life labour and capital are quite tied down, and likely to remain so. Financial capital has become extremely mobile, but the physical kind is not. This distinction is often overlooked. Once a multi-billion-dollar microchip-assem bly plant has been built, it tends to stay put, even if government decides to tighten the tax screw a little. Labour, even of the highly skilled, jet-setting, multilingual kind, is less mobile than often supposed. Most people are anchored down by culture, family and neighbourhood. Governments reinforce these ties by regulating immigration (albeit with the aim of stopping outsiders coming in, not residents moving out). Even within the European Union, where official barriers to the movement of labour have b een removed, relatively few people move from country to country.

This is not to say that freer flows of capital and labour are having no effect. It may be true that in a world of perfectly mobile capital and labour there would still be taxes and public goods; it may also be true that in the real world, where capital a nd labour move more reluctantly, the scope for taxation and public spending is greater still. Nonetheless, increasing mobility of capital and labour could be having an effect-maybe not destroying the state's power to tax and spend, but significantly reduc ing it. In practice, is this happening?

Not really. Certainly it makes sense these days for revenue-seeking governments to shift the pattern of taxation. The basic principle of efficient ``fiscal ranching'' is obvious: tax immobile factors of production more heavily than mobile ones. That me ans going easy on taxing the rich; given sufficient reason, they can get away more readily than the less rich. Also, where possible, tax the income generated by ``sunk'' capital, but offer tax breaks for new investment, so as to attract footloose financia l capital. More broadly, since capital of all kinds is more mobile than labour, shift the overall balance of taxation so that it weighs more heavily on workers and less on owners of capital.

By and large this has been happening, though some governments have shown more imagination than others. Ireland, for instance, is a tax-free zone for ``artists'', which makes it an attractive home for best-selling authors and other rich creative types.

The tax authorities seem less inclined to offer special deals to, say, window-cleaners or postmen. Ireland has also been able to establish an extremely low-tax regime for inward direct investment in manufacturing and ``internationally traded services'', d espite EU rules to discourage this sort of thing. The government says it applies these rules equally to new investment and existing operations, thus violating one of the maxims of efficient ``ranching''; but since Ireland embarked on this policy at a tim e when it was a farm economy with hardly any incumbent manufacturing or traded-services companies, the principle was upheld at least in the spirit, if not the letter.

Ireland is a cheeky extreme, but corporate tax rates have been coming down across the advanced industrial economies and beyond. The fall is not particularly dramatic, but neither is it insignificant. On average, the standard rate of corporate-income t ax in the OECD countries fell from 43% in 1986 to 33% in 1995; the overall yield dropped by less, because cuts in the tax rates were offset by measures to reduce exemptions. The highest marginal rates of personal-income tax have come down as well: in the OECD, the average top rate was 59% in 1975 and 42% in 1995. But rates lower down the scale have been raised. The marginal tax rate for workers on two-thirds of average earnings was 32.6% in 1978 and 38.4% in 1995. The marginal tax rates for workers on ave rage earnings and twice average earnings also went up (see table 9). In addition, governments have increased taxes on consumption, such as value-added tax (VAT).

Overall, therefore, despite the gale of global market forces, the aggregate tax burden continues to rise. Taxes in the OECD were 34% of GDP in 1980, 36.3% in 1990 and 37.4% last year. Taxes in the European Union were 41.3% of GDP in 1980, 43.8% in 1990 and 45.4% last year (see chart 10). Citizens, it seems, need not worry just yet that their governments' ability to tax them in a good cause is going to be much infringed. It will take more than today's global market to stop them.

A hand on the tiller

So much for taxes and spending. What about the effect of capital-market integration on macroeconomic policy? Undeniably, the deregulation of capital flows and continuing advances in information technology have altered the way that changes in budget defici ts and interest rates, the two levers of macroeconomic policy, affect the economy. In some ways they have made it easier for governments to smooth out fluctuations in the economic cycle (which is the most that macroeconomic policy can hope to do); in othe r ways they have made it more difficult. But it is wrong to claim, as many do, that the new global market for capital has made it impossible for governments to discharge their responsibilities as macroeconomic managers, or even that governments' freedom, in some overall sense, has been seriously infringed.

Macroeconomic policy must yield to certain fundamental constraints, regardless of whether capital moves freely across borders. One such constraint is that when a government borrows, voluntary lenders must be compensated both for delaying their spending and for taking a financial risk. Beyond a certain point, if a government borrows more, this compensation (or interest rate) will have to rise. Another constraint is that any persistent expansion of the money supply will push up prices. Even in a world wi thout capital mobility, a government that tries to escape or ignore these constraints will come a cropper. The same is true in a world with capital mobility. For the most part, what changes is the come-uppance: whether it arrives this year or next year, w hether it manifests itself as a gathering disillusionment or a sudden crisis.

Table 11 (next page) summarises the implications of increased government borrowing (ie, a relaxation of fiscal policy) and a faster expansion of the money supply (a relaxation of monetary policy) in three distinct economic regimes. In the first, econom ies are closed to both trade and capital; in the second and third, they are open to both, but in one case the government fixes the exchange rate and in the other it lets it float (a choice that in a closed economy does not arise). The table shows that onc e the economy is opened up, the rules of macroeconomic policy do change. Broadly speaking, if the government chooses to fix the exchange rate, it will need to rely less on monetary policy and more on fiscal policy, as compared with the closed-economy case . If it chooses to let the currency float, the opposite is true: it will need to rely less on fiscal policy and more on monetary policy.

In only one substantial respect have free flows of capital narrowed governments' choices over economic policy. Maintaining a semi-fixed exchange-rate regime-an awkward hybrid, in which governments attempt to peg the exchange rate without surrendering m onetary policy exclusively to that purpose-turns out to be impossible in a world of mobile capital. That is the lesson of the most spectacular economic-policy upsets of recent times: the EMS, the Mexican peso, the Thai baht. With strong capital controls, it was sometimes possible to share attempts at stabilisation between fiscal and monetary policy. With weak controls or none at all, a sharper choice must be made: fix the currency and use fiscal policy for stabilisation, or let it float and use monetary p olicy to that end.

Discretion preserved

Important as it is to get this right, it hardly seems a huge infringement of economic sovereignty. Moreover, if the government applies that division of labour, it will find that in some ways its macroeconomic control is enhanced, not diminished, compared with what would happen in a closed economy. For instance, governments can borrow more in an open economy without raising interest rates and ``crowding out'' private investment. Alternatively, given floating exchange rates, monetary policy works more power fully because the demand-expanding effects of lower interest rates are boosted by the demand-expanding effects of a depreciating currency. True, if the government pushes its borrowing too far, the international markets may suddenly demand much higher inte rest rates-but in a closed economy, domestic investors would have been demanding higher interest rates much sooner. And if the government persistently expands the money supply by more than the growing economy requires, the result will be higher inflation with or without internationally mobile capital.

The view that capital mobility has left governments with about as much discretion in macroeconomic policy as they ever had is not universally shared. Some like to argue that governments can no longer keep interest rates as low as they might wish. With capital controls, governments can make credit cheap by administrative fiat. If capital is free to leave, this becomes impossible: savers' money will flow abroad in search of a better return. Low interest rates encourage investment, it is further argued, a llowing the economy to grow faster than it otherwise would. In fact, experience has shown this policy to be counter-productive: a cap on interest rates reduces saving. The demand for investment will rise, but a lack of funds will prevent it from being met . Yet if the government insisted on this approach, it would apparently need capital controls to do it. So capital mobility, it seems, has indeed infringed the government's powers.

Or has it? The policy of financial repression, as it is known, is similar in principle to taxing savings and using the proceeds to subsidise investment. Self-defeating as it may be to distort economic decisions in this way, if governments want to do it, they can, with or without capital controls. Without the controls, hitherto-implicit tax-and-subsidy arrangements become much plainer to see, perhaps obliging the government to explain what it is up to. That might be inconvenient, but it seems irrational to attack capital flows for making governments more accountable for their actions.

It is also an exaggeration to say that the financial markets' judgment of bad policy is now more brutal than it used to be. Britain's forced devaluation of sterling in 1967, after a period of inflation and overborrowing, was at least as traumatic as th e currency's ejection from the European monetary system in 1992, after a period of inflation and overborrowing. That earlier humiliation happened despite strict controls on capital flows and far smaller financial markets. As many developing countries cont inue to prove, economies that are open to trade but not to capital flows are about as prone to exchange-rate troubles as economies that are open to both. Here, as elsewhere, the fundamental rules apply.

Thanks to greater capital mobility, macroeconomic policy is certainly more complicated than it was before, as the table indicates. This allows more scope for mistakes, and when mistakes are made they will sometimes (only sometimes) be punished more qui ckly. But this is much like saying that if the bank extends your overdraft you have more choices than before, and therefore more scope for making mistakes; and that if you do make a mistake, the consequences are likely to be more embarrassing. When a bank extends your overdraft, it does not reduce your ability to run your own financial affairs.

For the time being, then, the state's powers to tax, spend and borrow seem under no serious threat: on the face of it, quite the opposite. But bear in mind that globalisation may have much further to go. On several economic measures-such as the share of foreign capital in domestic investment, cross-border flows of investment in relation to national output, flows of people in relation to population-the world was more closely integrated before 1914 than it is now, in some cases much more so.

In politics, too, the world was in some ways more integrated before the first world war. Ernest Gellner, writing in the early 1980s about nationalism, asked readers to imagine two maps of the world, one pre-industrial, one modern:The first map resemble s a painting by Kokoschka. The riot of diverse points of colour is such that no clear pattern can be discerned in any detail, though the picture as a whole does have one. A great diversity and plurality and complexity characterises all distinct parts of t he whole . . . When it comes to painting the political system, the complexity is not less great than in the sphere of culture . . . Look now instead at the ethnographic and political map of an area of the modern world. It resembles not Kokoschka but, say, Modigliani. There is very little shading; neat flat surfaces are clearly separated from each other, it is generally plain where one begins and another ends, and there is little if any ambiguity or overlap . . . We see that an overwhelming part of politic al authority has been concentrated in the hands of one kind of institution, a reasonably large and well-centralised state . . . When we look at the society controlled by this kind of state, we also see why this must be so. Its economy depends on mobility and communication between individuals, at a level which can only be achieved if those individuals have been socialised into . . . the same high culture.

This last point-that the demands of economics shape both culture and politics-suggests an interesting possibility. Perhaps the history of the nation state has come to a turning point after all. In future, it might be argued, it is the need for ``mobility and communication'' between economies, rather than within them, that will gain the upper hand. This demand will give rise to an increasingly homogeneous global culture, just as its predecessor gave rise to homogeneous national ones. This in turn will not only promote economic integration but will also, however slowly, blur the political boundaries between nations. In future the contrasts among the neat Modigliani surfaces will begin to diminish, and the colours start to merge.

Lately the world has seen at least as much fracturing as merging-witness ex-Yugoslavia and the former Soviet Union. But these events could be construed as a necessary if painful precondition for the continuing advance of today's dominant global culture , which is capitalism. And what is the European Union-or even NAFTA, APEC, Mercosur and the rest-if not an attempt by governments to reach a joint accommodation with international capitalism on terms they find acceptable? The EU in particular, and the oth er regional trade agreements too (albeit to a much smaller extent), are instruments of political as well as economic integration.

In the shorter term, and once again assuming that governments are interested mainly in preserving their power, these regional trade groupings serve an obvious purpose. This survey argued earlier that as labour and capital become more mobile, government s will be drawn into competition with each other-and that, other things being equal, the effects of this competition will be to roll back the state. As the figures on taxes and spending show, at present this is still little more than a gleam in free-marke teers' eyes. Still, prudent governments will be doing what they can to stop it from becoming more. And what do firms threatened with too much competition try to do-especially if there are only a few of them, and one is willing to take the lead? Set up a c artel, of course.

Government by cartel

The official name for this alternative to competition among governments is ``harmonisation''. Its main political advantage is that, unlike competition, it keeps governments big. In its early days the EU was dedicated to the liberal principle of competitio n among governments: the rule of thumb was ``mutual recognition'' of each other's different standards and regulations. In time, this approach yielded to harmonisation as the preferred way to build the single European market, gradually encroaching upon eve r more areas of national economic policy. In a similar way, the United States has developed a preference for regional trade agreements such as NAFTA over the traditional multilateral approach to trade liberalisation, as carried on by the GATT and its succ essor, the WTO. These deals have allowed it to negotiate side-agreements on labour and environmental standards, and thereby bring its trading partners' policies more closely into line with its own. Now calls are mounting for harmonisation of standards to be extended into the WTO itself.

Whatever the future of supranational government, this trend has a paradoxical aspect. On the face of it, governments surrender ``sovereignty'' when they take part in such agreements, which makes them controversial everywhere. But, as explained above, a different interpretation is equally possible: that governments are pooling power in order to retain and increase it, in just the same way that a firm in a cartel gives up the freedom to sell all it could in order to gain a share in the group's fatter mon opoly profits. So far the state's freedom of action has barely been touched by the global market; should it become more circumscribed, expect more rule by cartel.

In a market economy, not everyone prospers

Ballad of the global worker

AS AMERICA'S continuing debates over NAFTA, trade with China, trade with Japan and trade with the EU show, trade policy is the most politically sensitive of all the issues raised by globalisation. The effects of international market forces on taxes, publi c spending, financial markets and supranational governance are all rather subtle compared with their effect (at least as people perceive it) on jobs and wages. And if globalisation is tending, however slightly, to reduce the capacity of the state when it comes to taxes and spending, it is enlarging the demands on the state when it comes to the labour market. That is why some argue that capitalism is heading for a crisis similar to that of the 1930s: the gap between what it can do and what it must do is gr owing intolerably wide.

Undoubtedly, the advanced economies have an urgent and seemingly intractable problem on their hands: what to do about their unskilled workers. In America the problem takes the form of poverty wages, in Europe of unemployment. The underlying cause appea rs to be the same. There are more unskilled workers than jobs at decent wages for them to do. In America, wages are allowed to fall far enough to match the demand for unskilled labour with the supply. Unemployment is low and millions of new jobs have been created; but low-wage workers earn too little to live comfortably, and their incomes in real terms have fallen for 20 years. In Europe, wages and other employment costs are buoyed by minimum-wage laws and comparatively generous welfare benefits, so unski lled workers are less poor than in America; but unemployment is far higher.

Trade is widely blamed, especially in America, for this downward pressure on the employment prospects of the unskilled. How can workers in the West compete with those in the third world who are paid a tenth of rich-country wages and who, these days, ar e using rich-country capital and technology? And yet, in this form, fear of trade is quite misguided. Across the world, the relationship between pay and productivity is very strong (see chart 12). Developing-country workers who are paid a tenth of rich-co untry wages usually turn out to be about a tenth as productive as well. As their productivity rises, their pay will go up too. So the trade problem, if there is one, has nothing to do with disparities between pay and productivity-that is, with competitive ness-in North or South.

But another possible connection between trade and wages needs to be taken more seriously. The developing countries' comparative advantage is in unskilled labour; America and Europe, relatively speaking, are much better supplied with skilled workers and capital. The effect of imports from poor countries is therefore akin to expanding the supply of unskilled labour in the West, which causes its price there to fall. If this is so, then trade with the developing world may make the advanced economies as a w hole better off, but at the expense of lower pay for the unskilled.

There is nothing wrong with this argument in principle. Note that the case for liberal trade does not depend on its being false. If low-cost imports had indeed driven the wages of the unskilled lower, that would be nowhere near enough to establish the case for increased trade protection-any more than finding that technological progress had widened the gap between skilled and unskilled workers would make it right to ban technological progress. Raising barriers to imports, which would make both rich coun tries and poor countries worse off in the aggregate, would be an insanely costly way to deal with the problem of the low-paid.

Simple but untrue

Surprisingly, the consensus among the economists who have examined the matter is that, plausible as the simple trade-and-wages story sounds, it is not in fact true. The reason, as Jagdish Bhagwati of Columbia University was the first to point out, is that a crucial link in the supposed series of events is missing. If low-cost imports had been the cause of falling wages for the unskilled, those imports must first have lowered the relative price of the competing American-made goods. That is the channel thro ugh which, according to the plausible theory just outlined, trade would drive wages down. But, as several studies have found, the relative prices of those goods have not in fact fallen.

So if there is a connection between trade and rising wage inequality in the United States, it must work in some other way. The issue has suffered no lack of scrutiny in recent years, but unhelpfully the range of estimates for the effect of trade on wag e inequality now runs from almost none at the low end to 100% at the other.

Adrian Wood of the University of Sussex argues for the upper extreme. He says that import competition has already driven many low-wage activities out of existence, so checking the prices of goods made in the North to see whether they are lower than the y used to be rather misses the point. He also points out that import competition has spurred the introduction of labour-saving technology, another effect missed by the conventional theory.

Those who go for the lower end of the spectrum argue that trade between the United States and the third world remains very small in relation to America's overall output (which is the relevant measure for judging trends in national wages). They criticis e trade-and-wage pessimists for making it sound as though the United States and other rich countries were small economies, opening up to a new world of trading opportunities with the South. The opposite is closer to the truth: the United States is the eco nomic giant, and the South, even in the aggregate, too small to have much effect.

Reviewing the many different studies, Bill Cline of the Institute for International Economics in Washington, DC, finds that many estimates of the worsening effect of trade on American wage inequality cluster around 10-20%. His own new calculations put the figure much higher, at 50%; but in tossing this figure into the ring, Mr Cline makes a simple and important point that other observers seem to have neglected up to now.

All these estimates measure the effect of trade against the change in American wage inequality. But that change is itself the net outcome of forces pushing in opposite directions: forces tending to increase inequality, such as trade, immigration, the r elative decline of the minimum wage, weaker trade unions, technological change and others; and forces tending to reduce inequality, notably the dramatically expanded supply of skilled workers. As unskilled workers become a smaller proportion of the workfo rce, their relative wage should rise to reflect their scarcity. Clearly, if the effect of trade is to be seen in its proper perspective, it should be measured not against the net change in equality, but against the gross disequalising forces.

Seen this way, even Mr Cline's bigger-than-average estimate of the effect of trade shrinks towards insignificance. His figure of 50% for the net effect of trade on inequality falls to just one-eighth when measured instead against all disequalising forc es taken together. And the middle estimate of the earlier studies-that trade explains 10-20% of the net change-turns into an effect of just 2.5-5%.

All the same, the list of possible trade-and-labour linkages continues. Many non-economists, for instance, are convinced that international competition will cause a different kind of race to the bottom, driving down labour standards such as rules on wo rking hours, safety in the workplace, trade-union recognition and so on. Yet this need not happen. To think it must is to accept a variant of the competitiveness fallacy discussed earlier.

From the employer's point of view, labour standards are just another labour cost. They can be afforded just as higher wages in the North can be afforded-that is, to the extent that they are justified by higher productivity. When a government enacts a n ew labour standard, it is in effect calling for other labour costs to be lowered enough to offset the extra burden. This could happen in several ways: lower wages, a depreciating currency, or subsidies to employers. Trade, or the economy's degree of openn ess, makes no difference.

Dani Rodrik of Harvard has recently questioned this argument, albeit in a rather narrow way. He maintains that trade exposes firms to more competition, which in turn makes them more sensitive to price, which in turn makes their demand for labour more s ensitive to wages and other labour costs. This greater sensitivity-an increase in the elasticity of demand for labour-weakens workers' bargaining power. So when workers and employers come to divide any monopoly rents that may exist in the business (becaus e, say, workers have accumulated job-specific skills), labour will get less and the owners more. Or if you think of employment standards as a tax on labour, trade will shift the incidence of the tax a little way towards workers rather than employers. Thro ugh its effect on elasticity, trade will also increase the volatility of employment, meaning more rapid turnover of jobs (a factor that other researchers have tied to rising inequality). As for how much any of this matters in practice, your guess is a goo d as Mr Rodrik's.

The machine that changed the world

Enough. The debate comes down to this: the popular arguments that say trade destroys jobs, lowers wages and forces down labour standards are economically illiterate. But there are subtler arguments to show that trade is likely to have some effect on the r elative wages of the unskilled. According to most estimates, these effects are small compared with the other forces that have been driving the wages of the unskilled lower: a reasonable top limit for the effect of trade might be one-eighth of the total. O ther forces, notably advancing technology, are likely to play a much larger role.

Yet, after all that, it really does not matter whether it is trade or technology that is pressing down on the wages of the unskilled. In every important respect, trade is just another kind of technology. Think of it as a machine that adds value to input s. In America, trade is the machine that turns computer software, which America makes very well, into video recorders, basketballs and other things which it also wants, but does not make quite so well. Trade does this at a net gain to the economy as a who le. If somebody invented a contraption that could do this, it would be regarded as a miracle. Fortunately, trade has already been invented.

In the end, what gives the lie to protectionist proposals on trade, and to the academic studies that strain so hard to lend them respectability, is that there is no corresponding zeal to bring other kinds of technology to heel. This difference is purel y one of politics, not economics. If trade is squeezing the wages of the unskilled, so are other sorts of advancing technology, only much more so. Yes, you might say, but to tax technological progress or put restrictions on labour-saving investment would be ridiculous: that would only make everybody worse off. Indeed it would, and exactly the same goes for trade-whether this superior technology is taxed (through tariffs) or over-regulated (in the form of international efforts to harmonise standards).

The only difference is that with trade, unlike other kinds of technology, it is just about possible to pretend that the only losers from restrictions will be foreigners. That fallacy was exploded a couple of centuries ago; but in politics its appeal r emains undimmed.

Smaller government need not hurt the weak

Beyond the welfare state

FROM the Platonic guardian's point of view, it hardly matters whether it is more trade or better technology that is putting a strain on the labour market: both are good for national income, and both have much the same bad side-effects. Real-life governmen ts usually take it for granted that technological advance is a good thing, but make a big fuss about trade and globalisation-far bigger than a wise and disinterested government would. This in itself is strong proof that governments should not be assumed t o be Platonic guardians.

For the unskilled workers who have lost their jobs or suffered cuts in real pay over the years because of trade or technology, the distinction between the two causes matters even less. What counts for them is simply that they are falling behind even as most others make faster progress. Market liberals tend to justify free trade and technological progress by pointing out that the increase in aggregate income they generate is big enough to compensate the losers and still leave society as a whole better o ff. That is a good start, but hardly conclusive if, in practice, compensation never comes. So what, if anything, should governments do?

It may be instructive to look at what actually happens in countries with small and big government respectively, and compare results. Relatively small government, as in the United States, seems to promote high overall living standards and low unemployme nt, but at the cost of considerable poverty and related social ills (notably crime). In Europe's economies, thanks to more generous transfers, the problem of poverty seems less acute, yet unemployment is much worse, bringing social ills of its own.

Conceivably, the problem will bring about its own solution if market forces are allowed to take their course. As wages of skilled workers rise, people will invest more time, money and effort in acquiring skills, reducing the supply of unskilled labour and thereby raising its wage. Also, as skilled labour becomes more expensive relative to unskilled, employers will have more of an incentive to hire the unskilled. More investment will then be directed to economising on skilled rather than unskilled worke rs, which again will tend to raise the relative wage of the unskilled. In some industries this appears to be happening already.

Critics say that this simply spreads the problem of unemployability more widely, but that confuses two issues. There is no problem of ``unemployability'' as such; there is a problem of low productivity among the unskilled. Economic growth can leave the unskilled unable to earn a decent living in the market. Technology would worsen ``unemployability'' across the economy as a whole only if the demand for output were fixed. If it were, rising productivity would mean that fewer workers were needed-not just in particular industries, but across all industries. In fact, history suggests that technology raises both supply and demand without limit.

So if the pattern of investment changed towards economising on skilled rather than unskilled labour, the end result would be higher average wages and a bigger share of that increase for the unskilled, all at unchanged overall levels of unemployment. Un fortunately America's experience suggests that the demand for unskilled labour, relative to its supply, will need to increase much faster than it has done for the past two decades. Automatic adjustment in the market is unlikely to be of much help to the a rmy of unskilled.

Don't subsidise, educate

If the market fails to come to the rescue, can the case for small government still be reconciled with a civilised society's wish to improve the lot of its least fortunate? The answer is yes. Two broad strategies suggest themselves, both consistent with li beral principles. One is better education, which would raise the productivity of some of the unskilled and increase the scarcity of the rest, raising incomes in both cases. The other is lower taxes for those at the bottom of the income distribution. There is much to be said for radical thinking under both headings.

In most countries calls for better education are a rallying-cry of the left: a way of asking for more public spending on education. Yet, as noted earlier, there is no evidence that extra spending produces better results. State-run education systems see m to suffer many of the problems experienced in other state-owned, worker-managed industries, notably lack of competition and ``capture'' by the workers. Where private and state-run systems operate alongside each other, the contrast in both these respects is plain. The difference is more to do with methods and attitudes than with physical resources.

Therefore, at the very least, smaller government and better education are by no means inconsistent. Supposing that governments continued to pay for education, standards could probably be widely raised by moving to a more competitive form of delivery-sa y by giving parents purchasing power by means of vouchers in a market for education. But there is a good case for going further and sharply reducing both public spending on education and the taxes needed to pay for it. If parents had to spend their own mo ney on their children's education, they would demand correspondingly more from their schools. Thus motivated, they would be likely to get better results than governments.

Hold on, you might think; what about the poor? At least state education ensures that they do not lapse into illiteracy. Even if private education meant better schools, it would do nothing to improve the prospects of the least well-off. This is true (t hough not as self-evident as it seems, remembering how much of he state's existing spending on education and most other services is channelled towards the non-poor). But the answer in any case is not to maintain education as a nationalised industry, any m ore than it makes sense to nationalise the bread or the housing industry to ensure that the poor have food and shelter. The answer is to give the poor purchasing power in the market for education, again by means of vouchers.

Since education below university level has fairly strong public-good characteristics, even competing governments in the Utopia of perfectly mobile people and capital would subsidise it, and help the poor to buy more. The case for heavy public spending on universities, in contrast, seems weak. Since most of the benefits from such education flow to the student rather than to society at large, it seems right to make the student bear the costs, at the time or later-which will encourage the student to cons ider those costs when choosing a course.

Education is one of the biggest public-spending programmes in most advanced economies. It accounts for 5.1% of GDP in America, 4.7% in Britain, and an average of 5.4% in Europe. Returning it to the private sector would therefore allow taxes to be cut s ubstantially. But education is only a start. Other expensive programmes could also be severely pruned-again on the principle that it is better for citizens to exercise their own purchasing power, not have governments do it for them. In most advanced econo mies, health care and pensions would be the next biggest targets of any determined effort to roll back government.

In health care, as in education, the aim would again be to give taxpayers their money back and let them buy for themselves in the market what they need. In some respects, admittedly, health is a more difficult and complicated issue than education. Peop le have to rely on doctors to tell them what medical services they need, so it is harder for them to act as informed buyers. Health care is also an insurance product. But are governments demonstrably better at dealing with such problems than markets? Is a government doctor more likely to be trustworthy, and to keep the patient's interests uppermost, than a private one?

America's mostly private system of health care is certainly not without its flaws. Until recently, it has proved inordinately expensive-though providing in return the lavish standards of medical attention that Americans seem to want. Britain's state-ru n National Health Service is comparatively cheap, but not exactly patient-oriented. If you are in only moderate pain and in no danger of expiring, then you will generally have to wait, even though the treatment would greatly improve your life.

Given the difficulty of finding the right balance of inputs and outputs, whatever the system, the case for privatising health care may turn less on its effects on the quantity and quality of health care and more on the benefits of lower taxes. If it ac hieved nothing else, greater private provision (allowing, as in education, for measures to give the poor purchasing power in the health-care market) would allow tax rates to fall.

Perhaps the best opportunity of all to cut public spending and taxes is pensions. The cost of taxpayer-financed Social Security in the United States is more than $300 billion a year, equivalent to 5% of GDP. In some European countries, the cost of state pensions is more than 10% of GDP (see chart 13) America's system, like those in most other advanced economies, works on a ``pay-as-you-go'' basis, meaning that the contributions citizens pay are treated as general revenue. Instead of being saved and inve sted, as they would be if the citizens were saving for their own retirement, these taxes are applied like any others to pay for overall government spending. This is true in nearly all rich countries, despite the various accounting fictions of trust funds, surpluses, deficits and whatnot. For taxpayers, this way of providing for pensions is bad news.

The industrial countries' populations are ageing, which means the ratio of retired people to workers is rising. In future, either taxes will have to go up to provide the same real level of pensions, or else pensions will have to be cut. In either case, people paying social-security taxes now can expect to receive a very low implicit return on their ``saving''. Martin Feldstein, president of America's National Bureau of Economic Research, calculates that the implicit real rate of return on social-securi ty taxes has fallen to 1.5% a year. Historically, the pre-tax real rate of return on privately invested capital in America has been 9% a year. (If that seems high, remember that the 5-6% return usually quoted is net of corporate-profits tax). In other wor ds, social-security taxes could be replaced by much smaller personal saving, so long as this added to national saving, and hence to the economy's stock of capital.

Pensions are different from the other big-ticket items of welfare-state spending. In much of Europe at least, a sudden and deliberate privatisation of health care may seem unthinkable just now; in both Europe and America, privatisation of primary and s econdary education also seems out of the question (though university education is already substantially private in America, and Britain is set to make a first small move that way). But public pensions provision is almost certainly going to be dismantled-t he only question is when and exactly how. This appears to be widely understood, and before long may well command wide acceptance.

But even if public pension schemes are privatised promptly, lower taxes will follow only after a decade or two. The switch to private funded schemes, with everybody saving for their own retirement and the poor again helped to do so through public subsi dy, will take years. During the transition, people retiring without a private fund will still need to receive a state pension. That cost will have to be met out of taxes, as it is now. There is also a risk that, when the tax savings materialise later on, governments will grab them and spend them on something else.

A smaller role for the state in education, health and pensions would lead to greater efficiency and, in one way or another, make the typical citizen better off. But what about the unskilled or unproductive citizen? Improving his purchasing power in the market by means of vouchers or other subsidies would, some argue, at best leave him no worse off than now. This is wrong, on two counts. First, the same amount of money as spent by the government now would buy better results, thanks to competition among p roviders. Second, much lower public spending means much lower taxes, offering scope for greatly improving the earnings prospects of the unskilled.

Good for the rich, good for the poor

It is a misconception that the poor have little to gain from tax cuts, even leaving aside the trickle-down effects of growth. In Europe, and to a smaller extent in the United States, low-income workers do fall within the tax net, paying amounts that are s mall in absolute terms but large in relation to their income, in or out of work. If they work, they may well pay both income taxes and social-security taxes, and their employers are also likely to be paying a payroll tax. The burden of all these taxes com bined forms a wedge between the already low value of the worker's output and the income he receives.

Suppose for one insane moment that public spending could indeed be radically cut, from roughly half of the economy in Europe and roughly a third of it in the United States to, say, a quarter-that is, to about the size of the public sector in America in 1960. That would mean labour taxes of all kinds could be eliminated completely for low-wage workers and deeply reduced for everybody else.

Helpful though that would be, it might still do too little to improve the earnings prospects of unskilled workers. Although low-wage workers do pay labour taxes of several kinds, most of the benefit of big tax cuts in this area unavoidably goes to the be tter off. Also, supposing vouchers were used to give low-wage workers buying power in the new markets for education, health care and pensions, the beneficiaries would face an implicit tax as their income went up and the value of the vouchers was reduced. So there might be a need to do more. In this scenario, however, there would be no lack of resources with which to do it.

In a new book (see box), Edmund Phelps of Columbia University has proposed a radical scheme to subsidise low-wage labour in the United States. In the 1960s Mr Phelps and Milton Friedman, working independently, were the first to postulate the idea of a n atural rate of unemployment. Mr Phelps's new work is in the same tradition of market (``neoclassical'') economics. His argument for big subsidies to the low-paid is consistent not just with respect for the power of market forces but also, and more broadly , with a classical liberal view of the world. The idea is particularly striking because it shows how to reconcile a rigorously ``market-friendly'' approach to policy with effective intervention by the state in pursuit of social goals.

A big idea for smaller government

Under Mr Phelps's plan, firms would be paid a graduated subsidy for employing low-wage labour: $3 an hour for each worker employed at a wage cost (before the subsidy) of $4 an hour. Gradually, as the pre-subsidy wage went up, the subsidy would taper away, until at a pre-subsidy wage of about $12 an hour it disappeared. These payments would not raise wages one-for-one; nor, despite being paid to firms rather than workers, would they stop wages from rising at all. The subsidy, just like a tax, would be shar ed between employers and workers according to conditions of supply and demand in the labour market. Nonetheless, in some combination, wages for the low-paid would be bid up and unemployment among them would be reduced.

Because Mr Phelps's proposed subsidy is very large, amounting to at least 75% of the pre-subsidy wage cost for the lowest-paid workers, the effects would be correspondingly dramatic: the pay and employment prospects of the least productive would be tra nsformed. By the standards of contemporary budget meddling, the gross cost would be enormous too: at today's prices, between $100 billion and $125 billion a year. That is roughly 1% of GDP, or not much less than the cost of the current Medicaid programme (which pays for health care for the poor). So this is nothing if not a big idea.

However, there would be some direct budgetary savings. Mr Phelps's plan would supersede (and greatly improve upon) the Earned Income Tax Credit, which currently costs $35 billion a year; spending on Medicaid itself would be lower (perhaps by $15 billi on a year); and so would the cost of Assistance for Families with Dependent Children (saving another $10 billion). Revenue gains from higher employment and assorted other feedbacks would reduce the long-term cost further. Ultimately, Mr Phelps argues, the plan would be roughly self-financing-though in the short-term other taxes would have to rise to cover the cost. He works carefully through this and other practical problems, including the temptation to under-report wages and over-report the number of emp loyees. At the end of it all, despite the drawbacks and uncertainties, he makes a persuasive case.

In what way, you might ask, is this plan any different from a myriad other schemes to spend outlandish amounts of public money to no good purpose? If big government has been shown not to work, it seems mad to make it even bigger. But the modern welfare state is working very badly, and Mr Phelps's scheme is intended in due course not to add to it but, in large part, to replace it:I have been careful not to assume that the government will first dismantle the welfare system or the whole panoply of entitlem ents. Let capitalism compete with the welfare system by giving it for the first time since the 1920s a level playing field. Let employment subsidies have at least some small fraction of the vast sum spent on transfer payments. Then, as the employment subs idy succeeds in shrinking welfare's `market share', it will be seen as very nearly self-financing . . .

It would be much better, however, if the initiative described here was merely the first step in the direction of re-establishing and revitalising the conception of government on which the country was founded. Such a scenario is at least imaginable . .

. With the less productive back on their feet, there will be hope of a reassessment of the government as an insurance company and equaliser of incomes . . . We may find that we have also helped to restore the founders' conception of a limited and minima l government and helped to preserve the capitalism that the founders provided.

In Washington, few if any people, whether sceptics or sympathisers, would agree with Mr Phelps that this scenario is ``at least imaginable''. Because of its initial cost, his scheme lies quite outside the realm of political possibility. Yet give the same politicians a few hundred small ideas for spending money uselessly (or worse), and over the space of a decade they will have no trouble at all in adding $100 billion a year to public spending.

The real problem with Mr Phelps's plan is not that the short-term costs are too high: it is that they are too clear. His scheme violates at least two laws of practical transfer-seeking politics: raise taxes little but often, and disguise higher taxes a s lower ones. If you follow these rules, no enlargement of the state is unthinkable. The Phelps plan shows that liberal economics need not be anti-poor; but sadly it offers no grounds for hoping that the state can in fact be rolled back. Such is the tyran ny of practical politics.

Can democracy tame the state?

The future of liberty

GROWTH of government in the advanced economies over the past 40 years has been persistent, universal and counter-productive. The evidence to date is that democracy is indeed incompatible with freedom, at least in a form that the classical liberals might h ave recognised. This is not to imply that any other form of government is more compatible with freedom: on the contrary, non-democratic governments are almost certain to be less so, either now or in the end. Democracy constrains the power of the state-not as effectively as one might wish, but to some extent. In authoritarian systems, including those which appear for a time to be benevolent, this check is missing. When it comes to protecting citizens' liberties, democracy is the only serious candidate. But the lack of good alternatives makes its record no less disappointing, and its prospects no less disturbing.

What seems most worrying is the perception in so many western nations that the market reforms of the past 10 or 15 years went too far, and that it is time to reaffirm the role of the state. In many developing countries, and above all in the former comm unist countries, market reforms have indeed rolled back the state-not so much because their citizens chose that course but because their states simply collapsed. In the West, however, progress towards smaller government has been more apparent than real. O n close examination, even the most dedicated reformers-Ronald Reagan in America and Margaret Thatcher in Britain-did not achieve a great deal. In America the growth of government was checked, at best; in Britain the state had shrunk a little by the end of the 1980s, but one more term of Conservative government restored it to the dimensions of 1979. Elsewhere in the West the state kept growing, checked only occasionally by fiscal crisis. But such fiscal emergencies should not be relied upon for anything bu t temporary relief. In the past, governments' medium-term response to fiscal crisis has been higher taxes, not lower spending.

In an earlier section, this survey identified some causes of democratic failure (the democratic state's built-in tendency to expand under its own power), such as the pursuit of the median voter, transfer-seeking, and political and bureaucratic self-int erest. However, the renewed popularity of leftist and/or centrist political parties suggests that something more is afoot. Rising taxes may be unpopular, as they always have been, but the demand for collective solutions to economic and social problems see ms, if anything, stronger than before. The upward pressure on public spending has not relented: the new left, just like the old left (and, for that matter, the right) intends to respond.

Being new, it has new ideas. Rather than spend more and raise taxes to meet the cost, it will devise more government regulations, or invent new non-tax revenues, or develop ``partnerships'' with the private sector. Is poverty a problem? Avoid increasi ng benefits, because that will require higher taxes. Raise the minimum wage instead (or set one for the first time, as in Britain). Short of revenue? Rather than raise taxes, increase charges for government services, which will probably count as negative spending in the public sector's accounts. Need new roads and bridges? Get them built by the private sector and then lease them back. It is bound to cost more in the long run, because of the financing charge wrapped up in the lease, but it makes public-spe nding figures look better in the short term. Schemes of this kind are so much the rage that, if they could be added back into a comprehensive measure of government intervention, the slowing in the growth of the state during the past few years might actual ly disappear.

America Inc, Britain plc

Michael Oakeshott, a philosopher and historian, described two ideas of government that have been in contention for centuries. The first, rooted in ancient Greece, is the state as civil association. On this view, the state's job is to help people live thei r own lives according to their own ideas, imposing no goals of its own on its citizens. Contrasting with that is the idea of the state as enterprise association. On this view the state has aims of its own (to raise the incomes of all its subjects, say, to establish economic equality among them, to conquer neighbouring lands, to glorify God): government directs the enterprise in order to achieve these goals. The first view puts individuals at the centre, the second society. From the first view comes classi cal liberalism-and from that the constitution of the United States (as it was understood until the 1920s). From the second comes socialism, and many varieties of modern conservatism.

It appears that the second view has prevailed. When presidents and prime ministers talk of winning on the battlefield of global competition, of what America, or Britain, or Germany needs, they talk as managers of enterprises with goals of their own. Th e notion that the pendulum is swinging back from too much reliance on markets to a wiser understanding of the role of government only shows how entrenched the idea of the enterprise state has become. Citizens respond to this vision: they want to be in the winning team, led by a captain they can rally behind. In the West, it seems, original sin has been superseded; instead, people come into the world with an original burden of obligation to the social enterprise, a debt to their fellow citizens that is not of their own making and that they can never discharge. Though not without its comforts, it is a kind of bondage. It augurs well for big government.

These pages brought to you by: