Down with deflation!

Paul Seabright

The power of central bankers – about which Edward Luttwak wrote in the LRB of 14 November – arises not just from their control over important aspects of economic policy, but also from the acceptance by the rest of us of what they may legitimately do in the exercise of this control. Until recently, our acceptance of the notion that central bankers should be committed to price stability has been entirely uncritical; and price stability (not low, but zero inflation) is what the European Central Bank will be required to maintain. But now that European Monetary Union suddenly looks a real, even an imminent possibility, a skirmish has broken out among economists about whether price stability is what monetary policy should be required to achieve.

There are two reasons to take seriously what might otherwise seem the purely pedantic question of whether governments should aim for zero inflation or a ‘reasonable’ rate of, say, 3 or 4 per cent. One is that, at first sight, several countries provide cautionary tales of what can happen when fear of inflation becomes obsessive. Germany is an obvious example: its recession of the early Nineties (subsequently exported, like so many other German products, to the rest of Europe) was due to the Bundesbank’s failure to understand that the extraordinary event of reunification required a more relaxed attitude to inflation if growth were not to suffer (a more sinister explanation is that the Bundesbank understood perfectly well, but didn’t care). Another worrying case is France, whose unemployment rate remains almost double that of Britain a decade after it adopted German monetary policy by effectively hitching its exchange rate to that of the Deutschmark. (One can multiply examples of countries that tighten monetary policy at severe cost to employment in the expectation that unemployment will fall once inflation is under control, and then find themselves waiting unconscionably long for entry to the Promised Land.) It’s easy – and in France’s case, almost irresistible – to blame arthritic labour markets and a romantic Marxist blindness to the sectional selfishness of trade unions. But Latin labour markets and Germanic monetary policy are an uncomfortable combination (Catholic behaviour plus Protestant conscience); and it’s not obvious that the labour markets are the easier to reform. Persistent mass unemployment – if it really is more likely under tight monetary policy – is by any standards a vast cost to pay for zero inflation. Is the allure of price stability great enough to compensate?

Even if the goal itself is a worthy one, is it right always to err on the side of monetary tightness? In Britain responsible newspapers are currently calling for rises in interest rates, for fear that the Government’s inflation target of 2.5 percent may be breached before the end of this Parliament. They are doing so in spite of the recent large rise in the value of sterling, which is likely both to make inflation slow down of its own accord and stifle the recovery in output and employment. Are there good reasons to think that the dangers of missing the inflation target should count for more than the dangers of an excessive tightening of demand? Is the reference by the Financial Times on 15 November to ‘the battle to slay the beast of rising inflation’ timely and prudent, or ludicrously over the top?

A second reason to find the debate fascinating is what it reveals about the way instinct and evidence confront each other in a profession that likes to pride itself on its technocratic objectivity. There is overwhelming evidence – from both case studies and comparative statistics – that very high rates of inflation are extremely damaging to economic welfare in the present, and to investment and growth in the future. I have visited many factories in the former Soviet Union where money has been largely abandoned as a medium of exchange because of hyperinflation and the collapsing banking system, and where all investment and most of management time and effort are devoted to finding and stocking goods that will be useful in barter exchange – everything from aero-engines to potatoes. But there is no evidence that moderate rates of inflation are damaging. Or rather, there are a few conflicting pieces of evidence, none of them individually compelling, plus a great many theories, hunches and plain assertions. What is fascinating is that the flimsiness of the evidence does not lead to tentativeness or moderation among the debaters: inflation appears to stir some strange and deep passions.

What does the evidence actually show about the effect of a little inflation on growth? National growth rates in the European Union are currently in the range of around 1 to 2.5 per cent per year; inflation rates range from zero to around 2.5 per cent. Annual growth rates in East Asia range currently from 4 per cent in Hong Kong, through 6 per cent in the Philippines and Taiwan, 7 per cent in Korea and Singapore, to 8 or 9 per cent in Indonesia, Thailand and China. All of these countries have inflation rates between 3.5 and 7.5 per cent, with the single exception of Singapore (1.5); they have not been using monetary policy to rein in growth for fear of a little inflation. Again, Poland and the Czech Republic (the star performers of Eastern Europe) are growing at around 4-5 per cent, with inflation rates of 20 and 8 per cent respectively.

Such crude correlations are, of course, easily misleading. But a recent econometric study by Michael Sarel (admittedly published in that notorious hotbed of irresponsible radicalism, the IMF Staff Papers) argues on the basis of cross-country evidence that inflation appears to have quite different effects when it is low and when it is high. Sarel estimates that there is a ‘structural break’ at around 8 per cent annual inflation, above which inflation is unambiguously bad for growth but below which it is even slightly beneficial. Another study, by Robert Barro, finds that at levels below 15 per cent the effect on growth, though negative, is statistically insignificant (since the study was published in the Bank of England Quarterly Bulletin, however, its author has thought it prudent to cite the average results – determined entirely by the experience of countries with inflation higher than 15 per cent – as ‘more than enough to justify the Bank of England’s keen interest in price stability’). The importance of the ‘structural break’ theory is its implication that the clear evidence on the evils of inflation at high levels is irrelevant to the question whether low inflation is sufficiently damaging to justify stamping it out altogether. The only studies that ask this question directly find no evidence that it is.

Appealing to structural breaks would seem like a quibble if there were no reasons in principle to expect inflation to have qualitatively different effects at low and high levels. There’s an old argument, to the effect that inflation (which taxes holdings of money) should, like other taxes, be set at modest levels – better to tax many things a little than a few things very hard and other things not at all. This argument is particularly relevant to some former Communist countries that have poorly functioning tax systems and which (like Poland, who in this respect has ignored foreign advice) have quite reasonably chosen to live with middling inflation until they complete their fiscal reforms.

But it is a second argument in favour of moderate inflation, particularly as expressed in a recent paper by George Akerlof, William Dickens and George Perry, that has stirred up the current controversy. Some years ago, James Tobin suggested that moderate inflation might ‘oil the wheels’ of an economy by facilitating adjustments in relative wages in different sectors (necessary to encourage people to shift jobs in response to changing technology and opportunities). Real wages in some sectors typically need to fall – even in an economy that is, on average, growing. But inflation allows this to be achieved by gradual erosion rather than by painful direct cuts in money terms, which Tobin claimed would be resisted. Akerlof et al have now found evidence from US labour market surveys that this does indeed happen: there is a very obvious ‘bunching’ of wage settlements at zero when one might expect them to be dispersed more smoothly at both positive and negative levels. The actual graph of settlements looks like a camel scratching its back rather desperately against a palm tree planted at zero, with only a thin tail trailing behind. If there is indeed a deep resistance to cutting money wages, Akerlof et al estimate that reducing inflation to zero might permanently raise the unemployment rate to a stable level of around 8 per cent, as against just under 6 per cent with an inflation target of 3 per cent. That’s more than two million extra Americans out of work, for ever.

The reaction to these arguments by defenders of anti-inflationary orthodoxy has been histrionic, not to say hysterical. The Economist (which also recently described inflation as ‘the great beast’) ran a leader in September entitled ‘Who’s afraid of inflation?’ under a picture of a wolf disguised as a sheep, warning against the ‘strong temptation ... of governments to nudge growth along by relaxing their monetary guard’. The word ‘temptation’ appears four times in the article. Michael Prowse in the Financial Times (describing the Akerlof view as a ‘disease’) claimed that it was ‘patently implausible’, even ‘insulting’, to suggest that workers could be fooled by inflation into accepting real wage cuts while refusing to take money wage cuts. More moderately, Mervyn King of the Bank of England devoted part of this year’s ESRC lecture, and an article in the Guardian, to warning against Akerlof, because although there was no evidence on the question either way, ‘monetary stability would surely do more good than harm.’ Curiously, his lecture did not mention the labour market evidence that underpins these views.

The revolt within economics against the naive Keynesian expansionism of the post-war era has been based largely on a welcome rejection of the idea that governments can deceive citizens permanently into ignoring the decline in the value of their currency. So a degree of professional scepticism towards Akerlof is understandable, even if the hint of panic seems bizarre. In the long run, it is now said, workers cannot be fooled. But there may be reasons for resisting money wage cuts (while accepting a gradual erosion of their value through inflation) even if workers are not fooled, for real-wage adjustments take place very differently under the two processes. Inflation erodes real wages smoothly and gradually, without any need for the humiliating summons to the boss’s office. Money wage cuts happen suddenly, discontinuously, disrupting in a single gesture the existing relativities between the wages of one group and those of others against whom that group compares itself. There is massive evidence that we care about our incomes not just because of the goods they will buy, but also because of what they signal about the perceived worth of our contribution relative to that of others. Adjustments through inflation can happen without branding our work and status as suddenly devalued, even if over time they may lead to disgruntlement. Cutting money wages (besides upsetting the calculations we make based on committed outgoings that are fixed in money terms) adds to the cost of adjusting real-wage relativities the wholly gratuitous cost of picking out groups of workers for humiliation one at a time.

At present, the Association of University Teachers is organising strike action in protest against the universities’ current pay offer, which is below the inflation rate: who could doubt that the strike response would be much more solid if inflation were lower and we were asked to accept a wage cut? (Adrian Wooldridge once suggested that academics should use their ultimate weapon, the footnote strike, in which we would all refuse to submit footnotes with our articles. It’s hard to know which prospect was more pleasing, that of the country grinding to a halt through a shortage of footnotes, or that of the army being called in to provide emergency footnote cover.)

If economically important decisions are interdependent in the way I’ve suggested, a little inflation might be a valuable co-ordinating mechanism to make the adjustments more smoothly. In a similar way, moving to British Summer Time helps us all to get up a little earlier in the summer months, something we appreciate and which we could in theory try to do on our individual initiative. We would never actually do it by ourselves, however, because our timetables are too dependent on what others around us are doing. Would Michael Prowse feel insulted by the paternalism of British Summer Time, or suggest we know it’s always winter in the long run?

If anything, the instinct of paternalism is more visible on the side of orthodoxy in this debate. A recent paper by Martin Feldstein argues that price stability would indeed be desirable for the United States, because inflation creates distortions through the imperfect indexation of the tax system, particularly in its treatment of saving and capital gains. To the obvious riposte that it might therefore be better to reform the tax system, he cites ‘technical and administrative difficulties’, the costs of overcoming which he chooses not to calculate, let alone compare to the unemployment costs of bringing inflation down. But, he continues, ‘there is a more fundamental concern that an indexed tax system might lead to less public support for anti-inflationary policies.’ In other words, inflation might hurt less, so we might fear it less. What an appalling prospect.

When the Aids epidemic began there were those who objected (and not only in the Catholic Church) to encouraging the use of condoms on the grounds that it would soften the effects of the disease and make us fear it less. The parallel may seem far-fetched, but it is curious how much of the language of the inflation debate brings to mind moral injunctions to personal restraint. A common theme is the suggestion that positive rates of inflation are somehow addictive, that (in the words of Arthur Okun, recently quoted approvingly in a Bank of England review of the issue) steady inflation is a ‘mirage’. Or as the then Governor of the Bank said sternly in a 1992 lecture on ‘The Case for Price Stability’: ‘Whenever inflation is viewed as acceptable, it is possible to settle for an alternative which is just a little higher.’ It leads on to the hard stuff, you see.

I never believed this argument when it was offered to me in my teens, and I don’t believe it now. All heroin addicts, I was confidently informed, had started off by taking pot (the same could be said of milk). Doubtless all hyperinflations have started off as mild inflations, but the proportion of mild inflations that have led on to hyperinflation has, historically, been very small. Or perhaps it has only been thanks to the iron self-control of central bankers that we flabby moderates haven’t gone over the precipice? Iron self-control was also, as it happens, a great preoccupation of those Victorian writers on the upbringing of children who sought to eradicate the evil of self-abuse. William Acton, for example, in his best-selling Functions and Disorders of the Reproductive Organs, offers pitiful descriptions of the physical and mental degeneration attributable to masturbation in boys, noting that ‘such boys are to be seen in all stages of degeneration, but what we have described is but the result to which they all are tending’ (Acton’s emphasis). He suggests that the way to continence lies in ‘that power of the mind over outer circumstances which we call “a strong will” ’. It is, he says, ‘a matter of habit. Every victory strengthens the victor ... The whole force of his character, braced and multiplied by the exercise of a lifetime, drives him with unwavering energy along his chosen course of purity.’ The language is uncannily familiar: every ‘courageous’ decision by the authorities to raise interest rates strengthens their hand – as it were – in the fight against ‘the great beast’. Anti-inflationary habits must be ‘entrenched’, ‘monetary discipline’ requires ‘painful choices’. Until the end of the 19th century the colloquial term for achieving orgasm was ‘to spend’. The Economist certainly hates spending.

Even Acton’s depressing reference to ‘the exercise of a lifetime’ finds an echo in the anti-inflation literature. Another recent piece by Michael Prowse (headlined ‘Monetary Vigil’, and warning the Federal Reserve Chairman, Alan Greenspan, against ‘the siren voices telling him that the battle against inflation is won’) concludes with the chilling observation: ‘After nearly twenty years of struggle the US is on the verge of achieving price stability. It would be tragic if the Fed threw in the towel just when success was within reach.’ Did he say twenty years? Silly me, I thought the whole point of monetary restraint was that the pain was short and the gain was long.

Are there no reasons to think inflation is addictive? As it happens, there are some. Or rather, there is a theory, profoundly influential within the economics profession, which has been unanimously accepted for so long that we have failed to notice that it fails to fit the facts of recent political history. This is the theory that there is ‘inflation bias’ in the political process. Once citizens and firms have made decisions based on expectations of a given inflation rate, the government is tempted to engineer unexpected increases in demand to raise output at the cost of making subsequent inflation higher than anticipated. When the fraud is discovered, citizens revise their expectations of inflation upwards, so it takes a bigger increase in demand to have the same effect, and thus a faster rise in prices. And so on.

This theory makes the addiction view of inflation rest on two mechanisms: one for explaining why it will always be higher than its target; and one for explaining why the target will tend to be periodically revised upwards, and will in any case be set higher than it ought to be. The two mechanisms are quite separate. The first seems plausible enough much of the time, if only because it is in the nature of governments to promise more than they can deliver. But the second is believable only if you think that the political gains to a government from listening to the unemployed (who benefit from demand increases) exceed the gains from listening to citizens whose incomes and savings are eroded by inflation. And the fallacy in that idea was exposed over a decade ago, not by a Nobel Prizewinner in economics but by an Oxford chemistry graduate with a second-class degree called Margaret Thatcher. Until the 1983 election it was virtually the unanimous wisdom of economists that no UK government could be re-elected with unemployment standing at over a million. Mrs Thatcher showed them what they should have known already: namely, that the unemployed are not, by and large, marginal voters (though pretty marginalised in most other respects). At the election of 1983, real wages for those in work were higher than they had ever been, not by accident but as the natural consequence of a policy that brought price inflation down at a faster rate than the wage inflation that had been struggling to keep up. Mrs Thatcher’s lesson was learned by François Mitterrand, among other politicians, and Western European electorates have since then been tolerating high unemployment for longer than would previously have been thought possible.

If the only benefits of inflation came from systematic deception of the electorate, then the view of moderate inflation as inescapably addictive would make more sense (though it would also imply that achieving price stability should take much less than twenty years once the willingness to struggle for it is there). In fact, policies towards inflation emerge, like most policies, from a much more complex process of electoral calculation. It’s hard to know whether this process leads to systematic biases one way or the other, but that’s exactly why the issue is traduced by being discussed in terms of simple moralities. But then this is the era of moral revival in politics, so perhaps we should expect no less. When appointing a central banker we look nowadays for a safe pair of trousers (officially so since the Economist’s own editor had a go at the job), and it can’t be long before corporal punishment is brought in for any Governor of the Bank of England who allows inflation to exceed 2 per cent. A Blair government will be no less excited than its predecessor by price stability’s twin appeal to moral self-restraint and to the wallets of Middle England. But it’s hard to believe that the countries of East Asia (most of which preach moral self-restraint without the need for such charades) would be growing even faster if they had listened to the Governor’s lecture on price stability.

A well-known journalist once explained to me that he couldn’t be bothered to follow the detailed arguments on many controversial issues, because all one needed to know was ‘which side the wankers are on’. Just such reasoning seems to be at work in the current debate. I don’t know whether the right inflation target is zero, or 5 per cent, or (my hunch) some target that depends on how bad unemployment is right now, and on what’s happening to the real exchange rate. I’d much rather the inquiry were conducted using logic and evidence, and engaged seriously with the analysis that Akerlof and his colleagues have made of the way labour markets work. But if you want to know where my instincts lie, in this, as in most things, they’re with the wankers every time.