Plate iii in print

The Tradeoff

Phelps; the Phillips curve and the natural rate — inflation buys lower unemployment only by surprise, and never for long.

A surprise inflation can coax out jobs for a season, by paying workers in money worth less than they thought; but once they see the trick and demand wages to match, the jobs melt away and only the rising prices remain — so the bargain, repeated, buys nothing but inflation.

— the argument of this plate, after Edmund Phelps, 1967

The quantity theory says inflation lowers the value of money; but for a generation, policymakers believed inflation could also buy something precious in return — lower unemployment. The apparent evidence was the Phillips curve, and the question it posed was whether a central bank, by tolerating a little inflation, might hold unemployment permanently low: a tempting bargain, and one that seemed to turn the power over money's value into a power to manage the real economy. Edmund Phelps, and independently Milton Friedman, showed the bargain to be an illusion, and in exposing it revealed why the power over money's value is so much more constrained, and so much more fraught, than it first appears. The next step is to see why inflation cannot be traded for jobs.

In 1958 A. W. Phillips, examining nearly a century of British data, found a striking regularity: in years when unemployment was low, wages and prices rose quickly; in years when unemployment was high, they rose slowly or fell. The relationship traced a smooth downward-sloping curve — high inflation paired with low unemployment, low inflation with high — and it looked for all the world like a stable tradeoff, a menu of combinations from which a government might choose. If a society wanted lower unemployment, it could have it, at the price of higher inflation; if it wanted lower inflation, it could have that, at the price of higher unemployment. To the policymakers of the 1960s this appeared to be a usable lever: run the economy hot, accept a few points of inflation, and keep unemployment gratifyingly low. The central bank's power over the value of money seemed, on this reading, to be also a power to deliver jobs — to buy employment with inflation.1

1. In 1958 A. W. Phillips found, in a century of British data, that low unemployment came with fast-rising wages and prices, high unemployment with slow. The relationship traced a downward-sloping curve — a stable-looking tradeoff, a menu of inflation-unemployment combinations. To 1960s policymakers it seemed a usable lever: tolerate a little inflation, keep unemployment low. The power over money's value looked like a power to buy jobs with inflation.

Phelps asked the question the simple curve had skated over: why would inflation reduce unemployment at all? There is only one possible mechanism, and it is unsettling. Inflation reduces unemployment by fooling workers. When prices rise faster than workers had expected, the wages they agreed to in advance turn out to be worth less than they thought; labor is cheaper, in real terms, than the workers realize, so firms find it profitable to hire more, and unemployment falls. But this depends entirely on the workers being deceived — on the inflation being a surprise. The lever works not because inflation is good for employment but because unexpected inflation quietly cuts real wages below what workers would have accepted with open eyes. And a trick that works only by deception cannot work for long.2

2. Phelps asked why inflation would reduce unemployment at all. The only mechanism: it fools workers. When prices rise faster than expected, agreed wages are worth less than workers thought; labor is cheaper in real terms than they realize, so firms hire more and unemployment falls. But this depends on the inflation being a surprise — it works by quietly cutting real wages below what workers would knowingly accept. A trick that works only by deception cannot last.
⁂

Phelps drew the consequence that demolished the simple tradeoff. Once people come to expect the inflation, the trick stops working. Workers, having been fooled once, build the expected inflation into their wage demands — they negotiate raises that anticipate the rising prices — so their real wages no longer fall, firms no longer find labor unexpectedly cheap, and unemployment climbs back to where it began. But the inflation does not climb back; it remains. The economy, Phelps argued, has a natural rate of unemployment, set by real features of the labor market — the friction of matching workers to jobs, the structure of skills and search, not the rate of inflation — and it returns to that natural rate whatever the inflation. You cannot hold unemployment below the natural rate for long; the attempt only ratchets inflation higher. The long-run Phillips curve, far from sloping gently downward, is vertical: every rate of inflation is consistent with the same natural rate of unemployment. The tradeoff is real only in the short run, and only by surprise; in the long run there is no tradeoff to exploit at all.3

3. Once people expect the inflation, the trick fails: workers build expected inflation into wage demands, real wages no longer fall, and unemployment returns to where it began — but the inflation remains. The economy has a natural rate of unemployment, set by real features of the labor market, to which it returns whatever the inflation. The long-run Phillips curve is vertical: any inflation rate is consistent with the same natural rate. The tradeoff is real only in the short run, and only by surprise.

The theory made a bold prediction, against the settled wisdom of its day: that a government which kept trying to exploit the tradeoff — running inflation to hold unemployment below its natural rate — would reap not permanently low unemployment but accelerating inflation, and in time high inflation and high unemployment together. The simple Phillips curve said that last combination was impossible, since inflation and unemployment were supposed to move in opposite directions. The 1970s delivered it exactly: stagflation, double-digit inflation standing alongside high unemployment, as years of buying jobs with inflation left only the inflation, with expectations now thoroughly built in. The episode vindicated the expectations-augmented Phillips curve and discredited the naive tradeoff, and its lesson burned itself into the profession and into the conduct of central banks: inflation cannot purchase lasting employment, and a central bank that tries will end with the worst of both.4

4. The theory predicted, against the wisdom of the day, that exploiting the tradeoff would yield not low unemployment but accelerating inflation, and eventually high inflation and high unemployment together — which the simple curve called impossible. The 1970s delivered exactly that: stagflation, double-digit inflation beside high unemployment, the inflation now built into expectations. The episode vindicated the expectations-augmented curve and taught central banks that inflation cannot buy lasting employment.

This transformed the understanding of the central bank's power over the value of money. If inflation cannot buy lasting lower unemployment, then inflation has no durable benefit to weigh against its costs — its only lasting effect is to lower the value of money and to impose the burdens of inflation: the silent tax on cash, the erosion of savings, the arbitrary redistribution between debtors and creditors. The power over money's value, then, is not a lever for prosperity; turned to the chasing of jobs, it yields nothing but inflation. The conclusion is a sobering one. It means the right long-run use of the power is simply to hold the value of money steady, and that the temptation to do otherwise — to inflate for a short-run gain — is a trap that ends in stagflation. Yet if the theory is known, why would any central bank fall into the trap? Because the short-run gain is real and immediate while the long-run cost is deferred and diffuse — a temptation built into the very structure of the power. That temptation, and the institutional device that answers it, is the matter of the next plate. The next plate is the temptation.5

5. The consequence: if inflation cannot buy lasting employment, it has no durable benefit against its costs — only a lower value of money and the burdens of inflation. The power over money's value is no lever for prosperity; chasing jobs with it yields only inflation. So the right long-run use is to hold money's value steady, and inflating for short-run gain is a trap. Why would a bank fall for it? The gain is immediate, the cost deferred — a temptation built into the power. Next: the temptation.

The Phillips curve seemed to offer the central bank a lever — buy lower unemployment with a little inflation. Phelps and Friedman showed the lever an illusion that works only by surprise: inflation lowers unemployment only by fooling workers into accepting wages worth less than they expected, and once the inflation is expected the effect vanishes, leaving only the higher prices. The economy returns to its natural rate of unemployment whatever the inflation, so the long-run tradeoff is no tradeoff but a vertical line — as the stagflation of the 1970s confirmed against the conventional wisdom. The lesson for money's value is profound: inflation buys nothing that lasts, so its only durable effect is to erode money's worth, and the power to inflate is not a tool of prosperity but a temptation to be resisted. The next plate is the temptation.