The Economist
2007/10/18
JOHN MAYNARD KEYNES thought that economics should be a technical profession. “If economists could manage to get themselves thought of as humble, competent people on a level with dentists,” he wrote, “that would be splendid!” If any economists have deserved this white-coated reputation for proficiency in their trade, it is probably those in charge of central banks. Yet the chief of their number, Mr Greenspan, instead achieved near-divine status as chairman of the Fed. Rather than being granted technocratic anonymity, he was sought out for his wisdom on almost any subject. He was feted as a maestro and a seer. The more mystifying his famously Delphic utterances, the more his powers seemed to grow. It was an odd way to run an economy.
……
Central bankers have a lot to lose if they slacken their grip. To see just how much, it is well worth delving into the abstruse technicalities of the statistical link between inflation and unemployment known as the Phillips curve. Bill Phillips, a New Zealander who taught at the London School of Economics, discovered a stable relationship between the rate of inflation (of wages, to be precise, rather than consumer prices) and unemployment in Britain over a long period, from the 1860s to the 1950s. Higher inflation, it seemed, went with lower unemployment. To the economists and policymakers of the 1960s, keen to secure full employment, this offered a seductive trade-off: lower unemployment could be bought at the price of a bit more inflation.
Flattening and flattering
In the late 1960s, however, the Phillips curve suffered devastating assaults—first from theory, then from fact. Separately, two American economists, Milton Friedman and Edmund Phelps (who both later picked up Nobel prizes, partly for this work), pointed out that the trade-off was only temporary.
In his version, Friedman coined the idea of the “natural” rate of unemployment—the rate that the economy would come up with if left to itself. Now economists are likelier to refer to the NAIRU (non-accelerating inflation rate of unemployment), the rate at which inflation remains constant.
Suppose that at first unemployment is at the NAIRU, u* in chart 3, and inflation is at p0. Policymakers want to reduce unemployment, so they loosen monetary policy: that stimulates spending, so that unemployment goes down, to u1. Inflation rises to p1, along the initial short-run Phillips curve, PC1. But that raises inflationary expectations, so that workers demand higher wage increases and real wages rise again. Firms shed labour, returning unemployment to u*, but with a higher inflation rate, p1. The new short-run trade-off is worse, with higher inflation for any level of unemployment (PC2). In the long run the Phillips curve is vertical (LRPC).
Remarkably, Mr Friedman and Mr Phelps made their criticisms before they were vindicated by the facts. But the facts were not long in coming. As inflation and unemployment rose in the late 1960s and 1970s, the curve steepened, or even sloped upwards (see chart 4, left-hand panel).
Robert Lucas, another eventual Nobel laureate, took the theoretical assault a stage further. The Phillips curve, said Mr Lucas, was a mere statistical regularity—a “reduced-form” relationship. It had no basis in theories about the behaviour of workers and firms. If workers and firms understood how the economy worked and thus could anticipate the effects of changes in policy, their behaviour would alter accordingly. Even in the short run, the curve would be mainly vertical.
All this left the Phillips curve in theoretical and empirical tatters. Thirty years on, however, both theory and fact have changed. The original Phillips curve, the observed trade-off between inflation and unemployment, has changed yet again. It is no longer steepening, as it was in the 1970s, but has been flattening. In the 1990s and early 2000s, as unemployment came down, inflation did not take off (chart 4, right-hand panel).
Why is the curve so much flatter? If you look at a Phillips curve for the 1990s and 2000s, all it really tells you is that inflation has become more or less constant. Because this has lasted for several years and because inflation depends ultimately on monetary policy, central banks can claim credit for this. They have won credibility, anchoring inflation expectations so that movements in actual inflation are damped. Because expectations are so central to inflation, communication has become an increasingly important part of monetary policy. Getting the message across depends as much on what central bankers say as on what they do.
……
