Speaking to the FRBK’s Economic Policy Symposium in Jackson Hole, Ben Broadbent, Deputy Governor for Monetary Policy, sets out why, historically, UK monetary policy failed to respond directly to labour market developments; why, in 2013, “the MPC chose to condition policy explicitly on unemployment… and how things have progressed since then.”
Ben begins with some history. In 1968, Milton Friedman warned that the negative relationship between wage growth and unemployment – the so-called “Phillips curve” – only holds when inflation expectations are stable, that monetary policy was in the end “neutral” (it has no long run effect on real quantities) and that policy makers should confine themselves to a purely nominal target. The experience of the 1970s, when inflation rose rapidly despite higher unemployment, bore out Friedman’s point and UK monetary policy was then directed to a series of nominal objectives, settling on an inflation target in 1992. There was no mention of any real objective and, although official interest rates have been clearly cyclical under inflation targeting, they responded to output not (at the margin) to unemployment.
During that time, however, as inflation expectations were stabilised, the Phillips curve re-appeared in the data. Indeed, the relationship between unemployment and wage inflation between 1993 and 2012 is remarkably similar to the one that Phillips himself identified in the final half century of the gold standard (1861-1913). So the fact that UK interest rates have been unresponsive to unemployment is not because it’s been an unreliable guide to inflation. The reason, Ben argues, is that unemployment has tended to lag output over the cycle and because, at least until the crisis, the MPC felt confident about the rate of growth of the economy’s underlying supply capacity. If potential output is growing at a steady rate, actual GDP growth is a sufficient statistic for changes in the output gap and future inflation: policy makers acted as if “stabilising inflation and stabilising the real economy amounted to the same thing”.
These assumptions no longer hold. Since the financial crisis “labour productivity has stagnated”, this has inevitably made the MPC “less confident about any forecast of productivity growth over the next few years” and meant that “output data are no longer sufficient statistics for inflationary pressure”. Ben argues that labour market data “give one a better steer about the evolution of spare capacity than output growth alone” and it was against this backdrop that, in August 2013, the MPC linked policy to unemployment in the first phase of forward guidance.
Since then, the MPC has been surprised by two things. Employment has grown “significantly faster” and “nominal pay growth has been much weaker” that we’d expected: during the first half of 2014, annual growth in average earnings has been around four standard deviations lower than one would have expected, given the level of unemployment and the Phillips-curve relationship between the two during the first twenty years of inflation targeting. This suggests that potential output may well have increased, but it’s been in the form of greater labour supply rather than greater productivity, as the MPC had anticipated.
It also means that it becomes harder to communicate publicly what matters for policy: when demand, productivity and labour supply can all vary, policy makers need to consider output, employment and wages. In the end, of course, “credibility depends on the attachment to the ultimate objectives – primarily the inflation target, in the case of the UK – not to any particular feedback rule”. It may well be that, in time, supply growth again settles down to some stable rate. In the meantime, “we have little option but to say what central bankers have always said, that meeting the inflation target depends on a ‘range of indicators’”.