Full employment objectives for central banks are common and long-standing. They are harder to assess than inflation objectives, and evolve as our understanding of the economy improves.
Many central banks, including the RBA, have ‘full employment’ as part of their mandate, in addition to inflation control. In most cases, again including Australia, this has been true for decades. Indeed, it is well understood that even a central bank with an inflation-only mandate would still want to have a view on full employment, because a labour market being away from that point will influence inflation in future, potentially pushing it away from target. This was the point Mervyn King made back in 1997, that even an ‘inflation nutter’ central bank would still care about employment and unemployment.
But how do central banks know if they have achieved a full employment goal? An inflation target can be specified as a desired rate or range of rates for inflation, and perhaps a horizon over which to achieve it. ‘Full employment’ has, however, always been a bit more nebulous. Typically, policymakers and academic economists alike have framed ‘full employment’ as being the lowest rate of unemployment that is still consistent with inflation remaining at its target. That way, the two mandates are defined not to conflict. ‘Full’ does not mean ‘everybody who wants one has a job’, and it is not necessarily as low a rate of unemployment as other parts of society might prefer.
The evolution of economic understanding of the relationship between inflation and unemployment began in the middle of the previous century, when A W Phillips noted an inverse relationship between growth in money wages and the unemployment rate in the UK. In other words, an unemployment rate below its ‘natural rate’ – an ‘unemployment gap’ – was associated with faster wages growth. The ‘Phillips curve’ concept later expanded to be between price inflation and unemployment rather than money wages growth and unemployment. More recently, the RBA has also related these ‘gaps’ to growth in unit labour costs (which includes all employee compensation, not just wages, expressed relative to output produced) rather than wages or prices directly. This approach stems from the ‘markup model’ framework where marginal costs and labour market slack are related, rather than prices and slack. As we have previously noted, other central banks do not put anywhere near the emphasis on this model that the RBA does.
Later work focused on the role of expected inflation, holding that unemployment would only decline below its ‘natural rate’ if people were ‘surprised’ by inflation. If unemployment was below the ‘Non-Accelerating Inflation Rate of Unemployment’ (now you know why economists call it the NAIRU), inflation (or wages growth) would not just be higher, but would be ever-increasing, because people would come to expect the current rate of inflation to continue, and then you needed to surprise them some more. So if inflation or wages growth was increasing, that was taken as evidence that the unemployment rate was below the NAIRU and the labour market was tighter than ‘full employment’. A steady rate of unemployment or wages growth was seen as a signal that the labour market was roughly in balance.
The RBA’s recent work on establishing where full employment might be has instead taken the view that inflation expectations in Australia are reasonably anchored. So instead of looking for increasing inflation or wages growth as a sign that the labour market is tighter than full employment, any deviation of inflation from target, or wages growth from what the central bank judges to be consistent with inflation remaining at target, is seen as a sign that the labour market is away from full employment. The NAIRU isn’t really dead in the RBA’s framework: they just dropped the “A”.
Observant readers will note that the judgement about whether wages growth is consistent with inflation at target requires a judgement about the mapping from wages growth to labour costs to inflation target. This involves a judgement about trends in productivity growth that is subject to both debate and data revisions.
The RBA has also fleshed their assessment process out with a ‘checklist’ or ‘dashboard’ of indicators of full employment and a separate one for the output gap. This has not been without criticism from labour market experts, though to be fair the dashboards have been improved over time and are now less sensitive to historical averages. It has also integrated at least some of the insights from labour economics that frame the labour market as an exercise in matching workers and jobs, with all the frictions that might arise from that. Still, the exercise has a little of the flavour of the RBA’s use of ‘checklists’ in the early 1980s, before the inflation targeting era. That experiment did not end well.
In the end, though, the RBA’s framework is all about taking signal from inflation and wages data and reading back to the labour market. Consider two possible views of economic trends. One is that trend growth in capacity – that is, potential output – is just above 2%, and the sustainable rate of unemployment is around 4.6%, the RBA’s latest estimate. Another, closer to our own view, would be that potential output growth is more likely to be in the 2¼-2½% range and the sustainable unemployment rate is more like 4¼%. The labour market data on its own will not help you distinguish between these two sets of assumptions. The signal from inflation and wages growth will be the deciders.
The upshot is that the RBA has put a lot of effort into measuring where full employment is. Neither it, nor the profession more broadly, have done as much on building theory about why it is where it is, or where it might go from here, including whether it could be deliberately moved to make even lower unemployment feasible. And despite the dropping of the “A”, the intellectual framework has not evolved much over the past decade. It is still all about inferring gaps from Kalman filtering the relationship between output, labour market data and price and wage outcomes.
There are alternatives, which both policymakers and the research community would be familiar with but have generally not pursued, sticking instead to the ‘gap-oriented’ approach to assessing the balance between demand and supply. One alternative – though by no means the only one – was even expounded in a Bank of England publication more than a decade ago (full disclosure: the author, Professor Roger Farmer, was one of my PhD thesis examiners). There is also more to be gleaned from search-oriented analysis of the labour market.
The analysis of where full employment is therefore far from settled. Further changes in central bank frameworks can be expected as more research is done. Hopefully this future research avoids incrementalism and considers the implications of how labour markets actually work in the 21st century. Meanwhile, the NAIRU might have changed form, and dropped a letter here or there, but it is not dead to central banks.
