In economic circles, the relationship between wage growth and unemployment is an important factor. Many will focus on the relationship between short-term unemployment and wage growth, but a paper released by the Bank of England highlights that long-term unemployment is also an important factor in the equation. Given the fact that wage growth is slowing in Australia, and long-term unemployment is rising, these findings are important.
The relationship between wage growth and unemployment is a key trade-off concerning monetary policy makers, as labour costs form a critical part of the inflationary transmission mechanism. One important question is how the composition of the unemployment pool, and specifically the share of long-term unemployment, affects that tradeoff. Detachment from the labour force is likely to increase with unemployment duration, so that the long-term unemployed search less actively for jobs and therefore exert less downward pressure on wages. If so, short-term unemployment may pull down on wage inflation more than long-term unemployment does. In this situation, policymakers might anticipate a period of high wage growth if short-term unemployment starts to fall to low levels even if the long-term unemployment rate remains elevated.
But there may be complications arising from the integral dynamics of unemployment. In this paper it emerges that the estimated disinflationary effects of long-term unemployment hinge on whether or not wage growth becomes less sensitive to unemployment as the latter rises – a form of non-linearity. One reason why the negative relationship between wages and unemployment might become flatter at high levels of unemployment is that workers may tend to resist cuts in their nominal wages. When unemployment is low, wage growth tends to be high as firms compete for a scarce pool of resources. But due to worker resistance to wage cuts the reverse might not hold to the same extent, with a relatively large increase in unemployment needed to reduce wage growth during a recession.
Why does this non-linearity matter for the measured effect of long-term unemployment on wage growth? It is because long-term unemployment inevitably lags behind movements in short-term unemployment as it takes time for the new unemployed to move into the long-term category. So high levels of long-term unemployed are only associated with lengthy periods of high unemployment. A flattening off of the relationship between wages and unemployment at high levels of unemployment would then imply that long-term unemployment does little to reduce wage inflation further. The apparently different effects of short and long-term unemployment on wage inflation could therefore be merely as a result of timing rather than labour market detachment among the long-term unemployed.
By modifying statistical models of labour market dynamics to incorporate this insight, this paper finds that there appears to be much less difference between the short and long-term unemployed in terms of their marginal influence on wage behaviour than is suggested by the recent literature. When the non-linearity described above is not taken into account, estimation results corroborate the finding already established in the literature that it is predominantly the short-term unemployed that matter for wage inflation. Long-term unemployment in this specification tends to have no statistically significant effect on wage inflation. When the non-linearity is taken into account, long-term unemployment has a much larger effect on wage inflation. For some of the specifications considered, the data fail to reject the hypothesis that short and long-term unemployment rates have equal effects on inflation. In some instances, the models even suggest that long-term unemployment creates more of a drag on wage growth than short-term unemployment does, all else equal. Statistical uncertainty makes it difficult to draw a very precise conclusion, but the results in this paper caution against excluding long-term unemployment from estimates of aggregate labour market slack as is suggested by much of the recent literature. Both the short-term unemployment rate and the long-term unemployment rate are likely to contain useful information for judging the degree of wage pressure in the economy.
