RDP 2008-05: Understanding the Flattening Phillips Curve 6. Conclusions

It is now 50 years after Phillips first observed the relationship between unemployment and wages, variants of which now occupy a critical position in the intellectual framework underpinning monetary policy. Recently, policy-makers have observed that fluctuations in activity do not appear to be as inflationary as in the past, which is borne out by our estimates of reduced-form Phillips curves. This paper has attempted to summarise some of the common arguments cited regarding why this has occurred, using the standard new-Keynesian Phillips curve as an organising framework. Our estimates suggest that there has also been a flattening in this ‘structural’ model, that is, there has been a change in the price-setting behaviour of firms. In particular, it appears that the duration between price resetting may have lengthened. Many of the common explanations for changes in the price-setting behaviour of firms are related to globalisation. While globalisation may alter the relationship between the output gap and marginal costs, it is unclear why it would alter the link between marginal costs and inflation in a way that corresponds to a flattening of the Phillips curve. In a structural model, the deep parameters in the Phillips curve should be invariant to changes in the conduct of monetary policy. However, one potential explanation is that lower-trend inflation resulting from the improved conduct of monetary policy may account for the more infrequent price resetting and hence the flattening Phillips curve, a possibility which is not accommodated in the benchmark new-Keynesian model. In all, it appears that after 50 years there is still considerable work to be done in order to fully understand the relationship between aggregate activity and inflation.