Regional and Sectoral Concerns

In the previous chapter on the Phillips curve we indicated that some of the problem came from the focus on aggregate unemployment without regard to its distribution over the regions of Europe. In this chapter we tackle that issue directly and whether the dispersion of output growth or unemployment rates within a country has a direct effect on the determination of inflation for the country as a whole in the context of a nonlinear relationship. This discussion is by no means the first in this area. Nonlinearity of the Phillips curve has been tested in numerous analyses (see Laxton et al. (1995), Laxton et al. (1999) and Linzert (2005) among others). The whole issue itself is also quite old. Lipsey (1960) remarks (p. 19) ‘If one wishes to predict the rate of change of money wage rates, it is necessary to know not only the level of unemployment but also its distribution between the various markets of the economy.’ (emphasis in original). While Lipsey does not attempt any estimates, Archibald (1969) offers some for the UK, where the variance of both regional and industry unemployment are shown to have a positive effect on wage inflation. Extending this to the US gives more problematic results with quarterly data.1 However, it is Brechling (1973) who introduced the nonlinear aggregation hypothesis which basically formulates the problem and suggests ways of testing the proposition. He also carried out some empirical tests with the US data. The results of the tests were somewhat disappointing from the hypothesis’ point of view and, maybe, therefore explain why the aggregation case has not been revisited with any intensity since.2