THE ECONOMETRICS OF MACROECONOMIC MODELLING

Econometric evaluation of Nordic structural employment estimates

While early models treated the NAIRU as a quasi fixed parameter, cf. the open economy Phillips curve NAIRU of Chapter 4, the ICM framework provides the intellectual background for inclusion of a wider range of supply-side and socioeconomic structural characteristics. Such factors vary over time and across countries. There has been a large output of research that looks for the true struc­tural sources of fluctuations in the NAIRU. This includes the joint estimation of wage - and price-equations (see Nickell 1993 and Bean 1994 for surveys), but also reduced form estimation of unemployment equations with variables rep­resenting structural characteristics as explanatory variables. However, despite these efforts, the hypothesis of shifts in structural characteristics have failed to explain why the unemployment rates have risen permanently since the 1960s; see Cross (1995), Backhouse (2000), and Cassino and Thornton (2002).

An alternative approach to the estimation of the NAIRU is based on some sort of filtering technique, ranging from the simplest HP filter, to advanced methods that model the natural rate and trend output jointly in a ‘stochastic parameter’ framework estimated by the Kalman filter (see Apel and Jansson 1999 and Richardson et al. 2000). A common assumption of these studies is that the (stochastic) NAIRU follows a random walk, that is, its mean does not exist. As discussed in Section 4.3, in connection with the Phillips curve NAIRU, this may represent an internal inconsistency, at least if the NAIRU is to represent the mean rate of unemployment in a dynamically stable system. However, proponents of the time varying NAIRU approach could claim that they capture the essence of the natural rate dichotomy, since only supply-side shocks (not nominal or demand shocks) are allowed to affect the estimated NAIRU process.

In this section, we show that the idea of a time varying NAIRU can be evaluated with conventional econometric methods.[47] The basic insight is that the amount of variation in the NAIRU ought to match up with the amount of instability that one can identify in the underlying wage - and price-equations. Because of its practical importance and its simplicity, we focus on OECD’s ‘NAWRU’ method.