THE ECONOMETRICS OF MACROECONOMIC MODELLING

The NAWRU

The NAWRU indicator has been used extensively by the OECD and others on several important issues, including policy evaluation and estimation of potential output and the structural budget balance; see Holden and Nymoen (2002) for a discussion. Elmeskov and MacFarland (1993) and Elmeskov (1994) define the non-accelerating wage rate of unemployment, NAWRU, in terms of a stylised wage-pressure equation

Д2^ = - ct(Ut - UNAWRU), ct > 0, (6.45)

where UNAWRU is the NAWRU level of unemployment. In words, it is assumed that wage inflation is affected in a linear way by the difference between the actual level of unemployment and the NAWRU. Equation (6.45) can either be
seen as a vertical wage Phillips curve (dynamic homogeneity is imposed); or as representing the heuristic dynamics of the wage curve model. The linear functional form is not essential, but is used in exposition and in applications of the method.

Based on an assumption that UNAWRU is unchanged between consecutive observations, (6.45) is used to calculate the parameter ct, for each observation separately

ct = - A3wt/AUt. (6.46)

Substituting the observation dependent parameter values ct back into (6.45) the NAWRU is calculated as:

UNAWRU = Ut - (AUt/A3wt)A2wt. (6.47)

In all four Nordic countries, actual unemployment has risen since the early 1970s, first in Denmark, more recently in the other countries. The raw NAWRU estimates as given by equation (6.47) are very volatile (see Holden and Nymoen 2002, figure 2), and published NAWRUs are based on HP filtering of these raw NAWRU estimates. Figure 6.2 records the NAWRUs that are cited in policy analysis discussions—see OECD Economic Surveys for Norway and Sweden, OECD (1997a,6).

Figure 6.2. Actual rates of unemployment (U) and NAWRUs for
the four Nordic countries

For all countries, the NAWRU estimates indicate a corresponding increase in structural unemployment. Hence accepting this evidence at face value, one is led to the conclusion that the rise in unemployment is associated with a structural change in the labour market. However, Solow’s 1986 critique of natural rates that ‘hops around from one triennium to another under the influence of unspec­ified forces ... is not natural at all’, clearly applies to NAWRUs.[48] Hence, in the following we investigate whether the dramatic changes in Figure 6.2 can be rationalised in a satisfactory way.