THE ECONOMETRICS OF MACROECONOMIC MODELLING

Inflation in open economies:. the main-course model

The chapter introduces Aukrust’s main-course model of wage - and price­setting. Our reconstruction of Aukrust’s model will use elements both from rational reconstructions, which present past ideas with the aid of present-day concepts and methods, and historical reconstructions, which understand older theories in the context of their own times. Our excursion into the history of macroeconomic thought is both traditional and plural­istic. The aim with the appraisal is to communicate the modernity of the set of testable hypotheses emerging from Aukrust's model to interested practitioners.

3.1 Introduction

As noted in the introductory chapter, an important development of macro­econometric models has been the representation of the supply side of the economy, and wage-price dynamics in particular. This chapter and the next three (Chapters 4-6) present four frameworks for wage-price modelling, which all have played significant roles in shaping macroeconometric models in Norway, as well as in several other countries. We start in this chapter with a reconstruc­tion of Aukrust’s main-course model of inflation, using the modern econometric concepts of cointegration and causality. This rational reconstruction shows that, despite originating back in the mid-1960s, the main-course model resembles present day theories of wage formation with unions and price-setting firms, and markup pricing by firms.

In its time, the main-course model of inflation was viewed as a contender to the Phillips curve, and in retrospect it is easy to see that the Phillips curve

won. However, the Phillips curve and the main-course model are in fact not mutually exclusive. A conventional open economy version of the Phillips curve can be incorporated into the main-course model, and in Chapter 4, we approach the Phillips curve from that perspective.

The main-course model of inflation was formulated in the 1960s.[12] It became the framework for both medium-term forecasting and normative judgements about ‘sustainable’ centrally negotiated wage growth in Norway.[13] In this section we show that Aukrust’s (1977) version of the model can be reconstructed as a set of propositions about cointegration properties and causal mechanisms. The reconstructed main-course model serves as a reference point for, and in some respects also as a corrective to, the modern models of wage formation and inflation in open economies, for example, the open economy Phillips curve and the imperfect competition model of, for example, Layard and Nickell (1986: Sections 4.2 and 5). It also motivates our generalisation of these models in Section 6.9.2.

Central to the model is the distinction between a sector where strong com­petition makes it reasonable to model firms as price takers, and another sector (producing non-traded goods) where firms set prices as markups on wage costs. Following convention, we refer to the price taking sector as the exposed sector, and the other as the sheltered sector. In equations (3.1)-(3.7), we, t denotes the nominal wage in the exposed (e) industries in period t. Foreign prices in domes­tic currency are denoted by pft, while qe, t and ae, t are the product price and average labour productivity of the exposed sector. ws, t, qs, t, and as, t are the corresponding variables of the sheltered (s) sector.[14] All variables are measured

(i = e, s).

qe, t = pft + (3.1)

pft = 9f + pft-1 + U2,t, (3.2)

ae, t = gae + ae, t-1 + u3,t, (3.3)

We, t - qe, t - ae, t = me + U4,t, (3.4)

ws, t = We, t + U5,t, (3.5)

as, t = 5as + as, t-1 + u6,t, (3.6)

Ws, t - qs, t - as, t = ms + U7,t. (3.7)

The parameters gi (i = f, ae, as) are constant growth rates, while (i = e, s) are means of the logarithms of the wage shares in the two industries.

The seven stochastic processes v1>t (i = 1,..., 7) play a key role in our reconstruction of Aukrust’s theory. They represent separate ARMA processes. The roots of the associated characteristic polynomials are assumed to lie on or outside the unit circle. Hence, u1t (i = 1,..., 7) are causal ARMA processes, cf. Brockwell and Davies (1991).

Before we turn to the interpretation of the model, we follow convention and define pt, the log of the consumer price index (CPI), as a weighted average

of qs, t and qe, t-

pt = Фqs, t + (1 - Ф)qe, t, 0 < ф < 1, (3.8)

where ф is a coefficient that reflects the weight of non-traded goods in private consumption. [15]