RDP 8910: An Analysis of the Determinants of Imports 1. Introduction

In this paper the empirical relationship between import volumes, domestic activity and relative prices is examined. To date, most attempts to estimate this relationship have modelled the demand for imports as a function of domestic activity and a relative price term, using standard regression techniques. Such models have generally been unsuccessful in terms of their tracking performance, particularly in recent periods. Part of the problem may be the omission or inappropriate choice of explanatory variables. However, a major problem is that most previous studies of import demand in Australia have not taken explicit account of non-stationarities in the time series. The usual techniques of regression analysis may result in biased and inconsistent coefficient estimates when the variables in question are non-stationary. In this paper, however, we take explicit account of non-stationarities in the time series data by applying cointegrative techniques.^[1]

Section 2 outlines the broad trends and major cycles in imports, activity and relative prices in the 1970s and 1980s. Section 3 discusses the appropriate choice of explanatory variables and looks at alternative relative price variables and demand variables (both trend and cyclical). Our choice of functional form is explained in Section 4, with an outline of the methodology used in Section 5. Section 6 presents the results of the preferred equations. Some conclusions are offered in Section 7.

Footnote

Two papers which have used cointegrative techniques to model the demand for imports in Australia are Cairns (1989) and Hall et al. (1989). Preliminary work by Warwick McKibbin and Julie Cairns provided the initial impetus for this study. [1]