This article examines the long-run relationship between Malaysian aggregate imports and the components of final demand expenditure and relative prices using Johansen multivariate cointegration analysis. An error correction model is proposed to model the short-run response of imports to its determinants. Annual data for the period 1970 to 1998 are used. The longrun relationship between aggregate imports and the macroeconomic components of final demand expenditure namely public and private consumption expenditure, investment expenditure and exports, is investigated because the different components of final demand might have different import contents. If this were true, the use of a single demand variable in the aggregate import demand function would lead to aggregation bias. The results of the analysis show that the components of final demand expenditure and relative prices are all important in determining aggregate demand for imports in both the long-run and the shortrun. I. Introduction Many empirical studies on the behaviour of aggregate Malaysian imports have been carried out (Mohammad 1980; Semudram 1982; Awang 1988; MIER 1990; Tang and Mohammad 2000). An even more extensive literature exists on the estimation of various import demand elasticities for other countries (Heien 1968; Houthakker and Magee 1969; Hughes and Thirlwall 1979; Gafar 1988; Giovannetti 1989; Ziets 1993; Mah 1993 and 1994; Bahmani-Oskooee and Rhee 1997; Abbott and Seddighi 1996). The Malaysian studies may be categorized as single equation studies or as part of macroeconometric models of the Malaysian economy. Other features of the studies are as follows. First, a single demand variable is used in the import demand function, thus implicitly assuming that the import content of each component of aggregate demand is the same. Secondly, the specification adopted for the import demand function is the log-linear form together with a partial adjustment model to describe the adjustment of actual imports to the desired longrun equilibrium imports. It is assumed that there is an underlying equilibrium relationship between the quantity of imports and the explanatory variables included in the model. Two issues may be raised here. The use of a single demand variable may lead to misspecification especially if the different components of aggregate demand expenditure have different import components. The use of partial adjustment framework imposes a restrictive lag structure on the regression equation. If the restriction is incorrect, it would lead to dynamic misspecification and predictive failure (Abbott and Seddighi 1996). In addition, if the underlying assumption that there exists a longrun relationship is false, attempting to estimate the traditional formulation is invalid (Mah 1994, p. 291). The primary objective of this article is to estimate the Malaysian aggregate import demand function using cointegration and error correction modelling methods. The long-run relationship between aggregate imports and relative prices is investigated using Johansen multivariate cointegration analysis. The previous studies, with the exception of Tang and Mohammad (2000) were undertaken before it was fashionable to apply the methodology of cointegration and error correction modelling (ECM). Recent advances in the time series methodology allow us to test for the presence of an equilibrium relationship between the variables that appear in the import demand function. Engle and Granger (1987) have shown that even though an economic time series may wander over time, there may exist a linear combination of the variables that converges to an equilibrium, that is, the variables are cointegrated. Examples of application of cointegration and ECM to import demand studies are found in Mah (1993, 1994), Abbott and Seddighi (1996) and Bahmanee-Oskooee and Rhee (1997). This study is justified by the following considerations. First, the different components of the final demand expenditure have different import contents. …