Modeling the US demand for imports through cointegration and error correction

The purpose of this paper is to provide new estimates of the aggregate demand for both total and non-oil merchandise imports of the United States over the last two decades. To this aim we have applied recent and by now widely used cointegration techniques in order to estimate the long-run equilibrium relationship. Then we have imposed this long-run result as an error-correction mechanism to the dynamic model. The estimate of the long- and short-run income and price elasticities of imports is important to assess the potential role of the United States as ''engine'' of world growth, as suggested by the ''locomotive'' theory. Furthermore, it is important to evaluate whether and to what extent the United States faces an external constraint on growth and how exchange-rate movements affect the trade balance. This paper is structured as follows: after a brief introduction, Section 2 provides a review of the theoretical framework of the imperfect substitutes model on which ground lies our estimates. In Section 3 unit root properties of the data, the Engle-Granger two-step procedure and the Johansen methodology for estimating and testing cointegrating vectors are presented. Section 4 contains the econometric evidence on short-run income and price elasticities. In Section 5 the analysis is restricted to import demand of non-oil products. Empirical results are summarized in the last section.