Linear and non linear growth models using mixed modeling: An application on European import volumes

The purpose of this paper is the identification of appropriate growth models for trade volumes as a policy tool. The methodology utilized is based on linear and nonlinear mixed modeling. The specifications tested are the linear, the exponential, the logarithmic and the logistic model. The focus lies on the imports of Europe from the world. We present two pilot cases corresponding to different levels of aggregation in terms of country groups and product categories, thus emphasizing the differences between aggregate and disaggregate approaches. The core econometric finding suggests that no clear superiority can be attributed to a single growth model specification on either level of aggregation. Therefore, the implications of each specification on policy decision making is discussed and a recommendation on the use of such models for policy making is made. The growth models are further employed for the purpose of trend extrapolation, to initiate a discussion on the role and responsibility of transport policies implemented today based on alternative future scenarios 20 years ahead.