An exploratory study on the effect of trade data aggregation on international freight mode choice

As part of a comprehensive regional or global maritime trade modeling framework, it is often necessary to determine the modal share of commodity flow between countries. Typically, this is resolved by either using a constant empirical value (such as the commodity value-weight ratio) or applying statistical models to determine mode choice or share. Whilst the latter method tends to provide a more rational solution, it is constrained by the appropriate choice of trade data classification level during model development. This paper therefore explores the potential impact of trade data aggregation (in terms of trade data classification level) on commodity mode choices derived from discrete choice models. Using international commodity flows between continents, discrete choice models can be developed to compare the modal split between air and sea transport. By considering four different international trade routes and three trade classification levels, the paper shows that detailed commodity trade classification may not necessarily lead to better results. Using the Harmonized Commodity Description and Coding System (or HS System) as an example, it was found that it may be more effective developing models using HS two-digit commodity trade data for the purpose of international freight mode choice studies for the purpose of regional or global maritime trade modeling.