This paper presents an econometric approach to model the dynamics in the air cargo market. Data published by the national customs is used to forecast the demand for air cargo per commodity and origin to destination-flow in Europe and the United States. Univariate techniques such as the Weighted Majority Algorithm, neural networks and the VECM are applied to model the macroe-conomic effects on air cargo demand. We extend our insight in the demand dynamics by modeling relations between demand flows using multivariate models. The use of multivariate techniques include the FAVAR and Bayesian Variable Selection on a time-varying parameters VAR-model. The multivariate techniques generate significantly better forecasts than known forecasts in the air freight industry.
AGIFORS 54th Annual Symposium: An Industry in Transformation
Department of Econometrics

Versnel, L. (2014). Forecasting air cargo demand. Presented at the AGIFORS 54th Annual Symposium: An Industry in Transformation. Retrieved from