Air demand forecasting for passengers and freight in Italy: A comparison of two statistical models

Air transport forecasting has received significant attention in the literature. Furthermore, economic growth and population are significantly associated with the aviation industry. Moreover, the time series of air passenger and freight demand usually exhibit a complex behaviour with high volatility and irregularity, particularly when considering the economic factors associated with freight demand. In this research, we implemented two different statistical methods, namely the SARIMAX model and the structural time series approach, to fit and forecast both the air passenger and freight demand, considering economic variables and population as regressors. Both methods can deal with seasonality and trend; interestingly, the structural time series model can also estimate the cycle by decomposing the time series using the Kalman filter method. We applied the two methods to monthly data obtained from the Italian national website Assaeroporti for the period from January 2000 to December 2023. We computed predictions for the passenger and freight demand up to 2035, with a monthly and yearly resolution. For this aim, it was necessary to implement separate time series models for the economic regressors and population to plug in corresponding forecasts in the demand models. The results could be particularly useful for optimizing air traffic infrastructure and guiding strategic investment, particularly in the planning and adoption of sustainable aviation technologies (e.g., electric and hybrid-electric systems, new sustainable fuels).