Air traffic modelling based on a long memory approach Air Traffic Modelling

Abstract

This paper deals with modeling air traffic data using a long memory class of models that use fractional integration. Two datasets have been considered: monthly global Revenue Passenger Kilometers and the number of monthly flights in Europe. The objective of this paper is to investigate whether Covid-19 has had a temporary or permanent impact on the air traffic trends. To do so, we investigate the orders integration of the series. Both datasets produced the same results: the trend was mean reverting when considering data before Covid-19, but the shock was so strong and long-lasting, that it produced a change to non-mean-reversion results after Covid-19. That said, if it is desirable to bring the air traffic trend back to its values before Covid-19, it will require intervention on the part of authorities or external factors since the series will not return by themselves to their original long term projections.