T.P. Moschovou, A.G. Giannopoulos
This paper contains an investigation of the COVID-19 impacts on freight flows and the handling of uncertainty in freight forecasting models, based on data from Greece. It collects and analyses, over a 7-year period before and during the pandemic, data for freight transport operations and some related factors in order to macroscopically examine any statistically significant changes in their values over time. This period was judged necessary in order to establish the pattern of fluctuations in the relevant data during the non-pandemic years and thus make the visual comparison with the previous period and the years during the pandemic, more clear. First, the paper tests the impact of the pandemic as expressed by the number of daily COVID-19 cases on freight flow variables in order to find the dynamic behavior of these variables and trace their reactions over time. This analysis is made by using the Vector Autoregressive Model (VAR). By implementing VAR modelling, we analyzed the dynamic relationship between freight transport volumes and other factors such as GDP, the industrial production index, exporting transactions and the number of coronavirus cases. The main result of the model analysis and the employment of impulse response functions revealed that the unexpected shock of COVID has a negative reaction to the economy and the freight transport volumes and a rather short-term limited duration disruption effect on the growth of exports as well as on the industrial production index, of approximately eight months. Secondly, the paper discusses how, unpredicted events like the pandemic, influence the uncertainty inherent in freight transport modelling and formulates a novel freight modelling framework procedure based on scenario building, regular monitoring and data updates on a permanent basis.
Keywords: coronavirus; freight transport; vector autoregressive (VAR) models; uncertainty; freight forecasting