The Impact of Seasonality on Sensitivity of Frequent and Occasional Motorway Users

This study investigates the impact of seasonality on motorway travel demand and how motorway users' behaviour responds to changes in travel demand in the previous year, toll prices, fuel prices, the Croatian GDP per capita, the European Union GDP per capita, and tourist arrivals in Croatia between summer and non-summer periods. In addition, with a focus on car transactions, this study also distinguishes how different types of motorway users (frequent and occasional motorway users) change their travel demand patterns concerning changes in the explanatory variables throughout the year. The study employs unique historical traffic and toll price data collected from one of the Croatian motorway concessionaires between 2014 and the end of 2019, with a total sample of 2,520 observations. Using the Weighted Least Squares (WLS) method, this study employs separate models to estimate the elasticities for each type of motorway user and season. Even Chow tests suggest structural differences between models, the results show a significant and positive impact of travel demand in the previous year on travel demand in the current year in all the models. This aligns with the historical records of Croatian motorway traffic and tourist arrivals. The remaining explanatory variables show that frequent motorway users negatively respond to increased fuel prices with higher sensitivity in the non-summer period than in the summer period. Occasional motorway users also exhibit a negative response to the increase in fuel prices on travel demand in the summer, with an insignificant impact during non-summer, indicating other factors may affect travel demand. Unlike extant research’s findings, this study’s results show that an increase in toll prices positively affects occasional motorway travel demand during the non-summer period. However, the small coefficient magnitude and traffic growth despite toll increases suggest motorway travel demand is weakly responsive to toll prices. The positive relationship may reflect occasional users as more tolerant of higher costs during the season of lower congestion, indicating that tolls have not reached their maximum levels. For robustness checks, the OLS findings are consistent with the WLS results despite heteroscedasticity issues. The remaining tests of GMM, FE, and RE models suggest GMM DIFF method as the most reliable for assessing three models, but with concerns about instrument validity in Model 3. Besides providing the first evidence of travel demand on the Croatian motorways, the main contribution of this study to the literature and practice is showing that consideration of seasonality is an essential parameter in transportation knowledge since the same type of motorway user has different sensitivities towards travel demand in the previous year, toll and fuel prices during the year. Also, the thesis provides new insight into the importance of separate observation of motorway users by frequency of usage due to their different sensitivity levels concerning changes in explanatory variables.