An Optimal Forecasting Method of Passenger Traffic in Greek Coastal Shipping
An Optimal Forecasting Method of Passenger Traffic in Greek Coastal Shipping

Summary/Abstract: Purpose: The main goal of this study is to exact an optimal forecasting method by answering the research question: which is the best model for capturing short-term seasonal components of passenger traffic in Greek coastal shipping? Design/methodology/approach: There are not a lot of scientific efforts in forecasting passenger traffic in Greece. In order to fill this gap, we tried to find an optimal forecasting method, by comparing Box-Jenkins ARIMA, smoothing and decomposition methods. As Greek coastal shipping consists of several concentrated submarkets (lines) we remained in fourteen popular itineraries (including total passenger traffic). Taking into consideration the high seasonality and no stationarity that characterizes those routes we limited our analysis to Winter’s triple exponential smoothing, to time series decomposition method, to simple seasonal model and to seasonal ARIMA models. Findings: The analysis results show that in fourteen popular coastal routes Winters’ multiplicative method, simple seasonal model and decomposition multiplicative trend and seasonal model have the best integration to the time series data. No coastal line led to better results by seasonal Box-Jenkins ARIMA models. Research limitations/implications: The results should be treated with caution since COVID-19 pandemic does not allow safe conclusions for the forecasting period 2020-2022 in GCS. However, the forecasting results of the first quarter of 2020, when pandemic had not fully prevailed, gave encouraging results with little deviations between predicted and actual values. Originality/value: Greek coastal shipping is one of the biggest in Europe serving a large number of passengers and having a large part of the total shipping fleet. It plays an important role for Greek economy and society, as it connects the majority of inhabited islands to mainland. The finding of an optimal forecasting method of passenger traffic is very significant for both business and government policy. Decisions on the number of routes served by shipping companies, on ships by coastal line (number and size), on companies' pricing policy, on public service obligations, on state port infrastructure policy and on the amount of state funding for barren lines are typical examples.

• Details
• Contents