Port congestion and container freight rate dynamics: forecasting with an RBF neural network

Abstract

With its safe and efficient characteristics, container transportation has become vital for advancing the global economy. However, port congestion has become a significant obstacle to the container freight price system’s stability. There is currently no dependable engineering solution to guarantee the stability of the maritime transport system in a port congestion scenario. Therefore, decision-makers must comprehend the changing characteristics of the container freight index in the context of port congestion. Using the Shanghai container freight index as a proxy, this paper investigated the effect of port congestion on container freight rates, proposing a container freight index forecasting model. This study compiled congestion data from the Shanghai, Busan, Los Angeles, and New York ports from January 1, 2016 to January 1, 2023, to predict a Shanghai container freight index (SCFI). With its high-precision fitting effect, the RBF neural network effectively predicted the change in SCFI, and the R2 reached 96%. We also confirmed the transfer effect of SCFI using the time-lag correlation model in a large congestion environment. The research results give container shipping organizations a decision-making foundation for planning shipping strategies and mitigating market risk.