AbstractsBusiness Management & Administration

Abstract

Ocean shipping of dry bulk commodities interacts in a market with a high degree of instability. According to Stopford (1997): "Forecasting has a poor reputation in maritime circles; it is never right." Some of the experts within the shipping theory state that dry bulk shipping charter rates are more or less unpredictable, and because of that it is extremely difficult to know when to make large investments, such as buying new ships. Are they right, or can ARIMA or ARIMAX models be able to do good forecasts of the Baltic Supramax Index (BSI)? The BSI is a weighted average of quotations by leading broking companies around the world of trip charter rates on a number of widely traded benchmark routes. ARIMA modelling is a well-known and much-used tool for forecasting time series. In this thesis, I compare forecasts of the log of the Baltic Supramax Index using a univariate model, ARIMA(1,1,1); and a multivariate model, ARIMAX(1,1,1), with the log of the McCloskey World Steam Coal Price Index as the explanatory variable. The accuracy of the forecast of these two models did not differ significantly. The one-step ahead forecasts were quite good, but the dynamic forecast did not result in so good short-term predictions as hoped. The Baltic Supramax Index is a time series with a significant trend, especially after the dry bulk boom in 2003, and that makes the series really hard to predict. This paper can be divided into four main parts. The first part of the paper gives an introduction of the shipping market in general, Western Bulk, the dry bulk market in particular, the time series and the dry bulk boom. The second part explains the Box-Jenkins methodology and the modelling with ARIMA and ARIMAX models. The third part present the process of the model selection. The last part explains the forecasting process, followed by a comparison of the forecasts that were made from the chosen models.