City Research Online

Abstract

This thesis investigates three salient areas of interest in the structure of freight rates in the shipping market, with a particular focus on the tanker and dry-bulk sectors, using recent econometric and time series techniques. The questions asked are: 1) do spot freight rate levels follow a fractionally integrated process, as opposed to being stationary or non-stationary, as had previously been proposed; 2) does spot freight rate volatility also follow a fractionally integrated process; and 3) do freight rates exhibit conditional skewness and kurtosis? It then evaluates the impact that these factors have on the risk exposure of market participants. These concepts are further tested in terms of their respective forecasting performance, relative to other more standard econometric techniques. An ongoing issue in the shipping literature is whether spot freight rate levels follow a stationary or non-stationary process. This thesis provides another dimension to this discussion by arguing that spot freight rate levels follow a fractionally integrated process. The rationale behind this argument is the fact that the supply and demand dynamics in this market mean that although freight rates are mean-reverting overall, the process of mean-reversion occurs with a delay, which is exactly how one would expect a fractionally integrated process to behave. Although in-sample results were promising in that fractionally integrated models are found to outperform their stationary and non-stationary counterparts across sectors and vessel sizes, out-of-sample forecasts indicate that models that assumed stationarity or non-stationarity outperformed these models, depending on the sector and vessel size. Additionally, the thesis extends this debate to the volatility of these spot freight rate levels, where it is proposed that volatility also follows a fractionally integrated process. In-sample results from the estimation of Generalised Autoregressive Conditional Heteroscedasticity (GARCH), Integrated Generalised Autoregressive Conditional Heteroscedasticity (IGARCH) and Fractionally Integrated Generalised Autoregressive Conditional Heteroscedasticity (FIGARCH) models indicate that FIGARCH models outperformed the other two models across all sectors and vessel sizes, however, when calculating the respective out-of-sample Values-at-Risk for each 18 vessel type, non-parametric models are found, in most cases, to outperform their parametric counterparts across sectors and vessel sizes. This thesis finally examines whether freight rates exhibit conditional skewness and kurtosis, where the shape of the supply function in the shipping freight markets indicates that these would not be constant over time, as is assumed by other standard models. Results for the in-sample period indicate that the Generalised Autoregressive Conditional Heteroscedasticity with Skewness and Kurtosis (GARCHSK) models outperformed GARCH and FIGARCH models. This being said, when calculating the respective out-of-sample Values-at-Risk for each vessel type, non-parametric models are found, in most cases, to outperform their parametric counterparts across sectors and vessel sizes.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HB Economic Theory
Departments:	Bayes Business School Doctoral Theses Bayes Business School > Bayes Business School Doctoral Theses

Export

Downloads