Volatility modelling in the shipping industry

Abstract

This thesis focuses on volatility estimating, modelling, and forecasting by using daily data from the shipping freight market. The topic is studied via two chapters, each with a different focus. After introducing the methodologies to be adopted in this thesis in Section 2, Section 3 introduces the first chapter which explores the performance of three types of mathematical models in volatility prediction by running a horserace among the forecasting precision of five different models, including the Two-component model, the traditional GARCH model, the GJR-GARCH model, the EGARCH model, and the heterogeneous autoregressive (HAR) model. In addition, the forecasting precision is evaluated via 4 different metrics, RMSE, MAE, MAPE, and Mincer-Zarnowitz regress test’s R2 . Using the data of seven indices of the shipping freight rate index, ranging from August 1998, to August 2013, this chapter finds that the Two-component model has the best prediction power for the first five indices, while the HAR model ranks second. In addition, the GARCH model has better forecasting capability than the EGARCH model. This is confirmed by Hansen (2005)’s superior predictive ability (SPA) test and Diebold and Mariano (1995)’s pair-wise test. The second chapter investigates whether the oil index affects the modelling and fore casting of the volatility dynamics for the seven different shipping indices. Using the same dataset as in the first chapter, the addition of the Brent oil index or the West Texas Inter national (WTI) crude oil index is incorporated into the HAR as well as the GARCH-family models to explore whether oil index plays a statistically significant role in volatility mod elling and forecasting of the seven shipping freight indices. The in-sample estimation results show that incorporating the oil index could better specify the volatility dynamics for the ii shipping freight indices especially for the HAR-X model. Furthermore, the augmented HAR and GARCH-family models are found to outperform their original models in predicting the future volatility. Overall, the results of my thesis aim to provide linkage between crude oil prices and shipping index returns that benefits short-term participants, such as ship owners, charterers, and operators. Accordingly, the use of the appropriate model leads to better fore casting results, thus other players who indirectly participate in the shipping industry, such as investment banks and policy makers, can also optimize their strategies in this market.

Date of Award	Nov 2021
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Xiaoquan Liu (Supervisor)