Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.567025
Title: Essays on the econometric modelling and forecasting of shipping market variables
Author: Pourkermani, Kasra
Awarding Body: University of Newcastle Upon Tyne
Current Institution: University of Newcastle upon Tyne
Date of Award: 2012
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
This thesis uses econometric modelling and forecasting to investigate a number of important topics associated with economic and financial aspects of the global shipping market. The thesis is made up of five chapters. Chapter 1 introduces the structure of the shipping market; it covers a wide range of topics, including the shipping sub-markets, shipping stock and shipping market information. It introduces the different types of freight rates involved, and discusses the economics behind the formation of spot and time-charter freight rates. It also introduces the new-build ship market and explains some of the different shipbuilding models. In addition, it discusses the market for second-hand ships. Finally, it reports and discusses the correlations of different shipping variables with each other and with the S&P500 stock market index. Chapter 2 focuses on forecasting the freight rate for ship operators. Since time-charter rates depend on market participants’ expectations about future spot rates, under market efficiency the ship operator should not be able to make abnormal profits by choosing a specific chartering strategy. The chapter investigates whether this is true by exploring the economic value of freight rate forecasts, using a regression-based recursive switching approach based on two sets of macroeconomic and commodity data. The ship operator is assumed to allocate the ship between a trip-charter and time-charter market according to forecasts of the quarterly excess freight rate. The Handymax and Capesize classes of ship are analysed, the analysis showing that this type of investment strategy does not generate significantly abnormal profits for the Handymax class, but does for the Capesize class. Forecasting with commodity variables is more profitable than forecasting with macroeconomic variables. Chapter 3 quantifies and discusses the volatility of index returns in the dry bulk freight rate market for freight traders and investors. The daily freight rate indexes of three ship classes, Baltic dry index (BDI), Baltic Panamax index (BPI) and Baltic Capesize index (BCI) from 14 January 2000 to 14 January 2010 are analysed. Some of the findings from applying variations of autoregressive conditional heteroskedasticity (ARCH) models suggest that the volatility of shocks is very persistent and that a unit root might exist in the conditional variance. No evidence of any asymmetry in the conditional variance is found. Volatility forecasting for one day ahead and multiple days ahead is also performed using a variety of ARCH models. At the end of the chapter the risk exposure of the freight rate index is assessed using the Value at Risk (VaR) technique. In Chapter 4 it is argued that if risk premiums are time-varying and correlated with macroeconomic variables, macroeconomic variables might have forecasting power for shipping stock returns. This issue is investigated using the recursive regression-based approach of Pesaran and Timmermann (1995) and it is concluded that allowing for different combinations of macroeconomic variables generally does not help forecasting. This may be because the model selection criteria do not seem to work efficiently when there is a structural break in the data. The model which includes all variables (AV) is found to be the best performing model. A data set is employed which includes four shipping stocks and the S&P500 index for comparison, and this shows that a trading strategy using the AV model generates 93% to 500% more wealth than a buy-and-hold strategy. When the explanatory variables are analysed individually, the US Treasury bill and NYMEX oil price are shown to have the most forecasting power. Chapter 5 concludes the thesis. It presents a review of the original findings and puts forward recommendations for future research.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.567025  DOI: Not available
Share: