An econometric analysis of the forward freight market

The success or failure of a derivatives (futures or forward) contract is determined by its ability to perform its economic functions efficiently, and therefore, to provide benefits to economic agents, over and above the benefits they derive from the spot market. These economic functions are price discovery and risk management through hedging. A considerable amount of empirical research has been directed towards examining these functions in different financial and commodity derivatives markets. The evidence however, on the over-the-counter FFA market is very limited. This thesis therefore, by investigating these issues provides new evidence in the literature for a forward market with some unique characteristics such as the trading of a service. Our empirical results can be summarised as follows. First, the FFA contracts perform their price discovery function efficiently since forward prices contribute to the discovery of new information regarding both current and expected spot prices. Furthermore, most FFA contracts contribute in the volatility of the relevant spot rate, and therefore, further support the notion of price discovery. Second, the introduction of FFA contracts has not had a detrimental effect on the volatility of the underlying spot market. On the contrary, it appears that there has been an improvement in the way that news is transmitted into prices following the onset of FFA trading. Third, FFA prices fail to reduce market risk to the extent evidenced in other markets in the literature and, hence, the FFA market does not perform its risk management function satisfactorily; this is thought to be the result of the lack of the cost-of-carry arbitrage relationship of storable assets that keeps spot and derivatives prices close together. Fourth, there seems to be a positive relationship between bid-ask spreads and expected price volatility in most FFA trading routes. Finally, in the routes where the cointegrating vector is restricted to be the lagged basis, the VECM generates more accurate forecasts than the VAR model and in the routes where the cointegrating vector is not restricted to be the lagged basis the VAR generates more accurate forecasts than the VECM model.