Canonical correlation of shipping forward curves

Abstract

The behavior and interrelations between the main shipping forward curves are analyzed using multivariate statistics after removing the volatility distortions dictated by the Samuelson hypothesis. Principal Components Analysis and Canonical Correlation analysis were used to demonstrate how the task of explaining the various shipping forward curves can be simplified substantially and how very high correlations can be achieved between shipping forward curves. The conditions under which correlations are higher are discussed as well as the various applications of these results using case studies. Applications include trading from a hedge fund perspective, cross hedging any physical exposure in illiquid markets and portfolio optimization. Conditioning as a tool is also examined to demonstrate how more reliable correlation results can be obtained for cross hedging or other purposes, and how the best trading opportunities can be unveiled conditional on recently observed data. Tanker valuations are carried out using the adjusted forward curves with the RAFL ship valuation model. The results are very close to transaction prices for relatively modem vessels while deviations in older ships are explained with regards to phase out regulations and other factors. The ship value volatility and consequently the valuations of typical options are substantial and increase as a percentage of the ship value with age. These results have to be considered seriously in shipping transactions that include optionalities which are very common.