A study of trip and time charter freight rate indices: 1968--2003

This paper addresses the question of whether the time series for the trip and time charter indices between 1968--2003 and 1971--2003 are identically and independently distributed, and whether they have nonlinear dependence. The test method used was based upon the work of Brock, Dechert and Scheinkman and is called the BDS test. The results show that the indices were not random and identically and independently distributed, and indeed nonlinear dependence exists. As a consequence of this result, linear and other traditional models are not suitable for modelling the distributions; the distributions for the indices are not normal, but instead show skewness, kurtosis and non-periodic cycles, all of which we have provided estimates for. Putting the results of the present study together with previously published papers by the same authors, which have covered a range of distributions including the second-hand shipping market, the trip charter and time charter markets, it has now been demonstrated that distributions that are frequently studied in marine economics show fat tails and long-term memory. Non-normality in these distributions has been identified by other authors in previous studies also. However, previous authors have not suggested any alternative to the conventional methods of analysis and modelling. The authors of the present paper propose the use of the rescaled range analysis (Hurst exponent) which was brought into prominence by Benôit Mandelbrot. It is further proposed that future work in this field should be directed towards forecasting.