Old ideas may still be new: a system identification approach to tanker freight modelling

A system identification approach is used in order to model and identify the structural relationship between freight rates in the tanker industry and a set of exogenous inputs. Our motivation results from the limited data availability and the prohibitive theoretical complexity of economic models for the evaluation of managerial decisions and risk management. The combination of statistical analysis and economic insight leads to an innovative multidisciplinary approach for modelling competitive economic systems. We calibrate the model with real data from 1980–2002, achieve estimation and identification of the system and fully track the directional changes in freight rates. After conducting performance evaluation an innovative hybrid model is introduced and system performance is maximized both within and out-of-sample. Finally we discuss potential uses of this model for policy analysis, managerial investment decisions and risk management. Copyright © 2007 John Wiley & Sons, Ltd.