Forecasting Volatility of Tanker Freight Rates Based on Asymmetric Regime-Switching GARCH Models

This paper investigates the performance of various conditional volatility models to forecast the second moment of tanker freight rates. Justified by existing theoretical and empirical evidence, we focus on asymmetric Markov regime-switching models to study the major global routes for long-haul trade of crude oil during the sample period from June 2000 to May 2015. Moreover, in contrast to a number of existing studies, we examine seasonally adjusted freight rates. We find that regime-switching GARCH models outperform their single-regime complements in terms of in-sample fit and out-of-sample forecasting accuracy. In particular, the asymmetric MRS-EGARCH and MRS-APARCH exhibit superior in- and out-of-sample performance. To additionally examine the applicability in freight risk management, we compare Value-at-Risk and Expected Shortfall forecasts. Our results show that accounting for volatility regimes and asymmetry does not enhance the performance of one-day-ahead forecasts of either risk measure for both long and short trading positions.