The aim of this paper is to investigate the existence and nature of seasonality (deterministic or stochastic) in tanker freight markets and measure and compare it across sub-sectors and under different market conditions (expansionary and contractionary) for the period January 1978 to December 1996. The existence of stochastic seasonality is rejected for all freight series while results on deterministic seasonality indicate increases in rates in November and December and decreases in rates from January to April. Seasonality is found to be varying across markets depending on vessel size and market condition. Seasonality comparisons under different market conditions, an issue investigated for the first time in the econometrics literature using Markov Switching models, reveal that seasonal rate movements are more pronounced when the market is recovering compared to smaller changes when the market is falling. This is well in line with the low and high elasticity of supply expected in expansionary and contractionary periods of shipping markets. The results have implications for tactical shipping operations such as budget planning, timing of dry-docking, vessel speed adjustments and repositioning. As expected, the out-of-sample forecasting performance of these Markov Regime Switching models is lacking somewhat, a result which is thought to be a consequence of having to predict ‘states’ simultaneously with mean values.