There are several ways to model seasonality in a time series. Traditionally, trend-cycle decomposition such as the Holt-Winters procedure has been very popular. Also, until today applied researchers often try to account for seasonality by using seasonal dummy variables. But of course, in a stochastic process it seems unreasonable to ...