Définition
Seasonal adjustment removes the effect of infra-annual periodic fluctuations due to calendar and seasonal effects, in order to highlight the most significant evolutions of the series, which are contained in the trend and the irregular component.
For instance, toys sales always rise between November and December, due to Christmas. On raw data, this periodic effect would mask the underlying evolution a given year. The seasonally adjusted series, from which the Christmas effect has been removed, can display a decrease in sales, corresponding to a grimmer year.