To date, there has been little investigation of the impact of seasonal adjustment on the detection of business cycle expansion and recession regimes. We study this question both analytically and through Monte Carlo simulations. Analytically, we view the occurrence of a single business cycle regime as a structural break that is later reversed, showing that the effect of the linear symmetric X-11 filter differs with the duration of the regime. Through the use of Markov switching models for regime identification, the simulation analysis shows that seasonal adjustment has desirable properties in clarifying the true regime when this is well underway, but it distorts regime inference around turning points, with this being especially marked after the end of recessions and when the one-sided X-11 filter is employed.
business cycles Markov switching models Seasonal adjustment