Constructing a Markov-switching turning point index using mixed frequencies with an application to French business survey data

Abstract

This paper proposes an indicator for detecting business cycles turning points incorporating mixed frequency business survey data. It is based on a hidden Markow-Switching model and allows for the detection of regime changes in a given economy where information is displayed monthly, bimonthly and quarterly. Adapting existing indicators such as Hamilton (1989) and Gregoir and Lenglart (2000) to this frequency mix constitutes the main contribution of the present work. The proposed methodology is applied to the French economy. Using balances from different business surveys, this indicator measures the probability of being in an accelerating or a decelerating phase. The indicator is compared over the past with a reference dating established upon the business cycle component of GDP e xtracted by a Christiano-Fitzerald filter. It exhibits quite clearly and timely regimes changes of the French outlook. In this case the mixed frequency methodology adapted from Gregoir and Lengart yields better performance than the Hamilton-based indicator. Considering the adequacy with the reference dating over the past, the French turning point index (TPI) provdies an accurate signal on the current outlook.