Detecting turning points in the inflation cycle

Based on a non-parametric method we construct a leading indicator for turning points in the inflation cycle. We extract the cycle by a band pass filter and identify the turning points by the Bry Boschan algorithm. The selection of variables in the leading indicator is based on turning point matching and dynamic correlation, while the weights are adjusted for cross-correlation between the variables. We apply the method to core inflation in the Netherlands and the euro area. We show that turning points in the core inflation cycle explain inflation forecast errors. Moreover, we show that since the pandemic the cycle of core inflation has mainly been driven by external factors and that the inflation cycle will remain below trend well into 2024.

Keywords: Euro area inflation; Trend-cycle decomposition; Leading indicators; (Turning point) forecasts; Real-time analysis
JEL codes C82; E31; E37; E58

Working paper no. 808