A Hodrick-Prescott filter with automatically selected jumps
11.07.2024
Paolo Maranzano (Department of Economics, Management and Statistics, University of Milano-Bicocca and Fondazione Eni Enrico Mattei); Matteo Pelagatti (Department of Economics, Management and Statistics, University of Milano-Bicocca)
C22, C63, E32, J21
Trend, State-space form, Unobserved component model, Structural change, LASSO, Business cycle, Employment
The Hodrick-Prescott filter is a popular tool in macroeconomics for decomposing a time series into a smooth trend and a business cycle component. The last few years have witnessed global events, such as the Global Financial Crisis, the COVID-19 pandemic, and the war in Ukraine, that have had abrupt structural impacts on many economic time series.
Moreover, new regulations and policy changes generally lead to similar behaviours. Thus, those events should be absorbed by the trend component of the trend-cycle decomposition, but the Hodrick-Prescott filter does not allow for jumps. We propose a modification of the Hodrick-Prescott filter that contemplates jumps and automatically selects the time points in which the jumps occur. We provide an efficient implementation of the new filter in an R package. We use our modified filter to assess what Italian labour market reforms impacted employment in different age groups.
***
Citazione suggerita: P. Maranzano, M. Pelagatti, ‘A Hodrick-Prescott filter with automatically selected jumps’, Nota di Lavoro 18.2024, Milano, Italia: Fondazione Eni Enrico Mattei.