SMU SOE Seminar (November 25, 2022): Spectral Centroid Targeting with the HP Filter

Please click here if you are unable to view this page.

TOPIC:

SPECTRAL CENTROID TARGETING WITH THE HP FILTER

ABSTRACT The Hodrick-Prescott (HP) filter is one of the most widely used trend-cycle decomposition technique in applied macroeconomic research, and its one-sided or real-time version has recently gained popularity as its use is recommended by Basel III regulations in calibrating the countercyclical capital buffer. The performance of the filter, like all nonparametric methods, depends critically on a tuning parameter that controls the degree of smoothing. However, most of the empirical practice with HP filter follows a naive ‘golden rule’ to set this smoothing parameter as a single number, 1600, when quarterly data is used. In this paper, we provide a systemic asymptotic analysis on how the consistent trend estimation/removal of two-sided and one-sided HP filters depends on the smoothing parameter for a broad class of data generating processes, and then propose a data-dependent approach to determine the tuning parameter. In particular, we show that the limiting processes of the filtered trend and cycle components can be expressed as weighted averages of sinusoids where the weights depend explicitly on the tuning parameter. Then, we introduce a measure named `spectral centroid’ which effectively characterizes the central frequency/period of a covariance stationary process and derive a formula relating the spectral centroid of the filtered cycle and the smoothing parameter. This formula allows us to solve for the smoothing parameter from the targeted central period of the filtered cycle and the actual data. We show in empirical analysis that the proposed method provides a useful conversion of smoothing parameter when the HP filter is used in various types of tasks where researchers may be interested in targeting cycles with different lengths in their estimation (e.g., it is known that financial cycles are often much longer than business cycles), or where the real-time version of the filter is preferred to its more well-known two-sided version. Click here to view the CV.
		PRESENTER Ye Lu University of Sydney
		RESEARCH FIELDS Econometric Theory Time Series Analysis Financial Econometrics
		DATE: 25 November 2022 (Friday)
		TIME: 1.00pm - 2.30pm
		VENUE: Meeting Room 5.1, Level 5 School of Economics Singapore Management University 90 Stamford Road Singapore 178903

© Copyright 2024 by Singapore Management University. All Rights Reserved.