VC: A method for estimating time-varying coefficients in linear models

Arbeitspapier

This paper describes a moments estimator for a standard state-space model with coefficients generated by a random walk. A penalized least squares estimation is linked to the GLS (Aitken) estimates of the corresponding linear model with time-invariant parameters. The VC estimator is a moments estimator that does not require the disturbances be Gaussian, but if they are, its estimates are asymptotically equivalent to maximum likelihood estimates. In contrast to Kalman filtering, no specification of an initial state or an initial covariance matrix is required. While the Kalman filter is one-sided, the VC filter is two-sided and uses more of the available information for estimating intermediate states. Further, the VC filter has a clear descriptive interpretation.

Object Details

• Language: Englisch

References and Relationships

• Bibliographic citation: Series: Munich Discussion Paper ; No. 2019-3

Classification and Topics

• Classification: Wirtschaft ; Single Equation Models; Single Variables: Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes ; Model Construction and Estimation ; Model Evaluation, Validation, and Selection
• Subject: time-series analysis ; linear model ; state-space estimation ; time-varying coefficients ; moments estimation ; Kalman filtering ; penalized least squares

Contributors, Places and Time

• Event: Geistige Schöpfung
• (who): Schlicht, Ekkehart
• Event: Veröffentlichung
• (who): Ludwig-Maximilians-Universität München, Volkswirtschaftliche Fakultät
• (where): München
• (when): 2019

Further information

• DOI: doi:10.5282/ubm/epub.59143
• Handle: http://hdl.handle.net/10419/224120
• URN: urn:nbn:de:bvb:19-epub-69765-2
• Last update: 10.03.2025, 11:43 AM CET