Change search
ReferencesLink to record
Permanent link

Direct link
Forecasting macroeconomic time series with locally adaptive signal extraction
Stockholm University, Faculty of Social Sciences, Department of Statistics.
2010 (English)In: International Journal of Forecasting, ISSN 0169-2070, Vol. 26, no 2, 312-325 p.Article in journal (Refereed) Published
Abstract [en]

We introduce a non-Gaussian dynamic mixture model for macroeconomic forecasting. The locally adaptive signal extraction and regression (LASER) model is designed to capture relatively persistent AR processes (signal) which are contaminated by high frequency noise. The distributions of the innovations in both noise and signal are modeled robustly using mixtures of normals. The mean of the process and the variances of the signal and noise are allowed to shift either suddenly or gradually at unknown locations and unknown numbers of times. The model is then capable of capturing movements in the mean and conditional variance of a series, as well as in the signal-to-noise ratio. Four versions of the model are estimated by Bayesian methods and used to forecast a total of nine quarterly macroeconomic series from the US, Sweden and Australia. We observe that allowing for infrequent and large parameter shifts while imposing normal and homoskedastic errors often leads to erratic forecasts, but that the model typically forecasts well if it is made more robust by allowing for non-normal errors and time varying variances. Our main finding is that, for the nine series we analyze, specifications with infrequent and large shifts in error variances outperform both fixed parameter specifications and smooth, continuous shifts when it comes to interval coverage.

Place, publisher, year, edition, pages
2010. Vol. 26, no 2, 312-325 p.
Keyword [en]
Bayesian inference, Forecast evaluation, Regime switching, State space modeling, Dynamic mixture models
National Category
Probability Theory and Statistics
URN: urn:nbn:se:su:diva-49253DOI: 10.1016/j.ijforecast.2009.12.011ISI: 000276707900007OAI: diva2:380178
authorCount :2Available from: 2010-12-20 Created: 2010-12-13 Last updated: 2010-12-20Bibliographically approved

Open Access in DiVA

No full text

Other links

Publisher's full text

Search in DiVA

By author/editor
Villani, Mattias
By organisation
Department of Statistics
In the same journal
International Journal of Forecasting
Probability Theory and Statistics

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 56 hits
ReferencesLink to record
Permanent link

Direct link