Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/307949 
Authors: 
Year of Publication: 
2022
Citation: 
[Journal:] Empirical Economics [ISSN:] 1435-8921 [Volume:] 63 [Issue:] 5 [Publisher:] Springer [Place:] Berlin, Heidelberg [Year:] 2022 [Pages:] 2655-2674
Publisher: 
Springer, Berlin, Heidelberg
Abstract: 
In this paper, I extend the standard specification of the empirical similarity (ES) model of Gilboa et al. (Rev Econ Stat 88:433–444, 2006) to account for changes in parameters. I implement this by allowing for a combination of component ES models in the spirit of Gaussian mixture models. The predictive power of the modified model, along with that of the standard specification, will be assessed and compared to the baseline models consisting of autoregressions and Markov-switching autoregressions within a simulation exercise. Finally, we also compare the predictive ability of models using data on quarterly US real GDP growth. The results indicate that in situations of a more complex regime-switching behavior and a moderate to high autocorrelation in series, modified ES model demonstrates a better empirical fit. In addition, results of the empirical example show that modified ES models can better predict more extreme observations.
Subjects: 
Empirical similarity
Regime switching
ML estimation
Markov-switching autoregression
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by Logo
Document Type: 
Article
Document Version: 
Published Version

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.