Test Structural Breaks Bai-Perron settings
Posted: Fri Nov 02, 2018 8:44 am
Hi,
I want to test a time series for structural breaks using the Bai-Perron method. For this I want to reconstruct the results from the paper "Regime Changes in International Real Interest Rates: Are They a Monetary Phenomenon?" from Rapach and Wohar (https://www.jstor.org/stable/3839151). In this paper they estimate whether the real interest rate contains regime changes. The data are publicly available along with the GAUSS code at: https://sites.google.com/slu.edu/davera ... blications but I will still attach an excel file with the final tax-adjusted real interest rate.
The settings for their test are based on Bai and Perron (2003 - Multiple structural change Models: A simulation analysis).
Unfortunately it is not entirely clear for me which settings they use and how I can use the same in EVIEWS to get an exact replication of their result. I know that they use trimming percentage of 15 and allow for heteroskedasticity and serial correlation. But which settings in EVIEWS do I have to use to replicate their output?
This is the result they get with additional information in footnote 7:
Thank you.
I want to test a time series for structural breaks using the Bai-Perron method. For this I want to reconstruct the results from the paper "Regime Changes in International Real Interest Rates: Are They a Monetary Phenomenon?" from Rapach and Wohar (https://www.jstor.org/stable/3839151). In this paper they estimate whether the real interest rate contains regime changes. The data are publicly available along with the GAUSS code at: https://sites.google.com/slu.edu/davera ... blications but I will still attach an excel file with the final tax-adjusted real interest rate.
The settings for their test are based on Bai and Perron (2003 - Multiple structural change Models: A simulation analysis).
Unfortunately it is not entirely clear for me which settings they use and how I can use the same in EVIEWS to get an exact replication of their result. I know that they use trimming percentage of 15 and allow for heteroskedasticity and serial correlation. But which settings in EVIEWS do I have to use to replicate their output?
This is the result they get with additional information in footnote 7:
- Results RW.png (290.27 KiB) Viewed 4867 times
Thank you.