EViews.com

I guess it would be easier to make the slight modification to the add-in to fetch other values from R, than it would be to go to another language. Although that, of course, depends on your knowledge of the languages.

I am just starting to learn R. I've used Eviews and Stata for most of my research. I have programmed the Bai (1999) test in Stata, but wanted to see how well the BaiPerron test worked in Eviews, mainly because of the easy of Eviews. I would have no clue how to adjust the add-in.

I'm assuming you want to run the Add-in...(if not, disregard the comments below).

There is documentation in the EViews 7.1 supplement. Here's the 30 second version.

If you have EViews 7.1 and R installed on your system, go to the webpage for the Add-ins and install the Bai-Perron Add-in. If you don't have EViews 7.1 installed, do so. If you don't have R installed, see the documentation for EViews 7.1 on the steps you'll need to follow to install and configure R.

Estimate your equation in EViews, then use the Proc/Add-ins menu to call your routine.

Hi, i just finished to run the B&P test (it give me two break in the data), but i want to see the diferent coefficients for the variables....How can i see them?

Thanks

Diego

I guess, run the regression by itself over the smaller subsamples.

EViews Gareth wrote:I guess, run the regression by itself over the smaller subsamples.

Ok, yes i did that but i thought there was another way to see it in the test. Thanks

I'm able to run the add-in but the statistics reported are nearly useless when presenting results. I have no come across a paper citing Bai-Perron without reporting the F statistics. Is there a 30 second guide for going into the add in and capturing the F statistics

I'm not sure what you mean by the F-statistics. The R package that we use, strucchange, does have a routine for calculating F-statistics, but they're just the collection of Chow breakpoint Wald F-statistics that can be calculated in EViews via the Quandt-Andrews test built into EViews.

That helps. Thanks.

I have used the Bai-Perron addin before and it have worked fine. Therefore I know for a fact that I am not missing anything on my computer, the only problem should be in my code below. I have updated to the latest versions of R, EViews 7.2, the addin, the package strucchange.



I have attached the file that I am using. What am I doing wrong, it has worked before with similar tasks but not this time. I really cannot understand why.

The first specification won't work because you'll get too many singularities. The second specification worked fine for me.

Why can't the test be used on ARMA terms?

I believe the R package doesn't support it.

It doesn't. The standard algorithms for computing all of the regressions in a Bai-Perron framework requires simple linearity of the specification.

Hi. I am new to eviews. I have to find the breakpoint date on stock data. I have downloaded eviews 7.1 and created my workfile with my two variables "date" and "closingprices". I downloaded R as required. I have run my regression test using the "estimate equation" option. I obtained the result window however when i click on Bai perron breakpoint test under Addin. It asked for a subsample size in which i entered the total no of observations of my sample. But when i click Ok, an error message appears stating that i havent installed R and advises to review the documentation regarding installation. I looked at the pdfs under help section but couldnt find any. What should i do to run the test?