Level shift estimation

If you are seeking RATS code for implementing a particular technique or replicating results from a paper, post your request here. Be sure to include complete citations for any papers or books.

Level shift estimation

Unread postby ag_2018 » Wed Jan 06, 2021 7:15 am

Hi,
I was wondering if anyone can help me coding the empirical example given in the following paper:

Harris, D., Kew, H., & Taylor, A. R. (2020). Level shift estimation in the presence of non-stationary volatility with an application to the unit root testing problem. Journal of Econometrics, 219: 354-388.

I have attached the US data they use for their estimation. In particular, I am having some trouble estimating equation 3.8 in the paper; and table 5 and 6.

Any help would be most appreciated!
Attachments
HKTdata1.xlsx
US GDP data will need to be expressed in growth form
(19.84 KiB) Downloaded 262 times
ag_2018
 
Posts: 9
Joined: Fri Nov 02, 2018 12:49 pm

Re: Level shift estimation

Unread postby TomDoan » Wed Jan 06, 2021 2:55 pm

3.8 (in fact, the whole analysis) is very similar to the NPREG.RPF example, except that it uses time as the explanatory variable in the non-parametric fit on the squared residuals.
TomDoan
 
Posts: 7240
Joined: Wed Nov 01, 2006 5:36 pm

Re: Level shift estimation

Unread postby ag_2018 » Wed Jan 06, 2021 8:23 pm

Thanks Tom.
I am finding the estimation difficult and was hoping if you/someone could help in the calculation of the series 'sigma(t)' in 3.8. Not sure how to carry out the kernel smoothing. Any pointers would be much appreciated!
Table 5 is the variance profile calculation due to Cavaliere and Taylor (2008) which is to determine whether there is non-stationary unconditional volatility; separate estimation altogether.
Many thanks in advance!
ag_2018
 
Posts: 9
Joined: Fri Nov 02, 2018 12:49 pm

Re: Level shift estimation

Unread postby TomDoan » Fri Jan 08, 2021 2:50 pm

This does an example of the calculation of the WLS estimator for a single example. The calculation of the time-varying variance is the same regardless of the regression---the bandwidth depends only upon the number of data points.
Attachments
wlsexample.rpf
(1.15 KiB) Downloaded 280 times
TomDoan
 
Posts: 7240
Joined: Wed Nov 01, 2006 5:36 pm

Re: Level shift estimation

Unread postby ag_2018 » Tue Jan 12, 2021 12:08 pm

Thanks Tom! Much appreciated.
ag_2018
 
Posts: 9
Joined: Fri Nov 02, 2018 12:49 pm


Return to Looking for Code?

Who is online

Users browsing this forum: No registered users and 1 guest

cron