st: burt - a new zandrews like command also does LM unit root tests withbreaks

[Mark Feldman <statalist528h5q@feldmark.com>]

Hi,

I have been working on a unit root test in STATA that I think anyone who uses the zandrews
command might be interested in. The zandrews command seems to have a number of bugs, probably
inherited from the original RATS code, as Christopher Baum pointed out about a year ago in response 
to a query of mine.  (These are documented in the below mentioned PDF file.)

I have been writing a command I call "burt" (Broken Unit Root Test) which I think corrects these and
adds some functionality. In particular, it can search for an arbitrary number of break dates (not
just one) and also implements Lagrange Multiplier style unit root tests within the same framework of
command line options. Lumsdaine and Papell (1997) extended the Zivot and Andrews (1992) tests to 2
breaks and there is a similar history of papers outlining Lagrange Multiplier tests up to 2 breaks,
so there are some critical values available up to 2 breaks. Although there are none available for 3
or more breaks, I believe that you can sometimes gain insight into your data from the graphs (with
estimated trend and detrended series) that burt will produce.

Burt is still under development, but I have already used it to reproduce as many published results
of Nelson and Plosser (1982) data as I could get my hands on (from the papers I mention above.) You
can see a summary of these tests and get further information here:

http://www.feldmark.com/code/

I don't have a STATA help file yet, but I do have a pretty complete PDF document at the above
location. It contains all the info I will eventually put into a help file plus some background and
a number of implementation notes I made while doing this. (Various potential collinearities that
had to be avoided and a description of how it handles observations lost to lags and differences
which I has been conspicuously absent from all of the papers I read.) Everything (except the help
file) should be just as you expect, if you run

net from http://www.feldmark.com/code/en/

There are probably still bugs, but I am actively developing this, so will do my best to fix anything
that anyone tells me about. I hope this is useful.

Best Regards,

Mark Feldman
feldmark@feldmark.com

References
----------
Lumsdaine, Robin and David H. Papell (1997) “Multiple Trend Breaks and the Unit-Root Hypothesis,”
The Review of Economics and Statistics, Vol. 79, No. 2 (May 1997), pp. 212-218.

Nelson, C. R. and C.I. Plosser (1982) “Trends and Random Walks in Macroeconomic Time Series: Some
Evidence and Implications”, Journal of Monetary Economics, 10, pp. 139-62.

Zivot, E. and D. W. K. Andrews (1992) “Further Evidence on the Great Crash, the Oil-Price Shock, and
the Unit-Root Hypothesis”, Journal of Business & Economic Statistics, American Statistical
Association, vol. 10(3), pp. 251-270, 1992.

END




*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/