st: Non-working SROOT options

["Francis, Richard N" <rnfrancis@utep.edu>]

Hi Statalist,

A heads-up . . . this is a problem associated with the functionality/failure of a user-written program rather than a user-programming problem per se.

A user-written program from SJ v9, n3, pp. 422-38 (ST-0172) provides enhancements to a seasonal unit root test developed by Hylleberg, et al. (J of Econometrics, v44, 1990).

Page 426 of this article provides the syntax, including the options, for the -SROOT- command.

However, I am finding that the -season(varlist)- option for the -SROOT- command is not functional in a stand-alone manner.

For example, the following syntax works great:

sroot varname, trend season(varlist)


The following syntax also works great:

sroot varname, trend 


However, the following syntax fails:

sroot varname, season(varlist)


Stata provides the following error message associated with the failure:

5.15\ not found
r(111);


The discussion on page 426 of SJ (v9, n3) concerning the -season(varlist)- option indicates that this option may not be used with the -noconstant- option.

Note that the examples above avoid this combination, yet the command continues to fail.

Although what I know about Stata fits neatly within the eye of a small needle, I did set the trace feature 'ON' and narrowed the problem to a section of the code which retrieves critical values.  

In a nutshell, the various model specifications for the testing procedure require varying critical values.

It 'APPEARS' that the code fails while attempting to retrieve the critical values for the stand-alone use of -season(varlist)- .

The stand-alone use of -season(varlist)- is legitimate and appropriate, as this specification appears in the original Hylleberg (1990) article on page 226 (J of Econometrics, v44, 1990) in the third horizontal group of auxiliary regressions/model specifications in Table 1A.

I am unsure how to rectify this problem, but thought that posting this message could possibly assist other Stata users, as well as the editorial staff at Stata.

I apologize if this is the inappropriate method for resolving the issue.

Thank you,


Rick



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/