Re: st: count time series model

[graupel75@gmail.com]

Hello Ikuho,

The easiest thing is to run the Poisson or negbinom regressions using
-glm- which allows Newey-West error structure:

glm count a b c, family(Pois) link(log) vce(hac nwest n)

where the n can be estimated with -ivreg2- (from SSC) using the bw(auto) option.

IIRC the same models can be obtained using -arpois- (SSC), but
optimization and reporting are less well developed.

Hope this helps; there's surprising little published on such models!

cheers-
Andrew Lover

___________________________________________________
Epidemiologist
Centre for Infectious Disease Epidemiology Research (CIDER)
Saw Swee Hock School of Public Health
National University of Singapore
*
*   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/