Hi all,
in the archive I found a helpful posting of Scott Merryman:
http://www.stata.com/statalist/archive/2005-02/msg00264.html
His code develops a numlist specification for the optimal lag length in the
context of a panel unit root test such as <levinlin> and <ipshin>. The code
proposed by Scott allows me to construct a local macro named opt_lag, which
includes the optimal lag length for each panel. Having modified the code for
my purposes (variable names and so on), I type:
. levinlin migration, lag(`opt_lag')
Everything works fine, which means I get some output. However, the following
gives an error:
. ipshin migration, lag(`opt_lag')
too many macros
r(920);
My panel has 240 cross-sectional units. I suppose this is too much for
<ipshin> when individual lag lengths are specified. When I use the code on a
smaller panel (e.g. 20 cross-sectional units), the code also works for
ipshin.
My question is: Is it possible to include 240 different lag lengths with
ipshin? The help file only says: "If a list of lags is provided (perhaps
with a local macro), its length must match the number of individuals in the
panel." This is the case in my application.
Thanks for any advice,
Falko.
--
Sparen beginnt mit GMX DSL: http://www.gmx.net/de/go/dsl
*
* 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/
