All,
With my own data, I'm able to control for higher-level clustering effects
by specfiying the -cluster()- option in addition to the lower-level
variable I specify in -absorb()- using -areg-. This is an extremely useful
facility to call upon.
For reasons presumably due to computational complexity, I can't do this
with -ivreg2-. From the -country- variable in the -garmit- panel dataset
of 18 countries that I have (available on request),
. tab country
country | Freq. Percent Cum.
------------+-----------------------------------
Australia | 34 5.56 5.56
Austria | 34 5.56 11.11
Belgium | 34 5.56 16.67
Canada | 34 5.56 22.22
Denmark | 34 5.56 27.78
Finland | 34 5.56 33.33
France | 34 5.56 38.89
Germany | 34 5.56 44.44
Ireland | 34 5.56 50.00
Italy | 34 5.56 55.56
Japan | 34 5.56 61.11
Netherlands | 34 5.56 66.67
New Zealand | 34 5.56 72.22
Norway | 34 5.56 77.78
Sweden | 34 5.56 83.33
Switzerland | 34 5.56 88.89
UK | 34 5.56 94.44
US | 34 5.56 100.00
------------+-----------------------------------
Total | 612 100.00
I created a dummy variable called -europe-, indicating whether or not a
country is European. This is an obviously crude higher-level clustering
effect, but it serves for the purpose below:
. ivreg2 growthpc trade fdi unem left spend captax labtax, bw(2)
cluster(europe) robust
cannot use HAC kernel estimator with -cluster- option
r(198);
. ivreg2 growthpc trade fdi unem left spend captax labtax, bw(2)
cluster(europe)
cannot use HAC kernel estimator with -cluster- option
r(198);
I can't see anything in the -ivreg2- help file that explains why these
combinations of options are illegal, so any help on this would be greatly
appreciated!
CLIVE NICHOLAS |t: 0(044)7903 397793
Politics |e: [email protected]
Newcastle University |http://www.ncl.ac.uk/geps
*
* 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/
