Testing equality of Poisson regression coefficients in two equation

[Allen Smith]

Hi all,

I use the following code to test equality of coefficient in two Poission regression equations, while State report equation is not found.

poisson exi6 hft_st rd_sale i.industry i.exchange  i.year 
est store Y1
poisson exr6 hft_st rd_sale i.industry i.exchange  i.year 
est store Y2

suest Y1 Y2, vce(cluster firm_id)
test [Y1_mean = Y2_mean]: hft_st

I receive the error as following:

. test [Y1_mean = Y2_mean]: hft_st
equation Y1_mean not found
r(303);

However, Stata can provide results when I run the following. I assume Stata actually stores the equation but has some problems for comparing. Could anyone help me deal with this? Many thanks.

. test hft_st

 ( 1)  [Y1_head1exi6_raw]hft_st = 0
 ( 2)  [Y2_head1exr6_raw]hft_st = 0

           chi2(  2) =    5.02
         Prob > chi2 =    0.0813

.

[Clyde Schechter]

After -poisson-, -suest- doesn't store those equations with the names you are using. You are using the names that -suest- would have used after -regress-. What you want is:

test [Y1_exi6 = Y2_exr6]: hft_set

Whenever you use -suest-, you have to pay attention to the names it shows you in its output. They differ with different kinds of regressions.