Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# st: cmp and condition numbers

 From "Bromiley, Philip" To "statalist@hsphsun2.harvard.edu" Subject st: cmp and condition numbers Date Mon, 30 Apr 2012 23:30:28 +0000

I'm trying to estimate a simultaneous system with three continuous and one discrete variable using cmp.  I have been unable to get it to estimate properly - lots of not concave and backed up messages and then it crashes saying it has hit a discontinuous or flat region.

Cmp warns me that I have an ill-conditioned regressor matrix and reports high condition numbers for each of the equations (40 to 1000).  However, when I run the equation with regress, I don't get high VIF's, and get a much lower condition number.

Would someone know the reason for such a discrepancy?  Any suggestions would be welcome.

Phil

Here is a simple example to illustrate the condition number difference.

webuse laborsup, clear

cmp setup
replace fem_inc = fem_inc - 10
cmp (kids = fem_inc male_educ)     (fem_work = male_educ),        ind(\$cmp_cont \$cmp_cont)
*to increase the correlation among the x's, I add a random number to all of them
g x1=rnormal() * 100
g fem_inc1=fem_inc + x1
g male_educ1=male_educ + x1
g fem_work1=fem_work + x1

cmp (kids = fem_inc1 male_educ1)      (fem_work1 = male_educ1),        ind(\$cmp_cont \$cmp_cont)

reg kids  fem_inc1 male_educ1
estat vif
cndnmb3  fem_inc1 male_educ1

reg kids  fem_inc male_educ
estat vif
cndnmb3   fem_inc male_educ

Philip Bromiley
Dean's Professor of Strategic Management