[Gretl-users] matrix-based ols calculations in presence of multicollinearity

Hi, Are the following commands supposed to produce the same results? indeed, they actually don't. <hansl> open greene19_1.gdt # some series series term1 = 0.3*TUCE + 4 series term2=2 list xlist = const TUCE term1 term2 matrix x = {xlist} matrix y = {GPA} # Model 1 ols GPA xlist --simple # Model 2 eval inv(x'*x)*x'y # Model 3 eval inv(x'x)*x'*y # Model 4 eval mols(y, x) </hansl> OLS-built function (model 1) works fine and eliminates the extra variables that cause multicollinearity. Same for the OLS matrix-built function (model 4) which halts on the ground that "Exact or near collinearity encountered" But, models 2 & 3 give different results (i) from the OLS-built function and (ii) one from the other. I also tried to use 'invpd()' as presented in the Guide (Listing 17.3: OLS via matrix methods (p. 153)); same results as Models 2 & 3. Am I missing something? Best, Artur