I ran an SUR model in R using systemfit package, but I need to take the summary result and use it for imposing two restrictions:
- the sum of the intercepts of the three equations should be equal to 24.
- the sum of the coefficients of each explanatory variable over all the activities is equal to zero.
How do I run an SUR model with these two restriction?
SUR estimates for 'eq1' (equation 1)
Model Formula: Y1 ~ X1
Estimate Std. Error t value
(Intercept) 8.12788487 1.22300735 6.64582
X1Rural -1.03309385 0.38727326 -2.66761
X1chdlessthan6 1.67758437 0.39414610 4.25625
SUR estimates for 'eq2' (equation 2)
Model Formula: Y2 ~ X1
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.7163935 0.9913563 0.72264 0.470431
X1Rural -0.7947226 0.3139194 -2.53161 0.011834
X1chdlessthan6 -0.5248445 0.3194905 -1.64275 0.101419
SUR estimates for 'eq3' (equation 3)
Model Formula: Y3 ~ X1
Estimate Std. Error t value Pr(>|t|)
(Intercept) 19.9376084 1.5112688 13.19263 < 0.000000000000000222
X1Rural 0.3432570 0.4785531 0.71728 0.47372572
X1chdlessthan6 -1.3580186 0.4870459 -2.78828 0.00561697
I tried linearhypothesis
, but I was getting an error saying The hypothesis "Y1_(Intercept)+Y2_(Intercept)+Y3_(Intercept)=24" is not well formed: contains bad coefficient/variable names.