Page 1 of 1

Restriction with SUR model(AIDS case)

Posted: Fri Dec 16, 2011 10:10 am
by eigenjin
I am totally newbie in RATS, I have not much knowledge on RATS.

I am going to estimate the Almost Ideal demand system with SUR method.
to consistent with the economic theory, the demand system satisfied the symmetric and homogeneity condition

For example. in case we have a 4 equation such as
eq1 = const +1 p1 +2 p2 +3 p3 +4p4 + log(TE/P)
eq2 = const +5 p1 +6 p2 +7 p3 +8p4 + log(TE/P)
eq3 = const +9 p1 +10 p2 +11 p3 + 12p4 + log(TE/P)
(I deleted last equation to estimate)

restriction is that sum(1:4)=1 , sum(5:8), sum(9:12) : homogeneity
and 2=5, 3=9, 7=10 : symmetric condition

To do this, I set the equation

equation eq1 w1
# constant lp1 lp2 lp3 lp4 lTP
equation eq2 w2
# constant lp1 lp2 lp3 lp4 lTP
equation eq3 w3
# constant lp1 lp2 lp3 lp4 lTP
equation eq4 w4
# constant lp1 lp2 lp3 lp4 lTP

sur(iterations=100) 3
# eq1
# eq2
# eq3

My question is following
I know the "equate" option in SUR model. I have read that reference manual.
It is not clear the description in the reference manual.


Can you explain more about how to impose the symmetric and homogeneity restriction on it?

Re: Restriction with SUR model(AIDS case)

Posted: Fri Dec 16, 2011 12:12 pm
by TomDoan
You probably want to look at the CONSUMER.RPF example (page UG-141 in RATS v8) which uses NLSYSTEM rather than SUR. Even though the model itself is linear, it's easier to handle the constraints by using the non-linear setup.

Re: Restriction with SUR model(AIDS case)

Posted: Fri Dec 16, 2011 2:42 pm
by eigenjin
I appreciate your comments.
It is very helpful