# Restrictions on 2SLS problems and SHAZAM

I had some problems running Shazam as follows.

C, I, Y are three endogenous variables. G and R are two exogenous variables.

There are three equations, two with constants and coefficients, but one is like this:

Y=C+I+G


I wonder if I can still use the 2SLS function and so I wrote the code like this:

2SLS C Y (R G) / DN
2SLS I R (R G) / DN
2SLS Y C I G (R G) / RESTRICT
RESTRICT C=1
RESTRICT I=1
RESTRICT G=1
END


But, it doesn't work. The results are different to my analytical results, which I'm pretty sure are correct.

edit retag close merge delete

Sort by » oldest newest most voted

The problem may be that with 2SLS the identities should not be included in the estimation. So the final equation that states the identity should be removed -- you then have two equations in the system.

If you are trying to estimate a system of equations then you can follow the example commands given in the SHAZAM User's Reference Manual Version 10 (page 331 of the Black cover manual; page 335 of the purple cover manual).

Note that the description states: All identities have already been substituted.

more