[R] Instruments for endogenous interaction term in Simultaneous Equation Model

Sukant Arora sukant.iitr at gmail.com
Fri May 20 11:48:13 CEST 2016


I want to estimate the following SEM model:

Y1 = a0 + b0*D + c0* Y3 + d0 * Y3 * D + control variables
Y2 = a1 + b1*D + c1 *Y3 + d1 * Y3 * D + control variables
Y3 = a2 + b2*D + c2 *Y1 + d2 * Y1 * D + e2*Y2 + f2 * Y2* D + control

where, D is dummy variable, Y1,  Y2 and Y3 are endogenous variables and
Y1*D, Y2*D, Y3*D represent the interaction terms.

Suppose I have Z as an instrument for Y2. Then, can I use Z*D as instrument
for Y2*D and similarly for others?


	[[alternative HTML version deleted]]

More information about the R-help mailing list