Economic Significance for 2SLS estimates

[Lubomir Litov]

I was wondering if there is a conventional method of calculating the economic significance of 2SLS estimates.

If X is the dep. variable, Y is the endogenous independent variable, Z is the instrumental variable, & beta(Y) is the 2SLS second-stage estimated coefficient on the instrumented endogenous variable, will it be reasonable to calculate the economic significance magnitude (to one standard deviation change in the independent endogenous variable) as follows:

(I) beta(Y) * st.dev.(Y) / median(X) or
(II) beta(Y)*st.dev.(Predicted Y|Z) / median(X).

Thank you in advance.

[George Ford]

if ln(X), then the coefficient is a pretty good indicator of economic significance.

you could use

margins, at("meanY" "meanY+sd")

and compute the difference