Dear all, I have run a test for endogeneity with IVREG2 with endog option. I used the lagged values as the instruments for the suspected endogenous variables. For example, say I used
xtivreg2 y x1 x2 x3 x4 (x5 x6 x7 x8 = x5 t-1 x6 t-1 x7 t-1 x8 t-1), fe r (endog x5 x6 x7 x8)
Stata gives me a number of results, For example
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic): 5.063
Chi-sq(1) P-val = 0.0244
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic): 7.678
(Kleibergen-Paap rk Wald F statistic): 1.273
Stock-Yogo weak ID test critical values: <not available>
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments): 0.000
(equation exactly identified)
-endog- option:
Endogeneity test of endogenous regressors: 8.621
Chi-sq(4) P-val = 0.0713
Regressors tested: x5 x6 x7 x8
------------------------------------------------------------------------------
My first question is how do I interpret the results, can I say that four variables which I tested are actually endogenous? and the instruments are valid?
My second question is what do I do next if I find these variables are endogenous? How do I modify my regression equation to run the actual model to get the desired result, should I incorporate the instruments along with the endogenous variables in the original regression equation?
Any comment is highly appreciated.
Ujjwal Kumar Das
Postgraduate researcher
University of Leeds
xtivreg2 y x1 x2 x3 x4 (x5 x6 x7 x8 = x5 t-1 x6 t-1 x7 t-1 x8 t-1), fe r (endog x5 x6 x7 x8)
Stata gives me a number of results, For example
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic): 5.063
Chi-sq(1) P-val = 0.0244
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic): 7.678
(Kleibergen-Paap rk Wald F statistic): 1.273
Stock-Yogo weak ID test critical values: <not available>
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments): 0.000
(equation exactly identified)
-endog- option:
Endogeneity test of endogenous regressors: 8.621
Chi-sq(4) P-val = 0.0713
Regressors tested: x5 x6 x7 x8
------------------------------------------------------------------------------
My first question is how do I interpret the results, can I say that four variables which I tested are actually endogenous? and the instruments are valid?
My second question is what do I do next if I find these variables are endogenous? How do I modify my regression equation to run the actual model to get the desired result, should I incorporate the instruments along with the endogenous variables in the original regression equation?
Any comment is highly appreciated.
Ujjwal Kumar Das
Postgraduate researcher
University of Leeds
Comment