The coefficient for endogenous is 0.313 with statistically significant.
Thank you so much!
The command 'estat endogenous' following an ivregress estimation will give the test statistic for the Durbin-Wu-Hausman test. The null hypothesis is that the coefficient on the residuals from the first stage regression is zero in the structural model, or that the variable in question is not endogenous. Therefore, if you get a significant test statistic, you are able to regress this null, and conclude that there is evidence that the variable being instrumented for is, in fact, correlated with the structural error and endogenous.
The command 'estat endogenous' following an ivregress estimation will give the test statistic for the Durbin-Wu-Hausman test. The null hypothesis is that the coefficient on the residuals from the first stage regression is zero in the structural model, or that the variable in question is not endogenous. Therefore, if you get a significant test statistic, you are able to regress this null, and conclude that there is evidence that the variable being instrumented for is, in fact, correlated with the structural error and endogenous.
Dear Prof. Michael,
Yes, I did. My question related to Two-Stage Least Squares, does the high-value (0.313) for the coefficient of endogenous is normal? If not, what is the interpretation for that?
Thanks indeed!
Best regards.
Yes, there is nothing wrong with statistically significant high-value coefficient of endogenous variable. It only means that the cost, expenses, etc of the endogenous variable is 31.3% higher than the other variables (exogenous).
However, if the value of the coefficient is negative, we say that the cost, expenses etc of the endogenous variable is lower, smaller than the exogenous variables.........
Please note that the 2SLS coefficients are usually quite different from the OLS coefficients.
**
Good luck
- Badges
- Science topic