I am working on 75 companies with 05 Independent variable (corporate governance parameter) and 10 dependent variable (firm performance measurement variable), 05 control variable and 10 industry dummy variables.
To resolve endogeneity error in regression model using Dynamic GMM, how to select instrument variable for analysis?
Shubhanker Yadav, the Instrument Variable estimator can help us distinguish endogenous variables by comparing them with OLS. If an explanatory variable is endogenous OLS is biased and inconsistent. If the Gauss-Markov assumptions hold, then OLS and IV should give us the same answer, at least within sampling error. We can test for this through Hausman test where we consider coefficients on a single explanatory variable of interest that might be endogenous. If you conduct the test and reject the null, it means that the 2SLS and OLS estimates differ from each other significantly. If our instrument is valid as assumed, this must mean that your k variable is endogenous.
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments