Dear ,
I would highly appreciated any help in that
The problem of endogeneity arises when the factors supposed to affect a particular outcome possibly depend themselves
on that outcome. Apart from reverse causality, endogeneity problems could also arise because of omitted variable bias and/or measurement error.
The Granger causality test is the test for causality, but it just can be applied for the time series data, while my data set is unbalanced panel, and I am looking for a way to perform this test on my data set. On the other hand, and from the second type endogeneity view, if the Housman test show that there is a correlation between the errors and the explanatory variables, what is the solution to this case if there is no instrumental variables to use in the model.
The final question, is it possible to have any sort of causality between the independent variables themselves, if this is the case what is the solution in this case?.
I would highly appreciated any help in that
The problem of endogeneity arises when the factors supposed to affect a particular outcome possibly depend themselves
on that outcome. Apart from reverse causality, endogeneity problems could also arise because of omitted variable bias and/or measurement error.
The Granger causality test is the test for causality, but it just can be applied for the time series data, while my data set is unbalanced panel, and I am looking for a way to perform this test on my data set. On the other hand, and from the second type endogeneity view, if the Housman test show that there is a correlation between the errors and the explanatory variables, what is the solution to this case if there is no instrumental variables to use in the model.
The final question, is it possible to have any sort of causality between the independent variables themselves, if this is the case what is the solution in this case?.