Question

If some of the omitted variables, which you hope to capture in the changes analysis, in fact change over time within entity, then the FE estimator a. will be unbiased only when allowing for heteroskedastic-robust standard errors. b. may still be biased. c. will only be unbiased in large samples. d. will always be unbiased.

Answer 1

d. will always be unbiased.

Explanation:

When some of variables are omitted the estimator on indulge charge regressor may still be unbiased. If there is a special case where T = 2 then the FE estimator will be unbiased. Fixed effect regression model has different intercepts.