Question

True or False:

1) ____ The "exclusion restriction" imposed on the instrumental variable z implies that z is excluded from the causal model of interest.

2) ____ The panel data estimator that uses first differencing to eliminate the "unobserved effect" will be less efficient than the estimator that applies deviations from means (time-demeaning) when the panel includes more than two periods and the idiosyncratic error is homoscedastic and serially uncorrelated.

3) ____ A graph plotting the time evolution of the outcome variable for the "treatment" and "control" groups can provide a useful visual assessment of the plausibility of the "common trends" assumption.

4) ____ The "clustering" procedure allows the errors to be arbitrarily correlated within each cluster but it assumes zero correlation between clusters.

Answer 1

1) True.

When exclusion restriction is imposed on a variable, it means that those variables are not in some of the equations.

2) True.

When panel includes two periods, and when the idosyncratic error is serially uncorrelated, panel data that uses time demeaning method is more accurate.

3) True.

In the absence of treatment, estimators assume that the difference between treatment and control is constant or fixed over a periodic of time. Group specific difference will be zero to hold common trend assumption.

4) False.

Clustering procedure allows error to be correlated but correlation between the clusters in not zero.

Kindly leave a comment for queries, if any. Upvote if u like it.

Thank you.