Increasing the number of overidentifying restrictions can cause severe biases in two stage least squares estimators.

Question

Quizplus · Accepted Answer

Increasing the number of overidentifying restrictions can lead to a phenomenon known as "overidentification bias" in two-stage least squares (2SLS) estimators. This occurs when additional restrictions are imposed on the second stage equation beyond what is required for identification. These additional restrictions can cause the estimator to become biased and inconsistent. Therefore, it is important to carefully consider the appropriate number of restrictions to use in 2SLS estimation.

Increasing the Number of Overidentifying Restrictions Can Cause Severe Biases