Robust IV Inference with Clustering Dependence

Linear IV models with clustering dependence are widely used in empirical studies, although the common solution, the cluster covariance estimator, often produces undesirable inferential results, especially with weak instruments. In this paper, I propose a method that is robust to both weak IV and (potentially heterogeneous) clustering dependence. The proposed method is based on the idea of Fama-MacBeth estimation, with group-level estimators being a truncated version of the unbiased IV estimator. Truncation stabilizes the group-level estimator by ensuring bounded second moments, thus improving finite-sample performance in weak instrument settings. Asymptotic validity is shown under both strong and weak IV sequences, as well as under general requirements. The proposed method is applied to study the effect of city compactness on population density.