Ruei-Chi Lee

Rutgers University-New Brunswick · Economics · 2026–27 job market

Fields

Econometrics

Job market paper

Overfitting Inference with Instrumental Variables

Overfitting is traditionally viewed as a problem to be mitigated in empirical economics. This paper shows that it can instead be a blessing for inference in high-dimensional instrumental variables (IV) models. We revisit the factor IV framework, in which the instruments and the endogenous variable share common latent factors, and show that the first-stage signal is dense, providing a theoretical justification for recent empirical evidence against sparsity. Moreover, the implied first-stage coefficients shrink toward zero as the number of instruments diverges, exhibiting “local-to-zero” behavior commonly used to describe settings with many weak instruments. Building on this characterization, we propose an IV estimator based on ridgeless regression combined with sample splitting. This procedure remains well defined when the number of first-stage regressors exceeds the sample size. We derive the asymptotic distribution of the estimator under regimes with (very) many weak instruments, and show that, when the number of instruments is sufficiently large, its asymptotic variance coincides with that of an oracle factor IV estimator that directly observes the latent factors. Monte Carlo simulations indicate favorable finite-sample performance relative to existing high-dimensional IV estimators. Empirically, we apply the proposed procedure to estimating the elasticity of substitution between immigrants and natives in the United States.