A Grid-Rate Condition for Valid Uniform Inference

Abstract

Conducting uniform inference on a continuous functional F defined on a compact subset X of R^d involves specifying L_n^d nodes for estimation and the construction of confidence bands. While asymptotically valid inference requires L_n to increase with n, existing fixed-L rules of thumb and heuristic data-driven approaches lack formal justification. This paper shows that, for functions within a Donsker class, the simple grid-growth condition r_n^(1/4)/L_n -> 0, equivalently L_n grows faster than r_n^(1/4), is sufficient for valid inference on twice continuously differentiable functions whose estimators satisfy r_n^(1/2)(F_hat - F) = O_p(1).