Working Paper: CEPR ID: DP10390
Authors: George Kapetanios; Lynda Khalaf; Massimiliano Marcellino
Abstract: Robust methods for IV inference have received considerable attention recently. Their analysis has raised a variety of problematic issues such as size/power trade-offs resulting from weak or many instruments. We show that information-reduction methods provide a useful and practical solution to this and related problems. Formally, we propose factor-based modifications to three popular weak-instrument-robust statistics, and illustrate their validity asymptotically and in finite samples. Results are derived using asymptotic settings that are commonly used in both the factor and weak instrument literatures. For the Anderson-Rubin statistic, we also provide analytical finite sample results that do not require any underlying factor structure. An illustrative Monte Carlo study reveals the following. Factor based tests control size regardless of instruments and factor quality. All factor based tests are systematically more powerful than standard counterparts. With informative instruments and in contrast with standard tests: (i) power of factor-based tests is not affected by k even when large, and (ii) weak factor structure does not cost power. An empirical study on a New Keynesian macroeconomic model suggests that our factor-based methods can bridge a number of gaps between structural and statistical modeling.
Keywords: Factor Model; Identification; Robust Inference; IV Regression; New Keynesian Model; Principal Components; Weak Instruments
JEL Codes: C12; C13; C52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| factor-based tests (G41) | control size (L25) |
| factor-based tests (G41) | more powerful than standard counterparts (Y50) |
| factor-based tests (G41) | maintain power with increasing k (E11) |
| factor-based tests (G41) | not adversely affected by weak factor structures (C52) |
| factor-based procedures (C38) | bridge gaps between structural and statistical models (C59) |
| factor-based tests (G41) | superior performance in size control and power (L25) |