Working Paper: CEPR ID: DP14415
Authors: Dante Amengual; Xinyue Bei; Enrique Sentana
Abstract: We study score-type tests in likelihood contexts in which the nullity of the information matrix under the null is larger than one, thereby generalizing earlier results in the literature. Examples include multivariate skew normal distributions, Hermite expansions of Gaussian copulas, purely non-linear predictive regressions, multiplicative seasonal time series models and multivariate regression models with selectivity. Our proposal, which involves higher order derivatives, is asymptotically equivalent to the likelihood ratio but only requires estimation under the null. We conduct extensive Monte Carlo exercises that study the finite sample size and power properties of our proposal and compare it to alternative approaches.
Keywords: generalized extremum tests; higher-order identifiability; likelihood ratio test; non-Gaussian copulas; predictive regressions; skew normal distributions
JEL Codes: C12; C46; C58; C22; C34
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Information matrix's nullity > 1 (C67) | Validity of traditional tests (C52) |
| Proposed score-type tests (C52) | Performance of likelihood ratio tests (C52) |
| Type of distribution used (C46) | Performance of tests (C52) |
| Model specifications (C51) | Outcomes of tests (C52) |
| Proposed tests outperform standard likelihood ratio tests (C52) | Testing power and reliability (C12) |