大阪大学　数理・データ科学教育研究センター
Center for Mathematical Modeling and Data Science,Osaka University

Directionally Differentiable Econometric Models with Application to Rosenberg's Conditional Heteroskedasticity Models

Jin Seo Cho (Yonsei University)

1. ホーム
2. セミナー・集中講義
3. セミナー
4. Directionally Differentiable Econometric Models with Application to Rosenberg's Conditional Heteroskedasticity Models

大証寄附研究部門セミナーシリーズ第30回

Directionally Differentiable Econometric Models with Application to Rosenberg's Conditional Heteroskedasticity Models

Jin Seo Cho (Yonsei University)

We relax the differentiability condition for standard econometric models to the level of directional Gateaux differentiability and analyze asymptotic distribution of extremum estimator. We show that its asymptotic distribution can be represented as a functional of Gaussian process indexed by direction. Our analysis also treats the differentiable models as a special case of directionally differentiable models.

For data inference, we refine standard likelihood ratio, Wald, and Lagrange multiplier test statistics. These refinements also permit the presence of nuisance parameters. Further, from this, these test statistics are shown to be asymptotically equivalent if the null models are differentiable and do not have boundary parameters.