Instrumental Variable Estimation of Structural Equation Under Covariance Restriction: A Partial Reduced Form Approach

• 영문명: Instrumental Variable Estimation of Structural Equation Under Covariance Restriction: A Partial Reduced Form Approach
• 발행기관: 한국계량경제학회
• 저자명: Yun-Yeong Kim Joon Y. Park
• 간행물 정보: 『한국계량경제학회 학술대회 논문집』2008년 하계학술대회, 1~39쪽, 전체 39쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 2008.08.30

영문 초록

We consider an estimation of structural equation under covariance restriction when the exclusion restrictions and exogeneity of variable are not plausible. The residual as instrument is exploited for the consistent estimation without identification of the other structural equations, if there is an exogenous variable which just appears in the estimated equation. The efficient instrumental variable estimator minimizing the asymptotic variance is derived which is asymptotically equivalent with the FIML estimator under the normality and exact identification. We suggest a Wald test for the inference. Substantial simulation works confirmed our theoretical expectations.

목차

1 Introduction

2 The Identification Assumption

3 Efficient IV Estimation

4 The Efficiency of IV Estimator

5 Inference and Specification Test

6 Monte Carlo Experiments

7 Conclusion

키워드

covariance restriction residual as instrument instrumental variable estimation identifying variables partial reduced form