TY - JOUR
T1 - Optimal Linear Instrumental Variables Approximations
AU - Escanciano, Juan Carlos
AU - Li, Wei
N1 - Funding Information:
Research funded by the Spanish Programa de Generación de Conocimiento , reference number PGC2018-096732-B-I00 .
Funding Information:
Research funded by the Spanish Programa de Generaci?n de Conocimiento, reference number PGC2018-096732-B-I00.Research funded by the Spanish Grant PGC2018-096732-B-I00.
Publisher Copyright:
© 2020 Elsevier B.V.
PY - 2021/3
Y1 - 2021/3
N2 - This paper studies the identification and estimation of the optimal linear approximation of a structural regression function. The parameter in the linear approximation is called the Optimal Linear Instrumental Variables Approximation (OLIVA). This paper shows that a necessary condition for standard inference on the OLIVA is also sufficient for the existence of an IV estimand in a linear model. The instrument in the IV estimand is unknown and may not be identified. A Two-Step IV (TSIV) estimator based on Tikhonov regularization is proposed, which can be implemented by standard regression routines. We establish the asymptotic normality of the TSIV estimator assuming neither completeness nor identification of the instrument. As an important application of our analysis, we robustify the classical Hausman test for exogeneity against misspecification of the linear structural model. We also discuss extensions to weighted least squares criteria. Monte Carlo simulations suggest an excellent finite sample performance for the proposed inferences. Finally, in an empirical application estimating the elasticity of intertemporal substitution (EIS) with US data, we obtain TSIV estimates that are much larger than their standard IV counterparts, with our robust Hausman test failing to reject the null hypothesis of exogeneity of real interest rates.
AB - This paper studies the identification and estimation of the optimal linear approximation of a structural regression function. The parameter in the linear approximation is called the Optimal Linear Instrumental Variables Approximation (OLIVA). This paper shows that a necessary condition for standard inference on the OLIVA is also sufficient for the existence of an IV estimand in a linear model. The instrument in the IV estimand is unknown and may not be identified. A Two-Step IV (TSIV) estimator based on Tikhonov regularization is proposed, which can be implemented by standard regression routines. We establish the asymptotic normality of the TSIV estimator assuming neither completeness nor identification of the instrument. As an important application of our analysis, we robustify the classical Hausman test for exogeneity against misspecification of the linear structural model. We also discuss extensions to weighted least squares criteria. Monte Carlo simulations suggest an excellent finite sample performance for the proposed inferences. Finally, in an empirical application estimating the elasticity of intertemporal substitution (EIS) with US data, we obtain TSIV estimates that are much larger than their standard IV counterparts, with our robust Hausman test failing to reject the null hypothesis of exogeneity of real interest rates.
KW - Hausman test
KW - Instrumental variables
KW - Linear approximation
KW - Misspecification
KW - Nonparametric identification
UR - http://www.scopus.com/inward/record.url?scp=85089099776&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85089099776&partnerID=8YFLogxK
U2 - 10.1016/j.jeconom.2020.05.002
DO - 10.1016/j.jeconom.2020.05.002
M3 - Article
AN - SCOPUS:85089099776
SN - 0304-4076
VL - 221
SP - 223
EP - 246
JO - Journal of Econometrics
JF - Journal of Econometrics
IS - 1
ER -