【主讲】王学新 (助理教授，厦门大学王亚南经济研究院)
【主题】Instrumental variable estimation via a continuum of instruments with an application to estimating the elasticity of intertemporal substitution in consumption
【时间】2018年11月27日 (周二) 15：30-17：00
【地点】上海财经大学经济学院楼701室
【语言】英文
【摘要】The endogeneity of regressors is a dominant feature of linear regression model, which continues to be a workhorse in econometrics. This paper introduces a new instrument variable estimation procedure based on a continuum of instruments. The proposed estimator is labeled as NIV, its weighted version as WNIV. These estimators are easy to compute, resembling the classic k-class IV estimators, but no any user-chosen number is needed. They are consistent and asymptotically normal distributed under weak instruments and heteroskedasticity. Monte Carlo simulations demonstrate that the proposed estimators have excellent finite sample properties, outperforming alternative estimators in a wide range of cases. We then apply the new estimation procedure to estimating the elasticity of intertemporal substitution (EIS) in consumption, which is of central importance in macroeconomics and finance. For UK quarterly data from 1970.3-1999.1, the WNIV estimate of EIS is 0.7, and statistically different from zero. For US quarterly data from 1947.3-2008.4, the WNIV estimate is as high as 2.0, and statistically different from zero. Moreover, these estimates are robust to model transformation, and different sets of conditioning variables. These empirical results are strikingly different from the alternative estimators, which are quite small, even very close to zero. These findings give strong support to the practice of model calibrations in macroeconomics and finance, where the EIS in consumption, as a crucial input parameter, is required to be large.
