Qiankun Zhou  (My name can be pronounced as Chiankun Chou)

I am a 5th year PhD candidate in Economics at the University of Southern California. My research interests are theoretical econometrics and applied econometrics, with an emphasis on panel data econometrics, especially on dynamic panel data econometrics. 

I received my B.S. in Statistics from Hunan Normal University (2000-2004) and my M.A. in Statistics from Peking University (2006-2008). I have worked as a lecturer at Hunan University of Science and Engineering for two years (2004-2006) and as a Research Associate at Singapore Management University for two years (2008-2010).

Detailed CV can be found here.

I will be attending the January 2015 AEA meeting in Boston.

You can find my teaching statement here, and my research statement here.

Job Market Paper:
Abstract:  We study the identifi…cation and estimation of panel dynamic simultaneous equations models. We show that the presence of time-persistent individual-specifi…c effects does not lead to changes in the identi…fication conditions of traditional Cowles Commission dynamic simultaneous equations models. However, the limiting properties of the estimators depend on the way the cross-section dimension, N, or the time series dimension, T, goes to in…nity.
We propose three limited information estimator: panel simple instrumental variables (PIV), panel generalized two stage least squares (PG2SLS), and panel limited information maximum likelihood estimation (PLIML). We show that they are all asymptotically unbiased independent of the way of how N or T tends to infi…nity. Monte Carlo studies are conducted to compare the performance of the PLIML, PIV, PG2SLS, the Arellano-Bond type generalized method of moments and the Akashi-Kunitomo least variance ratio estimator. We demonstrate that the reliability of statistical inference depends critically on whether an estimator is asymptotically unbiased or not.

Abstract:  It is well known that the two stage least squares (2SLS) estimator is inconsistent for models with many instruments, and the usual statistical inference is no longer valid for models with many exogenous regressors. In order to obtain a consistent 2SLS estimator and make valid inference, we need to use a small number of relatively important instruments and exogenous regressors. In this paper, we propose a two-step procedure, Lass-after-Lasso, to select instruments and exogenous regressors. In the first step, we apply Lasso to select instruments for the endogenous variable, and use these selected instruments to construct the predicted endogenous variable. In the second step, we replace the endogenous variable by its predicted value obtained in the fi…rst step, and use Lasso to select exogenous variables. The 2SLS estimator based on these selected instruments and exogenous variables is consistent and asymptotically normally distributed. Monte Carlo simulation con…firms the 2SLS estimator using the Lasso-after-Lasso procedure has good …finite sample performance in estimation and testing.

Prof. Hsiao Cheng (Chair)
Department of Economics, USC, Los Angeles, CA 90089.
Phone: 213-740-2103; Email: chsiao@usc.edu
Prof. M.Hashem Pesaran
Department of Economics, USC, Los Angeles, CA 90089.
Phone: 213-740-3510; Email: pesaran@usc.edu
Prof. Hyungsik Roger Moon
Department of Economics, USC, Los Angeles, CA 90089.
Phone: 213-740-2109; Email: moonr@usc.edu