Nuclear norm regularized estimation of panel regression models

In this paper we investigate panel regression models with interactive fixed effects. We propose two new estimation methods that are based on minimizing convex objective functions. The fi rst method minimizes the sum of squared residuals with a nuclear (trace) norm regularization. The second method minimizes the nuclear norm of the residuals. We establish the consistency of the two resulting estimators. Those estimators have a very important computational advantage compared to the existing least squares (LS) estimator, in that they are de fined as minimizers of a convex objective function. In addition, the nuclear norm penalization helps to resolve a potential identifi cation problem for interactive fixed effect models, in particular when the regressors are low-rank and the number of the factors is unknown. We also show how to construct estimators that are asymptotically equivalent to the least squares (LS) estimator in Bai (2009) and Moon and Weidner (2017) by using our nuclear norm regularized or minimized estimators as initial values for a nite number of LS minimizing iteration steps. This iteration avoids any non-convex minimization, while the original LS estimation problem is generally non-convex, and can have multiple local minima.

[1]  Serena Ng,et al.  Principal Components and Regularized Estimation of Factor Models , 2017, 1708.08137.

[2]  Trevor Hastie,et al.  Statistical Learning with Sparsity: The Lasso and Generalizations , 2015 .

[3]  A. Tsybakov,et al.  Estimation of high-dimensional low-rank matrices , 2009, 0912.5338.

[4]  Martin Weidner,et al.  Nonlinear Panel Models with Interactive Eects , 2014 .

[5]  Martin J. Wainwright,et al.  A unified framework for high-dimensional analysis of $M$-estimators with decomposable regularizers , 2009, NIPS.

[6]  Mingli Chen.  Estimation of nonlinear panel models with multiple unobserved effects , 2016 .

[7]  Xiaohong Chen,et al.  Estimation of Nonparametric Conditional Moment Models with Possibly Nonsmooth Generalized Residuals , 2009 .

[8]  A. Onatski.  Determining the Number of Factors from Empirical Distribution of Eigenvalues , 2010, The Review of Economics and Statistics.

[9]  P. Phillips,et al.  Identifying Latent Structures in Panel Data , 2014 .

[10]  Alberto Abadie,et al.  The Economic Costs of Conflict: A Case Study of the , 2003 .

[11]  M. Tani,et al.  Immigration Policy and Remittance Behaviour , 2017, SSRN Electronic Journal.

[12]  Jens Hainmueller,et al.  Comparative Politics and the Synthetic Control Method , 2014 .

[13]  Pablo A. Parrilo,et al.  Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization , 2007, SIAM Rev..

[14]  Alexei Onatski,et al.  Asymptotics of the principal components estimator of large factor models with weakly influential factors , 2012 .

[15]  G. Imbens,et al.  Matrix Completion Methods for Causal Panel Data Models , 2017, Journal of the American Statistical Association.

[16]  Cheng Hsiao,et al.  A Panel Data Approach for Program Evaluation — Measuring the Benefits of Political and Economic Integration of Hong Kong with Mainland China , 2012 .

[17]  Frank Schorfheide,et al.  Shrinkage Estimation of High-Dimensional Factor Models with Structural Instabilities , 2013 .

[18]  F. Wolff,et al.  The Local Effects of an Innovation: Evidence from the French Fish Market , 2017 .

[19]  David Powell.  Synthetic Control Estimation Beyond Case Studies: Does the Minimum Wage Reduce Employment? , 2016 .

[20]  Liangjun Su,et al.  Shrinkage Estimation of Dynamic Panel Data Models with Interactive Fixed Effects , 2015 .

[21]  Marcelo J. Moreira,et al.  DECISION THEORY APPLIED TO A LINEAR PANEL DATA MODEL , 2009 .

[22]  J. Bai,et al.  Panel Data Models With Interactive Fixed Effects , 2009 .

[23]  P. Bickel,et al.  SIMULTANEOUS ANALYSIS OF LASSO AND DANTZIG SELECTOR , 2008, 0801.1095.

[24]  Terence Tao,et al.  The Dantzig selector: Statistical estimation when P is much larger than n , 2005, math/0506081.

[25]  Peter Schmidt,et al.  GMM estimation of linear panel data models with time-varying individual effects , 2001 .

[26]  C. Adams.  Identification of Treatment Effects from Synthetic Controls , 2017 .

[27]  M. Pesaran.  Estimation and Inference in Large Heterogeneous Panels with a Multifactor Error Structure , 2004, SSRN Electronic Journal.

[28]  J. Bai,et al.  Determining the Number of Factors in Approximate Factor Models , 2000 .

[29]  M. Moslehian.  Ky Fan inequalities , 2011, 1108.1467.

[30]  Marc K. Chan,et al.  Policy Evaluation with Interactive Fixed Effects , 2016 .

[31]  Martin Weidner,et al.  DYNAMIC LINEAR PANEL REGRESSION MODELS WITH INTERACTIVE FIXED EFFECTS , 2014, Econometric Theory.

[32]  Seung C. Ahn,et al.  Panel Data Models with Multiple Time-Varying Individual Effects , 2013 .

[33]  Laurent Gobillon,et al.  Regional Policy Evaluation: Interactive Fixed Effects and Synthetic Controls , 2014, Review of Economics and Statistics.

[34]  Victor Chernozhukov,et al.  Inference for heterogeneous effects using low-rank estimations , 2018, 1812.08089.

[35]  Seung C. Ahn,et al.  Eigenvalue Ratio Test for the Number of Factors , 2013 .

[36]  Kathleen T. Li,et al.  Inference for Factor Model Based Average Treatment Effects , 2018 .

[37]  J. A. Rodríguez,et al.  Linear and Multilinear Algebra , 2007 .

[38]  M. Weidner,et al.  Linear Regression for Panel with Unknown Number of Factors as Interactive Fixed Effects , 2014 .