Large Panels with Common Factors and Spatial Correlation

TLDR: In this paper, the authors consider the statistical analysis of large panel data sets where even after condi- tioning on common observed eects the cross section units might remain dependently distrib- uted.

About: This article is published in Journal of Econometrics.The article was published on 2011-04-01 and is currently open access. It has received 618 citations till now. The article focuses on the topics: Spatial dependence & Estimator.

Frequently asked questions

Q1. What have the authors contributed in "Large panels with common factors and spatial correlation" ?

This paper considers methods for estimating the slope coe¢ cients in large panel data models that are robust to the presence of various forms of error cross section dependence. Initially, this paper focuses on a panel regression model where the idiosyncratic errors are spatially dependent and possibly serially correlated, and derives the asymptotic distributions of the mean group and pooled estimators under heterogeneous and homogeneous slope coe¢ cients, and for these estimators proposes non-parametric variance matrix estimators. The paper then considers the more general case of a panel data model with a multifactor error structure and spatial error correlations. Under this framework, the Common Correlated E¤ects ( CCE ) estimator, recently advanced by Pesaran ( 2006 ), continues to yield estimates of the slope coe¢ cients that are consistent and asymptotically normal. The authors are grateful to the Editor ( Cheng Hsiao ), an Associate Editor and three anonymous referees, Badi Baltagi, Alexander Chudik and George Kapetanios for helpful comments and suggestions. 

Q2. What is the main aim of this paper?

The main aim of this paper has been to consider estimation of a panel regression model under a number ofdi¤erent speci cations of cross section error correlations, such as spatial and/or common factor models. 

Q3. What is the advantage of using the above approach over standard spatial techniques?

Another advantage of using the above approach over standard spatial techniques is that, while allowing for serially correlated errors, it does not entail information on the time seriesprocesses underlying "it, so long as these processes are covariance stationary. 

Q4. What is the need for a consistent estimation of uit?

Estimation of uit is needed for computing tests of error cross section independence, or when the objects of interest are the coe¢ cients of the spatial process, eit. 

Q5. What is the purpose of the third set of experiments?

In the third set of experiments, the authors make a numberof robustness checks, to see the extent to which their estimators are e¤ective in dealing with various specialconditions, such as when errors are serially correlated, there are sizeable spatial error correlations, or when thepattern of cross section dependence varies over time. 

Q6. What are the main reasons for the uctuations in macroeconomics?

In macroeconomics, several studies have argued businesscycle uctuations could be the result of both strategic interactions as well as aggregate technological shocks(Cooper and Haltiwanger (1996)). 

Q7. What is the attraction of the CCE type estimators in these contexts?

The attraction of the CCE type estimators in these contexts lies in the fact that they do notrequire a quanti cation of the exact relative position of the units in space, which is required by the SHAC typeestimators. 

Q8. What is the main attraction of the CCE type estimators?

This important property of CCE type estimators is not necessarilyshared by estimation methods that use principal components (see Bai (2009)), since time variation in the degreeof cross section dependence can yield inconsistent estimates of the principal components. 

Q9. What are the common factors in the equations?

In the above equations, d1t and d2t are observed common factors, f1t, f2t, and f3t are unobserved common e¤ects, and eit are idiosyncratic errors. 

Q10. What is the significance of the results in Table C5?

results reported in Table C5 suggest that CCE estimators are also robust to possible time variationsin the degree of cross section dependence. 

Q11. What is the focus of this paper?

The focus of this paper is on estimating the slope coe¢ cients i, their cross section means, = E( i), and unit-speci c errors, uit. 

Q12. What is the difference between the spatial and the multifactor approaches?

The spatial approach assumes that the structure of cross section correlation is related to location anddistance among units, de ned according to a pre-speci ed metric.