Three Essays on Testing for Cross-Sectional Dependence and Specification in Large Panel Data Models

<p>This dissertation consists of three essays on testing for cross-sectional dependence and specification in large panel data models. The first two essays are based on the papers joint with Prof. Badi H. Baltagi and Prof. Chihwa Kao; the third essay is based on the working paper joint with Prof. Lee. The first essay considers testing for Sphericity with non-normality in a fixed effects panel data model. The second essay considers testing for cross-sectional dependence in heterogeneous large N and large T panel data models allowing error serial correlation. The third essay considers the tests of specification in large N and large T dynamic panel data models.</p> <p>The first essay proposes a test for sphericity in a fixed</p> <p>effects panel data regression model which is robust to non-normality of the</p> <p>disturbances. It builds up on the work of Chen, Zhang and Zhong (2010) who</p> <p>use $U$-statistics to test for sphericity of the variance-covariance matrix</p> <p>in statistics. Since the errors are unobservable, the residuals from the</p> <p>fixed effects regression are used. The limiting distribution of the proposed</p> <p>test statistic is derived. Additionally, its finite sample properties are</p> <p>examined using Monte Carlo simulations.</p> <p>The second essay considers the problem of testing cross-sectional dependence in</p> <p>large panel data models with serially correlated errors. It finds that</p> <p>existing tests for cross-sectional independence encounter size distortions</p> <p>with serial correlation in the errors. To control the size, it</p> <p>proposes a modification of Pesaran's CD\ test to account for serially</p> <p>correlation of an unknown form in the error term. We derive the limiting</p> <p>distribution of this test as $\left( N,T\right) \rightarrow\infty$. The test</p> <p>is distribution free and allows for unknown forms of serial correlation in the</p> <p>errors. Monte Carlo simulations show that the test has good size and power for</p> <p>large panels when serial correlation in the errors are present.</p> <p>The third essay considers the tests of specification,</p> <p>including the tests for serial correlation and the tests of overidentifying</p> <p>restrictions, for large dynamic panel data models.\ The test statistics are</p> <p>built upon the two-step GMM estimations using three different instrument</p> <p>matrices: the block-diagonal matrix with a full set of all available</p> <p>instruments, the block-diagonal matrix with a subset of all available</p> <p>instruments and the collapsed matrix with a subset of all available</p> <p>instruments. It shows that the conventional Sargan's test of overidentifying</p> <p>restrictions (Arellano and Bond (1991)) does not approximate to the chi-square</p> <p>distribution, when the number of instruments used, is relatively as large as</p> <p>$N$; therefore it proposes corrected Sargan's tests with different instrument</p> <p>matrices. The limiting distributions of all the tests of specification are</p> <p>derived as $N$ and $T$ go to infinity simultaneously. Power properties are</p> <p>discussed under varieties of alternatives, The results show that the tests for</p> <p>serial correlation are powerful against different alternatives, and the power</p> <p>of corrected Sargan's tests only increases as $N$ increases. Monte Carlo</p> <p>simulations confirm our theoretical findings, especially showing that the</p> <p>corrected Sargan's tests have the correct size. Besides, it suggests using the</p> <p>collapsed instrument matrix for the practical testing purpose.</p>

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Economics

Year

2016

Author dc:creator

• Peng, Bin

Contributors dc:contributor

• Chihwa Kao

• Cross-sectional Dependence
• Large N Large T
• Panel Data Models
• Serial Correlation
• Sphericity
• Tests of Specification
• Social and Behavioral Sciences

Repository record dc:identifier

https://surface.syr.edu/etd/524

OAI identifier oai:identifier

oai:surface.syr.edu:etd-1524