Abstract
It is known that the analysis of short panel time series data is very important in many practical problems. This paper calculates the exact moments up to order 4 under the null hypothesis of no serial correlation when there are many independent replications of size 3. We further calculate the tail probabilities under the null hypothesis using the Edgeworth approximation for \(n=3, 4\) and \(5,\) when the structure of the pdf (probability density function) of the test statistic is in essence different. Finally, we compare the three types of tail probabilities, namely, the Edgeworth approximation, the normal approximation and the exact probabilities through a large scale simulation study.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Baltagi BH, Song SH (2006) Unbalanced panel data: a survey. Stat Pap 47:493–523
Born B, Breitung J (2011) Testing for serial correlation in fixed-effect panel data models. Working paper, Institute of Economics. University of Bonn, Germany, May 2011
Cox DR, Solomon PJ (1988) On testing for serial correlation in large number of small samples. Biometrika 75(1):145–148
Czado C, Song Peter X-K (2008) State space mixed models for longitudinal observations with binary and binomial responses. Stat Pap 49:691–714
Hjellvik V, Tjøstheim D (1999) Modelling panels of intercorrelated autoregressive time series. Biometrika 86(1):145–148
Moran P (1948) Some theorems on time series: the significance of the serial correlation. Biometrika 35(3):255–260
Mudelsee M (2001) Note on the bias in the estimation of the serial correlation coefficient of AR(1) processes. Stat Pap 42:517–527
Phillips PCB (1978) Edgeworth and saddlepoint approximations in the first order noncircular autoregression. Biometrika 65:91–98
Peiris S, Rao CR (2003) On testing for serial correlation in large number of small samples using tail probability approximations (extract). Bulletin of the International Statistical Institute (ISI), 54th Session, Berlin, pp 232–233
Peiris S, Rao CR (2004) On testing for serial correlation in large number of small samples using tail probability approximations. Commun Stat Theory Methods 33(8):1767–1777
Perera DI, Peiris MS, Robinson J, Weber NC (2006) Saddlepoint approximation methods for testing of serial correlation in panel time series data. J Stat Comput Simul 76(11):1001–1015
Perera D, Peiris S, Robinson J, Weber N (2008) The empirical saddlepoint method applied to testing for serial correlation in panel time series data. Stat Probab Lett 78:2876–2882
Rai SN, Abraham B, Peiris S (1995) Analysis of short time series with an over-dispersion model. Commun Stat Theory Methods 24(2):335–348
Song HS (1996) Consistency and asymptotic unbiasedness of \(S^2\) in the serially correlated error components regression model for panel data. Stat Pap 37:267–275
Stuart A, Ord K (1994) Kendall’s advanced theory of statistics, 6th edn, vol I. Edward Arnold, New York
Acknowledgments
The author wishes to thank two anonymous referees and the editor for their comments and suggestions to improve the quality, presentation and readability of this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Peiris, S. Testing the null hypothesis of zero serial correlation in short panel time series: a comparison of tail probabilities. Stat Papers 55, 513–523 (2014). https://doi.org/10.1007/s00362-012-0495-5
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00362-012-0495-5