The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

A Comparison of Bayesian and Maximum Likelihood Estimations in a SUR Tobit Regression Model

SUR 토빗회귀모형에서 베이지안 추정과 최대가능도 추정의 비교

  • Lee, Seung-Chun (Department of Applied Statistics, Hanshin University) ;
  • Choi, Byongsu (Department of Multimedia Engineering, Hansung University)
  • 이승천 (한신대학교 응용통계학과) ;
  • 최병수 (한성대학교 멀티미디어학과)
  • Received : 2014.08.13
  • Accepted : 2014.10.07
  • Published : 2014.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Both Bayesian and maximum likelihood methods are efficient for the estimation of regression coefficients of various Tobit regression models (see. e.g. Chib, 1992; Greene, 1990; Lee and Choi, 2013); however, some researchers recognized that the maximum likelihood method tends to underestimate the disturbance variance, which has implications for the estimation of marginal effects and the asymptotic standard error of estimates. The underestimation of the maximum likelihood estimate in a seemingly unrelated Tobit regression model is examined. A Bayesian method based on an objective noninformative prior is shown to provide proper estimates of the disturbance variance as well as other regression parameters

Greene (2004a,b), Lee와 Choi (2014) 등의 연구에서 토빗 회귀모형의 최대가능도 추정은 표준오차를 과소추정한다는 것이 알려졌고, 그 원인의 하나는 오차항 분산의 과소 추정에 있다고 한다. 오차항 분산의 과소 추정은 회귀계수에 대한 가설 검정 및 구간추정에 영향을 미칠 뿐 아니라 독립변수들의 주변효과를 추정하는데에도 영향을 미치게 되므로 토빗 회귀모형에 대한 적절한 분석이 수행되려면 최대가능도 추정의 과소 추정 문제를 해결하여야 한다. 일반적으로 무정보 사전분포에 의한 베이지안 추론 방법은 빈도학파들이 요구하는 효율성을 갖는 경우가 많다. 본 연구에서도 무정보 사전분포에 의한 베이지안 추론을 적용하여, 베이지안 방법론이 SUR 토빗 회귀모형에서 최대가능도 추정의 과소 추정 문제를 해결할 수 있는 하나의 대안이 될 수 있다는 것을 보였다.

Keywords

References

  1. Amemiya, T. (1984). Tobit models: A survey, Journal of Econometrics, 24, 3-61 https://doi.org/10.1016/0304-4076(84)90074-5
  2. Bruno, G. (2004). Limited dependent panel models: A comparative analysis of classical and Bayesian inference among econometrics packages, Computing in Economics and Finance. Society for Computational Economics, http://editorialexpress.com/cgi-bin/conference/ download.cgi?db name=SCE2004%&paper id=41.
  3. Chib, S. (1992). Bayesian inference in the Tobit censored regression model, Journal of Econometrics, 51, 77-99.
  4. Cowles, M. K., Carlin, B. P. and Connett, J. E. (1996). Bayesian Tobit modeling of longitudinal ordinal clinical trial compliance data with nonignorable missingness, Journal of American Statistical Association, 91, 86-98. https://doi.org/10.1080/01621459.1996.10476666
  5. Daniels, M, D. and Kass, R. E. (1999). Nonconjugate Bayesian estimation of covariance matrices and its use in hierarchical models, Journal of American Statistical Association, 94, 1254-1263. https://doi.org/10.1080/01621459.1999.10473878
  6. Efron, B. and Hinkley, D. V. (1978). The observed versus expected information, Biometirika, 65, 163-168.
  7. Greene, W. H. (1990). Econometric Analysis, Macmillan, New York.
  8. Greene, W. (2004a). The behavior of the maximum likelihood estimator of limited dependent variable models in the presence of fixed effects, Econometrics Journal, 7, 98-119. https://doi.org/10.1111/j.1368-423X.2004.00123.x
  9. Greene, W. (2004b). Fixed effects and bias due to the incidental parameters problem in the Tobit model, Econometrics Reviews, 23, 125-147. https://doi.org/10.1081/ETC-120039606
  10. Greene, W. (2012). Econometric Analysis. 7th edition, Pearson.
  11. Hamilton, B. H. (1999). HMO selection and medicare costs: Bayesian MCMC estimation of a robust panel data tobit model with survival, Health Economics and Econometrics, 8, 403-414. https://doi.org/10.1002/(SICI)1099-1050(199908)8:5<403::AID-HEC455>3.0.CO;2-D
  12. Huang, C. J., Sloan, F. and Adamache, K, W. (1987). Estimation of seemingly unrelated Tobit regression via EM algorithm, Journal of Business & Economic Statistics, 5, 425-430.
  13. Huang, H. C. (1999). Estimation of the SUR Tobit model via the MCECM algorithm, Economics Letters, 64, 25-30. https://doi.org/10.1016/S0165-1765(99)00075-0
  14. Huang, H. C. (2001). Bayesian analysis of the SUR Tobit model, Applied Economics Letters, 8, 617-622. https://doi.org/10.1080/13504850010026069
  15. Joreskog (2004). Multivariate censored regression, available at www.ssicentral.com/lisrel/column12.htm.
  16. Johnson, S. G. and Narasimhan, B. (2014). Package Cubature, http://ab-initio.mit.edu/wiki/index.php/Cubature.
  17. Lancaster, T. (2000). The incidental parameter problem since 1948, Journal of Econometrics, 95, 391-413. https://doi.org/10.1016/S0304-4076(99)00044-5
  18. Lee, S.C. and Choi, B. (2013). Bayesian interval estimation of Tobit regression model, The Korean Journal of Applied Statistics, 26, 737-746. https://doi.org/10.5351/KJAS.2013.26.5.737
  19. Lee, S.C. and Choi, B. (2014). Bayesian inference for censored panel regression model, Communications for Statistical Applications and Methods, 21, 192-200.
  20. Louis, T. A. (1982). Finding the observed information matrix when using the EM algorithm, Journal of the Royal Statistical Society: Series B, 44, 226-233.
  21. Natarajan, R. and Kass, R. R. (2000). Bayesian methods for generalized linear mixed models, Journal of the American Statistical Association, 95, 227-237. https://doi.org/10.1080/01621459.2000.10473916
  22. Tanner, M. A. and Wong, W. H. (1987). The calculation of posterior distributions by data augmentation (with discussion), Journal of the American Statistical Association, 82, 528-550. https://doi.org/10.1080/01621459.1987.10478458
  23. Tierney, L. and Kadane, J. B. (1986). Accurate approximations for posterior moments and marginal densities, Journal of American Statistical Association, 81, 82-86. https://doi.org/10.1080/01621459.1986.10478240
  24. Wichitaksorn, N. and Choy, S. T. B. (2011). Modeling dependence of seemingly unrelated Tobit model through copula: A Bayesian analysis, Thailand Econometrics Society, 3, 6-19.

Cited by

  1. The restricted maximum likelihood estimation of a censored regression model vol.24, pp.3, 2017, https://doi.org/10.5351/CSAM.2017.24.3.291