An Alternative Estimation Method Based on Alpha Skew Logistic Distribution for Parameters of Censored Regression Model

Year 2019, Volume: 2 Issue: 2, 16 - 20, 28.12.2019

İsmail Yenilmez , Yeliz Mert Kantar

Abstract

In the case of censored data, it is often seen that the error distribution is skewed and multimodal. Ordinary least squares (OLS) estimator, which often gives biased and inconsistent results for censored data, and Tobit estimator, which is frequently used in censored data estimation and gives inconsistent results when some assumptions are not met, are also problematic in the presence of skewed and multimodal distribution of error terms. A new estimator is proposed as an alternative to the two conventional estimators used in the case of censored data. For censored regression model, anestimation method, known as partial adaptive or quasi-maximum likelihood estimator, has been introduced based on the alpha skewed logistic distribution, which is a flexible error distribution. According to the bias and mean square error (MSE), new estimator is superior forestimating the coefficients of the censored regression model under the skewed and multimodal error distribution

Keywords

partially adaptive estimator, maximum likelihood estimator, censored data, tobit, ordinary least squares estimator.

References

• [1] Tobin, J. "Estimation of Relationships for Limited Dependent Variables." Econometrica, 26, 24-36, 1958.
• [2] Goldberger, A. S. Econometric Theory. New York: Wiley, 1964.
• [3] Kennedy, P. A Guide to Econometrics (Fifth ed.). Cambridge: MIT Press. pp. 283–284. ISBN 0-262-61183-X, 2003.
• [4] Pagan A, Ullah A. Nonparametric econometrics. Cambridge University Press, Cambridge, 1999.
• [5] Powell J. L. "Least absolute deviations estimation of the censored regression model." J Econom 25:303–325, 1984.
• [6] Powell J. L. "Symmetrically trimmed least squares estimation for Tobit models." Econometrica 54:1435–1460, 1986.
• [7] Caudill, S. B. "A Partially Adaptive Estimator for The Censored Regression Model Based on A Mixture of Normal Distributions." Stats Methods Appl, 21:121-137, 2012.
• [8] McDonald, J. B., Xu Y. J. "A Comparison of Semi-parametric and Partially Adaptive Estimators of the Censored Regression Model with Possibly Skewed and Leptokurtic Error Distributions." Economics Letter, 51(2),153-159, 1996.
• [9] Lewis, R. A. and McDonald, J. B., "Partially Adaptive Estimation of the Censored Regression Model." Economic Reviews, 33 (7), 732-750, 2014.
• [10] Yenilmez, I., Kantar, Y.M., "A Partially Adaptive Estimator for the Censored Regression Model Based on Generalized Normal Distribution." Proceedings of the 3rd IRSYSC-2017, Konya, Turkey, 2017.
• [11] Yenilmez, I., Kantar, Y.M., Acitas, S., "Estimation of Censored Regression Model in the case of Non-Normal Error." Sigma J. Eng & Nat Sci 36 (2), 513-521, 2018.
• [12] Usta, I., Kantar, Y.M.,Yenilmez, I., "Estimation of Censored Regression Model with Maximum Entropy Distributions." Proceedings of the 4th IRSYSC-2018, İzmir, Turkey, 2018.
• [13] Martínez-Flórez, G., Bolfarine, H. and Gómez, H. W., "The Alpha-power Tobit Model." Communications in Statistics Theory and Methods, 42:4, 633-643, 2013.
• [14] Arabmazar, A., Schmidt, P., "An investigation of the robustness of the Tobit estimator to non-normality." Econometrica 50:1055–1069, 1982.
• [15] Greene, W. H., "On the asymptotic bias of the ordinary least squares estimator of the Tobit model." Econometrica 49:505–513, 1981.
• [16] Hazarika, P.J., Chakraborty, S., "Alpha-Skew-Logistic Distribution." IOSR-JM, 10(4):36-46, 2014.