Improving Standard Moment Estimators of Beta Random Variables

改進Beta隨機變數之標準動差估計式

李瑞祥(Ray-Shine Lee)

《經濟論文》 47卷4期 (2019/12) Pp. 547-570

摘要

The beta distribution of the first kind, including two shape parameters, is a flexible curve specification in studying the classical moment method of statistical principles. The research of this paper, originating with a need similar to that in econometrics, further finds a sequence of explicit high-order moment estimators for the beta distribution. In addition to the trials of weighting different moment estimators, this research also examines a deserving-emphasis condition for estimating the classic four-parameter beta distribution, and permitting moment-equation substitution.

關鍵字

Beta distribution ； The method of moments ； Maximum likelihood method ； Jackknife estimator

並列摘要

第一類型貝塔分佈，包括兩個形狀參數，是學習統計學原理中古典動差方法的一種易於伸縮曲線設定。本文的研究源自計量經濟學之相似需要，進一步發現貝塔分佈具有一系列外顯高階動差估計式。除了對不同估計式進行加權試驗外，這項研究還考察了一個值得重視的情況，估計經典的四參數貝塔分佈，並允許動差方程替換。

並列關鍵字

Beta分佈 ； 動差法 ； 最大概似法 ； Jackknife估計式

參考文獻

Andrews, D. W. K. (2000), "Inconsistency of the Bootstrap When a Parameter Is on the Bound- ary of the Parameter Space," Econometrica, 68, 399–405.

Carnahan, J. V. (1989), “Maximum Likelihood Estimation for the 4-Parameter Beta Distribu- tion," Communications in Statistics – Simulation and Computation, 18, 513–536. 99

Carrasco, M. and J.-P. Florens (2014), "On the Asymptotic Efficiency of GMM," Econometric 66 Theory, 30, 372–406.

Chang, S.-K. (2011), “A Computationally Practical Simulation Estimation for Dynamic Panel Tobit Models," Academia Economic Papers, 39, 1–32.

Chen, L.-Y. and S. Lee (2018), “Exact Computation of GMM Estimators for Instrumental Variable Quantile Regression Models," Journal of Applied Econometrics, 33, 553–567. 99