, Volume 3, Issue 3, pp 231–249 | Cite as

Robust Estimation of Tail Parameters for Two-Parameter Pareto and Exponential Models via Generalized Quantile Statistics

  • Vytaras Brazauskas
  • Robert Serfling


Robust estimation of tail index parameters is treated for (equivalent) two-parameter Pareto and exponential models. These distributions arise as parametric models in actuarial science, economics, telecommunications, and reliability, for example, as well as in semiparametric modeling of upper observations in samples from distributions which are regularly varying or in the domain of attraction of extreme value distributions. New estimators of “generalized quantile” type are introduced and compared with several well-established estimators, for the purpose of identifying which estimators provide favorable trade-offs between efficiency and robustness. Specifically, we examine asymptotic relative efficiency with respect to the (efficient but nonrobust) maximum likelihood estimator, and breakdown point. The new estimators, in particular the generalized median types, are found to dominate well-established and popular estimators corresponding to methods of trimming, least squares, and quantiles. Further, we establish that the least squares estimator is actually deficient with respect to both criteria and should become disfavored. The generalized median estimators manifest a general principle: “smoothing” followed by “medianing” produces a favorable trade-off between efficiency and robustness.

robust estimation tail index Pareto model exponential model generalized L-statistics 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Adler, R.J., Feldman, R., and Taqqu, M., (eds.) A User's Guide to Heavy Tails: Statistical Techniques for Analyzing Heavy Tailed Distributions, Birkhüser, Boston, 1997.Google Scholar
  2. Arnold, B.C., Pareto Distributions, International Cooperative Publishing House, Fairland, Maryland, 1983.Google Scholar
  3. Bingham, N.H., Goldie, C.M., and Teugels, J.L., Regular Variation, Cambridge University Press, Cambridge, 1987.Google Scholar
  4. Brazauskas, V. and Serfling, R., “Robust and efficient estimation of the tail index of a single-parameter Pareto distribution,” North American Actuarial J. 4, 12-27, (2000).Google Scholar
  5. Brazauskas, V. and Serfling, R., “Small sample performance of robust estimators of tail parameters for Pareto and exponential models,” J. Statist. Comput. Simul. in press. (2001).Google Scholar
  6. Choudhury, J. and Serfling, R., “Generalized order statistics, Bahadur representations and sequential nonparametric fixed-width confidence intervals,” J. Statist. Plann. and Inf. 19, 269-282, (1988).Google Scholar
  7. Dekkers, A.L.M. and de Haan, L., “On the estimation of the extreme-value index and large quantile estimation,” Ann. Statist. 17, 1795-1832, (1989).Google Scholar
  8. Falk, M., “Extreme quantile estimation in δ-neighborhoods of generalized Pareto distributions,” Statist. Prob. Letters 20, 9-21, (1994).Google Scholar
  9. Falk, M. and Marohn, F., “Von Mises conditions revisited,” Ann. Statist. 21, 1310-1328, (1993).Google Scholar
  10. Galambos, J., The Asymptotic Theory of Extreme Order Statistics, Wiley, New York, 1978.Google Scholar
  11. Gather, U., “Robust estimation of the mean of the exponential distribution in outlier situation,” Commun. Statist.—Theor. Meth. 15, 2323-2345, (1986).Google Scholar
  12. Gomes, C.P., Selman, B., and Crato, N., “Heavy-tailed distributions in combinatorial search,” in G. Smolka, (ed.), Principles and Practice of Constraint Programming CP-97, Lecture Notes in Computer Science 1330 121-135. Springer, New York, 1997.Google Scholar
  13. Johnson, N.L., Kotz, S., and Balakrishnan, N., Continuous Univariate Distributions, Vol 1, 2nd ed, Wiley, New York, 1994.Google Scholar
  14. Kimber, A.C., “Trimming in gamma samples,” Appl. Statist. 32, 7-14, (1983a).Google Scholar
  15. Kimber, A.C., “Comparison of some robust estimators of scale in gamma samples with known shape,” J. Statist. Comput. Simul. 18, 273-286, (1983b).Google Scholar
  16. Koutrouvelis, I.A., “Large-sample quantile estimation in Pareto laws,” Commun. Statist.—Theor. Meth. A10, 189-201, (1981).Google Scholar
  17. Lehmann, E.L., Theory of Point Estimation, Wiley, New York (transferred to Wadsworth, 1991) 1983.Google Scholar
  18. Quandt, R.E., “Old and new methods of estimation and the Pareto distribution,” Metrika 10, 55-82, (1966).Google Scholar
  19. Reiss, R.-D., Approximate Distributions of Order Statistics, Springer, New York, 1989.Google Scholar
  20. Resnick, S.I., “Heavy tail modeling and teletraffic data (with discussion),” Ann. Statist. 25, 1805-1869, (1997).Google Scholar
  21. Robbins, H., “A remark on Stirling's formula,” Amer. Math. Monthly 62, 26-29, (1965).Google Scholar
  22. Saleh, A.K., Md. E., and Ali, M.M., “Asymptotic optimum quantiles for the estimation of the parameters of the negative exponential distribution,” Ann. Math. Statist. 37, 143-151, (1966).Google Scholar
  23. Sarhan, A.E., Greenberg, B.G., and Ogawa, J., “Simplified estimates for the exponential distribution,” Ann. Math. Statist. 34, 102-116, (1963).Google Scholar
  24. Sen, P.K., Sequential Nonparametrics, Wiley, New York, 1981.Google Scholar
  25. Serfling, R.J., Approximation Theorems of Mathematical Statistics, Wiley, New York, 1980.Google Scholar
  26. Serfling, R., “Generalized L-, M-and R-Statistics,” Ann. Statist. 12, 76-86, (1984).Google Scholar
  27. Willemain, T.R., Allahverdi, A., DeSautels, P., Eldredge, J., Gur, O., Miller, M., Panos, G., Srinivasan, A., Surtihadi, J., and Topal, E., “Robust estimation methods for exponential data: A monte-carlo comparison,” Commun. Statist.—Simula 21, 1043-1075, (1992).Google Scholar

Copyright information

© Kluwer Academic Publishers 2000

Authors and Affiliations

  • Vytaras Brazauskas
    • 1
  • Robert Serfling
    • 2
  1. 1.Department of Mathematical SciencesUniversity of Wisconsin-MilwaukeeMilwaukeeUSA
  2. 2.Department of Mathematical SciencesUniversity of Texas at DallasRichardsonUSA

Personalised recommendations