Smoothing the Hill Estimator

For sequences of i.i.d. random variables whose common tail 1 – F is regularly varying at infinity wtih an unknown index –α < 0, it is well known that the Hill estimator is consistent for α–1 and usually asymptotically normally distributed. However, because the Hill estimator is a function of k = k(n), the number of upper order statistics used and which is only subject to the conditions k →∞, k/n → 0, its use in practice is problematic since there are few reliable guidelines about how to choose k. The purpose of this paper is to make the use of the Hill estimator more reliable through an averaging technique which reduces the asymptotic variance. As a direct result the range in which the smoothed estimator varies as a function of k decreases and the successful use of the esimator is made less dependent on the choice of k. A tail empirical process approach is used to prove the weak convergence of a process closely related to the Hill estimator. The smoothed version of the Hill estimator is a functional of the tail empirical process.

[1]  Paul Deheuvels,et al.  Kernel Estimates of the Tail Index of a Distribution , 1985 .

[2]  U. Stadtmüller,et al.  Generalized regular variation of second order , 1996, Journal of the Australian Mathematical Society. Series A. Pure Mathematics and Statistics.

[3]  Alan H. Welsh,et al.  Adaptive Estimates of Parameters of Regular Variation , 1985 .

[4]  D. Mason,et al.  A tail empirical process approach to some nonstandard laws of the iterated logarithm , 1991 .

[5]  S. Resnick,et al.  Consistency of Hill's estimator for dependent data , 1995, Journal of Applied Probability.

[6]  Sidney I. Resnick,et al.  Testing for independence in heavy tailed and positive innovation time series , 1995 .

[7]  D. Mason,et al.  A strong invariance theorem for the tail empirical process , 1988 .

[8]  D. Mason,et al.  Central limit theorems for sums of extreme values , 1985, Mathematical Proceedings of the Cambridge Philosophical Society.

[9]  S. Resnick,et al.  The qq-estimator and heavy tails , 1996 .

[10]  S. Resnick.  Extreme Values, Regular Variation, and Point Processes , 1987 .

[11]  Peter Hall,et al.  Using the bootstrap to estimate mean squared error and select smoothing parameter in nonparametric problems , 1990 .

[12]  Tailen Hsing,et al.  On Tail Index Estimation Using Dependent Data , 1991 .

[13]  A. Dekkers,et al.  Optimal choice of sample fraction in extreme-value estimation , 1993 .

[14]  S. Resnick,et al.  Second-order regular variation, convolution and the central limit theorem , 1997 .

[15]  Liang Peng,et al.  Comparison of tail index estimators , 1998 .

[16]  D. Mason,et al.  Intermediate- and extreme-sum processes , 1992 .

[17]  E. Haeusler,et al.  On Asymptotic Normality of Hill's Estimator for the Exponent of Regular Variation , 1985 .

[18]  J. Geluk,et al.  Regular variation, extensions and Tauberian theorems , 1987 .

[19]  Malcolm R Leadbetter,et al.  Tail and Quantile Estimation for Strongly Mixing Stationary Sequences , 1990 .

[20]  B. M. Hill,et al.  A Simple General Approach to Inference About the Tail of a Distribution , 1975 .

[21]  D. Mason.  Laws of Large Numbers for Sums of Extreme Values , 1982 .

[22]  Sidney I. Resnick,et al.  Estimating the limit distribution of multivariate extremes , 1993 .

[23]  P. Billingsley,et al.  Convergence of Probability Measures , 1969 .

[24]  W. Vervaat.  Functional central limit theorems for processes with positive drift and their inverses , 1972 .

[25]  P. Hall.  On Some Simple Estimates of an Exponent of Regular Variation , 1982 .

[26]  Charles M. Goldie,et al.  SLOW VARIATION WITH REMAINDER: THEORY AND APPLICATIONS , 1987 .