1: \begin{abstract}

2: In this paper, a novel approach to the problem of estimating the heavy--tail exponent 

3: $\alpha>0$ of a distribution is proposed.  It is based on the fact that block--maxima of size $m$

4: of the independent and identically distributed data scale at a rate of $m^{1/\alpha}$.

5: This scaling rate can be captured well by the {\it max--spectrum} plot of the data that leads

6: to regression based estimators.

7: Consistency and asymptotic normality of these estimators is 

8: established under mild conditions on the behavior of the tail of the distribution. The 

9: results are obtained by establishing bounds on the

10: rate of convergence of moment--type functionals of heavy--tailed maxima.  Such

11: bounds often yield exact rates of convergence and are of independent interest.

12: Practical issues on the automatic selection of tuning parameters for the estimators 

13: and corresponding confidence intervals are also addressed.  Extensive numerical

14: simulations show that the proposed method proves competitive for both small and

15: large sample sizes and for a large range of tail exponents. The method is shown to be more 

16: robust than the classical Hill plot and is illustrated on two data sets of

17: insurance claims and natural gas field sizes.

18: \end{abstract}

19: