Semiparametric tail-index estimation for randomly right-truncated heavy-tailed data

Purpose The purpose of this paper is to propose a semiparametric estimator for the tail index Pareto-type random truncated data that improves existing ones in terms mean square error. Moreover, we establish its consistency and asymptotic normality. Design/methodology/approach To construct root squared error (RMSE)-reduced index, authors used underlying distribution function given by Wang (1989). This allows us define corresponding process provide weak approximation one. By means functional representation using approximation, normality aforementioned RMSE-reduced estimator. Findings In basis on function, proposed new estimation method distributions randomly right-truncated data. Compared with ones, behaves well both bias RMSE. A useful empirical allowed Originality/value (empirical) introduced, introduced established.