Maximum likelihood estimators in a statistical model of natural catastrophe claims with trend (Q2488463)

The catastrophe claims form a sequence \(\{X_1,\dots,X_n\}\) of independent r.v.s with CDFs \(F_i\). In the Nevzorov record model \( F_i=F^{\gamma^{i-1}}\) only record indicators \(I_i=\mathbf{1}\{X_i>\max\{X_1,\dots,X_{i-1}\}\}\) are observed. In the three-parameter model \(F_i(x)=\exp(-\gamma^{i-1}(Ax)^{-\alpha})\), \(x>0\) and \(X_i\) are observed. The maximum likelihood estimators for the ``trend'' parameter \(\gamma\) are constructed in both models and their asymptotic normality is demonstrated. Analysis of data on hurricane (US) and taifun (Japan) events claims is considered.