The use of spectral analysis in insurance cycle research

Abstract

Purpose

Aims to address a number of issues related to the use of spectral analysis in the study of insurance cycles.

Design/methodology/approach

Spectral analysis has seldom been used in the study of insurance cycles. This may be due to the fact that no statistical test is readily available for rejecting the hypothesis that a spectrum is significantly different from random uncorrelated noise in a context in which the period of the alternative is not known. This article suggests one such test.

Findings

In evaluating the proposed test, the relevant critical points, when the number of observations is small, and provided the power of the test is also explored to identify three cyclical processes: a sine process with noise, a second‐order autoregressive process, and the rational expectations process suggested by Cummins and Outreville.

Originality/value

The article provides the first comprehensive analysis and discussion of spectral analysis in the context of insurance‐cycle research.

Citation

Venezian, E.C. (2006), "The use of spectral analysis in insurance cycle research", Journal of Risk Finance, Vol. 7 No. 2, pp. 177-188. https://doi.org/10.1108/15265940610648616