Multiperil Rate Making for Property Insurance Using Longitudinal Data

Abstract

Summary In property insurance, a contract often provides the policyholder with protection against damages to the insured properties that arise from a variety of perils. We propose a multivariate framework for pricing property insurance contracts with multiperil coverage in a longitudinal context. Specifically, a two‐part model is employed to accommodate the excess of 0s and heavy tails in the insurance loss cost, and a Gaussian copula with a structured correlation is used to capture the dependence within and between perils, as well as their interaction. Using the government property insurance data from the state of Wisconsin in the USA, we show that the multiperil claim model has important implications in both experience rating and risk margin analysis.

Risk analysis of a multivariate aggregate loss model with dependence

“…Models that allow dependence between claim frequencies and claim sizes have also been developed recently. For example, Gschlößl and Czado (2007), Frees et al (2011), andGarrido et al (2016) took a regression approach where the claim frequency is treated as an explanatory variable in the regression model for the claim sizes; Boudreault et al (2006), Cossette et al (2008), and Marri and Furman (2012) assumed that the inter-claim times and claim sizes are dependent; Czado et al (2012), Frees et al (2016), Cossette et al (2019), and Oh et al (2020) employed bivariate copulas to model the dependency relationship between the number of claims and the average claim amount; Shi and Zhao (2020) used a copula to model the relation between the frequency and the individual severity directly; Yang and Shi (2019) proposed a multivariate framework for pricing property insurance contracts with multiperil coverage in the longitudinal context by using copulas to capture the dependence within and between perils.…”

Predictive Analytics and Risk Management of Property Insurance

“…Property insurance is marketed through open perils coverage, i.e., it covers losses from all causes not explicitly excluded in the policy. Standard exclusions include damages from natural disasters such as earthquakes and oods (Yang et al, 2019).…”

Assessing Hail Risk for Property Insurers with a Dependent Marked Point Process

“…This viewpoint has been much appreciated by recent studies in the field of insurance analytics. Among the key functional areas identified by Frees (2015), research using granular‐level data is found most active in ratemaking (see, for instance, Guillen et al., 2019; Shi et al., 2016; Yang and Shi, 2019) and loss reserving (see Antonio and Plat, 2014; Badescu et al., 2019; Pigeon et al., 2013 among others). In contrast, this work fits more in the area of claims management where insurers focus on strategies in cost reduction after occurrence of insured events.…”