In non-life insurance, territory-based risk classification is useful for various insurance operations including marketing, underwriting, ratemaking, etc. This paper proposes a spatially dependent frequency-severity modeling framework to produce territorial risk scores. The framework applies to the aggregated insurance claims where the frequency and severity components examine the occurrence rate and average size of insurance claims in each geographic unit, respectively. We employ the bivariate conditional autoregressive models to accommodate the spatial dependency in the frequency and severity components, as well as the cross-sectional association between the two components. Using a town-level claims data of automobile insurance in Massachusetts, we demonstrate applications of the model output–territorial risk scores–in ratemaking and market segmentation.

We develop a Bayesian model for continuous-time incurred but not yet reported (IBNYR) events under four types of secondary data, and show that unreported events, such as claims, have a Poisson distribution with a reduced arrival parameter if event arrivals are Poisson distributed. Using insurance claims as an example of an IBNYR event, we apply Markov chain Monte Carlo (MCMC) to the continuous-time IBNYR claims model of Jewell using Type I and Type IV data. We illustrate the relative stability of the MCMC method versus the Gammoid approximation of Jewell by showing that the MCMC estimates approach their prior parameters, while the Gammoid approximations grow without bound for Type IV data. Moreover, this holds for any distribution that the delay parameter is assumed to follow. Our framework also allows for the computation of posterior confidence intervals for the parameters.

The class of spectrally positive Lévy processes is a frequent choice for modelling loss processes in areas such as insurance or operational risk. Dependence between such processes (e.g. between different lines of business) can be modelled with Lévy copulas. This approach is a parsimonious, efficient and flexible method which provides many of the advantages akin to distributional copulas for random variables. Literature on Lévy copulas seems to have primarily focussed on bivariate processes. When multivariate settings are considered, these usually exhibit an exchangeable dependence structure (whereby all subset of the processes have an identical marginal Lévy copula). In reality, losses are not always associated in an identical way, and models allowing for non-exchangeable dependence patterns are needed. In this paper, we present an approach which enables the development of such models. Inspired by ideas and techniques from the distributional copula literature we investigate the procedure of nesting Archimedean Lévy copulas. We provide a detailed analysis of this construction, and derive conditions under which valid multivariate (nested) Lévy copulas are obtained. Our results are discussed and illustrated, notably with an example of model fitting to data.

Human longevity is changing, but at what rate? Insurance claims are increasing, but at what rate? Are the trends that we glean from data true or illusionary? The shocking fact is that true trends might be quite different from those that we actually see from visualised data. Indeed, in some situations the upward trends (e.g. inflation) may even look decreasing (e.g. deflation). In this paper, we discuss this “trends in disguise” phenomenon in detail and offer a way for estimating true trends.