Insurance Risk Classification Research Articles

The Applications of Generalized Poisson Regression Models to Insurance Claim Data

Predictive modeling has been widely used for insurance rate making. In this paper, we focus on insurance claim count data and address their common issues with more flexible modeling techniques. In particular, we study the zero-inflated and hurdle-generalized Poisson and negative binomial distributions in a functional form for modeling insurance claim count data. It is shown that these models are useful in addressing the problem of excess zeros and over-dispersion of the claim count variable. In addition, we show that including the exposure as a covariate in both the zero and the count part of the model is an effective approach to incorporating exposure information in zero-inflated and hurdle models. We illustrate the effectiveness and versatility of the introduced models using three real datasets. The results suggest their promising applications in insurance risk classification and beyond.

Non-Life Insurance Risk Classification Using Categorical Embedding

This article presents several actuarial applications of categorical embedding in the context of non-life insurance risk classification. In non-life insurance, many rating factors are naturally categorical, and often the categorical variables have a large number of levels. The high cardinality of categorical rating variables presents challenges in the implementation of traditional actuarial methods. Categorical embedding that is proposed in the machine learning literature for handling categorical variables has recently received attention in actuarial studies. The method is inspired by the neural network language models for learning text data and maps a categorical variable into a real-valued representation in the Euclidean space. Using a property insurance claims we demonstrate the use of categorical embedding in three applications. The first shows how embeddings are used to construct rating classes and calculate rating relativities for a single insurance risk. The second concerns predictive modeling for multivariate insurance risks and emphasizes the effects of dependence on tail risks. The third focuses on pricing new products where transfer learning is used to gather knowledge from existing products.

When is utilitarian welfare higher under insurance risk pooling?

This paper considers the effect of bans on insurance risk classification on utilitarian social welfare. We consider two regimes: full risk classification, where insurers charge the actuarially fair premium for each risk, and pooling, where risk classification is banned and for institutional or regulatory reasons, insurers do not attempt to separate risk classes, but charge a common premium for all risks. For iso-elastic insurance demand, we derive sufficient conditions on higher and lower risks' demand elasticities which ensure that utilitarian social welfare is higher under pooling than under full risk classification. Using the concept of arc elasticity of demand, we extend the results to a form applicable to more general demand functions. Empirical evidence suggests that the required elasticity conditions for social welfare to be increased by a ban may be realistic for some insurance markets.

The efficiency of voluntary risk classification in insurance markets

AbstractIt has been established that categorical discrimination based on observable characteristics such as gender, age, or ethnicity enhances efficiency. We consider a different form of risk classification when there exists a costless yet imperfectly informative test of risk type, with the test outcome unknown to the agents ex ante. We show that a voluntary risk classification in which agents are given the option to take the test always increases efficiency compared with no risk classification. Moreover, voluntary risk classification also Pareto dominates a regime of compulsory risk classification in which all agents are required to take the test.

Reconsidering insurance discrimination and adverse selection in an era of data analytics

This article demonstrates how replacing age- and gender-based pricing variables with telematics data in auto insurance risk classification systems minimises insurance discrimination and increases cream skimming adverse selection. The study explains how incorporating telematics data into insurance pricing schemes reduces pricing heterogeneity, consistent with the anti-discrimination objective of Aristotelian equality. It also describes how anti-discrimination prohibitions of age- and gender-based insurance pricing can result in regulatory adverse selection; traditional adverse selection, where asymmetric information favouring applicants can result in an overpopulation of high-risk drivers in risk pools; and cream skimming adverse selection, where asymmetric information favouring telematics-based insurers supports premium discounts that attract safer drivers, prompting an underpopulation of low-risk drivers among non-telematics insurers. The study explains how insurers minimise their vulnerability to cream skimming by quickly entering the pricing “arms race” with their own telematics-based products and how incorporating telematics data increases the efficiency of risk classification systems.

Modern regulatory frameworks for the use of genetic and epigenetic underwriting technology in life insurance

Following Florida's recent prohibition on the use of genetic information in life insurance risk classification, this paper analyzes insurance regulatory law and policy pertaining to the use of modern molecular biotechnology and InsurTech--particularly in the developing field of epigenetics, which studies the area above and around a human cell's immutable DNA, and the changes in gene expression brought about by a person's behavior and experiences. Public policy to date, from the health insurance prohibitions in the 2008 federal Genetic Information Nondiscrimination Act (GINA) to the 2020 GINA-like Florida prohibitions for life insurance, seeks to redress concerns that genetic information identifying pre-dispositional risk of disease could be used to penalize otherwise healthy insureds who have not manifested a diagnosable health condition. In contrast, epigenetic information, derived from a noninvasive saliva specimen, can be used to identify current health and wellness statuses, both negative and positive, that result from behavioral choice (such as tobacco, alcohol and drug use, as well as diet and exercise), with more precision than traditional, incumbent paramedical testing methods. By providing more accurate information about insureds' current and changeable health status, while avoiding the perceived unfairness of immutable genetic information, the use of epigenetic testing in life insurance underwriting would conform to the fundamental regulatory norms of information transparency and symmetry; accurate risk classification; sound risk pool construction; and a resulting alignment of interests between insurers and insureds

Insurance loss coverage and social welfare

ABSTRACTRestrictions on insurance risk classification may induce adverse selection, which is usually perceived as a bad outcome, both for insurers and for society. However, a social benefit of modest adverse selection is that it can lead to an increase in ‘loss coverage’, defined as expected losses compensated by insurance for the whole population. We reconcile the concept of loss coverage to a utilitarian concept of social welfare commonly found in the economic literature on risk classification. For iso-elastic insurance demand, ranking risk classification schemes by (observable) loss coverage always give the same ordering as ranking by (unobservable) social welfare.

Cream Skimming: Innovations in Insurance Risk Classification and Adverse Selection

AbstractWe demonstrate how innovations in insurance risk classification can lead to adverse selection, or cream skimming, against insurers that are slow to adopt such pricing innovations. Using a model in which insurers with insufficient pricing data cannot differentiate between low‐ and high‐risk policyholders and therefore charge both the same premium, we show how innovative insurers develop new risk classification data to identify overcharged low‐risk policyholders and attract them from rival insurers with reduced prices. Less innovative insurers thus insure a growing percentage of high‐risk customers, resulting in adverse selection attributable to their informational disadvantage. Next, we examine two cases in which “Big Data” innovations in risk classification led to concerns about cream skimming among U.S. auto insurers. First, we track the rapid adoption of credit‐based insurance scores as pricing variables in personal auto insurance markets. Second, we examine the growing popularity of usage‐based insurance programs like telematics, plans in which insurers use data on policyholders’ actual driving behavior to set prices that attract low‐risk customers. Issues associated with the execution of such pricing strategies are discussed. In both cases, we document how rival insurers quickly adopt successful innovations to reduce their exposure to adverse selection.

Insurance loss coverage and demand elasticities

Restrictions on insurance risk classification may induce adverse selection, which is usually perceived as a bad outcome. We suggest a counter-argument to this perception in circumstances where modest levels of adverse selection lead to an increase in ‘loss coverage’, defined as expected losses compensated by insurance for the whole population. This happens if the shift in coverage towards higher risks under adverse selection more than offsets the fall in number of individuals insured. The possibility of this outcome depends on insurance demand elasticities for higher and lower risks. We state elasticity conditions which ensure that for any downward-sloping insurance demand functions, loss coverage when all risks are pooled at a common price is higher than under fully risk-differentiated prices. Empirical evidence suggests that these conditions may be realistic for some insurance markets.

Empirical Evidence on the Use of Credit Scoring for Predicting Insurance Losses with Psycho-social and Biochemical Explanations

An important development in personal lines of insurance in the United States is the use of credit history data for insurance risk classification to predict losses. This research presents the results of collaboration with industry conducted by a university at the request of its state legislature. The purpose was to see the viability and validity of the use of credit scoring to predict insurance losses given its controversial nature and criticism as redundant of other predictive variables currently used. Working with industry and government, this study analyzed more than 175,000 policyholders’ information for the relationship between credit score and claims. Credit scores were significantly related to incurred losses, evidencing both statistical and practical significance. We investigate whether the revealed relationship between credit score and incurred losses was explainable by overlap with existing underwriting variables or whether the credit score adds new information about losses not contained in existing underwriting variables. The results show that credit scores contain significant information not already incorporated into other traditional rating variables (e.g., age, sex, driving history). We discuss how sensation seeking and self-control theory provide a partial explanation of why credit scoring works (the psycho-social perspective). This article also presents an overview of biological and chemical correlates of risk taking that helps explain why knowing risk-taking behavior in one realm (e.g., risky financial behavior and poor credit history) transits to predicting risk-taking behavior in other realms (e.g., automobile insurance incurred losses). Additional research is needed to advance new nontraditional loss prediction variables from social media consumer information to using information provided by technological advances. The evolving and dynamic nature of the insurance marketplace makes it imperative that professionals continue to evolve predictive variables and for academics to assist with understanding the whys of the relationships through theory development.

풍수해보험요율체계의 문제점 및 개선방안

Generally, a standard of insurance rate is rationally graded or applied according to size of risk(degree of risk) but present natural disaster insurance applies same insurance rate regardless of measuring size of risk(degree of risk) and classifies each 230 cities and districts into one risk grade. Accordingly, the current insurance rate is unfairly discriminatory as it is supposed to be graded and applied with the standard of insurance rate regarding size of risk. This thesis is to investigate and research flood insurance risk classification system of each country such as the US and Europe and it, on basis of research, suggests a natural disaster insurance manual rate system which can be graded and applied according to size of risk. Therefore, we, according to degree of risk, propose the newly designed class rate system as we regard the inundation depth and inundation area as a weight for rate discrimination.

Statistical concepts of a priori and a posteriori risk classification in insurance

Everyday we face all kinds of risks, and insurance is in the business of providing us a means to transfer or share these risks, usually to eliminate or reduce the resulting financial burden, in exchange for a predetermined price or tariff. Actuaries are considered professional experts in the economic assessment of uncertain events, and equipped with many statistical tools for analytics, they help formulate a fair and reasonable tariff associated with these risks. An important part of the process of establishing fair insurance tariffs is risk classification, which involves the grouping of risks into various classes that share a homogeneous set of characteristics allowing the actuary to reasonably price discriminate. This article is a survey paper on the statistical tools for risk classification used in insurance. Because of recent availability of more complex data in the industry together with the technology to analyze these data, we additionally discuss modern techniques that have recently emerged in the statistics discipline and can be used for risk classification. While several of the illustrations discussed in the paper focus on general, or non-life, insurance, several of the principles we examine can be similarly applied to life insurance. Furthermore, we also distinguish between a priori and a posteriori ratemaking. The former is a process which forms the basis for ratemaking when a policyholder is new and insufficient information may be available. The latter process uses additional historical information about policyholder claims when this becomes available. In effect, the resulting a posteriori premium allows one to correct and adjust the previous a priori premium making the price discrimination even more fair and reasonable.

Containing the Promise of Insurance: Adverse Selection and Risk Classification

This essay extends to adverse selection the critical attention provided in prior work to moral hazard. Like moral hazard, adverse selection is an old insurance concept that was adopted, formalized and generalized by economists developing the economics of information. As with moral hazard, insurance economics has addressed the phenomenon of adverse selection largely from the insurers' point of view. This essay examines how insurers create and shape adverse selection. At least in the context of insurance risk classification, there is much to be gained in thinking of adverse selection as a 'dual' problem (similar to moral hazard), meaning that actions to address adverse selection problems can lead to the de-pooling effect that motivated the actions in the first place. The first part of the essay sets forth the case for understanding adverse selection as a dual problem and highlights alternatives to insurance risk classification. The second uses three historical examples to explore moral justifications for insurance risk classification. The three examples are the 19th century controversy over age-based pricing in fraternal insurance, the mid 20th century controversy over experience rating in unemployment insurance, and the late 20th century controversy over efforts to exclude battered women from life, health and disability insurance pools. These examples demonstrate that, rather than being a neutral, technical solution to a structural dynamic inherent in the insurance relationship, risk classification reflects moral commitments.

The Econometrics of Risk Classification in Insurance

We present in this article some questions related to risk classification. These are discussed depending on the information used—either data on conditional characteristics or also including data on claim histories or on endogenous insurance demand by the agents.

233007 (M10, B20) A genetic approach to the modelling of sickness rates, with application to life insurance risk classification: Chadburn R.G.,City University London, Department of Actuarial Science and Statistics, Actuarial Research Paper No. 111, May 1998

233007 (M10, B20) A genetic approach to the modelling of sickness rates, with application to life insurance risk classification: Chadburn R.G.,City University London, Department of Actuarial Science and Statistics, Actuarial Research Paper No. 111, May 1998

Efficiency and Fairness in Insurance Risk Classification

Efficiency and Fairness in Insurance Risk Classification

The Impact of Imperfectly Categorizing Risks on Income Inequality and Social Welfare

The use of imperfect information to categorize risks is a pervasive (and controversial) phenomenon in the insurance industry. Although this activity reduces the 'average' extent of vertical price discrimination, it is shown in this paper that both highand low-risk types may prefer, ex ante to knowing the category to which they will be assigned, that such information be suppressed. Furthermore, for the Lorenz criterion it is shown that under certain restrictive demand conditions the imperfect categorization of risks leads to an increase in inequality. The paper employs the Wilson E2 equilibrium concept. L'impact d'une categorisation imparfaite des risques sur l'inegalite du revenu et le bien-etre social. L'utilisation de renseignements imparfaits pour categoriser les risques est un probleme important qui porte a controverses dans l'industrie de I'assurance. Meme si ce phenomene reduit le degre de discrimination verticale par les prix, en moyenne, I'auteur montre que a la fois les personnes dont le coefficient de risques est faible et elev6 (et ce avant de savoir a quelle categorie on va les assigner) peuvent fort bien preferer que cette information soit supprimee. De plus, 1'auteur montre que sous certaines conditions de demande restrictives, la categorisation imparfaite des risques conduit a un accroissement du degre d'inegalite mesure par le critere de Lorenz. Ce memoire utilise le concept d'equilibre dit de Wilson E2. If insurance firms are unable to distinguish between the risk types of their clients, but consumers know their own probability of loss (i.e., if asymmetric information persists), then the complication of adverse selection may arise. Adverse selection occurs when the class of risks (or set of classes) that purchases insurance differs 'adversely' from that anticipated by the firm. For example, if a firm offers a contract to low-risk types at their actuarially fair rate, then high-risk types will also purchase it, with the result that the firm will earn (expected) losses. Hence the problem of adverse selection restricts the ability of firms to offer efficient insurance contracts to low-risk types. It has been demonstrated by Rothschild and Stiglitz (1976) that, provided the proportion of high-risk types in the population is above some critical level, a This paper is a revised version of chapter 3 of the author's PHD thesis entitled 'The economics of risk classification in insurance.' London School of Economics and Political Science, 1982. I am indebted to my supervisor, Lucien Foldes, and an anonymous referee for useful comments. The author alone is responsible for any errors herein. Canadian Journal of Economics / Revue canadienne d'Economique, xvil, No. 3 August / aotlt 1984. Printed in Canada / Imprimd au Canada 0008-4085 / 84 / 557-568 $1.50 C) Canadian Economics Association This content downloaded from 157.55.39.215 on Wed, 31 Aug 2016 04:39:57 UTC All use subject to http://about.jstor.org/terms