Grant: Discovery Grants Program - Individual
Simon Fraser University
$18,000.00
Insurance ratemaking with consideration of dynamic credibility
EF 2023-24
Résumé
The overarching goal of my new research program is to develop statistical methods for the determination of insurance premiums that reflect individual heterogeneity of risk. In practical terms, the proposed research will identify and fill important gaps between actuarial literature and practice for experience-based ratemaking. To account for individual heterogeneity of risk, the traditional credibility premium formula, which is used widely, is a weighted sum of prior premium and historical average of past claims where prior premium is determined based on the observable rating factors. Despite the prevalence of such form of credibility premium, this traditional form has the following issues: ? It treats every past claim equally so that discrepancy can exist between the actuarial credibility premium and premium based on the bonus--malus system in practice, which puts more weight on recent claims. ? It needs to be extended for the case where multiple types of claims are observed for the same insured. In this case, we cannot preclude possible dependence among the multiple types of claims and credibility weight may also depend on the type of claims. ? In terms of the compound risk model, the observed claim can be divided into the frequency and severity components, which also gives rise to an issue to incorporate possible dependence between the frequency and severity components in the credibility premium. Therefore, over the next 5 years, my team will address the following short--term objectives: ? Objective 1 - To avoid instability in the optimization of the random effects models, we propose a version of expectation-maximization (EM) algorithm that optimizes the joint likelihood sequentially and provides more stability and efficiency in the optimization scheme, which would result in better estimates of regression coefficients and posterior distribution of the random effects. ? Objective 2 - We will perform empirical analyses with actual datasets to convey the implications and provide modified versions of credibility formula for possible applications in practice with rationale. ? Objective 3 - Given that the choice of correlation structure of random effects is equivalent to prior elicitation of the random effects, we propose the use of Bayesian sensitivity analysis that enables us to determine the appropriate serial dependence structure based on statistical reasoning with observed data rather than subjective judgement. Taken together, the expected outcomes of this work will provide advanced ratemaking methods with credibility that enable the insurance companies to reflect observed and unobserved heterogeneity of the insured appropriately. More broadly, the ratemaking methods with credibility could be applied to a wide range of fields in natural sciences and engineering in which a study incorporates longitudinal data, such as prediction of annual agricultural production based on the historical productions and characteristics of the land.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2021-02622
Name nom
Jeong, Himchan
Department département
Statistics & Actuarial Science
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2021
Award amount
18000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1205
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Mathematical sciences
Area of application fr
Sciences mathématiques
Research subject code
3011
Research subject group en
Statistics and probability
Research subject group fr
Multivariate analysis
Research subject en
Statistique et probabilité
Research subject fr
Analyse multivariée
Application title
Insurance ratemaking with consideration of dynamic credibility
Keyword
Bonus-malus system; Credibility; General insurance; Individual prospective ratemaking; Random effects models
Application summary
The overarching goal of my new research program is to develop statistical methods for the determination of insurance premiums that reflect individual heterogeneity of risk. In practical terms, the proposed research will identify and fill important gaps between actuarial literature and practice for experience-based ratemaking. To account for individual heterogeneity of risk, the traditional credibility premium formula, which is used widely, is a weighted sum of prior premium and historical average of past claims where prior premium is determined based on the observable rating factors. Despite the prevalence of such form of credibility premium, this traditional form has the following issues: ? It treats every past claim equally so that discrepancy can exist between the actuarial credibility premium and premium based on the bonus--malus system in practice, which puts more weight on recent claims. ? It needs to be extended for the case where multiple types of claims are observed for the same insured. In this case, we cannot preclude possible dependence among the multiple types of claims and credibility weight may also depend on the type of claims. ? In terms of the compound risk model, the observed claim can be divided into the frequency and severity components, which also gives rise to an issue to incorporate possible dependence between the frequency and severity components in the credibility premium. Therefore, over the next 5 years, my team will address the following short--term objectives: ? Objective 1 - To avoid instability in the optimization of the random effects models, we propose a version of expectation-maximization (EM) algorithm that optimizes the joint likelihood sequentially and provides more stability and efficiency in the optimization scheme, which would result in better estimates of regression coefficients and posterior distribution of the random effects. ? Objective 2 - We will perform empirical analyses with actual datasets to convey the implications and provide modified versions of credibility formula for possible applications in practice with rationale. ? Objective 3 - Given that the choice of correlation structure of random effects is equivalent to prior elicitation of the random effects, we propose the use of Bayesian sensitivity analysis that enables us to determine the appropriate serial dependence structure based on statistical reasoning with observed data rather than subjective judgement. Taken together, the expected outcomes of this work will provide advanced ratemaking methods with credibility that enable the insurance companies to reflect observed and unobserved heterogeneity of the insured appropriately. More broadly, the ratemaking methods with credibility could be applied to a wide range of fields in natural sciences and engineering in which a study incorporates longitudinal data, such as prediction of annual agricultural production based on the historical productions and characteristics of the land.