Research Projects

This project is studying a broad range of parameters that are essential to the estimate of the repair/replacement cost of different parts of the condo infrastructure.

This is an annual project that tracks the trends and thoughts of risk managers across time.

To calibrate the extra mortality resulting from COVID-19, aggregated data is collected monthly from individual and group life insurers in Canada. The data collected from the companies (13 companies contribute individual data and 11 of the companies also provide group data) is focused on the number and amount of all death claims as well as the number and amount of death claims with COVID-19 as cause of death. From the aggregate results, we will be able to determine the excess claims incurred by month starting in March 2020 and compare these to the claims where the cause of death reported was COVID-19. 

The main objective of the current research project is to design a holistic hedging scheme which is sufficiently flexible to apply to various guarantee specifications and underlying mutual fund types. It will incorporate many risks such as equity risk, interest rate term structure risk, policyholder behaviour risk, mortality risk, basis risk, and model risk.

This study will result in the 70th annual report of the intercompany mortality experience for Canadian standard ordinary life insurance policies. It will cover the one-year period between policy anniversaries in 2018 and 2019. This will be the first study to collect data quarterly as well as annually. The quarterly study is focused especially on an early warning of the impact of COVID-19 on life companies. 

This will be the seventh study of T100 lapses and the fourth for UL level cost of insurance. The focus of the study is to assess the impact on lapses rates, particularly at high durations, of structuring policies on lapse support. The study will cover experience during the calendar years 2013–2019.

This project proposes a general framework that describes a bivariate-population mortality system which allows two populations at disparate development stages to be explicitly considered through an underlying process of common mortality trend (CMT). As one significant application, also one of the main contributions of the project, the bivariate-population CMT model is used as the basic learner in a computationally friendly framework to effectively borrow information from multiple populations for the mortality prediction of a target population. The framework is based on an innovative idea of model average. Extensive experiments with the Human Mortality Database (HMD) confirm that the newly proposed CMT-based prediction method yields a significant improvement in the prediction performance compared with a set of benchmark prediction models. 

To calibrate the extra mortality resulting from COVID-19, aggregated data is collected monthly from individual and group life insurers in Canada. The data collected from the companies (13 companies contribute individual data and 11 of the companies also provide group data) is focused on the number and amount of all death claims as well as the number and amount of death claims with COVID-19 as cause of death. From the aggregate results, we will be able to determine the excess claims incurred by month starting in March 2020 and compare these to the claims where the cause of death reported was COVID-19. 

Mortality improvement (MI) assumptions are used by actuaries across all areas of practice. Historically, each practice area has developed its own guidance in Canada. The project aims to develop consistent guidance for MI assumptions that will be adopted by different practice areas. 

The research examines the arguments regarding a basic minimum income, considering the programs launched in response to COVID-19. It will result in a well-reasoned and documented academic paper on this subject that the CIA can use as it contemplates what its public position should be with respect to a basic minimum income plan.

The project investigates the combination of measures of extreme value theory to project the behavior of rainfall through time. The project uses precipitation projections from 2006 to 2017 generated to train a model which is used to project rainfall until 2100. From this, the project studies the expected change in behavior of extreme rainfall. These projections affect insurable risks such as floods and droughts and can be combined with damage curves to project claims. 

This project proposes an innovative distributional forest estimating the whole (conditional) probability distribution of the insurance claim variable, by which the calculation of various risk measures, such as variance and value-at-risk of future claim costs, is straightforward. The proposed distributional forests are made up of flexible, parametric models, which make them not only capable of capturing complicated dependency structures of the data, but also easy to interpret and perform sensitivity analysis with. Insurance data on both the individual and the company level will be used to illustrate application of the proposed approach.

The purpose of this study is to investigate and benchmark how the Canadian P&C insurance industry is utilizing Predictive Analytics, Machine Learning and AI, and examine potential areas for enhancement. 

The aim of the research project is to determine how to diversify flood risk in Canada and to measure if any lack of diversification needs to be accounted for into the premiums and reserves. Another key objective of the project is to determine the impacts of climate change on the viability of flood insurance in Canada. To attain these objectives, the researchers designed a Canadian occurrence-impact flood model over thousands of watersheds where they use a combination of climate and statistical models to model fluvial flood risk.

Proper dialogue around intergenerational equity is severely hindered by three factors: (1) the lack of uniform definitions and language, (2) practitioners’ lack of familiarity with existing metrics for intergenerational equity, and (3) limited availability of modern, forward-looking metrics. The objective of the current project is to catalogue relevant definitions of intergenerational equity, identify existing metrics, and develop new ones suitable for occupational pensions in stochastic applications.

The goal of this research project is to add a quantitative element to the discussion of the merits of plan mergers and consolidation, by providing a framework and an evaluation mechanism under which the welfare impacts of various deals can be assessed. The framework will include: (1) an economic scenario generator that allows for inflation, interest rate and asset dynamics; (2) a model of economies of scale; (3) a model for longevity risk; and (4) a procedure to condense information into simple welfare measures. Use of the framework will be demonstrated with specific hypothetical pension plans. 

Canadians will be surveyed on their awareness of the risks associated with retirement. The focus will be on perception vs. reality when it comes to areas such as longevity, probability of needing long-term care, etc., in order to identify levels of preparedness, awareness, and potential areas of improvement for the public and policy makers.