Desjardins-Rotman : Misrepresentation in ratemaking variables

When building predictive ratemaking models in Property & Casualty insurance, the quality of the models and the adequacy of the premiums we charge to clients are dependent on the quality of the data used when building those models. In certain cases, we know that some variables have the potential of being misrepresented in our data, be it because of omissions from the clients, late reporting of new information on some variables, etc. We would like to use our partnership with the Rotman MMA program to explore ways to identify and quantify this misrepresentation and thus improve the results of our ratemaking models.

Faculty Supervisor:

Dmitry Krass

Student:

Partner:

Desjardins Assurances Générales

Discipline:

Mathematics

Sector:

Finance and Insurance

University:

University of Toronto

Program: