Machine Learning for Default Prediction of Private Pension Administrator

In this research project, we will partner with the Financial Services Regulatory Authority of Ontario (FSRA) to enhance its default prediction model for private companies administering pension plans in Ontario. Our goal is to enhance the current model’s timeliness in predicting default of private companies by addressing the lack of publicly accessible information from these private entities. We will bridge the existing gap using unsupervised machine learning algorithms to identify look-alike public companies and using natural language processing techniques to extract relevant information from alternative data sources such as websites and social media. The enhanced model will provide FSRA better insights into the plans’ funding adequacy and foresights on any potential solvency issues to better protect the pension plans’ beneficiaries.

Faculty Supervisor:

Stephen Thomas;Ryan Riordan

Student:

Cecilia Ying

Partner:

Financial Services Regulatory Authority of Ontario

Discipline:

Business

Sector:

Finance, insurance and business

University:

Queen's University