September 30, 2023
Staff research
-
-
Predicting Changes in Canadian Housing Markets with Machine Learning
We apply two machine learning algorithms to forecast monthly growth of house prices and existing homes sales in Canada. Although the algorithms can sometimes outperform a linear model, the improvement in forecast accuracy is not always statistically significant. -
Anonymous Credentials: Secret-Free and Quantum-Safe
An anonymous credential mechanism is a set of protocols that allows users to obtain credentials from an organization and demonstrate ownership of these credentials without compromising users’ privacy. In this work, we construct the first secret-free and quantum-safe credential mechanism. -
Should Banks Be Worried About Dividend Restrictions?
A regulator would want to restrict dividends to force banks to rebuild capital during a crisis. But such a policy is not time-consistent. A time-consistent policy would let banks gradually rebuild capital and pay dividends even when their equity remains below pre-crisis levels. -
Tattle-tails: Gauging downside risks using option prices
Options markets offer unique insights into the changing risks different assets face, which helps us better understand the broader risks to the Canadian economy. We show how option prices help reveal that investors did not anticipate large downside risks to either major Canadian banks or economic growth during the March 2023 financial sector system stress, a period when policy-makers and investors were unsure of what the future held for Canada’s economy. -
Digitalization: Definition and Measurement
This paper provides an overview of digitalization and its economic implications. We assess the scope of digitalization in Canada as well as the challenges related to its measurement. -
Forecasting Risks to the Canadian Economic Outlook at a Daily Frequency
This paper quantifies tail risks in the outlooks for Canadian inflation and real GDP growth by estimating their conditional distributions at a daily frequency. We show that the tail risk probabilities derived from the conditional distributions accurately reflect realized outcomes during the sample period from 2002 to 2022.