Understanding the Propensity to Refinance Model
HouseCanary’s Propensity to Refinance forecasts the likelihood that an existing mortgage will be refinanced within a chosen time horizon (most commonly, the next 3 months). It leverages all the real estate and mortgage data that HouseCanary is known for and is refreshed on a monthly basis to remain current with dynamic markets.
What is Propensity to Refinance?
Propensity to Refinance is a predictive percentile from 0% to 100% that estimates how likely a property is to be refinanced in the near future. Higher scores indicate a greater chance of a refinancing event.
This score is used across HouseCanary platforms to help clients:
Identify properties likely to be refinanced soon
Prioritize outreach and marketing efforts
Set up automated alerts based on custom thresholds
How the Model Works
Our model uses machine learning-based survival analysis trained on decades of real estate data. Some inputs include:
Property and mortgage characteristics
Neighborhood-level market activity and pricing trends
Historical transaction patterns
Economic and behavioral indicators
Debt pressure within the neighborhood
The model is refreshed monthly.
How Accurate Is It?
Internal testing shows the top 1% of properties are 2x more likely to refinance compared to a random group of properties.
What is considered a good probability?
A Propensity to Refinance percentile of 99% is considered “Very High”, 98% is “High”, and 95% is “Medium”. Any score below 95% is considered to have a “Low” probability of being refinanced within 3 months.
Where to See It
You can find Propensity to Refinance in our Portfolio Monitor product, where you configure alerts to monitor changes that exceed your defined thresholds. We expect to introduce it to other products, including CanaryAI, soon.