In 2003, iEmergent started creating forecasting models that examined multiple datasets with millions of data points. Through this, iEmergent uncovered patterns and drivers of mortgage market behavior that were unique to communities throughout the U.S. What emerged was a different way of looking at mortgage market data.


Many traditional mortgage industry volume forecasts are calculated using top-down, optimal-utility methodologies. Assumptions are made based on broad supply-side market behaviors. In contrast, iEmergent applies a “demand-driven” approach that captures the changing household-buyer patterns that define housing patterns over time. 

Bottom-up Approach – iEmergent’s forecasts are built on the premise that “one size does not fit all” when it comes to the U.S. housing market. Rather than take a top-down approach that distributes loans and dollar volume to sub-national markets, iEmergent constructs its forecasts from the bottom up. The building blocks of the forecast are the 74,000 census tracts defined by the census; each one is its own market, with different behaviors, patterns, and mortgage loan activity.

The Homebuyer is Key – Instead of using traditional broad-based, supply-side economic indicators to drive the forecasts, iEmergent’s approach focuses on longer-term, home buyer-demand behavior. What that means is that our forecasts analyze how a variety of economic, demographic, and housing-related factors will impact the number of households that are ready, willing, and able to buy a home.  Over the past seven years, this number has shifted dramatically, due to a reduction in credit availability, a struggling economy that brings unemployment, lower wages, and a drop in consumer confidence, and the effect that foreclosures, home values, modification programs, and refinance activity have on if, how, and when households make the transition to homebuyers.

2015 Mortgage Pool

The Driver Behind Mortgage Market Opportunity 

The behavior of mortgage markets is incredibly complex – in fact, there are hundreds of indicators, trends, patterns, and events that impact how and why a market behaves as it does.

iEmergent's PMGRAfter analyzing millions of loan records over decades, iEmergent recognized specific patterns of behavior that captured the complexity of these multiple factors.  What emerged was the Purchase Mortgage Generation Rate (PMGR),or the rate at which an individual market generates purchase mortgages.  The PMGR is the predictive driver that we apply to household pools and adjust to account for changes in legislation, broad-scale economics, and other prominent indicators.

The development of the PMGR allowed iEmergent to provide outcomes forecasts that actually quantify projected loans and dollars for markets, instead of simply listing all of the indicators and information that should be considered when determining what is a market's future mortgage opportunity, and leaving it up to others to make the connection between those factors.


iEmergent dedicates a great deal of time, energy, and effort to the constant improvement of its forecast models, which are measured in terms of performance accuracy and efficacy in all types, sizes, and levels of mortgage markets. 

Accuracy at the national market level is important.  However, it is the granularity of iEmergent's forecasts that is most essential to helping organizations use the intelligence to make the right decisions at the market level.  That is why iEmergent's validation studies focus on accuracy -- at the state, county and even the census tract level.

Through rigorous backward-looking, forward-blind model validations that compare previously forecasted volume projections to actual historical loan origination volumes for a known year, iEmergent tests model after model to determine which one achieves the greatest overall accuracy and efficacy. This concept is explained in iEmergent Forecasting Model Valuation Analysis – 2004-2012.

Upon the release of the 2014 Loan Application Record (LAR) data from HMDA, iEmergent will perform another in-depth valuation analysis for 2010-2014 using iEmergent models 2.0 and 2.1.

The table below identifies iEmergent's weighted mean percent accuracy at various market levels, using the iEmergent 2.0 model (as explained in the valuation analysis).

iEmergent Forecast Accuracy