The Innerworkings of Whole Loan Valuation

Whole loan valuation involves more than running a model. Data quality, assumptions, market conditions, and loan-level analysis all shape what a portfolio is actually worth.

What does it really take to produce a valuation that holds up under scrutiny?

In Episode 4 of The AiCR Exchange, Joe Furlong sits down with Dean Hurley, who leads MIAC’s Whole Loan Valuation Group. Dean walks through MIAC’s structured valuation process from start to finish and explains why the math alone does not produce a market price. 

How does whole loan valuation work? 

Whole loan valuation at MIAC follows a structured, repeatable process at two levels. At the highest level, every valuation follows the same sequence: receive the data tape, load it into ETL software for data review and normalization, send source and strat reports and data exception reports to the client, confirm the data is accurate, load into cash flow generating software, run the valuation, and send it through supervisory review before it goes back to the client. No shortcuts are taken regardless of timeline pressure. That process has been refined over decades. 

Why does data quality matter so much in whole loan valuation? 

Data quality shapes everything that follows in the valuation process. If the data going in is wrong, the output will be wrong regardless of how sophisticated the model is. Wholesale lenders and investment banks often receive messy, inconsistent data from multiple originators in different formats. MIAC spends significant time on data scrubbing, normalization, and error checking before any assumptions are applied or cash flows are run. When data is missing or appears erroneous, MIAC communicates that to the client directly, explains what assumptions would need to be made, and confirms with the client before proceeding.  

How does MIAC set assumptions for whole loan valuation? 

Once the data is clean and confirmed, the analyst loads the project into MIAC’s cash flow forecasting software, Vision, with the necessary assumptions. Vision takes inputs including voluntary prepayment speeds, defaults, severities, recovery timing, and yields to produce cash flows for each individual loan. MIAC’s proprietary Conditional Response Model, known as CORE, uses logistic regression techniques built on large borrower analytics databases to estimate prepayment and default assumptions with accuracy, particularly for residential loans. For missing or uncertain data, MIAC communicates the assumptions being made to the client and confirms before the valuation proceeds. 

How are spreads determined in whole loan valuation? 

Spread determination involves multiple data sources working together. MIAC’s system carries daily interest rates for swaps, treasuries, and TBAs. Average life is calculated for each loan, and the appropriate benchmark rate is interpolated from that. Spread assumptions are then drawn from rate sheets on new originations, third-party commercial surveys from sources like Cushman and Wakefield, and direct access to pricing data from insurance company-owned hedge funds and other institutional clients that MIAC values product for on an ongoing basis. That combination of sources creates triangulation, and MIAC compares them to determine the spread for each product type. Return on equity based models provide an additional check. 

Is whole loan valuation math or judgment? 

Both. The models produce an output based on assumptions, but market price is not always what the model says it should be. Demand matters. When investor appetite for a product type is strong, prices can trade significantly above what a pure mathematical model would indicate. Dean describes learning this early in his career when portfolios MIAC modeled were consistently losing bids. The issue was not the math. It was that the market assumptions did not reflect what buyers were actually willing to pay. MIAC’s approach is to start with objective model outputs and move toward market assumptions when client data or market activity signals that pricing has moved. A defensible valuation reflects both the math and the market. 

What is stress testing and how does it differ from standard valuation? 

Stress testing runs macro factor scenarios through the valuation models rather than producing a single market price. MIAC’s models accept inputs including unemployment, HPI, inflation, and GDP and produce credit and yield outputs under multiple scenarios. For DFAST purposes, MIAC uses the official scenarios from the OCC and the Fed. For CECL, current expected credit loss, MIAC runs macro factor scenarios to produce credit loss projections across the portfolio under base, adverse, and severely adverse conditions. The output is not a price but a range of credit loss estimates showing how the portfolio behaves across different economic environments. 

Where is the whole loan market headed? 

Dean Hurley’s outlook as of early 2026: the market is in a period of positive but slowing growth. GDP growth in the 3% range for 2026 is likely, declining to 1.8 to 2% in 2027. The market is pricing in two rate decreases, one in summer and one in fall of 2026, which should translate to gradually declining mortgage rates assuming the yield curve does not steepen further. That environment points to increased residential mortgage origination through 2026 and 2027. Demand for whole loan product continues to exceed supply, keeping prices elevated. Black swan events aside, the two-year outlook is constructive for originators and sellers of residential mortgage product. 

Frequently Asked Questions About Whole Loan Valuation 

What is whole loan valuation? 

Whole loan valuation is the process of determining the market value of individual mortgage loans or loan portfolios. It involves data normalization, assumption setting, loan-level cash flow modeling, spread determination, and supervisory review. The output is a price or range of prices that reflects both the mathematical model and current market conditions. 

What is the difference between whole loan valuation and stress testing? 

Standard whole loan valuation produces a market price based on current assumptions and conditions. Stress testing runs macro economic scenarios through the same models to produce credit loss projections or price changes under adverse conditions. Stress testing is used for DFAST, CECL, and regulatory capital planning purposes rather than transaction pricing. 

What data is needed for whole loan valuation? 

Whole loan valuation requires a data tape containing loan-level information including borrower characteristics, loan terms, and performance data. The quality and completeness of that data directly affects the accuracy of the valuation. Missing or inconsistent data requires assumption setting and client confirmation before the valuation can proceed. 

What is MIAC’s Whole Loan Valuation Group? 

MIAC’s Whole Loan Valuation Group handles residential, commercial, and consumer whole loan valuations as well as structured products and risk retention analysis. The group uses MIAC’s proprietary Vision™ cash flow software and CORE prepayment and default model to produce loan-level valuations for lenders, investment banks, insurance companies, hedge funds, and other institutional clients.