Jump to main content

Mortgage Servicing Rights Valuation with an ALM Focus

A White Paper and Case Study on QuantyPhi’s MSR Valuation Model

May 2024
By Adam Stone, President of QuantyPhi

Read the White Paper


Many credit unions have a program in place to fund mortgage loans to their members and sell them, typically to agencies like Fannie Mae and Freddie Mac, who then package the loans together and sell them as securities to investors. Credit unions retain the servicing rights to the mortgage, meaning they collect principal and interest payments on the loans and pass them through, as well as deal with paperwork, collections, and all the hardships that come with servicing a loan. For the troubles of servicing a loan, credit unions get a monthly servicing fee, typically around 25 basis points of the remaining balance of the loan. The servicing fee decreases proportionally to the paydown of the loan as balances decline.

On their quarterly call reports, credit unions must report the dollar amount outstanding of real estate loans that have been sold in which they retain the servicing rights. Credit unions must also report the dollar amount of Mortgage Servicing Rights (MSR) recorded as an asset on their call report under Other Assets. The value of MSRs must be the Net Present Value (NPV) of all expected future income from servicing the loans.

QuantyPhi has built a sophisticated model to calculate the NPV of a credit union’s MSR portfolio, with an emphasis on interest rate risk.

Key highlights and benefits of QuantyPhi’s MSR model include:

  • Each loan is examined individually.
  • Advanced process to determine forecasted prepayment speeds, combining multiple prepayment projections and market consensus on large pools of outstanding mortgages with excellent geographic dispersion. Projections are based on underlying prepayment projections on pools of tens of millions of individual mortgages with a current amount measured in the trillions.
  • Prepay projections are segregated by loan term (10-year, 15-year, 20-year, and 30-year), loan origination year, and loan interest rate to determine the best prepayment rate for each individual loan based on the loan’s specific characteristics.
  • The model uses a dynamic discount rate that combines the loan spread at origination date and the current yield curve to calculate a present value for each individual loan.
  • Results are provided showing a base prepayment, zero prepayment, +300 bps, +200 bps, +100 bps, -100 bps, -200 bps, and -300 bps interest rate scenarios. This provides a good range of expected valuations should rates change one way or another.
  • QuantyPhi provides a detailed final report including our process, results, and recommendation.