The Company
Dallas, Texas
Bluebonnet Savings Bank, headquartered in Dallas, Texas, is a financial institution with more than $2.5 billion in assets, specializing in mortgage-backed securities.

Situation
Stone Analytics performed a study for Bluebonnet to review and suggest ways to improve the profitability of its purchases of portfolios composed of Mortgage Backed Securities. A primary concern facing Bluebonnet is the fact that interest rate market conditions change radically. Although most mortgages are for 30 years, the changing conditions result in fluctuation in the series of payments, or prepayment "surprises" related to refinancing by the borrowers.

As interest rates decline, homebuyers tend to refinance their mortgages, or - from the point of view of the mortgage holder - to prepay. This creates a risk to holders of mortgage portfolios with higher face interest rates. The above chart indicates the movement of interest rates over the last ten years, the sample period.

Approach
Bluebonnet supplied Stone with information for two different types of mortgage portfolios, fixed and adjustable rate (ARM). These portfolios formed the basis of the development of a series of prepayment models aimed at improving the selection of new portfolios and the off-loading of existing ones. Due to the size of the Bluebonnet portfolio, a one basis point improvement in the prepayment rate on transactions of this size can generate profits between $125K and $150K per month.

Results
Using this information, Stone analysts constructed a series of prepayment models that enable Bluebonnet to better choose which portfolios to buy or sell. In doing so, Stone Analytics surveyed and evaluated currently available market data that could be used to improve predictability of "surprises" and made recommendations that might improve the evaluation of future portfolios.

The graph above characterizes the gains that were made with the prepayment model approach. Using the model to rank order mortgage portfolios by probability of fast versus slow prepayment presented the opportunity to avoid bad portfolios and choose good ones.

Implementation
This analysis was used to improve internal procedures and to expand the reviewed dataset for portfolio selection.