Exam Prep- Developing Key Assumptions for IRR Analysis


By Clay Wells


In the latest edition of the FDIC’s Supervisory Insights (Vol. 11, Issue 2), the article “What to Expect During an Interest Rate Risk Review” discussed the importance of having accurate and vetted Asset/Liability model  assumptions. It stated this topic would be one of the focal points of upcoming examines, and our clients have confirmed this assertion.
The same Supervisory Insights published for Winter 2014, fittingly contained an article titled “Developing the Key Assumptions for Analysis of interest Rate Risk”. One line describes what we here at AMG have been telling our clients for years:
“…use of unsupported or stale assumptions is one of the most common IRR issues identified by FDIC examiners”
Model assumptions are the most important factor relating to accuracy and meaningful IRR analysis in any Asset/Liability report. In the past, some institutions relied on third party assumptions that were derived from industry standards, and therefore had no significant connection to the actual behavior of an individual institution.  In the last few years, it has become necessary for banks to incorporate 
assumptions into their A/L model based off their own historical behavior in 3 assumption areas:
     1)      Asset Prepayment
     2)      Non-maturity deposits (deposit retention rate and deposit betas)
     3)      Driver Rate (at AMG we use current offering rate in our BancPath model)
At AMG, Asset Prepayment is determined using Mortgage Backed Securities that are similar in term and rate to each individual loan in the bank’s portfolio. A factor is then applied to refine the accuracy of the prepayment.
Non maturity deposit assumptions include betas (deposit pricing relative to changes in interest rates) and deposit retention and decay rates. These are determined using historical deposit data either obtained through longevity on our system (we need a minimum number of periods to make it statistically valid) or though past deposit data provided by the customer.
We use the current rate the bank has booked loans and paid on CDs in the prior period as our Offering Rate. Driver Rates have historically been problematic in that they introduce “basis risk” into the modeling process, leaving one more bias to be resolved when analyzing results. Using your actual offering rates eliminate this additional risk and reflects the banks pricing behavior accurately.




In closing, be sure to inquire with your Asset/Liability provider about the method used in creating assumptions, and be sure to provide them with the necessary historical data to make those assumptions robust. This will lead to a more meaningful, accurate Asset/Liability model and give you more confidence next time examiners come knocking.