At least once a week we get a question about core deposit decay assumptions.  It’ll typically be worded something like this:

How do the numbers we put in for the decay assumptions effect the IRR modeling?

It a good question; core deposit modeling (specifically decay) is a pretty hot topic these days because it can have a big impact on your projected long-term interest rate risk.

The decay rates essentially are an assumption about the average life of your non-maturity deposits.  They will have the most impact on your bank's EVE measurement.  The longer we model these deposits to be, the more base EVE for the bank.  We calculate EVE be taking the present (or economic) value of all assets less the present (or economic) value of all liabilities, EVE = EVA - EVL.  Calculating the value of all assets and liabilities is a reasonably straightforward exercise except when it comes to core deposits.  We have a beginning balance and a rate, but we're missing the term structure (i.e. they're "non-maturity" deposits).  The decay assumptions you provide give us an assumed term structure.

I know this is a short answer to a very involved question.  I often refer to the subject of core deposit valuation as the classic “black-hole” topic.  I’ve been three day conferences where this is all that was talked about (I know I need to get a life!).  There are a few firms out there that specialize (read $$$) in measuring core deposit value.  One is McGuire Performance Solutions, check out their site for more information.  I’ve seen Bill McGuire speak on the subject of core deposit valuation, it can be interesting stuff.  He recently did a Q&A for the Office of Thrift Supervision in their Quarterly Review of Interest Rate Risk, 3rd Quarter 2006.