We are having our annual banking conference this Monday, Tuesday and Wednesday. As such, the article in this section of the publication revisits an educational piece that we ran a few years ago. We will return again on Thursday and Friday of this week with fresh material, but in the meantime, please enjoy this blast from the past as you enjoy your morning coffee.
Since 1996, banks have been required to more quantitatively identify and measure balance sheet risk. As a result, banks have raised asset-liability modeling to an art form; fitting into the gap between expressionism and neo-classicism.
By basing their models on a phalanx of assumptions, forging complex links and correlations, and potentially inputting bad data, bankers have opened themselves up to a heavy dose of "model risk." Rapid shifts in interest rates can highlight "faulty" model assumptions and shortcomings.
While most examiners can tell you their own set of horror stories about a given bank's model, suffice it to say that it is not uncommon for a bank to think they are "interest rate neutral," when they are really asset sensitive.
To eliminate this issue, bankers should continually test and validate asset liability models. In addition, since the shortcomings of most asset liability models can be found in the underlying assumptions, those assumptions should also be tested on an ongoing basis. In this area, most of the problems lie in duration assumption, with regard to fixed rate commercial loans and deposits. The key here is to have a model that is robust enough to validate the underlying assumptions.
Finally, bankers should examine correlations of prepayments, deposit decay, run-off and deposit inflows. Categories for both loans and deposits should be broken down to a granular level until statistical differences can be eliminated. Further, these categories should be tested at least annually.
A bank is both blessed and vexed with cheap computing power and a powerful array of data to make this work take less than 40 labor hours per year, if organized correctly.
The resource investment will not only serve to reduce model risk, but will give everyone on the ALCO committee a better understanding of the inner workings of the bank.