Q: Lynn, tell me about yourself?
Lynn: I’ve been in banking since 1983 and in asset/liability management since 1986, serving time as an analyst, senior analyst, ALM manager, ALM consultant, etc. I’ve been a Senior Consultant with Raymond James Asset Liability Management Group for the past 10 years, working from an office in Huntsville, Alabama.
Q: Interest rate risk has been part of the examination process since 1996, yet from 2011 and 2019 when interest rates were at stagnant lows, in some instances interest rate management took a back seat to other banking functions. Yet in 2023 we saw why ALM management matters with Silicon Valley. Have you seen an increased focus on ALM management in the past two years?
Lynn: Yes, obviously there’s been a focus on liquidity following the bank failures, but also more scrutiny on the ALM Report assumptions in general. The assumptions are, for the most part, attempts to predict consumer behavior, whether it be pricing, loan prepayments, betas, or the decay rates or average lives of non-maturity deposits. The failed banks got their assumptions wrong, with devastating consequences.
Q: ALM reports can be confusing, there is a lot of data on those pages. What is one piece of advice you can provide for those trying to dig through those reports?
Lynn: Don’t get paralyzed by the complexity. Start with one scenario and understand what it’s telling you. It’s easiest to just consider the Static Base Case, i.e., your balance sheet holds level (no growth or runoff) and interest rates stay flat at the current rates. Look at the projected income statement – does it look realistic? Then consider the assumptions that went into this. You had to estimate the rates you’d get on new loans, new bonds, and what you’d pay for deposits and possibly borrowings. You had to estimate prepayments on loans and bonds. You had to estimate your net overhead and loan loss allowance. Once you’re comfortable with this “flat rate scenario,” move on to the others.
Q: When we are looking back at the past two years of rapidly rising interest rates, do you feel that those instantaneous rate shock scenarios captured their balance sheets really did. Or was beta set too low and didn’t accurately reflect what happened in the market?
Lynn: From our back-testing, we can say that the overall betas across the dramatic rising rate period were pretty accurate – but you didn’t see a smooth upward slope of deposit rates. If you split the last rising rate cycle in half (really, most any rate cycle), you’ll find that the first half betas are much lower than the last half betas – there’s relatively little change and then there’s a big “catch-up” in the last half of the cycle.
Q: Back when I was an examiner, the interagency standard was to have model outputs up to a positive and negative 400 basis points instantaneous rate shock. However, we saw rate increases that far exceeded that over the past few years...
Lynn: I don’t think the fact that the most extreme scenarios were “only” +/-400 bp meant banks couldn’t see the impact of significantly higher interest rates. We usually see fairly “linear” changes in net interest income as we move from +100 to +200, +300, +400 bp. If your margin narrows as rates rise, the picture painted by modeling through a +400 bp shock should be adequate to tell you whether you need to restructure or hedge.
Q: During the period where rates were skyrocketing and we had the three large bank failures, one thing I would ask bankers is are you losing depositors?
Lynn: Generally, our clients have done a good job of hanging on to deposits. As community banks, they know their customers, are very visible and available, and were very proactive in reassuring them that they were “local” and were practicing a very “old school” brand of banking that was quite different from the failed banks’ business models.
Q: What is the one thing that you feel will make the most impact on directors and officers from understanding assumptions?
Lynn: Assumptions are as important as the raw data in Asset/Liability Management. Although we occasionally see “bad data” – think inaccurate rates or incomplete data on complex products (usually loans) – it’s far more likely that bad forecasts follow from bad assumptions.