One of the inherent flaws of modeling complex systems (such as bank balance sheets) is the tendency to run too many scenarios to try to account for as many potential outcomes as possible. We create a "can't see the forest for the trees" situation.
With literally thousands of variables, there is a real danger to over-simplifying the environment in trying to reduce it to a clean and simple model that gives us "right" and "wrong" answers to our strategy questions. To combat this, we keep adding scenarios until pretty soon we drown in the depth of the analysis, and we come out less informed than when we started. I have seen banks that run hundreds of different interest rate scenarios so that they can be sure they have modeled every possible scenario and can be properly prepared. But, guess what? They still won't get it right, and will have done nothing but irritate and confuse the bank's decision makers.
We ran a post back in 2012 that discussed the "which rate scenarios to model" question. I stand by that post, as it still covers what we believe are the essential scenarios without going overboard. However, that still leaves 21 separate rate scenarios. When evaluating a bank's exposure, which is most important?
In discussing this question earlier this week, we had an interesting conversation. We came back to a concept that should be familiar to both investors and credit risk managers. We need to evaluate not only the outcome given a scenario, but also the probability of that scenario happening. So, can rates fall by another 300 basis points? Technically, yes, as the 30 year bond still yields above 3.00%. That would mean rates falling across the yield curve to 0%, and the 30 year Treasury yielding about 0.20%. I'm guessing your net interest income looks VERY ugly in this scenario. Does that mean we need to immediately put on a hedge to protect us from a 300 basis point rate drop? No, of course not, because the probability of that happening is extremely low.
So, given that basic approach, which scenarios should be getting extra attention? That will depend on your interest rate conviction, but for what it is worth, here are the scenarios that I personally have been focusing on with clients:
- Flat rates: Even though bank balance sheets have extended out on the curve, the majority of both assets and liabilities still reside on the shorter end. While there will be some fluctuations within a small range, these rates will not move meaningfully until the FPOMC begins hiking the Fed Funds target rate. We are likely still a year away, meaning flat rates are the most likely scenario over the next year. In addition, I think it is always wise to start by evaluating what kind of risk is "baked in" if everything stays the same.
- Rates ramp up by 200 basis points in 1 year: There is no magic to this scenario - it simply matches the basic pace at which the Fed has increased rates in the past. Until told otherwise, I am expecting well telegraphed rate increases of about 0.25% per Fed meeting until the rate is back to "normal" (whatever that may mean now). With 8 FOMC meetings per year, that works out to 2.00% over 12 months. Now, could the pace be different this time? Of course it could, especially given how extreme the Fed's accommodation has been through this cycle. But, this is the best idea we have at the moment. Also, we know that the FED is not starting this increase immediately. However, if we continue to monitor this scenario, we should have an idea of how our balance sheet will react once the move starts.
- Curve flattening: As Grant explained in his post, we have a decent chance of seeing long term rates stay low even if the Fed does hike overnight rates. In fact, this is exactly what we "typically" see during a Fed tightening. I also think this is worth watching because it is precisely where many banks are vulnerable at the moment. Higher short term rates will directly impact funding costs, and a lesser increase out on the curve means that the longer asset base will not benefit to the same degree.
My approach is to spend a little more time on these three scenarios, and then look through the others just make sure there is nothing too ugly in the forecast. And, of course, this list will change as circumstances change. The point is, though, that we at some point need to pick a few outcomes to which we can manage, and continue monitoring all of the others. This allows us to manage the risk while having some scenarios in which if we are right, we can optimize the balance sheet to benefit and generate a better return.