“Which banks are prepared for a rate spike?” (American Banker Online, March 4, 2013). I’m always drawn to headlines like this; it refers to interest rate risk after all and that’s why we’re in business.
It’s a good read and I think it’s fair assessment of today’s banking landscape. But the included interactive graph is…well…terrible. I’m a data-guy and I had a hard time deciphering it. The data is also a bit misleading.
The term “Duration Gap” is being used very loosely
First, they aren’t measuring and graphing duration at all. They are measuring the dollars of net short-term assets. That’s a far cry from duration. The whole point of the article is to show that some banks are better prepared than others to handle rising rates. It would be great if they examined real duration gap. A bank with a larger positive gap between total asset duration and total funding duration means it is likely to experience some pressure as rates rise. The article sites three distinct time frames. The first was mid-2004 when the Fed began to tighten. The second was mid-2007 when the Fed started providing emergency liquidity and then dropping rates soon after. And then finally they show data as of 3rd quarter 2012.
We take a different approach; we actually do model duration (and therefore duration gap) for all banks in the country each quarter. Here’s a much cleaner view of what’s happening in the industry. I think it’s pretty easy to see the outlier here. Clearly banks between $500M and $1B in total assets have a much wider duration gap now then they did in the recent past. I think this is a pretty good indication that they are “reaching for yield” (see prior posts here and here.)
“They don’t reprice at all”
The second problem with the analysis in this article is that it ignores the impact of non-interest bearing deposits. While the author admits that these cash flows can be volatile, “[they] do not directly impact duration gap measures since they do not reprice at all.” On the contrary, maybe they don’t impact the “net short-term assets” measure being used, but they definitely impact total funding duration and duration gap.
Yes, these deposits don’t technically reprice, but if (when) rates rise most experts agree that a measurable amount of that money will move. This behavior means that at least a portion of these balances is shorter-term than you might think. “Call it surge deposits, parked money, or whatever you like. It’s not going to reprice - it’s going to be gone,” says Kurt Schneckenburger. Kurt has worked with Olson Research for over 30 years and he’s an expert when it comes to bank financial analysis and A/L modeling. He clarifies, “…in fact this time around it may be gone even faster.” He thinks there are two reasons for this. “The first is the most obvious, we’re all starving for some sort of return, any hint of higher rates looks better than 0.00%.” He says there are also changes in technology which can help folks monitor rates paid on one account versus another (referring to online brokerage accounts, mobile apps, and financial consolidation services like Mint) . “In some cases the dollars may shift out of the account automatically.”
So while these funds don’t reprice, they are likely to become more expensive if depositors change their mix. This effectively shortens total funding duration which means greater risk to rising rates.