SHARPENING THE PENCIL ON MARGINAL COF

Just about every community banker we know is struggling with their cost of funds right now. Perhaps that is why so much talk is constantly swirling over "incremental" or "marginal" (hereinafter referred to as marginal) cost of funds and how best to calculate the number. To begin, marginal cost measurement attempts to capture the cost of each additional dollar borrowed to fund the next loan (or other asset). It assumes the cost of funds remains unchanged. The theory behind this in the banking industry is that the marginal cost of acquiring new funds decreases the larger the bank becomes (i.e. large banks have more flexibility to tap into the capital markets than smaller banks). All of that is generically true, which leads us to how banks often calculate their marginal cost of funds. For many banks, the solution lies in taking a rate sheet from the nearest FHLB and figuring that is as good a proxy as any. It is readily available, easily referred to and offers a maturity spectrum to cover almost any loan structure. The problem in using the FHLB rate sheet to calculate marginal cost of funds is that it is very wide the mark. Perhaps the biggest problem with this approach is that it simply ignores the fact that less than 5% of the typical bank's funding structure comes from FHLB advances. By adding in brokered deposits, the percentage doubles to 10%, but that is still a long way from being close to approximating overall funding costs. As the data shows, the vast majority of funding for most community banks isn't wholesale, but rather comes from the bank's own branch network. So if the branch is the primary contributor to funding, can we devise a way to approximate marginal cost of funds that isn't so skewed away from the core business? The answer is yes, but we will need more information, so let's introduce a few more data points. For the typical community bank, roughly 15% of bank funding is non-interest bearing; the average branch costs $1mm to open; the average community bank has 7 branches; the average branch cost includes overhead, FDIC expenses, phones, ATM machine, etc. Assuming the average branch costs $600k to $800k to operate each year and that the breakeven per branch is about $20mm in deposits, we can run some further analysis. Once calculated, we then go back 6 quarters or so into the FDIC data (banks with assets $1B or less) and break them into quartiles to prepare for back testing. By layering in the average Federal Funds rate over the quarters in question, we then have the basis for a market comparison. We used Fed Funds because it is easy for bankers to capture and monitor. After running all of these components through the modeling process and crunching the data, we find that taking the cost of funds from the call report and adding in roughly 26bp more produces a good "rule of thumb" bankers can use to approximate marginal cost of funds. So, how does your bank compare? The best performing 25% of banks have consistently produced a marginal cost of funds about 51% of Fed Funds. This compares to the next 3 quartiles of 65% of Fed Funds, 74% and 87%. So, the next time someone asks you what the marginal cost of funds is for your bank, just add 26bp to your most recently reported cost of funds and call it a day. While "rules of thumb" should be fully understood within a bank's own context, missing out on originating a loan because assumed funding costs are grossly overstated is probably much worse. So, the next time someone in your bank says marginal cost of funds is the same as the FHLB advance; rate pull out a pencil, check the call report cost of funds level and add 26bp to it.