Net interest margin (NIM) is one of the most over-utilized metrics in banking. As we have pointed out in the past, if you include all the failed banks over the last ten years, the statistic is about 20% predictive of underperformance. Thus, if you manage your bank trying to get the largest NIM possible, you are likely to produce less profit, not more. Of course, all things being equal you want wider NIM loans than not, but all things are rarely equal. Not only does NIM not take into account risk and cost, but we want to point out one more significant factor that NIM does not address that leads banks astray.
Beyond Adverse Selection
Because NIM does not take into account risk or cost, banks that focus on NIM may not only be adding lower risk-adjusted returning loans to their portfolio but doing so at a higher rate than average. Banks that focus on NIM are often adversely selected and utilized by borrowers that have been turned down or have received higher pricing/less favorable terms from other banks. As such, banks that focus on NIM may end up with a riskier portfolio of loans. During 2008 through 2010, it is no surprise that banks that failed had higher than average NIM largely driven by construction and land loans. Yes, the NIM was high, but so was the risk. However, there is another major shortcoming of NIM.
The Holistic Balance Sheet View
Banks are not in the business of managing loans; they are in the business of managing their whole balance sheet. While loans are likely the largest portion of your balance sheet, cash holdings are probably material. The average community bank holds 8.6% of their total balance sheet in cash while they only need to fund 0.7% of their total balance sheet each month to cover net new loan growth and deposit runoff. Now granted banks need to include a buffer to handle cash flow volatility but that is why banks have multiple liquidity lines in place.
Banks that turn down that quality loan at a 2.40% credit spread because they are holding out for a 3.00% credit spread makes little sense if they are holding 8.6% of their balance sheet in cash.
As the old banking saying goes – “You can’t pay your shareholders in net interest margin.” That is to say you would rather have a $3mm loan at Libor plus a 2% risk-adjusted credit spread than a $200,000 loan at a 3% credit spread plus $2.8mm of cash sitting around. The $3mm loan gives you $229,615 of earnings to the bottom line after taking into account risk, while the second scenarios give you only $69,016 of net earnings assuming the current forward curve of Fed Funds. Thus, your shareholders get over three times the earnings making that higher quality, lower margin loan than making that much smaller loan at a wider margin but having the remaining balance sitting around in Fed Funds (or at the Fed).
That $3mm loan is likely to a more creditworthy borrower that will protect your balance sheet in the next downturn and may have a lower expected loss rate than the $200,000 loan. Further, that $3mm loan comes with the probability of deposits and future business where the excess cash sitting at the Federal Reserve does not.
Putting This Into Action
We like looking at a risk-adjusted return on equity (or return on assets) first and then looking at the risk-adjusted net, after-tax profit (#2 on the graphic below) second at the instrument level to help make a decision.
If you are going to use a net interest margin, use a risk-adjusted number and consider using an average margin on your cash balances and securities and manage towards that. This will give a metric that better represents the return on the whole balance sheet.
Banks that try to optimize their risk-adjusted margin will be fine so long as they try to optimize their cash and securities holding in conjunction. At least every month, management should review loan production, estimate the next thirty days cash needs, and then optimize their cash and securities to meet their liquidity, asset-liability and return needs (in that order). By using this framework, banks will find that they can produce a higher performing balance sheet.
Submitted by Chris Nichols on October 30, 2018