A “pay-for-risk” model means that the goal of the business is to price risk at a premium so your revenue covers the risk and leaves you with an acceptable profit. If you believe that the business of banking is a pay-for-risk model, then we have some bad news for the current state of banking. The mature economic expansion is not a good time for bankers who expect to be rationally compensated for the risk they take. In this article, we identify two specific areas (credit and interest rate risk) where banks are often not earning enough return for the risk they are currently taking.
By using a simple variation of “analysis of covariance,” we show that in absolute levels and by historical averages that banks are currently not being adequately compensated for credit or interest rate risk. Analysis of covariance evaluates whether the average of a dependent variable (in our case, revenue) is equal across levels of a dependent variable (in our case, credit or interest rate risk).
The easiest way of conducting and explaining the analysis of covariance is to compare various levels of interest rate and credit risk and consider if revenue varies with different levels of the specific risk.
If banks are appropriately compensated for credit risk, then lower credit quality loans would command wider spreads. To analyze this covariance pattern, we need to compare spreads on loans with various credit quality and identify both the absolute level of credit premium and the trend over time. Unfortunately, there are very few sources that uniformly compare average industry credit spreads for different credit quality loans. Banks use different ratings and measuring across industry is problematic. What one bank considers an “A”-rated credit is a “B” or “C” credit to another bank.
Fortunately, S&P Global Market Intelligence tracks bank loans across the country and assigns each loan a risk ranking on a 16-point risk scale. Using this data we can measure the gap between A-rated loans (high-quality commercial grade) and D loans (barely acceptable commercial credit). All loans are currently originated in the market. This information is presented in the graph below.
The average spread for A credit is currently 2.04% over LIBOR and the average credit spread for D credit is 2.68% over LIBOR. The gap in spread between A and D loans across the banking industry is only 64 bps through 2018, versus 129 bps in April 2016 (when the data becomes available). This gap in pricing is low by historical standards and continues to decline. We conclude that banks are not properly compensated to take credit risk at the low end of the spectrum. There may well be sufficient return to justify an A credit at 2.04% spread, but our loan modeling shows that D credit at 2.68% spread is a poor ROE if priced through a credit cycle. The average A credit at 2.04% spread results in a 16.5% ROE loan (depending on loan size, additional cross-sell opportunities and loan category). However, the average D credit results in only a 3.9% ROE at 2.68% spread. Now is an inappropriate time for banks to deploy a pay-for-risk business model in banking.
While the D grade loan has an immediate higher yield, it is important to price loans through a full credit cycle. Because of the age of the current expansion, and the average commitment term for loans, it is highly likely that every loan booked today will need to endure a recession. While a D credit in today’s strong economic cycle at a 2.68% spread seems like a good return on capital, any credit stress on the loan will result in substandard ROE.
Interest Rate Risk
The same analysis of covariance for absolute levels and for historical trends can be applied to interest rate risk. We can consider if banks are currently adequately compensated for interest rate risk. We can analyze various points on the yield curve to discern if revenue is driven by interest rate risk. The most common two points to analyze on the term structure of interest rates is the two-year and ten-year Treasury yield. The graph below shows the spread premium between these two rates for the last ten years.
The spread premium between the two and ten year Treasury yield dipped to 38 bps and continues to decline with every Federal Reserve interest rate increase. The spread is by far the lowest it has been in the last ten years. Most community banks will commit to two years of interest rate risk but should banks also extend fixed-rate loans to five or ten years. It is puzzling that the difference in yield is so small. Ten-year duration is almost two business cycles with any number of possible interest rate outcomes over that 120-month term. Another way of considering the interest rate risk is that the differential between the two and ten-year part of the curve is separated by only two Fed rate hikes and five rate hikes are planned in the next 18 months alone. Extending duration to five or ten appears to be a poor bet for banks in this unusual interest rate environment.
It appears to us that banks are currently not appropriately compensated for either credit or interest rate risk. If your bank subscribes to pay-for-risk banking model, now is a time to decrease that risk because the market is mispricing it. So what are banks to do? The alternative to pay-for-risk banking model is relationship management banking which involves compensation for services and advice. We look to cover this topic next week so stay tuned.
Submitted by Chris Nichols on June 18, 2018