Where Resilience Gives Way to Concentration

A regional bank can look relatively steady in a broad macroeconomic stress test and still carry a serious exposure to one property market or funding path. The subject here is the point at which regional-bank resilience gives way to concentration. The practical task is to separate four questions—how much is exposed, how sensitive it is, when pressure arrives, and whether pressure becomes a realized loss—before treating a calm aggregate statistic as reassurance.

CRE stress is a timing-and-segment problem

The Federal Reserve’s regional-bank comparison gives the first warning signal: commercial real estate made up 41.64 percent of regional banks’ total loans in 2019Q4, versus 16.75 percent for large banks. In the 2023Q4 appendix, the corresponding shares were 44.24 percent and 13.24 percent. The difference is large enough that the same property-market downturn can matter more to a regional bank’s overall balance sheet even if the bank’s individual loans are well underwritten.

But a portfolio share is not a loss forecast. It answers how much of the bank is connected to CRE, not whether the properties are occupied, whether rents cover debt service, when the loans mature, or how much collateral value would be available in a workout. A useful diagnosis keeps four measures separate. Exposure is the share or dollar amount tied to the sector. Sensitivity is the estimated response of a loss rate to a variable such as unemployment or CRE prices. Refinancing timing determines when a borrower must replace a low-rate loan at current terms. Realized loss occurs only when the deterioration is recognized through a charge-off or another loss event.

Those measures can move in different directions. A bank may have high CRE exposure but low current charge-offs because borrowers are still current and maturities are several years away. A loan may have a weaker collateral position but no realized loss while the borrower can continue paying. Conversely, a smaller portfolio can create a large problem if its loans mature together, its properties are concentrated in one weak market, or its borrowers cannot refinance at prevailing rates. The diagnostic question is not simply whether CRE is large. It is which segment, what contract, what location, and what date connect the exposure to cash flow and capital.

The maturity date is especially important because refinancing changes the borrower’s payment burden even if the principal balance has not changed. A loan made when interest rates were low may have been serviceable with one rent and vacancy profile. At maturity, the borrower may face a higher rate, a lower property valuation, tighter underwriting, or all three. The bank must then manage the exposure under changed terms, and any restructuring or other workout can change the timing and size of the reported effect. A current loan is therefore evidence about present payment performance, not evidence that the refinancing risk is absent.

The Federal Reserve’s regression evidence adds a second layer. Regional banks are generally less sensitive than large banks to many aggregate variables, but construction and multifamily CRE are notable exceptions. Regional-bank net charge-off rates in those segments are more sensitive to aggregate CRE prices or unemployment. The combination of a larger CRE share and higher sensitivity in selected segments is more important than either fact alone. It says that a regional bank can be less reactive to the average macro variable while being more vulnerable to deterioration in a particular part of the property market.

Other CRE categories need not follow the same pattern. Office, multifamily, warehouse, hotel, retail, and construction properties have different demand, supply, income, and financing conditions. A single “CRE risk” number hides those differences. The Federal Reserve’s result is also a group-level comparison; it does not provide a bank-by-bank map of where each property sits or how each borrower will respond. The evidence supports closer segment and location analysis, not a claim that every regional CRE book is impaired.

The OCC’s Spring 2026 risk perspective supplies current market context.