Why the Same Shock Hits Banks Differently

The 2026 stress test applies one severely adverse scenario to a group of large banks, but it does not produce one uniform bank outcome. This chapter explains why: the scenario meets different loan portfolios, business lines, securities positions, and counterparty exposures, then passes through different model channels. Understanding those interfaces lets a reader interpret variation as evidence about exposure and mechanism—not as unexplained noise or a permanent ranking of which bank is safest.

The contrast is visible in the Federal Reserve’s results. Under the same hypothetical conditions, projected minimum CET1 ratios ranged from 6.7 percent for First Citizens to 32.2 percent for Charles Schwab Corp. The spread is large, but its meaning is conditional. It belongs to this scenario, this horizon, this set of starting exposures, and this model system. It does not establish that one bank will always outperform the other under every future shock.

The same scenario meets different balance sheets

A macroeconomic variable becomes a bank loss only through an exposure. A fall in house prices matters through mortgages, home-equity lending, collateral, and other related positions. A decline in commercial-real-estate prices matters through the banks that lend against or invest in that market. Wider corporate spreads can affect banks with commercial lending, securities, or trading exposure in different ways. The scenario supplies the pressure; the balance sheet supplies the channels through which pressure arrives.

The Federal Reserve’s aggregate loan-loss rates show why loan type is not a minor detail. Domestic first-lien mortgages had an aggregate projected loss rate of 1.5 percent, while credit cards had a rate of 17.1 percent. Commercial and industrial loans had a 9.0 percent rate, and domestic commercial real estate an 8.8 percent rate. These figures do not say that a credit-card loan is always riskier than a mortgage, or that a commercial-real-estate portfolio will always perform worse. They describe how those categories responded within the 2026 severely adverse scenario and its modeled relationships.

The differences are about more than the label on a loan. Credit cards are exposed to household income and employment conditions, but they also have a different pricing and revenue structure from a first-lien mortgage. Commercial and industrial loans depend on the financial condition of firms and the sectors in which they operate. Commercial real estate depends on property values, cash flows, financing conditions, and the characteristics of the underlying properties. A common macro path therefore enters portfolios with different sensitivities, maturities, collateral arrangements, and sources of repayment.

Business mix changes the income side as well. The Federal Reserve reports that pre-provision net revenue tends to be higher relative to assets at banks focused on credit-card lending because credit cards generally produce higher net interest income relative to other forms of lending. But those banks also tend to have relatively high loss rates. A higher revenue measure can therefore coexist with a higher exposure-driven loss measure. The first does not cancel the second automatically.

This is the key interface between a bank’s business model and the stress-test result. A bank is not simply “more profitable” or “more exposed” in the abstract. Its revenue sources may provide a larger cushion under one path, while its asset mix creates larger losses under the same path.