The US banking sector has moved from a rate windfall to a rate digestion cycle. In 2022 and early 2023, higher policy rates lifted asset yields faster than deposit costs, expanding net interest margins and producing unusually strong net interest income for money-center and regional banks. That phase is over. The investment debate now is whether margins stabilize before credit costs erode the earnings base.
The setup is more nuanced than a simple bearish call on banks. Industry capital is stronger than in prior credit cycles, large banks have diversified fee engines, and eventual Federal Reserve easing could lower deposit funding costs. But the market is right to demand a higher risk premium from lenders with high commercial real estate exposure, weak noninterest-bearing deposit franchises, or loan books tilted toward unsecured consumer credit. In bank equity valuation, a 50 basis point change in through-cycle return on tangible common equity can justify a material change in price-to-tangible-book multiples.
Net interest margins are past the easy-money phase
US bank net interest margin, or NIM, peaked as the Fed funds rate moved from near zero to 5.25%-5.50%, but the lagged cost of deposits has been catching up. FDIC data showed the industry NIM near 3.17% in the first quarter of 2024, down from the prior-year period as funding costs rose faster than earning-asset yields. The direction matters more than the level: margins are no longer being pulled higher by repricing loans, and many banks are now absorbing the cost of retaining deposits.
JPMorgan Chase remains the cleanest example of scale advantage. Its 2023 net interest income rose to roughly $89 billion, supported by higher rates and the First Republic transaction, but management has repeatedly warned that deposit margins are not a permanent annuity. Bank of America, with a large fixed-rate securities portfolio accumulated during the low-rate period, has been more exposed to asset-yield lag; its NII declined year over year in early 2024 as higher deposit costs collided with slower securities reinvestment. Wells Fargo, still operating under an asset cap, has had less balance sheet flexibility and reported NII pressure as deposit repricing continued.
The key analytical point is that NIM is now a distribution, not an industry average story. Banks with sticky operating deposits, low loan-to-deposit ratios, and strong treasury-management relationships should defend margins. Banks that funded loan growth with higher-cost CDs, brokered deposits, or wholesale borrowings will experience a more abrupt earnings squeeze. Money-market fund assets above $6 trillion have permanently raised the reference rate for depositors, even if the Fed cuts later.
Deposit beta is the earnings variable that still has not fully normalized
Deposit beta measures how much of a Fed rate increase is passed through to depositors. In the first year of tightening, beta looked artificially low because households and businesses were slow to reprice cash balances. That changed after the regional banking stress in March 2023, when uninsured depositors became more rate-sensitive and more aware of counterparty risk. The result has been a steady migration from noninterest-bearing deposits into interest-bearing accounts, certificates of deposit, and money funds.
For equity investors, the question is not whether deposit costs fall when the Fed cuts. They probably will. The question is how quickly funding costs decline relative to asset yields. Variable-rate commercial loans reprice down almost immediately. Securities yields reprice only as portfolios mature. Consumer loan yields are sticky, but new origination volumes slow when credit standards tighten. Banks with a high share of floating-rate commercial and industrial loans may actually see initial NII pressure in the first phase of easing unless they can cut deposit rates quickly.
This is why management commentary on noninterest-bearing deposit mix is more important than headline loan growth. A bank that grows loans 4% but funds them with 4.75% CDs may be destroying economic value if credit risk is rising. Conversely, a bank with flat loans, stable operating deposits, and improving securities reinvestment yields can produce better risk-adjusted earnings. In the current cycle, balance-sheet quality is more valuable than balance-sheet growth.
Credit quality is normalizing, but the tail risk is concentrated
Credit losses are rising from unusually low post-pandemic levels, not yet from a recessionary base. Industry net charge-offs remain far below global financial crisis levels, but the direction has deteriorated. Consumer credit has been the first pressure point. Credit card delinquencies and charge-offs at major issuers moved back above 2019 levels for some borrower cohorts, particularly lower-income consumers who exhausted excess savings and faced cumulative inflation in rent, food, insurance, and auto payments.
The labor market remains the swing factor. As long as unemployment stays contained, credit card and auto losses can be managed through pricing, tighter underwriting, and higher reserves. If unemployment rises meaningfully, unsecured consumer portfolios can reprice from normalizing losses to cyclical losses quickly. That is why investors should separate diversified banks with prime card exposure from lenders more dependent on subprime auto, private credit-style specialty finance, or high-yield consumer installment loans.
Commercial real estate is the more asymmetric risk. The issue is not all CRE; industrial, logistics, data center, and high-quality multifamily collateral still have demand support. The stress is concentrated in office buildings, older urban properties, and assets financed at cap rates that no longer work with today’s refinancing costs. The Mortgage Bankers Association estimated that roughly $929 billion of commercial and multifamily mortgages were scheduled to mature in 2024, creating a refinancing wall at a time when office vacancy rates in major markets remain elevated and property values have reset sharply.
For banks, the highest-risk profile is a regional or community lender with CRE loans above 300% of risk-based capital, a large office component, and limited ability to absorb nonperforming asset formation through pre-provision earnings. The market has already penalized many of these banks, but the accounting recognition cycle can lag the economic cycle. Extensions, modifications, and appraisal resets can push losses into 2025 and beyond.
Capital is stronger, but earnings power is less forgiving
The strongest argument for owning bank equities is that the system entered this phase with better capital and liquidity than prior cycles. The largest US banks operate with common equity tier 1 ratios well above pre-2008 norms, and annual stress tests have forced management teams to model severe unemployment, market shocks, and real estate declines. Large banks also benefit from diversified revenue: investment banking, trading, wealth management, payments, and custody can offset some NII compression.
Regional banks have a narrower margin of safety. They are more dependent on spread income, less diversified by fee revenue, and more exposed to local CRE markets. Unrealized losses in available-for-sale and held-to-maturity securities portfolios have declined from peak stress as rates stabilized, but they still constrain flexibility for some institutions. Selling low-yield securities to improve liquidity can crystallize capital losses, while holding them suppresses asset yields.
Reserve coverage is another differentiator. Banks that built allowances early can let credit costs rise without a shock to earnings. Banks that under-reserved during the benign loss period may need to increase provision expense just as NIM compresses. This operating leverage works both ways: modest credit deterioration can have a large impact on earnings per share when pre-provision net revenue is already declining.
Valuation: banks are cheap for a reason, but not all discounts are equal
Bank valuation should be anchored in sustainable return on tangible common equity, not in near-term earnings per share alone. A simple residual income framework is useful: price-to-tangible book value should approximate the spread between ROTE and cost of equity, adjusted for growth. A bank earning a 12% ROTE with a 10% cost of equity and 3% long-term growth can justify roughly 1.3 times tangible book. If credit costs and funding pressure pull ROTE to 8%, the fair multiple can fall below 0.8 times tangible book even without a capital crisis.
This framework explains the market split. JPMorgan can trade at a premium to tangible book because it has superior deposit scale, fee diversity, and a lower perceived cost of equity. High-quality regionals with stable core deposits and manageable CRE exposure can be attractive below tangible book if normalized ROTE remains near 10%. But banks trading at 0.6 times tangible book are not automatically bargains; the discount may be signaling future credit marks, dividend risk, or structurally lower profitability.
From a sector rotation perspective, banks become more interesting when three conditions align: the yield curve steepens through lower front-end rates, deposit costs begin to decline, and credit data stop worsening. A bull case for 2025-style bank performance would involve Fed cuts that reduce funding costs without a recession severe enough to drive large loan losses. A bear case would be a hard landing in which lower rates arrive only because unemployment rises and CRE losses accelerate.
What investors should watch over the next two quarters
The most valuable bank indicators are increasingly granular. Headline EPS can be flattered by reserve releases, securities gains, or expense timing. Investors should focus on balance-sheet signals that indicate whether earnings quality is improving or merely being managed.
- Noninterest-bearing deposits: Stabilization would indicate deposit migration is slowing and funding costs are near a peak.
- Interest-bearing deposit cost: A plateau or decline before asset yields roll over would support NIM stabilization.
- Criticized and classified loans: Rising criticized CRE balances often precede charge-offs by several quarters.
- Net charge-off guidance: Consumer lenders should be judged against pre-pandemic loss rates, not 2021 troughs.
- CRE maturity disclosure: Office exposure by vintage, loan-to-value, and geography matters more than total CRE balances.
- Capital return: Buybacks below tangible book create value only if credit marks are adequately reserved.
The best bank stocks in this phase are not the ones with the fastest loan growth. They are the ones that can defend deposit economics, absorb credit normalization, and still earn above their cost of equity.
Conclusion: the sector is investable, but selectivity is mandatory
The US banking sector is not facing a uniform solvency event; it is facing an earnings-quality test. Net interest margins are adjusting to a world where deposits have a market price again, and credit quality is shifting from pristine to late-cycle normal. That combination will expose weak franchises but can also create attractive entry points in banks with durable funding, conservative underwriting, and capital flexibility.
My base case is a slow margin trough rather than a sharp collapse, with large banks continuing to outperform lower-quality regionals on risk-adjusted returns. Credit costs should rise, particularly in credit cards, auto, and office CRE, but the system has enough capital to absorb a controlled normalization. The equity opportunity is therefore not a broad beta trade on banks. It is a franchise-quality trade: own institutions that can keep ROTE above cost of equity through the cycle, and avoid lenders whose apparent discount to tangible book is really a warning label on future credit losses.