The rate decision is only the beginning
A policy-rate move changes the opportunity cost of money, but a deposit account is also a payment service and a funding relationship. Customers value , convenience, insurance eligibility, access and yield in different proportions. Banks balance the cost of retaining deposits against lending opportunities, liquidity needs and alternative funding. As a result, there is no universal deposit rate that moves in lockstep with a central-bank announcement.
The New York Fed’s April 2023 study used bank-holding-company regulatory reports and documented delayed repricing, substitution toward time deposits and more rate-sensitive borrowing. It examined data through Q4 2022, not the current cycle. Its reported Q4 2022 average effective federal funds rate of 3.7% and interest-bearing deposit rate of 1.4% are historical observations for the study’s population, not today’s rates or the yield available to every saver. [1]
Beta requires a rate, denominator and interval
is the change in a defined deposit rate divided by the change in a defined reference rate over the same interval. A deposit-rate increase of 40 against a 100-basis-point benchmark rise is a 40% beta. A marginal beta concerns a selected interval; a cumulative beta compares with a chosen starting point. A very small benchmark change can make the ratio unstable, and a reversal can make a cycle-to-date ratio misleading.
Total-deposit cost includes noninterest-bearing balances in the denominator; interest-bearing-deposit cost does not. Both normally use period interest expense and average balances with an appropriate annualization convention. A posted , which incorporates compounding and may have eligibility conditions, is a different measure. These distinctions are mathematical, not reasons to choose whichever series gives the more favorable result.
For a hypothetical bank with $1 billion of total deposits, of which $700 million bear interest at a simple annual rate of 3%, annual deposit expense is $21 million. Total-deposit cost is 2.1%, while interest-bearing cost is 3.0%. Both are correct for their definitions. Neither gives the marginal cost of attracting the next dollar.
Migration can raise cost without any price changing
Continue the example: $100 million moves from zero-interest checking into the same 3% savings product. Total deposits stay at $1 billion and no account’s posted rate changes. Interest-bearing balances rise to $800 million and annual expense becomes $24 million. Total-deposit cost rises from 2.1% to 2.4%, or 30 , entirely because of mix. Treating that increase as pure product repricing misdescribes the mechanism.
Runoff is different from internal migration. If $100 million instead leaves the bank and is replaced with borrowing at 4.5%, the replacement costs $4.5 million annually. If the bank shrinks assets rather than replaces funding, foregone asset income and any disposal effects enter the analysis. There is no single runoff cost independent of the balance-sheet response.
Lag changes actual expense, not just the narrative
In a separate rising-rate illustration, assume $1 billion of unchanged interest-bearing deposits and a 100-basis-point benchmark increase at the start of a year. The deposit rate rises 30 for the first six months and 60 basis points for the next six. Incremental first-year expense is $1 billion × (0.003 × 0.5 + 0.006 × 0.5) = $4.5 million. The ending annualized run rate is $6 million. Applying the ending beta to the whole year would overstate the actual first-year increase by $1.5 million.
The same cumulative beta can therefore coexist with different reported earnings depending on timing. Certificate-of-deposit maturity schedules, promotional expirations, negotiated rates and account movement determine when expense changes. A quarterly average can conceal sharp changes late in the period.
Scroll horizontally to see all columns.
| Hypothetical calculation | Result |
|---|---|
| First six months: $1bn × 0.30% × 0.5 | $1.5m |
| Second six months: $1bn × 0.60% × 0.5 | $3.0m |
| First-year incremental expense | $4.5m |
| Year-end annualized incremental expense | $6.0m |
Falling rates need not reverse the movie
Consider a separate hypothetical bank with $1.2 billion of earning assets, including $600 million of floating-rate loans that immediately lose 100 of yield. Annual interest income falls $6 million. Its $800 million of interest-bearing deposits receive a 50-basis-point cut only after three months. First-year expense savings are $800 million × 0.005 × 9/12 = $3 million. First-year net interest income therefore falls $3 million, equivalent to 25 basis points of margin on unchanged $1.2 billion average earning assets.
This simplified sensitivity assumes no balance changes, floors, hedges, fees, credit losses or other repricing. A deposit-rate floor near zero limits further reductions; fixed-term accounts may not reset until maturity. A different asset mix can reverse the outcome. Bank OZK’s Q2 2026 management comments similarly describe the possibility of loan yields falling before deposit costs after rate cuts. That statement is the company’s expectation for its own balance sheet, not an industry forecast. [5]
Dated industry observations, with separate populations
The FDIC’s August 25, 2026 release reports that Q2 2026 industry net interest margin was 3.32%, up 1 from Q1, and domestic deposits rose 0.8% quarter over quarter. The reporting population was 4,238 insured commercial banks and savings institutions. Those aggregates are not a median bank, a neobank-only series, a household savings yield or an estimate of . Margin also reflects asset yields and mix, so its movement cannot be attributed solely to deposits. [2]
For a different measure, the FDIC’s March 16, 2026 national-rate table listed savings at 0.39%, interest checking at 0.07% and 12-month CDs at 1.52%. Its national rates weight available insured-depository-institution and credit-union rates by domestic-deposit shares. They are historical national measures used in the regulatory rate-cap framework, not best available offers, all-account realized costs or an October 2026 quote. A single cross-section cannot establish beta without comparable starting values and benchmark changes. [3]
Competition includes products outside the bank
Savers may compare deposits with Treasury securities and money market mutual funds, while also weighing transaction access, minimums, fees, maturity and protection. A money market deposit account is a bank deposit; a money market mutual fund is an investment product. Similar names do not establish identical , price behavior or insurance treatment. A higher quoted return may compensate for different terms rather than represent a free improvement.
For banks, a high-rate deposit may still be cheaper or more flexible than the next available funding source. Conversely, a cheap-looking channel may require acquisition payments, platform fees or operational capacity. Advertised coupon cost, all-in marginal funding cost and long-run relationship value are three separate measures. Competition can affect both the rate paid to existing customers and the quantity willing to stay.
Margins and liquidity are connected but not interchangeable
Net interest margin is net interest income divided by average earning assets, conventionally annualized. It is not divided by deposits or total assets. A bank can improve margin by shrinking low-yield assets while still facing a funding problem; it can also accept a lower margin in exchange for more reliable . The value of a low-cost deposit franchise depends on balances remaining over time. The New York Fed’s April 2023 valuation analysis discusses why accounting and economic measures can diverge. [4]
An analytical funding bridge separates price, volume, mix and timing, then links them to asset repricing and available liquidity. Holding customer balances fixed while severely underpaying competing alternatives can overstate resilience. Combining maximum repricing and maximum runoff without explaining which customers do each can double-count stress. These are model-consistency issues rather than universal forecasts.
What evidence changes the interpretation
Observed renewal rates, product-to-product transfers, customer concentration, negotiated pricing and wholesale replacement reveal more than a headline beta alone. Comparisons across banks require the same benchmark, interval, deposit denominator and annualization. Acquisitions, unusual quarter-end balances and promotional campaigns can distort apparent trends.
The conclusion is conditional: rate direction matters, but timing and funding composition often explain why two banks or two savers experience the same policy move differently. The numerical examples above are isolated hypothetical scenarios, not additive estimates, observed company results or predictions. The empirical observations retain their original dates and populations.
Sources
- New York Fed, Deposit Betas: Up, Up, and Away?; April 11, 2023, data through Q4 2022Official sourceBack to text: ↑
- FDIC official distribution, Q2 2026 Quarterly Banking Profile release; August 25, 2026SourceBack to text: ↑
- FDIC national deposit rates; March 16, 2026 historical observation retrieved in searchOfficial sourceBack to text: ↑
- New York Fed, How Do Interest Rates and Depositors Impact Measures of Bank Value?; April 10, 2023Official sourceBack to text: ↑
- Bank OZK, Q2 2026 management comments (management interpretation, institution-specific)SourceBack to text: ↑