Banking · Strategy and competition

The $1.3 Trillion Fight for America’s Idle Cash

AI agents could help customers earn more on spare cash without changing their everyday bank. That could force banks to compete harder for their savings.

Sourced analysis · Research as of November 15, 2026 · Published 11/15/2026

Conceptual tug-of-war: JPMorgan Chase, Bank of America, Citi, Wells Fargo and PNC face Revolut, X, Muse and OpenAI across a box marked Idle Cash.
Conceptual illustration of competition for surplus cash; it does not depict a partnership or confirmed product plans. Banks: JPMorgan Chase, Bank of America, Citi, Wells Fargo and PNC. Challengers: Revolut, X, Muse and OpenAI.

The Miller family in Sacramento, California, keeps $100,000 in cash. They want $25,000 available for bills, emergencies and unexpected spending. The other $75,000 could earn more, but they want to keep the bank they already use.

That leaves a decision worth about $3,000 a year.

A savings account paying 4.40% annual percentage yield would earn $3,300 on that $75,000 if the rate held for a year. At the Federal Deposit Insurance Corp.'s 0.38% national savings benchmark, it would earn $285. The $3,015 difference is before taxes, service fees and transfer costs. These rates were checked November 15; neither is a claim about what the Millers' own bank pays. Example rate: Pibank · FDIC benchmark.

An AI cash agent would offer to do more than point out the difference. With permission, it could move money the family doesn't need immediately, watch upcoming payments and bring cash back before a bill comes due. The family could keep its everyday account while shopping elsewhere for interest.

How much money might be available for that service? Our central estimate is about $1.3 trillion in household cash that could earn a more competitive return. It is a model built from Federal Reserve data and assumptions about how much cash families can spare. It isn't a measured pile of neglected savings, or a forecast that $1.3 trillion will leave banks.

Banks could raise rates on the balances most likely to leave. They could offer their own automation or move customers into their savings and investment products. Existing cash-management tools already do some of this. AI providers would have to show that they can make better decisions, move money reliably and leave customers ahead after fees.

A better rate is only part of the job

The example uses the highest APY among the accounts reviewed: 4.40%, offered by Pibank. That exceeded the FDIC benchmark by 4.02 percentage points. Other accounts have different rates and transfer terms.

That comparison needs care. The FDIC's October 19 release uses September 30 observations and a $2,500 savings-product tier. It doesn't measure the average return on every dollar Americans have in savings. A family already earning 4% has far less to gain than one earning 0.38%.

Savings account examples · November 15, 2026

New eligible U.S. personal applicant, $100,000 held for one year; rates assumed unchanged. These five reviewed accounts illustrate differences in rates and terms; they are not a list of every U.S. offer.

PiBank · Pibank Savings4.40% APY · $4,400/year

Minimum $0; no advertised balance-rate cap identified in reviewed evidence. No monthly maintenance fee. Verified 2026-11-15. Wire transfers; sending/receiving bank may charge. No Bill Pay or Zelle. Not verified for autonomous transfer integration.

Peak Bank · Envision High Yield Savings4.27% APY · $4,270/year

Minimum $100; new accounts only; 3.40% on entire balances at $500,000 and above. No monthly maintenance fee. Verified 2026-11-15. Opening funds held five business days. Not verified for autonomous transfer integration.

Zynlo Bank · ZYNLO Tomorrow Savings Account4.10% APY · $4,100/year

Minimum $0.01; no advertised balance-rate cap identified in reviewed evidence. No monthly maintenance fee. Verified 2026-11-15. Bank holds, transfer limits and settlement apply; autonomous integration not verified.

Bread Savings · Bread Savings High-Yield Savings Account4.05% APY · $4,050/year

Minimum $100; no advertised balance-rate cap identified in reviewed evidence. No monthly maintenance fee. Verified 2026-11-15. Bank holds, transfer limits and settlement apply; autonomous integration not verified.

Live Oak Bank · Live Oak Bank Personal Savings4.00% APY · $4,000/year

Minimum $0; no advertised balance-rate cap identified in reviewed evidence. No monthly maintenance fee. Verified 2026-11-15. Bank holds, transfer limits and settlement apply; autonomous integration not verified.

FDIC national rates: savings 0.38%, interest checking 0.07%, money market deposits 0.63%. Release October 19, based on September 30 data. FDIC source. Insurance limits are per depositor, per insured bank, per ownership category; balances at the same underlying bank aggregate. Rates can change; bank approval and transfer terms apply.

Transfers matter as much as the advertised rate. Pibank's savings page, for example, describes wire transfers and excludes Bill Pay and Zelle. Another bank involved in a wire may charge a fee. A strong rate doesn't establish that an account is suitable for frequent automated transfers.

Peak Bank shows why an agent would need to read the terms, not just rank advertised rates. Its new-account rate of 4.27% falls to 3.40% on the entire balance at $500,000 and above. Even interest earned can push an account across that threshold. Peak also specifies an opening-funds hold. Peak's terms.

The calculator below accounts for reviewed balance tiers. It excludes an Axos promotion because the ongoing rate after the promotional period wasn't verified, and Vio because its rate feed and issuer page disagreed. The accounts checked aren't a large enough sample to establish a national high-yield savings median.

For a customer with $1,000 to move, the dated 4.02-point spread adds only $40.20 a year before costs. At $50,000, it adds $2,010. Fees and transfer delays can make a service unattractive for the first customer even if it is worthwhile for the second.

Exhibit 1

How much more could your cash earn?

At national savings rate

$190

At the selected example account

$2,200

Additional gross interest

$2,010

The selected account is an example for the calculation. Choose another reviewed account below to compare its rate and terms.

Illustrative annual comparison · surplus only

Genuinely surplus balance
$75,000
Current estimated annual interest
$285
Alternative estimated annual interest
$3,300
Additional gross annual interest
$3,015
Annual product, agent and transfer fees
$0
Estimated net annual benefit
$3,015

Only the surplus is compared; retained cash earns the same in both cases. Fees default to zero and may be unknown. Enter charges from both banks and any service. Zero entered fees do not certify fee-free transfers. Larger deposits can exceed insurance limits once accrued interest and other accounts are included.

Eligibility, liquidity and insurance details

No minimum balance is required to open the account or to maintain the stated APY. Variable rate; the interest rate and APY can change at any time without notice. U.S. personal applicant; subject to bank identity and account approval. No direct-deposit or membership condition identified in reviewed rate evidence. No advertised rate cap identified except Peak's tiers. Research comparisons stop at $500,000; this is model coverage, not a claim about the bank's maximum permitted account balance.

Wire transfers; sending/receiving bank may charge. No Bill Pay or Zelle. Not verified for autonomous transfer integration. FDIC; statutory ownership-category limits apply. All five modeled products are bank deposits. Securities and investment funds have different risks and are excluded from this calculator.

Read the bank source
Hypothetical one-year deposit illustrations, before tax. Selected product evidence observed 2026-11-16; FDIC September 30 benchmark released October 19. APY includes compounding. APY-equivalent daily projection reassesses tiers daily; issuer crediting methods can differ. Product fees shown separately, without modeling their interest drag. No credentials, account connection or sensitive information required.

Where the $1.3 trillion comes from

The question isn't how much Americans have in the bank. It is how much they can move without disrupting their finances, and how much of that money is earning too little to justify staying.

The Fed's household accounts put deposits at $14.206 trillion in the second quarter of 2026. That category is broader than ordinary bank accounts: it includes currency, time deposits and short-term investments. The same households held $147.036 trillion in total financial assets. Money-market fund shares, another $5.111 trillion, are counted separately and aren't added to our deposit estimate.

Exhibit 2

Wealthier households hold most deposits

DFA deposit category: $14.21T. Household financial assets: $147.04T. The category includes currency, time deposits and short-term investments.

Bottom 50%$0.78T · 5.5%

67.62 million households · mean deposits $11,498 · 23.7% of group financial assets

50th–90th$5.09T · 35.8%

54.06 million households · mean deposits $94,153 · 14.4% of group financial assets

90th–99th$5.01T · 35.3%

12.17 million households · mean deposits $412,192 · 9.1% of group financial assets

99th–99.9th$1.87T · 13.2%

1.21 million households · mean deposits $1,544,032 · 6.6% of group financial assets

Top 0.1%$1.45T · 10.2%

0.14 million households · mean deposits $10,588,694 · 5.7% of group financial assets

Federal Reserve DFA, Q2 2026; downloaded November 15. Its Deposits category includes checkable deposits and currency, plus time deposits and short-term investments; it is broader than immediately movable bank deposits. Source and definitions. Wealth percentiles, not account-balance tiers. Means are group totals divided by estimated group households; not medians.

The distribution matters. The bottom half of households by wealth held about $777 billion of deposits. Households between the 50th and 90th percentiles held $5.090 trillion; the next 9% held $5.015 trillion. The wealthiest 1% held $3.323 trillion. These are wealth groups, not account-balance brackets. They cannot tell us which families have money available to move.

For that, the 2025 Survey of Consumer Finances offers a closer look. We examined checking, savings and money-market deposit accounts, excluding money-market mutual funds, brokerage call accounts, prepaid balances and CDs.

Using the survey weights and all five versions of the public data, those deposits totaled an estimated $6.966 trillion in 2025 dollars. About 25.3 million families had at least $50,000 across the measured accounts. They held roughly 85.4% of the balances.

Those families include people saving for tax bills or a home purchase. Their accounts also include credit-union deposits. A large balance alone tells us nothing about the rate paid or how soon the money will be needed.

SCF balance tiers and methodology

2025 public extract; 4,360 families, five implicates each. Point estimates average five separately weighted calculations. WGT is already divided by five; recovered full weights are used within each implicate. Dollars are 2025 survey dollars. Median is the average of five weighted medians. No replicate-weight confidence intervals were calculated. Rounding and survey measurement limit precision.

Bank and credit-union transaction/savings deposits

Under $10,000

73.80M families · $0.19T
Mean $2,532 · median $1,502
2.7% of balances · owners 88.6%

$10,000–$25,000

22.10M families · $0.34T
Mean $15,364 · median $14,840
4.9% of balances · owners 100.0%

$25,000–$50,000

13.77M families · $0.49T
Mean $35,344 · median $34,920
7.0% of balances · owners 100.0%

$50,000–$100,000

10.80M families · $0.74T
Mean $68,649 · median $66,662
10.6% of balances · owners 100.0%

$100,000–$250,000

9.05M families · $1.35T
Mean $149,142 · median $142,540
19.4% of balances · owners 100.0%

$250,000–$500,000

3.08M families · $1.06T
Mean $344,376 · median $325,200
15.3% of balances · owners 100.0%

$500,000–$1 million

1.51M families · $1.07T
Mean $710,225 · median $712,338
15.3% of balances · owners 100.0%

$1 million and above

0.86M families · $1.73T
Mean $2,016,533 · median $1,404,500
24.8% of balances · owners 100.0%

Broad liquid transaction accounts, including funds and brokerage cash

Under $10,000

72.23M families · $0.19T
Mean $2,646 · median $1,630
2.3% of balances · owners 97.6%

$10,000–$25,000

22.34M families · $0.34T
Mean $15,426 · median $14,862
4.1% of balances · owners 100.0%

$25,000–$50,000

13.26M families · $0.47T
Mean $35,370 · median $34,752
5.6% of balances · owners 100.0%

$50,000–$100,000

11.12M families · $0.76T
Mean $68,533 · median $65,860
9.1% of balances · owners 100.0%

$100,000–$250,000

9.22M families · $1.39T
Mean $150,459 · median $142,920
16.6% of balances · owners 100.0%

$250,000–$500,000

3.62M families · $1.23T
Mean $340,870 · median $318,242
14.7% of balances · owners 100.0%

$500,000–$1 million

1.89M families · $1.32T
Mean $696,272 · median $696,534
15.8% of balances · owners 100.0%

$1 million and above

1.28M families · $2.66T
Mean $2,077,996 · median $1,393,264
31.8% of balances · owners 100.0%

Deposit measure = CHECKING + SAVING + MMDA; excludes MMMF, CALL, PREPAID and CDs. It is not commercial-bank-only. Liquid measure = LIQ; includes deposit money-market accounts, money-market funds, call accounts and prepaid balances. Under-$10,000 counts include nonowners and zero balances. $1 million boundary belongs in the top tier. Full wealth/income distributions and asset ownership measures are in the reproducible analysis artifact.

Our model applies that 85.4% share to the Fed's broader $14.206 trillion deposit category. That gives about $12.14 trillion for families meeting the balance threshold. The survey and household accounts use different dates and definitions. Combining them this way is an assumption, not an official Fed estimate.

The central case then assumes that 30% of those balances remain available after bills and liquidity needs. Of that amount, it assumes 35% earns an uncompetitive rate and can be moved. That produces $1.275 trillion, rounded to $1.3 trillion in the headline.

Neither assumption can be checked against public account records. The second must also remove currency and other holdings that an agent couldn't transfer. There is no direct measurement of that deduction here. Lowering either assumption reduces the opportunity.

Adoption is another step. If agents influence 30% of that pool within three years, they affect about $382 billion. If 60% of the affected money leaves its original bank, departures total about $229 billion. The rest stays at the bank, perhaps earning a higher rate or moving to another account there.

Across the conservative and accelerated cases, the cash available for competition ranges from $607 billion to $2.185 trillion. The amount leaving its original bank ranges from $36 billion to $819 billion. These are scenarios without assigned probabilities.

At the unchanged 4.02-point spread, annual customer gains before costs would range from $3.7 billion to $43.9 billion once those adoption levels were reached. The central case is $15.4 billion a year, not cumulative income over three years. Halving adoption halves that result. A two-point spread would roughly halve the central gain as well.

Exhibit 3

How much household cash could leave its bank?

The adjustable central case keeps every filter separate. The size threshold identifies families by their combined deposits, not individual accounts; the persistent-surplus filter then reserves operating liquidity.

Household deposit category (DFA)$14.21T

Observed broad category, including currency and time/short-term balances

Held by families meeting the balance threshold$12.14T

85.4% SCF-derived share; cross-dataset calibration assumption

Persistent surplus$3.64T

30.0% assumed to remain after bills and liquidity buffers

Low-yield cash available to move$1.27T

35.0% assumed below competitive yield and transferable

Affected by agent competition$382B

30.0% assumed adoption within three years

Deposits leaving the original bank$229B

60.0% assumed movement outside originating bank

Bars share one linear scale. Small bars are intentionally small; their labeled amounts provide the full values. External movement is loss to the originating bank, not necessarily loss to the banking system.

Underpaid/accessibility also excludes existing competitive yields, restrictions and balances that cannot use the modeled offers. This is a combined assumption; public data cannot measure its components.

Observed baseline: Federal Reserve DFA deposit category (including currency, time deposits and short-term investments), Q2 2026, $14.21T. Balance-size filter: 2025 SCF bank/credit-union checking, savings and deposit money-market accounts; applying its share to DFA is an unvalidated calibration assumption. Every subsequent percentage is an analyst scenario assumption, not measured idle cash or a forecast. Dollar amounts are stocks at a hypothetical three-year horizon; customer benefit is the annual run rate at that horizon.

Larger banking totals answer different questions. Commercial-bank deposits were $19.862 trillion in the Fed's November 13 release, covering the week ending November 4 on a seasonally adjusted basis. The FDIC reported $20.722 trillion at insured institutions for June 30, including about $4.045 trillion in domestic noninterest-bearing deposits. Both include business money; their coverage and dates differ. Neither measures household idle cash. H.8 · FDIC industry profile · FDIC BankFind data.

Businesses have substantial cash of their own. The Fed's second-quarter accounts show corporate businesses with $2.786 trillion in checkable deposits and currency and $504 billion in time and savings deposits. Noncorporate businesses held $571 billion and $1.697 trillion, respectively. Those balances are excluded here. Payroll, supplier payments and existing treasury arrangements make them a separate market. Financial Accounts, tables S11.1.s and S11.2.s.

Banks have ways to keep the money

JPMorgan Chase can offer an everyday account, brokerage services and investment products within the same group. Its investment-services guide describes SmartCash. The service would move excess cash from an eligible Chase account into Premium Deposit through an eligible brokerage account. It would replenish the bank account when the balance falls below a target.

The guide describes weekly transfers and uses conditional wording about availability. We did not independently confirm a launch or test enrollment. Even so, the design shows how a bank could help customers earn more without surrendering the entire relationship. Whether the group keeps a deposit or gains an investment account depends on where the money ends up. J.P. Morgan services guide, SmartCash section.

Bank of America has reason to be selective about higher rates. Its Consumer Banking division held 47% of period-end deposits and paid an average deposit rate of 0.51% in the third quarter. That is an average funding cost for the division, not an advertised savings APY.

Paying more across those accounts would increase costs on balances that might otherwise stay. Merrill and relationship rewards give the bank other ways to retain customers and their surplus. Customers would still need to compare what they receive with offers elsewhere. Bank of America Q3 filing.

Wells Fargo can approach affluent customers through its consumer and wealth businesses, including Premier banking. It could help a household keep its everyday account while shifting extra cash to another account or investment at the company. That requires useful digital tools and reliable transfers; the bank's size doesn't establish that it can automate the job well. Wells Fargo Q3 filing.

Citi has a different deposit mix. At September 30, about $1.033 trillion of its $1.505 trillion in deposits was institutional, while wealth deposits were about $419 billion. Its filing describes Services clients using several integrated products. Managing corporate liquidity may matter more to Citi than competing for a household's spare savings. Citi Q3 filing, deposits discussion.

Exhibit 4

Four banks face different choices

JPMorgan Chase

$2.781T firmwide deposits; $1.104T Consumer & Community Banking, September 30, 2026.

Where the deposits come from
Consumer and small-business segment; global firmwide total is a different scope.
How it could keep customer money
Banking, brokerage and Premium Deposit could keep customer cash within the company.
What it could offer
Offer target-balance transfers; guide describes SmartCash with eligibility qualifications and weekly transfers.
What could go wrong
Weekly transfer timing and selective eligibility may leave openings for outside services.
Read the filing

Bank of America

47% of period-end deposits in Consumer Banking; 0.51% average rate paid there in Q3 2026.

Where the deposits come from
Consumer Banking cost is an average funding rate, not an advertised savings APY.
How it could keep customer money
Merrill and relationship rewards offer ways to retain customers without raising rates across all accounts.
What it could offer
Compete for surplus cash using banking and investment benefits. This is a possible strategy, not a confirmed plan.
What could go wrong
Higher rates can increase costs on existing deposits as well as attract new money.
Read the filing

Wells Fargo

Q3 2026 average Consumer Banking and Lending deposits $825.1B; Wealth and Investment Management $108.7B.

Where the deposits come from
Quarterly averages, not directly comparable with other banks' period-end figures.
How it could keep customer money
Premier banking and wealth advisers could help customers decide where to keep surplus cash.
What it could offer
Connect everyday accounts with wealth-management cash tools. This is a possible strategy, not a confirmed automated service.
What could go wrong
Customers need to qualify for the products and trust the transfers before moving more money to the bank.
Read the filing

Citigroup

$1.505T period-end deposits; about $1.033T institutional; Wealth about $419B, September 30, 2026.

Where the deposits come from
Global institutional operating relationships differ from U.S. household checking.
How it could keep customer money
Services integrates payments and liquidity management; Wealth serves a separate client base.
What it could offer
Automating corporate cash management may matter more than household transfers. This is analysis, not a quantified forecast.
What could go wrong
Corporate payment systems and rules for transfers between countries make household cash-transfer assumptions unsuitable.
Read the filing
Latest Q3 2026 filings reviewed November 15. Definitions and dates shown in each card. Strategic recommendations are analysis; no qualitative numerical scores or comparable market-share ranking is claimed.

These figures don't support a ranking of the banks' AI readiness. JPMorgan's consumer division includes small businesses; Wells Fargo's displayed figures are quarterly averages; Citi has a large global institutional business. No common measure shows how many customers at each bank use autonomous cash management.

What an independent service must prove

Rivo's disclosures show some of the practical choices. Its FAQ describes connections through Plaid, Treasury investments through Jiko Securities and banking services through Jiko Bank. Its Copilot makes one-way transfers; Autopilot includes replenishment. Earnings currently support one primary bank account, according to the company.

Rivo charges 0.05% monthly, or about 0.60% annually. It says transfers back take one to three business days and lists a $15,000 daily manual-withdrawal limit. It also acknowledges that automation cannot eliminate overdrafts. These are company disclosures, not results from an independent product test. Rivo FAQ.

For the Millers, a return transfer that takes several days would affect how much they could safely move. A fee of 0.60% would also consume part of any extra return. Its extra return would need to justify the fee and the time spent waiting for transfers.

Treasury investments bring different protections from bank deposits. They aren't FDIC-insured, and their prices can change before maturity. Money-market fund yields use a different convention from deposit APY. On November 13, four-week Treasury bills had a 3.95% bank-discount yield and a 4.02% coupon-equivalent yield. Fidelity's SPAXX reported a seven-day yield of 3.56% on November 15. These aren't interchangeable one-year returns and are shown separately from savings APYs. Treasury rates · Fidelity fund yields.

Exhibit 5

How an agent would keep money available for bills

  1. Paycheck arrives

    Confirm settled funds, pending debits and permissions.

  2. Forecast bills and liquidity

    Use scheduled payments and conservative spending allowances.

  3. Keep a safe operating balance

    Hypothetical example: $100,000 cash, $25,000 retained.

  4. Allocate surplus cash

    Up to $75,000; consider yield, fees, insurance and transfer access.

  5. Monitor upcoming obligations

    Reserve more cash when income or bills change.

  6. Return required funds

    Initiate early enough for holds, settlement and bank cutoffs.

Conceptual flow, not a tested product or reported household. Dollar example is hypothetical. Transfers are not assumed instantaneous. A forecasting error, bank hold or late transfer can cause a missed payment even when the yield calculation is correct.

An AI provider also has to improve on automation customers can already get. Fidelity's Cash Manager supports target balances and funding-account rules. Wealthfront offers automated savings above a customer-set maximum, while warning that overlapping withdrawals can still cause overdrafts. Moving cash when a balance crosses a threshold doesn't, by itself, require AI. Fidelity · Wealthfront.

The useful comparison is the money customers keep after fees, how accurately the service anticipates payments, and how often transfers fail. Connecting more accounts won't establish that it performs those tasks better.

Permission is a separate issue. The Consumer Financial Protection Bureau says a court stayed compliance dates for its personal-financial-data-rights rule. Access to account data shouldn't be treated as unrestricted permission to move money. Banks have legitimate concerns about credentials, consent and responsibility for errors; those concerns alone aren't evidence that a bank is protecting its interest margin. CFPB rule status.

Exhibit 6

How much cash moves under three sets of assumptions

Conservative

Cash available for a better rate
$606.9B
Balances affected
$91.0B
Deposits leaving the original bank
$36.4B
Annual customer gain before costs
$3.7B
Annual customer gain after entered costs
$3.7B

Surplus 20.0% · underpaid/access 25.0% · adoption 15.0% · external migration 40.0%

Central

Cash available for a better rate
$1.27T
Balances affected
$382.4B
Deposits leaving the original bank
$229.4B
Annual customer gain before costs
$15.4B
Annual customer gain after entered costs
$15.4B

Surplus 30.0% · underpaid/access 35.0% · adoption 30.0% · external migration 60.0%

Accelerated

Cash available for a better rate
$2.18T
Balances affected
$1.09T
Deposits leaving the original bank
$819.4B
Annual customer gain before costs
$43.9B
Annual customer gain after entered costs
$43.9B

Surplus 40.0% · underpaid/access 45.0% · adoption 50.0% · external migration 75.0%

Annual gain assumes the entered spread applies uniformly to affected cash for a full year. It is a sensitivity, not measured customer income. Rate compression, eligibility, taxes and transfer delays can reduce it. Cash retained by the original bank equals affected balances minus deposits leaving it.

How higher rates and lost deposits could affect banks
Retained deposits: added annual interest expense
$2.29B
Lost deposit balances
$229.42B
Replacement funding principal
$172.06B
Replacement annual interest expense
$6.88B
Avoided old annual deposit interest
$229.4M
Retained annual asset-management fees
$45.9M

Illustrative assumptions: old deposit cost 0.10%; replacement funding 4.00%; 20% of departing cash retained as affiliated investment assets earning 0.10% annual fees. Unreplaced funding $57B implies an asset/liquidity response whose income effect is unmodeled. Bank-to-bank transfers redistribute funding. System-wide NII also depends on asset yields, balance-sheet responses and where securities-sale proceeds land; customer gains do not equal bank losses.

Analyst scenarios, not probabilities or predictions. Same DFA baseline and selected SCF size filter in every column. Conservative and accelerated filters remain visible and fixed; changing the threshold, spread or fees recalculates all three. Central-case sliders also update the waterfall. No adoption trajectory or cumulative three-year income is inferred.

The bill still has to clear

Extra interest for customers doesn't translate dollar for dollar into lost bank earnings. A bank may pay more to retain deposits, replace money that leaves, or reduce the assets those deposits funded. It might also earn fees when customers shift money into its investment products.

Take $100 billion in affected balances. If 60% leaves, raising the rate on the remaining $40 billion by 1.5 percentage points adds $600 million in annual interest expense. Replacing three-quarters of the departing $60 billion at 4% adds $1.8 billion in gross annual funding expense. That excludes the interest saved on deposits that left. Asset income and investment fees would need separate calculations. This isn't a forecast of net interest income, the interest a bank earns minus what it pays.

For the Millers, the test is simpler. Can a service earn enough on their $75,000 to cover its costs while keeping their $25,000 available when they need it? If their existing bank can do that, they have less reason to hire an independent agent. If another service does it better, they could move the surplus and keep their everyday bank.

Before giving a service control, the Millers need to know when money will return and who handles a mistake.

Who would you trust to move your surplus cash: your bank, an independent service or yourself? What would you need to see before handing over control?

Sources and methodology

The Millers are a fictional family used to illustrate the numbers. This is sourced analysis, not interview-based reporting or a tested-product review. Research was retrieved November 15, 2026. Company capabilities are attributed to their disclosures; strategic judgments and market filters are explicitly analytical.

The DFA download supplies Q2 2026 wealth-group deposits, financial assets and household counts. The 2025 SCF public extract supplies balance tiers using all five implicates and survey weights. The Fed's extraction program defines the categories and confirms that WGT equals the original survey weight divided by five. The reproducible script recovers full within-implicate weights and averages five point estimates; it does not estimate replicate-weight uncertainty.

The rate engine uses dated, approved product observations with issuer URLs and checked eligibility rules. It holds rates constant and uses APY-equivalent daily accrual, including balance-tier changes. Actual issuer crediting, taxes, changing rates, transfer delays and fee timing can change realized returns. No universal eligibility or insurance coverage is implied. The rate examples retain their November 15 research date; the calculator excludes stale offers.

The DFA label Deposits includes currency, time deposits and short-term investments. The market model treats exclusions from that broad category, survey-to-DFA scaling, surplus liquidity, underpayment/access, adoption and external migration as assumptions. External displacement means departure from the originating bank; purchases of securities can leave proceeds elsewhere in the banking system. Business balances and existing fund holdings are not added to the household opportunity. Detailed source receipts, category definitions, balance distributions, exclusions and sensitivity outputs accompany the editorial package.

Get the week's biggest rate changes in savings, CDs and mortgages, free.

Or run your Money Map in 90 seconds.

Send me the weekly SwitchWize rate briefing: the biggest changes in savings, money market, checking and CD rates. Unsubscribe in one click. We never share your address.