How Banks Calculate Deposit Interest: Daily Balances, Average Daily Balances, Compounding, Tiered Rates and APY

Quick answer: a deposit-interest engine converts an account’s changing balance and rate rules into daily or periodic interest, then credits that interest according to the account contract. The calculation can use a daily-balance method or an average-daily-balance method, apply fixed or variable rates, handle balance tiers or stepped promotional periods, and compound credited interest into future principal. Annual Percentage Yield (APY) then converts the resulting interest into a standard annualised measure so deposit products can be compared more consistently.

The advertised rate is only one input. What a depositor actually earns depends on balance through time, compounding, tier rules and when interest starts and stops accruing.

Ownership boundary: calculation is not pricing strategy

Bukit Timah Tutor already has How Banks Choose Deposit Rates, which explains why a bank sets one rate rather than another. It also has How Banks Model Deposit Behaviour.

This article owns a different job:

Once the rate rules are fixed, how does the bank calculate the interest correctly from the account’s actual balance history?

1. Interest is a function of principal, rate and time

The simplest one-period formula is:

Interest = Principal × Rate × Time.

If S$10,000 earns 3% per year for exactly one year under simple interest, the interest is S$300.

Real deposit accounts are harder because principal changes. Customers deposit and withdraw money. Rates can vary. Interest may compound. Some balances earn different rates in different tiers. The engine must therefore operate on a time series rather than one static principal.

2. Daily periodic rate

If the stated annual interest rate is r, a common daily periodic rate for a 365-day basis is:

d = r / 365.

At 4% per year:

d ≈ 0.04 / 365 ≈ 0.000109589, or about 0.0109589% per day.

US Regulation DD provides a useful public rule example: institutions using the covered methods calculate interest on the full principal balance for each day using a daily rate of at least 1/365 of the interest rate, with permitted leap-year treatment. See Regulation DD section 1030.7.

3. Daily balance method

Under the daily-balance method, the bank applies the daily rate to each day’s principal balance.

Suppose an account pays 3.65% per year and uses a 365-day basis. The daily rate is exactly 0.01% in this teaching example.

The account has:

• S$10,000 for 10 days;
• then a S$5,000 deposit, giving S$15,000 for the next 20 days.

Ignoring compounding inside the month for simplicity:

First 10 days: 10,000 × 0.0001 × 10 = S$10.

Next 20 days: 15,000 × 0.0001 × 20 = S$30.

Total interest = S$40.

The algorithm respects when the additional deposit actually arrived instead of pretending S$15,000 was present for the whole month.

4. Average daily balance method

The same account history can be summarised with its average daily balance (ADB):

ADB = Σ daily balances / number of days.

For the 30-day example:

ADB = (10,000×10 + 15,000×20) / 30 = 400,000 / 30 ≈ S$13,333.33.

Applying the 0.01% daily rate across 30 days:

13,333.33 × 0.0001 × 30 ≈ S$40.

When implemented consistently, daily-balance and average-daily-balance approaches can produce the same result for a constant rate over the period. Regulation DD recognises both methods. See its model clauses for balance-computation methods.

5. Compounding changes future principal

If interest is credited into the account and future interest is calculated on the increased balance, interest compounds.

For principal P, annual nominal rate r and n equal compounding periods:

Future value = P(1 + r/n)ⁿ

for a simplified one-year fixed-rate case.

With S$10,000 at 4% nominal compounded monthly:

10,000(1 + 0.04/12)¹² ≈ S$10,407.42.

The depositor earns about S$407.42 rather than S$400 because credited interest itself begins earning interest.

The FDIC’s consumer education material uses this same core concept: once interest is added to principal, later interest can accrue on both. See FDIC — Compound Interest.

6. Interest-crediting frequency and compounding frequency are related but should be read from the contract

An account can calculate interest daily but credit it monthly. Another can use a different stated frequency. The calculation engine therefore needs separate fields for:

• rate effective date;
• daily/periodic accrual rule;
• balance computation method;
• compounding treatment;
• interest-crediting date;
• account maturity or renewal where relevant.

A system that stores only “4% savings rate” does not contain enough information to reproduce the account.

7. APY converts rate and compounding into one annualised yield

APY is designed to reflect the total interest that would be earned over a standard annual horizon given the account’s interest and compounding structure.

Regulation DD’s general formula is:

APY = 100 × [(1 + Interest/Principal)^{365/Days in term} − 1].

See Appendix A — Annual Percentage Yield Calculation.

For a one-year account with daily compounding at nominal 4%:

APY ≈ (1 + 0.04/365)³⁶⁵ − 1 ≈ 4.081%.

The APY exceeds the nominal rate because of compounding.

8. APY and APY earned are different measurements

Advertised APY describes the account under standard assumptions. APY earned on a periodic statement reflects the interest actually earned relative to the account’s average daily balance during that statement period.

Regulation DD’s periodic-statement formula annualises:

Interest actually earned / average daily balance

over the number of days in the period. See APY Earned for Periodic Statements.

If the customer withdrew half the balance mid-month, actual interest falls even though the account’s advertised APY did not change.

9. Tiered rates create piecewise mathematics

Suppose an account pays:

• 3.0% on the first S$10,000;
• 3.5% on the next S$40,000;
• 4.0% above S$50,000.

There are at least two common conceptual tier designs.

Whole-balance tier: once the account reaches a threshold, one rate applies to the full eligible balance.

Marginal tier: each rate applies only to the portion of the balance inside that band.

Under a marginal-tier design, S$60,000 earns:

• 3.0% on S$10,000;
• 3.5% on S$40,000;
• 4.0% on S$10,000.

A system that accidentally applies 4.0% to all S$60,000 has changed the product economics. Regulation DD contains separate APY treatment for tiered-rate accounts because the calculation depends on the institution’s tiering method.

10. Stepped rates are different from tiered rates

A stepped-rate account changes its rate over time rather than by balance.

Example:

• 3% for the first three months;
• 4% for the next three months;
• 5% thereafter.

The interest engine must apply the correct rate to the correct dates. APY then annualises the combined result. Regulation DD Appendix A includes explicit stepped-rate examples.

11. Variable rates require effective-dated state

If a savings rate changes from 3.5% to 4.0% on 15 June, the bank should not recalculate the first half of June at the new rate unless the account terms explicitly require that treatment.

The database therefore needs an effective-dated rate history:

[start date, end date, rate, tier rules, product version].

That rate history must join correctly to the daily balance history. Interest calculation is therefore partly a temporal-database problem.

12. Minimum-balance rules need the same balance definition

If an account requires a minimum balance to earn interest, the bank needs a consistent definition of that balance. US Regulation DD requires covered institutions to use the same method for determining a minimum balance as for the balance on which interest is calculated, unless an additional method is unequivocally beneficial to the consumer.

This prevents a logically inconsistent product in which one balance method determines eligibility while another less favourable method determines earnings.

13. Interest accrual start and stop dates matter

Two accounts with the same balance and rate can earn different interest if deposits begin accruing on different dates or if withdrawals stop accrual at different times.

The interest engine should therefore retain value dates and accrual dates separately where required. Regulation DD, together with funds-availability rules, provides public examples of when interest on covered accounts must begin accruing and states that interest accrues until the day funds are withdrawn.

14. Rounding is a control problem

Daily interest often produces fractions of a cent. If the bank rounds each day to cents, the result can differ from carrying greater precision internally and rounding only when interest is credited.

A robust engine should define:

• rate precision;
• daily interest precision;
• balance precision;
• rounding mode;
• when rounding occurs;
• how residual fractions are treated.

The CFPB’s official commentary on Regulation DD recognises carrying daily rates and daily interest to multiple decimal places for APY calculations. See CFPB Appendix A commentary.

15. A good interest engine should reproduce the statement from raw events

The strongest validation test is not “the formula looks right.” It is reconstruction:

1. Start with opening principal.
2. Replay every deposit and withdrawal by effective date.
3. Apply the correct daily rate and tier rules.
4. Accrue interest at full internal precision.
5. Apply compounding/crediting according to product terms.
6. Reproduce closing balance and interest credited.
7. Recalculate APY/APY earned independently.

If an independent calculation cannot reproduce the statement, the bank needs to know whether the error sits in the transaction history, product rules, rate table, date logic or arithmetic.

16. Creative-work lens: the snowball only grows from what is actually attached

Compounding is often compared with a snowball rolling downhill. The image is useful but incomplete. A real deposit account does not roll down a perfect slope: customers add money, withdraw money, cross tiers and experience rate changes. The snowball metaphor explains why interest can earn interest; the ledger explains the actual amount.

That is the recurring eduKate mathematics lesson: intuition gets us to the mechanism, but exact state and chronology decide the answer.

17. The deposit-interest algorithmic pipeline

1. Load the product’s effective-dated rate and tier rules.
2. Reconstruct daily principal balances from transactions.
3. Determine interest eligibility and minimum-balance status.
4. Calculate the applicable daily/periodic rate.
5. Apply daily-balance or average-daily-balance logic.
6. Apply tiered or stepped rules correctly.
7. Accrue interest at defined internal precision.
8. Compound/credit interest on the contractual schedule.
9. Update principal after credited interest.
10. Calculate APY or APY earned where required.
11. Generate statement interest and disclosures.
12. Reconcile with the general ledger and independent recalculation.

18. Failure modes

• Ending-balance shortcut. Interest is calculated from month-end balance instead of daily principal history.
• Rate-history overwrite. A new variable rate is applied retroactively to old days.
• Tier confusion. Whole-balance and marginal-tier methods are mixed.
• Stepped/tiered confusion. Time-based rate changes are treated as balance thresholds.
• Compounding mismatch. Credited interest is added to principal at the wrong frequency.
• Early rounding. Daily cents are rounded too soon and accumulate systematic error.
• APY/APR confusion. Deposit yield is mixed with credit-cost terminology.
• Accrual-date error. Deposits or withdrawals start/stop earning on the wrong business date.

19. Diagnostics and falsifiers

• Can the account’s interest be reproduced from raw daily balances?
• Does moving a deposit one day earlier increase interest by the expected daily amount?
• Does a leap year produce the intended daily-rate treatment?
• Does a balance crossing a tier by S$1 change only the intended portion?
• Does the APY exceed nominal rate when compounding should make it do so?
• Does a mid-period rate change affect only the dates after its effective time?
• Can statement APY earned be independently reproduced?
• Do daily balance and average daily balance implementations agree when their assumptions are equivalent?

Suppose someone claims, “A 4% interest rate means S$10,000 always earns S$400 per year.” A falsifier is an account with daily compounding, changing balances, tier rules or a partial-year holding period. The nominal rate is not the complete cash-flow algorithm.

20. Verification and update triggers

• unit-test known balances and rates;
• replay transaction histories through an independent calculator;
• test month-end, year-end and leap-year boundaries;
• test every tier edge exactly below, at and above the threshold;
• verify rate effective dates after product changes;
• compare credited interest to ledger postings;
• recalculate APY disclosures after compounding or fee changes;
• retain historical product versions so old statements can still be reconstructed.

Connections across the finance-and-banking algorithms lane

• Deposit-rate optimisation — decides the rate; this article calculates the resulting interest.
• Deposit behaviour — models how customers respond to those rates.
• Loan amortisation — applies related compounding and recurrence mathematics on the borrowing side.
• Transaction reconciliation — raw balance history must be correct before interest can be correct.

Research anchors

• Regulation DD — Payment of Interest.
• Regulation DD Appendix A — APY Calculation.
• Regulation DD — APY Earned.
• CFPB — Truth in Savings, Regulation DD.
• FDIC — Compound Interest.

The deeper lesson

Deposit interest is the mathematics of state through time. The balance changes. The rate can change. The account can cross tiers. Interest can become principal and compound. APY translates the resulting path into a common annual measure. A strong calculation engine therefore does not remember only the balance and the rate. It remembers the chronology that made them.

Educational note: This article explains deposit-interest mathematics and public disclosure concepts. It is not a recommendation for any deposit product and does not replace the actual terms or law governing a particular account.