A credit analyst needs a true average balance for an applicant's account, but the statement only lists days with transactions.
Take the last closing balance of each day, carry it forward on days with no entry, add up all daily balances and divide by the number of days in the period.
A complete statement for the period with running balances, the period start and end dates, and the opening balance on the first day.
Average daily balance for each month, the lowest daily balance, and the number of days the balance stayed below a chosen level such as the EMI.
To calculate average daily balance (ADB), take the closing balance of every day in the period, add them up and divide by the number of days. On days with no transaction, the balance did not change, so carry the last closing balance forward. Use calendar days, not working days.
What are the steps?
- Note the opening balance on the first day of the period.
- For each day with transactions, take the balance after the last entry of that day.
- For each day with no entry, repeat the previous day’s closing balance.
- Add all the daily balances.
- Divide by the number of calendar days in the period.
A quicker method for long statements: group days with the same balance, multiply each balance by the number of days it stayed, add the results and divide by total days.
How do you handle days with several entries or no entry?
| Situation | What to use |
|---|---|
| Several entries on one day | Closing balance after the last entry |
| No entry that day | Previous day’s closing balance |
| Weekend or bank holiday | Previous closing balance, still counted as a day |
| Value date differs from transaction date | Follow the lender’s policy; most use the balance column as printed |
Intra-day balances do not count. If ₹2,00,000 comes in at 10 am and leaves at 4 pm, the day closes at the old balance.
Worked example: a 30-day month
An applicant’s statement for a 30-day month shows three balance levels:
| Days | Balance held | Number of days | Balance × days |
|---|---|---|---|
| 1st to 10th | ₹50,000 | 10 | ₹5,00,000 |
| 11th to 22nd | ₹12,000 | 12 | ₹1,44,000 |
| 23rd to 30th | ₹80,000 | 8 | ₹6,40,000 |
| Total | 30 | ₹12,84,000 |
ADB = ₹12,84,000 ÷ 30 = ₹42,800.
Now compare a three-date sample of the 5th, 15th and 25th: (₹50,000 + ₹12,000 + ₹80,000) ÷ 3 = ₹1,42,000 ÷ 3 = ₹47,333. The sample overstates the true average by ₹4,533, because only one of its three dates fell inside the 12-day low stretch.
A simple month-end figure would show ₹80,000, almost double the ADB.
How is ADB different from a sample-date average, and which do lenders prefer?
ADB uses every day, so a short spike barely moves it. A sample-date average is faster to work out by hand but depends on which dates are chosen. Many lenders use sampled dates for speed and ADB where software does the arithmetic, but each lender’s credit policy decides.
Why does a month-end average hide a dip before the 5th?
Many EMIs are debited in the first week of the month. In the example above the balance was healthy on the 1st to the 10th, but if a different applicant’s balance dropped every month between the 1st and the 5th, a month-end figure would never show it. Two extra checks help:
- The lowest closing balance in each month.
- The number of days the balance stayed below the proposed EMI.
Across 12 months and several accounts, this means hundreds of daily balances, which is why credit teams usually rely on software for it.
What to read next
Read bank statement analysis accuracy signals to see where balance behaviour fits among other signals, and building a synthetic balance sheet from an MSME bank statement. For the definitions behind ABB, AMB and ADB, start at what are ABB, AMB and ADB.
TransactIQ reads bank statements for lenders and returns balance behaviour among its 40+ engineered credit signals. See the bank statement analyzer.