Use the statutory sequence
For tax year 2026–27, section 516 of the Income-tax Act, 2025 requires the relevant computed total income or amount payable or refundable under the Act to be rounded to the nearest multiple of ₹10. First ignore paise. Then increase to the next multiple where the remaining rupee unit digit is five or more; otherwise reduce to the lower multiple. Do not substitute ordinary spreadsheet rounding to the nearest rupee as an intermediate step. Income Tax Department: amended Income-tax Act, 2025, sections 1 and 516.
Establish which amount is being rounded
The Act, effective from 1 April 2026 and reviewed here as amended by Finance Act, 2026, combines total-income and payable/refundable rounding in section 516. This is a finishing rule for the relevant statutory amount, not permission to round every receipt, expense or withholding entry to ₹10. Retain detailed source precision, complete the proper computation, identify its statutory endpoint and apply the rule there. This article does not determine a tax rate or replace the rules for computing surcharge, cess, interest or credits.
Preserve a three-stage bridge
Keep the unrounded amount, the whole-rupee amount after discarding paise, and the final multiple of ten. The difference between the first and last figures is a reconciliation adjustment, not new income, expenditure or tax relief. A ₹4.99 tail is the useful boundary test: discarding paise leaves four, so the amount goes down. Rounding to the nearest rupee first would incorrectly turn it into five and push it up.
Do not round components before aggregation
Two components can each have a final digit below five while their combined amount crosses a ten-rupee boundary. An automated return working should therefore distinguish raw components from the relevant total. Protect formulas from an extra round-to-rupee function and avoid binary floating-point boundary errors by using decimal arithmetic or paise-based integers. These are implementation controls; section 516 does not prescribe a software brand or function name.
Reconcile a refund as a positive refund amount
First establish whether the final working produces a liability or a refund under the applicable computation. Apply this guide’s positive-amount examples to that amount rather than relying on a programming language’s treatment of negative half values. Verify that the final displayed figure, export and statutory payment/refund working agree. A rounding difference should not conceal an unrelated credit mismatch or an omitted interest amount. The examples isolate rounding only and do not represent complete tax assessments.
Applicable primary references
Income Tax Department: amended Income-tax Act, 2025, sections 1 and 516
Six worked illustrations
All facts and amounts below are fictional. Assumptions describe the model, not additional statutory rules.
Illustrative example 1: Below the five-rupee boundary
Computed total income ₹8,24,564.99 becomes ₹8,24,564 after ignoring paise. Its unit digit is four, so section 516 gives ₹8,24,560. The difference from the original amount is minus ₹4.99.
Illustrative example 2: At the five-rupee boundary
Computed total income ₹8,24,565.01 becomes ₹8,24,565 after ignoring paise. Its unit digit is five, so the result is ₹8,24,570, an increase of ₹4.99 from the original amount.
Illustrative example 3: Payable amount with an existing ten-rupee multiple
Assume the properly computed amount payable is ₹17,230.99. Ignoring paise gives ₹17,230, already a multiple of ten. The final amount is ₹17,230; it must not become ₹17,240 merely because the original included paise.
Illustrative example 4: Refund rounding
A properly determined refund of ₹3,455.75 first becomes ₹3,455 and then ₹3,460. The increase from the unrounded refund is ₹4.25. This arithmetic does not establish whether the underlying refund claim is legally admissible.
Illustrative example 5: Premature component rounding changes the result
Assume two components of the relevant final amount are ₹104.40 and ₹104.40. Their sum is ₹208.80, which becomes ₹208 and then ₹210. Rounding each component to ₹100 before summing gives ₹200, understating the correctly rounded aggregate by ₹10.
Illustrative example 6: A low-value payable
A properly computed payable of ₹4.99 becomes ₹4 after paise are ignored and then ₹0. A payable of ₹5.00 instead rounds to ₹10. The ₹10 difference illustrates the statutory boundary, not an exemption from computing or reporting income.
Ten-step verification table
| Step | Action | Evidence |
|---|---|---|
| 1 | Fix the applicable period | Use the Act governing the tax year and verify section 516. |
| 2 | Identify the statutory endpoint | Total income, payable or refundable amount, rather than a raw component. |
| 3 | Complete the underlying computation | Deductions, rates, surcharge, cess, interest and credits as applicable. |
| 4 | Retain source precision | Unrounded amount and supporting schedules. |
| 5 | Ignore paise first | Whole-rupee figure produced without rounding paise upwards. |
| 6 | Inspect the rupee unit digit | Five or more upwards; below five downwards. |
| 7 | Keep exact multiples unchanged | No extra ₹10 addition where the integer is already divisible by ten. |
| 8 | Test aggregate logic | Components combined before the relevant final rounding. |
| 9 | Reconcile direction and difference | Liability/refund classification and rounding adjustment. |
| 10 | Verify exported figures | Final output agrees with the supported statutory working. |
Sixteen practitioner FAQs
1. Is section 516 the current reference for 2026–27?
It is the rounding provision in the amended Income-tax Act, 2025 used in this guide.
2. What happens to paise first?
They are ignored before testing the remaining whole-rupee figure.
3. Does ₹4.99 round up because it is nearly five?
No. Ignoring paise leaves four for this statutory sequence.
4. What happens where the rupee unit digit is five?
Increase to the next higher multiple of ten.
5. What if the whole-rupee figure already ends in zero?
It is already a multiple of ten and remains unchanged.
6. Should every invoice be rounded to ₹10?
Section 516 does not make each source invoice a separate final tax-computation endpoint.
7. Can expense components be rounded before aggregation?
Do not use premature component rounding in place of rounding the relevant computed amount.
8. Does the rule also address refunds?
Yes, section 516 expressly covers amounts payable or refundable.
9. Does a rounded refund prove entitlement?
No. Determine the refund under the applicable provisions first.
10. Does this article compute surcharge or cess?
No. Those calculations must be completed correctly in the underlying tax working.
11. Can ordinary spreadsheet rounding produce a wrong boundary result?
Yes, especially if paise are rounded to rupees before the ten-rupee step.
12. Why retain the discarded paise in a working paper?
They explain the reconciliation between the detailed computation and final statutory figure.
13. Is a rounding difference a separate deduction?
No. It is the effect of the statutory finishing rule.
14. How should software handle refund signs?
Establish the positive refund amount and apply the rule deliberately rather than assuming a language-specific negative-rounding convention.
15. Are binary floating-point values always reliable at boundaries?
Use decimal or integer-paise controls and test the relevant boundaries.
16. What should be investigated if exported figures still differ?
Trace extra rounding, credits, interest and mapping differences instead of attributing every mismatch to section 516.
Conclusion
Apply section 516 by ignoring paise first, then rounding the relevant total income or payable/refundable amount to the nearest ₹10. Preserve the calculation bridge.