Most of what India borrows pays for what it already borrowed
The 2026-27 Budget borrows ₹16.96 lakh crore and spends ₹14.04 lakh crore servicing past debt. Two years ago that ratio was 71 paise in the rupee. It is now 83.
The Union Budget for 2026-27 provides for borrowings and other liabilities of ₹16,95,768 crore and interest payments of ₹14,03,972 crore. Interest is therefore 82.8% of new borrowing — roughly 83 paise of every rupee borrowed goes to servicing what was borrowed before.
The ratio was 70.9% in the 2024-25 actuals. It has risen in each subsequent estimate.
Where a budget rupee comes from, and where it goes
The Finance Ministry publishes two pie charts every year showing the composition of a budget rupee in paise. Set side by side as a flow, they show something the pies separately do not: the single largest source of money is borrowing, and the second largest use is the interest on borrowing.
Borrowings and other liabilities are 24 paise of every rupee the government handles — more than income tax (21), more than corporation tax (18), more than GST and other taxes (15). On the other side, the states’ share of central taxes takes 22 paise and interest takes 20.
View data table
| Rupee comes from | Paise | Rupee goes to | Paise |
|---|---|---|---|
| Borrowings & other liabilities | 24 | States’ share of taxes | 22 |
| Income tax | 21 | Interest payment | 20 |
| Corporation tax | 18 | Central sector schemes | 17 |
| GST & other taxes | 15 | Defence | 11 |
| Non-tax revenue | 10 | Centrally sponsored schemes | 8 |
| Union excise duties | 6 | Finance Commission & other transfers | 7 |
| Customs | 4 | Other expenditures | 7 |
| Non-debt capital receipts | 2 | Major subsidies | 6 |
| Civil pension | 2 | ||
| Total | 100 | Total | 100 |
Interest is catching up with borrowing
A government can borrow and still be gaining ground, so long as what it borrows exceeds what it owes on past borrowing by enough to fund something new. The gap between those two lines is, near enough, the primary deficit — the deficit before interest. In this Budget it is ₹2,91,796 crore.
That gap has narrowed sharply. In the 2024-25 actuals, interest was 70.9% of the fiscal deficit. In the 2026-27 estimates it is 82.8%. The borrowing is not mostly buying new things; it is mostly rolling forward the cost of old ones.
View data table
| Year | Interest payments | Fiscal deficit | Interest as % of deficit |
|---|---|---|---|
| 2024-25 Actuals | ₹11,15,575 cr | ₹15,74,431 cr | 70.9% |
| 2025-26 Budget Est. | ₹12,76,338 cr | ₹15,68,936 cr | 81.4% |
| 2025-26 Revised Est. | ₹12,74,338 cr | ₹15,58,492 cr | 81.8% |
| 2026-27 Budget Est. | ₹14,03,972 cr | ₹16,95,768 cr | 82.8% |
The pie’s rupee is not the Centre’s rupee
Here is a discrepancy worth understanding rather than explaining away. The pie says interest is 20 paise in the rupee. The table in the same document reports interest of ₹14,03,972 crore against total expenditure of ₹53,47,315 crore — which is 26.3%. Same document, same year, same line item.
Both are right. The document says why, in a footnote on each pie: total receipts and expenditure in the pies are inclusive of the states’ share of taxes and duties, which have been netted against receipts in the table. The pie is drawn on a larger rupee — one that includes money the Centre collects and passes straight to the states. The table is drawn on what the Centre actually has to spend.
So when a headline says interest takes a fifth of the Budget, it is quoting the pie. Measured against the money the Union government itself commands, interest takes closer to a quarter. Neither figure is spin; they answer different questions. But they are not interchangeable, and they are routinely quoted as if they were.
View data table
| Denominator | Value | Interest as % |
|---|---|---|
| Fiscal deficit (new borrowing) | ₹16,95,768 cr | 82.8% |
| Total expenditure, table basis | ₹53,47,315 cr | 26.3% |
| The budget rupee, pie basis | includes states’ share | 20% |
GST fell three paise, corporation tax rose one
Comparing the 2026-27 rupee with 2025-26’s: GST and other taxes fell from 18 paise to 15, the largest single move on either side. Income tax fell from 22 to 21. Corporation tax rose from 17 to 18, excise from 5 to 6, non-tax revenue from 9 to 10, and non-debt capital receipts from 1 to 2. Borrowings held at 24.
On the spending side the changes were smaller: defence rose from 10 paise to 11, central sector schemes from 16 to 17, while the Finance Commission and other transfers fell from 8 to 7 and other expenditures from 8 to 7. The two largest items — the states’ share at 22 and interest at 20 — did not move.
A caution on reading these: the pies are rounded to whole paise, so a one-paise move is not necessarily a one-per-cent change, and could in principle be a rounding boundary rather than a real shift. The three-paise fall in GST is large enough to be safe from that objection. A one-paise move is not.
Budget estimates are intentions, not outcomes. Everything for 2026-27 on this page is a Budget Estimate — what the government proposes to raise and spend. The 2025-26 revised estimates on the same table already differ from its budget estimates. Actuals will differ again.
Two denominators. Section 03 is the whole point of this dive: the pie and the table do not use the same rupee, and the government says so in a footnote most readers never see. Any figure quoted from this page should carry its denominator with it.
“83 paise of every rupee borrowed” is a ratio, not a flow. Money is fungible; no rupee is tagged. The statement means interest payments equal 82.8% of the year’s new borrowing, not that specific borrowed rupees are handed to bondholders. The distinction matters if you want to say something about causation.
A rising interest-to-deficit ratio is not automatically a crisis. It can also mean the deficit is shrinking relative to the accumulated stock of debt, which is what you would expect during consolidation. India’s fiscal deficit is budgeted at 4.31% of GDP, down from earlier years. The ratio in section 02 is doing two things at once, and this dive does not attempt to separate them.
What this cannot tell you. Nothing here addresses whether the borrowing is worthwhile. Effective capital expenditure is budgeted at ₹17,14,523 crore — more than the fiscal deficit — which is the government’s answer to that question, and assessing it needs a different investigation.
- Budget at a Glance 2026-27, Ministry of Finance, Government of India. Both rupee compositions, the receipts and expenditure table, and the GDP estimate.
- Union Budget portal — the full budget documents, including the Expenditure Budget and Receipt Budget that sit behind the summary used here.
Percentages of GDP, of expenditure and of the deficit are ImpactMojo’s arithmetic on published rows and are shown in the data tables. The paise figures are the government’s own and are reproduced unchanged. Fiscal deficit and “borrowings and other liabilities” are the same row in this document; primary deficit reconciles exactly as fiscal deficit minus interest payments.
ImpactMojo Data (2026). “Most of what India borrows pays for what it already borrowed.” ImpactMojo Data Dives. Retrieved from https://impactmojo.in/DataDives/borrowing-to-pay-for-borrowing.html
Have a dataset worth digging into?
Data Dives are independent investigations built from public data. If you know a number that deserves a closer look, tell us.
Pitch a Data Dive →