A state bond yields more than a central bond mostly because it is hard to sell, and only a few basis points because it is a state
The short answer
A state development loan is a state government's own bond, auctioned by the RBI most Tuesdays and secured on the state's Consolidated Fund under Article 293(1). The central government does not guarantee it; the RBI's own research calls the guarantee investors assume implicit. It yields more than a central security of the same maturity, by 30 basis points in 2024-25 and 50 in 2025-26 on the RBI's measure. Measured across 2,327 state securities in 130 RBI auctions from April 2024 to September 2026, that spread is mostly a price for liquidity, not for credit: the auction date explains 70 per cent of its variation and the state 13; it is widest beside the 10-year central benchmark, about 70 bps in 2025-26, and about 15 beyond 20 years; and between states it follows issue size more closely than debt, which adds about 1.2 bps per 10 points of GSDP. State bonds are 35 per cent of the government bond stock and 5.6 per cent of outright trading. Since April 2026 the RBI's Benchmark Issuance Strategy has raised re-issues from 12 to 62 per cent of state securities sold. The spread has not narrowed yet.
The usual explanation of why a state bond pays more than a central one is credit: states are weaker borrowers than the Union, so lenders charge them more. It is the natural guess, and the weekly auction results, which the RBI publishes in full, do not bear it out as the main part of the price. What they show is a premium that rises and falls for every state at once, is largest where the central alternative is easiest to sell, and differs between states by a few basis points, mostly according to how much paper each state brings to market. That is the signature of a price for liquidity.
This page sets out what the bond is, what does and does not stand behind it, and then the measurement: every state and central auction result from April 2024 to September 2026, and the same instrument run on 2020-21. It builds on the arithmetic of price and yield, the yield curve and the auction and the retail bid, which are covered on their own pages.
A state's own bond, sold through the central bank's window
A state development loan is an ordinary dated bond with a state government as the borrower. It pays a fixed coupon every six months and repays at par on maturity, and it is issued, registered and transferred under the same Government Securities Act 2006 as a central bond. What makes it a state's is the security behind it: Article 293(1) of the Constitution lets a state borrow within India upon the security of its own Consolidated Fund, within whatever limits its legislature sets.
The machinery belongs to the RBI. Under agreements made under section 21A of the RBI Act it manages the debt of 28 states and three union territories, Delhi having joined on 5 January 2026. Each quarter it publishes an indicative calendar of how much each state intends to borrow; each week it announces the securities on offer; and on the auction day, a Tuesday on 124 of the 130 days in the sample, it runs a separate auction for every state security on its E-Kuber system, with payment the next working day. Between 2 April 2024 and 22 September 2026 that produced 130 weekly results and 2,368 state securities offered by 31 states and union territories.
Three details of those results matter for everything that follows. Each state security is small: the median amount notified in 2025-26 was ₹1,000 crore, against ₹14,000 crore for a central security in a central auction. The retail reserve is larger, up to 10 per cent of each security against 5 per cent in a central auction, with any single non-competitive bid capped at 1 per cent. And the state can refuse the price: 41 securities in the sample drew bids that the issuing state declined in full. The auction is discriminatory, each successful bidder paying its own bid, which is how a 2017 RBI staff study described it and what the results show: the weighted average yield sits below the cut-off in 1,956 of them, equal to it in 371, and above it in none.
Who holds the bonds explains a great deal. At end-March 2026 the RBI Bulletin's ownership table puts banks at 36.05 per cent of the ₹73,03,451 crore of state securities outstanding, insurers at 23.19 per cent, provident funds at 23.11 per cent and pension funds at 5.43 per cent. Foreign portfolio investors held 0.01 per cent and the RBI 0.49 per cent. These are holders who buy to keep.
| State development loan | Central government security | |
|---|---|---|
| Borrower and security | A state, on its own Consolidated Fund (Article 293) | The Union, on the Consolidated Fund of India (Article 292) |
| Usual auction day | Tuesday, paid the next working day | Friday, paid the next working day |
| Non-competitive reserve | Up to 10 per cent of each security | Up to 5 per cent of each security |
| Median size of one security in one auction, 2025-26 | ₹1,000 crore | ₹14,000 crore |
| Number of securities | 833 new ones created in 2025-26 alone | 122 make up the entire stock at end-March 2026 |
| Outstanding, end-March 2026 | ₹73,03,451 crore | ₹1,25,67,045 crore |
| Held by provident funds | 23.11% | 4.50% |
| Held by foreign portfolio investors | 0.01% | 2.95% |
| Held by the RBI | 0.49% | 17.59% |
| Share of outright trading, 2025-26 | 5.6% | 85.5% |
| Bank capital risk weight | Zero | Zero |
The guarantee that is assumed, not written
Whether the central government stands behind a state bond has two answers that are easy to run together, one legal and one practical.
The legal answer is short. Article 293(2) allows the Government of India to guarantee loans raised by a state, within the limits Parliament sets under Article 292. None of the RBI's auction notices or results for these bonds mentions a guarantee. The RBI's own researchers describe the guarantee that investors assume as implicit: a 2017 RBI staff memo that found no significant link between state spreads and state deficits or debt put it down to "investor perception of implicit central government guarantee" and to the regulatory treatment of state bonds as free of credit risk, and a 2023 RBI Bulletin study makes the same point.
The practical answer is longer, and it is why the perception is reasonable. The first support is consent. Under Article 293(3) a state that still owes the central government money cannot borrow without its consent; the RBI's table of state borrowing opens with the gross sanctions given under that clause, ₹13,63,263 crore for 2025-26 against the ₹12,76,430 crore states raised, and the RBI's calendar of 25 June 2026 says the auctions depend on that approval. The centre does not guarantee state debt, but through consent it controls how much of it is issued.
The second is the RBI itself, which services the debt as the state's banker. The third is money set aside: states hold a Consolidated Sinking Fund and a Guarantee Redemption Fund at the RBI, ₹2,66,224 crore and ₹51,992 crore at end-March 2026. The fourth is short-term credit with hard edges: a special drawing facility against those funds and treasury bills, ways and means advances with an aggregate limit of ₹61,008 crore since 9 January 2026, and overdraft for no more than 14 consecutive days and 36 days a quarter. If a state's overdraft exceeds its advance limit for five consecutive working days a second time in a year, the RBI stops making payments for it. In 2025-26, 19 states and union territories used the drawing facility, 11 the advances and 10 the overdraft.
The fifth is regulation. A bank's holding of state securities carries a zero risk weight, the same as a central bond, while a loan that a state merely guarantees carries 20 per cent (RBI Basel III master circular, paragraph 5.2.2; the RBI's concentration risk directions of 28 November 2025 still treat state government exposures as eligible for a zero weight). State bonds also count toward the statutory liquidity ratio and are accepted as collateral at the RBI's liquidity window.
None of the written layers obliges the central government to pay a state's bondholders if a state's finances failed over years rather than days. That gap is what the word implicit covers, and whether the market charges for it is a question the auction results can answer.
Measured: one spread, moving for every state at once
The RBI prints each state auction's cut-off yield on Tuesdays and each central auction's on Fridays, which is enough to measure the spread without any licensed benchmark curve. The rule used here: for every state security sold, take the central security auctioned within seven days before or after whose remaining maturity is closest, provided it is within a year; where one qualifies on each side, average the two. That pairs 514 of the 2,327 state securities, mostly between 7 and 20 years, with an average maturity gap under half a year.
Two checks come first, because a spread is only as good as the yields on each side of it. Each state re-issue's printed price and yield were recomputed from its coupon, maturity and settlement date on the half-yearly 30/360 convention: 590 of 591 reproduce to the fourth decimal; in the exception the printed price and yield disagree, the yield implying 99.38 against the 99.34 printed, and the yield is what is used. Every central cut-off used passed the same test against its own printed price (224 of them) or, for a newly created bond with no price, against its weighted average yield and, where the bond reappears, the coupon it then carries (32).
Pairs cluster near ten years, where both markets issue most, so the raw average over-weights that segment. Re-weighted to each year's actual maturity mix, the measure lands within three basis points of the RBI's published spread, and the inter-state range on fresh 10-year paper, the highest cut-off less the lowest on the same day, lands within half a basis point of the RBI's.
| 2024-25 measured | 2024-25 RBI | 2025-26 measured | 2025-26 RBI | |
|---|---|---|---|---|
| State securities sold | 833 | 835 | 1,050 | 1,055 |
| Of which re-issues | 99 | 100 | 217 | 217 |
| Amount raised, ₹ crore | 10,71,110 | 10,73,310 | 12,72,003 | 12,76,430 |
| Weighted average spread over comparable central securities | 28 bps | 30 bps | 47 bps | 50 bps |
| Inter-state spread, fresh 10-year issues | 3.5 bps | 4 bps | 8.3 bps | 8 bps |
The 10-year segment sat between 33 and 42 basis points for five quarters, then jumped to 71 in July to September 2025 and has stayed above 70. Over the two years to March 2026 the ratio of all states' liabilities to GDP moved from 28.4 to 29.2 per cent (RBI, State Finances, Statement 20). The inter-state range roughly doubled, from 3.5 to 8.3 basis points, and that is the most that could be about individual states; the common move was about 34. Across all 514 pairs, knowing the auction date accounts for 70 per cent of the variance of the spread, and knowing the state for 13 per cent.
The auction data identify the timing of the 2025 step, not its cause, and supply alone does not explain it: across the ten quarters, the amount sold per auction and the 10-year spread are only loosely related (r = 0.23), and the heaviest quarters, the January to March rush each year, show no matching jump. A shock common to every state's finances would also move all states together, so the time pattern alone cannot prove the premium is liquidity. Two further patterns do the discriminating: its shape across maturities, and who pays more on the same day.
The spread is widest where the central bond is most liquid
A premium for the chance of default grows, if anything, with the time over which a default could happen. This one does the opposite. In every year measured, the spread is widest for 7 to 12-year paper and smallest beyond 20 years: 70 against 15 basis points in 2025-26, and 35 against 14 in 2024-25.
| Remaining maturity | 2024-25 | 2025-26 | 2026-27, April to September | Share of 2025-26 issuance |
|---|---|---|---|---|
| Under 7 years | 30 bps (24) | 55 bps (43) | 69 bps (21) | 10% |
| 7 to 12 years | 35 bps (100) | 70 bps (107) | 74 bps (33) | 34% |
| 12 to 20 years | 30 bps (56) | 48 bps (62) | 57 bps (32) | 29% |
| 20 years and over | 14 bps (13) | 15 bps (15) | 27 bps (8) | 27% |
The reason sits on the central side. The 10-year central benchmark is the most liquid bond in the country: in January to March 2026 that single security was 50.8 per cent of all outright trading in government securities, and the ten most traded central securities 78.7 per cent (Finance Ministry quarterly debt report, Table 5.2). A 10-year state bond is priced against that. A 30-year state bond is priced against a 30-year central bond that insurers and provident funds also buy to hold to maturity, where the liquidity gap between the two is small, and so is the spread.
What the state adds: a few basis points, priced on size before debt
To isolate what is specific to a state, compare states with each other on the same day. When three or more states sell the same fresh tenor in the same auction, each state's cut-off less the average of the group is its premium for that day, and everything common to the market cancels. There were 76 such groups in 2024-25 and 2025-26, and 24 states appear in at least 5 of them.
The premia run from -4.0 to +6.9 basis points. They do rise with debt: the correlation with liabilities to GSDP is +0.53, with a permutation p-value of 0.008, about 1.8 basis points per 10 points of GSDP. They fall more tightly with the amount a state borrows: the correlation with its logarithm is -0.69 (p = 0.0003). Fitted together, size stays strong (p = 0.0004) and debt shrinks to about 1.2 basis points per 10 points (p = 0.015); the two explain 60 per cent of the differences between states. The debt service burden, interest as a share of revenue receipts, shows no relation at all (r = -0.29, p = 0.17).
Individual states make the pattern concrete. Tamil Nadu, the largest seller of state paper in the window, carries liabilities of 30.2 per cent of GSDP and spends 20.8 per cent of its revenue receipts on interest, one of the heaviest burdens in the country, and it paid 3.5 basis points below the day's average. Sikkim and Puducherry, among the smallest issuers, paid +6.1 and +6.9. Punjab is the exception that shows credit is not ignored entirely: a large issuer with the highest debt ratio among the large states, 46.9 per cent, it paid +4.7.
| State | Comparisons | Premium, bps | Liabilities, % of GSDP, March 2025 | Interest, % of revenue receipts, 2025-26 | Borrowed at auction, ₹ crore |
|---|---|---|---|---|---|
| Gujarat | 7 | -4.0 | 17.8% | 12.4% | 90,330 |
| Tamil Nadu | 32 | -3.5 | 30.2% | 20.8% | 2,92,856 |
| Maharashtra | 10 | -3.0 | 18.6% | 11.5% | 2,60,080 |
| Karnataka | 7 | -2.3 | 25.4% | 15.6% | 1,96,426 |
| Madhya Pradesh | 15 | -1.5 | 30.8% | 9.8% | 1,47,450 |
| Haryana | 16 | -1.4 | 30.2% | 20.5% | 1,02,000 |
| Andhra Pradesh | 19 | -0.6 | 35.5% | 16.1% | 1,60,477 |
| Odisha | 8 | -0.4 | 20.6% | 2.8% | 31,780 |
| Rajasthan | 26 | -0.3 | 37.0% | 13.6% | 1,53,215 |
| Uttar Pradesh | 9 | -0.3 | 30.0% | 9.7% | 1,12,755 |
| Kerala | 7 | -0.2 | 36.2% | 20.9% | 1,03,454 |
| Telangana | 6 | +0.0 | 27.3% | 8.4% | 1,43,049 |
| Goa | 7 | +0.5 | 29.1% | 10.1% | 2,550 |
| Uttarakhand | 5 | +0.6 | 25.3% | 11.2% | 25,609 |
| Himachal Pradesh | 7 | +0.8 | 43.6% | 15.9% | 17,023 |
| Manipur | 6 | +1.3 | 40.7% | 5.6% | 3,500 |
| West Bengal | 11 | +1.3 | 40.0% | 18.3% | 1,63,492 |
| Assam | 16 | +1.4 | 27.7% | 9.4% | 39,513 |
| Bihar | 12 | +1.4 | 38.0% | 8.8% | 99,456 |
| Meghalaya | 6 | +3.4 | 39.8% | 5.3% | 4,769 |
| Nagaland | 6 | +3.8 | 49.4% | 7.1% | 4,150 |
| Punjab | 8 | +4.7 | 46.9% | 22.4% | 85,176 |
| Sikkim | 7 | +6.1 | 36.3% | 8.8% | 4,601 |
| Puducherry | 5 | +6.9 | 26.0% | 7.2% | 2,800 |
So the market does price the state, a little, and what it prices first is how much of the state's paper exists, which is itself a liquidity attribute: a large issuer's bonds are easier to find a buyer for. The debt effect is real and small. From the least to the most indebted large state, Gujarat to Punjab, it is worth about 3.6 basis points, against a common move of about 34 in 2025. That matches the RBI's earlier findings. The 2017 memo found no significant role for deficits or debt in the auctions of 2015-16 and 2016-17, and a larger role for the size of each auction; the 2023 Bulletin study, covering 26 states from 2015 to 2022, found issue size, system liquidity and the pandemic significant and did not include fiscal ratios at all.
Held to maturity, rarely traded
The liquidity gap itself is measurable. In 2025-26 outright trading came to 1.22 times the central stock and 0.14 times the state stock: a rupee of central bonds changed hands about 8.5 times as often as a rupee of state bonds.
| Quarter | Central dated | Treasury bills | State | State share of outright | State share of repo |
|---|---|---|---|---|---|
| Q4 2024-25 | 33,14,878 | 3,93,138 | 3,05,225 | 7.6% | 10.3% |
| Q1 2025-26 | 48,07,259 | 4,38,513 | 2,58,936 | 4.7% | 14.4% |
| Q2 2025-26 | 34,26,701 | 3,50,357 | 2,37,764 | 5.9% | 16.3% |
| Q3 2025-26 | 34,46,293 | 3,73,329 | 2,05,464 | 5.1% | 15.2% |
| Q4 2025-26 | 31,28,440 | 3,79,301 | 2,67,711 | 7.1% | 16.4% |
Three things keep it that way. The holders buy to keep: banks, insurers and provident funds between them own over four fifths of state securities, much of it held to maturity, and the RBI's 2023 study traces the market's illiquidity to fragmented issuance and large held-to-maturity books. The paper is fragmented: states created 833 new securities in 2025-26 alone, while the entire central stock at end-March 2026 was 122 securities, the largest with ₹2,41,038 crore outstanding, against a typical state issue of ₹1,000 crore. And state bonds are used as collateral more than they are traded: they were 16.4 per cent of repo volume in January to March 2026 against 7.1 per cent of outright volume.
A buyer who may need to sell before maturity pays for that. The bid is thinner, the wait for the other side longer and the price concession larger, and a bond that is expensive to leave has to offer more to be bought. The spread is that cost, paid in yield from the first day. A holder who keeps the bond to maturity collects it without ever paying the exit cost; a holder who sells early hands part of it back.
Stress and supply move the common part
The same instrument run on the RBI's releases from December 2019 to March 2021 shows how the common part behaves under strain. The RBI held no central auction between the end of January and 9 April 2020, so March 2020 itself cannot be matched; what can be seen is the months either side.
| Period | 10-year segment spread | Sold per auction, ₹ crore | Inter-state range |
|---|---|---|---|
| December 2019 to February 2020 | 59 bps (55) | 12,351 | 4.4 bps |
| March 2020 | no central auction | 18,023 | 22.3 bps |
| April to June 2020 | 75 bps (39) | 12,867 | 10.5 bps |
| July to September 2020 | 61 bps (59) | 14,564 | 7.6 bps |
| October to December 2020 | 70 bps (44) | 15,340 | 12.7 bps |
| January to March 2021 | 92 bps (61) | 18,893 | 9.3 bps |
In the first pandemic quarter the 10-year segment rose from 59 to 75 basis points, and the inter-state range jumped to 22 basis points in March 2020, so stress raises differentiation as well as the common premium. But the widest quarter of the year was January to March 2021, at 92, when states sold ₹18,893 crore per auction, the most of any period in the table. From October 2020 the RBI bought state bonds in the open market for the first time, ₹30,000 crore in three tranches, and the spread held near 70. Over that year the states' liabilities rose from 26.6 to 31.0 per cent of GDP, the largest one-year rise in the RBI's series back to 2008. The spread's peak did not arrive with the worst fiscal news. It arrived with the heaviest supply. The RBI's 2023 study reaches the same place by regression: larger auctions widen the spread, system-wide liquidity conditions move it, and the pandemic added about 11 basis points.
The benchmark issuance strategy, and what it has changed so far
If the spread is mostly the cost of fragmented, thinly traded paper, the direct remedy is fewer, larger securities, and that is what the RBI has now started to build. Its quarterly state borrowing calendar of 2 April 2026 (press release 2026-2027/18) introduced a Benchmark Issuance Strategy on a pilot basis from 2026-27 for nine states: Andhra Pradesh, Bihar, Chhattisgarh, Kerala, Madhya Pradesh, Maharashtra, Rajasthan, Telangana and Uttar Pradesh. They borrow in specific benchmark tenor buckets on a pre-announced calendar. The calendar of 25 June 2026 (press release 2026-2027/551) extended it from July to Delhi, Himachal Pradesh, Jharkhand, Manipur, Meghalaya, Odisha, Punjab, Sikkim, Uttarakhand and West Bengal, so that 18 states and Delhi now issue this way. The RBI's own Annual Report for 2025-26 lists persuading the remaining states as an agenda item. Any account of state bonds written before April 2026 describes a market of one-off issues; from April to September 2026 most of the money was raised by re-opening existing bonds.
The change in issuance is already large. Re-issues of existing securities were 12 per cent of the state securities sold in 2024-25 and 21 per cent in 2025-26; from April to September 2026 they were 62 per cent by number and 65 per cent by amount. Fresh 10-year issues by two or more states on the same day, which happened on 31 auction days in 2025-26, happened on 3 in the first half of 2026-27.
The change in the spread is not yet visible. Across all states the 10-year segment averaged 74 basis points from April to September 2026, against 70 in 2025-26. Among matched securities, the pilot states' average moved by -3.2 basis points while states outside the strategy moved by +8.0, a difference in the direction the mechanism predicts, but from small samples with a changed maturity mix over less than six months.
| Group | 2025-26 | April to September 2026 | Change, bps |
|---|---|---|---|
| Nine pilot states (strategy from April 2026) | 57.3 bps (80) | 54.2 bps (40) | -3.2 |
| Ten jurisdictions added from July 2026 | 71.9 bps (29) | 74.2 bps (25) | +2.3 |
| States outside the strategy | 54.3 bps (118) | 62.3 bps (29) | +8.0 |
The mechanism predicts that a benchmark shows up first in trading, as more turnover and tighter quotes in the re-opened securities, and only then in the price at auction. The Finance Ministry's trading figures for April to June 2026 were not yet published when this was written, so that test is still to come.
What the spread is, and is not, a signal of
The spread of a state bond over a central bond of the same maturity is, first, the price of being harder to sell than the most liquid bonds in the market; second, a reading of conditions common to all state paper, supply above all, then the appetite of the long-term holders and the level of stress; and third, a small state-specific tilt that follows issue size before debt. It is not a default probability, not a ranking of state finances and not a verdict on any one state. A widening across all states says something about the market for state paper. A state's fiscal position is better read from its own budget ratios.
The common ways of misreading it follow from that. Quoting the 10-year spread as the spread picks the widest point on the curve. Comparing a state cut-off with a central yield from another day, or another maturity, mixes a market move or a curve slope into the answer. Treating the implicit guarantee as an explicit one confuses a well-supported belief with a legal claim. And a figure that reflects mostly liquidity carries a different lesson for a holder who will keep the bond to maturity, who earns it without paying the exit cost, than for one who may need to sell. Reading a number for what it is made of, rather than for what it is called, is the discipline that separates analysis from a label, and it is the part of fixed income the curriculum spends its time on.
Frequently asked questions
What is a state development loan?
A dated bond issued by a state government on the security of its own Consolidated Fund, under Article 293(1) of the Constitution. The RBI manages the debt as the state's banker and runs a separate auction for each state security, usually on a Tuesday, with payment the next working day. It pays a fixed coupon every six months, repays at par and is governed by the Government Securities Act 2006, as a central bond is.
Does the central government guarantee state development loans?
Not in law. Article 293(2) allows the Government of India to guarantee a state's loans, but nothing in the terms of these bonds invokes it, and the RBI's own research calls the guarantee investors assume implicit. What exists instead is a set of written supports: the centre's consent to each state's borrowing under Article 293(3), the RBI as banker and debt manager, the states' sinking and guarantee funds at the RBI, capped short-term advances, and a zero risk weight for banks. They make a missed payment mechanically hard in the short run. They are not a promise by the centre to pay a state's bondholders.
Why do state bonds yield more than central bonds of the same maturity?
Mostly because they are much harder to sell. At end-March 2026 state securities were 35 per cent of the government bond stock, yet in 2025-26 they were 5.6 per cent of outright trading, and a rupee of central bonds changed hands about 8.5 times as often as a rupee of state bonds. The spread is widest beside the 10-year central benchmark, the most traded bond in the market, and narrowest beyond 20 years, which is the shape of a liquidity premium, not a credit premium.
How large is the spread?
The RBI puts the weighted average spread of state issuance over comparable central securities at 30 basis points in 2024-25 and 50 in 2025-26. Measured from the auction results it depends heavily on maturity: about 70 basis points for 7 to 12-year paper in 2025-26 and about 15 beyond 20 years. From April to September 2026 the 10-year segment averaged about 74.
Do states with more debt pay more?
A little. Comparing states that sold the same tenor in the same auction, the premium rises by about 1.2 basis points for every 10 percentage points of debt to GSDP once the size of the state's borrowing is allowed for. Size matters more: about 2.4 basis points less for every tenfold increase in what a state borrows. Across the 24 states measured, the whole range was about 11 basis points.
Why did the spread widen in 2025-26?
The auction data show when, not why. The 10-year segment rose from about 42 basis points in April to June 2025 to about 71 in July to September, for every state at once, while the states' combined debt ratio moved by 0.8 of a percentage point over the year. Supply per auction does not explain the quarterly pattern in this period. A cause common to all state paper fits the data; a change in any one state's finances does not.
What is the Benchmark Issuance Strategy?
An RBI framework, piloted from April 2026 with nine states and extended from July 2026 to nine more states and Delhi, under which a state borrows in specific benchmark tenor buckets on a pre-announced calendar, largely by re-opening existing securities rather than creating new ones. Re-issues rose from 12 per cent of state securities sold in 2024-25 to 62 per cent from April to September 2026. Its effect on the spread is not yet measurable.
Can an individual buy state development loans?
Yes. Up to 10 per cent of each state security in an auction is reserved for non-competitive bids, which individuals can place through the RBI's Retail Direct portal and which are allotted at the auction's weighted average price. State bonds also trade after issue, but thinly: a holder who may need to sell before maturity should expect the exit to cost more than it would for a central bond.
Does a wider spread mean a state has become riskier?
Not by itself. Most of any move is common to every state, so it is information about the market for state paper: supply, the appetite of the long-term holders, and stress. The part that differs between states is a few basis points and follows issue size more than debt. A state's fiscal position is better read from its budget ratios directly, such as liabilities to GSDP and interest to revenue receipts.
Where can the numbers be checked?
The RBI publishes a full result for every state auction, usually each Tuesday, with each security's notified amount, bids, cut-off yield and weighted average yield, and a full result for every central auction on Fridays. Pairing each state security with a central security of similar maturity sold within a week reproduces the RBI's annual spread figures to within 3 basis points once the pairs are weighted to the year's maturity mix.
Position stated as at 23 September 2026. The RBI changes auction terms, state borrowing calendars, the ways and means scheme and the reach of the Benchmark Issuance Strategy by press release, and bank capital treatment by direction. Check the RBI's current releases, the latest quarterly calendar and the current capital adequacy directions before relying on anything here.
How the figures were produced. Every state and central auction result and auction notice the RBI published as a press release from April 2024 to September 2026 (130 weekly state results, 2 April 2024 to 22 September 2026) and from December 2019 to March 2021 was fetched from www.rbi.org.in and parsed; the fetcher is _workspace/marketdata/a164-evidence/a164-fetch-rbi.py, which stores the release text, and every figure is computed in tools/build-article-164.py. State securities: cut-off yield as printed; remaining maturity from the tenor, or from a re-issued security's original issue date, with settlement from the auction notice; 590 of 591 re-issue price and yield pairs reproduce on the half-yearly 30/360 convention. Central securities: cut-off yield as printed, maturity from the announcing notice or the Finance Ministry's list of securities outstanding at end-March 2026, used only if the printed price reproduces the printed yield (to the fourth decimal, or within one basis point for a bond in its first coupon period) or, for a new bond with no price, if its cut-off equals the coupon it later carries and sits within 15 basis points of its weighted average yield; 3 results from 2020 failed and were excluded, none from 2024 to 2026. Matching: the nearest remaining maturity within one year among central securities auctioned within seven days either side, averaged when both sides qualify. Averages are weighted by amount allotted; yearly spreads are re-weighted over four maturity buckets (under 7, 7 to 12, 12 to 20, and 20 years and over) in proportion to that year's issuance. Inter-state range: highest less lowest cut-off among states selling fresh 10-year paper on the same day, averaged over days. State premium: same day, same fresh tenor, three or more states, each state's deviation from the group mean, averaged by state; the 24 states with at least 5 comparisons are shown (with at least 3 there are 28, and the correlations with debt and with size are +0.51 and -0.69). Permutation p-values use 100,000 random re-pairings with seed 20260923; the test for one variable holding the other shuffles that variable whole, seed 20260924. Turnover is a year's outright volume over the average of five quarter-end stocks. Nothing here is simulated.
Not verified this session. The relevant paragraph of the RBI's 2025 capital adequacy directions for commercial banks was not read; the zero risk weight is taken from the 2013 Basel III master circular and from the 2025 concentration risk directions' reference to state exposures eligible for it. The RBI does not publish the method behind its weighted average and inter-state spreads; the figures here reproduce them closely, which is not proof of the same method. The auction data do not identify the cause of the 2025-26 widening. Whether any state development loan has ever missed a payment was not checked and is not asserted. The RBI's document server refused automated downloads, so the Annual Report and State Finances were read in their HTML editions on www.rbi.org.in. That state bonds are eligible collateral at the RBI's liquidity window is taken from the RBI's primer dated 1 April 2020. The Finance Ministry's trading figures for April to June 2026 were not yet published. RBI staff studies carry their authors' views, not necessarily the RBI's.
Bharath Shiksha is an educational publisher and not a SEBI-registered investment adviser or research analyst. Nothing on this page is a recommendation to buy, sell or hold any security. State names identify issuers in published data, and every figure describes past auctions.
Ready to go deeper than this article?
Bharath Shiksha is a 90-volume curriculum across 6 stages, from chart reading at ₹14,999 through capital raising, or the full bundle at ₹1,49,999. Reading a spread for what it is made of, liquidity, supply and credit in their measured proportions, is taught as method across the fixed income volumes, never as a view on an issuer.
Take the free diagnostic →