Debt’s Second Bill

What India’s and Japan’s published bonds reveal about April 2026–March 2027

Evidence cut-off: 30 September 2026. The comparison follows the securities outstanding on 31 March 2026 over the subsequent fiscal year.

At the end of March 2026, India’s debt manager listed ₹12,183,645 crore of dated central-government securities. Of that published pool, ₹435,820 crore was due to mature by the end of March 2027: 3.58%. Japan’s finance ministry listed ¥1,060.2995 trillion of General Bonds after removing fiscal-expenditure Treasury Bills. Of that pool, ¥103.735 trillion was due over the same year: 9.78%. These are shares of each government’s specified securities, not percentages of national debt or predictions of financial trouble.[1][2]

The obvious objection is that a bond can become more expensive before it matures. A floating coupon may reset while its principal stays outstanding. That objection is particularly useful here because India’s published list names five floating-rate bonds. The four whose coupons could plausibly reset inside the year total ₹397,014 crore; we count every rupee of them as a possible reset. Adding that entire stock to maturities, without counting any principal twice, puts India’s one-year maturity-or-possible-reset share at no more than 6.84% of its dated-security pool. Japan’s 9.78% is maturities alone. Any qualifying Japanese coupon reset would raise its corresponding union, not lower it.[1][2][3]

The difference is 2.95 percentage points in this bill-excluded comparison. It survives a parallel test that adds each issuer’s published market bills: India’s upper share becomes 12.18% and Japan’s scheduled-maturity share 13.38%. The remaining gap is 1.20 points.[1][2] The result answers a narrow question: in these two named opening stocks, how much principal comes due or might have its nominal coupon reset between 1 April 2026 and 31 March 2027? It does not tell us which country has more fiscal room, which government will pay more interest, or which bond is safer.

A coupon can move before the bond comes due

A fixed-rate bond commits its issuer to the stated coupon until redemption. Higher market rates make refinancing more expensive than it would otherwise have been as the old principal falls due; the new coupon need not exceed the old one. A floating-rate bond can transmit a rate change sooner through its coupon. These events are different: a coupon reset is not a requirement to repay principal. To compare their timing without adding rupees to yen, we measure each as a share of its own issuer pool and count a bond once if it both matures and resets within the year.

India’s 122 listed dated issues consist of 109 fixed-coupon government-security labels, eight fixed-coupon sovereign green bond labels and five named floating-rate bonds. The selected list names no inflation-indexed bond. That is a classification of this finite published schedule; it is not a reading of every individual legal deed or a claim about all Indian public liabilities.[1][3] All five floating bonds survive beyond March 2027. One of them, FRB 2035, has a documented coupon period lasting until January 2030. Its ₹350 crore therefore cannot be a next-year reset under the stated terms. The other four total ₹397,014 crore. Their current individual reset dates were not established, so counting all four deliberately overstates or equals the possible next-year reset stock.[3]

The Japanese issue register includes retail floating-rate bonds and inflation-indexed bonds. We have not turned either into a Japanese reset point estimate. Inflation-linked principal changes with a price index; it is not the same as a floating nominal coupon. For the ordering tested here, the Japanese maturity figure alone is the lower bound. More complete reset dates would improve an estimate of Japan’s union, but cannot make its verified maturity share smaller.[2][4]

Each row uses its own opening-stock denominator. Japanese maturity shares are lower bounds for a maturity-or-reset measure. The bill variants are parallel sensitivities, not harmonised sovereign balance sheets. India’s pools exclude nonmarketable and external debt and other liabilities. Japan’s General Bonds exclude Fiscal Investment and Loan Program (FILP) Bonds, Financing Bills and other liabilities. Values are rounded from the audited issue and schedule figures; details and source discrepancies are in the accompanying calculation notes.[1][2]

Bills matter because their short original terms bring opening-stock principal back within a year. India’s market Treasury Bills are issued at 91, 182 or 364 days; the ₹741,709 crore in the March stock is therefore included once in the bill-added test. Japan’s workbook lists ¥44 trillion of fiscal-expenditure TBs, all redeeming inside the year. India’s 14-day intermediate bills are a separate nonmarketable category and remain outside the selected pool. Neither panel adds future deficits or repeatedly counts the bills that might be issued when old bills roll over.[1][2][5] Mixing India’s bill-included row with Japan’s bill-excluded row would answer no consistent version of this test.

Even an intentionally severe Indian classification check leaves the bill-added ordering intact. Count all five floating bonds, including FRB 2035; assign the ₹4 crore gap between India’s displayed issue rows and printed schedule total to immediate exposure; then assign the separate ₹15 crore difference between the quarterly report’s provisional dated balance and its maturity-table stock the same way. India’s upper share reaches 12.18% at two decimal places, still below Japan’s 13.38% maturity share in the bill-included pool.[1][2] Those ₹4 crore and ₹15 crore differences have no verified explanation in the retained source and are disclosed, not quietly treated as rounding. Japan’s issue sums also differ slightly from its displayed totals; its workbook explicitly warns about rounding.[1][2]

A schedule is only one of the bills

Suppose both governments had the same maturity share. The budget consequences could still differ. The new coupon on refinanced debt depends on the rate at issuance and the amount actually refinanced. Existing floating bonds have their own reset rules. Government revenue, other spending and fresh borrowing matter too. A maturity share is therefore neither a forecast of next year’s cash interest nor a measure of a rate shock’s net budget effect. India’s own older debt status paper already publishes a re-fixing measure, and BIS analysis separately considers short-rate, inflation and foreign-currency exposures. The dated reconstruction here shows how far the named resets and bills can move this particular comparison.[5][6]

The published FY2026–27 budget plans provide context on a different measurement boundary. India’s Union Budget Estimates set net interest payments at ₹1,403,972 crore and revenue receipts at ₹3,533,150 crore, or 39.74%. Receipts include tax net to the Centre after states’ share and non-tax revenue. Japan’s FY2026 General Account draft framework puts interest payment and discount expenses at ¥13,037.1 billion against ¥92,725.2 billion of tax plus other revenue, or 14.06%. The later Debt Management Report confirms the interest figure as an initial-budget amount; the draft revenue rows have not been reconciled here to the enacted budget.[7][8]

The two ratios describe their stated revenue plans, not realised-year outcomes. Japan’s numerator includes JGB interest, borrowing interest and Financing Bills discount; India’s net appropriation also extends beyond the dated bonds in the table. Japan’s published June supplementary framework adds spending and bond financing. The accounts and interest categories differ, so these ratios cannot establish an effective rate on the selected bonds or show that either maturity schedule caused the budget difference.[7][8]

Currency raises another distinct question. The selected Indian dated securities and market bills are described as domestic rupee market borrowing; the selected Japanese bond classes have yen face and payment terms. The Japanese list also includes CPI-linked yen obligations, which should not be confused with foreign-currency debt. These are class and issue-label findings, not a matched measure of either government’s entire currency exposure.[1][4][9] The distinction between foreign creditors and foreign currency matters in India: 26.5% of its sovereign external debt at end-March 2025 was reported as rupee-denominated. That older, narrower stock cannot be inserted into a March 2026 India–Japan currency ranking. A matched whole-central-government foreign-currency share has not been established here.[10]

There is also a public-sector countercase to the apparent comfort of a longer issuer schedule. A central bank can hold a long bond while funding part of its balance sheet with liabilities whose interest rates move sooner. Changes in its income and remittances can then matter to the treasury. This mechanism is well established in BIS analysis; long bonds held outside the central bank can instead expose private investors to price losses when yields rise.[11]

The country records show why the countercase matters, but cannot quantify it for our one-year sets. India’s debt manager says the RBI held 17.59% of GoI dated securities at end-March 2026. It does not allocate those holdings to the issues maturing or possibly resetting in the next year, or expressly state the table’s valuation basis. RBI’s broader domestic-investment account mixes central-government, state and other securities and uses accounting valuations. Its standing deposit facility pays interest on eligible deposits, while the current RBI Direction says the cash reserve ratio (CRR) balances maintained by scheduled commercial banks earn no interest.[12][13]

Japan’s central bank lists ¥527.6805 trillion face value of interest-bearing JGB holdings at the same date, but the issue file does not separate General Bonds from FILP Bonds. Its eligible excess reserves earned interest, and part of the BOJ’s net income was remitted to government.[14] Neither country’s totals identify which one-year securities sit with its central bank or how a rate change would alter net fiscal cost. Different liability pools, valuation bases and whole-bank income flows cannot be subtracted from the face-value issuer schedule.

In the published March 2026 pools, India’s maturity share plus every named plausible nominal coupon reset remains below Japan’s maturity share. Adding each pool’s specified market bills in parallel, then pushing the small disclosed Indian source differences against the result, does not reverse the ordering. A different government perimeter could; that comparison remains open. So does the question of who ultimately bears a rate move. Answering it would require compatible issue-level central-bank holdings, actual reset calendars, the funding liabilities behind purchases and comparable budget flows. The reconstruction bounds next-year maturities and possible coupon resets for named issuer contracts. It cannot settle the second bill’s final payer.