RBI's June 2026 Financial Stability Report (FSR) is an unusually transparent piece of central-bank writing - one that discloses its assumptions, names risks, and resists temptation of mistaking a rising market for a healthy one.
Household debt: It has risen to 45.5% of GDP, up four percentage points in a year, and the report takes comfort in the fact that borrowers are better rated than before. That is true and worth noting, but it answers only who is borrowing, not why.
On the report's own figures, the answer is increasingly consumption, which now accounts for nearly half of all household debt. A reader might wonder whether this reflects financial deepening or households using credit to manage flat incomes. The report embraces the first interpretation and leaves the second to be inferred.
Asset quality: A multi-decadal low of 1.8% in bad loans is a genuine achievement, broad-based across bank groups. It coexists, however, with a farm portfolio carrying a 5.1% bad-loan ratio, and more than a third of the system's impaired assets. The report's data reveal a similar pattern at the other end of the income scale, where lower-income borrowers account for a disproportionate share of fresh defaults even as the retail book improves overall. None of this undermines the headline figure, which is real. It suggests that a system-wide average, however encouraging, is best read alongside the dispersion beneath it.
Gold lending: Loans against household jewellery are growing at roughly 42% a year, reaching $54 bn and 17.4% of non-bank lenders' retail portfolios. RBI attributes the trend to rising gold prices, which have strengthened collateral buffers, while noting that a price correction could weaken them. Both points are true, and together they underscore that safety and risk are two sides of the same variable.
The report's data also show the growth concentrated among existing borrowers re-pledging gold to roll over debt, a pattern consistent with sensible refinancing or with mounting borrower strain. The two are difficult to distinguish without loan-level repayment data, which RBI has not yet published.
A similar dynamic appears in fintech small-ticket lending, where non-bank platforms have taken a majority share of a fast-growing, largely unsecured segment with delinquencies above 6%. Both would benefit from fuller treatment in future editions.
Funding: Credit growth has accelerated to 14.5%, led by state-owned banks, even as deposit growth lags - and RBI traces this to savers shifting from FDs into MF industry.
Connecting these two observations more explicitly would sharpen an otherwise valuable point. The report is similarly even-handed, if somewhat reticent, on why public-sector lenders are setting the pace. Readers are left to weigh scale and improved underwriting against the possibility that an implicit sovereign backstop permits greater risk-taking, without the report itself adjudicating between the two.
Stress-testing framework: This is the report's strongest feature. Assumptions are disclosed in full, many of them conservative - provisioning is front-loaded, dividends are assumed to continue through the modelled stress - and a reverse stress test usefully shows that a shock of 4.4 standard deviations would be needed to breach the capital floor.
This is a welcome benchmark. It would sit more prominently, and do more interpretive work, if placed alongside the headline 'severe' scenarios, which are not otherwise calibrated against a probability or a historical comparative number.
Where the report does disclose individual outcomes, the picture is a shade less uniform than the aggregate suggests: 4 banks holding 12% of system assets, and 15 of 174 NBFCs fall short of their minima under the severest scenario - a detail that rewards attention alongside the reassuring average.
The contagion analysis, which assumes bilateral exposures remain unchanged over a 2-yr horizon, finds limited knock-on losses, a natural extension, consistent with the report's emphasis on rising interconnectedness, would be to model the funding-side contagion that characterised the ILFS episode, where the transmission ran through withdrawn credit lines rather than direct losses.
Cybersecurity risk: Ranked by RBI's respondents as the leading concern for the year ahead - readers would benefit from the same quantitative treatment the report applies so thoroughly elsewhere. At present, it is flagged but not modelled.
Insurance Life-insurance grievances have declined by a fifth over 5 yrs, which the report reads as improved conduct.
General-insurance grievances have nearly tripled over the same period, and the report's own language here is notably direct - 'systemic shortfalls in claims management, service quality and product communication'. Having used that phrase, the document moves on without returning to it in the Overview, where a reader might expect a finding of that weight to surface.
None of this detracts from the report's core achievement. The arithmetic is sound, disclosure genuine, and RBI deserves credit for publishing material that allows this kind of scrutiny.
The suggestion, offered in the spirit the report invites, is modest: a published severity benchmark alongside the reverse stress test already calculated, a funding-side extension to the contagion model, and a touch more prominence in the Overview for findings - on gold, on fintech, and on general-insurance conduct - that the data support rather more strongly than the summary lets on.
A report this forthcoming has little to lose, and a good deal to gain, from carrying its own candour through to the front page.
