// analysis
When gold rises, so does the debt
Higher gold prices let Indian households borrow more. From jewellery boxes to bank balance sheets, anatomy of a heavily collateralised credit boom.
On 20 August 2026, Aditya Birla Capital announced its entry into India’s gold-loan market. The group plans to open 200 to 300 dedicated branches by March 2027 and roughly 1,000 within three years. This is no longer a peripheral activity confined to a handful of pawnbroking specialists. One of India’s largest financial groups is building a nationwide network to turn household jewellery into liquidity.
The announcement gives a concrete shape to a shift already visible in the data.
Since March 2024, loans backed by gold have grown at a 42.4% compound annual rate, almost twice as fast as non-housing retail credit overall. The Reserve Bank of India, or RBI, now ranks gold loans as the largest segment in that category. More importantly, it identifies an unusual engine: growth is coming mainly from existing customers using higher gold prices to obtain larger loans and roll over debt. (Aditya Birla Group, 20 August 2026, RBI, Financial Stability Report, 30 June 2026)
The mechanism fits into one sentence: the collateral becomes richer on paper; the borrower does not necessarily do so.
A rise in the metal mechanically reduces debt relative to the jewellery’s value. It makes the loan safer for the lender while releasing additional borrowing capacity for the household. Income has not increased. Repayment capacity may be unchanged. Yet the same chain or bracelet supports a larger loan.
Gold is therefore doing more than preserving precautionary savings. It is becoming a credit line recharged by the market.
A higher gold price opens a larger credit line
Take jewellery valued at 100 and a loan of 60.
The loan-to-value ratio, or LTV, is 60%. If gold rises by 20%, the jewellery is now worth 120 while the debt remains 60. LTV falls to 50%. Keeping the lender’s internal 60% ratio would allow principal to rise to 72. The household extracts another 12 of liquidity without pledging one more gram and without earning more income.
The regulatory maximum can create more headroom, but it should not be confused with the amount actually disbursed. Each lender applies its own risk policy. For a loan repaid in one bullet at maturity, future interest also enters the regulatory calculation.
That separation makes the cycle procyclical:
Gold price rises
↓
LTV falls mechanically
↓
Borrowing capacity increases
↓
New loan or rollover
↓
Household debt increasesThe first step objectively strengthens the collateral. The last increases the household’s exposure. Both can happen at the same time.
RBI has identified the accelerator
The most important number is not 42.4%. It is the central bank’s explanation for it.
RBI compares fresh originations with loans still outstanding. The gap indicates that expansion is primarily being driven by existing borrowers. They are using higher metal prices to obtain more credit and to renew maturing debt. The pattern is especially visible at NBFCs, RBI-regulated non-bank finance companies, where originations far exceed those at public and private banks. (RBI, paragraphs 1.71 to 1.75)
Rollover is not automatically a concealed default. A loan can be properly settled and replaced by another. The analytical issue lies elsewhere: the old contract disappears while the household’s economic debt can persist or increase.
Manappuram Finance’s commercial documentation illustrates the mechanics unusually clearly. In its online gold-loan service, a customer seeking a higher amount may take a new loan. The system uses part of the new proceeds to close the old account and credits the difference to the customer’s bank account. Only one loan remains officially live. The jewellery has never left the branch. (Manappuram Finance, Online Gold Loan FAQ)
Since RBI’s new framework took effect on 1 April 2026, renewal of a bullet-repayment loan requires accrued interest to be paid first. That prevents interest from being rolled indefinitely without control. It does not stop principal refinancing, or the extraction of an additional amount when LTV allows it.
The boom therefore looks less like a mass arrival of new borrowers than an increase in the financial power of the same collateral.
₹8.4 trillion of retail gold loans
The World Gold Council estimates that, at the end of May 2026, Indian banks held ₹5.1 trillion of retail gold loans, about US$54 billion, up 105% year on year. NBFCs held ₹3.3 trillion, or US$34.5 billion, up 70%. Adding the two published series gives approximately ₹8.4 trillion, or US$88.5 billion. That total is our arithmetic, not a separate RBI aggregate. (World Gold Council, India Focus Q2 2026)
The World Gold Council calls gold loans the second-largest retail segment after housing. RBI calls them the largest non-housing retail segment. The descriptions are consistent.
The potential collateral base is enormous. The World Gold Council estimates that Indian households hold roughly 31,000 tonnes of gold, worth US$3.4 trillion. This is a stock estimate rather than a census, and not all that gold can enter the formal lending system. It nevertheless conveys the scale of the dormant asset on which lenders are trying to build. (World Gold Council, Swarnim Udaan 2047)
The trend sits within a broader increase in household leverage. Household debt reached 45.5% of GDP at the end of September 2025. Non-housing retail loans represented 58.4% of household borrowing in March 2026, while consumption loans accounted for nearly half of the total. Those ratios do not establish a debt crisis. They show that gold lending is growing in an economy where household credit is already shifting towards non-housing uses. (RBI, household sector)
The new rule protects the lender and releases capacity
RBI harmonised the regime for loans against gold and silver in June 2025. Commercial banks, co-operative banks and NBFCs had to comply no later than 1 April 2026.
For consumption loans, the ceiling depends on the customer’s total loan amount:
| Total loan amount | Maximum regulatory LTV | Maximum principal at 18% for 12 months* |
|---|---|---|
| Up to ₹2.5 lakh | 85% | 72.0% of jewellery value |
| Above ₹2.5 lakh and up to ₹5 lakh | 80% | 67.8% |
| Above ₹5 lakh | 75% | 63.6% |
* l0g calculation using simple interest, unchanged gold price, no fees or interim payment. One lakh equals 100,000 rupees.
The third column exposes a crucial detail. For a bullet loan, in which principal and interest are paid at maturity, LTV is not calculated only on the cash disbursed on day one. RBI requires lenders to use the total amount payable at maturity.
At 18% simple interest over twelve months, principal equal to 80% of the jewellery would become a claim equal to 94.4% before gold moved at all. To comply with an 80% ceiling, initial principal cannot mechanically exceed 67.8%, unless interest is paid along the way or a different repayment structure is used.
The framework adds several safeguards:
- the reference price is the lower of the previous day’s close and the preceding 30-day average;
- only the intrinsic metal value counts, excluding stones, workmanship and sentimental value;
- purity and net weight must be documented in the borrower’s presence;
- LTV must remain compliant throughout the loan;
- a consumption bullet loan is capped at twelve months;
- above ₹2.5 lakh, the lender must assess repayment capacity;
- renewal or top-up is permitted only for a standard loan, within LTV limits, and after accrued interest is paid on a bullet loan. (RBI, Lending Against Gold and Silver Collateral Directions, 2025)
The framework reduces some conduct risks. It may also support volume because small loans can now reach 80% or 85%, instead of the former uniform 75% ceiling. In June 2025, Crisil estimated that close to 70% of NBFC gold-loan portfolios had ticket sizes below ₹5 lakh and that initial principal LTV on small bullet loans could rise towards 70% to 75%, even after future interest was included. (Crisil Ratings, 13 June 2025)
The rule therefore protects the lender from a silent build-up of unpaid interest. At the same time, higher ceilings on small tickets allow more credit against the same gold.
Why the balance sheets still look clean
Growth of 42.4% looks like a warning. Current risk data do not describe a banking crisis.
The World Gold Council reports portfolio LTVs of roughly 55% at banks and 60% at NBFCs in March 2026. Higher gold prices pushed those ratios down even as balances expanded. Average collateral therefore covers loans by a wide margin. RBI also observes improving consumer-loan asset quality overall. Across the whole NBFC sector, not gold loans alone, the net non-performing asset ratio has fallen to 0.8%. (World Gold Council, Q2 2026, RBI, NBFC sector)
Crisil studied twelve-month bullet loans at several specialist NBFCs. In its sample, about 90% were repaid by the end of the term, largely because of pre-closures. Of the amounts still unpaid at contractual maturity, more than three quarters were settled before auction. Typically, less than 3% of disbursed loans ended in a sale of the pledged jewellery. The study does not cover every Indian lender and is not a comprehensive regulatory statistic. It nevertheless helps explain why historical lender losses have been very low. (Crisil Ratings, 24 June 2026)
The structure is robust for four reasons:
- credit generally starts below the regulatory ceiling;
- gold can be marked to market frequently;
- the lender can act before collateral becomes insufficient;
- the asset is liquid and saleable, unlike specialised machinery or a disputed trade receivable.
That strength does not undermine the article’s thesis. It makes it more interesting. A mechanism can increase household debt while improving lenders’ collateral ratios at the same time.
The jewellery stays in the vault; the receivable travels
A gold loan looks local: one customer, one branch and one piece of jewellery in a vault. Its funding does not necessarily stop there.
Between April and June 2026, India’s securitisation market reached roughly ₹600 billion, up 22% year on year, according to Crisil. Gold loans accounted for 31%, ahead of vehicle loans at 26%. They became the largest asset class of the quarter in transactions tracked by the rating agency. (Crisil Ratings, 6 July 2026)
“Securitisation” covers two routes in Crisil’s data.
The first uses certificates backed by a pool of loans. The second, overwhelmingly dominant for gold, is direct assignment: the lender sells a pool of receivables to an investor without necessarily issuing a tradeable security through a vehicle. Crisil includes both forms in its structured-finance perimeter. During the quarter, 87% of transferred gold loans used direct assignment.
Banks invested in 90% of issuances across all asset classes. Crisil says public-sector banks are particularly attracted to gold loans because of negligible historical credit losses and capital-treatment benefits.
The necklace therefore does not have to cross India and enter another bank’s vault. What travels is the receivable: the right to collect principal and interest, supported by jewellery held and administered within the lender’s network.
RBI rules prohibit lenders from re-pledging customers’ gold to borrow themselves. They explicitly allow one lender to finance another against the underlying receivables. The boundary is clear: the same physical gold should not become collateral at several material layers, but the credit it secures can circulate through the financial system.
That plumbing creates a useful loop for NBFCs:
Household jewellery
↓
Loans originated by NBFC
↓
Receivables sold to a bank
↓
Cash returned to NBFC
↓
More gold loans originatedThe loop does not automatically create excessive leverage. It does turn a very old form of family savings into feedstock for modern balance-sheet funding.
The reverse scenario
When gold rises, LTV improves without a repayment. When gold falls, the same mechanism runs backwards.
For debt of 60 against jewellery initially worth 100, before interest:
| Gold move | Collateral value | LTV |
|---|---|---|
| 0% | 100 | 60.0% |
| -10% | 90 | 66.7% |
| -20% | 80 | 75.0% |
| -30% | 70 | 85.7% |
| -40% | 60 | 100.0% |
Interest makes the table harsher. At 18% for twelve months, debt of 60 becomes 70.8 at maturity. With gold unchanged, LTV already reaches 70.8%. After a 20% decline, it rises to 88.5%.
Crisil searched for the worst decline over rolling 90-day windows in twenty-five years of daily rupee gold prices. The maximum in its study was close to 20%, while declines above 10% occurred in roughly 2% of windows. That is neither an absolute boundary nor a forecast. The agency used the history to test how quickly lenders could call, collect or auction collateral. (Crisil Ratings, 24 June 2026)
Test the mechanism
The calculator separates principal from the amount due at maturity, applies a shock to collateral and measures the repayment needed to restore the selected cap. It does not model fees, interest rebates or the real recovery timetable.
// l0g tool
Stress-test a gold-backed loan
Change the gold price, rate and term. The model separates today’s principal from the amount due at maturity because RBI rules include accrued interest in the LTV of bullet-repayment loans.
Mechanical scenario, not a forecast. Simple interest; no fees, interim repayments, interest rebates, auction discount or early lender action. One lakh equals 100,000 rupees. No entered data is stored.
Method, formulas and sources
The amount due at maturity is principal multiplied by 1 + annual rate × term/12. Shocked collateral is the initial value multiplied by 1 + the price shock. Final LTV divides the amount due by that value. Maximum principal is calculated so principal plus interest stays below the selected cap if gold is unchanged.
- Maturity debt = principal × (1 + annual rate × months / 12)
- Post-shock LTV = maturity debt ÷ shocked jewellery value
- Maximum principal = collateral × LTV cap ÷ interest factor
- Cap-restoring repayment = debt - shocked collateral × cap
- Reserve Bank of India, Lending Against Gold and Silver Collateral Directions, 2025
- Reserve Bank of India, Financial Stability Report, 30 June 2026
- Crisil Ratings, gold-loan price-correction stress test, 24 June 2026
Model v1.0.0 · 2026-08-26
The household absorbs the risk before the bank
A prudent LTV and a timely auction can protect the financier very effectively. That protection does not eliminate the economic cost. It transfers it to the owner of the jewellery.
If the loan is not settled, the lender can sell the collateral after notice. RBI requires a public procedure, with a reserve price of at least 90% of current value for the initial auction attempts and 85% after two failures. Any surplus must be returned to the customer within seven working days after full auction proceeds are received. A shortfall may be recovered if the loan contract allows it. (RBI, auction rules)
From the lender’s perspective, the sale rapidly converts jewellery into repayment. From the household’s perspective, a temporary liquidity need becomes the permanent loss of an asset.
That difference is why lender loss rates alone cannot measure the product’s household risk. The auctioned item may have been accumulated over years, transferred within a family or held precisely for difficult periods. Its market value repairs the balance sheet. It does not necessarily restore its patrimonial function.
The near-term risk may therefore take a quieter form than a wave of bank losses:
- households repeatedly rolling principal;
- interest paid to retain access to the same collateral;
- borrowing capacity increasingly tied to the gold price;
- jewellery sales when income no longer covers the loan;
- a growing share of receivables transferred into banks.
A rising metal price keeps that chain comfortable. It allows the lender to remain over-collateralised and the customer to renew. The relevant reversal is not limited to a gold crash. A long period of flat prices would be enough to stop the automatic expansion of borrowing capacity while interest continued to accrue.
What now deserves monitoring
The raw number of loans will not reveal fragility on its own. More precise indicators matter.
Actual disbursement LTV. The 75% to 85% ceilings are limits, not the portfolio average. A sustained rise in origination LTV would reduce the time available after a price decline.
The share of genuinely new customers. RBI has already identified a market led by existing borrowers. A widening divergence between new accounts, originations and outstanding balances would indicate more renewal and cash extraction.
Interest paid before renewal. The 2025 framework requires it for bullet loans. Whether interest is paid from income rather than another borrowing source is an economically important boundary that public data do not readily reveal.
Auctions and surplus refunds. Rising jewellery sales would show household stress before large lender losses appeared.
NBFC funding. Direct assignment recycles balance sheets quickly. The share of gold loans in structured transactions, the purchasing banks and post-transfer pool performance all deserve scrutiny.
Substitution away from unsecured credit. RBI finds that personal-loan growth slows among customers holding both a personal loan and a gold loan, especially in riskier categories. That shift can reduce lender risk by turning unsecured debt into asset-backed debt. For the household, it puts family wealth directly in the line of fire.
Gold still protects, but it is working harder
India’s gold-loan market is not currently a documented banking bomb.
Average LTVs are low, bad loans remain contained, auctions usually protect principal and RBI has strengthened valuation, documentation and customer safeguards. The available evidence argues against an easy imminent-crisis story.
It shows something more mechanical and potentially more durable.
India is converting a growing part of its domestic gold stock into credit infrastructure. A higher metal price improves balance sheets, attracts new lenders, allows existing customers to refinance larger amounts and supplies NBFCs with receivables that banks want to buy.
Every link can be rational. The household receives liquidity without selling. The NBFC originates a heavily collateralised loan. The bank buys an asset with very low historical losses. The regulator imposes buffers and an auction process.
The system nevertheless creates a new dependency: part of household borrowing capacity is now indexed to the value of family jewellery.
As long as gold rises, that dependency resembles strength. Collateral grows, LTV falls and credit can be renewed. When the metal stops doing the work, one variable remains that the gold price never replaced: the household’s disposable income available for repayment.
Sources and method
The core data come from the Reserve Bank of India’s Financial Stability Report of 30 June 2026, the harmonised gold and silver collateral directions updated on 29 September 2025, and the World Gold Council’s India report for the second quarter of 2026.
Securitisation and performance figures come from Crisil Ratings and reflect the agency’s stated perimeters and samples, including specialist NBFCs for the repayment study. The l0g calculator and stress tests use static simple-interest formulas. They forecast neither the gold price nor lender or borrower behaviour.
For further reading: how to read the gold market and how to read a consumer-credit securitisation.
This analysis is not investment advice.
// cite this analysis
l0g, “When gold rises, so does the debt”, l0g.fr, published August 26, 2026, updated August 26, 2026, https://l0g.fr/en/analysis/when-gold-rises-so-does-debt/
$ cd ../analysis