Research / Gold & Recession Behavior
The Hypothesis
Everyone, regardless of income, prefers investing in gold for contingency planning.
One of three founding questions behind this project. Not whether people like gold, but whether it functions as a deliberate financial safety net, across every income bracket.
Why ask this here rather than assume the answer? Because the economics underneath it are striking.
South India isn't just fond of gold. It dominates India's gold economy. The World Gold Council puts South India at 40% of the country's entire gold jewellery demand, more than any other region, with Chennai a key centre. Southern India also shows the highest ownership rate nationally (76%) and the highest recent-purchase rate (60% bought within the year). Not a stereotype: the documented centre of gravity of Indian gold demand. (World Gold Council, "Understanding regional, income and demographic differences")
Women's gold holdings, in aggregate, are a monetary-scale phenomenon, not just a household habit. Widely cited estimates put Indian women's collective holdings above the official reserves of several of the world's largest sovereign holders. The scale is the point: one gold chain and one wedding at a time adds up to something that matters nationally. Which is why who decides and why gold are the same question.
And it isn't only households doing this. The RBI has been one of the world's most active central-bank gold buyers since the pandemic: roughly 244 tonnes added between 2020 and 2024, past 880 tonnes by 2025, second globally behind only China. Analysts frame it exactly as our respondents frame their own gold, as a hedge against currency risk and a form of trust that doesn't depend on any one institution. (Business Standard; Trading Economics, India Gold Reserves)
So the behaviour we're asking a Chennai household about is also, right now, the RBI's monetary strategy. When a central bank with access to every instrument on Earth reaches for gold for the same stated reason, the household instinct looks less like superstition and more like the same logic at a smaller scale.
One widely circulated figure, that Tamil Nadu households hold more gold than Germany or Italy hold in reserve, we couldn't trace to a verifiable primary source, so we're leaving it out. If we find the original, we'll add it.
Four ideas from behavioural economics do most of the explanatory work here.
When the future is uncertain, people set money aside as a buffer. Kimball's 1990 Econometrica paper showed the strength of that motive is measurable. The question here is whether gold becomes the buffer where formal insurance and credit are harder to reach. (Kimball, 1990)
Thaler's term for how people sort money into separate mental "accounts," even when that's inefficient. Gold bought for a daughter's wedding and never touched under strain is the purest form of it: not savings, but a promise kept in a physical object. (Thaler, 1999)
Gold pays no interest and no dividend, so by cold financial logic it should lose to almost anything else. It keeps winning because people aren't only pricing return. They're pricing trust, volatility, and recognizability, and gold scores high on all three.
The honest complication: is this preference, or just having enough spare income to buy any gold at all? Campbell's 2006 work on the "limited participation puzzle" showed people can be priced out of an asset class they'd otherwise choose. Untangling the two is much of what this tests. (Campbell, 2006)
A factory-floor household, an auto driver, and a business owner get different instruments, but the same underlying question. Verbatim examples:
"Many families buy gold regularly. Why do you think gold is preferred over stocks or mutual funds?"
"Do you trust banks more, or do you trust gold more?"
"Does your household own or regularly buy gold? What is the main reason for keeping it?"
"Suppose you had ₹100,000 that you would not need immediately. Where would you feel safest keeping it: gold, a bank deposit, cash, or somewhere else? Why?"
"If someone gave you ₹1 lakh today as a gift, what would you do with it?"
"When your family thinks about preserving wealth, what feels most reassuring or culturally natural (gold, property, business, market investments, or a mix), and why?"
Every interview is recorded, transcribed, and coded before it becomes a data point. Full process on the Research page.
"The gold that sells for ₹1 lakh today will become ₹1.5 lakh tomorrow... even if we invest in a bank outside, we have to do it when we have a budget of 50,000 or 2 lakhs."
Muthu, auto driver
"My grandmother used to save up gold, my mom did as well, and that's actually what paid for my wedding and my education."
Josef Chellathurey
"If you have ten thousand rupees, you can simply invest in gold. You can buy it like a coin... if you think you can buy land, you will need amounts, there will be a minimum requirement to buy square feet. We can't buy one or two feet, right? That's why this is grams."
Yugandhar
"If they have a daughter, they buy gold for the child's safety. They need the gold for the girl's marriage anyway."
Padmavathi
90 respondents across three income segments, interviewed in and around Chennai between May and June 2026. Every figure below traces back to a recorded, transcribed, and coded conversation.
71%
name gold as their most trusted financial vehicle (64 of 90 respondents).
44%
tie gold explicitly to emergencies, hedging, or have actually pledged it (40 of 90).
100%
of middle-income respondents have pledged or pawned gold before (30 of 30).
0
low-income respondents currently hold any gold, despite wanting to.
The 27-point gap between trusting gold and framing it as a safety net is the most interesting number in the study.
Gold holdings by income segment
Preference is constant across segments. Ability to act on it is not.
What separates gold from every other vehicle in respondents' accounts is how fast it converts to cash, with no paperwork or credit check. Middle-income respondents put it almost identically: you use it for emergencies, and if there's a problem you mortgage it. Every one of them has. Here the contingency logic isn't a preference; it's a documented behaviour.
If gold were only a workaround for limited access, it should vanish at the top. It doesn't. Respondents with portfolios across real estate, bonds, and fixed deposits still name gold as what they trust for immediate liquidation. One funded his education and wedding from gold his grandmother and mother saved. Another delegates gold buying to his wife and prefers real estate personally, yet still ranks gold first for preserving family wealth.
71% trust gold; only 44% tie it explicitly to liquidity, safety, or pledging. That 27-point gap matters: a substantial minority frame gold around appreciation, tradition, or how a family is seen. Gold does more than one job, and the hypothesis captures one.
Not one low-income respondent currently holds gold. But the preference doesn't weaken here: they describe the identical logic as an aspiration they can't yet act on. Reading that zero as indifference gets the finding backwards.
One narrative recurs in every segment: gold held for a daughter's eventual wedding. That gold is contingency planning, a decision women disproportionately drive, and proof of a household planning years ahead to celebrate well, all at once. Three hypotheses describing one behaviour from three angles.
The conclusion
A central bank hedging currency risk and a low-income household with no bank relationship both reach for the same asset, for recognizably similar reasons.
Behaviour described, including by the households themselves, as cultural habit is better understood as a rational response to constrained access, incomplete information, and real social signalling. The hypothesis holds where access allows it to.
The main constraint is size. Three samples of 30, from in and around Chennai across two months, are not statistically representative of South India or of any income group. These findings describe patterns in this dataset, not a claim about South Indian households generally.
The three questionnaires are also not the same instrument: a long fixed script, a shorter one, and an open-ended conversation where a topic not coming up doesn't mean no view on it. Cross-segment comparisons are directional, not like-for-like, which is why we pooled only genuinely comparable questions.
Recording and translation quality varied; ambiguous answers are flagged rather than resolved by inference. The middle-income sample was entirely male, so findings on women's gold influence there are husbands describing wives. A future round should close that gap, extend beyond Chennai, and follow a full festival and wedding season.
World Gold Council. "Understanding regional, income and demographic differences", from the India Gold Market Series.
Business Standard. "Reserve Bank of India's post-Covid gold addition second highest globally" and Trading Economics, India Gold Reserves.
Thaler, R.H. (1999). "Mental Accounting Matters." Journal of Behavioral Decision Making, 12(3), 183–206.
Kimball, M.S. (1990). "Precautionary Saving in the Small and in the Large." Econometrica, 58(1), 53–73.
Campbell, J.Y. (2006). "Household Finance." Journal of Finance, 61(4), 1553–1604.
Dvara Research. "Savings in Gold by Low-Income Households."
"Gold in household portfolios during a pandemic: Evidence from India." Journal of Empirical Finance.
Next: beyond Chennai, women's own accounts in the middle-income segment, and a full festival and wedding season instead of two months. Subscribe to follow it.
Subscribe for updates