The Ledger Before the Bank: Reciprocity and Social Norms in Informal Financial Practices in India

THE LEDGER BEFORE THE BANK: RECIPROCITY AND SOCIAL NORMS IN INFORMAL FINANCIAL PRACTICES IN INDIA
DR. KOMAL B. SHARMA1
¹ Assistant Professor, Centre of Excellence, School of Law, Gujarat University, Ahmedabad 380009, India (ROR: https://ror.org/017f2w007) iD (ORCID: https://orcid.org/0000-0001-9121-5362)
Corresponding author: Komal B. Sharma (e-mail: sharma.komal0096@gmail.com).
ABSTRACT At weddings across Gujarat and much of North and Central India, guests’ hand over cash, and a designated relative writes the amount in a register against the giver's name. Decades later, when the giver's own daughter or son marries, the host family consults that register and returns a comparable, usually larger sum. Households treat these registers as records of what is owed to them and what they owe. Most of the research literature reads this practice as ritual, kinship maintenance, or status display. This study argues that it is also a financial instrument, and that reading it as one is analytically productive.
The present study is a framework that characterizes recorded reciprocal gifting as a deferred-repayment, socially collateralized savings instrument with an event-contingent maturity, an implicit return set by norm rather than contract, and enforcement through reputation within a bounded network. Six parameters define the instrument: principal, tenure, implicit return, default risk, liquidity, and transferability. On each, the present study contrasts it with the formal instruments it is often assumed to have been replaced by, the fixed deposit, the chit fund, and the life insurance policy, and shows that the register dominates on some dimensions and loses badly on others. This is not a story of a primitive practice waiting to be modernized. It is a story of two instruments with different risk profiles, held simultaneously.
Because no ledger corpus has yet been assembled, it specifies a mixed methods protocol: digitization and coding of household registers spanning multiple decades, construction of a giver-recipient panel, estimation of implicit returns against consumer price inflation, and semi-structured interviews on norms and sanctions. Eight testable propositions are derived from the framework. The study closes by arguing that digital payment adoption is unlikely to kill the register, the register was never primarily about payment, but may quietly change who can see the ledger, and that this is worth watching.
INDEX TERMS Reciprocal Gift Exchange, Informal Finance, Social Collateral, Quasi-Credit, Neota, Chandlo, Household Finance, Gujarat, Marriage Payments
JEL CODES D14, G51, O17, Z13

1. INTRODUCTION
A man attends a wedding in a small town in Saurashtra. He walks to a table near the entrance, hands over an envelope with ₹2,100 in it, and gives his name. Someone writes it down in a bound notebook. He does not get a receipt. There is no interest rate quoted, no maturity date, no counterparty risk assessment, no regulator. He will probably not think about the money again for fifteen years. Then his own daughter will marry, and the family whose wedding he attended will open the same notebook, look up his name, and send someone with an envelope containing ₹5,100.
Ask him what he did that day and he will say he attended a wedding. He will not say he made an investment. But look at the cash flows, and it is difficult to describe the transaction any other way.
The practice has names. In North and Central India it is neota or nyota, and the registers that record it are kept and consulted across generations. In Gujarat the terminology is looser: vyavhar covers the general obligation, and chandlo is used colloquially for the cash a guest gives at a wedding, though the same word also names a specific engagement ritual in which the bride's father marks the groom's forehead with vermilion and hands over token money. The overlap is worth flagging early, because the two are not the same transaction. The engagement chandlo is vertical, it flows from one affinal family to another as part of the alliance. The reception chandlo, the one written in the notebook, is horizontal. It flows between peers, and it is expected back.
That distinction matters more than it might seem, and it explains part of why this practice has been under-studied. The research literature on Indian marriage payments is enormous, and it is almost entirely about dowry. Dowry is a vertical transfer, it is asymmetric, it is coercive at the margin, it is illegal, and it produces harm severe enough that studying it needs no justification. The horizontal, recorded, reciprocal transfer sitting alongside it in the same wedding, often involving comparable aggregate sums, gets mentioned in passing and then dropped. It is treated as background, as the social texture against which the interesting transfer happens.
Meanwhile, a separate literature in development economics has spent thirty years modeling exactly this kind of arrangement in the abstract. Coate and Ravallion [1] asked how reciprocal transfers survive without commitment. Kranton [2] showed that reciprocal exchange can be self-sustaining even when market alternatives exist. Thomas and Worrall [3] named the object of study “quasi-credit”, transfers that are neither pure gift nor pure loan, repaid in state-contingent ways. Ligon, Thomas and Worrall [4] fitted limited-commitment insurance models to Indian village data. Fafchamps and Lund [5] traced the networks that carry these flows in the rural Philippines.
Almost none of this literature engages with the fact that in large parts of India, households have been keeping written records of exactly these transfers for a century or more.
This is a strange gap. Economists modeling informal insurance have had to work hard to infer transfer histories from survey recall, which is noisy and short. The registers are the transfer history. They are already written down, in the household's own hand, spanning decades, with names, amounts, dates, and occasions. Any household with a marriage in the last fifty years probably has one in a cupboard.
The study argued on three parts. First, that recorded reciprocal gifting is usefully modeled as a financial instrument, and that doing so identifies a specific configuration, deferred repayment, event-contingent maturity, norm-set implicit return, social collateral, near-zero liquidity, partial heritability, that does not correspond to any instrument in the formal Indian financial system.
Second, that this configuration explains its persistence. The register survives not because households lack access to banks, many of the households keeping registers hold fixed deposits, insurance, and mutual funds, but because it does something those instruments do not. It converts an unpredictable, socially mandatory, lumpy expenditure into a claim on a network with an obligation to help meet that same expenditure. It is matched to its liability in a way a fixed deposit is not.
Third, that the household registers constitute a usable and currently untouched longitudinal dataset, and that the main obstacle to studying them is not availability but the fact that nobody has thought to ask.
2. LITERATURE REVIEW
2.1 The gift as obligation
The foundational text is Mauss [6]. Mauss's claim, against the then-standard view of gifts as voluntary and disinterested, was that gift exchange in the societies he surveyed involved three obligations that were not optional: to give, to receive, and to reciprocate. Refusing any of the three was an act of hostility. The gift looked free and was not.
Malinowski [7] had already documented the kula ring in the Trobriands, in which valuables circulated on fixed paths with an accounting that participants carried in their heads. Sahlins [8] sorted reciprocity into three types by social distance: generalized (kin, no accounting, no expectation of return), balanced (equivalent return within a defined period), and negative (extraction). Recorded wedding gifting sits squarely in Sahlins's balanced category, which is the category his framework has least to say about, because balanced reciprocity looks so much like trade that the anthropological interest drains out of it.
Gouldner [9] argued that a norm of reciprocity operates as a near-universal moral rule and functions as a starting mechanism for social relations, it lets strangers begin exchanging before trust exists, because the first gift creates an obligation that substitutes for trust.
Bourdieu [10] made the observation that matters most for this study. He noticed that gift exchange requires a time interval. Return the gift immediately and you have refused it — you have converted a relationship into a transaction. The delay is not a friction. The delay is the point, because it is what distinguishes the gift from the trade, and what keeps the relationship alive during the gap.
Bourdieu also thought the participants had to misrecognize this. The exchange works, on his account, only if people do not consciously experience it as exchange. The Indian registers are an awkward case for that claim. There is nothing misrecognized about writing the amount down in a book and looking it up fifteen years later. Households know exactly what they are doing. Whether this makes the Indian case an exception to Bourdieu, or a refutation, or simply evidence that his model was over-specified for the Kabyle, is an open question this study does not resolve, but it is worth someone resolving.
2.2 Reciprocity as economics
Development economics arrived at the same object by a different road, motivated by a puzzle: rural households in poor economies face large idiosyncratic income shocks and no insurance market, and yet consumption is smoother than income.
Townsend [11], using ICRISAT panel data from three Indian villages, tested full risk-sharing and rejected it, but found that consumption tracked village aggregates far more closely than own income. Something was pooling risk. Coate and Ravallion [1] modeled that something as an implicit contract sustained without external enforcement, where deviation is punished by exclusion from future transfers, and showed that the arrangements this supports are second-best, not efficient.
Kranton [2] proved something more provocative: reciprocal exchange systems can be self-sustaining, and can persist even when markets are available, because participation in reciprocal exchange lowers the returns to market participation for others, which sustains the reciprocal system. Systems like this do not necessarily die when banks arrive.
Ligon, Thomas and Worrall [4] formalized limited-commitment insurance and fitted it to village data. Thomas and Worrall [3] coined “quasi-credit” for transfers that resemble loans but with repayment terms that are neither fixed nor fully state-contingent. Fafchamps and Lund [5] showed empirically in the Philippines that risk-sharing runs through networks of friends and relatives rather than the village as a whole, and importantly, that gifts and informal loans, not formal credit, do the smoothing.
Platteau [12] pushed back on the whole program, arguing that mutual insurance in traditional communities is an elusive concept, often more about status and patronage than about risk pooling, and warning against economists reading their own models into other people's institutions. That warning applies to this study and is noted.
Besley, Coate and Loury [13] analyzed ROSCAs, the closest formal cousin. Karlan, Mobius, Rosenblat and Szeidl [14] developed “social collateral” the idea that a social connection has value that can be pledged, and that the pledgeable amount depends on network structure. Greif [15] had shown much earlier, with the Maghribi traders, that reputation within a bounded coalition can enforce contracts that no court will.
Every one of these studys is about the mechanism the registers implement. None of them mentions that the mechanism keeps written records.
2.3 Marriage payments in India
Rao [16] treated dowry as a price and used hedonic methods to explain its rise through marriage-squeeze demographics. Anderson [17] surveyed the dowry and brideprice literature and the puzzle of why dowry rises with development in India while brideprice falls elsewhere. Chiplunkar and Weaver [18] provided long-run estimates of dowry incidence and value across Indian marriage markets, finding it near-universal over the second half of the twentieth century despite prohibition since 1961.
Bloch, Rao and Desai [19] come closest to this study's territory. They analyzed wedding celebration expenditure in rural Karnataka as conspicuous consumption, a signal of status directed at the bride's future in-laws and the village. Their framing is signaling, and the expenditure is a sunk cost that buys reputation.
This is the point where the literature and this study diverge. Bloch, Rao and Desai model wedding spending as consumption. The register says at least part of it is not. If a substantial share of what a household spends on a wedding is recovered from the same guests at the guests' own weddings, then aggregate wedding expenditure figures, including the alarming ones about marriage loans at 10 to 37 percent interest, and about households borrowing to marry daughters, are measuring gross outflow while the household is thinking about net.
Nobody has netted it out. That is a measurement problem with policy consequences, and it is one of the reasons to do this study.
Lewis [20], in his study of a village near Delhi, did document neota directly, including the registers. Srinivas [21] recorded the density of gift obligation in Rampura. The ethnographic record exists. It has simply not been connected to the financial literature.
2.4 Recorded reciprocity elsewhere
The Indian case is not unique, and the comparative literature is better developed.
Yan [22] studied gift flows in a village in Heilongjiang, China, over a full year, and documented the lidan the gift list, kept for weddings and funerals. His analysis of renqing (human feeling, obligation) shows a system almost structurally identical to neota: recorded, expected back with increment, and central to a household's standing. Yan's book is the closest existing model for what an Indian study would look like, and it is thirty years old.
Japan has kōden, condolence money given at funerals, recorded, and partially returned in the form of return gifts. Korea has bujo-geum (경조사비), wedding and funeral money, with registers, and a well-documented cultural anxiety about the accounting. In each case, researchers have treated the practice as a serious object.
The Indian literature has not caught up. There is no Indian equivalent of Yan [22].
2.5 Money is not fungible
Zelizer [23] argued against the economist's assumption that money is fungible, showing that people systematically earmark money by source and purpose, and that a rupee received as a wedding gift is not psychologically or socially the same rupee as one earned as wages. It cannot be spent on anything, and it cannot be defaulted on in the same way.
Guérin (2014), working in rural Tamil Nadu, showed how households juggle multiple debts of different moral quality simultaneously, prioritizing repayment not by interest rate but by the social meaning of the debt and the relationship carrying it. Households will service a socially heavy zero-interest obligation before a formally contracted high-interest one.
If Zelizer and Guérin are right, then a household's decision to keep participating in the register despite having a bank account is not irrationality or inertia. The register money and the bank money are different kinds of money, held for different reasons, and asking why the household holds both is like asking why it holds both insurance and savings.
3. RESEARCH GAP
Pulling the four literatures together, the gap is specific.
Anthropology has the practice but not the finance. Neota, chandlo and vyavhar appear in ethnography as kinship and ritual. The registers are noted. Nobody has asked what the implicit return is, or how tenure is distributed, or what happens on default.
Economics has the finance but not the practice. Thirty years of quasi-credit and limited-commitment insurance models, calibrated on survey recall data, while the primary records sat in households' cupboards, unexamined.
The Indian marriage literature has the site but the wrong transfer. Dowry is studied exhaustively. Horizontal recorded reciprocity, occurring at the same event, gets a sentence.
The comparative literature has the template but not the case. Yan [22] did this for China. India has nothing equivalent, despite a practice that is older, more widespread, and better documented in writing.
And there is a measurement problem nobody has flagged. Wedding expenditure is reported gross. If a meaningful share is recoverable, then estimates of the household burden of Indian weddings, which drive policy arguments about marriage loans, indebtedness, and the case for expenditure caps, are overstated by an unknown amount. The register is the only instrument that can tell us by how much.
The present study addresses, recorded reciprocal gifting has never been characterized as a financial instrument, and the household registers that document it have never been used as data.
4. OBJECTIVES AND RESEARCH QUESTIONS
4.1 Objectives
1. To characterize recorded reciprocal gifting as a financial instrument, specifying its parameters in terms commensurable with formal instruments.
2. To situate the practice within a broader typology of traditional household financial arrangements, of which recorded gifting is one component rather than a curiosity.
3. To specify a replicable protocol for converting household registers into analysable longitudinal data.
4. To derive testable propositions about implicit return, tenure, network structure and enforcement.
5. To assess what formal financial deepening and digital payments do to the practice — and to state clearly what would count as evidence either way.
4.2 Research questions
RQ1. What are the parameters, principal, tenure, implicit return, default risk, liquidity, transferability, of recorded reciprocal gifting, and how do they compare with the fixed deposit, the chit fund, and the endowment insurance policy?
RQ2. What is the repayment norm in practice? Is the return nominal, inflation-indexed, status-indexed, or something else, and does it vary with kinship distance, elapsed time, and the relative wealth of the parties?
RQ3. How is the obligation enforced without legal contract, and what happens when a household defaults, dies out, migrates, refuses, or simply cannot pay?
RQ4. Does the register function as insurance, as savings, as a status ledger, or as several of these at once for different households and different entries?
RQ5. Is formal financial access associated with exit from the register, or with continued participation alongside formal instruments? What does the transition to digital payment do to recording practice, and to who has access to the record?
5. CONCEPTUAL FRAMEWORK
5.1 Definition
Recorded reciprocal gifting is defined here as a transfer of value from a giver to a host household on the occasion of a life-cycle event, entered in a written register maintained by the host, creating a socially enforceable expectation of a return transfer of at least equivalent value on the occurrence of a comparable event in the giver's household, at an unspecified future date.
Four features distinguish it from adjacent practices. It is horizontal, between households of broadly comparable standing, unlike dowry. It is recorded, unlike generalized kin reciprocity. It is event-contingent, the return is triggered by an occasion, not a date. And it is unenforceable at law; no court will hear the claim.
5.2 The six parameters
Principal. The amount given, set by the giver, but bounded above and below by norm. The lower bound is what would be read as an insult given the relationship. The upper bound is what would be read as showing off, or as an attempt to place the recipient under an obligation they cannot discharge. The band between them is narrow and well understood by participants, which is itself interesting: it means the “investment decision” has very little discretion in it. Auspicious increments (₹501, ₹1,001, ₹2,100, ₹5,100) are near-universal, which is a small anomaly worth noting, the number is chosen for its last digit before it is chosen for its magnitude.
Tenure. The interval between outward and return transfer. Not chosen, not known in advance, and bounded only by demography. If the giver's household has an unmarried child of six, tenure will be roughly twenty years. If the giver has no children, tenure may be infinite, which is to say the transfer was a gift after all. This is the instrument's defining oddity: its maturity is a function of the counterparty's family structure, which the investor can observe but not control.
Implicit return. The ratio of return transfer to original transfer, annualized over the realized tenure. Not contracted. Set by convention, and the convention appears, this is a hypothesis, not a finding, to track something like nominal social standing rather than a price index. The relevant question for RQ2 is whether the realized ratio beats CPI over the same interval. If it does, the register is a positive-real-return instrument, which would be remarkable for something nobody calls an investment. If it does not, participants are accepting a real loss for the relationship, which is also worth knowing.
Default risk. The probability of no return. Decomposable into at least four channels: demographic (giver has no marriageable children), migration (giver's household leaves the network), refusal (rare, socially catastrophic), and inability (the giver's household is impoverished when the occasion arrives, and the obligation is quietly forgiven). The last channel is the one that makes this insurance rather than credit, an obligation that is forgiven when the counterparty is in distress is not a debt.
Liquidity. Effectively zero. The claim cannot be sold, discounted, borrowed against at a bank, or called early. A household holding ₹8 lakh of claims across two hundred entries cannot convert any of it into cash for a medical emergency. This is the instrument's worst feature by a wide margin, and any account of why households hold it has to explain why they tolerate this.
Transferability. Partial and asymmetric, and this is where the instrument gets genuinely strange. The claim is heritable, it passes to sons, and the register is passed down, but it does not pass cleanly through a daughter's marriage, since she leaves the household that holds the claim. The obligation is also heritable: a household can inherit a debt to a family its patriarch never met. The register is therefore an intergenerational balance sheet, and its heritability rules are gendered in the same direction as everything else in the property system.
5.3 The register as contract technology
The register does three things at once and they are worth separating.
It is a memory device. Human recall over twenty years across two hundred counterparties is not reliable. The book removes the need to trust memory, which removes a whole class of disputes.
It is a verification device. Because the record is held by the recipient and consulted publicly at the time of return, often read aloud, in front of family, the giver can check it. A recipient who under-records is caught. A giver who claims more than they gave is caught. The record's public consultation is what makes it credible, not the mere fact of writing.
It is a commitment device, and this is the subtle one. Once written, the obligation is out of the household's head and into an object that outlives the household head. His sons will find it. Written obligation is harder to quietly forget than remembered obligation. The book binds the future household to the past household's promises, which is precisely what limited-commitment models say is hard.
Coate and Ravallion's [1] central problem is that reciprocal arrangements are limited by what people will voluntarily honour. The register does not solve that, nothing makes the obligation legally enforceable, but it raises the cost of dishonouring, by making the dishonour legible, dateable, and specific. “You gave nothing” is deniable. “Here is the page, here is your father's name, here is ₹501 in 1987” is not.
Greif [15] showed reputation enforcing contracts among the Maghribi traders through correspondence networks. The register is the household-level version: a private, decentralized, written reputation system with no central node.
5.4 A typology of traditional household instruments
Recorded gifting is one instrument in a portfolio. Studying it alone risks the error of treating it as a curiosity rather than a component. The table below sets out a working typology, it is a framework proposal, and the empirical work would refine it.

TABLE 1. Typology of Traditional Household Instruments
Practice Financial function Tenure Return Liquidity
Neota / chandlo register Deferred-repayment savings, matched to a known future liability Event-contingent, 5-40 yrs Norm-set, likely positive nominal None
Mameru / mosalu (maternal uncle’s obligation) Kin-based mandatory transfer; risk-shifting to the natal family Generational Reciprocated across generations None
Bhishi / committee (ROSCA) Rotating credit and forced savings Cycle length, fixed Zero to negative nominal; positive in access terms Position-dependent
Streedhan / gold ornaments Inflation hedge plus emergency collateral Perpetual Tracks gold High (pawnable)
Land or gold transferred to daughters Intergenerational wealth transfer outside inheritance law Generational Asset-dependent Low
Death-rite and illness contributions Mutual insurance against catastrophic lumpy cost Event-contingent Approximately unity None

Read as a portfolio, the logic is legible. The household holds a liquid inflation hedge (gold), a forced-savings vehicle (bhishi), a mandatory kin transfer (mameru), and a liability-matched claim on its network (the register). What it does not hold is anything with a contracted return and a stated maturity. That is what the formal system supplies, and it is why the formal system supplements the portfolio rather than replacing it.
6. RESEARCH METHODOLOGY
6.1 Design
Sequential mixed methods. Quantitative strand first (ledger digitization and analysis), qualitative strand second (interviews), with the qualitative strand designed after preliminary quantitative patterns are known so that it can interrogate them rather than merely accompany them.
The design is exploratory-descriptive. There is no identification strategy here and none is claimed. The first task is to establish what the parameters are. Causal questions come later, and would need a different design.
6.2 Site and sampling
Three sites, chosen for contrast rather than representativeness:
• Site A: an urban community in Ahmedabad with high formal financial access and high digital payment penetration.
• Site B: a semi-urban town in Saurashtra.
• Site C: a rural taluka.
Within each site, sampling is necessarily purposive and network-based. Random household sampling will fail, because access to a family's register requires trust that a stranger with a clipboard does not have. Entry through community elders, marriage-hall operators, and the men who traditionally staff the recording table at receptions is more realistic. Target: 12–15 households per site, 36–45 total, each contributing at least one register.
Snowball sampling has an obvious problem here that must be stated rather than hidden. The register is a network. Sampling through the network means the sample is drawn from within the object of study, and the resulting network measures will be biased toward density. This is a real limitation, partially addressable by seeding from several unconnected entry points per site and reporting steed provenance for every household.
6.3 The ledger corpus
The target is registers spanning at least three marriage events per household, ideally covering two generations, so that the same counterparty name appears as both giver and receiver.
Practical constraints, learned from anyone who has tried to work with household documents: registers are fragile, they are not lent out, and photographing them in a family's home in one sitting is the only realistic mode. Budget for a copy stand, diffuse lighting, and a fixed-focus phone rig rather than a scanner. Expect Gujarati script, inconsistent hands, marginal annotations, crossings-out, and pages where the ink has gone. Expect at least one household to produce a register and then decline to let it be photographed after seeing the consent form, and do not treat that as a failure.
6.5 Quantitative measures
For each matched pair (outward transfer i at time t, return transfer j at time t + τ):
• Nominal ratio: R = amountj / amounti
• Tenure: τ in years
• Implicit nominal annualized return: r = R(1/τ) − 1
• Real return: deflated by the CPI–Rural or CPI–Urban series for Gujarat, as appropriate to site
• Benchmark spread: r less the SBI one-year term deposit rate prevailing at t, held to maturity τ
• Gold benchmark: r less the annualized return on gold over the same interval
Descriptive analysis first: distributions of R and τ, and how R varies with τ, kin distance, geographic distance, and site. Then a regression of log R on log τ, kin distance dummies, site, event decade, and relative household standing, reported as description, with no causal interpretation, because the sample will not support one.
Network analysis on the giver- recipient bipartite graph: degree distribution, reciprocity rate, clustering, and, most interestingly, whether the size of the transfer scales with the network centrality of the counterparty, which is the direct test of Karlan et al.'s [14] social collateral prediction.
6.6 Qualitative strand
Semi-structured interviews, 8–10 per site, split between register-keepers aged 55+ and household members aged 25–40. Two protocols.
The elder protocol asks: who keeps the book and why that person; how the amount is decided; what happens when someone gives too little; whether a return has ever been refused, and what followed; whether an obligation has ever been forgiven, and on what grounds; what happens to the book when the head dies.
The younger protocol asks: whether they know the book exists; whether they have seen it; whether they expect to keep it; what they would do if an entry came due and they did not have the money; whether they would rather the money were in a bank.
The gendered access question needs its own treatment. The register is, by strong convention, kept by men. Women's knowledge of the household's outstanding claims and obligations is an empirical question nobody has asked, and it bears directly on what happens to the practice as household financial decision-making changes. Interview women separately, and not through their husbands.
6.8 Validity and its limits
Survivorship. Registers that survive belong to households that stayed put, stayed solvent, and stayed intact. Households that migrated, dissolved, or fell out of the network are absent by construction, and those are exactly the households where default happened. Estimated default rates from surviving registers are therefore biased downward, probably severely. This is the single most serious threat to the study and cannot be fixed by design. It can only be stated, bounded where possible, and carried through every claim.
Recording bias. What is written may not be what was given. Kind gifts may go unrecorded. Very small amounts may be omitted. Large amounts from important people may be inflated. Cross-checking one household's register against a counterparty's register, where both are available, is the only real check, and it will be available for a minority of entries.
Recall. Kin relations, event dates and household standing are reconstructed through interview, decades after the fact. Confidence coding on every inferred field is not optional.
Generalizability. Three sites in Gujarat. Community-specific norms are strong. Nothing here generalizes to India, and the study should not pretend otherwise.
7. ANALYSIS AND DISCUSSION
No data has been collected. What follows is analytical: it works out what the framework implies, illustrates the arithmetic with a stylized case, and converts the implications into propositions the protocol can test. Nothing in this section is a finding.
7.1 Stylized arithmetic
This example is constructed to illustrate the calculation. The numbers are invented for that purpose and describe no actual household.
Suppose a giver transfers ₹501 at a wedding in 1990 and receives ₹5,100 in return in 2020. Nominal ratio R = 10.18, tenure τ = 30 years, implicit nominal annualized return r = 10.18(1/30) − 1 ≈ 8.05%.
Indian CPI inflation over that period averaged somewhere in the region of 7–8% annually. So the real return on this transaction is somewhere near zero, possibly slightly positive, possibly slightly negative, depending on the deflator chosen and the exact years. A one-year term deposit rolled over for the same thirty years would have done comparably or slightly better. Gold, over 1990–2020, would have done considerably better.
Three things follow, and they are more interesting than the number.
First, if the norm reliably produces a return of roughly this order, then the convention is doing something remarkable: an unwritten social rule, applied by people who are not calculating, is approximately tracking a price index over thirty years. Nobody sat down and indexed it. It indexed itself, presumably because the auspicious-number ladder (₹101 → ₹501 → ₹1,001 → ₹2,100 → ₹5,100 → ₹11,000) ratchets upward in response to the same cost pressures that drive the index. That is a mechanism worth a study of its own.
Second, on a pure return basis, the register is unremarkable. A household optimizing for return would hold gold. So return is not why they hold it.
Third, and this is the argument, return is the wrong metric. Consider what the ₹501 was actually hedging.
7.2 Liability matching
The household's problem is not “how do I grow ₹501.” It is “my daughter will marry in roughly twenty-five years, the cost of that wedding will be set by prevailing social norms at that time, not by my income, and I do not know what those norms will demand.”
That is a liability with three nasty properties: it is large, it is socially mandatory, and its size is determined by a variable, future community expectations, that no financial instrument is indexed to.
A fixed deposit pays a contracted rate. If wedding norms inflate faster than deposit rates, the deposit under-funds the liability. Gold tracks gold. Equity tracks equity. None of these are indexed to what my community will expect me to spend on a wedding in 2050.
The register is. The return transfer is set by the same norms, applied by the same community, at the same moment, that will set the cost of the wedding it is meant to fund. If norms inflate, the return inflates with them, because the return is a norm-governed payment from the same normative system.
This is the framework's central claim, and it is stronger than “the register is a savings account.” The register is the only instrument available to the household that is denominated in the same unit as the liability it funds. It is a natural hedge. Its unit of account is not the rupee; it is the social obligation.
That reframes the liquidity problem too. A household cannot liquidate its register claims for a medical emergency true, and bad. But the register was never the emergency fund. Gold is the emergency fund. The register funds weddings. Each instrument in the portfolio is matched to a liability, and the illiquidity of the register is not a defect relative to its purpose. It is a defect relative to a purpose it does not have.
Proposition 1. Household participation in the register is better predicted by the household's expected marriage liability (number and age of unmarried children) than by its formal financial access.
7.3 Comparison with formal instruments

TABLE 2. The Register Compared with Formal Instruments
Dimension Register Fixed deposit Chit fund Endowment policy
Return Norm-set, uncontracted Contracted Negative nominal, positive in access Low, contracted
Maturity Event-contingent, uncontrollable Chosen Cycle Chosen
Liability matching Direct, same normative unit None None None
Liquidity Zero High (with penalty) Position-dependent Low (surrender loss)
Default risk Demographic, migration, forgiveness Near zero (insured) Organizer risk, high Insurer risk, low
Collateral Social Sovereign Social plus organizer Regulatory
Transferable Heritable, gendered Yes No Yes
Legal recourse None Full Partial Full
Regulated No RBI Chit Funds Act 1982 IRDAI
Confers standing Yes No Marginal No

Two entries in the last two rows do the explanatory work. The register confers social standing that no formal instrument does, the act of giving is itself consumed as reputation, so the household gets a flow benefit before the principal returns. And it is entirely outside regulation and law, which is a cost (no recourse) and a benefit (no tax, no reporting, no scrutiny of cash).
Proposition 2. Households will maintain register participation at unchanged real levels even after acquiring formal instruments, because the instruments are not substitutes. Following Kranton [2], the register is not a transitional arrangement awaiting a bank.
7.4 Norm-set returns and the ratchet
Proposition 3. The realized nominal ratio R is increasing in tenure τ, at approximately the rate of consumer price inflation, and the implicit real return is distributed tightly around zero.
Proposition 4. The realized ratio R exceeds unity in nominal terms in the overwhelming majority of matched pairs. Returning the same nominal amount is a legible insult, and the auspicious-number ladder provides no downward step, so the norm has a hard nominal floor at the previous amount and a socially enforced ratchet above it.
Proposition 4, if it holds, has a consequence worth noticing. The instrument is nominally incapable of loss. A household that gives ₹501 will never receive ₹300. In a country where household inflation expectations are poorly anchored and financial literacy is low, an instrument that cannot lose nominal value is behaviorally attractive in exactly the way fixed deposits are attractive and equity is not, and for the same reason, which is not a good reason.
Proposition 5. The ratio R is increasing in the relative wealth of the returning household and in kin proximity, and decreasing in geographic distance.
7.5 Enforcement and forgiveness
Proposition 6. Observed non-return is concentrated in demographic failure (no marriageable children) and migration out of the network, and outright refusal is rare enough to be anecdotal.
If Proposition 6 holds, the enforcement mechanism is doing something that the limited-commitment literature [1], [4] predicts is hard. Their models generate substantial deviation because deviation is individually rational when the continuation value is low. The register raises the cost of deviation through legibility, but that is not obviously enough.
The likelier explanation is that the models under-weight what Zelizer [23] and Guérin (2014) emphasize: the obligation is not experienced as debt. Defaulting on a bank is a financial event. Failing to return a chandlo is a statement about who you are, delivered in public, to everyone whose opinion structures your children's marriage prospects. The sanction is not exclusion from future transfers. The sanction is that your daughter is harder to marry.
Proposition 7. Forgiveness of the obligation is concentrated in cases where the obliged household has suffered an adverse shock, and is granted without a demand for partial payment.
Proposition 7 is the load-bearing one for the insurance interpretation. An obligation that is enforced when you are fine and forgiven when you are not is, by construction, state-contingent. That is Thomas and Worrall's [3] quasi-credit, implemented socially and written in a notebook. If the qualitative data supports Proposition 7, the register is insurance. If forgiveness turns out to be rare and grudging, it is closer to a zero-interest loan with reputational collateral, and the insurance reading fails.
7.6 The measurement consequence
If the register recovers a substantial share of wedding outlay, then every gross wedding-expenditure figure in the Indian policy literature overstates the household's net position.
This bears directly on Bloch, Rao and Desai [19]. Their signalling model treats celebration expenditure as sunk. If a fraction is recoverable, the signal is cheaper than modelled, and the equilibrium level of expenditure their model rationalizes is not the same equilibrium. The signalling logic does not collapse, a partially recoverable signal is still costly, and the recovery is delayed by decades, which is itself a cost, but the parameter values change, and so does the welfare arithmetic.
It also bears on the policy conversation about marriage loans. Reports of households borrowing at 10 to 37 percent to fund weddings are read as evidence of a debt trap. If those households simultaneously hold recoverable claims against their guests, the picture is a liquidity mismatch, not insolvency: the household is asset-rich in an illiquid, unpluggable asset and borrows expensively against a liquidity need. That is a different problem with a different solution, and it points at a policy intervention nobody has proposed, which is a mechanism to make register claims pledge able.
Proposition 8. The share of a wedding's cash outlay covered by contemporaneous incoming gifts, netted against the household's own outstanding obligations, is large enough that gross expenditure materially misstates net burden.
Whether “large enough” means twenty percent or seventy percent is exactly what the data would settle, and the honest position is that the framework does not predict it.
7.7 What digital payment does
The obvious hypothesis is that UPI kills the register. It probably does not, and the reason is that the register was never primarily a payment record.
If the analysis in 7.3 is right, the book's functions are memory, verification and commitment. UPI provides none of them. A UPI transaction log is scattered across the recipient's app, not organized by giver, not annotated with occasion or relationship, not consultable by the giver, not readable aloud at a wedding, and not inheritable by a son who does not have the phone. Payment digitizes. The ledger does not, unless someone deliberately builds a ledger.
Two effects are more likely than death.
Fragmentation. Cash at a table with one man and one book produces a single authoritative record. UPI to four different family members' handles produces four partial records and no authority. The obligation persists; the evidence degrades. Disputes should rise.
Access shift. The register is kept by an elder, usually male, in a physical object anyone in the household can pick up. A UPI history lives on a personal device behind a personal PIN. That is a quiet transfer of control over the household's obligation ledger from the household to the individual holding the handset, and given who holds handsets and who keeps books, the direction of that transfer is not neutral. The person who knows what the household owes may change.
Neither of these is a prediction the framework can prove. Both are things the qualitative strand should look for, and the second is the one that would matter most if true.
7.8 Against romanticizing this
The framing in this study, traditional instrument, natural hedge, elegant enforcement is seductive and should be resisted at the point where it becomes approval.
The register is coercive. Participation is not optional. A household that cannot afford ₹2,100 gives ₹2,100, because the alternative is a public statement about its standing that its unmarried daughters will pay for. Platteau's [12] warning applies exactly here: what an economist models as voluntary risk pooling may be experienced by the participant as an unavoidable levy.
It is exclusionary by design. Social collateral only collateralizes within the network. The network is bounded by caste, community, and kin. A household outside it cannot buy in. Karlan et al.'s [14] social collateral is, restated in the Indian context, a mechanism that makes credit available in proportion to how well-connected you already are, which is another way of describing why the people who need it most have the least of it.
It is gendered. The claim is held by men, recorded by men, inherited by men, and used to fund a transfer of a woman out of the household. Any account of the register that does not say this is incomplete.
And it is regressive in its incidence. The nominal floor means the poorest household in a network gives approximately what the median household gives, because the floor is set by the network's norm, not by the giver's income. As a share of income, the poorest household pays the most. It is a flat levy inside an unequal population, which is the definition of a regressive tax, implemented by nobody, enforced by everybody.
This study's claim is that the register is analytically a financial instrument. That claim is entirely compatible with it also being an instrument of stratification.
8. CONCLUSION
The argument is that a practice everyone in Gujarat knows about and nobody has studied as finance is, in fact, finance, and that its parameters do not match any instrument available in the formal system.
The distinctive feature is not the return, which is probably mediocre, or the enforcement, which is well understood in the abstract. It is the liability matching. The register is the only asset a household holds that is denominated in the same unit as the wedding it is meant to pay for. Its value is set by the same community, by the same norms, at the same moment, as the cost it offsets. No fixed deposit does that. No insurance policy does that. That is why households with bank accounts still keep the book, and it is why they will probably keep keeping it.
The empirical case rests on an observation that is almost embarrassing in retrospect. Development economists have spent three decades inferring transfer histories from survey recall in order to test models of reciprocal exchange, while the transfer histories themselves, written, dated, named, spanning generations, sat in cupboards in the villages they were surveying. Yan [22] used the Chinese equivalent thirty years ago. India, which has more registers, older registers, and a larger literature on marriage payments than anywhere, has not.
The next step is not more theory. It is forty households, a copy stand, and an ethics approval.
What the study could deliver, if the registers are as good as they appear: the first estimate of the implicit return on India's oldest household investment instrument, the first network map built from primary records rather than recall, and a net figure for what an Indian wedding actually costs the household paying for it. That last one has policy consequences, because the current figure is gross, and the difference between gross and net is the difference between a debt trap and a liquidity mismatch.
The risk is that the study finds nothing surprising, that returns cluster around inflation, defaults are rare, networks are dense, and everyone already knew. That would be a real result too. Institutions that work are also worth documenting, and this one has been running for a century without a regulator, a court, or a rate of interest.
9. LIMITATIONS
Survivorship bias in the ledger corpus is severe and unfixable by design; default estimates will be biased downward by an unknown amount. Network-based sampling draws the sample from inside the object of study. Three Gujarat sites support no generalization beyond themselves, and community norms vary enough that even that is generous. The matching of returns to originals across registers, decades and orthographies will fail for some unknown fraction of entries, and separating match failure from genuine default is a problem the protocol identifies but does not solve. The design is descriptive; no causal claim is available. And the third-party consent problem, that the registers profile hundreds of people who cannot be asked , is mitigated by pseudonymization and aggregation, not eliminated.
REFERENCES
[1] S. Coate and M. Ravallion, “Reciprocity without commitment: Characterization and performance of informal insurance arrangements,” J. Dev. Econ., vol. 40, no. 1, pp. 1–24, 1993.
[2] R. E. Kranton, “Reciprocal exchange: A self-sustaining system,” Am. Econ. Rev., vol. 86, no. 4, pp. 830–851, 1996.
[3] J. P. Thomas and T. Worrall, “Gift-giving, quasi-credit and reciprocity,” Rationality Soc., vol. 14, no. 3, pp. 308–352, 2002.
[4] E. Ligon, J. P. Thomas, and T. Worrall, “Informal insurance arrangements with limited commitment: Theory and evidence from village economies,” Rev. Econ. Stud., vol. 69, no. 1, pp. 209–244, 2002.
[5] M. Fafchamps and S. Lund, “Risk-sharing networks in rural Philippines,” J. Dev. Econ., vol. 71, no. 2, pp. 261–287, 2003.
[6] M. Mauss, The Gift: The Form and Reason for Exchange in Archaic Societies (W. D. Halls, Trans.). London, U.K.: Routledge, 1990 (orig. 1925).
[7] B. Malinowski, Argonauts of the Western Pacific. London, U.K.: Routledge, 1922.
[8] M. Sahlins, Stone Age Economics. Chicago, IL, USA: Aldine, 1972.
[9] A. W. Gouldner, “The norm of reciprocity: A preliminary statement,” Am. Sociol. Rev., vol. 25, no. 2, pp. 161–178, 1960.
[10] P. Bourdieu, Outline of a Theory of Practice. Cambridge, U.K.: Cambridge Univ. Press, 1977.
[11] R. M. Townsend, “Risk and insurance in village India,” Econometrica, vol. 62, no. 3, pp. 539–591, 1994.
[12] J.-P. Platteau, “Mutual insurance as an elusive concept in traditional rural communities,” J. Dev. Stud., vol. 33, no. 6, pp. 764–796, 1997.
[13] T. Besley, S. Coate, and G. Loury, “The economics of rotating savings and credit associations,” Am. Econ. Rev., vol. 83, no. 4, pp. 792–810, 1993.
[14] D. Karlan, M. Mobius, T. Rosenblat, and A. Szeidl, “Trust and social collateral,” Q. J. Econ., vol. 124, no. 3, pp. 1307–1361, 2009.
[15] A. Greif, “Contract enforceability and economic institutions in early trade: The Maghribi traders' coalition,” Am. Econ. Rev., vol. 83, no. 3, pp. 525–548, 1993.
[16] V. Rao, “The rising price of husbands: A hedonic analysis of dowry increases in rural India,” J. Polit. Econ., vol. 101, no. 4, pp. 666–677, 1993.
[17] S. Anderson, “The economics of dowry and brideprice,” J. Econ. Perspect., vol. 21, no. 4, pp. 151–174, 2007.
[18] G. Chiplunkar and J. Weaver, “Marriage markets and the rise of dowry in India,” J. Dev. Econ., vol. 164, 2023.
[19] F. Bloch, V. Rao, and S. Desai, “Wedding celebrations as conspicuous consumption: Signaling social status in rural India,” J. Human Resources, vol. 39, no. 3, pp. 675–695, 2004.
[20] O. Lewis, Village Life in Northern India: Studies in a Delhi Village. Urbana, IL, USA: Univ. of Illinois Press, 1958.
[21] M. N. Srinivas, The Remembered Village. Berkeley, CA, USA: Univ. of California Press, 1976.
[22] Y. Yan, The Flow of Gifts: Reciprocity and Social Networks in a Chinese Village. Stanford, CA, USA: Stanford Univ. Press, 1996.
[23] V. A. Zelizer, The Social Meaning of Money. New York, NY, USA: Basic Books, 1994.
[24] I. Guérin, “Juggling with debt, social ties, and values: The everyday use of microcredit in rural South India,” Current Anthropol., vol. 55, no. S9, pp. S40–S50, 2014.

PEER REVIEW & PUBLICATION HISTORY
Block Stage / Status Event / Action Timestamp
BLK-001 Editor decision: Accepted Editor — decision recorded 20 Jul 2026 10:14
BLK-002 All Reviews Received Reviewer — Accept 20 Jul 2026 10:12
BLK-003 Under Peer Review Reviewer — Invited & Accepted 20 Jul 2026 10:08
BLK-004 All Reviews Received Reviewer — accept 19 Jul 2026 17:19
BLK-005 Under Peer Review Reviewer — Invited & Accepted 18 Jul 2026 19:02
BLK-006 Reviewer Invitations Sent Initial reviewer screening passed 18 Jul 2026 16:14
BLK-007 Reviewer Invitations Sent Initial reviewer screening passed 18 Jul 2026 16:14
BLK-008 Reviewer Invitations Sent Initial reviewer screening passed 18 Jul 2026 16:14
BLK-009 With Editor in Chief Initial editorial screening passed 18 Jul 2026 16:12
BLK-010 Screening Editors Screening Editors 18 Jul 2026 16:08
BLK-011 Plagiarism Check iThenticate Plagiarism Check 18 Jul 2026 16:08
— Submitted Manuscript submitted / received via online portal 18 Jul 2026
Immutable audit trail (blockchain-anchored). Each block records a verified step in the peer-review workflow; blocks are listed most-recent first. Received 18 July 2026; accepted 20 July 2026.
© 2026 The Author. Published by Datarsoft Publishing House (Datarsoft Tech Private Limited). Open-access article under CC BY-NC 4.0.