Household money pots
A Socratic walk-through of household money pots — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why do people who share every meal keep separate pots of money and defend them against each other?
Two people share a bed, a kitchen, a mortgage and the raising of children. They will divide the last of the milk without a thought. And they keep separate current accounts, know exactly whose money paid for what, and can become genuinely angry about a transfer between the two.
The usual reading is that this is a residue of distrust, or a hedge against the marriage failing. But test the assumption underneath it: that a household with shared aims should behave like one decision-maker with one purse. Should it?
Reasoning it through
REASONING #Start with what a pot is, mechanically. It is not a store of value — money is fungible and the bank does not care which account it sits in. What differs is who decides: whose consent is needed to spend from it, and who has to explain afterwards.
That reframing does the work. Once money enters a common pot it stops carrying the mark of who earned it, and with that mark goes the earner's claim on how it is spent. Pooling is not neutral bookkeeping; it is a transfer of decision rights.
Does the identity of the holder actually change outcomes? If a household behaved as one decision-maker it could not — the same total income should produce the same basket regardless of whose hands it passed through. The classic tests are policy changes moving a payment from one parent to the other without changing the total. The British reform that shifted child support from a father's tax allowance to a payment made directly to the mother is the case usually cited, and the reported finding was a shift in what households bought, toward children's and women's goods. I recall that result rather than reading it and quote no magnitude — but the direction is the point, and it means the single-purse model is wrong as a description.
Several distinct jobs then come into view, worth separating because they predict different behaviour.
The first is claim preservation. Keeping money identified is how an earner keeps influence over household spending, which predicts separate accounting is most defended by whoever holds the weaker position elsewhere — less legal claim on joint assets, worse outside options.
The second is insurance against exit. Separate money survives dissolution, a partner's job loss, or a partner who begins controlling access to funds, so it is priced against the probability the arrangement ends. This predicts separation is commoner where dissolution is easier, or in remarriages where each partner's obligations point at different people.
The third is commitment. A pot named for something is harder for either party to raid, a device used against one's own future self as much as against the other. Naming a fund makes a future refusal easy: not "I would rather not", but "that is the boiler money".
The fourth is the cost of monitoring. A fully pooled account requires either accounting or trust in unaccounted spending; separate pots buy each person a zone with no explaining — cheaper than an audit and cheaper than the insult of proposing one.
None of these requires distrust. Three of the four are things two people who like each other would still build.
The analogy
THE ANALOGY #Think of two firms in a long-running joint venture. They fund a shared account for the project's costs, and each keeps its own treasury. Nobody reads that as suspicion; it is how each side keeps a say in the venture, the ability to walk if it fails, and a zone of spending it need not justify. The venture account is not evidence of union and the separate treasuries are not evidence of doubt — they are different instruments doing different jobs.
Firms write contracts specifying what happens on dissolution, whereas a household's terms are mostly unwritten and legally implied; and firms have no equivalent of the norm that accounting between intimates is itself an accusation, which is why some of the arrangement stays unspoken.
Clarifying the model
THE MODEL #A neighbouring idea needs distinguishing. There is a well-described phenomenon in which a single mind files money by source and purpose, then judges spending against that file's balance rather than against total wealth — which is why a named fund is harder to raid. That is attention inside one head, and the fixed point of difference is that its constraint is psychological and self-imposed. Here there is a second party, and the label does something the single-mind account cannot explain: it is enforceable, because someone else can object. A household pot is an institution rather than a mental file, defensible against a person rather than merely against an impulse.
What would show this account wrong? It claims arrangements track bargaining position and exit risk rather than sentiment, so it predicts pooling responds to the legal environment — rules on marital property, on what a separation divides — and to remarriage and relative earnings, more than to reported satisfaction. The refuting observation is clean: if pooling varied with reported closeness while staying flat across legal regimes, remarriage status and earnings gaps, the bargaining and insurance story is wrong.
And the caution about evidence. Couples choose their arrangements, and not at random: those who separate their money may already differ in commitment, in expectations, in the assets they arrived with. So a cross-sectional finding that separate finances predict instability is uninterpretable — the arrangement may be a symptom rather than a cause. Designs that get past this watch the same couple through a change nobody chose — a reform to marital property law, a change in how a benefit is paid — or compare first marriages with remarriages, where the claims structure differs for reasons other than affection. Absent those, the direction of causation is open, and I leave it open.
Lastly: all of this describes what arrangements do, not what anyone should adopt. These are population tendencies drawn from households in very different positions.
A picture of it
THE PICTURE #How to readThis repurposes a work-tracking board: the columns are not stages of a process but the pots themselves, and the cards are the jobs each pot performs rather than tasks to be moved. Nothing travels between columns — that is the point. Read down a column for what a household buys by keeping that pot distinct: shared obligations with no negotiation per purchase, or an unexamined zone plus an exit position plus a preserved claim on decisions, or a commitment binding both parties. The two personal columns are drawn identically on purpose; where a real household's are unequal, that inequality is the bargaining position the argument is about.
What became clearer
WHAT CLEARED #The separate accounts are not a residue of distrust that intimacy should have dissolved. They are the household's institutions, and each does a job a single merged purse cannot: keep an earner's claim on decisions legible, hold something that survives the relationship ending, make a future refusal easy by naming money in advance, and buy each person a zone of spending nobody has to justify.
The starting assumption was the error. A household is not one decision-maker, and we know it because moving a payment from one parent to the other changes what the household buys. Once there are two decision-makers with mostly-shared and partly-separate interests, a structure of pots is what you would expect them to build — and defending one is a claim about authority rather than a statement about love. What remains unsettled is the direction of causation between arrangement and outcome, because couples choose the arrangement and almost nobody has watched one change for a reason the couple did not pick.
Where to go next
ONWARD #- How arrangements shift when one partner's earnings fall sharply, and whether the pots are renegotiated or drained.
Key terms
TERMS #| Term | What it means |
|---|---|
| Income pooling | combining partners' incomes into a common fund rather than keeping them identified. |
| Earmarking | labelling a sum for a purpose, making it harder to spend on anything else. |
| Unitary household model | the assumption that a household acts as one decision-maker with one budget; contradicted by evidence that who receives income changes what is bought. |
Every term the collection defines is gathered in the glossary.