Six friends drive home from a weekend at a bach, and one of them is quiet in the passenger seat, doing sums on a phone. She paid for the house, and the petrol on the way up, and most of the groceries, because someone had to, and now she is working out who owes what and writing the messages: the friendly, slightly awful texts that ask five people she loves to send her forty dollars each. Nobody appointed her the bank. She just has the card details saved and a tolerance for being the one who asks, and so the entire coordination cost of six people’s weekend has quietly become one person’s Sunday evening.
It is tempting to read this as simply what being the organised one costs, a tax you pay for a personality trait, and to reach for the natural fix, a better tool. Split the bill in an app, log the expenses, pick one of the dozens that exist. The sums get easier, the messages get templated, and the quiet Sunday admin shrinks to something bearable. I held that reading for a long time.
It also misses where the cost comes from. The reason one person ends up holding a shared thing is not that the others are lazy or that her spreadsheet is bad. It is structural, and it has a name that is almost ninety years old.
The cost of doing it through the market
In 1937 Ronald Coase asked a question that sounds too simple to be famous: if the market is such an efficient way to organise activity, why do firms exist at all, instead of everyone contracting with everyone for each task as it arises? His answer was that using the market is not free. Every transaction carries a cost, the work of finding the other party, agreeing a price, writing and enforcing the terms, and when those costs get high enough it becomes cheaper to pull the activity inside an institution, a firm, where someone coordinates it by direction instead of by a thousand little bargains. (R. H. Coase 1937)
The part worth keeping is the question Coase forces you to ask about any coordination at all: who should hold this cost, the market, an institution, or someone else? Hold that question, and a whole class of ordinary problems looks strange, because the answer they have landed on is none of the above. The cost has been left on individuals, each absorbing privately a thing that no market cleared and no institution was ever built to hold.
The same shape, six times
The friend in the passenger seat is the smallest version. Scale it up and the shape repeats.
Pay someone back, and the cost of moving money is either absorbed once by a shared rail or privatised as a fee and a delay onto everyone who moves it. Brazil built the rail and India grew one: Pix is run by the central bank (Banco Central do Brasil 2020), UPI by a nonprofit consortium of banks (National Payments Corporation of India 2016), and nobody thinks about either any more than you think about the road when the bus is on time. New Zealand left money to the card networks and bilateral deals until a mandate arrived in late 2025 (Ministry of Business 2025), and in the meantime the Commerce Commission put the bill for card fees at around $1 billion a year, the running cost of a rail nobody built. (RNZ 2025)
Share a flat, and the money is not complicated, but with three or four people on three or four different banks and no shared account, the coordination has nowhere to live but inside one flatmate, who becomes the institution the flat never formed.
Own something, and the proof that it is yours stays scattered across the people who sold it to you, so you reassemble it alone, from scratch, at the three moments it turns urgent: a claim, a sale, and a handover after a death.
Put money in a KiwiSaver fund, and you own a slice of thousands of companies, more than 4,000 investments through one widely held growth fund (Simplicity 2026), and the running story of what you own is told to no one, so each owner privately carries the whole cost of understanding a thing they were told they need not.
Shop for groceries in a market two companies mostly control, a market where the Commerce Commission has concluded that “competition is not working well for consumers” (Commerce Commission 2022), and the true price is never printed, so every shopper runs privately, in their head, the comparison the market has arranged to be impossible.
Try to get ready for the disaster the country keeps promising, and the official advice hands you a category, keep three days of supplies (Get Ready (NEMA) 2026), and leaves you to translate it into your own household, which is the exact step at which most people give up. Only about two in five New Zealanders feel well prepared for a disaster. (National Emergency Management Agency 2022)
Every one of these is a coordination cost that a shared layer could absorb once, for everyone, and that instead sits on individuals because the layer was never built. It is one law in six costumes. When a coordination cost has nowhere institutional to live, it does not disappear. It gets privatised onto whoever is standing closest, paid in the least visible currency there is: time, attention, and a low background hum of being the responsible one.
The layer that answers for itself
So build the layer. That is where every one of these ventures points, and it is the right instinct, and it is also where I want to be careful, because a shared layer is not a free good either. Dan Davies gave the sharp version of the danger in 2024: a system built to absorb decisions becomes what he calls an accountability sink. (Dan Davies 2024) Once a responsibility is handed to a rulebook or a platform, no particular human owns the outcome any more, and the person it fails ends up arguing with a process designed so that nobody has to answer them. The same centralisation that lifts a coordination cost off your shoulders can hand you back a worse one: not the work of chasing your flatmate, but the work of being on hold to a company that will not tell you why the payment did not land. The cost is not gone. It has changed from labour into powerlessness.
And whoever builds the missing layer well does not merely relieve the cost, they capture the position. Ben Thompson’s distinction is the one to hold here: a platform lets suppliers keep their own customers, while an aggregator owns the customer relationship and commoditises everyone beneath it. (Ben Thompson 2015) The thin middle that nobody built becomes, the moment someone builds it well, the most valuable seat in the market. Which is why the honest question is never only whether the layer exists, but who holds it, and what they are allowed to do from the middle.
Which structure should hold it
Elinor Ostrom spent her life on that question. Against the economists who assumed a shared resource must be privately owned or state-run or it will be wrecked, she showed that communities repeatedly build durable institutions to hold common problems themselves, given the right design: clear rules, real monitoring, a way to settle disputes, and accountability that lands on someone. (Elinor Ostrom 1990) The word that matters in her work is institution, not app. A shared layer is not a piece of software that appears once the engineering gets cheap enough. It is a set of rules about who is responsible, and that is the unglamorous part, and it is precisely the part that gets skipped.
Which is why these costs stay on individuals long after the technology to lift them exists. The hard part was never the building. It is deciding who answers when the shared thing fails, because the moment you build the layer you inherit the accountability the individual used to carry alone, and almost nobody wants to sign for that. So the cost sits where it always sat, on the person closest to it, not because the layer is impossible but because leaving it unbuilt keeps the blame comfortably nowhere. The friend in the passenger seat is not carrying a problem with no solution. She is carrying the one thing a solution would have to take responsibility for, and until something does, it stays exactly where it is, in her lap, on a Sunday, one message at a time.