Who pays for the account nobody keeps

A century ago, in The Economics of Welfare, Arthur Pigou gave economics its most durable diagnosis of why markets produce the wrong amount of things. His example was smoke: a factory burning coal imposes a cost on the households around it — soot, laundry, illness — that never appears on the factory's own books. The factory weighs its private cost against its private benefit and, on that arithmetic, runs the furnace at full tilt. But the true cost to society is larger than the cost to the factory, by exactly the amount that lands on the neighbors. The factory produces too much smoke not because anyone is villainous but because it faces only part of the cost of its own decision. Pigou's word for the part that lands elsewhere was the externality, and his insight was that whenever the person who acts does not bear the full cost of acting, the market will get the quantity wrong — every time, and in a predictable direction.

Hold that structure against the decision not to keep an account. An institution decides — approves, denies, flags, ranks — and faces a choice about whether to record how and on what. Recording costs something: engineering, storage, discipline, the friction of building a decision that can be reconstructed instead of one that merely fires. Skipping it saves all of that, now, visibly, on the institution's own books. And the cost of having skipped it? That lands almost entirely on someone else. It lands on the person the decision fell on, who cannot contest what was never recorded. It lands on the regulator who cannot examine what was not kept. It lands on the future — the next audit, the next dispute, the next version of the institution that inherits a decade of decisions it can no longer explain. The saving is private and immediate. The cost is external and deferred. This is Pigou's factory exactly, and it produces exactly Pigou's result: too little of the account, systematically, for reasons that have nothing to do with anyone's character.

Why exhortation cannot fix a mispriced good

The reason this matters is that it tells you which cures will fail. When a shortfall is a matter of virtue, you address it with virtue: you exhort, you train, you appeal to professional pride. When a shortfall is an externality, virtue is beside the point — because the actor producing too little is already behaving rationally given the costs he faces, and asking him to behave irrationally against his own incentives is asking for a favor, not a fix. You will get compliance from the conscientious and nothing from everyone else, and the market as a whole will stay exactly where the incentives put it. A factory does not stop making smoke because you remind it that smoke is bad. It stops when the smoke starts appearing on its own books.

This is why a decade of asking institutions to be more transparent, more explainable, more accountable, has moved the aggregate so little. The asking is aimed at a virtue problem. The actual problem is an externality: the institution that skips the account is not confused about the value of accountability, and it is not wicked. It is responding, correctly, to a price structure in which the account costs it something and its absence costs it nothing — because the absence is paid for by parties who were not in the room when the trade was made and have no way to send the bill back. You cannot moralize your way out of that. You can only re-price it.

The institution that skips the record is not confused about the value of an account. It is responding correctly to a price in which keeping one costs it something and skipping one costs it nothing — because someone else pays.

Making the decider internalize the cost

Pigou's own remedy was the tax that bears his name: a charge on the polluter equal to the cost it imposes on others, so that the external cost becomes an internal one and the factory, now facing the true cost of its smoke, chooses the socially right amount without being asked to be a saint. The general principle is broader than the tax, and it is the one that matters here. To fix an externality you do not appeal to the actor. You change what the actor faces, so that the cost he was exporting comes home. Once the full cost of a decision is on his own books, the accounting that led him to skip the record inverts on its own, and he keeps it — not out of newfound virtue, but out of the same self-interest that led him to skip it before.

What does internalizing look like for the missing account? It looks like every mechanism that puts the cost of unaccountability back on the party that chose it. It looks like a legal standard under which a decision that cannot be reconstructed is presumed indefensible — so the absence of a record is a liability the decider carries, not a saving it banks. It looks like a procurement rule under which a vendor who cannot produce an account does not get paid, moving the cost of opacity from the buyer back to the seller who created it. It looks like a burden of proof that, when a machine decision is challenged, sits with the institution to show the decision was sound — which it can only do from a record, so the cost of not having one falls, at last, on the one who declined to keep it. Each of these is a Pigouvian move in disguise. None of them asks anyone to value accountability more. They arrange things so that the party who saves by skipping the account is also the party who pays for its absence — and once those two are the same party, the shortfall closes itself.

The deep point is that accountability is not undersupplied because the world underrates it. It is undersupplied because its cost and its benefit sit on different balance sheets — the saving on the decider's, the cost on the affected party's and the future's — and a good split that way is undersupplied by the same iron logic that overproduces smoke. The tax on opacity is real, but for most of its life it is a tax someone else pays, which is precisely why it fails to discipline the decision that incurs it. The work of governance here is not to preach the value of the record. It is to re-route the bill — to make sure that the party who enjoys the saving of skipping the account is the same party who receives the invoice for its absence. Do that, and you will not have to argue for accountability at all. It will be the cheaper option for the one who decides, which is the only place the argument was ever going to be won.

— Dispatches · Summit Cognitive

---