if the factory holds the right
—
—
—
the second hundred · metaphor 196
When two people can actually talk, does it matter who started out in the right? Or does the quarrel land in the same place no matter who owns the grievance — with the only thing that ever really mattered being how hard it is to strike the deal?
We reach for ownership as if it settled things. It's my right, you started it, the rule is on my side — as though establishing who holds the entitlement also fixes the outcome. And under one condition it doesn't. If the two of you can bargain at no cost, you will trade your way to whatever arrangement makes the pie biggest, and then split the pie. Who held the right decides who pays whom — but not what finally gets done.
The catch is the phrase at no cost. Real bargaining has friction — lawyers, distrust, time, the sheer effort of a conversation nobody wants to have. When that friction is small, ownership is almost beside the point; the efficient deal happens either way. When the friction is large, the deal dies, and now whoever happened to hold the right simply keeps the status quo. The allocation stops being a footnote and becomes the whole ballgame. It was never really about the right. It was about the cost of settling.
The invariance
Think of the disputed activity as a dial — how much the factory runs, how loud the party is, how far the fence extends. Turning the dial up helps one party and hurts the other, so somewhere there is a setting that makes the total gain largest: the point where one more unit of help to the offender is exactly cancelled by the harm it does the neighbour. That's the efficient level, and it depends only on the two parties' real values — not on who holds the right.
Now hand the right to one side. That fixes the starting point, not the destination. If the factory may run freely, it starts by running too much; the resident, hurt more than the factory gains at that margin, will pay it to ease off — and both come out ahead by moving to the efficient level. If instead the resident holds a right to quiet, the factory starts shut, and now the factory pays for permission to run up to that same level. Same dial setting, opposite cheques. Coase's point: with costless bargaining, the allocation of the right changes the distribution of money and nothing about the real outcome.
So when someone says the fight is about principle — who was wronged, who owns the entitlement — and the two sides can actually negotiate, the principle is quietly settling only the ledger. The world ends up the same shape either way.
What to try
Start at transaction cost = 0 and toggle the right-holder. The two comparison cells show identical outcomes: both land on q*. That is the theorem you can watch — the highlighted column moves, the result doesn't. Only the money in each cell's fine print flips direction.
Now raise the transaction cost. Nothing happens at first — small friction still leaves room to split the gains. Then you cross a threshold and one cell goes red: deal blocked. From here the two allocations diverge — whoever holds the right keeps the status quo, and the world's shape now depends entirely on who that was. Push the cost past the surplus for both sides and every path freezes at its starting point. The “harm exceeds value” preset flips the efficient answer to shut it down, and you can watch free bargaining reach even that — the resident buys the factory's silence — until friction makes that trade, too, impossible.
The mapping
The structure shows up wherever a right and a real cost sit on opposite sides of a table. A noisy neighbour: the law says who may do what, but if you're on speaking terms you'll settle on the arrangement that suits you both and trade a favour for it. A messy divorce: when lawyers are cheap and the couple can talk, assets flow to whoever values them most regardless of who's “entitled”; when the fight turns bitter and costly, the initial legal position sticks and value is destroyed keeping it there. Carbon and pollution: the whole case for tradable permits is Coasean — fix the total, let the right to emit change hands, and it migrates to where it's worth most, no matter who got the first allocation.
The lesson isn't that ownership never matters. It's where it matters. When talking is easy, arguing about who's in the right is arguing about the receipt. When talking is hard, that argument is the only thing standing between the world you have and the better one you can't reach. Fix the friction and the entitlement takes care of itself; leave the friction and no clever assignment of rights will save you.
Read as life lessons
When you can bargain freely, the entitlement decides who pays, not what gets done. Winning the argument about ownership can leave the actual outcome untouched.
The thing that blocks the good outcome isn't a bad allocation — it's the cost of the deal: distrust, ego, lawyers, the conversation nobody will start. Lower that, and the rest follows.
The theorem cuts both ways. Once bargaining is costly, the initial right sticks — so if you can't fix the friction, fight hard over who gets the entitlement, because now it's decisive.
The mapping, exactly
| Mathematics | Life |
|---|---|
| the activity level q | The disputed thing itself — how much the factory runs, how loud the party gets, where the fence sits. |
| total surplus W(q) | The combined well-being of both parties — the pie they are, knowingly or not, trying to make biggest. |
| the efficient q* | The arrangement that actually serves them both best, set by their real values, not by the rule book. |
| who holds the right | Who owns the grievance or the entitlement — the starting point, and the direction the cheque runs. |
| transaction cost T | The friction of settling: distrust, lawyers, time, pride — everything that makes the conversation expensive. |
| surplus > T ⇒ deal | The deal happens only when the gain from fixing things outweighs the pain of the negotiation. |
The honest model
Two parties, one dial q. The factory's benefit from running at level q is B(q)=a·q − q²/2 (each unit worth a little less than the last); the resident's harm is C(q)=c·q. Total surplus is W(q)=B(q)−C(q), a hump peaking at the efficient level q*=a−c — the panel finds it by maximising the real curve, not by quoting a formula. Left to itself the factory would run to q_F=a (ignoring the harm); a resident's right starts things at q₀=0.
A move from the status quo to q* raises the pie by exactly W(q*)−W(q₀) — the surplus at stake. The parties will pay up to that to make the move, so the deal goes through precisely when W(q*)−W(q₀) > T. When it does, both allocations end at the same q* (invariance); when friction beats the surplus, each allocation stalls at its own q₀ (divergence). Every number in the instrument — the peak, the two status quos, the stake, both verdicts — is recomputed live from the sliders. Nothing is scripted.
Where the metaphor tears
“Zero transaction costs” is a limit almost nothing reaches. Real disputes come loaded with distrust, missing information, and the sheer effort of talking to someone you're angry at. Coase's own point was the opposite of the slogan named after him: because friction is everywhere, the initial allocation of rights and the design of institutions matter enormously. The frictionless case is a foil, not a forecast.
Invariance is about the dial, not the wallet. Being handed the right versus having to buy it can be the difference between comfort and ruin — and if you're too poor to make the efficient payment, the “efficient” outcome may never be one you can actually reach. A theorem that's silent on distribution is silent on most of what makes these fights hurt.
The model treats the grievance as a quantity with a price. But some rights are refusals to trade at all — bodily safety, a child's welfare, a promise's meaning. Put those on the bargaining table and you haven't found the efficient level; you've changed what kind of thing they are. Where the good resists being priced, Coase has nothing to measure.