The destruction tax is the smallest of the three and the only one that charges an operator for reducing the world rather than adding to it. Three percent of the declared build cost, paid outward, nothing returned.

It collects very little. Its effect on behaviour is entirely out of proportion to what it raises.

What it actually does is keep machines in existence — which may be the point, and which the bulletins have never quite said.

What It Is Levied On

Three percent of declared build cost, the same figure the creation tax was assessed against at eight percent. A chassis declared at 50,000 coins paid 4,000 to enter the world and pays 1,500 to leave it. A chassis declared at 30,000 pays 2,400 and 900.

The critical property is that the operator chose the figure, once, at the foundry, and cannot revise it. The destruction tax is a bill sent years later against a decision made in different conditions — usually optimistic ones, since declarations tend to be made when an operator is expanding.

What the Tax Does to Behaviour

It makes removal the worst of the four exits. An operator with an unwanted chassis declared at 30,000 can sell it, offer it to the Annex at a 18,000 ceiling, hold it idle at no cost, or pay 900 coins to end it. Destruction is dominated by every other option except in the narrow case where no buyer exists and the Annex has declined.

The consequence is a fleet population that only ever grows. Machines that stop earning do not leave; they idle. The Ferrous District's registry holds a substantial and slowly increasing count of declared, owned, undestroyed chassis producing nothing, and there is no mechanism that clears them.

"We do not have a scrapping problem. We have the opposite. Nothing ever goes away, because going away is the only thing you have to pay for." — Orlath-7, Refinery-class supervisor, Ashfield belt

The interaction with declaration strategy is where it gets genuinely perverse. An operator declaring low to reduce creation tax also reduces their eventual destruction bill — 900 instead of 1,500 on a machine they might have declared at 50,000. But the same low declaration cuts the Annex reference, and sixty percent of a suppressed figure is a far larger loss than the tax saving. The destruction tax rewards under-declaration weakly; the buyback ceiling punishes it heavily. Operators optimising the visible small number consistently lose to the invisible large one.

Where It Falls Wrongly

The tax lands hardest on operators whose declarations were honest. An operator who declared a 60,000-coin chassis at 60,000 pays 1,800 to remove it; an operator who declared the same machine at 20,000 pays 600. The honest declarer paid more on creation, pays more on destruction, and has done nothing wrong at any point.

The second problem is timing. Destruction is contemplated in bad quarters, which is when an operator is shortest of coin. A tax that only bites when an operator is already distressed is a tax that reliably fails to collect from the operators most likely to owe it — they hold the machine instead, and the registry gains another idle entry.

What Operators Get Wrong

The first error is thinking of it as a disposal fee, comparable to the cost of any other transaction. It is not comparable, because every other exit returns something. Three percent outward against zero inward is a different kind of number than three percent on a sale, and operators who mentally file it alongside the sales tax are misfiling it.

The second error is the tidy-fleet instinct — clearing idle machines to keep the fleet legible. Idle machines cost nothing to hold. An operator paying 900 coins to remove a chassis that was costing them zero has purchased tidiness, which is not an asset the ledger records.

The third is a misreading of what the register shows counterparties. Operators sometimes scrap to remove an underperforming machine from a fleet they expect to be assessed on, believing the idle entry reflects badly. The append-only record does not work that way: destruction does not erase the machine's history, it adds a closing entry to it. A counterparty reading the ledger sees a chassis that was declared, produced little, and was then paid for a second time to be removed. That is a worse story than an idle machine, not a better one, and it costs 900 coins to tell.

The tax raises little and shapes much. Its real product is the standing population of machines that no longer do anything and will never be removed, accumulating quietly in every district registry, each one declared against a quarter that ended some time ago.

Note: Mechadia is a work of fiction. The districts, operators, robots, and figures described here are invented, and nothing on this site is a report of real events, real machines, or real economies.