#1406 1610 · Corporation of Trinity House and private lighthouse patentees · Maritime aids to navigation
English lighthouses ran privately for centuries: free at sea, billed at the port
问题
A lighthouse cannot bill a passing ship, so economists from Mill to Samuelson said no private builder could profit
背景
For over a century the lighthouse was economics' favorite impossible business: John Stuart Mill, Sidgwick, Pigou and Samuelson all used it as the canonical good that only the state could provide, because the keepers 'cannot reach out to collect fees from the skippers' of passing ships. The example did real analytical work: it stood for every service whose benefit cannot be denied to non-payers.
When Ronald Coase examined the actual records, the opposite emerged. Trinity House — the seamen's guild chartered by Henry VIII in 1514 — built no lighthouse at all from 1610 to 1675. Private promoters did, obtaining Crown patents (later Acts of Parliament) to build lights on their own land and levy tolls on ships 'presumed to have benefited', often with supporting petitions from shipowners and masters who had paid for wrecks. Even the Eddystone lighthouse was private: Winstanley's tower of 1696 (which took his life in the 1703 storm), then a 99-year lease that left 100 percent of profits to its private builders from 1709.
换别人会怎么做
Declare it a public good and wait for a ministry to fund it — England's lights got built two centuries before any ministry took them on, because a patent defined the toll and the port enforced it.
他们看到了什么
You cannot sell light at sea — but you never had to. Every ship must dock, and at the dock a toll per ton is cheap to collect. Price the unpriceable service at the chokepoint no user can skip.
那一手
The patent fixed the price — a toll per ton per voyage for each lighthouse passed, e.g. 1.25 pence or 0.25 pence per ton — but collection never happened at sea. It happened at the ports, where private agents, usually the customs officers stationed there anyway, collected the toll as ships docked. Published books told each master which lights his route passed and what he owed, making the liability predictable and auditable per voyage. Trinity House often held the patent and leased the building to private capital; by 1820, of 46 lighthouses in England and Wales, only 11 had been built by Trinity House itself and 34 by private promoters.
为什么管用
The patent fixed price and scope in advance, so nothing had to be negotiated with a moving ship; the customs officer at the port was already stationed there, making the marginal cost of toll collection near zero; the published toll books made each voyage's liability computable before departure, which is what made shippers accept it; and demand came from the payers themselves — owners petitioned for lights because grounding losses dwarfed the tolls. Trinity House's lease structure (a fixed rent to the guild, all profits to the builder) put construction risk on private capital while the guild monetized its monopoly right. The economics were durable enough that the state needed nearly 1.2 million pounds to buy the private lights out.
值了多少
By 1820, 34 of 46 English and Welsh lighthouses were privately built; buying them out (1836-42) took six years and nearly 1.2 million pounds
什么时候会失灵
The toll was set by patent, not by a market, so monopoly patentees overcharged captive traffic and complaints accumulated until the 1834 committee forced consolidation. Disputes arose over which lights a route had actually passed, and ships using unpatrolled ports could free-ride. Chokepoint billing only works where usage predictably flows through chokepoints you can cover — the lighthouse worked because every ship must dock where customs sits.
后来呢
Coase's 1974 paper made the lighthouse the standing counterexample to 'public goods require public production' in economics teaching; UK aids to navigation are still funded by light dues paid by shipping at ports into the General Lighthouse Fund, with shipowner committees scrutinizing the budget.
资料来源
- [1]The Lighthouse in Economics (Journal of Law and Economics 17(2), 357-376)University of Chicago Press, 1974gwern.net
- [2]Coase and Transaction Costs Reconsidered: The Case of the English Lighthouse SystemEuropean Journal of Law and Economics (Candela & Geloso), 2019rcandela.com