How lighthouses became the first example of a public good — and what economists got wrong

In 1698, an English engineer named Henry Winstanley began building a lighthouse on the Eddystone Rock, fourteen miles south of Plymouth, at his own expense. The rock barely cleared the water at low tide. There was no precedent for a permanent structure there. What Winstanley had was a private arrangement that included statutory rights to collect dues from passing vessels when they docked in port — one penny per ton, levied at the bottleneck where ships and records converged.

Over the five years before the Great Storm of November 1703 destroyed the lighthouse and killed Winstanley inside it, those dues totalled roughly £4,721. His costs — construction, two rebuilds, maintenance — came to roughly £7,814. He hadn’t recovered his investment. But something else had happened that would matter far more to economic theory: private provision of lighthouse services, with a functioning revenue model, had been demonstrated in practice. The “impossible” thing was already working before most of the economists who would later cite the lighthouse as their canonical illustration of impossibility were born.

They never checked.

What every economics student learns

The standard public goods argument begins with Paul Samuelson’s 1954 paper, “The Pure Theory of Public Expenditure,” which provided the formal apparatus: a good is pure public if consumption by one person does not reduce its availability to others (non-rivalry) and if no one can be excluded from its benefits (non-excludability). Lighthouses sat at the centre of this framework. You couldn’t stop ships from seeing the beam. You couldn’t bill a ship at sea. Therefore private provision couldn’t recoup its costs. Therefore the market would under-provide. Therefore government must step in.

But the argument is older than Samuelson’s formalization. John Stuart Mill put lighthouses in Principles of Political Economy in 1848: “since it is impossible that the ships at sea which are benefited by a lighthouse, should be made to pay a toll on the occasion of its use,” the state must build and maintain them. Henry Sidgwick, in the 1883 first edition of his Principles of Political Economy, picked up Mill’s lighthouse and generalised it to cover “a large and varied class of cases” where private exchange fails to capture collective benefit. A.C. Pigou put it in The Economics of Welfare in 1920 as a classic instance of the uncompensated positive externality — a service generating social benefit that private revenue cannot capture. Samuelson absorbed it into his introductory textbook, Economics, and it stayed there through the 19th edition with Nordhaus in 2009.

More than 160 years. No one in that entire chain of transmission, from Mill through Samuelson, went to look at an actual lighthouse.

The theoretical argument itself is coherent enough to have made the oversight easy. A ship at sea encounters a lighthouse beam. The beam illuminates the coast whether the ship is there or not; one ship’s use doesn’t diminish it for the next. No collection can occur at sea — there’s no way to identify which ships benefited, no mechanism to intercept them, no billing process compatible with maritime reality. So the market can’t price it. So the market won’t build it. The logic is tight, the conclusion seems to follow, and the lighthouse seemed like a clean empirical illustration of an abstract principle. Nobody looked too carefully at whether the illustration was accurate.

The public goods taxonomy

Samuelson's binary — private goods (excludable, rival) versus public goods (non-excludable, non-rival) — was always a simplification. James Buchanan's 1965 paper "An Economic Theory of Clubs" (Economica, Vol. 32) introduced club goods: excludable but non-rival, consumed collectively by members who can screen out non-payers. A toll road, a streaming service, a private club — excludable by design but not depleted by individual use. Elinor Ostrom's Governing the Commons (Cambridge University Press, 1990) added a third category: common-pool resources that are rival but not easily excludable, and that communities have historically managed without either privatisation or government control. The lighthouse debate turns partly on which category lighthouses actually occupy — Coase would later argue they sit closer to Buchanan's club goods, and that reclassification is central to the full argument.

A crack visible from the theory alone

The non-excludability argument had a structural weakness that didn’t require archives to notice. It required only following the theory’s own logic one step further than Mill did.

Ships cannot be charged at sea. Correct. But ships are also, in almost every practical case, heading to or from a port. Ports are precisely where ships can be identified, their tonnage verified, their route reconstructed. A per-ton charge collected when a vessel docks — levied by port authorities, remitted to the lighthouse operator — is exactly the kind of collection mechanism the theory assumed was impossible. Light dues of precisely this structure had been standard instruments of maritime administration in England since the sixteenth century.

The argument that lighthouses were non-excludable didn’t rest on evidence that no collection mechanism existed. It rested on an implicit assumption that charging must occur at the point of use — at sea — which is indeed impossible. But collection doesn’t have to happen at sea. It has to happen somewhere ships can be tracked, and ports are that somewhere. The theory treated “unchargeable at sea” as equivalent to “unchargeable,” and nobody in 150 years of citations paused to ask whether port-based collection might invalidate that inference.

The example wasn’t cited because someone investigated English lighthouse history and found it confirmed the theory. It was cited because the theory seemed to require that private provision couldn’t work, and the lighthouse seemed like a plausible illustration of that conclusion. The illustration was accepted as evidence without being checked as evidence. The empirical claim — that no working collection mechanism existed — was treated as already settled by the theoretical framework, not as something that needed verification. This is a different kind of error from getting a fact wrong. It’s using the theoretical conclusion as a filter that prevents the empirical question from being asked at all.

Ronald Coase, in 1974, asked the obvious question.

Coase goes to the archives

Coase’s paper, “The Lighthouse in Economics,” published in the Journal of Law and Economics, did something none of his predecessors had bothered to do: it went to the primary sources. Parliamentary records, Trinity House accounts, English lighthouse history from the 17th century forward. What Coase found was that the textbook story was simply wrong.

Private entrepreneurs had built and operated lighthouses in England from at least the 17th century onward. They did so under letters patent from the Crown, which granted rights to collect light dues from vessels passing specific stretches of coastline. Those dues were levied at port — a per-ton charge, collected by port authorities, remitted to the lighthouse operator. The dues were not voluntary and the rates were not negotiated. But they were collected — and they were sufficient to sustain private investment in structures that the theory insisted couldn’t attract it.

The Eddystone case illustrates the mechanism. After Winstanley’s death in 1703, John Lovett acquired the lease of the rock. The Eddystone Lighthouse Act of 1705 formally established his right to collect one penny per ton from vessels that had passed the lighthouse, collected when they docked. John Smeaton’s stone tower — commissioned 1756, completed 1759 — became one of the celebrated engineering works of the century, financed by dues that Winstanley had proved could be collected. Lighthouse rights were bought, sold, and bequeathed for commercially meaningful sums, indicating that private actors genuinely valued the income streams they generated.

The 1836 Lighthouse Act, which consolidated English and Welsh lighthouses under Trinity House, is routinely read as the moment the market failed and the state rescued it. Coase’s reading is different: Parliament intervened not because private provision had broken down but because the patchwork of private patents had produced an inconsistent tangle of dues rates, with ships paying different amounts for equivalent services depending on which stretch of coast they passed. The 1836 Act was administrative rationalisation, not market rescue.

The implication for the textbook example was stark.

The free-rider problem that economists had insisted made private provision theoretically impossible had in fact been operationally addressed, imperfectly and with state backing, for roughly two centuries before Parliament intervened. The canonical example of market failure was historically a case of private provision.

Trinity House

Trinity House was founded by Royal Charter in 1514 by Henry VIII, originally as a fraternity of mariners concerned with navigation and seamanship. Its transformation into a lighthouse authority was gradual and unplanned: over the 17th and 18th centuries it held some lights directly, collected dues on behalf of private patentees on others, and served as the Crown's intermediary in the complex business of maritime charities and navigational aids. By 1836, Trinity House was the obvious administrative vessel for consolidation — not because it had been running a public service in any modern sense, but because it had accumulated the institutional knowledge and legal standing to do so. Its history illustrates, better than any theoretical argument, that the line between "private" and "public" lighthouse provision had always been a spectrum, not a binary — and that the binary was always a fiction of the economists who needed a clean example, not a description of the actual system.

The problem with Coase’s correction

So far, Coase’s argument seems decisive: the textbook example was historically wrong, private provision worked, economists should stop citing the lighthouse as evidence of market failure. That conclusion is what the paper is most often cited for — in libertarian critiques of market failure theory, in law and economics scholarship, in public choice literature. It’s also the wrong conclusion.

The lighthouse dues system was not a market. A market, in the relevant sense, requires that buyers can choose whether to purchase, that sellers can compete for buyers, and that prices emerge from some interaction of supply and demand. None of these conditions held for English lighthouse dues.

Crown patents granted specific operators monopoly rights over specific stretches of coastline. A ship that sailed past the Eddystone lighthouse owed John Lovett’s heirs one penny per ton when it docked in Plymouth. It had no alternative lighthouse. It had no alternative route. It had no opt-out. The dues were mandatory levies enforced by port authorities under statute, not payments made by willing buyers to competitive sellers. David van Zandt, writing in the Journal of Legal Studies in 1993, called Coase’s government/private binary too crude: the actual system was a public-private hybrid that neither label captured. Elodie Bertrand, in the Cambridge Journal of Economics in 2006, was more direct: the feature that made English lighthouse provision work was not private entrepreneurship but the state machinery that made collection enforceable. Without crown patents, without parliamentary confirmation of dues rights, without port authorities legally obligated to collect and remit, the revenue model Coase documented would not have existed.

Coase’s counter-argument is that all private property depends on state enforcement of rights — so the fact that crown patents required state authority doesn’t make lighthouses “non-private” in any interesting sense. But this conflates two categorically different things. State protection of a competitive market — enforcing property rights between competing actors who can win or lose customers — is not the same as state creation of a monopoly, where one actor receives the exclusive statutory right to tax a captive customer base with no recourse and no alternative.

The difference is concrete. In a taxi market, the state enforces contracts, prevents theft, licenses vehicles for safety. That’s state support for competition. In a taxi franchise, the state grants one company the exclusive right to operate taxis in a given city — no competition, no price discovery, no exit option for passengers who dislike the service or the price. The English lighthouse dues system was the franchise, not the market. In a franchise, the fee is whatever the monopolist can extract from a captive customer base. In a market, the fee is whatever competition constrains it to. These are not two points on a spectrum; they produce categorically different price levels, and different incentives to maintain what you’ve built.

Dan Bogart, Oliver Buxton Dunn, Eduard J. Alvarez-Palau, and Leigh Shaw-Taylor examined English lighthouse efficiency directly in a 2022 paper in Economic Inquiry. Their finding: lights under more private control charged higher fees and had higher operating costs, while lights with greater local representation and funding were cheapest to maintain. That’s monopoly rent extraction, not the efficiency dividend a competitive market would produce. The “private” provision Coase documented was generating outcomes consistent with statutory monopoly — because it was statutory monopoly.

Both the textbook story and Coase’s correction are wrong. The textbook story says private provision was impossible; Coase shows it existed. Coase’s counter-story says private provision worked by market mechanisms; the evidence shows it worked by state-enforced monopoly. Neither story fits the actual history.

What the discipline does with history

What Coase found is unsettling enough. What happened after he published it is worse.

Coase’s 1974 paper appeared in a peer-reviewed journal. It was written by a scholar who would later win the Nobel Prize. It directly addressed the most-cited example in introductory public economics and found the example to be historically wrong. The correction received attention in law and economics scholarship.

In the broader economics curriculum, it had no discernible effect.

Paul Samuelson continued citing the lighthouse in every subsequent edition of Economics, through the 19th edition with William Nordhaus in 2009 — thirty-five years after Coase’s correction was published. The textbook that had transmitted the example for five decades kept transmitting it after the correction appeared. The lighthouse example is still in introductory economics textbooks today.

The transmission chain — Mill (1848), Sidgwick (1883), Pigou (1920), Samuelson’s textbook from at least 1961 onward — moved the lighthouse forward not as a verified claim but as inherited authority. Nobody in that chain consulted an archival record. Each economist cited it because a respectable predecessor had. That is a peculiar epistemic practice for a discipline that presents itself as empirical.

But the graver charge is not that nobody checked for 125 years. It’s that after Coase did check and published a direct correction, the textbook tradition continued unchanged. The failure is not ignorance. It’s indifference.

Nobody retired the lighthouse example over a question of historical accuracy, because the example is too useful to discard. It explains a concept cleanly, it’s vivid, and it has been used so long that questioning it reads as pedantry rather than scholarship. The textbook canon doesn’t discard examples when they turn out to be wrong. It discards them when better examples become available.

The lighthouse is not an isolated case. The same structure — theoretical example adopted as canonical illustration, transmitted by authority, found wrong when checked — appears in at least two other foundational disputes in introductory economics.

That is what the economics textbook canon does with history. It uses history as illustration, not as evidence — and when the illustration turns out not to match the history, the illustration survives.

Other contested canonical examples

The barter-origin-of-money narrative has appeared in introductory economics textbooks since Adam Smith's Wealth of Nations (1776), which described primitive societies trading arrows for venison before money simplified exchange. David Graeber's Debt: The First 5,000 Years (Melville House, 2011) argues that there is no archaeological or anthropological evidence for barter-based pre-monetary economies — that what historians and anthropologists actually find is gift exchange, tribute, and credit, with commodity barter appearing mainly between strangers and enemies, not as the foundation of economic life. Economists had asserted the barter-to-money sequence without examining the historical record; when the record was examined, the sequence didn't appear.

Garrett Hardin's "The Tragedy of the Commons" (Science, 1968) introduced what became another canonical illustration: that commonly held resources are inevitably destroyed by individual self-interest, requiring either privatisation or government control. Elinor Ostrom spent a career examining how actual communities managed actual commons — irrigation systems in Spain and Japan, lobster fisheries in Maine, mountain meadows in Switzerland — and documented that Hardin's "tragedy" was not the historical norm. Communities had developed rules, enforcement mechanisms, and institutional structures that sustained commons over centuries without collapsing into Hardin's predicted outcome. Ostrom won the 2009 Nobel Prize in Economics for this work. The textbook version of Hardin's argument remains in common use.
The lighthouse, the barter myth, and the tragedy of the commons share a structure: a theoretical example adopted as canonical illustration, transmitted by authority rather than verification, and found to be historically questionable when someone eventually looked. None of the corrections required exotic methodology. They required checking.

The current lighthouse on the Eddystone Rock was built in 1882. It is operated by Trinity House, a public body. It has no tollbooth. No ship passing it owes dues to anyone. It looks exactly like what economists always said a lighthouse should look like: publicly maintained, universally accessible, free at the point of use.

Neither account explains how it got there. The textbook needed the lighthouse to demonstrate that markets fail; Coase needed it to demonstrate that they don’t. What the history actually shows is that private provision came first, ran on statutory monopoly rather than market competition, and that Parliament’s 1836 rationalisation produced a public institution that inherited infrastructure built under a system neither account had described accurately.

The lighthouse still marks the same rock. Still useful. The justification for who provides it, and why, has simply drifted loose from the facts of its own history — and kept getting cited anyway.

If this is what happened to the most widely-taught, most extensively-debated example in the theory of public goods, it is worth asking how many of the other canonical illustrations in introductory economics have been examined with anything like the same rigour.

Gen AI Haftungsausschluss

Einige Inhalte dieser Seite wurden mit Hilfe einer Generativen KI erzeugt und/oder bearbeitet.

Medien

Susanne Jutzeler, suju-foto – Pexels

Wichtige Quellen und Referenzen

Ronald Coase, “The Lighthouse in Economics,” Journal of Law and Economics, Vol. 17, No. 2, pp. 357-376, 1974.

Paul Samuelson, “The Pure Theory of Public Expenditure,” Review of Economics and Statistics, Vol. 36, No. 4, pp. 387-389, 1954.

John Stuart Mill, Principles of Political Economy, John W. Parker, London, 1848.

Henry Sidgwick, The Principles of Political Economy, Macmillan, London, 1883.

A.C. Pigou, The Economics of Welfare, Macmillan, London, 1920.

David van Zandt, “The Lessons of the Lighthouse: ‘Government’ or ‘Private’ Provision of Goods,” Journal of Legal Studies, Vol. 22, No. 1, pp. 47-72, January 1993. https://chicagounbound.uchicago.edu/jls/vol22/iss1/4/

Elodie Bertrand, “The Coasean Analysis of Lighthouse Financing: Myths and Realities,” Cambridge Journal of Economics, Vol. 30, No. 3, pp. 389-402, 2006. https://academic.oup.com/cje/article/30/3/389/1730007

Dan Bogart, Oliver Buxton Dunn, Eduard J. Alvarez-Palau, and Leigh Shaw-Taylor, “Organizations and Efficiency in Public Services: The Case of English Lighthouses Revisited,” Economic Inquiry, Vol. 60, No. 2, pp. 975-994, 2022. https://onlinelibrary.wiley.com/doi/10.1111/ecin.13060

James M. Buchanan, “An Economic Theory of Clubs,” Economica, Vol. 32, No. 125, pp. 1-14, 1965.

Elinor Ostrom, Governing the Commons: The Evolution of Institutions for Collective Action, Cambridge University Press, 1990.

Garrett Hardin, “The Tragedy of the Commons,” Science, Vol. 162, No. 3859, pp. 1243-1248, 1968. https://www.science.org/doi/10.1126/science.162.3859.1243

David Graeber, Debt: The First 5,000 Years, Melville House, New York, 2011.

Paul Samuelson and William Nordhaus, Economics, 19th edition, McGraw-Hill, 2009.