The Market and the Lighthouse.
Public Goods in Historical Perspective.
Erik Lindberg, Associate Professor
Department of History
Box 628, 751 26 Uppsala
Paper delivered at the
Economic History Society Annual Conference, 3 – 5 April 2009, University of Warwick
Ronald Coase’s controversial paper “The Lighthouse in Economics” is widely cited as demonstrating that private enterprise can produce public goods. In this paper I re-examine Coase’s claims and add empirical evidence from two national lighthouse systems between 1600 and 1910 and inquire under what circumstances private ownership might be preferable to public, and why private enterprise in the provision of public goods in the end was largely abolished both in Britain and in Sweden in the nineteenth century.
The Market and the Lighthouse. 1
Public Goods in Historical Perspective. 1
Coase on the lighthouse 2
The interpretation of Coase in economics 3
The problem of the lighthouse in economics 6
The organization of the lighthouse services in theory 7
Lighthouse systems in practice: England and Sweden 14
Conflicts of interest 15
Contractual forms 18
The transition from “private” to “public” provision of lighthouse services 21
Conclusion: Lessons from the lighthouse 22
Coase on the lighthouse
Ronald Coase’s controversial paper "The Lighthouse in Economics" is widely cited as demonstrating that private enterprise can produce public goods. His article was an attack on the perceived wisdom about the nature of public goods and a reassessment of what governments should do and what markets could achieve. He quotes several famous nineteenth and twentieth century economists who had used the lighthouse as an analytical tool in discussing the proper role of government. J.S. Mill, Henry Sidgwick, A.C. Pigou and Paul Samuelson: all had emphasised that the lighthouse is a quintessential public good. A lighthouse is in economic theory defined as a public good because the difficulties of obtaining payment for its use. The light emitted from the lighthouse warns all ships sailing on the sea, regardless of whether they pay for its service or not. Since ships can take advantage of the information sent out by the lighthouse without coming into contact with the lighthouse operator, much less paying him for the service, this would discourage private entrepreneurs from building and maintaining lighthouses. Therefore, the lighthouse service fulfils the requirements of a public good: non-excludability and non-rivalness in consumption. Coase, however, set out to demonstrate that lighthouse services for an extensive period of human history were provided by private enterprise rather than governments and came to the conclusion:
[C]ontrary to the belief of many economists, a lighthouse service can be provided by private enterprise… The lighthouses were built, operated, financed and owned by private individuals, who could sell the lighthouse or dispose it by bequest.1
Coase claims that all these previous economists had been wrong and asks himself:
[H]ow is it that these great men have, in their economic writings, been led to make statements about lighthouse which are misleading as to the facts, whose meaning, if thought about in a concrete fashion, is quite unclear, and which, to the extent that they imply a policy conclusion, are very likely wrong.2
Further, he argues that “[w]e may conclude that economists should not use the lighthouse as an example of a service which could only be provided by the government”.3
In this paper I re-examine Coase’s claims and add empirical evidence from two national lighthouse systems between 1600 and 1910. I try to answer under what circumstances private ownership might be preferable to public, and why private enterprise in the provision of public goods in the end was largely abolished both in Britain and in Sweden in the nineteenth century.
The interpretation of Coase in economics
Coase’s article on the lighthouse is often quoted as having demonstrated that government involvement in the production of public goods is minimal and limited to the establishment of general property rights in society. What Coase essentially said in his original article is, however, far from clear, to judge from how his results have been interpreted. Peacock argues that:
the ‘parable of the lighthouse’ inaugurated by Coase has much to tell us about the limitations of public goods theory… One is bound to agree with him that a lighthouse service can be provided by private enterprise alone within the existing framework of property rights (Peacock 1979, pp. 127, 132, emphasis in original).
Barrowclough summarizes Coase’s findings as:
Coase and Peacock have shown that from the sixteenth to the nineteenth centuries private sector entrepreneurs successfully provided lighthouse services… The Crown allocated monopoly rights, to ensure no confusing duplication of signals in one area; it enforced externalities, such as important navigational landmarks; and it regulated prices… A sophisticated payment system operated, with dues varying according to ship size and port of origin.4
Cowen interprets Coase’s article as:
an important study of how what was until recently commonly cited as a public good incapable of production on the market (because consumers could not be excluded) was indeed produced through methods of exclusion (fees charged at ports).5
Similarly, Libertarian Nation Foundation endorses Coase’s view of a market based private enterprise driven lighthouse system:
Lighthouses are another example. For decades, standard economic textbooks had loftily explained that lighthouses could never be supplied privately, because ships at sea benefit from the light whether or not they pay. But one day free-market economist Ronald Coase decided to do some research, and discovered that in fact lighthouses in Britain had in the past been supplied privately for many years... Once again, entrepreneurs who stood to make a profit were motivated to devise innovations undreamed of by pessimistic academic economists.6
A final example is Independent Institute’s endorsement of Coase’s conclusions:
The Independent Institute’s logo was inspired by Ronald Coase’s renowned 1974 essay in the Journal of Law and Economics, “The Lighthouse in Economics,” … Until that time, conventional wisdom from John Stuart Mill to Paul Samuelson had claimed that the lighthouse was the quintessential “public good,” which allegedly had to be provided by government due to the inherent free-riding of those who could not be charged for the services being provided. Coase showed, however, that in Britain, “contrary to the belief of many economists, a lighthouse service can be provided by private enterprise...” Only later did the British government consolidate all lighthouse services under its own monopoly in order to eliminate competition and directly reap the financial benefits developed by private entrepreneurs. In addition to exposing the fallacies of a favorite public-goods rationalization, Coase’s essay rescued the lighthouse as a symbol of courage, enlightenment and independence.7
Few attempts to critically re-examine Coase’s conclusion has since the publication of the article been done. The most important exception is van Zandt, who argues that Coase’s use of the terms “private” and “government” are too vague and unspecified to be useful in analysing the choice of institutional structure for the provision of goods and services:8
the stark dichotomy between ‘private’ and ‘government’ provision of goods and services that stalks many general discussions of public policy is, at best, a useless abstraction and, at worst, a barrier to understanding how goods and services are provided in the real world.9
Van Zandt re-examined Coase’s evidence and found more government involvement than Coase’s article had suggested. He concluded that lighthouses were provided on a continuum between private and government provision. Further, Bertrand takes van Zandt’s criticism of the Coasean analysis one step further and argues:
[W]hen examining the same historical experience, we are led to doubt Coase’s conclusions: the system was not private, and above all not efficient, as he suggests, in the sense that is was not well adopted to needs.10
Bertrand re-examined the same data Coase built his conclusions on, but argued that the British lighthouse system was mixed with a significant amount of government involvement.11 The mechanism for making users pay was far removed from a market mechanism where people voluntarily enter agreements. Bertrand argues that the relationship between ship owners and lighthouse keepers was not really bilateral. The Crown not only fixed light dues (in most cases), but also imposed their payment directly on ship owners, if necessary by threatening them with prison sentences. Finally, Taylor argues that the lighthouse system endorsed by Coase was a result of “old corruption” and that “[t]he state… had caused the problem by granting rights to levy light dues to individuals…12 Coase himself did not give credit to his critics (especially Paul Samuelson who did not agree with the Coasean analysis), and declared in 1997 in a Reason Magazine interview that the criticism of his analysis of private ownership in lighthouses was “humbug because you could say that there’s no private property in houses by that logic, since you cannot transfer your rights to a house without the examination of the title and registration and without obeying a whole series of regulations”.13 Apparently, Coase still believed that he had shown that the lighthouse service was provided by a private market mechanism with the government’s role limited to defining a whole range of regulations. As Barnett and Block point out, Coase thus failed to distinguish between a situation where the relevant payments were compulsory, and a market situation in which all commercial interaction takes place on a strictly voluntary basis.14 Although Coase was wrong when claiming that the provision of lighthouse service was just like any other market, van Zandt endorses Coase’s second major conclusion, that private provision under grant (the lease system) was more efficient than direct provision by the government:
The institutions of patents… did have a decided administrative advantage over straight government provision of the service… That, however, does not mean “private” provision as that term is generally understood.15
In a similar vein as van Zandt, Fischel argues that the private system described by Coase provided lighthouse services “at least as well as a program that relied exclusively on government provision and general taxation to fund them”.16 Bertrand tries to qualify this conclusion and argues that the private system was a failure:
Coase chose the elements that tend to prove that a private lighthouse system could exist and would be appropriate. Choosing other elements seems to show that the English system was mixed, expensive and defective.17
Although Coase’s main point, that the market for lighthouse service was just like any other market was based on an underestimation of the role of government, the questions remain: how should we understand the “private” element under public grant and what were the relative efficiencies between pure public and the mixed system of lighthouse service provision? And if the rental system was as efficient as claimed by Coase and van Zandt (relative to direct provision by the government), why was it abandoned? Or if the rental system was inefficient as claimed by Bertrand, why was in not abandoned earlier, much less chosen at all?
The problem of the lighthouse in economics
Lighthouses are not only important for economic theorists trying to understand how public goods should be managed; lighthouses also have a most concrete and serious side, since well managed lighthouses save lives. Lighthouses provide information in dangerous waters. Mariners have always used landmarks to determine their location and appropriate routes for sailing. On open seas, the sun, stars and eventually the compass and navigational maps provided shippers with measures of location. Closer to land, natural and man-made structures were visible in daylight, and it was a significant step to put lights on such structures and organize the logistics to make sure that the lights were lit at appropriate times. The presence of lighted land marks was important because ships needed this service in bad weather or in the dark, precisely when the dangers at sea were the greatest. The information the lighthouses provided also needed to be reliable. Due to human error and neglect an expected lit landmark unexpectedly not functioning, might lead ships in danger to disaster. Lights might burn out and reflective mirrors becoming dirty, leaving the ships without guidance in the dark. Further, false lights, lit by wreckers, would sometimes lead ships astray and into land rocks crushing its hull. Therefore, a human organization was needed to assure that lights were sufficient in quantity and quality. On a methodological level, Coase was criticising the Samuelsonian research program in economics, exemplified by how Paul Samuelson reached his conclusion about the infeasibility of private provision of public goods. Samuelson presented a “mathematical exposition of a public expenditure theory”, demonstrating the “fatal inability of any decentralized market or voting mechanism to attain… optimum”.18 This “fatal inability” is based on two properties characteristic of public goods that distinguish them from private goods: non-rivalness and non-exclusiveness in consumption. By arguing that non-exclusiveness did not characterise the provision of the essential lighthouse service, Coase did not only change the basis for policy prescriptions concerning government activities.19 He also was very critical of how economists, in particular Samuelson, at the time handled externalities. Externalities, as a theoretical concept, were treated as market failures necessitating the directing hand of the state. But in the real world, Coase argued, markets and institutions will always fall short when measured against idealized norms that assume away real world problems. Thus, Coase wanted to create a program of comparative institutional analysis. This approach, however, has its own problems, as is evident by the confusion Coase’s article has caused. As emphasised by Fischel, Coase also opened up a property rights approach to economic problems where the very notion of property normally is disregarded: how can there be private ownership in something from which consumers cannot be excluded?20 Coase’s urge that more detailed studies of lighthouse provision is needed to better show “the richness of the social alternatives between which we can choose” took some time to respond to, possibly because he is widely accredited with proving that private enterprise could provide the lighthouse service just like any other good on the market. Nonetheless, the property rights approach to studying the lighthouse problem has not gone very far.
The organization of the lighthouse services in theory
The main problem with private lighthouse provision concerns the possibilities to exclude non-users from users. This was also the main point for Coase, as he tried to show that lighthouses were not non-exclusive, i.e., that it was possible to make users pay for the service by charging them once they arrived in a port. While natural land marks were provided free by nature, lighthouses were erected by human beings, and unless they acted out of charity, they needed to be compensated in one way or other. Incidentally, charitable provision of lighthouse service was a major reason behind the founding of early medieval lighthouses, as we will see later.21 J.S. Mill explained the problem with exclusion as “[i]t is impossible that the ships at sea which are benefited by a lighthouse should be made to pay toll on the occasion of its use”22 For a lighthouse owner, the problem is obviously, as Samuelson puts it that “lighthouse keepers cannot reach out to collect fees from shippers”.23 Therefore, “a businessman could not build it for profit, since he cannot claim a price from each user”.24 Coase’s criticism against Samuelson and Mill among others boils down to the question whether it is possible to charge users for the lighthouse serve: “As I read these writers, the difficulty of charging for the use of a lighthouse is a serious point with important consequences for lighthouse policy”.25 For a private lighthouse keeper, the difficulty to make a credible threat to a passing ship that he will withhold the service unless he is paid makes it difficult to provide such services. Sidgwick pointed out the difficulties of excluding ships from the seas: “Some utilities, from their nature, are practically incapable of being appropriated by those who produce them or would be willing to purchase them. For instance, it may easily happen that the benefits of a well-placed lighthouse must be largely enjoyed by ships on which no toll could be conveniently imposed”.26 In other words, could the lighthouse keeper force all ships to pass a turnpike toll collector who collected the fee the public goods aspect of lighthouses would disappear. But since the sea is huge it is a practical incapability to charge all passing ships and many ships will free ride on the commons. Excluding free riders is simply too costly. Van Zandt suggests that free riders could be stopped: the lighthouse operator would simply place his employees in small boats along the outer range of the lighthouse’s light, strike a deal with the passing ships and signal to the operator to light the lighthouse.27 He continues to argue that “while this possibility sounds farfetched, it points out that excludability is often tied to the precise legal institution surrounding the provision of a good… It might even be that the legal rule of freedom of the seas actually reduced seafaring traffic by suppressing the provision of navigational guidance so necessary for safe transport”.28 Coase claims that some exclusion was possible, if imperfect, since the ships eventually had to leave the commons and enter a port. At the ports, it was possible to levy tolls on ships “presumed to have benefited from [the lighthouse].”29 This is the centre of Coase’s analysis, and he strengthens his conclusion by giving the numbers of private lighthouses in England based on the 1834 Committee of Lighthouses. The committee stated that there were 71 public general lighthouses in total, of which 14 were run by private individuals and organizations.30 The majority of the lighthouses were run by Trinity House and Coase claims that the reason why all lighthouses eventually were consolidated under Trinity House was because it was thought to lower light dues.31 Trinity House of Deptford Strond originated as one of many associations of seafarers, organized to control pilotage in harbours and provide financial aid for old seamen and their families. After 1679, Trinity House was the sole issuer of rights to erect lighthouses in England and Wales.32
Coase made many important qualifications in pointing out that private parties historically had been involved in providing public goods. The question is: how should we analyse this mix of public/private production? Since the production of lighthouse services shifted from a lease system to a system of direct government control, focus for the analysis should be the shift of contractual arrangements over time. The purpose here is not to define an optimal contractual scheme33, but to make an economic analysis of the schemes, and highlight their rationale and problems. Several critics of Coase have noted that his definition of private is too vague. Under Coase’s definition, lighthouse services was no different than any other goods and services. The government’s role was limited to establishing and enforcing property rights to resources and enforcing contracts freely agreed to. Therefore, according to Coase, lighthouses arose spontaneously as shippers and ship owners petitioned the Crown to allow a private individual to construct a lighthouse and charge a fee from the petitioning shippers. Coase argued that the sole role of government was to establish property rights in the ownership of the lighthouse and that “the property rights were unusual only in that stipulated the price that could be charged.”34 Further, “[t]he charges were collected at the ports by agents for the lighthouses. The problem of enforcement was no different for them than for other suppliers of goods and services to the ship owner.” Coase concludes that “[t]he lighthouses were built, operated and financed and owned by private individuals, who could sell the lighthouse or dispose of it by bequest”.35 Van Zandt suggests, as mentioned above, that instead of working with a dichotomy of ‘private’ and ‘public’, lighthouse services can be produced on continuum between the poles of pure private provision with no government support to full government provision out of general revenues. The English lighthouse system was mixed, and we need to understand how different levels of government intervention affected the efficiency of the lighthouse system. van Zandt suggests the following definitions of property rights schemes for lighthouse provision:
(1) private provision with no government enforcement of property and contracts rights;
(2) private provision with government enforcement of property and contract rights only;
(3) private provision with government fixing rates, granting monopolies and enforcing collection of specified user levies;
(4) government provision from collection of specified user levies; and
(5) government provision from general revenues.36
Paul Samuelson’s analysis of pure public goods is an implicit comparison between van Zandt’s categories (1) and (5), which would endorse the arrangements under (5), because of the problem of excluding users unwilling to pay for the service. Coase, in his conclusions seems to endorse arrangements under (2), although his own description of the historical British lighthouse system perhaps better fits under category (3). Van Zandt argues that “Presumably…society has found that across a broad range of goods governmental enforcement of property and contract rights is desirable. This is because there are strong reasons to locate the right to use force in one entity.”37 This, of course, does not exclude the possibility that governments can be subject to corruption and other pressures. According to Bertrand’s analysis, this is exactly was bedevilled the mixed lighthouse system in England.38 As Bertrand notes, Coase definition does not exclude a significant governmental intervention, something Coase ruled out.39 As Fischel argues, the difficulties of excluding beneficiaries should not obscure the fact that lighthouses (the actual buildings) could be privately owned, and a system of payment (light dues collected at ports) could be invented that rewarded their owners for providing this non-excludable service.40 Fischel argues that the presence of significant private capital in the lighthouse business provided services “at least as well as a program that relied exclusively on government provision” and that “the obligation to pay port duties to fund their services were important means by which entrepreneurial capital was invested in this critical service.”41 Although van Zandt’s classification of varying degrees of government intervention is useful, it does explain the choices the government had in engaging itself in lighthouses: the choice between direct ownership based on salaried officials or private provision under royal grants and how these choices varied over time. In the selection of the contractual arrangements for the production of lighthouse service and collection of light dues, the Crown faced a principal-agent problem. In return for the provision of lighthouse services, the government gave right to individual to collect light dues, which, of course, were supposed to cover the costs for running the lighthouse. The principal-agent literature can be used to illustrate van Zandt’s scheme of different levels of state involvement, the evolution of choice of contracts and their shortcomings.42 The literature postulate three basic types of contracts—rental, share, and wage contracts—that are on a linear continuum, with extremes of a pure rental contract and a pure wage contract, where the agent and the principal respectively absorb all the risk. In a rental contract for a lighthouse, the entrepreneurs would pay a fixed yearly rent to the government for the right to collect light dues and keep the remaining revenue; in a share contract, the government would lease the right to collect light dues to a collector for a share of the revenue; and in a wage contract, the government would pay a fixed wage to its lighthouse keepers and due collectors in return for delivery of all revenue. I have not found any examples of share contracts in the literature; rental and wages contracts seems to be the two forms used in governing the light provision. Output is a function of the efforts of the agent, which were unobservable to the principal, and a random variable, reflecting the fact that light due yields varied from year to year with economic fluctuations and political upheavals that were beyond the control of the Crown and the lighthouse entrepreneurs. Economic fluctuations that induce the variability in revenue beyond the control of Crown and lighthouse entrepreneur would also influence the choice of contract. If the agent is less risk-averse than the principal, risk sharing implies that greater variance will shift contracts along the continuum away from wage contracts towards rental contracts (the agent’s share increases). The distribution of risk preferences is verified by the historical literature, where several authors have told stories of entrepreneurs who risked their capital in hope of earning great wealth.43 Costs to the principal of supervising or monitoring the agent also play a role. Lighthouse entrepreneurs need to be monitored to ensure that they are correctly reporting revenues and perform their lighthouse service. Wage contracts create incentives for an agent to over-report costs and under-report output, making a rental contract more likely when these costs are high. If monitoring costs decline, the choice of contract should shift towards a wage contract, as the principal stands to gain additional income from absorbing more risk as long as the incentives are maintained. Since there are economies of scope in monitoring employees and the use of capital, and this feature may move the principal towards wage contracts, as evident from the rise of salaried government agencies in the 1800s. In the modern world, governments almost exclusively pay the employees of their tax-collecting bureaucracies a fixed wage or salary to collect taxes. If the task of collecting taxes is well known (i.e., the government is at no informational disadvantage), collection can be monitored cheaply, and the government can borrow to smooth the fluctuations in revenue, then a fixed wage is an appropriate incentive to motivate employees. By absorbing all of the risk from revenue fluctuations, the wage contract also delivers the highest expected revenue to the government. However, these conditions were rarely attained in the past. For the collection of indirect taxes, for example, the many European governments in the seventeenth and eighteenth centuries did not choose a state-run salaried bureaucracy. Instead, it usually leased the right to collect taxes to a syndicate for a fixed number of years.44 This evolution of contractual forms suggests that the Crown-lighthouse owner relationship can be analysed as a principal agent problem. The principal agent literature helps to explain the Crown’s choice of contracts and the various problems associated with the contractual choice. A pure rental contract implies that the agent (lighthouse owner) absorb all risk. A pure wage contract implies that the principal (the government) absorb all risk. The key to understand the economics of lighthouse provision is to understand the shift of contractual forms. A rental system, although convenient for the government in some respects, could make some individuals rich, at the expense of the general public. Why would the government choose a form of public good provision that were expensive and inefficient? For the reformers of the 1830’s the answer was clear. As Taylor suggests, it was seen as a result of “old corruption” and thought to be against the interest of the public interest.45 Taylor suggests that the debate in England in the 1830s about how lighthouses should be managed concerned the rental system: “Public ownership would eliminate profit and surplus collection, and centralization would lead to economies in the costs of management”. Was the rental system based on corruption or was it a rational response to the difficulties facing underdeveloped states in history? Hague and Christie argue that the collection of light dues was “farmed out” to individuals, and that “all such activities were open to wide abuse, but sufficient money did find its way to the state coffers to enable the system to carry on in this manner until the affairs of the state became too complicated”.46 Coase believed that a general tax (i.e., the system suggested by Samuelson’s analysis) would “tend to reduce somewhat the efficiency with which the lighthouse service was administered.”47 Bertrand disputes this conclusion and suggests that the light dues levied from individual ship owners led to over charging and excess profits.48 For Coase, the main drawback with governmental supply of lighthouse services was that the government did not build enough lighthouses. Clearly, he implied a form of government failure on the part of the British authorities. He approvingly quotes Harris on this point: “The difficulty was that those who were motivated by a sense of public service did not build the lighthouses. As Harris says later: ‘Admittedly the primary motive of the lighthouse projectors was personal gain, but at least they got things done’”.49 The evidence for this government failure on the part of Trinity House is the fact that in the period 1610-1675, no lighthouses were erected by Trinity House, but at least ten were built by private individuals. For Coase, this was a crucial evidence for two conclusions: governments are not always responsive to the opinions and preferences of the public, and profit seeking individuals, although motivated by selfish goals, could provide the much needed public goods.