The idea of a Pigouvian tax traces back to the classic 1920 book by Arthur C. Pigou, The Economics of Welfare. The concept is straightforward. There are certain situations, like those involving pollution, where the unregulated producer of a good does not need to take the social costs of pollution into account. As a result, the private costs of production are not equal to the social costs. In such a situation, one can make a case for the government to impose a tax--a "Pigouvian tax"--which forces the producer to pay for the social costs that it is imposing.
In Part II, Chapter IX of the 1920 book, Pigou discusses "Divergences Between Marginal Social Net Product and Marginal Private Net Product." He offers a clear statement of what modern economists mean by a Pigovian tax, but in the course of explaining the economic logic also provides some ammunition to those who fear that government may have quite a difficult time in applying such taxes. I quote here from the version of Pigou's book available at the always-useful Library of Economics and Liberty website.
Pigou starts the chapter by discussing situations in which social and private costs might not be aligned for reasons of lack of competition, like monopoly, or in situations like landlords and tenant-farmers where it may be difficult to negotiate who will pay for investing in improvements to productive capacity and who will get the benefits. He then moves to the arguments over what modern economists refer to as externalities and public goods. Pigou writes:
"Here the essence of the matter is that one person A, in the course of rendering some service, for which payment is made, to a second person B, incidentally also renders services or disservices to other persons (not producers of like services), of such a sort that payment cannot be exacted from the benefited parties or compensation enforced on behalf of the injured parties. ...
"Among these examples we may set out first a number of instances in which marginal private net product falls short of marginal social net product, because incidental services are performed to third parties from whom it is technically difficult to exact payment. Thus, as Sidgwick observes, "it may easily happen that the benefits of a well-placed light-house must be largely enjoyed by ships on which no toll could be conveniently levied."Again, uncompensated services are rendered when resources are invested in private parks in cities; for these, even though the public is not admitted to them, improve the air of the neighbourhood. The same thing is true—though here allowance should be made for a detriment elsewhere—of resources invested in roads and tramways that increase the value of the adjoining land—except, indeed, where a special betterment rate, corresponding to the improvements they enjoy, is levied on the owners of this land. It is true, in like manner, of resources devoted to afforestation, since the beneficial effect on climate often extends beyond the borders of the estates owned by the person responsible for the forest. It is true also of resources invested in lamps erected at the doors of private houses, for these necessarily throw light also on the streets. It is true of resources devoted to the prevention of smoke from factory chimneys: for this smoke in large towns inflicts a heavy uncharged loss on the community, in injury to buildings and vegetables, expenses for washing clothes and cleaning rooms, expenses for the provision of extra artificial light, and in many other ways.Lastly and most important of all, it is true of resources devoted alike to the fundamental problems of scientific research, out of which, in unexpected ways, discoveries of high practical utility often grow, and also to the perfecting of inventions and improvements in industrial processes. These latter are often of such a nature that they can neither be patented nor kept secret, and, therefore, the whole of the extra reward, which they at first bring to their inventor, is very quickly transferred from him to the general public in the form of reduced prices. The patent laws aim, in effect, at bringing marginal private net product and marginal social net product more closely together. By offering the prospect of reward for certain types of invention they do not, indeed, appreciably stimulate inventive activity, which is, for the most part, spontaneous, but they do direct it into channels of general usefulness."There's a lot of content packed into that paragraph. Pigou is pointing out that private and social costs are likely to be misaligned in various public goods (lighthouses, parks in cities, forests), as well as in the effects of pollution, and in the effects of scientific innovation. In these cases, economic theory suggests the possibility that by using taxes related to the social costs of certain actions (like pollution) or subsidies (he calls them "bounties" related to the social benefits of certain actions, like innovation. Pigou writes:
"It is, however, possible for the State, if it so chooses, to remove the divergence [between social and private costs] in any field by "extraordinary encouragements" or "extraordinary restraints" upon investments in that field. The most obvious forms which these encouragements and restraints may assume are, of course, those of bounties and taxes. Broad illustrations of the policy of intervention in both its negative and positive aspects are easily provided. ...
"The private net product of any unit of investment is unduly large relatively to the social net product in the businesses of producing and distributing alcoholic drinks. Consequently, in nearly all countries, special taxes are placed upon these businesses. Marshall was in favour of treating in the same way resources devoted to the erection of buildings in crowded areas. He suggested, to a witness before the Royal Commission on Labour, "that every person putting up a house in a district that has got as closely populated as is good should be compelled to contribute towards providing free playgrounds."The principle is susceptible of general application. It is employed, though in a very incomplete and partial manner, in the British levy of a petrol duty and a motor-car licence tax upon the users of motor cars, the proceeds of which are devoted to the service of the roads. It is employed again in an ingenious way in the National Insurance Act. When the sickness rate in any district is exceptionally high, provision is made for throwing the consequent abnormal expenses upon employers, local authorities or water companies, if the high rate can be shown to be due to neglect or carelessness on the part of any of these bodies."
In short, Pigou back in 1920 was offering an economic justification for taxes on alcohol and gasoline, as well as for property taxes used to support local parks and amenities. This all holds together as a matter of theory and logic. But as usual when moving from theory to policy, the devil is in the details. Here, I'll point to two sets of concerns that arise from reading Pigou.
First, it's important to remember the decisions about which Pigovian taxes and bounties will be set up that the decisions are not made by the disinterested angels of our better nature, or even by economists, but rather by politicians. Later in the book, Part II, Chapter XX, is titled "Intervention by Public Authorities," and there Pigou offers this important distinction between the theoretical case for Pigouvian taxes and the practical reality:
In any industry, where there is reason to believe that the free play of self-interest will cause an amount of resources to be invested different from the amount that is required in the best interest of the national dividend, there is a prima facie case for public intervention. The case, however, cannot become more than a prima facie one, until we have considered the qualifications, which governmental agencies may be expected to possess for intervening advantageously. It is not sufficient to contrast the imperfect adjustments of unfettered private enterprise with the best adjustment that economists in their studies can imagine. For we cannot expect that any public authority will attain, or will even whole-heartedly seek, that ideal. Such authorities are liable alike to ignorance, to sectional pressure and to personal corruption by private interest. A loud-voiced part of their constituents, if organised for votes, may easily outweigh the whole. This objection to public intervention in industry applies both to intervention through control of private companies and to intervention through direct public operation. On the one side, companies, particularly when there is continuing regulation, may employ corruption, not only in the getting of their franchise, but also in the execution of it. ... On the other side, when public authorities themselves work enterprises, the possibilities of corruption are changed only in form. ... [Here Pigou quotes the US-based Report to the National Civic Federation on Municipal and Private Operation of Public Utilities.] "Every public official is a potential opportunity for some form of self-interest arrayed against the common interest."A second concern is about the potential breadth of the Pigovian argument. Many modern economists would accept the theory of a Pigouvian tax applied in the case of pollution or a Pigouvian bounty applied to subsidizing scientific research, with some degree of hesitancy over the political economy concerns. But Pigou mentions a number of other cases in which social and private costs may diverge. For example:
When construction of factories imposes costs on neighborhoods
"... when the owner of a site in a residential quarter of a city builds a factory there and so destroys a great part of the amenities of the neighbouring sites; or, in a less degree, when he uses his site in such a way as to spoil the lighting of the houses opposite: or when he invests resources in erecting buildings in a crowded centre, which, by contracting the air space and the playing-room of the neighbourhood, tend to injure the health and efficiency of the families living there."
When women work in factories
"Perhaps, however, the crowning illustration of this order of excess of private over social net product is afforded by the work done by women in factories, particularly during the periods immediately preceding and succeeding confinement; for there can be no doubt that this work often carries with it, besides the earnings of the women themselves, grave injury to the health of their children. ... [P]rohibition of such work should be accompanied by relief to those families whom the prohibition renders necessitous."When those who buy new products make other people envious
"For, in some measure, people's affection for the best quality of anything is due simply to the fact that it is the best quality; and, when a new best, superior to the old best, is created, that element of value in the old best is destroyed. Thus, if an improved form of motor car is invented, an enthusiast who desires above all "the very latest thing" will, for the future, derive scarcely any satisfaction from a car, the possession of which, before this new invention, afforded him intense pleasure. In these circumstances the marginal social net product of resources invested in producing the improved type is somewhat smaller than the marginal private net product."When certain industries like agriculture help to make people suitable for military training
"The private net product of any unit of investment is unduly small in industries, such as agriculture, which are supposed to yield the indirect service of developing citizens suitable for military training. Partly for this reason agriculture in Germany was accorded the indirect bounty of protection."When a city plan makes it necessary to demolish some buildings
"Thus it is coming to be recognised as an axiom of government that, in every town, power must be held by some authority to limit the quantity of building permitted to a given area, to restrict the height to which houses may be carried,—for the erection of barrack dwellings may cause great overcrowding of area even though there is no overcrowding of rooms,—and generally to control the building activities of individuals. It is as idle to expect a well-planned town to result from the independent activities of isolated speculators as it would be to expect a satisfactory picture to result if each separate square inch were painted by an independent artist. No "invisible hand" can be relied on to produce a good arrangement of the whole from a combination of separate treatments of the parts. It is, therefore, necessary that an authority of wider reach should intervene and should tackle the collective problems of beauty, of air and of light, as those other collective problems of gas and water have been tackled. ... Furthermore, it may, if desired, be extended to include land on which buildings have already been put up, and may provide "for the demolition or alteration of any buildings thereon, so far as may be necessary for carrying the scheme into effect.""When advertising just cancels out the effect of other advertising
"[I]t may happen that the expenditures on advertisement made by competing monopolists will simply neutralise one another, and leave theindustrial position exactly as it would have been if neither had expended anything. For, clearly, if each of two rivals makes equal efforts to attract thefavour of the public away from the other, the total result is the same as it would have been if neither had made any effort at all."As these examples (and Pigou offers others) suggest, the idea of Pigouvian taxes and bounties applies in any and every situation where someone can make an argument that someone else is affected by a market transaction--even if just makes someone feel bad when someone else buys a new product. It seems to me that almost every public policy argument can be framed in terms of avoiding social costs or gaining social benefits. Again, remember that these arguments are not being made among pure-hearted truth-seekers, but rather in a political setting.
Thus, a question arises of how one disciplines the process of deciding when the Pigouvian logic applies. For example, it's common on one side to hear arguments that there is a case for Pigouvian bounties to subsidize the research and development that leads to new innovations and higher productivity. On the other side, it's common to hear arguments that there is a Pigouvian case for limiting robots or other new innovations so that they don't impose costs on existing workers. But a set of policies that simultaneously encourage and discourage innovation runs a real risk of expressing our ambivalent feelings about new technology in way that is close to incoherent.
A related problem in thinking about Pigovian taxes arises when choosing the tax rate. For example, in the case of alcohol there is some evidence that moderate consumption may have health benefits, through a reduction in blood pressure. However, inappropriate and excessive consumption of alcohol can also lead to drunken driving, violence, fetal alcohol syndrome, and other consequences. Thus, it seems as if the appropriate Pigouvian tax on alcohol should be to subsidize the light social drinker, but to impose a high tax on drinkers who impose high social costs. When the effects of an action on third parties are heterogenous in this way. choosing an appropriate Pigouvian tax becomes tricky, and society may well feel a need for use of alternative or complementary policy tools.
Like many economists, I favor certain taxes and subsidies on Pigouvian grounds. But it's worth remembering that the arguments in such cases are not just technocratic, but ultimately involve value judgments about political economy and social welfare.