1 The verbal descriptionof anexternal effect – that is, a direct effect onanother’s profit or welfare arising as an incidental by-product of some other person’s or firm’s legitimate activity – would seem adequate to convey its meaning. Its nature is made yet clearer, however, by examining the notion of ‘internalizing’the external effect.
The basic idea is that of transforming the incidental by-product into a joint product that is priced on the market. I have been told by a number of Argentinians that, before the turnof the century, cattle were slainonthe ranches for their leather only. Their flayed carcasses were left to rot but, if found in time, they could be used as fresh meat by the poor peasants. Apparently, only the leather had a market price, the meat being a by-product, or external effect, of leather production – a favourable spillover of the leather industry for those peasants who happened to be inthe vicinity.1
Suppose, however, that the humanpopulationbeganto multiply more rapidly thanthe cattle population, that the taste for meat grew, that meat beganto be stored in refrigerators and that, most important of all perhaps, the meat could be exported to distant markets. Domestic meat would become scarce and, therefore, a market for it would come into being. It would then cease to be a spillover, an unintended by-product in the process of obtaining hides for leather. It would take its place as a good inits ownright, a joint product with leather. Whatever the separate demands for meat and leather are like, the long-run competitive equilibrium output is optimal, as the cattle population is expanded to the point at which the sum of the market prices are equal to the marginal cost of cattle production. The external effect has been internalized into the pricing system.2
1 Notwithstanding which the number of cattle slain could be optimal if, at the margin, the value of the meat was zero. We discuss this point further in the next chapter in terms of ‘allocative significance’.
2 It may seem unnecessary to remark that the possibility of internalizing an external effect (or, in the absence of internalization, correcting for optimal outputs) does not mean that the creation of an adverse external effect need not make things worse. Yet, students do sometimes argue as though this is so; as though, so long as optimizing by one method or another takes place, the creation of adverse external effects may be viewed with equanimity. The introduction of an adverse external effect into the economy is a bad thing no matter how the economy adapts to it. By internalizing the bad, or by optimizing the output that produces the bad, we are doing no more than making the best of a bad job.
We are certainly not as well off as we should be if this bad had not appeared on the economic scene.
QUAH: “CHAP18” — 2007/1/25 — 07:59 — PAGE 100 — #2 Internalizing spillover effects arises also in the case of external diseconomies that are internal to the industry. Common examples of the latter category are deep-sea fishing, in which any additional fishing boat above a certain number reduces the catch of each of the existing fishing boats in the fishing grounds, or traffic congestion, in which every additional vehicle above a certain number causes delay to each of the existing number of vehicles using a given highway system.
Internalizing this sort of spillover would require that a positive market price be imputed to the currently unpriced though scarce resource – the area of the sea in the first case, the highway in the second. Once such a resource is priced, it will be used more economically. The analogy of scarce land used in the production of, say, cornis exact. If priced correctly, which implies that ina competitive industry the rent of this scarce resource be maximized, the competitive equilibrium output that emerges is also the optimal output.3
Another example, though one in which internal accounting prices are substi-tuted for market prices, is that of two separately owned but adjacent factories, A and B. Factory A produces shoes and is powered by an old-fashioned coal engine, which emits so much smoke as to seriously affect the output of factory B, which produces chocolate bars. The manager of factory B remonstrates with manager A, but to no effect. The daughter of the owner of factory A and the son of the owner of factory B decide to get married, in consequence of which the two factories come under common ownership and control, and the couple live together happily ever after. The cost of the smoke, reckoned in terms of the damage inflicted on the output of factory B, is no longer a spillover gener-ated by A and suffered by B. It is now unambiguously a cost to the joint A–B enterprise and, as such, ways and means of reducing it will be sought. Either anti-smoke devices will be installed in factory A, or else, if cheaper (and assuming the smoke damage to B’s output varies directly with A’s output), A’s output will be reduced to the point at which the value of the marginal damage to B’s output, added to the marginal cost of shoe production in A, is equal to the market price of A’s shoes. Thus, the smoke ceases to become a spillover effect, but becomes a properly costed item that is internalized into the costing system of the A–B merger.
2 The number of spillover effects that can be internalized into the pricing mech-anism or into the costing systems of firms is, however, limited. Among those that cannot easily be internalized through the market are many of the by-products of modern industry and of the hardware it produces. One thinks, in this connection, of traffic noise and various forms of pollution arising from the spread of sewage,
3 Assuming a period during which there is one scarce fixed factor and one factor that is variable in supply at a constant price, the average cost curve eventually slopes upward. A curve drawn marginal to this average cost curve cuts the demand curve at the optimal output. At this output, the difference between average cost and marginal cost times output gives the amount of the rent to the fixed factor – the maximum rent possible ina perfectly competitive market inwhich the price of the product is treated as a parameter.
QUAH: “CHAP18” — 2007/1/25 — 07:59 — PAGE 101 — #3 Internalizing externalities 101 garbage and radioactive wastes; also of the post-war phenomenal growth of diseases of the nerves, heart and stomach, caused by high-tension living, the most ubiquitous by-product of sustained technological advance. Why cannot such spillovers be so internalized? The answer is simple: in order for a competitive mar-ket for such spillovers to emerge, certain conditions have to be met which, in the nature of the physical universe, cannot be met. First, the potential victim of these adverse spillover effects must have legal ‘property rights’ in, say, their ownership of some quantum of quiet and clean air which, if such rights were enjoyed, they could choose to sell to others. Second, in order for such rights to be enforceable, it would be necessary to demarcate a three-dimensional ‘territory’ about the person of each potential victim in order to identify the intrusions of others and take appro-priate legal action. Third, in order for a monopolistic situation not to arise, each of these three-dimensional properties withina givenarea, which canbe rented for particular purposes (say, to accommodate the noise or pollution of someone’s activity), must be a close substitute for the others.
The first conditioncould, of course, be met inthe sense that all forms of pollution could be outlawed in the absence of specific agreements between the parties con-cerned. But because the second condition cannot be met in the world we inhabit, there is difficulty in demarcating each person’s property, and a consequent dif-ficulty in identifying the trespasser and the extent of the trespass. Nor can the third conditionbe met, for, inthis hypothetical scheme of things, the right to use one man’s ‘territory’, within some given area, is no substitute for that of another man. Each man within the area has his own three-dimensional territory and, since the noise to be created by the new activity enters in some degree into all of such territories, the enterprise has to reach agreement with each one of them. None cansubstitute for the other. Unless all agree, the permissionof those who do is worthless.
If it were otherwise, if one territory could be substituted freely for another, as could plots of land inanagricultural area, anappropriate market price would arise from the competition of the sellers. The physical universe being what it is, however, each potential seller is in a completely monopolistic position for, without his particular consent, the necessary arrangement for the whole of the affected area cannot be concluded. The reader will detect a similarity between this hypothetical problem, posed by the third condition, and that facing a railroad company having to buy every mile of land through which the track has to run. The cost of acquiring rights where a large number of landowners are involved could be prohibitive were it not for legislation compelling the sale of rights on terms which the courts decide are reasonable. Another instance, occasionally reported by the press, is that of a single householder or small business holding out against a property company that is attempting to buy up a specific area of land as part of some new development scheme.
We must, then, resign ourselves to the prospect of never being able to internalize these important environmental spillovers within the market economy; that is, of not being able to create a market for them – which is, of course, one of the reasons why cost–benefit methods are required to evaluate them.
QUAH: “CHAP18” — 2007/1/25 — 07:59 — PAGE 102 — #4 3 Some further light is cast onthe nature of spillover effects by briefly observ-ing the connection between them and collective or public goods. Environmental spillovers usually affect a large number of people withinanarea. But whereas the positive spillover has been defined as an unintended beneficial effect on others arising from some legitimate economic activity, the benefits conferred on the community by a collective good are those that are deliberately created. It may be noted, moreover, that the collective good may be optional for members of the community, an example being a public park which allows each person to choose how many hours, if any, he wishes to spend there.4Anon-optional collective good, incontrast is one inwhich each personperforce receives some amount of it, an example being the rainfall over a certain area that is caused by ‘seeding’the clouds above it. Insuch a case, the amount of rainfalling onthe land of each personcan be more or less thanhe would prefer. It may evenbe so large anamount that, on balance, it becomes a net loss to the recipient – an adverse spillover.5
4 It is not to be supposed, however, that the evaluation of positive spillovers in a cost–benefit calculation is confined to collective goods. What are often referred to as public goods – whether or not they are publicly financed – need not also be a collective good as defined. A hospital or a railroad is not, strictly speaking, a collective good: once constructed, the services produced by either can be separately allocated to each of a number of persons for their own particular use.
For the collective good, the benefit in any period of any one unit of the good is equal to the aggregate of the benefits enjoyed by all affected by it. If, therefore, the amount of the collective good is variable, the net benefit to the community is maximized by increasing the amount of it until its marginal (aggregate) benefit is equal to its opportunity cost.
The same is true if, instead, we are initially addressing ourselves to the con-struction of some public good that is not, strictly speaking, a collective good. For instance, if we are to determine the longest distance to be covered by a proposed rail link that would connect a number of towns and villages over the time span contemplated, we have to aggregate the benefits of each potential passenger over that time span for each alternative length of the railroad – adequately measured in each case by the areas under the expected future demand curves – and compare them with their relevant opportunity costs.
Once the chosen rail link has been established, however, the service it provides has to be treated as a private good. Hence, in operating the service, net benefits are maximized if every person pays the marginal cost he incurs which, in the absence
4 There can, however, be a problem of congestion if the number of people increases relative to the number, or the size, of the facilities provided.
5 Among those who onbalance gainfrom the givenartificial rainfall, there canbe those farmers whose crops receive too much rain, in the sense that the benefit to them of the marginal inch of rain is negative. The optimal condition, however, requires that rain be increased until the sum of the benefits and losses of the marginal inch of rain is equal to the cost of producing it.
QUAH: “CHAP18” — 2007/1/25 — 07:59 — PAGE 103 — #5 Internalizing externalities 103 of congestion, could be zero, provided that the daily or weekly overheads are covered by the consumer surplus.
Inthe case of a public good such as a hospital, incontrast, once the optimal size of the hospital has been constructed, the marginal cost of each patient is positive and will generally vary between one patient and another.
QUAH: “CHAP19” — 2007/1/25 — 19:07 — PAGE 104 — #1