While the above example represents, a positive externalit), negative ones also occur. For example, the cost of producing glue may include noxious odors in tbe neighborhood of the glue factory. (Remember "cost" represents the value of vvhat is foregone, notjust money). Here the neighbors forego clean air. Therefore, the total social cost of glue production, which measures what the whole society gives up to produce glue, consists of the firms private input cost plus the dollar value of tbe associated external environmental and health costs. Most environmental problems are instances in which the social cost of producdon exceeds the private cost. The difference between the two is the external cost. Thus, car exhausts impose costs on pedestrians; firms discharging effluent mto a river impose costs on fishermen downstream; off-shore oil drilling affects the enjoyment of sunbathers and surfers; and airplane noise prevents people from sleeping. To the private costs of each such activitj, we would add the dollar value of the negadve externaldes listed above to calculate the full social cost.

The presence of externalides, whether posidve or negative, may result in an inefficient allocadon of resources because the uansactors usually consider only their own private costs and benefits in making decisions. Figure I presents the situation for the compedtive glue industry. The competitive output, Qc , will be that at which the supply curve, representing the private marginal cost of manufacturing glue, strikes the demand curve. The industry pays litde or no attention to the external health and pollution costs because the firms do not themselves bear these costs. They are borne by outsiders. If we add these external costs onto the private ones, w-e get the marginal social cost curve.

which measures the marginal cost to the whole society associated wdth various outputs of glue. The optimal (or efficient) amount of glue production is Qc . This output occurs where the marginal social cost curve just strikes the demand curve, here representing the social as well as private benefit from glue production, since positive externalities are absent. To see that is really optimal, consider what would happen if a little more glue than Qe were produced. In this case, the value to society of an added unit of glue as read from the demand curve is less than the marginal cost to society of producing it as read from the marginal social cost curve. Therefore, we would want some resources to flow out of glue production, until the value to society of the last unit of glue just equals die marginal cost to society of making it. Similarly, to the left of Q , the value to society of more glue as read from the demjuid (marginal benefit) curve, exceeds the marginal social cost of producing it Therefore, we would want more glue and consequendy fewer environmental amenities.

Note, however, that there is absolutely no implication that the efficient pollution level is zero. The efficient level of glue producdon at Qg involves some external pollution costs. Such pollution is efficient because at Qc society values added glue output as much as added environmental amenities. Therefore, welfere cannot be increased by producing more of one and less of the other.

In general, the optimal amount of pollution is so rarely zero that most economists simply say the optimal amount is not zero: Generally, the cost of removing the very last speck, particle, or molecule of pollution is enormously high, too high to justify the cost of doing so. In such cases, its best to leave some pollution and use the clean-up resources to produce some-

where their value is higher. For example, not want to spend $10 to remove a particle of pollution that does $6 worth of damage. Nonetheless, in some cases the cost of removing the last speck is sufficiendy low that the opdmal amount of polludon is indeed zero. The smallpox virus, for example, is one natural type of polludon that has been wiped out completely; the cost of killing the last virus appears to have been sufficiendy low to justify the effort.

Figure 11 displays the situadon for positive externalities. Here the marginal social benefits are the sum of private benefits plus the dollar value of the posidve externaldes. Of course the marginal cost curve represents both social and private cost, since were assuming the absence of negadve externahdes. Again the efficient level of output is Qe, where the marginal cost to society just equals the marginal benefit. Left to its own devices, however, the compeddve market would supply the lesser amount, Q. Therefore, at Qc the good is underproduced because society values addidonal output more than the cost of producing it.

MEASURING EFFICIENCY LOSS

It is somedmes possible to measure the dollar value of certain efficiency losses. Take, for example, the case of firms emitting pollutants into a river, affecting downstream fishing. Here marginal social cost exceeds marginal private cost, so that an efficiency loss ensues under compedtion. In Figure III, the efficient output Qe occurs where marginal social cost = marginal social benefit. The competitive market provides Qc instead. The welfare increase associated with producing Qe instead of Qc is given by the shaded area ABC.

This can be seen as follows. The reduction incoul social cost from the output reduction is the area under the marginal social cost curve, ABQ-Q.. The loss of social benefit equals the area under the marginal social benefit (demand) curve, ACQ-Qe. Since total social cost decreases by more than total social benefit the difference between the two, area ABC, represents the net welfare improvement from the reduced output of market goods and the increased provision of environmental amenities.

COASE THEOREM

If youre thinking as an economist, you might have asked two questions. First, in the case of noxious glue odors, if society values environmental amenities more than the glue, why dont the neighbors get together and pay the glue factory to cut its emissions? Similarly, in the czise of private house painting, if society values more frequent painting more than the cost of providing it, why dont the neighbors get together and subsidize private house painting?

The answer to these two questions lies in transactions costs. In principle, the parties could indeed make such deals. Nonetheless, the organization costs of doing so may swamp the benefits if theyre very dif fuse. For example, air pollution in the Los Angeles airshed comes from so many different sources and affects so many different people that the transactions costs of such negotiations would ove ower any potential benefits. Moreover, since environmental improvement is basically a public good, many people" would refuse to cooperate, preferring instead to do nothing and receive free benefits.

Nonetheless, according to the Coase Theorem of Nobel Prize winner Professor Ronald Coase, if transaction costs are zero and if property rights are unambiguously defined and tradeable, the private market will automatically produce an efficient allocation of resources, no matter to whom the rights are assigned. Under these assumptions in the refinery example, suppose that an oil company owns a propert) right to emit unlimited pollution into the Los Angeles air basin. the pollution damage is greater than the costs oi abatement, people would pay the company to cut back. Wlien would they stop? Clearly when the dollar damage caused by the last unit of pollution just equals the cost of stopping it. Otherwise, some deal can be made making both the oil company and the general populace still better off. When no further deals are possible, that is to say, when gains from trade have been exhausted, the efficient level has been achieved by definition.

!

Now suppose that people have an unqualified right to clean air. Then so long as the dollar value to the oil company of using the airshed for waste disposal exceeds the dollar value people place upon addidonal clean air, the oil company will buy some of their clean air rights. When will this process stop? Only when the value to the company of discharging another unit of waste just equals the dollar value of the environmental damage caused from emitting it. Otherwise, further mutually beneficial trade is possible so that the efficient level will again ensue. Therefore, as these two examples show, it doesnt matter to whom the airshed property rights are assigned; an efficient level of pollution will be reached.

Of course, whether the property right is assigned to the citizens or to the oil company will make a big difference in their income and welfare. Efficiency, however, ignores all such effects and considers only whether gains from trade are exhausted. In other words, economic theory is not concerned with who captures the benefits, only that all potential benefits are captured by someone. The identity of these people is irrelevant for the purposes of the theory.

INTERNAUZING THE EXTERNALITIES

How can a market characterized by external costs such as environmental degradation be made to work more efficiendy? The economists answer is usually that external costs should be internalized, so that private firms take all social costs into account in their decisions by making the firm, instead of outsiders, pay these costs. This could be done, for example, by imposing an excise tax just equal to the marginal external cost. In the glue example, the idea is to shift

the marginal private cost curve up through the imposition of an excise tax so that it completely coincides with the marginal social cost curve. In other words, the marginal social cost curve and the post-tax supply curve would be identical after such a tax. For example, if each unit of glue production causes external damages worth $2 per unit of glue, then the excise tax should be $2 per unit. This proposition is illustrated in Figure IV where the imposition of a tax equal to the external costs of glue production automatically reduces competitive output to the efficient level, Q.

STANDARDS AND POLLUTIONS TAXES

In general the government has not looked to the market for the solution to the problem of environmental pollution. Instead emissions standards requiring all firms to cut their output of pollution by some percent or another are usually imposed. Such methods are inefficient because the resulting level of pollution control can inevitably be achieved at a lesser cosL We make our point in Table I showing the marginal cost of pollution abatement for two equally sized firms. 100% abatement would, of course, mean a complete elimination of pollution, while 0% abatement would mean uncontrolled emissions. An abatement of 44% would mean that the firm is cutting its output of pollutants by 44% of what they would be under uncontrolled conditions.

Looking at Firm A, we see that the marginal cost of abatement at the 47% level is $6. This means that increasing abatement from 46% to 47% will cause the firms costs to rise by $6. Looking at firm B, we see that increasing abatement from 47% to 48% will raise the firms costs by $13. The Table also shows that the marginal cost of abatement rises with the level of abatement. This makes sense. A litde cleanup comes cheap, but as the firms become cleaner (abate more), it becomes increashigly expensive to remove additional amounts of pollutants. We also see that it costs Firm more to abate than it does Firm A.

Suppose - as is usually the case - that the pollution authorities require all firms to meet a fixed pollution standard, say for example, a 50% reduction in their emissions. Then from the Table, the marginal cost of the last unit of abatement will be $9 for firm A and $15 for firm B. But this is inefficient: we can have the same amount of abatement at a lower total cost as follows. Suppose that the low cost company. Firm A, abates by another 1%, and that the high cost company, Firm B. reduces its abatement by 1 %. Then since the firms are of equal size, total abatement will not