Property, Power, and Public Choice

Property, Power and Public Choice.  copyright, 1987 and 1997, A. Allan Schmid.


  Another source of interdependence arises from a situation that will be termed joint impact goods (JIG) or marginal cost equal zero goods. Some goods can enter two or more persons' utility irreducibly.1 The term "irreducible" emphasizes that the good is described in physical terms. These goods are used but not consumed. Thus, A can utilize the good without subtracting from the quantity of the good available for B. In terms of the cost function, this relationship can be stated as follows: The marginal cost of another user is zero over some range (MC=0).2 This is the polar extreme of economies of scale where marginal cost declines and is less than average cost, but not zero. Note that the cost function is expressed in number of users and not units of the good. The relationship must exist for two or more people, but it need not apply to an infinite number of people. Interdependence arises over pricing policy. Who pays fixed cost and who pays only marginal cost?

The MC=0 characteristic may be combined with several other characteristics of goods and people that create different varieties of interdependence. These other characteristics are cost of avoidance, pre-emption exclusion cost, and degree of jointness. These will be briefly introduced below and given separate consideration later.

The classic example of a joint-impact good is national defense. Whatever level and kind of national defense is present in the United States is available to all residents. One person may feel more secure if the army has no nuclear weapons, while another person prefers a nuclear army. But, if they live in the same nation, they cannot simultaneously obtain their preference. Whatever exists is costly to avoid. While the same physical good can enter two or more persons' utilities, the satisfaction derived may differ greatly. If tastes differ, there is conflict, and property rights determine whose taste counts. If two persons have different tastes for some non-optional joint-impact good, they can negotiate to adjust the commonly available quantity and quality and the price each pays. But they must deal with each other rather than independently buying different amounts from producers. Consumers of non-joint-impact goods may, independently of other consumers, adjust the quantity taken to the price offered by the producer, but such adjustment is impossible with nonoptional joint-impact goods. If A has the initial right to choose, such right is a potential source of income to A if B tries to avoid that particular level of joint impact by bargaining.

Further, if the good is desired by B, but exclusion is not too expensive, A's ownership may generate income for A even when A's own use value exceeds the cost of supply--for example, A could drive his car to the office and take B along without cost but charge B for the ride. How much of the fixed cost can be charged to B given B's willingness to pay?

If the military force of the United States increases the feeling of security (or insecurity) of neighboring countries, this illustrates jointness is a matter of degree. For many people of the same taste in the United States, each one can enjoy equally without subtracting from the good available to others. But the strength of a neighbor does not give the same kind of security as one's own military. Therefore the goods are not perfect substitutes, and people in the country that has a strong neighbor may apply a discount to the physical good that enters their utility function. Jointness is usually a matter of degree and sometimes leads to a mixture of public and private financing.

If whatever total supply of military force exists is available for all (even if in differing degrees), the marginal cost for another user is zero. (Note that this is not the same as saying that the marginal cost of another physical output unit such as number of soldiers or amount of deterrence is zero.) The issue then is whether the marginal user is allowed to pay zero. Who is regarded as the marginal user is a matter for public choice.

The capacity for joint impact is found in many goods. For example, whatever ambient air quality is available for A to breathe, B can breathe it too at no extra cost. The fact that B can avoid it by carrying her own bottled supply does not alter the potential jointness. B may be a smoker and unable to perceive changes in air quality that are quite troublesome to a nonsmoker. Each cannot simultaneously and independently have just the ambient air characteristic that she wishes, except at a cost of avoidance. You can have corn to eat and I wheat and all we must pay is their individual costs of production. But, if you like classical and I like rock music, I not only need to buy my own records but also must pay the cost of insulating my ears from the sound of your nearby speaker.

Whatever variety of goods is available for purchase on the market is available to all. If selection is narrow, it is narrow for all. Variety often raises per-unit costs, as was seen in the case of economies of scale. One person may prefer fewer goods at lower unit cost, while another prefers a larger variety at higher costs. But there is no way for both to exist simultaneously for a national economy (although coexistence can be achieved on a limited scale in local markets and differences in stores). Another example of this type is the availability of television program channels. If A can choose among the three major U.S. networks (or three government channels in France), the same range of choice can be made available to all, even if some are satisfied with one or two at an appropriate saving in prices of advertised products, taxes, or cable fees.

Many visual experiences have joint impacts. A's house exterior and lawn give her more pleasure than is engendered by B's viewing them, but there is also jointness. If B would prefer a different view from her window than is afforded by A's taste in house and landscaping, there is no way to please both simultaneously if they deal independently with producers of housing and turf.

If A is inoculated for polio, this good decreases the probability of B contracting the disease since its incidence in the population is decreased. The good is probably discounted by B and is not as valuable as being vaccinated herself, but there is a degree of jointness. Education is similar in that the human capital of the educated person can benefit that person plus others who are better off for living in a community of educated people. Also note that the classroom and laboratory space are not joint-impact goods, and their marginal cost is not zero.

In the section to follow, several dimensions of joint impact goods will be analyzed. The source of interdependence and thus the kind of controlling right will be seen to vary with the degree of jointness, cost of avoidance, and pre-emption.


A joint-impact good enters two or more persons' utility, but the physical products may differ somewhat raising questions of relative cost share. The substitutability of a unit of the good created by another person for a unit initiated by yourself can vary. Examples noted above include military protection, education, and immunization.

Albert Breton defines a "nonprivate good" as one that "though not available equally to all, has the property that the amount available to one individual does not reduce that available to others by an equal amount" (1965, pp. 175-87). This definition suggests that jointness is a matter of degree. There is a debate in the literature over whether the most useful distinction is between incompatible-use goods (sometimes called private) and joint-impact goods (called public) as polar cases with most actual goods falling somewhere in between; or whether the incompatible-use good is a knife edge, with most goods having some degree of jointness. The former is argued by Richard Musgrave and the latter by Paul Samuelson. Samuelson believes that most goods have some degree of jointness in use or what he calls "consumption externality." His examples include cinemas, electric lines, and roads. Samuelson believes that the private good can be defined cleanly. The remainder of goods form a continuum of jointness within the category of joint-impact goods.

The argument is significant for welfare theorists because it affects the scope for application of the global welfare maximization propositions (to be examined in Chapter 11). The implications for forming hypotheses linking property rights alternatives to substantive performance are less clear. The analytic consequences of applying a different value (discount or premium) to the same physical good versus specifying the difference as two separate goods need further study. The author's preference is for the latter because physical differences can be observed, but it is hard to distinguish whether values attached to a given good are due to differences in accessibility or in preference. It is clear that empirical analysts must be very careful in specifying the good they are studying. For theoretical work it may be all right to draw a diagram whose horizontal axis is labeled "joint-impact good" (such as education), but this gross specification will not do for empirical work.


There are two other characteristics of JIG that affect human interdependence -- cost of avoidance and the degree to which the availability to person A limits the choice of quality or quantity available to B who has positive, but different utility. The interdependence issue focuses not only on cost sharing but also who gets to choose the quantity and quality of the JIG when a physical unit is to be produced or purchased.

Output and cost relationships are studied by varying quantity of the good. But, quantity is far from a simple concept. This chapter has shown the need for distinguishing another physical unit of a good from another user of a given physical unit. Further distinctions are necessary to understand sources of interdependence. The literature speaks of joint supply and joint consumption. While a given physical unit is potentially available without added cost to another user, it may or not be actually utilized or its level of use can be varied. The cost of this variation is defined as avoidance cost.3 A non-optional or high avoidance cost joint-impact good is one whose quantity of use can't be varied by an individual, but is determined by the available physical unit. The quantity available becomes the quantity used. The avoidance dimension of quantity is measured in such things as frequency and time of use in reference to people.

Quantity is always with reference to a set of quality characteristics that define the good. One variety of interdependence depends on whether people can simultaneously utilize different qualities. A pre-emptive good is defined as one where qualities used can't be varied by an individual independent of the use of others. The pre-emptive dimension is measured in physical characteristics of the good such as size, weight, speed, etc. These characteristics define quality.

The interaction of consequences of avoidance cost and pre-emptions are summarized in Table 1. Consider first a joint impact good which is unavoidable (non-optional) and whose existence is pre-emptive of some other physical qualities and quantities (Table 1, line 1). For any given quality and quantity, two or more people who place positive value on the good will conflict over cost sharing. Each would prefer to be the marginal user and pay zero. But, the pre-emptive character creates additional interdependence. If tastes differ, the individuals will not each be able to optimally adjust the physical quantity utilized, which is the same for all, to the given price . The quantity (or quality) purchased is not optimal because someone's choice is pre-empted by another's. The institutional structure controls who gets to choose the good to be purchased when only one variety can exist.

Ambient air quality and national defense are examples. Suppose the incompatible use questions have been settled by factor ownership.  Suppose breathers want to buy more rights than they now own. Whatever is the air quality for one breather is unavoidably utilized by other breathers and there will be an argument over production cost sharing. The good is omnipresent (unavoidable) and frequency of use can't vary by individuals. Further, one ambient quality level is pre-emptive over others since two qualities simultaneously are impossible. So there is conflict over the right to choose (purchase) the quality to be produced.  Different quality levels are incompatible in the sense that both can't exist simultaneously. Even if breathers A and B have positive utility, they may want to purchase different qualities. If A is free to choose her optimal level, B is constrained and vice versa. Income distribution is not determined solely be factor ownership or income, but by rights in collective choice (public finance if purchased or produced by government).

The positively valued, unavoidable, and low pre-emption good category may be an empty set. Intuitively, it would seem that if B has to utilize a good available to A, it would be impossible for B to utilize another level of the good. The characteristics that makes it unavoidable make it impossible to utilize more of it than does another person. For example, if a dam or dike is built along a stream, it gives a certain level of protection to everyone in the flood plain. If a person who wants a higher level of protection builds a dike just around their own property, it is not additive, but rather makes the commonly available supply worthless to that person. Put another way, there is an avoidance cost to escape the common supply and supply your own higher level.

While a good can be JIG among one group, it can be IUG with another group. So while outdoor music is JIG among neighboring music lovers, it is IUG between them and lover of quiet. As we have already seen, IUG is controlled by factor ownership. If person A play loud music in his/her back yard, the sound is IUG. Person B for whom music has negative utility wants to be paid if the good exists at all, but the bid will run the other way if the neighbor owns the IUG choice.

Turning now to avoidable (optional) JIG goods whose existence is pre-emptive of other physical quantities and qualities, the interdependence among those with positive utility is both a matter of cost share and who chooses (Table 1, line 3). The interdependence is less severe here because one individual can adjust utilization in some dimension such as time (avoidance), but, not in other dimensions such as size or speed. Consider a road between two points. What are its dimensions of utilization? It is optional in the sense that a person can avoid use altogether or can make any number of trips independent of others. In the quality dimension, the road may be one or more lanes, etc. There are more alternatives here than in the unavoidable, pre-emptive case (1), but people are still interdependent with respect to who chooses the quality and pre-empts the other. Each added lane is non-marginal and quite costly. If tastes related to speed and congestion differ, the right to choose the physical quantity (or quality) of the good at a given price is critical to performance even if it can be avoided all together and even if marginal cost of another user of the given physical quantity is zero. Other examples include various goods involving lines such as electric lines or cable TV. People do not have a complete inventory of their wealth unless it includes their right to influence public and private agencies which decide such things as quality of roads, cable TV systems, and telephone, electric and gas utility lines.

 TABLE 1: Joint Impact Among Members of a Group

Interdependence Created By Interaction of Avoidance Cost and Degree of Pre-Emption


UNAVOIDABLE                      PRE-EMPTION             INTERDEPENDENCE                
(Quantity of use)                            (Quality of good)

1. B-postive utility                               High                              Who chooses*                                 Example: Air
 '                                                                                                and cost share.

2. B-positive utility                              Low                               Empty set --

_________________________________________________ ______________________________________


3. B-positive utility,                             High                             Who chooses and                               
    but different from A                                                             cost share.  B can adjust frequency.

Example: Road

4. B-positive utility                              Low                             Only cost share (and cost                    Example: Cinema
'                                                                                             (per person and variety trade-off)

Note: The "B" in column one refers to the utility of person B, who is affected by A's actions. "A" could be a group, the members of which have marginal cost equal zero.
 *Who chooses in line 1 is even more important tht in line 4, since the quantity is unavoidable.

Non-pre-emptive and avoidable goods only create cost share conflicts among those with positive utility (Table 1, line 4). For example, a cinema is MC=0 up to capacity, but it is self contained and avoidable and non-pre-emptive. One person can go to one cinema once a week while another can go every night or to another cinema. Each person can adjust their frequency of utilization and quality independently. There is conflict over how to share the fixed cost if they go to the same theater. And if one person prefers a different picture, it does affect the per person cost of each. This is the interdependence trade-off between per person cost and variety explored in the previous chapter on economies of scale. Another parallel example is letter and package delivery, although MC is declining and not zero.

To summarize, with JIG the issue of cost share (price) can't be separated from choice of output level. If physical quantity is pre-emptive and can't be adjusted to price, any given price to all will not be acceptable to people with different preferences (even if the good is avoidable and frequency of use can be adjusted).

In contrast to standard welfare economics, the point here is not that unavoidable and pre-emptive goods prevent global optimal resource combinations, but that unless the relevant property rights are previously determined, optimality has no meaning. The conflict over who chooses when utilization is inherently inter-dependent is a fundamental distributive question as is factor ownership of incompatible use goods.


There is another reason to emphasize the potential availability for joint impact: Where exclusion is relatively cheap, a user may be denied access even if his use adds nothing to resource cost. Some goods with joint-impact can be private property and allow the exclusion of others. An example is broadcast television signals. A signal, once it exists, can be used by all who have receivers at no extra cost. But, with some expenditure, a scrambler can exclude users who do not pay the producer. The so-called social welfare and efficiency considerations of allowing ownership and exclusion when marginal costs are zero will be explored in Chapter 11. Here we will just note the implications for how one party can affect another.

The previous discussion of marginal cost being less than average cost can be extended to the case of marginal cost being zero. Many of the same points apply as were discussed in Chapter 4 with economies of scale.4 People do care how total cost is shared even when marginal cost is zero. The right to exclude in either case is an important factor in income distribution.

Consider, for example, the right to exclude people from making copies of scholarly journals. Once a journal is produced, the cost of making a copy for an additional user is minimal. If the original subscribers pay enough to support the journal, photocopies do not directly affect them. Yet it is in the interest of the publishers to control photocopying. And, if the publisher's profits could be controlled, the regular subscribers would appreciate the photocopier's helping to reduce their subscription costs. A U. S. court case declared that the publisher did not have the right to exclude a library from making extra photocopies of a journal that it subscribed to (Weinberg 1975). More recently, Kinko lost a suit and now pays royalties for copyrighted material. The same issue arises with copying musical and video recordings when producers are forced to accept marginal cost pricing (often zero). This is termed "fair use" although the term is value presumptive (see Schmid 1985).5 The value of the right to exclude is, however, limited by policing costs (Gadbaw and Richards 1988).

A television program once recorded on videotape may be used by any broadcast station or cable system at marginal cost equals zero to the producer. The right of program suppliers to sell exclusive rights to programs raises their income, but at the same time a station and its viewers are denied a good that costs no more resources to create. The sale of exclusive rights to show a certain program in a broadcast viewing area was common in the U.S. When cable television was developed, it threatened the value (advertizing revenue) of these local exclusive rights if the programs were imported from a distant station. New technology requires new public decisions on property rights. In the United States, the Federal Communications Commission limited the importation of these signals and protected the value of the exclusive right in 1972 (Besen 1974).

Copyright questions also arise with respect to cable television. A television signal once beamed from a station's broadcasting antenna can be picked up by a cable television firm for delivery to its subscribers. This use adds nothing to the cost of producing the original signal. Yet the courts in the United States seem to be leaning to the granting of copyright to the originating station if it is a "distant station."6 This enhances the station's income at cable subscribers' expense. Those who think price is a matter of demand and supply with the latter being technologically determined should take note. Price is related to proprietary scarcity, and thus so is income distribution.

A related example is the right to sell radio and television rights to sporting events. Sports teams and facilities cost no more to create if seen only by spectators that are present or enjoyed by distant television viewers and radio listeners. A source of great private fortunes in the United States is the right to exclude broadcasters and thus their consumers from this joint-impact good where the addition of the electronic use adds no additional cost to creating the event. Some professional sports teams were profitable before television was invented, and the new value created by use of new technology made multimillionaires out of a lucky few. Such is the stuff that the rich are made of. This is the type of rights that separates the very rich from the middle class, who basically only own their labor power and a bit of accumulated savings.

Any group that feels its income is unjustly low may wish to try to have itself declared the owner of some joint-impact good with zero or minimal marginal costs for additional users, but where exclusion is not too expensive. Why should the publisher make extra profit because of new copying technology? Why should the owner of a sports team make extra profit when television is invented? Why not the poor? Or why should the users have to pay anything more than the marginal costs? To users, price above marginal cost would be the same as if the government were to put an equivalent tax on all photocopies, televised sporting events, and cable television transmissions. If the profits were used to stimulate still newer inventions, the distributional impact would differ. Thus, the rights question is not only who pays cost of production, but who gets rents after total cost is recovered.

To summarize, the situation of marginal cost equal zero indicates that rights in pricing policy give an opportunity to affect income distribution. The income from the right to charge for a marginal use which has no cost of production can be allocated without necessarily affecting the firm's breakeven point or the goods physical supply. Depending upon society's moral judgment, this added use can be given to the consumer free of charge (as in the case of the video cassette recorder), to the owner who controls the decision on the fixed costs (as in the case of sports teams) or to any member of the public. With JIG, society has the luxury of pricing added uses over some range without affecting physical supply. If allowed, the revenue above marginal cost might be used for a "better product." Even if this happened, not all consumers want the "better" product.

A common policy is to tie price to cost. If another use or user causes extra cost, the person causing the extra cost must pay. But, with JIG, there is no extra cost. What then should the extra user pay? The producer would like to shift from the focus on cost to demand, and just ask what can be collected and thus maximize producer net revenues. This raises the moral question of why the producer is any more deserving of the extra revenue than any one else--be it the consumer or welfare mother. MC=0 breaks the necessary connection between bearing costs and receiving benefits. Who is the marginal user eligible to pay only marginal cost is a matter requiring public choice of rights. A subsidy can not be defined independently of rights.

With increasing cost goods, enforcement of comptition takes care of the price differentiation question. But, with decreasing cost or MC=0 goods, many firms and competiton may not be desirable even where possible and this necessitates a public choice of price differentiation rights whether the good is privately or publicly produced. Several dimensions of joint impact goods have now been discussed. for quick reference, the interaction of avoidance and exclusions cost is summarized in Table 2.


Interaction of Avoidance and Exclusion Costs with Respect to Joint-Impact Goods


'                                                Avoidance Optional                                         Avoidance Non-optional

Low exclusion cost.              1. Cable television system                                        Empty set*

'                                           2. Access to existing electric, gas, and
'                                               telephone lines up to capacity.

'                                           3. Cinema seats up to theater capacity

High exclusion cost               1. Broadcast television                                        1. Defense

'                                           2. Outdoor fireworks                                          2. Ambient air (breathing)

'                                           3. Local roads                                                    3. Flood control

'                                                                                                                     4. Audible sound
*Is it possible for A to exclude B, but if A decides to make it available, B cannot avoid it? If B cannot avoid it, can A utilize it and exclude B? No, for if B cannot exclude herself, neither can A exclude B.


Readers familiar with the literature in public finance and welfare economics will realize that what is here called joint-impact goods goes by many other names. It will perhaps make the implications of these goods clearer if some of the similarities as well as the reasons for rejection of these other terms are made explicit. The most common term applied to these goods is that of "public goods." This term is rejected because it already suggests the policy conclusion before analysis of consequences can be made. Samuelson (1969, p. 108) says, "For the (n+l)th time, let me repeat the warning that a public good should not necessarily be run by public rather than private enterprise." That this warning has to be repeated indicates it was an unfortunate choice of terms. Samuelson's definition is essentially the one used here. He uses the synonym "consumption externality" as well, but it is also misleading because an essential feature is that the goods are enjoyed but not consumed. This is why the above discussion avoids the term "consumed" and refers to enjoyment and utilization (or disutility).

Musgrave (1969, p. 126) uses the term "non-rivalness in consumption." His definition is "Social goods are defined as goods, the benefits from which are such that A's partaking therein does not interfere with the benefits derived by B. " Except for the value-laden term "social goods," this is similar to the definition used here. Musgrave refers to a good's substitutability and the discount that is applied to what he calls "outside consumption." While A's partaking does not interfere with some degree of partaking by B, the utility may differ greatly, as Musgrave notes. In one place, Musgrave defined them as goods that "must be consumed in equal amounts to all." Equal amounts refers to equal physical units.

Buchanan's categorization of "public goods" emphasizes the degree of divisibility and the number of persons involved. He illustrates his definitions by saying, "The death of one mosquito benefits each man simultaneously, and is thus equally available to each man" (1968, p. 11). He suggests that "the theory of public goods can be meaningfully discussed only when the units are defined as 'those which are jointly supplied' and when 'equal availability' and, less correctly, 'equal consumption' refer only to jointly supplied production units or inputs, which may and normally will embody widely divergent final consumption units, measured by ordinary quality and quantity standards" (p. 54). He emphasizes that because of the indivisible (pre-emptive) character of the goods, it is impossible for each individual to equate marginal valuations to tax charges (or market prices), and thus, for any given tax, some individuals would want more and others fewer governmentally provided goods. These points are consistent with the above discussion.

Goods in "joint supply" are not the same as joint-impact goods. The theory of joint supply was well developed by Marshall in the classic case of the activity of sheep raising producing both meat and wool, which is quite a different situation than that discussed here since joint supply products when produced by large numbers of suppliers can be independently purchased and consumed, with each person adjusting quantity to the price. Many public investments produce goods in "joint supply"--for example, a dam and reservoir produce irrigation water, hydro-power, flood control, and recreation. Individuals can adjust their own consumption to price although the MC of water use for hydro power is zero if the water is already produced for irrigation uses. There is an issue of allocation of the overhead or joint cost.

Another distinction is necessary. A joint-impact good is not the same as what is usually meant by the term "interdependent utilities." Earlier we discussed the fact that people can derive satisfaction from knowledge of the consumption of others. In that respect all goods conceivably could enter the utility of more than one person. It is not the good that enters, but rather knowledge of the good's use by others. The consequences of this type of interdependence must not be confused with those of joint-impact goods. Samuelson (1969, p. 109) notes that "little is left of the 'true property' of private goods in U, if people through altruism or envy have in their u1 the bread consumption of other people, namely x2 in u2." In the negative sense, all conspicuous consumption goods are subject to possible Veblen effects, where A feels worse after seeing B's ostentation.

There are differences of opinion on the relationship of jointness and externality. The discussion here differs from the general concept of spillover or externality, such as where the activity of steel production produces both steel and pollution. (Mishan, 1969, calls this case "private goods with external effects.") This general concept is in fact a misconception of the ordinary case of incompatible uses of a resource such as the air for either waste disposal or breathing. Of course, from the point of view of a buyer of rights to clean air (perhaps from a factory that owns the air), this clean air once created is here considered as a joint-impact good that can be utilized by more than one person. If the unit of product, is not carefully defined, we will find the same nominal good used to illustrate contradictory points.

In summary, the relationship between incompatible-use and joint-impact goods can be made more explicit, as shown in Table 3. An example of item 1 is when A eats corn, B cannot eat the same unit. The users are incompatible. Another example is when A consumes clean air as an input in the production of steel and B cannot consume clean air because it has smoke in it. Uses are incompatible. This fact sometimes is conceptualized confusingly as an exceptional externality where A engages in an activity that produces steel and also smoke, but it is simply an incompatible-use controllable by factor ownership.

An example of item 2 is when A consumes national defense and B must consume it also, but B places a negative value on it. National defense is in this case a non-optional joint-impact good. Use is compatible, but utility of use is not (some would call this a negative externality). The issue is who gets to own and choose the good in all of the above cases.

An example of item 3 is also national defense, but this time while B must consume, B places a positive value on it. This is a non-optional joint-impact good, but there is no technological incompatibility. Item 4 is illustrated by broadcast television. Person A consumes and B can consume if she wishes. It is a joint-impact good with no incompatibility. The issue in 3 and 4 is who pays how much when several use it. In case 4, if no B wishes to utilize the good, there is no conflict, in contrast to situation 3.


Relationship of Incompatible-Use and Joint-Impact Goods


1. A consumes X1, B cannot consume X1                                       Incompatible-Use Goods

2. A utilizes X2, B must utilize X2  (negative value)                            Non-optional Joint Impact Goods

3. A utilizes X 3, B must utilize X3  (positive value)                             Non-optional Joint Impact Goods

 4. A utilizes X4, B can utilize X4                                                        Optional Joint Impact Goods

Various other terms and distinctions appear in the literature. Some focus on activities, some on resources or products, some on production, and some on consumption. Each has analytic utility, and the last word remains to be written.


The above points can perhaps be summarized and further implications drawn in the context of an example mentioned above. Consider a certain geographical area where two groups of residents are considering how to protect themselves from outside enemies. One feels safe with a certain sized army equipped with 100 nuclear bombs. The other group feels safe with only ten nuclear bombs and, further, feels unsafe with 100. This is a nonoptional joint-impact good, and it is not possible for both levels of this good to be simultaneously available. In contrast, with consumption of say, apples, one group could eat one apple per day and the other could eat two (though both cannot eat the same unit). Units of bombs do have a positive marginal cost, but units of users do not. Incidentally, it is also possible that there is a third group that prefers no army and does not fear the enemy. One person's good may be another's bad.

Defense happens also to be a good with relatively high exclusion costs. This creates the possibility of free riders and may prevent a private producer from receiving bids sufficient to produce the good even when people are willing to pay. But it is important to understand the different consequences of high exclusion costs and high avoidance costs (see Burkhead and Miner 1971, pp. 29-31). With goods whose use is optional and different qualities simultaneously possible, the issue is who pays how much. But with non-optional and pre-emptive goods, the issue can also be who gets to choose the good. These situations will be illustrated intuitively below and more technically in Chapter 11.

The different implications of exclusion and avoidance costs can be illustrated by making the unreal assumption that a magic meter exists to measure each person's marginal valuation. Now the high exclusion cost creates no problem for preference revelation. Each person can be charged the value of his or her marginal utility for the commonly available amount of the good. Is that the only source of interdependence?

The answer can be made clear if we consider again the case of incompatible-use goods. When consumers are price takers, each can adjust the quantity taken so that each person's marginal value equals the price. With a non-optional joint-impact good, this situation is impossible. Whatever exists for one will be utilized by all, like it or not. If preferences differ, it is impossible for each person to equate marginal value to price. But, if we have a magic meter, it would be possible to charge different prices to each person so that each person's marginal value equals price for the commonly available supply (Lindahl equilibrium). Price could be negative. Between two or more people who have a positive value for the nonoptional product, it is a joint-impact good, but at the same time it can be an incompatible-use good between them and third parties with negative valuations. This point is illustrated if we assume that 100 nuclear bombs have zero or negative value to the pacifists and positive value to the insecure. It does not settle the problem for the pacifists to pay nothing. They want to be free of the overprotection or be compensated for it. Even when each person's marginal valuation is known, there must be a public decision on whose preferences will count when they conflict. Who has to pay whom to obtain or avoid an effect? The cost of production of nuclear bombs is given not only by the physical production function but also by the property rights decision of how and whether to count the negative utility of those who do not want the product. B's lost utility is an input into nuclear bomb production, but whether it is a cost, and how much, depends on property rights. Cost is a social phenomenon whereby shares in output are determined.

It is frequently assumed that it is sufficient for human interaction if everything is owned. But that assumption is deceiving because what must also be considered is how every effect of human interdependence is accounted for. If we look only literally at units of commodities, we will miss the many ways people can affect each other, especially in the case of joint-impact goods. In the present case, it appears that everything is owned. Someone owns the factory to make nuclear bombs and the inputs for their manufacture and deployment. Buyers own their income. But this factor ownership does not describe the totality of the relationships among people. Tastes can conflict in ways other than who owns a certain input factor. In the present case, there is a conflict among those who want zero, ten, and 100 nuclear bombs. The key right is the question of whether A owns the right to fulfill his tastes when they conflict with B. Does A own the right to purchase ten nuclear bombs and expose B to the discomfort of underprotection or does B own the right to have 100 and expose A to the disutility of overprotection? This question exists even if A did not have any obligation to help pay for the nuclear bombs.

It might appear that joint-impact goods are the best of all possible worlds since they can be enjoyed without being consumed. But, alas, when there are differences in taste, there is conflict, just as in incompatible goods. Incompatible goods require a right to determine who gets the particular unit. Joint-impact goods require a right to determine who gets to choose the commonly available units. (They also require a right to determine distribution of production costs among people with different positive marginal utilities, as discussed in Chapter 11.) Markets reflect the balance of rights that are present but fail to reflect those that are not present. We can speak of the failure of a particular market (or administrative system) with its particular set of rights to reflect the tastes of a given person or group, but not of market failure in general. One group's market failure is another group's preferred world.


Characteristics of goods take on importance only in relation to the kind and distribution of interests among people. Consider the case where the value of nuclear bombs beyond ten is zero to group A, but equal to production cost up to 100 for B. If A buys ten first, B only has to pay for 90. If B acts alone to buy 100, A will not need to buy any. In this case, B will try to get A to help pay for at least the average cost of ten, perhaps threatening not to buy any if A does not help. If A guesses the extent of B's tastes, this may not be a credible strategy on B's part. The outcome is indeteriminate except to say it depends on the relative bargaining strengths and skills of the parties. (For additional discussion see Chapter 11, especially Figure 6.) The level of output is very likely to depend on the agreement for cost sharing.

Bargaining in competitive markets (large numbers of buyers and sellers) is a misnomer (Morgenstern 1972). Consumers either buy or not. If enough do not buy a particular good at the initial offer price, either quantity supplied or price falls. There is no face-to-face negotiation or conversation between buyers and sellers. Haggling is a phenomenon associated only with small numbers situations, large purchases, and imperfect markets. In that case, one person's behavior depends on the expected reaction of the other party. The result is indeterminate in the usual economic model. Where people act independently and impersonally, their price-quantity behavior is relatively consistent over time and not greatly affected by the marketing transaction itself (though some learning may take place). A consumer presents a demand schedule in the usual sense. But when people meet face to face in a strategic game, consistent behavior is a disadvantegeous practice for the party using it. A person has no demand schedule since bids vary with perception of the strategic situation at the moment. However, the directional result may be predicted by game theory and knowledge of the opportunity set, the taste for gambling, and the possibility of benevolence and malevolence. Just knowing that people prefer more to less is not sufficient behavioral information to predict the outcome of game-theoretic situations (Shubik 1984, Ch. 19).

The perceived fairness of today's bargain influences willingness to bargain in the future. If A insists he has no demand for a joint-impact good and B later discovers that this was false, B may refuse to cooperate in the future. If both parties recognize this utility of maintaining credibility and trust, they may choose to reveal their demand and avoid strategic nonrevelation. Then again, they may become involved in escalating series of mistrust and deception.


The purpose of these classifications of interdependence is to provide a basis for studying how alternative institutions (rights) control the interdependence and thus affect performance. Therefore, it is important for empirical analysis that the inherent character of a good not be confused with the particular rule chosen for it.

For example, in Table 1, flood control is considered a joint-impact good, not a public or private good. The form of ownership, type of transaction, boundary of firm, level of administration, and pricing rule, are institutions that affect performance in the context of interdependence created by marginal cost being zero over some range combined with the degree of exclusion cost and avoidance cost. Institutional rules can alter the incidence or consequence of joint-impact features. It is possible to take an incompatible-use, low exclusion-cost good with positive marginal cost and not charge the marginal user. But this does not make it a joint-impact good. Performance cannot be understood if institutional structure and inherent goods features are confused. If a person is not satisfied with the level of a joint-impact good, several alternatives might be pursued each with different implications for cost sharing and who chooses:

1. Accept grudgingly. Many, goods have some degree of joint impact where the initiation of supply by one person enters into another's utility without added cost. In the great bulk of these, the law gives the right to choose to whoever acts. If A puts up a modern house in a traditional neighborhood, it is frequently his right to do so and the neighbors enjoy the view without paying or must simply accept exposure to the disutility. They might try persuasion or even make some monetary offer, but in the bulk of the cases this is simply not worth the effort (which might be considerable because of the free-rider problem discussed previously). Segregation is sometimes used to group people with similar tastes, but this creates other conflicts.

2. Buy more yourself (assuming you are not offended by unequal cost shares). Where the good has positive value and you regard the total supply as inadequate, you may simply pay producers to provide more. For example, A may have been spraying her yard for mosquitoes with only limited effect. She may decide next to pay a commercial firm to spray around the neighborhood.

3. Bargain. You can pay others to purchase more or less of the good. (This assumes that others have no obligation to purchase any. This assumption can be reversed and bargaining proceed from there.) For example, you might pay your neighbors to spray their own yards, paint their houses, mow their lawns, etc.

If a person wants more of a nonoptional joint-impact good where all persons have positive value for it, the problem is cost sharing and perhaps who chooses the quality and physical quantity. If A is providing some to himself and B, but B wants more, B can either produce or bargain with A to produce more. An example of such a bargain is the classic case of beekeepers and apple growers. Beekeepers may use a certain number of bees to produce honey, but apple growers may want more for pollination. For a given number of bees and honey, the MC of another use (for pollination) is zero. Bees are a nonoptional, joint-impact input with positive value for both honey and apple producers. But it cannot be assumed that, as they begin to bargain and explore their interdependence, honey producers will be willing to expand if apple growers do not also offer to help pay for the present supply of bees. Does the apple grower have a right to receive free the joint product of the honey producer's bees and only pay for the marginal cost of more bees; or does the honey producer have the right to expect others to help pay for the bees' joint uses? In the small-number case, the situation is fraught with game strategy. Malevolence may develop, and the joint-impact good may not be produced despite its mutual benefit. This may be especially true if apple producers can't agree on the number and quality of bees to bargain for.

Where there are opposite tastes, the right for A to buy more or pay B to use more creates an exposure for C if the goods utilization cannot be avoided. The payment can be a matter of private contract, or a high demander may petition the government to offer a payment financed by taxes. Most governments offer monetary incentives to citizens to increase their purchase of certain goods (not all of which are joint-impact goods, however).

Even where everyone regards more physical units of the non-optional good as having positive value, there can be conflict over the physical quantity and quality of the good. The relevant rights are contained in such things as minimal or maximal consumption requirements.

4. Owners of a low exclusion cost JIG may sell it. Note that a good may be JIG among one set of users but incompatible between that group and a second group. If the owner in the first group can sell to someone in the second, others in the first group who had formerly not been excluded lose as the good is transformed by incompatible use. Government may therefore make the right non-exchangeable or set rules for decisions to sell by all joint owners.

5. Government provision. Even when all benefit from the good, there is argument over cost sharing and who chooses the quantity and quality of the good. Rights pursuant to tax laws are important factors in who pays how much for joint-impact goods and thus in the distribution of welfare. As in the case of private bargaining, failure to agree may mean the mutually beneficial good is not produced.

Government provision causes everyone to purchase more. Whether others actually utilize more depends on the costs of avoiding the available supply. The total supply of nuclear bombs cannot be avoided. The total variety of television programs and channels can. The offer of publicly provided education or innoculation services can be avoided, but laws are frequently passed to require schooling and some vaccinations, though the source can be private. Some people will not notice or be able to utilize the good, but if taxes are used, they become unwilling riders in the sense that they have to pay for something they do not want. In the case of nonoptional goods, even if they pay nothing, they may be unwilling riders if they do not put a positive value on the commonly available supply. Even when a person pays no taxes, he is affected by the tax expenditures of others. Where joint-impact goods are present, it benefits a person to have similar tastes as his neighbors. One might expect people to make a certain investment in creating similar tastes in their fellows. But such investment runs into free-rider consequences. Exclusion and joint impact are often interrelated.

6. Regulate. In the case of nonoptional joint goods, a person negatively affected can turn to administrative transactions, which in effect give A the right to be free of a certain level of utilization by B or, in the case of positively valued goods, regulation may give A the right to expect B to maintain at least a certain level of utilization on her own.7 Regulation shifts factor ownership and the beneficiary pays nothing, but the person regulated has a cost.

The regulatory alternative directs the interdependence of joint-impactness in a particular way. An "unwilling rider" is a person who has to pay a tax for an unwanted governmental good or who by regulation has to provide privately and/or utilize a good at an unwanted level. A person exposed to regulation requiring a limit to private utilization may in parallel fashion be termed an "unwilling abstainer" (prohibition of alcoholic beverages is both a literal and figurative example). For example, a zoning regulation may prohibit the use of flat modern roofs, which are offensive to traditional tastes. In general, land-use regulations are concerned with a variety of utilization patterns that are joint-impact goods and keep land use within certain bounds. If tastes differ, unwilling abstainers who cannot use their resources as they wish are the result.

Where you wish to increase the utilization of a certain good, you may seek a law requiring a minimum utilization. For example, the city may require all homeowners to clear their walks of snow. This requirement applies even to those who never leave the house except in a car. The supply of sidewalks has joint impacts, and one person cannot leave it covered if another is going to be able to walk along it. A law may require all dog owners to get rabies shots for their dogs. Another law might require certain human innoculations. In the case of education, the law may require one to stay in school until a certain age is reached. But one person's right is another's exposure. If tastes differ, someone is an unwilling rider. This is true whether A has the right to expect that B will stay in school for a certain period or whether B has the right to expose A to the disutility of living in a world where her neighbors cannot read or write. The right to choose a privately or publicly provided joint-impact good is an important factor in the distribution of welfare.

Each of these five alternatives relates to a different composition of property rights and probably a different performance depending on the specifics. Rights in addition to competitive rules affect who gets to choose the kind and amount of joint-impact goods and the sharing of costs.

Those readers schooled in the market failure approach will note that it focuses the issue on market versus administrative transactions in order to achieve consumer sovereignty and achieve optimum output. The approach here focuses explicitly on who pays and raises the additional question of who gets to choose the product. It suggests that the key question is not how to achieve general consumer sovereignty, but which consumer is to be sovereign. Ownership of money and factors of production do not completely control this type of interdependence. In the situation of JIG, the rights in market pricing and tax incidence are critical as well as regulation of consumption and production.

The continuum of topics in Chapter 4 and 5 can now be summarized. Some goods have production (cost) functions such that there is not a marginal relationship linking input and output. This means that cost of output loses some of its meaning and there is no physical relationship to use as the bases for prices. There is something in common among goods marked by economies of scale, joint impact, overhead cost, and team production (input complementarity). The unit of measure and perspective differs, but all involve an interuption of the link between input and output often referred to as lumpy or indivisable.

1. Economies of Scale: A particular physical unit of the same output can't be traced to a particular unit of input. (The source of this is fixed cost, team production, or complements.)

2. Joint Impact Good: A particular user of the same output can't be traced to a particular unit of input.

3. Overhead Cost: A particular use from among multiple outputs can't be traced to a particular unit of input. (This is discussed in Chapter 11.)

All of these situations require a pricing rule (property right) to allocate the advantages of doing something together. Prices of outputs or inputs (including wages) can't be derived from the production function. Cost is not simply a physical phenomenon, but is a matter of public choice of shares. Subsidy can't be defined independently of rights.


 1. The definition (but not the term) is that of Samuelson (1969, p. 108). He illustrates the need to distinguish the physical good from a person's sensing the good by noting that the definition applies to fireworks in the sky, rather than fireworks as registered on the retina of A's eyes versus fireworks observed by B. The latter definition would define joint-impact goods out of existence. Note that the good perceived by the different users can be quite different. This makes no difference for analysis as long as there is some expression of demand for the given physical commodity. The quantity unit is number of users for a given physical quantity and quality. Congestion must be accounted for since the marginal cost of another user is not zero if quality of the good changes with another user.

2. Joint-impact goods should not be confused with free goods, which may have a positive marginal cost of production but which are not scarce. Neither should they be confused with scarce goods that have positive marginal costs but that are provided free to the user as a matter of policy. Joint-impact goods are the opposite of incompatible-use goods where marginal cost is not zero. Note that a durable good can be joint impact if used sequentially though incompatible for simultaneous use. This means rights of stewardship are important to preserve the potential flow of benefits.

3. Mishan (1981 p. 431) defines an optional collective good as one "which a person may take all he wishes (up to what is available) at the going price . . ." This is ambiguous. Mishan gives the following examples (p. 438): television transmission, the services of parks, bridges, light houses, street lights, museums, theatres, and galleries. He is probably speaking of the frequency of use but is never explicit. He does contrast variation in telephone service in the short run (number of calls) with long run adjustments in the system. Mishan then defines a non-optional collective good as one "which bestows itself in some particular amount on each of the beneficiaries." His example (p. 439) of times during the year that clouds are seeded and the amount of rainfall precipitated indicate that this is a dimension of physical units. The reference to amounts is ambiguous because it does not distinguish to avoidance dimension from the pre-emption dimension. Other texts ignore the distinction altogether (Boadway and Bruce 1984, Ch.4).

4. When marginal cost is zero and when there is a sufficient plant to serve one person, all can be served. To create additional firms and plants in order to get competition (to control bargaining power) is to increase costs without increasing output (people served).

5. Columbia Broadcasting System v. Teleprompter, 476 F.2d 338, 349 (2d Cir. 1973). The United States Supreme Court earlier ruled that copyright did not apply to cable carriage of signals from nearby broadcast stations where the terrain made it difficult for some viewers to obtain a satisfactory signal. Fortnightly Corp. v. United Artists Televisionc Inc., 392 U.S. 390 (1968). Also see Posner (1972a). For a view arguing against copyright liability and program exclusivity see Chazen and Ross (1970).

6. Where this is a transferable right for A, when B wants to change the utilization, A can receive bids from B, rather than having to pay B, as with the bargaining option.

If you have any questions or comments, please email

Back to Main Page