In the summer of 1954, CE (Ed.) Lindblom and I were invited as guest scholars to the then-flourishing Rand Corporation in Santa Monica, California. Our assignment was to carry out a generalized overview of Rand’s research, particularly that done by its economists. It was soon apparent that these economists considered that their primary contribution was simply to elaborate, variously, the elementary notion of opportunity cost. TANSTAAFL (there ain’t no such thing as a free lunch) was, and remains, as a first principle to be mastered on the way toward rudimentary economic understanding. Something of value can only be secured if something else of value that could be produced does not come into being. Naive and utopian notions that good things can be created from nothing must be dispelled.
It soon became apparent to us, however, that Rand’s economists were often, themselves, naive in their implicit presumption that the decision authorities to whom the research findings were addressed were guided by relative "social" value of the alternatives within the relevant choice sets. Rand’s clients, the military authorities, and primarily the United States Air Force, did not choose among their weapons systems options on the basis of comparative nationwide or even systemwide objectives; these authorities chose among such options on the basis of their own objectives which may not have coincided with those aggregate efficiency norms, as measured by the economists.
This example suggests that the simple version of opportunity cost, although possibly valuable as a first step, is not sufficiently sophisticated to be of much explanatory use. It is necessary to examine the words "opportunity" and "cost" more carefully. My small book, Cost and Choice: An Inquiry in Economic Theory (1969), was an exploratory inquiry.
The word opportunity suggests the presence of alternatives as they are confronted by a chooser or decision maker. When we say that someone has an opportunity, say, to go to university, we imply that the person in question could, if she so chooses, reject this opportunity and do something else other than go to university. For the person without financial means, we say that she does not have such an opportunity, by which we mean that she has no choice along this dimension.
Opportunity is, therefore, intimately related to choice, and it has little or no meaning outside of some context of choice. Recognition of this point implies, in turn, that "cost" also when used with "opportunity" must be related directly to the choice setting. "Cost" becomes, quite straightforwardly, the value of the next best alternative that is sacrificed or given up in order to secure that which is chosen.
It follows from this elementary logical relationship that a choice, as confronted by a decision maker, cannot be informed by any scale of comparative values that are external to this decision maker. Opportunity cost, therefore, must necessarily be reckoned in a subjective utility dimension; it cannot be represented in some objectively measurable commodity or resource dimension.
Once this point is accepted, the elementary TANSTAAFL principle seems to be in jeopardy. Suppose that the money or numeraire value of a lunch is, say, five dollars. This principle correctly implies that something that is worth five dollars in the economy that might have been produced has not been produced. "Society" has, somehow, given up the "opportunity" to use five dollars’ worth of resources in some way other than providing the lunch.
Well and good. But whose choice has been involved here? Whose evaluation on resources has been determining in making the lunch available? For the person to whom the lunch is offered without charge, TISTAAFL (there is such thing as a free lunch) rather than TANSTAAFL applies. The foregone opportunity involved in choosing to take the free lunch is at least close to zero and is in no way connected to the resource value embodied in the lunch as provided. But who, then, has suffered the opportunity cost that might be reflected somehow in this latter value? Suppose that the free lunch was financed by a tax levied on the citizenry generally. Who chose to impose the tax?
Consider a familiar democratic process. The tax is imposed by a majority of the members of an elected legislature. We may examine the calculus of a decisive member of the majority coalition that makes the fiscal choice here. Such a person does not face the opportunity cost of the lunch financed from tax revenues in any sense remotely related to the measured objective economic value of the lunch itself. There is, of course, an opportunity cost involved in this legislator’s choice measured either by her evaluation of some alternative item of collective outlay or by her evaluation of taxpayers’ disposition of the five dollars.
The choice calculus just discussed depends critically, however, on the presumption that the legislator is decisive in determining the collective outcome. Consider, however, the setting where a single legislator is one among many in a large body. This person is now called upon to vote for the imposition of the tax necessary to finance the free lunch program. What is the opportunity cost of a vote for the free lunch? It is obvious that this choice may involve little or no reckoning that is even remotely related to the resource value actually involved in making the lunch available collectively from tax sources. That which is sacrificed in a vote for the lunch is, quite simply, a vote against the free lunch. And, in a large group, the value placed on either side of the account may be negligible. In such a setting, as for the final recipient of the lunch, TISTAAFL rather than TANSTAAFL may be more descriptive of the setting. The member of a large-number group who is asked to participate in a collective choice, whether this be a large-number legislative body or a referendum among all voters, finds that the opportunity cost of voting expressively is low indeed (Brennan and Buchanan, 1984; Brennan and Lomasky, 1993). The voter may find that the utility loss involved in giving up a negative vote is almost nonexistent, and if she is moved at all by considerations for the potential recipient of the free lunch, a positive vote may emerge with little or no reckoning of the ultimate resource value that the lunch embodies.
Although the emphasis is seldom expressed in this way, the differences in the opportunity-cost setting offers an effective means of distinguishing "public choice" from "private choice." In the latter, that is, in private choice, as exemplified in the stylized market setting, the individual chooser, whether this be a buyer or seller, bears the total incidence of the choice. The person who pays five dollars for her own lunch bears the full opportunity cost, as measured by the anticipated utility loss from having to forego the enjoyment of that which is given up when the outlay on the lunch is chosen. In the idealized market setting, in which there are no externalities or spillover effects on those who are not direct parties to exchange, there is no value or utility loss suffered by others, just as there is no value or utility gain or benefit. Indeed, this is precisely what is conveyed by the adjective "private" when appended to "choice." In this idealized market setting, TANSTAAFL applies fully and without qualification.
The linkage between the two sides of the choice account, so to speak, is broken once the incidence of any decision extends outside the direct exchange between contracting parties. In this setting, any choice becomes "public" to some degree. But the existence of external or spillover effects does not, at least directly, have implications for the opportunity cost of the choice alternatives. The setting is modified in the sense that the person confronted with choice does not secure either the full utility gains or the anticipated utility losses of the decision to be made. On the other hand, precisely because two sides of the choice calculus are reckoned in utility terms, the decision maker may well include her own evaluation of others’ utility losses. The breakdown of the effective linkage between the two sides of the choice account, as introduced by the presence of external or spillover effects, can be used to generate predictions about behavioral changes only under the postulate that choosers are, in fact, more influenced by their own experienced utility benefits and losses, as anticipated, than they are by the benefits and losses that their choices might impose on others. In other, and familiar, terms, if incentives matter, any attenuation of the effective incidence of choice must have predictable behavioral consequences.
As analysis moves beyond the effects of externalities on choice behavior in market-like exchanges and to the institutional structures that explicitly involve "public choices," that is decisions to be made on behalf of, and applicable to, all members of a collective unit, the attenuation of incidence becomes the central feature. A person, no matter what her role, who explicitly chooses for others cannot internalize the benefit and loss utility flows enjoyed or suffered by those person who actually experience the effects of choices made. At best, such a public choosing agent can base her calculus on a translation of these utility flows into her own utility dimensions.
In sum, the degree of "publicness" in any choice may be measured by the attenuation of the incidence of the effects of the decision taken, with idealized market choice at the one extreme of the imagined spectrum here and externally imposed collective choice at the other. Returning to the simple lunch illustration, the person who buys her own lunch is identified as being at the one limit, while the person who pays no taxes at all but is the effective decision maker that imposes taxes on others to finance the free lunch program is located at the other limit.
The subject matter of public choice, inclusively defined, as a research program or subdiscipline, concentrates attention on the choice calculus of persons who are located between these limits, and more specifically on this calculus within those institutional settings where the alternatives are explicitly public or collective. In any such choice calculus, it is immediately evident that TANSTAAFL is not fully applicable, if the two sides of the choice account are understood to be chooser-experienced ex ante utility gains and losses.
The opportunity cost faced by an agent choosing for the whole collectivity, or participating in such choice in any capacity, cannot, by definition, include the anticipated utility losses suffered by others from failure to choose the relevant alternative to that which is chosen. This agent can always, to an extent, if she is so inclined, enjoy the equivalent to the free lunch, with no reckoning for the utility losses suffered by others.
The "market failure" logic that was central to the welfare economics of the mid-twentieth century decades was based on the Pigovian distinction between private cost and social cost. Departures from the idealized limit of market exchanges with no external or spillover effects were identified to warrant politicized correction. Contributors to this strand of literature failed to recognize that any proposed politicized correction, based on the same behavioral models as those allowing for the identification of market failure, would itself embody attenuation of the incidence of the effects of any action (Buchanan, 1962). Public choosers, or choosers for the public, in any capacities (as voters, elected representatives, bureaucrats) cannot, by definition, choose among options on other than calculations of their own anticipated utility gains and losses. The effective scalar cannot, again by definition, be equivalent either to the utility gains and losses experienced by those affected by the actions, or to some imagined dimension measured by economists’ reckoning of "social" cost.
Many of the contributions from the related subdisciplines or research programs variously described under public choice may be reinterpreted as applications of the basic principle of opportunity cost, properly understood, to varying choice settings. The commonly encountered statement that "institutions matter" says little more than that differing settings for choice present choosers with differing opportunities. The anticipated value of that which is foregone when a choice is made, which is the proper meaning of opportunity cost, can never be objectified and quantified in such fashion as to make specification of the parameters of choice unnecessary.
