II. THE FOUNDATIONS: SELF-INTEREST, EXCHANGE, AND VOLUNTARISM

The purpose of this section is not to demonstrate that Buchanan was convinced that the main motive behind individual behavior is the pursuit of their self- or private interest. This is well known and would not be original. We are rather interested in explaining, and documenting, why Buchanan found “voluntarism”—which he then equated with “[t]he willing approval in the payment of taxes for all normal expenditures of the state” (1952b, p. 601)^{Footnote 3}—crucial and how he came to connect it to self-interest. And, from this perspective, it appears that we have to go back to the very beginning of his career, at the end of the 1940s, and to the essays he wrote while he was still a student at the University of Chicago: “A Theory of Financial Balance in a Federal State”—his 1947 essay for “Economics 362” (Federal and Local Taxation)—and “The Problem of Fiscal Inequality in a Federal State”—his doctoral dissertation, which he defended in 1948.

In these essays, Buchanan analyzed the “dilemma of federalism” (1947) or the “federal fiscal dilemma” (1948) that results from the differences in the fiscal capacities that exist between the various member states that comprise a federation, the related inequalities that result from these differences, and the “pressure of the subordinate units of a federation against their limited financial resources” (1947, p. 1). However, Buchanan refused to choose between, on the one hand, a federal regime with inequalities and different fiscal capacities and, on the other hand, no inequalities and homogenized fiscal capacities but without federalism. His objective was not to preserve “the federated political structure” (1948, p. 4) but to remove the problem that threatens it—the differences in capacities and the inequalities that lead the “subordinate units” to ask for more money from the central government. The solution he envisaged to remove the problem consisted in using a principle of fiscal justice, “as old as Aristotle” and “contained in every formulation of ‘justice’ in taxation from the latter part of the 17th century” (1948, p. 38),^{Footnote 4} and that aims at an “equal treatment for equals.”^{Footnote 5} To him, “[w]ith its application the so-called dilemma of federalism is theoretically solved” (1947, p. 17). Or, “the fundamental fiscal inequality problem of a federal state can be theoretically resolved without further resort to those more complex and disputed principles of fiscal justice and their application” (1948, p. 43).

Now, quite obviously, to know if “equals” are treated “equally” in fiscal terms, one needs to know how much individuals pay in taxes and how many public goods and services they have at their disposal. This means that, to solve the federal fiscal dilemma, Buchanan chose a principle of justice that implies that individuals receive “benefits” in exchange for the taxes they pay. Or, in other words, Buchanan chose a principle of fiscal justice that goes with a theory of public finance based on a “benefit principle.” And, in those years, the economists who shared those views on public finance were also those who adopted a “voluntary exchange” theory—Knut Wicksell, Erik Lindhal, Emil Sax, Adolph Wagner, and also the members of the so-called Italian School of Finance, Antonio De Viti de Marco, Maffeo Pantaleoni, and Ugo Mazzola.^{Footnote 6} Therefore, Buchanan’s interest in a specific principle of fiscal justice directed him to “voluntary exchange” theories in public finance.^{Footnote 7}

This is the central idea upon which all the “voluntary exchange” theorists, at least in Buchanan’s understanding as it appears in his dissertation (1948) or in his first published article (1949), agreed: all fiscal processes are exchanges or transactions that take place between the government and the taxpayers, and this is the major—if not only—justification for taxes and any fiscal contributions the government can ask taxpayers to pay. In other words, as Buchanan said in his dissertation, taxes are a “return for services rendered” (1948, p. 44), “the price of the protection provided [by the state]” (1948, p. 44); a few pages later, he insisted that “taxes or contributions paid are exchanged for services rendered by the political unit” (1948, p. 52; emphasis added). And, almost word for word in the “version” published in 1949, taxes are the “payment made by individuals out of their economic resources in exchange for services provided” (1949b, p. 498; emphasis added). Complementarily, it can be said that taxes are not a “burden” or “net subtractions from social income, never to be returned” (Buchanan, Reference Buchanan1949b, p. 500), as it is assumed they are in the alternative, “organismic” (1948, 1949a, 1949b) approach he identified in public finance.^{Footnote 8} To him, it is “impossible to speak of the ‘burden of taxation’ without considering, at the same time, the benefits from expenditure made out of such taxation” (1949b, p. 501; emphasis added). By excluding benefits—that is, by seeing taxes as a burden only—the organismic approach bases taxation on each individual’s ability to pay. By contrast, the benefit approach bases taxation on the individual’s willingness to pay. Therefore, benefits are precisely the reason that explains why transactions with the government are voluntary: taxes are paid because individuals receive benefits from paying them.

It is crucial to understand the nature of those benefits—that is, of the services rendered by the political unit or what taxes are paid for—to sustain our argument about voluntarism at the post-constitutional level. In effect, in the first place, Buchanan thought that the twofold idea of benefits and exchange with the government has its roots in the social contract approach of Thomas Hobbes, John Locke, and Hugo Grotius,^{Footnote 9} indicating that the taxes were paid in exchange for specific services: the protection of “human and property rights” (1948, p. 44). This could legitimate the thesis that taxes are paid because individuals have signed a social contract and gives ground for the social contract thesis about coercion. However, Buchanan did not make much use of the social contract foundations of the benefit/voluntary exchange approach and quite rapidly abandoned the references to social contract theorists; Hobbes, Locke, and Grotius were not cited or quoted in his article on public finance published in 1949.

In addition, it seems that Buchanan did not restrict the goods and services supplied by the government to the protection of property rights only and envisaged a wider range of public goods and services; this seemed to be the case in 1949, and became clearer in his article on marginal cost pricing (1951) and even clearer in his article on highway pricing (1952a). Finally, and more significant, Buchanan put forward another justification to defend a benefit/voluntary exchange approach: the transactions individuals engage in with the government are not specific compared with any other transaction between “private” individuals. Buchanan stressed that point in his first article, published in 1949, but was clearer in his dissertation, in which he not only insisted on the idea of exchange but also noted the implied similitude with private transactions: “The fundamental position taken by this theory is that the individual’s financial relations with government are basically analogous to any private economic transaction; they consist of an exchange relationship” (Buchanan Reference Buchanan1948, pp. 51–52). Thus, taxes are the price individuals pay to obtain the goods and services supplied by the state. This does not exclude a social contract, but the social contract does not appear to be necessary. Individuals simply pay taxes because they buy something from the government, even without a social contract, the reason being, as is the case in private transactions, that individuals pay a price because they receive a good in exchange. This refers to self-interest. Individuals pay taxes because their private, or self-, interest is satisfied.

One point must nonetheless be clarified to confirm that Buchanan actually meant self-interest when he was talking about benefits. Retrospectively, this might not seem necessary, in light of Buchanan’s well-known defense of self-interest. But a certain ambiguity seems to characterize his writings. Buchanan indeed spoke about the “collective willingness of individuals to purchase” (1949b, p. 499) public services, and referred to the “the collective equalization of benefits and taxes” (1949b, p. 502) and to “collective wants” (1948, p. 52) or “collective desires” (1949b, p. 505). And, even more ambiguously, he explained that “[i]deally, the fiscal process represents a quid pro quo transaction between the government and all individuals collectively considered” (1949b, p. 499). But this should not lead us to think that he did not reason in individual terms. First, he explicitly criticized the organismic approach for being anti-individualistic;^{Footnote 10} and, complementarily, clearly stated that the benefit principle and the voluntary exchange approach are individualist.^{Footnote 11} Second, to Buchanan, the collectivity, however, is nothing more—or less—than the sum of the individuals and does not—cannot—exist independently from the individuals. As a consequence, “[i]f benefits from public services accrue to individuals as a group (and this is impossible to deny), it follows that specific benefits are received by particular individuals” (Buchanan Reference Buchanan1949b, p. 500). Indeed, through benefits, Buchanan clearly put self-interest at the center of the relationship between the state and the individuals.

Certainly, these benefits are difficult, to say the least, to evaluate for each individual—essentially because they are subjective. And accordingly, it will be “extremely difficult, if not impossible” to determine the exact tax each beneficiary would have to pay in proportion “to his gain in subjective utility resulting from the free provision of the service” (Buchanan Reference Buchanan1951, p. 175). But this aspect of the discussion is secondary. One should not confuse a practical with a fundamental problem. In that case, the difficulty of evaluating individual benefits is not substantial, but simply empirical. What matters is that each individual gains something from paying their taxes. And, from this perspective, there is no doubt that benefits exist. Buchanan thus wrote that the individual’s “satisfactions are increased by the government services provided him and theoretically this increase in utility can be reduced to value terms” (1948, p. 56). Indeed, his defense of benefits, and accordingly of “voluntary exchange,” was not pure speculation, the logical conclusion of a reasoning without empirical dimension. To Buchanan, the existence of benefits and his belief in an individualist benefit principle were not simply grounded in a philosophical conviction. The existence of benefits are a fact,^{Footnote 12} which can be verified empirically and can be observed. Indeed, to Buchanan, the very fact that individuals pay their taxes or vote for politicians who propose programs involving public expenditures and taxation proves that they benefit from paying these taxes. Otherwise, if there were no benefits, one could not understand such behaviors. This point has rightly been emphasized by Wicksell, as Buchanan noted (1949b, p. 500).

In addition, the observation that individuals do pay their taxes empirically confirmed the reality and effectiveness of voluntarism and its connection to self-interest—individuals do voluntarily pay their taxes because they benefit from the services these taxes finance and because they realize private gains from trade. The price individuals are ready to pay—and their demand at this price—indicates the upper limit of the price they are willing to pay, exactly as is the case with private goods. Of course, the greater the difference, the larger the individual surplus. And, sometimes—in particular, when they are able to “conceal their true preferences”—certain individuals “will be able to secure the proportionately larger share of the gains to be made by the ‘trade’” (1960, p. 238) But the size of the individual surpluses is not a problem. What matters is only the existence of the surplus, and, more precisely, a surplus that allows mutual gains from trade. Then, as long as individuals gain something from the trade—with the government or with other individuals—they voluntarily pay their taxes or pay for the public goods they want to consume or internalize the external effects their actions generate. Or, to put it differently, the Pareto frontier will be reached.

Buchanan then shifted to a discussion in terms of efficiency. This was not at the core of his first works (1947, 1948, 1949a, and 1949b). It was only in 1950 that he came to refer to efficiency and to resource allocation. And it was only in 1951 that he explained why a Wicksellian scheme, based on exchange and mutual gains from trade, guarantees an optimal allocation of resources because Wicksell’s theory is “in accordance with the principle of unanimity” (1951, p. 177). In effect, and quite obviously, “unanimity” means that “no one is worse off if some allocation of the required tax can be found which is acceptable to everyone” (1951, p. 177). He even explained in a paper about how to finance highways that one could use the individuals’ willingness to pay for what they consume, even in the case of impure public goods with external effects like roads. One could ask users to pay a price for these goods exactly as was the case for private goods. In that case, however, the “correct price” for highway services should take interdependencies into account and should be equalized “to the marginal social cost incurred in providing a unit of that type of service” (1952a, p. 100). Individuals would pay for what they consume and should pay for the costs imposed on others. That was a strong statement, meaning that relying on the individuals’ willingness to pay would not only allow for the gathering of a sufficient amount of money to finance the provision of the goods but also the covering of the difference between the private and the social costs.

III. “THE LIMITS OF SELF-INTEREST”: MISALLOCATION OF RESOURCES, MARKET FAILURES, AND SOCIAL DISINTEGRATION

Thus, in accordance with the voluntary exchange theory, Buchanan believed that individuals voluntarily and spontaneously act “in the direction of group or social optimality, such as defined by the Pareto conditions” (Buchanan Reference Buchanan1967, p. 114). They pay for what they consume —and he even went as far as saying that the individuals’ willingness to pay includes the external effects their behavior has on others. And, as we tried to show in the preceding section, this type of behavior is perfectly explained in terms of self-interest. It is, therefore, because individuals gain from their “pro-social” behaviors that they adopt them. In other words, it is therefore because they are self-interested and their self-interest is satisfied.

His confidence in human beings, and in self-interest, never really disappeared. But Buchanan was not a blind believer in self-interest. He admitted that self-interest should be controlled. He made the clearer statements—at least, the most well known—in that direction in the 1970s. He then wrote about “the limits of self-interest” ([1976] 1979, pp. 210–213), explaining that “[e]ven in those aspects of economic intercourse that involve no externalities or spillover effects in the Pigovian sense, some limits must be imposed on the working of pure self-interest” ([1976] 1979, p. 210; emphasis added). The word “pure” indicates that he did not target self-interest per se. Indeed, he criticized the excess of self-interest—“the observed behavior of the modern American is excessively ‘self-interested’” (1978, p. 364)—and its narrowness, repeatedly speaking of the “narrowly defined self-interest” (1978, pp. 365, 366, 367) rather than of self-interest in itself. He even stressed the need to accept “a minimal set of moral standards” ([1976] 1979, p. 211), “standards which may not, in specific instances, be consistent with objectively measurable self-interest” ([1976] 1979, p. 211) and without which a “viable social order” ([1976] 1979, p. 212) would not be maintained. Precisely, in the 1970s, he believed that “[s]uch acceptance [was] by no means assured” ([1976] 1979, p. 212). Indeed, Buchanan insisted on these excesses because he saw them as the origins of the deterioration of the situation of his country. He was therefore describing a specific context linked to a specific type of behavior, and held these views as a consequence of such a context (Fontaine Reference Fontaine2012). But Buchanan did not come to believe that self-interest was problematic in the early 1970s only. Since the 1950s, and all throughout the years that preceded the arrival of those “new barbarians” (Buchanan Reference Buchanan1970), Buchanan emphasized the limits of self-interest in some of his previous works and, to be more precise, in his very first writings.

One would recall that, as was mentioned in the preceding section, the essays he wrote in 1947 and 1948 were devoted to fiscal justice and the need to treat equally equal individuals.^{Footnote 13} One would also recall that the need to base public finance on individual benefits —that is, on self-interest, in our interpretation—was a consequence of the use of this principle of justice. But, complementarily, a large part of Buchanan’s analysis consisted of showing that self-interest could have negative consequences on the functioning of the economy. He made this point in his 1947 essay and in his 1948 dissertation, and then proposed a more synthetic demonstration in “Federalism and Fiscal Equity,” his second published article (1950a).

In this paper, Buchanan discussed the consequences that cannot but result from the different fiscal capacities that exist between states within a federal state. From this perspective, federal systems have a specific, and twofold, problem to face. The first is a static problem which comes from the fact that states with lower fiscal capacities are not able to provide the same amount or quality of public goods and services or are obliged to impose greater fiscal pressures to reach the same level as rich states. Therefore, individuals who live in states with lower fiscal capacities are not treated like their equals in states with higher fiscal capacities; “[p]ersons earning the same income and possessing the same amount of property” would “be subjected to a much greater fiscal pressure in Mississippi than in New York, solely because of residence in Mississippi” (Buchanan 1950a, p. 591). This is problematic because it violates the principle of fiscal justice Buchanan adopted and the geographical neutrality that should result from its application.

Then, and this is the second problem, fiscal injustice also has negative dynamic consequences. In effect, individuals, acting out of self-interest, will migrate from poor states to the rich ones: “If ‘equals’ are thus pressed more in one area than in another, there will be provided an incentive for migration of both human and non-human resources into the areas of least fiscal pressures” (1950a, p. 589). One may note, at this point, that Buchanan did not make any reference to the costs implied by migration; in other words, he implicitly assumed that individuals were fully mobile. As a consequence, they would logically move to the “areas” in which the net benefits from paying taxes are the greatest. Nothing can be done about that. It is the logical or mechanical consequence of the normal behavior self-interested individuals adopt. Or, in other words, this is how a market economy functions. But this kind of normality could have rather dramatic consequences, according to Buchanan. He gave the following explanation. If “the fiscal balance for equals is not made equivalent for all areas of the economy” (Buchanan 1950a, p. 589)—that is, if the allocation of resources was based on “economic criteria alone” (Buchanan 1950a, p. 589; emphasis added)—mobility would contribute to increasing the interstate differences in fiscal capacities. This would create a really important economic problem—in Buchanan’s words, “a considerable distortion of resources” (Buchanan 1950a, p. 589). And, progressively, because individuals follow what their self-interest tells them to do, the federal structure would disappear. This is the result of the normal functioning of a market economy: “[t]he laissez faire result will be the ultimate centralization of a large share of effective political power” (Buchanan 1950a, p. 599; emphasis in original). This is an evolution that is not acceptable for someone attached to federalism, as Buchanan claimed he was.

That a market economy—a system in which private choices are made by self-interested individuals—does not always lead to a socially optimal allocation of resources was not a premature claim made by a young scholar unsure of his ideas. Buchanan did repeat, quite frequently, that under certain circumstances, markets were not capable of efficiently allocating resources. He wrote many (unpublished) notes in the first part of the 1950s, in which he insisted on the fact that, “when there exist external diseconomies in utilization, competitive pricing does not lead to an optimum” (Buchanan Reference Buchanan1954a, p. 3), or “competitive pricing could never take into account the spillover effects” (Buchanan Reference Buchanan1954a, p. 3). In effect, competitive markets “tend to equalize private or individual marginal enjoyments but since interdependence is present these tend to be greater than social marginal enjoyments” (Buchanan Reference Buchanan1954a, p. 5). To reach an optimal allocation of resources, “it is necessary that the prices of highway services be set equal to the marginal social costs of providing such services” (Buchanan Reference Buchanan1954a, p. 6); that is, it is necessary that the price “includes the incremental costs (or reduced enjoyments) imposed upon other road users!” (Buchanan Reference Buchanan1954a, p. 6).

Similarly, in “Consumption Interdependence and the Interpretation of Social Cost” (1954b), Buchanan repeated that “[u]nder certain conditions it appears evident that the argument is not valid even if the somewhat questionable case of external economies in production is excluded from consideration. In the presence of either external economies or diseconomies of consumption, the competitive economy should not appear to allocate resources properly” (1954b, p. 2). Thus, implicitly because he did not make specific comments about individual behaviors and simply described how markets work, Buchanan argued that self-interested individuals do not make socially optimum choices. They may well voluntarily pay for what they consume but that is not enough to guarantee an optimal allocation of resources. Eventually, in the published version based on these notes, “Private Ownership and Common Usage: The Road Case Re-Examined” (1956), Buchanan was also clear that markets could not lead to an efficient allocation of resources in the presence of externalities.

Does the failure of markets imply that we must abandon a Wicksellian framework based on a benefit principle and its related voluntary contributions? Does it mean that we must resort to coercion to oblige individuals to pay taxes, to act in a collectively efficient way, to act “pro-socially”? Or, does it mean that decentralized mechanisms are no longer efficient? No. This is precisely wherein lie the differences between Buchanan and other welfare economists and this is what we show in the next section: how Buchanan escaped the contradiction between voluntary cooperation and market failures by reconciling market failures with his ambition to define a positive welfare economics based on what individuals want.

IV. MARKETS, POLITICS, AND THE SIZE OF GROUPS: THE CONTINUITY OF DECENTRALIZED MEANS OF ACTION

Through the works of, among others, Samuelson (Reference Samuelson1954, Reference Samuelson1955), Bator (Reference Bator1958), and Musgrave (Reference Musgrave1959), economists paid more and more attention to market failures in the second half of the 1950s and inevitably associated the phenomenon with a more general or global failure of any form of decentralized mechanisms to allocate resources efficiently. And, accordingly, the intervention of the state was viewed as the only possible solution to correct those failures and inefficiencies and to promote an efficient—Pareto-optimal—allocation of resources.

This line of reasoning, this connection of arguments, seemed all the more logical, given that, at that time, in the late 1950s and early 1960s, the idea that governments could also fail was still in limbo.^{Footnote 14} Rare, to say the least, were those who had already envisaged that a comparative institutional analysis was required before drawing any conclusion in terms of which institution—the market or the state—was relatively more performant than the other in terms of how to promote an efficient allocation of resources in the presence of externalities or public goods. Among the exceptions were two economists who played an important role in Buchanan’s career and with whom he was acquainted: Ronald Coase (1960)—one of his colleagues at the University of Virginia from 1958 to 1964—and Roland McKean (Reference McKean1964) —one of Buchanan’s friends, an acquaintance from when he went to the RAND corporation in 1954 and a UVA economist too. Indeed, rare in the profession, the idea was nonetheless admitted in the small group of economists who constituted the Thomas Jefferson Center that Buchanan founded with Warren Nutter in 1957 at the University of Virginia. No surprise that Buchanan was also opposed to the intervention of the state.

Buchanan, however, did not make exactly the same claims as Coase and McKean about government failures. He did not seem to be (at least in his earlier works, written in the 1950s and 1960s) fundamentally and primarily interested in the costs and (in)efficiency of the intervention of the state.^{Footnote 15} He was not using efficiency arguments aginst the intervention of the state because of its supposed inefficiency. The point he made against the recourse to the solution favored by new welfare economists was—one could say, once again—ethical and was based on a defense of individualism. To him, those approaches defending the intervention of the state were the equivalent of the organismic approaches he had denounced in 1949. They define the society as a “whole” (Buchanan Reference Buchanan1954c, p. 116), building a “social value scale” from individual preferences but their approaches are not genuinely individualistic. Indeed, their “proper approach to social welfare functions appears to begin with the frank admission that such functions are social decisions, not individual” (Buchanan Reference Buchanan1954c, p. 118). Then, quite straightforwardly, those approaches were normative: first, because the objective to reach in the economy—the allocation of resources to promote—is defined in the normative terms of Pareto optimality; second, because the Pareto-optimal allocations are defined in reference to a social welfare function, artificially built by the economists; and third, and accordingly, because the gap between the current situation and the ideal—Pareto-optimal—allocation is also evaluated by the economists. Therefore, all the process that guides the intervention of the state to correct market failures is mastered by “omniscient” (Buchanan 1959c, p. 126) economists who impose their external norm on the individuals.^{Footnote 16}

To Buchanan, any perspective based on the intervention of the state necessarily implies or means that the state has the right or legitimacy to act on behalf of the individuals and, complementarily, as emphasized above, to ignore the individuals, to solve problems independently from the individuals. Such an approach could not be accepted because it contradicted the individualist foundations that Buchanan had chosen for his analysis in 1949, had repeated in 1954 (1954c, in particular), and that he repeated again in 1959. Indeed, no one but the individuals themselves—this means not even the economists (see Buchanan Reference Buchanan1962b; Buchanan and Tullock, Reference Buchanan and Tullock1964)^{Footnote 17}—has the right to tell the individuals what to do. Any approach to economic problems must start with and from the individuals. One must let the individuals decide what they want to do. For instance, if an economist calculates that an allocation of resources is efficient and corresponds to an optimum, but that the individuals are not ready to accept this allocation of resources, this first means that the allocation is not optimal for the individuals and, second, implies that one cannot force the individuals to act in a way that would lead to this allocation of resources:

He then wrote that “[u]sing a presumptive efficiency criterion, he suggests a possible course of action which the group may take” (1959, p. 130),

“[T]he economist can only conclude that the presumptive efficiency criterion was wrongly conceived and the hypothesis based upon it falsified” (1959, p. 131). Welfare economics should be positive, rather than normative, based on what individuals want and what they do.

It is therefore what individuals do that indicates whether or not a presumably efficient allocation of resources is effectively efficient; that is, if a market actually fails or not. A market failure exists if and only if one can observe individuals acting to remove or solve it. Otherwise, there is no failure. Then, from this perspective, one understands that the failure of markets cannot be generalized as a failure of decentralized means of coordination.^{Footnote 18} If a market fails to allocate resources efficiently and if individuals really perceive the allocation is inefficient, then they will engage in another form of collective action and devise private arrangements to solve their problem. Such private and collective arrangements are obvious extensions of market exchanges—and therefore more a proof that decentralized mechanisms work rather than an index that they fail. They correspond to the behaviors that Buchanan considered as ‘politics,’ ‘political action,’ or ‘collective action.’^{Footnote 19} In his views, politics complements economic market transactions when they fail. This is what Buchanan started to put forward in the mid-1950s, with papers like “Social Choice, Democracy, and Free Markets” (1954c) and “Individual Choice in Voting and the Market” (1954d), and frequently repeated over the years, especially in the 1960s.

Then, the next question is: Why does politics represent a non-contractual, post-constitutional solution to market failures?

V. SELF-INTEREST, ETHICS, AND THE SIZE OF GROUPS

Answering the question requires us to go back to a distinction Buchanan made between how individuals behave on markets and how they behave in politics. He indeed found erroneous and misleading those approaches that did not distinguish precisely between these two environments and that assumed that behaviors in different environments or contexts were identical. Ignoring the difference meant failing to capture an essential feature: the attitude towards others and the capacity to take (or not) into account the consequence of one’s actions on others. And, obviously, this is important from the perspective of a discussion of market failures, since market failures occur because individuals do not take others into account. Obviously, this was what Musgrave assumed and Samuelson too. And this was also a mistake that Kenneth Arrow had made in his 1951 book Social Choice and Individual Values. The reference to Arrow makes sense because the first time Buchanan stressed this distinction between economics and politics was in his review of Arrow’s book (1954c).

The same year, another of Buchanan’s articles was published in which he insisted that, on markets, when they make economic decisions, individuals behave as if they were alone or independent from others or as if there existed no interdependencies with other individuals: “[t]he individual tends to act as if all the social variables are determined outside his own behavior, which, in this subjective sense, is non-participating and therefore nonsocial” (Buchanan Reference Buchanan1954d, p. 336). This is precisely the reason that explains market failures. By contrast, when individuals act “collectively,” as is the case when they vote, it could reasonably be assumed that individuals are aware of the existence of interdependencies and that they take them into account. Once again, Buchanan put that claim forward in the mid-1950s. He thus wrote, in 1954, that “[i]t seems probable that the representative individual will act in accordance with a different preference scale when he realizes that he is choosing for the group rather than merely for himself” (Buchanan Reference Buchanan1954d, p. 336). A few years later, he gave a similar account of this distinction, writing that, on markets, “[n]o considerations of the ‘public’ or the ‘social’ interest are assumed to enter this … calculus” (Buchanan Reference Buchanan1962a, p. 20). By contrast, in collective action, “each individual … tries to identify himself with the community of which he is a member and … tries to act in the genuine interest of the whole group” (Buchanan Reference Buchanan1962a, p. 22; emphasis added). And he added an interesting remark about the fact that individuals who spontaneously and voluntarily internalize the external effects of their behavior can be said to follow a “Kantian-like rule of action” (Buchanan Reference Buchanan1962a, p. 22; emphasis added). There is no need for coercion.

One important point for our analysis and the role of self-interest is that Buchanan did not replace self-interest with ethics. In effect, although there exists a difference in terms of behavior between markets and politics, this does not, however, mean that their motivations differ. On the contrary, motivations are identical, in the sense that individuals are supposed to act out of self-interest in politics—that is, when they engage in collective action—as well as on markets. In other words, Buchanan’s distinction between economics and politics and between two types of behavior did not mean a criticism or imply a rejection of self-interest. Actually, following a “Kantian-like rule of action” is consistent with behaving self-interestedly. Indeed, to Buchanan, even if individuals behave differently in different contexts and happen to follow a Kantian rule of action, they nonetheless remain self-interested. Buchanan was clear about that when he criticized the analyses based on the assumption, made by “Pigou and his followers,” that individuals “respond to different motives when they participate in market and in political activity” (1962a, p. 23). This assumption of a “bifurcated man” (Buchanan Reference Buchanan1962a, p. 23) or of “behavioral dichotomy” (p. 24) must be rejected.^{Footnote 20} It is “clearly extreme” (p. 24) and “naive” (p. 25), because it implies that individuals are “inconsistent” (p. 25) in their behaviors and motives. To Buchanan, if one assumes that individuals are self-interested—and Buchanan does not see how a different assumption could be made if one wants to make sense of individual behavior—then consistency requires that we stick to this assumption that individuals are self-interested. Market failures may be a consequence of self-interest but this does not imply a condemnation of self-interest. Buchanan clearly departed from the view adopted by (new) welfare economists.

Individuals may adopt a Kantian-like rule of action and therefore engage in collective action because they have signed a social contract. That would be a plausible interpretation of this idea of “collective action.” This would be all the more plausible, given that 1962 was also the year of publication of The Calculus of Consent. However, there are no indications in the works we are studying that Buchanan believed that individuals adopt such pro-social behaviors because they have signed a social contract. Even if one admits that a social contract explains that individuals follow an ethical rule of action and take the consequences of their action into account, it also seems that Buchanan believed that post-constitutional solutions were possible. For instance, in “What Should Economists Do?,” published in 1964 but written in 1963, Buchanan did not refer to any social contract, but he did claim that individuals will resort to a constitutional solution only if they are not able to find a non-constitutional solution to their problems. Examining the case of draining a swamp, he noted:

Most of the time, when they have a problem to solve, “[i]ndividual citizens will be led, because of the same propensity, to search voluntarily for more inclusive trading or exchange arrangements” (1964, p. 219).

During the same period, Buchanan started to work on “numbers” in 1964, when he was working on his “economic theory of clubs” (1965a). In a letter sent to George S. Tolley (20 January 1965), Buchanan wrote that his paper on clubs was going to be published soon and also that he was working on various papers that were all—but one—linked to externalities and free riding: “Externality, Jointness, and Numbers” and “Cartels, Clubs, and Free Riders.” On 16 January 1965, he wrote a paper significantly entitled “A Probabilistic Approach to the ‘Free Rider’ Problem.”^{Footnote 21} The title echoes “A Probabilistic Approach to the Free Rider’s Choice,” the title of a section in The Demand and Supply of Public Goods ([1968] 1999, pp. 88–89), but this draft was actually the preliminary version of “Ethical Rules, Expected Values, and Large Numbers.” In this article, Buchanan no longer envisaged cooperation (and its corollary, free riding) and studied “[t]he societal dimension of free riding” (Fontaine Reference Fontaine2014, p. 13) or, for that matter, the societal dimension of cooperation. In that case, “societal” refers to “numbers,” the size of the group to which individuals belong.^{Footnote 22} The size of groups was becoming a key variable in understanding cooperation and free riding. And the argument was that individuals do follow the Kantian-like rule of action in small groups, in which there is no social contract—at least, Buchanan does not mention that a social contract could exist in those groups. Thus, to put it in other words, coercion is not required in small groups even though there is no social contract and even though individuals remain self-interested.

More precisely, Buchanan explained why self-interest does not necessarily exclude ethics and why individuals may self-interestedly benefit from following a moral rule that leads them to spontaneously internalize the effects of their actions on others. More precisely, he claimed that individuals could always choose between two kinds of rules: “the moral law” or “the expediency criterion”; that is, ethics is simply a matter of following one type of rule rather than the other. It does not mean behaving self-interestedly or not. In effect, the choice to follow one rule rather than the other, the “critical determinant” that “influences an individual’s choice among ethical rules [is …] the size of the group” (Buchanan Reference Buchanan1965b, p. 1), because it “depends upon his own predictions about the behavior of others” (1965b, p. 3). Therefore, in small-number environments, individuals anticipate that their behavior will affect others and self-interestedly choose to “follow the moral law,” while, in large-number environments, they “follow the expediency criterion.”

As a consequence, it seems clear that, in small groups, because they follow the moral rule, individuals behave pro-socially. Or, to put it in different words, in small groups, individuals behave strategically, mainly because they anticipate or predict that others’ behaviors are also based on the same moral rule. Free-riding behaviors are unlikely. Each individual therefore accepts to bargain, when asked to do so, with those who suffer from external effects: “[i]n small-number groups, individuals suffering damages from external diseconomies imposed by the behavior of others can directly initiate arrangements through which they offer appropriate compensations in exchange for some contraction in the scope of the activities in question” (1966, p. 38); externalities can be settled directly and spontaneously through private bargaining. There is therefore no need to oblige individuals to pay for the external effects they cause. In that case, bargaining difficulties and transaction costs are not obstacles to the possibility of devising “market-like contractual arrangements” (1973, p. 69). Coercion is absolutely not necessary.

The situation is different in large-number groups, where the expediency criterion that individuals follow leads them to ignore the costs (or benefits) they impose on others. They do not behave strategically, or strictly follow their self-interest and treat others as if they were a part of their environment. This type of behavior generates what Buchanan names the “large-number ethical dilemma,” because of which the initiation of direct action and private bargaining is unlikely (1966, p. 38), and because of which market failures remain a problem. In those cases, however, it is not because individuals self-interestedly follow the “expediency criterion” that government intervention and therefore coercion should be used. What has to be done is to change “the rules through which activities are allowed to take place” (1966, p. 38). Buchanan thus suggested that a change in the rules of the game—that is, a change in the constitutional rules—is likely to solve the problem of social cost (see a more detailed analysis in Buchanan [Reference Buchanan1973]). This is the only relevant solution, and it consists of devising a constitutional framework within which individuals will be led to follow the moral rule rather than the expediency criterion.

There is another instance in which coercion might not be necessary. This corresponds to the cases when groups are heterogenous—the same group may be composed of cooperators and of free riders—or when the society as a whole is composed of heterogenous groups—groups of “enlightened” may co-exist along with groups of “nonenlightened” individuals. Buchanan does not conclude that, in those instances, coercion is necessary. To him, cooperators, either individually or groups of them, may find it in their interest to tolerate free riders (1968). Such tolerance is even normal, usual in every society: “[t]he individual members of the nonenlightened group are not expected to share in such community ‘responsibilities,’ and they may even be tolerated in activities that are often acknowledged to be offsetting to the group-benefiting activities of the others” (1968, pp. 357–358). This is true up to a certain point only. Potential cooperators may be less and less willing “to provide gratuitous transfers” (1968, p. 358). When free riders are too numerous or too dangerous for cooperators, it may become necessary to resort to constitutional rules and to a social contract. Once again, the social contrat appears to be a second, compared with the first, solution that rests on ethics and voluntary cooperation.

This is confirmed by a note Buchanan prepared for the Public Choice conference of 1970 and entitled “The New Barbarians.” He explained there were three possible means to deal with pollution, or “public bads,” that free riders create. The first is a Kantian solution that corresponds to the case in which individuals try to “take others’ well-being into account in all aspects of their own behavior” (1970, p. 2). The second is a Smithian solution, which consists of rearranging ownership rights “in such a way as to cause persons, behaving selfishly or in their own private interest, to act in what we may legitimately call the public or social interest” (1970, p. 4). The third solution would be that “we … through explicit or public or collective choice, invest in the provision of the ‘public good’ that represents depollution” (1970, p. 6). But the third solution, the third mechanism, is required if the first two are impossible to implement. Significantly, Buchanan wrote:

Indeed, in a world in which individuals spontaneously behave pro-socially, in which individuals adopt a Kantian-like behavior, there are no problems of pollution; there are no “public bads” and therefore there is no need for a social contract.