In 1965 James Buchanan (1919–2013) and Friedrich Hayek exchanged letters about an article Buchanan had recently published in Ethics. In the course of the correspondence, both Buchanan and Hayek expressed their enthusiasm for the recent work of John Rawls (1921–2002). Buchanan wrote, “I agree on the work of John Rawls. Gordon Tullock and I discovered him, or he discovered us, I forget which, shortly after our book was published, and we have since got to know him personally. I agree that he is doing the most interesting work in philosophy and I hope that he has some impact.”Footnote 1
A striking aspect of the early reception of Rawls is that he was warmly embraced by figures such as Hayek and Buchanan, who subsequently became famous as the intellectual leaders of the global market-liberal revival of the late twentieth century. While Hayek's role in the rise of this “neoliberalism” has been well documented, Buchanan was also an influential advocate of restricting the capacity of democratic collective action to regulate the economy and redistribute resources.Footnote 2 This seems some distance from the more familiar political images of John Rawls as either offering a quasi-Hegelian rationalization of the mid-twentieth-century welfare state or a radical departure from it in the direction of equalizing private property ownership or perhaps even socialism.Footnote 3 Yet Hayek remarked that, although he regretted Rawls's use of the term “social justice,” Rawls's presentation of principles of justice as setting out a framework of rules for institutions and action was “more or less what I have been trying to argue.”Footnote 4 In his earlier letter to Hayek, Buchanan recollected that he had been encouraged when visiting London to hear lectures by the sociologist W. G. Runciman on the BBC Third Programme “which accepted without question Rawls's approach to justice.” “If Rawls's ideas can begin to affect such men as these,” Buchanan continued, “there is perhaps some hope.”Footnote 5
Buchanan had written to Runciman about the lectures, expressing admiration for his use of the “justice-as-fairness” framework. Buchanan explained that although he and his coauthor Gordon Tullock took a slightly different path from Rawls in their 1962 book The Calculus of Consent, “in the final analysis, we reach the same point.” In writing that book, he stressed, “we were not familiar with Rawls's work, except only incidentally, but since that time we have worked closely with him.”Footnote 6 Indeed, Rawls was even nominated for membership of the Mont Pèlerin Society (MPS)—the invitation-only organization of intellectuals famously convened by Hayek to remake the principles of classical liberalism.Footnote 7
Viewed philosophically, this neoliberal enthusiasm for Rawls could certainly be argued to rest on a telling underestimate of the egalitarian implications of Rawls's project. But as Barbara Fried has observed, although Rawls is now regarded as offering one pole of a grand political argument between social democrats and neoliberals, A Theory of Justice was composed before it was evident that this would become such a central dividing line in political argument in Western democracies in the 1970s and 1980s. As a result, Rawls himself actually said surprisingly little about the neoliberal critique of social justice.Footnote 8 Although it subsequently became apparent to later neoliberals and Rawlsians that their political thought differed in critical ways, a historical account that tracks the emergence of both Rawlsian liberalism and neoliberalism reveals a more complicated story, in which shared influences and assumptions generated overlapping theories that took time to distinguish themselves from one another.
The relationship between Rawls and Buchanan provides the clearest window into this complex theoretical unraveling, since they not only shared important methodological similarities but also maintained a friendly intellectual exchange across their careers.Footnote 9 Buchanan in particular continued to make frequent references to Rawls as a kindred spirit, even as the gap between Rawlsian liberalism and Buchanan's own immediate political allies became more visible. In 2005, looking back on his career and his efforts to defend “classical liberalism” as a “realistic utopia,” Buchanan said, “in this whole enterprise, I have, from the 1960s, felt a strong affinity for the efforts of John Rawls who was embarked on a quite similar enterprise, although our two efforts have been interpreted quite differently.”Footnote 10 Geoffrey Brennan, a close collaborator and friend of Buchanan's, recollected that at a Liberty Fund conference in 1982 Anthony Flew mounted a fierce attack on Rawls “and was astounded at the severity of Buchanan's response. Buchanan was as defensive of Rawls as he was enraged by John F. Kennedy.”Footnote 11
This affinity between Buchanan and Rawls has been noted in passing by a number of commentators, and explored analytically by certain political theorists, but its significance for our historical understanding of both Rawls and Buchanan remains to be established.Footnote 12 In this article we first examine the shared intellectual context that framed the early work of both Buchanan and Rawls and brought them together as seeming philosophical allies during the 1960s. We then analyze a significant theoretical divergence between them in the 1970s—which we see as related to their contrasting responses to the politics of those years and which generated an early version of the debate about ideal theory and realism in political thought. Indeed, part of the interest of a comparative study of Buchanan and Rawls lies in observing how these two theorists developed contrasting positions out of a common stock of intellectual resources.
Political economy, games and the ethics of competition
The emergent historiography on Rawls makes the important point that Rawls's theory of justice was a product of the 1940s and 1950s rather than the 1960s—the essentials of what subsequently became A Theory of Justice had been settled long before its publication in 1971. Rawls's protean intellectual trajectory has been shown to have begun with liberal Protestant theology in the 1940s and then shifted in the 1950s towards a Wittgenstein-influenced ethical naturalism grounded on interpersonal recognition and community.Footnote 13 A similar story can be told with respect to Rawls's engagement with political economy. The economics that underpinned A Theory of Justice blended together the cutting-edge welfare economics, game theory and social choice theory of the mid-twentieth century.Footnote 14 Like Rawls's, James Buchanan's writings were also initially a response to the debates among economists in the 1940s and 1950s about rational behavior, market failure and collective action.
In broad terms, Buchanan and Rawls demonstrate the way in which normative theorists of the mid-twentieth century sought to revive and legitimate their projects by drawing on economic theory. As has been discussed elsewhere, two vectors of economic argument were significant for postwar political theory and for Buchanan and Rawls in particular: welfare economics and game theory. First, in welfare economics during the 1930s and 1940s there had been a backlash against the traditional utilitarian analysis that identified how to maximize social welfare through cardinal utility measurements. This approach now appeared to rest on a mistake, since it assumed that it was possible to compare the utility levels of different individuals when this was not, or so it was argued, scientifically possible. Instead, a new wave of welfare economists—led by Abram Bergson, John Hicks, Nicholas Kaldor and Paul Samuelson—argued that only so-called “Pareto-optimal” improvements could be justified by economists, since policies that met this criterion would improve the position of someone without making anyone else worse off. Simultaneously, fearing that the Pareto criterion was ultimately a relatively weak and uninformative guide for public policy, some of these “new welfare economists” argued that it would also be scientifically legitimate to scrutinize the economic implications of different normative decisions about social welfare. It was this opening of economics to normative theory that Kenneth Arrow famously disrupted when he argued in his 1951 Social Choice and Individual Values that it was not in fact possible to reach a coherent collective choice about social welfare that reflected the individual preferences of the members of a given community.Footnote 15 The early work of Buchanan and Rawls therefore emerged in the context of a debate about whether normative theory and economic analysis could adjudicate on distributional issues, which in turn raised the further question whether public policy could ever be said to be grounded on the free consent of citizens.
Second, Buchanan and Rawls drew on game theory, which emerged in the 1940s as a flexible set of techniques for analyzing the interaction between rational agents that could be applied to a number of disciplines.Footnote 16 As historians of game theory have noted, its rise to prominence reflected the particular technocratic preoccupations of Western democracies governed by militarized states making the transition to the new circumstances of the Cold War after 1945. This new age of technological governance raised acute questions about rationality and choice while new interdisciplinary forms of research in the social sciences—drawing on the mathematical techniques of the hard sciences—were advanced on a much larger scale through the patronage of states and think tanks.Footnote 17 The publication of John von Neumann and Oscar Morgenstern's Theory of Games and Economic Behavior in 1944 stimulated many social scientists and philosophers, among them Buchanan and Rawls, into deploying the techniques of rational choice in their own fields. For Buchanan and Rawls, there was a fascinating similarity between designing the rules of a game and the social contract tradition in political theory, which also sought to formulate general principles to structure individual decision making.
However, this latter aspect of the thought of Buchanan and Rawls was also inspired by a third source of economic ideas which has attracted less attention than welfare economics and game theory: the economic theory and social philosophy of Frank Knight. Knight was a creative but also elusive thinker, who did not conform to any of the straightforward categories used to classify political thought.Footnote 18 Along with Jacob Viner and Henry Simons, he was a founder of the so-called “Chicago school” of economics, a critic of the New Deal, and the teacher of a generation of economists who would play important roles in the later rise of neoliberalism such as Buchanan, Milton Friedman and George Stigler. Yet Knight was also a staunch critic of a merely economic analysis of human behavior and expressed grave doubts about the long-term viability of a social order grounded on a pure form of capitalism. Indeed, Knight felt an affinity with conservatism, believing that social science had fostered a hubristic belief that “it may be possible to change the character and constitution of society in a way comparable to the modern transformation of technology through science.”Footnote 19
Amid a complex, dizzying body of work Knight made several pioneering points that would be of great significance for subsequent debates in political theory and which help to explain what Buchanan and Rawls had in common intellectually. Both Buchanan and Rawls attested to the importance of Frank Knight in the development of their thinking.Footnote 20 In Buchanan's case, Knight had taught him at Chicago while he was a graduate student after the Second World War and Buchanan paid generous tribute to Knight as his intellectual role model.Footnote 21 Rawls recollected reading Knight in the early 1950s along with other economists such as Samuelson, Hicks, and Morgenstern and von Neumann as he germinated his ideas, a claim that is supported by the archival evidence of his unpublished lectures at Princeton of 1951–2.Footnote 22
Close readers of Rawls are familiar with the distinction between risk and uncertainty made by Knight, which played a significant role in decision theory more generally and in Rawls's formulation of the original position in particular. As Knight put it, risk is when the outcome is unknown but the probabilities of different outcomes occurring are known, whereas uncertainty is a situation in which both the outcome and the probabilities of different outcomes are unknown.Footnote 23 It was this latter condition that Rawls thought applied to the choice of principles of justice in the original position—since it was not possible to quantify the chances of the parties in the original position ending up as the worst off in society they would (Rawls argued) maximize the position of the least advantaged subject to the equal liberty and fair equality of opportunity constraints.Footnote 24 But this was just one well-known example of a larger intellectual debt that Rawls owed to Knight—a debt also owed by Buchanan. First, Knight was a strong critic of the maximization of utility as an adequate characterization of the goals of liberalism. Rather than envisaging the classical liberal tradition as one that regarded “liberty for the individual and mutuality in the relations between individuals as instrumental, i.e., as conditions for maximizing individual want satisfaction,” Knight argued that “liberty and mutuality” were best regarded as “moral ideals” that were ends in themselves which guaranteed freedom as an “inalienable right” and entrenched equality of status in modern commercial societies.Footnote 25
Second, Knight argued that both market economies and democratic polities were vulnerable to the accumulation of inequality over time. Knight made an observation about inequality that was to be very influential for a number of political theorists, notably Rawls and Hayek, but also for Buchanan and numerous other subsequent writers on distributive justice.Footnote 26 Knight argued against desert as a justification for economic inequality on the ground that the “capacity to labor productively” was ethically no different from inherited wealth since it too was a product of unchosen “inheritance, luck, and effort of acquisition.”Footnote 27 Indeed, Knight thought it was simply a matter of luck that some individuals found themselves in a position “to furnish services which are in demand” and it was therefore implausible that such lucky individuals should have “an ethical claim to a superior share of the social dividend, except to the extent that the capacity is itself the product of conscientious effort.”Footnote 28
In both economics and politics, argued Knight, it was in fact “personality power, ability to persuade and to organize,” that was simultaneously the most critical skill for commercial and electoral success and “one of the forms of power most unequally distributed by nature.”Footnote 29 As a result, in both markets and democracies Knight envisaged a cumulative growth of inequality and monopoly, with large corporations driving out competitors, a wealthy elite growing ever more affluent, and demagogues constructing political machines and manipulating public opinion.Footnote 30 Both fields, said Knight, were arenas which were fundamentally about a competitive struggle for power, with participants driven by “the psychological sense of emulation or rivalry,” although he was inclined to see politics as plagued by a more dangerous form of this competition since in business the will to power was somewhat constrained by the concrete material realities of providing a service to customers.Footnote 31
Third, Knight developed a powerful analogy between economics (and later politics) and playing games. Knight alighted on this parallel in the early 1920s, before the later rise of game theory or Wittgensteinian philosophy, both of which are usually regarded as the key sources of interest in games among mid-twentieth-century political philosophers.Footnote 32 As we have seen, Knight saw participation in business and politics as to a large extent motivated by the desire to outdo rivals rather than to satisfy wants or render public service. It therefore occurred to him that these fields could be analyzed as competitive games and subjected to a normative rather than a purely social-scientific assessment according to how “fair” the rules of the game were. Knight argued that games lose their interest and degenerate without a set of fair rules that are rigorously enforced.Footnote 33 The difficulty posed by Knight was how to determine these rules: “the problem of agreeing upon a process and a mechanism for securing agreement on particular issues, i.e. on a ‘constitution.’” This “problem of social self-constitution” required free agreement through open discussion. This ruled out, Knight insisted, both the use of coercive force and of “persuasion” in formulating political norms.Footnote 34
Knight had grave doubts that ordinary citizens were capable of this level of constitutional deliberation without turning to political leaders for guidance. Yet Knight was also pessimistic about the capacity of leaders to rise to the occasion, not least because they would be far from disinterested observers: “advocacy of extensive reform is practically the solicitation of the position of king on the part of the reformer.”Footnote 35 Knight did not offer a worked-out alternative to this possibility, but he thought that the impartial authority of appropriately credentialed intellectual experts, alongside, and analogous with, religious leaders could play a valuable role in designing fair social institutions. Such experts would work together disinterestedly, “absolutely renouncing interest in individual prominence and power,” since Knight distrusted above all else the public intellectuals of his own time who masqueraded as specialists but in fact were engaged “in a competitive solicitation of the ‘ignorant masses’ for support against each other.”Footnote 36 Although Knight never offered a systematic theory of constitutional choice, his interest in devising a mechanism to ensure the fairness of social rules was a resonant one, which both John Rawls and James Buchanan were to take very seriously.
The fourth idea explored by Knight that would play out in the later work of both Rawls and Buchanan was the role of “ideal theory” in the analysis of social institutions.Footnote 37 Knight defended the specification of highly abstract, purposefully unrealistic ideals in political and economic theory. He used Max Weber's notion of “ideal-types” to defend economic theory from the criticism of institutionalist economists that economics wrongly disregarded the role of social context in shaping market behavior. Knight maintained that the value of economic theory was precisely that it abstracted away from the complexities and contingencies of social reality.Footnote 38 As Knight put it, “a ‘science’ of human behavior, to be relevant or practically significant, must describe ideal and not actual behavior, if it is addressed to free human beings expected to change their own behavior voluntarily as a result of the knowledge imparted.”Footnote 39 Curiously, then, Knight presented economic theory as valuable because it embodied a regulative ideal that could be compared with historical reality and subsequently made the target of efforts at social reform, in effect shifting from a Weberian social-scientific notion of idealization to a more explicitly normative one without acknowledging the change of register. Indeed, he favored elaborating a normative set of ethical ideals that were necessarily unrealistic but still offered a guide to action:
Does the judging faculty actually work by reasoning out alternatives and deciding which is preferable, or does it not rather formulate ideals and compare actuality and potentiality with these, and with each other indirectly, by so comparing them with an ideal? No doubt both methods are used, and are useful; but we contend that in regard to the larger and higher questions, the ultimate problems of moral and social life, the formulation of ideals is a necessary step. There is a place, and a vital place, for an “absolute” science of ethics. Its dicta will not be really absolute, for they never cut loose entirely from the real world and its possibilities of growth and transformation, and they will always grow and change. But at least they are not “merely” relative; they must be beyond the immediately attainable, and will often lie in the field of the actually impossible, patterns to be approached rather than objectives to be achieved.Footnote 40
As we will show below, the subsequent work of Rawls and Buchanan can be read as making apparent a tension within Knight's work between his staunch advocacy of “ideal theory” on the one hand and his conservative skepticism about ambitious projects of social reform on the other.
Behind the veil
As Buchanan suggested in his letter to Hayek, Buchanan and Rawls first got to know each other after the publication of Buchanan and Gordon Tullock's seminal work, The Calculus of Consent, in 1962. On Buchanan's account, The Calculus of Consent was drafted between September 1959 and April 1960, with the final version sent to print in February 1961.Footnote 41 In his memoirs, Buchanan recollected that although the book was “essentially worked out independently of the comparable construction of John Rawls,” he and Tullock did read Rawls's 1958 paper “Justice as Fairness” during the book's composition.Footnote 42 After the publication of The Calculus of Consent, Rawls wrote to Buchanan to praise the book and to note that his own nascent work had embarked on a similar approach in seeking to select “a system of basic institutions” rather than “specifying criteria for specific allocations,” and in following “the main idea of the contractarian tradition, i.e. some sort of unanimity rule.”Footnote 43 Shortly thereafter Buchanan and Tullock created the research network that would eventually become the Public Choice Society, which was initially called the Committee on Non-Market Decision-Making and held its inaugural conference in April 1963. Rawls attended the second meeting of the Committee on Non-Market Decision-Making in 1964.Footnote 44 At this meeting, Rawls presented a paper that summarized his nascent theory of justice to an audience that included Buchanan, Tullock, Anthony Downs, Mancur Olson, William Riker, and John Harsanyi.Footnote 45 The first reference to Buchanan in Rawls's published work occurs around this time as well, in his 1963 paper “Constitutional Liberty and the Concept of Justice,” which has a footnote on the similarities between The Calculus of Consent and Rawls's approach, but also indicates that The Calculus of Consent appeared only after the paper was written.Footnote 46
Why did Buchanan and Rawls see themselves as engaged in a similar project? One reason is that both had been quick to appreciate that the game-theory debates about decision making under conditions of risk and uncertainty initiated by von Neumann and Morgenstern were of great potential significance for political theory. The 1940s and early 1950s saw the development of a large economic literature—which ran far beyond the field of game theory—on how individuals should rank alternatives when it was unclear what the outcome of a specific choice would be. A key finding of this literature was that it was possible to use probability theory to translate ordinal utility rankings into a numerical scale that economists could use to calculate which decision between a set of alternatives would maximize an individual's expected utility.Footnote 47 The economist and philosopher John Harsanyi was the first to notice that the technical apparatus of this discussion could be used to answer the dilemma posed by Kenneth Arrow to welfare economics, namely how was it possible to move in a liberal society from diverse individual preferences to an overall agreement on how to advance social welfare? Harsanyi showed that the puzzle could be solved by reinterpreting the ethical choice about the preferred distribution of income in society as a hypothetical decision by individuals about the distributive pattern they would prefer if they did not know what their own position within the social structure would be. From this impartial perspective—in which individuals are assumed to have an equal chance of ending up anywhere in the income distribution (and also of having different “subjective attitudes and tastes”)—Harsanyi argued that the rational choice of distributive principle would be to maximize expected average utility.Footnote 48
Both Buchanan and Rawls took up this intuitive idea of modeling political agreement between diverse individuals. However, as Andrius Gališanka and Katrina Forrester have pointed out with respect to Rawls, it was Frank Knight's earlier preoccupation with games that offered Buchanan and Rawls a distinct path away from strategic interaction towards an attempt to design rules that the players would regard as fair while also granting them space for individual choice and the pursuit of legitimate self-interest.Footnote 49 Both Buchanan and Rawls drew on Knight's analogy with games to work out how to draw up fundamental constitutional principles—as with the formulation of the rules of a fair game, constitutional principles, they said, should be drawn up before particular players can anticipate how certain rules might advantage them.Footnote 50 As we have seen, Knight's work had made some initial steps towards proposing disinterested deliberation over rules (albeit by experts) as a way of generating governing political principles.Footnote 51 Both Buchanan and Rawls drew on contractarian ideas in the history of political thought to elaborate on this insight by introducing a constitutional deliberation or decision to identify fair principles for political organization.
In strict chronological terms, Buchanan and Tullock were the first in print in 1962 with a worked-out version of their theory of constitutional choice. Fundamental to their theory was the stipulation that the rational individual faced with a choice about which constitutional principles to adopt should be “uncertain” about what their “precise role will be in any one of the whole chain of later collective choices that will actually have to be made.” Such an individual would therefore not “have a particular and distinguishable interest separate and apart from his fellows”; such an individual “will not find it advantageous to vote for rules that may promote sectional, class, or group interests” because they cannot predict which side of such rules they will end up on. Individuals should approach constitution making as equals and agree unanimously on the principles that will structure the constitution.Footnote 52 In his copy of The Calculus of Consent Rawls wrote next to this passage: “This is equivalent to the Harsanyi-rule.”Footnote 53
Rawls's original position, although it had been evolving across a number of papers during the 1950s, only reached its settled form in his 1967 paper “Distributive Justice.” His 1958 “Justice as Fairness,” which as we have seen Buchanan and Tullock were aware of while writing The Calculus of Consent, contained the idea of parties reaching agreement on principles to govern society, but in this iteration of the original position the parties still retained knowledge of their current social position. Rawls pictured an already established society in which agreement was sought on the principles by which social complaints might be judged. Risk or even uncertainty entered Rawls's theory at this stage insofar as he stipulated that once such principles had been agreed they would be binding on all future complaints, so he argued that individuals would be unlikely to propose a principle that would advantage themselves in the present because they wouldn't know how it would affect them in the future.Footnote 54 But this was not the full-dress version of the veil of ignorance.
In the aforementioned 1963 paper “Constitutional Liberty and the Concept of Justice,” Rawls considered the possibility of a situation in which existing people may find themselves in a situation in which they bear “equal risks” of a bad outcome—becoming enslaved having been captured during a war between city-states—which is an improvement on being killed upon capture. He suggested that their acceptance of the “real and equal risk” under the right conditions would render the scheme “fair or at least not unfair.”Footnote 55 In his second 1963 paper, “The Sense of Justice,” Rawls set out the original position at the beginning of the paper in a way that mirrored the 1958 paper but then added that “it is assumed that there is an absence of information; in particular, it is assumed that the parties do not know their social position, nor do they know their peculiar talents and abilities.”Footnote 56 The veil of ignorance finally received its name, shape and settled position in the 1967 paper “Distributive Justice.”Footnote 57
It will not come as a surprise to those familiar with these debates that choice under uncertainty seemed to these thinkers to point to specific recommendations, but these turned out to be different for each author. Strictly speaking, Buchanan and Tullock offered a procedural argument that did not seek to prescribe the constitutional principles that would be chosen through such a procedure—rather their concern was to stress the importance of unanimity rather than majority rule or another decision rule as the only unobjectionable means of arriving at such principles. There was therefore a sense in which Buchanan and Tullock had developed a contract theory that was analogous to the Pareto principle: on their account, all citizens had to agree to the constitutional rules, and “consensus” on those rules could be said to prevail when “there exist no particular changes on which all citizens can agree.”Footnote 58 This was ultimately, said Buchanan, a “marginalist” rather than a “total” “interpretation of the contract theory,” which sought to evaluate specific improvements to existing constitutional structures rather than to prescribe the construction of an entirely new polity.Footnote 59
The cost of this approach—as Rawls recognized—was that it privileged existing institutional structures and distributional arrangements rather than questioning their basis. Rawls's notes in the margin of his copy of The Calculus of Consent expressed his worries about the status quo bias inherent in this approach. Next to the passage from Buchanan quoted above about the social contract addressing only those changes on which all citizens can agree, Rawls wrote, “But not satisfactory if take existing distribution and knowledge.”Footnote 60 In The Calculus of Consent, Buchanan and Tullock did give some hints about how they saw their approach to constitution making playing out politically, perhaps revealing how narrowly they interpreted their references to “uncertainty” over one's future position. Since “constitutional prohibitions against many forms of collective intervention in the market economy have been abolished within the last three decades,” the “rational individual” may in fact seek at the level of constitutional choice “the imposition of some additional and renewed restraints on the exercise of such legislative power.”Footnote 61
Tullock gave an early glimpse of their differences with Rawls over their favored distributive pattern in correspondence in 1964, noting that, based on the acceptance in The Calculus of Consent of some form of moderate income insurance, Tullock's favored distribution was less egalitarian than the one Rawls had in mind. Tullock supported “a free market economy with mildly progressive taxes but a floor below which incomes cannot fall except for convicted criminals” rather than maximizing the position of the disadvantaged.Footnote 62
Nonetheless, the similarity of the methodology of the early work of Buchanan and Rawls is striking, especially in light of the differing political orientations they came to represent. Both sought to develop a contractualist method that could generate rules to guide institutional structures rather than individual behavior or policy decisions (though as we will highlight in the next section, Rawls saw his work as primarily offering a theory that investigated judgments about justice rather than offering recommendations for political action).Footnote 63 To accomplish this, the veil of ignorance had a dual role to play (which Harsanyi had already made explicit in the 1950s). The veil had a technical significance—it was seen as necessary to solve an otherwise intractable problem: what would people with diverse preferences (or the possibility of changing them) all find unobjectionable? And it also had an ethical significance in that it modeled the “impartial and impersonal attitude” necessary for non-self-interested decision making.Footnote 64 Rawls and Buchanan, more or less explicitly or happily, followed Harsanyi there.
Even at this earliest stage of their work, normative values played a fundamental role in Rawls's and Buchanan's rational-choice modeling of agreement. To be sure, the normative ideas played different roles. For Buchanan and Tullock, the contracting parties, while assumed capable of something akin to the Rawlsian sense of justice, were meant to consider only their self-interest behind the veil. In addition, Buchanan and Tullock emphasized the technical or “practical” nature of the uncertainty assumption. But in justifying the adoption of the veil, they made a reference to Hayek's explicitly normative defense of the prohibition to model general rules on particular cases.Footnote 65 Moreover, not only did they admit in The Calculus of Consent that the device of choice allowed them to model the normative force of consent but also, as we will see in the next section, Buchanan accepted that the device modeled a certain view of the moral status of individuals. Meanwhile, Rawls explicitly wanted the veil to model “fairness” and “justice” and, in addition, the contracting parties were to be reasonable. But he was also aware of the veil's technical significance in allowing parties to reach an agreement.Footnote 66
What is the point of contractarianism?
In spite of the similarity of the early work of Buchanan and Rawls, after the publication of A Theory of Justice it became clearer that there were important differences between them. This was in part a result of the impact of the politics of the 1960s on their respective political views. For Buchanan, the politics of the 1960s were traumatic—he was deeply disturbed by the radical currents of the decade, by their impact on university campuses, and by the leftward tilt in US social and fiscal policy and the assertive liberalism of the Supreme Court.Footnote 67 The Calculus of Consent was (as Buchanan himself put it) a defense of “a political structure not grossly divergent from that envisioned by the Founding Fathers and embodied in the United States constitution.”Footnote 68 Buchanan's faith in these structures and in the contractarian apparatus of The Calculus of Consent had been diminished in the decade since its publication, and this cast a more negative hue on his reception of A Theory of Justice. In Rawls's case, while his published political views were submerged in his larger theoretical edifice, it seems clear that his assessment of the contentious debates and social change of this period was far more sympathetic than Buchanan's.Footnote 69
These differences first became evident when Buchanan reviewed A Theory of Justice. Buchanan observed that although, on the basis of Rawls's earlier papers, he had “looked forward to publication of Rawls's long-promised treatise,” he in fact found himself “less sympathetic” to Rawls than he thought. He sent a draft of the review to Rawls, noting that “Gordon Tullock says that the review reads as being more critical than I actually am, but there are, I think, significant differences between us.”Footnote 70 Buchanan identified two reasons for his disappointment, and each repays closer scrutiny.
First, Buchanan felt that Rawls's derivation of the difference principle from the original position exceeded the brief of contractarianism as Buchanan understood it. While Buchanan accepted that contractarianism might rule out certain institutions—a “discriminatory” franchise, “restrictions on entry into professions,” indeed any departure from Rawls's first principle of justice—he doubted that it was possible via the contractarian approach to agree that one specific distributional rule was the ideal or just one. Rather, it seemed more plausible that several such rules might qualify as fair in contractarian terms.Footnote 71 As Buchanan observed privately in a later letter to Rutledge Vining, he regarded Rawls's identification of a single just distributional principle in A Theory of Justice as “a genuinely tragic error.”Footnote 72 More broadly, Buchanan felt that contractarianism was fundamentally about the “non-specification of outcome by external observers,” by which he meant the crafting of fair procedural decision rules, rather than specifying the substantive policy decisions or distributional pattern that would emerge within the contract. This was in part because Buchanan understood drawing up a social contract as analogous to determining the rules for economic exchange. As with agreeing on the legal framework for market transactions, Buchanan thought that “the contractarian social philosopher” should seek a rule-governed social order within which diverse individuals could freely agree between themselves on a variety of possible, mutually agreeable outcomes. Interestingly, though, Buchanan also understood such a procedural approach as an implication of what he called “the individualistic or Kantian precept that human beings are to be treated as ends never as means,” and thus as an attempt to treat people as equals.Footnote 73
Rawls offered a characteristically emollient reply to Buchanan. He stressed that he had learned a lot from Buchanan and Tullock (“I fear that the references to it were not ample enough [in A Theory of Justice]”). On this first criticism, he accepted that Buchanan's point about the (unlikely) derivation of the difference principle had “great force” and noted that the principle was itself intended to be open to “considerable latitude of interpretation” when it was deployed at the legislative stage.Footnote 74 Rawls himself later explained in Justice as Fairness that a unique agreement on the difference principle cannot be guaranteed and, in light of the exchange, we can speculate that Buchanan's criticism may have played a role in Rawls's thinking here.Footnote 75 But in another letter to Buchanan, written in 1976 in relation to Buchanan's critique of Robert Nozick, Rawls also emphasized that his contractarianism was only ever to allow for determinate conclusions as a result of a specific move: the principles were always to be selected in explicit comparison with an alternative set of principles.Footnote 76 This did allow Rawls to reach beyond Buchanan's “non-specification of outcome by external observers.”
Nonetheless, Rawls recognized that differences remained with Buchanan. The most significant was that Rawls sought “to find a central place for the value of community” in Kantian idealism and hence “to soften the individualist tendency of contract theory.” Rawls added understatedly: “you may not want to do this.”Footnote 77 Rawls's emphasis on a communitarian aspect of his contractarianism may seem surprising, but as Mackenzie Bok has observed, the concept of community in fact played an important role in Rawls's thought (and represented an element of his moral philosophy that was quite different from Buchanan's). For Rawls, “true community” consisted in a mutual recognition of moral personhood through the acceptance of shared governing principles of justice. This, Rawls thought, was a surer basis for the foundation of justice than rational self-interest. As Bok has shown, this commitment can be traced back through Rawls's engagement with Wittgensteinian ethical naturalism to his early Protestant Christianity.Footnote 78
Buchanan—a strong secularist—was in fact keen to dissociate Rawls from any connection with religious discourse. Buchanan preferred to interpret Rawls's theory “as an extension of the ethics of classical liberalism, as opposed to a modern secularized version of the ethics of Christian charity. The imputation of elements from the second of these ethical structures into the Rawlsian dialogues has been the source of much confusion.”Footnote 79 Buchanan described his own underlying philosophical position as a “normative individualism” which “locates the sources of value exclusively in the consciousness of the individual,” and accordingly sought to debunk the notion that politics should concern itself with concepts such as the common good or the public interest.Footnote 80 This was a position that—in spite of Buchanan's own creative reading of Rawls—was ultimately somewhat different from Rawls's focus on a shared scheme of social cooperation.
The second reason Buchanan gave for his growing doubts about Rawls related to his sense of how his own political position had shifted since writing The Calculus of Consent. Buchanan now worried that idealized views of justice, when invoked in real-world political controversy, could be socially disruptive. Given that it was inevitably interests that would determine even the construction of constitutional rules, Buchanan was concerned that rules that were perceived by some to be “unjust” on idealistic grounds would thereby be subject to destabilizing social conflict, with no clear means of determining precisely “when criteria of ‘justice’ override the stability of law.” Buchanan left no doubt about what he had in mind here when he added, “Would John Rawls allow Earl Warren or his successors to make judgments on their own versions of ‘justice as fairness’ and, in the process, to disregard the embodied predictability of existing constitutional order?” Earl Warren's tenure as chief justice of the Supreme Court was, of course, notable for a number of landmark civil rights rulings. Buchanan glossed his question as indicating that he believed that “misguided idealism, operating in disregard of constitutional precepts,” was “a major source” of what he regarded as the social and legal “chaos” of the 1960s and 1970s.Footnote 81 In the 1960s, Buchanan argued, “the judiciary began to assume the role as guarantors of ‘justice’ in some idealistic sense” and, as a result, bore significant responsibility for the decay that Buchanan detected in the constitutional order of the United States.Footnote 82 The Warren Court, he had earlier remarked, “simply declares to be ‘constitutional’ that which it conceives to be ‘good’ and ‘true,’” thus undermining any attempt to control the disrespect for the rule of law that Buchanan thought was spreading throughout American academia and society.Footnote 83 While clearly sparked by his own political preoccupations, Buchanan's more general suspicion about “idealism” was nonetheless one that would be echoed by the later realist critics of Rawls.Footnote 84
Rawls responded that A Theory of Justice offered a moral theory in the sense that it sought to “account for moral judgments” rather than “explain actions or institutions.” Hence it aimed principally to “account for our judgments of which constitutions are just, and why.” Rawls regretted that Part II of the book might be misleading insofar as it suggested that he aimed to offer a “theory of political action and institutions.”Footnote 85 Should the Supreme Court apply “idealistic” justice directly, then? Politely setting aside the specific point about Earl Warren—“I assume here that Earl Warren is a stand-in and one is considering the general question, leaving aside the merits for this purpose of particular decisions”—Rawls did not agree that justice as fairness would entail a Supreme Court that disregarded precedent and predictability in law. His answer to Buchanan's specific question—about the judiciary being empowered to deploy justice as fairness come what may—was no: ideas of justice would “come into play” in judicial decision making where the law was ill-defined or where “standards of fairness” had to be invoked, not as principles that could dislodge or override precedent.Footnote 86
But behind this specific question lay a difference of emphasis, which drew on divergent aspects of Frank Knight's thought. Whereas Rawls was ultimately committed to the notion of “ideal theory” that Knight had set out—“the function of theory is to formulate and clarify some conceptual pattern of action to the end that men may better conform to it”—Buchanan stressed instead the insight also pressed by Knight that “there are very weighty presumptions in favor of a generally ‘conservative’ position” insofar as radical, idealistic changes to the status quo would be socially destabilizing.Footnote 87 This latter point was the position that Buchanan developed at greater length in his full-scale work of contractarian political theory, The Limits of Liberty (1975). The Limits of Liberty reflected Buchanan's growing social anxieties or, as Brian Barry remarked in a review, it was characterized by a “tone of rancorous detestation for most of what has happened in the USA since 1960.”Footnote 88 The book emerged from a series of workshops that Buchanan had participated in at his Public Choice Center at Virginia Tech in the early 1970s on the relationship between anarchy and social order, a research topic that reflected the worries that Buchanan and his colleagues felt about the apparent decay of social norms during the 1960s.Footnote 89 One of the most forceful features of The Limits of Liberty was Buchanan's cautionary words about idealized models of reform that did not recognize the fragility of the status quo. Invoking Frank Knight, Buchanan set out to diagnose the tensions that had rent the American constitutional contract but to avoid “description of the ‘good society’ independent of either that which exists or the means through which attainment might be secured.”Footnote 90
Buchanan therefore argued that his ideas were “simultaneously more and less ambitious than those of Rawls.” They were more ambitious since he sought to renegotiate the social contract in an existing society in which the parties to the constitutional deliberation were “not equals at the stage of deliberation and … are not artificially made to behave as if they are.” But they were also less ambitious in that Buchanan did not offer any clearly defined principles that he thought would emerge from these contractual deliberations and indeed he repeated that he thought that Rawls's principles would be socially disruptive if they became the basis for concrete institutional proposals “in the pragmatically-oriented arena of day-to-day politics.”Footnote 91 As we have seen, for Rawls such a stance was problematic since it did not address the biases and inequalities of the status quo—something he thought could only be done by a consideration of social ideals. Rawls later emphasized to Buchanan that those who, like Buchanan, started “from the status quo” also needed to work out “a conception of the kind of society we would want gradually to reach … moving from a non-ideal situation to an ideal situation.” He added, “The principles one follows in doing this are extremely rough and it is difficult to make any firm statements. I did not attempt to do so.” But he continued, “The only point I would make here is that a contractarian view should, to be complete, include both cases and say how they are connected.”Footnote 92
Treating people as equals
Buchanan was closely engaged with the controversies about the fiscal consequences of the welfare state in the United States during the 1970s, notably the emergence of the so-called “tax-payers’ revolt” which most famously resulted in California passing Proposition 13 in June 1978. Proposition 13 amended the state constitution to place limits on the level of property taxes and to require legislative supermajorities for any future tax increases.Footnote 93 Buchanan was a firm supporter of such constitutional constraints as a mechanism to rein in levels of tax and public spending.Footnote 94 His writings on political theory also increasingly reflected this preoccupation with the justification of fiscal redistribution, and this led Buchanan to grapple afresh with Rawls's egalitarianism.Footnote 95
In the most emblematic of these writings, Buchanan reinterpreted the difference principle not as “providing a philosophical–ideological basis for egalitarian income–wealth transfers superimposed on a market order,” but as grounded on what Buchanan termed a “Hobbesian” defense, namely that, in the absence of a sufficiently generous social settlement, the society that emerged from the social contract would be vulnerable to disruption by the refusal of the least advantaged to cooperate in production or abide by the law. Buchanan made it clear which sorts of contemporary social disruption he had in mind, citing the “increase in crime” and “industry-wide strikes, notably those in public utility and public service enterprises.” Such “erosion in the structure of legal order” could on this account perhaps be staved off by the application of the difference principle (though Buchanan indicated that he was also tempted by a rival strategy, that of more stringent enforcement of rights to secure social order, since he thought it was possible that current levels of social disorder simply reflected “a loss of political will” on the part of the authorities as opposed to Rawls's diagnosis that “the precepts of a just society are not present”).Footnote 96
Rawls wrote an appreciative letter to Buchanan about the draft of this article that Buchanan had sent him, suggesting that, while it was not precisely what Rawls had in mind, Buchanan's argument came close to the considerations of stability that Rawls saw as entering his theory at the second stage of the justification of his two principles. Rawls offered his own version of Buchanan's point (and here we get a rare insight into Rawls's own political views):
I guess I do think that the social product is very much a joint product; and that while the poorer individuals & groups don't threaten to withdraw co-operation as a tactic, as a rational strategy, so to speak, they do in fact withdraw their co-operation in other ways: by crime, by self-destructive & compulsive behavior and ways of life (e.g. by drug addiction); by acquiring far less constructive capacities for working successfully & co-operatively; by just not developing themselves from loss of self-respect (I am convinced that much low IQ scores is the result of large and disorganized families, etc). So while all this is not done as a threat, it perhaps has much the same destructive effect. The upper classes greatly miscalculate its consequences; though I believe they should take it into account for other reasons as well.Footnote 97
Rawls's analysis of this social injustice was therefore some distance from Buchanan's ultimate diagnosis of decaying state authority, and he was careful to reserve his position with respect to the ultimately normative justification for a more egalitarian society. Indeed, in light of the critiques of Rawls that followed, it is notable that he was clear (in 1975) that he did not see the United States as a fair system of cooperation.Footnote 98
Nonetheless, it was also evident that the differences between Rawls and Buchanan on redistribution were more complicated than a standard conceptual contrast between egalitarian liberalism and neoliberalism might suggest. Although reluctant to endorse extensive redistribution as a necessary outcome of contractual deliberations, Buchanan accepted that it could certainly be one of a range of legitimate agreements made at the contractual stage. More fundamentally, in the late 1970s and early 1980s Buchanan drew on Knight's point that political and economic games must seem fair to their participants to be sustainable to argue that there should be a substantial attempt to equalize the endowments of market participants before competition begins. High inheritance taxes and government investment in education were necessary, Buchanan maintained, in order to give the market system a basic structural fairness. Libertarians such as Robert Nozick or Murray Rothbard who did not accept the need for “some rough equality in premarket positions” and later attempts to “knock off the edges of post-market extremes,” said Buchanan, were in fact weakening their case for the market.Footnote 99 “A libertarian position,” Buchanan argued, “is not inconsistent with an egalitarian one.”Footnote 100 Buchanan interpreted Rawls precisely as offering an argument for equalizing pre-market endowments along these lines, in a way that Buchanan thought was reminiscent of another early Chicago economist, Henry Simons.Footnote 101
But there was—as Rawls was aware—a significant problem with how Buchanan conceptualized pre-market endowments. Like Rawls, Buchanan accepted Knight's intuition that inequality in productive ability was a form of unchosen advantage analogous to inherited wealth. But unlike Rawls, Buchanan did not think that inequalities that arose from differences in productive ability should be the subject of inequality-reducing policies. This was a curious and under-argued distinction for Buchanan to make. As Andrew Lister has observed, the reason why Knight's equation of social and natural inequality was adopted by figures on the neoliberal right such as Hayek was precisely to deploy it as a reductio ad absurdum. Since it was manifestly ridiculous to equalize natural endowments, argued Hayek, it was hard to justify equalizing social class differences without also being logically committed to this unappealing option.Footnote 102 Rawls saw the force of this point—he was willing to bite the bullet and accept that social justice should address both sources of inequality.Footnote 103 Buchanan's position, in contrast to both Hayek's and Rawls's, was that it was justifiable to treat differently the disequalizing forces of initial class position and natural talent.
In his notes on Buchanan's writing on this topic, Rawls found it hard to discern why Buchanan thought this. Buchanan had stated that it would be impossible in the social “game” for every player to have equal capacities and instead recommended that social arrangements should be so structured as to advance what was in effect Rawls's notion of “fair equality of opportunity.” Buchanan regarded it as unfair “perverse handicapping” that equals in productive capacity might start in different social positions and so some would first have to catch up with the others to obtain the equal treatment they were due. Rawls noted, “could it be crucial here, though unstated, that the perverse handicapping of the market can be corrected; genetic draw cannot? But that would be a bad arg.”Footnote 104 But this was indeed what Buchanan thought. As he baldly stated elsewhere, differences in productive capacities “cannot be equalized by a reassignment of nonhuman assets or claims.”Footnote 105
Conclusion
The value of this comparison of Rawls and Buchanan is in part that it gives us a different perspective on Rawls's relationship with the emergent neoliberal challenge. As Barbara Fried observed, and we reported above, the polarity between neoliberalism and egalitarian liberalism had in effect not yet been historically constructed when Rawls wrote A Theory of Justice. As Fried also argued, the tendency to retrospectively present that polarity as a contest between Rawls and Nozick is misleading, since Nozick's hard line on the state's distributive role is not in fact the position taken by most economic libertarians, who on the contrary accommodated some social welfare provision into their theories.Footnote 106 We can see this quite clearly in the case of Buchanan, who wrote a review of Anarchy, State and Utopia in which he ruled out Nozick's rights-based approach as a blind alley for liberals and defended instead a contractarianism in which rights are derived from contractual agreement, a move that permitted space for agreement on redistribution. “Could we have reached agreement?” Buchanan argued, was a more productive and soluble question to pose than “did this violate someone's natural rights?”Footnote 107 To that extent, Rawls and Buchanan were on the same side against Nozick. The theoretical debate therefore becomes a more subtle one about precisely which forms of redistribution are justifiable rather than whether it is in principle permissible (and also becomes a debate about how one might use a contractarian methodology to resolve that question).
But it is also of historical interest to observe that Rawlsian liberalism and neoliberalism initially emerged as entwined critiques of mid-twentieth-century political economy. Although both intellectual currents have been criticized as unwelcome attempts to extend economic reasoning into the domain of the political, the point of the early writings of figures such as Buchanan and Rawls (and, for that matter, Hayek) was to register dissatisfaction with the aggregating, utilitarian logic of the field of economics and to press for an alternative approach that focused on assessing the broad social structures and rules that governed human conduct rather than individual policy decisions.Footnote 108
This shared appetite for rules over governmental discretion initially brought Rawls to the margins of a neoliberal movement slowly growing in intellectual coherence and confidence. However, as the politics of the 1960s and 1970s unfolded, clear differences began to emerge. In comparison with Buchanan, the political change of that period—the civil rights movement, the Vietnam War, welfare-state expansion—evoked little public comment from Rawls. But his political theory proved to be fertile ground for a new moral idealism that sought to advance civil and social rights.
Buchanan spoke for many on the neoliberal right when he bemoaned the apparent decay of state authority and moral discipline revealed by this more assertive politics of individual rights and liberties. His commitment to a realist analysis of politics led him to a rather conservative response to the novel political alignments of the late twentieth century, many of which precisely revolved around emancipating individuals from unjust constraints on their liberty. Paradoxically, it was Rawls's ideal theory that proved to be a more robust vessel for the transmission of a new social emphasis on personal autonomy into political philosophy.
Acknowledgments
We are grateful to Max Shock, Thijs Kleinpaste and Robert Cheah for research assistance; to Elizabeth Beckman for help with navigating the Buchanan Archive; to the Leverhulme Trust for funding Ben Jackson's research at the Hoover Institution Archives; and for comments to Teresa Bejan, Simon Caney, Cécile Laborde, Daniel Luban, David Miller, Sophie Smith, Annette Zimmermann, the editors and reviewers of Modern Intellectual History, and the participants at the Historical Rawls conference, Oxford, May 2017. We owe particular thanks to Alan Hamlin for his detailed written comments on an earlier draft.
