Imagine a theory that is self-realizing: a theory that applies to the world as a result of having been applied to the world, intervening in a way that changes reality in its image. At first sight, such a thought seems rather fanciful. Would such a theory not be true by default? Would it not be as much about itself as about the world? Clearly, the spectre of Berkleyan idealism beckons, with the mind-independence of the “things” in the world being called into question.

But then, as social ontologists and social epistemologists keep reminding us, social reality cannot easily be conceived of as mind-independent. It is precisely the reflexivity inherent in social objects, and in our efforts to research such objects while forming part of them in one way or another, which appears to distinguish a theory of market behaviour from a theory of high-energy physics.

The question posed by the title of the present volume – Do economists make markets? – seems thus to a large extent rhetorical. Of course they do, although not in every instance, and never exclusively so. Economists have long recognized the reflexivities that reside in any attempt to predict, model, and behaviourally frame economic phenomena (e.g. Grunberg and Modigliani Reference Grunberg and Modigliani1954; Muth Reference Muth1961; Gauci and Baumgartner Reference Gauci and Baumgartner1992; Frank, Gilovich, and Regan Reference Frank, Gilovich and Regan1996; Arce Reference Arce2007; cf. Davis and Klaes Reference Davis and Klaes2003). Economic game theory in particular has been a forum for some of the most subtle discussions in the social sciences on the properties of self-realizing theory (Nash Reference Nash1950; Bicchieri Reference Bicchieri1993, 127–75).

How then should we appraise the assertion put forward in Do economists make markets? (DMM), that in economics we can find episodes where the use of an options pricing model “brought about a state of affairs of which it was a good empirical description” (MacKenzie in DMM, p. 66), where the organization of an auctions market “may be treated as a practical realization of the model of pure competition” (Garcia-Parpet in DMM, p. 20), where experimental economists conspire “in the making of homo economicus” in the wild when they participate in the design of real-world markets (Guala in DMM, p. 153), where statistics helps articulate collected data “in a peculiar way, and through that articulation, which is nothing but a description, a new property of the whole [e.g. the county average of a particular harvest, MK] comes into being” (Didier in DMM, p. 293), where therefore economics “participates in the actualization of a world in which it becomes or is true” (Callon in DMM, p. 337)?

Let us begin such appraisal by following some of the analyses in DMM a bit more closely. Peter Holm's chapter recounts how a particular image of the market, informed by neoliberal doctrines of marketization, becomes realized in the design of fishery quota trading schemes. Assume your aim is to cap the total annual amount of fish caught, and you seek to implement this cap not through externally imposed quotas on individual fisheries but through a market mechanism that is left to determine those quotas endogenously. Basic economics, rehashed in fisheries management research papers, will tell you that you need to set up private property rights in tradeable quotas. Based on the initial allocation of those quotas, each fishery would then decide whether to make money by catching fish or by selling their quota to a competitor. If the market mechanism works – “and why wouldn't it?” (Holm, not without sarcasm, in DMM, p. 236) – quotas will be redistributed from less to more efficient fishermen to the point where no further mutually advantageous trades are possible and hence, an optimal allocation of individual quotas has been reached. As a result, the traditional fisheries sector will end up transformed in the image of economic theory.

Holm spells out the disruptive effects of this transformation in great detail: a whole way of life, grown over centuries and at the heart of local communities, gives way to a brave, albeit somewhat stereotypically portrayed, new world of profit-maximizing entrepreneurs. He also lists the considerable mobilization of resources necessary to bring this reification of the market about, which includes a politically charged process of implementing the new regime in the place of what came before.

Overall though, one gets the impression that little more is involved, methodologically, than what one would expect in the context of a civil engineering project. Engineers too transform the world in their image. This points us to a curiosity lurking beneath the surface of this volume, that also extends to its companion volume (Callon, Millo, and Muniesa Reference Callon, Millo and Muniesa2007), which both trace their intellectual ancestry to Callon's (Reference Callon1998) Laws of the markets. Why is it that scholars who not so long ago led the field of science and technology studies have now turned their attention to finance and economics, forming a curious alliance with economic sociologists, and calling for a new anthropology or micro-sociology of markets?

The chapter by Philip Mirowski and Edward Nik-Kah offers a tentative answer, by suggesting important affinities between the new “social studies of markets” literature and neoclassical economics. Commenting in particular on the actor–network theory branches of this literature, they note a disdain for traditional sociological explanation (in the tradition of for example Granovetter's (Reference Granovetter1985) “embeddedness” thesis) that is reminiscent of the status of such explanation in economics. More importantly, they detect a refusal to critically engage with the content of economic theories, which are instead accepted at face value.

This is a remarkable shift away from conventional disciplinary attitudes found among sociologists and anthropologists vis-à-vis economics. It is a shift however that is fully consonant both with Callon's original call to relinquish any sharp distinction between “economics” and “the economy” in favour of studying the co-evolution of both from a non-evaluative anthropological perspective, and with the Edinburgh school's long-standing commitment to the symmetry principle in the social study of (any) science, including now economics (Bloor Reference Bloor1976: 7).

Mirowski and Nik-Kah complement their critique of the conceptual apparatus of Callon cum suis with an empirical critique of one of the key case studies of the field, that of the US Federal Communications Commission spectrum auctions. Since Francesco Guala, whose work has in many ways been the original source to inspire this exemplar, offers a further, and more nuanced, discussion of these auctions in the present volume, a brief look at the broader historical context of the spectrum auctions will be useful to help elucidate what is at stake (cf. Campbell and Klaes Reference Campbell and Klaes2005).

About a century ago, an influential view prevailed that broadcasting was a natural monopoly due to the problems caused by potential interference between stations seeking to broadcast at the same wavelength (Coase Reference Coase1959: 6). As a result, the nascent US broadcasting industry was subject to strict state regulation of the use of the frequency spectrum, under powers which, since 1934, came to reside with the Federal Communications Commission (FCC). The FCC exercised these powers through the granting and revoking of broadcasting licences, guided as it were by considerations of the public interest. In practice, this meant that the Commission, inadvertently or not, exercised considerable influence on broadcasting content. Ronald Coase, of later Chicago school fame, found himself troubled by this. An internationally renowned expert on the telecommunications industry at the time, he had closely studied broadcasting regulation in the UK, where the British Broadcasting Corporation, as a state monopoly, refused to broadcast Winston Churchill's pre-war arguments against appeasement, out of considerations of the public interest (Coase Reference Coase1950).

Coase was convinced that the allocation of broadcasting licences be better left to the market, by for example auctioning them off. In the course of preparing this argument, he ended up suggesting that it would in fact not matter how licences were distributed initially since market forces would subsequently reallocate them to their highest valued uses regardless. While this suggestion became later codified as the “Coase theorem” and gave rise to considerable controversy, what caused even greater controversy at the time was Coase's suggestion to allocate broadcasting licences via the market in the first instance, in particular since he advocated his policy proposals in a variety of highly visible consultancy and policy arenas such as the RAND corporation, and in testimony to Congress. Nevertheless, it took more than three decades for these views to gain sufficient influence for them to be put into practice, with the first (non-broadcasting) spectrum auctions finally allowed to be held in 1994 (see Hazlett Reference Hazlett1998).

Have economists made the first spectrum market? In an important sense they have, as contributors to debates that helped bring about these auctions, and, perhaps more importantly, as educators who over the decades shaped the intellectual toolkit of anybody laying claim to market expertise in those debates. These avenues of intellectual influence, traditionally the domain of historians of ideas and social historians, are relatively well understood. They contrast sharply with the quite different roles of economists investigated in the more provocative chapters of DDM, where those economists help realize the spectrum auctions through direct input in their design that ensures their implementation reflects, and thereby confirms, key aspects of auction theory.

Whereas economists in the former role may be regarded as intellectual ambassadors or “preachers” (Stigler Reference Stigler1980), they become engineers in the latter role. The self-realizing potential of economic preaching is borne out by Timothy Mitchell's chapter in DMM, where he follows the ascendancy of Hernando de Soto and his economic prescriptions for developing economies. De Soto has been advocating a programme of propertization of assets hitherto left without the boundaries of the working of the economic system, not unlike Holm's fisheries reform, and indeed Coase's spectrum licensing reform. In the words of the author, de Soto has thereby given decisive shape “to the work of sociotechnical mechanisms that reorganize how people live, the political claims they can make, and the assets they can control” (Mitchell in DMM, p. 248). In preaching, the economist acts as an invited or uninvited adviser to policy makers, who in turn may decide to listen to the advice given, to go on and implement it or even hire the economist into roles with direct policy input. Note that implementation, in this context, will ultimately mean translating policy proposals into statutory and administrative legal code.

With the economist as engineer, things change in a rather subtle way: the code in question is now no longer legal but algorithmic in nature (Lessig Reference Lessig1999). As Guala recounts in his chapter, the last decade of the 20th century saw a branch of economics develop with the ambition to finally establish economics as a “proper” science. Received wisdom had more or less accepted that economics would never be able to reach the methodological rigour of the natural sciences due to the difficulties involved in setting up controlled experiments in its domain of investigation. Experimental economists sought to overcome this handicap. But in the course of implementing ever-changing experimental designs, they found themselves designing and implementing market “machines” (see also Guala Reference Guala2001), thereby quickly moving from the mere testing of theory to literally “realizing” such theory: those market devices were in fact built to embody the theories and models that formed the basis of the market mechanism they were meant to implement. It was “by designing and implementing an adequate mechanism that the engineer ensures that rational choice models can work” (Guala in DMM, p. 149).

Guala is adamant to point out in his chapter that this focus on the economist as engineer should not be regarded as the whole story. The role of the organizations who hire market engineers may be as decisive, and the processes of implementation are subject to considerable political pressures and accommodations on the part of all relevant stakeholders involved. The economist-engineers play thus at most one role among a whole range of competing influences. In their empirical criticism of the use of the spectrum auctions as an exemplar for the self-realizing potential of economics, Mirowski and Nik-Kah press precisely this point: how decisive has the role of economists and their theories in fact been in the final auction design?

Mirowski and Nik-Kah's discussion is not the only critical voice in this volume. In fact, for an edited volume, one finds a refreshing density of critically engaging cross-commentary between individual chapters. By and large, this feature succeeds in turning the attention of the reader to the finer points of disagreement between the various authors, in particular when it comes to assessing the potentially self-realizing dimensions of economics. Take what may well be the paradigm case: Donald MacKenzie's analysis of the role of the Black–Scholes options pricing model in actual options trading (cf. MacKenzie Reference MacKenzie2006). Here we have a mathematical model about options pricing, published in 1973, which allows the calculation of options prices on the basis of estimates of future volatility. We also have, in 1973, the first organized options exchange opening in Chicago. As it happens, some traders begin to inform their trading behaviour by the model. MacKenzie finds that as this practice becomes widespread, traders who disagree with the model price get priced out of the market. As a result, model prices become actual market prices.

This situation obtained on options markets until the 1987 stock market crash, which some commentators have attributed at least in part to the uncritical adoption of the options pricing model by major market players to steer automatic trades. The crash seems to have undermined trust in the theory, with the apparent by-product that its theoretical prices no longer tally up with market prices. In short, what MacKenzie presents is some of the best evidence we have got for the self-realizing potential of economic theory as a phenomenon that relies on certain benign conditions being met for the self-realization to take hold. Clearly, option pricing theory served, for a period, as the official notion of market rationality and thereby became a self-realizing strategy in the trading game. If, in a social dilemma game, you know that your opponent will play the Nash strategy, you have little choice but play that strategy too, thereby helping the Nash account of the game turn into a true account of that game.

Nevertheless, as Emmanuel Didier rightly points out in his chapter, it may require some stretch of imagination to think of this as an instance of self-realization in the same vein as the performative speech act of uttering “I promise . . .”, which brings the promise into being. Publishing a model does not, as such, make the model true of the world, for example. Surely, Didier (in DMM, p. 302) asserts, economists cannot simply create facts out of thin air.

But MacKenzie (in DMM, p. 80) does have a point when he urges us to consider whether “the widespread use of [a] theory or model will make the world it posits more or less real”. This brings us to the real puzzle raised by the present volume. For here, in a nutshell, we are faced with the central proposition of a long-standing tradition in social ontology and epistemology according to which theoretical entities exhibit self-realizing potential in the sense that collective belief in a theoretical entity “realises” that entity.

Consider the following scenario. A branch of science has successfully been operating on the premise that its domain of investigation can be described in terms of an ontology of basic entities h and some fundamental interaction between the h. At some stage though, a series of experiments appears to establish that the h are considerably more diverse in nature than hitherto thought. Instead of just two or three different kinds of h, everything seems to indicate that there are in fact hundreds of different h. Along comes a group of theorists who offer a plausible theoretical model that reduces this complexity to a new ontology of basic entities q. Compared to the previous ontology though, it turns out much more difficult to find any observational evidence for the new entities. In fact, these entities do not seem to be observable at all in isolation. And yet, the new theory is simple and helps accounting for a vast range of phenomena in a highly general way. It becomes in fact the new commonly accepted way of looking at the domain of investigation in question, to the point where school children are taught that the world is made of q, where students of science are trained in understanding and applying the various models of q, and where scientists and engineers manage to successfully manipulate certain phenomena on the basis of their understanding of the various properties ascribed to q.

We could now go on seeking to establish the implications of describing and explaining the world in terms of h rather than q. Maybe the h ontology is less powerful than the q ontology in guiding our manipulations of the world we live in, or maybe the two ontologies are incompatible with each other in certain respects and could not therefore both concurrently be deemed true of that world. Be this as it may, it seems that in an important sense, the q are brought in existence by scientists talking about q in an affirmative way, and linking them with what we know about the world when we manipulate it so that q-talk becomes part of such manipulation, all in the absence of any separately observable q (cf. Pickering Reference Pickering1984, in particular pp. 180–95).

This is of course the account of the Edinburgh school in the sociology of scientific knowledge (SSK) of how communities of scientists and those sustaining them contribute to the realisation of theoretical entities, models, theories, and research traditions (Bloor Reference Bloor, Niiniluoto, Sintonen and Woleński2004). Does q-talk realise the q out of ‘thin air’ though in this account? Nothing could be further from the proposition that is on the table. Only under a set of benign conditions may q-talk become self-realizing, and these demand congruence between Searle's world of “brute facts”, the biological and psychological make-up of individual human beings, and the communities of cognition that authorize particular accounts of the world as real at any given moment in time. Should these communities of cognition root for an image of the world of brute fact that lacks such basic congruence, chances are that it will prove unsustainable.

The provocative point advanced by the Edinburgh school is that the only sound scientific realism is social constructivist in the self-realizing sense described above, in physic and elsewhere. Certain conditions (“benign”) need to be met for q(uark)-talk to be self-realizing. In the abstract scenarios of game theory, one might readily identify them formally. It takes Pickering a dense book to spell them out historiographically for a particular episode in the history of high-energy physics. But the claim as such is a general one, to hold for any scientific claim to knowledge. Other conditions, different from those that helped realize q-talk, need to be met for h(adron)-talk to be self-realizing. High-energy physic moved from h to q, and Pickering demonstrates that this occurred solely as a result of a shift in language games and measurement technology. Metaphysically therefore, one can be committed to a non-social substrate of entities and processes which, jointly with the benign (social and biological) conditions, stabilize q- and h-talk, while at the same time being committed to the q and h as social objects.

The puzzle implicit in DMM is therefore the following: Why go hunting down, in an SSK fashion as this volume by and large does, episodes in economic history broadly conceived, that are particularly striking in their demonstration of the self-realizing potential of economics if the current state of theoretical debate and empirical research in SSK leaves little doubt that this is how science operates in general, in each and every successful invocation of a theoretical concept? Part of the answer may be that the self-realizing processes of the conventional view, of theoretical entities as mind-independent objects, are still, and in particular in economic philosophy, entrenched enough for the question of whether economists make markets to come across as counter-intuitive at first sight, and in this respect the present volume is thought-provoking. But with the envisaged appointment of one of the authors of Nudge, perhaps the most enthusiastic embrace to date of economic micro-engineering of individual choice, to head the US Office of Information and Regulatory Affairs of the new Obama administration, this volume is very timely, too.