8th April 2013

Are markets matching Callon and Roth?

The last meeting of our “Copenhagen market group”[i] was devoted to an increasingly influential stream within current economics, namely “market design”. The discussion left me with the somehow perplexing puzzle I am trying to unfold in this post: isn’t this type of economics almost too close to the ‘markets as calculative collective devices’[ii] approach developed by Michel Callon and colleagues so influential among us -non-economists market researchers- in the last years?

During the meeting we discussed two articles (here and here) written by the 2012 Nobel Prize winner Alvin Roth. As Roth explains, see also his very clear Nobel Prize speech, his and his colleagues’ work has been dedicated to very practical problems. A favourite situation is this: on the one hand, medical students looking for the best hospital to do their specialism, on the other hand, hospitals wanting to hire the best young doctors. Problem: hospital officers were pushing so much that they were selecting students years before they finished their undergraduate studies and even started thinking about their possible specialism. The problem was solved with the introduction of a centralized clearing mechanism equipped with an algorithm- that is a chain of formalized decisions- which paired young doctors’ and hospitals’ lists of priorities. The problem started again some years later when couples (yes couples) were introduced in the equation. As Roth (humorously?) explains:

In a simple match, without couples or other complications, all of the applicants would have preferred the applicant proposing match, and no applicants who were matched or unmatched at the outcome of one algorithm would change employment status at the outcome of the other (Roth: 1360)

Roth has not only worked pairing medical students and hospitals, but also children and schools, and even kidney donors and receivers. These are the types of challenges tackled by market designers. Equipped with the tools provided by game theory (where the key name seems to be Roth’s co-Nobel winner Lloyd Shapley) and experimental economics, economists are working to find the best “matching” mechanisms to solve these types of dilemmas.

This is, no doubt, a quite particular type of economics. Resembling some ideas of an earlier Nobel Prize winner, Herbert Simon, market designers see economics as a “science of the artificial” (also this), that is closer to engineering, design, or medical knowledge than natural sciences. Here markets are no more a quasi natural phenomenon in search of its mathematical laws, but rather practical situations that need to be carefully designed and, Hayek’s heroes, prices, are left to play  a somewhat secondary role. Actually: market design seems to work in both traditional auctions and also in moneyless exchange, such as the already mentioned kidney donation, school choice (or who knows, tables and guests at a wedding party). Of course, we could ask here, what we have internally called the “Frankel question”: can these situations be called markets at all? Are not these matching problems instead a more general case of operational research or even central planning? As far as I can see, the difference seems to be that unlike matching containers and ships or other forms of planning, market designers focus their attention on situations where both sides of the equation are choosing-maximizing agents (but not necessarily sellers and buyers), that, accordingly, can be modelled with the tools provided by game theory.

In other words, in this increasingly influential branch within economics, markets are not natural but artificial encounters, they are not necessarily about price but about calculation, economists are not external observers, but active agents in making these carefully designed encounters possible, and these encounters need, now and then, to be reframed or re-designed. Doesn’t this sound closer to something I have heard before? It does. Actually, Roth himself has seen the resemblance, and, in a recent post on his quite active blog, he says:

Yesterday’s prize to Paul Milgrom for his work in market design (among other things) brings to mind a curious critique (and criticism) of economics in the economic sociology literature, namely that economics is “performative”, in the sense that economic theories influence the real economy to become more like economic theory […] As far as I am aware, the term originated in linguistics to distinguish those cases in which saying is also doing. Thus saying “it will rain tomorrow” is not performative, but saying “I apologize” is: when you say it, you have done it, saying it makes it happen. So the basic idea applied to economics is that e.g. creating an option pricing formula might change the way options are priced. Designing a kidney exchange might change the number of patients who get kidney transplants. Or in Paul’s case, designing auctions might change the way the FCC sells radio spectrum licenses.

So far so good. There are, though, a couple of misunderstandings. Roth continues:

The criticism, such as it is, seems to take two forms. The first is that, since economics is performative, it isn’t a ‘real’ science which describes things as they are. The second, often more between the lines, is that this is just part of the way that economics has been sucking the meaning out of life ever since the invention of agriculture and trade […] Of course, that economics is performative is a criticism that economists, especially market designers, might take as a compliment. (It’s a little like criticizing body builders for working hard to have big muscles, and not just settling for the ones they could get without cheating by exercising).

Anyone who has spent some time reading Callon and colleagues knows this is not the case. To say that economics is performative is not a criticism of economics not being a real science, but, if any, a call to assume and not deny the practical situated character of economic knowledge. And about sucking the meaning out of life…This is more what people say about Callon et al. than what Callon et al. say about the poor economists. But, even if we leave these misunderstandings behind, some questions remain.

Are the new2 economic sociology and market design the same? Avoiding the obvious methodological splits that separate a highly formalized and a rather descriptive-reflexive ethnographic approach, there are still important conceptual differences. The ideal situation for Roth’s designers seems to be “give me some choosing things[iii] to match and I will rise a technologically equipped market”, while for Callon – especially in his work connecting his thoughts on technical democracy, hybrid forums and markets- the ideal situation is where what is traded, who can participate in the exchange, and who and what is equipping the market encounter are collectively and heterogeneously defined. Civilizing (Callon) and engineering (Roth) markets are therefore two different programs of market design. More practically, for instance, in school choice, for Roth et al. what a good school is or who can choose or what is chosen while matching school places is defined before the market. In Callon et al.’s view such an arrangement would not only match pre-calculating families and schools but it would include the consequence of making the involved agents calculative, changing accordingly the way they understand and deal with education. In other words, in Callon’s view markets are never only about matching, or matching would need to be understood also as a mode of per-formatting new agencies and things.

But, despite these differences, it seems like finally engineer economists and engineer sociologists are finding a common starting point. Isn’t that scary? I don’t think so. This is a much better place to start and try a dialogue that is not so limited by pre-existing disciplinary boundaries (see also Callon here). Let’s agree: markets are not pre-social metaphysical forces that need to be left alone, but they are practical arrangements that can be more or less, better or worse, designed[iv]. In those cases where there is an already functioning market or quasi-market mechanisms (for instance: school choice or carbon trade) let’s try to make them work the best we can. In other words, social researchers should not only criticize marketization but also spend time, energy and knowledge on engineering and/or civilizing these complex arrangements. This is, I think, a nice pragmatic starting point for market researchers at large.

But is this just good?  No it isn’t, there is also a serious flaw. These two streams of market research seem to share a somewhat excessive optimism about markets as devices that can solve social and environmental issues. As a product of neoliberal Chile, I would happily pay for not having to make choices in areas like health insurance, pension funds, schools or long distance phone carriers. And certainly many people have argued that these and other sectors (have you heard about trains in the UK?) are not necessarily working better years after features such as competition, choice and providers able to select or exclude their potential users have been introduced. Market design risks becoming the face of the latest round of social and environmental reforms (for instance: emissions trading or the announced Job Match interface in the UK). And the new reformers seem to believe something like: it is not that markets were necessarily a bad social policy but that they were not properly designed. But, shouldn’t we also be experimenting with other ways of doing things? I am not saying that markets are always bad, but that the same brilliant ideas currently oriented at designing better markets could also be spent devising other forms of solving our common problems. In my opinion market civilizers and engineers will become fully respectable technicians the day they are also able to advise something like: “thank you for contacting me, but here you don’t need a market”.

[i] The group is more formally called “Thought in progress group: organizing and valuing markets” and its meetings have had the more or less consistent participation of Christian Frankel, Susse Georg, Peter Karnøe, Trine Pallesen, Ann-Christina Petersen Lange, Rasmus Ploug Jenle, and Satu Reijonen.

[ii] Or, to use a more recent formulation, the markets as ‘socio-technical agencement’ program.

[iii] What would be the Latin term for “things that choose” instead of “res cogitans”?

[iv] For a much less optimistic view see Mirowski & Nik-Khah here, who, polemically suggest that Actor-Network theory and Market design share a common ancestry in Operational Research. In an article written in a less confrontational tone, Mirowski explains that the movement toward engineering markets is a much broader turn within economic thought of the last couple of decades. In his words: “One of the great challenges for intellectual historians of the future will be to explain how it came to be that a professional academic orthodoxy that had eschewed most considerations of the specificity of markets […] then neatly negotiated a 180º turn, and managed to convince a broad array of outsiders that they possessed special expertise to construct all manner of actual usable markets, tailor-made for their narrowly specified purposes” (Mirowski, 218).


As part of our inter-network collaboration this entry is co-posted with estudios de la economia. Image by inhabitat, used under a Creative Commons license.