While it is generally true that economists “failed to foresee” the recent near-collapse of the world’s financial system, the contention involves so many assumptions and qualifications that its point easily gets lost in claim and counter-claim. It is not so much that economists “did not see the crash coming”: instead, the real problem is that the crash was the direct outcome of policies aligned with the way in which, since the 1970s, most economists had come to think about competition and regulation. And the way they had come to think was a direct consequence of what they had been taught as students. Economists are therefore part of the problems that became obvious in 2008, and so not obviously part of any solution.
This is not a story that sits easily with the view of economics promoted by the annual Riksbank Prize in Economic Sciences, which was awarded in mid-October 2014 to Jean Tirole for his work on regulation and competitive processes. Tirole has been a leading figure in the study of industrial organisation since the 1980s. This post does not seek to advance arguments, positive or negative, about the merits of this work. Rather, I want to draw attention to the arguments advanced for the award. In the press release accompanying the announcement of the 2014 award, the Committee stated that, up until the 1980s, “research into regulation was relatively sparse”, and that “traditional economic theory” does not deal with markets “dominated by a few firms that all influence prices” – the case of oligopoly.
There are two problems with this account of the development of economic thinking about markets, competition and industry structure. The first is that it is quite wrong about the history of economic analysis. The analysis of oligopolistic competition was well-advanced by the 1930s, but was sidelined in the 1950s by an approach to competition and markets that saw no need for anything other than the concepts of perfect competition and monopoly in market analysis. This approach has proved remarkably resilient as an ideology of market functioning, exemplified in a recent article by Peter Thiel in the Wall Street Journal (“Competition is for losers”, 12 September 2014). Moreover, the Committee awarded the 1982 Riksbank Prize to George Stigler for his vigorous promotion of exactly this ideology, which they now treat as the “traditional” approach.
Secondly, the issues of asymmetric information arising in the regulation of oligopolies and natural monopolies, issues to which Tirole has addressed much of his work, have proved remarkably resistant to the application of contemporary economic theory. It can be plausibly argued that all of the major British utility privatisations involved at least one major structural defect, and different in each case. Rail privatisation failed so catastrophically that the government had to dissolve the first industry model and allow a new one to develop. Many of the mistakes made in designing a new “competitive” structure arose from the contractual relations established between its various elements; exactly the area that Tirole’s work is supposed to have addressed. And as translator of Tirole et al. Balancing the Banks (2010) I could extend this argument to the area of financial regulation, but that is not my purpose here.
The problem is rather the way in which economists think of competition, and the way in which this is transmitted into a public realm. There were really no formal models of competitive processes until the later nineteenth century, when Cournot’s treatment of duopoly began to have some impact. In 1838 Cournot had imagined a situation where two producers of bottled water shared the market; he examined the way in which two producers would come to share this market. This was in a context where industry structure was not seriously considered to be an aspect of price formation outside of the idea that “competition” was preferable to “monopoly”. In 1912 Pigou introduced the idea of “monopolistic competition” (Wealth and Welfare Bk. II Ch. X), but it was primarily in the United States during the 1920s that ideas about price and competition in different industry structures developed. This was for the very good reason that, since the 1890s, there had been vigorous public debate around the “Trust Question” and the impact of Supreme Court rulings on cases related to the Sherman Act of 1890. A significant literature on railway rates and regulation also developed at this time. (Nobody involved in the privatisation of British Railways during the 1990s seems to have been acquainted with the vast literature on railway administration and regulation).
“Perfect competition” was first named as such in the context of an “imaginary society” and itemised in Frank Knight’s Risk, Uncertainty and Profit (1921, pp. 76ff.). In 1927 Knight published a translation of Max Weber’s Allgemeine Wirtschaftsgeschichte, and in the mid-1930s ran a Chicago seminar on Weber’s methodology attended, among others, by Milton Friedman and Edward Shils. It is not therefore a stretch to suggest that this itemisation of “perfect competition” was intended as a Weberian ideal type – a set of characteristics employed to organise thinking about social phenomena, not isomorphic with them, and certainly not a specification of “ideal” competitive conditions.
In the later 1920s Harold Hotelling developed ideas of competitive conditions that were suggestive of the kind of suboptimal conditions that “free competition” generates, such as a reduced range of choice and convergence on a (possibly rather average) standard product. At the same time, Edward Chamberlin completed a Harvard PhD thesis that began from Cournot’s duopoly model, eventually published in 1933 as The Theory of Monopolistic Competition. In fact, this was not as the title suggests an elaboration of Pigou, but an account of the persistence of oligopolistic conditions in modern markets. Chamberlin reduced perfect competition to one essential characteristic: that the perfectly-competitive firm was a price-taker. A monopoly was by contrast a price-maker. He then placed these two market characteristics at the poles of a continuum, and proposed that most market behaviour could be characterised as an attempt to secure monopoly conditions. Since it is very hard to envisage a perfect monopoly – there are generally near-substitutes at different price ranges, and various factors conspire to undermine monopoly powers – efforts by firms to secure control of markets were constantly challenged and undermined.
Nonetheless, the important message here was that “perfect competition” had very limited use as a conceptualisation of market conditions, for this was a condition firms actively sought to evade or escape. In so doing, market behaviour became oligopolistic, firms paying more attention to the behaviour of their competitors than to that of their customers. From this brief sketch it should be clear that Chamberlin opened up a conception of a market as a set of negotiations between agents seeking to create or retain autonomy of decision. The game theoretic implications of this are today obvious, underlined the following year by the publication of Heinrich von Stackelberg’s Marktform und Gleichgewicht (1934), which not only argued that competitive markets lacked a mechanism imposing a unique equilibrium of price or quantity, but that the more “perfect” the market, the more unstable it became.
That is a conclusion that resonates today; for deregulation and privatisation have been aimed at the creation of “ever more perfect” markets, presuming that in so doing the engine of economic growth would be fuelled. But as we have seen, Stackelberg was on to something: the freer the market, the greater the instability. The era of sustained stable postwar growth had many sources, but liberal economic policy comes a very long way down the list. The millennial fantasy that well-informed economic policy could put an end to the “cycle of boom and bust” that had developed in the 1970s only lasted as long as the current boom that was supposed to be the new normal. So what happened to the kind of thinking about markets and competition developing in the 1930s?
My attempt to answer the question can be accessed here: EUI-93. The development of neoclassical economics in the later 1930s, and its consolidation in the later 1940s and early 1950s, meant that the attention of economists shifted from real markets and industries to the properties of competitive models. Oligopolistic market models are relatively intractable in this framework, since price formation depends to a great extent on a range of factors such as the nature of the product or its raw materials, industry structure, geography, and income distribution, among others. In this situation, one appealing property of perfect competition is that it requires no knowledge of any of this. If the firm is a price taker, decisions made by the firm and industry structure are irrelevant. Strictly speaking, there can be no neoclassical theory of the firm, since the decisions of the firm merely reflect market conditions. Correspondingly, in a neoclassical world deviations from perfect competition are represented by “imperfect competition”, a category representing only the presence of features that prevent the realisation of perfect competition.
None of this would matter that much if it were made clear to students that perfect competition is a way of beginning to think about competition, prices, and markets that will subsequently be left a long way behind; that it is a logical step in their training, not an ideal condition. Further, that markets are characterised by agencies actively seeking to control them; so attempting to analyse markets in terms of perfect competition makes no sense at all. To understand market processes, one needs to command a great deal more than an understanding of market models. Alfred Marshall conceived the training of economists as a process in which students assembled a mental “toolbox” from which they could take instruments for the analysis of economic problems and the formation of policy. He implied that judgement was needed, both in the choice of tools and in their application; that, therefore, the training of economists did not involve the simple inculcation of technique, but the cultivation of powers of judgement. “Prediction” is not really so important here; rather, an informed view on likely outcomes in given situations. And it would help if a healthy regard for the unpredictability and instability of markets ranked rather higher in such thinking.