Jump to ContentJump to Main Navigation
The Knowledge We Have Lost in InformationThe History of Information in Modern Economics$

Philip Mirowski and Edward Nik-Khah

Print publication date: 2017

Print ISBN-13: 9780190270056

Published to Oxford Scholarship Online: June 2017

DOI: 10.1093/acprof:oso/9780190270056.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 21 February 2020

The Standard Narrative and the Bigger Picture

The Standard Narrative and the Bigger Picture

(p.31) 2 The Standard Narrative and the Bigger Picture
The Knowledge We Have Lost in Information

Philip Mirowski

Edward Nik-Khah

Oxford University Press

Abstract and Keywords

Many believe there has been an “information revolution” in economics, but almost no one has taken the effort to explain its timing and content. Information was added as an American Economic Association subject category in 1976; it is therefore almost exclusively a postwar phenomenon. This chapter explores the conventional narratives that are often broached in orthodox economics concerning the rise of information in economics, and discovers that generalist intellectual historians have not much improved upon these urban legends. Epistemic asymmetry between economist and agent is often elided. By contrast, we stress questions touching on the epistemology and ontology of information.

Keywords:   agent ignorance, epistemic universalism, knowledge vs. information, genealogies of information, economist ignorance

Economies are built out of information. This has been true from the Stone Age to our knowledge economy today.

Eric Beinhocker1

As explained in the previous chapter, we start from the premise that economists possess no special expertise in the mysteries of human nature, or at least no more than the man on the street. Rather, the prime directive of professional economists has been the attainment of scienctific status for the field; and what this has meant after World War II is that they must participate in the most important trend in the modern sciences—namely, that they have elevated information to the pinnacle of their theoretical structures. Hence much of what has been misunderstood or misconceived as controversies about “rationality” have been, in fact, struggles over how to insinuate “information” into the canonical model of microeconomics, which dated back to the 1870s.

The fact that “information” has transformed the natural sciences from the 1940s onward is a proposition now widely taken for granted among scientists and historians. Some of the more philosophically inclined, dissatisfied with a basic trinity of matter, energy, and information, go so far as to attempt a further reduction (p.32) of all three into the One True Entity.2 We need not defer to its enthusiasts to that extent in the current context; it is enough to paraphrase the popular historian James Gleick, who said that after the 1940s, information was the axis around which the entire world began to spin.3 While we will only touch upon a few small facets of that massive transformation of the natural sciences, the timing of its appearance is of paramount importance. That information revolution preceded the watershed event that many media theorists take for granted—namely, the appearance of the Internet and everything now associated with it. The key period turns out to have been the years surrounding World War II, and it is from there our own narrative will soon begin.

Over and above the issue of timing, it is indispensable to realize that recourse to “information” in neoclassical economics could never have been a matter of effortless appropriation. Inevitably, it could never have been simple larceny, if only because there was no simple, unique thing to appropriate. Information had proved that it “can add colors to the chameleon, / Change shapes with Proteus for advantages, / And set the murderous Machiavel to school.”4 This multiplicity was only one of the many ways the history starts to look different, once we situate the advent of “information” at its center. One rather obvious effect is that we begin to glimpse that “rationality” remains a rather empty concept, at least until one begins to have some idea what a theorist means when she suggests an agent “knows” something. Another eye-opener comes in noticing that the referent of information in economics changes appreciably over time, largely in reaction to changes in the cultural themes swirling around theories of information and knowledge. Perhaps the most bracing shock consequent upon shifting the frame is a dredging to awareness of the gradual transition from economists’ regarding information as an unalloyed good to praising ignorance as the appropriate state of a dedicated market participant. In what brave (p.33) new world can the economist Gary Gorton unself-consciously write: “We show that preserving symmetric ignorance in liquidity provision is welfare maximizing and strictly dominate symmetric and even perfect information.”5 Or, to paraphrase, blessed are the meek and stupid, for they shall promote liquidity and stability.

Never before have the high priests of efficiency been so happy to proclaim that a sucker is born every minute. Economists nowadays hold it as gospel, but must not let on that is so. One major motivation for writing our book is to provide a meditation upon why and how such an incongruous trajectory has pirouetted to meet its antithesis. Why must an economics of information end up propounding a doctrine that economic agents should be ignorant? While the causes are many and the detours abundant, there is one profound philosophical question that needs to be broached before we set out on our narrative.

There is a bit of commonplace wisdom found in the community of generalist intellectual historians that the era of the 1970s–1990s was one of contextualization and historicization of ideas, as well as the exploration of various formats of “identity politics.” David Hollinger, an intellectual historian and a major proponent of this view, summarizes the trends in a half-facetious way as “Kuhn, antiracism, feminism and Foucault.” Whatever one might think about such broad-brush generalizations, he also explicitly insists that the economics profession was uniquely immune from the entire trend:

One could write an exceedingly detailed history of this entire episode without mentioning a single economist. This exclusion applies even to those economists who sometimes write about large and hard-to-solve problems for transdisciplinary audiences, such as Robert Heilbroner, Albert Hirschman or Amartya Sen… . Epistemic universalism was never seriously challenged.6

(p.34) One might be inclined to agree that most economists never actually gave the feminists or Foucault the time of day (although it appears many were assigned Kuhn some time in their student years, only to cement their belief that economics had attained “normal science” status), but it is a serious misunderstanding of the historical record to accord the economics discipline immunity from cultural trends in epistemology. The reason Hollinger and his colleagues have missed this connection is that economic arguments about identity and truth were played out in a space somewhat removed from the generalist magazines and newspapers that often provide such historians with their major sources of evidence. We shall demonstrate in this volume that economists were also swept up with popular concerns that truth may appear to be contingent and dependent upon personal identity during the same period, but that those disputes were carried out primarily in technical discussions of mathematical models in specialized journals; furthermore, the issues of identity that drew the most attention were not those of race and gender but, rather, the minatory perseverance of an epistemic divergence between the professional economist and the quotidian agent, with lesser (but not unsubstantial) attention paid to epistemic divergences between agents.7

The disparagement of the sagacity of Homo economicus has direct bearing on the asymmetry between agent and analyst. There might be a sense in which the legitimacy of the modern economics profession almost demands that their representative rational agent needs to be rendered less knowledgeable (or, to court offense, more stupid) than the orthodox economist himself. To allow the agent to know substantially more than the economist is transparently a nonstarter, since in that instance the pretense to special economic expertise comes to naught. One might then suspect that the sweet spot is for the agent to know exactly everything the analytical economist knows—one observes this in modern assertions that the (p.35) agent “knows the true model of the world”—which conveniently turns out to be identical to the orthodox neoclassical model—but that option also has its own severe drawbacks. These include the overwhelming evidence of empirical violation of rational choice models; but more to the point, there are the compromised ambitions of the economist to provide privileged normative prescriptions for the economy. If everyone else happens to be situated on the very same epistemic footing, then the man in the street really has no reason whatsoever to attend to economists and their prognostications; and, quite starkly, economists have nothing relevant to impart to them.8 Can the student appreciate that, for neoclassical economists to hold their heads high in society, and believe they can model information and its consequences, it is imperative their imagined representative agent be stupider than the median economist?

The repercussions might even be momentous. What happens when the person on the street comes face to face with this kind of scorn (if that ever comes to pass)? Will the contempt of the economist be met with a different sort of disdain by the public?

What Everyone Knows

Once upon a time—say, around the era of David Ricardo and Karl Marx—political economy was primarily concerned with the production of national physical wealth. This “classical” notion tended to hang on long into the twentieth century, well after the invention of neoclassical economics in the 1870s. Nevertheless, there was no denying that within neoclassical economics, notions of exchange had displaced those of tangible production as the primary topic of interest; this informed the definition of economics articulated by Lionel Robbins that its proper subject was the (p.36) “allocation of scarce means among given ends.” But subsequently, something rather extraordinary happened around the middle of the twentieth century, gaining momentum as the century waned. More and more, economics at the cutting edge (as opposed to the textbooks) became relatively cavalier about treating trade as static allocation, and instead became all wrapped up in the image of The Market (or else the agent) as a processor of information or knowledge.

Appeals to information and knowledge pop up almost everywhere these days: the efficient markets hypothesis in finance, common knowledge in game theory, rational expectations in macroeconomics, asymmetric information in principal/agent theory, adverse selection in mechanism design, focal points in behavioral economics, and so on. For many economists, these elaborate constructions of the market consequences of information are the main reason the modern economic orthodoxy is always an improvement on your father’s economics.9 We provide the following paean to modern success, lifted from a recent blog:

It is a strange fact that many social scientists feel economics to some extent stopped progressing by the 1970s. All the important basic results were, in some sense, known. How untrue this is! Imagine labor without search models, trade without monopolistic competitive equilibria, IO or monetary policy without mechanism design, finance without formal models of price discovery and equilibrium noise trading: all would be impossible given the tools we had in 1970. The explanations that preceded modern game theoretic and information-laden explanations are quite extraordinary: Marshall observed that managers have interests different from owners, yet nonetheless are “well-behaved” in running firms in a way acceptable to the owner. His explanation was to credit British upbringing and (p.37) morals! As Stiglitz notes, this is not an explanation we would accept today.10

People with advanced training in economics seem to think they know what this transformation has been about and why it has happened. One indicator might be that “information” was added as a subject category in the American Economic Association subject taxonomy in 1976; this in itself suggests a postwar heritage. As Kenneth Arrow once put it: “one of the biggest differences between 1950 and 2000 is the much greater role now given to … knowledge and information… . Nobody would have denied the importance of these in 1950, but the tools to handle them did not formally exist then.”11 But, in fact, one could just as easily, and with justice, assert the opposite: certainly we possess many more “tools,” but the impression that clarification followed formalization needs to be approached with a modicum of salt. Interestingly, in the interview, Arrow did not bother to identify even one specific hallmark doctrine within the orthodoxy that owed its existence to the stipulated sea change. Further, Arrow hedged his bets by citing “knowledge and information,” and this begins to reveal the profound ambivalence and unapologetic befuddlement that often besets anyone attempting to make sense of the history of this literature. Rather than concede the distinction between knowledge and information as meaningful or significant for economics, Arrow rather haughtily responded to a request to clarify the distinction with the dismissal, “I am afraid the topic does not inspire me.”12 The Muse of Information moves in mysterious ways for neoclassical economists.

Those fortunate enough to have experienced Philosophy 101 will come away wary of people who conflate knowledge and information with abandon, much less defer to those who treat knowledge as if it were something that needs no definition, since everyone who thinks he or she has it, does indeed have it. Information enthusiasts (p.38) outside of economics manage to signal their awareness of the importance of some crude distinctions: “Information can be moved around easily in the products that contain it … but knowledge and know-how are trapped in the bodies of people and networks that these people form.”13 It seems fairly commonplace to concede that knowledge is something not readily reified, and yet many economists still seem to resist this insight. Indeed, a historical wariness toward this tendency is one of the sentiments we hoped to evoke indirectly by means of this book’s title. In this regard, the oft-neglected Kenneth Boulding was a far better guide to these issues than Kenneth Arrow:

Knowledge, however, has a dimension which goes beyond that of mere information or improbability. This is a dimension of significance which is very hard to reduce to quantitative form. Two knowledge structures might be equally improbable but one might be much more significant than the other.14

Boulding, somehow, did not make it into the modern economist’s Hall of Fame.

The prudent intellectual historian cannot help but be struck by the fact that when some judicious souls began to write surveys of this newfound enthusiasm among economists for information in the 1980s and 1990s onward, they came away bearing impressions of something akin to the Tower of Babel.15 One such attempt at a history published in 1995 began by quoting George Akerlof on the initial rejection of his famous “Lemons” paper: “They were afraid if ‘information’ was brought into economics, it would lose all rigor, since in that case almost anything could be said—there being so many ways that information can affect an equilibrium.” The authors then concluded their survey with: “we encounter an (p.39) over-abundance of results and/or equilibria; almost anything can happen.”16 “Anything goes,” on its face, would hardly seem the prescription for a successful intellectual tradition.

One could confront the impressions of disarray in the historical archive in many different ways, but our task here is to set out to tame the blooming buzzing confusion by getting a clearer idea of what the economists did and did not accomplish. First off, we need to establish that there really was something new under the sun in economics back in the twentieth century. As Dan Schiller perceptively queried over a quarter century ago: “Why wasn’t the status of information a major topic in economic theory in 1700, 1800, or 1900? Why was it only in the postwar period that the economic role and value of information took on such palpable importance?”17 The easy retort—that it was absent-mindedly “overlooked,” exiled to peripheral vision until WWII—simply will not wash. Neither will the bad habit of searching for precursor statements from previous eras, anticipations which were never really there. Rather, we shall insist that there were specific historical conditions in the 1940s that stood as necessary prerequisites for the seeming inevitability of the juggernaut of the economics of knowledge. But the era from the 1940s to 1970s is a blind spot for the modern economist—at least when it comes to the economics of information.

What manner of impressions about the history of the economics of information does the modern student receive from her theory teachers, even if only as capricious stage-setting preliminaries? More often than not, there she encounters some assertion that nothing really happened before the 1960s or 1970s. Some may cite George Stigler’s (1961) paper as kicking off the literature, or else that of William Vickrey (1961), and perhaps the aforementioned 1970 “Lemons” paper by Akerlof (1970). What is noticeable about these supposed landmarks is that, by and large, their actual models turn out to be unrelated to pretty much anything that follows (p.40) in the curriculum our student undertakes. For instance, Stigler’s sequential search for lower market price among dispersed differentiated purveyors has been pretty much forgotten. Instead, the class in the economics of information mostly assumes one of two formats: either a meandering survey of picky small variations on the benchmark von Neumann–Morgenstern individual expected utility model to capture something called “uncertainty”; or if the class is skewed toward Nash game theory, then the topics tend to revolve around how agents seek to mislead, dominate, and otherwise confuse their opponents, with topics called “asymmetric information,” “cheap talk,” “adverse selection,” “private values,” “signaling,” “moral hazard,” “common knowledge,” and the like. The macroeconomists are partial to the former tweaks to decision theory, whereas microeconomists—and particularly market designers—favor the latter pastime of Liar’s Poker. Over time, it seems the game theory advocates may have come to numerically dominate the decision theorists, and as that happened, their own favored potted history has grown more austere and sparse, ignoring the vast mass of information theorizing of that era in favor of their one purportedly canonical general model of information, the Bayes-Nash model first propounded by John Harsanyi in 1967. In both cases, the contested notion of “rationality” keeps getting amended willy-nilly to supposedly handle information, while past approaches get erased. There is a third genealogy of the economics of information that tends to be favored in business school classes outside of economics departments proper, especially in finance, which is instead structured around the efficient markets hypothesis: in that potted history, it is The Market itself that stands as a superb processor of information in its own right. There, cognition and rationality are dispensed with altogether.

First, even at this most superficial level, we can observe that there has been no single canonical model of the economics of (p.41) information at almost every step along the way. The student haphazardly gets arbitrarily schooled in one manifestation or another, depending on the angle of approach to their own specialization. Hence, the conventional notions of the history of information in economics tend to be a little incoherent. Furthermore, the potted orthodox histories turn out to be singularly uninformative about the deeper issues driving this need to privilege “information,” or perhaps “knowledge,” in economic theory over time. For instance, why would someone think the unique best way to approach the question of knowledge about the economy was through an inductive definition of “uncertainty”? The words sound like they have a family resemblance, but it doesn’t take a philosophical genius to realize they are hardly identical. As Thomas Schelling once observed back in 1962, “There is a tendency in our [theories of] planning to confuse the unfamiliar with the improbable” (p. vii). One wag later said that you can always tell a decision theorist because of his struggles with the undecidable: he seemingly can’t tell the difference between irrationality and ignorance.

Second, existing orthodox histories of economics do a poor job of explaining what effects the economics of information have wrought. Because we do not automatically share the orthodox obsession with “progress,” it will be crucial to establish as historians whether or not this nouveau fascination with knowledge actually has thoroughly reconstructed the heartland of orthodoxy, or instead, as so often happens, merely clutters up the periphery with sprawl. Here again, there has been no general consensus on this issue among the orthodox. At one end of the continuum, Joseph Stiglitz has hinted from time to time that the previous world of neoclassical economics was turned upside down by the economics of information:

There is no single new Law of Economics… . The world is not convex; the behavior of the economy cannot be described as (p.42) if it were solving a (simple) maximization problem; the law of supply and demand has been repealed.18

We know of hardly any orthodox economists today who would subscribe to any one of those precepts. Conversely, at the other reach of the continuum, various respected economists have hastened to reassure the rest of us that nothing has really changed after all, and that, contrary to most impressions, the ongoing info-fascination was just a minor variation on the age-old wisdom of neoclassical economics.19

Both narratives turn out to be equally implausible; and this ushers us toward the nub of the present philosophical conundrum. How can it be that so many economists have come to believe there has been some sort of Great Transformation of Economics into a Science of Knowledge in the last half-century—the dawn of a higher wisdom buttressing an information economy—and yet be utterly incapable of producing even a sparse, clean consensus on the hallmark doctrines of the New Order?20 Why are they so convinced that “information” constitutes the panacea for so many of their problems? Could it be due to the internal logic of their economic model, or has the pressure come primarily from outside the discipline?

Risking ridicule, or maybe just a standard paradox of self-reference, in this book we seek to pose the question: How do these economists know that knowledge has become ineluctably central to their discipline? This paradox will lead us to an even thornier question: How is it possible that a neoclassical economic theory, committed to a thoroughly ahistorical, noncontextual theory of equilibrium (and notoriously weak on how that equilibrium is attained), could provide an adequate account of the process by which knowledge is gained, interpreted, and understood? Can the dynamic character of (p.43) knowledge and interpretation be anything more than a square peg forced into the round hole of given preferences?

Of course, we come equipped with no magic Philosopher’s Stone to settle these thorny questions once and for all; nor would we ever want to. But we do think we can do a better job in organizing the history of economic thought around information, as a prelude to a coherent answer. We will therefore avoid the standard practice of structuring narratives around all those various “applied” areas of economics (e.g., macroeconomics, finance, decision theory, signaling, and adverse selection); rather, we divide the story into the different ways that information and knowledge themselves have been conceived and elaborated though time. In other words, rather than dutifully following the economists in their arbitrary sequences of applied models, we will adopt a more epistemic and ontological approach. The questions that will serve as our protocols will be:

  • Who or what is supposedly doing the thinking?

  • Is there any real cognition going on?

  • What, precisely, qualifies as information and where did the model inspiration come from?

  • Have informational considerations been effortlessly grafted onto the prior neoclassical model of fixed utility functions and fixed endowments, or has the bequeathed canonical model become compromised over time?

These questions will lead us to argue that the older, potted, orthodox histories of the economics of information have been misleading at best, and at worst have been utterly confused in some of their more commonplace variants. Our alternate narrative should help the student understand how the profession has arrived at this curious impasse, where ignorance of their own history has (p.44) permitted economists to believe the most extraordinary things about knowledge and their place in it, without any serious consensus among their peers.

And more to the point, fortified with this history, the student will be much better prepared to recognize the possibility of a New Economics, if and when she encounters it. She may even experience the joy of learning something.


(1.) Eric Beinhocker, INET, Oxford University, blurb for Hidalgo 2015.

(2.) See the authors in Davies and Gregersen 2011 for instance, or Lloyd 2006.

(3.) See Gleick 2011; Agar 2012, chap. 16; and Kline 2015.

(4.) William Shakespeare, Henry VI, Part Three, Act III, Scene ii, lines 191–193.

(5.) Dang et al. 2012. This is doubly ironic, given that Gorton was partially responsible for the promulgation of the financial instruments that blew up AIG in the financial crisis. See Mirowski 2013, pp. 209–210.

(6.) See Hollinger 1997, pp. 345–346; emphasis added.

(7.) This trend was first described in the underappreciated work of Esther-Mirjam Sent (1998).

(8.) This problem of reflexivity began to surface in the 1950s, in a number of social sciences, about the possibility of “self-fulfilling prophecies.” On this, see Hands 1990. Interestingly, Norbert Wiener refused to comment on this literature when asked (Kline 2015, p. 135).

(9.) This ignores a rather sad reverse phenomenon, that once some novel intellectual trend becomes apparent in modern economics, someone somewhere attempts to read its content back into the classical economists, usually Adam Smith, seeking to insist that economists really knew about it all along. For an example explicitly attempting this retrospective whitewash with regard to “information,” see Prendergast 2007.

(11.) The Arrow quote is taken from Colander et al. 2004, p. 292.

(12.) See Daniel Klein 2005, p. 105. The next sentence reads: “In my old-fashioned positivism, concepts have meaning only in the context of a model (which may be very general), and I can’t think of one which will accommodate this distinction.”

(13.) Hidalgo 2015, p. 79.

(14.) Boulding 1968, p. 142. Of course, other economists insisted upon similar distinctions; one such theorist, George Shackle, was also banished from orthodoxy (Basili and Zappia 2009).

(15.) We would point here to Babe 1994; Machlup and Mansfield 1983; Lamberton 1998; Levine and Lippman 1995. Things have not much improved in the interim, judging by Leppälä 2015.

(17.) Schiller 1988, p. 32.

(18.) Stiglitz 1985, p. 22.

(20.) Twenty-five years on, David Kreps’s (1990, p. 578n) warning is still good advice: “The terms of information economics, such as moral hazard, adverse (p.246) selection, hidden action, hidden information, signaling, screening and so on are used somewhat differently by different authors, so you must keep your eyes open when you see any of these terms in a book or an article… . As a consumer of the literature, you should pay less attention to these labels and more to the ‘rules of the game’—who knows what when, who does what when.” The only codicil one might add is to replace “consumer of the literature” with “epistemically challenged member of the economics community,” which better captures the repressed paradox.