From Rational Choice to Reflexivity: Learning from Sen, Keynes, Hayek, Soros, and most of all, from Darwin

Download full paper

This paper is closed for comments.

Abstract

This paper identifies the major failings of mainstream economics and the rational choice theory it relies upon. These failures were identified by the four figures mentioned in the title: economics treats agents as rational fools; by the time the long run equilibrium arrives, we are all dead; the social, political and economic institutions that meet most urgent human needs most effectively could not have been the result of rational choice, but their “spontaneous order” needs to be explained; human uncertainty and reflexivity prohibit a predictively useful rational choice approach to human affairs, and even limit its role in institution design. What unifies the perspectives of all four of these critics of neoclassical economics, however, is their implicit reliance or on need for a Darwinian perspective on human affairs.

Posted for comments on 3 Jul 2013, 10:29 am.
Published

Comments (2)

  • Constantinos Repapis says:

    In this paper Professor Rosenberg produces a wide ranging narrative that contrasts the view of human society and its institutions as the outcome of deliberate rational human action with the alternative view eloquently summarised by Hayek, as “spontaneous order”, or order emanating from human action but not of human forethought. The nodes that form the structure and substance of this narrative are the four names of the title: Sen, Keynes, Hayek, and Soros. What unites these nodes, as a leitmotif, is an articulation of their theoretical positions on social order as manifestations of “Darwinian processes”. By Darwinian processes, Professor Rosenberg means the proliferation of conventions and institutions that have attributes favourable for society, even though these institutions or conventions are not the outcome of deliberate rational design. He sees the contributions of these authors within this structure as cumulative, with Soros at the end of this process. Then he contrasts Soros’ analytical framework with rational choice theory, and discusses how these two theories explain movements of the stock market. But this précis does not fully represent the true scope of the paper, as a number of other fields (e.g. functionalism) are discussed as we progress between the different authors and their contributions. Professor Rosenberg’s clear familiarity with these topics and authors is evident in the dense language he uses to codify their contributions and place them in his narrative. The, at times, fast-paced use and definition of key concepts, makes it difficult for the more general reader, like myself, to follow parts of the argument and there is the constant fear from the perspective of the reader, that he is not doing full justice to the arguments of the author. As a consequence, and primarily because of my limited knowledge of some of these authors and their work, I will confine this review on two aspects of Hayek’s conception of spontaneous order in relation to this paper that I found particularly interesting.

    The first point relates to this antithesis between rationalism and spontaneous order. Hayek was in no way a believer that human action was not guided by rational principles. However he realised that an individual who is rational also knows the natural limits of his deductive powers. The individual then realises that he occupies a place in relation to institutions that antedate him, and will most likely, continue to exist after him. This conception is distinct to that of the individual as a calculating rational machine whose powers and abilities know no bounds. These two distinct types of rational human action are discussed in the following quote:

    “While the generic utilitarianism of Hume thus rests on a recognition of the limitations of our reason and expects its fullest use from a strict obedience of abstract rules, the constructivist particularist utilitarianism rests on the belief that reason is capable of directly manipulating all the details of a complex society” (Hayek, 1964, 88)

    This second form of rationalism, Hayek argues, leads to a logical conclusion that is as erroneous as it is dangerous. This is that

    “rationalism in this sense is the doctrine which assumes that all institutions which benefit humanity have in the past and ought in the future to be invented in clear awareness of the desirable effects they produce;…[rationalism implies] that we have it in our power so to shape our institutions that of all possible sets of results that which we prefer to all others will be realized” (Hayek, 1964, 85).

    And he adds

    “It is from this kind of social rationalism or constructivism that all modern socialism, planning and totalitarianism derives” (Hayek, 1964, 85).

    Therefore, the second ‘constructivist’ tradition in rationalism is a much broader movement than contemporary microeconomic rational choice theory. But, and I agree with Professor Rosenberg, microeconomic theory is one of its offshoots. However, this opens up an interesting question that logically follows from the above argument, but remains unexplored in Rosenberg’s paper. If, modern rational choice theory is part of this constructivist rationalist agenda, then those institutions that it creates -mostly markets these days in developing countries or new types of product markets in the developed world- are also logically constructed institutions and not outcomes of some Darwinian evolutionary process. I.e. these new markets are not manifestations of ‘spontaneous order’. Are we to see them as dysfunctional social constructs that go against the grain disrupting the natural processes in society? This would have been an interesting thread to follow in the last part of his paper where issues of uncertainty and reflexivity are discussed. It would also be interesting as it would approach Hayek’s work in this topic from a novel direction, one in which the market as a mechanism of social organisation is not on the defensive as it was in the 60’s and 70’s.

    A related second point has to do with money. Professor Rosenberg notes in several occasions that rational choice theory cannot explain the existence of money. There have been several scholars who have noted that economic analysis cannot easily meaningfully integrate money in an Arrow-Debreu general equilibrium model. Instead of following this line of argument, for which there is an extensive literature, it may be better to attack the problem head on and consider what is money, and what kind of institutions we have or we would like to have in relation to money. Hayek has written extensively on this topic. He defined money as a quality that a good has if it is to be used as a medium of exchange. Things may be more or less accepted as a medium of exchange and this would make them more or less be seen as ‘money’ (see Hayek, 1978, 160-163). This provides a clear, precise, logical, and ultimately a-historical and socially decontextualized definition of money. But, Hayek continues, money is not a ‘tool’ that has been invented and given to us by an original inventor (the state in many such narratives), its existence is the outcome of “spontaneous generation of such undesigned institutions by a process of social evolution” (Hayek, 1978, 145). Money is the prime example of this process with law, language and morals being “other main instances” (Hayek, 1978, 145). He then continues to argue against the current status quo of money being supplied by the government as sole legal tender, and instead favours private issuers of competing currencies to take over, creating effectively a competitive market of currencies in a country. People would reward the good issuers by trusting them with their savings and flee the bad ones that over-extend lending (see Hayek, 1978). The details of this scheme need not concern us here. What I find interesting is how Hayek proceeds from the definition of money, through its inception as an example of “spontaneous order”, to suggesting a revolutionary new institutional framework that would logically be a vast improvement to the type of monetary institutions that currently operate, and operated through much of recent history. This leap into an unknown new world supported only by deductive arguments on the logical and expected consequences of this framework on society is an interesting change if we superimpose this analysis to that of the previous paragraph. The deeper issue here is that reality is defined both by processes of selection and institutional survival, as well as new conceptions that occasionally are manifestations of abstract deductive thought. It is, to me, unclear why some such constructs are more successful than others, but it is an important question. It is equally unclear why some failed institutions survive over long periods due to inertia or an inability of society to replace them with better ones. I would be very interested in reading Professor Rosenberg’s ideas on this point, which I think is related to the central tenor of his argument.

    References:

    Hayek, F.A. (1964), “Kinds of Rationalism”, published in Hayek, 1967, Studies in Philosophy, Politics and Economics, London: Routledge.

    Hayek, F.A. (1978), “The Denationalization of Money”, in Hayek, 1999, Good Money Part II. London: Routledge

  • Dorian Jullien says:

    It is fair to say that this paper by Alex Rosenberg is in continuity with his well-known and long lasting criticism of the folk psychological underpinnings of rational choice theory. In a nutshell, the folk psychological underpinnings of rational choice theory, i.e., that individuals undertake intentional actions on the basis of beliefs and desires about consequences, prevent economics from being truly ‘scientific’, i.e., to explain and/or predict economic phenomena.

    Here, Rosenberg brings the ‘big guns’ – Sen, Keynes, Hayek, Soros and Darwin – to substantiate and strengthen his criticisms. In line with the previous comment, I will focus on one of the authors discussed by Rosenberg, Amartya Sen, in order to do full justice to the latter with respect to the arguments of the paper.

    More precisely, I will:
    contextualize the rational fool argument within Sen’s broader intellectual project (I);
    make a comment on the connection that Rosenberg draws between Sen and some work in psychology (II);
    compare Sen’s discussion of the Darwinian view of progress with Rosenberg’s Hayekian discussion of the market price system (III); and
    discuss an eventual complementarity between the way Sen uses the term “reflexivity” in his conception of rationality, and the place of Soros’ reflexivity and uncertainty in Rosenberg’s overall argument (IV).

    Section I
    The ‘rational fool argument’ that Rosenberg focuses on is representative of Sen’s broader considerations on economic rationality, which are primarily concerned with the positive/normative dichotomy in economics. Paraphrasing Rosenberg, the argument can be accurately summarized by saying that: not only rational choice theory does not explain a great deal of actual human behaviors, but also, that it would be foolish to actually abide by the dictates of rational choice theory. The first part of the criticism is positive, while the second is normative. This has to be understood in reference to what is sometimes called ‘the dual status of rationality’ in economics, i.e., the same conception of economic rationality is supposed to explain/predict and prescribe choices. This, in turn, is supposed to allow positive economics and normative economics to be separated, and yet to communicate. Sen does not argue against the duality of economic rationality per se. He rather attacks the standard way by which this duality is theorized. His analysis aims at making explicit how positive and normative aspects of economic rationality are entangled, to criticize this entanglement, and then to propose an alternative entanglement, so to speak. To do so, he starts by distinguishing between the (traditional) “direct” and “indirect” uses of rational choice theory in economics. The former characterizes the substance of rationality in order to “think and act wisely and judiciously, rather than stupidly or impulsively” (Sen 2002a, p.42). The latter describes and predicts actual behavior, first by characterizing the substance rationality (i.e., the direct use), “and then assuming that actual behavior will coincide with rational behavior” (Sen 2007, pp.20-21). What Sen condemns is that “[t]he two uses – directly normative and indirectly predictive – are closely linked. Indeed, the latter is basically parasitic on the former, but not vice versa” (Sen 2002a, p.43). Part of Sen’s intellectual project of bringing economics and ethics back together consists in turning this scheme upside down so to speak, by developing the normative direct use of rational choice theory from its positive indirect use. In other words, he theorizes the substance of rationality with insights from actual behaviors, interpreted through ethical considerations. This has to be understood within the broader project of scrutinizing the “triple entanglement: of fact, convention and value”, in economics and other policy-oriented social sciences (Sen 2005, p.112; see also Putnam and Walsh 2012).

    Section II
    Interestingly, Rosenberg connects Sen’s criticisms with the ones made by psychologists, such as Daniel Kahneman, Amos Tversky, and Gerd Gigerenzer. However, two subtleties could undermine such a connection. On the one hand, although Kahneman (1994) does indeed make some reference to ethical considerations by referring to the work of Sen, his work with Tversky along with the related behavioral economics do not share Sen’s stance on the positive/normative dichotomy in economics, especially with respect to economic rationality. In a nutshell, this type behavioral economics calls for a disconnection between the direct normative use and the indirect positive use of rational choice theory by conserving the former and theorizing only about the latter from experimental and empirical observations. On the other hand, in psychology, the ‘Heuristics and Biases’ research program of Kahneman and Tversky and the ‘Ecological Rationality’ research program of Gigerenzer are opposed on the issue of the normative soundness of rational choice theory. The latter argues against it, like the rational fool argument, but proposes theorizing from insights borrowed to evolution theory, not ethics. Therefore, the fast and frugal heuristics of Gigerenzer are more in line with Rosenberg’s Darwinian perspective than the heuristics and biases of Kahneman and Tversky. Thus the discordance between these two research programs in psychology might be of importance for the overall argument of the paper. For this reason, to put both in the same basket is questionable.

    Section III
    What, then, to make of the rational fool argument, understood within Sen’s broader intellectual project explained above, with respect to Rosenberg’s Darwinism? To answer this question, I propose to contrast one of Rosenberg’s claims with some reflections by Sen “On the Darwinian View of Progress” (2002b). In this paper, Sen opposes two evaluative frameworks to assess natural as well as social outcomes of evolution: “the quality of species”, which is based on a Darwinian view of progress and takes degrees of fitness as a central criterion, and “the quality of life”, which is based on, roughly speaking, his conception of rationality, and takes survival to reasoned scrutiny as a central criterion. For instance, if social evolution has made it that Humans “do reasonably value the survival of all the species” (p.488), even the ones that are ““unfit” and “unselected”” (ibid), then environmental concerns are better understood in terms of human reasoning than in terms of the Darwinian view of progress. Now, when Rosenberg discusses the Hayekian view of the market price system as performing “a function indispensable not just to modern life but to all human life beyond the Pleitoscene” (p.10), which would have never “been forseen by humans, no matter how rational” (ibid), he is clearly taking the market price system as an outcome of a Darwinian evolutionary process; a ‘social’ outcome, one could add. So far so good. But when this outcome is seen as “a solution to the problem people face that is so ingenious it automatically and successfully responds to […] subversion attempts” (pp. 10-11), we arguably have here a “quality of species” evaluation based on a “Darwinian view of progress”, in Senian terms. There might be a problem here with respect to the coherence of Rosenberg’s overall argument with the rational fool argument outlined previously, not on the (positive) characterization of the phenomena, but on its (normative) evaluation. As Sen (2002b, p.494) puts it:

    “It is a non sequitur to argue that since our ability to reason may have evolved through survival advantage, it can be used only for that purpose. Our faculties are not, in general, specifically tied to a single purpose. […] No matter how and why our ability to reason may have developed, we can use it as we like, and scrutinizing the criterion of reproductive success or survival advantage as a yardstick of progress is among its possible uses.”

    This is important because:

    “The dichotomy between leaving the deprived and the miserable to nature, and using public action to try to help them, remains important in the contemporary world. Indeed, the significance of the contrast may well have increased in recent years, with the growing tendency to let impersonal forces – the market mechanism, for example – have their way. The bankruptcy of the Second World has often been interpreted not simply as the failure of a particular system of intervention, but as the impossibility of designed improvement of all kinds” (ibid, p.498)

    The point here is to stress the importance of clarifying normative criteria of evaluation, especially while conducing positive Darwinian analyses, because the edge is rather thin and sloppy between the two. It is not because a phenomenon occurred through necessary and sufficient blind evolutionary forces that, once this occurrence is acknowledged as such, it is nonsense to assess its outcomes through non-evolutionary related criteria. As Sen (ibid, p.497) also points out, Darwinian analyses have the tendency to bias evaluation toward:

    “silence on the case for adjusting the world to suit our needs. That gap in attention is the direct result of judging progress by the nature of species, rather than by the kind of lives they can lead – which would have immediately drawn attention to the need to adjust the external world. From that common Darwinian point, the activist view proceeds toward genetic manipulation, whereas the more passive view suggests trusting nature. Neither directs us toward reforming the external world in which we live.”

    Even though this is not exactly well in line with Rosenberg’s Hayekian story about the market price system, it is well in line with his further discussion of the failure of functionalism, and the need to sometimes “think of the people as the environment and […] institutions […] as the things that survive” (p.13) and “exploit human characteristics” (ibid).

    Section IV
    However, as Rosenberg highlights in his discussion of Soros on ‘uncertainty’ – the fallibility of humans especially regarding their expectations – and ‘reflexivity’ – preferences are not independent from expectations in a complex way –, actions directed at changing the external world are ‘tricky’, to say the least. Sen’s use of the word “reflexivity” is different from Soros’. Nevertheless, there might be an interesting complementarity here. “Reasoned scrutiny”, which is how Sen labels his broad conception of rationality, imposes two main constraints on choices for them to count as ‘rational’. These constraints concern the ‘reasons’ held by individuals regarding their choices, which (1) have to survive thorough ‘reflexive’ scrutiny by the individuals themselves (Sen 2002a), and which (2) have to be able to enter ‘reasoned public scrutiny’ for discussion, though not necessarily for the purpose of being universally accepted (Sen 2009). Now, on a Sorosian account, reflexivity and uncertainty are combined in a way that human actions make the economic environment unpredictable. This unpredictability is explained by the Darwinian nature of uncertainty and reflexivity, as Rosenberg argues. Thus, in a way, there is no knowledge of the outcomes of uncertainty and reflexivity, but there is the knowledge of such an absence of knowledge. This is relevant for at least the individualistic part of Sen’s reasoned scrutiny, where the latter’s conception of ‘reflexivity’ lies. It means that no reasons can survive thorough scrutiny in a definitive manner. Sen defends this idea from an ethical perspective, which could then be strongly supported by the Darwinian conception of uncertainty and reflexivity defended by Rosenberg. For the collective part of Sen’s reasoned scrutiny the implications are not so clear. Let’s consider the following summary of Sen’s criticism against the Darwinian view of progress, which, on his account:

    “ignore[s] the quality of life of human beings and other animals; it undermines the importance of rationally evaluating our priorities and trying to live according to them; and it draws our attention away from the need to adjust the world in which we live. This, in turn, tends to encourage either activism in genetic manipulation (as in the eugenics movements), or a passive reliance on spontaneous progress (more in line with Darwin’s own pronouncements). But in neither case is much attention paid to the dependence of the quality of our lives on the nature of the adjustable external world.” (Sen 2002b, p.500)

    Given that reasoned public scrutiny aims, in fine, at changing the world to improve quality of life, the implications of uncertainty and reflexivity are that, even in cases where the world actually is adjustable, it is impossible to adjust it to purposefully preconceived objectives. The question remains, though, as to whether – or more precisely how – the knowledge of such impossibility could, in turn, help us in conceiving a better way of making adjustments. In other words, if knowledge of evolution has helped us in our aim of understanding and changing the world, could the knowledge of reflexivity and uncertainty contribute as well?

    References:
    Kahneman, D. (1994). New Challenges to the Rationality Assumption. Journal of Institutional and Theoretical Economics, 150(1), 18–36
    Putnam, H., & Walsh, V. (eds.). (2012). The End of Value-Free Economics. Routledge
    Sen, A. K. (2002b). “Introduction: Rationality and Freedom”, in Rationality and Freedom. Harvard University Press.
    Sen, A. K. (2002b). “On the Darwinian view of Progress”, in Rationality and Freedom. Harvard University Press. Originally publishd in 1992, London Review of Books, 14; republished in Population and Development Review, 1993.
    Sen, A. K. (2005). Walsh on Sen after Putnam. Review of Political Economy. 17:1, 107-113
    Sen, A. K. (2007). Why Exactly is Commitment Important for Rationality. In F. Peter & H. B. Schmid (Eds.), Rationality and Commitment (pp. 17-27). Oxford University Press
    Sen, A. K. (2009). The Idea of Justice. Penguin Press.