Hayek and the Use of Physics in Economics: Towards a Progressive Synthesis?

Abstract

F. Hayek is known as an economist and a political philosopher who provided a specific social theory based on the idea of spontaneous order. This order state comes from a process that Hayek called ‘catallaxy’ and that results from a particular phenomenon of emergence. In this context, social systems such as money, market or language are presented as spontaneous orders, resulting from a particular evolutionary process of society. Although Hayek is well known to be an opponent of importation of physical concepts into economics, this article proposes a parallel between the Hayekian theory and a new field called ‘econophysics’. More precisely, I will show how the way of characterizing the phenomenon of emergence by econophysicists might, to some extent, be an analytical characterization of the Hayekian emergence. This claim will be explained and illustrated with the presentation of econophysical works describing money and language as a spontaneous order taking the form of a power law. The last section will discuss the debates about the methodological relevance to associate Hayek’s theory with an area of knowledge coming from physics and the potential\economic implications of such a rapprochement.

JEL: B25, B41, P11

Keywords

Econophysics Hayek emergence complex system

Introduction: The Hayekian Framework

Friedrich von Hayek (1899–1992) was an influential economist and a forward-looking commentator about economic complexity. Although Hayek was a famous intellectual who worked on monetary cycle theory (1929, 1931 & 1939), he also contributed in neuroscience (1952), political philosophy (1973) and epistemology (1945). In economics, Hayek is well known for his social theory based on the idea of spontaneous order that he called ‘catallaxy’. As Sawyer (2005, p. 43) emphasized it, Hayek ‘described all higher-level social phenomena as an emergent order from individual actions. For him, social order cannot result from anything resembling conscious control or planned rational design. According to Hayek (1967), the incomplete knowledge about the situation and the numerous non-simple interactions between individuals lead to let order emerge out of this complexity (Ball, 2003). Therefore, he proposed a bottom-up perspective within it is only by the individualist or compositive method that we can give a definite meaning to social processes that often are more than merely the sum of their parts (Hayek, 1944, p. 30).

Economic systems are presented as a ‘highly complex affair’ (Hayek, 1967, p. 71) from which a macro result called ‘spontaneous order’ emerges. The Hayekian way of conceptualizing the economic systems make them candidate for the complexity treatment because they are assumed to be composed by multiple components (agents) interacting in such a way as to generate a macro property (spontaneous order) that is more than the mere sum of the individual agents’ behaviour. The epistemic justification of this conceptualization refers to the impossibility to aggregate all information related to a specific economic situation. By noting that an ‘economic problem is a problem of utilization of knowledge not given to anyone in its totality’, Hayek (1945, p. 520) conceived uncertainty as a lack of knowledge (Ball, 2003).¹ Knowledge we have about economic situations is always fragmented, uncertain and complex as Hayek wrote it:

The peculiar character of the problem of a rational economic order is determined precisely by the fact that the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess. (Hayek, 1945, p. 1)

That means that ‘we cannot expect that this problem will be solved by communicating all this knowledge to a central board which, after integrating all knowledge, issues its orders. We need decentralization’ (Hayek, 1945, p. 524). This decentralization takes a specific form since the only way of governing the society is to let self-regulating structures emerge in society² (Hayek, 1945). In other words, the lack of knowledge about economic systems should invite governments not to intervene in society.³

The notion of emergence and spontaneous order are very important in the Hayekian framework because they appear to be universal and repetitive in society: the market, money, language or even ethics are looked on as spontaneous orders because they do not result from a human design (decided by a nation or a government), but rather the consequence of an evolutionary process. This concept of spontaneous order is directly inspired by the Darwinian idea of evolution⁴: Once emerged in a society, a spontaneous order evolves into a more and more abstract system (Dostaler, 2001, p. 87). Although Hayek justified his argument for ‘spontaneous order’ from an epistemological point of view (absence of knowledge), he did not give an analytical form describing the emergence of this order. Hayek has no model or analysis showing how decision-making under uncertainty will generate a spontaneous order. He simply assumed that such a solution will be attained because uncertainty creates possible opportunities for people. This point still generates a lot of debates in the economic literature (Aimar, 2002; Butos, 2003; Légé, 2007).

In terms of modelling, Hayek (1942, p. 288) strongly disapproved the idea that economic uncertainty could be reduced to a physical law.⁵ However, the Hayekian framework is quite ambiguous in its conceptual links with hard sciences. For instance, by arguing that the emergence of a spontaneous order is the only justified way of characterizing the economic complexity, Hayek implicitly considered that economic phenomena can be described through a concept of Darwinian evolution (Légé, 2007). In the same vein, the Hayekian epistemology is very ambiguous about the possibility of making predictions. Since the knowledge about a society is fragmented and dispersed throughout individuals, economic (or other kind of) knowledge cannot be centralized in any one mind or institution. The knowledge we have on a situation is always limited and contextualized—therefore, our ability to make accurate predictions of future single events is non-existent and we are just able to develop a limited knowledge about patterns or order of events. Despite he recognized that economists use prices as empirical observations, Hayek deplored the confusion that creates between objective and subjective data (Gray, 1982). For him, all social phenomena are deeply subjective, so that knowledge originating in empirical observations is only partial knowledge of an economic system (Hayek, 1945, p. 523).

Hayek (1942, p. 288) rejected the idea of reducing the economic reality to a causal or statistical law, but he was not totally opposed to the use of statistics—as reminded in the next section, he simply argued that this kind of knowledge is limited and incomplete (Hayek, 1937, 1945). However, as Légé (2007, p. 35) emphasized, the Hayekian framework is sometimes abstruse on this point since Hayek used a causal reasoning to explain inflation and unemployment. In the same vein, despite his rejection of the universal law (Hayek, 1937), Hayek provided a social theory based on the notion of emergence that, paradoxically, appears as an implicit universal law for complex economic systems. This article does not aim at debating these issues (see Aimar, 2002; Butos, 2003; Légé, 2007 for further details on these debates), but it rather shows how the existence of these debates pave the way to a potential conceptual bridge between the Hayekian framework and a new area of knowledge coming from physics that is called ‘econophysics’. In line with the Hayekian framework, econophysics (as I will explain it in the following section) deals with the emergence of a spontaneous order taking the form of a statistical invariance. This specificity of econophysics calls for a re-evaluation of the link between Hayekian framework and the use of physics in economics.

Physics and Economics in the Eleventh Century

In the 1990s, physicists had turned their attention to economics by applying ideas, models, conceptual and methodological methods of statistical physics. The movement’s official birth announcement came in a 1996 article by Stanley, Afanasyev et al. (1996).⁶ Econophysics was at that time still a young and ill-defined field, but nowadays, the field is institutionalized and well established (Gingras & Schinckus, 2012).

The influence of physics on economics is nothing new. A number of writers have studied the influences of physics on economics: Mirowski (1989) detailed the contributions of physics to the development of mathematical economics, while Ingrao and Israel (1990) highlighted the role of mechanics in the conceptualization of equilibrium in economics.⁷ Although there exists a lot of methodological\conceptual debates on the emergence of this field,⁸ calling into question its contributions to economics, one can mention two major differences between econophysics and the previous influences of physics on economics: (a) Econophysics is a new way of thinking economic phenomena through the lens of complexity and emergence (Rickles, 2008; Schinckus, 2013b). This time, it is the physicists that are going beyond the boundaries of their discipline, using their methods to study various problems thrown up by social sciences (previous influences of physics on economics took the forms of an importation of physical concepts into the existing economic theory, while econophysicists ignore or deny the economic theory.

These two specificities of econophysics call for a new methodological analysis of the links between economics and physics. Because econophysics and the Hayekian framework both associate economic complex systems with an emergent spontaneous order, this article proposes a conceptual parallel between the notion of emergence used in econophysics and the Hayekian idea of emergence. A contribution of this article is also to show that the methodological links between physics and economics are moving through the history of economic thought.

In a sense, I present here the econophysics as a compatible frame to the Hayekian theory since the first offers an operational solution to the theoretical argument given by the latter. After reminding that complexity and emergence are at the heart of econophysics, I will show that this key concept used in this field clarifies the notion of emergence proposed by Hayek.

The Emergence of Econophysics

The birth of econophysics is usually associated with the publication of Mantegna (1991) in which the authors compared the occurrence of extreme variations on the financial market with the occurrence of earthquakes whose observations can statistically be described through a power law. Mantegna and Stanley (1999) and Jovanovic and Schinckus (2013) explained that this initial article opened the door to an increasing number of empirical works observing power laws in socio-economic phenomena: Mantegna and Stanley (1994), Yamamoto and Miyazima (2004), Lux (2009), Bak, Paczuski, & Shubik (1997) and Gabaix, Gopikrisnan, Plerou and Eugene (2003) observed that the large fluctuations on the financial markets can be captured through a power law while Levy et al., (1995, 2000) and Claes, Frenken and Hellervik (2006) confirmed the conclusion made by Pareto (1897) one century before by showing that economic data can statistically be characterized by a power law. In the same vein, Amaral et al. (1997) explained the annual growth rates for the US manufacturing companies can also be described through a power law, whereas Axtell (2001), Luttmer (2007) and Gabaix (2009) claimed that this statistical framework can also be used to characterize the evolution of the firms size as a variable of their assets, market capitalization or number of employee. These ‘size models’ have afterwards been applied for describing the evolution of the cities size (Eeckhout, 2004; Gabaix, 1999).

The specificity of these patterns evoked above, identified by physicists, refers to their statistical form since they can be expressed as what we call a ‘Levy process’. A Lévy process, named after the French mathematician Paul Lévy, is a time stochastic process with stationary and independent increments,⁹ càdlàg paths.¹⁰ More precisely, Lévy worked on a generalization of the Gaussian statistical framework by developing a new class of distribution called Lévy α-stable. Lévy’s α-stable movements are processes whose accretions are independent and stationary and follow an α-stable law of type P(X > x) = x^–α, in which it is possible to observe constancy of the parameter α. Although this statistical framework is well known by economists (since this kind of regularity was observed for the first time by Pareto for the distribution of wealth), they do not use it for a technical and methodological reason: Stable Levy processes generate infinite variance and the variance is probably the key variable of finance since it refers to risk (for further details on this point see Jovanovic & Schinckus, 2013, 2017; Schinckus, (2013a) for a technical solution to this problem).

From an economic point of view, these statistical regularities are often associated to what we call ‘stylized facts’ which refer to ‘empirical facts that arose in statistical studies of financial\economic time series and that seem to be persistent across various time periods, places, markets, assets etc’ (Chakraborti, Chakraborti, Muni Toke, Patriarca, & Abergel, 2011, p. 994). For econophysicists, economic systems are composed of multiple components (no learning agents) interacting in such a way as to generate the macro properties for systems (Rickles, 2008, p. 4). Concretely, econophysicists work with a huge number of economic data from which they let emerge a statistical order characterizing the macro properties of the systems.¹¹

Econophysicists associate emergence with the observation of persistent patterns in accordance with Holland (1998, p. 295) who wrote that ‘emergent phenomena are, typically, persistent patterns with changing components that often satisfy macrolaws’.¹² By associating emergence with no deducible and no reducible macro laws, econophysics rejects the reductive explanation (McCauley, 2004): econophysical macro laws are not predictive (i.e., deduced or anticipated) because they are simply observed from empirical data. In this context, the only prediction refers to the future emergence of a phenomenological regularity implying that one remains unable to predict the evolution of these complex systems.¹³ In a sense this methodological aspect is coherent with the data-driven philosophy which drives econophysics (McCauley, 2006) since it is not possible to define the a priori form of the emerging invariance without observing empirical data (Craver, 2006). The irreducibility and the no deducibility of econophysical macro laws suggest the use of a strong emergence in which the gap between micro components and macro level is not analytically defined. In this context, as Schinckus (2013b) explained it, econophysicists use the term ‘emerge’ (or notion evoking the idea of emergence) as a ‘filler term’ to conceptualize the gap between the micro and the macro levels.

The Improbable Link Between Econophysics and the Hayekian Framework

By writing his ‘Theory of Complex Phenomena’, Hayek (1964) initiated a large field of research about complexity in social sciences. The links between the Hayekian framework and the complexity paradigm have been documented by a prolific literature (as, e.g., Caldwell, 2004; Gaus 2006; Koppl, 2000; Lavoie, 1989; Rosser 1999; Vriend, 2002). In line with all these works, I emphasize, in this section, the common features between econophysics and the Hayekian tradition. More precisely, the first subsection will explain how econophysics can describe the emergence of some Hayekian spontaneous order such as language, money or the evolution of the market. Although econophysicists did not use a political argument when they developed their models, the second subsection will discuss the political (dis)similarity between econophysics and the Hayekian framework. Finally, I conclude this article with a methodological debate about the possibility of connecting econophysics with the Hayekian tradition.

The Argument of Emergence

Hayek provided a specific response to the understanding of complex systems: the emergence of a spontaneous order in complex system composed by a high number of interacting agents. This stance implicitly refers to the notion of emergence used in econophysics in which it appears as a macro phenomenon arising from non-simple interactions between lower level entities¹⁴ (Rosser, 2003).

By considering that social systems are ‘more than the sum of their parts’, Hayek (1944, p. 30) emphasized the fact that economics is concerned with emergent phenomena and complex systems. He defined self-organizing emergence as a phenomenon based on multiple agents interacting in such a way as to generate the macro properties of economic systems (Hayek, 1942). However, as mentioned above, Hayek did not give an operational definition to this phenomenon. The gap between the micro and the macro level remains obscure and the verb ‘emerge’ could be associated with other ‘filler terms such as activate, cause, produce and represent’ (Craver, 2006, p. 360). As evoked in the previous section, Hayek called ‘catallaxy’ the emergence of a self-organized order in the economic systems and he presented this phenomenon more as a political argument than a scientific one: Social order must result from the ‘free will’ of actors and not from a utopian rational planner.¹⁵

Econophysicists also use ‘emerge’ (or concept evoking the idea of emergence) in the same way as ‘filler term’ to conceptualize the gap between the micro and the macro level. Although econophysicists use of the notion of emergence appears as a fitter term, they propose, in contrast to Hayek, a statistical characterization of the self-organized (spontaneous) order which is surprisingly in accordance with the definition Hayek gave to this concept.

By considering the statistical macro patterns as a novel quality of physical or social systems, econophysicists implicitly consider these complex systems as entirely constituted by composite structures that are not mere aggregates (neither definitional extension) of the simple ones. This notion of emergence of statistical macro laws suggests the existence of a gap between micro and macro scales consistent with classical emer-gentism. I have defined in the first section what is called ‘strong emergentism’, in which an emergent property cannot be reducible/deducible; neither predictive (Kim, 2006; Morgan, 1923). This phenomenological econophysics keeps this radical enigma about the occurrence of emergent properties. In this context, the ability to identify a power law in a complex system does not mean that one can predict the behaviour of this system. Indeed, the predictability must be understood in a specific way: These macro laws are not predictive (i.e., deduced or anticipated) because they must be observed from the past empirical data. The only thing scientists are able to predict is that a statistical invariance will appear in specific complex systems, but they are unable to deduce the evolution of these complex systems. The emerging invariance is not deductible from the initial defined rules, but it rather inductively results from an accumulation of empirical data observed in the past.

The increasing computational power of computers observed in the 1980s was accompanied by a growing expansion of storage capacities. Scientists quickly understood that computers offered an important source of knowledge. Computerized simulations are more and more used and the identification of macro patterns in historical data became easier. The growing storage capacities of computers allowed modellers to deal with large databases paving the room for a better statistical analysis. Nowadays, empirical data are collected for everything and several complex systems appear to exhibit the same kind of macro patterns taking the form of power laws. Concretely, the way of identifying this output consisted in checking visually on a simple histogram that the frequency distribution of the quantity of x appears as a straight line when plotted on double logarithmic axes. If a distribution approximately falls on a straight line, then one can consider that the distribution follows a power law, with a scaling parameter α given by the absolute slope of the straight line. Such visual investigation has guided econophysicists empirical research (Jovanovic & Schinckus, forthcoming; Mantegna & Stanley, 1999) and can be illustrated with Figure 1.

Figure 1.

Source: (Gabaix, 2009, p. 276).

All these statistical observations show a linear relationship on a log–log plot, meaning that the numbers on both the axes increase by a power of ten with each tic on the axis. In other words, variables expressed on the two axes can be related through the following equation: ln p(x) = – αln x + c, where α and c are constant: While the first is the slope of the line (and then to the sensitivity of variations), the latter is ‘an uninteresting’ scale parameter (Newman, 2004, p. 1) (i.e., referring to the unit of measure used in the observations). This formula can also be reformulated (by taking the exponential of both sides) as a power law: p(x) = Cx^–α, where the α is the characteristic exponent of the power law (this parameter is an indicator of stability since it refers to the sensitivity of potential variations).

These empirical observations are not due to a (un)happy coincidence since this linear relationship has also been identified in a lot of contemporary events: the magnitude of earthquakes (Newman, 2004), citations of scientific papers (de Solla Price, 1965; Redner, 1998), web hits (Adamic & Huberman, 2000), telephone calls (Aielloet, Chung, & Lu, 2000), copies of books sold (in the USA) (Hackett, 1967), diameter of moon crater (Neukum & Ivanov, 1994), etc. In other words, this pattern appears in the observation of social and natural phenomena leading some authors to consider this linearity as a law of nature.

Although the increasing use of computers was a necessary condition for associating power laws with dynamic complexity, it was not a sufficient condition. This conceptual association also required a theoretical justification which has been proposed by a member of the SANTA FE INSTITUE (SFI): Per Bak et al. (1987) who developed what he called ‘self-organized criticality’. Bak (1948–2002) was a Danish theoretical physicist specialized in phase transitions, who worked at the Brookhaven National Laboratory at that time. He became a member of the SFI in 1987. According to Bak, the linearity, visually identified on the histogram related to the occurrence of a phenomenon, is the expression of the complexity of this phenomenon: ‘This simple law is impressive in view of the complexity of the phenomenon’ (Bak, 1994, p. 478). Afterwards, the author wrote,

This is an example [plot with occurrences of earthquake] of a scale-free phenomenon: there is no answer to the question ‘how large is a typical earthquake?’ Similar behaviour has been observed elsewhere in Nature [. . .] The fact that large catastrophic events appear at the tails of regular power-law distributions indicates that there is ‘nothing special’ about those events, and that no external cataclysmic mechanism is needed to produce them.

In other words, we have a ‘self-organized criticality’ in which, ‘slowly driven dynamic systems, with many degrees of freedom, naturally self-organize into a critical state obeying power-law statistics’ (Bak, 1994, p. 480). The basic idea of self-organized criticality is that certain phenomena maintain themselves near a critical state. A telling example of that situation is a quiet sand pile, in which the addition of one grain generates mini-avalanches. At some point, these mini-cascades stop, meaning that the sand pile has integrated the effect of this additional gran. The sand pile is said to reach its self-organized critical state (because the addition of a new sand grain would generate the same process). Physicists talk about ‘critical state’ because the system organizes itself into a fragile configuration based on a knife-edge (the addition of only one sand grain would be enough to modify the sand pile). Bak et al. (1987) showed that the dynamics of critical state (i.e., modifications of the sand pile) follow a power law distribution.

Epistemologically, econophysics is founded on the belief in the universality of the statistical properties, that is, the fact that they reappear across many and diverse phenomena (McCauley, 2004). This statistical universality that can be characterized by power laws is at the heart of the econophysics (Bouchaud, 2002; Stanley, Gabaix, Gopikrishnan, & Plerou, 2007). All these power laws used by the econophysicists, to characterize the evolution of complex systems, also describe the emergence of some human spontaneous orders like language or money, for example.

Loreto and Steels (2007, p. 759) wrote an interesting story about the theories of language. According to them, the formal and static Chomskian definition of language¹⁶ is giving the way to the view that language can be looked on as a complex adaptative system. Once language is considered as a complex system, statistical physics suddenly becomes relevant for studying the emergence and the evolution of this process. Gong, Minett and Wang (2006) provided a model of emergence and evolution of a human language. They used a computational model based on the agent which acquired compositional linguistic knowledge by detecting and learning recurrent patterns in the linguistic input. The authors explained that the emergence of a human language (i.e., with meaningful sentences) can be explained only if the popularity¹⁷ of the individuals follows a power law distribution. Moreover, Gong, Minett and Wang (2006) reminded that the development of linguistic features also follow power law distributions (see Baronchelli, 2006; Soares, Corso, Luneca, 2005; Wichmann, 2005) for an introduction to the ‘mechanics of language’). In accordance with the conceptual reasoning proposed by Hayek, to present the language as a spontaneous order, econophysicists showed that a language emerges when some regularities appear in the actors’ behaviour.

Money is another important topic for Hayek who considered its emergence as a self-organized order. Several econophysicists (Bak et al., 1997; Farmer, Shubik, & Smith, 2005; Yasutomi, 1995) studied the emergence of money through the lens of complexity.¹⁸ Donangelo and Sneppen (2000) and Shinohara and Gunji (2001) studied the dynamics of exchange in a system composed of many interacting agents until an emergence of money appeared. They showed how this process can be described through a scaling relation between a number of exchanges and time. While Donangelo and Sneppen (2000) used non-Gaussian statistics whose fluctuations of exchanges are quantified by anomalous Hurst exponent¹⁹ (i.e., scaling laws); Shinohara and Gunji (2001) developed a reciprocity model in which the interactions between agents is asynchronous²⁰ and the duration of the good M (i.e., time needed by M to become money) follows a power law (Shinohara & Gunji, 2001, p. 146).

Although they seem to be conceptually opposed, econophysics and Hayekian frameworks, both consider the notion of emergence as a process characterizing the state of complex systems.²¹ Actually, econophysicists and the Hayekian tradition share the common key assumption according to which the dynamics of economic phenomena can be described in terms of spontaneous order that takes the form of an emergent property resulting from the complexity of micro interactions. In a sense, with econophysics, the Hayekian idea of self-organized order finds a more analytical formulation through the observation of a power law. This idea of spontaneous order echoes the ‘free will’ argument of agents that I will discuss in detail in the next section.

The ‘Free Will’ Argument

Although econophysicists usually do not justify their approach by using a political argument, the statistical formulation of the emergence they developed seems to be directly in line with the liberal and individualist view of economic system enhanced by Hayek.

As Rosser (2003, p. 3) explains “econophysics formulations implicitly assume heterogeneous actors’, that is, a large numbers of interacting subunits that display free will”in line with what Amaral et al., 1997 developed. This link between the probabilistic nature of the laws of the statistical mechanics and the free will was been developed by Majorana (1942) and summarized by Bouchaud and Potters (2003) as follows:

The statistical framework is particularly well adapted to describe the free will (real or supposed) of individuals. Each actor is at liberty to act as he/she wants according to personal and complex reasons that are difficult to understand. The collective behaviours, observed in their anonymous totality, have a regularity such as we can find some laws, causality and invariant like we do in physics.

Therefore, the application of physics to economics in its econophysical form appears to be compatible with the Hayekian acceptation of the word since the first offers a conceptual framework for describing an emergent order based on ‘free will’ and an interactive reductionism (Colander et al., 2008). This interactive reductionism contrasts with the atomistic reductionism, used in the neoclassical economics within reality, must be explained in terms of rational expectations. This opposition with the economic mainstream is very important because it involves some empirical implications: In basing all economic macro phenomena on the rational representative agent, economists implicitly set the macro level equal to the micro level. The consequence is that all macro concepts as ‘market’, ‘systemic risk’ or ‘financial crisis’ are misunderstood in the economic theory. In this perspective, it is simply impossible to describe (and understand) an economic crisis like we faced in 2008. In the opposite, econophysicists focus their works on interactions between the overall complex system and its parts. Since the economic activity is interactive in essence, this perspective is more appropriate to understand the connections between all parts of the economic systems (firms, banks, households). In this perspective, the analysis of a crisis phenomenon (and its repercussions on investment or consumption) then becomes possible.⁸ This interactive reductionism enhanced by econophysicists is directly in line with the Hayekian idea; according to which, the spontaneous orders in society arises from the interactions between individuals (Hayek, 1952).

The Methodological Debate: Conceptual Similarities with Different Epistemologies

The methodological positions of Hayek generate a lot of debates (see for, example, Aimar, 2002; Butos, 2003; Légé, 2007; Tuerck, 1995). While sharing the subjectivist and methodological dualist positions of Menger and Mises, Hayek diverges from them on matters of epistemology—more precisely, he accepted the Popperian philosophy of science; according to which, the hallmark of any scientific theory is its openness to empirical falsification (Hayek, 1964).²² He is an opponent of the belief in a universal law in the economic phenomena (1974), but he implicitly considers the emergence as a universal law since it characterizes the process of all spontaneous order (market, language, money, etc.). Although Hayek often reminded the distinction between social sciences and exact sciences, he acknowledged that society and biology are based on the same concept of evolution. His anti-mechanical perspective based on a scientific subjectivism describes human mind as an apparatus of classification; however, as Tuerck (1995) emphasized it, his view of spontaneous orders can be considered as a mechanistic process.

At the first glance, Hayek’s framework is incompatible with econophysics because he thought that complex economic systems could not be studied from a physical point of view. Moreover, Hayek is a firm opponent of the use of quantitative technique which aims at discovering universal laws in economics. However, as debates about Hayekian theory stressed it, there is a difference between what Hayek explicitly claimed about hard sciences and his way of using them. This ambiguity is pronounced in the Hayekian warning about the use (and the abuse) of objective statistics: Although Hayek ‘was arguing against the mindless transference of the methods of the natural sciences into the social science’, he was not in favour of a more hermeneutical economics (Caldwell, 2004, p. 52). Statistics are not enough to explain a complex phenomenon, but Hayek acknowledged they can contribute in the understanding of this kind of phenomena (Hayek, 1974, p. 148). Hayek did not totally reject the use of objective probabilities. He just claimed that these probabilities provide limited and incomplete knowledge about reality (Hayek, 1945, p. 523). For Hayek, statistical regularities in economics must not be understood as equivalent to causal laws in the physical sciences, precisely because underlying relationships between variables cannot be observed.

In accordance with Hayek, econophysicists acknowledge that there is no universal law for social phenomena, as McCauley reminds it, ‘Econophysicists [. . .] would like to ignore the details and hope that a single universal “law of motion” governs markets, but this idea remains only a hope, not a reality’ (McCauley, 2004, p. 4). Although econophysicists import their ‘universal laws’ from physics, they know that the underlying phenomena they are studying are not governed by physical laws, but rather by human issue. McCauley (2004) added that the best they can reasonably do is to describe what happens with some general patterns. In a Hayekian perspective, these general macro patterns that often take the form of power laws must be seen as a rearrangement of familiar elements in a systematic manner where only predictions about future configurations of the pattern (but never about a single event) can be estimated—as Hayek wrote it: ‘predictions will also refer only to some properties of the resulting phenomenon, in other words, to a kind of phenomenon rather than to a particular event’ (Hayek, 1967, p. 15).

As mentioned before, the only thing econophysicists can predict is the emergence of an order\invariance, but the prediction of a specific form for a particular system remains impossible. Power laws are macro patterns describing an order of events, but these laws do not claim to predict specific forms that future instances of evolution may take (if it was the case, all econophysicists would be rich by applying their models). In line with the Hayekian tradition, power laws can then be looked on as a general classification, a ‘partial map’ of the external world.

These similarities are interesting and they call for more question: Is econophysics totally compatible with the Hayekian tradition? How do econophysicists derive different implications from their conceptual similarities with the Hayekian tradition? This question will be studied in the following subsections through the analysis of two dimensions: (a) an epistemological one; and (b) a political economic one.

Different Epistemologies

Econophysicists consider economic\financial phenomena as complex systems composed by a large number of interacting elements. While economists are well known to find their models on the methodological individualism using the mere addition as a process generalizing their knowledge,²³ econophysicists rather consider that ‘something happens’ between the micro and the macro level of complex systems. However, this emergence of macro results must be captured\explained to describe ‘what is happening’ between these levels. Whereas Hayek does not provide any analytical formulation of the emergence associated with the spontaneous order; econophysicists characterize the emergent order as an asymptotic reasoning. Specifically, the emergence of a macro pattern is considered as a novel (not expected) and robust (regularly observed) result describing the dynamics of micro elements whose number of interactions goes to infinity. In order words, power laws appear as a novel behaviour by taking the limit, where n is the number of observations in accordance with,

\lim_{n \to \infty} T_{2} = T_{1}

T₁ refers to the power law observed at the macro level, while T₂ rather characterizes the (unknown or unnecessary) description of micro interactions. In this schema, a more encompassing (macro) theory T₁ reduces a specific (micro) theory T_2, if the laws of T₁ can be asymptotically derived from the observations\iterations of T₂. This way of characterizing the notion of emergence is inspired by Batterman (2002) who promoted the development of an ‘asymptotic explanation’ (Batterman, 2002, p. 3). Through this definition, Batterman (2002) claimed that many of why-questions based theories are explanatorily deficient to understand how universality can arise (by universality, Batterman refers to ‘a feature of the world—namely, in certain circumstances distinct types of systems exhibit similar behaviors’, Batterman, 2002, p. 9).²⁴ This universality takes the form of a specific statistical relationship (power law) whose explanatory nature appears to be inductive: The use of such statistical explanation for characterizing a particular event is directly relative to a given knowledge (i.e., generalization of power laws observed in the past) that would pre-exist to these specific events (Hempel, 1965). In other words, the statistical explanation has quite an explanatory power since it ‘obviously depends heavily upon the concept of statistical relevance’ (Salmon, 1984) p. 14).²⁵ Although econophysicists and the Hayekian tradition agree on the fact that complexity of micro interactions generates an emergent order that is ‘more than the sum of its parts’, these two conceptual frameworks also have a different epistemological relationship with the notion of emergence: While econophysicists consider that this emergence can be captured through an asymptotic reasoning, the Hayekian tradition keeps the idea that this process cannot be analytically described. In other words, the first work with an epistemological emergence is in contrast with Hayek who rather works on an ontological emergence.²⁶

‘Free Will’ and Regulation

Even though econophysicists and Hayek describe the evolution of a complex system as the result of free will of micro elements (agents), they do not really use the concept of free will in the same way.

Models proposed by econophysicists offer conceptual tools to understand, and therefore, adjust the dynamics of these systems. In this context, the concept of ‘free will’ must be nuanced and defined as the liberty of agents in a given system whose rules can be adjusted by the understanding of these dynamics. In doing so, econophysicists use the notion of free will as a starting point (i.e., a methodological tool) for their reasoning. In contrast, Hayek promoted an ontological free will that will determine the dynamics of the complex system by letting emerge a spontaneous order, but whose characterization cannot be used as conceptual tools for the definition of a particular regulation. In other words, Hayek is an opponent to the regulation that could reduce the free will of actors, whereas econophysicists rather focus on ‘local free will’ contextualized in a particular system. However, the existence of this local free will of agents in a given system is not necessary in opposition with regulation for econophysicists. Although, the literature dealing with regulation and econophysics is still in its infancy, one can mention the existence of some works promoting the ability to identify a specific macro pattern in complex systems to regulate the functioning of economic\financial systems (McCauley, 2004). In the same vein, Artemi (2009), for example, considered the potential contributions of econophysics to political decision-making; Gonzalez, Hidalgo and Barabasi (2008) or Song, Qu, Blumm and Barabasi (2010) who applied the econophysics perspective to understand and regulate the human mobility patterns and, finally, Balcan et al. (2009) or Meloni, Arenas, & Moreno (2009) rather studied the diseases mobility.

Conclusion

Econophysics is a new field of research that generates a lot of methodological debates (Gallegati, Keen, Lux, & Ormerod, 2006; McCauley, 2006; Rosser, 2006). These debates also call for a re-evaluation of the link between economics and physics. Few epistemological links between the Hayekian tradition and econophysics have been emphasized in the literature.²⁷ The main contribution of this article is to propose this kind of link by providing a methodological parallel between these two areas of knowledge.

After having introduced the social Hayekian theory in which the notion of emergence plays a key role, I showed that the econophysical emergence generally characterized by the emergence of power laws could appear as an analytical formulation of the Hayekian perception of emergence. In addition to that the notion of free will is an important element in the reasoning proposed by both the approaches. In other words, these two frameworks appear, to some extent, complementary: Hayek is a social theorist of emergence by explaining why this phenomenon can appear in economic systems, whereas econophysicists develop models to explain how emergence works. However, beyond these conceptual similarities, the last section of this article also nuanced this rapprochement by emphasizing two differences between econophysicists and the Hayekian tradition: (a) the way of describing the emergent order; and (b) the use of the notion of free will. Although econophysics is often presented as a field arising from statistical physics, without links to economics, this article shows, paradoxically, that the emergence of econophysics echoes to old issues in economics that a very influent economist, Hayek, explored more than 50 years ago.

Footnotes

Notes

References

Adamic

, & Huberman

(2000). Power law distribution of the World Wide Web. Science, 287(1), 2115–2216.

Aielloet

, Chung

, & Lu

(2000). A random graph model for massive graphs. Proceedings of the 32nd Annual ACM Symposium on Theory of Computing, New York, 171–180.

Aimar

(2002). Commentaire sur economie et connaissance de F. Hayek. Cahiers d’Economie Politique, 2(43), 105–118.

Amaral

, Buldyrev

S.V.

, Havlin

, Leschhorn

, Maass

, Salinger

M. A.

, … Stanley

M.H.R.

(1997). Scaling behavior in economics: I. Empirical results for company growth. Journal de Physique I, 7(2), 621–633.

Artemi

(2009). From econophysics to politics (Working Paper). Scientific High School. Universita’ Per La Terza Eta, Orvieto, Italy.

Ausloos

(2014). Toward fits to scaling-like data, but with inflection points & generalized Lavalette function. Journal of Applied Quantitative Methods, 9(2), 1.

Axtell

(2001). Emergent cities: Microfoundations of Zipf’s law. Science, 293(1), 1818–1820.

Axtell

, & Florida

(2006). Emergent cities: Microfoundations of Zipf’s law (Working paper). Center of Social and Economic Dynamics, Washington, USA.

Bak

(1994). Introduction to self-criticality. In Cowan

, Pines

and Meltzer

(Eds), Complexity: Metaphors, Models, and Reality (476–482). Santa Fe: Santa Fe Institute.

10.

Bak

, Paczuski

, & Shubik

(1997). Price variations in a stock market with many agents. Physica A, 246(1), 430–453.

11.

Bak

, Tang

, & Wiesenfeld

(1987). Self-organized criticality: an explanation of 1/f noise. Physical Review Letters, 59 (4), 381–384.

12.

Balcan

, Colizza

, Gonc¸alves

, Hu

, Ramasco

J. J.

, & Vespignani

(2009). Multiscale mobility networks and the spatial spreading of infectious diseases. Proceedings of the National Academy of Sciences, USA, 106, 21484–21489.

13.

Ball

(2003). The Cambridge history of twentieth-century political thought. Cambridge: Cambridge University Press.

14.

Baronchelli

(2006). Statistical mechanics approach to language games (Ph.D Dissertation). Retrieved from Sapienza University of Rome.

15.

Batterman

(2002). The devil in the details: Asymptotic reasoning in explanation, reduction and emergence. Oxford: Oxford University Press.

16.

Bouchaud

J-P.

(2002). An introduction to statistical finance. Physica A, 313(3), 238–251.

17.

Bouchaud

J-P.

, & Potters

(2003). Theory of Financial Risk and Derivate Pricing. Cambridge: Cambridge University Press.

18.

Butos

(2003). Knowledge questions: Hayek, Keynes and beyond. The Review of Austrian Economics, 16, (4), 291–307.

19.

Caldwell

(2004). Hayek’s scientific subjectivism. In Wood

& Wood

(Eds), Friedrich A. Hayek: Critical assessments of leading economists (2nd series, vol. 1–4). London: Routledge.

20.

Chakraborti

, Muni Toke

, Patriarca

, & Abergel

(2011a). Econophysics review: Empirical I. facts. Quantitative Finance, 11(7), 991–1012.

21.

Chomsky

(1965). Aspects of the theory of syntax. Cambridge: MIT Press.

22.

Chomsky

. (1972). Language and mind. New York, NY: Harcourt Brace Jovanovich.

23.

Claes

, Frenken

, & Hellervik

(2006). A complex network approach to urban growth. Environment and Planning A, 38(10), 1941–1964.

24.

Colander

, Follmer

, Haas

, Godber

, Juselius

, Kirman

, … Slot

(2008). The financial crisis and the systemic failure of academic finance. 98th Dahlem Workshop on Modeling of Financial Markets, Dahlem, Germany.

25.

Craver

(2006). When models explain. Synthese, 153(2), 355–376.

26.

De Solla Price

D.J.

(1965). Networks of scientific papers. Science, 149(3683), 510–515.

27.

Dempster

(1999). Austrians and post keynesians: The questions of ignorance and uncertainty. The Quarterly Journal of Austrian Economics, 2(4), 73–81.

28.

Donangelo

, & Sneppen

(2000). Self-organization of value and demand. Physica A, 276(1), 572–580.

29.

Dostaler

(2001). La liberalisme d’Hayek. Paris: La Découverte.

30.

Eeckhout

Jan

. (2004). Gibrat’s law for (all) cities. American Economic Review, 94(1), 1429–1451.

31.

Farmer

J.D.

(1999). Physicists attempt to scale the ivory towers of finance. Computer Science & Engineering (IEEE), 1(6), 26–39.

32.

Farmer

, Shubik

, & Smith

(2005, September). Is economics the next physical science? Physics Today, p. 37–42.

33.

Gabaix

(1999). Zipf’s law for cities: An explanation. Quarterly Journal of Economics, 114(3), 739–767.

34.

Gabaix

(2009). Power laws in economics and finance. Annual Review of Economics, 1(1), 255–294.

35.

Gabaix

, Gopikrisnan

, Plerou

, & Eugene

S.H.

(2003). A theory of power law distributions in financial market fluctuations. Nature, 423(1), 267–270.

36.

Gallegati

, Keen

, Lux

, & Ormerod

(2006). Worrying trends in econophysics. Physica A, 370(1), 1–6.

37.

Gaus

G.F.

(2006). Hayek on the evolution of society and mind. In Feser

(Ed.), The Cambridge companion to Hayek (pp. 232–258). Cambridge: Cambridge University Press.

38.

Gong

, Minett

, & Wang

(2006). Language origin and the effect of individuals’ popularity. Proceedings of IEEE conference on Evolutionary Computation, Vancouver, Canada.

39.

Gonzalez

, Hidalgo

, & Barabasi

(2008). Understanding individual human mobility patterns. Nature, 453(3), 779–782.

40.

Gingras

, & Schinckus

(2012). The institutionalization of econophysics in the shadow of physics. Journal of the History of Economics Thought, 34(2), 109–130.

41.

Gray

(1982). Hayek and classical liberalism: A bibliographical essay. Literature of Liberty, 5(4), 19–101.

42.

Hackett

(1967). 70 years of best sellers, 1895–1965. New York, NY: Bowker Company.

43.

Hayek

([1948] 1980). The Use of Knowledge in Society. In Individualism and Economics Order. Chicago: The University of Chicago Press.

44.

Hayek

F.A.

(1929). Monetary theory and the trade cycle. London: Jonathan Cape.

45.

Hayek

F.A.

(1931). Prices and production. London: Routledge & Sons.

46.

Hayek

F.A.

(1937). Economics and knowledge. Economica, 4(1), 33–54.

47.

Hayek

F.A.

(1939). Profits, interest and investment. London: Routledge and Kegan Paul.

48.

Hayek

F.A.

(1942). Scientism and the study of society. Economica, 9(1), 267–291.

49.

Hayek

F.A.

(1944). The road to serfdom. London: Routledge.

50.

Hayek

F.A.

(1945). The use of knowledge in society. The American Economic Review, 35(4), 519–530.

51.

Hayek

F.A.

(1952). The sensory order: Inquiry into the foundations of theoretical psychology. Chicago: Chicago University Press.

52.

Hayek

F.A.

(1964). The theory of complex phenomena. In Bunge

(Ed.), The critical approach to science and philosophy (pp. 332–349). London: MacMillan.

53.

Hayek

F.A.

(1967). The constitution of a liberal state. Il Politico, 32(1), 455–461.

54.

Hayek

F.A.

(1973). Law, legislation and liberty, Volume1: Rules and order. Chicago: University of Chicago Press.

55.

Hayek

F.A.

(1974). The pretence of knowledge. American Economic Review, 79(1), 3–4.

56.

Hempel

(1965). Aspects of scientific explanations. In Hempel

, Aspects of scientific explanation and other essays in the philosophy of science (pp. 331–489). New York, NY: Free Press.

57.

Holland

(1998). Emergence: From chaos to order. New York, NY: Helix Books.

58.

Ingrao

, & Israel

(1990). The invisible hand: Economic equilibrium in the history of science. London: MIT Press.

59.

Jimenez

(2004). Econophysics: From game theory and information theory to quantum mechanics. Physica A, 348(3), 505–543.

60.

Jovanovic

, & Schinckus

(2013). The history of econophysics as a new approach in modern financial theory. History of Political Economy, 45(3), 443–474.

61.

Jovanovic

, & Schinckus

(2017). Econophysics and financial economics: An emerging dialogue. New York: Oxford University Press.

62.

Kim

(2006). Emergence: Core ideas and issues. Synthese, 151(2), 547–559.

63.

Kirman

(1992). Whom or What Does the Representative Individual Represent? Journal of Economic Perspectives 6(2), 117–136.

64.

Koppl

(2000). Policy implications of complexity: An Austrian perspective. In Colander

(Ed.), The complexity vision and the teaching of economics (pp. 97–117). Cheltenham: Edward Elgar.

65.

Kutner

, & Grech

(2008). Report on foundation and organization of econophysics graduate courses at faculty of physics of University Of Warsaw and department of physics and astronomy of the Wrocław University. Acta Physica Polonica A, 114(3), 637–647.

66.

Lavoie

(1989). Economic chaos or spontaneous order: Implications for political economy of the new view of science. Cato Journal, 8(3), 613–635.

67.

Légé

(2007). Hayek: Penseur génial ou incohérent? L’Economie Politique, 4(36), 46–59.

68.

Leroux

, & Nadeau

(2002). Ordre spontané et ordre organise chez Hayek. Cahiers d’Economie Politique, 2(43), 31–32.

69.

Levy

M.H.

, Haim

, & Sorin

(1995). Microscopic simulation of the stock market: The effect of microscopic diversity. Journal de Physique I, 5(8), 1087–1107.

70.

Levy

M.H.

, Haim

, & Sorin

(2000). Microscopic simulation of financial markets: From investor behaviour to market phenomena. San Diego: Academic Press.

71.

Loreto

, & Steels

(2007). Emergence of language. Nature, 3(2), 758–760.

72.

Luttmer

(2007). Selection, growth, and the size distribution of firms. Quarterly Journal of Economics, 122(3), 1103–1144.

73.

Lux

Th.

(2009). Applications of Statistical Physics to Finance and Economics. In Rosser

Barkley

(Ed.), Handbook of Research on Complexity (pp. 213–258). Cheltenham: Edward Elgar.

74.

Maas

Harro.

(2005). William Stanley Jevons and the making of modern economics. New York, NY and Cambridge: Cambridge University Press.

75.

Majorana

(1942). Il valore delle leggi statistiche nella physica e nelle scienze sociali [The value of Statistical Laws in Physics and in Social Sciences]. Scientia, 36(1), 55–66.

76.

Mandelbrot

(1963). The Variation of Certain Speculative Prices. Journal of Business, 36(4), 394–419.

77.

Mandelbrot

(1966). Forecasts of Future Prices, Unbiased Markets, and “Martingale” Models. Journal of Business, 39(1, Part 2), 242–255.

78.

Mandelbrot

(1997). Fractals and scaling laws in finance. New York, NY: Springer Verlag.

79.

Mantegna

(1991). Levy walks and enhanced diffusion in Milan stock exchange. Physica A, 179(1), 232–242.

80.

Mantegna

R.N.

, & Stanley

(1994). Stochastic process with ultraslow convergence to a Gaussian: The truncated lévy flight. Physical Review Letters, 73(1), 2946−2949.

81.

Mantegna

R.N.

, & Stanley

(1999). An introduction to econophysics. Cambridge: Cambridge University Press.

82.

McCauley

(2004). Dynamics of markets: Econophysics and finance. Cambridge: Cambrigde University Press.

83.

McCauley

(2006). Response to ‘worrying trends in econophysics. Physica A, 371(1), 601–609.

84.

Meloni

, Arenas

, & Moreno

(2009). Traffic-driven epidemic spreading in finite-size scale-free networks. Proceedings of the National Academy of Science, USA, 106, 16897–16902.

85.

Ménard

Claude

(1981). La machine et le coeur. Essai sur les analogies dans le raisonnement économique. In Lichnerowicz

, Perroux

& Gadoffre

(Eds), Analogie et connaissance. Paris: Éditions Maloine.

86.

Mirowski

(1989). More heat than light: Economics as social physics, physics as nature’s economics. New York: Cambridge University Press.

87.

Morgan

C. L.

(1923). Emergent evolution. London: Williams and Norgate.

88.

Morrison

(2015). Reconstructing reality. Oxford: Oxford University Press.

89.

Neukum

, & Ivanov

(1994). Crater size distributions and impact probabilities on Earth from lunar terrestial-planet, and asteroid cratering data. In Gehrels

(Ed.), Hazards due to comets and asteroids (pp. 359–416). Tucson: University of Arizona Press.

90.

Newman

M.E.

(2004). Power laws, Pareto distributions and Zipf’s law. Contemporary Physics, 46(5), 323–351.

91.

Pareto

(1897). Cours d’Économie Politique Professé a l’Université de Lausanne (Vol. I, Vol. II).

92.

Redner

(1998). How popular is your paper? An empirical study of the citation distribution. European Journal of Physics B, 4(3), 131–134.

93.

Rickles

(2008). Econophysics and the complexity of the financial markets. In Collier

& Hooker

(Eds), Handbook of the philosophy of science, Volume 10: Philosophy and complex systems. New York, NY: North Holland Elsevier Editions.

94.

Rosser

Jr. (1999). On the complexities of complex economic dynamics. Journal of Economic Perspectives, 13(4), 169–192.

95.

Rosser

Jr. (2003). Complexity in economics. New York, NY: Edward Elgar.

96.

Rosser

(2006). The nature and future of econophysics. In Chatterjee

Arnab

and Chakrabarti

Bikas K.

(Eds), Econophysics of Stock and other Markets (pp. 225–234). Milan: Springer.

97.

Salmon

(1984). Scientific explanation and the causal structure of the world. Princeton: Princeton University Press.

98.

Sawyer

(2005). Social emergence. Cambridge: Cambridge University Press.

99.

Schabas

(1990). A world ruled by number: William Stanley Jevons and the rise of mathematical economics. Princeton, NJ: Princeton University Press.

100.

Schinckus

(2009). Economic uncertainty and econophysics. Physica A, 388(20), 4414–4423.

101.

Schinckus

(2010). Econophysics and economics: Sister disciplines? American Journal of Physics, 78(4), 325–327.

102.

Schinckus

(2013a), How do econophysicists make stable levy processes physically plausible. Brazilian Journal of Physics, 43(4), 281–293.

103.

Schinckus

(2013b). Between complexity of modelling and modelling of complexity: An essay on econophysics. Physica A, 392(1), 3654–3665.

104.

Schinckus

(2013c). Econophysics, a new step in the evolution of physical sciences. Contemporary Physics, 54(1), 17–32.

105.

Schinckus

(2017). Econophysics and Complexity Studies. Ph.D. Dissertation, Universiy of Cambridge.

106.

Shinohara

, & Gunji

(2001). Emergence and collapse of money through reciprocity. Applied Mathematics and Computation, 117(1), 131–150.

107.

Soares

, Corso

, & Luneca. (2005). The network of syllables in Portuguese. Physica A, 335(3), 678–684.

108.

Song

, Qu

, Blumm

, & Barabasi

(2010). Limits of predictability in human mobility. Science, 327(2), 1018–1021.

109.

Stanley

H. E.

, Afanasyev

, Nunes

Amaral. L.

, Buldyrev

, Goldberger

, Havlin

, Leschhorn

, Mass

, Mantegna

, Peng

C-K.

, Prince

, Salinger

, & Viswanathan

(1996). Anomalous fluctuations in the dynamics of complex systems: From DNA and physiology to econophysics. Physica A, 224(1), 302–321.

110.

Stanley

, Gabaix

, Gopikrishnan

, & Plerou

(2007). Economic fluctuations and statistical physics: Quantifying extremely rare and less rare events in finance. Physica A, 382(1), 286–301.

111.

Tuerck

(1995). Economics as mechanism: The mind as machine in Hayek’s sensory order. Constitutional Political Economy, 6(3), 281–292.

112.

Vriend

N. J.

(2002). Was Hayek an ace?. Southern Economic Journal, 68(4), 811–840.

113.

Wichmann

(2005). On the power-law distribution of language family. Journal of Linguistics, 41(2), 117–131.

114.

Yamamoto

, & Miyazima

(2004). Power-law behaviour in social and economical phenomena. Physica A, 344(1), 757–763.

115.

Yasutomi

(1995). The emergence and collapse of money. Physica D, 82(1), 180–194.