What is complexity economics, why is it heterodox, and what are its policy implications?

Please cite the paper as:
Wolfram Elsner, (2017), What is complexity economics, why is it heterodox, and what are its policy implications?, World Economics Association (WEA) Conferences, No. 2 2017, Economic Philosophy, Complexities in Economics


Complexity economics has developed into a powerful empirical, theoretical, and computational research program in the last three decades, advancing more realistic economics. It converges with long-standing heterodox schools, and its theoretical and empirical findings are consistent with older heterodox research interests and predictions. Economic complexity is characterized by path-dependence, idiosyncrasies, some self- organization capacity, structural emergence, and certain statistical distributions in economic topologies and motions, as complex economic systems move between building order and phases of sudden disorder. In agent-based systems, underlying ―intentionality of agents includes improving their performance, reducing perceived complexity, and generating social institutions. Boosted by the financial crisis 2008ff., a surge to explore complexity-economics’ policy implications has emerged. This chapter will briefly review the literature on economic
Complex Adaptive Systems (CAS) and derive implications for economic-policy interventions and the state to act upon socio-economic complexity. From an ―evolution-of-cooperation perspective, we exemplarily derive some more specific policy orientations, specified ―framework-policy or ―interactive-policy approaches, embedded in a conception that we call ―new meritorics. We consider some required structures and capacities that a modern effective state, capable of a strong and persistent, but learning and adapting ―complexity policy, should have.

Keywords: , , , , ,

Recent comments


7 comment

  • Dr Dhiresh Kulshrestha Associate Professor Economics - Mody University of Science and Technology- Laxmangarh -Sikar ( Rajasthan)- India says:

    Complexity economics has played a vital role as a powerful tools. Theoretical, it is right concept regarding the economic policies and its implications. Policies are good and based on economic development of the nation, but its implication is not good. If I take example Optimization theory in economics – It is a technique to finding the best results under the given conditions. Conceptually, it is good but implications should not be based on Social Cost Like water contamination, Air pollution and global warming. It is not the economics principle, the concept of the business to earn the maximum profit at minimum cost, but not on the cost of social, health and society. In the long run, The so-called development will convert into the negative sense and create the more health issues in the society. It is not heterodox, it is the right implications of policy. In India’ capital and China is biggest example those are fighting with the air pollution from the last decade, and day after day the problems are increasing and negative impact on economy. my deep congratulation to Prof. Wolfram Elsner for the contribution with the help of this paper as “Revising the concept of regulation”. Once again congratulations Prof Wolfram Elsner

  • wolfram elsner says:

    Thanks for the comment, Dhiresh.

    The divorce between individual and collective rationalities and the phenomenon of increasing “social costs” are clear indications of the fact that the reductionist understanding of (socio-) economies by the neoclassical economic “mainstream” is too simplistic and, with its focus on unique static equilibria (even if equilibrium paths), increasingly misses relevant parts of the socio-economic and natural-environmental realities and developments.

    Policies based on that philosophical (or better: ideological) basis — caught in and obedient of its axioms, assumptions, mechanisms, reduced price-quantities world, allowed or tabooed themes and policy measures, following the totem of “perfect”competition (the “totem of the x”, to cite the famous Leijonhufvud paper of the 1970s), but in fact following the real interests of the ruling oligarchy — will never be able to deal with the most basic problems of humankind, but rather, will restrict the allowed space even tighter and worsen the problems.

    The issue you are touching upon, Dhiresh, is, whether a “complexity policy” is “heterodox ” or not. You are right that we should be interested in, and chasing after, not primarily “heterodoxies” vs “orthodoxies”, but what is “real”, real-world, “right” or “wrong”, what are the “right” things to do. But being after the “right” policy is not at all in contradiction to being “heterodox”.

    My criterion of what is the “right” thing to do, would be the Veblenian criterion of problem-solving, still to be operationalized in a thousand fields and new situations continuingly coming up. A big task will always be to operationalize what societal problem-solving concretely means. And I like to argue with mainstreamers about “right” or “wrong” policies, problem solving or ceremonial measures etc., in the context of complexity economics and complexity policies.

    And yet, we are in a world where we need to say that such search for “right” solutions, “real-world” perspectives, and problem-solving policies in fact is infeasible by sticking to the interpretation patterns of the neoclassically inspired, and often neoliberal mainstream. Not the least, this has the important implication that we can refer to the abundant history of economic thought (HET), that has been devaluated and tabooed by the mainstream, from the classics to the rich “heterodox” traditions and current plural heterodox perspectives.

    this was one of the main impetuses of my paper, I would say …

    Thanks again for the inspiring comment.


  • Ping Chen, Professor at Peking University & Research Fellow at Fudan University, China says:

    Eisner asked a critical question: why complexity economics is considered as heterodox by mainstream economics? My answer would be two possibilities. First, its policy implication may against the orthodox ideology, such as the efficient (self-stabilizing) market and laissez- fair policy.
    Second, its mathematical novelty shocks the standard of mainstream publications, since the conventional editors are hard to evaluate competing articles with non-familiar mathematics. Exclusion of these works as heterodox economics justified that the mainstream economics
    is more like “theoretical theology” (Duncan Foley) rather than “empirical science.
    I should extend Eisner’s question by classifying which complexity theory can be integrated into orthodox economics, but which self-organization model likely belong to heterodox economics.
    Take the case of POWER LAW. If you discover financial fluctuations can be described by power law, and government can do nothing about it. I believe Fama would easily accepted it, just like he accepted Levy Distribution in his 1970 paper on efficient market. However, some author here argued that firm size follows a power law, which has important implications in anti-trust law. Some Chicago economists may strongly question its relevance to economic policy.
    I have some experiences in selective implications of complexity models. Financial economists seem easily embrace Mandelbrot theory of fractal Brownian motion in financial market. But macro economists strongly oppose the existence of economic chaos. When I discovered empirical and theoretical evidence of economic chaos from monetary indexes, I was fiercely opposed by Chicago school but welcomed by Austrian school, Why? Because Friedman believed that monetary movements were exogenous so that central bank has independent power to interfere monetary market. Friedman even claimed that a strong Fed leader could prevent the Great Depression in his history of monetary policy. In contrast, Hayek believed that monetary movements were endogenous and central bank may enhance business cycles. The evidence of monetary chaos clearly supports Hayek theory of endogenous money, but rejects Friedman theory of exogenous money. The historical lesson of 2008 financial crisis indicated that Hayek was right but Friedman was wrong, since China did not suffer much from 2008 crisis even though its central bank is not independent like western banks. See. Ping Chen, “Empirical and Theoretical Evidence of Economic Chaos,” System Dynamics Review, 4, 1-2, 1-38 (1988); Ping Chen, Economic Complexity and Equilibrium Illusion, Routledge (2010).
    In this sense, CAS approach did demonstrate rich features in collective behavior, but not yet challenge fundamental policies in mainstream economics. The reason is simple. Both system dynamics and agent-based models have many equations and many variables, just like large econometric models. You can easily change policy implications by adjust any parameter or adding new variables. Therefore, they are useful tools as educated guess or teaching economics, but not powerful weapon in policy-making. If Santa Fe Institute and their co-workers could simulate different scenario of monetary policy and compare their simulation with real cases in Central banks at US, Euro, or Japan, I would expect much stronger policy impact of CAS approach to mainstream economics.
    In short, the road from complexity economics to mainstream economics is DIRECT FACING Fundamental Issues of OUR TIME, such as financial crisis, transition depression, climate change, the origin of poverty and war. Otherwise, COMPLEXITY ECONOMICS so far only made small progress in mathematical modeling, but its policy impact is yet to beyond the scope of old EVOLUTIONARY ECONOMICS or SYSTEM DYNAMICS. We should try harder than post-Keynesian economics.

  • Malgorzata Dereniowska says:

    Dear Wolfram,

    I read your paper with great interest. In your discussion of heterodox, or rather should we say “real-world” complexity economics, you raise a number of issues that are pregnant topics of interests for both economists and philosophers. I appreciate your perspectives-taking and problem solving orientation, and enlivening the debates from the history of economic thought in the context of demanding contemporary challenges.

    Two questions:
    1. You talk about collective rationality in several places – even “superior collective rationality”. This notion raises many philosophical questions, e.g., about the ontological status and properties of collective entities and agency. These matters in turn have implications to thinking complexities and institutional change. I am curious what does collective rationality mean for you?
    2. You point to the importance of pragmatic approach through continuous inquiry, learning, adaptation, and participation. You also emphasize the role of strong state, normative criteria, and a clear determination of public goods and public objectives. In this context, what would be the (scope of) proper legitimate public processes?

    A comment: there is a sparkling resemblance between the debate in complexity economics about equilibrium theory, and the debate in ecology and ecosystem theory about equilibrium and disequilibrium paradigms. Ecology for a long time characterized natural systems through categories of integrity and stability. The main premise of equilibrium ecology was that ecosystems tend to equilibrium state. Furthermore, it was assumed that the equilibrium point is stable, i.e. after disruption the system returns to its previous state. These assumptions had bearing on the understanding of organization and dynamics of ecosystems, implying the unchangeability of their structure and functions over time, or that natural systems organization and dynamics are somewhat independent from history. This approach did not consider the irreversibility of effects (such as changes in the structural composition of the ecosystem after disruptions, species loss, etc.). Many concepts used in conservation biology and sustainable development—such as ecosystem health, ecosystem integrity – are based on the assumption of equilibrium state. Around the 70’ a paradigm shift began, as many ecologists have noted that disturbance is more of a norm, and that tending to climax (a mature, stable state) is not so typical to ecosystems. Although it cannot be said that the non-equilibrium paradigm (disequilibrium ecology, ecology of instability, chaos theory, etc.) is a dominant paradigm, certainly it has initiated serious philosophical, theoretical and methodological debates—still disputed and of course much more complex than the picture painted here with the wide brush. One of the fascinating aspects of these debates is their direct relevance to environmental policy and the fact that both stability and instability models in ecological theory co-shape the (divergent) meaning ascribed to environment-related concepts, such as ecosystems, and impact the policy approaches.
    I find these parallel debates on complexity in economics and environmental sciences emblematic of the collective patterns of thinking and underlying philosophies of science, which is also a reminder of the importance of looking at the history of the concepts that we use in our discourses (especially as they have policy relevance), and drawing lessons through interdisciplinary exchange.

  • wolfram elsner says:

    Comment on Ping Chen (comment on Marlgorzata Dereniowska later):

    Thanks, Ping, for your detailed comment, to which I largely agree. My (minor) points on this are these:

    — First, my intention is to have CE considered heterodox by HETERODOXIES themselves rather than by the mainstream. In order not to have CE captured by the minstream but “engrossed” by heterodoxies rather.

    — Second, you are putting a most crucial question, which I did address only in indirect ways (struggling to show, HOW CE may be heterodox against the HET perspective): You are addressing the formal components that make it easy, or not, for the mainstream to reinterpret CE into their equilibrium/optimality frame. your case of right-skewed distributions of degree centrality is a good case: mainstreamers would interpret them as “correct” and “optimal” “market outcomes”, while we would argue that they have severe impacts on system stability and crises, calling for a renewed anti-trust policy (e.g., in the financial sector), among other things. Your financial-sector and monetary-policy example is crucial here, and most convincing.

    — Third, the core of your formal argument is something I missed indeed: system dynamics and ABM have many variables and equations that are not selective per se in terms of policy orientations following — so that mainstreamers may easily select and shape their “complex” models and their simulations into more harmless general-equilibrium directions.

    — Fourth, yes again, CE did not invent a new mathematics, and we need to have clear real-world orientations and broad approaches in modeling and model simulations in order to derive substantial, substanially new, and distinctive policy orientations for the future.

    Thanks for those clarifying amendments to my paper.

    Cheers over to the land of the middle,

    • Ping Chen says:

      Eisner, historically, it is the fundamental character of the underlying system that select proper mathematic tools. For example, it is the physics observation in cosmology that select the specific type of non-Euclidean geometry. I think you have better understanding of fundamental issues in economics, so that you can have stronger voice in judging and selecting mathematical tools in complexity studies for addressing fundamental issues in evolutionary/institutional economics and macro policies.
      Best wishes,

      Ping Chen

  • wolfram elsner says:

    Hello Gosia :-),

    thanks for comments and questions. Very insightful parallel that you are drawing to earlier ecological science. you are raising three major issues:

    — First, the isssue of the “collectivity” and superiority of rationality. I have been arguing in the frame of games, and particularly of ubiquitous social dilemmas (SD). This formal frame allows us to clearly distinguishing a common and a collective problem, and respective solutions. A coordination game, for instance, requires for its evolutionary solution just “parallel” behavior, and it is in the immediate individual(istic) interest to be coordinated with the other(s), although it cannot be determined in advance, in which of the (often many) attractors the system will settle (on driving on the left or right side of the street, contributing to a team product or not, etc.). The problem is just a “common” one and so is the solution a “common” one, which just requires adhering to a social rule (which, as said, is in everyone’s immediate interest). Coordination probelm — coordination as solution — social rule for coordination — a common problem and a common solution. The SD is different, as the evolutionary solution across repeated SD requires an active sacrifice of the potential one-shot maximum. The solution among short-run hyper-“rational” (individualistic) agents thus requires more than just a social rule, as the adherence to it is NOT in everyone’s immediate indicidualistic interest (but the very deviation from it is). Thus a “collective” problem. (Equally clever agents, though, will not let themselves be exploited, and together run into the one-shot Nash euquilibrium, thus will not get their maximum, which later may serve as an evolutionary motivation for agents to learn and habituate cooperation.) A social institution thus is a rule plus an (at least endogenous) sanction mechanism (or some external punishment mechanism on top), namely the credible threat of a later retaliation and thus of joint worsening of payoffs. Note that social learning of it requires some “semi-conscious” habituation, as the social institution cannot be acquired hyper-“rationally”. The simple evolution-of-cooperation single-shot solution I used shows that the more intricate problem of an SD requires the acquisition (social learning, and in fact coevolution) of a longer time horizon, a longer-term “culturally” acquired rationality, which only makes the learning and habituation of the social institution of cooperation feasible. This is why I call the broader rationality “collective”, and as it is problem-solving for a far more “wicked” problem than a more simple coordination problem, I call it superior. SD, a wicked problem — learned and habituated cooperation as solution — social institution — a collective probelm and collective solution. For more, see my paper on “the theory of institutional change revisited”, JEI 1-2012, 1-44.

    — Second, scope of proper legitimate public process, a weaker point in my paper, indeed. You have seen the broader socio-economic issues that I think emerge from CAS as a new policy paradigm, the deep structural measures to be taken in the fields of incentive structures, culturally acquired time horizons, mobility, arena and network structures etc. At the systemic level, the obvious awareness of the public agent to (much more intelligently!) intervene into defined framework structures, permanently and sometimes massively, which justifies the ideas of a “strong” state, understood in new ways. All that also substantiates the idea of a broader and more inclusive (“participatory”) policy process, very much in the sense of the pragmatistic policy conception of the 1930 and 40s. But I would be happy, if some philosopher or political science people or both (someone like you :-)) would take up this loose end of my paper and develop a policy conception (and a conception of politics and the state) for such a complexity policy paradigm …

    — Third, your comment on the earlier “equilibrium ecology” is most interesting as well. Sure, if you leave the entire environment of the spaceship earth (as related to the energy inflow from the sun) alone, it would always remain in its “equilibrium” somehow, as it would adapt to all changes, and all change would come “endogenously” (excluding a clash with some comet). This seems to be trivial and we would have no criteria to say, whether really it is in equilibrium or not, where it would head to, and whether this then would be an equilibrium path or not. So the older perspective seems to have been naive. Whether we consider the activities of recent “homo sapiens sapiens” and its huge “anthropocenic” and destructive changes on the constitution of the natrual environment as “exogenous”, is another “philosphical” question. But independent of endogenous or exogenous impacts, you are right, we cannot really claim that the complex natural system is in any “equilibrium” (unless we consider claim it for the most abstract level of its entirety, as indicated above). As we can see (and as we would expect from a CAS perspective) all kinds of irreversibilities of processes of, say, a reduction of genetic diversity, occur. As you say, this is another point, where CE economists and philosophers of science would have common things to develop. In the end, I do not see the natural CAS in any state of “integrity”, “equilibrium” or “stability”, particularly not under the impact of homo sapiens sapiens and of his/her flawed current socio-economic system. So the policy issue is as urgent for the natural environemntal CAS as it is for the socio-economic CAS — as you indicate. Obviously a common task for complexity philospohers, complexity economists, and other complexity sciences.