In 1898, Thorstein Veblen posed a fundamental question: "Why is economics not an evolutionary science?" Following dramatic advances in evolutionary biology in the latter half of the nineteenth century, this was a question that a number of economists were beginning to ask, including Alfred Marshall, one of the founding fathers of neoclassical economics. (1) While Veblen's insights led to the emergence of one of the two main strands of American institutional economics in the twentieth century, Joseph Schumpeter (1950) offered an "evolutionary economics" which had at its core entrepreneurship and innovation, both technological and organizational. However, despite the considerable influence of American institutionalism in the early decades of the twentieth century and the high profile of Schumpeter's writings, economics never became an "evolutionary science."
The objective of this article is to argue that economics will have to become a "complex systems science" before the valuable insights of institutionalist and evolutionary economists can be incorporated into the mainstream. First, the complex adaptive economic system, as defined in John Foster (2005), is compared with the view of a system that is embodied in standard economic theory. Second, this difference is illustrated through an examination of the role of the familiar production function construct in economics. As Zvi Griliches and Jacques Mairesse (1995) stressed, the production function began its life as a macroeconomic construct in the work of, most notably, Charles Cobb and Paul Douglas (1928). With this in mind, the macroeconomics of John Maynard Keynes is then revisited to examine the extent to which his non-neoclassical approach contains features that we might expect to find in modern complex economic systems theory. As is well known, modern macroeconomics with its neoclassical micro-foundations has largely set aside the revolutionary insights of Keynes (1936). So it is instructive to see if these insights can be placed within a complex adaptive system perspective, raising both the possibility of a new macroeconomics and of new links with institutionalist and evolutionary economics, as foreshadowed in Foster (1987).
The Resurgence of Evolutionary Thinking in Economics
In the post WWII period, evolutionary economic thought gradually became confined to the margins of economics as neoclassical theory began to form the analytical core of the modern discipline. However, in the early 1980s, there was something of a resurgence of interest in evolutionary perspectives on economics, following the publication of prominent books by Kenneth Boulding (1981) and Richard Nelson and Sidney Winter (1982). What these contributions gave rise to was a coherent "neo-Schumpeterian" approach to evolutionary economics that grew out of studies of innovation, technological change and the behavior of firms and centered on the analytics of "replicator dynamics" in the presence of variety (or heterogeneity). (2) This dynamic mathematical representation of competitive selection was borrowed directly from neo-Darwinian evolutionary biology (Fisher 1930) and suitably adapted for economic application settings. (3)
Although aspects of this new evolutionary economics could be represented in formal mathematics and closely connected to vast and informative empirical literature on entrepreneurial and innovation processes in a range of industrial settings, it never succeeded in penetrating the mainstream of the economics discipline to any significant extent. Despite this, lip service was sometimes paid to "Schumpeterian" ideas through modifications of conventional economic models, particularly in the field of economic growth. For example, Philippe Aghion and Peter Howitt (1998) provided an extended "Schumpeterian" model of "endogenous growth" building on the neoclassical growth model. Replicator dynamics are absent for reasons that seem clear: the conventional economic paradigm cannot easily accommodate such dynamics and, in any event, the timeless construction of conventional economic theory renders it malleable enough to provide a theoretical explanation of almost any economic behavior that we might observe.
So, although neo-Schumpeterian evolutionary economics has become a thriving heterodoxy in economics, as evidenced by the powerful contributions contained in Kurt Dopfer (2005), its proponents have not been able to change how economists think, except in marginal ways. This is because, as a coherent "system of thought," neo-Schumpeterian evolutionary economics has not been able to compete with what is a very powerful conventional paradigm, particularly since the latter became expressed in game theoretic (i.e., strategic) terms. But another reason why acceptance has not been widespread lies in the fact that many economists have sensed that, although the replicator dynamic model works well in many biological settings, it may not be an adequate enough analytical tool to deal with all of the behavioral complexities that are found in economic systems. Even in biology, the selection processes that lie behind such models are now seen by many researchers as only part of an evolutionary story in which endogenous "self-organization" is at least as important (Depew and Weber 1995).
Self-organization has its origins in the study of complex physio-chemical systems. (4) These are viewed as "dissipative structures," importing "free" energy and exporting entropy and, in so doing, maintaining structure that is kept away from thermodynamic equilibrium. Such systems are complicated but, at the same time, exhibit order. Daniel Brooks and E.O. Wiley (1986) were two of the first biologists to argue that the evolution of living systems could also be viewed as self-organizing if their capacity to absorb and utilize information is taken into account. Foster (1997) and Ulrich Witt (1997) went on to consider what might be added in order to capture the process of self-organization in economic settings. However, it became apparent that the self-organization approach to economic evolution is as "one-sided" as the earlier approach that relied only on the operation of selection mechanisms--both needed to be set in a more general analytical framework.
The complex adaptive system perspective that has emerged over the past decade looks below the processes of self-organization and selection that we observe to discover the generic rules that give rise to such processes. These rules form structures that are networks of connections between elements that are ordered in particular ways. All such structures are dissipative and, thus, have to obey a set of physical rules but economic structures involve important additional rules (Potts 2000; McKelvey 2002). Complex systems science involves the examination of the connective properties that exist between the elements of a system under investigation to discover how particular sets of rules give rise to ordered complexity. To this end, a range of analytical techniques can be employed (e.g., graph theory, network analysis, power laws, diffusion models, agent-based simulation and, of course, replicator dynamics). (5) Depending on the kind of system that we are dealing with, the order we observe can range from deterministic chaos through to the purposeful creation and destruction of connections in complex adaptive systems. (6) Gradually, an analytical toolkit has been developed that can be used in a range of different contexts, both to understand the historical dynamics of complex systems and in some cases, to predict their likely time paths. (7)
Although there have been significant advances in the application of complex systems science in evolutionary biology, (8) psychology, sociology (9) and management, (10) there has been relatively little progress in economics (Markose 2005). This is largely because modern neoclassical economic principles, being strictly timeless, offer a vast array of sophisticated analytical tools that can be employed to mimic virtually all aspects of observed economic behavior. It is easy for economists to build models on the strong premise that human beings are cognitively sophisticated (or "rational") and, thus, there seems to be little need to think of economic systems in new ways. This is practically the same justification that has always been offered by economists when choosing to ignore the important insights of institutional economists. Not surprisingly, it is comparatively straightforward to identify connections between institutional economics and the complex system perspective on economics (Potts 2000).
Simplistic Perspectives on Economic Systems
The public and policymakers seem to be much less enthusiastic about "modern" economics than they were in the days when it was a less sophisticated (i.e., a less mathematical) discipline. (11) Some economists will tell you that this is because economics has become more about the operation of "free markets." For them, this loss of popularity is not because of the presence of complicated mathematics but because hard-nosed neoclassical economic doctrine has never been palatable to those wishing to exploit or "free ride" on others. It is true that, at least since the time of Adam Smith, economic doctrine based on the promotion of self-interest and free markets has not always been popular among those enjoying economic rents on the political right or those subject to exploitation on the left, irrespective of whether it is presented in words or equations. Such a defense implies that we are dealing with an ideology rather than a science; or at least, what is a science is being viewed from an ideological perspective by some observers. (12) From such a perspective, the beauty of complicated mathematics is that it appears to the public as a product of scientific discourse. In television and film, economists invariably appear in front of a black or white board on which are written sets of equations. But...