Complexity and the Economy Read online

  [ xviii ] Preface

  either of us to design an 8-bit adder we’d have to bone up on digital electronics and do this from scratch. Yet we had designed an algorithm that could design

  such circuits automatically by evolution. I found the idea of this remarkable, and of the papers assembled here this is one I am greatly taken by. It demonstrated evolution in action, and evolution by a different mechanism—by

  combination, or successive integration.

  Somehow, I thought, all this had to fit with how an economy evolves, indeed

  how an economy forms in this first place. As I worked on technology, I real-

  ized that while the economy creates technology, more important, technology

  (the collective of technologies we use to meet our human needs) creates the

  economy. So the economy is not just a container for its technologies, it is an expression of them. As these technologies changed, and as whole new bodies of technology entered, the economy changed. It changed in what it did and how

  it did it, and it changed in the arrangements and institutions that fitted to the new ways of doing things. The economy, in other words, changed in structure.

  I wrote all of these findings up in a book, The Nature of Technology: What It Is and How It Evolves, that appeared in 2009. It was well received, particularly by professional engineers, and has gone into several languages. Some of the

  papers collected here were way stations on the path to this book and one was

  directly part of it. This work on technology took me 12 years from inception to completion, and I found it fascinating. Of particular wonder were the mechanisms by which the collective of technology evolved, and the realization that

  technology is a thing with considerable logical structure. Technology, I believe, studied in itself, is every bit as complicated and structured as the economy, or the legal system. And it is an object of considerable beauty.

  The various lines of research that have made up this intellectual journey

  seemed to me at the time disparate and unconnected. But if I look back on

  them now, and on the work of other colleagues at Santa Fe and elsewhere,

  I see that what was forming from all this slowly and gradually was an approach to economics. I’d summed up my earlier understanding in a 1999 article in

  Science,6 and the editor insisted I give this different approach a name. I called it “complexity economics.” Looking back now, the features of complexity economics are clear. The economy is not necessarily in equilibrium; in fact it is usually in nonequilibrium. Agents are not all knowing and perfectly rational;

  they must make sense of the situations they are in and explore strategies as

  they do this. The economy is not given, not a simple container of its technologies; it forms from them and changes in structure as this happens. In this way the economy is organic, one layer forms on top of the previous ones; it is ever changing, it shows perpetual novelty; and structures within it appear, persist 6. W. B. Arthur, “Complexity and the Economy,” Science, April 2, 1999, 284, 107–109.

  Pr eface [ xix ]

  for a while, and melt back into it again. All this is not just a more poetic, humanistic view of the economy. It can be rigorously defined, and precisely

  probed and analyzed.

  I’m often asked how this new approach fits with standard economics. Isn’t

  it simply a variation of standard economics? And won’t it be absorbed seam-

  lessly into—“bolted on” to (in economist Richard Bronk’s phrase)—the neo-

  classical framework? My answer on both counts is no. This different approach

  is not just the use of computers to do agent-based modeling, nor of adding a

  deeper understanding of technology change to endogenous growth models. It

  is economics done differently, economics based on different concerns—par-

  ticularly on how nonequilibrium works—an economics where the problems

  are different and the very idea of a solution is also different.

  One way to see this is to recognize that standard neoclassical economics

  comes out of a particular way of looking at the world. Neoclassical economics

  inherited the Enlightenment view that behind the seeming disorder of the

  world lay Order and Reason and Perfection. And it inherited much from the

  physics of the late 1800s, in particular the idea that large numbers of interacting identical elements could be analyzed collectively via simple mathemati-

  cal equations. By the mid-1900s this led in turn to a hope that the core of

  economic theory could be captured in simple mathematically expressed prin-

  ciples and thereby axiomatized. Some parts, such as macroeconomics or the

  theory of institutions, might have to be left out, but the core of the field could be ordered and tamed, and reduced to mathematics.

  That program was at best only partially successful. It certainly cleaned

  up much of the sloppy logic that had passed as theory before, and led to a

  fresh respect for the workings of markets and for the inherent advantages of

  the capitalist system. But it also, I believe, led to a stiffness in thinking, to a righteousness in what was permitted as economic theory and what was not,

  and to a closedness to other ideas. Shut out were the effects on the economy

  of politics, of power, of class, of society, of fundamental uncertainty, and of formation and creation and development. In the end it could be argued that

  the program—at least the extreme hyper-rational version of it—failed. If it

  needed Popperian testing, its ideas were falsified spectacularly in 2008 and

  the aftermath of the financial meltdown. Nobody could claim that the market

  had lost half of its worth in a short time because companies had suddenly

  lost half of their usefulness; the companies were much as before. And nobody

  could claim either that unemployment rates of 20% and upward in some of

  the European economies were due to the suddenly changed preferences of the

  labor force; people wanted jobs just as before. In 2009 the Economist magazine noted wryly that Wall Street was not the only victim of the financial crash,

  standard neoclassical economics had collapsed along with it.

  On reflection, it shouldn’t be surprising that this highly purified form of

  economic thinking ran into difficulties. One lesson Western thought has had

  [ xx ] Preface

  to learn slowly in modern times is that if we try hard enough to reduce anything to pure logic—for example if we try to pin down a final meaning of such

  concepts as Truth, or Being, or Life, or if we try to reduce some field such as philosophy or mathematics (or economics for that matter) to a narrow set of

  axioms—such attempts founder. The world cannot be reduced to pure logic

  and caged within it. Sooner or later it slips out to reveal its true messiness, and all such projects fail.

  Slowly replacing the pure order of neoclassical economics is a new respect

  for reality, shared by many researchers in economics. Behavioral economics

  is one such approach being pressed forward; the psychology of markets is

  another. So too are theories of development that rely increasingly on under-

  standing institutions and the workings of technology. And so too is the

  approach offered here which now has very many practitioners besides our ini-

  tial group at Santa Fe.

  One of things that has surprised me, and pleased me enormously, was

  that many of the “modern” themes in this approach fit well with ideas in

  Schumpeter, and Smith, and Mill, and Marx, and Keynes, and with the ideas of

  the institutionalists and political economists that followed. They too saw the e
conomy as emerging from its technologies, as changing structurally, as not

  necessarily being in equilibrium, and with its decision-makers facing funda-

  mental uncertainty. These connections have not yet been formally made; they

  are more like threads of thought that link these new ideas with ones discussed in the past. But they do show economics rediscovering some of what it lost.

  We are beginning to have a theoretical picture of the economy in formation

  and in nonequilibrium.

  The papers collected here were written from when I first went to Santa Fe

  in 1987 until the present day. There is some overlap among them; but that is

  inevitable. Several of the papers were written to make the main ideas available to wider audiences; and also to explore these ideas from different viewpoints.

  The papers build on the work of many other people in economics, complexity,

  and other fields, in particular my Santa Fe colleagues John Holland, Stuart

  Kauffman, David Lane, and Richard Palmer. And they build also on the work

  of people not closely connected with our original Santa Fe group: in particular, Peter Allen, Rob Axtell, Josh Epstein, Alan Kirman, and Lee Tesfatsion, who

  all have contributed to this new approach. The papers here owe a great deal

  to the neoclassical formulation—that’s after all what I was trained in. Many

  are full-blown analyses, others are essays. They are arranged more or less by

  theme rather than by when they were written, but the research papers are

  mostly in the first half of the book and the essays in the second. They can be read in any order and I encourage readers to follow whatever sequence appeals

  to them—and certainly to reach out to the work of others in this area.

  Taken together, a theme or framework for thinking emerges from the

  papers here. In the place of agents in well-defined problems with well-defined Pr eface [ xxi ]

  probabilistic outcomes using perfect deductive reasoning and thereby arriving at an equilibrium, we have agents who must make sense out of the situation

  they face, who need to explore choices using whatever reasoning is at hand,

  and who live with and must adjust to an outcome that their very adjustments

  may cause perpetually to change.

  In 1996 the historian of economic thought David Colander put forward

  an allegory in which economists a century ago stood at the base of two

  mountains whose peaks were hidden in the clouds. They wanted to climb the

  higher peak and had to choose one of the two. They chose the mountain of

  well-definedness and mathematical order, only to see when they had worked

  their way up and finally got above the clouds that the other mountain, the one of process and organicism, was far higher.

  Many economists have started to climb that other mountain in the last few

  years. I will be interested to see what we will find along the way.

  W. Brian Arthur

  Palo Alto, California

  January 2014

  [ xxii ] Preface

  ACKNOWLEDGMENTS

  Many people have heavily influenced the ideas here. I thank in particular

  John Holland, Stuart Kauffman, David Lane, and Richard Palmer—the core

  of our team on the Economy as an Evolving Complex System Program in

  1988–1990 at the Santa Fe Institute—from whom I learned so much and with

  whom I greatly enjoyed myself. I have also learned much over the years from a

  wide cluster of practitioners in several fields: complexity, economics, biology, physics, mathematics, and computation. It’s impossible to mention them all,

  but to single out a few, I’d like to thank Peter Allen, Ken Arrow, Rob Axtell, Chris Barrett, Eric Beinhocker, Larry Blume, Buz Brock, Richard Bronk, David

  Colander, Steven Durlauf, Josh Epstein, Yuri Ermoliev, Doug Erwin, Doyne

  Farmer, Magda Fontana, Walter Fontana, John Geanakoplos, Herb Gintis,

  Yuri Kaniovski, Alan Kirman, Roger Koppl, Chris Langton, Steve Lansing,

  Blake LeBaron, Sander van der Leuuw, Phil Mirowski, Melanie Mitchell,

  Wolfgang Polak, David Reisman, Nate Rosenberg, Daria Roithmayr, John

  Rust, Tom Sargent, Lee Smolin, Martin Shubik, William Tabb, Paul Tayler, and

  Leigh Tesfatsion—all of whom have influenced me.

  None of our early work at Santa Fe would have happened without the

  help and backing of Kenneth Arrow, Philip Anderson, George Cowan, Murray

  Gell-Mann, and David Pines. And at Santa Fe, Ronda Butler-Villa, Ellen

  Goldberg, Ginger Richardson, Geoffrey West, and Chris Wood have been of

  constant help and encouragement.

  Most of papers here were written when I was at Stanford or the Santa Fe

  Institute. Others were written under the auspices of PARC (formerly Xerox

  Parc), IBM Almaden, and Fuji Xerox Palo Alto Lab (FXPAL). For financial sup-

  port, or simply for access to facilities where I could chew the end of a pencil and think and write, I’m grateful to several people and sources: John Reed and Henry Lichstein of Citibank; Ernesto and Andrea Illy of IllyCaffè; Andy Grove

  and Les Vadasz of Intel; Bill Miller and Michael Mauboussin of Legg Mason

  Capital Management; the Sloan Foundation; John Seeley Brown and Walt

  Johnson of PARC; Jim Baker of FXPAL; Paul Maglio and Jim Spohrer of IBM

  Almaden; Bertil Andersson, Jitendra Singh, Guaning Su, and Jan Vasbinder

  of Nanyang Technological University; Mike Keller of Stanford University

  Library, and the excellent staff of St John’s College Library at Santa Fe. I also thank James Bailey, Randy Burge, John Chisholm, Mick Costigan, Bo Ekman,

  Richard Rhodes, and Paul Saffo for much help and cheerful encouragement.

  Naturally, I take responsibility for all the ideas here; no intellectual transgres-sions should be attributed to any people or institutions named on this page.

  It has been a pleasure to work with my son Ronan Arthur who read and

  commented on several of the papers, and my editors, Joan Bossert and Louis

  Gulino at Oxford, who did much to make this book a reality. Finally, for sup-

  port, advice, and simple good fellowship over the years, I thank my friends

  David Lane, Cormac McCarthy, Jim Newcomb, and Martin Shubik.

  [ xxiv ] Acknowledgments

  CHAPTER 1

  Complexity Economics

  A Different Framework for Economic Thought

  W. BRIAN ARTHUR1

  This paper serves as an introduction to many of the themes that follow; it provides a framework for complexity economics. Complexity economics builds on the proposition that the economy is not necessarily in equilibrium: economic agents (firms, consumers, investors) constantly change their actions and strategies in response to the outcome they mutually create. This further changes the outcome, which requires them to adjust afresh. Agents thus live in a world where their beliefs and strategies are constantly being “tested” for survival within an outcome or “ecology” these beliefs and strategies together create.

  Economics has largely avoided this nonequilibrium view in the past, but if we allow it, we see patterns or phenomena not visible to equilibrium analysis. These emerge probabilistically, last for some time and dissipate, and act at the “meso-level” of the economy (between the micro- and macro-levels). We also see the economy not as something given and existing but forming from a constantly developing set of technological innovations, institutions, and arrangements that draw forth further innovations, institutions, and arrangements. Complexity economics thus sees the economy as in motion, perpetually “computing” itself—perpetually constructing itself anew. Where equilibrium economics emphasizes order, determinacy, deduction, and stasis, comp
lexity economics emphasizes contingency, indeterminacy, sense-making, and openness to change.

  The paper was written especially for this volume. It builds on a 1999 essay of mine in Science on complexity and the economy (Chapter 12 here).

  1. Intelligent Systems Lab, PARC, and External Professor, Santa Fe Institute. I thank Ronan Arthur, Richard Bronk, David Colander, Doyne Farmer, Magda Fontana, Ole

  Peters, David Reisman, and William Tabb for valuable comments.

  Over the past twenty-five years, a different approach to economics has been slowly birthing, and slowly growing— complexity economics. Complexity

  economics holds that the economy is not necessarily in equilibrium, that com-

  putation as well as mathematics is useful in economics, that increasing as well as diminishing returns may be present in an economic situation, and that the

  economy is not something given and existing but forms from a constantly

  developing set of institutions, arrangements, and technological innovations.

  The approach got its start largely at the Santa Fe Institute in the late 1980s but now has many practitioners,2 and it raises several questions. What does

  this different way of thinking about the economy offer? How exactly does it

  work and where does it fit in? Will it replace neoclassical economics, or be

  subsumed into neoclassical economics? And under what logic, if any, does it

  operate?

  My purpose in this paper is to answer these questions, especially the last

  one. In doing so I will not attempt to provide a survey or guided tour, rather I want to provide a framework—a coherent logic—for thinking about this

  new approach. I will argue from first principles and will build from two earlier essays of mine (Arthur 1999, 2006) as well as the work of many other people

  to illustrate the key points.3

  I will argue that this new approach is not just an extension of standard eco-

  nomics, nor does it consist of adding agent-based behavior to standard mod-

  els. It is a different way of seeing the economy. It gives a different view, one where actions and strategies constantly evolve, where time becomes important, where structures constantly form and re-form, where phenomena appear