Complexity and the Economy

Home > Other > Complexity and the Economy > Page 33
Complexity and the Economy Page 33

by W Brian Arthur


  on all scales, large and small. Thus emerge periods of change triggering further change, periods of high volatility, followed by periods when little changes and little needs to be changed, periods of quiescence. This is GARCH behavior.

  Let me now summarize. What we found in our artificial stock market is that,

  providing our investors start near the rational-expectations academic solu-

  tion, this solution prevails. But this is a small set of parameter space. Outside this, in the complex regime, self-reinforcing beliefs and self-reinforcing

  avalanches of change emerge. A wider theory and a richer “solution” or set

  of behaviors then appears, consonant with actual market behavior. The

  rational-expectations theory becomes a special case.

  In the standard view, which has come down from the Enlightenment, the

  economy is an object. It is complicated but can be viewed mechanistically.

  Subject and object—agents and the economy they perform in—can be neatly

  separated. The view I am giving here is different. It says that the economy itself emerges from our subjective beliefs. These subjective beliefs, taken in aggregate, structure the micro economy. They give rise to the character of financial markets. They direct flows of capital and govern strategic behavior and negoti-ations. They are the DNA of the economy. These subjective beliefs are a-priori or deductively indeterminate in advance. They co-evolve, arise, decay, change, mutually reinforce, and mutually negate. Subject and object cannot be neatly

  separated. And so the economy shows behavior that we can best describe as

  organic, rather than mechanistic. It is not a well-ordered, gigantic machine. It is organic. At all levels it contains pockets of indeterminacy. It emerges from subjectivity and falls back into subjectivity.

  t He end of certain t y in economics [ 181 ]

  CHAPTER 12

  Complexity and the Economy

  W. BRIAN ARTHUR

  This essay summarizes my thinking on complexity and the economy in 1999. It is a pre-cursor to (and heavily overlaps) the introductory chapter of this volume, but I include it here because it introduces the term “complexity economics” for the first time. The article appeared in Science, April 2, 1999, 244: 107–109.

  Common to all studies on complexity are systems with multiple elements

  adapting or reacting to the pattern these elements create. The elements

  might be cells in a cellular automaton, ions in a spin glass, or cells in an

  immune system, and they may react to neighboring cells’ states, or local mag-

  netic moments, or concentrations of B and T cells. Elements and the patterns

  they respond to vary from one context to another. But the elements adapt

  to the world—the aggregate pattern—they co-create. Time enters naturally

  here via the processes of adjustment and change: As the elements react, the

  aggregate changes; as the aggregate changes, elements react anew. Barring the

  reaching of some asymptotic state or equilibrium, complex systems are sys-

  tems in process that constantly evolve and unfold over time.

  Such systems arise naturally in the economy. Economic agents, be they

  banks, consumers, firms, or investors, continually adjust their market moves,

  buying decisions, prices, and forecasts to the situation these moves or deci-

  sions or prices or forecasts together create. But unlike ions in a spin glass, which always react in a simple way to their local magnetic field, economic elements (human agents) react with strategy and foresight by considering out-

  comes that might result as a consequence of behavior they might undertake.

  This adds a layer of complication to economics that is not experienced in the

  natural sciences.

  Conventional economic theory chooses not to study the unfolding of the patterns its agents create but rather to simplify its questions in order to seek analytical solutions. Thus it asks what behavioral elements (actions, strategies, and expectations) are consistent with the aggregate patterns these

  behavioral elements co-create? For example, general equilibrium theory asks

  what prices and quantities of goods produced and consumed are consistent

  with (would pose no incentives for change to) the overall pattern of prices

  and quantities in the economy’s markets. Game theory asks what moves or

  choices or allocations are consistent with (are optimal given) other agents’

  moves or choices or allocations in a strategic situation. Rational expectations economics asks what forecasts (or expectations) are consistent with (are on

  average validated by) the outcomes these forecasts and expectations together

  create. Conventional economics thus studies consistent patterns: patterns

  in behavioral equilibrium that would induce no further reaction. Economists

  at the Santa Fe Institute, Stanford, MIT, Chicago, and other institutions are

  now broadening this equilibrium approach by turning to the question of how

  actions, strategies, or expectations might react in general to (might endog-

  enously change with) the aggregate patterns these create ( 1, 2). The result—

  complexity economics—is not an adjunct to standard economic theory but

  theory at a more general, out-of-equilibrium level.

  The type of systems I have described become especially interesting if they

  contain nonlinearities in the form of positive feedbacks. In economics, posi-

  tive feedbacks arise from increasing returns ( 3, 4). To ensure that a unique, predictable equilibrium is reached, standard economics usually assumes

  diminishing returns. If one firm gets too far ahead in the market, it runs into higher costs or some other negative feedback, and the market is shared at a

  predictable unique equilibrium. When we allow positive feedbacks, or increas-

  ing returns, a different outcome arises. Consider the market for online ser-

  vices of a few years back, in which three major companies competed: Prodigy,

  Compuserve, and America Online. As each gained in membership base, it

  could offer a wider menu of services, as well as more members to share spe-

  cialized hobby and chat room interests with—that is, there were increasing

  returns to expanding the membership base. Prodigy was first in the market,

  but by chance and strategy America Online got far enough ahead to gain an

  unassailable advantage. Today it dominates. Under different circumstances,

  one of its rivals might have taken the market. Notice the properties here: a

  multiplicity of potential solutions; the outcome actually reached is not pre-

  dictable in advance; it tends to be locked in; it is not necessarily the most

  efficient economically; it is subject to the historical path taken; and although the companies may start out equal, the outcome is asymmetrical. These properties have counterparts in nonlinear physics where similar positive feed-

  backs are present. What economists call multiple equilibria, nonpredictability, lock-in, inefficiency, historical path dependence, and asymmetry, physicists

  comPlexi t y and t He economy [ 183 ]

  call multiple metastable states, unpredictability, phase or mode locking, high-energy ground states, non-ergodicity, and symmetry breaking ( 5).

  Increasing returns problems have been discussed in economics for a long

  time. A hundred years ago, Alfred Marshall ( 6) noted that if firms gain advantage as their market share increases, “whatever firm first gets a good start will obtain a monopoly.” But the conventional static equilibrium approach gets stymied by indeterminacy: If there is a multiplicity of equilibria, how might one be reached? The process-oriented complexity approach suggests a way to deal

&nb
sp; with this. In the actual economy, small random events happen; in the online

  services case, events such as random interface improvements, new offerings,

  and word-of-mouth recommendations. Over time, increasing returns magnify

  the cumulation of such events to select the outcome randomly. Thus, increasing returns problems in economics are best seen as dynamic processes with random events and natural positive feedbacks—as nonlinear stochastic processes.

  This shift from a static outlook into a process orientation is common to com-

  plexity studies. Increasing returns problems are being studied intensively in

  market allocation theory ( 4), international trade theory ( 7), the evolution of technology choice ( 8), economic geography ( 9), and the evolution of patterns of poverty and segregation ( 10). The common finding that economic structures can crystallize around small events and lock in is beginning to change policy in all of these areas toward an awareness that governments should avoid both extremes

  of coercing a desired outcome and keeping strict hands off, and instead seek to push the system gently toward favored structures that can grow and emerge

  naturally. Not a heavy hand, not an invisible hand, but a nudging hand.

  Once we adopt the complexity outlook, with its emphasis on the formation

  of structures rather than their given existence, problems involving prediction in the economy look different. The conventional approach asks what forecasting model (or expectations) in a particular problem, if given and shared by all agents, would be consistent with (would be on average validated by) the actual time series this forecasting model would in part generate. This “rational expectations” approach is valid. But it assumes that agents can somehow deduce

  in advance what model will work and that everyone “knows” that everyone

  knows to use this model (the common knowledge assumption.) What hap-

  pens when forecasting models are not obvious and must be formed individu-

  ally by agents who are not privy to the expectations of others?

  Consider as an example my El Farol Bar Problem ( 11). One hundred people must decide independently each week whether to show up at their favorite

  bar (El Farol in Santa Fe). The rule is that if a person predicts that more that 60 (say) will attend, he or she will avoid the crowds and stay home; if he predicts fewer than 60, he will go. Of interest are how the bar-goers each week

  might predict the numbers of people showing up, and the resulting dynamics

  of the numbers attending. Notice two features of this problem. Our agents

  will quickly realize that predictions of how many will attend depend on others’

  [ 184 ] Complexity and the Economy

  predictions of how many will attend (because that determines their attendance). But others’ predictions in turn depend on their predictions of others’

  predictions. Deductively there is an infinite regress. No “correct” expectational model can be assumed to be common knowledge, and from the agents’ viewpoint, the problem is ill defined. (This is true for most expectational problems, not just for this example.) Second, and diabolically, any commonalty of expectations gets broken up: If all use an expectational model that predicts few will go, all will go, invalidating that model. Similarly, if all believe most will go, nobody will go, invalidating that belief. Expectations will be forced to differ.

  100

  90

  Numbers Attending

  80

  70

  60

  50

  40

  30

  20

  10

  0

  0

  20

  40

  60

  80

  100

  Time

  Figure 1:

  Bar attendance in the first 100 weeks.

  In 1993, I modeled this situation by assuming that as the agents visit the

  bar, they act inductively—they act as statisticians, each starting with a variety of subjectively chosen expectational models or forecasting hypotheses. Each

  week they act on their currently most accurate model (call this their active

  predictor). Thus agents’ beliefs or hypotheses compete for use in an “ecology”

  these beliefs create.

  Computer simulation (Figure 1) showed that the mean attendance quickly

  converges to 60. In fact, the predictors self-organize into an equilibrium ecology in which, of the active predictors, 40% on average are forecasting above

  60 and 60% below 60. This emergent ecology is organic in nature, because

  although the population of active predictors splits into this 60/40 aver-

  age ratio, it keeps changing in membership forever. Why do the predictors

  self-organize so that 60 emerges as average attendance and forecasts split

  into a 60/40 ratio? Well, suppose 70% of predictors forecasted above 60 for a

  longish time, then on average only 30 people would show up. But this would

  validate predictors that forecasted close to 30, restoring the ecological bal-

  ance among predictions. The 40%/60% “natural” combination becomes an

  comPlexi t y and t He economy [ 185 ]

  emergent structure. The Bar Problem is a miniature expectational economy with complex dynamics ( 12).

  One important application of these ideas is in financial markets. Standard

  theories of financial markets assume rational expectations—that agents

  adopt uniform forecasting models that are on average validated by the

  prices these forecast ( 13). The theory works well to first order. But it doesn’t account for actual market anomalies such as unexpected price bubbles and

  crashes, random periods of high and low volatility (price variation), and the

  heavy use of technical trading (trades based on the recent history of price

  patterns). Holland, LeBaron, Palmer, Tayler, and I ( 14) have created a model that relaxes rational expectations by assuming, as in the Bar Problem, that

  investors cannot assume or deduce expectations but must discover them.

  Our agents continually create and use multiple market hypotheses—indi-

  vidual, subjective, expectational models—of future prices and dividends

  within an artificial stock market on the computer. These “investors” are

  individual, artificially intelligent computer programs that can generate and

  discard expectational hypotheses and make bids or offers based on their

  currently most accurate hypothesis. The stock price forms from their bids

  and offers and thus ultimately from agents’ expectations. So this market-in-

  the-machine is its own self-contained, artificial financial world. Like the bar, it is a mini-ecology in which expectations compete in a world those expectations create.

  Within this computerized market, we found two phases or regimes. If

  parameters are set so that our artificial agents update their hypotheses slowly, the diversity of expectations collapses quickly into homogeneous rational

  ones. The reason is that if a majority of investors believes something close to the rational expectations forecast, then resulting prices will validate it, and deviant or mutant predictions that arise in the population of expectational

  models will be rendered inaccurate. Standard finance theory, under these

  special circumstances, is upheld. But if the rate of updating of hypotheses is increased, the market undergoes a phase transition into a complex regime and

  displays several of the anomalies observed in real markets. It develops a rich psychology of divergent beliefs that don’t converge over time. Expectational

  rules such as “if the market is trending up, predict a 1% price rise” that appear randomly in the population of hypotheses can become mutually reinforcing: If enough investors act
on these, the price will indeed go up. Thus sub-

  populations of mutually reinforcing expectations arise, agents bet on these

  (therefore technical trading emerges), and this causes occasional bubbles and

  crashes. Our artificial market also shows periods of high volatility in prices, followed randomly by periods of low volatility. This is because if some investors discover new profitable hypotheses, they change the market slightly,

  causing other investors to also change their expectations. Changes in beliefs

  therefore ripple through the market in avalanches of all sizes, causing periods

  [ 186 ] Complexity and the Economy

  of high and low volatility. We conjecture that actual financial markets, which show exactly these phenomena, lie in this complex regime.

  After two centuries of studying equilibria—static patterns that call for no

  further behavioral adjustments—economists are beginning to study the gen-

  eral emergence of structures and the unfolding of patterns in the economy.

  Complexity economics is not a temporary adjunct to static economic theory

  but theory at a more general, out-of-equilibrium level. The approach is making itself felt in every area of economics: game theory ( 15), the theory of money and finance ( 16), learning in the economy ( 17), economic history ( 18), the evolution of trading networks ( 19), the stability of the economy ( 20), and political economy ( 21). It is helping us understand phenomena such as market instability, the emergence of monopolies, and the persistence of poverty

  in ways that will help us deal with these. And it is bringing an awareness that policies succeed better by influencing the natural processes of formation of

  economic structures than by forcing static outcomes.

  When viewed in out-of-equilibrium formation, economic patterns some-

  times fall into the simple homogeneous equilibria of standard economics.

  More often, they are ever changing, showing perpetually novel behavior and

  emergent phenomena. Complexity therefore portrays the economy not as

  deterministic, predictable, and mechanistic but as process dependent, organic, and always evolving ( 22).

 

‹ Prev