sfwork logo

Lessons from Complexity Theory

"The East Asian Economic and Financial Crisis: Lessons from Complexity Theory"

Asia Pacific Economic Cooperation Economic Committee

APEC Economic Outlook Symposium Xiamen, China, 16 - 17 May 1998

by Dr Mark McKergow

Mark McKergow PhD MBA is an independent consultant with Mark McKergow Associates, 6 Auburn Road, Bristol BS6 6LS, United Kingdom. Following a doctoral degree in Physics and a career in the electricity industry, he is one of the leaders in applying insights from the new sciences to corporate settings. He consults to major UK corporations including British Energy, Halifax Group and British Nuclear Fuels, and has lectured in the UK, Europe and South Africa. His articles and book reviews have appeared in several journals including Long Range Planning and Organisations & People. Dr. McKergow was invited to prepare this paper for discussion at the APEC Economic Outlook Symposium. As he was unable to attend in person due to other commitments, the paper was presented on his behalf by Mr. Dan Ciuriak in his capacity as representative of the Chair of the Economic Committee.

Introduction

What do s freeway traffic jam, stock market fluctuations, a tractor accident and the layout of characters on a typewriter keyboard have in common? Answer: they are all examples of, or results from, "complex systems".

Complexity theory has been one of the most interesting scientific developments of the 1990s. Taking insights and inputs from mathematics, biology, computing, economics and evolution amongst others, scientists have begun to grapple with the structures and development of complex systems. Applications are being sought in the fields of economics (Ormerod, 1994) and management (McKergow, 1996). In this note, I will describe briefly the characteristics of a complex system in the sense that scientists understand them, and use four particular aspects of complexity to hold up a light to the ongoing economic and financial crisis in a number of East Asian economies. This light will not reveal answers, but can help us to arrive at some good questions. These "Lessons from complexity theory" may be summarised as:

  1. Searching for "The Cause" of the crisis is futile. Think instead about the way ahead.
  2. Don't ask whose fault is was. Ask instead about the influences to be brought to bear on the future.
  3. Don't ask "What's going to happen?" Ask "How would we recognise a useful way ahead?".
  4. Don't seek a grand plan for recovery. Seek small steps combined with positive feedback to grow the recovery.

Complexity Theory

There are a number of meanings to the word "complexity", and it will be as well to be clear about how we are using it here. We do not mean simply "complicatedness". A system or artefact can be complicated or intricate without have any "complexity" attributes. Peter Senge, in his book The Fifth Discipline, makes the point well in drawing the distinction between "detail complexity" (for instance, highly intricate machines or management systems) and "dynamic complexity" (where inter-relationships between elements in a system are based on feedback loops, and are many and varied).

In this context, the complexity we refer to is similar to Senge's dynamic complexity. There are some attributes which are associated with complex systems. Such systems are self-referential (i.e., their elements interact in a systemic fashion) and contain elements of both positive and negative feedback. They are non-linear, so that a small change can lead to much larger effects in other parts of the system and at other times. Systems which display these traits have some interesting properties:

  • They are non-deterministic. Wholly accurate predictions of future states cannot be made, however well the current and past states are known.
  • They show emergent properties - patterns which result from the overall action of all the elements of the system, are not obvious even if the system interactions are precisely known, and may be rich and unexpected.

Perhaps surprisingly, complex systems can be made up of many relatively simple parts. An example might be ants in a nest, molecules of air, birds in a flock, people in an organisation, or elements in a computer program. It has been shown, however, that even if all the details of the individual parts are known (as has been demonstrated using certain types of computer program), the overall system can still be Complex and non-deterministic, and can produce effects which are real and observable, but cannot be determined by examining a single element out of context.

There is not space to review complexity science further here. Further reading for the interested layperson can be found in Waldrop (1994) and Casti (1994).

The economic and financial crisis that has hit East Asia certainly fulfills the criteria above as the result of a "complex system". So what can the study of complexity do to help us to understand the situation and guide us about future developments? I do not propose to tackle the economics of the situation directly. Rather, I will take four angles for which complexity gives us some pointers:

  • "Causes" of the crisis
  • Who is in control of the situation?
  • Possibility space
  • Increasing returns and positive feedback

Causes of the East Asian economic crisis

In researching this piece, I have been struck by the number of factors that have been identified by various writers as contributing in one way or another to the crisis. These include, inter alia:

  • The floating of the Thai baht against the US dollar
  • Shifts in the Hong Kong, China stock market
  • Investment choices and decision-making in Southeast Asia
  • The differences between Asian and Western "values"
  • The prevalence of family and clan-focussed business structuring in the region
  • Foreign currency speculators
  • Autocratic government
  • Market protectionism
  • Frictions surrounding Myanmar's entry into ASEAN
  • Indonesian borrowing from Japan and South Korea
  • State-directed economic development in Malaysia, Korea, Singapore and Chinese Taipei
  • Moves away from state-directed economic development
  • The halting nature of Japan's recovery from recession in the early 1990s
  • Strength of the US economy and currency, coupled with large volume of dollar-denominated loans made to Southeast Asian economies

There is also disagreement as to the extent to which the difficulties faced by the economies involved are the same. Is this one crisis, or several? Will the resolution to the crisis be the same for all concerned?

The list goes on.......but how sensible is it to construct such a list? Complexity theory has shown us how sets of dynamic interlinks and feedbacks combine to produce an unpredictable and developing outcome. But to what extent can the crisis be laid at the door of any one of these? It can't.

In the UK, there is a popular radio serial called "The Archers", about a rural community in the English Midlands. John Archer, a bright 24 year-old, has just been killed in an accident where his tractor overturned and crushed him. His distraught parents are currently trying, in a mood that will be familiar to many of us, to work out "Why?". If only the tractor's wheel alignment hadn't been faulty.....If only his father hadn't shouted at him that fateful morning....If only he hadn't been turned down by his girlfriend the day before.....If only he hadn't been working too hard....If only his mother had gone to look for him earlier.....This list could go on as well.

Which is the "actual" cause of the accident? In complexity terms - none of them and all of them. In such cases, cause and effect become intertwined, in ways which may be hard to unravel. And, even if we could unravel them, the outcome would still not be predictable. In complex systems, like their close relations from chaos theory, outcomes are highly sensitive to even infinitesimally small changes in the system parameters. So, it makes no sense to ask, other than for fun in the bar afterwards, "What was the cause of the accident?". We must beware the overapplication of formal logic - if applied to the wrong question, it can lead us into tight and unwelcome corners (Kosko, 1994).

Turning to the Asian crisis, the above suggests that, rather than seeking to identify the cause, better questions to ask might be:
What actually happened?
What contributed to the situation?
What combination of events preceded the crisis?

This may shed light on the ways in which events combined to produce the crisis, and would provide historians with some data to perhaps identify similar patterns in the future. This might help to prevent a future crisis. But it may not help resolve this one. In my work with the UK nuclear energy industry, we have been known to say that the circumstances leading up to an incident may help prevent the next one - but it sure doesn't help clean up this one!

Who is in control of the situation?

This question is sometimes phrased in terms of "Whose fault is it?". For similar reasons to those outlined above, the complexity answer is - nobody and everybody.

In a complex system, we observe how rich and varied patterns of behaviour can arise from the interactions of many "agents", each with a relatively simple behaviour. Yet the net effect may be contrary to the wishes of the agents. An example is the phenomenon of "phantom traffic jams" on freeways. When the road is busy, a small action (for example, one car braking sharply to avoid another) can result in other drivers having to brake briefly. These drivers then go on their way - but back down the motorway more and more drivers are braking, by greater and greater degrees, to avoid those in front, until the freeway is at a standstill. Each driver approaching the scene sees a jam, which progresses slowly for a distance and then frees up. From a higher vantage point, for example a police helicopter, we can see a standing wave of traffic passing up the freeway, as cars join the back, work their way to the front, and then go on.

Who is in control of such a phenomenon? No-one. It is the result of the interactions. Everyone has an influence - maybe small, but an influence, and can potentially act to change the situation. There is a principle in cybernetics - a forefather of complexity theory - called Ashby's Law. It states that the part of the system with the most influence is the part that has the highest number of potential states - ie, the most flexibility. The degree of influence of the players is, of course, not necessarily the same. And some players may deploy their influence by NOT acting, when it is known they have the potential to act. So complexity guides us away from question of control and blame, towards questions like:
Who are the players?
What influence do they have?
What influence could they have, if they chose to?

One of the interesting properties of cybernetically-interlinked complex systems is that, in principle, any of the players with some kind of influence may have a critical effect on the whole. This potential impact is most definitely not in proportion to the size of the initial influence. Just as in the infamous "butterfly effect", small beginnings can lead to significant ends.

Possibility Space

One of the ideas from the mathematical side of complexity theory is that we can imagine a "space" - a many-dimensioned array - where ALL the possible future states of the system are arranged, rather like a library of possibilities. One way to imagine this is as a library of books. In his short story The Labyrinth, Jose Luis Borges postulates the existence of a library containing all the books which could ever be written in Roman script. This would be a truly gargantuan place. The vast majority of the books, of course, would be completely meaningless in any language. In one volume, 300 pages of the letters "pdfe" would be repeated. It contains the book which described YOUR life! It also contains the book which describes your life up to this point, and then turns into a complete fiction. This library is enormous.

One of the findings of complexity theory is that, huge though such spaces must be, it can be profitable to search through them, using criteria to quickly sort out useful possibilities from the useless. In order to do this, some kind of boundary is defined, to set out the possibility space, so that the search algorithms can go to work. A similar process might be applied to our consideration of East Asia. Rather than simply trying to define the most likely outcomes, we might ask:

What is the range of possible outcomes?

What are the bounding cases - the most unlikely scenarios we can think of?
What characteristics could we use to begin to identify useful routes through these possibilities?
How would we know the difference between a strong recovery and an unsustainable lurch forward?
By concentrating on establishing what would be a "useful" scenario (and also what would not!), we can then start to look around and find elements which are already in place in the world. We can then start to put into place other, missing, elements, and refine the way forward. Otherwise we may end up like the man who rebuilt his house so that the sun wouldn't get in his eyes while he worked - rather than simply adjust his workdesk!

Increasing returns and positive feedback

One indicator of complexity is the presence of positive, as well as negative, feedback. Negative feedback serves to return a system to its equilibrium state, and is well understood - indeed classical economics is founded on it. Positive feedback is well understood in scientific terms, as the self-generated escalation of a systemic variable. Controversial economist Brian Arthur (1990, 1995) has examined the role of positive feedback effects in economics. He concludes that such effects are indeed present, and can lead to situations such as "lock-in", where the benefits of standardisation around a particular technology can lead to its persistence even after "better" alternatives are developed. (The best known example is the ubiquitous QWERTY keyboard, designed to slow down the operators of 19th century machines, and now an apparent fixture.)

Complexity leads us to conclude that large changes can be produced by a small initial event combined with positive feedback. This is contrary to the "conventional" thinking that large changes require commensurately large actions -- for example, to kick-start the economy, the Government of Japan has recently announced a massive fiscal stimulus package amounting to 3% of GDP; this is already being criticised by economists and observers. What is the best way to grow an oak tree - to plant a single acorn, around which guards are stationed, with electric fences, spotlights, round-the-clock forestry experts in attendance, waiting to see whether it comes up......or to plant many acorns, expecting only a few to grow, and then to nuture the ones that actually grow, knowing that one is enough.

Complexity therefore leads us away from designing grand plans for recovery, and instead gets us to ask:
What positive feedback mechanisms are available to help advance the situation?
What small actions can be taken to prime these mechanisms?
How many actions, each costing small amounts of money, can we devise?
What has not yet been tried?"

Conclusion

We have explored some basic aspects of complexity theory. We have seen how the science world's exploration of complex systems can offer some guidance. This comes in terms of pointing out that some questions are unanswerable, some are unhelpful, and others may provide useful pointers. Complexity theory is a rich source of inspiration, and there are may other angles which the reader may know of, or find, which could be helpful. I chose these four.

In considering the way forward, we can summarise the Lessons from Complexity by comparing the ways to seek progress in an ordered world, and in a complex one.

"Ordered" world

"Complex" world

The cause and effect of actions can be determined precisely.
Certain parties have "control".
There is only one way out - first we must find it.
Large effects require enormous co-ordinated efforts.
The future can be planned.

Cause and effect are intertwined, and cannot be determined in advance - don't pretend they can.
All parties have influence.
There are many possible ways out - we should focus on finding their starting points.
Large effects can come from small starts and positive feedback.
The future "emerges" from the combined actions of the players.

This note has not attempted to provide answers. These must come from those who know the territory better then I. Whatever answers we may find, and however the situation develops, one thing is certain. Despite the assumptions in classical science and economics about reversibility, the world is in fact undergoing a continuous and irreversible process. There is no way back. To want to return to a past situation is as unrealistic as wanting to be 5 years old again, wanting to unscramble your breakfast eggs, or wanting the atom bomb never to be invented. We may want these things - and yet we cannot have them. We must go on, to marvel at our own five year olds, to have boiled eggs tomorrow, to devise treaties for disarmament. The future, whatever it holds, however wonderful or dreadful it may be, always starts........now.

References

Arthur, Brian: "Positive Feedbacks in the Economy", Scientific American (February 1990), pp 90 - 92 (1990)
Arthur, Brian: Complexity, Vol.1, No 1, pp 20 - 25 (1995)
Casti, John: Complexification, Abacus (1994)
Kosko Bart: Fuzzy Logic, HarperCollins London (1994)
McKergow Mark: "Complexity Science and Management - What's in it for business?", Long Range Planning, Vol 29, pp 721 - 727 (1996)
Ormerod Paul: The Death of Economics, Faber and Faber (1994)
Senge, Peter: The Fifth Discipline, Century (1991)
Waldrop, M Mitchell: Complexity, Penguin (1994)

Acknowledgements:

Thanks to Jenny Clarke of MMA and Scott Lichtenstein of Henley Management College for comments and draft reading.