Guest Post: More on the Complexity Economics Panel at Bretton Woods
New Economic Thinking… on reading economic data
The recent INET meeting at Bretton Woods organized by the Institute for New Economic Thinking (INET) with support from people like George Soros, brought together a large group of leading creative economists and ecological systems scientists, to push the envelope in discussing how to reorganize the world economic system. Choosing the same site for which the world changing Bretton Woods conference after WWII was named was both intentional and seemed quite appropriate. A strong “vision” seemed to be emerging, even if the details of how to advance it were not so clear. All might agree that nature’s economies seem to work better than ours, and we need to change something. The concluding conversation between Paul Volker and George Soros presented some of the broad issues, nicely moderated by Gillian Tett, and was quite interesting.
I liked the session on “Complexity Economics” particularly. It really is necessary that people begin thinking of the economy as a complex system, and that they learn the lessons of ecology in order to understand how the economy can be transformed to be as stable and healthy as ecologies. The concluding comments were also a very nice lead-in to my own approach, that to meet the challenge of using complex systems modeling, what we need are better conceptual models for reading economic data. A good example of new data needing new concepts was discussed in another session by Duncan Foley and I insert his first graph below.
Finding how to model emergent behaviors in a complex world is the problem. What model do you use when all you know is the economy is clearly not following your theory? Slightly edited below is my comment on the conference posted to the INET site and at the linked History of Economics Playground blog.
I thought what was most provocative at the INET-BW conference were the ending comments at the complexity economics session. The panel explained the problem of using the available economic data to develop complex system models as a lack of good conceptual models to use. With the flood of new data sources and emerging changes in system complexity and behavior, what seems to be limiting the development of models is not just hardware, software and data, but also significantly a lack of ideas for what complex relationships to model.
It interests me particularly in that translating observed behavior into modeling concepts is one way to describe what has been my main subject for some time. How to read complex system behavior from available data involves considering natural systems somewhat like organisms, that behave on their own in open environments, that you study to watch how their organization develops and changes. You might start with provocative data as presented by Prof. Duncan Foley at the conference showing the US economy not following expected behavior during the economic collapse of 2008 (Fig 1).
What Foley first saw in the data was evidence of misguided monetary policy, resulting in a rapid recovery in the share of National Income due to finance, failing to result in a recovery for the producing (Value-Added) economy. I might agree with that, while also noting that the popular idea that the collapse was caused by finance was evidently mistaken too. The decline in the productive economy started well before finance collapsed, and was not reversed by rescuing finance either. From a dynamics point of view, the tipping point of the downturn was actually the inflection point in both curves in 2004, and was followed by a major transfer of wealth from the producing economy to finance. It seems likely that was what actually “broke the bank” in the end, that the producing economy was being sapped of earnings. That raises questions about what mysterious force “emerged from nowhere” in 2004, and caused a slide of wealth between two sectors that had previously been in balance.
The main point here, though, is just showing how making these kinds of critical observations about the directions of accumulative change raises questions about organizational change in the behavior of the system, without any theory for it. That directly gives you lots of very relevant modeling ideas to test.
That something started depressed the producing economy faster than the financial economy as early as 2005, becomes a leading indicator of the collapse, too, pointing to some sort of natural structural cause. There were the oddly exploding prices for food and fuel resources occurring around the world at that time, for example, that might be connected. The scientific problem is that these kinds of departures from established economic theory show the economy changing in a way that established theory doesn’t allow for, and you don’t know what part or parts of the established framework of theory to discard in looking for the real explanation. What emergent phenomena require is new theory and models for which you have no history.
To build models of emergent phenomena we are faced with having to narrow down the vast array of design options that nature deals with, without having her way putting them to the test. I find it very productive to start with learning how to read the systemic behaviors of nature, without a theory, by watching her emerging new forms of organization develop. Sure, people are made very uncomfortable by even the idea of studying emerging systems without any theory thought of as driving them… Natural systems clearly are independently energetic and organized in their behavior, though, somehow. Nature seems to make sense of them, and we need to learn. There are approaches and methods to help getting over the misgivings.
The discussion in the INET-BW Complexity Economics session was from the usual viewpoint that building theoretical models of the economy was necessary for studying the “non-theoretical” economy. Most of modern science is also centered on the study of theory, rather than nature. The natural economy, however, does itself display organized non-theoretical systemic behaviors, and does successfully work by itself, somehow. I’d offer that “studying both” theory as theory and nature as nature is the better paradigm. One might start from either, and go back and forth, connecting theory with the observed organic behavior of the non-theoretical system, generating better questions for inventing new and more useful theories. It would greatly enhance the use of theoretical models as diagnostic tools for gauging the health of the organic system too.
If you consider the economy as a complex natural organism, then data taken from it displays how its natural organization is behaving, by itself, so the contest becomes how to identify changing and emerging organization in how the non-theoretical system operates, and devise ways to alter our models. My core method starts with the implication of energy conservation, that new organization for energy using systems requires an accumulative process of development that can be spotted by aggregating data displaying growth curves.
An excellent and important example using that principle is in the now 10 year phenomenon of global resource demand increasingly exceeding supply. Demand is persistently growing exponentially, and supply is now only growing linearly, and the previous price relationship has now completely broken down. I have a short essay about to appear in New European Economy, discussing it as “A defining moment for investing in sustainability.“ The challenge is that the apparent collective response of world food and fuel markets to excessive demand is essentially to panic, sending the floor price of all food and fuel resources on an exponential path, still unresolved.
That does call for response, but here it’s presented just as one good example of the difference it makes to have a working systems diagnostic method, informed by both physics and economics. It becomes a kind of forensic study, which serves to expand the subjects of science from mainly a study of theories and models to include the study of complex systems in their native form too.
Philip F. Henshaw’s innovative systems science work goes back to 1970’s, evolving into a fairly practical new general method, using physics principles as diagnostic tools, for investigating complex natural systems that develop by growth. Phil has a B.S. in physics from St. Lawrence University, an MFA in environmental design from the University of Pennsylvania, and a substantial accumulated body of original research and publication. He does consulting, research and writing as HDS systems design science. Further information is on his website www.synapse9.com.