January 16, 2007
Standard economics and the ‘evolution’ thesis can coexist
Is the discipline of economics built on sand? Most economists would answer with a resounding “no”. But most must also know that the economy is not characterised by perfect foresight and equilibrium, but by trial and error and evolution. That was the intuition of the Austrian economists, Joseph Schumpeter and Friedrich Hayek. But this vision has had next to no influence in the discipline itself. This gap between how economists think and what economies are is evident to any careful observer. But hitherto nobody has closed the gap between rigorous theory and broad vision. This, argues McKinsey’s Eric Beinhocker in a brilliant, thought-provoking and wide-ranging book, published last year, is about to change.* Welcome, he argues, to the world of “complexity economics”, computer-based simulations and more realistic assumptions. Mr Beinhocker has a measure of the complexity of the modern economy – the number of distinct products, or “stock keeping units”. In a stone-age culture the number was a few hundred. In today’s New York, he suggests, the number may be 10bn. Moreover, not just most of those products but the complex system that invented, designed, produced and sold them is largely the result of just the last 250 years out of 2.5m years of human evolution. The remainder of Martin Wolf’s column can be read here (FT.com subscribers only). Discussion from our guest economists is free.
Guillermo de la Dehesa: I must say first that I have not read the book by Eric Beinhocker but only Martin’s excellent article. My general comments, based on my early reading of previous analysis by J. Barkley Rosser and Kirby L. Kramer (1999) and by Paul R. Krugman (1996) are the following:
First, I am in general rather sceptic in general about all these “unifying theories” that intent to become applicable to all sciences and disciplines and which promise fundamental new insights, a paradigm shift or even a scientific revolution, like the four C’s: Cybernetics, Catastrophe, Chaos and Complexity but I recognised that they represent very important advances for the science in general. This attitude is due most probably to the fact that I do not have enough mathematical knowledge to even grasp them. The fact is that they have created fascinating metaphors such as the “butterfly effect”, “fractals”, “artificial life”, “edge of chaos” or “self-organised critically”, but, at the same time, they have been around for a long time since the early sixties and late seventies (in Brussels, Stuttgart and much later in Santa Fe), but they have not yet become what they expected to achieve. They all come from hard sciences, mainly mathematics (out-of-equilibrium dynamics with zero or multiple equilibria and game theory) physics (interacting particle systems) and also biology (evolutionary biology) and they use large computers but they seem to have great difficulties to be applied to economics, maybe because it is even more complex that the former ones, due to the problem of the interaction of humans beings, which do not exist in those previous sciences.
According to these theories, we live in a world of “bounded rationality”, that is, rational choice that takes into account the cognitive limitations (both of knowledge and computational capacity) of the decision maker, through a detailed and systematic empirical study of human decision making behaviour in laboratory and real world situations. And not so much of “rational expectations” given that there are many cases in which the postulates of neoclassical economics do not hold (i.e. there is money illusion and there are differences between the industry price and the general price level etc.).
Second, it is true that the idea that the economy is complex and that may not be characterised by perfect foresight and equilibrium comes from Hayek who was an independent developer of complexity theory in a way resembling its current forms (but without the use of computers). The reason why he applied it to economics was that he had significant communication with Ilya Prigogine and Herman Haken, (the founders of the Brussels and Stuttgart Schools).
Third, one of the deeper arguments made by these theories, based on evolutionary biology and self-organizing structures and systems, is that there are processes by which a global structure can emerge from strictly local effects and these processes represent how life evolved out of non living molecules and how multi-cellular organisms evolved out of single-celled organisms. Thus, the process of economic development and economic institutional evolution is seen to arise from such self-organizing emergent structure phenomena. The “self-organized critically” approach is also known as the “sand-pile model” which dynamics are like those in a pile of sand being built by randomly dropped grains of sand. As the pile of sand grows, it gets further away from its long-run equilibrium of being flat (out of equilibrium dynamics) and from time to time a single drop of a grain of sand triggers an avalanche that restructures the sand pile to a new state of self-organized critically.
Fourth, there are two implications of these theories and models for present economic theory: They are “strong Keynesian” (versus the “weak Keynesian” models of Mankiw and Romer). Their chaotic or out-of-equilibrium dynamics and the existence of multiple equilibria (by which systems can converge to equilibrium or not at all) imply that neither rational expectations nor Walrasian market clearing, which are two cornerstones of the “New Classical” model, are realistic assumptions. And there are some implications for economic policy:
The dismissal of rational expectations and markets that clear continuously, plus empirical uncertainty and multiple equilibria, makes life very difficult for economic policy makers because they do not know what is the appropriate model or adjustment mechanism, due to cognitive limitations or computational difficulties. Even their efforts to carry a stabilization policy may itself generate other complex dynamics which can become unpleasant. Thus there is a clash between the Austrian School followers, who do not find it so difficult because they believe that the economy will again end up, in a natural way, self-organized effectively and the Post Keynesian who argue that it depends on the institutional structures of the economy that may bound its tendency to complex dynamic fluctuations.
Fifth, nevertheless, the real world is showing that there are, apparently, mechanisms which keep economies within certain bounds most of the time, if not necessarily convergent on equilibria, on much less optimal equlibria, so that the economy can fluctuate reasonably well within some bounds, (“the corridor of stability” as shown previously by Leijonhufvud) possibly in a quite complex manner, but will behave in a dysfunctional manner if it exceeds those bounds.
Paul Krugman, who confesses to be an “evolutionary biology groupie”, shows that, by contrast to the popular and better known but light biologists, most of the greatest evolutionary biology scientists (such as George Williams, William Hamilton or John Maynard Smith) do not depart from the neoclassical paradigm. According to him, economics is basically, about first, what individuals do, that is, about individual actors and individual behaviour and not classes or correlation of forces or interests. Second, is also about self-interested individuals, people who care about themselves and their families. Third, is about intelligent individuals who do not neglect obvious opportunities for gain and fourth, is about the interaction of such individuals and about “invisible hand” processes in which the collective outcome is not what individuals intended.
He then asks himself, what is evolutionary theory of serious biologists about? The answer is that evolutionists share three of the four concerns of economics. Their field is about the interaction of self-interested individuals (the selfish gene), who are often thought of organisms trying to leave as many offspring as possible or genes trying to propagate as many copies of themselves as possible. Even when a bird finds a predator approaching and issues a warning cry, that puts itself at risk, but may save its neighbours, it is still a selfish action because most of its neighbours are likely to be relatives. The main difference is that economists routinely suppose that these individuals are smart and rational choosing the best strategy, evolutionists assume that organisms and genes are myopic. He then considers that both economics and biology are much more similar than economics and physics and it is wrong to attribute the inspiration of economic theory only on physics. For instance, economic agglomeration models works the same than the genes of the female peacocks
Even more, even though evolution is a theory of gradual change, of myopic dynamics, in practice, most evolutionary theorists focus on the presumed end result of such a dynamics: an equilibrium in which individuals maximize their fitness given that other individuals do. Thus, they resort to individual maximization and equilibrium. Why? They adopt the useful fiction that individuals are at their maxima and that the system is in equilibrium because, even if they now can study complex dynamics using computer simulations and that these ones tell them that they cannot safely assume maximization and equilibrium, these computations are so extremely complex and difficult that eventually they make use of maximization and equilibrium as modelling devices to cut through complexities. Evolutionary dynamics makes life of the researcher much harder, not easier.
Thus, we should be waiting for the future dynamics of computer scientists to be able to use and understand properly what is all about complexity and evolutionary theory.
Paul R. Krugman (1996) “What Economists can learn from evolutionary theorists”, A talk given to the European Association of Evolutionary Political Economy, Mimeo November
J. Barkley Rosser and Kirby L. Kramer (1999) “On the complexities of complex economic dynamics”, Journal of Economic Perspectives, Fall, Vol. 13, No. 4
Posted by: Guillermo de la Dehesa | January 22nd, 2007 at 11:01 am | Report this commentGuillermo has made a powerful posting, which I regard as an elucidation of the practical limitations of the evolutionary approach to economic analysis.
I do not disagree with anything he says. In particular, he is right to underline the point I made in my article, namely, that the great evolutionary theorists use very similar models to those of market economists. This is so because evolution is solving much the same problem as a business operating in an economy: how to obtain enough resources to survive. The big difference is only that human beings can plan a response, while organisms evolve.
I also agree that the economy is almost certainly too complex to model in a much more ‘realistic’ way. So, paradoxically, simple models with powerful assumptions often capture more of what is going on, more comprehensibly, than evolutionary, computer-based models.
My own conclusion is the one I made. This is the right way of thinking about how the economy works over time; and it explains how competition works. But it is not a substitute for standard analysis. We will never be able to capture economic phenomena perfectly. What we have to do, instead, is use intellectual approaches that capture the aspect that is crucial for our particular purposes, while understanding its limitations. There is, therefore, no one “right” model. But we do need to understand the implications and limitations of what we are doing in any particular case.
Posted by: FT Forum - Martin Wolf | January 27th, 2007 at 3:29 pm | Report this comment