Since Bill Easterly stuck his head above the parapet last week and referred to complex systems in response to Paul Collier, the “complexity” community has been up in arms. In a quick reference to complexity, Easterly wrote:
A popular topic in the aid blogosphere this week was not about Haiti or Ivory Coast or south Sudan but about complex systems, i.e. systems that cannot be reduced to a simple mathematical or statistical model, where actions often have unintended effects.
He then (rightly in my opinion) questioned Collier’s ability to predict the consequences of supporting a coup, including Collier’s chain of predictions about the actions of various levels of Ivory Coast army officers. With what confidence can the course of a civil conflict be predicted? Not much if previous conflicts are any guide.
Although I agree with Easterly’s analysis, I did see his use of complexity as a touch out-of-place. The article has generated some great discussion in the blogosphere on what complexity is and whether it is relevant to aid. It was watching the idea be thrown around and in the course of a few days some order appears to have emerged in the understanding of the concept. I feel that I have learnt a bit myself, so I’ve put together the below on what I thought were the most interesting points or issues.
First, complexity can be simple. As Philip Auerswald said:
[T]he core insight of the study of complexity …. is this: systems that are not just reduced to, but actually defined by, simple mathematical models, have the potential to generate extremely…well, complex behaviors.
For examples of this, I recommend Thomas Schelling’s Micromotives and Macrobehavior as a first stop.
Second, and to me the most interesting point, comes from Auerswald’s likening of Easterly’s use of complexity to previous attempts by economists to tie complexity theory into Hayek’s concept of spontaneous order. While I consider that Hayek’s work provides some early thinking on complexity, Auerswald suggested that bringing these ideas together is a failed undertaking from the outset, with Hayek’s concern being about calculation in the presence of randomness, not the emergence of complexity in the absence of randomness. That distinction is important, and the problem identified by Hayek was fundamentally one of calculation (although the calculation problem was about more than just randomness). However, I would argue that there is a more important distinguishing feature. Hayek saw a benevolent force (or invisible hand) creating a more efficient system than humans could create by central planning. A complexity theorist might argue that there is no such benevolent force and that human interference may be required for an efficient outcome (this distinction was the central conclusion of a paper by Kilpatrick). If we look at Easterly’s argument under this distinction, it is much more Hayekian than of modern complexity theory.
The third point, and flowing directly from this distinction, is that if a system is complex, that (in itself) does not mean that we should not touch the system. We can see this in some of the work that has come out of the Santa Fe Institute. When you start throwing around concepts such as path dependence, increasing returns and out-of-equilibrium dynamics, one can argue that the current state of affairs is not ideal and a few “tweaks” from above might help. I don’t generally agree with that argument, but you need something more than “it’s too complex” to respond.
Fourth and finally, I don’t think this third point matters in the current debate. Complexity is not the major problem. As stated by David Ellerman in response to Easterly’s update on where the complexity debate is at:
The mistake in applying complexity theory to human relationships such as the education, management, development aid, and helping in general is that the basic problem is NOT that the human “systems” are complex, “messy,” nonlinear, etc. The basic problem …. is that success lies in achieving more autonomy on the part of the doers, and autonomy is precisely the sort of thing that cannot be externally supplied or provided by the would-be helpers. This is the fundamental conundrum of all human helping relations, and it is the basic reason, not complexity, why engineering approaches and the like don’t work.