Andrew Haldane (via Arnold Kling writes about one of my favorite subjects: complexity. Ezra Klein also has a post about complexity that makes absolutely no sense, and I’m going to combine analysis of both articles. At the risk of being redundant, I’m going to quote the same passage as Kling:
Complex control of a complex system is a recipe for confusion at best, catastrophe at worst. Complex control adds, not subtracts, from the Knightian uncertainty problem. The US constitution is four pages long. The recently-tabled Dodd Bill on US financial sector reform is 1,336 pages long. Which do you imagine will have the more lasting impact on behaviour.
Taking a cue from evolution, the most effective regulatory design is to have a few hard rules, and then allow the people to search the infinite design space of all possible Business Plans that conform with those rules to find optimal strategies, and then allow the market mechanism select for and amplify those business plans which suit consumer tastes. I believe that this is the model that Mr. Haldane has in mind…and I agree.
However, is it complexity that broke our banking system? Well, it’s worth a look at what complexity means in the context of networks. When I think about complexity, I like to use a simple definition that states that a network is complex when adding another node at the margin increases the number of interactions within the network exponentially. For a more in-depth discussion of complexity in networks, see this article.
That definition certainly fits our banking system, which is deeply interconnected, and is also deeply connected with NGDP growth levels. Indeed, no matter what your story of the recent recession, one thing is clear — troubled assets (or what would become troubled assets with falling NGDP) had a contagion effect, and found their way onto the balance sheets of many different primary and secondary financial institutions, and spread to affect other institutions that weren’t financial in nature. Of course, there is an important caveat to this. Densely connected networks are very robust, and due to the nature of the network being densely connected, a change in one part of the network has an exponential effect on other parts of the network…and once the network hits a tipping point, it falls into complexity catastrophe very quickly. But there is an important limit to the growth of these types of networks in the economy, their degrees of freedom. The fact that large institutions have exponentially declining degrees of freedom is the reason nimble upstarts are able to out-compete large, established firms in a “David and Goliath-type” contest.
This should have created a natural limit on the growth of the individual firms within the financial “ecosystem”…however, the “ecosystem” was made of up “Evergreen trees“. By that, I mean the financial system is dealing with homogeneous inputs, and generally homogeneous output — money. Our money system is extremely efficient…so efficient that in the world’s Forex markets alone, $4tn is traded daily. This, in effect, “creates a forest of all evergreen trees”, which as noted in the previous article, makes the ecosystem brittle and susceptible to various sorts of catastrophe. So the problem is indeed systemic, as it is a feature of the very money system we use, which is unconsciously designed to maximize efficiency by an implicit top-down, concentration structure.
So Mr. Haldane is correct in saying that dealing with a complex system by constructing a complex system that surrounds and intertwines it is definitely not an optimal strategy. However, implicating the complexity of the system as the cause of the problem (as Ezra Klein does) is shockingly wrong-headed. Instead, we should implicate the mechanisms which serve to maximize efficiency, and create homogeneity in the cause of problems. In particular, our money system.
So in the “Too Brittle to Sustain” model, it is the desire to maximize efficiency that is at the very root of the problem (not, however, the proximate cause). It is about time we take a lesson from the wisdom of ancient societies, and design alternative money system which complement (not replace) or current system. By instituting this type of diversity, we can better operate within the “window of viability” which govern all natural systems.
Need proof? Switzerland is legendary for its stability, however, there is nothing inherently “Swiss” about it. They operate a very robust complementary currency system, which works countercyclically to stabilize their economy. They fared very well in the recent recession. I know I risk taking this point too far, as their independent currency certainly plays a role…but again, it plays to my bias =].