Excerpt

WILLIAM TRACY Our last academic panel of the afternoon is focused on economic architectures, which will include some discussions of networks, institutions, and structures within the economy. Obviously, these are all people that we heard from already in the program this weekend. I want to push each of you to think about the areas that really need to be developed next. Particularly, feel free to sort of transcend that positive economics/normative economics boundary as some of your talks did last time. So, not just what’s needed for economics, but what we need to do in the way that we think about economic systems to better the world. Maybe, with that very small request, I’ll start with you, Brian.

W. BRIAN ARTHUR Okay. I want to enlarge on something I said last night. Actually, it’s come up in some of the other panels. In spite of a lot of what we know in economics, there still have been financial failures, economic failures, policy collapses, all kinds of things. The Euro was launched as a common currency quite a few years ago in Europe, and that turned out to be something of a disaster for Greece and Portugal and Italy, possibly also Ireland.

These things were not particularly well anticipated. When policies or policy systems are put in place, what economists tend to do is simulate and try to figure out what are going to be the consequences of that policy. That’s a perfectly sensible thing to do.

But if you think in terms of equilibrium, as I said, there’s a subtle bias with all of us economists. If things are in equilibrium, there’s no way to improve on that, otherwise it wouldn’t be an equilibrium, so we don’t think in terms of people gaming such systems.

With complexity economics, if we threw out this assumption of equilibrium, suddenly everything becomes a web of incentives that can be gamed or can be exploited in some way. Again, I don’t want to be cynical, but I love this quotation from Janet Napolitano. In 2005, she was governor of Arizona, and people were bringing up the idea of a fifty-foot wall. So this is an older idea than the current administration. And she said, “Show me a fifty-foot wall and I will show you a fifty-one-foot ladder.”

So, my question and challenge for economics, or for complexity economics, or even for SFI—it’s not a huge challenge, but it’s doable and it could be part of some program in the future—is, can we anticipate how such systems could be gamed, or played, or taken advantage of, and thereby collapse?

This is something medical professionals understand extremely well. In virology, in any sort of disease control, certain other organisms can get into the system, take it over, and cause severe difficulties, if not total collapse and death.

How would we get systems to do that? These would be intelligent systems, systems that can learn. A trivial example I thought of was, “Well, I could set up a model of fifty-foot walls. And then, trivially, I could feed it the idea that people could wheel up ninety-two-foot ladders or thirty-nine-foot ladders, or maybe fifty-two-foot ladders. And that would converge and discover that fifty-one-foot ladders would do the job just about right.”

Now that’s cheating. It’s trivial because the whole qualitative idea’s been thought of. So my question, I want to put it to you—I thought a lot about this. I’ve written a paper called “All Systems will be Gamed.” This was shortly after the collapse of 2008/2009.

BACK TO Complexity Economics