Simulating the Market
By: Onur OzgodeHere is an interesting piece by a French theoretical physicist on how one might want to do security when it comes to the economic catastrophes. His starting point is no surprise: that neo-classical model of the markets do not make sense b/c they are not realistic. That is not how markets work…
Well, probably many of us would agree with him, but what makes this point not interesting is precisely “we” would also include the economists he blames. This is probably an important point to make, though not it is not an interesting one as far as positive knowledge goes: I am not sure if economists think in terms of classical Walrasian neo-classical market models of general equilibrium of supply and demand. It is true that economists will teach this in their undergraduate courses and will speak in these terms to the public or the politicians. Also probably policy advisors, technocrats, experts, and politicians find this language helpful because we are all familiar with it. But I do not think this is how they operate. I do not think they wake up and look at their screens and see if the market is in equilibrium or not. First of all, general equilibrium models left their place to partial equilibrium models. Second, with rational choice theory and other decision making models economists turned to new analytical techniques to analyze the economy since the 1950s. So, to me it seems that the language of market equilibrium is a heuristic discursive device for certain actors who are responsible for informing the public on the economy. This is however is not an unimportant task as we are right now seeing there is a huge gap between what actually these experts do and how they talk about what they do. One might argue that this is why the bailout failed, precisely because it was not a bailout but a rescue attempt to save the economy conceptualized as a vital system. It would also be interesting to think about how as the Fed was trying to manage the crisis what they called systemic risk and therefore rescued, and what they called moral hazard and how they let it sank. There seems to be a high level of imprecision and non-correspondence between the language and terminology of the Fed and what they do…
Second, if we actually come back to his interesting argument, we see that what he is proposing is not so far from the kinds of things we saw at OEP. His basic point is that we cannot rely on these ideal typical models and rather we should ground our regulatory practices on a realistic model of science instead of a formal one. Given that he is a physicist it is not surprising that his solution is “simulate the market” as opposed to model it. The funny thing with this proposition is that the two are not contradictory practices, but rather are mutually complementary. This is one of the interesting things with the experts who are in the business of simulation. Often times because they get their models from other experts already black-boxed, they forget that simulation rests on a diffuse network that includes both the theoreticians and applied scientists. Without models, i.e. a formalist scientific practice, simulations would not make any sense; it would simply be guess work. (We might want to think if there is a similar process in the case of catastrophe enactments and exercises that Andy, Stephen and others on this blog have been thinking about. In other words, how do we or rather the experts themselves know that we are basing our simulations on right kinds of scenarios? Are scenarios types of models in correspondence to the computer simulations?)
Finally, the article gives us interesting clues on how simulation is used to govern the economy in comparison to the early attempts we saw at the OEP. He gives 3 examples:
1) An agent model being developed by the Yale economist John Geanakoplos, along with two physicists, Doyne Farmer and Stephan Thurner, looks at how the level of credit in a market can influence its overall stability.
2) Professor Westerhoff and colleagues have used agent models to build realistic markets on which they impose taxes of various kinds to see what happens.
3) Charles Macal and colleagues at Argonne National Laboratory in Illinois and aimed at providing a realistic simulation of the interacting entities in that state’s electricity market, as well as the electrical power grid.
These can be interesting experiments to follow to see where practices of governing might go that are linked to the types of problems we would like to tackle on VSS…