The Institute for New Economic Thinking was founded two years ago with funding from George Soros. It held its first ever conference here in Cambridge, where possibly the greatest combination of leading economists ever assembled for one meeting debated how to reorientate the subject so that it did a better job of predicting, understanding and ideally avoiding calamities such as the financial crisis of 2007-09. Their website has a wonderful set of resources for learning about economics and hearing from leading thinkers.
The INET blog has a fascinating interview with Judy Klein, an historian of economic thought, about the links between military research and economics. I knew about some of this but it goes far deeper than I realised. Professor Klein is the author of a book called A History of Time Series Analysis 1662-1938, which I admit is not everyone’s bed-time reading.
The US military, during World War Two and the Cold War, hired economists to help solve problems such as optimising the targeting of bombs and missile trajectories. Modern economics is the “science” of optimisation but much of the development work was done for the military and then applied to economics. The maths of these techniques was applied to economics (and operations research and management science more generally, which share the underlying economics methodology).
Examples include the way that adaptive expectations macroeconomic models “originated in attempts during WWII to model information flows between gunner and analog computer in the lead computing gun sights of B-17 bombers.” The next generation, rational expectations models, arose from digital computing and the work by Richard Bellman, who developed dynamic programming as a way to “solve the Air Force problem in the late 1940’s of how to allocate scare nuclear bombs to competing targets in a potential multistage strike on the Soviet Union.” (Later there was no shortage of nuclear bombs so optimisation became unemployment. Once you can blow up the entire world more than once you have a different kind of problem to solve).
What I already knew was that the development of game theory, pioneered by one of the all time clever people John von Neumann, had its roots in thinking about nuclear strategy. Von Neumann and Oskar Morgenstern wrote a classic text called “Theory of Games and Economic Behaviour” in 1944. This covered zero-sum games, of which global thermonuclear war is a pretty salient example (actually it’s a massively negative sum game). The book also introduced post-calculus maths (topology) to economics, for which generations of graduate students have not thanked him. Game theory was then applied to the problems of deterrence, credibility and reputation that are at the heart of how two nuclear powers avoid destroying each other but also avoid giving in to nuclear blackmail. The fact that the world is still here is not so much due to game theory as the willingness of key people to ignore the theory at critical times such as the Cuban Missile Crisis in 1962.
Game theory went on to become a central part of economics, because it deals with interdependent decisions i.e. what is optimal for me depends on what is optimal for others. John Nash (of the film A Beautiful Mind) was another brilliant mathematician who contributed the “Nash equilibrium” concept to such situations.
The idea of “rocket scientists” in finance comes from stochastic control, a further development of Bellman’s optimal control work using Itō calculus, where the control problem involves uncertainty. Itō was (he died in 2008) a brilliant Japanese mathematician who was celebrated for his contribution to financial maths but (according to Wikipedia) could barely remember even inventing the technique that bears his name. Stochastic control is used to guide the path of intercontinental ballistic missiles, hence rocket scientist. The maths works beautifully (but with great difficulty for most graduate students) for financial analysis and is used in the rigorous derivation of the Black-Scholes formula done by Robert Merton (Black and Scholes didn’t know how to do stochastic calculus).
I was once able to solve simple stochastic control problems but I am now, to say the least, a little rusty. But then I don’t teach macroeconomics any more and the dynamic stochastic general equilibrium (DSGE) models that use this approach were completely useless in forecasting or even understanding the financial crisis. Which is why the INET was set up. A knowledge of the writings of John Maynard Keynes’s writings and basic national accounting get you much further in practice than abstruse maths.
I once worked with an equity analyst who covered the giant Russian company Gazprom (and to a first approximation, Gazprom IS Russia). A Ukrainian PhD in nuclear physics, he was once referred to by a colleague (respectfully) as a rocket scientist. He paused, looked unimpressed and said with magnificent disdain, “Rocket science, trivial.”