Skip to content →

More Like a Bow and Arrow Than Like a Transistor… has a proposal

Can Science Help Solve The Economic Crisis?…

for a group of scientists and economists to collaborate in solving the economic crisis.

I have a comment in the debate there:…

which contains this:


This is a noble proposal, but I remain a bit of a skeptic with respect to the ability of a cohort of scientists and economists to find a scientific solution to the problems of our economy. Economies are living organisms, about as old as the oldest profession, and rebuilding the economic system from scratch is a problem in engineering and social engineering, not in science. Humans and scientists don’t have a good history as regards social engineering.

Science is reductive, and seeks to establish the laws which govern physical systems. To do so, scientists carry out repeatable experiments. For an experiment to be approximately repeatable, history has to be unimportant, and so the system has to couple very weakly to the rest of the universe. You can figure out the statistics of a coin flip because initial conditions of the coin are more or less irrelevant. The hand of the coin flipper, the temperature of the room, the location on earth, don’t matter too much because the coin is coupled weakly to the rest of the world, and hence to history.

Engineering is constructive, and tries to build little almost-isolated universes (usually called machines) which obey the laws of physics and chemistry and more or less do what we want them to do, in certain regimes. You can do engineering without knowing too much science but by having a good empirical understanding of how matter behaves (building a hammer or chisel or lever or bow and arrow), or you can engineer it using scientific laws (using quantum electrodynamics and solid state physics to build transistor radios).

Now let’s think about economics and the behavior of markets and the people who comprise them. First of all, we don’t have the scientific basis for engineering economics. Second, it will be very difficult to find the scientific laws governing the behavior of economies, because there are very few isolated economic machines. There is one large economy, more or less. Since economies aren’t isolated, you can’t carry out the repeated experiments that science requires. History is important in economics, and in human behavior in general. You cannot replicate the initial conditions over and over again, as you can with a coin flip. Credit markets tomorrow won’t behave like credit markets last year because we have learned what happened last year, and you cannot get back to the initial conditions of a year ago. Human beings and societies learn; physical systems by and large don’t.

Though I came to finance as a scientist, my experience working in the financial arena has taught me to be very humble of scientific claims and of big universal ambitions. Whenever we make a model of something involving human beings, we are trying to force the ugly stepsister’s foot into Cinderella’s pretty glass slipper. It doesn’t fit without cutting off some of the essential parts. You have to understand that you do need models—you can’t think about finance and economics without math and models but you have to understand too that models are not the world, at least not in the social sciences, and so the models have to be simple, shallow even.

I would add that the economy we have been part of has not been treated as an efficient market, even by the economists who claim that it is. Every time the economy has suffered a threat, the Fed has eased credit to reignite it, but whenever it has boomed they have never tried to quench it very hard. It is an oddity that the only solution people can think of now to the credit crisis, which was in large part caused by too much easy credit rather than by scientism or mathematics, is to ease credit again and have the government spend money on infrastructure and public works. This is probably the right thing to do, but it shows you how counterintuitive and complex the system is, and how important history in fact is: first time around easy credit is the problem, second time around it’s the hoped-for solution.

So, let’s try to figure out how to do things better, with words and thoughts and with math too. More uniform regulation, more transparency, less leverage are going to be critical. I’d guess that in the short run the robust solution will lie in a proposal written with words rather than with equations, and be more like a bow and arrow than like a transistor.

Published in Models