Saturday, June 16, 2007

flailing

Not Li'l solipsist

I read an interesting article over at The Republic relating to forecasts of economy, human behaviour, triggers to calamities, etc. It is redacted below.

The conclusion that I drew was: don't listen to the boffins! They are just flailing like little babies.

Theoretical physicist Mark Buchanan’s Ubiquity: The Science Of History . . . Or Why The World Is Simpler Than We Think (Phoenix, 2000), is a book whose implications are so powerful they’ve given me vertigo. Buchanan makes a very strong argument that, despite all appearances to the contrary, human history may be governed by very simple mathematical laws. Unfortunately, those laws would guarantee history’s complete unpredictability, and they would also fundamentally contradict the economic and political theories informing our decision-makers.

The study of non-equilibrium statistical physics shows that critical states seem to emerge whenever systems are kept far from equilibrium in conditions where the forces of chaos and order are in constant flux, and where the components of the system exert some influence on each other’s behaviour. It seems that at least some features of our collective behaviour can be reasonably explained by these concepts.

Might as well just guess

Research suggests that both stock market fluctuations and the distribution of global wealth follow power laws. This is an unnerving finding, as it flies in the face of traditional economic theory,...the effects of a relatively minor economic variable may produce anything from unnoticeable effects to such calamities as the 1929 stock market crash and the collapse of the “tiger” economies of Southeast Asia in 1997.

This helps explain why economic forecasts are so consistently proven wrong. Buchanan writes that in 1993 the OECD “analyzed forecasts made between 1987 and 1992 by the governments of the USA, Japan, Germany, France, Italy, and Canada, as well as those of the International Monetary Fund and the OECD itself. Their conclusions? Not only were each of these organizations’ predictions abysmally inaccurate, but they would have made better predictions for inflation and gross domestic product if they had scrapped all their sophisticated economic models and simply guessed that the numbers in each year would be unchanged from the last.” Similar studies have produced the same results. This has grave implications. Given that the study of economics presupposes equilibrium, and given that our nations’ powerbrokers rely upon economists for their understanding of the world, it seems that our leaders’ most basic assumptions about the economy are dangerously wrong-headed.

Indeed.

5 comments:

Unknown said...

Interesting article.

Makes me think that whenever you want to make something more predictable - you should make it a Science.

After this, by printing out a couple of books on the subject, you change how people think. And that's when our collective behavior kicks in.

What you end up at the end is a bunch of rules and laws, the rest you call illegal activities.

go figure.

Unknown said...

Sadly, the social sciences.... aren't.

Humans are too irrational to model accurately.

That said, it is a little comforting that we can be explained in aggregate after the fact :)

Making something a science for the purposes of imposing boundaries on people's thought is a neat idea though, although not a new one. Greg Egan's Teranesia novel briefly touches on the idea of using the same approach to limit the effectiveness of social movements.

Anonymous said...

Nassim Taleb has written extensively on this over the years. The way he draws it is that our modeling generally assumes we live in Mediocristan, but in reality we live in Extremistan.

His tone can be...interesting. And I don't quite get his fascination with Mandelbrot. But the message is powerful and resonant.

solipsist said...

Thanks for the comments, and adding to my must-read list.

Anonymous said...

Cute picture