Sunday, September 30, 2007

Book review: The Social Atom

Mark Buchanan: The Social Atom - Why the rich get richer, cheaters get caught. and your neighbor usually looks like you

The main thesis discussed in this book is that social science and economics is flawed because it looks at people as individuals, instead of groups behaving in patterns. Buchanan claims that if you look at individuals as atoms interacting with each other. you'll realize that people behave in ways (patterns) that can be observed to be like other patterns occurring in nature, and thus it's possible to predict what is going to happen in e.g. the market.

All of this is of course interesting, and Buchanan does make a good case for it, but I have some serious issues with the book.

If you've read Buchanan's book, Nexus, it'll come as no surprise to you that Buchanan suffers from a severe case of physicist glorifying, believing that they are experts at overturning dogma in existing non-physics fields. This comes through even more clearly in The Social Atom, where Buchanan makes it clear that he doesn't think much of economic theory, which is generally based on flawed premises. Buchanan does well at explaining why, while giving a simplified introduction to economic concepts like "rational individuals".

The problem is that he seems to not realize that his simplified version is not the full version, and that economists are fully aware of the problems with these concepts.

No economist, except perhaps members of the Chicago School, really believes that people are/act rational (or that markets are really transparent), but because economic models are incredible complex, it's been necessary to operate under these flawed assumptions, as to be actually be able to use the models at all. In other words, while the assumptions are flawed, it makes it possible to make an approximation of reality.

As computers have become more and more powerful, it has become possible to re-evaluate these assumptions, and dismiss the flawed ones. This could of course only be done, as the reality underlying those concepts became apparent through computer models/analysis, which is the very process Buchanan describes in his book. In other words, Buchanan complains of economics using simplified models/abstractions, while explaining the very process of possible for economists to not use them.

It would be like me complaining about physicists using a simplified model of the universe (or the atom, or gravity, whatever), while explaining how it is now finally possible to make a more complex (and correct) model, and what that model shows.

Another problem I have with the book is that it's light on science, and while it often mentions that something (the market, people in cities etc.) behaves in a certain way that's similar to a patten observed in nature, it usually doesn't explain why it behaves that way. That's problematic when you're writing a book about a descriptive field, which economics is.

All in all, I'm quite lurk warm towards the book. It has some interesting ideas, and it's pretty well written. However, if you, like me, have any type of background in economics, you'll get distracted by the flaws and oversimplifications. On top of that, you'll probably have heard about most of the experiments mentioned in the book.
In the end, I won't say that you shouldn't read it (unless you are an economist), but I won't recommend it either.

Labels: , , ,


Anonymous 'Tis Himself said...

In the end, I won't say that you shouldn't read it (unless you are an economist)

I am an economist so I probably shouldn't read it. Based on your review, I probably won't.

Economics is built on models. Even the basic paradigms, like supply and demand and guns or butter, are models. We've been trying to refine the models with some success. One reason why the Austrian School is so out of favor these days is their refusal to modify their models. Ideology ├╝ber alles. But most economists recognize we're describing models which hopefully have relevance in the real world.

December 26, 2009 2:04 PM  

Post a Comment

<< Home