This post is part of a series from the TED 2009 conference held in Long Beach, California from February 4-8th. You can read other posts in the series here, and the TED site will release video from the talk in the coming weeks or months. Because I’m putting these posts together very quickly, I will get things wrong, will misspell names and bungle details. Please feel free to use the comments thread on this post to offer corrections. You may also want to follow the conference via Twitter or through other blogs tagged as TED2009 on Technorati.
Bruce Bueno De Mesquita can predict the future. He’s not the Great Karnak, he tells us, nor does he examine the entrails of sheep. He believes that we can predict future events via “rational choice theory“, a technique he sees as vastly superior to the “seat of the pants wisdom” most politicians and economists use.
This theory, the tells us, applies to situations of persuasion and coercion – that means that’s it’s applicable to politics and business negotiations, not to markets. His work is rooted in game theory, which means math has come to politics, he warns us. (I think we’d worked that out with Nate Silver, thank you very much.) Game theory requires an assumption that people are rationally self-interested and aware that others have motivations and rational self-interest.
Everyone, he tells us, is rational, with the exception of two year-olds and schizophrenics. Corporations are rational as well – while it looks hard to convince a corporation to stop dumping toxins, we can do so if we can explain why it’s in their interests.
Understanding what the President wants to do regarding Iran misses the point. The President is influenced by his secretary of state. She’s influenced by her advisors on Iran. To understand US policy, we need to understand that whole complex web of relationships.
These relationships, he asserts, are too complicated for humans to understand without assistance. He points out that ten people can have 3.6 million interactions and suggests that only computer models can help us understand what decisions are likely to be made in these interactions.
In building these models, we need to know:
- who has a stake?
- what they say they want – not what they actually want in their heart of hears, but their strategically chosen positions
- how focused are they on these goals?
- how much power or clout do these people have?
This information can be filled in from public sources – the newspaper of the Economist, or from subject experts. You don’t need history, he tells us – just these easily available inputs?
Offering a 90% success model for these techniques, we looks at Iran:
- How secure is the theocracy?
- Where will the nuclear program go?
- Will Ahmedinejad remain in power?
He sees a decreasing chance that Iran will build bombs – instead, he sees scenarios that has Iran building only sufficient quantities of nuclear material for research purposes. He sees this as a liveable equilibrium for the US, and more likely to occur in the absence of international pressure. This analysis is based on a computer model of 87 Iranian decisionmakers, considering public polling data. He points out that virtually no one wants to test a bomb in Iran, and that his models of power structures suggest that the people who want to test a bomb are moving out of power.
The model is based on tracking the kremlinology inside Iranian politics – he offers a graph of rising and falling fortunes of figures like Ahmedinejad, moneyed interests and different groups of clerics. His money is on “the Quietists”, a group of clerics based in Qom with great influence and concerns that Iran is moving away from the values of the revolution.
de Mesquita tells us that people who say “That’s impossible” really mean “I don’t know how to do it.”