Modeling Madness
27 April 2009Some people try to light a candle. Some people curse the darkness. Me? Part of me wants to model the darkness.
I was led to this reälization upon reading the latest entry from zenicurean. In response to news reports about the latest swine-flu concerns, he writes
Plenty of first reactions appear to heavily involve doing things actual health care experts are not chiefly concerned about getting done, but that's how it always works, isn't it?
And I almost immediately thought about why those first reäctions are what they are. For example
- Officials want to be seen as doing something.
- People, including officials, often greatly over-estimate their understanding of issues that have (or seem to have) a significant bearing on general welfare.
- Officials with axes to grind are quick to find excuses for the grinding.
- Politicians can exploit the prejudices and desires of voters who are predisposed to support various measures (such as blocking foreign trade or travel, or subsidizing some profession).
So, could we pull this altogether, and surely other things that don't come so quickly to-mind, perhaps into a mathematical model, or perhaps into something less formal, that would have some predictive efficacy, or at least some distinctive explanatory efficacy?
Tags: politics, public choice, public policy, science, social science
What you are offering up here seems impossible. In economic terms, it's almost like asking a Keynesian to formulate a Hayekian model of the nuances of capital investment. A more simple approach would be to presume that all initial reactions, irrespective of any rational understanding of the situation, is that people tend to cover their asses. Whether it's economic crisis or the threat of pandemic flu, my empirical observation is that officials lie, people panic, politicians exploit, and the media wins.
Here in Tijuana, officials are claiming that everything is under control, half of the population is wearing surgical masks, politicians of the opposition party are snarking at the party in power, and the broadcast news ratings are soaring and newspapers are selling out.
Yet, there has not been a single case reported of anyone in Tijuana with this particular virus.
If you can model this, you're a certified genius.
It might well be impossible to get a useful model, but the bars by which I propose to judge such models are not perfect realism, but predictive and explanatory efficacy. The model can be something of a caricature, so long as it helps us either to predict behavior or to wrap our heads around the behavior that we observe.
I was more concerned with the possibility modeling the behavior of officials than of society more generally. I reälize, of course, that the behavior of officials cannot be modelled out of a wider context, but the behavior of the wider public would be treated as exogenous to the model — we neither predict it nor explain it, but just take it as a given. (If someone were later to extend the model to endogenize the behavior of other parts of society, then one applauds.)