Intellectual Property Law

Useful Models

By Mike Madison

From the Lone Gunman, quoting The Truth About Markets/Culture and Prosperity (UK/US title respectively), a”two-paragraph excerpt from a section discussing ‘large models purportedly descriptive of entire economic systems’ (pp. 193-194)”:

The error of principle?the reason these models will never be useful?is best exposed by Jorge Luis Borges? story of mapmakers who competed to build the best possible map. They eventually understood that the most accurate map simply replicated the world. The search for realism destroyed the purpose of the map. A map is valuable precisely because it simplifies and omits. Economic models are maps for the market economy. A map can be false but never true. Our criterion for selecting among maps that are not false is usefulness, and a map can be too detailed or not detailed enough. We seek the simplest map adapted to our purpose, and it is a different map if we are walking or driving: not better or worse, but more fitted for its use. The London Underground map is a brilliant design for its purpose but useless to pedestrians. The ?little stories?, or economic models, of this book are to be judged in the same way.

I once debated the relationship between the social sciences with some anthropologists. We adjourned to the pub, and someone bought a round of drinks: the discussion naturally turned to the reasons why. For the economists, the explanation was obvious: the practice of buying rounds minimized transaction costs, reducing the number of exchanges between the patrons and the bar staff. The anthropologists saw it as an example of ritual gift exchange and described the many tribes that had developed similar customs. I proposed a test between the competing hypotheses: did you feel cheated or victorious if you bought more rounds than had been bought for you? Unfortunately, the economists and the anthropologists gave different answers to that question.

No Tags

Full post as published by Madisonian.net on April 18, 2009 (boomark / email).