Found in 1 comment on Hacker News
christofd · 2009-03-28 · Original thread
You keep on insisting that "more data" is a bad idea. Economics has a long tradition of learning from Economic History and comparative observation. I can't resist quoting Google here, that "more data" is the way to go, as Norvig argues (slightly out of context; field is machine-learning): http://www.computer.org/portal/cms_docs_intelligent/intellig...

I'm more or less arguing for that the IMF, or any other org. in those waters, has a lot of rich data in their possession. Well, here's an example: I worked for three years on a government-funded survey on "barter networks in Eastern Europe", where I did the number crunching on the data set... regression analysis, multivariate statistics etc. This resulted in one of the first extensive studies on contract theory in barter networks (http://www.amazon.com/Contracts-Trade-Transition-Resurgence-...). Basically in absence of cash to keep trade flowing you barter. Now if the U.S. financial system breaks down to the point where nobody can pay their suppliers and vendors etc. they'll have to pay each other with parts and other barter derivatives, passing around IOUs. Turns out, somebody looked at this and it works on a grand scale and even across borders. Anyway, there are lots of crazy ideas that you can try out and experiment with, looking at a rich data set of failed nations. I remember how expensive it was compiling all that data on barter networks (granted, we did the legwork; data was not from the IMF).

Regarding "abuse through power aggregation in developing nations" I don't have an example right now, on how that directly applies to the U.S. situation.

Fresh book recommendations delivered straight to your inbox every Thursday.