« Who should be bounced from The Great Books series? | Main | The best sentence I read today »
What works in development economics?
Seven papers, all by first-rate authors on topics of critical importance. I haven't had time to read these yet but I can recommend them nonetheless. I'll print them out once I'm back from Japan in the meantime they are yours.
The pointer is from Chris Blattman.
Posted by Tyler Cowen on May 30, 2008 at 10:01 AM in Economics | Permalink
Comments
Tyler Cowen surely is awesome, he can even recommend papers without having read them. Only the wisest can do this stuff.
Posted by: Sécessionniste at May 30, 2008 11:38:05 AM
No, it is not wiseness of Tyler, but the reliability of Mr. Blattman that makes me believe that these are worth reading.
Posted by: Revisionist at May 30, 2008 1:24:27 PM
I read the "High Bandwidth Policies" and it seems to be on the right track.
Posted by: Matt at May 30, 2008 3:56:39 PM
It's like all the development economists are from Harvard or MIT. A bit worrisome from the viewpoint of intellectual diversity - these people only talk to each other and end up thinking the same thoughts.
Posted by: RW at May 30, 2008 5:07:12 PM
Regarding the high bandwidth paper.
The author talks about the possible paths to extend the variety of traded goods as being partner to a productive function change in thee trading partners.
So, if we imagine two trading partners with different wealth distributions, but the distributions are sparse. Why sparse? The authors hint at production functions that hold inertia. The two distributions tend to merge, the combination of the two (weighted RMS variance, and weighted linear mean) becomes more smoother.
Then they talk about directed graph representations of the ability of one wealth distribution to expand its product variety. Lot of data, really good.
Over time production functions of the trading partners diffuse domestically in response to differential demand from the other. Like molecules, shorthand. The authors say "We become what we trade"
Assume reversibility, we can say the two economies co-diagonalize by using exchanges as linear updates to a matrix diagonalizing procedure. So we can see that as the production function become close to equilibrium the next level of precision needed will go up as increasing the rank of the matrix goes up, which is going to be NlogN exchanges. In directed graph terminology, this is branching right and filling in left with a sub branch.
So, do you believe in state changes that directed graphs imply? Expect many restructures based on global trade, most of them good changes. We have a compulsion we cannot disobey, which is to get a more accurate count. So we decide between getting the next precision with the current production function, or breaking some of the aggregation and "re-graphing" to get a better density of fill.
Posted by: Matt at May 30, 2008 7:01:04 PM
