Saturday, June 21, 2014

Time Horizons -- Development Edition

For those of you wondering, FLG still holds fast to his time horizons theory*, which was applicable again in this EconTalk interview with William Easterly.

Russ: Well, when you are steering the car, you dowant to know the GPS (Global Positioning System) to be very accurate. You want to get a measurement every second. So, the more often, the better. Guest: Well, the more often, the more often you are going to be very badly wrong. Because the short run data are just too unreliable. What those data are useful for are really to give you much more of a long run verdict. Which system is better: a system based on individual rights or a system based on technocrats and dictators? And there I think the long run evidence is very conclusive. The gigantic reductions in child mortality have come mainly in free societies, and not in autocratic societies. It came much earlier and much bigger in free societies than they came in autocratic societies. That's what I think the data is telling you. And so the data is really weighing in on this big debate that Bill Gates is completely ignoring and avoiding, about: Is the autocrat of Ethiopia actually a big part of the solution to development in Ethiopia? Or is he actually a part of the problem? I think the big long-run picture says he's actually part of the problem, not the solution. Russ: Well, we had an episode with Morten Jerven on the unreliability of data, often--in the long run and in the short run. So, one of my concerns is that in many of these poor countries, data collection--it's not so great in the United States sometimes. So in a wealthy country that has lots of resources-- Guest:Yeah. But it's not just the long run and the short run. You can't equate the two. The great saving value of the long run is that no matter how big the measurement error is--and it is very big--but if it's just unbiased and just sort of fluctuating up and down even by a large amount, over the long run that tends to average out. And so you do tend to get a more accurate measure even with measurement error when you go to longer run data. That's one thing this book really insists on repeatedly, is that to really get the right evidence for what's working in development and what's not. You really need to go very long run and not over-react to kind of short run episodes of improvements in either development or child mortality or whatever. Those are just really not reliable. But the long runis reliable.

* For those of you who are new around these parts, FLG's time horizons theory is an evergreen theme here and the short version is that conservatives place more emphasis on the long run and search for root causes while liberals place more emphasis on the short run and proximate causes.  This then leads to a few corollaries, but FLG is too lazy to list them all right now.

No comments:

Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.