Sunday, May 06, 2007

A Good Question

Dani Rodrik asks a good question:

I asked the audience a favorite question of mine: would you rather be poor in a rich country, or rich in a poor country. I gave them the following terms for thinking about the question:

  1. Assume you care only about your own consumption
  2. Define poor and rich as someone who is the in the bottom or top decile of a country
  3. Define poor and rich country analogously as a country in the bottom or top decile of the distribution of per-capita incomes across countries.

The audience was divided evenly between those who would choose to be rich in a poor country and those who would rather be poor in a rich country. The real answer is that it is not even close. This has important implications for how we think about poverty reduction in the world.

Can you guess what the correct answer is?

I'm embarrassed to say - despite having taking the odd lecture on this sort of stuff - that I'm not actually certain of the answer; my guess is:

My guess is poor person in a rich country. My answer might have been different were it not for the terms (points 1-3 though).

Because of term 1, it's absolute consumption that you are concerned about. This makes the exercise slightly counter-intuitive as most of us are inclined to think in relative terms (and, indeed, happiness and health research suggests that one's relative position is actually quite important). We're also required to exclude things like social capital, which negates the potential benefits of living in a gated community in Lagos rather than in a poor part of Otara (to use an example from my home country.)

Term 2 requires top decile, whereas, in really poor countries, the very wealthy are a much smaller percentage of the population than this.

Term 3 'bottom decile of countries' knocks out most (if not all) of Latin America where the wealthy are very wealthy indeed.

Just out of interest, are we adjusting for purchasing power parity here, or using market exchange rates?
One thing I am certain of though, is that - as I note above - the conditions of the 'experiment' matter.

UPDATE: Phew, I got it right.

No comments: