Aid Thoughts

No, not poverty, but pareto improving:

However, the new index, like all current poverty measures, fails to address a major problem. It looks only at outcomes, not process. As I have suggested before, this is insufficient to tell us what is really happening to poor people. In Colombia, for example, displacement is an overriding problem, with internal refugees reaching levels expected in countries such as the Democratic Republic of the Congo.

As well as counting poor people, researchers should apply the following ethics test to development strategies: if you displace 100 families from their land, is it possible that poverty indicators will actually improve? If the answer is yes, which it often is, there is something wrong or limited with the indicators. Perhaps some brilliant academic can devise a measure of poverty reduction that acknowledges this issue.

Every poverty indicator or well being measure you can think of has an implicit welfare function supporting it - i.e. there is a value judgement being made: how much you value outcome X vs outcome Y, or how much you care about relative deprivation, etc. Most, as Glennie points out, are worried about aggregate improvements, and so often don't capture individual loses and gains. To heavily discount a poverty measure if someone has lost out suggests a concern that progress should be pareto-improving, which is econospeak for not making anyone worse off.

Indicators that are too concerned with this measure are likely to lead to paralytic policy making, but I suspect Glennie is less concerned about loses in welfare overall, and more with that of the already-poor. This is more reasonable, and suggests a more Rawlsian take on measuring poverty. Rather than accounting for process directly, wouldn't it make more sense just to make our poverty measures extra sensitive to severity, thus discounting trade-offs which leave very poor people even worse off, even if leads to a net improvement in traditional measures of poverty.

To answer Glennie's question - yes, I can think of situations where it would be possible that displacing people would improve poverty indicators. I don't recall ever hearing anyone claim that development was a series of strictly pareto-improving moves, nor even ones that lead to less inequality in the short run. That said, we should pay much more attention to poverty dynamics, as often aggregate improvements mask a lot of downward-movement (with some people escaping poverty and others falling back in). This might not be the same thing as looking directly at process, but it would make us a little more sensitive to policies that throw a lot of people back into poverty, even if leads to a net reduction.