News

Let’s avoid creating a dog’s breakfast of MDGs

Inclusive Economy,Poverty & Inequality28 Feb 2013Martin Ravallion

The poverty reduction goal already embodies inequality. Even if we agree that it under-values things, it is far from obvious that adding an overall inequality measure is the best corrective. We need to think clearly about what is missing and how best to measure it.

Setting goals is not inherently silly, but it can sometimes turn out that way. If a goal is neither too easy nor too hard it can motivate action, which is a good thing if the goal is desirable. As Stephan Klasen pointed out in a post on this blog, it is hard to say how much of a reduction in inequality is in fact desirable. It is possible for there to be too little income inequality from the point of view of reasonable concepts of development, and even distributive justice. That is clearly a problem for those who advocate a new Millennium Development Goal (MDG) for reducing “inequality.” It is important that there is a broad consensus about “which way is up”, but it seems unlikely we will reach such a consensus for inequality.

But let’s put that point to one side, and take it as given that there is too much inequality in the world today – I know I am one of many who think so. That still does not mean we need a new MDG for inequality. In setting goals we need to think carefully about what are “ends” and what are mainly “means.” When we muddle the two we risk ending up with a dog’s breakfast of goals, lacking any clear conceptual framework and full of double counting. Something can be instrumentally very important to things we care about but not be deemed an important explicit “goal” in its own right.

Take poverty. Along with many others, I believe that eliminating extreme poverty, as judged by the standards of what “poverty” means in the poorest countries, should be our main development goal going forward. So it is very clear which way is up. Of course, extreme poverty is not the only thing we care about, but who would say today that it is not something we should strive to eliminate? And there are good reasons to be optimistic about both a global agreement on eliminating extreme poverty, and for making huge progress toward that goal over the next 15 years (as I argue here).

Any standard measure of poverty will depend on both mean income and (roughly speaking) the extent of “inequality” in the distribution of income. That is one way inequality matters to poverty, and standard measures of poverty are typically quite responsive to changes in overall inequality within the range of country experience.

However, there is another important way inequality matters. Even if inequality does not worsen over time, a high initial level of inequality can make it harder to reduce poverty. High inequalities in certain dimensions (especially in human development, but also in access to production inputs, including finance) are one of the greatest challenges we face to sustained poverty reduction.

But that is not a good reason for adding an MDG on inequality unless we are also convinced that existing goals put too low a weight on inequality reduction and that adding an inequality goal is the best way to correct the problem. Probably the main reason we think that standard poverty measures (such as the “$1 a day” goal for poverty in MDG1) under-weight inequality is that they ignore social effects on welfare, by which I mean concerns about relative deprivation and social inclusion in specific contexts. These are real welfare effects (though quantifying them rigorously is not easy). However, I think we can be pretty confident that such social effects exist.

By implication, a common international poverty line, intended to have constant purchasing power, will not yield the same level of welfare in different societies. Higher expenditures will be needed in richer societies to avoid shame in public and to participate fully in social and economic activities. A measure of relative poverty that reflects these social effects on welfare will almost certainly put a higher weight on inequality than a standard absolute poverty measure.

That does not mean that we need an inequality MDG. Better to measure relative poverty directly, and add a goal for that. I don’t believe it makes any sense (except possibly in very rich societies) to do so by setting a poverty line at some constant proportion of the mean (or median), as is popular in Western Europe. Applying that method globally will definitely under-estimate the costs of social inclusion in poor countries. However, one can use what I call “weakly relative” poverty measures, which avoid this problem, and they are not difficult to implement on a global basis. (See my paper with Shaohua Chen here.)

Another thing that is clearly missing from standard poverty measures is inequality within the household. But (again) it is far from clear that a measure of overall inequality is the best corrective; indeed, intra-household inequality is typically missing there as well. Better to focus on more direct supplementary measures such as infant mortality and nutritional status.

The key point is that the poverty reduction goal already embodies inequality, and even if we agree that it under-value things like social inclusion or intra-household inequality, it is far from obvious that adding an overall inequality measure is the best corrective. We need to think clearly about what is missing and how best to measure it.

So far I have based my case on poverty reduction – the first of the MDGs. Similar arguments can be made for other MDGs, almost all of which will depend on both the overall mean income and the inequality in its distribution as well as other factors like the coverage and quality of public services (as explained more fully here for human development indicators). When we think about adding goals we do not want to add things that are already included.

If you are still not convinced, ask yourself whether we also need a new “economic growth” goal, or a “growth rate of the poorest x%”. We can probably agree that growth is instrumentally important to many of our existing development goals without wanting a new goal, with little or no genuine value-added. Fostering economic growth is a potentially important means, but it is not an end in itself; the same applies to inequality.