What’s going on with global inequality? Let’s ask Andy Sumner……

When I was asking around about last week’s Oxfam paper on the G20 and inequality, ace number cruncher Andy Sumner emailed to say that in his opinion, talking about inequality in terms of the Gini index (a single number for overall inequality) is terribly old hat, and these days, everyone is trying out different indicators to get at the fine detail of different kinds of inequality. So I went back to his July summary of the latest research on within-country inequality on Global Dashboard, and I have to say, it is brilliant. Some highlights:

‘What’s happening to within-country inequality isn’t immediately clear.

The new Solt database of the main measure of inequality (known as the Gini after an Italian Sociologist who developed it) was analysed by Ortiz and Cummins at UNICEF who concluded the evidence showed:

Rising inequality in Asia, 1990-2008 but falling inequality in Sub-Saharan Africa over the same time period.

And inequality in Latin America rose slightly 1990-2008 but fell between 2000-2008 and inequality was static in the Middle East and North Africa.

Ortiz and Cummins list a long set of countries where inequality significantly fell between 2000-2008. For example, inequality fell by more than 3 points in Thailand, Malaysia; Brazil, Peru, Argentina, Chile; Lesotho, Malawi, Ethiopia, Burundi, Mali, Sierra Leone, Burkina Faso, Uganda, Nigeria, Gabon.

However, a new paper by Chilean Economist Gabriel Jose Gabriel Palma does a detailed study of within country inequality between 1985 vs 2005 suggesting the Gini hides as much as it reveals.

Instead we need to look at each 10% of the population and what they get.

He finds that there is now a surprising similar picture in most countries, noting:

1. The great majority of regions and countries have a relatively similar distribution of income inequality because countries with low inequality at the outset (1985) have got more unequal and countries with high-inequality have got slightly more equal.

2. The middle classes generally get half of the economic pie wherever you look and the middle classes are incredibly successful about protecting their half.

3. Politics is increasingly a fight for the remaining half between the richest 10% and poorest 40% meaning the other half of the distribution is increasingly ‘up for grabs’ between the very rich and the very poor and who can win over the middle classes.

This might begin to explain some of the recent declines in inequality in Latin America as suggested in a paper by Birdsall et al., who argue that ‘social democratic’ regimes (eg Brazil, Chile and Uruguay) are more likely to reduce inequality than ‘left populist’ (eg Argentina, Bolivia, Ecuador, Nicaragua and Venezuela) and both are more likely to reduce inequality that non-left regimes (eg Colombia, Costa Rica Mexico, Peru) and that this is largely due to more social spending and more progressive spending especially so in the social democratic regimes (eg spending on cash transfers targeted to the poorest and greater increases in spending on health and education and increases in spending on basic services – in particular in education, greater increases in spending on primary and secondary schooling rather than on public universities.

These social democrats have strong support in the middle classes and this throws up the question posed by Birdsall et al:

Might the growing middle classes in countries like Chile and Brazil help lock in leftist social democratic political regimes (whether because or despite its concentration in the top quintile of households)?  There is no evidence that a large middle class is necessary let alone sufficient to these regimes.  But a growing global middle class does seem likely to reinforce effective government that manages moderate redistribution while retaining investor confidence in the likelihood of continuing growth and price stability. Put another way: When is the middle class large enough to become politically salient in supporting or at least tolerating the kind of social and other distributive policies that are good for them but turn out to be good for the poor—for example universal public education?

Food for thought – the middle classes as the new revolutionaries?’

Great stuff. It set me thinking about one of the findings of our paper – that whereas inequality is falling in many low and lower middle income countries (see chart), it is rising in all but four of the G20 countries. Might that be because the middle class is more likely to ally with the elite in the G20 countries than elsewhere and why could that be? Perhaps because their economies are integrated, or more dominated by financial institutions? Just idle speculation of course  - feel free to add y0ur own.

As for organizations like Oxfam, it seems to provide empirical support for our focus on ‘convening and brokering’ – the latest jargon for getting people from different social and economic backgrounds into a room, and helping them build alliances. Any other thoughts?

What should aid focus on, poor people or poor countries?

Finally got round to reading the paper that’s been making waves in wonk-land, ‘Global poverty and the new bottom billion: Three-quarters of the World’s poor live in middle-income countries’, by Andy Sumner from the UK’s Institute of Development Studies. In a classic bit of number crunching, Andy takes a fresh look at where poor people now live, and comes to a striking conclusion:

‘In 1990, we estimate that 93 per cent of the world’s poor people lived in low income countries (LICs). In contrast, in 2007-8 we

wonk of the month

estimate that three-quarters of the world’s approximately 1.3bn poor people now live in middle-income countries (MICs) and only about a quarter of the world’s poor – about 370mn people – live in the remaining 39 low-income countries, which are largely in sub-Saharan Africa.’

So much for Paul Collier’s ‘bottom billion’, living in Africa and/or fragile states. In fact most of them live in stable, middle-income countries, and that finding applies across economic, nutritional and multi-dimensional poverty measures (although intriguingly, not to education, where underperformance remains more weighted towards the LICs).

So what? On one level, most of the change results from former LICs graduating to MIC status (the paper also provides a really useful overview of the ridiculously complicated world of overlapping but slightly different country classifications – low income, least developed etc). Over the last decade the number of LICs has fallen from around 60 to just 39. In particular, the graduation of India, Nigeria, Indonesia and Pakistan account for most of the change in poverty distribution.

So the finding is to some extent an artifice of country classification  – poor people live in roughly the same countries as in 1990, but those countries have got a little bit richer, and so joined the bottom tier of MICs. But there are at least two ways that the paper challenges our thinking on development:

1. Direction of travel: the upward march of countries like India doesn’t stop when they graduate to MIC status. GDP per capita is likely to increase steadily over coming years. If poverty persists, it will become much more the result of internal inequality, and the solution one of internal redistribution. In those circumstances, the paper asks ‘whose ‘responsibility’ are the poor in MICs – donors or governments or both?’ The challenge for aid becomes dual – how to catalyse mass poverty reduction in the remaining LICs, but how to support and stimulate that struggle for redistribution in the MICs. In Andy’s words:

‘This might mean that long-term poverty reduction requires more focus on structural economic transformation (assessed perhaps by the percentage of employment in agriculture) or a social transformation to a low level of inequality (assessed by gini coefficient and implied emergence of a middle/consuming class), or political transformation (assessed by tax revenue as percentage of GDP and the implied accountability that follows).’

2. Stability: the paper finds that 61% of the world’s poor live in stable MICs. Only 23% of the world’s poor lie in fragile states, roughly evenly split between LICs and MICs. So if you are using poverty as your compass, the development industry’s increasing focus on fragile states (and attention to issues of security and stability) looks a bit overdone. Two thirds of poor people live in stable settings where issues such as politics, inequality, livelihoods, essential services can be debated and resolved without the threat of civil war.

What does all this mean for DFID? Well, in recent years it has focused much more on LICs and conflict-affected countries, and that process is likely to continue with the findings of the current bilateral aid review, which is expected to lead to DFID’s withdrawal from dozens of countries, many of them MICs. Andy’s paper suggests that might not be the right answer, at least in terms of poverty reduction. Andy’s boss at IDS, Lawrence Haddad certainly thinks so, judging by his recent defence of giving aid to India.

On the other hand, working in MICs is not so much about money as influence, and aid donors (not just DFID) have always found that a much harder nut to crack. Maybe DFID should be prepared to abandon 75% of the world’s poor to their domestic political processes, and focus instead on the 25% where it can make a difference?

For more on the paper see the Economist, Global Dashboard and Lawrence Haddad