Warning: quite long. If preferred, download as a PDF

Any serious conversation about the nature and causes of poverty in Africa, and its reduction, must begin with an attempt to measure the scale of the problem. The Make Poverty History campaign, and the huge wave of debate, campaigning and publicity on the topic that seemed to take over the world in mid-2005, made repeated recourse to the more scary-sounding statistics on poverty, most notably that “more than one billion people around the world live on less than one dollar a day.” But once you start to delve a little deeper into the argument, it becomes obvious counting the poor is a lot harder than it seems.

This matters, because measures of the number of poor - and how it’s changed over time - are crucial to the debate over the best course of action. Take the long-running dispute between the Financial Times’ Martin Wolf and political economist Robert Wade. Wolf started the barney by asserting that the number of people living in extreme poverty had declined by 200 million since 1980, the period associated with privatisation and globalisation in developing countries, and therefore these policies must be good for poverty reduction. Wolf’s evidence for the figures was a World Bank study. Sounds pretty tight. But Wade retorts that the World Bank’s estimate is flawed. In fact, look behind almost any disagreement on the way forward for poverty reduction, and you’ll find a disagreement over figures.

If we can’t even agree on how many poor people there are, how are we ever to plan for poverty reduction? What does “extreme poverty” mean, anyway? Who are “the poor?”

Defining poverty

The measure most commonly used is the aforementioned - people living on less than a dollar a day. In fact, this common phrase is a complete misstatement. First, it’s not one dollar at all, but $1.08. That may not sound like much, but that 8 cents can vary estimates of poverty enormously. Second, it’s not dollars in today’s money. It’s actually Purchasing Power Parity 1993 Dollars. In other words, it’s the money required in any given country, in 1993, to buy an amount of goods that you could buy for a dollar in the US. It’s then adjusted to make for today’s prices. So “living on less than a dollar a day” really means, “living on less money per day than it would take to buy what you can buy for a dollar in the US.”

Realising this makes a huge difference to your perception of the $1/day figure. In no sensible world could $1 a day be seen as a decent income. But if you imagine an African country with a heavily devalued currency, $1 might sound like enough to buy, say, a few loaves of bread, as opposed to just one.

Although there are objections to the PPP dollar measure, it’s generally accepted as a reasonable attempt to compare countries. But the actual details of the PPP measurement are controversial. The standard “basket of services” that was used in 1993 to assess PPP dollars - i.e. what the value of a dollar in goods terms was - included not just essentials such as food, clothes and shelter, but advanced services too - like massages. This wouldn’t matter, except that the relative cost of essentials and luxuries is very different in different countries. In the US, a 30-minute massage costs a lot more than a bag of grain. In many African countries, the opposite is true. So a measure including such services in the “basket” has the effect of over-estimating the ability of the poor in Africa to buy essentials (see Wade, p3-11).

So why don’t we just use a new basket? The problem is, when you change the measure, you move the goalposts and make comparison with the old system impossible. This has already happened once, with the PPP assessment based on 1993 figures, that generated the $1.08 poverty line. Before that, a different measure of PPP was used, that had generated the original poverty line of $1. That was based on measurements taken in 1985. One of Wade’s key criticisms of the World Bank study Wolf quotes is that it tries to compare figures taken with the 1985 PPP measure with figures taken with the 1993 measure. The problem could be solved if the World Bank went back and calculated the figures before 1993 with the new measure, but for some reason, it hasn’t. To prevent any further problems like this, most development economists say we’ll have to stick with the measure we use now, but be more aware of its limitations.

The PPP dispute is just one of the massive number of disagreements over all our methods of measuring poverty. Here are some more:

Number of poor vs. proportion of population who are poor

Even if the number of poor people in the world has gone up over the past two decades, so has the world’s population - by a lot. So it’s still quite possible the proportion of the world’s people who are poor has still gone down. Does this matter? In fact, this dispute is central to the wide range of opinions on whether things are getting better or worse. Most studies suggest that the proportion of people in the world who are poor - “poverty incidence” - has gone down. But the number of poor may have gone up, and it’s that that campaigners tend to focus on (Wade, p3).

Poverty vs inequality

Here again there’s disagreement both about what’s happening, and about whether it matters. Many studies suggest that overall, inequality between countries has gone down over the last 25 years. These studies, horribly complex things, usually use something called the “Gini coefficient,” a standard measure of inequality in any set of figures.

However, others counter with two caveats. First, while overall inequality may have decreased, polarisation - the gap between the two extremes - has almost certainly increased (Wade, p14). This is the meaning behind statements like:

While the share of the world’s resources in the hands of the top ten percent people has increased, the share held by the bottom ten percent has decreased.

Second, critics argue, while inequality between countries may have decreased, inequality within countries appears to have increased. So for example the gap between the US and China has got much, much smaller, but the gap between the poorest and richest Chinese has got much, much bigger. Any attempt to measure the actual inequality of all the individuals in the world requires a mixture of the two measures, and such measures are decidedly ambiguous about how the last twenty years have gone (Dowrick & Akmal).

What’s more, there’s also the question: does it matter? What should concern us is how many people are poor, unable to reasonably feed, clothe and educate their families, protect themselves against disease, and so on. Increasing inequality could just mean the rich have got richer. As long as the poor have got richer as well, where’s the problem?

Of course, this is an argument that goes well beyond development theory. Domestic politics in every western country features disputes over whether inequality matters - as Tony Blair rather craply put it, “why is it a problem if David Beckham makes more money?”

But critics argue that it’s particularly relevant to the development debate. First, almost all domestic political systems assume that inequality does matter, but the same ideas haven’t been applied to development. The poverty line in any western country is defined as a share of average income, but the standard measures of poverty in developing countries are all based on actual, not relative, wealth.

Secondly, the dangers of inequality may be larger internationally. Domestically, problems widely believed to be linked to inequality include crime and civil disorder. Internationally, the argument goes, the risks include terrorism, war, and massively increased migration to the developed world, legal and otherwise.

Household Surveys vs National Accounts

Both poverty and inequality measures require knowledge of consumption rates, that is, some measure of how much money people actually have to spend. This is carried out through household surveys, involving having households record their income and outgoings for a period of time. These surveys are vital to the PPP poverty measurements.

There are various problems with household surveys, but the most serious is the duration of recall, i.e. how long you ask people to record. Asking people to record seven days tends to lead to higher estimates of income than asking them to record 30 days. One study revealed that if you used a seven-day recall in India, the number of people classed as poor would halve! (Wade, p7)

However household surveys tend to give higher measures of poverty than governmental accounts. This matters, because many measures of world inequality use accounts instead of household surveys for their figures (household surveys are only used to study income distribution in a country, not overall income itself). The few studies that use income estimates derived solely from household surveys report higher poverty and inequality, and worsening trends in both (Wade, p15, quoting Branko Milanovic).

Pesky success stories: China and India

Above I said, “Many studies suggest that overall, inequality between countries has gone down over the last 25 years.” There’s one other, dirty great caveat on that which needs to be outlined. In the last 25 years China and India have grown tremendously, and together, they account for over 30% of the world’s population. This has a massive effect on the figures: so much so that, while measurements of inequality which weight each country by population show reductions in inequality, measurements that give each country equal weight don’t (Wade p14).

So what? They’re real countries, aren’t they? The problems are twofold. First, in simple statistical terms, such large figures overshadow the story in other areas. So we can’t learn anything useful about Africa from global figures so skewed by the world’s two largest countries. Second, China and India are very particular types of LDC, with particular needs and challenges. There’s a risk of policy conclusions being drawn from their success and applied elsewhere where it isn’t appropriate.

So what if you take China and India out of the figures? The results are then pretty unclear, with no clear trend either upward or downward in either inequality or poverty.

Poverty: about more than money?

So far, we’ve talked about poverty purely in terms of income. Well, obviously, you might think. Poverty means no money, doesn’t it? Actually, the whole idea of measuring the poor by income has come under heavy fire under recent years from theorists such as Nobel prize-winner Amartya Sen, who argue that a range of other factors either affect whether people are poor, or can be used as better measures of poverty than income (OUP, p35-42).

The other elements of this “multi-dimensional” model of poverty are:

  1. Participation. Also known as “voice,” this means access to decision-making, influence over politics, and influence in the community.
  2. Empowerment. Related to but not the same as participation, this is about people’s control over their own lives. So, for example, a women’s group that gives women a chance to find out about the reasons for recent declines in their income, and tells them about their options, is empowering because it increases their control over their situation.
  3. Vulnerability to shocks. Even someone whose income is enough to keep them at a reasonable level of comfort could face disaster in the event of a price crash, failed crop, or illness. Measures which help people plan for such events - from government-run social security, to private insurance, to tinned food - have the effect of improving people’s security, even though they don’t affect income measurements.

These factors are notoriously hard to quantify, but are hugely important to planning poverty policies. Another set of factors are seen as useful indicators, alongside income, of poverty:

  1. Life expectancy
  2. Literacy
  3. Education enrolment

Amongst a few others (including an income-based measurement of poverty), these are included in the UN’s Human Development Index. In addition, there’s the Human Poverty Index, which is weighted more towards poverty.

Multi-dimensional measurements of poverty are hard to compare and work with, and economists mostly hate them. But they’re an important part of the case for the general state of poverty having got worse in the last 20 years. AIDS, for example, has slashed life expectancies across Africa, which only indirectly affects income measures but would be seen by many as part and parcel of the African poverty problem.

Finally, it’s pointed out that income-based measures ignore the sources of food and other essentials that come from outside the market. Food is grown, foraged, and bartered, and these activities are hard to measure.

The need for common measures

The range of measures available is undoubtedly a positive thing - it makes us more aware of the problems of those measures that are used most frequently. Indeed, given that the only comprehensive poverty studies regularly done are by the World Bank, and there’s some evidence of the World Bank selecting its measures in line with political factors, more, rather than fewer, studies are needed. But there’s also a risk of multiple measures - that all the players in a debate are using those that support their opinion. How many times have you seen one newspaper columnist declaring that things are just getting worse for the world’s poor, and another declaring things are just getting better? For a meaningful debate, we need some agreement on measures.

You can’t blame the likes of Make Poverty History for using catchy, if potentially misleading, figures such as “more than a billion on less than $1/day” (incidentally, the actual World Bank figure is 1.2bn; but the problems listed above mean the figure is almost certainly higher). They’ve got a point to make. But the media should know better. If a fraction of the analytical rigour that goes into the reporting of business-friendly economic statistics, such as interest and inflation rates, went into the presentation of development data, the conversation would be much, much richer. For my part, I’m determined - on this site and in the documentary - to be clear about the methodology and sources of figures cited.

This much we know

Realising the extent to which even the basic measures of poverty are controversial, it would be easy to throw hands up in the air and give up. And of course, attempting to study the relationship between poverty and other factors, such as globalisation, throws up even more problems. Fortunately, there are a few things which we can draw out with certainty.

1. China and India are growing really, really fast
2. Other LDCs, particularly Africa and Latin America, are growing significantly slower than MDCs
3. Too many people are poor

Whether you see poverty and inequality as having worsened or improved over the last 25 years, and however you draw the measure of the numbers living in extreme poverty now, no-one is suggesting things are good enough. Given the generally extremely positive economic progress of MDCs in the last 15 or so years, there’s simply no moral justification for the continued existence of extreme poverty - or rather, there’s no excuse for poverty reduction not to be top of the political agenda.

Useful Sources

“Globalization, Poverty And Income Distribution: Does The Liberal Argument Hold?” Robert Wade

Introduction to Poverty and Inequality Open University 2002

“Explaining contradictory trends in global income inequality: a tale of two biasses” Steve Dowrick and Muhammad Akmal

“The Two Faces of Globalization: Against Globalization as We Know It” Branko Milanovic

“Why are the Critics so Convinced that Globalization is Bad for the Poor?” Emma Aisbett