We measure relative poverty in rich countries; absolute poverty in poor ones – what if we combine them?

Ravallionodd divide between poor countries, where absolute measures are more often used (eg $1.25 a day, the current international poverty line) and rich countries, which tend to use measures of relative poverty. For example, in Western Europe (including the UK) the poverty line is set at a constant proportion (typically 60%) of median income. Ravallion calls this a “strongly relative line”. Ravallion argues that both approaches make sense – an absolute line in establishing who is able to feed and clothe themselves, a relative line in determining whether people feel socially excluded or not. In very poor countries, absolute survival plays a bigger role, but as countries’ average wealth rises, so social inclusion, and thus relative poverty, becomes more salient. However, Ravallion argues that social exclusion matters even in the poorest imaginable country. So a sensible poverty line cannot simply rise and fall proportionately to average income, as in the strongly relative poverty lines used in Europe. Look at Ravallion’s graph. The strongly relative line keeps falling toward zero. The strongly relative lines will be too low (below survival levels) in poor countries. Ravallion says we need “weakly relative lines” which incorporate an absolute minimum, as in the bold line in his graph. Ravallion graph In a paper with Shaohua Chen published in the Review of Economics and Statistics, he puts some numbers on this (and lots of equation which were way above my head, so thanks to Martin for helping me make sense of it to write this post……). Up to an average national consumption of $2 per day, he sticks with absolute measures – anyone under $1.25 is poor. But then he switches to a relative measure – above $2 a day, the poverty line slopes upwards, reflecting the importance of social inclusion (see graph). The graph slopes up with a gradient of a third – less than the 60% of the UK strongly relative poverty line. Chen and Ravallion use data on national poverty lines across countries to set this schedule—the same data they had used to select the $1.25 a day line. What do they find? ‘The trend decline in the incidence of relative poverty has not been sufficient to reduce the number of poor by this measure, which rose from 2.3 billion to 2.6 billion over 1981 to 2005 (see table 1). The turning point is around 1987.’ This contrasts with the more standard measures of absolute poverty, which have the numbers of poor people falling both as a % and in absolute numbers to 1.4bn people by 2005. Why the disparity between the two (a rising absolute number of relatively poor, and a falling number of absolute poor)? As the global economy grows, poverty lines start to rise to reflect the higher costs of social inclusion, which puts a brake on the pace of relative poverty reduction, despite falling absolute poverty. More background in Martin’s World Bank blogpost.]]>

Subscribe to our Newsletter

You can unsubscribe at any time by clicking the link in the footer of our emails. For information about our privacy practices, please see our .

We use MailChimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to MailChimp for processing. Learn more about MailChimp's privacy practices here.


7 Responses to “We measure relative poverty in rich countries; absolute poverty in poor ones – what if we combine them?”
  1. That’s interesting – I never knew poor countries used the $1.25/$2 absolute poverty lines, I always thought that was a global measure used by UN affiliates to eliminate the complexities of measuring poverty between countries while accounting for exchange rates, varying median incomes, etc.
    Actually, come to think of it, I’m not clear on how much poverty data in a poor country is self-collected and reported, and how much is done through various NGO estimates and reports. Maybe I’m just not looking searching the right channels, but I don’t usually see poverty-related data from a national body; it’s almost always from an international reporter or observer.
    Maybe an all round better way to tackle relative poverty and social exclusion is to encourage national governments to collect their own data on poverty rates. After all, absolute poverty is universal and translatable, but social exclusion is very much dependent on national and local variables; it seems appropriate then to encourage national governments to try to collect this data and take action to remedy such a situation.

  2. I am talking about consumption expenditure rather than consumption.
    I am not sure that it would be true to say that there are any consumption measures of poverty – in the ordinary meaning of the word ‘consumption’ – which are commonly discussed in an international context.
    The World Bank global estimates are from surveys on consumption expenditure (or where that is unavailable, income) and imputed value of self-produced consumption.
    I question the appropriateness of the abbreviation ‘consumption’ both among economists and for a wider readership, including politicians.
    In reality no-one can know from the spending numbers
    a) what people consumed, or
    b) whether it was enough to meet needs.
    It may be that macroeconomists are, without describing their method, using other information to infer that prices for the poor, and needs, did not differ significantly across the time periods, countries, policies, economic conditions or whatever it is that they are comparing.
    But such inferences need to be explicit. Otherwise it is not clear why the economist is talking about benefits, or poverty, rather than about what the surveys are actually on. I don’t know what has happened to poor people’s prices, for instance, in terms of correlations with GDP per capita, or trade volume, or social spending.
    The FAO, unsurprisingly, use the word “consumption” in relation to what people consume, rather than in relation to spending.
    Once we use the more accurate word ‘spending’, it may become clearer what is being looked at.
    Economists have not collated either prices or needs for the poor over the last few decades, and so cannot know what people could afford, either with what they spent or with what they earned.
    To answer a slightly different question – do I mean using what are called consumption expenditure measures of poverty is risky? – the answer is yes, for more reasons than those.
    I have written about similar matters in a submission to the House of Lords Economic Affairs Committee inquiry on development aid:

  3. Martin Ravallion

    Reply to Matt Berkeley from Martin Ravallion
    As is made clear in the source papers for Duncan’s blog post, all the data on consumption expenditures we have used are deflated to allow for inflation (in making comparisons over time) and for differences in prices between countries (in making international comparisons). Of course the data are rarely ideal. But it is not correct to imply that prices are ignored. These are all real values.

  4. nilakshi

    I am not sure why we need to use the poverty line to measure social exclusion. isnt the gini and other measures of inequality sufficient for this purpose? I feel that the measure of poverty should really be just that – the number of people all over the world who are deprived in absolute terms. Otherwise i wonder if we risk diluting the picture and taking attention away from the critical issue.
    Duncan: but then how do you take into account the fact that notions of poverty in all countries include a bit of both (absolute and relative), with the balance shifting as an economy grows?

  5. nilakshi

    you mean because the national poverty line is based on a basket of goods and services – which is derived from the economy, culture and lifestyle of that country, and is different from country to country? well, i think that amount of relativity is probably inevitable and acceptable.. wasnt it Adam Smith who said poverty is ‘the inability to purchase necessities required by nature or custom’? adjusting for ‘custom’ will include a relative perspective i agree. But measuring poverty as relative to the rest of the in-country population – while useful for that country – at the global level may be bringing in a set of complexities that the situation does not require.