Some things governments can do to support development even without spending more money

February 3, 2010

Why militarizing aid in Afghanistan is a bad idea

February 3, 2010

Are women really 70% of the world's poor? How do we know?

February 3, 2010
empty image
empty image

Doing research for advocacy (which is a large part of my job) is a balancing act. The pressure to come up with clear findings and ‘killer facts’ that speak to policy-makers can easily tip over into something much more questionable.

I once challenged a colleague at another NGO on a ‘fact’ she was using on Bolivia.

London's Senate House, inspiration for Orwell's 'Ministry of Truth'

London's Senate House, inspiration for Orwell's 'Ministry of Truth'

‘Well, it’s politically true’, she replied with a grin. Should we use facts we know are wrong, because we like their message? Surely the answer has to be no.

I remembered this when mulling over a favourite NGO and UN factoid: ‘70% of the world’s poor are women’. Its source is decidedly murky: the 1995 Human Development Report is often cited, but it gives no reference for how it arrived at the figure, or way of checking it, and contributes to the confusion by saying in the main text ‘more than 70% are female’ (i.e. including girls) but then simply refers to ‘women’ in the executive summary. The 1994 IFAD report “The State of World Rural Poverty” is also named, with doubts expressed at the time about its numbers.

Other friends tell me it was around long before that, but no-one can point me to the original source. But in a way that’s beside the point – these ‘magic numbers’ take on a life of their own through sheer circulation and repetition. Note that this particular figure hasn’t changed in the intervening 15 years – that alone should ring alarm bells.

Let’s assume the HDR means women and girls. With the latest poverty stats, that adds up to 980m women and girls living below the $1.25 a day poverty line, and 420m men and boys. That either means that women-headed households are particularly poor, or that within households more women and girls are poor than men and boys. But as far as I’m aware, official income poverty figures don’t drill down below the household to say ‘in this family, this child/adult is poor, and this child/adult isn’t’ – that would be even harder to do for income (which is what the 70% stat refers to) than for consumption.

So either the number is based on long-lost research on the division of income within a (presumably) fairly small number of households scaled up to a global figure (and if so, please put me out of my misery and send me the reference), or the only explanation of the figures is that families headed by women explain the difference.

But to make the 70% figure stand up, you would need 560 million women-headed households (i.e. the difference between 980m poor women and 420m poor men), all of them below the poverty line. With an average of 2-3 kids per household, that is more than the total number of poor households – clearly the figure is nonsense.

Before you brand me as an incorrigible reactionary (OK, it’s probably too late for that…..), I’m not denying the existence or huge importance of gender inequality. On the contrary, gender is undoubtedly one of the world’s great faultlines of distribution and injustice. Ownership of assets is hugely distorted between men and women (probably much more than 70/30), as is time spent on reproductive activities (child rearing, cooking, cleaning etc). The same is probably true of consumption, and if we could find a credible way of measuring income inside the black box of the household, we might well find a big discrepancy there too. All the more reason to get our facts right, no?

street childrenSome ‘political truths’ diverge even further from reality. UNICEF campaigners on street children used to say that Brazil had 7-8 million of them – another stat that rapidly acquired a life of its own. Yet when researchers actually tried to count them, setting out in the middle of the night to the places where street kids sleep rough, they found fewer than 1,000 sleeping in Brazil’s two major cities, Rio and Sao Paulo, combined. Scaling up to the whole country, and scaling up by a factor of 3 in case they missed some, one researcher came up with a total of 38,000. That’s a lot of kids sleeping rough, but only about one two hundredth of UNICEF’s imaginary battalions.

Every expert (feminist economists, poverty researchers etc) I’ve consulted on this agrees the number is dodgy, and yet people just keep on using it, presumably because its message is one they want to promote. But isn’t that short sighted? Sure, any attempt to produce a simple, powerful narrative out of inevitably messy data entails some level of violence to the complexity of real life. And simple narratives are precisely what stick in people’s heads, improve policy, change attitudes and bring about change. Insist too much on the intellectual purity of the ‘everything is complicated, context specific and difficult’ camp and you will be right but completely ineffectual. But depart too far from reality, go too far down the road of ‘political truth’ and you undermine your own legitimacy – why should people believe anything else you say on the issue? I reckon ‘70% of the world’s poor are women’ crosses the line. But please someone, prove me wrong!

For more excavation on the 70% stat, see this paper by LSE professor Sylvia Chant.


    1. Memes only go so far – what you need is to look at “The Woozle Effect” – False Evidence By Citation.

      “Results of a weak study may be repeated so many times in different sources (e.g., professional journals) that they (undeservedly) achieve the status of a law.”

      — Eileen Gambrill, Propaganda in the Helping Professions (2012)

  1. Dear Duncan,

    Thank you for this thought-provoking blog and your questions. I too have been citing the 70 % (or is it 2 thirds-66 % figure, quoting from those UN reports you mention in your blog.

    Clearly, more research is needed, as well as generating more and better sex-aggregated data. I am, therefore, very happy that you, as head of research for Oxfam GB, are interested in this figure. I wonder whether you or your staff could not produce a better estimate of the number of poor women in about as long as it takes to write a blog. I agree with you that female poverty is not just a question of single-woman households (though these are usually among the poorest also in the North).
    We know that in many developing countries women are getting a raw deal also within families. We know this from the consumption surveys that you refer to, which show, for example, that when a family goes hungry women and children go hungry long before the men do. Such surveys exist for several African countries and for India, and surely for other countries as well. Would it really be that bad to estimate, from caloric intake, what share of a household’s income goes to women, and use that to estimate poverty rates for women and men separately (see Amartya Sen)? Sure, that’s just an estimate, but science is all about estimation, critique, and re-estimation.

    At the same time, a discussion concerning the 70% figure and its origins, cannot be used to make a general point about addressing on gender justice issues in the context of fighting poverty. I am not suggesting that this is the point you are trying to make with this blog. However, it could be understood in that vein. As you pointed out, there are some hard facts that support the feminization of poverty:
    -you point out unequal wealth and unequal work divisions (without numbers btw.)
    -there is unequal access to education, especially secondary and tertiary education which really makes all the difference in a women’s life, resulting in huge disparities in illiterateness rates
    – the lack of and access to medical care resulting in 500.000 unnecessary deaths in childbirth per year
    -there is the difference in child mortality between girls and boys
    – Finally, violence against women, including sexual violence, an unacceptable factor obstructing any woman’s development
    (For more facts see, UNIFEM, 2008. Gender Equality Now. The Unfinished Agenda. Balance sheet on progress and backlogs on Gender Equality)

    Underlying these facts is one fundamental issue: the abbhorent discrimination of women, based on the belief they are less worthy than men (or girls being less worthy than boys). That very rightlessness is hard to qualify in numbers. Or perhaps this could be an interesting subject as well for you and your colleagues. What I know, is these facts make a strong enough case for us to work on women’s rights and gender equality issues if we want to end poverty.

    Yours sincerely, Daniela

  2. Thank you for another good post Duncan!

    NGOs need to be accurate, precise and clear with their figures because the work is so important. Widely quoted but implausible figures may start off sounding impressive but will eventually be discredited which doesn’t help in the long run to solve any of the problems you are tackling. (I personally have never found the 70% of people in poverty are women figure to be believable and had wondered where it came from). This is something that has recently come back to haunt the UK government with it’s “45 minute warning” in the lead up to the war in Iraq, and the IPCC with it’s unrealistic estimate on the end of glaciers in the Himalayas.

    I hope your blog post will be a step in encouraging enthusiastic campaigners to be scrupulous with their accuracy. Trust can so easily be lost.

  3. Yes, Duncan, 70% of the poor being women is only ‘political truth’, we have known this for a very long time. But I cannot help wonder whether, rather than ask why people keep using it, it may have not been much more useful to explore the reasons why after so many years we still do not have the necessary methodologies, data, and resources to have adequate statistics on this. Perhaps this would have already been sorted out ‘if women counted’ (as the title of the 1989 book by the economist Marilyn Waring went). –

  4. Ken, your comment prompted me to have a look at the latest data from the UK and the US (found via the new and v useful and sites).

    In the UK, there’s reams of sex-disaggregated data on poverty for 2007/8 in the ‘Households Below Average Income’ publication, but the headline seems to be that, in terms of disposable income, 20% of adult females (4.84m) are in the bottom quintile compared with 17% of adult males (3.88m), and in Q2 it’s 20% of females (4.84m) compared with 18% (4.10m) of men (children are not sex-disaggregated). But it’s also worth noting the paper’s caveats around this:

    “In any analysis of gender, it must be remembered that [the paper] attempts to measure the living standards of an individual as determined by household income. This assumes that both partners in a couple benefit equally from the household’s income, and will therefore appear at the same position in the income distribution. Any differences in figures can only be driven by gender differences for single adults, which will themselves be diluted by the figures for couples. The lower level gender disaggregation in the family type classification is therefore likely to be more informative.

    “Research has suggested that, particularly in low-income households, the above assumption with regard to income sharing is not always valid as men sometimes benefit at the expense of women from shared household income. This means that it is possible that… results broken down by gender could understate differences between the two groups.”

    In the US, data from the Census bureau indicate that for 2006-8, 14.5% of females (21.7m) were below the national poverty line, compared with 11.7% (16.9m) males. The methodology in this case dictates:

    “If a family’s total income is less than the dollar value of the appropriate threshold, then that family and every individual in it are considered to be in poverty.”

    Given that statisticians in these two countries are challenged to produce nationally representative data that reflect intra-household asset allocations between men and women, the absence of such reliable data for developing countries is hardly surprising.

  5. Thanks very informative. From these numbers it looks about of the poor people in the UK/US , 55% are women , 45% men. Given that this does not include intra-household allocation and that the situation of women in the developing world is likely to be worse than this , I don’t think the 70% figures looks unrealistic. What’s the best guess from somebody who knows about this ?

    1. Sorry, but you have left out children entirely. Rates of poverty among children are far higher than among women and men in both the UK and the US.

  6. I like your statement Mr. Smith, yes, may be that 70% is not such a bad estimate at all! Of course it cannot be ignored that gender disparities and – inequalities do exist all over the world and that these, besides the fact that they show tremendous injustice, are an important constraint on development and growth. But is it really worth to devote so many words to this figure? Feminist scholars became aware of its shortcomings a long time ago, so at least the 70% figure triggered a lot of research and thinking. Mr. Green mentions Chant, that’s interesting because it’s precisely her who points out the limitations of statistics and numbers that focus on only one (quantitative) aspect of women’s poverty and unequal status, like income for example, or longevity. She therefore rightly calls for a more complex approach of measuring gender inequality and gendered poverty, bringing in the important dimensions of meaning and quality. To put it in her words: “Accepting that not every aspect of gendered privation is amenable to quantification and that indices will always require gender analysis to tell us about processes, it is vital to start cultivating a broader and more inclusive base for longitudinal comparisons of gendered privation, and to determine whether, how and in which particular forms a “feminization of poverty” is evolving.” This points to, yes, ‘more complicated, context specific’ and not simplistic ways of mapping the results and outcomes of development interventions. Fortunately today many development organizations feel this need and are in search for innovative and qualitative ways to monitor and evaluate in a meaningful way their efforts to advance gender justice.

    See Chant, Sylvia. Re-thinking the “Feminization of Poverty” in relation to Aggregate Gender Indices. In: Journal of Human Development, Vol. 7, No. 2, 2006. See also Fuwa, Bobihiko (2001) ‘A Note on the Analysis of Female-headed Households in Developing Countries, Mimeo. Agricultural Economics Department, Faculty of Horticulture, Chiba University. And Klasen, Stephan (2004) Gender-related Indicators of Well-Being. Discussion Paper No.102 (Goettingen: Georg-August Universität, Ibero-Amerika Institüt für Wirtschaftsforschung). (

  7. I am glad you raised this point, Duncan, though its not a new point. I have argued for at least 10 years that we should abandon this dubious 70% figure ( see my discussion of this in the UNIFEM reports, Progress of the World’s Women, 2000 and Progress of World’s Women, 2002).
    It is always going to be difficult to apportion household expenditure and consumption to different household members ( except perhaps for clothing).
    I think that measures of consumption poverty need to be supplemented by other measures of deprivation. I would like to see more emphasis on getting data on the proportion of adult women who have no money income of their own; and the proportion whose money income is less than the minimum wage, or less than a living wage, or less than the national or international poverty line. This reveals something about the extent to which women are impoverished in terms of lacking the power that money of your own brings.

  8. Like virtually all discussions of the feminization of poverty, the debate over the legitimacy of the 70% figure misses a crucial point. When it was discovered in the late 1970’s that female-headed families comprised a much larger proportion of the poor than they previously had (leading to the coining of the phrase “feminization of poverty”), it was universally overlooked (and, in the main, continues to be overlooked) that a significant part of that increase was a result of a decline in poverty, including the poverty of female headed families. Disadvantaged groups comprise a larger proportion of the very poor than of the somewhat poor. Hence, declines in poverty tend almost invariably to increase the proportion disadvantaged groups make up of the total poor, even as the poverty of disadvantaged groups also declines – as illustrated for example, in Table 1 of my “Can We Actually Measure Health Disparities?,” Chance 19(2) (Spring 2006) :47-51

    When it was predicted that in 1980 that by the year 2000 the poor would be entirely comprised of female headed families, no one recognized that at the point where a society verged on the total elimination of poverty one would expect all poverty to be concentrated in disadvantaged groups. Nor did anyone question why society would have an interest in there also being poor who were not in disadvantaged family units. In any case, at approximately the same time that the feminization of poverty was discovered, poverty ceased to decline, and poverty in fact never become more feminized than it was when the feminization of poverty concept was discovered.

    A key issue is illustrated in the following observation in Diane Elson’s March 19, 2010 comment:: “I would like to see more emphasis on getting data on the proportion of adult women who have no money income of their own; and the proportion whose money income is less than the minimum wage, or less than a living wage, or less than the national or international poverty line.” These data are indeed important and far more revealing of the circumstances of women than the proportion women comprise of persons with no money etc., just at (1) the proportion of women who are poor is far more important than (2) the proportion that women comprise of the poor. Further, (1) and (2) tend to be inversely related.
    See my “The ‘feminization of poverty’ is misunderstood,” The Plain Dealer Nov 11, 1987 (reprinted in Current 1988;302 (May):16-18 and Annual Editions: Social Problems 1988/89. Dushkin1988:; “The perils of provocative statistics,” The Public Interest 1991;102:3 14 (; “Comment on ‘McLanahan, Sorensen, and Watson’s ‘Sex Differences in Poverty, 1950 1980,’” Signs 1991;16(2):409-13 (; “Race and mortality,” Society 2000;37(2):19-35 (reprinted in Current 2000 (Feb)). See also the Feminization of Poverty sub-page of the Scanlan’s Rule page of (

  9. Let’s face it ducan and all the other posters, does it really matter, memes or others satstatics, thats a slippery downward slope, with a big ocean at the end. If you don’t want to drown, keep your head above the water, just love and help others where ever you can.
    No big mystery to solve there.

  10. A similar meme I came across recently is that women own 1% of the land in the world. This also appears to be a ‘statistic’ that started in the 1970s based on a guess. The best description of the source that I could find (in a quick search) is here:

    UNICEF state that women own between 11 and 27 percent of land in 5 sample countries in South America and probably 10% in parts of sub-saharan Africa:

    The only way that I can imagine the 1% statistic to be valid (and grossly misleading) is if it excludes all communally or jointly owned land, or land owned by trusts or limited companies. If this land is counted as not owned by women, then maybe only a small part of the worlds land is owned directly by single individuals, leading to a small percentage for both men and women.

  11. I think the real problem here is that both the author and the readers are allowing themselves to be driven by belief while under the mistaken impression that being close is good enough. Ultimately, the accuracy of those beliefs is irrelevant if you cannot provide any evidence for them. It’s not what you know, it’s what you can prove… you know, unless you openly admit to being more concerned with what amounts to propaganda than with facts.

    Science is NOT about estimation – it is about MEASUREMENT. Ideologies are about estimation and estimation is just an optimistic way of admitting that you’re just guessing. When you make such silly claims, you open yourselves up to attack and this is exactly how your ideological opponents will take you down – your use of fuzzy logic, misleading “facts” and un-sourced claims.

    Get your house in order.

  12. The figures in the Silvia Chant article do NOT refer to 70% of the world’s poor being women. But rather to the estimate that 70% of the world’s poor are FEMALE – i.e. women and girls. There’s a difference!

    Since the World Bank (late 2016) estimates that 50.2% of the world’s extremely income-poor are children, that would mean roughly 35% of the world’s poor are women, assuming that a clear majority of the adult poor are female.

    1. ‘assuming that a clear majority of the adult poor are female.’ What does that even mean when we also lament the lack of gender disaggregated data, and anyway, it’s extremely hard to measure the individual poverty levels within a household?

      If the number includes children, then it gets even sillier. If half those living in poverty are children, and let’s ‘assume’ as you put it, that they are roughly equal numbers of boys and girls (as they mainly don’t earn independent income, the measurement of their differential levels of poverty is well nigh impossible), then the disparity between adult men and women has to be twice as big as if the number refers to adults alone.In fact, you would have to have 45% of the world’s poor being adult women, and 5% being men, which along with a 25/25 split between boys and girls, gives your 70% figure. Completely bonkers.

Leave a comment

Translate »