Is ‘Getting to Zero’ really feasible? The new Chronic Poverty Report

March 21, 2014

What have we learned on getting public services to poor people? What’s next?

March 21, 2014

Killer factcheck: ‘Women own 2% of land’ = not true. What do we really know about women and land?

March 21, 2014
empty image
empty image

Cheryl Doss, a feminist economist at Yale University argues that (as with ‘70% of the world’s poor are women‘ ) we need to stop using the unfounded ‘women own 2% of the world’s View More: http://karissavantassel.pass.us/dossfamilyfarmland’ stat, and start using some of the real numbers that are emerging (while also demanding much better gender data).

For advocates, nothing is better than having a powerful statistic at your disposal to use in support of a cause. In the world of women’s property rights advocacy, there’s one statistic cited by advocates more than the rest:  Women own less than 2 percent of the world’s land.  It certainly is a great rallying cry to mobilize people in support of equal land rights.

If only the statistic were true.

Don’t get me wrong – worldwide, women do not own anywhere near the amount of property that men do. But this statistic and its variations, often repeated by women’s rights advocacy groups including Oxfam and Action Aid, is actually unsubstantiated and advocates are doing a disservice to their own cause by using it.

Women’s land rights are critically important. We know that they are correlated with increased empowerment and better outcomes for women and children. When women have a secure claim to land, they are less vulnerable when their husband dies or leaves.

Yet, using unsubstantiated statistics for advocacy is counterproductive.  Advocates lose credibility by making claims that are inaccurate and slow down progress towards achieving their goals because without credible data, they also can’t measure changes.  As some countries work towards improving women’s property rights, advocates need to be using numbers that reflect these changes – and hold governments accountable where things are static or getting worse. For example, the Demographic and Health Surveys, the leading source of nationally representative data on health and population, now ask women and men (of reproductive age) in ten African countries whether or not they own land. Their data shows a far more complicated and diverse landscape than is reflected by the statistic above.

African women farmersThe percentage of women reporting that they own land ranges from 11% in Senegal to 54% in Rwanda and Burundi.  But these numbers must be compared with those for men: The comparable figures for men are 28% in Senegal and 55% in Rwanda and 64% in Burundi.  The largest gender gap in land ownership is in Uganda, where the share of men who own land is 21 percentage points higher than that for women. The gender gaps are much larger if we consider only land that is owned solely (individually), by a man or woman, rather than include both sole and joint land ownership.  By having a more accurate picture of where women have more or less property rights, advocates can more effectively leverage their resources to achieve their goals.

Percentage of Women and Men (of reproductive age) Who Own Land

Country (year)

Household

 

Women

 

Men

% of households owning any agricultural land

 

% who own any land

(sole or joint)

% who own any land

(sole only)

 

% who own any land

(sole or joint)

% who  any land (sole only)

Burkina Faso (2010)

79

32

12

54

43

Burundi (2010)

86

54

11

64

50

Ethiopia (2011)

73

50

12

54

28

Lesotho (2009)

53

38

7

34

9

Malawi (2010)

80

48

23

NA

NA

Rwanda (2010)

81

54

13

55

25

Senegal (2010–11)

47

11

5

28

22

Tanzania (2010)

77

30

8

NA

NA

Uganda (2011)

72

39

14

60

46

Zimbabwe (2010–11)

63

36

11

36

22

Source:  DHS data compiled in Doss et al.

The good news is that data is slowly becoming available. In a recent paper that I coauthored, we use data from several new sources to examine the issue of women’s land ownership.  In addition, to the DHS data that we compiled above and that from the UN Food and Agriculture Organization’s women’s land rights database, we use the recent Living Standards Measurement Studies Integrated Surveys on Agriculture (LSMS-ISA).  While the DHS data tells us the percentage of women and men who own land, the LSMS-ISA surveys provide plot level data on ownership and management for six countries in Africa.  These surveys allow us to analyze what share of land is owned by women.  These are two very different perspectives on women’s land ownership, but they tell similar stories.

The plot level data allows us to consider whether land is owned by men, women, or jointly by men and women. In all six countries, women own less land than men, but the patterns differ across countries and vary depending on the definition of ownership that is used.

These new data sources point to key issues that need to be addressed, including the importance of defining whether we are interested in statistics on women (i.e. what percentage of women own land) or on land (i.e. what percentage of land is owned by women).  They also point to the importance of considering both sole and joint ownership.

As next steps, we can do two things.  First, we can stop using statistics that are unsubstantiated, even if they convey an important women rice farmerstruth.  Women do own less land than men – and we can provide more details for some countries.  But we don’t have a global figure.

So, what can we say?  The best approach is to make a general claim and then provide numbers on a specific country or region for which data is available. Here’s some possibilities:

  • Globally, more men than women own land. On average, across 10 countries in Africa, 39% of women and 48% of men report owning land, including both individual and joint ownership.  Only 12% of women report owning land individually, while 31% of men do so. (note that these data only include men and women of reproductive age.)
  • More of the privately owned land is reported as being owned by men than by women.  In Niger, only 9% of the land is reported as owned by women, 29% jointly by men and women and 62% by men.  In Tanzania, only 16% of the land is reported as owned by women, 39% jointly by men and women and 48% by men.  In Ethiopia, 15% of the documented land ownership is reported as owned by women, 39% by men and women jointly, and 45% by men.

Second, we can advocate for better data.  In surveys that ask, “Does anyone in the household own any land?”  a follow up question that asks “Who owns the land?” would allow us to exponentially increase our understanding of the gendered patterns of land ownership.   Yes, we need to consider the local context and what is meant by ownership; but this simple first step will go a long way toward providing an accurate statistic advocates can use to strengthen women’s land rights.

Cheryl Doss is a Senior Lecturer in African Studies and Economics at Yale University and a Public Voices Fellow at The Op-Ed Project.

From Duncan: Anyone know the origins of the 2% figure? Always instructive to track down the origins of these stats. Also, here’s a nice 2 minute video for International Women’s Day with (I think) some more reliable gender inequality stats (h/t Richard King).

 

24 comments

  1. Many thanks for posting this – really interesting. It highlights the problem that campaigners come up against again and again – the desire to have a neat, impactful statistic to describe an issue vs giving an accurate picture of an extremely complex problem which differs widely from context to context. It reminds of the problems civil society has had devising a statistic for the total amount of land grabbed globally. I actually wrote a blog about this on the Global Witness site last week which you can see here: http://www.globalwitness.org/blog/?p=144

  2. Thanks, interesting piece. I thought that in Ethiopia, all land was owned by the state … another myth?

    A quick interrogation of CABI’s database (c. 9 million abstracts) reveals 979 hits for [women+land+ownership].

    For Uganda CIFOR says it’s 7%:
    Gender and forestry in Uganda: policy, legal and institutional frameworks. Mukasa, C.; Tibazalika, A.; Mango, A.; Muloki, H. N.; Center for International Forestry Research (CIFOR), Jakarta, Indonesia, CIFOR Working Paper, 2012, 89, pp viii + 40 ., 48 ref.

  3. Very glad to see this post here!

    I have found an analysis of the source of this figure (actually a 1% version) here:

    http://www.theatlantic.com/sexes/archive/2013/03/women-own-1-of-world-property-a-feminist-myth-that-wont-die/273840/

    The key part is:
    in 2007 Krishna Ahooja-Patel … published a book called Development Has A Woman’s Face: Insights from Within the U.N.. In that book she attributes the formula to herself, and offers an unsourced sketch of the methods used, “based on some available global data and others derived by use of fragmentary indicators at the time, in the late 1970s.”

    The figures used for the formula were: women were 33 percent of the world’s formal workforce, and they were “only on the low income level in the pyramid of employment,” where—even in those lowly jobs, based on data from “several countries”—they earned 10 percent to 30 percent less than men. Therefore, “one could assume that women’s income is only one-third of the average income of men.” Since they were one-third of the workforce, and earned one-third as much as men, their total income was .33 * .33, or 11 percent. (She rounded it down to 10 percent.) In short, a guess based on an extrapolation wrapped round an estimate.

    What about the dramatic conclusion, that women “possess less than one-hundredth of world property”? She offers only this explanation: “if the average wage of women is so low, it can be assumed that they do not normally have any surplus to invest in reproducible or non-reproducible assets.” Hence, less than 1 percent. That’s it. In fact, she adds, “In reality the figure may be much lower.”

    Source? “Various UN Statistics.”

  4. Duncan, while your rigorous approach to statistical inequality data is laudable, there are some points we’d like to make collectively, in response to today’s blog, as feminists working in Oxfam to support women’s rights and gender justice – in partnership not only with our colleagues in our organisation and the sector, but with women’s movements in the Global South.

    – ‘Gender data’ is subjected to scrutiny in your posts with an intense attention that other data does not receive. The data on which Oxfam makes its claim that 84 people are as wealthy as the poorest half (Davos Report) is based on data collected from Credit Suisse and The World Top Incomes Database. We use these killer statistics in our work often, many are simplified and iron out nuances, and are therefore fallible and open to inquiry. It is in all our interests to use really accurate data. As gender advocates we are the first to need and want and call for more accurate statistics on gender inequality and the status and lives of women.

    – We feel that this blog is undeservingly critical of ‘gender advocates'; we are, as above, well aware of the need for more accurate statistics. We are operating in a world where many gender-disaggregated statistics are still not captured. And yet we are often pressurised by our organisations to provide ‘killer facts’ for use in development advocacy. We in Oxfam, like so many others, have called for better statistics repeatedly and for a long time, and on a huge range of themes; from land, GBV, mortality, on disasters, care work, etc.

    – The post could have also helpfully highlighted the efforts of gender advocates to advocate for the use of mixed research methods – that is, using qualitative data, alongside statistics and other quantative data -, to capture full, valid and useful evidence of gender differences or trends. This is essential in interpreting hard data on issues such as land ownership and their impact on gender equality and the empowerment of women.

    – Relating to the last point – but not limited to it as other quantitative data would also come into play here – we wish for researchers (including feminist academics seeking to help policymakers and practitioners bring about change for women) to interpret indicators and statistics in context. No focus on land and property ownership should leave out a focus on marital and other gender relations, because only then can we interpret the practical impact of land ownership on women and on gender relations. Women need truly independent property ownership to change gender power relations, and in the context of gender inequality in marriage, the family and wider society, titular land and property rights do not necessarily mean women are free to control or dispose of such property

    Stimulating debates in the form of blogs that perceive failures to ‘get’ the issues, alleged cover-ups or politically-naive strategies from those of us working to advance the rights of women and gender inequality do us, your colleagues, a huge disservice by discrediting our legitimacy. We would urge you to acknowledge how many of us have trodden this hard road of arguing for fuller and more rigorous research, data, and analytical thinking, to underpin and nuance the big picture of continuing marginalisation of women from power and resources in the world today.

    We would encourage you to consider seriously the risk that such posts oversimplify the issues we are confronted with day in and out. They can potentially harm the delicate alliances, positions, and incremental progress we make in partnership with women in the Global South to convince decision makers and the world at large of the serious gender inequality and injustices experienced by women.

    Ines Smyth, Caroline Sweetman, Thalia Kidder, Jane Remme, Emily Brown, Chloe Safier

  5. I appreciate the author’s concern for the way in which figures are often re-used and yet unsubstantiated, particularly in forming general statements for development texts. With respect to land ownership, this was also highlighted in the synthesis book produced by Debbie Budlender and myself on Women’s Rights and Access to Land, published by IDRC in 2011, which highlights the research findings from around 22 research projects conducted by African researchers across Sub-Saharan Africa. We recalled at the time of writing, the informal slogan of the UN Decade for Women (1975-1985)stating that, “while women do two-thirds of the world’s work, they receive 10 percent of the world’s income and own 1 percent of the means of production” and also questioned this. General statements aside, practitioners and gender theorists alike have been looking at the relation between women and land in much more complex terms for some time now. We know that large percentages of land are neither owned by individual men or women, but instead communally or by the state. The important areas to focus on are not just land ownership, but if women in particular have access and control over land and land-based resources and whether such access and control is enhancing their well-being and affording them control over their lives. Thus, as noted by Smyth, Sweetman and colleagues above, the importance of understanding gender power relations in the interpretation and analysis of any statistics.

  6. Thanks Duncan and Caroline for raising an issue that is worthy of a Blog in its own right – the usefulness or otherwise of “Killer Facts”. I have long been frustrated and annoyed at their increasing penetration of the Development Core. Do we really need to find ways to diminish complex issues into Twittersphere appropriate forms?

    My main issue here though is the unhelpful defensiveness of Caroline and the Oxfam team on the scrutiny of the data. Bad data is Bad data and often extremely unhelpful in educating the public on real issues. Do we really believe that a ‘1 in 5 people have a disability’ statistic is helpful in supporting those with real and important special needs? Maybe it is because I am an economist, but I want the disability effort focused on those holed up in houses for two thirds (or more) of their lives to receive support before spending money on those who are deaf in one ear or who have lost three fingers.

    I do not think this post in any way diminishes the value of mixed-methods, nor do I think it prioritizes one area of statistics over another. To suggest otherwise seems like nit-picking to me. Many data people like me take gender and womens issues really seriously. I think it is a great travesty that development agencies have failed to disaggregate all their data given its importance over the last 20 years. I also think that data scrutiny does not over-simplfy the issues. Quite the contrary, I think it highlights just how complex the issues are and the importance of using stats in a way that progresses the issues you fight so hard for. There has been a really interesting conversation going on about data and gender recently and I think we should see this instalment as a positive step in this debate, not a ‘discrediting’ of your (or the issues) legitimacy.

    1. Thanks Chris, On the value or otherwise of killer facts, I think they are most useful in establishing the broad framing of issues, especially early on in their trip down the ‘policy funnel’, but absolutely not for suggesting policy responses. So our inequality stats are great at getting the issue on the table, but you wouldn’t use them to reform your tax system, or to understand the nuanced impact of inequality on wellbeing

  7. A killer fact is such because it performs a dual role of being both a fact and crystallising a given judgement. It is especially important, therefore, for it to be accurate because it is designed for people to invest emotion into a cause and discovering it may be untrue in fact is doubly disabling.

    I too would be interested on the statistical analysis of this blog to justify your gender colleagues claim as to an inherent bias! Either they have been secretly stalking you waiting upon their moment or are making it up!

  8. As an academic working on data on women’s land and asset ownership, I am frequently asked for better global statistics. I wrestle often with the tension between the desire for global figures and the need to have data that reflects the local variations, even with a country. To understand the complexities at national and local levels, both qualitative and quantitative data are certainly needed, as Caroline notes. Understanding marital property rights and inheritance laws and norms are also critical. There is a lot of momentum at the moment for collecting better sex disaggregated data and my hope is it will feed into the important advocacy work that is being doing.

  9. All killer facts are true in the Discworld, where million-to-one chances pop up 9 times out of 10. In Terry Pratchett’s Discworld, and in the minds of many politicians and newspaper editors, facts change to fit the story. It is the story that matters.

    In this round world, there is such a thing as evidence, and it doesn’t always fit the story. But many people (apart from professional gamblers and bookmakers) need to be trained to assess uncertainty. We live in a world of uncertainty (which is also the name of a computer game to teach a skill needed to make decisions under uncertainty).

  10. Those of us who sent the earlier post wish to apologize for addressing the reply to Duncan rather than Cheryl Doss. Needless to say we appreciate her work and insights. It seems clear that we all agree on the need to question global stats, and look forward to a continued dialogue with both.

  11. Earlier this year my wife received a leaflet from a large INGO with one of the discredited gender facts. She read the ‘fact’ said “That’s rubbish!” and binned the leaflet. The rest of the leaflet may have been full of nuanced qualitative arguments and stories, but after the headline lie she just didn’t read on and that is the trouble with the blantantly untrue gender statements that just refuse to go away.

    I am very pleased to see this post by Cheryl Doss simply because it is the first I have seen here that does attempt to put real data into the issue (I can see the numbers here aren’t the definitive answer – but they are a very welcome start). I am surprised that Caroline and her colleagues are not supportive as they “are the first to need and want and call for more accurate statistics on gender inequality and the status and lives of women”. If they want to encourage data such as this then they need to be a lot more welcoming when it arrives.

    Can more accuracy do any cause “a huge disservice by discrediting our legitimacy”? This reminds me of the debates around climate change a decade ago. I remember one campaigner telling me about a new climate ‘killer fact’ – when I pointed out that it couldn’t possibly be true she replied “That doesn’t matter – it’s really impressive”. Soon after ‘Climategate’ came along and set the movement back a couple of years, but since then Climate Change is almost completly accepted purely because it’s data gets so much intense attention.

    I am sure the cause argued by gender advocates does have real ‘killer facts’, but until the old killer lies are out of circulation there is no room for the real story to emerge. Gender equality can only benefit from rigorous research data (which needs intense scrutiny) and robust analytical thinking, so thank you Cheryl (and Duncan) for starting it here.

  12. I really appreciate the response from Cheryl Doss, and I absolutely welcome her analysis which indicates that we are on the same page re the need for more detailed and disaggregated data to support our efforts for women’s rights . I strongly welcome research that tells us more about the true picture of women’s property and land ownership, and its connections to the empowerment of women and the work of Cheryl and her colleagues makes an essential contribution to this aim. The long years spent trying to manage the tension between lack of data and the thirst for headline data on the part of colleagues in advocacy, campaigning and fundraising is such an important issue and we are still struggling with it. Many feminists have long lamented the ways in which rigorous and complex research have been diluted and distorted in the interests of consensus building and direct campaigning messages. Anne Marie Goetz (now of UN-Women), writing in 1998 with Sally Baden, wrote a powerful critique of ‘killer facts’ including the UN’s claim that ”Women account for two thirds of all working hours, receive only one tenth of work income and own less than one percent of world property” that is the focus of the current discussion. They bemoan the fact that these kinds of statements are taken up in development discourse and repeated endlessly to the point of becoming clichés, as a justification for attention to gender inequality in access to resources. They argue that “whilst highly effective as an advocacy slogan.. the claim had no basis and thus had the potential to backfire and discredit feminist research”..(Baden and Goetz 1998 in (eds) Jackson and Pearson 1998,p 23)
    http://www.amazon.co.uk/Feminist-Visions-Development-Routledge-Economics/dp/0415157900).
    Sadly, it would seem that this kind of discrediting is still continuing. But as we come to know more about the ways in which the gendered distribution of assets, including land and other resources, have changed over time and vary between different countries and political contexts, we will be better equipped to advocate for policy and practice which is more effective in terms of achieving progress towards gender equality and sustainable development.

  13. Definitely useful for advocates, educators and researchers. Compelling piece for those seeking “better and further stats,” as we say here. It absolutely affirms the call to update stats of the ’70s and ’80s, which continue drive gender work and analysis, especially in this millennium. Of course, there is the complexity factor that difference analyst and advocates have always pointed to but also the relativity across nations and continents. Those should not however rule out the world, regional, subregional or even national aggregates, but definitely not those based on questionable stats.

  14. Very true. Another gender myth is : “Women produce all the world’s food with no resources”
    If that would be even remotely possible,it would solve a lot of problems!

  15. This is a very interesting discussion and I am pleased to see it on this blog. Cheryl Doss and her colleagues have made a very important contribution to our understanding of women’s land rights through their efforts to clarify the statistical evidence on women’s land ownership and add nuance to oft-quoted and over-generalised statistics.

    One thing not yet mentioned in the discussion here is the relevance of customary land rights. Cheryl rightly notes that much hinges on the definition of land ownership, but how this is defined by governments in their formal land legislation and how it is defined by different men and women in practice can be very different things.

    In all my field research on land rights in Africa, women have very commonly said they owned land if asked that question directly. If you then ask what they mean by owning land, the explanations will vary widely according to the local customs and practices and to their understanding of those customs and practices and their country’s formal land law. In the majority of cases the meaning would not be ‘ownership’ as it would be understood in the UK or the USA. And the relevance of even using that Western ‘ownership’ concept in a survey can be called into question if it is the customary land tenure arrangements that are socially legitimate – and if policy and field-based efforts to strengthen women’s rights within customary land tenure may be far more helpful to protecting women’s land rights than formal ownership-related policies such as land registration.

    Qualitative fieldwork enables this kind of careful unpacking of people’s perceptions of their land rights. However when national level household surveys on which most statistics on land ownership are based are carried out by enumerators with limited understanding of the complexity of issues around land rights and whose training is focused on getting the right ticks in the right boxes as quickly as possible so as to complete their target number of questionnaires each day, the data that emerges from the surveys – even when analysed with nuance – is always going to be limited.

    This is why the next step in improving our data on women’s land rights is not just to gender-disaggregate more statistics and improve quantitative survey methods and analysis, but also to advocate for the consistent inclusion of qualitative approaches on a par with the emphasis given to quantitative approaches. Then, once we have achieved that, we can advocate for openness to considering the most effective policies and strategies for supporting women’s land rights in specific contexts that may then be better understood.

  16. Cheryl Doss is correct to challenge us to be sure of the data we use. As others have mentioned above this could and should be applied to all our use of data not just that on gender issues.
    The alternative figures that Doss sites are, however, highly questionable as well. Elizabeth Daley has touched on some important points about this above. Having done large scale quantitative and other qualitative research work I have seen the confusion around the concept of ownership. And further what ownership means will vary in different countries (yes the state owns the land in many of the countries mentioned by Doss) and in different statutory and customary systems.
    A long term study in Tanzania by Askew, Maganga, Odegaard, and Stein has found in one area that of 520 households, interviewed who received certificates of customary rights of occupancy in a land formalisation process, all where held by men except 11 in joint title with spouse. That is 2% in joint title, if we take joint title to indicate 50/50 ‘ownership’ we have 1% of women ‘owning’ land.
    Even less believable than the Tanzanian figures Doss refers to are the Burundi figures. Burundi still has legislated discrimination against women in land rights and inheritance.
    Four weeks ago I was in a village in Tanzania looking at outcomes from what was seen as a successful land formalisation project, one that emphasised land rights for women. Hadija (not her real name) proudly showed my her land, explained how she was happy to have it registered in her and her husbands name, told me she got a loan based on now being registered as the ‘owner’ of the land. On returning to the village office I checked for Hadija’s name and land in the registry and copy of certificate held in the office. Hadija’s name was not in the registry, the land she had shown me was registered only in the name of her husband. Hadija’s name was on the certificate, not as an owner, but as part of the family along with her husbands sister.
    Getting the balance between accuracy and simple campaign messaging is never going to be easy.

  17. I have appreciated all of the comments on this blog. The question of what constitutes “ownership” and how it is defined and understood is one that is very complicated. The one thing that we can do in reporting these data is to always clearly state how ownership has been defined and measured in any particular study. This makes comparisons difficult — are the measures based on asking people if they own land, asking them if their name is on an ownership document, or actually looking at the documents themselves. Even the documents may not always confer the rights for women that we are interested in. Who was asked? Was it only the household head or was it individual men and women?

    One other aspect that is important in reporting these statistics is to pay attention to the unit of analysis. Are we talking about the percentage of women who own land, the percentage of plots that are owned by women, the percentage of titles on which women are listed as owners, or the percentage of owners who are women? These may all be useful, but they are different statistics and it isn’t appropriate to compare them as though they are presenting the same information.

  18. This has been a fascinating discussion to read. I find myself agreeing with parts of what everyone has been saying. Yes, concepts of “ownership” are highly variable, we need to consider customary rights, and bring in qualitative research. And it is unfair that gender data gets a higher level of scrutiny than for other issues. But given that we now have quite a bit of attention to gender, I’m nervous about overblown statements that might discredit the serious work on gender that both researchers and activists are doing. So I cringe when I see statements that “women produce 60-80% of the world’s food”, or these “killer facts” on land ownership.
    As a qual/quant gender researcher (and, full disclosure, someone who has worked closely with Cheryl and knows that she is fully aware of these nuances) I think my question is how we can convey nuance and still get attention. What can researchers provide to campaigners or development practitioners that is useful for getting attention to the issues (whether that attention is from the public, policymakers, donors, or program designers), and directing attention to the appropriate types of action?

  19. It seems to me that measuring the percentage of land owned by women and the percent of women who own land are two entirely different things, which is an issue that might be worth clarifying before debating the methodology used to measure these things.

  20. Does anybody know where co-ownership exists, if the woman automatically gets right to the land? It might be another myth, but I was of the belief that womens rights to inherit land were restricted?

  21. I’m sorry, but since you are still presenting opinion as facts, i cannot accept this apologie, it is not hearthfelt, it doesn’t show how bad a lie can be, you are basically saying that we need to lie to people, because if we don’t, people will do nothing.

    Thank you for your obvious contempt for all humanity.

Leave a comment