The problem seems very simple: there are lots of poor people. But solutions are hugely complex. “Yes. That's life …”, said Richard King, who tries to work out what some of those solutions should be. So we talked of the complexities of the aid business, of measuring poverty and of measuring the success or the failure of efforts to alleviate it.
Oxfam is a charity. From its early days after the Second World War, when distributing food to the hungry seemed response enough, it has matured to trying to attack the causes of poverty as much as poverty itself. It is a major player in the scheme of things, one of the many non-governmental organisations that with varying degrees of governmental backing tend to end up doing the work at the sharp end of disasters and development. Richard King is its policy research adviser.
He is not a statistician: “I am a geographer turned half-economist.” He is nevertheless one of those through whom the statistics flow and on whom falls the burden of analysing them and advising on policy conclusions from them.
Measuring poverty – and measuring the effectiveness of efforts to alleviate it – is not easy. What data does he spend his day looking at? “Lots of different stuff. A lot of my work is around food and climate change. Global numbers on hunger, on food prices, on harvest volumes and those kinds of things occupy quite a bit of my time. A lot of my colleagues work much more on analysing household data sets and micro data. The research team now is seven people, who are part of a larger campaigns and policy team.”
So how does he identify the areas where Oxfam should campaign most? Is that part of his job? “Yes, I guess I'm involved in the conversations.” So what does it involve? The answer turns out to be a combination of quantitative and data-driven analysis guided by local understanding. “It's a bit of a mixed approach. We've got offices on the ground in 90 countries, so understanding the very localised context specific, the nuts and bolts of what is appropriate to do on the ground, is fundamental to what we do, and is vital.” Societies are made up of human beings, and if you don't understand how they feel, the best quantitative data in the world falls down. The world is full of examples of schemes that made quantitative sense but that communities just did not buy. In a UK context, replacing slums with tower blocks is an obvious example. In a global context every community is different: “so a lot of influence on decision-making flows through our local officers and the understanding which they have”.
Data and analysis tell big stories; local understanding is needed as well
But ground work is only part of what Oxfam does. It tries to change minds as well as conditions. “In terms of advocacy and making your case to the public and to the people who make the decisions – yes, a lot of that is based on understanding the data. But there's also a need undertake power analysis. You have to understand what the basic problem is, but as well as that you have to understand where there are opportunities for change on various issues. Because if you are faced with an intractable problem with no prospect of anything moving on it then you are wasting your efforts.” One again it comes down to the context in which these things occur.
Some things are unquantifiable. A hungry child is one. But the statistics of hunger do talk. “Hunger is an interesting one because obviously there are some very powerful narratives that you can use to help make your case. An example is that one in eight of the world's population goes to bed hungry. That is based on UN Food and Agriculture Organization chronic undernourishment statistics. That kind of global statistic is good for talking to the public and to engaged policy-makers.” (There are more of them in the box below.) “But once you get down to field-level operations – what you do in a refugee camp – it doesn't help you much. Then you are much more down to such things as the anthropomorphic measurements, around being underweight and underheight, and what that means in terms of adequate nutrition.”
At which point your interviewer should declare an interest. In his early days he worked for Oxfam during a famine in Uganda. We were issued with height-for-weight charts to identify the children who were undernourished – chronic undernourishment obviously makes children thinner. Except that sometimes it does not: it gives them oedema – fluid retention – which accounts for the swollen bellies on children that you see in photographs. As the World Food Programme guidelines (see http://www.unhcr.org/45f6abc92.html) put it: “Children with [oedema] should always be classified as having severe acute malnutrition regardless of their weight-for-height or weight-for-age z-score or percent of median…. [T]he weight-for-height and weight-for-age indices should be ignored when deciding which children have acute malnutrition.” This made little practical difference to us on the ground. Since every child was hungry, we rather tended to feed them all.
“So things like GDP, per capita income, access to sanitation and healthcare and clean water – a lot of these global statistics help tell the story to the public more than they help us do the direct programming”, says King. “Many of them are the basis of the Millennium Development Goal targets, the international development goals that were set by the United Nations in 2000, so again they are useful in terms of telling a story, framing a narrative. There is now a post-2015 agenda that will succeed it. As we move towards that, there's going to be another look both at what are the most appropriate goals and what are the most appropriate indicators and how can those indictors be improved. A lot of them, powerful as they are, have problems and assumptions underlying them.” The problems can lie both in the concepts and the reliability and availability of the data.
“For example, GDP is clearly a useful measure. A lot of people talk in terms of it, and it does define what a middle-income country and what a less developed country is in World Bank terms. But does it make sense to base your development trajectory purely in terms of increasing a country's GDP no matter what? Obviously not. There are clearly a lot of issues with environmental externality and a lot of maldevelopment that can get captured in GDP figures.”
- Almost half of the world's wealth is now owned by just 1% of the population.
- The wealth of the richest 1% amounts to $110 trillion. That is 65 times the total wealth of the bottom half of the world's population.
- The bottom half of the world's population owns the same as the richest 85 people in the world.
- Seven out of ten people live in countries where economic inequality has increased in the last 30 years.
- The richest 1% increased their share of income in 24 out of 26 countries for which we have data between 1980 and 2012.
- In the US, the wealthiest 1% captured 95% of post-financial crisis growth since 2009, while the bottom 90% became poorer.
The aid business has its jargon like any other. Maldevelopment is one of its more obvious terms. “For example, cutting down a rainforest and selling the timber in any given year would boost that country's GDP and move it up in the world league table. Whether it makes sense in the long term, both for that country's own development – the resources that remain available to it – and what that means for global carbon emissions and all of our collective futures in terms of climate change … Clearly making any decision based on GDP alone is a bad idea.”
The rainforest example is an obvious one, but the debates are live and sometimes bitter. After a pause of some years the World Bank is again funding several giant hydro-projects on rivers in Africa, Nepal and southeast Asia. Its reasoning is that they will connect many communities to dependable electricity for the first time while keeping carbon emissions low. Against that has to be balanced the environmental impact and the effect on indigenous peoples. The World Bank, with some reason, calls such schemes “transformational”. Their own 2009 review of hydropower (http://documents.worldbank.org/curated/en/2009/03/12331040/directions-hydropower) noted the “overwhelming environmental and social risks”, but they have recently decided that the gains outweigh the risks. Rachel Kyte, the bank's vice-president for sustainable development, has said that the earlier move out of hydro “was the wrong message … That was then. This is now. We are back.” Many advocacy groups strongly disagree. As we said, complexities are the norm in trying to reduce poverty.
Back to ways of measuring poverty, which can also be complex. Income alone is not sufficient –though the commonly used “dollar a day” is a sometimes useful rule of thumb. Things called multidimensional poverty indices have been developed, notably the OPHI index by Alkire and Foster of the Oxford Poverty & Human Development Initiative (http://www.ophi.org.uk/policy/multidimensional-poverty-index) Briefly, their version takes 10 indicators, including nutrition, child mortality, years of schooling, access to clean water, sanitation and electricity; gives weightings to each; lumps them together into three sections of health, education and living standards; and considers a family poor if it is deficient in at least 30% of the indicators (see Figure 1; see also Significance, February 2012, for more details).
Are these useful to Richard King and Oxfam? Again the answer is yes and no. “In terms of global policy work, our global advocacy work, it is not something we overtly talk about but a lot of it is very interesting. Many of these multidimensional indices are good for headline grabbing, but as development professionals what is useful to us is unpicking the index and looking at the individual indicators within it. The reason is that in aggregating the different indices and coming up with weightings for them, a lot of it becomes quite subjective.” The OPHI index, for example, gives more weight to years of schooling than it does to access to clean water – a judgement that could be argued with. “It is good to have one number to tell a story around, but once you come down to nuts and bolts and start trying to actually do something with it, it is the individual indicators rather than the multidimensional ones which are more interesting. If you are making an education effort, or a clean water effort, then the other indicators are not terribly relevant to getting education or water provision changed.”
Which not to say that the multidimensional indicators are not useful: “They do help to focus attention on where there are pinch points of lots of different types of poverty which are very pressing. The other place where we have used the OPHI approach is on our own evaluation of our programmes. We have outputs reporting on how well programmes have worked out, whether they've met their objectives, but we're also doing some deep-dive effectiveness reviews of particular programmes, and there we have in many cases adapted the multidimensional approaches taken by OPHI and constructed our own multidimensional indices to evaluate particular programmes in particular countries.”
There are several of these Oxfam-devised measures. One is around resilience programmes. Many communities survive on the edge. It takes only a small change in one or two conditions to tip them over it into poverty. What, then, does it take not to be vulnerable to such small changes? “We look across different indicators to get a handle on that. Similarly, we have campaigns for women's empowerment. Evaluating those is very much based on the OPHI approach.”
Last year two American economists1 produced a book called The Robin Hood Rules for Smart Giving. It argues strongly for measuring the value of aid in purely monetary terms – of dollar back per dollar spent. Michael A. Lewis (Significance, August 2013) has pointed out some of the flaws in that reasoning – counterfactuals among them; but cost–benefit analysis is something that aid organisations such as Oxfam can hardly ignore.
“There are various value-for-money hoops, for want of a better word, that we have to jump through. We have to acknowledge, and many of our donors acknowledge, that pure cost–benefit analysis doesn't really allow you to count the more intangible or difficult thing to measure. If you are delivering mosquito nets it is quite easy to work out how many you have distributed to how many villages at what cost; but as an organisation we are not that much involved with direct service delivery like that. When you are involved in programmes which are more to do with women's political participation and empowerment, and equitable government, those sorts of things don't lend themselves so well to cost–benefit analysis.
“Some of the important things are in any case unquantifiable – you do some things because they are good per se; relieving distress is just something we ought to do. But I guess the other danger is the timescale over which you judge your effectiveness. Do you go back a year after the programme has closed and see how people are doing, or do you go back 10 years later? And how do you try to quantify all the knock-on benefits that might exist? So a lot of the stuff that matters in what we do is unquantifiable. Not that that stops us trying to find ways of quantifying it. It is certainly a lot more difficult than with direct service provision.”
So how do you measure the success, or the failure, of an aid project? Go back after 10 years and imagine what would have happened without it? “We do not systematically return 10 years after. We are doing more on randomised control trial approaches, where the counterfactual is a comparable community which hasn't seen an intervention.”
And what conclusion can be drawn from those trials? Do they shed light in the contrasting approaches we mentioned above – the huge transformational schemes of the World Bank, versus the small-scale – sometimes tiny-scale – local projects of organisations like the acclaimed Grameen Bank, which lends tiny amounts within villages to allow people to buy such things as a sewing machine to set themselves up in business, and which has won Nobel prizes in the process? Do smaller-scale, more local involvements give greater results, or is that oversimplifying?
“It is oversimplifying to some extent. Development always has to be from the bottom up, it has to be built on an understanding of the local context and in collaboration with local partners who have a much better understanding of the place and the people. That is our basic modus operandi. But at the same time we're never going to leverage any meaningful change if we are dotted here and there just doing isolated micro-projects. So a lot of our work is around how do you get that leverage, how do you scale up such programmes, how do different projects relate to one another, how do you get a knock-on effect. And, although it is bottom-up, how do you start to effect national rather than purely localised change.
Evaluating a project in Tanzania
Oxfam is working with local partners in four districts of Shinyanga Region, Tanzania, to support over 4000 small-holder farmers (54% of whom are women) to enhance their production and marketing of local chicken and rice. The producers are encouraged to form themselves into savings and internal lending communities and are given specialised training and marketing support.
Two outcome indicators were piloted to judge the success of the project. They were the percentage of targeted households living on more than £1.00 per day per head; and the percentage of supported women who are meaningfully involved in household decisions and are able to influence affairs in their communities.
A household survey (n = 457) and a women's questionnaire (n = 446) were administered to randomly selected chicken and rice producers in 86 intervention and matched comparison villages. To compare like with like, statistical analysis was undertaken using propensity score matching with exact matching by product type to control for observable differences.
Overall, there was no statistically significant difference in the proportion of households living above £1.00 per day per head. However, when examined by product type, a different picture is revealed. Households supported through the chicken value chain intervention are significantly better off, with the proportion being 63% versus 48% for the intervention and comparison groups, respectively (p-value < 0.05). However, this did not translate into a notable difference in their food security. It should also be noted that both a serious chicken disease and drought have hit the region, which have negatively affected both groups of producers.
Interestingly, women in the rice producing groups were assessed as having greater household decision-making power: 37% scored positively, against 17% in the comparison group (p < 0.05). The overall difference for both rice and chicken producers was not statistically significant. However, women in both groups scored better than their comparators in relation to the international self-efficacy scale (p < 0.01) and, at 45% versus 32%, were found to be more likely to own productive assets (p < 0.05).
“You hope for a cascade effect. A lot of our work is a combination of not just traditional long-term development programmes, as you might imagine, but also the advocacy side of things, so we build partners’ capacity, people we work with, to advocacy, lobby for change, get involved in decision-making forums as well as nuts and bolts on the ground.”
The World Bank claims that poverty in the world halved between 1990 and 2010. If that is true, it is a major achievement. The pessimist or cynic in me finds it hard to believe, so I sought Richard King's expert opinion.
“I think that the number deserves credit to some extent. Trying to quantify world poverty is obviously riddled with complications. A problem with a lot of these global figures is that the empirical basis for them can be quite weak. The World Bank figure relies on a lot of household surveys. A lot of those surveys in many countries are missing; a lot are quite old now; a lot are unreliable. Another issue is the purchasing power parity, which is not always credible, especially in China.” If you are working out how many people survive on a dollar a day, the currency exchange rate obviously matters, but so does what a dollar or its equivalent will actually buy in the country in question. “There is a lot of evidence that purchasing power parity isn't particularly credible for China at the moment, and obviously China weighs very heavily in the calculation, because of its very rapid progress and its huge population. So the World Bank figure is a commendable effort with a lot of clear weaknesses.
“The other difficulty with that figure is that it is a very binary threshold. It sets a dollar a day as the poverty line. If you look at people who have come out of the slightly higher poverty line of $2.50 a day there has been a lot less progress. There is still a lot of transient poverty, of people moving just above and below that line. And that is always going to be a problem with those very binary ways of measuring progress.
“That is the reason why Nancy Birdsall at the Centre for Global Development has come up with a proposition for using a country's median wealth as a better indicator of its population's poverty or wealth.”
What, then, of the future? Is poverty, however you care to define it – perhaps as the ability to lead a decent life – increasing or decreasing in the world?
“It is difficult to say. We are all reliant on these global numbers which, as we have seen, have their inherent problems. Undoubtedly there has been a lot of progress in many areas since the 1990s. Until recently we have been on the right track. We might have been moving more slowly than is ideal, but progress has generally been in the right direction.
“But one of the increasing concerns now is not so much absolute income poverty as the level of inequality; and we are entering an era where we are banging up against some quite serious resource constraints in terms of food, climate, water and land. We are yet really to see the implications that that is going to have. The future is going to be more precarious, and more volatile.
“Population pressures are pushing on those resources, but more important than the level of population is the level of consumption, which is clearly the bigger issue.”
Should we all, then, reduce our standard of living?
“We should make our lives less full of material possessions. It might mean working out what well-being really is. I wouldn't necessarily see that as reducing our standard of living; more as redefining what it means to live well.”