Return to MandE NEWS
1.1 A half-day meeting was held at ODI to discuss indicators of poverty and their operational significance. The context was provided by the ?DAC' goals and the debate concerning their use in the run up to the launch of the White Paper.
1.2 This note provides a record of the main themes raised in the eight short presentations and the discussion which ensued. A list of the presenters and attendees is attached, together with short summaries of presentations where these were available.
1.3 In his introduction, Simon Maxwell raised the dilemma of the apparently contradictory trends in discussion on poverty. On the one hand, international conferences and donor agencies are increasingly inclined to adopt global targets which are simple, quantitative and unambiguous: reduce poverty by half by 2015 being a good example. On the other hand, theoretical debate points to the increasingly complex conceptualisation of poverty, to the importance of issues like participation and social inclusion, and the imperative of privileging local perceptions.
1.4 Rosalind Eyben (Chief Social Development Adviser, DFID) continued by emphasising the potential value of using simple global messages to mobilise the international development community and partner governments. She emphasised that DFID was very concerned to work with other DAC members, and with the academic community, to ensure that qualitative, human rights and participative aspects are, however, taken into account.
1.5 A brainstorming session identified the following framework for debate:
1.6 This short summary follows these points. There was considerable consensus on many issues (see below), despite the range of disciplines of researchers present, looking at poverty in both the North and the South. (Names in brackets refer to presenters.)
2.1Everyone recognises that there is more to poverty than low income - physical security, autonomy, self-esteem, participation, social inclusion, and human development, are all important. A commensurately broad range of indicators is needed to capture the multi-dimensionality of poverty. Income and expenditure are relatively easy to measure, in the sense that we have accepted indicators (Foster-Greer-Thorbecke); however, many other dimensions can also be measured, and a portfolio of indicators is available. Some are location-specific, others seem to travel well.
2.2What could crudely be termed ?economic' and ?social' indicators are likely to be closely related in those situations where income can be used to purchase social goals, and will thus move together. In these cases, any one indicator can act as a proxy for the others. Thus, money metric measures (income or consumption) can serve as portmanteau indicators. However, this is not always true - there are cases where social and economic indicators will move in different directions, in which case no single proxy will serve. Income will not allow an ?untouchable' to drink water from the village well, or parents from a socially conservative area to send their daughters to school (Naila Kabeer). In these cases we need indicators of both economic and social dimensions to capture the complexities of context.
2.3Pramilla Krishnan's presentation underlined the complexity of any enterprise to measure poverty. Efforts should be made to understand the local context when self-assessments of poverty are used. Looking at subjective poverty between 1989-95 in a particular region in Ethiopia, she found an odd correlation that those who owned more land were likely to consider themselves as having become worse off. This may be due to a growing uncertainty in the security of landholding arrangements rather than to an increase in income poverty, for instance.
3.1Pip Bevan's presentation added a further dimension: ?why are we measuring poverty and for whom?' Reasons included: to set realistic global targets; to inform public policy; to assist donors and NGOs in targeting aid; to assess the impact on poverty of public policy; to develop explanations of poverty; to inform civil societies in rich and poor countries about the incidence of poverty; to inform and empower the poor; and to make a living.
3.2 David Booth's paper also examines why, and concludes that donors (notably the World Bank) have conducted poverty assessments to increase their understanding of who the poor are (enrich poverty profiling); and what their perceptions are of the constraints to realising market-based opportunities and thus what their priorities are for public action. Very recently, however, more attention has been given to systematically exploring the causes of poverty and vulnerability, going beyond a heterogeneous catalogue of insights into poverty. This promises to enhance the effectiveness of government and donor interventions to counter poverty. It also brings us to the ?How?'
4.1 There was a broad consensus that standard household surveys had their role, being feasible on a large scale at relatively low cost, but that this non-contextual approach was best complemented by context-sensitive participatory appraisal methods. Such a multi-stranded approach draws on a broader range of indicators, allowing a fuller understanding of poverty issues based on what the poor themselves say. But, by combining methods it also permits the triangulation of findings, by testing the reliability of understandings based on different kinds of data or investigative methods (Booth).
4.2 Support for a context-sensitive approach was based on the consensus that local conditions vary, that local perceptions matter, and that there are ethical reasons not to impose outside values. The prescription of participatory methods also generated a number of concerns, however. Some felt that they undervalued the time of the poor and of women in particular, without demonstrating clear tangible benefits for participants. Also, they could be poorly executed, where insufficient training resulted in the biases of fieldworkers (sometimes under pressure to show results) distorting findings (Hugh Goyder).
5.1There was consensus that there is value in drawing on the insights of researchers examining poverty in both the North and the South. The presentation by Jane Pryer underlined the striking and increasing similarities in the nature of urban poverty in the developed and developing worlds. In both cases adverse socio-economic indicators were correlated with poor health indicators: access to food, health-care, health and nutritional status and premature mortality. This suggests that there may be more commonality than might have been thought between indicators relevant to North and South, and thus considerable potential for an exchange of lessons.
5.2 This was also reflected in the presentation by Jonathan Bradshaw, talking on the subject of poverty in the North. He gave an example of the local study methods being used in the European Community Household Panel Survey being adapted (by Rick Davies) for research in Vietnam. He concluded by mentioning the Rowntree Poverty Conference, which takes place from 18-20 March 1998, marking the centenary of the Rowntree Poverty Study.
5.3 Concern was also expressed that in both the North and the South indicators might increase the level of understanding of poverty, yet this did not necessarily translate into action to benefit the poor. Arjan de Haan's presentation raised the question of who was responsible for social exclusion, which he defined as ?the process through which individuals or groups are wholly or partially excluded from the society in which they live'. This could only be understood by examining the processes and institutions by which social exclusion is reproduced. Others felt that the social exclusion debate can expand in all directions, and that it is more fruitful to narrow the discourse to participation versus exclusion, and the nature of the resources necessary for participation, including social capital.
6.1The extent of aggregation of indicators that was valuable and defensible and role of indicators in governing the sequencing of poverty reduction measures was discussed. Taking the latter first, one view was that it was not possible to fire on all cylinders at once – that different dimensions of poverty should be tackled in a particular sequence for maximum impact. Indicators, therefore, should operate within a dynamic context, providing evidence of whether the sequence of preconditions had been met for poverty alleviating measures to have the desired effect. An example would be the reluctance of tribal villagers in India to allow their children to attend school as long as they are worried about where the next meal is coming from. Once incomes have increased, and people begin to feel more secure about livelihoods, then a school is supported. More generally, there was consensus that indicators had an important role in tracking variability and vulnerability (exposure to risk). In telling us more about the dynamics of poverty, what Stephen Jenkins has called ?bottom-end churning', indicators may shed light on causes and solutions.
6.2While it was recognised that governments were not necessarily committed to poverty reduction, some stressed that it was important that the governments be engaged in the process of measuring as they are the ones deciding public expenditure. Without ownership of the process by governments, little will change.
6.3One thing that we do when we have indicators is to aggregate them. In particular, indicators may be used to reflect progress against national or international targets. For some, such targets are simplistic and reductionist, driven by the ?tyranny of replicability'. In particular, the DAC target of reducing absolute poverty by half by 2015, defining poverty in income terms, is a pretty poor guide to what we are trying to do and how we might achieve it.
7.1Simple targets have value, and politicians have a perfectly legitimate need to find simple messages that will galvanise public opinion. However, many saw a tension between selecting indicators which can be universalised across countries and the complex realities of poverty. Thus the DAC goals are seen as failing to reflect the multidimensionality of poverty, and thus entirely ill-equipped to indicate the progress that they are intended to inspire. This view leads to the anxiety that a great deal of money is going to be spent operationalising the indicators and monitoring progress, and yet this money is going to be spent measuring the wrong thing. Secondly, and more important in the long run, targets influence the way programme money is spent: if we choose the wrong targets, we are going to get the wrong programmes. And thirdly, decisions made in the DAC are likely to have huge knock-on effects for data collection and analysis in developing countries, especially for scarce professional skills. Should scarce resources be spent on measuring replicable international indicators of dubious development worth, or should they go to produce locally meaningful analysis?
7.2 An alternative view agrees that the DAC goals have the primary purpose of motivating governments (North and South) to make poverty reduction their priority, but rejects the anxieties as misplaced since it considers (a) the goals to be laudable and relevant; and (b) they are goals and not methods.
7.3 (a) DAC goals complement the target of reducing income poverty with four social development goals (universal primary education, gender equality, reduced infant, child and maternal mortality, and access to reproductive health for all). Progress against the income indicator should be considered alongside progress against the four social development goals, and taken together would give a good indication of progress in reducing multi-dimensional poverty. (The main weakness is the lack of a goal/measure for empowerment/social exclusion.)
7.4 (b) The DAC goals are measures of progress and are not intended to provide operational insights or guides to best practice. It remains the responsibility of national governments (and international aid agencies) to devise and implement the appropriate poverty reduction strategy (which would necessarily also address the DAC goals). This would, of course, be expected to be developed on the basis of local studies and multi-dimensional and participatory research and evaluation, of the kind discussed above.
7.5 There was general agreement that the DAC-led (OECD/UN/WB) process of devising goals and indicators should be more transparent to ensure that the outcome reflects the best thinking available. Lionel Demery invited participants to contribute to the process by contacting Brian Hammond at the DAC. One suggestion from the floor was that the developmental priorities reflected in the DAC goals should be compared with the revealed preferences of the poor in a few case study countries.
8.1 Rosalind Eyben summed up by saying that the meeting had reconfirmed the enormous knowledge base available in the UK. DFID would attempt to draw upon this in its attempts to influence the international agenda in favour of a rights-based approach to development cooperation.
Prepared by Aidan Cox. 13 November 1997
Please Note: Copies of brief papers presented at the workshop can be obtained from ODI, via Aidan.
Mande NEWS: ODI Workshop moved to this new address on January 19th
1998. Since then there have been
visits to this page. See
MandE Statistics for more details of the usage of
MandE NEWS.
Return to MandE NEWS