poverty analysis

Poverty elimination involves more than increases in income. Poor people are concerned about the lack of control they have over their lives. Many people are unable to access the resources made available by growth because of discrimination on the basis of gender, ethnicity, age and other social distinctions. This means that not only sustainable growth, but human rights and social inclusion must be at the heart of all development policies.The elimination of poverty, and hence also sustainable growth, can only be achieved through the engagement of poor people in the development processes which affect their lives.

However poverty assessment and statistics generally continue to focus on $1 a day measures at the household level. Poverty levels are assessed through measuring a combination of one or more of the following: income, assets, expenditures and consumption each assessed by a pre-determined checklist of measurable indicators. However these measures all face similar challenges of coping with nonmarket values and taking into account differences in sustainability of access to or control over assets and different sources of vulnerability to shocks and downward mobility.

Economic measurement challenges

Firstly both World Bank global poverty estimates using the $1 a day definition and national level poverty assessments and Housing surveys are notoriously methodologically problematic and analytically and politically contentious (Sillers nd; Reddy and Pogge 2003). In practice even cash incomes from work are often extremely variable even from day-to-day, very poor people rarely keep any records, and incomes need to be balanced also against costs of various types eg transport, childcare and equipment. Income measures of poverty have proved difficult and unreliable in practice (Grootaert .). In designing the PTTs therefore a number of key measurement issues will need to be addressed:

• How precise will measurement at the cut-off boundary need to be to classify people as poor or non-poor? eg how do we determine whether someone is living on 0.0009US$ or 1.0001US$ a day? This is here of more than just academic interest. For the person interviewed it may mean the difference between receiving and not receiving loans and services.

• How are international differences in national level poverty measures and the difficulties in establishing purchasing power parity to be addressed? To fail to do so might disadvantage programmes working in very poor countries (eg Bangladesh and Sub-Saharan Africa) even though the absolute poverty levels of clients deemed 'poor' by their national estimates might be the same as those deemed 'very poor' according to estimates in slightly better-off countries (eg Latin America and Eastern Europe)?

Return to top

Non-market incomes

• nonmarket incomes and nonmarket exchanges of goods and services within and between households

• the sustainability of such incomes over time due to differences in security of ownership or entitlements in relation to productive and other basic assets like land, house sites, market sites and different sources of vulnerability due to contextual economic, sociopolitical and environmental factors.

Return to top

Most current poverty
assessment tools are generally conducted
at the individual level, but collect
data for the household as a whole.
Total household measures are then
divided by numbers of people within
households.

However this fails to
consider intra-household differences
in access to and control over income,
assets, expenditure and consumption
allocation between women and men,
and between children, youths, adults
and the elderly.

An extensive body
of research has shown that merely
dividing data from the household
by numbers of household members (even
where household members can be accurately
identified in the contexts of household
instability, migration and polygamy
which characterise many poor and
rural communities) is not an accurate
predictor of the poverty status of
different individuals within households.

• How can tools designed to assess household-level poverty be adapted to incorporate intrahousehold gender and age inequalities to reliably predict individual poverty? Again this is of more than academic interest in view of the considerable potential of micro-enterprise interventions to strengthen the position of currently disadvantaged and vulnerable individuals within households: adolescent girls, women in unstable relationships and junior wives, elderly women, the disabled and so on.

Finally, and equally importantly, the USAID initiative will face particular problems of motivation, both at programme level, and also of respondents themselves. Programmes will inevitably have an interest in presenting the best picture possible, which in view of pressures for 'upward drift' to meet financial sustainability targets may lead to overestimation of very poor clients. Unless the Poverty Targeting Tools can be integrated into programme implementation and/or other poverty-related strategies eg poverty targeting itself, market research or poverty impact assessment, there will be considerable pressure to get the task done as quickly as possible, undermining reliability. Perhaps the most critical challenge will be encouraging respondents to be open and honest in their responses. Very poor people have limited time. In this case they may also have even more limited motivation to reveal all their sources of income, assets, expenditure and consumption if in this case ' very poor ' = 'good' and may prejudice decisions about their access to credit, training or other services. This creates for the USAID initiative a situation rather different from academic economic research or national poverty surveys. It will be crucial therefore, in the interests of reliability of data as well as the respondents themselves to consider:

• How can the Poverty Targeting Tools be most cost-effectively implemented without undermining reliability of information?

• How can the the investigation process be designed so that respondents will find it worth their while to answer questions reliably?

To adequately cover all these complexities, the range of questions and indicators to cover all the market and non-market 'economic' dimensions of poverty and the complexities of intra-household relations would need to be considerably extended. However counter to this is the practical problem of limited time and resources available to programmes and organisations being asked to demonstrate the effectiveness of their poverty targeting.

Meeting the challenge of reliability and accuracy will require identifying criteria for 'optimal ignorance' (Simanowitz, Nkuma and Kasim 2000) or ways of balancing:

• adequate and credible reflection of the complexities of the reality which very poor women and men face in coping with poverty.

• time and resource constraints of the implementing organisations and the need to limit diversion of resources from development implementation in the context of pressures for financial sustainability.

• costs and benefits to respondents in view of their own limited time and probably also interest in a process which may have negative as well as positive practical repercussions.

The best ways in which these challenges can be met will vary between contexts and types of organisation and enterprise intervention. The USAID Amendment will also require Tools to go beyond assessing numbers of very poor programme participants to including information about assistance received and estimating its costs. How is '%@ of assistance targeted' to be assessed? Does this include costs of delivery or only resources received? How is the amount received by savers to be compared with that received by loan recipients? How this should be done is not clear until more is known about the chains of implementation and these issues are not discussed here.

Return to top