The Small Enterprise Foundation (SEF) actively aims to work aggressively towards the elimination of poverty by reaching the poor and very poor with a range of financial services to enable them to realise their potential. So, it is imperative that SEF understand 1) how to reach poor individuals and 2) how to track changes in poverty over time. Internally, SEF employs two different tools (the Participatory Wealth Ranking system and the Progress Out of Poverty Index) in order to support these objectives. Figure 1 below highlights the key differences between the two tools. The following pages describe the tools independently and through a comparison.
The Participatory Wealth Ranking (PWR) is a poverty assessment tool that engages community members to share their knowledge and opinions. Through its facilitation, community members define poverty indicators and identify which households are experiencing poverty.
The process begins with a question, “what is poverty?” Community members must determine indicators of wealth and poverty.
Then, SEF engages the participants in a mapping exercise to draw a map of the entire community or village. Every house or plot in the area is numbered. Groups of approximately five participants each rank all of the households in the village in terms of these indicators. Households are grouped into different buckets from the most rich to the most poor. SEF repeats the process with multiple reference groups within the community, and records an average score for each household.
One of the results of this process is the ability to identify the households who are poor and very poor. SEF uses this information to create a target population within a village. Additionally, the act of listening to the community sentiments rather than prescribing poverty assessments helps SEF to gain local knowledge and develop trust in the area.
The Progress Out of Poverty Index (PPI) is a survey used for poverty measurement that identifies how likely a client or group of clients is to be living below previously identified poverty lines. It was developed by Grameen Bank in conjunction with the Consultative Group Against the Poor (CGAP) and the Ford Foundation. PPIs are used in more than 40 countries and hundreds of MFIs.
The PPI is a favourite tool for poverty measurement in the MFI community because it is inexpensive and quick to administer. But beyond its ease-of-use, it relies upon sound statistical modelling. That is, it is constructed alongside government reports on national incomes and expenditures. PPI questions are proven to be correlated with income and expenditure levels, which are used as a proxy for poverty. Scoring weights are determined based on the strength of each variable’s correlation with income and expenditure levels. The PPI is therefore a simple survey with sophisticated backing. Because the PPI reports statistical likelihoods, it is of course most accurate when used to assess large groups of clients, and is less accurate when used to understand and individuals’ likelihood of poverty.
SEF uses the PPI to develop a deep understanding of its clients’ poverty profiles by geographic region and by financial product. The PPI also contributes to the organisation’s robust data set on client poverty, which is used to better understand how different groupings of clients are accessing and benefiting from SEF’s services. Most importantly, perhaps, SEF uses the PPI to track clients’ progress over time. SEF collects baseline PPIs for all incoming clients, revisiting a sample of these clients after two or more loan cycles. This allows SEF to determine how clients’ incomes and expenditures are changing over time.
The PPI and PWR measure poverty in different ways, and therefore, they cannot be expected to produce the same results. But SEF was interested in examining how much overlap exists between the two tools. To understand this, SEF looked at PPIs and PWRs for a sample of 1,400 incoming clients.
At first SEF realised that correlation between individuals’ PPI and PWR was very low. More specifically, across the sample a correlation coefficient of only 0.11 was found. However, this analysis relied on individual-level findings, and as mentioned above, the PPI scores can be viewed as unreliable when assessing individual clients.
Therefore, SEF elected to compare PPI and PWR scores by creating groups of clients. Figure 2 below illustrates each group of clients (organised by PWR score), the number of clients in each group, and the average PWR and PPI scores within the group.
Using these groups, SEF observed that the likelihood of poverty seems to increase as the PWR score increases. This relationship is better depicted in Figure 3 below.
The graph above highlights a correlation coefficient between PPI and PWR of 0.946. This reveals that there is in fact a very strong positive relationship between the two poverty assessment tools.
PPI and PWR are two tools that approach poverty measurement very differently. PPI relies on government data whereas PWR relies on local knowledge and perceptions, for example. Because of the key differences in methodology, PPI and PWR can both be used to understand and interpret poverty data in different ways. Combined, SEF uses PPI and PWR to target clients, ensure outreach to the poor, and monitor and evaluate changes in poverty levels over time. Despite the fundamental differences in these approaches, PPI and PWR are highly correlated in a positive direction when assessing large groups of clients.
To engage in further discussion about the use of poverty tools at SEF, please e-mail SEF at info [at] sef [dot] co [dot] za.
|Two Approaches to Measuring and Understanding Poverty.pdf||309.05 KB|