• No results found

Analysis of Poverty Levels and Dynamics in Rural Kenya 9

2.3 The Setting and study sites .1 Poverty trends in Kenya

2.4.1 The Stages-of-Progress method

The Stages-of-Progress involves facilitated focus group discussions followed by household-level interviews. It is an adapted participatory poverty assessment method that relies on community-based poverty definitions to assess household welfare. This method is a relatively rapid, effective and participatory way to learn about poverty processes at both community and household levels. In addition, this method captures many of the advantages of quantitative approaches, including the ability to aggregate numerical information and can be applied in a modular manner, linking with other methods including household surveys. The Stage-of-Progress methodology is described in detail in previous studies (Krishna, 2006;

Krishna et al., 2006). We briefly describe the main steps.

Assemble a diverse and representative community group:

In each community, the focus group discussions involved 20-25 individuals from different households, who were knowledgeable about the community and households within their

village. Different groups of households within the communities were represented including poorer households.

Clearly present the objectives of the exercise

Clearly explaining the objectives of the study to the community groups is crucial in managing any expectations. Particularly, the fact that there would be no benefits or losses from speaking out freely and frankly and no development project to be implemented. This helps to remove any incentive anyone would have to misrepresent themselves or anyone else as being poor.

Define and describe poverty collectively

This step involved eliciting a common understanding of concepts of poverty based on a shared conception of ‗poorest family in the community‘. Once this was done, each community group defined locally applicable stages of progress that poor households typically follow on their pathways out of poverty. The group successively answers the question ‗What would this family do with additional resources‘? Which expenditures are the very first ones to be made? Until they reach the point at which the household would be considered prosperous. We are interested in the actual experiences of typical households  not the community‘s opinion of what a household should or should not do. Community groups were asked to identify the poverty cut-off and prosperity cut-off points on the progression of stages. The poverty cut-off denotes the stage after which a household is no longer considered poor. It is equivalent to the concept of the poverty line commonly used in conventional poverty studies.

Refer to a well-known signifying event or events to demarcate an earlier period

Well-known significant events were chosen to demarcate time periods being used in the study. The aim is to ensure that people across all community groups in the study zones discuss same reference time periods. For example, we used the El Niño rains in 1997 and pre-election period in 2007 as reference points for 1997 and 2007, respectively.

Ask about households’ poverty status today and in the earlier periods

Using the stages of progress developed as a yardstick, the position of each household in the community for each time period was determined by the community groups through consensus.13 The exercise involved going through each household in the community, one at

13 A complete list of all households in the village was prepared by the village representative (village elder in advance) and verified by the community group for accuracy and completeness. Verification of the list is usually done during the first day of the community meeting.

a time, and having the community group come to a consensus as to what stage the household is at the present time, what stage they (or their parents‘ household) were at some point in the past using the significant events. Three reference points — 1997, 2007 and 2009 — were selected for this study. The reference years were chosen to coincide with the periods for which panel data existed. There were relatively few disagreements regarding a household‘s position on the ladder and those that arose were resolved through discussion and debate among the participants.

Assign households to particular welfare categories

Based on their welfare status in each year, households were assigned to one of the four categories below, in relation to the poverty cut-off:

A. Poor then and poor now (Remained Poor) B. Poor then and non-poor now (Escaped Poverty) C. Non-poor then but poor now (Became Poor)

D. Non-poor then and non-poor now (Remained Non-Poor)

In this study with five data waves, there are many possible combinations. However, we look at the long (1997–2007) and short (2007–2009) period changes.

Ascertain reasons for change or stability for a random sample of households

The sampling procedures for follow-up are usually determined by the objective of the research. We selected a random sample representing 35% of households from each village, spread across the four poverty categories (remained poor; escaped poverty; became poor;

remained non-poor) for in-depth enquiries into the reasons associated with households‘

welfare trajectories at the community focus group discussions. In addition, for these households selected, their respective stages of progress for 2000 and 2004 were ascertained.

Follow up with household-level interviews to verify and go deeper into reasons for change The reasons indicated by the community group above were cross-checked separately through individual household interviews. This was done for a subset of the 35% to verify and go deeper into the reasons for change or stability, to triangulate and verify the group responses, but also it is possible that there are factors that were unknown outside the particular household. Approximately 45% of the households selected for ascertaining reasons for change at community level were followed up for in-depth household interviews.

We linked with the panel households at this stage, where the subset included all the panel

households in every selected community, and additional households were randomly selected in order to take care of households that may have dropped out from the panel.14

The household survey collected information on the chronology of events between 1996 and 2009 regarding livelihood strategies, positive events and negative shocks that had an impact on household well-being, particularly in terms of making them poorer or wealthier. In addition, the household level interviews constituted the final wave of the panel data and provided the link between income and SOP welfare measures. Subsequent discussions and analysis in this paper, comparing SOP and income-based welfare measures are based on the same set of individual households.