The P20 approach is flexible and adaptable. It is able to demonstrate the potential ways to address data and policies and to underscore the potential to improve data sources. A first step is to identify what data are currently available and to try to identify possible methods for disaggregating target populations by wealth. There are important differences between who might be identified as the P20 by one definition of living standards (income vs consumption vs assets vs multidimensional indexes) but these different definitions do not necessarily preclude an analysis by any method. There are trade-offs in the measurement of living standards but an analysis by different methods can highlight the different dimensions of exclusion that may impact lives.
In many databases, there are no clear methods for disaggregating averages by wealth status. In principle, datasets could be joined. This would require unique IDs that could be linked from one government dataset to another. This linking would not be public but could be a tool for analysis within the government.
As a first instance, this report sought to identify key data sources for potential analysis through a rough data landscape. This data landscape was first conducted drawing on international data sources. This includes a review of the World Bank’s Statistical Capacity Indicator which highlights data frequency and quality issues across economic, health, agricultural, demographic and education indicators. Open Data Watch’s Open Data Inventory contains similar information with some additional data types considered and insights about the openness and the types of disaggregation that are found in publications from national statistics offices. Furthermore, PARIS21’s Statistical Capacity Monitor provided additional insights about the data availability and the operations of the statistical system.
Following a review of those elements, the website of INSAE was searched for databases, reports and microdata. Additionally, a review was made of the currently available data through the International Household Survey Network (IHSN) website. EMICOV data and the DHS and MICS microdata are the key sources of available microdata. Other micro datasets which have not been included in this analysis but may be included in future analyses include the Enquete sur la transition vers la Vie Active (ETVA 2012, 2014), AGVAS 2008, AGVSAN 2013, and so forth. This approach brought to light key sources of potential data as well as challenges and gaps, which were discussed in further depth during key informant interviews and meetings with groups of stakeholders.
To produce internationally comparable data on consumption, the World Bank’s PovcalNet was evaluated. Thresholds were used to identify what poverty line includes 20% of the global population. Once these thresholds were found, PovcalNet was queried to identify the poverty headcount, the poverty gap and the average incomes of those below and above these lines. Because the poverty gap indicates the average distance below the poverty line people in poverty lie, it is a straightforward transformation to identify the average income of those below any poverty line that is selected. PovcalNet reports the average household income for the population, making it easy to calculate the average income of the rest of the population.
Disaggregating indicators by P20 classification is simple with DHS and MICS microdata. These surveys include several questions on household asset ownership, asking about ownership of refrigerators, radios, or bicycles, for instance. From this, the DHS or MICS creates a principal component analysis index of wealth and provides a report for the quintile of wealth for each household which can then be linked to each person. Identifying a threshold on the wealth index to distinguish the P20 from the rest of the population is very straightforward, allowing for easy summaries between the P20 and the rest of the population. For most comparisons between the P20 and the rest of the population, we have simply used the quintile disaggregation comparisons published by INSAE’s analysis of DHS and MICS. We have taken the bottom quintile as the P20 and have averaged the top four quintiles to calculate the status of the rest of the population.
There are challenges ensuring consistent definitions of the P20. Data from PovcalNet or EMICOV would identify a different set of people as in the bottom quintile when measuring consumption than the population that would be identified using the DHS and MICS wealth index. Additionally, the questions about household assets in DHS and MICS are changed from survey to survey meaning that people defined as being in the bottom quintile in one survey may not be in the next. The P20 approach is meant to be flexible rather than definitive. It is meant to provide an approach for thinking about populations at risk of being left behind rather than a definitive definition of who those populations are.
Exclusion and poverty are multifaceted. To highlight some basic dimensions of exclusion, we drew on the monetary dimensions of exclusion through the graph of the widening gap in consumption between the P20 and the rest. We then calculated the differences for stunting and birth registration drawing on microdata from MICS and DHS. After reviewing the SDG indicator data and recent publications from INSAE, additional data were added to the report as appropriate.
To develop maps of the P20 headcount, the geospatial data from the most recent Demographic and Health Survey (2017–2018) were used. As a household survey, the data are collected using a stratified sample from INSAE. This sampling indicates that data should be representative at the department level. Using the survey weights provided, it was possible to calculate the percentage of the population in the P20 in each department. DHS also publishes partly anonymised geospatial data on where survey clusters are located. These are randomly scattered within 5–10 kilometres of the centre of where enumerators conducted surveys with 1% of the GPS coordinates being randomised even further. The DHS microdata indicate the cluster to which each household belongs for a simple calculation of the percentage of households in each cluster that is in the P20. Because of the random sampling and the anonymisation process, the GPS coordinates and percentages for each cluster mapped should be taken as indicative rather than definitive.
In 2018, key informant interviews were conducted with the primary persons responsible for SDG reporting from the Ministry of Agriculture, Livestock and Fisheries, the Ministry of Health, the Ministry of Primary Education, the Ministry of Secondary Education and the Ministry of Higher Education, the Ministry of Youth, Culture and Artisans, and the Ministry of Water and Electricity.
In 2019, DI developed a relationship with the MdSC, who in October 2019, helped organise a trip for Development Initiatives to the Alibori and Borgou departments. MdSC is the national structure for strengthening the capacity of civil society organisations in Benin. They play a key role in the development of the civil society’s official response to the Ministry of Planning and Development’s SDG reports. The organisation has key partners in each department and holds periodic meetings for discussions around certain themes. MdSC organised meetings in Kandi and in Parakou with civil society members of the health cluster and separate meetings with the education cluster. Time and resource constraints prevented further visits to municipalities beyond Kandi and Parakou.
Alibori and Borgou departments were selected as areas of focus because they provided an opportunity to explore the potential to apply the P20 in different contexts from Cotonou. Borgou is home to the second largest city in the country but is in a different climate zone with different ethnicities and different economic considerations. Alibori is even further from the capital, is much more rural and is home to many municipalities with the highest levels of poverty. Further engagement with other municipalities and departments in the future would provide a richer picture still.
In Borgou, a departmental education official participated in part of the meetings on education. Separately, meetings were held with other civil society members, local health and education workers in public and private sector and local officials working on social affairs, civil registration and municipal planning.
Meetings with civil society in health and education with the assistance of the MdSC had between 1 and 25 participants. These meetings included a general description of who the P20 are and a desire to better understand the challenges the P20 in the district may face, thoughts about policy responses to their needs and thoughts about data. Additionally, respondents were generally asked if they felt that the P20 approach was useful. All who were asked said that the P20 approach would be useful with one vocal exception. One civil society member in Borgou said that the reality is that almost everyone is poor. Even the leaders of civil society face high levels of insecurity. It is only the elites and the ministers in the south that have confidence that their children will always be able to eat well.
Key informant interviews focused on a few key actors. As with the theme meetings, these were not highly structured and are meant to identify some key issues, the validity of the P20 approach and potential data sources. Two days were spent in meetings and key informant interviews in Kandi and two days were spent in Parakou.
In November 2019, a technical workshop was held with the participant list developed with assistance from technical and financial partners, MdSC the Ministry of Planning and Development. Technical experts were invited from the donor community, civil society, academia, ministries, INSAE and the Ministry of Planning and Development. Following a presentation on preliminary insights from DI and comments from the Ministry of Planning and Development, MdSC organised a ‘World Coffee’ style meeting asking for recommendations and challenges for the P20. There were approximately 50 participants with diverse employers and specialties. Participants were assigned to three different groups and were asked to provide feedback and recommendations as a group sequentially on education, health and cross-cutting issues. The results of their comments are in Appendix 3.
The MdSC was a valuable partner in the production of this report. They helped arrange for a broad amount of stakeholder engagement in Borgou and Alibori and designed and recorded notes for the report as seen in Appendix 3.
