This is a joint post with Alex Ezeh.
Maybe you missed it–but this past Sunday, November 18, was African Statistics Day.
It’s the 12th celebration of the day, and a good time to note that substantial gaps remain in the availability and quality of data on basic indicators of human well-being such as income, poverty and cause of death in Africa.
In spite of a decade of historic levels of international and national spending on health, a 2009 study found that 3 out of 46 countries in the WHO/AFRO region had population-level data on cause of death. Even seemingly comprehensive and definitive statistical compilations, such as the recently released Atlas of African Health Statistics, readily concede that their data is entirely reliant on weak country-level data collection and variable tabulation. Similarly, the development of national administrative information systems, in health as well as other sectors, has been intermittent and slow to improve despite national and international efforts over the years (the challenges of planning based on estimates generated from limited data were pointed out by CGD research fellow Victoria Fan in a blog earlier this year).
Fragmented Data Collection
-National Health-related Surveys
- Standard DHS: 2008/2009, 2003
- Malaria Indicator Survey: 2010
- HIV/MCH Service Provision Assessment: 2010, 2003
- Kenya Integrated Household Budget Survey: 2005/2006
- Census: 2009
-Project and Micro-data (Health Only)
- 15 IPA Projects
- 16 impact evaluations posted on 3ie
- 5 World Bank surveys since 2000 (Central Data Catalogue)
- Routine project monitoring data from PEPFAR, PMI, USAID, World Bank, etc.
-Vital Statistics (approx. 60% birth coverage)
While the results of weak national statistical systems are well-known and the subject of many aspirational declarations, we’re interested in African Statistics Day because we –CGD and African Population and Health Research Center (APHRC)–are joining forces on a working group to analyze the political economy challenges that underpin many countries’ notoriously low statistical coverage, quality and frequency.
Statistical system weaknesses stem, in part, from limitations in capacity, technical know-how and qualified human resources. Limited financial resources also have something to do with weak systems, but the explosion of data collection efforts in the region suggests this is not the main obstacle (see: Case Study Kenya). Our working group has identified a third, relatively unaddressed, obstacle to statistics development: misaligned political and institutional incentives within governments and created by donor assistance policies and practices (for example, here and here).
Examples of misaligned incentives abound. National statistics offices may collect and analyze data for a consumer price index, for example, but be barred from reporting accurate results for political reasons. Budget formulas or results-based funding systems can unintentionally create incentives to “beef up” numbers, as in systems where schools are paid per pupil enrolled and administrative information systems grossly over-report the number of students. Even when data has been collected by national statistics offices, many times with donor money, data sets are inaccessible to policymakers, researchers and civil society. Working Group member Gabriel Demombynes notes in a recent blog that in spite of technical and financial support from DfID, USAID, EU, DANIDA, the World Bank, and UNDP for the Integrated Household Budget Survey in Kenya, only a small group of researchers have access to the raw data. Further down the line but equally important are challenges related to ensuring that the data is useful for policy and decision makers as well as civil society.
We need to better understand the political economy challenges of data issues in Africa in order to develop more practical strategies to strengthen national statistical systems. The APHRC-CGD Working Group aims to better understand the relationship between the institutional arrangements governing national statistics systems and how they affect rigorous, efficient and timely production of relevant data for decision-making. The extent of national statistics offices’ autonomy and accountability is a major area for investigation (related issues are discussed in a recent blog by Francis Fukuyama). The second challenge the new group hopes to tackle is how to address the problem of limited accessibility to and use of data that is already being produced. Innovative open data systems (for example the Kenya Open Data Portal) provide a good starting point for this research.
This African Statistics Day—although we don’t think we are ready to celebrate yet—we would like to acknowledge past efforts to improve data collection, emphasize the connection between quality data and progressive African development—and finally urge donors and international institutions to continue to focus on improving data quality and access.