Data quality in Africa is near-universally poor, especially at a sub-national level. Organisations and individuals who care about development, poverty alleviation and social welfare should fund measures and programmes that improve data quality via technical assistance. With good data, ‘mysteries’ of African (under)development can be better addressed, and more people can be lifted out of poverty across the continent.
Introduction
There’s a general consensus that data quality in Africa is very, very bad — regardless of the source. The IMF, World Bank and the UN all report to have an abundance of socio-economic and demographic data on every corner of the world, including Africa. But there is plenty of reason to believe that Africa has a data quality problem that even the major global institutions can’t get around and that most data is just made up. The implications are real, both for how we understand what’s happening on the ground and for what evidence-based interventions actually make sense. For someone like me, running armchair regressions to answer questions I’m curious about, the stakes of using unreliable data are relatively low. For the people making far higher-stakes decisions on the back of the same numbers, they’re not.
The Importance
Consider decisions made by bureaucrats working in a foreign aid department for some major government. They might use (unreliable) data from an international organisation that reports infant mortality, PPP per capita income, and maternal mortality to decide how much money they should give to NGOs working on these respective matters. If the reported data is wrong, in any of these areas, we have an obvious problem. It’s also a complication for those working within the country, such as domestic civil servants. If you don’t know how many people live somewhere, let alone what they earn, there is a hard ceiling on what can be understood and on the scale of problems you can even begin to measure. In short, without reliable data, we are shooting in the dark.
There are real questions about what “funding data quality” actually looks like in practice. It might mean placing technically skilled personnel inside national statistics bureaus where the problem is worst. This could look something like the ODI fellowship specifically for data quality. It might mean regular on-the-ground data collection, or supporting census rollouts where they happen. The latter is especially important when considering that, in West Africa alone, Benin and the Gambia haven’t conducted a census since 2013 and Nigeria’s last census was recorded in 2006. This means that when such rollouts do eventually happen, a significant knowledge gap is likely to further compromise the reliability of the numbers. Nonetheless, ideally measures to fund data quality consists of all of the above. The point about its importance holds true regardless.
The title for this post is inspired by: Forecasting is Way Overrated, and We Should Stop Funding It — LessWrong
Summary
Data quality in Africa is near-universally poor, especially at a sub-national level. Organisations and individuals who care about development, poverty alleviation and social welfare should fund measures and programmes that improve data quality via technical assistance. With good data, ‘mysteries’ of African (under)development can be better addressed, and more people can be lifted out of poverty across the continent.
Introduction
There’s a general consensus that data quality in Africa is very, very bad — regardless of the source. The IMF, World Bank and the UN all report to have an abundance of socio-economic and demographic data on every corner of the world, including Africa. But there is plenty of reason to believe that Africa has a data quality problem that even the major global institutions can’t get around and that most data is just made up. The implications are real, both for how we understand what’s happening on the ground and for what evidence-based interventions actually make sense. For someone like me, running armchair regressions to answer questions I’m curious about, the stakes of using unreliable data are relatively low. For the people making far higher-stakes decisions on the back of the same numbers, they’re not.
The Importance
Consider decisions made by bureaucrats working in a foreign aid department for some major government. They might use (unreliable) data from an international organisation that reports infant mortality, PPP per capita income, and maternal mortality to decide how much money they should give to NGOs working on these respective matters. If the reported data is wrong, in any of these areas, we have an obvious problem. It’s also a complication for those working within the country, such as domestic civil servants. If you don’t know how many people live somewhere, let alone what they earn, there is a hard ceiling on what can be understood and on the scale of problems you can even begin to measure. In short, without reliable data, we are shooting in the dark.
There are real questions about what “funding data quality” actually looks like in practice. It might mean placing technically skilled personnel inside national statistics bureaus where the problem is worst. This could look something like the ODI fellowship specifically for data quality. It might mean regular on-the-ground data collection, or supporting census rollouts where they happen. The latter is especially important when considering that, in West Africa alone, Benin and the Gambia haven’t conducted a census since 2013 and Nigeria’s last census was recorded in 2006. This means that when such rollouts do eventually happen, a significant knowledge gap is likely to further compromise the reliability of the numbers. Nonetheless, ideally measures to fund data quality consists of all of the above. The point about its importance holds true regardless.
Conclusion
Good data will not resolve every challenge, but it would meaningfully improve the capacity of those trying to respond to pressing questions. Hence, those individuals and organisations that care about development should fund it. The case for better data is not limited to Africa. If current political trends in the rich world continue, it may become an equally urgent problem there too, albeit for different reasons.