Written by Hetan Shah, Executive Director of the Royal Statistical Society. Follow him on twitter @HetanShah
This is the latest post in our blog series on ‘What kind of ‘data revolution’ do we need for post-2015?’
Everybody wants a data revolution. Like every fashion in development, people jump onto the bandwagon and try to entice funders with their new shiny innovation. Big data, open data, multidimensional indicators, new global statistical institutions, we are awash with suggestions. Where can we best place any investment that is made in data?
My answer is an unfashionable one: invest in national statistical systems. We need to build national capability in official statistics as a core part of development. This sounds pretty boring compared to the other kinds of ‘data revolution’ propositions that are floating around, but there are good arguments for it.
Why national statistics?
For a country to develop, it needs quality information about its own society, economy and environment. This implies national statistics. As the Development Initiatives publication ‘Investments to End Poverty’ tells us, currently the quality of data about most poor countries is weak. There tend to be disconnections between what household surveys and national accounts tell us. The data are often not available in time to be of much use, and, in some cases, they are not trustworthy. All of these basic issues need addressing.
Looking beyond the basic measurement issues, there are a great many development challenges which require new ways to think about statistics. Can technology be used to provide data in a more timely and disaggregated way? How do we develop quality national accounts which take environmental assets into account? How can we better measure inequality within countries? What techniques allow us to move away from measuring inputs to measuring quality of institutions, for example, health or education? All of these will require investment in statistical systems.
Work by PARIS21 suggests that only 0.16% of ODA was spent on statistics in 2012. In comparison to the proportion of project funds usually allocated to the monitoring and evaluation of aid projects, this is a very low figure and we need to see it increase significantly.
How might such funding be spent? National statistical offices could receive more funding for staffing and for technical support and training. The training needs to be both at an elite level but also underpinning what has often been ‘learned on the job’. Countries need resources to have a conversation between users and producers of statistics to consider which measures need prioritising in their current context. Where appropriate, the creation of civil society national statistical associations could also be encouraged.
One practical opportunity will come through the collection of the data for whatever comes after the MDGs. Let’s not farm this out to international consultants or multilateral institutions, but use this as an opportunity to build national capacity in the collection of the data.
What about big data and open data?
There is a lot of hype around ‘big data’. This tends to mean using large unstructured datasets which are by-products of other processes – particularly mobile phones and internet based activity. Clearly this is going to be a growing area, but this kind of data is only helpful when it can be benchmarked with well functioning official statistics.
Similarly, open data is an effective agenda only when there are underlying data sets of reasonable quality to open up. The kind of accountability and transparency that open data has the potential to provide requires poor countries to collect that data at a national level.
We currently lack an evaluation of the impact of the MDGs on the statistical capability of poorer countries. There are those who argue that improvements are patchy and potentially unsustainable, and indeed that those statistics which were not highlighted in MDG goals such as national accounts, trade or price statistics have suffered as resources have switched into other areas. Such an evaluation might help us think through what changes are needed in the international architecture to provide more effective support to national statistical systems, especially in low-income countries. This seems a good starting point to help think about any investment decisions that might be made by the international community next year.
There is a major opportunity for the data revolution. We need to make sure any investment made has a long term sustained effect, and I think this can only be through supporting national statistical systems. There is a danger if we throw the money at other shiny data initiatives, we will blow our once in a generation opportunity to really improve data for development.