Huffpost, 19 June 2017
Big data is transforming the nature of business and profits world-wide. So, can it also transform international development?
The value of the EU data market was estimated at almost EUR 60 billion in 2016. It employs roughly 6.1 million workers. This expanding market is founded on the “digital breadcrumbs” of our modern society. Data brokers and analysts and the algorithms they design are wielding growing power throughout our economies and societies – they are used by recruitment agencies, dating sites and in election campaigns. Governments are tapping into databases owned by mobile phone companies, banks or internet giants like Facebook to track public opinion, target reforms and monitor performance.
Yet, while big data from satellite imagery, social media, web scraping and call detail records, to name a few, have some advantages for policy makers, to be really useful they must meet the quality and ethical standards that apply to official statistics. The OECD’s Chief Statistician, Martine Durand, best summarises the opportunity: “There are a growing number of public, private and civil society institutions involved in the production and use of data. Bringing together these different actors will help leverage the potential of big data in what UN expert advisors sometimes call the data revolution for sustainable development. At the same time, combining multiple sources of data will require clear legal, ethical and quality standards and protocols to foster trust.”
Development and humanitarian actors, among them members of the OECD’s Development Assistance Committee (DAC), are increasingly interested in how big data can help unleash faster and higher impact when it comes to responding to outbreaks of infectious diseases such as Ebola, co-ordinating emergency aid as seen after the 2015 earthquake in Nepal, increasing financial inclusion in Uganda, and ensuring development projects leave no-one behind by providing insights on the needs and location of the most vulnerable people.
To assess the views of DAC members on big data, we conducted a survey earlier this year. The response was clear: over half of the 30 DAC members are exploring the potential of big data for addressing development challenges; 50% of DAC members feel that more and better data is a priority for achieving the Sustainable Development Goals (SDGs). All this is very welcome.
With their 169 targets and some 230 indicators, the SDGs make it clear that more and better development data are needed. They are also driving a surge in demand for data. All data sources, including big data, will need to be tapped. Yet so far most countries, including many OECD member states, have not yet started collecting data for many SDG indicators. There are serious methodological challenges to solve, including the need to figure out a way to strike a balance between the production of data for global monitoring versus data for national policy making. What is clear though is that national statistical offices in developing countries are overwhelmed with these additional data requests. Can big data in this context do the trick? The answer is yes and no.
New data sources will complement – not replace —official statistics. The United Nation’s latest Action Plan for Sustainable Development Data, adopted in March 2017, seeks to apply new technologies and new data sources in mainstream statistical activities. It sets out guidelines on the use of new and innovative data – generated outside the official statistical system – in official statistics.
This brings us to the heart of the problem: official statistics in developing countries do not get the resources they deserve. They also tend to suffer from significant data gaps and low quality. Our survey showed that many DAC members feel that it is a challenge to base development co-operation decisions and monitoring on evidence from statistics and data because such data is often old, approximate and incomplete. Indeed, in their efforts to attribute development results to aid, donors often conduct unilateral, uncoordinated data collection.
On the other hand, despite saying they believe in the value of national statistical systems, donors are not stepping up investments in them. The latest official data show that between 2013-2015, DAC members’ aid for statistics has, at best, levelled off at about USD 250 million – a mere 0.25% of official development assistance. Ten DAC members accounted for 91.6% of this aid – by size of contribution: Canada, the European Commission, the United Kingdom, Sweden, Norway, Denmark, Australia, Germany, the United States and Korea. Yet the 2016 report on the State of Development Data Fundingestimates that about USD 3 billion needs to be invested annually in order for developing countries to meet SDG data demands. Despite the momentum around the SDGs and the data revolution, donors are failing to invest and co-ordinate – probably because of the high pressure at home to show short-term development results that often leads them to collect their own data.
Providers of development assistance are right to explore ways of embracing new data sources in the fight against poverty, exclusion and climate change. To make the most of big data for development, however, they will need to step up investments in developing the capacities of people and systems in developing countries, and in linking up to a larger data ecosystem. They will also need to bridge the digital divide that keeps many people from connecting to the opportunities of the Internet for trade, education, health and many other aspects of their lives. In OECD countries, mobile broadband penetration reached 95 subscriptions per 100 inhabitants in June 2016. By comparison, just 28.3% of Africans are Internet users. Digital dividends, as coined by the World Bank, are on the horizon when the Internet is universally accessible and affordable and empowers people.
The forthcoming OECD 2017 Development Co-operation Report: Data for Developmentrecognises that data is central to achieving the SDGs. Big data will play an important part in this effort, but only a part. Capable national statistical systems are a prerequisite for making the data revolution work for sustainable development. And they will only be able to do so if support is increased in a big and quick way to produce the transformations needed for lasting results.