“Information is the oil of the 21st century, and analytics is the combustion engine.”
– Peter Sondergaard
Big data is a vague term for a massive phenomenon that in recent years has become the center of much debate and controversy. A massive pool of data has resulted from the widespread use of mobile technology, internet traffic, and social networks, enabling people around the world to access banking services, employment information, medical services, and markets. Although there is no agreed-upon definition, “big data” is often characterized by the three V’s— high volume, high velocity, and high variety. While the list of V’s keeps increasing (i.e., veracity and volatility), competing definitions and understandings of the concept as well as its potential usage are also proliferating. Despite the confusion, one thing is clear: big data is here to stay. Every organization needs to understand what big data means to them, and how big data analytics can be deployed to enhance their work in a timely, efficient, and effective fashion.
Big data is a game changer for this century, what oil was to the last one (The Economist). Digital information is being extracted, refined, valued, bought, and sold in new and diverse ways, and everyone wants a piece of it (Gaining an Edge with Big Data). For the field of macroeconomics, there is exciting potential for how big data can produce new indicators, bridge time lags, support the forecasting of existing data sets, and provide innovative data sources to produce official statistics (IMF Staff Discussion Note). Public sector institutions and organizations have an interest in using big data and modern technologies to inform policy making. Here at the IMF, some of the applications of big data concern the assessment of competitiveness in the tourism sector through the A Week at the Beach Index, monitoring global financial flows and correspondent banking relationships through SWIFT data, and evaluating firm behavior based on Orbis data. Central Banks are using big data as an input for forecasting and nowcasting tools to support macroeconomic and financial stability assessments, assess the impact of policy communication and expectations for policy decisions through text mining, and collect information on, among other things, prices, fiscal indicators, and granular credit data. To harness the power of big data for policymaking, a holistic understanding of the opportunities must be accompanied by a thorough evaluation of the challenges and limitations that come with it.
Big data comes with big challenges. For policymaking applications, the quality assessment of indicators derived from big data is paramount, despite the strong demand for timely and granular data. “Big data poses a considerable legal challenge and requires specialized training that goes well beyond established econometric and statistical methods. The real challenge rests in assuring that the quality of the results is rigorous and credible so it can inform sound policy insights,” says Mamoon Saeed, a member of the IMF’s Information Technology Department. Currently, the IMF is using big data to uncover important real-time trends and insights as opposed to causal inference. “A systematic use of big data in policy analysis requires rethinking the institutional governance of information technology, and revamping long-standing practices in acquiring, disseminating and analyzing information,” adds Marco Marini, who is part of the IMF’s Statistics Department. Such changes will include new legal agreements, adapting cloud storage and related big data platforms, and acquiring an expertise in data science and machine learning techniques. The necessary skills will be acquired through a combination of training existing employees and hiring those with new skills. In short, a big data practice for policy analysis and economic surveillance in the long term will tip the skill balance and change the future of work everywhere.
Amid exciting prospects and significant challenges, the question arises: Can international organizations and public institutions ride the big data wave? They can, but not if they go it alone or too late. Big data is a dynamic phenomenon, the systems and networks generating it are ever evolving, and related challenges, limitations, and opportunities are ever changing. Saeed reiterates that “the world around us is changing at a faster pace. Industries are pioneering innovative ways of conducting business and shaping markets. Are we confident that our methods and indicators can cope with and capture these changes?” Organizations like the IMF recognize the need to go beyond individual and scattered applications of big data, build public-private partnerships to deliver measurable, scalable, and high-quality results, and facilitate peer learning across their membership. “Establishing sound partnerships, resolving legal issues, and acquiring the right skills and technologies are as important as statistical expertise, data representativeness, and methodological accuracy in harnessing the power of big data for better policymaking,” conclude Cornelia Hammer and Diane Kostroch, both of whom are in the IMF’s Statistics Department. The key to success lies in putting together a dynamic environment of people and processes that can take big data innovations forward and put them to work in a timely fashion without falling prey to bureaucratic inertia. ▪