24 November 2017

ECB: Policy analysis with big data

Benoît Cœuré, Member of the Executive Board of the ECB, told an audience that micro data collected by central banks themselves, in particular data on transactions between banks, have already proven to be an additional valuable guide for policymakers in devising policy responses.

Mr. Cœuré also shows that data generated through the greater use of technology in our daily lives have an enormous potential to help policymakers overcome prevailing constraints on timely data availability, understand better the consequences of their policies and calibrate them accordingly, while also creating challenges.

The recent financial crisis, and the euro area sovereign debt crisis that followed, were characterised by periods of increased heterogeneity, market fragmentation and sudden turns in economic activity. This often made it difficult for economic policymakers to understand and assess in real time the underlying forces driving economic behaviour. Both traditional statistical datasets and our models proved at times inadequate to support the decision-making process, reflecting long time lags, linear assumptions and the absence of more granular information.

These events increasingly boosted the efforts in policy circles to obtain timelier and richer data for policy analysis, in short big data. This push towards more granular information was not a revolution, however. It can be argued that big data, under different guises, have been used as an input into policymaking since Adolphe Quételet’s Mémoire in 1848. Since then, big data has been central to business cycle analysis, from the early work of Clément Juglar to the contributions of both Wesley Mitchell and the Cowles Commission, right up until today. According to central bank mythology, former Federal Reserve Chair Alan Greenspan would sit in his bathtub perusing sheets of statistics. And indeed, economic historians have analysed how the power of governments has been shaped by statistics – and vice versa.

The most recent push towards more granular data was thus an evolution rather than a revolution, triggered in part by the emergence of new opportunities – themselves a reflection of rapid technological progress – and the experience gained over several years of crisis management.

This evolution is already bearing fruit. Policymakers today have access to a large number of micro datasets, often very different in nature and scope. Some are the result of new financial regulations. Others are by-products of increased use of technology. What they have in common, however, is that, if used appropriately and responsibly, they can help policymakers to extract more timely and diverse economic signals, and thus are a meaningful complement to existing official data.

Central banks have made considerable progress in recent years in integrating big new datasets into their policy analysis and decision-making. Granular data collected by central banks themselves have, in particular, become an indispensable source of information for policymakers.

Mr. Cœuré  concludes: “Yet, it is probably fair to say that we are still exploring how to use much of the big data that new technology opens up to us. Although evidence is growing that online data may provide tangible benefits for short-term forecasting, more research is needed to ensure that the data are of sufficient quality and reliability to systematically inform policymaking.

“The potential of such data to enrich central bank analysis in the future is considerable, however, as are the challenges that come along with it. For this reason, I encourage all of you to continue your efforts and to work in the same collaborative spirit as you have done so far.”

Full speech

© ECB - European Central Bank