Big Data in Economics
The availability of massive datasets on economic choices of individuals and firms, from administrative and corporate sources, changes empirical research in the social sciences fundamentally.
Economists and statisticians develop new methods that provide new answers to old-standing empirical questions and allow for much faster and granular research. At the same time, new data and methods also change the way administrations and firms work, which creates new and challenging questions for public policy.
The keynote lectures shed light on various aspects of using big data in economics research. Raj Chetty (Harvard University) explains how administrative data such as tax filings can be used to understand the heterogeneous effects of macroeconomic shocks and public policies on individuals and to construct real-time diagnostics of the economy. Raffaella Sadun (Harvard Business School) focuses on how managers use the large amounts of data their firms generate in such areas as human resource management. Michael Lechner (University of St. Gallen) explores research opportunities that arise from combining machine-learning methods with large administrative datasets to identify the causal effects of public policies. A panel discussion will review opportunities and challenges of using big data in economic research in Germany, Austria, and Switzerland.
Monday 12 Sept. / 11:30 Michael Lechner „Causal machine learning“
In recent years microeconometrics experienced the ‘credibility revolution’ culminating in the 2021 Nobel prices for David Card, Josh Angrist and Guido Imbens. These developments, hopefully, lead to more reliable estimation of causal effects of certain public policies. At same time, computer science, and to some extent also statistics, developed powerful algorithms (Machine Learning) that are very successful in prediction tasks. The new literature on Causal Machine Learning attempts to unit these two developments, i.e., use Machine Learning to improve causal analysis. In this talk, I review some of these developments. Subsequently, I use an empirical example from the field of active labour market evaluation to show how these methods can be fruitfully applied to improve the usefulness of empirical studies w.r.t. to evaluating and improving policies. I conclude with some considerations about current shortcomings and possible future developments of these methods.
Monday 12 Sept. / 14:15 Raj Chetty "Using Big Data for real-time diagnostics in the economy“
Tuesday 13 Sept. / 14:15 Raffella Sadun "Using Big Data to analyze Management Practices“
Panel of the Core Conference
Wednesday 14 Sept. / 12:15-13:15
Datenzugang für die Forschung
- Moderation: Gert G. Wagner (DIW Berlin)
- Kerstin Schneider (Bergische Universität Wuppertal, Vorsitzende RatSWD)
- Tobias Thomas (Generaldirektor Statistik Austria)
- Kurt Schmidheiny (Universität Basel)
Moderation: Gert G. Wagner (Max-Planck-Institut für Bildungsforschung)