10th ECB Workshop on Forecasting Techniques: Economic Forecasting with Large Datasets

The European Central Bank (ECB) is holding its tenth Workshop on Forecasting Techniques in Frankfurt am Main on 18 and 19 June 2018.

This biennial conference provides a forum for new theoretical and applied work on forecasting. Macroeconomists today have more – and richer – data at their disposal than ever before. Many of these datasets are not only very large, but also unstructured, not rectangular, or irregularly spaced. Fully tapping the embedded information for forecasters and policymakers requires new methods and tools.

This conference will bring together experts from all fields to exchange new ideas on utilising large datasets for macroeconomic and financial forecasting and to put new insights from econometric and statistical theory into practice in the current macroeconomic environment.

Scientific committee

Barbara Rossi (ICREA – Universitat Pompeu Fabra, Barcelona GSE, CREI) and Carlos Montes-Galdon, Georg Strasser and Srečko Zimic (all ECB).

Attendance is upon invitation only.


Monday, 18 June 2018

* indicates the presenter

Registration and coffee

Welcome address

Matteo Ciccarelli, European Central Bank


Part I

Chair: Barbara Rossi, Universitat Pompeu Fabra

Keynote speech

Learning to analyze economic data using machine learning methods

Serena Ng*, Columbia University


Common factors, trends, and cycles in large datasets

Matteo Luciani*, Board of Governors of the Federal Reserve System
with Matteo Barigozzi, London School of Economics slides

Discussant: Siem Jan Koopman, Vrije Universiteit Amsterdam and Tinbergen Institute

Coffee break

Multivariate Bayesian predictive synthesis in macroeconomic forecasting

Knut A. Aastveit*, Norges Bank and BI Norwegian Business School
with Kenichiro McAlinn, University of Chicago, Booth School of Business;
Jouchi Nakajima, Bank for International Settlements;
and Mike West, Duke University slides

Discussant: Xuguang S. Sheng, American University


On the evolution of the United Kingdom price distributions

Kim Huynh*, Bank of Canada
with Ba M. Chu, Carleton University;
David T. Jacho-Chávez, Emory University;
and Oleksiy Kryvtsov, Bank of Canada slides

Discussant: Fabrizio Venditti, Banca d'Italia


Lunch and poster session

Poster Session

Poster 1: A new approach for detecting shifts in forecast accuracy

Jeremy Chiu*, Bank of England
with Simon Hayes, Bank of England;
George Kapetanios, Kings College London;
and Konstantinos Theodoridis, Cardiff University

Poster 2: Nonlinear dynamic factor models

Pablo Guerrón-Quintana*, Boston College
with Alexey Khazanov, Boston College;
and Molin Zhong, Board of Governors of the Federal Reserve System poster

Poster 3: Nonlinear dynamic factor models with interacting level and volatility

Siem Jan Koopman*, Vrije Universiteit Amsterdam and Tinbergen Institute
with Geert Mesters, Universitat Pompeu Fabra and Barcelona Graduate School of Economics;
and Bernd Schwaab, European Central Bank

Poster 4: Forecasting with many predictors using message passing algorithms

Dimitris Korobilis*, University of Essex poster

Poster 5: The global component of inflation volatility

Massimiliano Marcellino*, Università Bocconi
with Andrea Carriero, Queen Mary University of London;
and Francesco Corsello, Bank of Italy and Università Bocconi

Poster 6: A new approach to nowcasting with mixed-frequency Bayesian VARs

Francesca Monti*, Bank of England and Centre for Macroeconomics
with Domenico Giannone, Federal Reserve Bank of New York and Centre for Economic Policy Research;
and Andrej Sokol, Bank of England and Centre for Macroeconomics

Poster 7: Expectation formation following large unexpected shocks

Xuguang S. Sheng*, American University
with Scott R. Baker, Northwestern University;
and Tucker S. McElroy, U.S. Census Bureau

Poster 8: Adaptive discrete smoothing with an application to (high-dimensional) nonlinear panel data

Martin Spindler*, Universität Hamburg
with Xi Chen, New York University;
Victor Chernozhukov, Massachusetts Institute of Technology;
and Ye Luo, University of Florida

Poster 9: Adaptive state space models with applications to the business cycle and financial stress

Fabrizio Venditti*, Banca d'Italia
with Davide D. Monache, Banca d'Italia;
and Ivan Petrella, Warwick Business School and Centre for Economic Policy Research poster


Part 2

Chair: Michael Ehrmann, European Central Bank

Keynote speech
Invariance and causality for robust predictions

Peter Bühlmann*, Eidgenössische Technische Hochschule Zürich

Coffee break

Forecasting with a panel Tobit model

Frank Schorfheide*, University of Pennsylvania
with Laura Liu, Board of Governors of the Federal Reserve System;
and Hyungsik R. Moon, University of Southern California and Yonsei University slides

Discussant: Martin Spindler, Universität Hamburg


A multiple testing approach to the regularisation of large sample correlation matrices

Natalia Bailey*, Monash University
with Hashem Pesaran, University of Southern California;
and L. Vanessa Smith, University of York slides

Discussant: Massimiliano Marcellino, Università Bocconi




Panel discussion: (Macroeconomic) Forecasting with big data: What works? What doesn't? What's next?

Moderator: Giorgio Primiceri, Northwestern University

Peter Bühlmann, Eidgenössische Technische Hochschule Zürich
Francis X. Diebold, University of Pennsylvania
Serena Ng, Columbia University
Hal Varian, Google Inc.

Dinner (by invitation only)

Tuesday, 19 June 2018

Registration and coffee

Part 3

Chair: Isabel Vansteenkiste, European Central Bank

Keynote speech
Out of controls

Hal Varian*, Google Inc.


Macroeconomic nowcasting with big data through the lens of a sparse factor model

Laurent Ferrara*, Banque de France
with Anna Simoni, Centre de Recherche en Economie et en Statistiques and Centre National de la Recherche Scientifique slides

Discussant: Francesca Monti, Bank of England and Centre for Macroeconomics

Coffee break

Does modeling a structural break improve forecast accuracy?

Andreas Pick*, Erasmus Universiteit Rotterdam, Tinbergen Institute, De Nederlandsche Bank and CESifo Institute
with Tom Boot, University of Groningen slides

Discussant: Jeremy Chiu, Bank of England


Part 4

Chair: Geoff Kenny, European Central Bank

Keynote speech
Big data tools and small data surveys

Francis X. Diebold*, University of Pennsylvania


Economic predictions with big data: the illusion of sparsity

Giorgio Primiceri*, Northwestern University
with Domenico Giannone, Federal Reserve Bank of New York and Centre for Economic Policy Research;
and Michele Lenza, European Central Bank slides

Discussant: Pablo Guerrón-Quintana, Boston College

Coffee break

Large-scale dynamic predictive regressions

Daniele Bianchi*, University of Warwick
with Kenichiro McAlinn, University of Chicago, Booth School of Business slides

Discussant: Mike West, Duke University


Predictive density combinations with dynamic learning for large data sets in economics and finance

Herman van Dijk*, Erasmus Universiteit Rotterdam, Norges Bank and Tinbergen Institute with
Roberto Casarin, Università Ca' Foscari Venezia;
Stefano Grassi, Università degli Studi di Roma "Tor Vergata";
and Francesco Ravazzolo, Free University of Bozen-Bolzano and Norges Bank slides

Discussant: Dimitris Korobilis, University of Essex


Concluding remarks

Geoff Kenny, European Central Bank

End of workshop

General information

Conference venue

European Central Bank
Main building – Press centre, room C5.01
Sonnemannstrasse 20
60314 Frankfurt am Main

+49 69 1344 0

Conference language


Conference organisers

Marta Bańbura, European Central Bank

Carlos Montes-Galdón, European Central Bank

Barbara Rossi, Universitat Pompeu Fabra

Georg Strasser, European Central Bank

Srečko Zimic, European Central Bank


Participants are requested to arrange their own transfers from and to the airport, unless indicated otherwise.

Lunch and poster session venues

Press room foyer

Time frame for presenters

For each paper, the author will speak for 25 minutes and the discussant for 15 minutes. This will be followed by a general discussion lasting five minutes.


Ms Iris Bettenhäuser
Directorate General Research
Monetary Policy Research Division

Please note that this programme may be subject to change without notice.