Editorial Introduction to Special Issue on Large Data Sets
| Author | Anindya Banerjee |
| Published date | 01 February 2013 |
| Date | 01 February 2013 |
| DOI | http://doi.org/10.1111/obes.12016 |
1
©Blackwell Publishing Ltd and the Department of Economics, University of Oxford 2012. Published by Blackwell Publishing Ltd,
9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 75, 1 (2013) 0305-9049
doi: 10.1111/obes.12016
Editorial Introduction to Special Issue on Large Data
Sets
Anindya Banerjee
Department of Economics, Birmingham Business School, Birmingham B15 2TT, UK
( e-mail: a.banerjee@bham.ac.uk) and Banque de France, 75001 Paris, France
( e-mail: anindya.banerjee@banque-france.fr)
This Special Issue on Large Data Sets arose out of a call for papers which invited contribu-
tions in this broad area, both in micro-econometrics and macro-econometrics where large
data sets and associated techniques are used. Examples suggested in the call included mod-
els with factors, factor augmented vector autoregressions, structural vector autoregressions
with large data sets and panel data models.
After several rounds of refereeing, the final selection of eight papers reflects the rubric
broadly, although it is somewhat more narrowly focused than the original call. The empha-
sis here is upon large-sample macroeconometrics, while two of the papers use panel tech-
niques. One of these two papers is a data-intensive study of discrimination in football in
the English Premier League, which uses a sample of over a million ‘in-match events’,
drawn from 760 Premier League football matches over two seasons (2006–7 and 2007–8).
Several of the papers make important theoretical contributions to the literature, and all are
illustrated by empirical examples and data sets drawn from empirically relevant sources.
Six of the papers deal explicitly with factor models which have become an influential
tool in estimation, policy analysis and forecasting. The use of factor models, as ways of
incorporating and summarizing information from large data sets, has gained prominence
in recent years with the availability and gathering of increasingly large quantities of data
by governments, central banks and private organizations to capture all aspects of an evolv-
ing economy. The early influential papers include those by Bai (2003, 2004), Bai and
Ng (2002, 2004), Forni et al. (2000, 2004, 2005) and Stock and Watson (2002a,b), which
established the theoretical and inferential basis for large-sample analysis using factor mod-
els and provided the early empirical applications of the method. The literature has since
grown to encompass important contributions on incorporating factor models into so-called
structural vector autoregressions, prominently represented by Bernanke, Boivin and Eliasz
(2005) which introduced factor augmented vector autoregressions (FAVAR) and their use
in analysing the responses of macroeconomic variables to policy shocks. Markov Chain
Monte Carlo (MCMC) techniques have been employed to capture continuous and discrete
time variation in these models. Two excellent recent summaries of the entire field, which
continues to evolve rapidly, are by Stock and Watson (2006, 2011).
The motivation underlying factor models is that a few unobserved dynamic factors ft
may be taken to drive the co-movement of a high-dimensional vector of time-series vari-
ables Xt, where the dimension Nof the vector Xtwill typically exceed N. These factors
To continue reading
Request your trial