Page 302 - Special Topic Session (STS) - Volume 3

STS544 M. Camachoa et al.

Nowcasting and forecasting with Dynamic

Factors Models: Some experiences and lessons
2
1
M. Camachoa , R. Doménechb
1 Universidad de Murcia, Spain
2 BBVA Research and Universidad de Valencia, Spain

Abstract
Dynamic Factor Models (DFM) have become a useful econometric
methodology for GDP nowcasting and forecasting. We review the experience
in the use of these models at BBVA Research for a large sample of advanced
and emerging countries. Particularly when financial variables are included,
DFM forecast GDP growth at least as well as other alternative methodologies,
in a very parsimonious ways, allowing to present the contribution of each
variable to the innovation of forecasts in a very informative way. We also show
that DFM can be used to nowcast GDP growth when the estimates provided
by national statistical institutes are not useful indicators of the underlying
activity. Finally, DFM can be adapted to include valuable indicators of
economic activity at different time frequencies obtained using real-time big
data information such as retail sales or credit cards spending, improving
significantly the nowcasting performance of our models.

Keywords
Dynamic factor models, GDP, nowcasting, forecasting, financial variables, big
data. JEL Classification: E32, C22, E27.

1. Introduction
During the Great Recession in 2008 and 2009, policy makers and other
economic agents seemed eager to detect signals of its intensity and length,
and those of the subsequent the recovery. Some years later, the same type of
concerns attracted attention to the crisis of several emerging countries. And
more recently, in one of the most longest lasting expansions since the mid-
19th century in the US economy, there is a huge interest to anticipate the
signals of a potential downturn in future quarters. As for other peers in the
world baking industry, the anticipation of economic conditions is one of the
most important challenges for BBVA to position competitively and
strategically in its footprint.

 M. Camacho thanks the financial support of grants MINECO ECO2016-76178-P and Seneca
Excellence Groups 19884/GERM/15. R. Doménech thanks MINECO CICYT ECO2017-84632
and Generalitat Valen-ciana PROMETEO2016-097 for financial support. Contact:
r.domenech@bbva.com.

291 | I S I W S C 2 0 1 9