Fulcrum’s Nowcasting Models for Economic Activity:  A Primer

 

Gauging the state of economic activity in real time is an important priority for investors and policy makers.  This task is complicated by the fact that the broadest and most well-known indicator of economic activity, real Gross Domestic Product, is generally published on a quarterly basis, with about a month delay.  Moreover, initial estimates of GDP are generally revised over time as additional data become available.  On the other hand, many indicators related to economic activity (such as industrial production, employment, and surveys of consumer and business confidence) are published at higher frequency (monthly, weekly) which can be used to produce real time tracking estimates (known as “nowcasts”) of economic activity that are updated promptly as new information arrives.

Making sense of this often delayed and contradictory flow of news coming from economic activity requires a disciplined and data-driven approach.  Following an extensive academic literature, to which our own researchers have made important contributions (Antolin-Diaz, Drechsel and Petrella, 2017), we employ a state-of-the-art statistical tool known as Dynamic Factor Model (DFM).  Factor models employ a large number of economic activity indicators, including many like employment or confidence surveys, which are not part of the official GDP calculation.  From this large dataset they extract “common factors” which capture the bulk of fluctuations in the data.  Typically, one factor is extracted, which is interpreted as an index of underlying economic activity, and projected on GDP to produce a nowcast.  Evans (2005) and Giannone, Reichlin, and Small (2008) pioneered the use of Dynamic Factor Models for nowcasting. Fulcrum’s global nowcasts, as well as the New York Fed US Nowcast, employ dynamic factor models.

 

 

The Fulcrum Models allow us to track, on a daily basis, our best estimate of economic activity across the main economies using all available information released up to each point in time.  Whenever new information is released, the model will automatically update, and our estimate will “jump”, reflecting the news contained in the data.  Additionally, Fulcrum’s models provide us with an estimate of the longer-run, normal rate of economic activity, to which we would expect economic activity to return to after two to three years.

Fulcrum’s nowcasting models are an essential part of our quantitative toolkit.  They provide our discretionary strategies with a quantitative, disciplined assessment of the state of the economy, free from the prejudices, behavioural biases and anchoring typically found in human forecasters, constantly updating as new information arrives.

 

Read our most recent Macro Commentaries by Gavyn Davies

 

 


References
Antolin-Diaz, J., Drechsel, T., & Petrella, I. (2017). Tracking the slowdown in long-run GDP growth. Review of Economics and Statistics, 99(2), 343-356.
Evans, M. D. D. (2005): “Where Are We Now? Real-Time Estimates of the Macroeconomy,” International Journal of Central Banking, 1.
Faust, J. and J. H. Wright (2009): “Comparing Greenbook and Reduced Form Forecasts Using a Large Realtime Dataset,” Journal of Business & Economic Statistics, 27, 468-479.
Giannone, D., L. Reichlin, and D. Small (2008): Nowcasting: The real-time informational content of macroeconomic data,” Journal of Monetary Economics, 55, 665-676.