Wednesday, February 10, 2010

Nowcasting: recession ended in July 2009

There is always considerable interest in learning about current economic conditions, just see how financial news concentrates on the impact of the latest data releases on where the economy stands now. Actual data on GDP, for example, is released with a four month lag, and is subject to revisions. One could thus try to forecast current GDP in real time, nowcasting.

Boragan Aruoba and Francis Diebold assess the current state of nowcasting. There is rather little data available for nowcasting, as one needs data that is of higher frequency and quickly available. One can then work using dynamic factor analysis, which is about finding some commonness among them and interpreting this as current macroeconomic conditions. While this is now an established art and a real-time index is now maintained by the Federal Reserve Bank of Philadelphia, the Aruoba and Diebold paper is mainly about adding a nominal factor to the analysis that summarizes information about inflation.

These factors seem to tract quite well in real time historical data. This allows to draw some conclusions already about the current recession. First, it is unusually deep and long, which should not be a surprise too many. Second, it seems to have ended in July 2009. However, one could not exclude that a double-dip recession is in the works. Finally, there is still of increase in inflation visible.

I wonder whether this paper will be frequently revised to adjust for progress in the literature, something like nowreviewing nowcasting.

1 comment:

Anonymous said...
This comment has been removed by a blog administrator.