| By: |
Marco J. Lombardi (European Central Bank, Kaiserstrasse 29, D-60311 Frankfurt, Germany.);
Philipp Maier (Bank of Canada, International Department, 234 Wellington, Ottawa, ON, K1A 0G9, Canada.) |
| Abstract: |
We evaluate forecasts for the euro area in data-rich and ‘data-lean’
environments by comparing three different approaches: a simple PMI model based
on Purchasing Managers’ Indices (PMIs), a dynamic factor model with euro area
data, and a dynamic factor model with data from the euro plus data from
national economies (pseudo-real time data). We estimate backcasts, nowcasts
and forecasts for GDP, components of GDP, and GDP of all individual euro area
members, and examine forecasts for periods of low and high economic volatility
(more specifically, we consider 2002-2007, which falls into the ‘Great
Moderation’, and the ‘Great Recession’ 2008-2009). We find that all models
consistently beat naive AR benchmarks, and overall, the dynamic factor model
tends to outperform the PMI model (at times by a wide margin). However,
accuracy of the dynamic factor model can be uneven (forecasts for some
countries have large errors), with the PMI model dominating clearly for some
countries or over some horizons. This is particularly pronounced over the
Great Recession, where the dynamic factor model dominates the PMI model for
backcasts, but has considerable difficulties beating the PMI model for
nowcasts. This suggests that survey-based measures can have considerable
advantages in responding to changes during very volatile periods, whereas
factor models tend to be more sluggish to adjust. JEL Classification: C50,
C53, E37, E47. |