902 Macroeconometric Modeling for Policy
index adjusted for electricity and taxes) index would predict a 2006 inflation rate
of 0.3%, which is a more accurate forecast than any of the econometric or profes-
sional forecasts that were produced in 2005. The DDD forecast is not improved on
by Normetrics before January 2006, nor by Norges Bank before theMonetary Policy
ReportJanuary 2006 (from March). However, by this time the DDD forecast could
also be have been updated, most simply by replacing 0.3% by the actual rate of
inflation in 2005, which was 1.0%. This forecast is practically identical to Normet-
rics forecasts from before September 2006, and it is not beaten by Norges Bank’s
forecast before theMonetary Policy ReportFebruary 2006.
2006 is not the only year for which the Normetrics forecasts compete well with
the central bank’s forecasts for the monetary policy target variable. Juel, Molnar and
Røed (2008) note that, for a relatively long period of time, the automated forecasts
from an empirically validated inflation model, AIF above, have been better than
Norges Bank’s forecasts. Figure 17.10 illustrates the point.
The figure can be used to assess theex anteforecasts based on forecast errors for
the period 2004(2)–2007(3). The forecasted variable is the annual rate of CPI-AET
inflation. The graphs in the upper panel in the figure show the MFEs. A negative
MFE means that the inflation forecasts are on average higher than the actual infla-
tion rates in the period. The biases of the econometric model (AIF) forecasts are
Inflation reportsInflation reports01–1.0–0.50.02345
Length of forecast horizon (quarters)67891011AIFAIF
010.5MSFEMFE1.02345
Length of forecast horizon (quarters)67891011Figure 17.10 Forecasts from an empirically validated inflation model compared to inflation
report forecasts
First panel: mean forecast errors (MFEs) for core inflation (annual rate of change in CPI-AET). Inflation
report forecasts and AIF forecasts.Second panel: mean squared forecast errors (MSFEs).
Source:Inflation Report/Monetary Policy ReportFebruary 2004 to February 2007 and AIFs published at
http://folk.uio.no/rnymoen/forecast_ air_index.html.