external shocks, such as increases in commodity prices and exchange rate movements, are
passed on sequentially to consumers. These links are generally not captured by macro-
econometric models which might be the reason why our small and partial pricing model is
still capable of reproducing relatively similar results although spill-over effects, in particular
from activity to prices, are not captured. Second, the inclusion of non-energy commodity
prices (split into food and industrial raw material commodities) in the analysis reveals that
these are rather important determinants of euro area prices but have so far been left out of the
analysis in the existing literature. Third, the distinction between goods and services prices
gives a more refined view on the determinants of these prices, which may help to better
understand their developments.
Our results suggest that a shock to the nominal effective exchange rate of the euro has a direct
effect on most HICP components, but also indirect effects via the producer price components.
In general, the size and timing of these effects is similar to what has been found in the
literature at an aggregate level when taking into account that we have used the HICP
excluding unprocessed food and energy in our estimations. Meanwhile, energy and food
commodity and industrial raw material price shocks have a direct effect on producer prices
and are passed through indirectly, via these components, to consumer prices, except for food
commodity prices which also have a direct impact on processed food consumer prices. The
estimation results are consistent with the shares of imports of each sector by its output as
measured in input-output tables. An analysis of bootstrapped confidence intervals shows that
for most simulations, the confidence bands are relatively narrow in most cases and suggest
that the pricing chain is working in the expected way.
■ ECB
Working Paper Series No 1104
November 2009