full post-1966 period and is robust to various specifications. This result is obtained by
combining a VAR specification with a multivariate GARCH model.
The present study aims to contribute to this area of the literature by estimating a time-
varying AR-GARCH model of inflation producing measures of inflation uncertainty
for the euro area, and investigating the linkages between them in a VAR framework,
also allowing for the possible impact of the policy regime change associated with the
start of EMU in 1999. Among recent studies, O’Really and Whelan (2005) focus on
inflation persistence and find relatively little instability in the parameters of the euro-
area inflation process. Full-sample estimates of the persistence parameter are
generally close to 1, and the hypothesis that this parameter has been stable over time
cannot be rejected. Angeloni et al. (2006), using micro data on consumer prices and
sectoral inflation rates from six euro-area countries spanning several years before and
after the introduction of the euro, find no evidence of a shift around 1999. Finally,
Altissimo et al. (2006) note that, for aggregate data, the degree of inflation persistence
in the euro area appears to be very high for sample periods spanning multiple decades
but falls dramatically once time variation in the mean level of inflation is allowed for;
furthermore, the timing of the breaks generally coincides with observed shifts in the
monetary policy regime.
The choice of focusing in the present study on the Harmonised Index of Consumer
Prices (HICP) for the euro area as a whole reflects the ECB’s mandate to achieve
aggregate price stability in the Eurozone, the absence of instruments to fine-tune
monetary policy to cyclical fluctuations in individual EMU countries, and the
recognition that inflation differentials play a secondary role when calibrating the
safety margin for admissible inflation in the euro area (ECB, 2003). The availability
of reliable and easy-to-update measures of inflation uncertainty is particularly relevant
for monetary policy purposes (Goodhart, 1999, and Greenspan, 2003). As Soderstrom
(2002) notes, when there is uncertainty about the persistence of inflation, it is optimal
for the central bank to respond more aggressively to shocks than when the parameters
are known with certainty, in order to avoid undesirable outcomes in the future.
According to Shuetrim and Thomson (2003), for certain shocks, taking into account
parameter uncertainty can imply that a more, rather than less, activist use of the policy
instrument is appropriate. While this finding is specific to their model specification,
parameter estimates and shocks analysed, it contrasts with the widely held belief that
the general implication of parameter uncertainty is a more conservative policy.
Finally, Coenen (2007) argues that a cautious monetary policy-maker is well-advised
to design and implement interest-rate policies under the assumption that inflation
persistence is high when there is considerable uncertainty about its degree. Such
policies are characterised by a relatively aggressive response to inflation
developments and exhibit a substantial degree of inertia.
Surprisingly, studies on the relationship between inflation and inflation uncertainty
for the euro area as a whole are distinctly rare. Most of them are either based on
survey data (see, e.g., Giordani and Soderlind, 2003, and Arnold and Lemmen, 2008)
or adopt a country-by-country approach (see Fountas et al. 2004, and Apergis, 2004).
Our econometric framework is similar to that originally proposed by Evans (1991)
and recently used by Berument et al. (2005) and Caporale and Kontonikas (2009).
The latter authors analyse inflation uncertainty and its relation with inflation at
country level in twelve EMU member states, and report that there is considerable