Can inflation expectations in business or consumer surveys improve inflation forecasts?

Working Paper N° 348

JEL classification: 
Download (PDF)

Abstract

In this paper we develop a new model that incorporates inflation expectations
and can be used for the structural analysis of inflation, as well as for forecasting.
In this latter connection, we specifically look into the usefulness of real-time survey
data for inflation projections. We contribute to the literature in two ways. First,
our model extracts the inflation trend and its cycle, which is linked to real economic
activity, by exploiting a much larger information set than typically seen in this
class of models and without the need to resort to Bayesian techniques. The reason
is that we use variables reflecting inflation expectations from consumers and firms
under the assumption that they are consistent with the expectations derived from
the model. Thus, our approach represents an alternative way to shrink the model
parameters and to restrict the future evolution of the factors. Second, the inflation
expectations that we use are derived from the qualitative questions on expected price
developments in both the consumer and the business surveys. This latter source,
in particular, is mostly neglected in the empirical literature. Our empirical results
suggest that overall, inflation expectations in surveys provide useful information for
inflation forecasts. In particular for the most recent period, models that include
survey expectations on prices tend to outperform similar models that do not, both
for Belgium and the euro area. Furthermore, we find that the business survey, i.e.
the survey replies by the price-setters themselves, contributes most to these forecast
improvements.