have the same distribution of contract lengths as the Calvo model. This is
an important contribution in itself since the two approaches have until now
appeared to be distinct and incompatible at the theoretical level even if they
are sometimes claimed to be empirically similar (see for example Kiley (2002)
for a discussion). As we shall show, a simple Taylor economy can indeed be
a good approximation to a Calvo model, but only if the two are calibrated
in a consistent manner.
We develop our approach in a DGE setting following the approach of
Ascari (2000). The issue we focus on is the way a monetary shock can
generate changes in output through time, and in particular the degree of
persistence of deviations of output from steady-state. Much recent attention
has been devoted to the ability of the staggered contract approach of Taylor
to generate enough persistence in the sense of being quantitatively able to
generate the persistence observed in the data. Two influential papers in this
are Chari, Kehoe and McGrattan (2000) and Ascari (2000). Both papers are
pessimistic for staggered contracts. CKM develop a microfounded model of
staggered price-setting and find that they do not generate enough persistence
and conclude that the “mechanism to solve persistence problem must be
found elsewhere". Ascari focusses on staggered wage setting, and finds that
whilst nominal wage rigidities lead to more persistent output deviations than
with price setting, they are still not enough to explain the data. Based on
these conclusions, it is commonly inferred that in a dynamic equilibrium
framework, the staggered contracts cannot generate enough persistence.
In this paper, we follow Ascari in focussing on staggered wage-contracts.
However, we show that by allowing for an economy with a range of contract
lengths, the presence of longer contracts can significantly increase the degree
of persistence in output following a monetary shock. We calibrate the model
in a way that in either the CKM or Ascari setting would not generate much
persistence. We show that even a small proportion of longer contracts can
significantly increase the degree of persistence. For example, we consider
the case of a economy where 90% of the economy consist of simple 2-period
Taylor contracts, and 10% have 8-period Taylor contracts (the average is
2.6 quarters) and show that the economy has a marked increase in output
persistence. We also take an empirical distribution of contract lengths (from
1-8 quarters) for the US taken from Taylor (1993) and show that this will
generate a significant degree of persistence.
It has long been observed that in the Calvo setting there can be a sig-
nificant backlog of old contracts: for example, with a reset probability of