Kimball curvature parameter, serve to flatten the Phillips curve, discretion is more likely to
dominate timeless perspective policymaking when ω is large. Unlike in Figure 3, however,
the frequencies with which discretion dominates the timeless perspective are much larger at
all points of the parameter space and the ability for discretion to be better than the timeless
perspective occurs at much lower values of ξ, i.e., at much smaller levels of nominal price
rigidity.
6 Conclusion
In this paper I have shown that discretion can be superior to timeless perspective policymaking
and I have identified factors that contribute to this. Broadly speaking, discretion is more likely
to be superior to timeless perspective policymaking when the Phillips curve is relatively flat,
i.e., in models where nominal price rigidity is important or where factors such as Kimball
aggregation or firm-specific labor∕capital are present. These findings are important because
these very factors are becoming widely employed in the New Keynesian DSGE models used
to analyze monetary policy. Although a timeless perspective approach to policymaking may
have its attractions, one should not simply assume that timeless perspective policymaking is
superior to discretion.
One difficulty with comparing discretion to timeless perspective policymaking has been
finding a suitable metric for assessing performance. This difficulty arises because the time-
less perspective introduces auxiliary state variables that are absent from the time-consistent
equilibrium. Rather than simply assigning initial values to these auxiliary state variables or
using unconditional loss to evaluate policies, I propose evaluating policies using a measure of
conditional loss that integrates out the auxiliary state variables conditional upon the known
predetermined state variables. The measure of performance that I develop is easy to com-
pute, provides a consistent treatment of the initial conditions in the discretion and the timeless
perspective equilibria, and is consistent with the conditioning assumptions that describe the
associated optimization problems.
The goal of this paper has not been to criticize the timeless perspective as an approach to
policy design. Rather, because timeless perspective policies are suboptimal, the goal has been
to highlight that timeless perspective policies are not necessarily superior to other suboptimal
policies, of which discretion is a leading example. It is certainly unclear why a central bank
should commit to implementing a timeless perspective policy when that policy is inferior to a
time consistent alternative. The results in this paper suggest that studies analyzing timeless
24