S.G.B. Henry 923version of the NKPM augmented by anad hocequation for commodity price infla-
tion, their structural model is different. In their “stylized structural model” they use
a variant of a dynamic AD function composed of all interest sensitive spending, so
amalgamating consumption (the dynamics of which depend on habit persistence)
with investment (with dynamics dependent on adjustment costs). Their main find-
ing is that over the later sub-period monetary policy appears more stabilizing, a
finding at variance with those reported by SW.
In sum, the evidence from the results of this short, but fairly representative, sum-
mary of single-equation and small-complete models of the NKPM sort shows there
are findings supporting monetary regime effects in the US, but equally there are
findings which reach the opposite conclusion. In spite of this somewhat inconclu-
sive state of play, we take the view that much of the research reviewed here points
to the importance of more emphasis on econometric testing. SVAR studies, in par-
ticular, have reached diametrically opposite conclusions, in part because different
theoretical identifying restrictions are employed by different authors. More, rather
than less, empirical testing is one possible way out of this impasse. Also, a lesson
from the single-equation research is the probable gains of a more complete treat-
ment of the system as represented by (18.1)–(18.3) above, since there are evident
shortcomings in testing for regime change effects with single equations only.
18.2.3 The importance of uncertainty
Much of the research described so far, particularly that in section 18.2.2, has used
the rational expectations hypothesis. But there is a burgeoning literature, which
again is largely found in applications to the US, which emphasizes uncertainty
about the effects of regime change, transmission mechanisms and shocks hitting
the economy.
Among the alternative approaches under this heading, the first is the argument,
primarily associated with Sargent (1999), emphasizing changes in governments’
beliefs as central to inflation and unemployment behavior. In this approach, as
in the rest of the literature reviewed here, information is assumed to be symmet-
ric between private and public sectors, with neither side having an informational
advantage.^8 In the Sargent model, the crucial assumption is that the authorities
learn about the “true” economy over time. (This model is discussed more fully
in section 18.3.) An emphasis on more general forms of uncertainty is found in
the important work by Orphanides and associates on the effects of uncertain nat-
ural rates of unemployment, as in Orphanides (2001, 2002) and Orphanides and
Williams (2006). A related development has considered the effects of uncertain
rates of technical progress, including work on the effects of technology shocks
on monetary policy performance, an example of which is given in Gali, Lopez-
Salido and Valles (2003). Much of this line of analysis comes broadly under the
“policy mistakes” heading. Lastly, there is the “bad luck” view, which figured in
the previous section, and which argues that it was the volatility of exogenous,
non-policy, shocks that were primarily responsible for the high volatility of infla-
tion and growth in the 1970s and subsequent falls in these exogenous shocks that
were responsible for the improved US performance in later decades. The paper by