Shutting off the real interest rate response makes little difference to the cumulative
GDP response at 1 year, but often a large one at 3 years. In the 4 countries with the
largest decline in the GDP response is S2 relative to S1, the real interest rate response
increases in S2 relative to S1, although significantly so only in Germany and the UK.
Shutting off the real interest rate response causes the GDP response to shift up in all
these cases, and in particular in Germany and the UK. In fact, now the GDP response in
these two countries is larger in S2 than in S1.
Still, the real interest rate is an endogenous variable, hence deeper explanations are
needed. Two promising candidates come to mind. First, changes in monetary policy.
Several studies have documented a more aggressive anti - inflationary stance since the
early eighties, in the form of a higher coefficient on expected inflation in the Taylor rule, or
a higher coefficient on the output gap. The models of Galí, López-Salido and Vallés [2003]
and of Linnemann and Schabert [2003] predict that in this case the positive response of
GDP and private consumption to a government spending shock would be attenuated.
A second explanation is that financial markets have become more sophisticated in the
last two decades, and react more strongly to fiscal news. Formally, this can be formalized
as higher values of β1 and β4 in equation (6c); from (9), given γ2 < 0, this will decrease
the estimated effect of government spending on output. Both explanations await a deeper
empirical scrutiny