the difference between the cumulative impulse response of GDP in S2 and S1, at 1 and
3 years, with an asterisk if this difference is significant (if the difference is less than .10
in absolute value and insignificant, a “=” is entered). Rows 1 to 5 display the results
of as many different specifications: a linear time trend (the benchmark case displayed in Table 6), linear and quadratic trends, levels, stochastic trend36, and stochastic trend with
cointegration between spending and taxes (as it is well known the latter specification is
a way of imposing the intertemporal government budget constraint in the estimation).
Under all specifications, the key result remains: the only difference is that in Germany
under a stochastic trend and under cointegration the response of GDP is slightly larger
in S2 than it S1 at both horizons. On the other hand, the evidence is stronger in a
specification in levels: except in Australia at 1 year, now the response in S2 is significantly
smaller in all countries and at all horizons.
The pattern of the differences between the two subperiods remains largely unaffected
if one omits the 3 years between 1973:1 and 1975:4 (row 6), which were characterized
by large swings in GDP growth and government spending; if one assumes a break points
between the two subperiods of 1976:1 instead of 1980:1 (line 7); or if government spending
is ordered second (line 8). In this last case, typically the point estimates (not shown) of
the impulse responses at all horizons change by only a few hundredths of a percentage
point, reflecting the small correlation between the structural spending and net tax shocks.