10
GALSTYAN AND LANE
manner, with an increase in relative GDP per capita and a long-run trade deficit both
associated with real appreciation.
Columns (2) and (3) show the results when the sample is split between the G3 and
non-G3 groups. In each group, the main fiscal results from the full sample carry over in
qualitative terms. However, in line with the findings of Lane and Milesi-Ferretti (2002,
2004) and Galstyan (2007), the magnitudes of the coefficients are much larger for the G3
group. It is also noteworthy that the estimated semi-elasticity of the real exchange rate
with respect to relative government consumption for the G3 group is quite similar to
the estimate reported by Ricci et al (2008). Finally, we note that the control variables are
significant for the non-G3 group but are individually insignificant for the G3 group.
We turn our attention to the EMU sample in columns (4)-(6). The results are broadly
similar to the wider panel, with the exception that neither fiscal variable is individually
significant for the E4 group of larger member countries.
The estimates for the relative price of nontradables are reported in Table 3. An in-
crease in government consumption is significantly associated with an increase in the
relative price of nontradables for the EMU and E4 samples. In contrast, it is striking that
an increase in government investment is significantly associated with a decline in the
relative price of nontradables for each sample, with the exception of the E4 group. For
the non-G3 and non-E4 samples, the approximate coefficient estimate of -3.4 indicates
that a 1 percentage point increase in the ratio of government investment to GDP (which
corresponds to a one standard deviation increase in the panel data) is associated with a
3.4 percent decline in the relative price of nontradables. In relation to the control vari-
ables, we note that the relative price of nontradables is increasing in the level of GDP per
capita across all country groups.25
According to our model, government investment affects the structure of relative prices
through its impact on relative sectoral productivity: it generates real appreciation if pro-
ductivity is raised relatively more in the traded sector than the nontraded sector, and
conversely it is associated with real depreciation if the productivity gains are concen-
trated in the nontraded sector. We directly examine this mechanism in Table 4, where
we regress (A*N/AT) on GDP per capita and the level of government investment, using
the same DOLS(1,1) estimator. The results in Table 4 are consistent with those in Table
3, in that there is evidence that an increase in government investment disproportion-
ately raises productivity in the nontraded sector for most of the country groups. (As
expected, an increase in GDP per capita is associated with an increased in traded-sector
productivity relative to nontraded-sector productivity).
25In addition, the trade balance is positive and marginally significant for the non-E4 group.