section sample, in order to capture those individual country characteristics. A tentative
model is given by the following equation,
S it = βi + ðS it 1 + θBit I + U it (27)
it i it — 1 it — 1 it ,
where S is the primary surplus as a percentage of GDP, B is the debt-to-GDP ratio, the
index i denotes the country, the index t indicates the period and βi stands for the
individual effects to be estimated for each country i, in order to test:
i) if θ = 0, the budget surplus does not react to the level of public debt, then the
price level could be determined by the government budget constraint;
ii) if θ > 0, the government tries to increase the budget surplus in order to act in
react to the existing stock of public debt and comply with the budget constraint,
this could be seen as a sign of a regime of monetary dominance.
Table 2 reports the results regarding the estimation of equation (27). Notice that one can
not reject the hypothesis θ > 0, since this coefficient is indeed statistically different from
zero and positive. In other words, the EU-15 governments seem to act in accordance
with the existing stock of public debt, by increasing the budget surplus as a result of
increases in the outstanding stock of public debt. This is also consistent with a Ricardian
regime, where fiscal policy adjusts to the intertemporal budget constraint, preventing for
that reason the determination of the price level through the budget constraint.
The feasibility of the random effects model is assessed by the Hausman statistic, which
tests the null hypothesis that the random effects are not correlated with the explanatory
variables. In our case, and taking into account the fact that the test statistic is significant
at the 1 per cent level, the random effects model hypothesis is rejected, in favour of the
fixed effects model.
23