debt service and the factor for industrialised countries and emerging market
economies. Again, the evidence for emerging market economies appears stronger.
<Insert Table 7 here>
6. Robustness Analysis
Our results before make use of nonstationarity assumptions in the data and use
a nonstationary methodology to identify common behaviour between industrial and
emerging market countries. To test whether our results are robust in a different
methodology we adopt Arellano and Bond (1991) and Blundell and Bond (1998)
Dynamic Panel Data Estimator. This also takes account of potential endogeneity
amongst our variables since we use Generalised Method of Moments (GMM) and a
systems estimator for our instruments (lagged values of the variables themselves).
This approach also adds another dimension to the analysis in the previous section by
incorporating an explicit test of the standard neo-classical approach to fiscal policy
popularised by Barro (1979). This takes account of government’s desire to go into
deficit in situations of a temporary downturn in the economy or to a temporary
increase in government spending in order to smooth tax rates. We measure temporary
declines in government spending using detrended GDP and detrended real
government expenditure. Additionally we include temporary increases in government
expenditure on interest payments. This can be considered a proxy for interest rate
effects in the previous analysis.
<Insert Table 8 here>
21