Notes: p-statistic reported under coefficients; White heteroskedasticity-consistent standard errors.
Instruments: yt-1 and Xt-1, and yt, yt-1, Xt and Xt-1 for the US and Japan, The constant and country fixed-
effects not reported. In the case of Italy, the set of available information for λbEt-1yt+ and λaEt-1yt+ is
too small so we do not distinguish between pre and post 1992 - i.e. we impose λb = λa. Unbalanced
panel.
The country-by-country estimates are often imprecise, which partly reflects the short size of
the sample. We focus our discussion on the panel estimates, as shown in the columns labeled
“Pooled 1”. Table 2 summarizes the test results. These results confirm previous findings that
fiscal policy has become counter-cyclical since 1992. Yet, the finding that βa- > 0 indicates
that this shift occurred during upswings while fiscal policy remains approximately acyclical
during downswings. In other words, the SGP is having asymmetrical effects. These results
differ from those reported in the EU Commission (2006) which finds that prior to 1992 fiscal
policy was procyclical in bad times and acyclical otherwise, while it has become procyclical
in good times after the adoption of the euro. To check whether our results, the column
“Pooled 2” shows the results when we break the sample in 1999, so that “before” refers to the
period 1970-1998 and “after” to the period 1999-2006. The results remain practically
unchanged.10
Table 2. Tests
|
Fiscal policy counter-cyclical until 1991 |
βb < 0 |
No, acyclical |
|
Fiscal policy counter-cyclical after 1992 |
βa < 0 |
Yes |
|
Fiscal policy more counter-cyclical in Maastricht years |
βa < βb |
Yes |
|
SGP limits counter-cyclical policy in downswings |
βa- > 0 |
Yes |
|
Fiscal policy acts on demand to compensate for external |
θb < 0 |
Yes |
|
Fiscal policy acts on demand to compensate for external |
θa < 0 |
No |
|
Fiscal policy acts on costs to deal with external |
θb > 0 |
No |
|
Fiscal policy acts on costs to deal with external |
θa > 0 |
Yes |
Regarding the use of fiscal policy to deal with external competitiveness, we find that the
situation has been radically altered following the adoption of the Maastricht Treaty. Up until
1991, our results suggest that governments used fiscal policy to make up for demand shortfall
when they faced external competitiveness losses, and conversely tightened up fiscal policy
when external competitiveness was boosting demand. After 1992, instead, they tightened up
fiscal policy in the face of deteriorating competitiveness as if they were using their last
remaining macroeconomic instrument to weigh on costs. These results suggest that
governments are now willing to suffer short-term demand shortfalls to restore cost
competitiveness. Put differently, fiscal policy is not necessarily misused.
10 They are also unchanged if we compare the periods 1970-1991 and 1999-2006, dropping the
Maastricht years 1992-98. A possible explanation is that the Commission’s results are based on a
graphical analysis. Another possibility is that they look at the realtionship between changes in the
cyclicaly-adjusted budget and in the output gap while we look at levels.