Discussion
Jan in 't Veld
The chapter by Antonio Afonso and Peter Claeys provides a new method of calculating
indicators of fiscal policy. The objective is to uncover any underlying past trends behind
developments in public finances that may explain the current budgetary outlook. The
authors propose their fiscal indicator as an alternative to the standard method of cyclically
adjusting budget balances (CABs), which according to them does not properly reflect
discretionary shifts under control of government. They interpret their results in great
depth and conclude that a procyclical bias in fiscal policy has resulted in a ratcheting-up
in the size of governments. I do not disagree with the general thrust of their analysis of
what has gone wrong with fiscal policy in EMU in the recent years. Rather then
commenting on the specific aspects of four countries considered, I would like to give here
a more general discussion of the method and the resulting indicator with a view to its
usefulness for policy evaluation.
The strength of the chapter is in my view that the approach based on structural VARs
allows for a simultaneous determination of a measure of cyclical output and fiscal
balance. According to the authors their fiscal indicator has as advantage over standard
CABs that it captures the so called supply side effects of fiscal policy. It should be
pointed out that the criticism the authors put forward to justify their alternative approach,
i.e. the difficulties in interpreting the structural balance in the absence of economic
arguments to underpin the trend-cycle decomposition, applies chiefly to statistical trend
extraction procedures and less to the production function approach adopted by the
European Commission. Having said that, the proposed VAR-based indicators provide an
interesting alternative to the CABs. The production function approach does not take into
account the growth effects of fiscal policy due to taxation and productive spending, while
this indicator in principle could include this. As the authors point out, it is therefore more
consistent with the distinction between supply and demand side effects of fiscal policy
adjustments found in recent DSGE models, which unlike earlier RBC models not only
incorporate supply side effects through wealth effects and labour/leisure choice but also
give a larger role to demand side effects by including financial frictions such as liquidity
constraints.(Results from these models suggest fiscal policy has a significant demand side
effects and counter-cyclical fiscal policy has reduced output volatility).
A second advantage of the VAR-based indicators that the authors propose is that it allows
identification of confidence bands. The standard CABs have large cumulative uncertainty
due the various assumptions made in calculating output gaps but this is never shown
explicitly. Just like Bayesian-estimated DSGE models (with have a proper structural
identification), the SVAR method applied by the authors allows for an explicit assessment
of uncertainty.
SVARs are useful tools. However, it is important to be aware of their shortcomings in
properly identifying short and long run effects. It has been shown to be difficult to
159