5.3
A SVAR Model for Gauging Fiscal Indicators
There are a variety of reasons for which the cyclically adjusted balance does not properly
reflect discretionary shifts under the control of the government. Its use in assessing fiscal
balances is therefore debatable. Some problems are related to the properties of the
econometric filters that are being used.71 More importantly, we believe fiscal policy
contributes to the size of economic fluctuations. And it does so by adjusting a variety of
spending and revenue items. Recent general equilibrium theories of fiscal policies
provide a rationale for real economic effects of fiscal policies, and stress the prevalence
of its supply-side consequences over short-term demand effects. This is all the more
important for the assessment of the new Stability and Growth Pact. We develop an
indicator of discretionary fiscal policy stance that builds on the recent empirical literature
on the effects of fiscal policy using structural VARs, and combine this with evidence on
the cyclical behaviour of government budget. Next to its favourable properties, the
indicator is best seen as a first step in verifying recent theories of fiscal policy as well as
giving an instrument for assessing the quality of fiscal adjustments.
Fiscal Indicators
The notion of structural balance is based on the premise that total output fluctuates
around some unobserved trend that depends on the long-term potential growth path of the
economy. In combination with some assumptions on the cyclical behaviour of fiscal
policy, this allows deriving a cyclically adjusted balance. Common practice at the
European Commission, IMF or OECD regards the determination of cyclical variation in
output and the cyclicality of the budget as two distinct problems.
First, the output gap usually comes from some trend-extraction procedure with a
statistical filter applied directly to real output. This decomposition in trending and
cyclical components is usually done with a band-pass filter. Alternatively, the output gap
is calculated as the distance from actual to potential output where the latter is based on a
production function for the aggregate economy.72 Second, a bottom-up approach is
adopted for the derivation of the cyclical elasticities of the budget. The output elasticities
of government revenues are based on the taxation structure of each main sub-item73 - in
some cases accounting for collection lags - and the elasticity of the tax bases to output.
The spending elasticity is of relatively minor importance, as only the spending on
unemployment benefits is adjusted for the cycle. Other budget components are assumed
to be cyclically insensitive. Table 5.1 gathers the elasticities from OECD for the major
budget categories in the countries we study.74 As in most other European countries, the
71
72
73
74
Figure 5.2 already illustrates that differences between the various methods are certainly not minor.
The European Commission backs up a Hodrick-Prescott based decomposition with results from the
production function approach (European Commission, 1995). The OECD uses only the production
function method (Giorno et al., 1995). The IMF has no uniform strategy but the production function
method prevails for industrialised countries (IMF, 1993). Many other approaches abound. Methods
that use a Beveridge-Nelson decomposition or track output developments with unobserved
components are less common. Blanchard (1993) asks what the primary surplus would have been, had
the unemployment rate remained the same as the previous year. Chouraqui et al. (1992) compare
different moving benchmarks. Cohen and Follette (2000) use spectral analysis to isolate low
frequency changes in fiscal policy.
The OECD adjusts only social contributions, corporate, personal and indirect taxes.
Girouard and André (2005) update the elasticities in a previous OECD study by van den Noord
(2002).
126