The name is absent

cyclical elasticity of total net lending varies around 0.50. Most of the variation in the
budget comes from procyclical corporate and personal taxes.

Table 5.1 OECD Output Elasticities of Various Budget Items

Source: Girouard and André (2005).

Quite some uncertainty surrounds the computation of structural balances in this two-step
procedure. Depending on the skewness of the distribution of the moving-average weights
in the filter that is being applied and the phase of the economic cycle, trend output is
biased towards actual values especially towards the end of the sample. Another problem
is posed by structural breaks. Windfall revenues or unexpected spending are entirely
included in the structural balance if they have no economic effects. Filters distribute the
effects of a break forward and backward on the trend. But this problem is not limited to
statistical methods. Even if we use the production function or consider a deterministic
trend a reasonable approximation to potential output, incorporating shifts remains a
problematic issue. The production function approach moreover suffers from plenty of
assumptions that make cumulative uncertainty rather large.⁷⁵ The various assumptions on
budget elasticities are not as crucial for the cyclically adjusted balance, but are
nevertheless not less problematic. Implicitly, it is assumed that average budget elasticities
have a time-invariant linear relation to changes in the economy. We return to these
difficulties in a sensitivity analysis in section 4.4.

Towards an Economic Indicator of Fiscal Policy

The main difficulty in interpreting the structural balance is the absence from economic
arguments to underpin the trend/cycle decomposition. There is an implicit assumption in
the filtering methods on the frequency of the business cycle and hence on trend output
under average economic conditions. And while the production function approach builds
upon economic foundations, the dynamics are nonetheless driven solely by the longer-
term effects of investment feeding back on changes in the capital stock.⁷⁶

Macroeconomic models that allow for cyclical fluctuations around some steady-state
trending growth path can be found in the growing class of Dynamic Stochastic General
Equilibrium (DSGE) models with nominal rigidities. These models have by now been
extended to include fiscal policy. In the initial Real Business Cycle models, there are only
supply-side effects of fiscal policy that transmit through wealth effects and the
labour/leisure choice (Baxter and King, 1993). Micro-founded models based on sticky
prices provide a rationale for stabilisation policies, but even in the New Keynesian type of
models of fiscal policy, the supply side effects still tend to dominate demand side effects
of fiscal policy management (Linnemann and Schabert, 2003). A larger role for demand
side effects of fiscal policy is only found in models that introduce some further

These assumptions relate to its functional form, the presence of returns to scale, technological
progress, the utilisation rates of production factors and the use of auxiliary estimates.

Potential output is nevertheless assumed exogeneous in the production function approach.

127

<<   <   119   120   121   122   123   124   125   >   >>