adjusted balances (e.g. Alberola et al. (2003), OECD (2003), Forni and Momigliano
(2004)) or the impact on the level of cyclically adjusted primary balances (e.g. GaH and
Perotti (2003)).
4.3 A Stylised Framework
Several techniques have been developed to estimate the variations of budget aggregates
arising from the economic cycle.54 The conventional approach (e.g. EC, OECD, IMF) to
correct budget balances for fluctuations in economic activity starts from a notional
decomposition of the observed budget balance bt into two (unobserved) components: the
cyclically adjusted budget balance bst , often called "structural" balance, and a cyclical
component bat aimed at capturing the built-in stabilisers. To adequately estimate the
cyclical component bat , various methods have been developed by international
institutions such as the EC, the OECD, the IMF and the ECB. Within these approaches,
the structural balance bst is defined as the difference between the observed and the
cyclical balance, bst = bt - bat . Obviously, any other dimension of fiscal policy, even if it
is related with the cycle, shows up in the structural component.
However, if the focus is on the development of the underlying fiscal position (adjusted for
all temporary impacts irrespective of whether they are “economy dependent or policy
dependent”55 (Braconier and Forsfalt, 2004, p.4)) a direct calculation of the structural
balance as a “long-run component” via specific filtering techniques (see Brandner et al.,
1998) may be more appropriate. If so, the effects of the built-in stabilisers as well as
cyclically related discretionary measures are captured in the resulting "cyclical"
component bat = bt -bst .
To analyse the issues raised, we set up a framework that allows distinguishing between
several dimensions of fiscal policy, short-run vs. long-run, and active vs. passive. We
start with a quite general decomposition
bt = μt + bat + bdt + εt (4.1)
of the actual/observed balance bt into the core balance μt , two cyclically related
components - namely bat capturing the impact of the automatic stabilisers, and bdt
capturing the discretionary policy in response to the cycle - and a residual component εt
reflecting all remaining (temporary) effects ("fiscal noise"). To be more precise, we
specify
bat = αt ■ Ia (4.2.1)
bdt = γt ∙ Id (4.2.2)
54
55
However, all these techniques are subject to a number of methodological problems, notably defining
trend/potential output - a shortcoming that unfortunately is also valid for our approach.
I.e. pro- or countercyclical discretionary policy measures such as a temporary increase in the volume
of labour-market programmes due to political decisions.
104