Design of Fiscal Rules

On design, it is useful to start the discussion with a simple comparison between budget balance rules that are combined with expenditure rules and those which are not. Historical observation is consistent with the regression results in suggesting that in general budget-balance rules that are not combined with expenditure rules are less effective. A striking example of this is the United States experience: neither the Gramm-Rudman-Hollings (GRH) Act of 1985 nor its revised version in 1987 succeeded in significantly reducing the fiscal deficit.16 A further example is the Stability and Growth Pact (SGP), which has not so far led to sustainable positions being attained, notably in large EU countries. On the other hand, when the United States turned to an expenditure-based rule, the Budget Enforcement Act (1990-2002),17 a surplus was achieved and maintained for a time. Some EU countries (e.g. Netherlands, Spain, Sweden, Finland and Czech Republic) supplemented the SGP by national rules (in most cases including some expenditure ceilings) and also enjoyed success. There were, however, some failures. For instance, after France introduced multi-year objectives for real government expenditure in 1998, its structural fiscal position deteriorated continuously until 2003, at which time it came under the European excessive deficit procedure.

There is no one-size-fits-all rule applicable to every country but there seems to be a consensus that, to be effective, rules should have several features. In particular, they should be simple to manage, understand and monitor, while flexible enough to respond to the cycle. Against this background, there are several features of expenditure rules that can explain why they have often been associated with success: not only do they exclude cyclically volatile revenues but they can be (and often are) designed to let economic stabilisers work in a downturn and to save windfall gains during an upturn; 19 they are typically more transparent than all but the simplest budget balance rule; they allow spending ministers/ministries to be held accountable;20 and they make the availability of financial resources predictable for policymakers and programme managers.

An important issue in designing fiscal rules is their possible impact on the quality of public expenditure. Both expenditure rules covering total spending and budget balance rules can potentially cause allocative inefficiencies by biasing spending towards items that are politically sensitive and difficult to cut.21 Typically governments have responded by excluding some capital items from overall spending (as done notably by Golden rules in the United Kingdom and Germany), but this may make the rule more difficult to monitor as well as easier to circumvent. Moreover, there is an element of arbitrariness in excluding physical investment from the rule but not current spending with investment attributes, such as spending on education.
The time period over which the target is to be met is also important, not least in providing flexibility to deal with cyclical fluctuations. Although enforcing the rule on a year-by-year basis appears strict, many countries do just that, with varying degrees of success. Switzerland is an example of a country combining year-by-year enforcement with cyclical flexibility by targeting a balanced budget in cyclically adjusted terms. The United Kingdom pursues another approach: its budget-balance rule22 holds over the business cycle. Such a procedure, however, provides less accurate short-term guidance. As well, rules defined over the cycle or embodying some kind of cyclical adjustment require a subjective23 assessment to be made about the cycle’s start and end dates and/or the size of the output gap, which (together with data revisions) creates a degree of uncertainty about whether or not the rule was (or will be) met. The same objections apply to rules such as the SGP that allow normal procedures to be waived in conditions of pronounced cyclical weakness.

-Fiscal consolidation: lessons from past experience