Rules constrain policymakers’ autonomy in policy decisions. But what is the justification for limiting discretion in fiscal policy? While recognizing that history, tradition, and preferences—in addition to economic factors—may cause the optimal size of government and public debt to vary across countries, the political economy literature has shown how the political and institutional environment can lead to distortions in the conduct of fiscal policy, resulting in outcomes that may be undesirable from society’s point of view (see Appendix I).
Several factors can push policymakers toward structural budget deficits (Alesina and Perotti, 1995). Politicians faced with losing power may discount the future more heavily than private agents, or they may be induced to manipulate policy levers to facilitate their re-election. Conflicts among coalition partners may block the policy decision process, thus delaying fiscal reforms.
Certain features of the political process can also lead to excessive government spending. For instance, the literature on the so-called common pool problem highlights the distortions that can result from forms of collective decision making in which policy reflects the aggregation of individual decisions. A classic example is elected officials, spending ministers, or parties in a coalition government asking for projects that benefit their constituencies but fail to fully internalize the consequences of the higher taxes needed to finance such decisions. The result is excessively high spending (Weingast, Shepsle, and Johansen, 1981).
Fiscal rules have been proposed as a device to overcome these distortions. For instance, a deficit bias can be addressed through a legally binding balanced-budget rule. Alternatively, setting an aggregate spending ceiling (that is, an expenditure rule) forces individual spending bids to explicitly take into account aggregate resource constraints, thus reducing the common pool problem.15 The remainder of this section explores the design of fiscal rules in more detail.
Deficit and Expenditure Rules and Targets
Countries committed to a multiperiod fiscal framework typically anchor fiscal policy on a numerical rule relating to the budget balance, public expenditure, or the public debt (see Appendix II).16 The rule needs to support, but need not be identical with, the ultimate objective of the policymaker. For instance, an expenditure rule can be designed to safeguard a medium-term target for the budget balance by being consistent with the projected path of revenues.
Perhaps the simplest and most intuitive rule is a balanced budget or, more generally, a rule on the maximum level of the budget deficit. The rule recently adopted in Spain and the 3 percent deficit threshold of the Maastricht Treaty are examples of rules on the budget balance. Such a rule has the dual advantage of being easy to explain to the public and market participants and being relatively simple to monitor.17 An obvious drawback is that it does not address biases toward excessive expenditures, as higher expenditures can be financed through higher taxes. Another shortcoming is that fiscal policy would become procyclical: if revenues decline and expenditures rise in a recession, a discretionary tightening would be needed to keep the balance in check. In addition, the “activist” fiscal policy needed to abide by this rule might imply a tax policy that varies over the cycle. Adjustment taking place on the revenue side would conflict with principles of optimal taxation. On the other hand, if expenditures have to adjust, efficient medium-term expenditure management may become difficult, or capital expenditures may be excessively squeezed.18
Some of the difficulties of a budget balance rule can be remedied by targeting the structural balance.19 This rule requires taking a stand on the position of the economy in the cycle and the effects of the cycle on fiscal revenues and expenditures. Different methodologies can be used. If the government can change the methodology or parameters from year to year, then the scope for manipulation may be large. To be credible, the rule itself should specify the methodology and parameters of the cyclical adjustment; alternatively, an independent agency could be put in charge of estimating the cyclically adjusted deficit. In either case, the rule may be difficult to explain to the public, and if it is enshrined as a law, to a court. This would make the rule difficult to monitor and enforce. In addition, the methodology may be in-adequate, especially if there are structural changes in the economy or if failure to revise the rule leads to wrong policy decisions.
Instead of focusing on a country-specific goal for the balance, some countries set binding upper limits on expenditure growth as a key operational focus. Finland, the Netherlands, and Sweden are examples (see Appendix II). This type of framework directly addresses distortions leading to excessive spending and does not automatically lead to a procyclical fiscal stance because stabilizers on the revenue side are free to operate.20 This type of rule can also curb the tendency to increase public spending during upturns. In addition, an expenditure rule can be easily explained to the general public and market participants, provided that the control aggregate is clear.
One drawback of an expenditure rule is that it does not necessarily correct a tendency toward excessive deficits, for instance through large tax cuts or the systematic overprediction of revenues. However, empirical studies suggest that fiscal consolidation based on expenditure reduction tends to be long-lasting (Alesina and Perotti, 1995). In addition, the deficit risk can be overcome by anchoring the framework over the medium term—for example, by supplementing the binding expenditure rule with an explicit medium-term “target” for the budget balance, as in Sweden. While a target represents a weaker commitment than a binding rule, it may nonetheless ensure enough discipline. Because the target would have to be met only over the cycle, fiscal policy need not be procyclical. A slight variant to this approach is to adopt an explicit target for the stock of debt relative to GDP, arguably a more relevant yardstick for the sustainability of public finances since changes in the stock of public debt often differ from the budget balance. In the approach taken by Switzerland, revenue forecast errors are cumulated into a “notional” debt stock that must be reduced to zero over time (Appendix II).
All in all, the advantages of an expenditure rule may outweigh its drawbacks in the four countries under study, where the tax burden is high and the medium-term level of the deficit is already bounded by the SGP. To safeguard against systematic revenue underperformance or excessive tax cuts, such a rule may be combined with a medium-term stock anchor, such as a target for the ratio of public debt to GDP. This target should achieve the medium-term goals of reducing the tax burden, addressing future population aging, and ensuring an adequate margin for counter cyclical fiscal policy. The next section examines the broad design features of a fiscal framework based on an expenditure rule.
Issues in Designing Multiyear Expenditure Rules
Macroeconomic Assumptions and Time Horizon
Multiyear fiscal plans need to be based on a macroeconomic scenario that projects the evolution of the various expenditure categories and illustrates how they relate to overall economic trends. A crucial choice is whether to adopt a “realistic” or “cautious” scenario, an issue that also arises in drafting the yearly budget. The advantage of a cautious scenario is the likelihood of “favorable” surprises ex post, as cyclically sensitive spending components are likely to turn out lower than projected. Accordingly, each year there will likely be some room within the expenditure ceiling to finance new programs or deal with overruns in noncyclical expenditure categories. This may facilitate negotiations within the cabinet and help deal with or accommodate unexpected spending without violating the framework.21
A realistic macroeconomic scenario offers distinct advantages. A cautious scenario obfuscates the true fiscal goals of the government, defeating the purpose of the multiyear fiscal rule to make such goals explicit and build consensus around them. A cautious scenario may also encourage ministries to second-guess the likely scope for additional spending, resulting in unrealistic initial budgets. Furthermore, an overly cautious scenario may lead the authorities to use the additional room to engage in procyclical policy when the outturn is favorable. With a realistic macroeconomic scenario, a limited margin for new discretionary spending can be introduced as a contingency reserve.22
A longer time horizon means less discretion but also less flexibility. The duration of the legislative session (as used in the Netherlands) may be a natural time frame for rules that are self-imposed by a governing coalition because it acknowledges that such rules would not bind future governments. A four-year rolling horizon, as in the stability programs, would also be a sensible choice, provided that each update only introduces a ceiling for the additional year rather than modifying those for previous years. Using multiyear frameworks to constantly push adjustment into the future would, of course, be counter-productive.
Real Versus Nominal Rules
The multiyear framework could specify either the evolution of nominal spending or real spending. With a real rule, the spending ceiling for the year is transformed into a nominal ceiling using the latest inflation forecast.23,24 In contrast, with a nominal rule changes in the inflation outlook do not lead to revisions in spending ceilings. Accordingly, a nominal target implies lower real government expenditures when inflation is unexpectedly high. If higher inflation results from excessive domestic demand, reducing real government expenditure would help cyclical stabilization. Similarly, in the case of a permanent adverse supply shock, lower real government expenditure would help absorb the shock and stabilize the expenditure-GDP ratio. By contrast, if the supply shock were temporary, a more appropriate response would be to increase government spending, but the nominal ceiling would force a reduction instead. Thus, a nominal rule may be more appealing in countries where cyclical stabilization is an important concern—temporary supply shocks are unlikely, and automatic stabilizers on the revenue side are small.
The presence of automatic indexation rules on some spending categories may make a real rule more attractive. In a number of countries, pensions and other entitlements are indexed to expected inflation; thus, a real spending rule makes it easier to set (and to respect) the ceiling on entitlement spending.
The Choice of the Aggregate
In choosing which expenditure aggregate to target, it is desirable to use a comprehensive measure. A broad aggregate is most relevant from a macroeconomic perspective. In addition, defining expenditure narrowly makes it easy to circumvent the ceiling by introducing new expenditures as items not covered. An aggregate with a clear counterpart in national account statistics would also increase transparency and facilitate monitoring, which are key elements of a fiscal rule.
A comprehensive aggregate, however, has its drawbacks. For instance, if there is a history of containing spending growth by excessively compressing capital expenditures (as in the United Kingdom), excluding these expenditures may be desirable. In this case, the framework must also clearly spell out the criteria that separate investment from current expenditures. This limits the scope for arbitrary expenditure reclassifications designed to get around the rule. Interest payments could also be excluded on the grounds that they are not under the direct control of the government. This would not encourage creative accounting, given that this type of expenditure is easily identifiable.25 However, to the extent that the main goal of the spending rule is to ensure the compatibility of small deficits and lower taxes, capital and interest expenditure should remain in the control aggregate.
Another issue is whether to include cyclically sensitive expenditure items, mainly unemployment compensation. While comprehensiveness and ease of monitoring favor inclusion, cutting discretionary spending during a downturn to make room for higher unemployment outlays may be undesirable. On the other hand, if changes in unemployment appear to be permanent (as in the four countries of interest in the last 30 years), then it is necessary to adjust other expenditures permanently rather than letting the deficit increase. Excluding this item from the framework would not help achieve this outcome. Table 3.1 shows that unemployment spending is a small component of total spending in the four countries; but when a recession hits, the increase in this spending item can be nontrivial. In addition, unemployment spending was fairly closely correlated with the cycle in Germany, France, and Italy during 1980-2000, though not in Spain, where unemployment had a strong structural component. All in all, despite the possibility of structural changes in unemployment, it may be preferable to exclude this item from the spending ceiling to increase the ability of the frame-work to deal with cyclical fluctuations.
Characteristics of Unemployment Spending
(In percent of GDP)
Characteristics of Unemployment Spending
(In percent of GDP)
| France | Germany | Italy | Spain | ||
|---|---|---|---|---|---|
| Unemployment spending during 1980–2000 | |||||
| Average | 1.3 | 1.6 | 0.5 | 2.4 | |
| Maximum yearly increase | 0.3 | 0.4 | 0.2 | 0.5 | |
| Correlation with the output gap | –44.9 | –37.2 | –47.1 | 11.2 | |
Characteristics of Unemployment Spending
(In percent of GDP)
| France | Germany | Italy | Spain | ||
|---|---|---|---|---|---|
| Unemployment spending during 1980–2000 | |||||
| Average | 1.3 | 1.6 | 0.5 | 2.4 | |
| Maximum yearly increase | 0.3 | 0.4 | 0.2 | 0.5 | |
| Correlation with the output gap | –44.9 | –37.2 | –47.1 | 11.2 | |
Should the expenditure rule cover only discretionary expenditures, or also entitlements? In the four countries under consideration, entitlements are large and growing—because of demographic and economic trends and because of pressures to create new entitlements and expand existing ones. In addition, it is relatively easy to disguise new types of expenditures as entitlements. Including entitlements, however, creates problems for enforcement because automatic cuts in case of overshoots are not feasible (an issue addressed in the next section). All in all, including entitlements is necessary to make the spending rule effective in the four countries.
Compliance and Enforcement
When a country adopts a binding fiscal rule, the inevitable question is: How will the rule be implemented and enforced? Departures from the rule can occur ex ante if the yearly budget proposed to parliament does not conform with the multiyear rule, or ex post if departures occur at the budget implementation stage. An effective rule should ensure compliance both ex ante and ex post.26
The means of enforcing a fiscal rule ex ante depend on its statutory nature. Is it a constitutional amendment, on organic law, a regular law, or simply a political commitment? In the latter case, violations are legally possible but presumably result in lost credibility that may damage the government. A rule sanctioned by ordinary law can be violated if parliament approves a new law. In this case, a clear political commitment to the framework is therefore necessary to ensure that violations result in loss of reputation. Constitutional amendments and organic laws are usually more difficult to modify, thereby representing the strongest form of commitment. Because the commitment is so strong, however, budget rules written into constitutions may need to be vague to allow for flexibility to deal with unforeseen circumstances, resulting in less effectiveness. In the case of a spending rule, in which the numerical value of the rule is set every few years, a constitutional amendment is not a practical solution, and the rule would have to be entrusted to a political agreement, possibly strengthened by an ordinary law.
Concerning ex post compliance, the rules should specify how ex post deviations will be clawed back in subsequent years.27 This would provide an important anchor to the framework. To minimize deviations, it may be necessary to strengthen budget implementation provisions before introducing the rule.28 All countries have procedures to ensure that the actual budget is close to that which is approved by parliament, but these procedures can be more or less restrictive depending on the ease of transferring expenditures across different chapters, on whether spending has to be approved by a centralized authority (such as the ministry of finance), on the powers of the finance minister to block outlays in case of overruns, on the ease with which the budget law can be amended, and other aspects. In countries with modern public expenditure management systems, ministers are held accountable at various times in the budget process. Safeguards may include hearings and questioning by the legislature, as well as internal and external audits.29
Even though rules may not be broken legally, they may be broken de facto if not clearly written. According to Kopits and Symansky (1998), rules should clearly define the aggregate to be controlled and the escape clauses. Examples of rules that could benefit from greater specificity are a golden rule that leaves substantial room for interpretation of investment (as in Germany), and a “cyclically adjusted” budget balance rule that does not define how to perform the cyclical adjustment.30 The rule should also be transparent, specifying accounting and reporting standards to limit “creative accounting.”31 Simplicity also enhances the ability of the legislature and the public to detect deviations, making the reputation mechanism more effective. Rules that are rigid or that require costly measures to implement in certain circumstances are likely to be difficult to enforce because of the strong temptation to circumvent them. For instance, the second version of the Gramm-Rudman-Hollings law in the United States proved unworkable since it would have entailed drastic cuts in discretionary programs (Appendix II). From this perspective, rules that leave some flexibility to accommodate cyclical changes in the fiscal position may be easier to enforce. Additional flexibility can be introduced through contingency funds to be tapped in case of unpredictable events or emergencies, as in the case of the Spanish Fiscal Responsibility Law, or by allowing for (some) expenditures to be shifted across budget years (as in the United Kingdom’s Departmental Expenditure Limits framework; see Appendix II).
A particularly thorny but relevant issue for the four countries is how to deal with overruns in entitlements. With discretionary spending, last-minute cuts to bring spending back within authorized budgetary limits are always possible, but for entitlements, this is neither realistic nor desirable. In this case, a limited contingency fund could cover overruns resulting from overly optimistic cost projections. Once the resources in the fund are exhausted, then discretionary spending would have to be cut or the overrun would have to be clawed back in the following year. On the other hand, the expansion or introduction of new entitlement programs during the budget year should be forbidden unless they have been explicitly budgeted for under the spending ceilings.
Implementing Fiscal Rules in a Decentralized System
The existence of subnational governments with power to decide on important shares of public expenditure and public revenue and to accumulate debt may enhance the effectiveness of public spending, but it may also hinder the achievement of national fiscal objectives.32 The previous section of this paper examined a spending rule for the general government. Implementing such a rule in a decentralized system is not straightforward, and each possible approach has its advantages and disadvantages.
Among the countries covered by this study, the issue of subnational governments is particularly relevant in Germany, Italy, and Spain. In France, the institutional arrangements result in sufficient control on local finances to implement a general government spending rule. In general, enacting fiscal rules in decentralized systems requires a negotiated agreement between the center and subnational governments that, to be effective, has to be sanctioned by an explicit political commitment or enshrined in law.33 Such an agreement must also be supported by strong reporting requirements to allow monitoring, namely an effective information system that makes reliable and timely fiscal aggregates available at the subnational level.
What should be the content of a pact with subnational authorities? One option is to negotiate limits on spending by each level of government, effectively parceling out the general government ceiling. This type of agreement could include a limit on central government expenditure (net of transfers to subnational governments) and another on the expenditure of subnational governments as a whole. Expenditure ceilings could differ for levels of government and even within each level of government, depending on spending responsibilities and reflecting preferences across regions. This framework would allow automatic stabilizers to work at the level of subnational budgets provided that subnationals can borrow to finance deficits during downturns, or that financing from the center (transfers plus shared taxes) is adjusted to compensate for cyclical revenue shortfalls. The agreement could usefully specify rules for dealing with both ex ante and ex post deviations from the ceilings in order to assure compliance.
A drawback of this approach is the difficulty in reaching an agreement on how to apportion the aggregate spending ceiling, because a large number of entities may be involved. Furthermore, monitoring may be quite complex because subnational governments tend to report with longer lags than the central government, and public expenditure management systems are often less well developed at the subnational level.
An alternative option is to negotiate a budget balance requirement for subnational entities. This type of agreement would probably be simpler to negotiate and monitor than spending limits. If subnationals lack control on the revenue side—as is the case in Germany, Italy, and Spain—a deficit rule might effectively control spending growth over the medium term. However, depending on the formula used to allocate transfers, it could also result in a procyclical fiscal policy at the subnational level. In addition, in years in which revenues overperform relative to expectations, subnational government would be able to expand spending, possibly beyond levels consistent with the general government expenditure rule. To remedy these drawbacks, transfers from the central government (including shared taxes) could be designed to minimize cyclical influences on the revenue side, but this may be complicated and require an overhaul of the whole system of subnational finance.34
All in all, where subnational governments have substantial fiscal autonomy, designing and implementing fiscal rules is more complex than in centralized systems. The main trade-off is between a rule that is simple and easy to monitor (a balanced budget rule for subnational governments) and one that avoids procyclicality (a spending rule). The choice has to be made on a country-by-country basis.