Journal Issue

Fiscal Rules and the Budget Process

Gian Milesi-Ferretti
Published Date:
June 1996
  • ShareShare
Show Summary Details

I. Introduction

The 1970s and 1980s saw the emergence of large fiscal imbalances in a number of OECD countries, leading to rapid accumulation of public debt. This has led to concerns that political forces, together with the procedures whereby budgets are made and implemented, have an inherent bias towards deficits. These concerns have been heightened by the risk that a high level of public debt may constitute a threat to macroeconomic stability, and the difficulties encountered by some countries in permanently reversing the rising trend of the debt to GDP ratio. As a result, increased attention has been devoted to the possibility of adopting legislative measures that would constrain government fiscal policy decisions. This is exemplified by the proposal for a balanced budget amendment in the United States, by the Maastricht criteria on the size of public sector deficits and debt in the European Union, and by the suggestion of a “stability pact” to govern fiscal policy after EMU. This paper examines the rationale for the imposition of “rules” as a way to improve fiscal policy performance, with particular emphasis on the Italian case. It considers rules in a broad sense, encompassing legally established quantitative targets on fiscal variables such as budget deficits, government expenditure, and/or public debt, as well as rules and procedures governing the various stages of the budget process.

Explaining the differences in fiscal policy across countries at a relatively similar level of development has been a major challenge for economic researchers, and there is agreement that economic arguments alone cannot account for the size of these differences. A promising avenue of research has examined the impact of political and institutional factors on macroeconomic policy formation. In this context, government policy formation is treated as endogenous, reflecting the interaction of different agents (policymakers, political parties, private sector) with possibly conflicting interests, rather than being chosen by a benevolent planner maximizing a social welfare function (see, for example, Persson and Tabellini (1990)). In this literature, the institutional framework is important because it defines the “rules of the game” and contributes to shaping incentives and constraints of the various agents. Empirical evidence suggests that political and institutional factors can indeed contribute to the explanation of different fiscal policy choices (see, for example, Roubini and Sachs (1989) and Grilli, Masciandaro and Tabellini (1991)).

The theoretical models that have been proposed to explain fiscal policy choices are a natural starting point for an analysis of fiscal rules. In a broad sense, rules can be viewed as a modification of the institutional framework within which economic policy decisions are taken. Ideally, they should provide a set of incentives and/or constraints that makes fiscal policy actions closer to “desirable” outcomes. The literature on fiscal policy rules has developed along two complementary directions. The first has examined the consequences of legislative limits on fiscal variables, such as budget deficits or government spending (see, for example, Corsetti and Roubini (1994) and Roubini (1995)). The second (von Hagen (1992), von Hagen and Harden (1994, 1995), Alesina and Perotti (1996)) has instead focused on the rules and regulations governing the various stages of the budget process, emphasizing the importance of agenda setting, voting rules, and, more generally, of budgetary procedures in shaping fiscal outcomes. Empirical evidence on the importance of budget rules and budget institutions in explaining fiscal policy, while still at a preliminary stage, is promising. von Hagen (1991), Poterba (1994), Alt and Lowry (1994), and Bayoumi and Eichengreen (1995) find evidence of systematic differences in fiscal policy between US states with balanced budget rules of different “stringency”. von Hagen (1992) and de Haan and Sturm (1994) find that an index of centralization of the budget process contributes to explaining differences in average fiscal deficits, government debt and government expenditure in countries belonging to the European Union. Similar results are obtained by Alesina, Hausmann, Hommes and Stein (1995) using a comprehensive sample of Latin American countries.

This paper provides an evaluation of this literature, examines a number of proposals for reforms in the budget process that have been proposed in Italy (see, for example, Alesina, Marè and Perotti (1995), Commissione Spesa Pubblica (1994), Commissione per la Riforma dei Bilanci Pubblici (1994), Masera (1995a)), and discusses to what degree fiscal rules and reforms of the budget process may contribute to the reduction of fiscal imbalances. The main findings of the paper can be summarized as follows. Numerical fiscal rules can play a role in enhancing fiscal responsibility only insofar as they are adopted in a framework for budgetary reform that addresses existing incentive problems at other levels of the budget formation process. Otherwise they are unlikely to be effective, and could create incentives for creative accounting, and, more generally, for a reduction in the transparency of the budget process. There is also a risk that quantitative targets could hinder the flexibility of fiscal policy to respond to macroeconomic shocks. An alternative strategy that does not rely directly on numerical targets is to enhance the strategic position of the finance minister in the budget formation process, while restricting the scope for amendments at the Parliamentary level and enforcing hard budget constraints at the budget implementation stage. 1/

Clearly, reforms in budgetary procedures per se cannot address the problem of the growth in entitlements, which “explains” a sizable fraction of the divergent behavior of fiscal policy across OECD countries (Alesina and Perotti (1995b)). They can, however, strengthen the relative position of those actors in the budget process that are more committed to fiscal discipline and, through enhanced transparency, raise the awareness of the sources of fiscal problems and increase political accountability for profligate fiscal behavior.

The rest of the paper is organized as follows: Section II discusses political economy explanations for persistent budget deficits; Section III discusses the rationale for rules, and Section IV discusses the link between the budget process and budget outcomes. Sections V and VI present and compare various proposals for budgetary reform in Italy, and discuss more generally the limitations of budget reforms. Section VII concludes, and the Appendix sketches a simple model of excess spending with a deficit bias.

II. The Political Economy of Budget Deficits

When discussing the political economy of fiscal policy, it is natural to ask what benchmark for government behavior that should be adopted. The neoclassical theory of fiscal policy (see Barro (1979), (1989); Lucas and Stokey (1983); Lucas (1986)) stresses the importance of achieving tax smoothing; budget deficits should be used to cover temporary increases in government spending (such as, for example, those due to a war) while tax rates should be kept constant to minimize distortions. Keynesian models of aggregate demand management stress the importance of fiscal policy as a stabilizer: fiscal policy should be expansionary during recessions and contractionary during expansions in order to moderate business cycle fluctuations. The political economy literature provides possible explanations as to why governments may systematically deviate from these principles of fiscal policy. A general feature of these models is that the institutional set-up is taken as given; their focus is on how economic, political and distributional factors interact in a given institutional environment in shaping fiscal policy.

Alesina and Perotti (1995a) provide a useful classification of political economy models of fiscal policy: (1) models based on “fiscal illusion” with opportunistic policymakers and naive voters; (2) models of debt as a strategic variable; (3) models of distributional conflict; (4) models of geographically dispersed interests; (5) models of inter-generational redistribution; (6) models emphasizing the effects of budgetary institutions.

A thorough review of this literature is beyond the scope of this paper. What is of interest for our purposes is the mechanism which causes fiscal policy to “deviate” from its benchmark, however defined. The first class of models is in the spirit of the public choice literature: the key assumptions are that policymakers are opportunistic (that is, they care about electoral prospects, and not directly about private agents’ welfare) and use fiscal deficits to increase their electoral chances. Voters fail to understand the intertemporal budget constraint of the government--they overestimate the benefit of current expenditures and/or underestimate future tax burdens--and therefore do not “punish” politicians for fiscally irresponsible behavior. A second strand of literature emphasizes that the stock of debt has an effect on the policy choices of future governments, and can therefore be used to constrain its actions (Alesina and Tabellini (1990); Tabellini and Alesina (1990)). In this context, a deficit bias can arise because of a conflict about spending priorities between different political parties. This conflict, together with electoral uncertainty, implies that the current government does not fully internalize the cost of running budget deficits today, because the future spending that is going to be compressed may reflect the priorities of a different government. This deficit bias is increasing in the degree of political polarization (reflected in the difference between spending priorities) and in the degree of electoral uncertainty.

Models of distributional conflict (Alesina and Drazen (1991); Drazen and Grilli (1993)) have a somewhat different focus. They emphasize how conflict between different social groups (represented by parties, interest groups, coalition members) can delay the adoption of necessary measures to stem an increase in public indebtedness caused by some exogenous factor. The reason for the delay is that the groups cannot agree on burden-sharing for the necessary fiscal adjustment. These models predict that fragmented or divided governments and polarized societies would have more difficulty implementing fiscal adjustment measures than single-party governments and less polarized societies. Evidence presented in Roubini and Sachs (1989) and Grilli, Masciandaro and Tabellini (1991) for OECD countries and by Poterba (1994) and Alt and Lowry (1994) for US states is consistent with these predictions.

Public debt redistributes the tax burden across time, and can therefore be a vehicle of intergenerational redistribution. Cukierman and Meltzer (1989) and Tabellini (1991) provide interesting examples of cases when public deficits are chosen as a way to achieve intergenerational transfers. In their model, budget deficits effectively act as a form of “negative bequest”, transferring resources from richer future generations to the poorer current generation.

Finally, a number of contributions closer to the political science literature study the interaction between the organization of legislatures and fiscal decisions. These contributions stress in particular that representatives of geographically-based constituencies overestimate the benefits of public expenditure in their district with respect to its financing costs, since these costs are typically borne not only by the legislator’s constituents, but by taxpayers as a whole. This failure to internalize costs leads to excess government spending with respect to the “optimal” level (see, for example, Weingast, Shepsle and Johansen (1981)). 1/ These models, richer in institutional detail that the ones surveyed so far, are static, and generate a bias towards excess government spending rather than excess deficits. 1/ The Appendix presents a simple model based on this idea, that incorporates a deficit bias due to the possibility that today’s government will be replaced by a government with different spending priorities.

In conclusion, these models provide several complementary explanations for the persistence of fiscal imbalances and/or for excessive levels of government expenditure. From a normative perspective, the issue is whether this bias towards budget deficits and/or excess spending can be corrected by institutional reform, or, more specifically, by appropriately designed fiscal rules. The question is addressed in the next Section.

III. Fiscal Rules

In this section, we consider the potential role of “numerical” fiscal rules in ensuring fiscal responsibility. Numerical rules are typically imposed on the maximum size of budget deficits (balanced budget amendment, Maastricht criteria etc.). As in the literature on the rules versus discretion in monetary policy formation (Kydland and Prescott (1977); Barro and Gordon (1983), the debate on the usefulness of fiscal rules focuses on a critical trade-off between the elimination of a policy bias and the need to retain policy flexibility. 2/ In principle, the ideal rule, as in the context of monetary policy, would be state-contingent, so as to allow the authorities sufficient flexibility to react to shocks while at the same time removing any inherent bias towards excess fiscal imbalances, or excess inflation. However, it is commonly agreed that rules have to be simple in order to be credible, since contingent rules leave the door open to manipulation, given a private information problem. Furthermore, it would be impossible to enumerate every possible contingency.

Corsetti and Roubini (1994) explore these issues more formally. They present a simple model that extends Alesina and Tabellini (1990) in order to highlight the trade-off between deficit bias and margin for stabilization in the context of a closed and open economy. They first consider whether leaving the government a margin for stabilization policy (modeled as a “tax smoothing” role) contributes to worsening the deficit bias. They conclude that this is not the case. An interesting result they obtain is that the political deficit bias is enhanced in an open economy. Since in their model there is no default risk, in an open economy the government faces an infinitely elastic supply of funds at the given world interest rate, contrary to the case in a closed economy. This implies that additional borrowing to finance more expenditure is not discouraged by higher interest costs. 1/ Clearly, the scope for fiscal rules depends on the relative intensity of the deficit bias and the need for tax smoothing.

The nature of the trade-off between flexibility and deficit bias is apparent in some of the empirical work on the link between statutory fiscal restraints and budgetary outcomes within US States. ACIR (1987) and von Hagen (1991), among others, find that states with such restraints run smaller budget deficits, while Bayoumi and Eichengreen (1995) find that the cyclical responsiveness of state budgets is significantly reduced by the presence of fiscal restraints.

At a more practical level, the important issue is what fiscal variable should be the object of a rule. What definition of government or public sector should be considered? Should the rule be imposed on government spending, the fiscal deficit, or public debt? With regard to the question of “coverage” of a budget rule, there are two conflicting factors at play. On the one hand, it makes sense to design a rule that establishes norms for the most comprehensive definition of public sector, because this would limit greatly the ability of fiscal authorities to use creative accounting in order to exclude certain types of spending from the rule. On the other hand, the government may lack direct controls on decentralized spending centers, the larger the public sector aggregate being considered. This is especially true in federal systems.

A second, related, set of problems is due to the characteristics of conventional measures of fiscal deficits. It is well known that nominal budget deficits (inclusive of interest payments) may be a flawed measure of the actual fiscal stance, for a number of reasons. 2/ First, they do not take into account the effects of inflation on interest payments, therefore counting anticipated debt repayment as deficit. Second, in the presence of economic growth, the debt to GDP ratio can be kept constant even if the country is running a budget deficit. Third, seigniorage revenues are not included. Fourth, conventional measures of the fiscal deficit do not correspond to changes in the government’s net worth: this implies that privatization proceeds always improve the government fiscal position by reducing public debt, because the decline in public sector assets is ignored. Finally, contingent liabilities are not explicitly accounted for in the budget. Although in principle corrections could be made to the deficit figures in order to address these problems, in practice this would leave too much scope for discretion; indeed, the literature on policy rules suggests that these should be simple and easy to understand. It appears, therefore, that the shortcomings listed above may well represent the unavoidable cost of imposing a simple fiscal rule. Interestingly, most proposals for rules have focused on the budget deficit rather than government spending, although the latter may be more appropriate in the presence of excess spending biases of the type discussed at the end of Section II.

A final important issue to consider is whether fiscal rules should be imposed on budget formation (ex ante), budget outcomes (ex post), or both. An ex ante rule could take the form of a balanced budget amendment; an ex-post rule would envisage automatic measures to be taken whenever budget outcomes differ from some pre-determined standard (that could be itself determined by an ex-ante rule). In the context of ex post rules, an important issue arises with respect to the treatment of automatic expenditure components (entitlements). If the effect of an automatic rule is, for example, to cut discretionary spending whenever budget overruns occur, there may be undesirable consequences on the composition of the budget. However, this could increase the pressure to modify automatic spending mechanisms in order to ensure their consistency with the desired fiscal stance.

IV. Budget Institutions

The approach that examines budget institutions differ from the previously examined ones because it looks inside the “black box” from which budgetary outcomes emerge. There is an extensive political science literature on the effects of budgetary procedures on fiscal policy decisions. Although this literature is mostly inspired by the institutional framework and practical experience in the United States, it contains important theoretical insights for cross-country studies of fiscal institutions. 1/

There is a sense in which the literature on budget deficits and budget institutions is complementary to the literature on the political economy of budget deficits briefly surveyed in Section II. Namely, while the latter takes institutions as given and studies the impact of political and distributional factors on budget outcomes, the former takes political incentives as given, and examines the link between different budget institutions and budgetary outcomes. As pointed out by Alesina and Perotti (1996), there are two related objections to this approach: the first is that institutions are themselves endogenous with respect to the political situation, and the second is that political factors will shape budgetary outcomes regardless of budgetary institutions. Both objections are reasonable; with respect to the first, one can argue that institutions change only slowly over time, and that they can therefore be taken as given for short- to medium-term analysis. The second objection is probably more relevant; we postpone a fuller discussion to Section VI. Nevertheless, it seems reasonable to assume that changing the “rules of the game” may alter, at least partially, the strategic interaction between the different “actors” in the budget process and therefore have an influence on outcomes.

The budget process can be characterized as a system of formal and informal rules and regulations shaping the decision-making process that leads to the formulation of the budget by the executive, its passage through the legislature and its implementation (von Hagen and Harden (1995)). 1/ In order to understand the mechanics of the budget process and the scope for a change in budgetary procedures, it is important to take all three stages into account, and in particular the set of incentives and constraints the different actors (government; legislature; bureaucracy) are subject to at each stage. In addition, the degree of transparency and the informational content of the budget play an important role at all stages of the budget process, and will be discussed separately.

1. The formulation of the budget

This stage consists of the drafting of the budget law by the government. The formulation of the budget is divided into different sub-stages: the setting of budget targets and guidelines, the presentation of budget bids, the compilation of the budget draft, its reconciliation and finalization. The participants in this phase include the prime minister, the finance minister(s) and the “spending” ministers. Analogously to the literature on spending by district within the US legislature, it seems reasonable to assume that spending ministers have an incentive to seek the control of a sufficiently large amount of resources. 2/ Given that tax revenues are typically not “earmarked” for specific expenditures, these ministers may fail to internalize fully the cost of the (present or future) taxation necessary to cover their expenditures. By contrast, the prime minister and the finance minister are more likely to internalize the costs of additional spending, and are therefore more likely to be “fiscally conservative”. The model presented in the Appendix provides a simple formalization of these considerations. The key issue is how procedures in the preparation of the budget law shape the strategic interactions between these agents.

A fundamental feature of this stage is the degree of centralization in the formation of the budget draft. Greater centralization implies that the finance minister has a more dominant role in the formation of the budget law. By contrast, in a more decentralized process “collegial” mechanisms typically guide budgetary decisions. Table 1, reproduced from von Hagen and Harden (1994), summarizes the key steps in the government stage, distinguishing between three different degrees of “centralization” of the process. According to these authors, within the European Union, France and the UK are the clearest examples of budget centralization, while the Greek, Irish and Belgian systems are the most decentralized. Italy has an intermediate degree of centralization. There are three ministers in charge of the economy: the Treasury Minister (responsible for public expenditure), the Finance Minister (responsible for tax revenue) and the Budget Minister (responsible for planning and macroeconomic projections). The budget proposal is drafted by the Minister of the Treasury, but none of these three Ministers enjoys special or formal authority in budget negotiations. Furthermore, cabinet committees that bring together expenditure ministers in order to coordinate public expenditure across departments raise the possibility of forging strategic alliances between ministers in order to raise their spending bids.

Table 1.Structure of Government Stage*
Type of Procedure
Step1. Strategically

2. Decentralize

3. decentralized
G1Budget targets and guidelinesPrime minister (PM) or finance minister (MF)Cabinet on Proposal by MFCabinet
G2Budget bidsSpending ministries
G3Compilation of draftMF, in bilateral negotiationsMF, serving as intermediary between spending ministers and cabinetMF, simple collection of bids
G4ReconciliationPM or senior cabinet committeeSenior cabinet committee or cabinetCabinet

The nature of the distortion--failure by spending ministers to appropriately internalize the cost of their expenditure--clarifies why decentralized procedures are more likely to lead to a deficit bias. What structure of budgetary procedures would reduce the bias towards excess deficits or excess spending at this stage? There are two possible answers. One is to have the overall size of the budget determined in advance by some commitment mechanism, such as a numerical target. This target could be formally enshrined in the law, for example in the form of a balanced budget amendment or constitutional limits on the size of the budget deficits (see Section III). In this case, the bargaining within the cabinet would take place in the presence of an overall ceiling on available resources, thus containing free riding behavior. The second is to give more power to the finance minister vis-à-vis the spending ministers in the bargaining process. For example, the finance minister could direct the budget procedure and have a veto power on spending decisions, or the ability to adjust spending proposals. Also, conducting bilateral negotiations with spending ministers, rather than full cabinet negotiations, may reduce the scope for “reciprocity”--that is, favoring each other’s spending programs (von Hagen (1992)). von Hagen and Harden (1994) argue that small countries have typically opted for the first type of mechanism, while large countries for the second. The reason, according to these authors, is that a large country, given the greater complexity of its budget structure, needs an additional degree of flexibility in the process of budget formulation that rigid numerical targets cannot provide. 1/

In Italy, numerical targets for the central administration budget deficit are contained in the Document of Economic and Financial Planning (DPEF), which specifies fiscal targets for the next three years. 2/ Parliament does not formally approve this document, although it approves a “resolution” that defines the key policy guidelines and the quantitative targets to be met by the Legge Finanziaria (draft budget law). In the past few years, the amount of the state sector deficit incorporated in the DPEF has been considered de facto as binding for the Legge Finanziaria (LF), in accordance with new Parliamentary procedures.

As noted in Section III, an important issue is the coverage of a numerical fiscal rule. In Italy, the coverage of the Budget Law is limited to the State Sector (Central Administration); none of the budget documents focuses directly on general government expenditure (as required, for example, by the Maastricht treaty with regard to its criteria) or on the overall public sector. This, together with accounting practices, hinders transparency and creates incentives to transfer expenditures off the State Budget, so as to meet, albeit artificially, numerical targets on the deficit. Giavazzi (1993) and Commissione Spesa Pubblica (1994) give several examples of “creative accounting” procedures used for this purpose. As noted by Onofri et al (1994), an implication of these practices has been that the level of the public debt exceeds the sum of annual budget deficits by a substantial amount.

Another aspect of the budget formation stage, whose importance is more controversial, is the existence of a multi-annual framework within which budget decisions are taken. In principle, this aspect could be important because it could force the government to internalize the intertemporal effects of its revenue and expenditure decisions. The practical relevance of this aspect is, however, debatable, for two main reasons. The first (Alesina and Perotti (1996), Alesina et al. (1995)) is that multi-period budgeting leaves ample scope for a strategic use of budget projections-- for example, the incorporation of future policy measures whose fiscal consequences are not properly evaluated. The second reason has to do with the incentives that shape the government’s intertemporal choices. Clearly, political factors play a very important role in this area: governments that are less likely to be in power in the future would probably fail to internalize fully the additional future tax burden associated with current budget deficits. These incentives are likely to overcome the incentives for a fully “responsible” intertemporal behavior that a strict multi-period budgetary framework is going to impose. Indeed, von Hagen (1992) finds that an index designed to capture the degree to which current fiscal decisions are tied to a multi-period fiscal program is not significantly correlated with various measures of fiscal imbalances, such as average budget deficits or the level of government debt.

2. The parliamentary stage

This phase consists of the passage through parliament of the budget law, up to its final approval. There are two crucial aspects of this phase. The first is the scope of the amendments: for example, whether these amendments can increase spending and/or reduce revenues. The second, more general, aspect is the strategic relation between the government and Parliament, which is linked to the modalities of voting procedures on the budget.

Why are these factors theoretically important? What are the incentives for “profligate” fiscal behavior within parliament? Like spending ministers, members of parliament may propose amendments that benefit their constituencies, failing to internalize the fiscal cost associated with these measures. These incentives may be partially mitigated by party discipline, although this factor is likely to play a more modest role in politically fragmented systems. Budgetary procedures designed to lead to more responsible fiscal behavior may therefore limit the scope of amendments that change the overall budget balance, or, even more restrictively, that increase expenditures and/or reduce revenues. Since 1990, Italy has had a rule that Parliamentary amendments cannot worsen the overall fiscal balance --so that any expenditure increase or revenue reduction must be covered by additional resources (copertura) to finance it. However, this rule is widely acknowledged to be pro forma in many cases, often thanks to very optimistic assumptions about the yield of offsetting measures.

With regard to the strategic interaction between the government and parliament, the key issue concerns the political implications of an outright rejection of the budget. If this forces the resignation of the government, two strategic elements can be at play (von Hagen and Harden (1994)). On the one side, the government may be pushed to adopt proposals that are likely to receive broad support in Parliament. If the same deficit bias exists at the parliamentary level, this will tend to weaken the overall budget stance. On the other side, however, members of Parliament belonging to the governing party(ies) will be more unlikely to vote against specific budget proposals, or to propose amendments that would lead to the defeat of the budget. This strengthens the position of the government. von Hagen and Harden argue that the second element is likely to prevail on the first, implying that a vote on confidence on the budget strengthens the relative bargaining position of the government vis-à-vis parliament, thus facilitating the adoption of “sound” budget policies.

Table 2, which is drawn from von Hagen and Harden (1994), illustrates the different steps in the Parliamentary stage of the budget process, and identifies three different types of procedures, classified from restrictive to open. Among countries of the European Union, France and the UK have a system which is particularly restrictive with regard to the scope of the amendments that can be proposed. In contrast, the Belgian system has no limit on amendments, and amendments can force the government’s resignation only under very special circumstances (von Hagen and Harden (1994)). With regard to Italy, it should be noted that the number of amendments being proposed can be--and has been--staggering, thus imposing de facto constraints on the government at the budget preparation stage. Even though Parliament may not have directly adopted amendments worsening the overall fiscal balance, the need for compromise may have pushed the government to draft “weak” budgets in order to pre-empt Parliamentary opposition--particularly when the Government’s support in Parliament was weak and fragmented. In part because of the complexity of the interaction between government and Parliament, there remains disagreement on whether Parliament has played an important role in shaping fiscal policy decisions in Italy, in particular over the last few years. For example, Onofri, Pisauro and Siniscalco (1994) argue that the spending ministers have the main responsibility for fiscal profligacy while the Commissione di Riforma dei Bilanci Pubblici (1994) fingers Parliament as the main culprit.

Table 2.Structure of Parliamentary Stage*
Type of Procedure
Scope of amendmentsamendments cannot increase spending or reduce revenues, or certain amendments are not receivableamendments cannot change overall balanceno limits on amendments
Relation of upper and lower houseupper house has no budgetary powerslower house has prerogative over upper houselower and upper house have equal rights
Relation of government and parliamentGovernment can call vote of confidence, can impose voting procedure on parliament; amendments require consentAmendments may cause fall of governmentno special stipulations

3. The implementation stage

This stage refers to the execution of the budget law. There are two aspects of particular importance: how binding is the budget law, and how much flexibility is there to face unforeseen contingencies. Examples of the first are the power of the finance minister to restrain spending beyond the budgeted amounts, and the control of open-ended expenditure such as unemployment compensation and pensions. The degree of flexibility is related to the possibility of introducing supplementary budgets during the fiscal year, the ability to transfer resources between budget chapters, the existence of a budget reserve, and the possibility of setting aside unused funds for future expenditure (carryover). According to these criteria, Italy’s budget law is only weakly binding and the budget implementation system is inordinately flexible. There are no cash limits on spending ministries, the budget ministers lack the power to block expenditures, budget changes and transfers of expenditure between chapters are routinely adopted, and carryover procedures are extremely flexible.

A binding budget law clearly constitutes a commitment to budget discipline, because it imposes a hard budget constraint on spending centers. In this regard, practices that enhance budget flexibility, such as the ability to transfer resources between budget chapters or to carry over unspent funds to the next budget year, weaken budget constraints and hinder budget transparency by creating wedges between ex-ante appropriations and ex-post outcomes. Finally, the possibility of introducing substantive budget revisions, such as a supplementary budget, can also weaken fiscal discipline at the implementation stage. More importantly, this possibility also alters incentives at previous stages of the budget process, because it reduces the degree of commitment implicit in budget decisions. In practice, however, there can be discrepancies between ex-ante budget targets and ex-post outcomes because of open-ended expenditures or because of the effects of the cycle on tax revenues. In the presence of a rigid numerical target, it is important to consider what, if any, are the ex-post mechanisms designed to ensure its achievement--for example, automatic spending cuts and/or tax increases. 1/

Table 3, also reproduced from von Hagen and Harden (1994), exemplifies three different types of procedures at the implementation stage, classified according to their degree of restrictiveness. According to the index of flexibility developed by von Hagen (1992), France and Germany are the EU countries with the most restrictive budget implementation procedures, while, as already noted, Italy is the country with the most flexible ones.

Table 3.Characteristics of the Implementation Stage*
Type of Procedure
Expenditure managementDisbursement approval required, spending departments subject to cash limits, MF can block expendituresDisbursement approval required, and/or spending departments subject to cash limitsDisbursement approval required or full authority of spending departments.
Transfers of appropriationsWithin chapters onlyWithin chapters unrestricted, between chapters upon approval by MFUnrestricted
Substantive revisionsBy new law and rarely usedBy new law, commonly usedBy approval of MF

A more general issue that is germane to this stage but beyond the scope of this paper is the reform of public administration and of the system of expenditure management. Among the topics that have been debated one can cite the establishment of cost centers within ministries, the enhancement of managerial responsibility, increased emphasis on output rather than inputs in the provision of public services, etc. These reforms reflect a more general need to shift from a system of expenditure management and provision of public services based on a strict legal and procedural framework to one which relies heavily on incentives (Tanzi (1994b)). A number of countries have undertaken reform effort in this direction, most notably New Zealand. 2/

4. Information content and transparency of the budget

These aspects are particularly important, not only for accountability and ex-post control, but also for the incentives and constraints at play during the budget process. Indeed, public choice-type theories of fiscal deficits, based on fiscal illusion, emphasize that lack of transparency is used by politicians as a tool to hide from taxpayers the true costs and benefits of tax and spending decisions. Similarly, theories based on rational voter behavior emphasize the incentives of the government to use informational advantages strategically vis-à-vis voters, which more obscure budgetary procedures can enhance. 3/ Factors that reduce the informativeness of the budget process are the frequent use of off-budgetary funds, the omission from the budget of government loans and guarantees, the presentation of the budget in multiple formats, strategic use of budget projections, and, more generally, lack of transparency in areas such as expenditure administration and sources of revenue. When the budget process is informative, it becomes easier for the finance minister to impose hard budget constraints on spending ministers, for the Parliament to monitor the actions of the government, and for the government to ensure that budget decisions are properly implemented. Furthermore, public accountability for the actions of both government and Parliament is increased.

Among the countries of the European Union, von Hagen (1992) finds that Germany, France, and the UK have the most transparent budgets, while Belgium, Ireland, Portugal and especially Italy have the least transparent budgets, with extensive use of off-budgetary funds in the past. Alesina and Perotti (1995b) and Alesina, Marè and Perotti (1995) underline the potential for strategic manipulation of multi-year planning, associated with the use of the so-called bilancio tendenziale (no-measures budget projection) as a benchmark for measuring fiscal policy corrections. Given the ample recourse to one-off revenue measures in the budget (whose revenues are by definition excluded from the no-measures budget projection) and the strategic use of macroeconomic projections, the amount of actual fiscal adjustment is typically overstated. An additional factor of confusion is generated by the needlessly complicated budget preparation process, involving several stages that cover different definitions of government and different types of budget (cash and accrual basis). Even the approval of the Budget Law involves a multiplicity of documents: Legge Finanziaria (LF), Provvedimenti Collegati (PC), Nota di Variazioni (NV) and the Budget Law itself. 1/

5. The empirical evidence

But what is the empirical evidence on the link between budgetary procedures and budgetary outcomes? Two approaches are possible. The first, more descriptive in nature, consists in relying on case studies. 1/ The second approach consists in statistical analysis involving the construction of numerical indices summarizing the degree of centralization and restrictiveness of the various stages of the budget process, as well as its overall degree of transparency. Clearly, empirical analysis is plagued by a series of problems, starting with the difficulty of converting the arguments of the previous sections into an index that summarizes the most important cross-country differences. Furthermore, the power of statistical tests is very low given the limited number of observations.

von Hagen (1992) and von Hagen and Harden (1994) use both nonparametric tests and regression analysis to test whether a “structural index” of the budget process is correlated with a number of fiscal variables, such as the debt-to-GDP ratio, net lending, and government expenditure in European Union countries. The index is constructed on the basis of several indicators of centralization in budgetary procedures and transparency in the budget. 2/ They find a significant correlation of the coefficient with net lending and debt, but a weaker one with government expenditure. This latter finding is somewhat disturbing, given that a bias towards excess public expenditure plays an important role in the theoretical arguments in favor of changes in budget rules. A shortcoming of their empirical analysis is the lack of control for other economic and political factors that researchers have found to be correlated with fiscal performance.

de Haan and Sturm (1994) provide some preliminary evidence that the correlation between changes in the public debt to GDP ratio and budgetary procedures is robust to the introduction in the regression analysis of political and economic variables. They use pooled cross section/time series data for the period 1981–89 and an index for the centralization of the budget process based on von Hagen (1992). As more countries change their budgetary procedures, it would be interesting to test whether changes in budgetary institutions within countries are correlated with changes in fiscal performance. Clearly there would be a causality problem in interpreting such a correlation, since the changes in budgetary institutions may reflect a general shift towards more fiscal policy restraint, or a response to a fiscal crisis.

Alesina, Hausmann, Hommes and Stein (1995) examine the link between budget procedures and budget outcomes for a sample of Latin American countries, using information obtained by a survey on budget legislation and actual budget practices. Their results are consistent with those of von Hagen. In particular, they find that the existence of a binding macroeconomic program voted before the budget discussion, and limits to the power of the legislature to amend the budget have a significant influence on fiscal outcomes.

Another strand of empirical work has examined the relation between budgetary procedures and outcomes in US States (see Poterba (1995) for a survey). Poterba (1994) examines the dynamics of state taxes and expenditures during the late eighties. He finds that States with more restrictive state fiscal institutions undertake more rapid fiscal adjustment in response to unexpected deficits. He also finds that the response is more rapid when the same party controls the governorship and the state house, consistently with the theoretical presumption of the literature on divided government. Similar results are found by Alt and Lowry (1994) and by von Hagen (1991). 1/ Finally, Bohn and Inman (1996) find that end-of-year (ie, ex-post) balanced budget requirements are more effective than beginning-of-the year (ex-ante) constraints in limiting deficit financing.

V. Reform Proposals for Italy: A Synthesis

The need for fiscal adjustment measures has stimulated a debate on reforms to the budget process that increase transparency and facilitate the adoption of “fiscally responsible” measures. At the time of writing, however, no proposal for a reform of the budget process had been submitted to Parliament. With one exception, we briefly review here proposals that focus specifically on the Italian situation. These proposals can be broadly divided into two groups: a first one (sections V.1 to V.5) suggesting modification of budgetary procedures (changes to the “rules of the game”) and a second one (sections V.6 and V.7) favoring the imposition of rules for outcomes.

1. The proposal of the commissione per la Riforma dei Bilanci Pubblici (CRBP)

The proposal of the Commissione di Studio per la Riforma dei Bilanci Pubblici (Commissione Giannini-Finocchiaro), concerns the budget document and all phases of the budget process, with particular emphasis on auditing and ex-post control. It stresses the responsibility of Parliament for the persistence of fiscal imbalances in Italy, and accordingly envisages a system in which the role of Parliament in the budgetary process is substantially reduced.

The report notes that the budget document in its current form is needlessly detailed and obscure. It proposes a functional reclassification of expenditures, with sub-categories within each broad function. For each sub-category, the budget would indicate the amount of expenditure that is to take place only if sufficient resources are available (that is, depending on actual revenue and expenditure developments during the year). Parliament would vote on this document, while the usual administrative budget with the more detailed sub-division into chapters would be attached to the new budget document, but not voted on.

With respect to the government stage, the proposal suggests a reduction of the number of ministers to nine, with responsibilities for the budget concentrated in a single ministry, the Treasury. It does not address, however, the formation of budget proposals within the government or the relative bargaining power of ministers. The proposed changes to the parliamentary stage are much more drastic: the proposal would exclude any amendment that alters expenditure and/or taxes (including balance-neutral changes) whenever the government is opposed to them.

The changes in the budget implementation stage are associated with an overall reform of the State General Accounting Office (Ragioneria dello Stato), which would become more independent and at the same time more powerful, with its role extended to the ex-ante drafting of the budget. The General Accountant (Ragioniere Generale) would participate the budget formation process, albeit without formal voting powers. The Treasury would retain control over the behavior of spending ministries through an office within each ministry, headed by a “financial controller”. The role of this controller would be to ensure that new measures and projects within spending units conform with the budget and with management objectives, as well as to verify estimates of their total costs. Finally, the proposal envisages a detailed system of certification and controls on the budget of local authorities, undertaken by a new category of certified public accountants. The purpose of the proposal is to maintain a degree of central control over the behavior of local authorities.

2. The proposal of the Commissione della Spesa Pubblica (CSP)

This proposal, described in Onofri, Pisauro and Siniscalco (1994), has three main components: tighter constraints on expenditure decisions, modification of the political responsibility for the budget within the government and an increase in the overall degree of transparency. This would be accomplished through constitutional reforms, aimed at imposing a harder budget constraint on the government, together with reforms in budgetary procedures.

With respect to the budget document, the report proposes recording all expenditure authorizations in the budget of the ministry that is politically responsible for them, and including off-budget items, such as transfers to the social security agency (INPS) through the Unified Treasury Account in the budget, in order to reduce the scope for creative accounting. The report also proposes strengthening the supervisory role of the Audit Court (Corte dei Conti) by allowing it to intervene when budget laws are being discussed in Parliament.

At the government level, this proposal suggests the unification of the three existing budget ministries into a single Finance ministry. The Prime Minister’s position in the budget process should be strengthened so as to improve control over spending decisions. The proposal also notes the possibility of strengthening the role of the Finance Minister in the budget formation process.

At the parliamentary stage, the norms that restrict the presentation of amendments to those that do not worsen the overall budget balance should be strengthened, possibly with a constitutional law. Amendments to the budget law presented by the Government should receive prior authorization from the Prime Minister. The proposal also considers the possibility of excluding amendments to the budget law; it views this as a useful measure only insofar as it is accompanied by a strengthening of the relative position of the new Finance Minister in the budget formation process, since otherwise political pressures would merely be shifted to the budget drafting stage. The Finance Minister could also be given veto power over spending laws. The powers of Parliament with respect to government budget actions (such as, for example, the estimates in the Bilancio Tendenziale) should be strengthened, both ex ante and ex post.

With respect to budget implementation, it is noted that the Ragioneria has played an important role in deciding ex-post budget allocations. The systematic underprovision for the needs of INPS in the budget has been addressed by the Ragioneria through “advances” to INPS, financed by delaying expenditure authorizations for local authorities. The proposal suggests an increased role for the budget on a cash basis; more specifically, cash limits should be applied to budget allocations, and off-budget items (the unified Treasury Account, that currently handles the allocation of budget resources to government agencies and local authorities, and INPS) should be incorporated in the budget. This would bring the role of the Ragioneria back to its original mandate (handling of cash flows) and avoid the formation of “hidden debt”. In the budget, spending authorizations should be attributed to the actual expenditure centers (ministries), rather than, as now, being recorded in the Treasury’s budget.

3. The IMF report on expenditure control

The main features of this report are summarized in IMF (1995). The general recommendations are as follows. First, there should be an explicit goal for the level of public expenditure, rather than an exclusive focus on the budget balance. Second, there is a need for greater transparency in the management of public spending. Third, more weight should be attached to improving value-for-money and accountability. There are also more specific recommendations for different stages of the budget process. With regard to the budget drafting stage, the proposal suggests setting a ceiling in cash terms for general government expenditure, exclusive of interest, for the budget year. Bilateral discussions between each spending Ministry and the Treasury should follow, with the debate conducted in terms of shares within a given total level of public spending. The report also suggests to discontinue the practice of carry-overs, with the exception of some capital programs. There is a need strictly to enforce the existing requirement that no spending proposals can be put forward unless offsetting expenditure and revenue measures are also identified. Finally, there should be a contingency reserve with strict limits on its operations. With regard to the implementation phase, the proposal argues for a split of the cash target for total noninterest general government expenditure into three components: central government, including transfers to local governments and state agencies; a small contingency reserve; and residual noncentral government spending, including spending financed by local authorities and state agencies from their own resources. The report also refers to explicit cash limits to ensure that budget appropriations for each year are not exceeded, and proposes a reform in the operation of the Treasury account, through which all transactions of local governments and state agencies should be channelled. A number of other proposals look in more detail to the issue of expenditure monitoring and the role of cost centers. Finally, the report puts forward proposals to improve control and accountability, which are outside the scope of this paper.

4. The proposals in Alesina, Marè and Perotti (1995)

This paper puts forward a comprehensive series of proposals for a reform of the Italian budgetary process. The proposals address the following areas: (i) reform of the budget law and budget transparency; (ii) the government and the budget process; (iii) the legislative process; and (iv) budget implementation, control and managerial responsibility.

The authors propose a drastic simplification of budget documents. Existing legislation complicates the conduct of fiscal policy because it forbids the inclusion in the budget law or the LF of necessary legislative measures concerning taxation and public spending (see Section II.4). The authors propose to unify the four budget documents that are currently drafted into a single budget document. This would require a constitutional reform as well as legislative reform, to allow the government to introduce new taxes and expenditures in the budget law. The authors propose to replace the “Disegno di Legge di Bilancio a legislazione invariata” and the DPEF with two no-measures budget projections, one for unchanged legislation and the other for unchanged policies. These documents should have the same format of the budget law, and should be based on macroeconomic projections from “independent” sources, such as international organizations. Longer-term projections should be discontinued, at least in the context of this budget document.

With respect to budget transparency, the authors suggest enlarging the definition of government in the budget law to the public sector, thus reducing the scope for creative accounting. Also, the effects of budget measures should be calculated by an independent agency or at least certified by one. Finally, all liabilities of the public sector should be accounted for “above the line”. The budget should be presented on both an accrual basis and a cash basis (rather than only the former, as at present); it should be based on “cost centers”, and the number of chapters (over 4,000 in the 1995 budget) should be drastically reduced. Carryovers for current expenditures should be abolished, and transfers of resources between expenditure programs should be substantially reduced or eliminated.

At the government level, the authors propose to increase centralization by consolidating the three economics ministers into a single one, and strengthening the position of the Finance Minister. More specifically, the Finance Minister would decide the level of expenditure and the budget balance, and these could not be modified during budget negotiations. These negotiations would have to take place on a bilateral basis, and the final budget composition would be voted by majority rule, with veto power for the Finance Minister.

At the Parliamentary phase, the proposal suggests an initial vote on the budget size and balance, with a two-thirds majority required for voting down the size of the budget, with defeat for the budget proposal triggering the government’s resignation. Amendments that increase expenditures or the deficit would be disallowed.

The authors do not address in detail reforms of the budget implementation stage. They note, however, the need for a general reform of public administration, based on the design of appropriate incentives and constraints to induce “efficient” behavior consistent with expenditure control and an emphasis on output (as has been done, for example, in New Zealand--see Scott (1995)). This “ex ante” incentive structure is deemed superior to a web of regulations and legal restrictions.

5. The proposal of von Hagen and Harden (1994)

This proposal was not directed specifically to Italy, but rather to the countries of the European Union. Its salient feature is the establishment of a “National Debt Board” (NDB) that would be in charge of determining each year the “permissible” change in the level of the national public debt. The NDB would be an institution “independent from government but with democratic accountability” (page 316). Once the change in the public debt is determined, the budget process should have either a numerical target on spending or “strategic dominance” of the finance minister. If the NDB were not to be established, the authors argue that numerical targets should be determined within a multi-annual framework formally announced by the government.

The proposal is clearly inspired by the literature on central bank independence. Delegation of the decision on the permissible change to the public debt to an independent agency would, in the intention of the proponents, eliminate a deficit bias without eliminating the margin of flexibility required to respond to economic conditions. At the same time, the government would retain control over the overall size of the budget. Regardless of the feasibility of the NDB, the suggested reforms of the budget process that accompany the proposal (limited amendments, votes on the budget chapter by chapter, cash limits and limits on supplementary appropriations, among others) are clearly inspired by the factors described in Section IV. The two alternative routes being suggested are based on strict numerical targets within a more decentralized structure, or on a more centralized structure with more power to the finance minister.

6. The Masera proposal

Rainer Masera, Budget Minister from February 1995 through January 1996, proposed budget reforms in order to ensure adherence to Maastricht-type fiscal criteria during the so-called “third phase” of European Monetary Union (Masera (1995a, b)). 1/ He focussed on three aspects of budgetary procedures: budget documents, reforms in the Parliamentary stage of the budget process, and constitutional rules that enshrine the limitation of fiscal autonomy implied by the European Union treaties. With regard to budget documents, Masera (1995a, b) notes the lack of transparency of the Italian budget process, and the excessive disaggregation of expenditure items in the budget. He proposes a re-organization of the budget document, either by expenditure center or by function, thus facilitating the use of “zero-base budgeting”. With regard to the Parliamentary stage of the budget process, Masera argues that it is necessary to re-establish the principle that the government has formal decision powers with regard to economic policy, while Parliament has the power to reject the government’s choices, and thus force a government resignation. More specifically, he proposes to limit the scope of amendments to the budget law that can be presented in Parliament by forbidding amendments that increase expenditure. This limitation could be accompanied by other legislative changes establishing that current expenditures cannot rise faster than GDP or that the current account balance has to be non-negative. Finally, Masera argues that it is necessary to enshrine Maastricht-type criteria in the constitution to underscore the irreversibility of the third phase of EMU. This necessity derives from the fact that, while before the beginning of Phase III countries that fail to meet the Maastricht criteria are clearly penalized (through exclusion from EMU), enforcement mechanisms to ensure that such criteria are respected after monetary union has taken place are considerably weaker. Constitutional changes along these lines would be part of a process of harmonization of budgetary rules in countries belonging to the European Union.

7. The Monti proposal

This proposal has the objective of ensuring that the Maastricht criterion on the size of the budget deficit will be met, and is driven by the observation that budget outcomes in Italy have systematically been worse than planned (1995 being an exception). Monti therefore proposes an ex-post rule designed to ensure that deviations from budgetary targets imply an automatic fiscal correction. Namely, he envisages an automatic increase of the income tax should the budget deficit exceed the planned level. Other proposals along the same lines have instead suggested reliance on automatic expenditure cuts rather than tax increases as a response to budgetary overruns. Clearly, such a mechanism could also be the complement of an ex-ante budget rule, such as a balanced budget amendment or a rule establishing limits on the maximum size of the budget deficit, although the original proposal was not cast in these terms.

VI. Discussion

The general conclusion that can be reached from the discussion of Sections III and Section IV is that reforms of the budget process are likely to be successful in enhancing fiscal discipline only insofar as they can ensure responsible behavior at all stages of the budget process. We therefore compare the modifications to the different stages of the budget process that the proposals outlined in the previous section envisage. 1/

All proposals recognize the need for a change at the government stage of the budget formation. Several proposals (Alesina et al, CSP, CRBP) suggest the unification of the three existing budget ministries (Treasury, Finance and Budget). The first two also underscore the need to strengthen the position of the Finance Minister in the budget negotiation process. IMF (1995) and Alesina et al. (1995) also suggest bilateral negotiations between the Finance Minister and spending ministers within the framework of strict numerical targets on government spending. Finally, Masera (1995a) envisages a constitutional rule as a way to ensure a fiscally responsible budget.

With respect to the Parliamentary stage, a number of proposals suggest the strengthening of the rule establishing that amendments cannot worsen the budget balance (CSP, IMF). Other proposals envisage even stricter limitations on amendments, such as the prohibition of any amendment that increases expenditure (Alesina et al.; Masera; CRBP). Overall, the proposed reforms are broadly in line with those discussed in Section IV and with von Hagen and Harden (1994).

With respect to the implementation of the budget, IMF (1995) and Alesina et al. (1995) suggest the elimination of carryovers and the imposition of strict cash limits on expenditure centers. The latter element is also mentioned in CSP, alongside the elimination of off-budgetary items. The CRBP proposal puts more emphasis on controls, and suggests an increase in the power of the Ragioneria. CRBP and CSP contain different suggestions regarding a re-organization and reduction of budget chapters, while Alesina et al. discuss a more ambitious reform and drastic reduction of the entire set of budget documents.

The Monti proposal, envisaging automatic tax increases in the case of expenditure overruns, fits less clearly in one of the stages of the budget process. In evaluating this proposal, one must consider the incentives that it sets in place. In principle, a mechanical rule of this kind is more effective in creating incentives to avoid budgetary overruns, the more distortionary the source of taxation. To some degree, it could be argued that the income tax in Italy fulfills at least part of this criterion, given both its relatively high rates and the fact that it falls more heavily on certain types of taxpayers (dependent workers) than others. This would create a lobby in favor of respecting the budgetary targets. On the negative side, it is clearly possible to have budgetary overruns for exogenous reasons, such as, for example, an unexpected slowdown in economic activity; under such circumstances, the rule may prevent automatic stabilizers from working, thus exacerbating cyclical fluctuations. In addition to the lack of flexibility already mentioned in Section III, a more general issue concerning automatic rules is the likelihood that their effectiveness will be undermined by creative accounting. At a minimum, this suggests that a necessary condition for the effectiveness of automatic mechanisms of fiscal adjustment is an increase in the transparency of the budget that would reduce the scope for accounting manipulations. 1/

In summary, two general approaches to the problem of budget reform have been put forward. The first consists in strengthening the degree of centralization in the budget process, increasing the power of the finance minister both at the budget formation level and at the implementation level, while limiting the scope for amendments at the Parliamentary level. The second, complementary approach highlights the need for (constitutional) budget rules, that constrain ex-ante the behavior of spending ministers at the budget formation process and of parliament during its examination of the budget document.

The discussion of budget rules and procedures leaves a number of issues open. An important one concerns the role of automatic spending mechanisms (variously described as open-ended expenditure or entitlements). Alesina and Perotti (1995b) document how the divergent trend in public expenditure in OECD countries can be primarily attributed to the dynamics of transfers. They also show that successful fiscal adjustments have relied mainly on cuts in open-ended expenditure, while unsuccessful ones typically relied on tax increases. It is not clear how a change in budgetary procedures per se would affect the dynamics of this type of expenditure, given the limited amount of discretion these expenditure decisions involve. Indeed, they are often a legacy of past decisions taken when the growth rate of the economy was higher and the demographic trends substantially different. von Hagen and Harden (1994) find that the share of open-ended expenditure in GDP is correlated with their structural index and also with the long-term constraint index. While these findings suggest that a well-executed reform of the budget process may lead to fiscal measures that better incorporate intertemporal constraints, the issue is how to modify the automatic mechanisms that are already in place. This is clearly a government decision: a reform of the budget process can at most make the problems more transparent and streamline the decision mechanism that should lead to the adoption of corrective measures.

The discussion above is simply a reflection of the more general question raised at the beginning of section IV: namely, whether budget outcomes will reflect political decisions, regardless of rules and procedures. It is clear that budget rules and changes in budgetary procedures cannot deliver more responsible fiscal policy decisions per se, regardless of political will. This can be seen most easily by considering the proposals for strengthening the position of the finance minister: if that minister and the government are not committed to fiscal discipline, this change in budgetary procedures would not lead to better fiscal policy outcomes. Fiscal rules and reforms of the budget process can strengthen the relative position of those actors that are more likely to aim at fiscal discipline, provided such aim is there, and they can enhance the budget’s clarity and transparency, so as to increase political accountability for profligate fiscal behavior. Insofar as this increase in transparency increases awareness of the sources of fiscal problems, it can help to improve the quality of fiscal policy decisions, because it reduces the scope for using more hidden and costly ways to spend and tax. Furthermore, it could help build political support for altering automatic spending mechanisms that are ultimately unsustainable.

A related issue is the existence of political support for institutional reform. As can be seen from the discussion in Section II and IV, the inefficiencies existing in the budget process can play a strategic role by affecting the payoffs of the various participants in the process. 1/ Proposals for reforms in the budget process are therefore likely to encounter resistance from vested interests--and this may account for the lack of action in Italy on any of these proposals so far. Increases in the transparency of the budget may also encounter opposition insofar as they would involve disclosure of hidden debt and expenditures, thereby worsening the officially reported state of public finances. Such incentive to resist greater transparency is particularly strong in the runup to EMU, and could also be heightened if constitutionally mandated limits on budgetary imbalances were put in place.

VII. Concluding Remarks

This paper has presented theory and evidence on the relation between the budget process and fiscal policy outcomes. It has focused in particular on incentives and constraints that shape the behavior of various actors at different stages of the budget process, and discussed the scope for fiscal rules as a way to ensure desirable budget outcomes.

With regard to the Italian case, a reform of the budget process would be difficult to achieve without major legislative changes, possibly at the constitutional level. These changes need to reflect the fact that at present the budget process in Italy lacks transparency at all its stages and suffers from excessive reliance on formal rules, regulations and controls, with scant attention to the role of incentives and to the importance of clear accountability. The current problems plaguing the Italian budget process suggest that the simple reliance on a constitutional rule on budget deficit targets will have limited effectiveness; it would introduce an additional element of legal rigidity and create further incentives for creative accounting, unless it is accompanied by an overall reform of the budget process along the lines discussed. Ensuring that the government’s budget proposal will reflect prudent fiscal behavior is not particularly useful if Parliament can modify the size of the budget with relative ease, or if at the implementation stage the discretionary margins are such that systematic divergences emerge between ex-ante budgeted expenditures and ex-post actual ones. Lack of transparency in the budget process reduces accountability, makes it more difficult to monitor and control spending and tax decisions, and reduces the significance of numerical rules.

APPENDIX A Simple Model of Excess Spending and Deficit Bias

We present here a simple model, similar to the one in von Hagen and Harden (1995), that captures some of the key distortions in fiscal policy formation that can lead to excess spending and/or the emergence of fiscal deficits. The model focuses exclusively on the public sector, although it could be incorporated into a more general framework that explicitly considers private sector decisions.

Government spending decisions are “decentralized” at the level of n individual spending ministers. Each spending minister is assumed to have a “target” level of expenditure g*, representing some form of “ex ante optimality”. In addition, however, spending ministers are assumed to derive utility from the overall size of the budget at their disposal. Taxes T are distortionary, but each spending minister cares only about the distortionary costs that are borne by its “constituency”. The welfare function for an individual minister is given by:

the first term represents the cost of deviating from the “optimal” level of spending, the second represents the benefits from a larger budget and the third the distortionary costs of taxation. The “social” loss function is given by:

where m is Σmi is the measure of overall tax distortions. Suppose that total public expenditure is determined in a decentralized fashion, with every spending minister obtaining the requested resources. Assuming for simplicity that all spending ministers are of the same size, we obtain:

where the superscript “D” stands for decentralized. A benevolent social planner, that internalizes the distortionary effects of taxation, would instead choose a budget of size:

where SP stands for social planner. Comparing gSP with gD, it can be seen that there are two sources of excess spending. The first is due to the decentralized decision process, in which individual spending centers fail to fully internalize the effects of the tax burden necessary to finance their spending (m instead than nm in the denominator); the second is due to the utility each minister gets from a larger budget, γ.

Consider now a two-period version of the model. The first period is analogous to the one considered so far. in the second, there is a positive probability 1 - p that each minister (or the government) will be replaced. It is further assumed that the spending preferences of the new minister (government) can differ from the old, so that no utility from spending is derived by the first period minister in the second period if replaced. In this case, each minister’s utility function takes the form:

subject to the intertemporal budget constraint:

where R is the interest factor (equal to 1+r). Also, assume for simplicity that the interest rate and the rate of discount coincide, so that βR = 1. In this case, solving the model for the decentralized solution yields the following amount of first-period expenditure:

and the budget deficit

The deficit bias is caused by the fact that current ministers would like to tilt the government expenditure profile towards today (since they may not be in power tomorrow), while taxes will be smoothed over the two periods. The deficit bias is larger, the smaller the probability p that the ministers (government) will be in power tomorrow.


I am indebted to Alberto Alesina, Alessandro Leipold, Tim Lane and Nouriel Roubini for very useful discussions, suggestions and comments on earlier drafts. The opinions expressed are those of the author and do not necessarily reflect those of the International Monetary Fund.

Throughout the paper we use the term “finance minister” to denote the minister(s) in charge of overall budgetary policy.

These models can be applied to the determination of public spending within a government, where each “spending” minister fails to fully internalize the costs that more spending imposes on society at large.

Chari and Cole (1993) show that a dynamic version of this type of model can give rise to a deficit bias, in the presence of political uncertainty. The reason is that accumulated government debt reduces the bias towards excess government spending, in a fashion similar to Persson and Svensson (1989) and Alesina and Tabellini (1990). More generally, however, there are serious difficulties in constructing dynamic political economy models with forward-looking agents that are tractable and yet incorporate essential features of fiscal policy decisions. The main difficulty is due to the strategic interactions in voting behavior, since today’s vote determines the “state of the world” and therefore affects the incentives for future votes.

In the literature on rules versus discretion in the context of monetary policy, there is an inflation bias which results from a credibility problem in the relation between the policymaker and the private sector. In the case of fiscal rules, instead, the relevant bias is determined by political factors and does not reflect problems of time consistency.

This point is often considered in the context of the Maastricht debate. It is argued that the move towards a common currency would lower borrowing costs for high-debt countries by removing inflation and exchange rate risk premia on interest rates.

For a thorough discussion of these issues, see Blejer and Cheasty (1991).

For a brief discussion of this literature, see Alesina and Perotti (1996).

In addition to the three stages of the budget process mentioned above, there is a fourth one, ex-post control, which is beyond the scope of this paper. For a discussion with reference to the Italian case, see Tanzi (1994b).

This can occur either to increase personal prestige, or to seek a more important role for their ministry.

An example provided of a numerical target adopted in a large country that failed to live up to its purpose is allegedly the Gramm-Rudman-Hollings act in the United States.

This document contains two set of fiscal projections: the first is a baseline projection for the central administration in cash terms, under the assumption of no policy change. The second is the program projection which specifies and explicitly incorporates the policy changes to be implemented.

See the discussion of the Monti proposal in Section V. Clearly, these mechanisms are an important component of the credibility of an ex-ante fiscal rule.

For a description of the New Zealand experience with public sector reform, see Scott (1995).

See, for example, Cukierman and Meltzer (1986), Alesina and Cukierman (1990) and Rogoff (1990).

The Legge Finanziaria includes annual allocations for multi-year expenditures and specific expenditures to be authorized by new legislation; it can also incorporate new revenue-raising measures. However, it cannot contain structural changes to the tax or expenditure system: these have to be put forward in the so-called “Provvedimenti Collegati” (PC) which accompany the LF. The LF is necessary because according to the Italian Constitution the Budget Law itself has to be based on existing legislation, and cannot therefore include new taxes or new expenditure measures. Before its approval, it is then necessary to introduce the necessary legislative changes to implement the government’s budgetary objectives, and this is achieved with the LF and the PC. The changes to the existing law so introduced are then included in the Budget Law through the “Nota di Variazioni”. See also the discussion of budget documents and the scope for creative accounting in Section IV.1.

See von Hagen and Harden (1994), Appendix C, and other contributions in the same issue of European Economy.

See Tables 13 and especially von Hagen (1992) for details on the index. In the regression analysis pooled cross-section/time series data are used in order to increase the degrees of freedom. The nature of the index of budget institutions (a fixed number for every country) clearly makes it impossible to explain the time series behavior of fiscal variables.

See also the discussion in Section III.

Fresa (1995a, b) and Bilancia (1995) discuss legal aspects related to the Maastricht criteria.

It should be recognized that some of the proposals were not formulated to address an overall reform of the budget process. For example, IMF (1995) focuses on public expenditure management, while the Monti proposal is limited to ensuring no deviations between ex-ante budget targets and ex-post outcomes.

In US states, balanced budget requirements apply to the general fund only, leaving out other funds (capital, pensions, social insurance). Bohn and Inman (1996) find little evidence that balanced budget constraints tend to shift deficits to other fiscal accounts.

Milesi-Ferretti and Spolaore (1994) present a model where inefficiencies in tax collection can be used strategically for electoral purposes.

Other Resources Citing This Publication