Abstract
This paper reviews the trade-offs in Switzerland, focusing on challenges for fiscal policy coordination. It reviews the benefits and costs of a highly decentralized government, describes the Swiss institutional architecture, and analyzes Switzerland’s fiscal performance. It also discusses the specific policy challenges related to population aging, reviews the Swiss National Bank works on government financial assets and liabilities, describes the Swiss, Dutch, and the U.K. pension systems, respectively, on the regulation and supervision of the occupational pension pillar, recent reforms, and policy implications.
I. Coordinating Fiscal Policy in Switzerland: Issue, International Experience and Prospects1
A. Introduction
1. The redistribution of policy responsibilities from national authorities to other levels of government has attracted considerable attention in recent years. In a number of European Union member states, important prerogatives have been consolidated in supranational bodies or devolved to subnational entities (e.g. Belgium, France, and Italy). Decentralization is keenly debated in traditionally centralized countries, whereas reforms of intergovernmental fiscal relations have been envisaged or implemented in already highly decentralized countries, such as Germany, Spain, and Switzerland. Although the extent of fiscal decentralization typically reflects political considerations rooted in a country’s history, geography, and demography, economic incentives also play a central role, suggesting that intergovernmental fiscal relations need to respond to changes in the economic environment.
2. The economic literature traditionally contends that decentralization favors a better match between the supply and the demand for public goods and services. One reason is that citizens’ preferences and needs have a regional dimension that local policymakers are better able to observe. Another is that the proximity between policymakers and constituents enhances democratic accountability, and reduces the risk of discrepancy between policymakers’ objectives and incentives, and the interests of the community.
3. Yet decentralization also faces limits. First, it is difficult to design a clear-cut assignment of tasks between different decision levels, and costly overlaps often result. Second, coordination problems arise when choices by one jurisdiction affect citizens in another, or undermine the efficiency of national policies as regards macroeconomic stabilization, income redistribution, and long-term fiscal sustainability. Third, fragmentation makes it more difficult to realize economies of scale in the production of certain public goods and services. Fourth, issues of equity come up when subnational entities have different financial capacities.
4. This paper looks at these trade-offs in Switzerland, focusing on challenges for fiscal policy coordination, and on possible options to address them. Section B reviews the benefits and costs of highly decentralized governments. Coordination issues receive particular attention. Section C describes the Swiss institutional architecture, and provides a cross-country analysis of Switzerland’s fiscal performance over the last three decades. Section D discusses the specific policy challenges in the Swiss federalist state related to population aging. Some policy implications are drawn in Section E.
5. For Switzerland, the analysis points to the following main conclusions:
A. The diversity of objectives across government levels and a relatively weak intertemporal perspective make it difficult to set an ex-ante fiscal stance for the general government. This may explain the ex post procyclical fiscal policies. Subfederal entities are the main contributors to these tendencies.
B. While competitive fiscal federalism is effective in avoiding a proliferation of new spending programs, it is less ideally equipped to cope with pressures from existing (entitlement) programs, which implies clear fiscal risks in the face of population aging. Coordination among governments is important to avoid delays in implementing credible policy responses, the demise of the fiscal rules, or undue tax increases.
C. The reform of financial equalization and of the distribution of tasks between the Confederation and the Cantons—Neuordnung des Finanzausgleich or NFA2—is a welcome development. The NFA improves incentives to lower costs at the cantonal level, clarifies the tasks of federal and subfederal governments, and mandates coordination among Cantons in key areas, including education and health.
D. Looking forward, implementing the NFA without delays and alleviating coordination problems between federal and cantonal levels should improve the ex-post fiscal stance and the intertemporal profile of fiscal policy. In that regard, greater transparency on medium-term challenges and policy options at all levels of the general government could foster an ex-post coordinated response. Also, the extension of fiscal rules akin to the federal “debt brake” to more entities could strengthen the fiscal anchor of the general government, discourage free riding, and foster structural fiscal reforms.
B. The Benefits and Costs of Decentralization
Fiscal Federalism: Benefits and Limits
6. Devolving policy responsibilities to subnational governments can improve social welfare through a better matching of the demand and supply of public goods. On the demand side, needs and preferences for public goods and services (in quality and quantity) may vary across regions of a given country (or municipalities of a given state), reflecting different socio-economic conditions, economic structures, or geographical features (Oates, 1972). On the supply side, local policymakers have more accurate information, and are therefore better placed to tailor public goods to local conditions. Also, decentralization keeps policymakers close to their constituents, which enhances democratic accountability and reduces possible gaps between the pursuit of the common good and private objectives.
7. Yet decentralization has its limits. First, it is difficult to design a clear-cut assignment of policy responsibilities among different levels of government because most policy areas (e.g. education or hospital care) involve many dimensions (e.g. fixed investment, curriculum, salaries), which may not all be suitable for decentralization. The outcome may be costly overlaps across jurisdictions (e.g. redundant infrastructures), insufficient incentives for cost effectiveness (e.g. ill-designed grant financing), or unfunded mandates. Second, decentralization reduces the scope for economies of scale, which also reduces overall cost-effectiveness of providing public goods and services. Third, costly coordination failures are possible. They are discussed in greater detail below.
8. These limitations are highly country-specific. The extent of decentralization thus varies widely across the world (Figure 1). Switzerland emerges as one of the most decentralized government systems, surpassed only by Canada among industrial countries. The discrepancy between the revenue and expenditure shares of subnational governments (or vertical imbalance) is also smaller than in the world sample (Switzerland is located relatively close to the 45-degree line depicted in the right panel of Figure 1).
Decentralization Across the World
(98 countries—average 1972–2000)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Sources: Government Finance Statistics and IMF staff calculations.9. Despite significant differences of view on the merits of decentralization, economic theory suggests several widely accepted principles guiding the distribution of responsibilities between central and subcentral governments. Broadly speaking, functions concerning macroeconomic stabilization and redistribution—along with intrinsically “national” public goods such as defense and foreign relations—should be performed by the center, while other, more allocative, functions (such as education, public transport, sanitation, etc) are potential candidates for devolution to subnational entities. The reasons for keeping stabilization at the centre includes easier coordination with monetary policy, avoidance of free-riding problems, risk-sharing opportunities (through equalization transfers, and the national tax-transfer system), and “scale economies” (including better borrowing conditions, access to a larger and less mobile tax base, and smaller Ricardian effects3). The rationale for centralizing redistributive policies follows from the “adverse selection” problem created by household mobility within national borders (Stigler, 1957), as net contributors move to low-tax-low-transfers areas, and net beneficiaries concentrate in high-tax-high-transfers areas (Wildasin, 1991).
10. The actual distribution of tasks between national and subnational entities for a sample of industrial and developing economies reveals two notable trends (Table 1). First, all functions listed in the table are concurrent in at least one country of the sample, signaling both the high degree of country-specificity in the determinants of intergovernmental fiscal relations and the difficulty to design a clear cut assignment of tasks. Second, predominantly national public goods emerge as functions related to defense, foreign relations and the operation of the national economic union. As the sample does not include Switzerland, shaded entries identify the country’s position
The Distribution of Tasks between Central and Subcentral Governments in Selected Countries
Source: OECD/World Bank (2003), and IMF staff calculations. Note: Sample size varies between 38 and 43 countries depending on the task. Numbers show the percentage of countries where the task in line is performed by the government level in column, or concurrently (third column). For Switzerland, IMF staff relied on Dafflon (2004) and the Federal Constitution.The Distribution of Tasks between Central and Subcentral Governments in Selected Countries
| Functions | National | Subnational | Concurrent |
|---|---|---|---|
| Defense | 95.2 | 0.0 | 4.8 |
| Foreign affairs | 85.4 | 0.0 | 14.6 |
| Old age and disability pensions | 82.5 | 0.0 | 17.5 |
| Competition | 80.0 | 5.0 | 15.0 |
| Justice | 80.0 | 0.0 | 20.0 |
| Labor market | 76.2 | 2.4 | 21.4 |
| Higher education | 59.5 | 21.4 | 19.0 |
| Local police | 36.8 | 52.6 | 10.5 |
| Health | 30.0 | 7.5 | 62.5 |
| Social assistance | 29.3 | 22.0 | 48.8 |
| Transport infrastructure | 27.9 | 14.0 | 58.1 |
| Primary and secondary education | 21.4 | 47.6 | 31.0 |
| Sanitation | 12.2 | 63.4 | 24.4 |
| Waters | 9.8 | 56.1 | 34.1 |
| Local transport | 7.1 | 66.7 | 26.2 |
| Garbage collection | 7.3 | 82.9 | 9.8 |
The Distribution of Tasks between Central and Subcentral Governments in Selected Countries
| Functions | National | Subnational | Concurrent |
|---|---|---|---|
| Defense | 95.2 | 0.0 | 4.8 |
| Foreign affairs | 85.4 | 0.0 | 14.6 |
| Old age and disability pensions | 82.5 | 0.0 | 17.5 |
| Competition | 80.0 | 5.0 | 15.0 |
| Justice | 80.0 | 0.0 | 20.0 |
| Labor market | 76.2 | 2.4 | 21.4 |
| Higher education | 59.5 | 21.4 | 19.0 |
| Local police | 36.8 | 52.6 | 10.5 |
| Health | 30.0 | 7.5 | 62.5 |
| Social assistance | 29.3 | 22.0 | 48.8 |
| Transport infrastructure | 27.9 | 14.0 | 58.1 |
| Primary and secondary education | 21.4 | 47.6 | 31.0 |
| Sanitation | 12.2 | 63.4 | 24.4 |
| Waters | 9.8 | 56.1 | 34.1 |
| Local transport | 7.1 | 66.7 | 26.2 |
| Garbage collection | 7.3 | 82.9 | 9.8 |
Coordination Issues
11. Coordination issues are either horizontal—between jurisdictions on the same level—or vertical—between different levels of government. Horizontally, decisions implemented in one jurisdiction may affect welfare elsewhere, either through direct externalities (e.g. an international airport), or through location choices of individuals and businesses. Absent ex-ante coordination, decentralization may typically entail an underprovision of goods and policies with positive externalities (e.g. road network, recreation areas, public libraries, and other hardly excludable goods), and an overprovision of goods and policies with negative externalities (e.g. business-friendly zoning and infrastructures, “predatory” tax systems intended to attract certain categories of mobile taxpayers)
12. Vertically fragmented fiscal policy making may undermine the efficiency of the general government budget as a tool of macroeconomic stabilization. In a noncooperative setting, the fiscal stance of the general government is not set ex-ante in line with national objectives, but results from autonomous decisions by each policymaker taking the actions of others as given. As stabilization is typically a second-order consideration for subcentral entities (see above), their policies are more likely to result in procyclical fiscal impulses.4 Also, the smaller size of central tax-transfer systems (including financial equalization arrangements) inevitably limits horizontal insurance in the face of region-specific shocks (see e.g. von Hagen, 1999; and Fatàs, 1998) 5.
13. Social and income redistribution policies may suffer with decentralization, leading to pressure on central government finances to compensate. The adverse selection problem described above may encourage subnational entities to provide suboptimally low levels of redistribution. The central government finds itself in a weak position to refuse supplementary funding for such expenditure programs, known for their political sensitiveness (Sutherland, Price and Joumard, 2005). This point is particularly relevant in the context of population aging, as the costs of many social programs are creeping up whereas the adverse selection problem may be worsening. With limited willingness and/or ability for subnational entities to bear the rising costs of social programs, the central government thus becomes the natural “financier of last resort.”
14. A sensible vertical distribution of tasks can reduce the scope for coordination problems, but it cannot eliminate them. The main reason is that it is in general impossible to find one-to-one correspondences between policy targets and instruments. Typically, each instrument—i. e., a tax measure or a spending program—affects multiple policy targets—growth, employment, macroeconomic stability, income redistribution—on which each government level places a different emphasis. In such a setting, no clever assignment of tasks can ensure the best possible performance on all objectives, and policy coordination is needed to establish priorities. Coordination problems are most obvious with respect to macroeconomic stabilization (traditionally a national objective). For example, subnational governments may find it individually optimal to launch valuable investment projects when financing conditions are favorable. Yet this could add unwanted macroeconomic stimuli to a booming economy.
15. Decentralized fiscal policymaking may be less conducive to fiscal discipline, prompting many countries to impose restrictions on subnational borrowing (Figure 2).6 First, the possibility of a bailout by the center may weaken the perception of hard budget constraints at the subnational level. A related issue is the incentive for subnational governments to outsource fiscal stabilization to the center by demanding additional transfers. As already noted, outsourcing is more appealing, the greater the role of subnational government in politically sensitive programs. Second, subnational governments are individually less exposed to the unpleasant financial repercussions of excessive public debt accumulation such as higher expected inflation and interest rates (Beetsma and Bovenberg, 1999). Third, intergovernmental financial arrangements themselves may carry the seeds of fiscal indiscipline. Specifically, high dependency of subnational finances on intergovernmental transfers implies that policymakers fail to fully internalize the economic costs of public spending through taxation. Fourth, fiscal retrenchments are likely to be more difficult to implement in decentralized fiscal systems because decentralization adds a layer of complexity to the “war of attrition”7 over sharing the burden of adjustment. Finally, decentralization may reduce the transparency and timelines of fiscal accounts, potentially reducing the benefits of closer democratic control8 and the disciplining effect of financial markets.9
Borrowing Restrictions in Selected Countries
(relative frequency in the sample)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Source: OECD/World Bank (2003).Coordination Mechanisms
16. Policy coordination can take many forms, involving varying degrees of constraint on policymakers’ discretion. At the “soft” end of the spectrum, the open method of coordination (or OMC) operates through benchmarks characterizing best practices. An open exchange of information on policies provides implicit rewards and punishments—including through peer pressure—which ultimately affect policies. In its stronger form, the OMC may involve explicit “scoring” or “naming and shaming” practices. The OMC is useful in cases where institutionalized (i.e. ex-ante) forms of coordination are politically difficult to put in place, and where binding commitments are not necessary to improve outcomes—that is essentially when policy competition is expected to lead to a cooperative equilibrium.10
17. In decentralized fiscal systems, relatively strong ex-ante coordination mechanisms are often in place.
E. Horizontally, subnational entities sometimes join forces through special purpose entities (e.g. water distribution or school districts), binding cooperation agreements, and sometimes, outright amalgamation.11 These forms of coordination work best when incentives for cooperation are high, say to exploit economies of scale. Horizontal coordination on the provision of public goods carrying significant externalities can also be envisaged to the extent that both parties clearly perceive the higher costs of a noncooperative interaction. In contrast, harmful tax competition is unlikely to be credibly resolved through such mechanisms.12
F. Vertically, the idea is either to limit discretion at the subnational level by promoting national spending priorities (e.g. through conditional grants) and constraining tax parameters (e.g. through tax sharing or harmonization), or to encourage in various ways subnational governments to adopt budgets consistent with national macroeconomic objectives. In principle, medium-term financial plans covering all levels of government provide a useful framework for vertical coordination. This, however, requires extensive ex-ante negotiation at all stages of the budget process—from preparation to implementation—akin to centralized policymaking, and is therefore likely to be difficult in decentralized countries. In practice, vertical coordination often involves targeting rules imposed by (or negotiated with) the center (see Figure 3), and in some cases contractual agreements (such as the Accords de Coopération in Belgium). However, implicit policy coordination can be promoted through the adoption of uniform budget procedures, identical macroeconomic assumptions, and full disclosure of relevant information (see Figure 3).
Mechanisms of Vertical Coordination in Selected Countries13
(relative frequency in the sample)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Source: OECD/World Bank (2003).C. Swiss Fiscal Federalism at Work
18. This section gathers empirical evidence on Swiss fiscal policy over the last 3 decades. As macroeconomic coordination issues discussed above receive particular attention, the first subsection presents the institutional framework for coordination in Switzerland. The second subsection provides key stylized facts, while the third one ties these to an analysis of the determinants of fiscal policy in a cross-country setting. The fourth sub Section Discusses evidence of strategic interactions among fiscal policymakers.
Fiscal Policy Coordination in Switzerland
19. Swiss competitive fiscal federalism gives ample room for policy discretion at the subnational level. In Tiebout’s (1956, p. 418) words, “the consumer-voter may be viewed as picking that community which best satisfies his preference pattern for public goods.” In that model, policymakers have strong incentives to provide high-quality services at a minimum cost. Direct democracy also fosters citizens’ effective power to vote not just with their feet (as assumed in Tiebout’s quote) but also with their voice. The importance of horizontal competition inevitably limits the use of formal, ex-ante coordination mechanisms, especially those involving the center.
20. While the Federal Constitution grants considerable autonomy to the Cantons,14 it does not preclude extensive horizontal and vertical coordination. The general principle that the Confederation must “respect the autonomy of Cantons” (Article 47) effectively prevents a top-down approach to vertical coordination. However, the Constitution invites policymakers at all levels to cooperate (Article 44), which implies in particular that the Cantons may participate in federal decisions (Article 45.1), and that they must enforce federal laws (Article 46.1). The Constitution also prohibits unfunded mandates arising from federal law, and demands that a “fair” financial equalization mechanism be organized (Article 46.3). The Confederation is required to provide detailed information on policy initiatives to Cantons, and, when their interests are at stake, to consult with them (Article 45.2). Horizontally, the Constitution establishes a legal instrument supporting cooperation agreements among Cantons: the intercantonal conventions (Article 48). The Confederation can, within the limits of federal prerogatives, take part in such conventions.
21. The Federal Constitution imposes some tax policy coordination. Vertical coordination is achieved through a clear allocation of tax instruments to each level of government,15 and numerical caps on federal tax rates (Articles 128 and 130).16 Horizontally, the Confederation can enact laws prohibiting “unjustified” tax holidays, and harmonizing tax bases, collection principles, and legal procedures (Article 129). In contrast, rates, brackets and (within limits) exemptions are not subject to harmonization, leaving the door open to tax competition among Cantons.
22. The Federal Constitution is largely non-binding regarding vertical coordination on macroeconomic objectives. Specifically, Article 100.1 mandates the Confederation to ensure macroeconomic stability (smooth the business cycle and maintain price stability), whereas Cantons and Communes are only required to take into account the business cycle when setting their own fiscal policy (Article 100.4).
23. Other formal coordination mechanisms, however, are at work. They include specialized Cantonal Conferences. Among them, the Conference of Cantonal Finance Directors organizes horizontal coordination on financial matters to be discussed with the federal government, and circulates information on financial issues. Still the Conference does not operate as a forum of negotiation leading to genuine ex-ante coordination on cantonal policies. The jurisprudence of the Federal Court—Switzerland’s supreme judiciary authority—has emerged as an implicit mechanism of policy coordination, especially in tax policy matters (Dafflon, 2000).
24. The recognition of inefficiencies related to the lack of coordination was one important motivation behind the NFA. Key dimensions of the reform are to minimize the scope for coordination failures—by streamlining the distribution of tasks between Cantons and the Confederation—and to improve vertical and horizontal coordination.
25. The NFA significantly enhances horizontal coordination by defining 9 areas in which intercantonal conventions are mandatory, including hospitals, higher education, culture, and local transport. The rationale for making coordination compulsory in these matters is to eliminate free-riding behaviors and thereby make sure that all beneficiaries of a certain public good or service contribute to the costs of providing it (Dafflon, 2004).17 To prevent that outsiders undermine the benefits of intercantonal cooperation, it is envisaged that the Confederation could, at the request of a qualified majority of Cantons (18, and in some cases 21), automatically extend to all Cantons certain conventions deemed of national interest.
26. Financial equalization mechanisms implicitly discourage “predatory” (horizontal) tax competition. Currently, statutory tax rates enter negatively in the formula determining Cantonal indices of “financial capacity” (i.e. the lower the tax rates, the lower net equalization transfers). The NFA preserves a negative link between financial capacity and tax rates by linking the latter to the size of a harmonized tax base. Hence, aggressive reductions in tax rates driven by the desire to attract mobile taxpayers will continue to entail a reduction in net equalization transfers. In contrast, the new equalization system will be neutral on tax reductions affecting immobile tax bases—in essence, not driven by tax competition.
27. The scope for strong forms of vertical coordination remains limited. On the one hand, the NFA addresses some pressing coordination issues, whereas the federal Constitution does not provide necessary authority to the Confederation to impose such coordination. On the other hand, the large and heterogeneous group of Cantons is unlikely to welcome “contractual” forms of vertical cooperation.18
28. Perhaps one underestimated dimension of Swiss fiscal federalism is the implicit role of the open method of coordination. The OMC is arguably a natural feature of the Swiss system where each Canton (or even Commune) is seen as a “laboratory” for policy experiments and innovations “without widespread risks of failure” (Dafflon, 2000, Page 3). The introduction of statutory fiscal rules reminiscent of the federal “debt brake” in various Cantons19 could be interpreted as one ‘spontaneous’ manifestation of the OMC. Each government indeed has a chance to learn from the success (or failure) of others (including the federal level), encouraging a certain convergence of policies and institutions towards a freely accepted best practice (ex-post coordination). At the same time, the OMC avoids the strictures of a ‘one-size-fits-all’ solution, allowing for local variants of the model, and even for the possibility that certain entities do not adopt the model if it does not fit their needs or preferences.
Public Debt and Expenditure: Growing Problems
29. In comparison to other European countries, general government expenditure in Switzerland is low (Figure 4). Only the Baltic states, Ireland, and Spain (albeit only marginally) spend less relative to GDP. That outcome is generally attributed to the expenditure-containment role of direct democracy within a highly competitive federalism, where the tax base (firms and citizens) appears to be fairly mobile and responsive to tax policy changes.20 The Swiss system thus seems to effectively tie the size of government to citizens’ demand for public goods and services rather than to the policymakers’ willingness to spend, which can be a source of expenditure and deficit bias.
General Government Expenditure in Selected European Countries
(2004, percent of GDP)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Sources: Eurostat, and IMF.30. However, since the early 1990s, a protracted period of slow growth combined with increasing pressures on social spending have altered fiscal performance (Figure 5). The public debt-to-GDP ratio, which had been broadly stable over the 1970s and the 1980s, grew rapidly in the 1990s as the primary balance failed to offset the automatic debt accumulation stemming from interest payments (the “snowball effect”). The federal government was by far the largest contributor to the debt increase, owing to sustained primary deficits and to significant off-budget operations, including the recapitalization of public enterprises and support to their pension funds, and the financing of large-scale railway investment projects through a special fund.21 The growing importance of off-budget operations during the 1990s is particularly evident from the declining correlation between the overall balance and variations in the public debt.
Switzerland: Public Debt Developments (1970–2004)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Sources: IMF country desk data, and staff estimates.31. Aside from off-budget operations, structural expenditure pressures largely account for the deterioration of the fiscal situation. On the one hand, the rise in unemployment led to higher social spending—notably through the federal unemployment insurance—with a lagged impact on cantonal social assistance programs, as unemployment benefits are limited in time. On the other hand, rising real healthcare costs (see IMF, 2004) were reflected in higher current transfers from federal and cantonal governments—mainly health-insurance assistance to low-income households, and transfers to social security programs. Low incentives to enhance cost-effectiveness in existing expenditure programs—particularly those financed with intergovernmental transfers—further contributed to the lack of expenditure restraint (Giorno and Joumard, 2002). Overall, the current primary expenditure of the general government increased by almost 8 percentage points of GDP between 1990 and 2004 (Figure 6).
Switzerland: Public Expenditure Developments (1990–2004)
(Current primary expenditure in percent of GDP23)
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Sources: IMF country desk data and staff estimates.Assessing Fiscal Behavior(s)
32. To better grasp the macroeconomic determinants of Swiss fiscal performance, it is useful to place it in cross-country perspective, using a common methodology. To do so, econometric analysis allows checking whether a stable relationship exists between the fiscal stance and some of its key macroeconomic determinants—i.e. cyclical conditions and the concern for long-term solvency. It should be made clear upfront that the estimated model merely describes ex-post, average policy patterns for the period under review, and should not be interpreted as characterizing a systematic ex-ante “reaction function” of the policymakers. Given the relatively short time series available for fiscal policy in most countries (30 to 40 years at best), the literature generally relies on a parsimonious specification24 described by equation (1).
using a subscript t to designate year t, and where p captures the primary balance (in percent of GDP), gap, the output gap, and debt is the public debt-to-GDP ratio.25 The lagged dependent variable is introduced to account for likely persistence in the policymaking process so that λ is expected to be positive. Also, the estimated model allows for β to differ between good times (positive output gap) and bad times (negative output gap).26 A positive sign for β means that when GDP growth exceeds potential (i.e. the output gap variable increases), the primary balance improves, in line with a countercyclical response of fiscal policy. In the absence of discretionary fiscal measures, automatic stabilizers should lead to a positive β. In contrast, a negative β would point to a systematic tendency of discretionary fiscal measures to offset automatic stabilizers, signaling a procyclical bias in policymaking. The sign of γ is expected to be positive as governments concerned with long-term sustainability tend to increase the primary balance in response to a rise in the public debt ratio (Bohn, 1998).
33. The analysis proceeds in two steps. First, an average fiscal behavior is estimated on a panel including all EU countries (except Luxembourg) as well as Australia, Canada, Switzerland, and the United States. Second, individual country estimates for all federal countries in the sample are compared.27
34. The industrial countries gathered in this panel tend to run countercyclical fiscal policies in bad times, and exhibit a robust stabilizing response to the public debt. As fiscal policy affects the state of the economy, least-squares estimates of the β coefficients may be biased. The second column of Table 2 controls for that reverse causality, using instrumental variables techniques. It appears that the estimated coefficient is positive and significant only in bad times. In good times, the coefficient is negative and non-significant, indicating that automatic stabilizers are offset by procyclical discretionary fiscal impulses. Procyclicality in good times reflects the inherent difficulty for policymakers to resist expansionary pressures when revenues are strong. The positive and significant sign found for γ suggests that the intertemporal budget constraint is properly internalized. All other things being equal, a rising debt ratio triggers a fiscal retrenchment and vice versa.
Average Fiscal Behavior in a Panel of 18 Industrial Countries (1970–2004)
(dependent variable: primary balance in percent of GDP)
Panel least squares with country fixed effects (not reported).
Panel IV estimates, instrumenting the output gap with all other explanatory variables, one lag of these, and two lags of the output gap variables. Country fixed effects are included.
Test of whether federalism leaves the coefficient estimates unaffected. The test is perfomed on a variant that does not distinguish between positive and negative output gaps.
Average Fiscal Behavior in a Panel of 18 Industrial Countries (1970–2004)
(dependent variable: primary balance in percent of GDP)
| LSDV 1/ | IV-FE 2/ | ||||
|---|---|---|---|---|---|
| Primary balance (-1) | 0.72 | *** | 0.73 | *** | |
| Output gap (+) | 0.13 | *** | −0.09 | ||
| Output gap (-) | 0.25 | *** | 0.34 | *** | |
| Public debt (-1) | 0.03 | *** | 0.03 | *** | |
| F-test 3/ | |||||
| Federalism | 0.34 | 0.42 | |||
| R-squared | 0.78 | 0.78 | |||
| Number of obs. | 541 | 517 | |||
Panel least squares with country fixed effects (not reported).
Panel IV estimates, instrumenting the output gap with all other explanatory variables, one lag of these, and two lags of the output gap variables. Country fixed effects are included.
Test of whether federalism leaves the coefficient estimates unaffected. The test is perfomed on a variant that does not distinguish between positive and negative output gaps.
Average Fiscal Behavior in a Panel of 18 Industrial Countries (1970–2004)
(dependent variable: primary balance in percent of GDP)
| LSDV 1/ | IV-FE 2/ | ||||
|---|---|---|---|---|---|
| Primary balance (-1) | 0.72 | *** | 0.73 | *** | |
| Output gap (+) | 0.13 | *** | −0.09 | ||
| Output gap (-) | 0.25 | *** | 0.34 | *** | |
| Public debt (-1) | 0.03 | *** | 0.03 | *** | |
| F-test 3/ | |||||
| Federalism | 0.34 | 0.42 | |||
| R-squared | 0.78 | 0.78 | |||
| Number of obs. | 541 | 517 | |||
Panel least squares with country fixed effects (not reported).
Panel IV estimates, instrumenting the output gap with all other explanatory variables, one lag of these, and two lags of the output gap variables. Country fixed effects are included.
Test of whether federalism leaves the coefficient estimates unaffected. The test is perfomed on a variant that does not distinguish between positive and negative output gaps.
35. A second salient result is that, as a group, countries with a federal structure do not appear to behave differently on average than the rest of the sample.28 Table 2 reports tests of joint significance of interaction terms between explanatory variables and a dummy identifying federal countries (not displayed). The low values for the corresponding F-tests indicate that the null hypothesis of no effect of federalism cannot be rejected. In fact, country-specific estimates illustrate that federalism, irrespective of its cooperative—as in Austria, Belgium, and Germany—or competitive—as in Canada and the US—nature, does not mechanically entail macroeconomic coordination failures that result in procyclical fiscal policies and neglect for the intertemporal budget constraint (Table 3).29 On the one hand, procyclicality and short-term bias in policymaking may stem from constitutional features and political economy factors unrelated to federalism per se (Persson and Tabellini, 2004). On the other hand, as discussed before, coordination issues need not be overwhelming and can be addressed through various institutional mechanisms.
Fiscal Behavior in Selected Federal Countries (General Government, 1970–2004)
(dependent variable: primary balance in percent of GDP)
Note: IV estimates, instrumenting the output gap with other explanatory variables, one lag of these, and two lags of the output gap.Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
Fiscal Behavior in Selected Federal Countries (General Government, 1970–2004)
(dependent variable: primary balance in percent of GDP)
| Austria | Belgium | Canada | Germany | US | Switzerland | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Constant | −1.44 | *** | −2.63 | *** | −4.52 | *** | −0.91 | −2.16 | ** | −0.42 | ||
| Primary balance (-1) | 0.50 | *** | 0.81 | *** | 0.67 | *** | 0.50 | *** | 0.64 | *** | 0.77 | *** |
| Output gap | 0.38 | *** | −0.07 | 0.23 | *** | −0.12 | 0.33 | * | −0.08 | |||
| Public debt (-1) | 0.03 | *** | 0.03 | *** | 0.06 | *** | 0.02 | 0.04 | ** | 0.01 | ||
| Structural break? 1/ | no | 1982 | no | no | no | 1990, 2000 | ||||||
| R-squared | 0.57 | 0.88 | 0.90 | 0.29 | 0.74 | 0.54 | ||||||
| Number of obs. | 31 | 32 | 33 | 33 | 32 | 33 | ||||||
| Distinguishing cyclicality in good and in bad times | ||||||||||||
| Constant | −1.34 | *** | −2.40 | *** | −4.40 | *** | −1.09 | −4.87 | * | 0.77 | ||
| Primary balance (-1) | 0.50 | *** | 0.86 | *** | 0.65 | *** | 0.54 | *** | 0.58 | *** | 0.82 | *** |
| Output gap (+) | 0.34 | 0.00 | 0.08 | 0.10 | 1.36 | −0.43 | ** | |||||
| Output gap (-) | 0.48 | * | −0.40 | 0.33 | *** | −0.30 | −0.06 | 0.07 | ||||
| Public debt (-1) | 0.02 | ** | 0.02 | ** | 0.07 | *** | 0.02 | 0.07 | ** | −0.01 | ||
| Structural break 1/ | no | 1982 | no | no | no | 1990, 2000 | ||||||
| R-squared | 0.57 | 0.87 | 0.90 | 0.21 | 0.58 | 0.49 | ||||||
| Number of obs. | 31 | 32 | 33 | 33 | 32 | 33 | ||||||
Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
Fiscal Behavior in Selected Federal Countries (General Government, 1970–2004)
(dependent variable: primary balance in percent of GDP)
| Austria | Belgium | Canada | Germany | US | Switzerland | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Constant | −1.44 | *** | −2.63 | *** | −4.52 | *** | −0.91 | −2.16 | ** | −0.42 | ||
| Primary balance (-1) | 0.50 | *** | 0.81 | *** | 0.67 | *** | 0.50 | *** | 0.64 | *** | 0.77 | *** |
| Output gap | 0.38 | *** | −0.07 | 0.23 | *** | −0.12 | 0.33 | * | −0.08 | |||
| Public debt (-1) | 0.03 | *** | 0.03 | *** | 0.06 | *** | 0.02 | 0.04 | ** | 0.01 | ||
| Structural break? 1/ | no | 1982 | no | no | no | 1990, 2000 | ||||||
| R-squared | 0.57 | 0.88 | 0.90 | 0.29 | 0.74 | 0.54 | ||||||
| Number of obs. | 31 | 32 | 33 | 33 | 32 | 33 | ||||||
| Distinguishing cyclicality in good and in bad times | ||||||||||||
| Constant | −1.34 | *** | −2.40 | *** | −4.40 | *** | −1.09 | −4.87 | * | 0.77 | ||
| Primary balance (-1) | 0.50 | *** | 0.86 | *** | 0.65 | *** | 0.54 | *** | 0.58 | *** | 0.82 | *** |
| Output gap (+) | 0.34 | 0.00 | 0.08 | 0.10 | 1.36 | −0.43 | ** | |||||
| Output gap (-) | 0.48 | * | −0.40 | 0.33 | *** | −0.30 | −0.06 | 0.07 | ||||
| Public debt (-1) | 0.02 | ** | 0.02 | ** | 0.07 | *** | 0.02 | 0.07 | ** | −0.01 | ||
| Structural break 1/ | no | 1982 | no | no | no | 1990, 2000 | ||||||
| R-squared | 0.57 | 0.87 | 0.90 | 0.21 | 0.58 | 0.49 | ||||||
| Number of obs. | 31 | 32 | 33 | 33 | 32 | 33 | ||||||
Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
36. The ex-post fiscal stance in Switzerland appears highly procyclical, and disconnected from long-run solvency concerns. Unlike the two other countries with competitive federal arrangements, the Swiss fiscal stance on average does not contribute to output stabilization. It appears procyclical in good times, pointing to weak expenditure restraint (or resistance to tax cuts) in the face of temporary revenue gains. In bad times, the fiscal stance is acyclical, indicating a tendency to undertake discretionary adjustment efforts when the economy underperforms. That tendency is strong enough to offset automatic stabilizers. It is important to note, however, that the recent introduction of an annual taxation regime (in 2003) has enhanced the countercyclical role of tax revenues. Also, the fiscal rules adopted by the Confederation (debt brake) and a number of Cantons appropriately encourage a countercyclical fiscal stance. The likelihood of procyclical retrenchments and expansions should therefore be reduced in the future.
37. The Swiss primary balance does not appear to systematically respond to public debt developments. Such behavior does not exclude explosive debt paths, and may thus be inconsistent with a proper internalization of the intertemporal budget constraint. Although similar features emerge from German data, they appear more pronounced in the Swiss case. That said, the one-off nature of some large shocks on the Confederation’s debt in the 1990s combined with a low initial debt level might have limited the perceived need to aggressively revert such developments with higher primary surpluses, especially in a protracted slow growth episode. This might explain the instability observed in the estimated relationship over the 1990s (see below).
38. Within each government level, more detailed results support the view that subfederal entities—both the Cantons and the Communes as a whole—are mainly responsible for the destabilizing impact of the general government fiscal stance (Table 4). At the Confederation’s level, discretionary policy is broadly neutral to the cycle. This is irrespective of the economy’s position in the business cycle, pointing to more effective expenditure control (along with greater attention paid to macroeconomic objectives).
Switzerland: Fiscal Behavior by Level of Government (1970–2004)
(dependent variable: primary balance in percent of GDP)
Note: IV estimates, using the same instruments as in other estimations.Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
Switzerland: Fiscal Behavior by Level of Government (1970–2004)
(dependent variable: primary balance in percent of GDP)
| Federal | Cantonal | Municipal | ||||
|---|---|---|---|---|---|---|
| Constant | −0.29 | 0.17 | 0.06 | |||
| Primary balance (-1) | 0.61 | *** | 0.80 | *** | 0.69 | *** |
| Output gap (+) | −0.03 | −0.15 | *** | −0.09 | *** | |
| Output gap (-) | 0.02 | 0.02 | −0.01 | |||
| Public debt (-1) | 0.01 | 0.00 | 0.01 | |||
| Structural break? 1/ | 1991-2000 | no | no | |||
| R-squared | 0.38 | 0.63 | 0.70 | |||
| Number of obs. | 33 | 33 | 33 | |||
Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
Switzerland: Fiscal Behavior by Level of Government (1970–2004)
(dependent variable: primary balance in percent of GDP)
| Federal | Cantonal | Municipal | ||||
|---|---|---|---|---|---|---|
| Constant | −0.29 | 0.17 | 0.06 | |||
| Primary balance (-1) | 0.61 | *** | 0.80 | *** | 0.69 | *** |
| Output gap (+) | −0.03 | −0.15 | *** | −0.09 | *** | |
| Output gap (-) | 0.02 | 0.02 | −0.01 | |||
| Public debt (-1) | 0.01 | 0.00 | 0.01 | |||
| Structural break? 1/ | 1991-2000 | no | no | |||
| R-squared | 0.38 | 0.63 | 0.70 | |||
| Number of obs. | 33 | 33 | 33 | |||
Based on preliminary screening for possible breaks using the Cusum of squares test, and confirmed with Chow tests.
39. The policy response to public debt developments appears to have been weak at all government levels. However, the instability of the estimated relationship is confirmed for the federal level only, supporting the earlier conjecture of a deliberate policy by the Confederation not to revert aggressively the impact of the one-off shocks on its debt. In contrast, the relationship is stable for Cantons and Communes, unambiguously pointing to a systematic lack of intertemporal perspective at these levels. Hence, the combination of subfederal policies appears inconsistent with the stability of their aggregate debt ratio.
40. Data availability prevents a more detailed analysis of cantonal macro-fiscal behavior along the lines suggested by equation (1). Systematic analyses of cantonal fiscal performance therefore focus more on explaining the cross-cantonal variance in the size of government (see e.g. Feld and Kirchgässner, 2005) than on assessing their fiscal stance in the light of macroeconomic goals. Still, a number of empirical regularities identified in the literature are worth discussing.
41. First, while tax competition and direct democracy contain the size of cantonal governments, they do not automatically provide them with a fiscal anchor. Institutions that operate symmetrically on primary expenditure and revenues may not effectively constrain deficits and debt accumulation. Also, tying the hands of policymakers may become costly when policy responsiveness to changing circumstances is highly valued. With debt accumulation as the only degree of freedom, it is therefore hardly surprising that cantonal debt levels and debt accumulations are much more diverse than tax burdens (Table 5).
Switzerland: Descriptive Cantonal Data
Sources: Swiss authorities, University of St Gallen, and Feld and Kirchgässner (2005).Switzerland: Descriptive Cantonal Data
| Surface | Population | Net cantonal income in 2003 | Per capita income in 2003 | Nominal income growth 98-03 | Cantonal debt per capita (2003) | Change in debt per capita (90-03) | Change in debt per capita (98-03) | PIT burden 2003 | CIT burden 2003 | Large city | University | Statutory fiscal rule | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (Sq. Km) | (in 2002) | (million francs) | (francs) | (percent p. a.) | (francs) | (francs) | (francs) | (index) | (index) | () if project | |||||
| Aargau | 1,403.7 | 556,229 | 26,907.6 | 48,145.1 | 2.02 | 3,165.8 | 1,299.7 | 0.7 | 82.1 | 97.4 | no | no | (soft) | ||
| Appenzell-Ausserrhoden | 243.0 | 53,189 | 2,225.7 | 42,231.7 | 0.56 | 2,276.9 | −783.4 | −2,008.9 | 122.2 | 85.2 | no | no | soft | ||
| Appenzell-Innerrhoden | 172.5 | 14,995 | 625.7 | 42,632.7 | 0.58 | 2,297.8 | −1,645.8 | −1,177.7 | 107.8 | 73.4 | no | no | no | ||
| Basel-Landschaft | 517.5 | 263,194 | 13,631.7 | 51,848.9 | 1.55 | 11,551.8 | 6,250.3 | 4,137.6 | 89.1 | 107.8 | no | no | no | ||
| Basel-Stadt | 37.1 | 186,871 | 19,046.1 | 99,865.3 | 3.48 | 25,987.7 | 13,557.9 | 333.4 | 110.9 | 111.5 | yes | yes | soft | ||
| Bern | 5,958.9 | 950,209 | 41,285.1 | 43,066.0 | 1.97 | 10,483.8 | 6,504.9 | 2,039.7 | 114.1 | 102.3 | yes | yes | medium | ||
| Fribourg | 1,670.8 | 242,679 | 9,528.5 | 38,342.3 | 1.50 | 4,477.7 | 1,606.2 | −100.4 | 129.0 | 112.4 | no | yes | very strict | ||
| Geneve | 282.2 | 419,254 | 25,299.6 | 59,123.2 | 2.56 | 46,511.5 | 33,298.2 | 18,085.9 | 81.7 | 126.7 | yes | yes | no | ||
| Glarus | 685.2 | 38,380 | 2,462.3 | 64,318.1 | 2.35 | 6,507.9 | 3,144.1 | 1,498.8 | 134.3 | 112.7 | no | no | no | ||
| Graubunden | 7,105.2 | 186,105 | 8,711.9 | 45,565.2 | 1.38 | 5,742.9 | 3,436.6 | −1,253.3 | 109.7 | 121.1 | no | no | soft | ||
| Jura | 838.6 | 69,196 | 2,506.1 | 36,900.6 | 1.64 | 8,984.5 | 5,879.0 | 1,372.3 | 139.3 | 106.3 | no | no | no | ||
| Luzern | 1,493.5 | 352,311 | 14,538.3 | 41,218.9 | 1.44 | 8,092.7 | 3,286.0 | 765.3 | 132.4 | 89.4 | no | yes | strict | ||
| Neuchatel | 803.1 | 166,949 | 7,647.4 | 45,474.0 | 2.33 | 11,022.6 | 6,480.8 | 1,209.3 | 131.4 | 105.8 | no | yes | (in preparation) | ||
| Nidwalden | 276.1 | 38,897 | 2,377.3 | 61,933.6 | 2.40 | 6,264.0 | 1,600.5 | −278.5 | 80.0 | 76.0 | no | no | strict | ||
| Obwalden | 490.5 | 32,999 | 1,167.4 | 35,359.4 | 1.74 | 6,012.7 | 2,500.7 | −1,105.1 | 161.9 | 102.2 | no | no | no | ||
| Schaffhausen | 298.5 | 73,916 | 3,915.0 | 52,781.3 | 1.63 | 5,394.5 | 1,896.5 | −222.2 | 119.4 | 112.4 | no | no | no | ||
| Schwyz | 908.3 | 133,358 | 6,242.4 | 46,696.3 | −1.12 | 2,831.1 | −111.9 | −841.5 | 65.1 | 57.3 | no | no | (in preparation) | ||
| Solothurn | 790.7 | 246,504 | 10,785.6 | 43,902.2 | 1.53 | 5,813.1 | 2,692.6 | −1,003.2 | 115.7 | 96.4 | no | no | (strict) | ||
| St. Gallen | 2,025.6 | 455,193 | 19,866.5 | 43,388.0 | 1.88 | 3,829.8 | 1,305.3 | −397.8 | 111.1 | 110.9 | no | yes | very strict | ||
| Thurgau | 990.9 | 229,882 | 9,980.8 | 43,120.9 | −3.82 | 5,523.7 | 2,391.9 | −942.4 | 106.9 | 93.6 | no | no | no | ||
| Ticino | 2,812.5 | 314,563 | 12,263.7 | 38,745.1 | 5.84 | 5,063.3 | 853.7 | 42.4 | 61.7 | 91.2 | no | no | (soft) | ||
| Uri | 1,076.6 | 35,246 | 1,585.2 | 45,629.8 | 1.70 | 6,172.5 | 3,387.8 | 1,796.1 | 157.8 | 114.6 | no | no | no | ||
| Valais | 5,224.5 | 281,020 | 10,596.9 | 37,367.0 | 1.86 | 16,383.5 | 13,472.7 | 5,912.8 | 116.6 | 101.3 | no | no | (medium) | ||
| Vaud | 3,212.1 | 631,999 | 33,290.4 | 51,332.5 | 2.93 | 8,774.3 | 4,797.0 | −179.9 | 93.2 | 119.6 | yes | yes | no | ||
| Zug | 238.8 | 102,247 | 8,922.9 | 86,499.3 | 3.58 | 3,959.4 | 1,833.9 | −1,592.8 | 48.7 | 60.2 | no | no | no | ||
| Zurich | 1,728.8 | 1,242,488 | 83,950.4 | 66,050.1 | 2.65 | 7,282.2 | 2,477.0 | −1,314.6 | 93.0 | 103.8 | yes | yes | strict | ||
| Average | 1,587.9 | 281,457 | 14,590.8 | 50,443.8 | 1.77 | 8,861.8 | 4,669.7 | 952.9 | 108.3 | 99.7 | - | - | - | ||
| excluding Geneve | 1,640.1 | 275,945 | 14,162.4 | 50,096.6 | 1.74 | 7,355.9 | 3,524.6 | 267.6 | 109.3 | 98.6 | - | - | - | ||
| fiscal rules in 1998 | 2,599.2 | 337,342 | 15,160.0 | 43,868.6 | 1.66 | 4,605.9 | 2,068.1 | −550.8 | 109.5 | 107.6 | - | - | - | ||
| large city | 2,243.8 | 686,164 | 40,574.3 | 63,887.4 | 2.72 | 19,807.9 | 12,127.0 | 3,792.9 | 98.6 | 112.8 | - | - | - | ||
| excluding Geneve | 2,734.2 | 752,892 | 44,393.0 | 65,078.5 | 2.76 | 13,132.0 | 6,834.2 | 219.6 | 102.8 | 109.3 | - | - | - | ||
| university | 1,912.5 | 516,439 | 28,272.5 | 54,206.7 | 2.30 | 14,051.4 | 8,145.9 | 2,271.2 | 110.8 | 109.2 | - | - | - | ||
| excluding Geneve | 2,116.2 | 528,587 | 28,644.1 | 53,592.1 | 2.27 | 9,993.9 | 5,001.9 | 294.4 | 114.4 | 107.0 | - | - | - | ||
| high tax burden | 1,999.6 | 273,052 | 13,264.5 | 48,977.1 | 1.99 | 11,262.0 | 6,702.9 | 1,797.0 | 122.4 | 109.1 | - | - | - | ||
| excluding Geneve | 2,114.1 | 263,305 | 12,462.2 | 48,300.7 | 1.96 | 8,912.0 | 4,929.9 | 711.1 | 125.1 | 107.9 | - | - | - | ||
| Dispersion (Coef. of var.) | 1.17 | 1.05 | 1.20 | 0.30 | 0.94 | 1.03 | 1.47 | 4.12 | 0.26 | 0.18 | - | - | - | ||
Switzerland: Descriptive Cantonal Data
| Surface | Population | Net cantonal income in 2003 | Per capita income in 2003 | Nominal income growth 98-03 | Cantonal debt per capita (2003) | Change in debt per capita (90-03) | Change in debt per capita (98-03) | PIT burden 2003 | CIT burden 2003 | Large city | University | Statutory fiscal rule | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (Sq. Km) | (in 2002) | (million francs) | (francs) | (percent p. a.) | (francs) | (francs) | (francs) | (index) | (index) | () if project | |||||
| Aargau | 1,403.7 | 556,229 | 26,907.6 | 48,145.1 | 2.02 | 3,165.8 | 1,299.7 | 0.7 | 82.1 | 97.4 | no | no | (soft) | ||
| Appenzell-Ausserrhoden | 243.0 | 53,189 | 2,225.7 | 42,231.7 | 0.56 | 2,276.9 | −783.4 | −2,008.9 | 122.2 | 85.2 | no | no | soft | ||
| Appenzell-Innerrhoden | 172.5 | 14,995 | 625.7 | 42,632.7 | 0.58 | 2,297.8 | −1,645.8 | −1,177.7 | 107.8 | 73.4 | no | no | no | ||
| Basel-Landschaft | 517.5 | 263,194 | 13,631.7 | 51,848.9 | 1.55 | 11,551.8 | 6,250.3 | 4,137.6 | 89.1 | 107.8 | no | no | no | ||
| Basel-Stadt | 37.1 | 186,871 | 19,046.1 | 99,865.3 | 3.48 | 25,987.7 | 13,557.9 | 333.4 | 110.9 | 111.5 | yes | yes | soft | ||
| Bern | 5,958.9 | 950,209 | 41,285.1 | 43,066.0 | 1.97 | 10,483.8 | 6,504.9 | 2,039.7 | 114.1 | 102.3 | yes | yes | medium | ||
| Fribourg | 1,670.8 | 242,679 | 9,528.5 | 38,342.3 | 1.50 | 4,477.7 | 1,606.2 | −100.4 | 129.0 | 112.4 | no | yes | very strict | ||
| Geneve | 282.2 | 419,254 | 25,299.6 | 59,123.2 | 2.56 | 46,511.5 | 33,298.2 | 18,085.9 | 81.7 | 126.7 | yes | yes | no | ||
| Glarus | 685.2 | 38,380 | 2,462.3 | 64,318.1 | 2.35 | 6,507.9 | 3,144.1 | 1,498.8 | 134.3 | 112.7 | no | no | no | ||
| Graubunden | 7,105.2 | 186,105 | 8,711.9 | 45,565.2 | 1.38 | 5,742.9 | 3,436.6 | −1,253.3 | 109.7 | 121.1 | no | no | soft | ||
| Jura | 838.6 | 69,196 | 2,506.1 | 36,900.6 | 1.64 | 8,984.5 | 5,879.0 | 1,372.3 | 139.3 | 106.3 | no | no | no | ||
| Luzern | 1,493.5 | 352,311 | 14,538.3 | 41,218.9 | 1.44 | 8,092.7 | 3,286.0 | 765.3 | 132.4 | 89.4 | no | yes | strict | ||
| Neuchatel | 803.1 | 166,949 | 7,647.4 | 45,474.0 | 2.33 | 11,022.6 | 6,480.8 | 1,209.3 | 131.4 | 105.8 | no | yes | (in preparation) | ||
| Nidwalden | 276.1 | 38,897 | 2,377.3 | 61,933.6 | 2.40 | 6,264.0 | 1,600.5 | −278.5 | 80.0 | 76.0 | no | no | strict | ||
| Obwalden | 490.5 | 32,999 | 1,167.4 | 35,359.4 | 1.74 | 6,012.7 | 2,500.7 | −1,105.1 | 161.9 | 102.2 | no | no | no | ||
| Schaffhausen | 298.5 | 73,916 | 3,915.0 | 52,781.3 | 1.63 | 5,394.5 | 1,896.5 | −222.2 | 119.4 | 112.4 | no | no | no | ||
| Schwyz | 908.3 | 133,358 | 6,242.4 | 46,696.3 | −1.12 | 2,831.1 | −111.9 | −841.5 | 65.1 | 57.3 | no | no | (in preparation) | ||
| Solothurn | 790.7 | 246,504 | 10,785.6 | 43,902.2 | 1.53 | 5,813.1 | 2,692.6 | −1,003.2 | 115.7 | 96.4 | no | no | (strict) | ||
| St. Gallen | 2,025.6 | 455,193 | 19,866.5 | 43,388.0 | 1.88 | 3,829.8 | 1,305.3 | −397.8 | 111.1 | 110.9 | no | yes | very strict | ||
| Thurgau | 990.9 | 229,882 | 9,980.8 | 43,120.9 | −3.82 | 5,523.7 | 2,391.9 | −942.4 | 106.9 | 93.6 | no | no | no | ||
| Ticino | 2,812.5 | 314,563 | 12,263.7 | 38,745.1 | 5.84 | 5,063.3 | 853.7 | 42.4 | 61.7 | 91.2 | no | no | (soft) | ||
| Uri | 1,076.6 | 35,246 | 1,585.2 | 45,629.8 | 1.70 | 6,172.5 | 3,387.8 | 1,796.1 | 157.8 | 114.6 | no | no | no | ||
| Valais | 5,224.5 | 281,020 | 10,596.9 | 37,367.0 | 1.86 | 16,383.5 | 13,472.7 | 5,912.8 | 116.6 | 101.3 | no | no | (medium) | ||
| Vaud | 3,212.1 | 631,999 | 33,290.4 | 51,332.5 | 2.93 | 8,774.3 | 4,797.0 | −179.9 | 93.2 | 119.6 | yes | yes | no | ||
| Zug | 238.8 | 102,247 | 8,922.9 | 86,499.3 | 3.58 | 3,959.4 | 1,833.9 | −1,592.8 | 48.7 | 60.2 | no | no | no | ||
| Zurich | 1,728.8 | 1,242,488 | 83,950.4 | 66,050.1 | 2.65 | 7,282.2 | 2,477.0 | −1,314.6 | 93.0 | 103.8 | yes | yes | strict | ||
| Average | 1,587.9 | 281,457 | 14,590.8 | 50,443.8 | 1.77 | 8,861.8 | 4,669.7 | 952.9 | 108.3 | 99.7 | - | - | - | ||
| excluding Geneve | 1,640.1 | 275,945 | 14,162.4 | 50,096.6 | 1.74 | 7,355.9 | 3,524.6 | 267.6 | 109.3 | 98.6 | - | - | - | ||
| fiscal rules in 1998 | 2,599.2 | 337,342 | 15,160.0 | 43,868.6 | 1.66 | 4,605.9 | 2,068.1 | −550.8 | 109.5 | 107.6 | - | - | - | ||
| large city | 2,243.8 | 686,164 | 40,574.3 | 63,887.4 | 2.72 | 19,807.9 | 12,127.0 | 3,792.9 | 98.6 | 112.8 | - | - | - | ||
| excluding Geneve | 2,734.2 | 752,892 | 44,393.0 | 65,078.5 | 2.76 | 13,132.0 | 6,834.2 | 219.6 | 102.8 | 109.3 | - | - | - | ||
| university | 1,912.5 | 516,439 | 28,272.5 | 54,206.7 | 2.30 | 14,051.4 | 8,145.9 | 2,271.2 | 110.8 | 109.2 | - | - | - | ||
| excluding Geneve | 2,116.2 | 528,587 | 28,644.1 | 53,592.1 | 2.27 | 9,993.9 | 5,001.9 | 294.4 | 114.4 | 107.0 | - | - | - | ||
| high tax burden | 1,999.6 | 273,052 | 13,264.5 | 48,977.1 | 1.99 | 11,262.0 | 6,702.9 | 1,797.0 | 122.4 | 109.1 | - | - | - | ||
| excluding Geneve | 2,114.1 | 263,305 | 12,462.2 | 48,300.7 | 1.96 | 8,912.0 | 4,929.9 | 711.1 | 125.1 | 107.9 | - | - | - | ||
| Dispersion (Coef. of var.) | 1.17 | 1.05 | 1.20 | 0.30 | 0.94 | 1.03 | 1.47 | 4.12 | 0.26 | 0.18 | - | - | - | ||
42. Second, both the large number of Cantons and the extent of cantonal disparities make it difficult to internalize national macroeconomic goals, and weaken incentives to enter into binding cooperative arrangements. Intercantonal disparities are considerable on most counts, including size, income, growth, fiscal institutions and indebtedness (Table 5). In that context, it is particularly challenging for subnational policymakers to reconcile local interests with uniform, national objectives, making vertical coordination less easy. Horizontally, coordination is typically harder to sustain when players are numerous and diverse. For instance, predatory tax competition can more easily take place between a small, geographically isolated, rural Canton and a large, urban one, than between two entities of equal size and needs. A key reason is that the large Canton is unlikely to lose out much from its small neighbor’s policy. In contrast, two equally sized entities would be more sensitive to the mutually destructive effect of predatory tax cuts, and the mere fear of a race to the bottom (perhaps accompanied with pressures for centralized harmonization) might encourage some degree of coordination.
43. Mindful of the potential risk of fiscal indiscipline, a growing number of Cantons have adopted statutory fiscal rules.30 Empirical evidence suggests that these rules, based in law, do lead to stronger budget balances on average (Feld and Kirchgässner, 2005). As Table 5 illustrates, both the level of debt and the rate of debt accumulation over the period 1998-2003 were on average much lower in these Cantons. Currently, 9 cantonal governments operate under statutory “debt brakes” (against 5 in 1998—see Table 5) while 6 other Cantons have similar projects. However, one specific difficulty in implementing cantonal debt brakes (in comparison to the federal arrangement) is the assessment of cyclical conditions, as there are no data on cantonal output gaps.
Strategic Interactions
44. Beyond the difficulty to internalize the objectives of other government levels and the existence of institutions tying the hands of policymakers, the extent of coordination problems also depend on the type of interaction among decision makers. This subsection provides some evidence on these interactions in Switzerland.
45. Horizontally, a chief cause for concern remains a possible increase in tax competition. Lower transport and communication costs have increased the mobility of the tax base, and correspondingly, the incentives to use tax policy in a strategic way. Many empirical studies indicate that in competitive federations such as Switzerland, tax policies are set strategically.31 Table 6 illustrates this for corporate income taxation (CIT) in selected Cantons of western Switzerland (Dafflon, 2000). In all Cantons but one, CIT payments of a “typical” firm decreased between 1985 and 2003. Early cuts in Bern and Jura between 1985 and 1995 were emulated elsewhere—Geneva excepted—in the second sub-period (1995–2003). While the cross-cantonal dispersion of CIT payments was broadly constant between 1985 and 1999, it was halved between 1999 and 2003, suggesting a convergence towards lower tax levels.
Switzerland: Standardized Corporate Income Tax Payments in Selected Cantons
Sources: Dafflon (2000), and Federal Tax Administration. Notes: Entries in the table are numerical examples of taxes paid on a net profit of 320,000 Swiss francs by a firm with 2,000,000 francs in capital and reserves.Coefficient of variation across Cantons.
Switzerland: Standardized Corporate Income Tax Payments in Selected Cantons
| Average rates of change (per annum) | |||||||
|---|---|---|---|---|---|---|---|
| 1985 | 1995 | 1999 | 2003 | 1985-2003 | 1985-95 | 1995-2003 | |
| (in current Swiss francs) | (in percent) | ||||||
| Bern | 72,028 | 51,710 | 51,681 | 47,329 | −3.2 | −3.3 | −1.1 |
| Fribourg | 61,614 | 59,149 | 58,721 | 55,029 | −0.9 | −0.4 | −0.9 |
| Geneva | 59,641 | 59,607 | 63,342 | 63,567 | 0.5 | 0.0 | 0.8 |
| Jura | 85,582 | 58,723 | 58,549 | 56,416 | −3.2 | −3.7 | −0.5 |
| Neuchatel | 90,180 | 88,897 | 88,565 | 58,240 | −3.3 | −0.1 | −5.1 |
| Valais | 74,704 | 65,325 | 65,204 | 53,638 | −2.5 | −1.3 | −2.4 |
| Vaud | 66,914 | 64,091 | 63,976 | 59,183 | −0.9 | −0.4 | −1.0 |
| Average | 72,952 | 63,929 | 64,291 | 56,200 | −1.9 | −1.3 | −1.6 |
| Dispersion 1/ | 0.16 | 0.19 | 0.18 | 0.09 | |||
Coefficient of variation across Cantons.
Switzerland: Standardized Corporate Income Tax Payments in Selected Cantons
| Average rates of change (per annum) | |||||||
|---|---|---|---|---|---|---|---|
| 1985 | 1995 | 1999 | 2003 | 1985-2003 | 1985-95 | 1995-2003 | |
| (in current Swiss francs) | (in percent) | ||||||
| Bern | 72,028 | 51,710 | 51,681 | 47,329 | −3.2 | −3.3 | −1.1 |
| Fribourg | 61,614 | 59,149 | 58,721 | 55,029 | −0.9 | −0.4 | −0.9 |
| Geneva | 59,641 | 59,607 | 63,342 | 63,567 | 0.5 | 0.0 | 0.8 |
| Jura | 85,582 | 58,723 | 58,549 | 56,416 | −3.2 | −3.7 | −0.5 |
| Neuchatel | 90,180 | 88,897 | 88,565 | 58,240 | −3.3 | −0.1 | −5.1 |
| Valais | 74,704 | 65,325 | 65,204 | 53,638 | −2.5 | −1.3 | −2.4 |
| Vaud | 66,914 | 64,091 | 63,976 | 59,183 | −0.9 | −0.4 | −1.0 |
| Average | 72,952 | 63,929 | 64,291 | 56,200 | −1.9 | −1.3 | −1.6 |
| Dispersion 1/ | 0.16 | 0.19 | 0.18 | 0.09 | |||
Coefficient of variation across Cantons.
46. Vertically, an important issue is whether interaction patterns between the different layers of government carry risks for macro-fiscal performance. It is already clear from the analysis above that subnational fiscal policies are procyclical. On public debt sustainability, the key issue is whether the pattern of strategic interactions leads to an equitable burden sharing between the Cantons and the Confederation. Specifically, fiscal adjustments and structural fiscal reforms may be harder to achieve if each player engages in beggar-thy-neighbor strategies. For instance, downsizing federal welfare programs—such as unemployment insurance or transfers to asylum seekers—often merely relocates the pressure on social outlays to the cantonal and communal levels. Conversely, the federal level may come under pressure from subcentral entities to cope with structural spending increases in the health sector (IMF, 2004). To assess formally the extent and nature of vertical macrofiscal interactions, simple statistical tests have been performed using aggregate data for Cantons and Communes (Table 7).
Switzerland: Indicators of Fiscal Policy Interaction
Based on system (3SLS) estimation of equation (1) for the 3 levels of government, in which the lagged aggregate public debt and primary balance of the other two levels have been added.
Switzerland: Indicators of Fiscal Policy Interaction
| Granger causality tests: primary balance | |||||
|---|---|---|---|---|---|
| Federal | Cantonal | Municipal | |||
| Federal does not cause | - | 4.38 | ** | 2.40 | |
| Cantonal does not cause | 9.43 | *** | - | 0.37 | |
| Municipal does not cause | 4.87 | ** | 2.18 | - | |
| Granger causality tests: overall balance | |||||
| Federal does not cause | - | 2.94 | * | 1.72 | |
| Cantonal does not cause | 7.18 | *** | - | 0.14 | |
| Municipal does not cause | 4.21 | ** | 1.60 | - | |
| Policy interactions: primary balance response to other levels 1/ | |||||
| Primary balance (-1) | 0.51 | *** | −0.13 | 0.02 | |
| Public debt (-1) | 0.09 | ** | −0.02 | 0.00 | |
Based on system (3SLS) estimation of equation (1) for the 3 levels of government, in which the lagged aggregate public debt and primary balance of the other two levels have been added.
Switzerland: Indicators of Fiscal Policy Interaction
| Granger causality tests: primary balance | |||||
|---|---|---|---|---|---|
| Federal | Cantonal | Municipal | |||
| Federal does not cause | - | 4.38 | ** | 2.40 | |
| Cantonal does not cause | 9.43 | *** | - | 0.37 | |
| Municipal does not cause | 4.87 | ** | 2.18 | - | |
| Granger causality tests: overall balance | |||||
| Federal does not cause | - | 2.94 | * | 1.72 | |
| Cantonal does not cause | 7.18 | *** | - | 0.14 | |
| Municipal does not cause | 4.21 | ** | 1.60 | - | |
| Policy interactions: primary balance response to other levels 1/ | |||||
| Primary balance (-1) | 0.51 | *** | −0.13 | 0.02 | |
| Public debt (-1) | 0.09 | ** | −0.02 | 0.00 | |
Based on system (3SLS) estimation of equation (1) for the 3 levels of government, in which the lagged aggregate public debt and primary balance of the other two levels have been added.
47. Despite the limited size of its budget, the Confederation plays a distinct role in the system. Granger causality tests suggest that the Confederation’s budget is sensitive to the aggregate fiscal stance of the Communes and the Cantons, whereas only cantonal policies appear (weakly) affected by changes in the federal fiscal stance. As Granger-causality tests are sometimes weak and unreliable, system estimates of equation (1) for each entity have been performed, allowing for the lagged aggregate policy stance of the other two government layers to affect the current policy stance of each entity. Interesting empirical regularities emerge.
First, the Confederation’s budget appears related to subfederal balances in such a way that aggregate slippages (improvements) at these levels are followed by similar developments at the federal level. This is consistent with the idea of the Confederation being a “financier of last resort.”
Second, the Confederation appears sensitive to subfederal debt developments, and this in a stabilizing fashion consistent with the possibility that it bears the burden of adjustment to the intertemporal budget constraint of the general government.32
Third, the aggregate fiscal stance of Cantonal and Communal governments is largely unaffected by federal fiscal policy, suggesting that pressures on federal accounts are not transmitted to subfederal levels in a systematic way.
48. As a result, the Confederation may find it increasingly difficult to maintain fiscal discipline without broader support from other entities. First, transfers to Cantons and to the social security system account entirely for the rise in the Confederation’s current primary expenditure (1.3 percent of GDP) over the last 15 years (Figure 7). Second, the limited flexibility of the federal budget on the revenue side implies that the pressure to adjust falls in large part on current primary expenditure net of transfers to Cantons and Social Security. The Confederation’s share of “flexible” spending (i.e. goods and services, and capital spending) in total expenditure of the general government declined accordingly, especially after the implementation of the federal “debt brake” rule (right-hand panel of Figure 7).
Switzerland: Selected Expenditure Indicators for the Confederation
Citation: IMF Staff Country Reports 2006, 203; 10.5089/9781451807301.002.A001
Source: IMF Staff calculations.D. Challenges Ahead and Policy Options
49. The main challenge facing Swiss public finances concerns the long-term impact of population aging. The current system rests on mechanisms—tax competition, direct democracy, and fiscal rules—that tie government outlays to the population’s demand for public goods and services.33 As discussed earlier, this largely explains the moderate size of the government sector in Switzerland. However, aging-related pressures on existing expenditure programs change the dynamics of the budget at constant policy. Such structural pressures require adjustments for which the system may not be ideally prepared.
50. Competitive fiscal federalism by itself does not facilitate broad-based support for a policy package combining structural fiscal reforms and new financing. The likely delays in tackling the aging question would add to the medium term fiscal pressures. Frictions between the Confederation and the Cantons may arise for a variety of reasons, including potential asymmetries in demographics and migration flows, different degrees of fiscal flexibility, an asymmetric distribution of the political costs associated with a given adjustment package,34 and last but not least, institutional asymmetries in the distribution of the costs of doing-nothing.
51. By default, the Confederation has been the financier of last resort for Social Security, and it is effectively exposed to the bulk of spending pressures at unchanged policies.35 The social security programs—and especially old-age pensions, disability insurance, and means-tested health-insurance support to households—face no hard budget constraints of their own, relying heavily on transfers from the Confederation. This may encourage free riding on efforts to either provide additional financing to these politically sensitive programs, or to push for their reform.
52. Yet the Confederation is approaching the limits of budgetary flexibility. Federal tax increases are subject to double-majority (i.e., population and Cantons) referenda, margins for expenditure reallocation have narrowed, and the “debt brake” rule formally shuts off the deficit financing option. Cantonal budgets are somewhat more flexible, although the revenue side is constrained by horizontal tax competition, and the expenditure side is often bounded by mandatory referenda on new measures. Moreover, in many Cantons, “debt brake” arrangements formally prevent debt buildups.
53. The overall lack of fiscal flexibility has important policy implications. On the one hand, it forces budgets to remain within strict parameters that are difficult to change unless there is a broad-based support in the population to do so. Strong fiscal commitment should thus tilt the balance in favor of ambitious structural fiscal reforms. On the other hand, reforms tend to have a dynamics of their own driven by their often important distributive effects. The status quo bias in these areas is consequently large, making it difficult to implement any workable reform package in time to prevent the unraveling of the rules-based fiscal framework, and an undesirable drift in public debt.
54. Intergovernmental coordination can reduce these problems. First, increased horizontal coordination in areas such as health and education can bring about significant cost savings. Second, enhanced vertical coordination could reduce the free riding problem associated with the “financier-of-last-resort” role of the Confederation.
55. Informing the public on long-term policy options and trade-offs can help deliver an ex-post coordinated response without the strictures of formal commitments. This takes a special importance in the Swiss context, where strong forms of (ex-ante) vertical coordination are unlikely. Specifically, if voters have information on the policy options and trade-offs facing all government levels, policymakers may find it easier to obtain support in their respective constituencies for policies consistent with a comprehensive reform strategy. In that regard, the periodical publication of a long run fiscal sustainability report that clearly outlines the cost-of-doing-nothing and the policy trade-offs involved can be a useful instrument. Still, better-informed voters are no guarantee against free-riding attempts, especially if the size of the dissenting entity does not threaten the success of the broader strategy.
56. Imposing a hard budget constraint on the social security system could also downsize the free riding problem. For example, distinct solvency requirements for social security programs as a whole could reduce the role of the Confederation as the payer of last resort. Admittedly, the implicit commitment not to bail out an entity in charge of politically sensitive expenditures would hardly enjoy immediate credibility. However, this would have the merit to make the magnitude of the looming fiscal problem more evident, and as such, could trigger a broader appetite for reforms, especially if repeated financing crises were to occur.
57. At unchanged structural policies, aging-related pressures are likely to force the Confederation, and to a lesser extent the Cantons, to recover some fiscal flexibility. If the credibility of rules-based fiscal frameworks is to be preserved, then flexibility on the revenue side will have to be augmented proportionately to the lack of flexibility on the expenditure side. For the Confederation, the economically least damaging way to recover tax flexibility would be to allow for temporary adjustments in VAT rates when the debt brake is at risk and room for expenditure cuts is exhausted. Incidentally, such VAT changes would put a clear price tag on eventual delays in implementing structural fiscal reforms.36
58. However, greater fiscal flexibility may weaken the strong incentives for cost efficiency present in the Swiss system. To alleviate that problem, permitted VAT increases could be limited in size and over time, and only be allowed as a last resort measure to preserve the debt brake.
E. Conclusions
59. Significant differences exist in the macro-fiscal behavior of federal and subfederal entities, with severe procyclicality and lack of intertemporal perspective observed in the latter. At the subfederal level, the apparent failure to internalize the solvency constraint underscores the lack of a fiscal anchor. Mindful of the problem, a growing number of Cantons have introduced (or are planning to introduce) statutory fiscal rules.
60. Competitive fiscal federalism arrangements prevailing in Switzerland have contributed to contain the size of the public sector and public debt by discouraging new spending initiatives. However, the system’s resilience in the face of pressures from existing entitlement programs is less evident, carrying serious fiscal risks in the context of population aging. In particular, the federal level has so far absorbed the bulk of fiscal adjustment efforts, leaving it with limited flexibility to do more. Yet, under current trends, the center is also likely to bear a large share of aging-related outlays, particularly in the health, disability, and old-age pension sectors. Absent some intergovernmental coordination, the implementation of a credible policy response—a mix of entitlement reform, and revenue increases—could be delayed, leaving debt accumulation—and with it, the demise of the fiscal rule—as the only safety valve.
61. The recent reform of financial equalization and of the distribution of tasks between the Confederation and the Cantons provides better incentives for cost-effectiveness at the cantonal level, reduces areas of potential friction between federal and subfederal governments, and enhances horizontal coordination among Cantons. However, vertical coordination mechanisms remain weak, possibly slowing down the implementation of solutions to gathering spending pressures.
62. Looking forward, a steadfast implementation of the NFA and enhanced vertical coordination between federal and cantonal levels seem important to improve the ex-post fiscal stance and to guarantee that intertemporal constraints are properly internalized at all government levels. In that perspective, a menu of policy options is available.
63. First, greater transparency with a view to inform the public on longer-term challenges, and on policy options and trade-offs at all levels of the general government could provide stronger incentives for each entity to contribute to a solution (ex-post coordinated response). The long-run fiscal sustainability report currently under preparation is a major instrument to raise public awareness of long-run challenges.
64. Second, consideration could be given to extending “debt brake” arrangements to all decision levels to strengthen the fiscal anchor of the general government, reduce the risk of free riding on adjustment efforts, and spur structural fiscal reforms.
65. Third, to the extent that preserving a credible fiscal anchor is a priority, flexibility in the federal budget, and in particular VAT, may need to be increased. However, such flexibility should be exerted within strict parameters to preserve the strong incentives for cost efficiency present in the Swiss system.
Appendix
Data
The OECD/World Bank (2003) sample covers the following countries:
| ALGERIA | ARGENTINA | AUSTRALIA | AUSTRIA |
| BELGIUM | BOLIVIA | CAMBODIA | CANADA |
| CHILE | COLOMBIA | CZECH REPUBLIC | DENMARK |
| EGYPT | FINLAND | FRANCE | GERMANY |
| GREECE | HUNGARY | ICELAND | INDONESIA |
| IRELAND | ISRAEL | ITALY | JAPAN |
| JORDAN | KENYA | KOREA | MEXICO |
| MOROCCO | NETHERLANDS | NEW ZEALAND | NORWAY |
| PERU | PORTUGAL | SLOVAK REPUBLIC | SLOVENIA |
| SOUTH AFRICA | SPAIN | SURINAME | SWEDEN |
| TURKEY | UNITED KINGDOM | UNITED STATES | URUGAY |
| ALGERIA | ARGENTINA | AUSTRALIA | AUSTRIA |
| BELGIUM | BOLIVIA | CAMBODIA | CANADA |
| CHILE | COLOMBIA | CZECH REPUBLIC | DENMARK |
| EGYPT | FINLAND | FRANCE | GERMANY |
| GREECE | HUNGARY | ICELAND | INDONESIA |
| IRELAND | ISRAEL | ITALY | JAPAN |
| JORDAN | KENYA | KOREA | MEXICO |
| MOROCCO | NETHERLANDS | NEW ZEALAND | NORWAY |
| PERU | PORTUGAL | SLOVAK REPUBLIC | SLOVENIA |
| SOUTH AFRICA | SPAIN | SURINAME | SWEDEN |
| TURKEY | UNITED KINGDOM | UNITED STATES | URUGAY |
Cantonal “Debt Brake” Arrangements (Active or in Preparation)
Source: University of St Gallen, and IMF Staff calculations.Restrictiveness indices range between 1 (least restrictive) and 10 (most restrictive). When possible, the ACIR index was calculated following the definition provided in Sutherland, Price and Joumard (2005). IMF staff index was elaborated to best capture available information on cantonal rules. A value of 10 corresponds to a strict balanced budget requirement (8 for any other main budget target). Points are then subtracted depending on whether (i) deficits are allowed, (ii) targets are adjusted for the business cycle, (iii) carry over are allowed, and (iv) escape clauses exist. The corresponding spreadsheet is available upon request.
Cantonal “Debt Brake” Arrangements (Active or in Preparation)
| Status | Main target | Deficit allowed ? | Target adjusted for business cycle? | Adjustment strategy | Escape clauses? | Restrictiveness Indices | ||
|---|---|---|---|---|---|---|---|---|
| IMF staff 1/ | ACIR | |||||||
| Aargau | Projected | Balanced budget | Yes | Yes (no correction if recession, 10-year carry over if stagnation). | Mainly expenditure with a 5-year carry over. | Yes | 3 | 5 |
| Appenzell-Ausserrhoden | In effect | Balanced budget | Yes (cap on deficit of approved budget) | Yes | Mix with a 7-year carry over | Yes | 3 | 3 |
| Basel-Stadt | In effect | Debt ceiling / Expenditure growth | Yes | Yes | Mix with a 1-year carry over | Yes | 2 | - |
| Bern | In effect (2002) | Stable debt ratio / balanced budget | No | No | Mix with a 1-year carry over | Yes | 4.5 | 7 |
| Fribourg | In effect (1996) | Balanced budget | No | No | Tax | No | 10 | 9 |
| Graubunden | In effect (1998) | Balanced budget / expenditure growth | Yes | Yes | Mix with a 1-year carry over | Yes | 4 | 7 |
| Luzern | In effect (2001) | Balanced budget | Yes (cap on deficit of approved budget) | Yes | Mix with a 1-year carry over | No | 6.5 | 3 |
| Neuchatel | In preparation | - | - | - | - | - | - | - |
| Nidwalden | In effect (2001) | Balanced budget | Yes (cap over a 3-year horizon) | Yes (“exceptional circumstances”) | Mix with a 1-year carry over | No | 7 | 7 |
| Schwyz | In preparation | - | - | - | - | - | - | - |
| Solothurn | Projected | Balanced budget | Yes (cap) | No | Tax | Yes | 6.5 | 9 |
| St. Gallen | In effect (1929) | Balanced budget | Yes (cap) | No | Mix with a 1-year carry over | No | 8 | 7 |
| Ticino | Projected | Expenditure cap (relative to cantonal income) | Yes | Yes | Automatic expenditure cuts | Yes | 3 | - |
| Valais | Projected | Balanced budget | No | Yes | Mix with a 1-year carry over | Yes | 5.5 | 7 |
| Zurich | In effect (2001) | Balanced budget | Yes | yes (balanced budget over 8 years) | Mainly expenditure | No | 7 | 5 |
Restrictiveness indices range between 1 (least restrictive) and 10 (most restrictive). When possible, the ACIR index was calculated following the definition provided in Sutherland, Price and Joumard (2005). IMF staff index was elaborated to best capture available information on cantonal rules. A value of 10 corresponds to a strict balanced budget requirement (8 for any other main budget target). Points are then subtracted depending on whether (i) deficits are allowed, (ii) targets are adjusted for the business cycle, (iii) carry over are allowed, and (iv) escape clauses exist. The corresponding spreadsheet is available upon request.
Cantonal “Debt Brake” Arrangements (Active or in Preparation)
| Status | Main target | Deficit allowed ? | Target adjusted for business cycle? | Adjustment strategy | Escape clauses? | Restrictiveness Indices | ||
|---|---|---|---|---|---|---|---|---|
| IMF staff 1/ | ACIR | |||||||
| Aargau | Projected | Balanced budget | Yes | Yes (no correction if recession, 10-year carry over if stagnation). | Mainly expenditure with a 5-year carry over. | Yes | 3 | 5 |
| Appenzell-Ausserrhoden | In effect | Balanced budget | Yes (cap on deficit of approved budget) | Yes | Mix with a 7-year carry over | Yes | 3 | 3 |
| Basel-Stadt | In effect | Debt ceiling / Expenditure growth | Yes | Yes | Mix with a 1-year carry over | Yes | 2 | - |
| Bern | In effect (2002) | Stable debt ratio / balanced budget | No | No | Mix with a 1-year carry over | Yes | 4.5 | 7 |
| Fribourg | In effect (1996) | Balanced budget | No | No | Tax | No | 10 | 9 |
| Graubunden | In effect (1998) | Balanced budget / expenditure growth | Yes | Yes | Mix with a 1-year carry over | Yes | 4 | 7 |
| Luzern | In effect (2001) | Balanced budget | Yes (cap on deficit of approved budget) | Yes | Mix with a 1-year carry over | No | 6.5 | 3 |
| Neuchatel | In preparation | - | - | - | - | - | - | - |
| Nidwalden | In effect (2001) | Balanced budget | Yes (cap over a 3-year horizon) | Yes (“exceptional circumstances”) | Mix with a 1-year carry over | No | 7 | 7 |
| Schwyz | In preparation | - | - | - | - | - | - | - |
| Solothurn | Projected | Balanced budget | Yes (cap) | No | Tax | Yes | 6.5 | 9 |
| St. Gallen | In effect (1929) | Balanced budget | Yes (cap) | No | Mix with a 1-year carry over | No | 8 | 7 |
| Ticino | Projected | Expenditure cap (relative to cantonal income) | Yes | Yes | Automatic expenditure cuts | Yes | 3 | - |
| Valais | Projected | Balanced budget | No | Yes | Mix with a 1-year carry over | Yes | 5.5 | 7 |
| Zurich | In effect (2001) | Balanced budget | Yes | yes (balanced budget over 8 years) | Mainly expenditure | No | 7 | 5 |
Restrictiveness indices range between 1 (least restrictive) and 10 (most restrictive). When possible, the ACIR index was calculated following the definition provided in Sutherland, Price and Joumard (2005). IMF staff index was elaborated to best capture available information on cantonal rules. A value of 10 corresponds to a strict balanced budget requirement (8 for any other main budget target). Points are then subtracted depending on whether (i) deficits are allowed, (ii) targets are adjusted for the business cycle, (iii) carry over are allowed, and (iv) escape clauses exist. The corresponding spreadsheet is available upon request.
References
Alesina, Alberto, and Allan Drazen, 1991, “Why Are Stabilizations Delayed?,” American Economic Review, 81, pp. 1170 –88.
Balassone, Fabrizio, and Maura Francese, 2004, “Cyclical Asymmetry in Fiscal Policy, Debt Accumulation, and the Maastricht Treaty”, Temi di Discussione No 531, Rome: Banca d’Italia.
Barro, Robert, 1979, “On the Determination of the Public Debt,” Journal of Political Economy, 64, pp. 93 –110.
Bayoumi, Tamim, and Paul Masson, 1998, “Liability-Creating Versus Non-Liability-Creating Fiscal Stabilisation Policies: Ricardian Equivalence, Fiscal Stabilisation, and EMU”, The Economic Journal, 108, pp. 1026 –45.
Beetsma, Roel M.W. J., and A. Lans Bovenberg, 1999, “Does Monetary Unification Lead to Excessive Debt Accumulation?” Journal of Public Economics, 74, pp. 299 –325.
Bohn, Henning, 1998, “The Behavior of U.S. Public Debt and Deficits,” Quarterly Journal of Economics, 113, pp. 949 –63.
Celasun, Oya, Xavier Debrun, and Jonathan Ostry, 2006, “Primary Surplus Behavior and Risks to Fiscal Sustainability in Emerging Market Countries: A “Fan-Chart Approach”, IMF Working Paper No 06/67, Washington, D.C: International Monetary Fund.
Dafflon, Bernard, 2000, “Fiscal Federalism in Switzerland: A Survey of Constitutional Issues, Budget Responsibility and Equalization,” manuscript, University of Fribourg.
Dafflon, Bernard 2004, “Le désenchevêtrement des tâches et la péréquation financière: les chantiers de fédéralisme suisse après le 28 novembre,” Working paper No 385, Facultés des Sciences Économiques et Sociales, University of Fribourg.
Debrun, Xavier, and Hamid Faruqee, 2004, “Fiscal Behaviors in a Heterogenous Monetary union: the Case of EMU”, manuscript. Shorter version in Chapter II of World Economic Outlook, September 2004, Washington, D.C: International Monetary Fund.
Fatàs, Antonio, 1998, “Does EMU Need a Fiscal Federation?” Economic Policy, 26, pp. 165 –92.
Feld, Lars, and Gebhard Kirchgässner, 2002, “The Impact of Corporate and Personal Income Taxes on the Location of Firms and on Employment: Some Panel Evidence for the Swiss Cantons,” Journal of Public Economics, 87, pp. 129 –55.
Feld, Lars, and Gebhard Kirchgässner, and Christoph Schaltegger, 2003, “Decentralized Taxation and the Size of Government: Evidence from Swiss States and Local Governments”, CESifo Working Paper No 1087, Munich: CESifo.
Feld, Lars, and Gebhard Kirchgässner, 2005, “Sustainable Fiscal Policy in a Federal System: Switzerland as an Example,” Working Paper no 2005–16, Basel: CREMA.
Feld, Lars, and Emmanuelle, Reulier, 2005, “Strategic Tax Competition in Switzerland: Evidence from a Panel of the Swiss Cantons,” CESifo Working Paper No 1516, Munich: CESifo.
Feld, Lars, and Christoph Schaltegger, 2005, “Voters as a Hard Budget Constraint: on the Determination of Intergovernmental Grants”, Working Paper No 2005–21, Basel: CREMA.
Galí, Jordi and Roberto Perotti, 2003, “Fiscal Policy and Monetary Integration in Europe,” Economic Policy, 37, pp. 535 –572.
Giorno, Claude, and Isabelle Joumard, 2002, “Enhancing the Effectiveness of Public Spending in Switzerland,” OECD Economics Department Working papers, No 332, Paris: OECD.
Hayashi, Masayoshi, and Robin Boadway, 2001, “An Empirical Analysis of Intergovernmental Tax Interaction: the Case of Business Income Tax in Canada”, Canadian Journal of Economics, 34, pp. 481 –503.
International Monetary Fund, 2004, Switzerland—Selected Issues, IMF Country Report No 05/188, Washington, D.C: International Monetary Fund.
Keen, Michael, 1998, “Vertical Tax Externalities in the Theory of Fiscal Federalism”, IMF Staff Papers, 45, pp. 454 –85.
Keen, Michael, and Cristos Kotsogiannis, 2002, “Does Federalism Lead to Excessively High Taxes?” American Economic Review, 92, pp. 363 –370.
Kirchgässner, Gebhard, and Werner Pommerehne, 1996, “Tax harmonization and tax competition in the European Union: Lessons from Switzerland”, Journal of Public Economics, 60, pp. 351 –71.
Küttel, Dominique, and Peter Kluger, 2001, “Explaining Yield Spreads of Swiss Canton Bonds: An Empirical Investigation,” manuscript, University of Basel.
Oates, Wallace, 1972, Fiscal Federalism, New York: Harcourt.
Persson, Torsten, and Guido Tabellini, 2004, “Constitutional Rules and Fiscal Policy Outcomes,” American Economic Review, 94, pp. 25 –45.
Schaltegger, Christoph, and René Frey, 2004, “Les restrictions fiscales et budgétaires, un moyen de stabiliser les finances publiques,” La Vie Économique, No 2.
Stigler, George, 1957, “The Tenable Range of Functions of Local Governments”, in US Congress Joint Economic Committee, Federal Expenditure Policy for Economic Growth and Stability, pp. 213 –19.
Sutherland, Douglas, Robert Price, and Isabelle Joumard, 2005, “Fiscal Rules for Subcentral Governments: Design and Impact”, Economics Department Working Paper No 465, (Paris: OECD).
Tiebout, Charles, 1956, “A Pure Theory of Local Expenditures,” Journal of Political Economy, 64, pp. 416 –424.
Van Rompuy, Paul, 2005, “The Coordination of Budgetary Policies in Belgium: 15 years of Experience with the High Council of Finance,” manuscript.
von Hagen, Jürgen, 1991, “A Note on the Empirical Effectiveness of Formal Fiscal Restraints,” Journal of Public Economics, 44, pp. 199 –210.
von Hagen, Jürgen, 1999, “A Fiscal Insurance for the EMU?” in Beetsma, Roel and C. Oudshoorn (eds.), Tools for Regional Stabilization, Discussion Paper No 9903, Dutch Ministry of Economic Affairs.
Wildasin, David, 1991, “Income Redistribution in a Common Labor Market,” American Economic Review, 81, pp. 757 –774.
Prepared by Xavier Debrun (FAD). Part of the cantonal data used in this paper were kindly provided by Professor Lars P. Feld and the University of St Gallen.
The NFA was approved by federal referendum in November 2004 and is expected to come into force in 2008.
The downsizing of Ricardian effects results from the fact that liabilities created by centralized stabilization efforts are shared by the entire federation. Using Canadian data, Bayoumi and Masson (1998) estimate that stabilization policy is two to three times more effective (in terms of offsetting the shocks on real disposable income) at the federal level than at the provincial level.
In particular, when stabilization is not a concern, revenue windfalls are more likely to be spent, and adjustments are most likely to be carried out in bad times.
Studies surveyed by von Hagen (1999)—mainly on the US—suggest that on average, the federal tax-transfer system offsets only 10 to 15 percent of regional shocks on state disposable income (with a range of 7 to 40 percent). Fatàs (1998) argues that those studies overestimate the true insurance component “by a factor 3” as one needs to correct for the intertemporal smoothing provided by automatic stabilizers in each state.
For a recent and comprehensive survey of fiscal rules at the subcentral government level, see Sutherland, Price, and Joumard (2005).
Alesina and Drazen (1991) show how a “war of attrition” between two groups of equal size that disagree on sharing the costs of a fiscal adjustment delays fiscal stabilization until the expected costs of fiscal indiscipline become too high for one group.
Democratic accountability reduces potential discrepancies between policymakers objectives and social welfare, and thereby, the relevance of politically motivated deficit bias.
The disciplining role of financial markets in federations is generally thought to be quite weak. In the case of Switzerland, Küttel and Kugler (2001) indeed find that fiscal indicators such as debt and the budget balance do not affect cantonal yield spreads. In contrast, institutional factors (e.g. the extent of mandatory referendums on new expenditure programs) and geographical features (e.g. Cantons with large cities) appear to matter.
That argument helps explain why structural policy coordination in the EU is based on the OMC—the so-called Lisbon strategy. On the one hand, ex-ante coordination is difficult because of the Member States’ reluctance to see “Brussels” impinge on sensitive matters like labor market policies. On the other hand, the competitive dimension of liberalization measures makes ex-ante commitments less necessary as a way to foster mutually beneficial structural reforms.
Municipal mergers are not uncommon. In 1975, Belgium undertook a one-off merger operation, reducing the number of communes from 2498 to 589. In Switzerland, the process is ongoing; there were 3021 communes in 1990; 2899 in 2000; and 2763 on January 1, 2005.
Tax competition is harmful if it prevents government to raise sufficient tax revenues to finance the public goods and services demanded by the population.
In Switzerland (not included in the OECD/World Bank sample), the NFA moves away from conditional grants towards unconditional ones. Cantons have no medium-term budgetary frameworks, and the Confederation has no power to impose any coordination rule (see below). Finally, accounting standards between the Confederation and the Cantons are not the same.
Article 3 of the Constitution states that the Cantons are sovereign and exert all prerogatives not explicitly delegated to the Confederation. The limits of municipal autonomy and forms of coordination at the municipal level are a cantonal matter (Article 50) and will not be discussed in detail here.
Each government level has access to two sources of taxation, namely direct taxation (for all levels), and another, level-specific source (indirect taxation for the Confederation, tax sharing arrangement for the Cantons, and user charges for the municipalities). See Dafflon (2000) for further analysis.
Vertical tax competition arises when two government levels tax the same base (Keen, 1998). As a higher tax rate by say a subnational jurisdiction reduces the base (through supply-side effects), and thereby revenues of the center, the latter has an incentive to increase the tax rate as well, leading to excessive taxation in (the Nash) equilibrium. The constitutional caps on federal tax rates can be viewed as a commitment mechanism preventing such vertical competition. In game-theoretic terms, those caps make the Confederation a Stackelberg leader in the tax setting game, leaving Cantons with full discretion to set their own tax policy. It is, however, less straightforward to rationalize the cap on VAT—an exclusive tax of the Confederation—along those lines. It is important to note that horizontal tax externalities (which may lead to excessively low taxes) dominate vertical externalities if the tax base’s sensitivity to the state-level tax is not too high (Keen and Kotsogiannis, 2002).
This implicitly suggests that the threat of imposing user charges was not credible enough—for political or feasibility reasons—to encourage and sustain ‘spontaneous’ coordination through reputational mechanisms.
In Belgium for example, vertical Cooperation Agreements essentially involve 4 players (the federal government and federated entities). For a description of the Belgian system, see Van Rompuy (2005). Belgium also has an ad-hoc ‘concertation committee’ between federal and subfederal governments to pre-empt potential conflicts of interest. That said, some intercantonal conventions—e.g., on university financing—involve all Cantons, suggesting that there is scope for broad-based coordination through contractual instruments also in Switzerland.
Schaltegger and Frey (2004) describe statutory “debt brakes” at work or under discussion in a number of Cantons. See also Appendix Table A.1. Feld and Kirchgässner (2005) provide empirical evidence supporting the effectiveness on these rules.
See among others Feld and Reulier (2005), Feld, Kirchgässner and Schaltegger (2003), Giorno and Joumard (2002), Feld and Kirchgässner (2002), Kirchgässner and Pommerehne (1996).
Proceeds from the Swiss National Bank gold sales could in part reverse these adverse debt developments.
This panel depicts 15-year rolling correlation coefficients between the overall balance (in percent of GDP) and the first-differenced public debt relative to current year GDP.
Intergovernmental transfers are netted out.
Bohn (1998) relates that specification to the “tax smoothing” theory of fiscal policy (Barro, 1979).
Bohn (1998) provides the first systematic discussion of such model, with an application to the US. Galì and Perroti (2003) use panel data techniques to detect the impact of the Maastricht Treaty on policy behaviors in the European Union, while Celasun, Debrun, and Ostry (2006) further refine the panel approach in the context of emerging market economies, and discuss in detail the econometric issues involved. Data used here are from the OECD’s Analytical Database and IMF country desk.
The null hypothesis of nonstationarity was rejected for all variables except the public debt in Germany and Switzerland. This means that the regression results are not driven by common trends in dependent and independent variables.
Federal countries in the sample include Austria, Belgium (federalism introduced in 1989), Canada, Germany Switzerland, and the United States. Australia could not be analyzed separately due to insufficient data.
Table 3 also reports estimates without distinguishing between good and bad times given the few degrees of freedom and the lower precision of the estimates.
Almost all Cantons have constitutional limits on debt accumulation. In contrast to the US, where constitutional provisions restraining fiscal discretion have direct effects on day-to-day policies (e.g. von Hagen, 1991), statutory provisions appear more effective in Switzerland.
This result is consistent with the role of one-off debt shocks in the 1990s to explain the apparent lack of responsiveness of the Confederation to federal debt.
While direct democracy contains cantonal expenditure (through fiscal referendums), and strictly caps federal tax rates, tax competition limits the Cantons’ room for maneuver on the revenue side. Fiscal rules also provide a fiscal anchor at the federal level and in many Cantons.
This could be the case if the public perceives one specific level of government as having a particular responsibility in social security issues.
VAT is not only less distortionary than direct taxes, but it is also relatively low in Switzerland. Moreover, being an exclusive tax of the Confederation, the risk of harmful vertical tax competition is low.
