Sweden
Selected Issues
Author: International Monetary Fund
Publication Date:
05 Aug 2003
eISBN:
9781451835953
Language:
English
Keywords:
ISCR; CR; expenditure ceiling; revenue boom; finance minister; government expenditure; borrowing authority; finance committee; government duration; Social Democratic government
Download PDF (1.9 MB)

This Selected Issues paper analyzes the strengths and weaknesses of the fiscal framework of Sweden, and discusses some guidelines for their modification. It describes the Swedish experience under the new fiscal framework, and documents the increasing expenditure profile during the economic boom. It describes the basic rationale for fiscal rules, different types of fiscal rules, and also provides suggestions for improving the framework.

Abstract

This Selected Issues paper analyzes the strengths and weaknesses of the fiscal framework of Sweden, and discusses some guidelines for their modification. It describes the Swedish experience under the new fiscal framework, and documents the increasing expenditure profile during the economic boom. It describes the basic rationale for fiscal rules, different types of fiscal rules, and also provides suggestions for improving the framework.

I. Emerging Strains on the Swedish Fiscal Framework: Recent Developments and Future Prospects1

A. Introduction

1. Sweden’s successful fiscal framework is likely to face challenges in the years ahead. Sweden has enjoyed favorable fiscal outturns over the recent past, with a revenue boom fueling the build up of large surpluses, and allowing for tax cuts as well as expenditure increases. Sweden’s performance stands out in a comparative European context. Its fiscal framework, introduced in the 1990s, was underpinned by institutional reform, which culminated in the establishment of two fiscal rules in 1997, one mandating a general government surplus of 2 percent over the cycle, and the other setting nominal expenditure ceilings on central government expenditure. The framework was enhanced in 2000 with the introduction of a balanced budget rule for local governments, with the expectation that a deficit in one year be covered within the next two years. Until recently, the rules have worked very well. Favorable economic developments meant they did not really constrain policy. Recently, however, the end of the revenue boom and the onset of significantly lower economic growth caused the surplus to dwindle, and the framework of fiscal rules to come under strain. If the current period of slow growth continues, the robustness of Sweden’s fiscal framework will be seriously tested for the first time.

2. The emerging strains on the fiscal framework, stemming from a ratcheting up of public spending, suggest that it could be made more robust. The main source of the emerging pressures is that the growth of expenditures was not adequately curbed during the good times. Despite the framework of rules in place, the government took advantage of favorable cyclical conditions to ratchet up expenditure while still managing to respect the rules. This was especially true at the local government level, where rising consumption was financed by a revenue boom. Indeed, the balanced budget rule was maintained at the expense of a further expansion in the size of the government. A cursory review of recent budgets and outturns shows that expenditure was continuously revised upwards, especially for 2002 onwards—precisely the period that the fiscal rules began to come under pressure.

3. This paper analyzes the strengths and weaknesses of the fiscal framework in the light of recent experience and discusses some guidelines for their modification. The next section (Section B) describes the Swedish experience under the new fiscal framework over the past few years. It documents the increasing expenditure profile during the economic boom, which is beginning to strain the framework during the downturn. Section C discusses the basic rationale for fiscal rules in the first place—curbing politically-induced expenditure growth. It then goes on to discuss the different types of fiscal rules, how they should be designed, and how they interact with different institutional factors. Drawing on both the literature and the actual experience in Sweden over the past few years, Section D offers suggestions for improving the framework. All in all, while Sweden’s framework is a strong one, reflecting best practices in numerous respects and superior to that of many other EU countries, there is still room for improvement.

B. The Evolution of Fiscal Policy in Sweden, 2000–03

4. Sweden introduced a new fiscal policy framework in the late 1990s, underpinned by binding fiscal targets. In 1997, the authorities announced a target of a general government surplus of 2 percent of GDP over the cycle while simultaneously establishing comprehensive central government nominal expenditure ceilings set three years in advance. The ceilings capture almost all central government expenditure, including transfers and local government grants, but excluding interest payments. They also include pension payments which are technically outside the scope of central government. An expenditure path consistent with the surplus target is calculated. A margin designed for unanticipated cyclical factors and forecast uncertainties is also built in. The framework was enhanced in 2000, when the government implemented a balanced budget requirement for local governments, with the understanding that any deficits must be followed by corresponding surpluses within two years.

Fiscal developments

5. Sweden’s fiscal policy framework has been broadly successful so far. (Table 1) From 1998–02, the general government surplus averaged 2½ percent of GDP. The structural surplus was also significantly greater than 2 percent of GDP in 2000 and 2001 (Table 2). Moreover, both the revenue and expenditure ratio declined simultaneously. At the same time, the expenditure ceilings were always respected; they were taken seriously by policymakers and thus boosted policy credibility. Furthermore, the aggregate local government sector was in balance on average over this period.

Table 1.

Sweden: General Government Developments, 1998–2003

(In percent of GDP)

article image
Source: Ministry of Finance, Spring Budget 2003.

The shift to a tax guaranteed pensions raised the revenue and expenditure ratios by 0.5 percent of GDP from 2003.

The Church was Sweden was transferred from the local government to the household sector in 2000, meaning that both revenue and expenditure declined by around 0.5 percent of GDP.

Note that local government revenue is protected from any measures taken by the central government which impacts on the local government tax base through a change in the grant. Hence recent tax reforms and the abolition of special deductions for pensions did not affect the total local government revenue base.

Includes consumption by the pension system, 0.1 percent of GDP a year.

A pension reform meant that the financial position of the central government worsened as it took over non-age related pension obligations (survivor and disability benefits). To compensate for this, the pension fund transferred SEK 45 billion in 1999 and 2000, and a further SEK 155 billion (nearly 7 percent of GDP) was transferred in 2001.

Table 2.

Sweden: General Government Unadjusted and Structural Balance, 2000–06

(In percent of GDP)

article image
Source: Ministry of Finance, Spring Budget 2003.

6. Following a period of exceptional strength, fiscal surpluses have declined significantly. Owing to favorable revenue developments, Sweden was able to run large surpluses in 2000 and 2001. Tax revenue in 2001 was particularly high as it reflected exceptionally high corporate profits and capital gains on shares from 2000 during the stock market boom. Temporary revenue from taxes on funds distributed from the insurance company, Alecta, also boosted revenue in 2001. The period of exceptional revenue growth seems to have ended, heralding the return of more normal times. From its zenith of over 4½ percent of GDP in 2001, the general government surplus dropped to just over 1 percent in 2002, and is expected to decline further to less than ½ percent of GDP in 2003 (see Table 1). The remaining surplus is coming entirely from the pension system, as the central government slipped into deficit last year, a deficit that is expected to widen to more than 1¾ percent of GDP in 2003. The local government sector is also recording modest deficits, although, due to higher tax rates, it is expecting balance in 2003. Reflecting substantial discretionary loosening, the structural balance declined by almost 3½ percent of GDP in two years (Table 2). Although fiscal policy is moderately contractionary in 2003, discretionary expenditure is still rising by ½ percent of GDP.

7. The reversal of Fiscal fortune is due to a combination of lower cyclical revenue, tax cuts, and higher expenditure. In the three years to 2003, total revenue fell by 1½ percent of GDP; the decline is especially notable at the central government level given that local government taxes increased by around the same amount over this same period. The end of the revenue boom and a series of significant tax cuts (see below) contributed to this decline at the central government level. While the revenue boom had dissipated by 2002, the pace of spending actually increased. Indeed, the expenditure ratio is also estimated to have increased by more than 1½ percent of GDP in three years. Almost all of this is due to higher local government consumption, as higher transfers and investment were outweighed by declining interest costs, which fell by around 1½ percent of GDP over these years. The large increase in local government expenditure, as well as local government revenue, began in 1998 when local government took over drug financing from the central government; the pace quickened in recent years. On the transfer side, following a decline in the late 1990s coinciding with the economic boom, the recent increase is driven by illness-related transfers, owing to a dramatic rise in sick leave over the past few years.

8. The government implemented large tax cuts over the past few years. The substantial surpluses in 2000 and 2001 allowed the government to implement significant tax reform. Reform centered on compensating pension contributions by a reduction in the income tax. In 2000, income tax was lowered by SEK 12 billion (½ percent of GDP) through a one-quarter reduction of the general pension contribution and a higher central government tax threshold. Tax cuts of SEK 22 billion (1 percent of GDP) were implemented in 2001, three-quarters of which represented the second phase of the compensation for national pension contributions and an increase in the threshold.2 The third phase of this compensation took place in 2002, at the cost of SEK 20 billion (1 percent of GDP). Owing to the fiscal deterioration, the fourth phase has been postponed.

9. Discretionary spending increases, however, outweighed the tax cuts. The Social Democratic government explicitly ranked preserving the welfare state and provision of quality public services ahead of tax cuts. Indeed, following fiscal consolidation in the mid-1990s, there was a substantial pent-up demand for public services. According to the Swedish National Institute of Economic Research, the total cost of tax cuts and expenditure increases enacted over the period 1998–2004 was SEK 167 billion. Tax cuts accounted for only SEK 71 billion of this increase. Successive budgets tended to add more discretionary measures, which accounted for higher expenditure projections with each revision. Areas of increase included health care, subsidies to local government, and transfers for children and pensioners.

10. Transfers related to sick leave have grown sharply.3 A large portion of the increase in transfers comes from an unanticipated rise in illness related spending—sick leave pay and disability pensions. The total cost of sickness transfers is expected to reach 4.6 percent of GDP in 2003, up from 4 percent of GDP in 2000. Successive budgets consistently over-estimated the degree to which the costs associated with ill health would be reversed. The current goal is to halve the number of days compensated by sick pay by 2008. The Spring 2003 budget announced a number of measures to combat overly-generous incentives, such as marginally reducing the sick leave benefit. Recent trends show that while the increase in the number of sick days has leveled off, the number of people on disability pensions is increasing.

11. Local government revenue grew robustly over this period. Tax revenue was buoyed by a strong local economy and employment growth, and also by changes in income tax rules which increased the local government tax base.4 Costs began to outpace revenue in 2001 as the economy slowed. Nonetheless, revenue remained temporarily high during 2000–02, owing to large final tax settlements from the previous years. Total revenue still managed to grow by 6 percent in 2002, partly because social insurance payments are taxable, meaning that the economic slowdown does not fully impact on the tax base. The central government also raised the local government grant successively over the past few years, reflecting the emphasis on improving the quality of public services, especially in health and education (see Table 3). The central government also provided temporary employment support to the local governments, and transferred some income tax revenue to the municipalities and county councils.

Table 3.

Sweden: Increase in Local Government Grants, 1997–2002

(Level compared with 1996, SEK Billion)

article image
Source: Ministry of Finance.

12. The higher revenue fed through directly to higher local government consumption. The upward trend in local government consumption must be seen against the backdrop of the expenditure squeeze following the crisis at the beginning of the 1990s and the subsequent fiscal consolidation. Although the rising trend can be traced back to 1998, it has become more pronounced over the past few years. Local government consumption rose by almost 7 percent in 2001, and by even more in 2002, a result of a higher wage bill and the introduction of a central government mandated maximum daycare charge which forced local governments to bear more of the cost.5 Most of the increase in consumption was due to wages. Employment at the local government level also increased substantially, especially in 2002: the number of people working for municipalities and county councils increased by 1.1 percent, at the same time economy-wide employment stagnated.6 This pace of wage growth is expected to be maintained in 2003.

13. Local government finances have been broadly in balance over this period, but problems emerged in 2002. By and large, the municipalities and county councils have managed to respect the balanced budget requirement, at least in aggregate. A ½ percent of GDP deficit emerged in 2002, when revenue failed to keep pace with the rapidly expanding expenditure. Some local governments ran into acute financial problems. County councils were in worse financial shape than municipalities: in 2002, 62 percent of the latter, but only 28 percent of the former ran surpluses. Part of the problem is that additional central government mandates were not followed by adequate increases in grants.

14. The ratcheting up of local government expenditure, and the fear that revenue would grow less robustly in the future, led to a decision to raise local government tax rates in 2003. The rising expenditure path and the likelihood that the past revenue growth would not be sustained increased the risk that local governments would not be able to balance their budgets in the future. As a result, taxes were increased by SEK 8.5 billion (0.4 percent of GDP) in 2003. In total, 4 county councils and 70 municipalities increased taxes. Furthermore, the temporary employment support and the transfer of central government income tax to municipalities and county councils will be extended until 2004. Nonetheless, the increase in revenue is expected to slow down dramatically in 2004, as final tax settlements this time around will probably reduce revenue. Given that costs are not slowing, taxes may have to increase further in 2004 to maintain budget balance. In any event, the expenditure increases realized in recent years will no longer be sustainable.

15. Local government expenditure is increasingly strained by demographic pressures. Over 70 percent of the services provided by the municipalities depend on the size and age structure of the population, which will lead to an increasing expenditure burden over time. Age related costs are estimated to increase expenditure by around ½ percent a year over the next decade, while more substantial ageing costs will come after 2010. Already, 80 percent of municipalities are exceeding the budget for elderly care, while two-thirds report over-runs on disabled care costs.

The pattern of budgetary projections

16. Emerging fiscal weaknesses are evident from looking at the evolving budgetary projections. Sweden presents two budgets each year, the main budget bill for the upcoming year in the Fall, and a second one in the Spring. Until 2001, the Spring budget set aggregate expenditure ceilings, but this now takes place in the Fall budget bill (which is the principal budget bill for the upcoming year). Looking back over the previous eight budgets, from the Fall of 1999 to the Spring of 2003,7 the fiscal deterioration can be viewed in a more dynamic context as spending plans change in response to new developments. Given that key policy changes are often set out in advance, comparing the different projections for each year gives some sense of changing priorities and outlooks. Tables 4a and 4b compare the outturn data with the estimate from the earliest budget in the sample, that of Fall 1999, for 2000, 2001, and 2002. More detailed information from each budget can be found in Appendix Tables 1 and 2, which consider the evolving forecasts in each of the budgets over this period over 2000–03 (see also Figures 1 and 2 for the 2002 and 2003 projections).

Table 4a.

Sweden: Difference Between Fall 1999 Budget and Outturn. 2000–02 1/

(In SEK billion)

article image
Source: Ministry of Finance.

Outturn data differ from Table 1 due to change in national accounts in 2003 to include VAT on government consumption. This table excludes VAT on government consumption in the outturn to make it consistent with past budgets.

Table 4b.

Sweden: Difference Between Fall 1999 Budget and Outturn, 2000–02 1/

(In percent of GDP)

article image
Source: Ministry of finance.

Outturn data differ from Table 1 due to change in national accounts in 2003 to include VAT on government consumption. This table excludes VAT on government consumption in the outturn to make it consistent with past budgets.

Figure 1.
Figure 1.

Sweden: Changing Budgetary Projections, 2002 1/

Citation: IMF Staff Country Reports 2003, 231; 10.5089/9781451835953.002.A001

Source: Ministry of Finance1/ FA and SP stand for Fall and Spring budgets respectively.
Figure 2.
Figure 2.

Sweden: Changing Budgetary Projections, 2003 1/

Citation: IMF Staff Country Reports 2003, 231; 10.5089/9781451835953.002.A001

Source: Ministry of Finance1/ FA and SP stand for Fall and Spring budgets respectively.

17. Projections showed increasing surpluses in 2000 and 2001. By and large, this windfall gain from the revenue boom in the earlier years was saved; comparing the earliest forecast in the sample with the outturn shows that roughly 80 percent of the unanticipated revenue gain fed through to the general government surplus. For 2001 in particular, the revenue boom was not anticipated, as the revenue ratio is around 2 percent of GDP higher in the last projection than in the first; this reflects central government revenue developments. The strength of the boom was underestimated, especially its impact on capital taxes. The expenditure ratio in 2000 and 2001 was also less than projected in these years, driven by lower than anticipated unemployment benefits and interest payments.8

18. The fiscal deterioration took place gradually as successive budgets lowered their surplus projections over time. While the Fall 1999 budget projected a general government surplus of almost 3 percent of GDP for 2002, the outturn was almost 2 percentage points lower.9 Most of this is due to the central government: as earlier budgets predicted a surplus in the range of ½–1 percent of GDP, the outturn was a 1¼ percent of GDP deficit. The picture for 2003 is even more stark, with the projected general government surplus declining by 4 percent of GDP over the course of seven consecutive budgets. While most of this again is due to the central government, the pension system surplus was also revised down over time. The declining surplus projections reflect both forecasting errors related to the strength of the boom and new discretionary measures announced in successive budgets.

19. Revenue projections were consistently revised downwards in 2002 and 2003. For 2002, general government revenue was ¾ percent of GDP lower than first projected; this rises to 1 percent of GDP for the central government. A similar pattern emerges for 2003, with the difference between first and last projections reaching 1 percent of GDP for general government, and twice that for the central government. One possibility is that the authorities over-estimated the permanence of the revenue boom; this was true elsewhere in the European Union as well (OECD, 2002). They could also have over-estimated the strength and durability of output growth. Another complication was that tax cuts in each year were not decided until the Fall budget bill of that year.10 Pension fund revenues were also lower than projected, with forecast errors varying between ¼ and ½ percent of GDP in these years. Performance here was due to a lower than expected return on pension fund investments: lower interest rates reduced revenue from capital from 30 percent of total revenue in 1999 to under 10 percent in 2002. Only in the local government sector were forecast errors positive, and for 2003 this reflects a large tax increase which had not been anticipated.

20. Expenditure for 2002–03 was revised upwards with each new budget. The upward trend in transfers and local government consumption is clear from looking at the pattern of projections over time. While this rising trend can also be seen in 2001, it is more pronounced in later years. For 2002, as the revenue ratio was being revised downwards, the expenditure ratio kept rising—the outturn was more than 1 percent of GDP higher than the earliest projection in the sample. Around 70 percent of this upward revision was due to local government expenditure. However, transfers were also around ¾ percent of GDP higher than initially projected; this increase is partly mitigated at the central government level by a lower interest bill. The revisions to 2003 projections are even more stark, with an expenditure ratio 3¼ percent of GDP higher in the Spring 2003 budget than in the earliest projection. Again, the increase over the course of the projections is due to transfers (2 percent of GDP higher) and local government expenditure (1½ percent of GDP higher). Part of the reason for the larger increase in 2003 is that many of the discretionary measures introduced in successive budgets did not come into force immediately.

21. Local government expenditure was revised upwards steadily, which led to previously unplanned tax increases in 2003. As can be seen clearly in Figures 1 and 2, both revenue and expenditure at the local government level trended upwards over the course of the projections by ¾ percent of GDP in 2002, and 1½ percent of GDP in 2003. When the revenue boom petered out in 2002, the inability to cut spending meant that taxes had to be raised—which is exactly what happened in 2003. There were no plans to raise taxes until relatively recently. Throughout this period, the average tax rate for municipalities and county councils was 30.5 percent. The 2003 budget (Fall 2002) announced a 0.25 percentage point increase to 30.8 percent; this was again revised upwards in the Spring 2003 budget to 31.17 percent, an increase of 0.65 percentage points, the largest such increase in 25 years. Indeed most observers believe that taxes will have to be raised again in 2004 if the balanced budget rule is to be respected.

22. Combined with the balanced budget rule, the local government revenue base fosters procyclicality. As seen clearly over the past few years, the local government revenue boom has translated directly into higher spending.11 As revenue in 2002 and 2003 was revised upwards with each budget, so was expenditure. In essence, a cyclical revenue boom was spent, and the inability to cut spending during the downturn meant that taxes had to be raised to avoid violating the balanced budget rule. The local government balance in Sweden is known to be far less sensitive to the economic cycle than that of the central government (Schimmelpfennig, 2002). Whether this reflects an innate lack of revenue and expenditure cyclical variability12 or off-setting procyclical movements is an open question. In any event, unlike most countries, Sweden stands out in that its local government tax base is uniquely on income; indeed, of the sample of countries in Table 5, Sweden is the only one to rely only on this tax base. Given the cyclical tax base, the balanced budget rule increases the chances of procyclicality dramatically. Dahlberg and Lindstrom (1998) argue that local government consumption is based on permanent, rather than transitory, income. If this is consistent with the recent trend, then local policymakers would have had to believe that the revenue boom reflected permanent rather than cyclical factors.

Table 5.

Sweden: Percentage Distribution of Local Government Taxes by Tax Base 1/

article image
Source: OECD

Data from 1995.

Experience under the expenditure ceilings

23. Increases in discretionary spending were made possible by savings on cyclical components of expenditure during the boom. A number of commentators (Schimmelpfennig 2002; Roseveare, 2002; Heeringa and Lindh, 2001) believe that a major weakness of the Swedish fiscal framework is that cyclical margins under the expenditure ceilings have been used for discretionary spending. From 1998, savings from higher than expected growth and lower than expected inflation were used up in full to increase discretionary expenditure up to the ceiling, leaving almost no margin. The expenditure ceilings have been under threat every year since 1998. The 2002 ceiling was barely met. To prevent the expenditure ceilings being breached for the first time, the Spring 2003 budget announced a series of expenditure cuts and delays, saving 0.2 percent of GDP in 2003 and 0.4 percent of GDP in 2004. Even so, the margin for error is practically non-existent.

24. The ceilings have been respected, but some undesirable accounting practices have begun to creep in. The government bought forward portions of the EU contribution and the acreage subsidy to farmers (SEK 4 billion) to 2002 in anticipation of problems with the ceilings. Many commentators also noted increasing recourse to tax expenditures such as abatement of some social security contributions and general tax relief for municipalities. Specifically in 2003, the employment subsidy to local governments was treated as a decrease in revenue rather than a central government expenditure item, allowing the ceiling to be circumvented. Overall, tax expenditures that are treated as expenditure by the national accounts but are not subject to the ceiling increased to SEK 6 billion (0.2 percent of GDP), from a negligible amount in earlier years. Also, off-budget investment also reached SEK 5 billion (0.2 percent of GDP) in 2003. Looking ahead, credibility could be endangered given that for the first time, the Budget bill for 2003 failed to set an expenditure ceiling for 2005.

25. Transparency regarding expenditures could be enhanced. The expenditure ceiling is adjusted frequently for technical reasons. Differences between state budget and national accounts treatment of expenditure hinders transparency. For example, transfers to local governments under the revenue equalization scheme go through the central government, and so are under the ceiling, but are excluded from national accounts as they reflect internal transfers. Similarly, VAT on government consumption and investment is excluded from the budget, but included in the national accounts.13 Other factors prompting a difference include the treatment of EU transfers and EU fees as part of the budget, which the national accounts does not do. These technical adjustments typically have offsetting effects on revenue and expenditure. Since they are not really part of expenditure, but are treated as part of the budget, the ceilings often need to be adjusted for technical reasons. While valid, such adjustments reduce transparency (see Figure 3 for the difference between budget and national accounts expenditure definitions).

Figure 3.
Figure 3.

Sweden Central Government Expenditure, State Budget Versus National Accounts

(In Percent of GDP)

Citation: IMF Staff Country Reports 2003, 231; 10.5089/9781451835953.002.A001

Source: Ministry of Finance

C. Political Economy and Fiscal Rules

Deficit and expenditure bias

26. Sweden’s history of higher deficits in the past can be explained partly by political economy factors. Over the period 1980–1986, before the most recent reforms, Sweden ran an average general government deficit of 3½ percent of GDP; in the first half of the 1990s, this rose to 8 percent of GDP. While other, external, factors were at play here, the inability of the major agents to coordinate to control spending was a key factor. The basic problem is often seen through the prism of a common pool of fiscal resources, which each group has a claim on, while no single group has an incentive to constrain its spending demands. Such a model can easily explain both an expenditure and a deficit bias. For example, if allegiances are geographical, with the benefits of spending accruing to one region but with all regions paying an equal share of taxes, then the size of government becomes larger than optimal (Weingast, Shepsle, and Johnsen 1981). In a dynamic context, the same kind of conflict between groups can lead to the build-up of debt and delay adjustment in the face of a negative shock (Velasco, 1999).

27. Sweden’s proportional electoral systems increases the risk of poor fiscal outcomes. With different vested interests and weak enforcement mechanisms, coalition or minority governments—typical under proportional electoral systems—are seen as especially prone to larger deficit and expenditure biases. Moreover, these governments are more likely to be unstable and short-lived, making them less likely to consider the long-term impact of their actions. Accordingly, a number of political factors have been found to influence fiscal policy, including the number of parties in government (De Haan, Sturm, and Beikhuis 1999), the number of spending ministers (Kontoupolis and Perotti 1999), government duration (Grilli, Masciandaro, and Tabellini, 1991), frequency of minority governments (Edin and Ohlsson, 1991), and electoral systems that foster proportionality (Persson and Tabellini, 2003). With a proportional electoral system associated with minority governments forced to rely on opposition support, Sweden may have been vulnerable to these effects. Indeed, its acceptance of fiscal rules can be seen as an attempt to bind the hands of policymakers and reduce the deficit and expenditure biases.

28. Cross-country evidence suggests that spending revenue gains during good times are a leading cause of fiscal distress. Fiscal policy procyclicality should be seen as the norm across the world, rather than the exception, even in industrial countries (Talvi and Vegh, 2000). In case after case, the reason is the same: temporary increases in revenues are spent rather than saved during booms, forcing governments to cut spending or raise taxes in recessions to deal with intertemporal budget constraints. Again, the revenue windfall is viewed as common property, which, absent coordination, leads to inefficiently higher spending. Empirically, procyclicality is associated with political constraints and volatile output and tax bases. Talvi and Vegh (2000) even argue that it may be optimal to cut taxes in a procyclical manner if the aim is to fend off a plethora of spending pressures.

29. Consistent with recent developments in Sweden, evidence has shown that politically-induced procyclicality is prevalent among industrial countries. Lane (2001) shows that fiscal policy in the OECD is more procyclical when output is volatile and political power is dispersed. Hercowitz and Strawcynski (1999) demonstrate that high tax revenue in expansions allows short-sighted governments to increase spending which becomes difficult to reverse, especially in downturns; these factors contribute to a steady rise in the expenditure ratio. One key finding is that the savings in transfers resulting from lower unemployment benefits during good times are dissipated as spending on other transfers increases, which fits the recent Swedish experience. Skilling (2001) argues that fiscal policy is loosened in good times, and not tightened in bad times, and that this asymmetric response in the leading cause of debt accumulation. In his words, the “key problem for governments is reigning in profligacy in good times”. Again, he invokes political economy reasoning, arguing that such an asymmetry is more pronounced in divided governments and decentralized fiscal institutions. Linking with the rest of the literature, he finds that it is coalitions rather than single-party governments which are more prone to loosening in good times, and that budgetary centralization can curb fiscal profligacy (see below).

Political economy and fiscal decentralization

30. The common pool problem can also arise under fiscal decentralization. If fiscal devolution means that all expenditure is financed locally, then the only problem of the commons would be at the local level. However, fiscal decentralization typically involves more locally determined expenditure than revenue, with the gap being made up by central government transfers. The larger this vertical imbalance, the more the tendency of local government spending to be inefficiently high, as local policymakers can pass on some of the cost outside the region. This problem is compounded whenever subnational governments have substantial borrowing authority. Pisauro (2001) goes even further and argues that eliminating the vertical imbalance need not prevent local governments from running deficits, if they still think a bailout is a possibility.

31. Empirically, a greater reliance on transfers from the central government has been associated with weaker fiscal outcomes, but less so in OECD countries. In a comprehensive cross-country analysis, Rodden (2002) finds that large and persistent deficits occur when subnational governments are more dependent on transfers, and are free to borrow. Moreover, this confluence of factors is more likely in federal than unitary states. De Mello (2000) finds that dependency on transfers worsens fiscal outcomes in developing countries, but not in the OECD. Similarly, Drummond and Mansoor (2002) find no evidence of a negative effect of decentralization on fiscal outcomes in advanced economies as institutions tend to be strong. Eichengreen and von Hagen (1995) show that borrowing constraints tend to be imposed when there is greater dependence on the center; otherwise local governments would borrow excessively in anticipation of a bailout. Along similar lines, Stein (1998) finds that decentralization leads to larger government size in Latin America when it is associated with a high vertical imbalance and few borrowing constraints.

32. Local government in Sweden is large, and relies predominantly on its own revenue sources. As in other Nordic countries, local self-rule is a well established concept and the right of local governments to levy taxes is enshrined in the constitution. With responsibility for basic public goods provision, the size of local government is substantial in Sweden, comprising over 40 percent of total expenditure and 70 percent of government consumption. Compared with the rest of the EU, the share of local government revenue in total revenue is higher in Sweden. Most of this is financed from locally raised revenue and the ratio of grants to total local revenue is relatively low (Table 6). Given municipalities and country councils raise most of their own revenue, the problems associated with vertical imbalances would not be expected to apply.

Table 6.

Sweden: Comparison of Fiscal Decentralization Indicators Across Industrial Countries 1/

article image
Source: GFS.

Average over the 1990s.

33. However, a generous revenue equalization system leads to perverse incentives and a bias towards higher taxation. Sweden has in place a complicated equalization scheme to redistribute income among the local governments (see Box 1 for a description). This scheme, based on evening out tax bases, reduces the incentive of local governments to increase their own tax base. Indeed, for a locality charging the average tax rate, any direct gain is pretty much offset by either a reduction in the transfer received or an increase in the contribution paid out. If local growth boosts tax revenue, the municipality will lose 95 percent of this additional revenue through either increased fees or decreased contributions. With an effective marginal tax rate of 95 percent, incentives to boost income are limited. Furthermore, contributions from cost equalization will be reduced as the share of unemployed and low income residents fall. There is also little gain from enhancing tax administration or improving the effectiveness with which money is spent (Thakur and others, 2003). However, if the local government increases its tax rate, revenue will increase as any normal decrease in the tax base is offset by the revenue equalization scheme. Hence there is an in-built bias towards increasing taxes in the system, which fits with recent evidence.

34. The revenue equalization scheme also reduces the link between expenditure and local resources. While the overall vertical imbalance is low for Sweden, the scale of revenue equalization means that, for the poorer localities, the link between expenditure and own revenue is limited. While overall the share of grants in local revenue at around 20 percent is low by international standards, this varies from under 10 percent for some of the richer municipalities to 50 percent for the poorer ones (Johansson, 2003). Clearly, the extent of this scheme tempts some local governments to treat revenue as common resources, and increase spending accordingly. From the point of view of the richer municipalities, it reduces their incentive to comply with the balanced budget rule, knowing that much of their resources are not kept locally.

Local Government Revenue Equalization in Sweden

The aim is to redistribute resources from richer to poorer local governments. There are two types of redistribution: revenue equalization, which aims to even out tax bases, and cost equalization which accounts for different cost factors in each municipality and county council. Although mainly rules-based, the system is extremely complicated and not very transparent. What follows is a highly simplistic overview.

The revenue equalization compensates local governments with lower per capita tax bases. 1/ The tax revenue of each municipality and county council can be calculated as follows: the population multiplied by:

(Own tax base) × (Own tax rate) + (Average tax base − own tax base) × (corrected average tax rate)

where tax bases are computed in per capita terms, and the corrected tax rate is 95 percent of the average. The basic idea is that tax bases are smoothed among local authorities.

This scheme is supplemented by cost equalization. The model sums the cost implications of sixteen different factors—including elderly care, schooling, population, streets and roads, water and sewers, and cold climates—and derives a standard cost per capita in each municipality and county council. If this derived structural cost exceeds the average structural cost, then this local government receives a transfer equal to the excess cost.

1/ This explanation is from Thakur and others (2003), chapter 8.

35. A low vertical imbalance does not necessarily eliminate bailout risk. As documented in von Hagen and others (2000), the central government in Sweden bailed out two cities in the 1990s which were facing bankruptcy over the housing company indebtedness. By 1998, almost a third of the municipalities had applied at least once for help overcoming a financial crisis. In an empirical exercise, the authors traced the demand for bailouts to the size of municipal debts and found little role for vertical imbalance. Given that the local government sector provides the key public services, the central government may be especially vulnerable to bailout demands. Knowing this increases the local government’s temptation towards profligacy. In the context of Germany, Rodden (2000) demonstrates that a rational bailout risk leads weaker localities to raise expenditures in response to positive revenue shocks, but not to reduce them in the event of negative shocks. There is therefore a good case to be made for borrowing constraints, especially given the incentives under the revenue equalization system.

36. A discretionary element within a grants system or a revenue-equalization scheme increases bailout risk. Although mostly rules-based, the central government can grant supplementary transfers on a discretionary basis to municipalities in economic distress. However, this discretionary element is small, comprising around 2 percent of the revenue equalization grants (Johansson, 2003). Rodden (2001) argues that such a discretionary element—the third component of the German equalization scheme—is a ready-made mechanism for providing bailouts.14 As a result, it leads to excessively high deficits among local governments, as they believe they can pass the costs of their expenditure elsewhere. Even with limited discretion, governments can influence redistribution by changing the rules (Johansson, 2003)—this is facilitated by a complicated and non-transparent equalization mechanism.

The choice of fiscal rules15

37. Countries have resorted increasingly to fiscal rules to curb expenditure and deficit pressures. Drawing a parallel with monetary policy, many observers have argued that to prevent the excessive growth of deficits or expenditure, fiscal policy should be guided more by rules rather than discretion. Towards this end, industrial countries have adopted a diverse set of fiscal frameworks over the past decade, including the adoption of binding fiscal targets, improvements in fiscal transparency, and reforms of budgetary institutions. Sweden’s comprehensive three-pronged framework—comprising a cyclical surplus rule, nominal central government expenditure ceilings, and a local government balanced-budget requirement—is superior to the framework in many other countries.

38. Like Sweden, many countries target the overall balance, usually over the cycle to avoid procyclicality. Many observers have criticized simple balanced budget rules on the grounds that they are inflexible and induce procyclical behavior. For this reason, most of the countries that have adopted a deficit rule apply them over the cycle, to allow automatic stabilizers to operate; besides Sweden, other examples include Australia, New Zealand, Switzerland, and the euro area. One advantage of this type of rule is that it discourages optimistic growth projections relative to potential (OECD, 2002).

39. Deficit targets have a number of weaknesses. Targeting the overall balance induces procyclicality and does not provide adequate protection against expenditure pressures during good times. Moreover, targets can always be met by raising taxes, which could have adverse growth consequences (OECD, 2002). The flexibility granted by cyclical balance rules may come at the expense of credibility. Estimating the effects of the cycle or the structural balance is an imperfect exercise, relying on an array of technical assumptions. The benchmark by which fiscal policy is judged is thus not fully transparent. In Sweden, as in other countries, there is also an accountability issue as output gaps are not part of official statistics.

40. Expenditure rules are effective in promoting discipline as they attack the problem at its source. The expenditure rule is at the heart of Sweden’s fiscal framework. Given that the politically-induced deficit bias typically results from expenditure pressures, expenditure rules force the players to internalize budget constraints explicitly. Being relatively simple to understand—unlike cyclical balance rules—they foster transparency. Moreover, they can be designed to allow automatic stabilizers to operate without constraint, as the bulk of the automatic stabilizers operate on the revenue side.

41. Like Sweden, a growing number of countries have chosen various forms of numerical expenditure rules in recent years. There are many ways to design an expenditure rule; see Table 7 for a comparison between Sweden, Netherlands, and Switzerland. While Sweden opted for nominal ceilings, the Netherlands sets real ceilings four years in advance, transforming into nominal ceilings in the budgetary year with the GDP deflator. In Switzerland, the government sets overall spending limits while parliament chooses the composition (Danninger, 2002). In the United States, the 1990 Budget Enforcement Act (BEA)—which imposes caps on discretionary spending and forces any new increase in mandatory spending to be accompanied by offsetting reductions elsewhere—is credited with a dramatic fiscal turnaround in the last decade.

Table 7.

Comparison of Expenditure Rules

article image
Source: Danninger (2002).

42. While expenditure rules should be as broad based as possible, there are certain categories of spending that could be justifiably omitted. Interest payments could be excluded, as they are not directly controlled by the government. Unemployment benefits could also be excluded from the ceiling to avoid inducing procyclicality; this concern could be taken care of with a cyclical margin, however. Here, the temptation will arise to spend up to the ceiling even in favorable cyclical conditions, as was the case in both Sweden and the Netherlands. If public investment is deemed productive, then it may make sense to exclude it too from the rule16, especially since most fiscal adjustments target public investment rather than the politically-sensitive items like the wage bill and transfers. However, the distinction between current and capital expenditure is sometimes murky. Problems can also arise if certain items of expenditure are subject to large forecast errors, an example being sick leave costs in Sweden (Heeringa and Lindh, 2001). In general, expenditure rules are prone to the temptation to resort to creative accounting, although this has been limited in most countries with such rules.17 It also makes sense to set multi-annual targets, to avoid moving expenditure between years to circumvent the rule.

43. Expenditure rules can also be cast in real or nominal terms, or in terms of ratios or growth rates. A nominal target is transparent, and can keep expenditure under control if inflation is higher than expected (European Commission, 2003). On the other hand, the use of nominal ceilings has its weaknesses in that adjustment is needed when the price of government expenditures is higher than expected. Real targets, while less transparent, are not affected by inflation. Setting an expenditure target as a percent of GDP risks procyclical behavior. This problem can be overcome by setting the rule in terms of fixed growth rates, which again, can be either real or nominal. Growth rate targets tend to be more popular than ceilings—Belgium, France, and Denmark have relied on real growth targets while Germany, Ireland, Italy, and Luxembourg have targeted nominal expenditure growth. At the same time, lacking the credibility of a legislated numerical ceiling, these rules have been less successful in curbing expenditure growth.

44. Transparency should go hand-in-hand with well designed fiscal rules. Countries such as the UK, Australia, and New Zealand include transparency provisions in their fiscal frameworks which can loosen the tension between credibility and flexibility. Basically, although fiscal policies should be designed with the rules in mind, it is possible to deviate from them at the implementation stage so long as the governments explain why and how this is being done. Numerical rules may be set, but they are not necessarily legislated. The reputation effects of transparency lead to more accountability among policymakers. Sweden’s fiscal framework is quite transparent in this regard. However, the relationship between the expenditure and surplus rule is not totally clear. The surplus target is evaluated using national accounts data, where the treatment of expenditure is quite different from the state budget. As noted by Schimmelpfennig (2002), there is no explicit link between the expenditure ceiling and the surplus target.

45. For credibility, rules should ideally build in sanctions for non-compliance. Many fiscal rules are defined at the budget rather than the outcome stage. Examples of sanctions include an obligation to amend the budget law, automatic sequesters, or pecuniary sanctions imposed by a higher level of government or external monitor. To enhance their effectiveness, sanctions could have a legal or constitutional basis—such as in Switzerland—rather than a mere reputational one, as is the case in Sweden and most other countries. The European Commission (2003) contends that a weakness of most fiscal rules is that they tend to be based on political commitments rather than legislation. Nonetheless, a credible political commitment could also act as a disciplining device. Certainly, the Swedish government invested a lot of credibility in the expenditure ceilings.

46. Countries have different rules dealing with unexpected shocks during the implementation of the budget. Table 8 provides an analysis of the mechanisms EU countries have in place to deal with both positive and negative shocks. Sweden has limited provisions, including some ability to carry-over unspent funds into future years and also some borrowing facilities in the case of expenditure over-runs. If necessary, a supplementary budget must be presented during the fiscal year. Belgium, Denmark, Finland, Luxembourg, Netherlands, and Portugal all have rules to deal with negative shocks. Belgium, Luxembourg, and the Netherlands, have more formal rules for dealing with positive shocks to the budget. These rules typically detail how such shocks should be apportioned between financing, expenditure and taxation. For example, the Dutch system lays out clear rules for dividing surpluses into debt reduction and tax cuts, and deficits between expenditure adjustments and financing. Switzerland also has formal rules to deal with deviations from the budget.18 A rule to deal with a positive shock is a direct way of preventing the oft-observed pattern of procyclicality in good times.

Table 8.

Formal Rules to Deal with Unanticipated Shocks

article image
Sources: Hallerberg, Strauch, and von Hagen (2001). Swedish information from Ministry of Finance.

47. For local governments, the most pervasive fiscal rule is a borrowing constraint. Control of borrowing at the subnational level can take many forms, including sole reliance on market discipline, negotiating with the center, or resorting to clear rules (see Ter-Minassian and Craig, 1997). Restricting local government autonomy is more the norm in unitary than federal states. European Union countries and US states all apply some kind of borrowing restrictions. However, even if borrowing constraints keep budgets balanced, local governments can often obtain higher grants or shared taxes from the center. A balanced budget rule also risks generating procyclical behavior.

Fiscal rules and fiscal institutions

48. The adoption of fiscal rules went hand in hand with budgetary and institutional reforms. If there is insufficient coordination in the budgetary process, spending ministers may fail to internalize the costs of their demands on society as a whole. Appealing again to the common pool model, a more fragmented budget process, with more autonomous agents making spending decisions, will lead to a deficit or expenditure bias (von Hagen and Harden, 1996; von Hagen, Hughes Hallett, and Strauch, 2000). Sweden’s critical 1997 reform followed a European trend towards improved institutions designed to reduce fragmentation in the budgetary process19 (see Hallerberg 2003 for details). Not all of these countries underpinned their reforms with formal fiscal rules.

49. There are a number of strategies that can be followed in reforming budgetary institutions. In a nutshell, there are two basic paths to budgetary centralization (Hallerberg and von Hagen, 1999; Hallerberg, Strauch, and von Hagen, 2001; Hallerberg, 2003). Under delegation, the finance minister assumes a leading role in the budget process, from negotiation through design and implementation, and is vested with agenda setting powers and strong monitoring capacity. Under commitment, on the other hand, the different parties negotiate strict budget targets. Such targets typically take the form of binding spending targets for the individual ministries. Negotiation enables players to overcome the common pool problem and achieve budgetary centralization. In both cases, the coordination problem underlying the common pool model can be overcome. Delegation is more suited to single-party governments, where there is usually no major policy difference between the finance minister and other ministers. This level of trust will not exist in a coalition, making commitment the more logical choice. The threat of breaking up the coalition serves as the enforcement mechanism.

50. Best practices differ between delegation and commitments states.20 Delegation states should cede authority in both setting and implementing the budget to the finance minister. Commitment states should have clear multi-annual budget plans and fiscal rules to deal with unexpected shocks during the implementation of the budget. Legislated expenditure ceilings are especially useful in commitment states. Formal rules are not as important in delegation states, so long as a strong finance minister is empowered to make the necessary adjustments. Minority governments are typical in Sweden. In this case, the government must first propose a budget, and then negotiate with parliament to make sure the budget is approved. Hence Sweden is what Hallerberg (2003) refer to as a “mixed” system, where the ideal solution would look like a cross between the two polar cases of delegation and commitment. Under this framework, the finance minister must have a strong position in setting the budget, and there must be a negotiated agreement with the opposition. Therefore legislated expenditure ceilings are appropriate.

51. Reform in Sweden was prompted by weakening budgetary institutions. Sweden was traditionally a delegation state, for while it experienced mainly minority Social Democratic party governments, these governments could rely on almost unconditional support from the smaller Communists. This arrangement broke down in the 1980s as Communist support became increasingly conditional and internal conflicts within the governing party weakened the position of the finance minister and strengthened vested interests in parliament (Hallerberg, 2003). Moreover, increasing electoral competition over this period nudged expenditure upwards. In the aftermath of the banking crisis in 1992, the economy spun into deep recession, and the public finances deteriorated dramatically—the general government deficit approached 12 percent of GDP in 1993. Policymakers recognized that fiscal consolidation had to go hand in hand with institutional reform.

52. Expenditure ceilings formed the linchpin of the reform strategy. Ceilings were introduced in 1997, with 27 individual caps on expenditure items which, along with a margin, sum to the overall ceiling. The reforms also strengthened the role of the finance minister within the government. Under the new regime, the opposition can present alternatives, but they must present a complete new package, forcing parliament to think in terms of the total budget. To change the expenditure ceilings, the opposition in parliament must unite around an alternative proposal. It became much harder for the opposition to defeat the budget, as parties found it difficult to unite on an alternative (Heeringa and Lindh, 2001; Molander 2000). Since 1998, the government negotiates its budget with the opposition Left and Green parties.

53. Sweden’s reforms are compatible with best practices, but it does not have a rule to deal with unexpected shocks. Sweden’s reforms are appropriate for its type, strengthening the position the finance minister, and providing a framework for negotiating with the opposition. However, they did not incorporate a rule for dealing with unexpected shocks. While suitable for all “commitment” arrangements, it is especially important for minority governments in “mixed” systems to adopt this kind of rule, to avoid the temptation to placate the opposition simply to stay in power.

Are fiscal rules effective?

54. Part of the credit for the dramatic improvement in fiscal performance in the 1990s belongs to fiscal rules, and to better institutions in general. The average general government deficit across OECD countries fell by over almost 3 percentage points between the first and second half of the 1990s. In Sweden’s case, the turnaround in the overall balance was even more dramatic—it grew by almost 9 percent of GDP between the pre-rules period (1991–1997) and the post-rules period (1998–2002). Although there is no counterfactual, it is clear that Sweden’s strong framework played a large role in restoring solvency. Taking a more skeptical position, the European Commission (2003) argues that expenditure rules have typically not restrained expenditure growth given that they are either not ambitious enough, or easily abandoned. However, the Commission adopts a rather loose definition of expenditure rules, encompassing everything from a legislated ceiling to a vague growth norm. This finding only bolsters the case for a more formal fiscal framework when needed.

55. Stronger budgetary institutions are associated with better fiscal outturns. Sweden’s institutional reform impacted on the fiscal outturn in the post-reform years. A number of studies highlight a notable correlation between the degree of budgetary centralization and fiscal outturns, defined in terms of expenditure, deficits, and debt in European Union countries (Von Hagen, 1992; de Haan and Sturm, 1994; Von Hagen and Harden, 1996). Budgetary institutions matter in other parts of the world as well: there is a large body of evidence in the US highlighting the link between stronger institutions and more favorable fiscal outcomes at the state level.21 In Latin America, Alesina and others (1999) and Stein, Talvi, and Grisanti (1999) also show a link between more transparent and more hierarchical procedures with greater fiscal discipline. Brumby and Cangiano (2001) maintain that the recent wave of consolidation is likely to prove more durable than in the past given that it has been accompanied by deep rooted public expenditure management reforms.

56. Transparency in itself is also important. As noted, some countries chose to underpin their new fiscal frameworks with transparency, rather than rely on explicit targets. Sweden’s framework scores highly on transparency grounds. Drawing on an index of transparency for nineteen OECD countries, Alt, Lassen, and Skilling (2000) document a negative relationship between transparency and debt levels. They argue that transparency is more valuable in countries like Sweden where one party is in power most of the time.

57. Recent trends suggest that fiscal rules may not have addressed the problem of procyclicality. The pattern observed for Sweden in the last section—a revenue boom leading to a procyclical expansion in discretionary spending—was a common one. Across the European Union, while the level of primary expenditure decreased by around 9 percent between 1992–2000, it increased by 5 percent in the two subsequent years (European Commission, 2003). The new fiscal frameworks, even the better designed ones, failed to prevent expenditure increases during the recent boom and deteriorating balances thereafter. In other words, the rules failed to insulate the economies against the well-documented pattern of procyclical loosening during good times. Moreover, as in Sweden, countries tended to view tax receipts from booming equity markets as more permanent than they actually were.

58. Some evidence shows that countries with the weakest fiscal position are countries which do not follow the best practices appropriate to their type. Hallerberg and von Hagen (1999) argue that either delegation to a strong finance minister or adoption of negotiated spending targets is the key to healthy public finances. The connection between fiscal rules in themselves and budget discipline was not that strong in the late 1990s; what seemed to matter more was adopting the rule appropriate to the country’s particular institutions (Hallerberg, Strauch, and von Hagen, 2001). Sweden’s choice of fiscal rules was also appropriate for its “mixed” type (Hallerberg, 2003). So when some observers point out that legislated fiscal rules are not necessary for fiscal success, as both countries with and without rules both achieved significant consolidation (Kennedy, Robbins, and Delorme 2000), they are not adequately distinguishing between different institutional types, some more amenable to legislated targets than others. Indeed, the states with the loosest fiscal policies in the late 1990s were delegation states with weak finance ministers and strong parliaments, and commitment states without rules to deal with unexpected shocks in the implementation of the budget.

59. Local government borrowing restraints have been effective in ensuring budgetary balance, but sometimes at the expense of procyclical outcomes. Borrowing restraints across industrial countries typically lead to roughly balanced local government budgets. Within the European Union, Germany is a notable exception, where the local governments account for nearly half of the general government deficit (European Commission, 2003). As in Sweden, borrowing constraints could lead to procyclical fiscal policy. Rodden (2001) found clear evidence of procyclical revenues and expenditures among German local governments.

D. Prospects for the Future

60. While generally well-designed, Sweden’s fiscal framework could be made more robust. Sweden’s fiscal framework—geared towards controlling expenditure in a manner consistent with a medium-term surplus target—is generally well designed. The expenditure rule in particular is appropriate to its institutional type. The emphasis on transparency is also important. However, some weaknesses are evident, and rectifying these weaknesses—consistent with theoretical literature and international experience—would strengthen the framework.22

61. The following reforms would improve the central government fiscal framework:

  • The cyclical margin under the expenditure rule should not be used for discretionary spending. Evidence points to the need for a rule to deal with unexpected shocks, especially in good times, when growth is high and inflation is low. A ban on eroding the margins is a small step in the right direction.

  • A bolder reform would set a rule spelling out how to deal with unexpected shocks. As in other countries, Sweden could set a formal rule detailing how to respond to both positive and negative shocks. Coping with positive shocks is more pressing, given the emphasis in the literature and what happened in Sweden over the past few years. It is also in accordance with best practices for Sweden’s particular form of fiscal institution. Such a rule could specify how revenue shocks could be apportioned between saving, spending, or cutting taxes. A similar rule could be set up to deal with negative shocks, in a way that avoids procyclical behavior. This would also have the benefit of automatically making the link between the ceiling and the surplus target more explicit.

  • The expenditure ceilings could be set using the same methodology as the national accounts. This would increase simplicity and transparency, and would make for an easier link to the surplus target. It would also make it far more difficult to use tax expenditures and other methods of circumventing the ceilings.

  • The link between the expenditure ceiling and the surplus target could be made more explicit. It should be clear and transparent what compliance with the medium-term surplus target means on an annual basis.

  • Enforcement mechanisms could be strengthened. One option would be to follow Switzerland and move from reputational to legal enforcement. However, this may not be needed as the rules are both credible and transparent, insofar as the government has shown its commitment to them. However, it may be an option down the line if ceilings become difficult to maintain.

62. The following reforms would improve the local government fiscal framework.

  • The balanced budget rule should be retained. Despite problems of procyclicality, and questions about whether it is even needed given low vertical imbalance, the rule still serves a useful role. Bailout expectations are not negligible, given the politically sensitive nature of locally provided public services, and the nature of the revenue equalization scheme.

  • The cyclical sensitivity of local government revenue should be reduced. This would make local government finances more compatible with the balanced budget rule. Sweden’s local government tax assignment is not in line with best practices, which recommend less cyclically sensitive taxes.23 A radical overhaul of the tax system, based on shifting to a property tax base at the local level may not be feasible as it would lead to increased inequality and probably even more revenue equalization. Of course, the magnitude of Sweden’s local government spending may need to be financed by this kind of broad based revenue (income tax). In this case, a system of shared taxes or grants that corrects for cyclical variability could be designed. The government-appointed “Commission on Stabilization Policy for Full Employment in the event of Sweden joining the Monetary Union” made a few recommendations in this regard, including setting the tax base as the average of taxable income over the cycle, or over a given number of years. Grants could also be adjusted to achieve a stable revenue stream, but this would cut the link between expenditure and own-revenue.

  • The revenue equalization scheme should be modified. In a nutshell, it should operate in a way that does not eliminate the incentive to increase the tax base. It also should be simple, transparent, non-discretionary, and not amenable to frequent change.

  • Enforcement mechanisms could be strengthened. At present, deficits are supposed to be eliminated within two years, but there is little enforcement. The government could devise a system of penalties for failing to comply.

Appendix I

Appendix Table 1a.

Sweden: Fall and Spring Budget Projections, 2000–2003 1/

(In billion of SEK)

article image
Source: Ministry of Finance

FA and SF stand for Fall and Spring budgets respectively.