This Selected Issues paper contains two studies examining key issues for fiscal management and long-term fiscal sustainability in Spain. The first study discusses how best to ensure fiscal discipline at lower levels of government by examining the institutional setting and mechanisms that make this task particularly challenging in Spain’s highly devolved political and fiscal system. The second study seeks to analyze the potential macroeconomic impact of different approaches to deal with the fiscal costs of aging in Spain.


This Selected Issues paper contains two studies examining key issues for fiscal management and long-term fiscal sustainability in Spain. The first study discusses how best to ensure fiscal discipline at lower levels of government by examining the institutional setting and mechanisms that make this task particularly challenging in Spain’s highly devolved political and fiscal system. The second study seeks to analyze the potential macroeconomic impact of different approaches to deal with the fiscal costs of aging in Spain.

I. Fiscal Discipline at Lower Levels of Government: the Case of Spain1

1. Over the last twenty years, Spain has reached a very high degree of fiscal decentralization. In 1985, sub-national governments accounted for 14½ percent of expenditure; in the 2005 budget, the same figure is around 49 percent. This is among the highest in industrial countries. Excluding social security and debt service, the central government will control less than 30 percent of public expenditure in 2005.

2. Such a high level of devolution poses challenges for the conduct of fiscal policy. In 2003, 11 of the 17 regional governments ran a deficit despite the fact that the Budgetary Stability Law (BSL, Ley de Estabilidad Presupuestaria), which was first applied in that year, mandated a balanced budget or a surplus at the sub-national level. While the total regional deficit remained relatively contained in terms of national GDP, the relatively widespread violation of the budget target in the first year of the law’s application is seen by the government to raise questions about the framework’s effectiveness in ensuring fiscal discipline. The government is thus reconsidering the present legal framework, with a view to increasing ownership and observance by the regions, as well as providing explicit room for countercyclical action. In addition, the present regional financing system is being revisited, with some regions claiming that public health expenditure, which was fully devolved to all regions as from 2002, is under-funded, especially going forward.

3. This paper analyzes various aspects of regional fiscal finances in Spain. The first part of the paper describes the institutional framework, including the constitutional provisions on fiscal regional autonomy, the laws regulating regional budgets, and informal but all-important operating rules. The second part highlights the characteristics of the regional business cycle in Spanish regions compared with other European economies. The third section summarizes the experience of other countries in monitoring sub-national borrowings. The last section discusses some possible approaches to ensuring fiscal discipline at lower levels of government in Spain.

A. Institutional Framework

4. The process of fiscal devolution in Spain, which started with the promulgation of the Constitution in 1978, has been driven largely by political considerations. After a period of strong centralization during the Franco regime, the prevalent view within Spanish society was that the new democratic state should be characterized by a decentralized organization. The new Constitution enshrined the classical principles of federal democracy: the competences of the central government and the autonomous communities (Comunidades Autónomas, henceforth CCAA or regions) are clearly determined, locally elected representatives govern the CCAA, and the Constitutional Court resolves possible conflicts between the central government and the CCAA. The federal political organization enshrined in the Constitution has important economic implications.

5. The Constitution clearly specifies expenditure responsibilities. The Constitution specifies that CCAA may assume competence in several areas including education, health, public works in their territory, and the environment.2 The central government has exclusive competence in a set of functions, which include, inter alia, general coordination of economic activities, general finance and public debt, regulation of the financial markets, and the monetary system (Article 149 of the Constitution). Article 149 also states that: (i) any matter that is not defined by the Constitution as reserved to the central government belongs to the CCAA as stipulated in their respective Statutes of Autonomy (Estatutos de Autonomía); (ii) all matters not envisaged by the Statutes of Autonomy shall belong to the central government; and (iii) central government legislation shall be supplementary to CCAA’s laws and prevail in cases of conflict.3

6. In contrast to the exhaustive list of expenditure responsibilities, the Constitution sets only general principles on the financing of CCAA:

  • Financial autonomy. Article 156 enshrines the principle that CCAA must enjoy financial autonomy for the exercise of their functions. This principle is the legal basis for the devolution of taxes and for the existence of transfers to regions.

  • Financial solidarity among regions. This principle, enshrined in Article 2, must be guaranteed by the central government (Article 138), and provides the legal basis for the solidarity or sufficiency fund.

  • Coordination. Article 156 specifies that the right to fiscal autonomy must be exercised in conformity with the principle of coordination with the central government Treasury and in solidarity with other CCAA.

  • Limit to tax autonomy. The Constitution also establishes that the legal power to establish taxes belongs exclusively to the central government (Article 133). Consequently, the CCAA may establish and levy taxes only within the limits established by the

  • Constitution; in particular, CCAA cannot tax properties outside their jurisdiction, or levy taxes that hinder trade (Article 157), and taxes must be based on the principle of equity and progressivity (Article 31). Finally, CCAA cannot impose taxes on goods already taxed by the central government.

The implementation of these general principles was left to organic and ordinary laws.4

7. The asymmetric treatment of expenditure responsibilities, which are clearly detailed by the Constitution, and their financing, which is regulated only by general principles, has created an ‘institutional’ need for a periodic reconsideration of the financing agreements between the central government and the CCAA. After a transitory period between 1979 and 1986, the central government and the CCAA have regulated their financial relationships every five years in the institutional setting of the Fiscal and Financial Policy Council (Consejo de Politica Fiscaly Financiera; FFPC hereafter).5 The latest agreement, concluded in 2001, differently from the past, was intended to be permanent.

8. Against this backdrop, in order to consolidate the fiscal stabilization undertaken in the late 1990s, the government enacted the Budgetary Stability Laws (Ley de Estabilidad Presupuestaria, Ley 18/2001; and the Organic Law 5/2001; BSL, hereafter), which introduced the principle that the central administration (defined as the central government and social security) and the CCAA cannot run a deficit; in addition, the BSL established formal processes to enforce the fiscal stability targets (see below).6

9. Under the 2001 agreement, the sources of financing for the “common regime” CCAA are:7

  • Shared taxes. The CCAA are assigned the following amounts:

    1. 33 percent of personal income tax revenues. Moreover, CCAA have the discretion to increase the marginal personal income tax rate. In practice, no CCAA has used this possibility;

      In addition, the CCAA may introduce new deductions to the base of the personal income tax. CCAA have used this faculty often, introducing, inter alia, deductions to increase natality. However, the revenue consequence of these deductions are estimated to be limited.

    2. 100 percent of the tax on retail sales of hydrocarbon fuels; and

    3. 100 percent of special taxes on certain means of transportation.

  • Transferred taxes. CCAA receive the following amount without the possibility of modifying the tax rate or the legislation:

    1. 35 percent of VAT revenues;

    2. 40 percent of some excises; and

    3. 100 percent of taxes on electricity.

  • Regional taxes. CCAA have exclusive power to tax gifts and bequests, wealth, legal documents, and gambling wins. Starting in 2002, the CCAA have freedom to set the tariffs and the exceptions for these taxes.

  • Transfers from the central government. The major part of central government transfers to the CCAA is channeled through the sufficiency fund, which is intended to cover the gap between the mandatory expenditures and revenues accruing to the CCAA. The initial amount of the sufficiency fund was calculated as the difference between the estimated cost of the CCAA mandates and revenues calculated in the year 1999. Every year this amount is increased at the same rate as the central government’s tax revenues.8

  • If the other sources of financing, including own and shared taxes, were larger than a region’s expenditure responsibilities, the region is a contributor to the central government’s treasury. The amount devolved will be equal to the amount devolved in 1999 (the base year) increased by the rate of growth of the national taxes or the regional taxes, whichever is lower.9

  • The above formula is independent of the actual costs of providing services in the year after 1999. Regions which are efficient and lower the cost of providing the services, can retain the totality of resulting savings in the future.

  • Finally, the law mandates that the formula regulating the regional allocation of the sufficiency fund should be revised if new expenditure responsibilities or taxes are devolved to the CCAA. There is however no automatic revision after five years, as was customary under the previous agreements. In addition to the sufficiency fund, some CCAA also receive resources from other funds, including the Interregional Compensation Fund (Fondo de Compensación Interterritorial), which is the main instrument for regional development and is coordinated with regional transfers from the European Community (Table 2).

Table 1.

Personal Income Tax Brackets After the 1999 Reform

article image
Table 2.

Allocation of Resources from Sufficiency and Interregional Compensation Funds

(2005 budget, in million of euro)

article image
Source: Staff calculations based on 2005 budget. The regional GDP used to calculate the last column is based on an approximation from regional data for 2003 and the national GDP growth as in the budget document.

10. The 2001 financing agreement does not include (permanent) floors on regional financing that are tied with the economic cycle. In contrast with previous financing arrangements, the agreement was seen to be based on full fiscal “co-responsibility.” In other words, the central government and regions would both face the cyclical fluctuations in revenues. The law established, however, transitory financing floors, which were not as such designed to provide insurance for the regions against economic downturns. Specifically, it introduced the following:

  • Transfers to cover health expenditure would grow at least as much as (nominal) GDP through 2004.

  • In its first year of application (2002), no region would receive less financial resources than it did under the previous regime.

11. The law also provides some compensation for demographic shocks and solidarity. The former consists of a compensating formula that increases the share of the transfer to regions experiencing population growth substantially larger than that of the national average (fondo de nivelación). The solidarity factor established, inter alia, that poorer regions—those with income levels below 70 percent of the national average—would see their transfers increase by at least 20 percent more than the national average.

12. Because of historical developments, the Basque Country and Navarre have a separate financing system (régimen foral). Both regions enjoy considerable freedom in establishing and administering personal and corporate income taxes. Most expenditure responsibilities have been devolved with the only notable exception of pensions, which are still administered by the central social security system. Both regions contribute to the central government with a fixed share of GDP for the general expenditures that remain in the hand of the central government, including defense and nationwide infrastructure.

13. CCAA may borrow from the financial markets but are subject to the central government’s authorization. Formally, the FFPC establishes, by the first quarter of every year, the borrowing limits for each CCAA for the following three years.10 In addition, long-term debt may be issued only to finance capital expenditure and the debt burden may not exceed 25 percent of current revenues. However, in practice, the FFPC has set a balanced budget target for every CCAA.

14. Enforcement mechanisms are weak. If a CCAA does not comply with the budgetary stability laws it must present a three-year adjustment plan to the FFPC following the year of the violation. However, the central government has only two legal instruments to ensure adherence to the plan: (i) in the event that Spain were to be fined by the European Union for violation—under the excessive deficit procedure—of the Stability and Growth Pact, and that such violation was attributable to regional deficits, the fine would be charged to the errant CCAA in an amount proportional to its/their contribution to the excessive deficit; (ii) the central government must consider compliance with the BSL before authorizing the issuance of debt by a CCAA. The first enforcement mechanism is unlikely to be called upon since it is quite improbable that Spain will be fined for running a deficit larger than 3 percent of GDP in the foreseeable future. The second mechanism can be quite easily circumvented as CCAA can create semi-private companies, which are difficult to monitor. In addition, the fiscal outcomes for individual CCAA are available only nine months after the end of the fiscal year, so that adherence to the BSL can be verified only with some delay and a possible adjustment plan can start only about two years after occurrence of the violation.

15. The formal rules described above are only part of the institutional framework; informal or unwritten practices are as important in determining the functioning of decentralization in Spain. Informal features are important because the decentralization process has been largely motivated by political considerations as the country moved from a highly centralistic and undemocratic regime toward democracy (Pérez-Díaz, 2004).11 The main informal features of the system can be summarized as follows:

  • Regional authorities are unwilling to use their power to increase regional taxation. For instance, no CCAA has used the legal power to increase the marginal rates for the personal income tax. In contrast, all CCAA have used their power to extend deductions to the personal income tax base to particular groups. Moreover, only a few CCAA have used the faculty of imposing a surcharge on hydrocarbon fuels to increase funding for health expenditure.

  • Regional governments seem more subject to pressure of special groups than the national government. For instance, CCAA are more willing to restrict competition through regulation than the general government. A clear example is the special taxation and the regulation concerning large distribution chains in some CCAA.

  • Regional authorities have close links with the local financial system, given representation on the board of directors of local saving banks (cajas). Though strict regulations avoid direct conflicts of interest, informal avenues of influence remain potentially available.

  • Timely information on budget execution of individual CCAA is not released. This lack of transparency is partly due to understandings under which the information is provided to the central government.

  • Scarce interregional labor mobility has limited the scope for the race to the bottom in the provision of public services, which is the traditional view of decentralization (Oates, 1972), and for welfare enhancing tax competition, which has been highlighted by Brennan and Buchanan (1980).

  • Finally, the central government has at times depended on the parliamentary support of regional parties, whose main objective is the defense of regional interests.12 Moreover, regional authorities’ votes in the FFPC have tended to reflect, where applicable, nationwide party affiliation rather than pure regional considerations. This political configuration contributes to an unwillingness of the central government to go against its political allies at the regional level.

B. Regional Shocks

16. Different institutional frameworks may produce very different outcomes if the economic fundamentals are different. This section analyzes the extent of regional disparities and regional shocks and their persistence in Spain compared to other large European countries.

17. Spain presents larger regional disparities, defined as the standard deviation of regional (log) real income per capita, compared to other large European countries. Figure 1 illustrates the income disparity across regions in Spain, France, Germany, Italy, and the United Kingdom.13 There is furthermore no clear evidence of convergence in regional incomes over time (Decressin, 2002). In response to this persistent differential, which is viewed as a major challenge for the development of Spain as a whole, an Interregional Compensation Fund was created. The Interregional Compensation Fund is intended to finance public infrastructure and often complements transfers from the European Union.

Figure 1:
Figure 1:

Regional Income Dispersion

Citation: IMF Staff Country Reports 2005, 057; 10.5089/9781451934878.002.A001

18. In addition to a high degree of regional income dispersion, Spain also exhibits a relatively high volatility of regional incomes. Figure 2 shows that Spain’s regional income shocks, measured by the standard deviation of de-trended regional growth for each year, are among the highest in large European countries.

Figure 2:
Figure 2:

Regional Shocks

Citation: IMF Staff Country Reports 2005, 057; 10.5089/9781451934878.002.A001

19. Despite being relatively large, Spanish regional shocks are not particularly persistent by international comparison. In order to calculate the persistency of regional shocks, the following regression for the regions of a selected number of large European countries was run:


20. The coefficient β measures the persistency of a regional idiosyncratic shock, defined as the difference between annual regional growth and national growth controlling for the regional growth. If the parameter β is close to 1, regions which experienced ‘excessive growth’ with respect to the national average and their own trend will continue to experience higher growth during the following year. The size of the regional shock is the standard deviation of the regional growth in excess of the national average (GrowthregionalGrowthnationalt). The size of the national shocks is defined as the standard deviation of national income growth; the persistence of national shocks is defined as the coefficient β in the following regression:


Table 3 presents the results.

Table 3.

National and Regional Shocks in Large European Economies

article image
Sources: Staff calculations based on EUROSTAT data.Note: The size of regional shocks is defined as the standard deviation of regional income growth in excess of the national average and regional trend. To obtain the persistence of regional shocks, we run a regression of regional income growth in excess of national average and regional trend on its lag. The coefficients on the lag are reported in the third column and its t-statistics are reported in parentheses. The size of national shocks is defined as the standard deviation of national income growth in excess of the national trend. To obtain the persistence of national shocks, we run a regression of national income growth in excess of national trend on its lag. T-stat are in parentheses.

21. Volatile regional growth does not automatically translate into volatile tax revenues for local authorities. As is common in other countries, the structure of sub-national taxes has delivered more stable revenues to the CCAA than to the central government.14

22. In summary, Spanish regions exhibit large and lasting disparities in terms of income per capita, and a high volatility of regional growth, but a comparatively low persistence of these idiosyncratic shocks. This is consistent with a fair amount of regional specialization, with some regions highly specialized in tourism and others in agriculture. At the same time, this specialization may explain the comparatively temporary nature of the shocks, since agriculture and tourism are more subject to temporary weather-related shocks than sectors such as manufacturing. The scarce persistence of local shocks would suggest a potential role for “rainy day funds” in the CCAA (see below).

C. International Comparison

23. The purpose of this section is to evaluate Spain’s decentralization vis-à-vis international experience. Spain’s decentralization is notable for a number of reasons:

  • Decentralization, measured as share of sub-national spending in general government spending, has proceeded more aggressively in Spain than in other countries: from being one of the most centralized countries, Spain has become one of the most decentralized.

  • Sub-national expenditures have increased relatively more rapidly than those of the central government while sub-national revenues have lagged. This is a common trend in almost all OECD countries but has been particularly marked in Spain.

  • The share of sub-national employment in total public employment has increased greatly and now is well above the OECD average. With the devolution of health expenditure to the remaining CCAA, this share increased further starting in 2002.

24. The relatively low level of sub-national revenues in Spain is notable especially because Spanish CCAA have relatively larger taxing capacity than sub-national authorities in other large European countries. Joumard and Kongsrud (2003) calculate that Spanish sub-national governments could raise taxes equivalent to 2.9 percent of GDP in 1995 compared with only some 1½ percent of GDP in Germany, Poland, or the United Kingdom. However, sub-national governments often do not fully use the tax capacity assigned to them; for instance, according to the same study, sub-national governments in Korea, Finland, and Norway show a similar reluctance to using local taxes.

25. As in other countries, transfers from the central government have increased over time. However, differently from other countries, the transfer system provides an incentive to improve efficiency. In general, inter-regional equalization funds risk providing insufficient incentives to sub-national governments to improve the efficiency of tax collection and expenditure, as part of the financing gap is covered through transfers. The issue arises because the effort level of the agent (sub-national government) in collecting taxes and in using resources efficiently is not perfectly observed by the principal (central government). In this case, the optimal insurance literature suggests that the principal should give some incentive to the agent to improve efficiency. Spain’s transfer system aims to address the moral hazard by calculating the needs with respect to a base year and by leaving substantial taxing capacity to finance marginal expenditure. In this respect, Spain compares well with respect to other countries.15

26. The increase in central government transfers has also highlighted the importance of fiscal discipline at the regional level, which countries have managed in different ways. The possible control systems range from purely administrative approaches, in which local administrations are strictly controlled by the general government, to an arrangement without any formal central government coordination (Table 5 provides an OECD classification, based on the categories of Ter-Minassian, 1997; see also Box 1 on market-based discipline). In the middle, there are solutions in which sub-national governments have some level of discretion within some rules. Spain falls into the category in which sub-national governments formally cooperate because all the decisions concerning inter-governmental relationships are taken by the FFPC on which they are represented. In summarizing the experience of the largest European countries with fiscal rules for local government, Rattsø (in Dafflon, 2002) concludes that, overall, European countries have been able to avoid financially unsustainable situations at sub-national level through the use of rules and some discretion.

Table 4.

Comparison of Decentralization in Selected Countries

article image
Sources: Joumard and Kongsrud (2003), Spanish Ministry of Finance, staff elaboration.Notes: Data for the share of sub-national spending for Spain refer to the years 1982 and 2003, the data for Germany refer to 1991 and 2001, data for United Kingdom refer to 1987 and 2000. The data for revenues exclude transfers from other level of governments but include revenues from tax-sharing agreements. Data on sub-national employment refer to the most recent year available.
Table 5.

Instruments to Ensure Fiscal Discipline at Sub-national Level

article image
Source: Joumard and Kongsrud (2003).

27. Countries with fiscal rules for sub-national governments use different targets as well as different sanctions in case a sub-national government deviates from the rule. Table 6 summarizes the experience of several countries.

Table 6.

Fiscal Rules for Sub-national Governments

(Targets and sanctions)

article image
Sources: Joumard and Kongsrud (2003) and staff elaboration.Note: Spain appears in two cells of Table 6, because the current legal framework mandates the presentation of a three-year adjustment plan if a CCAA exceeds its deficit target; moreover, the law establishes that any fine imposed on Spain under the Stability and Growth Pact is shared with the non-complying region(s).

28. Beside the strict design of the rule, the success in maintaining fiscal discipline at sub-national levels depends on many other institutional factors, including whether (Rodden, 2003):

  • the central government can monitor all direct and indirect activities of sub-national governments, including activities through state-owned enterprises and entities;

  • the central government can credibly commit to enforce the rule(s), especially a firm no bail-out condition;

  • sub-national governments are relatively homogenous in economic strength and political power.

Spain meets these criteria only partially because the monitoring of extra-budgetary activities of CCAA is still imperfect, there has not been a long enough history to demonstrate a firm no bail-out commitment, and the CCAA are not homogenous.

Pure Market-based Enforcement Mechanisms

Among existing federations, the United States and Canada use almost pure market-based mechanisms to instill fiscal discipline at the sub-national level of government. While this solution has the appeal of relieving the central government of the onerous political task of monitoring local authorities, a pure market-based approach is effective only if the central government can credibly commit not to bail out sub-national governments in difficulty.

The central government can earn such credibility through a history of no bail-outs and strong institutions. The case of Canada and the United States provide examples to this effect. In both cases, sub-national governments (provinces and states) existed before the formation of a national government. Moreover, both central governments were “tested” in the XIX century as sub-national governments experienced serious financial crises. During the 1830s and 1840s, the federal government resisted pressures to provide loans to states in financial straits (Rodden et al. 2003).

As a consequence of the strong “no bail-out” commitment by the central government, voters (and tax-payers) in several affected states requested the imposition of rules to impose fiscal discipline on local politicians and avoid the tax hikes associated with irresponsible policies. As capital markets reacted favorably to these innovations, other states introduced similar restrictions. The credibility of the no-bail out clause in the U.S. was further reinforced by the Eleventh Amendment to the U.S. Constitution: “The Judicial power of the United States shall not be construed to extend to any suit in law or equity, commenced or prosecuted against one of the United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.” Under this provision, not only is the central government barred from bailing out states, it cannot even act as a third party to enforce the repayment of state debt. This hard-won credibility has its clear results: studies (Singh and Plekhanov, 2004) show that countries with a history of bail-outs of sub-national governments by the central government (e.g., the Netherlands, Ireland, Mexico, and Italy) exhibit disappointing fiscal outcomes at the local level for a prolonged period thereafter.

In the case of Spain, the possibility of pure market-based enforcement is however weakened by a number of factors. First, the CCAA do not have credible self-imposed fiscal rules; second, regional fiscal accounts are not sufficiently transparent; and third, the level of regional debt is still too low to provide a powerful market-based reaction. Given this, it is not surprising that the rating of regional debt by international agencies is almost the same for all regions. Finally, an indication that Spain is not suited to a pure market-based approach is provided by the fact that market participants themselves, who should provide an independent external discipline, mention the existence of the central government’s oversight as a primary factor for regional fiscal stability (Moody’s Investors Service, Spanish Regional Governments: System Outlook).

D. Conclusions and Policy Options

29. Spain’s current fiscal stability framework, enshrined in the BSL, has appreciable merits. Apart from its achievements at a more general level, including in promoting a broad culture of fiscal stability, it has been reasonably successful also with regard to the subject of this paper—fiscal discipline at lower levels of government. Although 11 regions out of 17 ran a deficit in 2003, these were (with one exception) all below 1 percent of the relevant regional GDP, and the combined CCAA deficit was less than 0.2 percent of national GDP. Moreover, the aggregate debt of CCAA was equivalent to only 5.8 percent of national GDP in 2003. However, spending discipline is not uniformly in place, there is appreciable scope for efficiency gains, and regional accounts suffer from a lack of transparency—including, importantly, with respect to the activities of public enterprises and entities.

30. Enforcement mechanisms for lower levels of government are weak. At the same time, the share of sub-national non-discretionary expenditure, which includes education and health, is large, and population aging is set to raise health spending further (see Chapter II of this Selected Issues paper). The main challenges in reforming the enforcement system for CCAA are:

  • The legal and institutional setting limits the scope of possible intervention by the central government. As a result, the latter cannot credibly commit to enforce rules, especially if these imply harsh consequences (default, fines, etc.).

  • CCAA have access to local financial markets and own public enterprises and other public entities so that market discipline does not appear strong and rules from the center can be circumvented.

  • Timely information on CCAA budget execution is not available (or readily provided) so that many enforcement schemes are de facto cumbersome to apply. For instance, regions that exceeded the deficit target in 2003 are only now presenting adjustment plans for 2005.

31. Some options, possible in other contexts, are not feasible in Spain because of the institutional setting, but disciplining and enforcement mechanisms could be improved along several dimensions within the present institutional framework, through:

  • The publication of timely and comprehensive information on CCAA fiscal outcomes. Full disclosure of fiscal accounts is fundamental for the exercise of effective public pressure by the citizens of a CCAA, who are the final judges of efficiency of local public expenditure. The exercise of such dissuasive pressure is dependent on early identification of unsound fiscal behavior, so as to increase public censure and reputational costs of profligate policies and stimulate appropriate corrective action.

  • The containment of CCAA borrowing. Theoretical arguments and international practices suggest that the demand management (anticyclical) role is best retained by the central government, which controls the appropriate policy instruments (Box 2). In addition, regional public debt is not large enough to provide an effective market disciplining device, and the credibility of the central government’s no bail-out commitment remains to be firmly established (Box 1). Finally, there is an asymmetry in the regulation of local borrowing, which could lead to higher-than-desirable debt over time. This is because the central government, through the FFPC, may authorize borrowing during downturns but does not have the legal authority to impose debt repayments during good times. There is thus an in-built risk toward debt accumulation over time.

  • The introduction of explicit expenditure limits at the regional level. These would be a natural extension of the norms of the BSL for the central government and would provide a transparent way of monitoring sub-national finances. Given Spain’s constitutional set up, such limits would likely need—as in the case of those adopted by many U.S. states—to be self-imposed.

  • The promotion of greater reliance on local taxes to finance discretionary expenditure. While it is important that adequate financing arrangements be in place to meet mandatory spending assignments, both structurally and over the cycle, discretionary sub-national spending is best financed by resources mobilized locally. In this manner, local authorities face stronger incentives to evaluate the benefits of additional expenditure versus the costs of higher taxation, and local accountability is enhanced.16 De iure, Spanish CCAA have considerable scope to raise resources by increasing taxes; however, de facto, they have been notably reluctant to exploit this faculty, tending to press rather for additional central government transfers. Consideration could be given to changing the incentives so that marginal expenditure is financed through local taxes; a possibility could be to tie some transfers to the use of local taxation.17

  • The provision of positive incentives to reward regions that save during good times. From an economic point of view, it is more efficient to save during good times than cutting expenditure during bad times. However, strong political pressures tend to discourage saving during upswings. The present system provides some incentives for saving by CCAA (see Section B), but these incentives could be enhanced through explicit transfers or rewards.

Should Sub-national Governments Run Countercyclical Fiscal Policies?

With almost 50 percent of public expenditure managed by CCAA (and as much as 80 percent excluding social security and debt service), it is legitimate to ask whether CCAA should share the burden of countercyclical fiscal policies with the central government. The role of countercyclical fiscal policy at sub-national level has been thoroughly studied in the context of the U.S. in the 1970s, when the fiscal role of states was becoming more relevant and the prevalent economic thinking strongly favored Keynesian countercyclical fiscal policy. The conclusions of this debate were that the discretionary countercyclical role should remain in the hands of the central authorities for several reasons (Oates, 1972):

  • Fiscal policy multipliers are limited for small open economies such as most sub-national entities.

  • Regional business cycles tend to be highly correlated so that there is scant need for region-specific countercyclical policies.

  • Capital and labor mobility can offset the effects of stabilization policy. The countercyclical effort of sub-national authorities can lead to negative long-term consequences to the extent that the increase in local expenditure in response to a local negative shock lessens the private sector’s incentive to relocate.

  • Sub-national governments cannot take into proper account the externalities that their countercyclical policy can have on other sub-national units, resulting in sub-optimal stabilization efforts.

  • Taxes that are usually controlled by sub-national governments, including property taxes, are by their nature less amenable to change and thus unsuitable as a discretionary tool over the business cycle.

  • Some expenditure that is executed by sub-national governments including education, health, and social services, is not susceptible to discretionary cuts, reducing the scope for active sub-national countercyclical fiscal policy.

  • A final argument—not however applicable to Spain—concerns the policy mix: central governments can coordinate monetary and fiscal policies, a possibility precluded to sub-national governments.

These arguments were originally discussed for the U.S. states, but many of them apply a fortiori for Spanish CCAA, which are relatively smaller and more integrated. In addition, other studies have found that local authorities, being closer to voters, find it difficult to run countercyclical fiscal policies during good times, when surpluses should accumulate. Evidence from the U.S. (Fisher, 1997) and from Russia (Kwon and Spilimbergo, 2005) shows that sub-national governments tend to spend more during periods of high economic growth, leading to procyclical fiscal policies. Finally, no European country has rules for sub-national governments that include explicit provisions for the business cycle (Balassone, 2004).

32. One option to address the latter issue could be provided by the introduction of rainy-day funds. Rainy day funds (RDF), which are part of the fiscal rules of almost all U.S. states, have proven useful in addressing deep but relatively brief fiscal crises at the state level (see Box 3)—to which, according to the analysis in Section B above on regional shocks in an international perspective, Spanish regions may be relatively prone. The benefits of RDF is that they provide a transparent mechanism to save during good times. The central government could provide incentives to the creation of voluntary regional RDF that satisfy some good governance characteristics by, for example, granting a premium interest rate to resources, which are allocated to the RDF up to a certain limit. Even a doubling of the interest rate would not be overly costly. The cost at full regime in the most expensive case in which all regions have adopted the RDF with a cap of 5 percent of expenditures can be estimated to amount to some €430 million, equivalent to 0.05 percent of GDP or less than 2 percent of annual transfers to regions.18

Experience with Rainy Day Funds in U.S. States

Following the recession of the early 1980s which imposed severe fiscal strains, several U.S. states introduced measures to handle the adverse impact of recessions on local public finances. The number of states with rainy day funds (RDF) rose sharply from 12 in 1982 to 38 in 1989, to 45 in 1995. The main purpose of RDF is to smooth public spending during recessions and, possibly, increase public savings over the business cycle. In absence of RDF, states, which are compelled by their constitutions or by statutory requirements to run balanced budgets, would have few instruments to avoid procyclical fiscal policies. The need to smooth expenditure has increased over time as the composition of state expenditure shifted toward non-discretionary spending (in the early 1960s roughly a quarter of state expenditure was on highways, versus roughly 45 percent on public welfare and education; in 2000, these shares were 8 and 65 percent respectively) and because states are increasingly reliant on more volatile source of revenues such as personal income taxes (these accounted for 12 percent of state revenues in the early 1960s and over 35 percent in 2000) (Garrett and Wagner, 2004).

After a prolonged upswing, the recession of 1990-1991 constituted the first test for the RDF. There is indeed evidence that RDF alleviated the effects of the recession, though the positive effects depended crucially on the design of the rules governing the RDF, including the savings and withdrawal requirements (Douglas and Gaddie, 2002). In addition, RDF are associated with larger public savings (Knight and Levinson, 1999). Preliminary results indicate that RDF have fared less well during the state fiscal crisis of 2003-2004, which has been more severe than the crisis of the early nineties. The main problem seems to be the limited resources of the RDF; total RDF balances fell from US$17 billion at the beginning of 2002 to US$8.5 billion in FY 2003, and many states depleted their funds in the middle of the crisis. This experience indicates that RDF cannot be the solution for prolonged fiscal crises, but the analysis in Section II of this paper has shown that, although CCAA are subject to large shocks, these tend to be relatively short-lived.

The design of RDF can differ considerably in the rules that govern the accumulation and withdrawal of funds as well as the size of the cap on accumulated funds. The rules for withdrawing funds from the RDF range from the observance of given economic indicators to the requirement of a (super or simple) majority vote in the legislature. In the case of a simple majority requirement, rainy day funds may be used during eminently “sunny days” by complacent legislatures. As regards the size of RDF, 5 percent of general expenditures tends to be a common cap; there is however no evidence that this size is optimal for all states. In conclusion, the success of a RDF depends ultimately on its features. RDF are often combined with other fiscal rules including expenditure or tax limits. Moreover, all states with the exception of Vermont have some form of balanced budget rule.


  • Balassone, Fabrizio, Daniele Franco, Stefania Zotteri. 2004EMU Fiscal Rules and Fiscal Decentralization.” Presupuesto y Gasto 35(2). pp. 6397.

    • Search Google Scholar
    • Export Citation
  • Brennan, G. and J. Buchanan. 1980. The Power to Tax: Analytical foundations of Fiscal Constitutions. Cambridge. Cambridge University Press.

    • Search Google Scholar
    • Export Citation
  • Dafflon, Bernard. Editor. 2002. Local Public Finance in Europe: Balancing the Budget and Controlling Debt. Edward Elgar. Cheltenham. United Kingdom.

    • Search Google Scholar
    • Export Citation
  • Decressin, Jörg. 2002. “Regional income redistribution and risk sharing: how does Italy compare in Europe?Journal of Public Economics 86: 287306.

    • Search Google Scholar
    • Export Citation
  • Douglas, James W. and Ronald Keith Gaddie. 2002. State Rainy Day Funds and Fiscal Crises: Rainy Day Funds and the 1990-91 Recession Revisited. Public Budgeting & Finance. Spring pp.1930.

    • Search Google Scholar
    • Export Citation
  • Fisher, Ronald. Editor. 1997. “Intergovernmental Fiscal Relations.” Kluwer Academic Publisher. London.

  • Garcés Sanagustín, Mario. Editor, 2004. “La Estabilidad Presupuestaria en el Derecho Español.” Ministerio de Hacienda, Instituto de Estudios Fiscales.

    • Search Google Scholar
    • Export Citation
  • García Díaz, Miguel Angel. 2003. “La Financiación autonómica de regimen común. Perspectivas después del acuerdo del 2001.” Madrid, Confederación Sindical de CCOO.

    • Search Google Scholar
    • Export Citation
  • Garrett, Thomas A., and Wagner, Gary A. 2004. State Government Finances: World War II to the Current Crises. Review Federal Reserve Bank of St. Louis Vol. 86(2): 925.

    • Search Google Scholar
    • Export Citation
  • Joumard, Isabelle and Per Mathis Kongsrud. 2003Fiscal Relations across Government Levels.” OECD Economic Studies.

  • Knight, Brian and Arik Levinson. 1999. Rainy day Funds and State Government Savings. National Tax Journal. Vol. LII(3): pp. 459472.

  • Kwon, Goohoon and Antonio Spilimbergo. 2005. “Russia’s Regions: Income Volatility, Labor Mobility and Fiscal Policy.” IMF Working paper. Forthcoming.

    • Search Google Scholar
    • Export Citation
  • North, Douglas. 1981. Structure and Change in Economic History. New York. Horton.

  • Oates, Wallace. 1972. Fiscal Federalism. New York: Harcourt Brace.

  • Pérez-Díaz, Victor. 2004. Los Puntos Débiles de la Vida Pública Española. Papeles de Economía Española N. 100.

  • Poterba, James M. 1994. “State Responses to Fiscal Crises: the Effects of Budgetary Institutions and Politics.” Journal of Political Economy. Vol. 102(4): pp. 799821.

    • Search Google Scholar
    • Export Citation
  • Report on the Oservance of Standards and Codes (ROSC) for Spain. 2005. Fiscal Affairs Department. IMF.

  • Rodden, Jonathan, Gunnar S. Eskeland, Jennie Litvack. Editors. 2003. “Fiscal Decentralization and the Challenge of Budget Constraints.” MIT Press. Cambridge, MA.

    • Search Google Scholar
    • Export Citation
  • Salinas Jiménez, Javier, editor. 2002. “El Nuevo Modelo de Financiación Autonónomica.” Instituto de Estudios Fiscales. Madrid.

  • Singh, Raju and Alexander Pleckhanov. 2004. How Should Subnational Borrowing be Regulated? Empirical Evidence from a Cross-Country Study. October. Mimeo. IMF.

    • Search Google Scholar
    • Export Citation
  • Ter-Minassian, Teresa. Editor. 1997. “Fiscal Federalism in Theory and Practice.” Washington. IMF.

  • Wurzel, Eckhard Georg. 2003. Consolidating Germany’s finance: issues in public sector spending reform. OECD Economics Department Working Paper No. 336.

    • Search Google Scholar
    • Export Citation

Prepared by Antonio Spilimbergo (FAD).


Recognizing the different aspiration and preparation for autonomy in different CCAA, the Constitution envisioned different speeds of devolution of fiscal responsibilities in different regions. Since the devolution of health expenditure to ten remaining regions January 2002, all regions have the same expenditure responsibilities.


See fiscal Report on the Oservance of Standards and Codes for Spain (hereafter ROSC, 2005) for a detailed account of the legal framework regulating the fiscal operations of the central government and CCAA.


The Organic Law on the Financing of the Regional Governments (LOFCA, Ley Orgánica de Financiacidn de las Comunidadades Autónomas) of 1980, which was revised in 2001, provides the latest legal framework for CCAA financing. The approval or modification of Organic Laws, which are regulated by Article 81 of the Constitution and concern fundamental rights, require an absolute majority in a final vote of the entire bill.


Formally, the FFPC, which was established by LOFCA in 1980, is only a technical and consultative body composed by economic ministries of the central government and representatives of the CCAA. In practice, the FFPC takes all decisions concerning the financial relationships between the central government and the CCAA. The decisions are taken by qualified majority in which the representatives of the central government and the combined representatives from the CCAA have the same weight. The voting patterns often reflect party lines; for instance, the September 1996 agreement for the period 1997-2001 was voted by the central government’s representatives and by representatives of the CCAA in which the governing party was in power, while representatives from Extremadura, Andalucia, and Castilla-La Mancha, which were governed by the opposition, voted against. In contrast, the latest (2001) agreement was voted unanimously by the central government and the CCAA.


Note that the BSL contains more specific norms for the central government, including the obligation to set an expenditure limit and to include a contingent line item in the budget, equivalent to 2 percent of total expenditure, to cover unforeseen non-discretionary spending. In contrast, the BSL does not contain expenditure norms for the CCAA. For a more detailed discussion of the BSL see Box 7 of the fiscal ROSC, 2005.


“Common regime” CCAA include all regions with the exception of the Basque Country and Navarre.


The year 1999 was chosen, because it was the last available year for which the information was available in 2001 when the agreement was signed. The rules to revise the amount for the sufficiency funds were quite complex, to the point that some observers claim that the technical rules for the determination of the sufficiency fund are simply an ex-post rationalization of political agreements between the central government and CCAA (see Salinas Jiménez, 2002).


Note that this introduces some asymmetry since the CCAA that are net contributors to the fund are able to keep a higher share of resources if regional revenues are higher than expected.


In principle, the FFPC may assign a different level of deficit or surplus to each region; in practice, every CCAA has been required to run a balanced budget. In April 2002, the FFPC established that every CCAA should run a balanced budget for the following three years.


The importance of informal rules for the working of institutions has been highlighted by North (1981). Aggressive decentralization as a reaction to a very centralized system is quite common. This may reflect the pent-up demand for direct democracy and the idea that a decentralized political system may provide insurance against the risk of undemocratic relapse. Alesina and Spolaore (2003) review the evidence of the correlation between decentralization and democratization.


This is only formally similar to the political situation in Germany in which the Bavarian Christian Democrat party was a fundamental component of the center-right coalition. In the German case, nationwide party ideology is stronger than regional identification while in Spain regional parties do not have nationwide counterparts with similar ideology. In this respect, Spain is more similar to the political setting of contemporary Italy and Canada.


Regions correspond to the NUTS2 classification of EUROSTAT. The sharp jump in the regional income dispersion in Germany reflects the German unification.


The calculation of the relative variability of tax revenues at the CCAA and central government levels is complicated by the fact that the taxes assigned to the CCAA have changed over time. For this reason, simple extrapolation from the past could be only indicative of future volatility. The experience of countries with a long tradition of decentralization, such as the U.S., suggests that the base of local taxes tends to become more volatile over time as the importance of local consumption based taxes diminishes.


In reforming the interregional equalization system, Germany has introduced incentives for Länder that perform well in terms of tax collection (Wurzel, 2003).


According to the OECD, the experience of Canada suggests that fiscal consolidation has been inversely related to provincial reliance on transfers from the federal government (Joumard and Kongsrud, 2003).


For example, the Italian system is designed so that the marginal expenditure equivalent to 10 percent of regional expenditure is financed through local taxes.


A possible drawback of RDF is that, under the European System of National Accounts 1995 (ESA95), rainy day funds are considered as below-the-line items, so that drawings from RDF are considered a financial transaction and not fiscal revenue. This does not matter for macroeconomic stability, but it means that resources drawn from the RDF could not be used to comply with the Stability and Growth Pact’s deficit limits.


The overlapping-generations model used in this study captures the effect of the Spanish pension rule. As noted in the text, the (closed) economy is populated by overlapping generations of households, atomistic firms, and the government. The general equilibrium structure is standard, but the model incorporates a stylized version of the pension rule whereby the old-age benefit is calculated based on wage earning in the last 15 years in the workforce. Details of the model follow.


The household’s objective is to maximize lifetime utility. A household does so by choosing consumption (c), leisure (l), and asset (A) accumulation optimally during its lifetime, which is characterized by two distinct phases: an active work life lasting T periods or years and a mandatory retirement lasting TR years.47 The utility function of a household born at time t can be expressed as follows:


where subscripts (superscripts) refer to the time period (age of the household). During their work life, utility is optimized subject to the budget constraint:


where assets next year are determined by adding to this year’s assets this year’s net return (first term) and savings (net wage income minus consumption); payroll (τ), income (τI), and consumption taxes (τc), and interest rates (r), and wages (W) are given.48 The total number of hours is normalized so that the sum of work (n) and leisure (l) equals one, that is lts=1nts.

In retirement lts=1 and wage income is replaced by an old-age pension (b) in the budget constraint


Note that the old-age pension for a household retiring at time t can be expressed as


where the average (gross) wage in the reference period (covering the last μ years before retirement) is “scaled down” by the replacement ratio (ψ).49

Two sets of conditions solve household’s problem under standard dynamic optimization techniques. As discussed in the main text, these are the a household’s consumption-leisure choice in a specific year (intratemporal first order conditions), and its consumption-savings decisions over time (inter-temporal first order conditions or Euler equation).50 Spain’s pension rule introduces three subperiods in the household’s work life, the first—comprising of the initial T-μ years prior to the reference period—corresponds to the standard conditions (Table 2). The second corresponds to the first μ-1 years of the reference period, when the consumption-leisure choice (intratemporal first-order conditions) also reflects the fact that wage earnings accrued in this subperiod provide additional utility during retirement because of their effect on the pension benefit, and explains the jump in work hours at the outset of this period.51 The consumption-saving decision remains unchanged. In the final year of the reference period, the consumption-saving decision reflects, nonetheless, the retirement of the individual in the following period (lt+TT+1=1). Finally, when the household retires, there is no labor supply choice by definition, and only the consumption-saving decision remains.52

Table 1.

Variable Definition and Notation

article image
Note: Superscripts (subscripts) indicates the age of theousehold (time period); stock variables are dated at the beginning of the corresponding year.
Table 2.

First Order Conditions—Household’s Optimization Problem

article image

The model poses a log utility function. To simplify an already complex model, a log utility function is specified,


in this study. As explained, in the main text this specification favor policies based on adjustments in consumption taxes over other taxes, because these do not distort household’s labor decision. In other words, following an increase in consumption taxes—but not for income or payroll taxes—the negative wealth effect that tends to increase labor effort as households attempt to recoup their ability to accumulate assets, exactly offsets the substitution effect which tends to reduce labor effort (to enjoy more leisure) because consumption becomes more expensive relative to leisure.

To define the macroeconomic equilibrium, aggregate household consumption (Cth), supply of labor (Nth), and assets (Ath) are needed at each point in time. These are obtained by aggregating an individual household’s optimal choices, as follows:


where the summation is over the population of all ages (Pts, except for aggregate labor supply, which excludes the retired population that by definition no longer supplies labor.


Firms maximize profits net of capital depreciation. They do so subject to a constant-returns-to-scale Cobb-Douglas production function with labor-augmenting technological progress,


where both output and factor markets are perfectly competitive, i.e., they face a given wage rate (Wt) and rental rate (rt).53 The first order conditions require that Wt (rt) equal the marginal product of labor (capital):


where δ is the rate of capital depreciation.

The Government

The government sets taxes to ensure long-run fiscal stability. The government collects payroll, income, and consumption taxes from households. Tax revenues are used to finance public consumption (G)—whose path remains unchanged as a share of GDP—and pension benefits, and to redeem government debt (D) according to the following budget constraint:

Dt+1=(1+rt)Dt+[GtτtI(rtAth+WtNth)τtcCth]+Σs=T+1T+TRbts .Ptsτt .Wt .Nth ,

where the (non-social security) primary deficit (term in brackets), and the social security deficit (last two terms) have been separated for clarity.


Equilibrium is defined as a situation that simultaneously places all households and firms on their maximizing paths, establishes the solvency of the government, and clears markets. Consider an initial government debt D0 ≥ 0, capital stock K0 > 0, and distribution of assets {A0s}s=1T+TR, such that D0+K0=A0h=Σs=1T+TRA0sP0s. The equilibrium is a collection of lifetime plans for both households born during the period of analysis (t ≥ 0),


and those born before then (t < 0)—households of ages 2 through T+TR at t = 0—and thus face “truncated” lifetime plans


a sequence of allocations for the firms {Ktf,Ntf}t=0,

and a sequence of relative prices of labor and capital and payroll, income, and consumption tax rates {τt,τtI,τtc,Wt,rt}t=0 such that

  • the government budget constraint is satisfied for t ≥ 0, and for a given government consumption plan {Gt}t=0;

  • the sequence of allocations {Ktf,Ntf}t=0 solves the firm’s optimization problem;

  • the lifetime plans for households born during the period of analysis {ct+s1s,lt+s1s,At+ss+1}s=1T+TR; t = 0,1,...,∞ solve their optimization problems, and the lifetime plans for households of ages s˜=2,...,T+TR at time t = 0 {css˜s,lss˜s,A1+ss˜s+1}s=s˜T+TR solve their truncated optimization problems;

  • the labor market clears, Nt=Ntf=Nth,fort0;

  • the asset market clears, At=Dt+Ktf=Ath,fort0; and

  • the output market clears, Kt+1=(1δ)Kt+YtCtGt for all t ≥ 0, where Yt=Ytf and Ct=Cth are the equilibrium aggregate output and consumption levels.54

A balanced-growth equilibrium is defined to calibrate the model. The economy exhibits a balanced-growth equilibrium—assuming a constant population growth rate (p)—when the government implements a fiscal policy characterized by a constant government expenditure to GDP ratio and constant tax rates (and thus a constant debt to output ratio). Along this balanced growth equilibrium path of the economy, all endogenous variables grow at constant rates (text table). The balanced-growth equilibrium can be represented as a steady state in “detrended” variables by transforming aggregate variables so as to eliminate the effects of technological progress and population growth.

Balanced Growth Path

article image

The model is calibrated to match the stylized facts of the Spanish economy (Table 3). This is done for the relevant features along an initial balanced growth equilibrium path, which constitutes the starting point of the analysis:

  • Standard parameter values in the real business cycle literature are used for the share of capital in production, and the discount and capital depreciation rates.

  • The leisure parameter is calibrated so that the fraction of time worked by a representative individual in the population is 0.275. Assuming that this individual sleeps 8 hours per day, the leisure-work decision is made for the remaining 16 hours. This translates into a total of 112 (=7x16) hours per week. Assuming that households work 40 hours per week, individual’s work 35.7 (=40/112) percent of time he/she is awake. Adjusting the fraction of time work by labor force participation—about 77 percent for those between the ages of 16-64—yields 0.275.55

  • The tax rates are calibrated so as to match effective rates observed in 1994-2004. Specifically, the payroll tax rate (τ) matches the observed ratio of social security contributions to wage income, and the consumption (τc) and income tax rates (τI) match, respectively, the ratios of indirect tax revenues to private consumption and direct tax revenues to GDP.

  • The value of the population growth rate (p) matches the average population growth rate for 1900-70.

  • The replacement ratio value is set at 0.60. Because of recent changes to the social security system in Spain, the replacement ratios for new pensioners differ from those retired under previous regimes. The parameter choice matches the ratios observed in the most recent pensioners’ cohorts.

Table 3.

Calibration of the Baseline Model (Initial Steady State)

article image

Given these values, the calibration exercise verifies that the resulting values of the endogenous variables in the initial steady state and the fiscal data ratios closely resemble those observed in the Spanish data.


  • Abío, Gemma, Eduard Berenguer, Holgar Bonin, Joan Gil, and Concepció Patxot, 2001, “Is the Deficit under Control? A Generational Accounting Perspective on Fiscal Policy and Labor Market Trends in Spain,” FEDEA Working Paper 2001-06.

    • Search Google Scholar
    • Export Citation
  • Alonso-Meseguer, Javier, and Simón Sosvilla Rivero, 2004, “Proyecciones del sistema educativo español ante el boom inmigratorio,” FEDEA Working Paper #2004-14

    • Search Google Scholar
    • Export Citation
  • Auerbach, Alan and Laurence Kotlikoff, 1987. “Dynamic Fiscal Policy- Cambridge University Press.

  • Bailén José M., and Joan Gil, 1998, “Transitional Effects of a System Change in Spain,” in Marmor, Theodore and Phillip De Jong (Eds), “Ageing, Social Security and Affordability”.

    • Search Google Scholar
    • Export Citation
  • Balmaseda, Manuel, and Patrocinio Tello, 2003, “The impact of legislative changes on contributory pensions,” Situación Spain, pp. 2940.

    • Search Google Scholar
    • Export Citation
  • Conde-Ruiz J. Ignacio and Javier Alonso, 2004. “El futuro de las pensiones en España: perspectivas y lecciones”. FEDEA.

  • Culyer, Anthony, 1988, Health Expenditures in Canada: Myth and Reality; Past and Future, Canadian Tax Foundation: Paper No. 82.

  • European Commission, 2001, “Budgetary Challenges Posed by Aging Populations: The Impact on Public Spending on Pensions, Health, and Long-Term Care for the Elderly and Possible Indicators of the Long-Term Sustainability of Public Finances,” (Brussels).

    • Search Google Scholar
    • Export Citation
  • Harrison G.W., S. Jensen, and T.F. Rutherford, 2000, “Using Dynamic General Equilibrium Models for Policy Analysis, Amsterdam, Elsevier.

    • Search Google Scholar
    • Export Citation
  • Heer, Burkhard and Alfred Mauβner, 2004, “Dynamic General Equilibrium Modelling: Computational Methods and Applications- Springer-Verlag.

    • Search Google Scholar
    • Export Citation
  • Herce, José, and Javier Alonso Meseguer, 2000, “La reforma de las pensiones ante la revisión del Pacto de Toledo,” Colección Estudios Económicos, No. 19, Servicio de Estudios, la Caixa.

    • Search Google Scholar
    • Export Citation
  • Hoffmaister, Alexander and Luis Cespedes, 2003, Measuring the output gap for Spain Mimeo, IMF

  • IMF, Country Report (Spain) No 04/89, 2004.

  • Instituto Nacional de Estadística, 2004, “Evolución futura de la población,” Chapter 6, in Tendencias demográficas durante el siglo XX en España, pp. 25794

    • Search Google Scholar
    • Export Citation
  • Jimeno, Juan F., 2000, “El sistema español de pensiones: previsiones de gasto y medidas de reforma”, Hacienda Pública Española, 1/2000.

    • Search Google Scholar
    • Export Citation
  • Judd, Kenneth, 1999, “Numerical Methods in Economics- MIT Press.

  • Kotlikoff, Laurence, 1998, “The A-K OLG Model: Its Past, Present, and Future,” NBER Working Paper Series, #6684.

  • Ministerio de Trabajo y Asuntos Sociales, 2001, Proyecciones a largo plazo de gasto en pensiones (ejercicio periodico)

  • Newhouse, Joseph P., 1977, “Medical-Care Expenditure: A Cross-National Survey.” Journal of Human Resources 12(1):115-125

  • Organization for Economic Cooperation and Development, 2001, “Fiscal Implications of Aging: Projections of Age-Related Spending,” Economic Outlook 2001/1, Vol. 69 (June), pp. 14567

    • Search Google Scholar
    • Export Citation
  • Serrano, Felipe, Miguel Angel García, and Carlos Bravo, 2004, “El Sistema Español de Pensiones: Un Proyecto Viable desde un Enfoque Económico,” Ariel Publisher

    • Search Google Scholar
    • Export Citation
  • United Nations, 2000, “World population prospects: the 2000 revision


Prepared by Mario Catalán, Jaime Guajardo, and Alexander W. Hoffmaister.


These studies were based on information available before the 2001 census, which showed a marked increase in immigration flows. At one extreme, previous projections envisaged Spain’s population to collapse by 20 percent in the first 50 years of this century, and the dependency ratio to skyrocket to 74 percent by 2050 (UN, 2000). At the other extreme, projections saw the population declining by only about 2 percent, and the dependency ratio rising to 56 percent by 2050 (Jimeno, 2000). Others saw the decline in the population and the rise in the dependency ratio to be less than 15 percent and 65 percent, respectively.


This refers to immigrants from non-EU countries, which recently constitute more than three-quarters of overall immigration. The demographic profile of immigrants from EU countries (primarily retirees from France, German, and the U.K.), however, tends to be markedly older than that of Spanish residents.


For details see ¶18-19, Country Report No 04/89, 2004. Herce and Meseguer (2000) discuss earlier reforms and reprint the original text.


The effect of moving to a fully funded system is not discussed, as it is beyond the scope of this study; Bailén and Gil (1998) discuss this case and quantify the costs of such a change.


Special regimes exist for farmers, fishermen, coal miners, and domestic employees. For details of the pension system, see Serrano, Garcia, and Bravo (2004).


This also includes a small amount of expenditures on disability pensions for those individuals older than 65 years. But it does not reflect the expenditure on the separate scheme for civil servants (pensiones de clasespasivas) which, although distinct, is governed by rules and regulations comparable to those of the general scheme. The bulk of the remaining expenditures are pensions for widows and permanently disabled workers, respectively about 19 percent and 12 percent of the total.


Benefits are topped off (complemento a mínimo) when the corresponding pension is less than the minimum pension established in the annual budget.


Participation data, however, are likely to overestimate the true number of individuals, who might choose to participated in more than one plan and thus be double-counted.


Personal income tax is paid only on pensions that exceed €500 (€750) monthly for pensioners filing individually (jointly).


These incentives remain even after applying the corresponding “reduction coefficients” for retirement before the statutory retirement age of 65.


The pension base—average gross wage during past 15 years—is computed as


where wages (W) are adjusted for inflation (CPI), except those from the two most recent contribution periods; the average is computed based on 14 annual installments (12 monthly plus additional payments in June and November).


Low female labor market participation and longer female life spans have meant that most elderly women receive only a widows’ pensions, of which about three-fourths are less than the minimum wage and average about 60 percent of old-age pensions.


Civil servants may choose between private and statutory insurance.


See Auerbach and Kotlikoff (1987). For a more recent survey of this literature see Kotlikoff (2000); numerical methods to solve these models are described in Heer and Mauβiner (2004), and Judd (1999).


When households retire they face only an inter-temporal condition as they no longer supply labor.


Thus, the population growth in the third century reflects a small rebound from its minimum (0.3 per annum) in the decade ending in 1991.


This reflects the assumption that 75 percent of the immigrant flow enters the labor force, i.e., is older than 20 years of age (see ¶2).


More precisely, the model assumes that the total annual growth rate of entrants is 0.5 percent. However, the actual decomposition between natives and immigrants is left indeterminate. Note that if the rate of growth of newborns and life expectancy are constant, the rate of growth of the total population and that of labor force entrants at age 22 are equal.


The burden of consumption taxes is particularly heavy on pensioners, because retirement is a “consumption-intensive” stage of life (individuals dis-save rapidly and sustain consumption). Moreover, the pensioners-bias of the consumption tax burden is larger in Spain because most pension benefits are income-tax exempt.


In this, and subsequent simulations, households are assumed to anticipate that, aside from TAYG, the demographic transition is unaddressed (i.e., there is no pension reform). In 2006, however, they are faced with a once-and-for-all regime change and henceforth reoptimize accordingly.


Note that the new steady state debt is lower than in the initial steady state because the one time increase in the tax rate of consumption is more than enough to cover all the cost of the demographic shock.


Specifically, T=42 and since life expectancy is constant retirement is two years shorter, TR=18. Considering that individual enters the work force at age 22, this implies increasing the (effective) retirement age to 64.


In other words, μ=40, but the retirement age is unchanged: T=40 and TR=20.


In the simulations, this loss is partially compensated by increasing the replacement ratio, ψ, so that pension benefits decline by 12 percent (instead of 18 percent) compared with TAYG steady state.


In other words, μ=42, and the retirement age has been raised: T=42 and TR=18.


Pension benefits are now 16 percent lower than in the TAYG steady state.


Table 2 provides definitions for the notation used in this section.


Specifically, T = 40 and TR = 20. Considering that individuals enter the labor force when they are on average 22 years old, this implies retirement at age 62. Households live 82 years with certainty, which is the implicit “life expectancy” at birth. This is a bit higher than Spain’s life expectancy, but given the time line of the model, this assumption anticipates that gains will continue converging to the highest levels in Europe.


The income tax is levied on labor income and asset earnings, and, for simplicity, these are assumed to be taxed at the same rate.


Consistent with the majority of old-age pensions in Spain, pensions are taken as not taxed. Moreover, the pension benefit is determined upon retirement and remains constant in real terms throughout retirement, that is btT+1=bt+sT1s for s = T + 2,...,T + TR. Also, there are no intergenerational bequests or inheritances: households are born (“enter the labor force,” s=1) and die (s=T+TR +1) with no assets, i.e., At1=AtT+TR=0.


When households retire they face only an inter-temporal condition as they no longer supply labor.


Upon retirement, the household’s optimization problem can be expressed recursively, and using a log utility assumption allows us to obtain a closed-form solution for the individual’s value function at the retirement age. Specifically, Vt+TT+1 is the household’s discounted indirect utility or continuation value when it is T + 1 years old and retires at time t + T. This function depends on its stock of assets at retirement, At+TT+1, future annual pension benefits bt+1T+1, future interest rates, and income tax rates. A forthcoming IMF Working Paper will provide details of the derivation of the value function (V). Note that VA,t+TT+1(Vb,t+TT+1) denotes the partial derivative of V with respect to At+TT+1(bt+TT+1).


The analysis starts with a full set of generations, T+TR, at time t = 0. Thus, during the first T+TR years a number of “truncated” optimization problems are associated with those households of ages s˜=2,...,T+TR that were notionally born before t = 0. These “truncated” optimization problems are defined in the forthcoming IMF Working Paper.


Perfect competition in output and product markets, and the constant-returns-to-scale assumption imposed on the production technology imply that, in general equilibrium, the number and the size of individual firms are indeterminate. Therefore, Ytf, Ktf, Ntf can be interpreted as aggregate variables.


The economy’s aggregate constraint Kt+1=(1δ)Kt+YtCtGt is obtained from the aggregate constraint of the household sector, the first-order conditions of firms, the market equilibrium conditions, and the government budget constraint. The aggregate constraint of the household sector at time t is given by



The labor participation rate assumption anticipates that it will continue—particularly in female participation—converging with EU levels.

Spain: Selected Issues
Author: International Monetary Fund