that these macroeconomic policies are better served by the European than the member-state levels. This said, the EU is scarcely involved in redistributive, let alone developmental, policies. The Common Agricultural Policy and the expanding Regional Policy funds may be considered as a form of European redistribution, but the scale of these programmes is very limited compared with the vast sums which member-state governments pour into their welfare or social security programmes (for a comparison, see Scharpf 1999; McKay 2001). Furthermore, the EU lacks proper fiscal resources (own taxation) to fund these programmes.
Second, apart from establishing the general budgetary framework within which the euro member-states must operate, the EU is also a very active regu- lator, primarily in areas of interstate commerce, trade or, as documented above, the environment. In their analysis of MLG in Europe, Liesbet Hooghe and Gary Marks documented the increasing relevance of the EU in areas of economic policy, social/industrial policy, legal-constitutional policy (justice and home affairs, policing and citizenship) and international relations/
external security. In each of these four policy sections, the EU has become an important regulator, even in some respects (for instance, external commer- cial relations) the only one (Hooghe and Marks 2001: 187–9). However, the regulatory impact of the EU should be put into context. For the EU has turned into an active regulator, particularly in the four policy sections men- tioned above, but it still relies on the goodwill of the member-states or their regions for implementing most of these regulatory policies (Börzel and Risse 2000; Kelemen 2000, 2004; Swenden 2004a). In this sense, the EU exempli- fies ‘administrative federalism’ par excellence, and it is no wonder that this peculiar distribution of functions was incorporated into Kelemen’s regula- tory theory of federalism.
4. The scope of federalism: expenditure and tax
Court). There is no clear rationale why criminal and civil legislation is central in some countries, but regional in others.
Hence, we can reasonably claim that for each of these countries develop- mentalorallocativeprogrammes are more likely assigned to the regional level than redistributive programmes. However, the number of developmental programmes that falls under regional control may be higher in some countries than in others. Similarly, the extent to which the centre is involved in redis- tributive programmes may vary from one country to another. To specify these differences, I introduce the notion of the scopeof federalism. The scope of federalism would be rather limited if the centre clearly outweighs the regions in the public policies which it controls. Put differently, we face a rather ‘centralized’ state (or federation). Conversely, a state in which the opposite holds is a rather decentralized one. The scope constitutes another continuum, alongside the dual-organic continuum that was specified in the previous chapter.
The scope of federalism could be measured in various ways. One method is to look at the number of fields in which the centre legislates and/or imple- ments and to contrast this with the policy areas in which the regions fulfil either or both of these tasks. Yet this method would raise a number of problems. For instance, how ‘decentralized’ is German federalism, when the most common method of distributing powers is to centralize legislation but to decentralize implementation? Furthermore, in Austria, Spain and Switzerland, some policy fields (such as education policy) are not assigned entirely to the central or regional levels of government. Rather, the centre and the regions are each in control of certain aspectsof education policy (F/R in Table 4.1). In addition, some policy fields may absorb more public money than others. Education policy, in which the regions usually play the largest role, is more costly than environmental policy, a policy in which central regulation dominates. Finally, some policies may be funded on the basis of own-source revenues, others may be entirely sustained by central grants.
When measuring the scope of federalism, I suggest taking two criteria into account: levels of expenditure (de)centralizationand corresponding levels of tax (de)centralization. The scope of federalism is very limited, when there is limited expenditure and tax decentralization; it is high when the regions enjoy extensive expenditure and tax-raising autonomy. States with high expenditure autonomy but little revenue-raising autonomy occupy an inter- mediate position. Although a theoretical possibility, high revenue-raising autonomy does not normally correspond with little spending autonomy (however, one could argue that in Switzerland the level of revenue-raising autonomy exceeds that of the corresponding spending autonomy).
Of both concepts, taxonomists seem to agree best on how to measure levels of expenditure (de)centralization. Table 4.2 lists the distribution of public expenditure programmes between the central, regional and local levels for each of our cases.
As we can read from Table 4.2, Switzerland and Germany have the most decentralized expenditure distribution. Regions and municipalities in both countries absorb respectively 68 and 65 per cent of all public expenditures.
The comparable figures for Belgium and Spain stand at respectively 40 and 49 per cent. In Austria, regional and local governments account for 36 per cent of all public expenditures, whereas the comparable figure for the UK stands at only 25 per cent. The data illustrate that of all the local governments listed, the Swiss spend the largest share of public money. The comparable figure for Germany is high as well. However, compared with Switzerland, the policy autonomy of the German municipalities is subject to stronger (regional) political oversight. In this sense, Swiss federalism is a federation of three, equally strong layers. In the other states, the expenditure autonomy of the local governments is relatively weak or the mechanisms of administra- tive or political oversight (usually by the regional instead of the central governments) are generally much stronger.
The scope of federalism cannot be measured by solely considering the extent to which public expenditure programmes are decentralized. We must also take a close look at corresponding levels of ‘revenue or tax (de)central- ization’. For instance, on the basis of their public expenditure levels, the German and Swiss federations seem equally decentralized. The German Länder and the Swiss cantons absorb roughly comparable shares of the pub- lic expenditure pie. Yet, unlike the Swiss cantons, the German regions have little revenue-raising autonomy to finance their regional expenditures. They cannot decide to expand these programmes without receiving extra money, for which they are partially dependent upon the centre. Furthermore, the centre may specify certain conditions as to how it wants that money to be spent.
Measuring levels of revenue or tax (decentralization) is a complex matter.
We must first specify what makes certain taxes central and others regional.
Table 4.2 Public expenditure distribution in a multilayered setting
BEL ESP GER ÖST SUI UK
Central 59.2 51.0 35.0 63.9 32.0 75.0
Regional 40.8 32.5 38.0 24.5 39.0 25.0
Local 16.5 27.0 11.6 29.0
Total 100.0 100.0 100.0 100.0 100.0 100.0
Note: For Belgium and the UK countries local and regional cells were merged, as no separate data for regional and local public expenditures were available.
Sources: (Québec Commission sur le désequilibre fiscal, 2001: figures for Belgium and Germany);
Thöni 1999: 106–7 (Austria); Pola 1999: 42–3 (United Kingdom); Dafflon 1999: 270; Ruiz- Almendral 2003: 43 (Spain). Year of accuracy for percentages listed: Belgium (1998), Spain (2002), Germany (1995), Austria (1995), Switzerland (1994), United Kingdom (1993).
This requires a brief review of the literature on fiscal federalism. Tax legisla- tion contains information on the base, the rate, the administrationof the tax and the distribution of the tax revenues, that is, who should be entitled to spend tax revenues (Braun et al. 2003: 19). Hence, we can think of a personal income tax for which the centre determines what qualifies as taxable income (base) and sets the rate. However, on that same tax, the regions may add a surcharge or, politically more beneficial, offer a rebate. Furthermore, the tax revenues may be shared between the centre and the regions, and the regions may be responsible for administrating the tax. Thus, we must consider the role of the centre and the regions in each of these aspects. For each tax we can then determine whether we are dealing with a fully centralized or a fully regionalized tax. Fully centralized taxes leave all decisions on the tax base and rate entirely with the centre; they are administered by the centre. All tax receipts exclusively benefit the centre which can use it to finance its consti- tutionally allocated tasks. Typically, taxes on foreign trade qualify as fully centralized taxes (Ter-Minissian 1997: 9).
In contrast, taxes are fully decentralized if the tax base, rate and adminis- tration are entirely regionalized and all tax revenues accrue to the regions (or municipalities). Taxes on housing property frequently qualify as fully decen- tralized. In reality, most taxes prescribe a joint or parallel effort at the federal and regional levels in any, several or all four aspects of a tax policy. Cost- saving, redistributive or macroeconomic goals may determine the tax regime that is chosen for each type of tax. Dietmar Braun and his ‘fiscal federalism’
team worked out a comparative taxonomy, mapping the variety of tax arrangements in a multilevel setting. In Table 4.3 I borrow from their termi- nology (Braun et al. 2003: 16–27).
Fully centralized or fully decentralized taxes fit in the first row: they are unilateralinsofar as one level of government is entirely responsible for deter- mining the tax base, rate and administration. In turn, tax receipts accrue entirely to the government which is levying the tax. Unilateral taxes offer
Table 4.3 Typology of tax regimes (on the basis of four indicators)
Tax Tax
Tax base Tax rate administration recipients
Unilateral C or R C or R C or R C or R
Concurrent C and R C and R C and R C and R
Shared C C C or R C and R
Overlapping C C and R C C and R
Source: Braun et al. 2003: 19 (amended for combination shared – tax administration from C into C or R). In Germany, shared taxes provide for federal administration; in Spain, shared taxes pro- vide for larger degree of regional administration; with C⫽central; R⫽regional.
the benefit of transparency, given that tax payers know exactly which government to hold accountable for the taxes they are paying. Conversely, governments have complete information to determine a tax rate that best corresponds with existing market conditions.
Taxes are concurrent if the central and regional governments establish independent tax policies with regard to the same tax base. For instance, citizens may have to pay a central and a regional personal income tax. With the exception of Switzerland, no federal or regionalized European state has a significant concurrent component in its tax system.
Sharedtaxes assume a central monopoly in determining the base and rate of the tax. Tax revenues are shared between the federal, regional (and possi- bly also local) governments. With the German context in mind Dietmar Braun and his team also specified the federal administration of the tax as one of its core features. Arguably, Spanish VAT comes conceptually close to a shared tax, but the Spanish regions are solely responsible for administrating the tax. I have adjusted Dietmar Braun’s classification scheme accordingly.
The distribution of the tax revenue between the various levels can be arranged in a fixed formula that is specified in a federal or central fiscal law.
Alternatively, a formula may result from an intergovernmental agreement between the federation and the regions. More robustly, distributional formula could be constitutionally anchored.
Finally, taxes are overlappingwhen the federal government determines the tax base and is responsible for levying the tax, but the regions are allowed to vary tax ratesand to share in the general tax receipts.
Of all four tax types, only fully centralized (unilaterally central) and shared taxes leave the regions without discretion in determining the base or rate of a tax. Regions may influence the rate and base structure or the distri- bution of such taxes through their participation in federal legislative policies or their required consent to constitutional amendments with such effect.
However, this constitutes more of a collective than an individual right to regional self-expression. Each region loses its autonomy to determine the rate or base of such taxes as it sees fit. Hence, shared taxes are certainly
‘more’ centralized than unilaterally regional taxes, but not to the point where we should categorize them as ‘fully’ central.
Table 4.4 lists the regional and local public revenues as a percentage of all public revenues, prior to central revenue transfers. We should treat these figures with caution, as the IMF statistics on which they are based typically categorize a number of shared taxes as ‘regional’. This somewhat distorts the data for Germany in particular, where shared taxes constitute a large share of the total tax mass. Furthermore no comparable data which would have disaggregated regional and local revenue sources was found. Keeping these considerations in mind, Table 4.4 demonstrates that the level of revenue or tax decentralization is highest in Switzerland and lowest in the UK.
Although the findings are consistent with the assumption that regionalized
states have a less decentralized fiscal structure than federal states, the Belgian case stands out as an exception. The Belgian regions have an unusually low level of revenue-raising autonomy. Leaving aside the Swiss example we notice that the levels of revenue decentralization are generally at the low end, certainly compared with the corresponding levels of expenditure decentral- ization that were listed in Table 4.3. Put differently, the regions and local government do not raise enough money to cover their expenditure needs.
When the regional governments do not raise sufficient money to cover their expenditure needs, a fiscal imbalance arises. With the exception of Switzerland, the regions are dependent upon central grants for a substantial part of their expenditure. Vertical should be distinguished from horizontal fiscal imbalances. The latter emerge when the capacity to meet expenditure needs varies from one region to another. This may be the case because the revenues from certain regional taxes are unequally spread (for instance, an oil tax can only benefit regions with several oil wells). Alternatively, hori- zontal imbalances could arise because per capita revenues and expenditures vary strongly from one region to another. For instance regions with high economic growth may receive higher tax revenues, and have to spend less on social assistance payments.
In Table 4.5, I have sought to illustrate the degree of vertical fiscal imbal- ances (VFI) for all of our case studies, by combining the data in Tables 4.2 and 4.4. The top line expresses regional and localpublic expenditures as a per- centage of all public expenditures aftercentral transfers (RE). The middle line indicates regional and local public revenues as a percentage of all public revenues prior to central transfers (RR). Vertical fiscal imbalances are calculated as the ratio of RR and RE. When the ratio is lower than 1.0, regionally or locally raised taxes do not cover regional or local expenditures, leading to a vertical imbalance. Conversely, a perfectly vertical balanceleads to a ratio of 1.0. However, if the ratio exceeds 1.0, the regional and local governments receive more money than they spend. A VFI arises in a sense that the central government may become dependent upon the regions and local govern- ments for financing part of its expenditures (a very rare case of ‘upward revenue sharing’).
Table 4.4 Public revenue distribution in a multilayered setting
BEL ESP** GER ÖST SUI UK
Central 92.7 87.6 51.5 79.8 29.2 95.9
Regional and Local 7.3* 12.4 48.5 21.2 70.8 4.1
Total 100.0 100.0 100.0 100.0 100.0 100.0
*Regional and local revenues as a share of total public revenues without social security revenues.
**Excluding Navarra and the Basque Country.
Source: As for Table 4.2.
Table 4.5 should be interpreted with some caution. If overall revenues exceed overall public expenditures or vice versa, a vertical imbalance cannot be adequately expressed by calculating the ratios of total (regional and local) revenues and total (regional and local) expenditures. Yet, if we assume that regions aspire to balanced budgets, total revenue and expenditure levels for any given period of time should be of comparable magnitude.
The table displays VFI of considerable size. The level of VFI is high in Belgium, but in Switzerland the combined revenue resources of the cantons and municipalities exceed the already comparatively high level of regional and local expenditures. The level of VFI seems low in Germany as well, albeit that the German regions are primarily financed by means of so-called shared tax revenues. Conversely, the limited fiscal autonomy of the British regions (and municipalities) is in tune with the embryonic or only partial regional- ization of their state. In spite of the significant policy autonomy of the Belgian and Spanish regions, they have remained largely dependent upon central grants.