Revenue Sharing and Tax-Base Sharing Tomas J. Courchene* July 2004 Language: English Prepared for the program on: Fiscal Management For Better Governance: Learning from Each Other A Joint Program of the Ministry of Finance, China, the Canadian Agency for International Development and the World Bank Institute Web: www.worldbank.org/wbi --------------------------------------------------------------------------------------------------------------------------------------- * The author is a professor at Queen s University. This presentation was made at the International Seminar on Intergovernmental Finance System, held in Chongqing, China, July 4-9, 2004. The seminar is one of series of events organized under the program, Fiscal Management for Better Governance. The program is directed by Dr. Anwar Shah (ashah@worldbank.org), Lead Economist and Program Leader on Public Sector Governance, World Bank Institute, 1818 H Street, NW, Washington, DC 20433, USA. For further information, please contact Chunli Shen (cshen1@worldbank.org). REVENUE SHARING AND TAX-BASE SHARING by Thomas J. Courchene Jarislowsky-Deutsch Professor School of Policy Studies, Queen s University and Senior Scholar Institute for Research on Public Policy, Montreal Presentation to the Intergovernmental Finance System Workshop Chongquing, China July 4-10, 2004 OVERVIEW Revenue Sharing is defined as one or more levels of government having access to a specific share of revenues collected by another level of government; Tax-Base Sharing occurs when two or more levels of government can access (levy taxes against) a common base Revenue sharing is often a substitute for intergovernmental transfers, where the magnitudes of the former are determined by assigning a share of (usually) federally-raised revenues to the states/provinces, whereas the magnitudes of the latter are based on criteria other than the amount of federal revenues (and can be quite discretionary) OVERVIEW: II Tax-base sharing generally implies greater coordination or harmonization and lesser compliance costs than would be the case if the two levels of government accessed the tax base independently; Revenue sharing ameliorates vertical fiscal imbalance by increasing revenues of sub-national governments, but it does not allow these governments to influence these revenues. On the other hand, tax-base sharing increases sub-national revenues while at the same time allowing them to influence their revenues at the margin (by increasing tax rates). Revenues associated with tax-base sharing are in principle unconditional revenues. Revenues associated with revenue sharing can be unconditional as well, but need not be, as our survey of international practice will demonstrate.
OVERVIEW III: ALLOCATION RULES Tax-based sharing normally implies that the resulting revenues will be allocated on the basis of the principle of derivation, on the basis of what was actually collected in (or derived from ) that jurisdiction Revenue sharing can be allocated in many ways: According to the principle of derivation, as above Equal per capita (i.e., implicitly equalizing vis-à-vis derivation) According to the principle of equalization (eg, poverty, income) According to some other principle (e.g., in the German federation, some revenues are earmarked only for the new Länder) Implication: The more redistributive is revenue sharing, the more that addressing vertical imbalance also addresses horizontal imbalance We now turn to some international examples of tax-base and revenue sharing, beginning with the German federation. German Revenue Sharing: PIT and CIT Personal Income Tax (PIT)/Corporate Income Tax (CIT) Revenues from PIT: 15% is allocated off the top for local governments (presumably on the derivation principle). The remainder is allocated roughly 50-50 between the federation and the Länder on a derivation basis. Revenues from CIT: Allocated 50-50 between federation and Länder on a derivation basis Note that this revenue sharing is complex jurisdictionally. Tax rates and tax parameters for PIT and CIT are set by the German parliament. Collection of these taxes is falls to the Länder, and the sharing formula needs the agreement of the Bundesrat (House of the Länder) German Revenue Sharing: VAT VAT is shared (in 2001) as follows: 50.5% to the federation and 49.5% to the Länder. Distribution of the Länder share is as follows: 75% on a per capital basis and 25% on an equalization basis where part of this 25% is in reverse relationship to the allocation of PIT and CIT revenues, and part is designed as anticipatory equalization in order to bring all Länder up to at least 92% of average per capita revenues prior to the formal equalization program (the inter-länder revenue sharing pool). The highly redistributive nature of the VAT allocation serves to ameliorate much of the horizontal fiscal imbalances in the German federation Check all this with the presentation by P. Bernd Spahn Canadian Revenue Sharing: I VAT ( Canada VAT is called the Goods and Services Tax or GST) In New Brunswick, Newfoundland and Nova Scotia there is a harmonized GST that integrates the former provincial sales taxes (PSTs) with the GST and it is collected by the national tax collection agency (Canada Revenue Agency). The allocation to the three provinces is on a derivation basis. Quebec also converted its provincial sales tax to a VAT format and now collects both its VAT and the federal GST. This is a reverse sharing as it were provincial to federal. The remaining provinces continue with their non-harmonized provincial sales taxes that have a base different than that of the federal GST. Some might prefer to call this tax-base sharing, since it is possible that either the provincial rates or the federal rate could change. However, I am restricting tax-base sharing to apply to the PIT and CIT, as shown later
Canadian Revenue Sharing: II Miscellaneous Revenue Sharing Some provinces share a portion of their PIT revenues with their cities Some provinces also share a portion of the gasoline taxes with their cities. This tends to be tied to spending related to urban transit and transportation generally Recently, all federal political parties have promised to share a major potion of the gasoline tax with cities (Provinces would have to agree with this). These revenues would be tied to urban transport. The precise distribution has not been settled derivation or equal per capita are the two leading options. The economic and political star of cities is in ascendancy so that one can expect much more in the way of tax-base and revenue sharing in the coming years. If this is done on a derivation basis then provision will have to be made for addressing horizontal balance across cities CANADIAN TAX-BASE SHARING: I PERSONAL INCOME TAXATION (PIT) Except for Quebec (which has its own independent PIT), all provinces now share the PIT tax base with the federal government. Ottawa has the right to set some basic parameters (definition of income), but the provinces have considerable flexibility in terms of various deductions and credits. Both can apply their own rate and bracket structures to the common (federally determined) base. Ottawa s tax rates are 16, 22, 26 and 29 per cent. Alberta has a flat 10% rate, Saskatchewan has three rates-- 11, 13 and 15 per cent, and so on. The revenues to the provinces are allocated on a derivation basis and are unconditional. Collection is done by the Canada Revenue Agency (free of charge to the provinces). There are provisions to ensure harmony across provinces, ie, no discriminatory provisions against non-residents CANADIAN TAX-BASE SHARING: II Corporate Income Taxation (CIT) Canada offers the provinces much the same opportunities for corporate income tax-base sharing as for the PIT (Indeed, this has been in place much before PIT tax-base sharing) All provinces except Quebec, Ontario and Alberta have taken advantage of this option. Again, these are collected (free of charge to the provinces) by the Canadian Revenue Agency. And for a charge, the CRA will also offer to collect the taxes of the three other provinces. By way of harmonization, all ten provinces have agreed to a formula for allocating profits across provinces so that the derivation principle can be employed for allocating revenues across participating provinces. Since both PIT and CIT tax-base sharing serve to widen horizontal fiscal disparities, these revenues are then subject to Canadian equalization which ensures that all provinces have at least the minimum standard level of per capita revenues CANADIAN TAX-BASE SHARING: III LESSONS FROM THE CANADIAN EXPERIENCE The history of Canadian tax-base sharing is instructive. It began in the 1950s with revenue sharing. Ottawa directed a given share of its PIT and CIT revenues to the provinces. This share was determined by Ottawa. Because the payments were allocated on the derivation principle (necessitated because Quebec already had its own separate PIT) this was simultaneously accompanied by an equalization formula to address the resulting horizontal or across-province imbalances. Later on, revenue sharing evolved into PIT piggy-backing where provinces could apply a single rate of tax to the tax collected by Ottawa, which recently evolved into provincial rate and bracket flexibility as elaborated earlier. Lesson 1: Revenue sharing can be an important prelude and catalyst for tax-base sharing or even tax decentralization; Lesson 2: Greater tax decentralization along derivation lines could not proceed if there were no equalization program in place. Lesson 3: While Quebeckers have to file 2 PIT forms, taxpayers in other provinces just file one form. Thus Canada s tax-base sharing is at the same time a decentralizaed and harmonized approach
Australia: Revenue Sharing The Australian federal government recently introduced a GST (i.e., a VAT) the entire proceeds of which are turned over to the Australian states. The revenues are allocated across states via the Commonwealth Grants Commission s equalization formula in a manner that comprehensively addresses both vertical and horizontal fiscal imbalances in the Australian federation. (See the later presentation by Alan Morris, Chair of the CGC) The GST rate can be changed only by unanimous agreement of the Commonwealth government and the governments of all the states. Key Lesson: The Australian states have finally latched on to an approach to revenue sharing that provides revenue growth in excess of GDP, an obvious advantage of sharing a growth tax like the VAT, in contrast to the discretionary approach to addressing vertical imbalance that traditionally characterized the Australian federation. MEXICO: GENERAL REVENUE SHARING The Mexican states (as of 1999 at least) are entitled to receive 20% of the combined federal revenues from income taxes, excises, oil royalties and the VAT. The allocation of these unconditional revenues across states is as follows: --45.17% is distributed on an equal per capita basis --45.17% is distributed on an historical basis (e.g., what states used to get under the previous order when they had some tax powers) --9.66% allocated in a way that compensates for (or offsets) the previous two allocations --Roughly 20% of these revenues received must then be passed down to local gov ts (General revenue sharing also exists in Argentina) Tax-base sharing: Two small taxes a property tax on automobiles (tenencia) and a tax on new cars (ISAN) while de jure federal taxes are collected by the individual states who then keep the proceeds (derivation principle). Effectively, the central gov t has transferred these tax bases to the states, which is a variant on tax-base sharing. INTERIM SUMMARY One could continue this comparative survey ad infinitum, since examples of revenue/tax-base sharing can be found in virtually all federations and indeed in most unitary states. Hopefully, the above examples have covered the rich variety of ways that tax-base sharing and revenue sharing have been implemented. The remainder of the presentation is devoted to assessing these two approaches to sharing in the larger context of alternative intergovernmental fiscal arrangements. The three one-on-one assessments are: A comparison between tax-base sharing and tax decentralization; A comparison between revenue sharing and intergovernmental transfers, and finally A comparison between revenue sharing and tax-base sharing. Tax-Base Sharing vs. Tax Decentralization To facilitate this comparison, assume that tax-base sharing (TBS) is represented by the current Canadian model for PIT, namely rate and bracket flexibility but a common approach to the rest of the PIT system, including collection Tax decentralization (TD) will be defined as full sub-national tax autonomy with respect to the PIT TBS, relative to TD, will preserve harmony (particularly avoid tax-base competition), will minimize taxpayer compliance (only one form to fill out), while at the same time will ensure revenue-raising autonomy. Under TD, one would need to mount policies/processes that would preserve the internal economic union (prevent discriminatory tax measures levied at out-of-province residents. TD would require substantial managerial capacity (assessing, collecting, auditing). This can be ameliorated by having a national collection agency with members from sub-national governments on the board of directors (like Canada s CRA). Nonetheless, TD may be required on nationalism grounds (Quebec), in which case TBS may be an important interim stage en route to full TD. Moreover, a too narrow interpretation of TBS may trigger a shift to TD
Tax-Base Sharing vs. Tax Decentralization continued By way of a forward-looking concluding comment, the 21 st century is ushering in two relevant changes to the choice between TBS and TD. The first is that tax bases are becoming more mobile, so that the optimal taxing jurisdiction is larger (higher in the hierarchy) than it used to be. But the information revolution means that coordination costs are falling so that the optimal expenditure jurisdiction is arguably shrinking (closer to the people). This means that there is a disjuncture between the loci of where revenues are collected and where expenditures decisions are made. In turn this means that sub-national taxation becomes problematical, which favours TBS over TD in terms of how to operate a decentralized federation in an increasingly mobile world. (Alternatively tax assignment or at least tax-collection assignment needs to be rethought in the context of increased mobility and the ease of by-pass). Example: François Mitterrand favoured Europe 1992 because there was now a chance to tax capital (at the EU level). But these would be France s (member countries ) revenues, not EU revenues Revenue Sharing vs. Intergov tal Transfers Both Revenue Sharing (RS) and Intergovernmental Transfers (IT) can accommodate vertical and horizontal imbalance. RS having the additional option of being able to be distributed on a derivation basis. This would provide recipient provinces with a greater sense of autonomy or ownership of these revenues. While RS is generally provided on an unconditional basis, IT is more likely to carry conditions and therefore be more intrusive of provincial/state spending priorities. Moreover, IT is more easily adapted to embody cost-sharing, such as 50-50 federal-provincial sharing for health costs. Hence, IT is ideally suited for establishing national principles pertaining to sub-national spending. As these principles become established and resonate with citizens, the conditions can gradually be relaxed and a RS system may become more attractive. RS, once established, is likely to be much less controversial because it is formula driven, a key issue since this is a zero-sum game Revenue Sharing vs. Intergov tal Transfers: continued Among the most obvious benefits of RS is that it is (almost certainly) likely to be more stable and less subject to central government discretion. Moreover, the future values of RS are easier to forecast if they are tied to a predictable revenue source. Relatedly, if the underlying revenue source is a broad-based tax (like the VAT in Australia) then the revenue growth will, over time, serve to lessen any potential for a vertical fiscal imbalance. Neither RS nor IT allows the recipient states (at the margin) to influence the amounts of their revenues. In other words, there is no revenue autonomy. (The exception relates to German-type RS where the states/länder are the revenue collectors public choice theory suggests that if the amount of own revenues retained by rich Länder is quite small (as it is) then they have incentives to use the tax collection system to cater to firms rather than to maximize revenues, a classic principal-agent problem). Finally, and forward looking, if greater tax decentralization (either TBS or TD) is a longer term goal, the RS represents an important interim step along this road, a clear advantage over IT which has a centralization bias. Revenue Sharing vs. Tax-Base Sharing Some of the relative pros and cons of RS and TBS have already been elaborated in the above slides. RS lends itself to more flexible allocational criteria than does TBS. Phrased differently, while both RS and TBS are able to address vertical fiscal imbalance, TBS will generally need to be supplemented by an equalization program in order to address horizontal fiscal imbalances On the other hand, TBS brings with it some sub-national ability to alter a province s own revenues, an essential ingredient for meaningful decentralization. More generally, TBS, if not tax decentralization will characterize decentralized federations while RS, if not IT, will characterize more centralized federations. One can find arguments in the literature suggesting that narrow tax bases are not well suited to either TBS or RS. Nonetheless, sources like gasoline taxes might be ideal on efficiency grounds for TBS/RS because it will be easy to make the case that the resulting revenues should be viewed as quasi user fees and, therefore, dedicated to transit infrastructure or the environment. Canada is pursuing this.
CONCLUSION This has been an analytical presentation with a focus on comparative experience drawn largely from developed federal systems. But China is the world s largest and most diverse unitary state, so that it is not obvious that experience in federal systems can be imported (See my opening presentation to this workshop) Hopefully, however, the selected range of comparative experience has been sufficiently diverse that the analysis will prove instructive, if not useful, in thinking about creative ways to address intergovernmental taxation and revenue relations in China.