9 Oil Revenue Assignments: Country Experiences and Issues
- Jeffrey Davis, Annalisa Fedelino, and Rolando Ossowski
- Published Date:
- August 2003
The assignment of oil revenues in multilevel countries raises a number of issues, including the right of subnational regions to raise revenues on natural resources, the ability of subnational jurisdictions vis-à-vis the central government to stabilize revenues in response to oil price uncertainty and volatility, inter-jurisdictional equity and redistribution, and the financing of a stable level of public services provided by subnational governments, as well as environmental concerns. There is also an overriding emotive or political economy consideration associated with the assignment of natural resource revenues—particularly with oil—reflected in demands for a direct share of the oil revenues from the regions where the oil fields are located. Factors in determining appropriate revenue assignments would also include macroeconomic stabilization, efficiency, and redistributive considerations traditionally addressed in the literature.
The paper describes arrangements to assign oil revenue to different levels of administration in a range of countries—ranging from large federations to relatively small unitary states. Based on country studies, the paper presents a simple typology of existing arrangements, including full centralization, full decentralization, and various revenuesharing schemes including the definition of various overlapping revenue bases.
While the paper considers a large number of oil-producing countries, the following countries are examined in more detail: Canada, Colombia, Indonesia, Mexico, Nigeria, Russia, United Arab Emirates, United States, and Venezuela—representing differing constitutional and political arrangements touched on above.
Arguments for centralization of oil revenues are based on a number of considerations. A central government can better absorb the uncertainty and volatility of oil prices because it usually has a broader tax base, less correlated with oil prices, than subnational jurisdictions. And if subnational governments do have other assigned taxes, oil-rich regions may have less incentive to use such bases if they are also assigned oil revenues—this can lead to internal beggar-thy-neighbor outcomes within federations, and a misallocation of factors of production. Moreover, a central government can contribute to horizontal equity by redistributing oil revenue between resource-rich and resource-poor regions. As a depletable resource, oil cannot be considered as a buoyant long-term source of revenue for subnational governments.2
The main arguments for assigning oil revenues to subnational jurisdictions are mainly political and turn on either constitutional stipulations that delineate the regional or local ownership of the natural resources, or the right to levy taxes on certain bases or sources of income. Another common arrangement—sharing revenues—is often the result of attempts by central governments to appease separatist tendencies in natural resource producing regions. We examine the merits and problems associated with such arguments in the context of possible arrangements for different types of countries.
The case of the small unitary state may appear trivial, as oil revenues are necessarily centralized in these countries. However, a number of considerations are important, and the preconditions for effective macroeconomic stabilization in such countries indicate the necessary conditions in larger unitary states or federations for decentralization of oil revenues to subnational jurisdictions.
The paper assesses the advantages and drawbacks of the various types of arrangements, including their ability to contribute to macroeconomic stabilization and fiscal discipline. It concludes that centralization of oil revenues would be a desirable arrangement, along with the assignment of other tax bases to subnational governments accompanied by a transfer system to address distributional or equalization concerns. An alternative solution is to assign more stable oil-tax bases to oil-producing subnational governments, such as production excises, and to design the transfer system accordingly to guarantee a minimum level of resources for all subnational governments to finance a stable level of public services. The least preferred solution is oil-revenue sharing. By taking away large amounts of revenue from the central government, it precludes and complicates macroeconomic management. By fully transmitting the volatility of oil revenue to subnational governments, it does not provide stable financing of local public services, and it usually does not manage to diffuse separatist tendencies, since oil-producing regions can still be better off by keeping 100 percent of their oil revenues.
II. Experiences in Oil-Revenue Assignments
Data on oil revenue disaggregated by level of government are scarce. For both sets of unitary and federal countries, Table 9.1 presents oil revenue collections as well as total revenues over the period 1997-2000. A major feature that is apparent is that oil revenues are more volatile than total revenues—in all major countries and groupings of countries (small and large, predominantly oil producing, and so on). Volatility of oil revenues, defined as the standard deviation relative to the mean, was roughly similar for unitary and federal countries. However, within each group there were considerable differences across countries. For unitary states, for instance, oil revenue volatility was higher for countries such as Iran, Indonesia, and Colombia than for Saudi Arabia or Kuwait—reflecting a lower ability to use quantity adjustments to compensate for price variations. A similar pattern was apparent for federal states, with Nigeria, the United States, and Venezuela being subject to greater oil-revenue volatility than Mexico, Russia, or the United Arab Emirates.
and Grants (In
percent of GDP)1
(In percent of
|Bahrain||13.0||26.3||24.6||15.1||51.9||Oil and gas|
|Colombia||2.6||38.3||27.7||2.8||9.4||Nonfinancial public sector|
|Ecuador||7.4||32.1||25.1||15.9||31.3||Nonfinancial public sector|
|Indonesia||5.6||32.7||17.5||11.5||31.1||Oil and gas. General government4|
|Iran||13.3||42.0||25.5||18.2||49.8||Oil and gas. Central government5|
|Kuwait||38.6||19.2||60.7||13.3||63.0||Oil and gas6|
|Saudi Arabia||24.6||27.4||33.6||15.5||72.0||Central government|
|Mexico||5.3||17.7||21.5||4.4||24.7||Public sector. Excludes excises on gasoline|
|Russia||3.8||26.4||13.1||14.2||28.8||Oil and gas. Federal government|
|United Arab Emirates||18.2||22.4||34.6||6.3||52.0||General government|
|United States||…||31.18||29.9||1.6||…||General government9|
|Venezuela||12.2||34.6||25.9||13.8||45.7||Public sector (excl. nonrecurrent operations)|
Average during 1997-2000.
Defined as the standard deviation in percent of the mean over the period 1997-2000.
Fiscal year starting on April 1.
Fiscal year starting on March 20.
Fiscal year ending on June 30.
Resource revenue in the provinces of Alberta and Saskatchewan.
Oil revenue in the state of Alaska.
Fiscal year ending on September 30.
Average during 1997-2000.
Defined as the standard deviation in percent of the mean over the period 1997-2000.
Fiscal year starting on April 1.
Fiscal year starting on March 20.
Fiscal year ending on June 30.
Resource revenue in the provinces of Alberta and Saskatchewan.
Oil revenue in the state of Alaska.
Fiscal year ending on September 30.
The data also confirm that the volatility of total revenues is greater when the share of oil revenues in total revenues is higher. This, which is seen clearly from Figure 9.1, has implications for states that lack non-oil revenue bases, as well as subnational governments that come to rely extensively on oil-revenue shares or assignments.
Figure 9.1.Volatility of Total Revenue in Selected Oil-Producing Countries, 1997-20001
Specific arrangements to assign oil revenues in unitary and federal countries can be broadly classified into four categories: full decentralization, full centralization, assignment of overlapping/shared tax bases, and revenue sharing. Table 9.2 shows a classification of a number of oil-producing countries according to their assignment of oil revenue. The general principles of tax assignment and revenue-sharing arrangements are set out in the Appendix.
|(until 2000)||(since 2001) (C)|
|Federal||United Arab||Canada||Mexico (C)|
|countries||Emirates1||United States||Nigeria (D)|
Upward revenue-sharing arrangement.
Upward revenue-sharing arrangement.
Most small unitary countries centralize oil revenues, often because local governments in these countries do not have important expenditure responsibilities or have other revenue sources. The challenge for these countries is to develop non-oil revenue bases and to deal with the stabilization and saving of volatile oil revenue at the national level.
Large unitary states (such as Colombia and Indonesia) tend to be pressed into revenue-sharing arrangements with subnational governments, but these lead to many difficulties, not least in terms of macroeconomic management. Moreover, oil-producing regions are generally dissatisfied with the arrangement, leading to potential political instability.
The same issues arise in federal countries, particularly in Nigeria. While expenditure responsibilities of subnational governments remain relatively stable over time, oil revenue transferred through revenue-sharing arrangements is highly volatile, leading to major fiscal management problems that these governments cannot address for lack of alternative revenue bases.
Two federal countries (Canada and the United States) assign oil tax bases to subnational governments (overlapping with the federal government) instead of sharing oil revenue collected centrally. This tax assignment creates more accountability for the governments concerned. The revenue disparities with respect to non-oil-producing regions can be addressed through the national equalization system (Canada). Only one country, the United Arab Emirates, fully decentralizes oil revenue and has an upward revenue-sharing arrangement.
These arrangements are discussed in more detail in the following sections.
III. Unitary States
Most small unitary states reflect full centralization model, where all oil revenues accrue to the central government. Proceeds of taxes on the extraction and production of oil are included in general revenue of the central government. Such countries include those in the Gulf Cooperation Council (GCC) region: Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and Yemen (the case of the United Arab Emirates is somewhat different and is examined with the federal countries). Most other large unitary oil-producing countries (Algeria, Azerbaijan, Indonesia—until 2000, Iran, Iraq, Libya, Norway, and the United Kingdom) also follow this model.
The advantages of full centralization of oil revenues are that (i) the central government fully absorbs oil-revenue fluctuations, which it is able to do because of more diverse tax bases uncorrelated with oil; (ii) interregional disparities can potentially be reduced if the central government is assigned oil revenues, as the latter may be in a better position to establish horizontal equalization mechanisms; and (iii) there may be less scope for a competitive “race to the bottom”3 for other subnational taxes than if oil revenues were assigned to regions or localities.4
Issues in Small Unitary States
A number of policy responses to oil-price volatility and exhaustibility of oil in small oil-producing states are noteworthy. First, most oil-producing states recognize the need to develop a significant non-oil revenue base to minimize the variance in total revenues. Such a base is also needed to compensate for a reduction in customs tariffs that have occurred during the past decade, either due to regional initiatives or the impending membership in the WTO of some of the oil-producing states. Countries such as Azerbaijan have made significant progress in the establishment of a value-added tax (VAT).
Second, some countries have attempted to address the issue of oil-price volatility by the establishment of a stabilization fund, with a view to smoothing revenues, hence expenditures, over time.5 Examples include Azerbaijan, Norway, Oman, and Venezuela. One of the main difficulties with such an arrangement is the possibility of establishing a parallel budget mechanism, often with less oversight than the regular budget. This difficulty is particularly pronounced at the subnational level, especially in developing countries. Norway’s responses to the potential difficulty are noteworthy: all resources coming into the oil fund are reported to parliament—which must also authorize any transfers from the fund to the budget. All spending is done from the regular budget.6
Third, the finite duration of oil extraction is sometimes addressed by setting up a savings fund, with a share of annual oil revenue being saved for future generations.7 An example is the Kuwaiti Reserve Fund for Future Generations (RFFG)—spending from which is expressly curtailed. However, the government was permitted to borrow from the RFFG because of the country’s reconstruction at the end of the Iraqi occupation. Neither a stabilization fund nor a savings fund necessarily stabilizes public finances or effectively raises public savings—these depend on the overall stance of fiscal policy. The potential difficulties noted above may far outweigh potential benefits and the introduction of such a fund should be very carefully assessed at the outset.
Trends in Larger Unitary States
At a time when small nation states are seeking to offset problems caused by the variance in oil revenues, there has been considerable pressure in the larger unitary states (and federations) to enter into revenue-sharing arrangements or assign oil-revenue bases to producing regions or localities.
In addition to the problems also faced by the small oil-producing regions, the sharing or assignment of oil revenues to the producing regions or localities also has to address a number of issues: (i) it is difficult to determine the relevant jurisdiction that should receive the “largesse,” especially if the relevant unit for revenue sharing is the district or county—typically an oil well in one district may draw on reserves technically under several district jurisdictions; (ii) the assignment of such revenues to producing districts or regions may exacerbate horizontal inequalities across districts or regions; (iii) as stated above, it may lead to distortions in economic development by allowing non-oil taxation to be lowered in a “race to the bottom” in oil-producing districts; and (iv) it makes the stabilization function of the central government more difficult to achieve.
Fiscal discipline may be difficult to enforce if subnational governments receive large amounts of fluctuating oil revenues. In developing countries, governments may not be able to resist the temptation to spend the “oil windfall” during periods of high prices—given the relatively low-income levels and the substantial developmental needs. Moreover, the transparency in expenditure management needed to ensure an effective stabilization fund would be difficult to establish at the central level and more so at the subnational level, where expenditure management procedures and reporting practices are likely to be rudimentary.
If regional governments provide significant amounts of services and infrastructure for the exploitation of natural resources, some charges may be considered a compensation for the associated costs, justifying that some share of the revenues (e.g., a proportion of royalties) could be returned to them under the benefit-tax linkage principle. There could also be production excises or severance taxes.8 These alternatives to the sharing of oil revenues may also be less volatile than oil revenues per se, and thus more appropriate as revenue sources for subnational governments.
While oil production is not a good source of subnational revenue, political economy realities might dictate that some form of regional taxation of natural resources is inevitable if there is distrust that the center will provide sufficient untied grants for regional expenditures. There are also a number of economic arguments (based on the benefit principle and externalities) that may justify the levy of taxes on such resources. In addition, since oil and gas are subject to income taxation, a piggybacking arrangement with this and other sources may be considered (this might be quite attractive and might not exacerbate regional inequalities if applied across all income sources). There may thus be a number of taxes or charges on oil and gas that could accrue to regional governments.
Three case studies are illustrative of some of the difficulties discussed above—including Indonesia, Colombia, and Nigeria (discussed later in Section IV on federations).
Until 2000, all oil and gas revenue accrued to the central government and was not subject to any revenue-sharing arrangement with provincial and district-level governments. Between 1997 and 2000 oil and gas revenues represented 33 percent of total revenue of the central government (5.6 percent of GDP).
The laws on decentralization, which went into effect in 2001, transferred expenditure responsibilities such as education, health, and local infrastructure to provinces and districts. They also instituted revenue-sharing arrangements for onshore oil and gas revenue, whereby 15 percent of oil revenue and 30 percent of gas revenue were to be transferred to the originating provinces, districts, and adjacent districts, starting in 2001.9 However, unlike in Canada and the United States, where the provinces own the natural resources and have the right to tax earnings from these resources, in Indonesia the latter are owned by the unitary state.
Given the political economy considerations in Indonesia, it was inevitable that recognition of residents’ demands for a share in the resource rents be accommodated in some fashion. However, sharing of oil revenues with producing regions within the country may not in itself satisfy the aspirations of regional separatists—since by definition they would be better off if they kept all the oil revenues.10 Thus, for the center and the producing regions it may be difficult to establish the politically acceptable level of resource sharing, particularly of oil and gas revenues, and there is plenty of scope for building up further resentments. National unity would thus have to rest on other factors, such as the services that the center could provide with greater efficiency, defense, and national identity.
Moreover, the sharing of oil and gas revenue is likely to further increase regional disparities in revenue capacities, as the sources of oil and gas revenue are concentrated in a small number of provinces and districts. It is estimated that districts in five provinces would likely receive over four-fifths of the local share, while those in the remaining 25 provinces would receive zero or near zero oil and gas revenue shares. Thus, there remains a need for “equalization transfers” for non-oil/gas-producing provinces or districts.11 Finally, revenue from oil and gas production is highly volatile and does not provide a stable source of financing.12
In sum, there are considerable risks in sharing oil and gas revenues in Indonesia. This could widen regional disparities; prove difficult to administer, particularly at district level, as volatile oil prices lead to a divergence between budgets and realized revenues; complicate the functioning of a grants system; add to “unsatisfied aspirations”; and significantly increase the central government deficit.
Colombia is a unitary republic, but with significant subnational governments. Considerable expenditure responsibilities, as well as some revenue, were devolved to districts and municipalities by the 1991 constitution—the provinces perform deconcentrated functions on behalf of the center (as in Indonesia). As a result, municipalities have taken over several important functions, and their expenditures represent over one-third of total nonfinancial public sector (NFPS) expenditure (35 percent on average during the period 1997-2000). The municipalities are mainly responsible for health and education expenditures,13 as well as for local capital expenditure.
Until 2001, the constitution specified compulsory intergovernmental transfers to subnational governments, and these were to reach 46.5 percent of the central government’s current revenues by 2002. This imposed a considerable constraint on the central government’s overall fiscal effort. Since subnational governments have not fully used their powers to collect own taxes, a large share of their revenue (close to one-half) comes from the central government or public enterprises, making them dependent on intergovernmental transfers. A constitutional amendment was approved in 2001, however, delinking transfers to subnational governments from current revenues of the central government.
In this context, oil revenue represents a nonnegligible share of total revenue of the NFPS, on average 8 percent during the period 1997-2000 (about 2½ percent of GDP).14 This revenue is shared among the various levels of government and public enterprises. Oil revenue comprises the operating surplus of the public oil company (Ecopetrol) and direct transfers from Ecopetrol to an oil savings and stabilization fund (FAEP).15 Part of the operating surplus is transferred to the central government, another part as royalties for subnational governments, mostly in producing regions, harbors, or regions crossed by pipelines. Resources from the FAEP fund are shared among all subnational governments and Ecopetrol, according to a formula established in the law (the formula takes into account population and tax revenue performance). In sum, a large share of oil revenue accrues to subnational governments, either through direct transfers to the FAEP fund or through royalty payments.
The law on the oil-revenue-sharing system was established to prevent political considerations from influencing use of the revenue. However, these oil-revenue shares (together with guaranteed transfers) have had a deleterious effect on macroeconomic stability by inducing some territorial governments to contract debt beyond their repayment capacity. Several subnational governments have been brought to the brink of bankruptcy, and in 2000, seven provinces and two municipalities had to restructure their debt. Moreover, the sharing of oil-related royalties favors relatively small and rich oil-producing subnational governments, exacerbating regional disparities (Ahmad and Baer, 1997).
Most federal countries use some form of revenue sharing or assignment of tax bases. In the revenue-sharing model, oil revenues are collected by the central government (or by a subnational government in the case of the United Arab Emirates) and redistributed to all or some levels of government according to a specific rule or formula. The alternative assigns specific tax bases to different levels of government—some of these might be overlapping bases.
There are numerous types of revenue-sharing arrangements for oil. Some apply the same rule or formula to share oil revenues as is used for other fiscal transfers, while others have a different rule. Some arrangements favor the derivation principle, whereby each subnational government’s share is related to the oil revenue originating in its territory. Others follow criteria such as population, needs, or tax capacity (this makes the resulting revenue share similar to a transfer).16 Some revenue-sharing arrangements provide relatively large amounts of revenue to subnational jurisdictions (these are rather decentralizing arrangements), others provide small amounts (rather centralizing arrangements) (see Table 9.2). The first category would include Colombia, Nigeria, Russia, and Venezuela. The second category would include Ecuador, Indonesia, and Mexico.
The main advantage of revenue sharing is that it is a convenient form of transferring fiscal resources to subnational governments, especially if oil is a major source of revenue. While administrative considerations often militate for centralization of tax assignments, including oil revenues, revenue sharing provides a way to redistribute funds to subnational governments. The sharing formula may also be tailored to address various concerns regarding equalization or compensation for special regional needs.
Revenue sharing, however, has major drawbacks with regards to macroeconomic management and overall fiscal discipline (Ter-Minassian, 1997b). Revenue-sharing arrangements reduce the revenue base that the central government can use to stabilize the economy or to adjust the fiscal policy stance—they may distribute revenues that are not commensurate to subnational governments’ spending needs—and they leave subnational governments vulnerable to potentially large swings in oil revenues. In addition, these arrangements tend to be politically controversial and unstable.
The alternative to revenue sharing would be assigning specific revenue bases (such as production excises) to the subnational level. Specific excises linked to oil production—instead of prices—may be relatively stable, as well as positively correlated with environmental damages. The United States and Canada represent examples of this variant.
Mexico (limited revenue sharing)
Oil revenue in Mexico represents a significant share of revenue for the public sector, close to one-third during 1997-2000. Over this period, revenues from oil extraction rights, oil excess return levies, and the net income of the state-owned oil company amounted to about 5.3 percent of GDP.
Although most of the oil revenue accrues to the central government, a small part is shared with the subnational governments according to a formula-based system established by the law on fiscal coordination.17 Specifically, lower level governments receive (i) 20 percent of so-called ordinary oil extraction rights; and (ii) 3.17 percent of the additional oil extraction rights. The former share is incorporated into a general fund, which is distributed to the states based on a fixed formula taking into account their characteristics.18 The latter share is earmarked for municipalities located in oil-producing regions or where oil is shipped abroad, to compensate for environmental damage caused by the extraction and sale of oil (Amieva-Huerta, 1997).
Revenue sharing from oil has not been a source of political contention in the recent past, mainly because the resources are channeled through a general pool of shared resources and distributed according to a transparent formula. Except for a small part transferred to municipalities involved in oil production and export, there is no direct link between oil-producing subnational governments and oil revenues in the general revenue-sharing arrangement. Moreover, because oil revenue represents a small part of the revenue of the subnational jurisdictions, the oil-revenue sharing has not been a source of fiscal instability.
Nigeria (extensive revenue sharing)
Oil is the main source of revenue in Nigeria, amounting to 82 percent of the total revenue of the general government, or 40 percent of GDP, in 2000.19 After so-called first charges20 are withheld, oil revenue is shared among the federal government and state and local governments according to an arrangement whereby the remainder (over 75 percent of gross oil revenue) is divided between the federal government and subnational governments. More specifically, the 1999 constitution assigns the control and collection of oil revenue to the federal government, but attributes at least 13 percent of the net oil revenue to the oil-producing states. In addition, about half of the net proceeds (after deduction of first charges) are redistributed to state and local governments according to a formula decided by parliament every five years. Excess proceeds over the budgeted revenue are also redistributed in the same way, after assigning 13 percent to oil-producing states.
The formula is clearly defined and based on an oil price assumption used in the budget documents. If the actual price exceeds the budgeted price, the excess proceeds are to be distributed. Excess proceeds need not be spent; they could be saved for spending in the following years. If prices are lower than envisaged, the actual amount is distributed.
State and local governments are highly dependent on revenue-sharing arrangements with the federal government.21 In 1999, 75 percent of state revenue and 94 percent of local government revenue came from redistributed revenue from the federal government, including for the former, their share of the VAT (85 percent of total proceeds). Most of this revenue was oil related. Federation revenue released to subnational governments rose from 7.4 percent of GDP in 1999 to about 15.3 percent of GDP in 2001.
The two key challenges posed by the Nigerian model of fiscal federalism are the conflicting claims over oil resources and the lack of fiscal discipline of subnational governments (IMF, 2001). As sharing arrangements for oil revenue combine the derivation and distribution principles, they remain highly contentious. Given large disparities in natural resource endowments, oil is concentrated in only a few of the 36 states, which have onshore or offshore oil production. Several oil-producing states claim that they should have total control over their own oil revenues. Others contend that the 13 percent share accruing to oil-producing states should also be applied to oil revenue from offshore production (which represents about 40 percent of oil revenue). Clearly, the sharing agreement has led to severe political problems with oil-producing states demanding an increasing share of oil revenues, and non-oil producers demanding greater redistribution of oil resources.
There is no legal mechanism to impose fiscal discipline on the lower tiers of government. The high oil prices in 2000-2001 led to a large increase in the distribution of financial resources to subnational governments, particularly to oil-producing states, without the corresponding assignment of new expenditure responsibilities. As these transfers were immediately spent, the rapid increase in aggregate demand contributed to a rise in inflation and a sharp depreciation of the currency in the parallel market.
The revenue-sharing arrangement leaves little room to maneuver on the fiscal side. It places the burden of macroeconomic adjustment on the federal government with control over less than half of the federation revenue.
Ideally, financial resources to state and local governments should be insulated from oil price fluctuations, and should be commensurate with the tasks they are assigned to perform and their capacity to effectively spend such resources. This would imply a shift from a transfer system, which relies on the sharing of volatile natural resource revenues, to one that provides for stable financing of the provision of at least a minimum set of essential subnational public services.
Oil and gas revenues are considerable in Russia. At the federal level, oil revenue is estimated to make up to about 25 percent of total revenue, and 45-50 percent including gas revenues. This includes oil- and gas-related revenue from profit taxes, VAT, excises, export duties, and some resource mineral taxes. Most taxes are collected by the federal government and shared, to a certain extent, with subnational governments according to either negotiated or fixed rate systems.22 However, natural resource taxes are mainly collected by subnational governments and are important sources of own revenue in oil-producing regions. These taxes include petroleum production royalties, charges for use of mineral deposits, and exploration fees.
Because natural resources, including oil, are highly concentrated in a few regions, decentralization of taxes on these resources creates large disparities in revenue between resource-rich regions and others. In 1997, the five richest regions, accounting for only 5½ percent of the population, collected 53 percent of all subnational government revenues from taxes, fees, and charges on natural resources (Martinez-Vazquez and Boex, 2000).
There have been strong pressures for local control over natural resources. Resource-rich regions with high per capita revenues have demonstrated a strong desire for greater autonomy, and particularly for control over revenue from their own resources. While economic arguments would favor centralizing oil rents to minimize large horizontal fiscal disparities, producing regions may resist (McLure, 1994a and 1994b). As in other oil-producing countries (e.g., Nigeria), revenue sharing may not be a fully satisfactory solution, as it would not protect regions from price fluctuations, and they would always be better off keeping a larger share. An alternative may be to assign part of the revenue base (e.g., production excises) and implement an appropriate and reliable fiscal transfer system.
Oil revenue accounts for a large share of central government revenue in Venezuela, on average 47 percent of total revenue during the period 1997-2000 (about 9 percent of GDP). Oil revenues include royalties, income taxes, and dividends, which all accrue to the central government.
Subnational governments play a relatively important fiscal role in Venezuela, with expenditure representing about one-third of central government expenditure. Most of subnational governments’ revenues come from revenue-sharing arrangements and transfers, making them dependent on the central government. It is estimated that about half of subnational government revenues are oil related. In addition to receiving 20-25 percent of total central government revenues through various schemes, subnational governments are entitled to 20-30 percent of oil royalties.
Revenue-sharing arrangements are a continuous source of political and fiscal contention. The central government believes that resources assigned to subnational governments are excessive, and that there is little control over the ways these funds are spent. State and local governments, on the other hand, would like to see their share of revenue increased. The issue of oil-revenue sharing will be discussed once again in the context of the draft law on public finance of state and local governments.
United Arab Emirates
The United Arab Emirates is a federation formed in 1971, comprising seven emirates.23 Each emirate retains a considerable degree of economic and political autonomy, and has full ownership of and control over its oil resources. Oil and gas revenue accounted for about half of total revenue of the consolidated government on average during 1997-2000 (18 percent of GDP).
Oil and gas revenue accrue to emirate governments through royalties, company profit transfers, and income tax receipts. For the largest emirate, Abu Dhabi, 58 percent of revenue came on average from crude oil royalties and taxes during 1997-2000 (about 15 percent of GDP).
The federal government is mostly financed by an upward revenue-sharing arrangement, whereby the oil-producing emirates—mostly Abu Dhabi and, to a lesser extent, Dubai and Sharjah—make cash or in-kind contributions. These contributions and services provided, which represent 60 percent to 106 percent of Abu Dhabi’s oil revenue, account for about two-thirds of federal government revenue. As cash contributions are negotiated each year between the federal government and each emirate, there is considerable smoothing between high- and low-price years, providing the federal government with a relatively stable source of revenue despite price fluctuations.
Overlapping Revenue Bases: North America
Assigning specific tax bases—possibly with some overlap between levels of government—represents an attractive alternative to revenue sharing. Canada and the United States are examples of this model, although Canada is close to full decentralization.
The United States assigns resource revenue bases to the states, which are sovereign under the constitution and own the resources (except on federally owned land), but with a sharing of the income tax base—following from its specific political and historical setting (McLure, 1994b). Alaska—where oil revenue represents about four-fifths of own revenue—presents an interesting example of the methods used.
On oil, the state of Alaska levies a property tax (at 2 percent on appraised value); a severance tax ranging from 12¼ percent to 15 percent on oil, subject to a minimum tax per barrel; royalties; a production tax surcharge for hazardous spill; and a corporate income tax. The corporate income tax is based on worldwide corporation net income apportioned to Alaska under a three-factor formula involving (i) the percentage of corporate sales and tariffs from Alaskan operations; (ii) the percentage of production from Alaska; and (iii) the percentage of property represented by Alaska holdings—at a maximum marginal rate of 9.4 percent. All state taxes and royalties are deductible for federal income tax purposes.
The main advantage of this type of system is that the subnational government is fully accountable for the fiscal policy choices related to oil revenues and their possible uses for spending or savings. Alaska has created a fund to save part of the oil revenue for future generations24 (as has the province of Alberta in Canada).
In Canada, taxes on natural resources, mostly oil, are also assigned to the provinces, and Canadian provinces levy a range of taxes (including minimum taxes) and royalties on natural resources.25 However, oil revenues are strongly concentrated in a few provinces, mainly Alberta where these revenues account for about one-quarter of revenue, and Saskatchewan where they represent about one-tenth (Krelove, Stotsky, and Vehorn, 1997). On aggregate, however, oil revenues account for only 3½ percent of the total revenue of subnational governments. While decentralization of oil revenue contributes to the disparities in fiscal capacity among provinces, the federal equalization system takes these into consideration by withholding any equalizing grants from the relatively rich provinces, including oil-producing Alberta.
However, decentralization and the imposition of oil and gas royalties have led to a heavy taxation of these industries (possibly deterring investment) relative to other industries. In addition, the oil-producing provinces have reduced sales and income taxes relative to other non-oil-producing provinces. Some Canadian scholars would prefer to reallocate natural resource taxes from provinces to the center, but recognize that this is unlikely to happen given the political constraints and the constitutional rights of the provinces.26
Canada and the United States have found means to compensate for the drawbacks of assigning oil revenues to subnational governments. First, oil-producing states in these two countries have created oil funds that aim at shielding their budgets from fluctuations in oil revenue and prevent excessive public spending. Second, the federal equalization system in Canada compensates for the large revenue generated by oil-producing provinces by excluding these provinces from the equalization framework. Third, the piggybacking arrangements for personal and corporate income taxes, whereby provinces can set their own tax rates on the federal tax base, contribute to mitigating the volatility of oil revenue by providing a somewhat more stable revenue source.
V. Policy Options
Based on country experiences, this section assesses the various policy options to assign oil revenue to the different levels of government, given the objective of financing a stable level of public services provided by subnational governments.
In an unconstrained world, it would be best to fully centralize oil revenue. This should be accompanied by (i) appropriate revenue assignments that give the subnational administrations control over some major tax rates at the margin (needed for accountability); and (ii) well-designed transfers with appropriate transparency and based on equalization principles—these sources of finance allow a stable level of public services to be provided by the subnational governments and the financing, if necessary, of large infrastructure projects. This would address the issues of oil revenue volatility and uncertainty, disparities in oil revenue among subnational governments, and fiscal discipline and accountability.
However, given the political realities in most countries, some degree of subnational control over natural resource bases or revenues is likely to be needed—this could take the form of revenue sharing or assignment of revenue bases.
Because revenue-sharing arrangements take away large amounts of revenue from the central government, they complicate macroeconomic management. Sharing a large revenue source—such as oil taxation—reduces the capacity of the central government to run countercyclical fiscal policies or implement effective fiscal adjustment. The central government may be left with little revenue to exert a significant macroeconomic impact, and subnational governments may not be willing to follow a policy determined at the central level, especially in the case of fiscal adjustment.27 Further, by fully transmitting the volatility of oil revenue to subnational governments, revenue sharing does not provide stable financing of local public services. Finally, revenue sharing may not always diffuse separatist tendencies, as it may be difficult to agree on the proportion of revenues to be shared—the oil-producing subnational governments can always be better off keeping 100 percent.
For these reasons, sharing tax bases, with stable elements assigned to subnational governments (such as specific production excises), may be preferable to revenue sharing. It may also be more desirable, for example, on environmental grounds, since revenue would be correlated with environmental damage. These revenue assignments encourage stronger accountability for the subnational governments concerned, which can determine oil tax rates and the use of oil revenue, and can potentially provide adequate financing. They would need, however, to be assessed against expenditure assignments, and transfer systems designed accordingly. Moreover, subnational governments remain exposed to oil-price volatility, and the oil-revenue base may be at times insufficient.
Hence, several conditions must be met to assign oil revenues to subnational governments:
Oil should not be assigned solely to subnational governments—that is, there should be some degree of overlap in the oil tax base.
The amounts of revenue involved should be kept relatively small, if possible.
If the assignment leads to large or fluctuating revenues, appropriate safeguards and transparency would be needed (such as in Alaska). This is not likely to be possible at the subnational level in most developing countries.
Subnational governments should be assigned other sources of stable revenues, for example, production excises and piggybacked income taxes supplemented by stable transfers from the center.
If these conditions are met, assigning oil tax bases to subnational governments may provide—for the reasons discussed above—an attractive alternative to oil-revenue sharing.
Subnational Taxation and Revenue Assignment
Arrangements assigning revenue-raising responsibilities to central and subnational governments (SNGs) should seek to promote economic efficiency, distributional equity, macroeconomic stability, and accountability.
First, arrangements should induce the least possible economic inefficiencies or distortions in the location of economic activities and individuals by limiting wide differences in the tax burden across jurisdictions, and by limiting excessive tax competition. To limit tax-induced movements of businesses, SNGs should be assigned a more immobile tax base.
Second, arrangements should favor distributional equity by providing all SNGs with a fair share of the national tax base, and by promoting positive incentives with respect to local economic development. Equalization grants may have to correct for an uneven distribution of the tax base or to support jurisdictions facing particularly high costs; but there are limits to the degree of equalization achievable. Thus, the ideal subnational tax base is one that is relatively evenly distributed across jurisdictions.
Third, arrangements should provide the central government with stabilization instruments, and conversely, shelter the budgets of SNGs from cyclical fluctuations. Thus, subnational taxes should be relatively insensitive to movements in GDP and national income. Table 9.3 below summarizes these criteria.
|Central Tax||Subnational Tax|
|Mobility of tax base||Mobile||Immobile|
|Distribution of tax base||Uneven||Even across SNGs|
|Sensitivity to GDP||High||Low|
In addition to these criteria, revenue-raising responsibilities should be closely associated with expenditure responsibilities. This links the benefits of public expenditure to their price, namely the taxes levied to finance them. Such a link promotes fiscal responsibility and accountability of subnational policymakers to their electorate. Instead of relying on transfers from the central government, SNGs have to adjust the marginal tax rates when they decide to increase expenditure.
According to these criteria, it is generally accepted that corporate income taxes, VAT, taxes on natural resources, and taxes on international trade should be assigned to the central government. Personal income taxes appear more suitable for—at least partial—assignment to the subnational level, since the mobility of individuals and households tends to be less than that of businesses. However, to minimize distortions and tax-induced movements of labor and capital, it is preferable that the definition of the subnational tax base be homogeneous throughout the country and that rate differentials among SNGs be kept relatively small. Single-stage sales and excise taxes can be assigned to SNGs, provided that the rates do not differ excessively among regions. Property taxes, business license taxes, and user fees for local services are also good potential subnational taxes, since their base is relatively immobile.
Tax Assignment and Revenue-Sharing Arrangements
Tax bases can be assigned either exclusively to one level of government (separation of tax bases) or shared among several levels (overlapping tax bases). In the latter case, the same tax base, for example personal income, would be used by both central and subnational governments (SNGs) as the base of their own taxation, which may have different tax rates and schedules.
Once a tax has been assigned to one level or another, there remains the possibility to share the tax receipts among levels of government (revenue sharing). This requires a definition of the share of receipts to be allocated to SNGs, and the appropriate formula for distributing this share among the different SNGs. Typically, the formula can be based either on the revenue that has been collected in each jurisdiction (derivation principle), or on other criteria such as population, expenditure needs (needs assessment), and tax capacity (resource base).
Box 9.1.Revenue Assignments and Fiscal Autonomy of Subnational Governments
|Own taxes||Base and rate under local control.|
|Overlapping tax bases||Nationwide tax base, but total rates (national and subnational governments) under SNG control.|
|Nontax revenues||Fees and charges under SNG control, but sometimes with specific provisions set by the central government.|
|Revenue sharing||Nationwide tax base and rates. A fixed proportion is allocated to SNGs, according to the derivation principle or some needs-or resource-based formula.|
|General purpose grants (block grants)||Determined by the central government, but SNGs are free to determine how the grant should be spent. The amount received may have an equalization component.|
|Specific or conditional grants||The central government specifies the expenditure program for which the funds should be spent.|
In addition to tax or revenue-sharing arrangements, grants from the central government to SNGs are often an important component of intergovernmental financial transfers. Grants can be unconditional transfers usable for general purposes, or can be conditioned by the grantor to specific uses, for example to finance specific programs or projects. Box 9.1 summarizes the various possible arrangements for assigning revenue to SNGs, which result in decreasing degrees of fiscal autonomy (see Norregaard, 1997).
AhmadEhtisham and KatherineBaer1997“Colombia,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 457–503.
AhmadEhtisham and JonCraig1997“Intergovernmental Transfers,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 73–107.
AhmadEhtisham and AliMansoor2002“Indonesia: Managing Decentralization,”IMF Working Paper 02/136 (Washington: International Monetary Fund).
AhmadEhtisham and Raju J.Singh2003“Political Economy of Oil-Revenue Sharing in a Developing Country: Illustrations from Nigeria,”IMF Working Paper 03/16 (Washington: International Monetary Fund).
AhmadEhtisham and VitoTanzieds.2002Managing Fiscal Decentralization (London: Routledge).
Alaska Permanent Fund Corporation2001An Alaskan’s Guide to the Permanent Fund (JuneauAlaska).
Amieva-HuertaJuan1997“Mexico,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 570–97.
BoadwayRobin and Paul A.R.Hobson1993“Intergovernmental Fiscal Relations in Canada” Canadian Tax Paper No. 96 (Toronto: Canadian Tax Foundation).
BoadwayRobin and AnwarShah1994“Fiscal Federalism in Developing/ Transition Economies: Some Lessons from Industrialized Countries,”in National Tax Association: Proceedings of the 86th Annual Conference on Taxation 1993 (ColumbusOhio) pp. 64–71.
BretonAlbert2002“An Introduction to Decentralization Failure” in Managing Fiscal Decentralizationed. byEhtishamAhmad and VitoTanzi (London: Routledge).
BrosioGiorgio2000“Decentralization in Africa” IMF Conference on Fiscal DecentralizationNovember20-212000. Available via the Internet: http://www.imf.org.
International Monetary Fund2001Nigeria: Selected Issues and Statistical Appendix IMF Staff Country Report No. 01/132 (Washington).
BrosioGiorgio2002Indonesia: Selected Issues IMF Staff Country Report No. 02/154 (Washington).
IpIrene K. and Jack M.Mintz1992Dividing the Spoils: The Federal-Provincial Allocation of Taxing Powers (Toronto: CD. Howe Institute).
JiménezJuan Pablo and F.Devoto2002“Argentina: Coordination of Subnational Borrowing” in Managing Fiscal Decentralizationed. byEhtishamAhmad and VitoTanzi (London: Routledge).
KreloveRussellJanet G.Stotsky and Charles L.Vehorn1997“Canada,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 201–25.
Martínez-VázquezJorge and JamesonBoex2000Russia’s Transition to a New Federalism (Washington: World Bank).
McLure Jr.Charles E.1983“Fiscal Federalism and the Taxation of Economic Rents,”in State and Local Finance: The Pressures of the 1980sed. byGeorge F.Break (Madison, Wisconsin: University of Wisconsin Press) pp. 133–60.
McLureJr.CharlesE.1994a“The Taxation of Natural Resources and the Future of the Russian Federation,”in National Tax Association: Proceedings of the 86th Annual Conference on Taxation1993 (ColumbusOhio) pp. 87–91.
McLureJr.CharlesE.1994b“The Sharing of Taxes on Natural Resources and the Future of the Russian Federation,”in Russia and the Challenge of Fiscal Federalismed. byChristine I.Wallich (Washington: World Bank) pp. 181–217.
McLureJr.CharlesE.1998“The Revenue Assignment Problem: Ends, Means, and Constraints,”Journal of Public Budgeting Accounting and Financial Management Vol. 9 (Winter) pp. 652–83.
McLureJr.CharlesE.2000“Tax Assignment and Subnational Fiscal Autonomy,”Bulletin for International Fiscal Documentation Vol. 54 (December) pp. 626–35.
McLureJr.CharlesE.2001“The Tax Assignment Problem: Ruminations on How Theory and Practice Depend on History,”National Tax Journal (Washington) Vol. 54 No. 2 pp. 339–64.
MieszkowskiPeter1983“Energy Policy, Taxation of Natural Resources, and Fiscal Federalism,”in Tax Assignment in Federal Countriesed. byCharlesE. McLure Jr. (Canberra: Australian National University) pp. 129–45.
MieszkowskiPeter and EricToder1983“Taxation of Energy Resources,”in Fiscal Federalism and the Taxation of Natural Resourcesed. byCharlesE. McLure Jr. and PeterMieszkowski (Lexington, Massachusetts: Lexington Books) pp. 65–91.
NorregaardJohn1997“Tax Assignment,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 49–72.
SchwartzGerd and ClaireLiuksila1997“Argentina,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 387–422.
TanziVito1996“Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomic Aspects,”Annual World Bank Conference on Development Economics 1995ed. byM.Bruno and B.Pleskovic (Washington: World Bank) pp. 295–316.
Ter-MinassianTeresa1997a“Intergovernmental Fiscal Relations in a Macroeconomic Perspective: An Overview,”in Fiscal Federalism in Theory and Practiceed. byTeresaTer-Minassian (Washington: International Monetary Fund) pp. 3–24.
The authors benefited from comments from Giorgio Brosio, Jeffrey Davis, Gene Handel, Charles E. McLure, Jr., and Rolando Ossowski, as well as from input from M. Alier, A. Bauer, D. Chua, T. Daban, S. Danninger, A. Espejo, U. Fasano, M. Guin-Siu, H. Hirschhofer, L. Leruth, A. Rambarran, S. Schwartz, A. Vivanco, and D. Zakharova on individual countries. All errors remain ours.
An interjurisdictional competition that would drive tax rates down and result in sub-optimal outputs of public services. See Breton (2002).
Mieszkowski and Toder (1983) find that efficiency losses due to the migration of capital and labor to energy-producing states in the United States (which benefit from oil revenues) may range from 2 percent to 9 percent of the revenue raised.
For a full discussion of the advantages and problems associated with oil stabilization funds, see Davis and others, Chapter 11 in this volume.
See Skancke, Chapter 12 in this volume.
This is also the case in Alaska.
Severance taxes are levied on the extraction or removal of natural resources.
In addition, at least 25 percent of central government revenue (net of revenue sharing) is to be transferred to all lower level governments through a general transfer mechanism, and additional grants may be distributed for specific purposes (IMF, 2002).
The equilibrium may be less than 100 percent, however, because of the costs associated with secession. See Ahmad and Singh (2003) for a discussion. In Indonesia in 2001, Aceh Province accepted the offer of an asymmetric deal based on retention of 55 percent of oil revenue and 40 percent of gas revenue.
For example, as of mid-December 1998, the price of benchmark Minas crude oil stood at about US$10 per barrel, compared with an assumed price of US$13 for purposes of the central budget, causing monthly average revenue to fall to US$600 million in the third quarter of 1998, compared with US$1 billion per month in the fourth quarter of 1997. The rupiah’s strengthening to about Rp 7,500 to the dollar from the Rp 10,600 assumed in the central budget compounded the fall in revenues in domestic currency. The revenues from oil rose dramatically during the rebound of oil prices in 1999-early 2000. The seesaw pattern continued during 2001-02.
As there have been difficulties in the assumption of teachers’ salaries by the municipalities, the education function has been only partially decentralized.
This share increased from 5½ percent in 1997 to about 14 percent in 2000.
The fund’s objectives are to save part of oil revenues and to stabilize oil-revenue fluctuations.
Subnational governments have considerable expenditure responsibilities, mostly the provision of education (elementary and middle schools), public health care and social services, basic social infrastructure, security, and utilities. Most of their revenues are in the form of conditional transfers and revenue-sharing arrangements with the central government, totaling about 7 percent of GDP.
In addition, states receive a 20 percent share of the excise tax on gasoline and of the overall VAT revenue. They must redistribute at least 20 percent of their shares to municipalities. Central government transfers are based on revenue estimates included in the annual budget, but are automatically adjusted each month to reflect actual revenues.
Oil revenues consist of (i) crude export earnings of the Nigerian National Petroleum Company (NNPC); (ii) profit taxes and royalties of oil-producing companies (usually joint venture companies with a government majority ownership); and (iii) domestic crude sales and upstream gas sales.
First charges comprise mainly the government share of the production cost of oil (“cash calls”) and priority projects of the national oil company, the external debt service, and the 13 percent allocated to oil-producing states.
State and local governments are responsible for providing education, health, public works, and local utilities and infrastructure. However, there is little information on sub-national governments’ budgets, much less on the precise level and composition of their expenditure. In general, it is assumed that subnational expenditures are equal to their revenues, but some subnational governments have accumulated considerable debt vis-avis banks.
For specific revenue-sharing arrangements, see Martinez-Vazquez and Boex (2000).
The seven emirates are Abu Dhabi, Ajman, Dubai, Fujairah, Ras Al-Khaimah, Shar-jah, and Umm Al-Qaiwan.
The Alaska Permanent Fund was created in 1976 to save part of the large oil revenue derived from the exploitation of the state’s oil resources. The fund receives about one-fifth of oil revenue, as well as other discretionary state transfers. It saves part of its revenues and distributes the other part to its residents in the form of an annual dividend. At end-2001, the fund totaled close to US$25 billion in net assets (about 95 percent of state GDP); see Alaska Permanent Fund Corporation (2001).
For a discussion of the role of natural resource rents on the design of the transfer system in Canada, see Boadway and Hobson (1993).
See Ip and Mintz (1992).
Although not related to oil revenue, in Argentina there were several “pacts” between the federal government and the provinces during the 1990s, aiming at sharing the burden of fiscal adjustment. However, these arrangements proved difficult to set up (not all provinces participated) and to enforce (mechanisms to that effect were weak). See Schwartz and Liuksila (1997), and Jiménez and Devoto (2002).