Revenue autonomy (or revenue decentralization) can be seen as a continuum of three types of revenues: (a) own source revenues; (b) shared revenues (based on derivation basis, but for autono-mous use); and (c) a variety of transfers. For the purposes of this discussion a distinction has been drawn between revenue autonomy (tax and “non-tax” revenues such as charges and fees) and in-tergovernmental transfers (to be addressed in the following section). Adopting this approach, rev-enues of SNGs may be divided into categories of decreasing local autonomy (Table 1). If SNGs have total or significant control over a tax, fee or charge as demonstrated by political control over the tax rate (necessary and sufficient) or base, it is a subnational (local) tax. If the SNG does not, at a minimum, have control over the rate of a tax, then the tax is not local. Thus, for example, that the central government may split revenues (tax sharing) with the SNG (where the central tax is collected) does not make it a local tax (Blochliger and King 2006).
7 This discussion draws on the work of Leif Jensen (2001) and Working Party 2 of the Committee on Fiscal Affairs of the Organisa-tion for Economic CooperaOrganisa-tion and Development (OECD). The Seminal Work is Paper No. 7 of the OECD Tax Policy Studies series, Fiscal Design Surveys Across Levels of Governments (Jensen 2001).
Table 1. Classification of Local Taxes by Degree of Local Autonomy High Revenue
Autonomy
No Local Au-tonomy
SNG sets tax rate and base. Highest degree of own source revenues.
Most often pertains to fees and charges;
refer to Tables 3 and 4.
SNG sets tax rate only.
Necessary and sufficient condition for cat-egorization as “own revenue” (piggyback-ing, tax base harmonization/conformity permitted).
SNG sets tax rate, but only within
centrally permissible ranges. A typical practice is to cap the top rate.
Tax sharing whereby central/local revenue split can be only changed with consent of SNG.
Can result when a local authority collects the tax and remits to the center
Revenue sharing with share deter-mined unilaterally by central authority.
100% control by center; this category is a source of much misspecification of what is a central vs. local revenue (GFS includes this category as a local tax).
Central government sets rate and base
of “SNG revenue.” May accompany political decentralization.
Source: Adapted from Jensen 2001.
Intergovernmental Grants
Similar to that for taxes and non-tax charges and fees, the framework for classifying grants is struc-tured according to the degree to which SNGs can control the use of funds transferred from the central government. As discussed in the briefing note Intergovernmental Transfers, grants may be either conditional or unconditional, and as with revenues, there is a continuum (Table 2).
Table 2. Classification of Intergovernmental Grants by Degree of Fiscal Autonomy
High Au-tonomy With Respect to Transfers
Low or No Lo-cal Autonomy
Unconditional Grant (which may be formula based or in the form of rev-enue sharing)
Highest degree. Note that whereas revenue sharing was given low marks in terms of “own”
revenue autonomy, if that shared portion is returned to the SNG where is was collected, it takes on the character of a very flexible transfer.
Conditional Non-Matching
The donor government gives the recipient SNG a sum with the stipulation that funds shall be used for a specific purpose. Given the fungible nature of this grant, it can offer a degree of fiscal autonomy.
Conditional, open-ended matching
Now, flexibility is diminished due to the match-ing requirement. Though there is still a degree of flexibility due to fungibility, with the match-ing requirement it is difficult to say that this form of grant gives an SNG much flexibility (autonomy).
Conditional, closed-ended
This is as specific as a grant is likely to become.
The grant is clearly designed to influence the composition of SNG spending. Such specific grants certainly have their merits (e.g., to adjust for net externalities), it is just that they do not add to the degree of SNG fiscal autonomy.
Source: Adapted from Table A-1 from the briefing note “Intergovernmental Transfers”
Expenditures
Just as getting clarity in the sorting out of (or assignment of) expenditures is probably the most difficult intergovernmental design, so too is the expenditure side the most problematic task for determining the degree of subnational fiscal autonomy. As a generalization of that problem, the issue is the degree to which the center “interferes” with and/or “mandates” how local revenues shall be spent. The issue is a most important one since local officials will balk at taking on the political risk of raising revenues unless they can demonstrate to citizens some service delivery quid pro quo. A policymaker loaded with central restrictions on the spending of local funds cannot make that service-tax cost link.
An additional challenge on the expenditure side is that autonomy may differ across different func-tional classifications (e.g., health, education, and infrastructure). Also, developing measures of autonomy over expenditures is difficult because of the many ways central government can control spending by SNGs. For example, in South Africa and Pakistan, local governments have discretion over some expenditure categories, but negotiating wages and work rules with municipal unions is centralized at the national level (Bell 2006). A further problem is that the center may prescribe administrative arrangements for subnational governments; for example, in Pakistan the center prescribes the provincial and local (district, town) organizational charts.
Table 3, which is illustrative, presents a typology for identifying the role of SNGs in service deliv-ery as developed by Bell (2006). The various dimensions of service delivdeliv-ery include policymaking, control of the civil service, standard setting, administration, and monitoring and evaluation. Note that to get a clear sense of the degree of expenditure autonomy, one would have to develop a Table 3 matrix on a sector-by-sector basis.8
8 And then develop—or at least have some good sense of—a weighted index of the relative importance of each sector. Thus, for example, a local government might have complete control over expenditure decisions relating to a relatively “small” function such as maintaining local roadways (e.g., which road to repair and when and how to do it), but very little control over a “ large” sector that dominates the budget (e.g., the center sets the teacher and staff salaries as well as determines who to hire, promote, and/or fire).
Table 3. An Illustration of the Elements of Subnational Fiscal Autonomy with Respect to Expenditures
Factor Influ-encing Degree
of Autonomy Description of the Factor Central Control and Low
Degree of Fiscal Autonomy High Degree of Subnational Fiscal Autonomy
Broad Control over Policy
Which government sets the main policy guidelines for a service (e.g., free primary edu-cation as a national policy)?
Center either makes or has final control over the local budget; and/or can over-ride provisions of the local budget.
Clearly delineated assignment of functions; SNG controls its own budget process and budget execution institutions (accounting systems, treasury operations, internal and external audit).
Civil Service
Control over the level of the wage bill and decisions with respect to hiring, promotion and firing.
Center determines (perhaps through negotiation) the level and structure of civil servant salaries and the conditions of employment.
Local control over civil servants who are engaged in the delivery of local public goods and services. Includes agreements and settlements on wages and employment conditions.
Standards Setting and Regulation
Which government sets the standards for the composi-tion of local public services and the regulations that may accompany SNG spending programs?
(i) Central standardization imposed in circumstances where there are no clear
“spillovers.” Which unit of government sets the stan-dards? (e.g., national tests of teacher certification may be justified for their externalities;
land use zoning and building codes all fail the externalities test and are typically local). (ii) Lack of consultation when there are external effects.
Local control consistent with compliance with the law, constitutional principles and international standards of human rights.
Administration Administration of service delivery on a day-to-day basis.
Mandating of internal administrative organization and day-to-day expenditure management (e.g., procure-ment practices).
Local authorities determine their own internal adminis-trative structures in order to adapt them to local needs and ensure effective manage-ment (see European Charter 1985).
Monitoring and Evaluation
Which government monitors and evaluates SNG perfor-mance.
The center has an appropriate and important role in moni-toring fiscal flows (spending and revenues). But monitor-ing shall not be confused with control.
Clarity in revenue and expenditure assignments is achieved and a responsibil-ity to report fiscal flows, including flows from external (foreign) sources, is accepted and practiced.