- Annalisa Fedelino
- Published Date:
- October 2010
Indonesia is often cited as an example of a “big-bang” approach to fiscal decentralization. Until the late 1990s, Indonesia’s decentralization efforts had proved timid, with deconcentration of service delivery to regional governments in the 1980s and 1990s being used to address dissatisfaction with lack of control over the allocation of public expenditures, especially in natural resource-rich regions.1 With the 1997–98 economic and political crisis, demands for increased regional autonomy and the need to stem deep political turmoil added urgency to the decentralization agenda, leading to more-rapid attempts at decentralization. Separatist movements gained strength again, particularly in resource-rich regions such as Aceh, which had felt deprived of their natural resources by the central government.
Responding to these pressures, the government quickly drafted and enacted the basic legal framework for decentralization during 1999–2001. This was an ambitious plan to hold together a country with more than 200 million people (the fourth most populous in the world), speaking more than 200 languages, and comprising in excess of 17,000 islands.
Extensive technical assistance from both the IMF and the World Bank stressed the importance of moving carefully on the proposed decentralization agenda. Several missions were fielded during this period (and advisors were posted at the Ministry of Finance) to provide advice on the key dimensions of the envisioned reforms, including on the specific services to be devolved to local government; design of own-taxes, revenue-sharing, and equalization transfers; and subnational borrowing. Appropriate sequencing of expenditure and revenue assignments, coupled with accountability mechanisms, was seen as crucial to minimizing possible risks from a hastened decentralization process.
|Type of government||Unitary|
|Population (million, 2007)||231.6|
|Area (million km2)||1.9|
|Levels of government||3|
|Provinces (including special regions)||33|
|Districts (regencies and cities)||510|
|Average district size (population)||498,153|
|Minimum district size (population)||14,065|
|IMF technical assistance missions on fiscal decentralization||1998, 1999, 2000|
Despite the call for cautious implementation, a major decentralization reform began in January 2001. All public services except those explicitly expected to be carried out by the central government were delegated to subnational governments. The framework also led to a sharp increase in transfers to subnational governments, and gave regional governments more scope for own-revenue collection. Since then, subnational governments have enjoyed almost complete authority over the spending of their fiscal resources. Subnationals now manage nearly 33 percent of total public expenditures (Table 8.2) and carry out about 50 percent of development expenditure.
|Total subnational revenue||40.3||8.2|
|Transfers from the central government1||33.3||6.8|
|Total subnational expenditure||34.5||6.9|
|Overall balance (before transfers)||n.a.||-5.5|
Transfers include shared revenue.
Transfers include shared revenue.
Initial implementation surpassed expectations, and no substantial disruption of public services occurred. By the end of the first year, over 16,000 public service facilities had been transferred to the regions, along with responsibility for about 2 million civil servants. Regional spending doubled from the previous year. The effects of the reforms can be seen in the significantly increased transfers to subnational governments, from 1.5 percent of GDP before the reform (FY1999/2000) to a peak of 6.8 percent of GDP in 2006.2
Since the 2001 reform, the number of subnational governments has increased. Currently, Indonesia comprises 33 provinces (provinsi), including special status regions (Nanggroe Aceh Darussalam, Papua, and West Papua), Yogyakarta Special Region (DIY Yogyakarta), and Jakarta Special Capital Region (DKI Jakarta); and 510 cities (kotapodya) and regencies (kabupaten), including six within the capital Jakarta.3 These compare with 27 provinces and fewer than 300 districts before the reform.
IMF Advice on Fiscal Decentralization
In the early stages of the government reform, IMF staff advice focused on proper and gradual sequencing of decentralization. A December 1998 staff mission first assessed plans for a law on decentralization, also covering services to be devolved. At that early stage of the debate, the staff team noted its fundamental concern with the revenue-sharing arrangements—at that point there was no agreement on the specific services to be devolved, taxes to be assigned, or borrowing arrangements; analysis of distributional effects across local governments was lacking, as were estimates of the transitional costs; and no detailed plans for the transition had been laid out. Given the political pressures to move forward, the mission recommended that legislation be limited to broad principles, without including a specific blueprint for implementation.
Political realities, however, dictated a fast-tracked decentralization, and two decentralization laws were quickly drafted. The Law on Regional Governance, under the authority of the Ministry of Home Affairs (MoHA), focused on enhanced administrative and political decentralization and included the only references to the devolution of expenditure responsibilities. Meanwhile, the Law on Fiscal Balance, being drafted by the Ministry of Finance (MoF), governed the distribution of resources across regions. Two separate ministries were drafting these laws under a very tight timetable resulting in the emergence of some inconsistencies. A second staff mission in early 1999 provided advice on the legal framework for decentralization, recommending that expenditure assignments should follow local capacity; transfers of resources should follow those of expenditures; and decentralization itself should be fiscally neutral. It also strongly encouraged the authorities to coordinate closely the political and institutional legislation being prepared by the MoHA and the MoF.
In line with the authorities’ plan to move expeditiously, both decentralization laws were approved by Parliament in May 1999. In addition, a new Regional Taxation Law (Law 34) was adopted in 2000, allowing local governments to introduce new taxes.
The Law on Regional Governance broadly defines assignments for different levels of government. First, the law drastically rebalanced power in Indonesia by breaking the hierarchical relationship between provincial and local governments; both levels now have the same direct hierarchical relationship with the central government. All public services were delegated to districts, except those expenditures explicitly reserved for the central government, including, among others, international policies, defense and security, judiciary, and monetary and fiscal policy. Districts were given responsibility for key public services such as public works, health, education, agriculture, and local infrastructure; and provinces were made responsible for the provision of services spanning districts, and held as a fallback if a district could not perform an assigned function.
Staff advised that the actual transfer of responsibilities be linked to the institutional capacities of the local governments to take on those new responsibilities. They also recommended that the devolution of expenditure to local governments be underpinned by a strengthening of budgeting, monitoring, and auditing arrangements at the local level to improve transparency and avoid misuse of public funds.
The scope for own-revenue collection expanded greatly with the 2000 Regional Taxation Law. Regional governments can add taxes through regulations approved by regional parliaments, without central government approval, as long as they abide by certain criteria: new regional taxes should not conflict with public interests, and their bases should not be subject to tax by the center. The central government has the authority to annul regional taxes breaching these criteria. However, local tax assignments remain inadequate, without major sources of own-revenues. In 1999, staff had strongly (but unsuccessfully) advised that the land and building tax, collected by the central government and shared with regions, be assigned to the local level, to provide the localities with a degree of own-revenue-raising responsibility.
The importance of revenue-sharing arrangements significantly increased with decentralization. The Shared Revenue Fund (DBH) is split into two parts, the Tax DBH and the Natural Resources DBH. The former includes income tax (20 percent), land and building taxes, and taxes on land and building transfers (80–90 percent).4 The Natural Resources DBH includes revenues collected by the central government based on natural resources (e.g., oil, gas, forestry, fishery, and mining), and is shared with subnational governments according to specified rates.5 The sharing formulae include some elements of redistribution, but not enough to affect the overall unequalizing impact of the system.
These arrangements, made after intense negotiations between the center and local governments, led to several weaknesses in the decentralization framework. Staff had advised against sharing oil and gas revenues on an origin basis because it would exacerbate inequalities in regional revenue capacities (three provinces got about 80 percent of the total local share). In addition, highly volatile prices and output would complicate the local governments’ budgeting of shared resources.
The system of grants was radically reformed with the 2001 decentralization. A General Allocation Fund (Dana Alokasi Umum, or DAU) was created as the cornerstone of the intergovernmental fiscal system. The grant constituted over 60 percent of all central transfers to regions; it replaced the system of transfers to regions composed of the Subsidy for Autonomous Regions and Presidential Instruction Grants for sectoral expenditure that existed before 2001. The law required the central government to transfer to the subnational level a minimum of 25 percent of central government revenues after tax sharing and minus energy subsidies. Regencies and cities received 90 percent of the DAU pool, with the provinces receiving the remaining 10 percent. The DAU allocation to individual provinces and local governments is, in principle, determined by formulae (separate ones for local governments and provinces). Several transitional elements were added to the DAU distribution in addition to the formula, including a base amount, used as a floor, and a contingency allocation to protect regions receiving lower funds than needed.
A special allocation grant (DAK) was established for earmarked transfers. The law states that DAK can be used for special needs of the regions, including emergencies, and for financing central priorities at the regional level. It initially covered five types of grants (education, health, rural roads and irrigation, public administration infrastructure, and forestry).6 Regions are supposed to apply to the central government for the grant, with proposals that meet the allocation criteria. Although DAK-related amounts are nominally small, their share in total grants (excluding revenue-sharing) from the central government increased from about 1.5 percent in 2001 to 10.5 in 2009.7
Staff had advised extensively on the grant design. They recommended an equalization grant system based on relative expenditure needs and revenue capacities—as reflected in the Law on Fiscal Balance. Staff had also recommended that specific grants for sectors and projects be focused on areas for which minimum service standards are societally important. Nevertheless, their main advice against the floor on general allocation fund transfers, set at a minimum of 25 percent of net domestic revenues, was not included in the law.8 The staff’s advice reflected a concern that the floor could pose significant macrofiscal risks in the short run (by leading to a higher central government deficit) at a time when districts were not in a position to use transfers adequately.
Despite some equalizing features of the intergovernmental fiscal transfers, regional inequalities had initially increased with decentralization. Wealthy regions were able to retain a high share of their natural resource revenue and were in a position to mobilize more own-resources. They also received higher general-purpose (DAU) transfers than implied by a formula based solely on a fiscal gap measure. As a result, inequality after transfers—as measured by the Gini coefficient—increased with decentralization (according to the World Bank Public Expenditure Review, 2007).
Public Financial Management Arrangements
Given Indonesia’s extensive countrywide network of treasury offices, staff advised the central government to offer to provide subnational governments with treasury services rather than leaving each province and district to set up its own arrangements for managing treasury operations. These services would include tax collection, payments, and accounting. In the end, the authorities opted for letting subnationals set up their own arrangements for budget and treasury management, despite staff’s position that strengthening public financial management should be a prerequisite for increasing decentralization. This is also one factor behind the difficulties with subnational reporting.
Some progress was made on subnational accounting. The MoHA drafted its own chart of accounts, which initially was different from the new chart being drafted for the central government. In line with staff advice, one single chart of accounts was eventually adopted (the two separate charts were merged, facilitating fiscal reporting according to a common format). However, it is unclear the extent to which common reporting formats have been adopted at the three levels of government.
Although the legal framework for fiscal decentralization covers local governments, its application remains largely limited to the central government. While the 2003 State Finances Law and the 2004 Treasury Law cover subnational governments, as do the 2007 Cash Management Regulations, in practice, these are observed only by the central government. The same applies to the 2004 Planning Law and the 2007 External Audit Laws. For external audit—important for accountability—only about 60 percent of subnational government accounts are audited regularly (World Bank, 2007).
The fiscal balance and debt rules for general government are not monitored regularly. Two fiscal rules (capping the deficit at 3 percent of GDP; and the gross debt at 60 percent of GDP) were adopted, as per the Fiscal Balance Law 33/2003. However, irregular reporting by subnationals on their fiscal balances and the unavailability of comprehensive data for subnational debt make it impossible for the central government to monitor compliance with the debt rule, and the deficit rule is monitored with an excessively long lag.
Both the Law on Regional Governance and the Law on Fiscal Balance allowed regional governments to borrow. External borrowing is limited to onlending from the central government. Rules-based controls for domestic borrowing mandated that the total stock of debt not exceed 75 percent of a region’s total receipts in the previous year’s state budget. The law also states that long-term loans can only be incurred for income-generating projects. The central government can withhold DAU transfers (and the shared DBH) if a region fails to meet its debt obligations to the central government. These provisions were largely in line with staff advice from the 1999 mission.
An Assessment of Recent Reforms and Remaining Challenges
Local governments have witnessed a surge in revenue transfers from the central government, but have struggled to increase spending rapidly enough to meet large social needs. Because oil and gas revenue are shared with local governments, transfers to the regions grew threefold from 2002 to 2008. These increased transfers, combined with increases in deconcentrated central government spending, a lack of clear delineation of spending authority between the levels of government, weak budgeting capacity of local governments, and a fear of corruption investigations, caused underinvestment and underspending to occur in many local governments. Nonetheless, subnational government spending continues to account for about one-third of public spending.
Subnational governments have been building up large fiscal surpluses. As a result, sizable regional government deposits have accumulated, although this trend slowed somewhat in 2007 and deposit accumulation remained low in 2008 (the stock at end-2008 was estimated to be about 0.2 percent of GDP, possibly suggesting improvements in budget implementation). However, significant differences exist among districts, with a few resource-rich districts accounting for most of the accumulated deposits.
Subnational government debt remains small.9 However, given the lack of adequate mechanisms for coordinated fiscal and cash management, the large bank account balances of subnational governments have complicated the conduct of monetary policy for the Bank of Indonesia, especially given that regional governments or provincial banks purchase short-term Bank of Indonesia securities.
A number of changes to the system of intergovernmental fiscal relations were introduced in 2008–09.10
- The need to increase the taxing power of regions has been recognized, including in ongoing discussions on a draft law on Regional Tax and Retribution. Some new taxes are proposed to be added, encouraging regions to set higher and progressive taxes on motor vehicles and introduce other related charges. Although the land and building taxes (earlier recommended by staff, as explained above) are not mentioned, the 2009 budget still recognizes the decentralization of these taxes as a future priority.
- Starting in 2009 (according to an earlier-envisaged provision), if the Indonesian crude oil price exceeds 130 percent of the price assumed in the budget, the additional revenue from oil and gas will be shared with regions as additional DAU on the basis of fiscal gap considerations.
- DAU allocation policies have changed significantly, and are now fully based on (1) a “basic allocation,” computed based on total salary of regional civil servants; and (2) the “fiscal gap,” defined as the difference between fiscal needs and fiscal capacity (both calculated based on specified parameters).
- The definition of net domestic revenue (26 percent of which funds the DAU) was changed; it now excludes subsidies (including for fuel), with the objective of sharing subsidy costs among levels of government. The introduction of this “burden-sharing” policy means that the total DAU pool in 2009 barely increased.
- Finally, the proportion of the basic allocation in the DAU formula is to be reduced every year, so that the fiscal gap will become increasingly relevant; and the DAU is to be based over time on standardized spending needs indicators. The phasing out of the hold-harmless provision should help reduce inequalities and make the transfer system more equalizing.
Despite these improvements, weak fiscal reporting capacity at the subnational level continues to limit fiscal policy management and coordination. Regional governments are required by law to regularly submit their budgets and to report on budget execution to the central government, but long delays occur—sometimes as long as two years on the execution side—and make it difficult to regularly monitor local governments’ financial operations.11 Therefore, fiscal policy is conducted on the basis of the central government balance. Although regular monitoring of the local governments’ fiscal stance is still feasible through examination of changes in regional government deposits held at commercial banks (the main source of financing for subnational governments), data have proved more difficult to compile than initially expected.12
More generally—and as highlighted by a 2007 staff mission—fiscal risks could arise from subnational governments in the future (Davis and others, 2007).13 Although immediate risks to the central government budget appear limited, they could rise over time as a result of (1) insufficient coverage of the government sector, which can distort the assessment of the fiscal stance; (2) a weak decentralization framework with limited information systems in place; and (3) high dependence of local governments on potentially volatile transfers. This volatility is to be somewhat moderated by the implementation of the new formula for calculating DAU transfers, which excludes fuel subsidies from the calculation base.14 On the positive side, legal limits exist on domestic borrowing by the regions, provided the relevant information for their enforcement is readily available. As indicated above, reporting and monitoring are still issues; the 2009 budget mentions steps that the government is taking to strengthen sanctions for violating regional loan provisions, as well as to improve coordination in planning, evaluating, and monitoring regional loans. These are welcome steps, although additional progress is warranted to ensure sound fiscal policy design and implementation at the general government level.
Following independence in 1945, decentralization attempts dating to the early days of the Republic were of set by the center’s reluctance to grant significant autonomy to the regions. Legislation was first passed in 1957 to revitalize regional autonomy but was stopped after regional unrest in Sumatra and West Java. A new law in 1974 attempting to revamp the process was never implemented because of concerns about the administrative capacity of regional governments.
This could have also been the result of the “hold-harmless” policy (stipulating that the regions will not receive lower transfers than in the previous year), as the number of regions grew. The hold-harmless provision was lifted in 2008 (when transfers amounted to 5.8 percent of GDP).
Cities and regencies are also called “level II regions” or districts. Their number has been increasing over the years, and is expected to continue to do so.
“Tobacco yield excises” are also shared with tobacco-producing regions (2 percent), but the arrangement is designed as an earmarked grant.
These rates vary from 15.5 percent of total collection for oil to 30.5 percent for gas revenues and 80 percent for general mining, and forestry and fishing. Special (higher) rates originally applied to resource-rich special autonomy regions; however, the peace deal with Aceh in 2005 gave Aceh virtually total control over all natural resource revenues.
The number of sectors funded has varied over the years, depending on the central government priorities (for example, the 2009 budget includes 13 sectors).
The central government also allocates deconcentration and coadministration funds to regions, funded from the central budget. Deconcentration funds are handed out to governors as representatives of the central government’s line ministries in the regions; coadministration funds are disbursed to regional governments or village administrations. The combined amount of these funds is far greater than DAK transfers. In principle, Article 108 of Law No. 33/2004 states that deconcentration and coadministration funds (which have objectives similar to those of DAK) are to be transferred in stages to DAK in the future.
The provision was actually removed by the executive but reintroduced by Parliament.
According to the 2009 budget, outstanding loans of regional governments at end-2007 (onlent foreign loans and domestic loans from the central government) amounted to 0.02 percent of GDP.
The decentralization laws were revised in 2004. Major changes included an increase in the share of oil and gas receipts to the regions by 0.5 percentage point by 2008, with the funds earmarked for education spending; “excess revenue” being kept with the center when oil prices rise by more than 30 percent above the budget price; basic grant allocation no longer being a minimum amount, and reduced by the extent to which fiscal revenue-generation capacities exceed fiscal needs (based on formulae); and reporting by regions of their borrowing and debt positions to the center twice a year under the threat of delayed transfer payments.
Submission of local government budgets to the MoF significantly improved starting with the 2008 budget, particularly as a result of the introduction of sanctions (delaying DAU and DAK transfers); by end-February 2007, 46 percent of local governments had submitted their 2007 budgets to the MoF. For the 2008 budget, that proportion had reached 82 percent by end-February. In 2009, three local governments were penalized, down from five in 2008 (out of 510 districts).
A mission from the IMF Statistics Department in April 2009 found that work had progressed on the pilot test of the Local Government Finance and Governance Reform project—implementing systems to improve the timeliness and quality of subnational fiscal data—but that there was likely to be a long lag before complete data become available for all local governments. Accordingly, the mission recommended that a simple, but robust, sample expansion approach be adopted to estimate data for the local government subsector.
The issue of weak and untimely reporting and its possible implications also featured prominently in the 2006 ROSC Report on Fiscal Transparency (IMF, 2006).
Based on the new transfer formula, the share of transfers to GDP is expected to decline in 2010, and over time, to about 5 percent of GDP.