Central development spending in the regions post-decentralisation

Abstract

The central government spends just under one-half of its development budget outside Jakarta. The centre provides more than one-third of total development spending carried out in the regions. Central development spending in the regions on primary and secondary sector activities and on education and culture is especially significant. The spatial distribution of consolidated public development expenditure is very unequal. However, central spending mitigates inequality in the geographic distribution of subnational development expenditure, at least to a certain extent. Agood deal of central government spending in the regions appears to be on tasks that have been turned over to subnational governments and is therefore illegal under current decentralisation legislation. Central support of specific decentralised activities could be implemented more transparently and efficiently through the government's specific purpose grant.

Acknowledgments

The authors are employed at the World Bank in Jakarta, under financing from the Dutch Trust Fund (TF050378). The views expressed here are those of the authors and should not be attributed to the World Bank or the government of the Netherlands. The authors would like to thank Bert Hofman, Wolfgang Fengler and two anonymous referees for useful comments on an earlier draft of this paper, Ahya Ihsan, Bambang Suharnoko and Bastian Zaini for valuable research assistance, and the Ministry of Finance for access to data.

Notes

In this paper, we use the terms ‘subnational’ and ‘regional’ interchangeably to refer to provinces and districts/municipalities (kabupaten/kota) collectively.

Ministry of Home Affairs Decree 29/2002 on Guidelines on Regional Financial Management, Responsibility, and Supervision and Form of Regional Revenue and Expenditure Budgets, Regional Finance Implementation and Administration, and Compilation of Regional Revenue and Expenditure Budgets.

Most analysts assume that central government routine expenditure in the regions is limited because it is restricted to that needed to support the deconcentrated offices of central ministries; since 2001 there have been very few such offices (see Law 22/1999 on Regional Administration, Government Regulation 39/2001 on the Implementation of Deconcentration and Government Regulation 52/2001 on the Implementation of Assistance Tasks). We have no data on such spending and do not address this question in the paper.

According to Law 22/1999, the central government's development spending in the regions on its own tasks may be implemented through two different mechanisms: deconcentration (dekonsentrasi) and assistance tasks (tugas pembantuan). Under deconcentration, the central government may delegate some of its functions to a province, where the latter is acting in its capacity as the centre's representative in the region. Under assistance tasks, the centre may ask subnational governments or villages to undertake functions on its behalf; this same system also allows provinces and districts to request villages to undertake some of their respective service responsibilities. In all cases, the delegating level of government must finance the delegated function from its own budget, and the employed funds are not deposited in lower-level government budgets. See also Government Regulations 39 and 52, both of 2001.

See, in particular, Law 22/1999 and Government Regulation 25/2000 on the Authority of the Centre and Provinces in Their Capacity as Autonomous Governments.

It is possible that provincial governments spend money within their jurisdictional boundaries on district functions. We have no knowledge of or data on such spending and do not consider the question further.

Jakarta is excluded from analysis here due to lack of data.

These estimates of central spending in the regions are based on 89% of total central development spending in 2001 and 84% in 2002. Preliminary data for 2003, based on 100% of realised central development expenditure, imply that the portion of central development spending carried out in the regions is just over 50% of the total.

We thank one of the reviewers of this paper for suggesting the example of road construction.

According to Law 22/1999, the central government has full authority only over foreign affairs, national security, justice, monetary and fiscal affairs and religion, as well as policies concerning national planning and macro national development, fiscal balance, the system of national administration and national economic institutions, development of human resources, efficient use of natural resources and strategic technologies, conservation and the institution of national standards.

See Government Regulation 25/2000.

For example, detailed data on Ministry of National Education programs and projects for 2002 show that the vast bulk of programs is related to primary and secondary school activities, which are the responsibility of districts.

See Dom (2002) for a description of the kinds of (mostly primary and secondary school) tasks that the Ministry of National Education supported with its development spending in 2001 and 2002, and a review of the funding mechanisms it used and/or planned to use to bypass district budgets.

Law 25/1999 stipulates that decentralised functions, as outlined in Law 22/1999, must be financed by subnational budgets (APBD). Atwo-year transition period, outlined in the same law, that permitted the use of APBN funds to directly finance decentralised functions expired in May 2001. Some sectoral legislation may conflict with this view of the illegality of central spending in the regions on decentralised tasks.

Law 22/1999 assigned 11 areas or fields (bidang) of responsibility to local governments: public works, health, education and culture, agriculture, communications, industry and trade, investment, environment, land, cooperatives and labour. The precise functions of districts within those broad areas of responsibility were not made clear, however. In addition, the law stipulated that local governments became responsible for all functions not implemented by the central or provincial governments, which had the effect of further obscuring local responsibilities. Government Regulation 25/2000 provided more detail on the functions of the centre and provinces, specifically concerning the 11 above-named fields, but did not further specify the tasks of local government.

The central government's contention that it performs some functions more effectively than subnational governments would do is not limited to central line ministries. One of the arguments made by some Ministry of Finance officials in Indonesia in defence of continued central administration of the property tax is that it has performed well under their authority and that its decentralisation would inevitably result in weaker operations and outcomes. In a review of this line of reasoning, Lewis (2003) provides evidence to suggest that only around 40% of Indonesian local-sector property tax potential has been realised under central government control, thus casting doubt on the centre's assertion.

Detailed data on per capita development spending by level of government are presented in table A1 in the appendix. Data on central spending in the regions are available only at the provincial level. District spending has been aggregated to the provincial level so as to be comparable with central and provincial expenditure data.

The Gini coefficient has been calculated based on the cumulative percentage of total expenditure versus the cumulative percentage of population, within provinces, as is the usual practice. The value of the coefficient may range between 0 and 1. Relatively greater levels of inequality are represented by relatively larger values of the Gini.

We implicitly assume here that the expenditure of subnational governments is a function of their revenue (and not the other way around, for example). Intergovernmental transfers make up the bulk of subnational government revenue and such revenue is an especially significant determinant of expenditure. Aregression of local government per capita expenditure during 2001and 2002 (pooled) on per capita transfers from the centre shows that over 95% of the variation in the former is explained by the latter. This at least supports the assumption that subnational expenditure is driven by revenue. Many questions might be raised about this quick example but this is not the place to explore such issues in depth. Chakeri and Lewis (2004) examine local government revenue budgets by type of revenue, and Lewis (2004) analyses the determinants of district expenditure (and other fiscal outcomes).

As of 2003, 100% of shared natural resource revenue has been counted towards the estimation of fiscal capacity in the context of determining DAU allocations.

More generally, see Lewis (2001, 2002), Fane (2003) and Hofman, Kadjatmiko and Kaiser (2004) for some analysis of the weak equalising effects of the DAU.

The distribution of funds through the DAK would necessarily be more transparent than the use of central development spending, since the former must be revealed formally via Ministerial Decree, whereas the latter is hidden inside unpublished departmental budgets. Greater efficiency of DAK-funded subnational spending is not assured. But the decentralisation of authority for funds linked to local service responsibilities, such as would obtain under the DAK system, is commonly advocated on economic efficiency grounds.

See Leigland (1993) for an examination of the dominant role of Bappenas in national development budget formulation and the historical difficulties associated with decentralising development planning to subnational levels.

The Ministry of Finance will take primary responsibility for formulating a unified budget (i.e. incorporating both routine and development-cum-capital components) starting in 2006. Bappenas will remain involved to an as yet unknown extent.