Chapter 5. Fiscal transfers in Asia: Pitfalls and opportunities for achieving the Sustainable Development Goals

Roger Shotton
Uyanga Gankhuyag

This chapter is based on a wider review of the same issues prepared by the authors for the United Nations Development Programme (UNDP), available at

To achieve the Sustainable Development Goals (SDGs), increased and better quality public spending is needed, as indicated in Figure 5.1.

Much of this public expenditure is mandated to subnational governments (SNGs) in Asia. For example, Gram Panchayats and Union Parishads, the lowest SNG tiers in India and Bangladesh respectively, are usually responsible for: building and maintaining village roads and bridges; water supplies; irrigation; early education; primary education and primary health facilities; as well as for managing various social welfare programmes. Vietnamese communes, Indonesian kapubatens and Mongolian soums all have similar responsibilities. The range of SDG-critical public expenditures then widens much further when the spending responsibilities of higher-tier districts, provinces or regional SNGs in these countries are also factored in (OECD and UCLG, 2016[2]).

Achieving the SDGs requires more and better public spending on their mandates by these SNGs. No matter how well prepared the sustainable development policies and plans are, there can be little progress unless these policies and plans are operationalised into locally appropriate budget-spending priorities. These budgets need to be executed efficiently, to reflect the right balance between investment and operational spending, to ensure sustainability of benefits. Finally, overall, budget resources must be allocated and spent effectively, efficiently, equitably and accountably.

Fiscal transfers to SNGs1 are the major source of financing for SDG-related expenditures for most SNGs worldwide, and certainly in Asia. They matter for achieving the SDGs in several ways. First, and most obviously, the volume of resources transferred will determine the levels of local spending on sustainable development priorities. Second, the manner of their allocation across SNGs will also affect territorial equity in spending and hence may promote – or undermine – progress on SDG 10 (Reduce inequality within and among countries). Third, and less obviously, fiscal transfers often also carry various unintended consequences that can shape local budget processes and priority-making decisions, and so directly affect the level and quality of local spending on SDG priorities.

Fiscal transfers can also be expressly designed to carry positive incentives to promote better local performance. To this end, there are now a number of initiatives in Asia that offer important lessons.

With the exception of large, wealthy metropolitan areas, the levels of revenues mobilised by subnational governments around the world are almost always well below the levels of spending mandated (UCLG, 2010[3]); this is particularly true of rural SNGs in Asia. This asymmetry arises from the fact that the major national revenue sources are managed more efficiently and equitably under central control, added to which there is often central resistance to decentralise even those revenues that might be better placed under local control. The relative importance of subnational spending in total government expenditure in Asia varies greatly – from a mere 4% of government expenditure in Cambodia to 85% in the People’s Republic of China (hereafter “China”) (OECD and UCLG, 2016[2]).2

Overall, in Asia, as seen globally, wealthier countries tend to display greater degrees of decentralisation of spending responsibilities. But this rough correlation is of course qualified by structural and political country contexts – the subnational share in overall government spending also reflects the history, size and configuration of subnational administrations, as well as the strength of the national political and policy drives to assign spending responsibilities and resources to the subnational level. Thus Viet Nam displays a relatively high degree of decentralisation of both expenditures and revenues (55% and 33% respectively), reflecting the historical importance of provincial administration and local party structures. Thailand, on the other hand, which is notably wealthier, remains relatively centralised (with only 17% of decentralised expenditures and 8% of decentralised revenues), reflecting the historical dominance of the central bureaucratic and military establishment, as well as the reluctance to empower subnational governments.

However, in all cases, there is a “fiscal gap” at the subnational level, which governments seek to fill through intergovernmental fiscal transfers. While the primary objective of fiscal transfers is to address these SNG fiscal gaps and supplement local spending capacities, there are also sometimes other policy objectives. One such common objective is to use transfers to equalise spending capacities “horizontally” across SNGs, recognising that “fiscal gaps” may vary considerably across the national territory.

Other transfers may have different objectives: to encourage SNG spending in specific sectors deemed of national priority; to address the costs of social, economic or environmental externalities faced by SNGs, or the costs of “spillovers” across jurisdictions; to address what may be a region-specific spending urgency or to implement a specific national government programme. Transfers may also be established for political reasons – e.g. to recognise local political demands to share in the benefits of natural resource extraction activities, perhaps to discourage tendencies toward secession; or to give parliamentary representatives a spending fund for their constituencies.

Fiscal transfers are established in different ways to meet these various objectives. There are many typologies of transfers (Boadway and Shah, 2001[4]), but for the purposes of this chapter, the following main types of fiscal transfer are:

  • Grants. These may be of two types: unconditional grants (UCGs) allowing for wide discretionary use by SNGs, or conditional grants, whose use is restricted to a particular type of expenditure (which may be defined by sector, by expenditure classification, by type of beneficiary, by geographical area, etc.). Very often, UCG transfers are also intended to perform an equalising role across SNGs.

    • One variant to be noted here is the “gap-filling” grant transfer, common in (former) socialist countries. This is a transfer to SNGs whose use has been largely pre-determined in the budgeting process, by virtue of the nesting of subnational budgets into the state budget (Ebel, Wallich and Bird, 1995[5]; Martinez-Vazquez and Boex, 1999[6]; Bahl, 2000[7]; Dabla-Norris and Wade, 2006[8]).3

  • Revenue-sharing (RS) arrangements. Where specific revenues are shared “downward”4 from central government to the SNG areas where the revenues derive from, on a percentage basis. These transfers are usually open for discretionary use, though in some cases their use is restricted. RS arrangements are very often related to natural resource extraction activities – e.g. oil and gas, mining, logging, hydropower, etc. (Bauer et al., 2016[9]).

There is considerable variation in both the magnitudes and combinations of these sorts of fiscal transfer arrangements across Asia.

Overall, fiscal transfers to SNGs in South Asia are dominated by various grant mechanisms (although Indian states and Pakistani provinces also receive revenue-sharing transfers from the federal government). Socialist/transition countries in South East and East Asia, such as China, Viet Nam, Lao People’s Democratic Republic (hereafter “Lao PDR”) and Mongolia, have historically featured complex revenue-sharing arrangements between levels, which in many cases are now being reformed into a combination of conditional grants, “gap-filling” grants and “downward” RS transfers. Indonesia has instituted a significant natural resource-related revenue-sharing mechanism alongside a mixture of unconditional grants and conditional grants as part of its “big bang” decentralisation reforms initiated 20 years ago. Thailand maintains sets of UCG and conditional transfers as well as RS based on commercial, business and other taxes and fees.

Not surprisingly, in federal states such as India and Pakistan, there is considerable variation in the transfer patterns within different states and provinces. However, even some unitary states such as China (Wong, 2007[10]; Man and Hong, 2011[11]; Wang and Herd, 2013[12]) and Viet Nam (World Bank, 2015[13]) also allow a substantial degree of autonomy to provinces in establishing sub-provincial transfer arrangements through their state budget laws.

Over time, in some countries such as India, Nepal or Viet Nam, fiscal transfer systems became more complicated, with a proliferation of different sectoral or programme conditional grant transfers, each with their own allocation criteria and procedures, reducing – in some cases for better, but often for worse – the degree of local discretion in spending, and complicating local planning, financial management and reporting. In India, however, this trend is now being reversed, with a shift towards discretionary UCGs, to promote local discretion and leverage the benefits of decentralised decision making (Government of India, 2015[14]).5

A number of countries have moved away from ad hoc transfer arrangements to allocations towards more stable, transparent and predictable rules-based arrangements for the financing of both the allocable pools and for the allocation of transfers to individual subnational governments. For example, in the Philippines, the Local Government Code (1991) specified a 40% share of national revenues going to the Internal Revenue Allotment (IRA) to be allocated to subnational governments (Government of the Philippines, 1991[15]). Since 1994, China has been implementing major reforms to its complex, vertically negotiated sharing arrangements, placing transfers to provinces within a more stable and transparent rules-based framework, although these reforms are still ongoing (Wong, 2000[16]; Wang and Herd, 2013[12]; Man and Hong, 2011[11]). In Indonesia, the Law on Fiscal Decentralization of 1999 dictates a specified share of the national budget be allocated to the Balancing Fund transfers to SNGs (ADB Institute, 2016[17]). More recently, in Cambodia the national pools for the two main UCG instruments, the District/Municipal Fund and the Commune/Sangkat Fund, are now linked to specific percentages of national government revenues; and since 2015, in Myanmar, the main UCG grant pool for transfers to states and regions is linked to gross domestic product (GDP) growth and allocations now made by formula rather than by the previous “gap-filling” arrangements (Shotton, Yee and Oo, 2017[18]).

Despite the reforms being made, transfer mechanisms across Asia are typically beset by a number of problems (Smoke and Kim, 2003[19]; White and Smoke, 2005[20]; Martinez-Vazquez, 2011[21]):

  1. 1. Inadequacy. Most fundamentally, the volume of resources budgeted for subnational government transfer pools are generally inadequate. Given the critical nature of most of the devolved services at risk of underfunding, this will be a serious constraint on achieving local SDGs. To some extent, this inadequacy of central budgetary allocations for fiscal transfers is simply a reflection of the overall budgetary constraints faced by most Asian countries. But it is also the result of two other factors. First, there is typically weak advocacy for SNG budget interests – as compared to that for central sector ministries and national programmes – in the national budget allocation process. Second, there is a general lack of information on the volume of resources SNGs require to properly fulfil their service delivery expenditure mandates, due simply to lack of basic “groundwork” research on service standards and delivery costs. In other words, the size of the real fiscal gap is often an unknown quantity (Figure 5.2).

  2. 2. Unreliability. Moreover, even the budgeted levels of transfers are sometimes not fully released to SNGs. This may be for several reasons: the actual central revenues allocated to the national pool are less than those estimated in the original central government budget – which itself may be due to an unforeseen economic downturn or to bad revenue forecasting; national budget priorities change in the course of the year (this is more of a problem where the arrangements for financing the national transfer pool have not been specified in law); or the central government is unable to approve the release of all the budgeted transfers within the fiscal year, due to SNG capacity, reporting delays or related treasury blockages.

  3. 3. Lack of clarity and co-ordination. Transfer mechanisms are sometimes instituted without clarity as to their policy objectives – e.g. this is frequently the case with revenue-sharing arrangements, as discussed below. Even where there may be clarity within central government, very often there is inadequate guidance to SNGs as to the spending scope, rules and procedures for spending resources transferred. There is also a tendency for transfers to proliferate in response to particular political or bureaucratic interests or their passing priorities but without overall co-ordination of their policy aims or monitoring of the total resource flow implications.

More generally, transfers often unintentionally create geographic disparities in public spending and may also transmit perverse incentives for local performance. The next sections look at these issues in more detail.

Achieving SDG 10 to reduce inequalities and achieve the goal of “leaving no one behind” means that public spending should be geographically equitable.

Overall, geographic spending patterns across the national territory reflect the sum of both spending from subnational government budgets on their devolved mandates as well as of deconcentrated spending by ministries and other central agencies from the central government budget.6 This chapter focuses on spending for those mandates devolved to SNGs, which are financed by own revenues and transfers, rather than on deconcentrated spending.7

SNG own-source revenues usually provide only modest budget resources in aggregate. Nevertheless, these may still cause substantial horizontal inequities between SNGs, given the inevitable variations in revenue bases between regions and more and less urbanised areas.

Fiscal transfers constitute the bulk of resources for most SNGs and their allocation has a direct impact on the equity of public resources and spending across SNG localities.

As noted earlier, in some countries, there is a proliferation of transfer flows. The equity of geographic distribution of these resources is the outcome of two factors: allocation arrangements and the size of the national pools for each of the various transfers.

Revenue-sharing fiscal transfers do not aim to promote horizontal equity. Instead, they aim to promote other objectives – simply supplementing SNG spending overall, addressing socio-economic externalities, or satisfying local political claims on resources being taxed. These transfers are therefore almost certain to result in inequities across SNGs, given the usually very uneven geographic distribution of the revenue bases concerned, such as income, profits and sales tax revenues, and natural resource-related revenues in particular. Revenue sharing will, therefore, very likely further compound the disparities that already arise from own-revenue assignments.

It is the role of unconditional grant transfer instruments to play the equalising role to address horizontal disparities between SNGs. UCGs are usually allocated to SNGs by a formula that aims to capture broad, proxy measures of relative spending need and fiscal capacity (Martinez-Vazquez and Boex, 2006[22]). There are two sides to relative need. On the one hand, SNGs will have different expenditure needs due to different population sizes, levels of development and poverty incidence, physical conditions of the area, and differing service unit delivery costs arising from differing population densities or degrees of remoteness.8 On the other hand, SNGs will have different levels of own-source revenues and shared revenues due to differing levels of economic development, urbanisation, size and the composition of tax bases.

Ideally, fiscal transfer systems are established to take all of these factors into account and ensure “equalisation” of need and fiscal capacity, but all too often, they fail to do so in practice, for various reasons.

A frequent source of inequity in transfer systems lies in the allocation criteria for grant transfers – and particularly those unconditional grant transfers that are intended to play a fiscal equalising role. Different problems may arise:

  1. 1. Inadequate scope of the allocation formula. Generally speaking, three types of “equalising formula” exist based on the ranges of factors taken into account (see Figure 5.3). In many cases, the allocation formula only considers relative expenditure needs; it does not consider the relative revenue generation capacities of SNGs [the Type (a) formula]. Even when own revenues are taken into account, the revenue-sharing transfers to SNGs may not be [the top formula of Type (b) that only takes into account SNGs’ own revenues].

  1. 2. Problems in the design and management of the allocation formula. There are a number of common errors in the design of allocation formulae for equalising UCGs:

    • assigning a weighting that is too low relative to the population sizes of SNGs, and so undervaluing the prime determinant of relative spending needs

    • using the absolute poverty index or similar index values, or other per capita values, without normalising these values by SNGs’ relative population sizes9

    • using variables that may to some extent duplicate each other (e.g. SNG land area and SNG distance from the capital, as proxies for costs of infrastructure and service delivery) and thereby overweighting the importance of those “need” factors relative to others, particularly population size

    • when inserting a “fixed element” in the formula to ensure there is a floor guaranteed to all SNGs, setting it at too high a level, thus limiting variance in total allocations.10

The upshot of such flaws is the introduction of a bias against more populated SNGs. There may well be reasons why less populous SNGs need relatively higher per capita allocations (if this correlates with remoteness, or lower population density, which in turn may increase relative spending need), but this should be addressed by using explicit criteria to capture such factors, and not by accident.

  1. 3. Inappropriate management of the formula. At a more practical level, formulae can be mismanaged or even wilfully manipulated by officials. Thus, for example, in Myanmar the allocation formula has only been applied to increments to the annual transfer pool – the aim being not to overly disturb historic relative allocations to states and regions. In Bangladesh, the results from the earlier formula used by the Planning Commission to make allocations to Union Parishads were found to have been substantially manipulated by local members of parliament in some areas.11

  2. 4. A special case: “Gap-filling” transfers. In a number of countries, in particular transition countries, the predominant fiscal transfer mechanism is through the “gap-filling” transfers previously mentioned. This form of transfer carries several perverse incentives, which will be examined further below. The focus on bilateral negotiations by a ministry of finance with individual SNGs that characterises this sort of mechanism also makes it very hard for central government to balance resource allocations across all SNGs in an equitable manner. To illustrate the outcomes of these processes: in 2017, in Mongolia, the ratios between the highest and lowest aimag per capita deficit grant allocations were 6:1 overall, but 14:1 for health and education transfers combined. In 2019, in Lao PDR, the range between the highest and the lowest levels of total provincial transfers per capita was 11:1.

As already noted, in many countries, multiple fiscal transfer instruments are established, often with different rationales and often by different parts of central government. Policy co-ordination is, therefore, difficult – even securing budget data and monitoring the overall resource flow patterns can often prove challenging.

For example, several countries have established revenue-sharing arrangements, particularly to share license, tax or royalty revenues deriving from natural resource extraction. Since these natural resources are distributed unevenly, this inevitably results in considerable geographic disparities in SNG revenues. Figure 5.4, for example, illustrates the variance in per capita mining-related revenues in Mongolia for 2019 (Panel A) and Indonesia for 2010 (Panel B). In Mongolia, at the extremes, Dornogovi aimag receives USD 13 per capita, while Orkhon receives nothing; earlier 2016 budget data indicate that the range in per capita revenues shared with the soums, the lowest SNG level, is much greater, at nearly 2000 : 1. In Indonesia, the Kaltim province received USD 65 per capita, and Papua and Kalsel around USD 30 per capita, while 17 provinces received virtually nothing.

Whether such wide disparities are justified depends partly on the policy objectives of these transfers. However, while policy reference is often made to such transfers being aimed at compensating for social and environmental externalities faced by the provinces due to extraction activities, there is little clarity about the types or levels of spending expected by SNGs to address these.12 In reality, in both countries, it is likely that the prime aim of these transfers is political – to satisfy demands of local communities and leaders who feel excluded from the benefits arising from these extractive activities.13

Whatever the policy rationale for sharing revenues, equity concerns dictate that there must be limits to the disparities in per capita funding that result from revenue sharing. The challenge presented by the volume of such RS transfers is that it then proves very hard to compensate disparities, through allocation of other transfers intended to equalise – even with allowing SNGs in natural resource extraction areas some higher level of per capita revenues to address special spending needs. There are two main reasons for this difficulty:

  • In many cases, as noted, the equalising transfer allocation formula is not designed to take into account different levels of shared revenues received by SNGs, but – at best – only own-revenue capacities.

  • However, even if these were taken into account, very often, equalising transfer resources may be merely inadequate to offset the disparities caused by revenue-sharing transfers (Figure 5.5).

As a result, in both Indonesia and Mongolia, the net result of all transfers combined still displays a substantial degree of inequity across subnational governments. To illustrate, Table 5.1 shows the data relating to per capita aimag transfers in Mongolia. Despite the moderating effect of the General Local Development Fund (GLDF) (the “equalising transfer”), which significantly reduces disparities generated by shared revenues, the net result is still a substantial 12:1 range from highest to lowest per capita value. If these are viewed at the lowest SNG level – the soum – the total transfer disparity is much greater.

Aggregate transfer allocation patterns display similar disparities in Indonesia, with the equalising unconditional grant transfer – the DAU (Dana Alokasi Umum) – merely being too small to offset the large disparities noted above.

Equity issues can also arise with the allocation of sector-specific or other earmarked conditional grants. In some countries, these grants are allocated by criteria related to the size of the existing infrastructure stock in the sector – e.g. in Viet Nam, health sector conditional grants are allocated to provinces based on several factors, including the number of hospital beds, introducing a bias in favour of better-endowed areas. Similarly, where conditional grants are allocated based on performance in the achievement of service delivery outcomes, this can also generate inequity in funding. For example, in India over the 13th Central Finance Commission period (2010-15), health transfers were linked to local infant mortality rate (IMR) outcome improvements. This led to massive disparities in funding (some states receiving INR 200 per capita, others less than INR 1) because there was not adequate recognition of the very different degree to which improvements in IMRs could be made in states at different points on the IMR spectrum – and so states starting at better IMR levels, which were relatively harder to improve, were penalised (Centre for Global Development, 2015[25]).

Several conclusions emerge concerning the equity outcomes of fiscal transfer systems. First, the lack of clear policy objectives of some transfers can make it hard to assess the overall equity of outcomes. Second, there is often little or no comprehensive oversight of the equity consequences of the entire set of all fiscal transfer flows combined. Third, “equalisation” grants are often inadequately resourced to compensate for other disparities, and/or they are inappropriately allocated.

These equity problems are very often hidden, simply because of the lack of readily available, comprehensive data on the whole set of centre-local fiscal transfers. It is important to establish a comprehensive mechanism to consolidate data on, and monitor, the various fiscal transfer flows that may be managed by different central government departments and agencies that do not necessarily co-ordinate with each other, particularly in countries without a national local government ministry or finance commission with a mandate to ensure such monitoring. A comprehensive mechanism to monitor different transfer streams – unconditional grants (usually intended to equalise), conditional grants, and revenue-sharing transfers (not intended to equalise), and own-revenues of SNGs – would allow for the monitoring of the combined equity effects of subnational public financing arrangements, and hence the implications for achieving SDG 10.

Even where such equity problems are identified, however, it can be hard to undertake reforms. In any reform to the fiscal transfer system toward greater equity, some SNGs will benefit, but others will lose from a change to the status quo. This will be a greater or lesser problem depending on the political power of subnational government leaders.14

In principle, the simplest way to address this is for the central government to increase the total volume of national transfer pools to a level where no SNGs will lose in absolute terms with a change in allocation criteria (the “hold harmless” approach). This was the case in Nepal, where a move from equal allocations to formula-based village grants meant that the less-populated villages would lose out. This problem was avoided by a massive increase in the national village grant pool. But not many countries have the resources or the political drive for devolution to make the large budgetary increases to the national grant pools needed for such reforms.

Elsewhere, the problem has been addressed by phasing in changes over time, to allow SNGs the time to adjust to the increases or decreases in annual budget resources brought by such reforms. This has been considered to ease proposed transfer changes in Indonesia where reforms to bring in greater horizontal equity have been stalled for many years.

There is a longstanding concern in public finance and public choice literature that fiscal transfers may weaken accountability mechanisms for local spending and/or discourage local revenue-raising effort (Pöschl and Weingast, 2013[26]). However, the evidence for these effects is somewhat mixed. With regard to the impact on local fiscal effort, there is indeed contrary evidence that unconditional grant transfers may actually encourage local revenue effort (Brun and El Khdari, 2016[27]; Lewis and Smoke, 2017[28]; Troland, 2014[29]).15

Less frequently discussed are the incentive effects that transfers may exert on the local budgeting process and determination of spending priorities. As already underlined, sustainable development policies and plans to achieve SDGs have minimal traction at the local level unless these priorities are reflected in local budgets and spending. It is the quality of the capital and recurrent budgeting procedures, and the explicit or default priority-setting arrangements within them, which determine the quality of the expenditure priorities that are chosen, and which in turn determines the quality of the public services delivered. Fiscal transfers may shape these budget processes – for better, but often for worse – in several ways.

This is the major policy feature of any transfer instrument: whether the transfer allows subnational governments wide discretion, as is the case with most revenue-sharing transfers and unconditional grants, or whether it may only be spent on centrally determined priority areas of expenditure, as with conditional grants. The rationale for such conditionality lies in the view that SNGs may otherwise be inclined to underspend on the areas in question. Where transfers are earmarked, there is, therefore, an overwhelming incentive16 for SNGs to spend those funds in the designated manner, or else face sanctions for failure to comply.

It is not easy to make general a priori judgements about the merits of imposing conditions or allowing discretion; the right balance will be very context-specific. On the one hand, if transfer-use conditions reflect a central expenditure blueprint which does not recognise variations in a local context or which precludes desirable local flexibility, then this may undermine the achievement of the SDGs. On the other hand, if, without such restrictions, SNGs are tempted to undervalue certain national priorities, or undermine service standards, then earmarking is positive.

While some conditionality will be necessary for some types of transfers, a transfer system dominated by conditional grants may have two undesirable effects for the sustainable development agenda.

  • Rigidities in budget options. First, the local sustainable development agenda requires substantial flexibility of spending between and within sectors, to allow tailoring of overall policies and plans to specific local contexts. An excessive degree of earmarking in the overall transfer system can limit – or even prevent altogether – the sort of discretionary choices that UCGs are able to make, and the budgetary flexibility they enjoy.

An extreme case of this rigidity is seen in the problems faced by SNGs in India until the 14th Central Finance Commission (CFC) reforms (see Table 5.2). This illustrates how Indian SNGs (even in states with the most advanced devolution policies) have been faced with transfer flows dominated by an array of highly conditional or tied grants.

The benefits of decentralisation are built on better local knowledge about local needs and realities, and the ability to adapt spending priorities accordingly. Rigidity in fund use through excessive earmarking will undermine the flexibility, and hence the effectiveness and efficiency of local spending. One example of this is seen in countries such as Bhutan, Lao PDR and Viet Nam where public investments must be consistent with those outlined in the five-year plan, perhaps prepared several years previously – precluding local ability to respond to unforeseen emergencies or opportunities that arise in the short term. Another example is seen in Indonesia, in the DAK (Dana Alokasi Khusus) grants. These are earmarked for investment spending, especially for capital budget expenditures. There is some evidence that SNGs are encouraged to invest in new facilities, even where rehabilitation of existing facilities is needed more and represents a more efficient use of funds, simply because the latter may not always qualify as “capital” expenditure (Lewis, 2013[31]; Lewis and Smoke, 2017[28]). Similarly, in Mongolia, the emphasis that Local Development Fund grants be used only for capital budget expenditure has encouraged subnational governments to make infrastructure investments they cannot then sustain in the long run since they are not able to make the corresponding recurrent budgetary allocations for the operation and maintenance of existing facilities.

  • Too many local planning procedures. A plethora of earmarked conditional grants encourages, or even obliges, SNGs and communities to engage in separate planning exercises for each of the funds, through various sector-specific community group and committee arrangements. This undermines the overall integrity of SNG planning and budgeting, and thus the general effectiveness and efficiency of spending; it may also lead to “participation fatigue”.

This is a problem that was long faced by SNGs in India, which have often had to organise parallel community-planning arrangements for the large number of centrally sponsored conditional grants in sectors such as water and sanitation, education and health (Government of India and Tata Institute, 2015[30]).

Conversely, excessive discretion can also be problematic. Subnational governments may face local pressures to make budget priorities that do not always fully match local developmental needs, or bow to pressures to spend more than optimally on local staff and administration, particularly where local planning and budgeting capacities and accountability mechanisms are weak.

For this reason, limitations are frequently placed on the use of unconditional grants (as in Cambodia, Mongolia and Kerala state in India) to ensure that a minimum part will be spent on development rather than administration. By contrast, in Indonesia, where there has been no such limitation on the use of DAU grants, there is some evidence of excessive spending on administration (Lewis and Smoke, 2017[28]).

Similarly, where there is no clear earmarking by expenditure component for conditional grant transfers for key sectors such as education and health, there can be a risk of excessive variance in spending patterns, which may undermine national service quality standards if there is insufficient guidance to SNGs. The Indian experience following the recent relaxation of grant conditionalities is instructive. The extreme interstate variance in component expenditure for secondary education conditional grants seen in Figure 5.6 raises questions, although this variance can perhaps partly be justified by different state contexts.

Here the key issue lies in how the national pool for each transfer instrument is determined each year in the national budget process. Where the national pool is determined on an ad hoc basis, then the size of the pool may vary considerably year by year. In consequence, allocations from the pool to individual subnational governments will also vary and will be hard to predict. Insofar as such transfers are a major revenue source for SNGs, this will make the annual budgeting exercise very difficult and may undermine efforts to make any serious budget priorities. Furthermore, it will also make medium-term planning very hard for SNGs and undermine efforts for more strategic multi-year spending plans at the subnational level.

This volatility will be greater still where the national pool is determined as a percentage of only one or two revenue streams, rather than from the entire national revenue base, which is inherently more stable. The volatility of the total allocable pool is particularly large if they are tied to revenues from economic activities subject to world market fluctuations, such as the mining of minerals, oil and gas. Where this annual volatility is extreme, the incentive effects can be especially undesirable. In “boom” years, high revenues can encourage wasteful showcase spending simply to use the funds; while in “bust” years, the SNGs may have to renege on commitments and cut back on essential services. All of this undermines efficiency and effectiveness. In Mongolia in boom years, the aimags and soums have undertaken generous social spending, but then in later bust years, this spending had to be drastically cut back, causing serious social problems (Bauer et al., 2016[9]).

Although rather neglected in the literature, at what exact point in their budget cycle SNGs are informed of the amount of the fiscal transfer they will receive in the upcoming fiscal year matters a great deal. Where SNGs are not informed until after they have finalised their budget and determined their budget priorities, this can lead to serious problems affecting the local budget prioritisation process.

Lack of advance notice of transfers undermines SNG incentives to review spending options seriously, from what is usually a long list of budget proposals coming from different sources, and to determine the optimal, affordable priorities in light of the total known budget resource ceiling. Therefore, when the centrally approved transfer is finally announced, SNGs have to scramble then to select the spending priorities that are affordable, with little time for proper review, consultation and comparative assessment of different spending options. This is an especially serious problem in countries operating “gap-filling” transfer arrangements, and where SNG budgets are “nested” within the overall state budget (see Box 5.1).

There is a further undesirable knock-on incentive effect in such countries. Where SNGs are not informed of the budget ceiling in advance, it is impossible for SNG authorities, in turn, to transmit budget ceilings either to SNG departments or to authorities at lower levels. This means that in the annual bottom-up planning and budgeting process, an excessively large volume of proposals is generated because originating departments or lower levels of administrations have no incentive18 to screen and eliminate options from the long lists, and to identify priority proposals before they are submitted to SNG authorities. In consequence, SNG authorities receive an overwhelmingly long, unfiltered set of proposals that have to be appraised and prioritised to determine what to include or exclude from the budget, making it an even harder exercise given the concise time allotted.

In the specific case of countries where gap-filling transfers dominate the transfer arrangements, SNGs have clear incentive to generate inflated sets of expenditure proposals with little real prioritisation, and underestimated revenue projections, in order to present the greatest deficit possible and thereby game the system to attract a greater transfer.

The arrangements for actual release of funds down to SNGs and the associated treasury and reporting requirements for SNGs to access these funds can have consequences for actual spending priorities.

In some cases, the fund flow route is so slow that SNGs only receive transfer funds very late in the fiscal year. Two examples from India illustrate the issue. An extreme case is the Backwards Region Grant Fund, from which grants were reported as sometimes arriving right at the end of the fiscal year or even well into the following fiscal year (i.e. one to two years late) (World Bank, Sida and Government of India, 2010[33]). Less extreme, but still serious, are delays registered in the allocation of health and education conditional grants through the treasury system. Reports on education conditional grants19 suggest that only some 50% of grants arrived in the first six months of the fiscal year and that up to 10% had not even arrived by the end of the year (Accountability Initiative and Centre for Policy Research, 2018[32]);

In such cases, when funds finally arrive in the SNG account, local officials will be tempted to reorder the original budget priorities so that funds can hurriedly be spent in time – resulting in suboptimal expenditures. Generally speaking, where subnational governments have little faith that such funds will arrive on time, the incentive to make a serious effort to plan and budget in the first place, and prepare considered implementation and procurement plans, is greatly diminished.

In countries that operate single treasury systems,20 two sorts of problems may arise: first, transferred funds can be stuck in local treasury offices, where problems in payment order documentation provided by SNGs may hold up their release. These delays are reported as common in Cambodia where commune and district officials must travel to the (often distant) Provincial Treasury Office in the hope of finding an official present that day to approve the release of funds. This not only wastes time, delaying budget execution but also encourages rent-seeking by the officials involved.

That aside, where central guidance is unclear as to what legitimate SNG expenditure responsibilities are, local treasury officials may be reluctant to approve even legitimate spending requests. Again, in Cambodia spending on construction from the unconditional grant transfer (Commune/Sangkat Fund) is well understood as legitimate by treasury officials, while recurrent spending on special services from this transfer is reportedly often questioned and approval denied, despite such spending on services being encouraged by the central government. These known treasury problems may perhaps be one of the factors underlying a widespread concern that Commune budgets are overly biased toward spending on construction works.

The reason why central governments often do not allow carry-over of unspent funds is to pressure SNGs into efficient execution of national budgetary resources. SNGs are often faced with serious budget execution constraints, however, through no fault of their own. This is particularly the case for the capital budget in rural, more remote SNGs where:

  • the funds regularly arrive late in the fiscal year, as illustrated above, leaving little time to spend before year-end

  • there are seasonal weather constraints, such as the long, intense monsoon season in Myanmar, and the long hard winter in Mongolia, which allow only a few months each year to undertake investment activities, especially in rural areas

  • there are problems and inevitable delays in securing supplies, contractors and technical support to implement development activities.

Unable to carry over unspent funds, SNGs may be encouraged to select the sort of spending priorities for inclusion in the budget that minimises the risk of underspending by year-end: investments in more accessible urban areas, rather than more inaccessible rural areas; a few, more manageable, large investments rather than many, smaller, dispersed investments; investments in facilities for which standard design blueprints are available rather than those which require site-specific design work. The resultant investment pattern may thus not always match local needs and sustainable development agenda priorities.

Similarly, especially where no carry-over is allowed and when funds only arrive late in the fiscal year, SNGs may be compelled to make rushed spending decisions to use the money in time, but may not always reflect the original budget priorities. That aside, there may also be little incentive for efficient spending if any savings are not retained by subnational governments but returned to the central government.

However, in allowing carry-over of unspent funds, caution is also required to avoid misuse. For example, provinces in Lao PDR are allowed to carry over unspent funds into the next year, but can then spend them in a manner that does not allow any real oversight, and which may all too easily undermine the effectiveness and transparency of spending.

Some of the pitfalls noted above stem from problems in the design of particular transfer instruments, but others derive from more systemic problems in the underlying national public financial management (PFM) arrangements.

It seems clear that any move towards a “rules-based” determination of national transfer pools, linked to specified shares of total national revenues, is positive in ensuring year-to-year predictability of SNG allocations and hence encouraging more realistic annual budgeting and priority setting. The stability and predictability are greater if these arrangements are formally embedded in law. Such rules, of course, are not always popular since they undercut the discretion of national policy makers. Such commitments also require central government confidence in its medium-term fiscal projections, based on the adoption of a reliable medium-term fiscal framework.

If stability and predictability of fiscal transfers is a preferred objective, it is also best to avoid creating oversized revenue sharing or other transfer mechanisms linked to a few individual revenues, which are inherently more variable from year-to-year than overall government revenues – particularly when these revenues are linked to potentially volatile natural resource commodities markets. Indeed the unpredictability at SNG level is further amplified through the allocation of such transfers on the derivation-sharing principle since the difference between estimated and actual revenues and the year-to-year variations in revenues derived from single SNG jurisdictions will likely be much greater than aggregate national variations.

There are, of course, always important reasons for some spending conditionalities in the transfer system. Indeed, where central governments are moving to devolve additional spending responsibilities to SNGs, it may often be prudent to do so initially through conditional transfers to help allay the understandable concerns of sector ministries. However, where possible, more – rather than less – local discretion is preferable, to ensure that the advantages of decentralising SDG-related mandates to SNGs are not undermined by overly “straitjacketing” local budget priority setting. Where it is necessary to introduce spending conditions, these should be kept as few and as simple as possible, with practical guidance to SNGs on the “eligible menus”. There should be a strategy to gradually relax these over time, as both SNG capacities and central monitoring capacities are expanded (Spahn, 2012[34]).

As far as possible, transfers should be designed and managed such that SNGs are given notice of their (likely) transfer amounts sufficiently in advance before SNGs finalise their own budget proposals. This advance “hard budget constraint” will allow subnational governments to review and appraise spending options and determine priorities in the knowledge of their total revenue ceiling, and thereby prepare a comprehensive spending plan based on that ceiling before they finalise their budget proposals. Lack of such advance notice of the forthcoming transfer amount is an incentive for wish-list budgeting without any priority-setting – always the easiest default approach, since selecting some proposals as priorities also means rejecting others, and alienating the local stakeholders concerned. Further, without some budget certainty, it is also much harder to engage stakeholders or communities in discussions around priorities year after year. There is also evidence to suggest that advance knowledge of resources to be transferred is an encouragement to local efforts to mobilise other sources of financing, notably from local revenue collection.

While there may often be perverse incentives embedded within transfers, fiscal transfers can also be expressly designed to transmit positive incentives to promote better SNG performance. Historically, the focus of such attempts has been to encourage local revenue-raising efforts. However, recent years have seen the emergence of performance-based grants (PBGs) with a deliberate focus on improved local governance and public financial management performance, many of which were initially piloted through donor-supported programmes in Asia and Africa.21 These PBGs are usually linked to existing grants (unconditional grants or conditional grants) and carry explicit incentives to encourage better subnational government performance in service delivery and governance. Early lessons show that they offer promising avenues to encourage better quality spending and service delivery for the local sustainable development agenda, with some cautions and caveats.

The key characteristics of PBGs are that they are given as a reward “top up” existing grant transfers based on the results of annual performance assessments of SNG performance. The assessment scores are then used to reward or sanction SNGs (by transferring more or less to them) depending on their performance.

Broadly, PBGs can be categorised into multi-sectoral PBGs and sectoral/thematic PBGs:

  1. 1. Multi-sectoral PBGs. These are PBGs linked to multi-sectoral UCGs. Here the performance criteria are generally “process” indicators related to governance, planning, budgeting, public financial management and transparency (Steffensen, 2010[35]).22 The common features of this mechanism as implemented in several countries in Asia and Africa are detailed in Box 5.2.

  2. 2. Sector-specific PBGs. These are PBGs linked to sectoral or thematic conditional grants. Here it is more feasible to link performance criteria to higher-order service delivery “output” indicators, although “process” indicators are also used. To illustrate:

    • Health sector PBGs have been introduced and implemented in a number of countries, such as Argentina, Brazil, India, Pakistan, Tanzania and Uganda (Musgrove, 2011[36]; UNICEF, 2013[37]; Fritsche, Soeters and Meessen, 2014[38]; Forgia and Baeza, 2015[39]). Funds are transferred to subnational governments and then further to health service units, based on measures of both general process performance and of health service outputs delivered.23

    • Ecological Fiscal Transfers have been implemented in Brazil, France and Portugal to reward subnational government performance in environmental protection, with performance measured against the size and quality of conservation measures by subnational governments (Cassola, 2010[40]; Borie et al., 2014[41]; Droste et al., 2017[42]; Loft, Gebara and Wong, 2016[43])

The impact of some sector-specific performance-based grants such as the health grants under Plan Nacer in Argentina has been well documented (World Bank, 2015[44]).

However, for multi-sectoral PBG mechanisms linked to unconditional grants, the evidence mainly relates to general trend improvements in the annual performance scores of subnational governments over the years. This performance is generally related to the degree of “process compliance” around planning and PFM processes and governance arrangements, and itself is not evidence of the improved quality of public spending or service delivery outputs and outcomes. There is, of course, a reasonable presumption that greater compliance will lead to these higher-order results, but hard evidence so far is scarce. This appears to be an area in which in-depth research and case studies would be useful.

While it is tempting to try to link performance to service output or even outcome performance, this is very challenging and often not feasible. For multi-sectoral PBGs linked to UCGs, this would involve comparing often very different service delivery mixes across SNGs (such variations being indeed a key rationale for decentralised decision making).

For some sectoral PBGs, notably in the health sector, using “output” or “outcome” indicators of performance is possible, but requires much preliminary groundwork and can still be challenging. Measuring the quantity and quality of service outputs requires a much greater, more costly and time-consuming fieldwork effort – while comparing outcomes across SNGs faces the challenge that the starting point for such outcomes will vary considerably across SNGs. PBGs linked to output or outcome performance, therefore also require a considerable baseline study effort to calibrate rewards accordingly. All that aside, many other extraneous factors come into play, outside the control of SNGs, such that the results may be more easily contested and seen as less legitimate than measures based solely on compliance.

However, even for “process” compliance, the essentials need to be in place. Laws and regulations against which “process” performance is assessed should be appropriate, clear and consistent, and subnational governments should be able to comply with these processes on their own initiative, independent of human or other resources outside SNG control – such as those provided by central governments or donors. Preliminary work needs to be done to determine reasonable standards and assess capacities.

Indicators to measure performance should be relevant, objective, verifiable, and few in number. They should also be based on reasonable rather than “ideal” standards and relate to relatively recent subnational government activities (in the past two years).

The relative size of the “top-up” performance-based grant reward is important in transmitting incentives. It is doubtful that those (mainly urban) SNGs that have substantial own-source revenues, and for which fiscal transfers account only for a small part of overall revenues, will be encouraged to improve performance with a PBG mechanism. For other SNGs, if the amount of the PBGs is too small, it will not provide an incentive to improve performance (the usual rule of thumb is to calibrate PBGs to about 15-20% of the “parent” unconditional grant or conditional grant fiscal transfer). Similarly, if too many or too few SNGs are rewarded with PBGs, then the PBGs will also lose the ability to incentivise governments (here, the rule of thumb is to reward about 30-70% of SNGs). Lastly, the formula for PBG allocation should be simple so that subnational governments can see the link between performance and reward.24

Given what is at stake in terms of resources gained or denied, performance assessments should be seen as objective and impartial by SNGs. There is, therefore, a strong argument for these to be undertaken independently. Most of the PBG mechanisms outlined above, supported by donors, out-source the assessments to private contractors or research agencies that send teams into the field, process the results and report to the Ministry of Finance. Out-sourcing provides independence and also allows the necessary personnel to be fielded to visit large numbers of SNGs in the relatively short time window required by the budget calendar imperatives, which would not be possible for a central government ministry.

However, these outsourcing arrangements are not without problems. Procurement and management of the contractors throughout the process itself take effort, as does the quality controls over the results of the different teams fielded. Keeping to the timeline to ensure that results are forthcoming in time to meet Ministry of Finance deadlines for budget submission for the following year is also often difficult. These all prove to be real challenges to sustaining performance-based grants, even in the presence of donor technical support; they are thus even more worrisome at the point they need to be fully taken over by the government.

The challenges outlined above around management of assessment processes constitute a problem for the long-term sustainability of PBG mechanisms. These challenges stem both from the technical management of the assessment process but also from the cost, if this cost is borne by the government. In principle, of course, these costs are perhaps far less than the benefits of the PBG mechanism in terms of improved PFM and service delivery and so should be justifiable. However, to date, there is little hard evidence, of the sort that might convince cost-conscious ministries of finance, to support this contention.

One avenue to ensure the long-term sustainability of such assessments could be through the broadening of the scope of the external audits of subnational governments. However, in many countries in Asia, there is as yet no effective external audit of SNGs. More generally, the full costs of implementing performance-based grant mechanisms, including assessment and monitoring costs, should be acknowledged and factored in, along with the fiscal resources needed for the actual PBG transfers.

All that said, there may be an argument that such performance-based mechanisms may have value as part of a temporary strategy to strengthen SNG performance, and are not needed for perpetuity once SNG performance has reached generally acceptable levels. At that point, any shortfalls in performance may be addressed through either ad hoc remedial support or legal sanctions.

Ultimately, the aim of these PBG mechanisms is to create incentives for SNG personnel to perform better. It is therefore essential that information about performance criteria and results be made public, for transparency reasons, to dispel suspicions of favouritism or influence in allocation of the PBGs, and also so that pressure can be put on poorly performing subnational governments by local citizens. Central governments have so far proven reluctant, however, to publish such information on the results, for fear of provoking controversy.

Finally, these mechanisms require that politicians and central government policy makers back up the results and be willing to resist the inevitable pressures from SNGs that failed to secure PBGs, or that feel that the results were unfair.


[32] Accountability Initiative and Centre for Policy Research (2018), “Rashtriya Madhyamik Shiksha Abhiyan (RMSA), Government of India 2017-18”, Budget Briefs, Vol. 10/2.

[17] ADB Institute (2016), Government Decentralization Program in Indonesia,

[24] Agustina, C. et al. (2012), Political Economy of Natural Resource Revenue Sharing in Indonesia, ARC, London,

[7] Bahl, R. (2000), Intergovernmental Transfers in Developing and Transition Countries: Principles and Practice, World Bank, Washington, DC.

[9] Bauer, A. et al. (2016), Natural Resource Revenue Sharing, NRGI and UNDP,

[4] Boadway, R. and A. Shah (2001), Intergovernmental Fiscal Transfers, Principles and Practice, World Bank, Washington, DC.

[41] Borie, M. et al. (2014), “Exploring the Contribution of Fiscal Transfers to Protected Area Policy”, Ecology and Society, Vol. 19/1,

[27] Brun, J. and M. El Khdari (2016), “The Incentive Effects of Conditional and Unconditional Transfers on Local Own Revenue Generation: Empirical Evidence from Moroccan Municipalities”, Études et Documents 6,

[40] Cassola, R. (2010), TEEB Case: Financing Conservation Through Ecological Fiscal Transfers in Brazil,

[25] Centre for Global Development (2015), Power to the States Making Fiscal Transfers Work for Better Health,

[8] Dabla-Norris, E. and P. Wade (2006), “The Challenge of Fiscal Decentralization in Transition Countries”, Comparative Economic Studies, Vol. 48/1995, pp. 100–131.

[42] Droste, N. et al. (2017), “Ecological Fiscal Transfers in Brazil – incentivizing or compensating conservation?”, Environmental Policy and Governance, Vol. 27/4.

[5] Ebel, R., C. Wallich and R. Bird (1995), Decentralisation of the Socialist State: From Command to Market, World Bank, Washington, DC.

[39] Forgia, J. and C. Baeza (2015), Performance-Based Intergovernmental Transfers: Brazil’s Family Health Program.

[38] Fritsche, G., R. Soeters and B. Meessen (2014), Performance-Based Financing Toolkit, World Bank, Washington, DC.

[14] Government of India (2015), Report of the Fourteenth Finance Commission,

[30] Government of India and Tata Institute (2015), Devolution Report 2015-16,

[15] Government of the Philippines (1991), The Local Government Code of the Philippines,

[31] Lewis, B. (2013), “Local Government Capital Spending in Indonesia: Impact of Intergovernmental Fiscal Transfers”, Public Budgeting and Finance, Vol. 33/1, pp. 76–94.

[28] Lewis, B. and P. Smoke (2017), “Intergovernmental Fiscal Transfers and Local Incentives and Responses: The Case of Indonesia”, Fiscal Studies, Vol. 38/1, pp. 111–39.

[43] Loft, L., M. Gebara and G. Wong (2016), The Experience of Ecological Fiscal Transfers: Lessons for REDD+ Benefit Sharing,

[11] Man, J. and Y. Hong (2011), China’s Local Public Finance in Transition, Lincoln Institute of Land Policy.

[21] Martinez-Vazquez, J. (2011), Fiscal Decentralization in Asia: Challenges and Opportunities, Asian Development Bank, Manila.

[22] Martinez-Vazquez, J. and J. Boex (2006), The Design of Equalization Grants: Theory and Applications, World Bank and Georgia State University.

[6] Martinez-Vazquez, J. and J. Boex (1999), “Fiscal Decentralization in the Russian Federation during the Transition”, International Studies Program Working Paper, 99–3 Georgia State University.

[36] Musgrove, P. (2011), Financial and Other Rewards for Good Performance or Results : A Guided Tour of Concepts and Terms and a Short Glossary, World Bank, Washington, DC, for Good Performance or Results - Short Glossary.pdf.

[2] OECD and UCLG (2016), Subnational Governments Around the World: Structure and Finance, OECD Publishing, Paris and United Cities and Local Governments, Barcelona,

[26] Pöschl, C. and B. Weingast (2013), The Fiscal Interest Approach: The Design of Tax and Transfer Systems,

[18] Shotton, R., Z. Yee and K. Oo (2017), State/Region Financing, Planning & Budgeting in Myanmar, Asia Foundation,

[19] Smoke, P. and Y. Kim (2003), Transfers in Asia: Current Practice and Challenges for the Future,

[34] Spahn, P. (2012), Conditioning Intergovernmental Transfers and Modes of Interagency Cooperation for Greater Effectiveness of Multilevel Government in OECD Countries,

[35] Steffensen, J. (2010), Performance-Based Grant Systems: Concept and International Experience, UNCDF.

[29] Troland, E. (2014), “Can Fiscal Transfers Increase Local Revenue Collection? Evidence From The Philippines”,

[3] UCLG (2010), Financing Local Government: The Challenges of the 21st Century. Second Global Report on Decentralization and Local Governance, United Cities and Local Governments, Barcelona.

[23] UNDP (2019), Fiscal Transfers in Asia: Challenges and opportunities for financing sustainable development at the local level, United Nations Development Program, Bangkok,

[37] UNICEF (2013), “Performance-Based Equitable Resource Allocation Model: The Punjab Experience”, Maternal, Newborn and Child Health Working Paper, UNICEF,

[1] United Nations (n.d.), Sustainable Development Goals, (accessed on 11 February 2020).

[12] Wang, X. and R. Herd (2013), “The System of Revenue Sharing and Fiscal Transfers in China”, OECD Economics Department Working Papers, No. 1030, OECD Publishing, Paris,

[20] White, R. and P. Smoke (eds.) (2005), East Asia Decentralizes: Making Local Government Work, World Bank, Washington, DC.

[10] Wong, C. (2007), “Budget Reform in China”, OECD Journal on Budgeting, Vol. 7/1,

[16] Wong, C. (2000), “Central-local relations revisited the 1994 tax-sharing reform and public expenditure management in China”, China Perspectives, pp. 52-63.

[44] World Bank (2015), Argentina : Can Performance Payments Improve Newborn Health, World Bank, Washington, DC.

[13] World Bank (2015), Making the Whole Greater Than The Sum Of The Parts: A Review Od Fiscal Decentralization in Vietnam, World Bank, Washington, DC.

[33] World Bank, Sida and Government of India (2010), First Independent Review of the Backward Regions Grant Fund - BRGF, World Bank, New Delhi.


← 1. Fiscal transfers are made by central governments to legally constituted subnational governments to whom responsibilities are devolved or delegated. Fiscal transfers are different from the flows of resources from central government ministries to deconcentrated local branches of these ministries, though fiscal transfers usually coexist with deconcentrated flows. This chapter focuses on fiscal transfers; however, some of its recommendations are also relevant for deconcentrated flows of finance to subnational administrations.

← 2. See also country subnational finance profiles at

← 3. In countries that have had a history of socialist economic management (e.g. China, Viet Nam, Lao PDR, Mongolia, and, until recently, Myanmar), significant similarities can be seen in the intergovernmental fiscal arrangements they have inherited. SNGs are frequently classified according to whether they are “deficit” or “surplus” SNGs. “Deficit SNGs” are those where the sum of own-source revenues and transfers is not enough to cover their expenditures, while in “surplus SNGs”, they are usually sufficient. “Gap filling” or “deficit” grant transfers are then provided on some negotiated basis to partly fill the gap, through bilateral reviews and negotiations with the Ministry of Finance. SNG recipients of such transfers are then subject to tighter controls and enjoy less discretion than surplus SNGs. There is ample documentation that this mechanism is replete with many negative incentives that undermine local budget priority setting, is very inequitable, and promotes non-transparent patronage relationships between different levels of government. It is also common that the budgets of each level are nested, in matryoshka fashion – which leads to the need for a much more lengthy and iterative budget preparation process for subnational governments, and may also introduce greater uncertainty into their subnational revenue estimates for expected fiscal transfers, with the attendant negative incentives.

← 4. Although it is conventional to consider revenue sharing as a form of fiscal transfer from central to subnational governments, in a number of socialist/transition countries in Asia (e.g. China, Viet Nam, Lao PDR, Mongolia) there has been a tradition whereby revenues are shared “upwards” by subnational governments – sometimes through a complex array of sharing agreements varying between SNGs.

← 5. Following the recommendations of the 14th Indian Central Finance Commission. These recommendations were for a five-year period from FY 2015/16 to FY 2020/21. At the same time, the states’ share in national revenues was increased from 32% to 42%.

← 6. There are also often other centre-to-local resource flows such as donor and non-governmental organisation (NGO) project spending and constituency grants, but these are generally much smaller in volume.

← 7. In principle, assuming devolved spending mandates are clearly distinguished from central spending mandates; then the geographic patterns in deconcentrated spending should not affect the equity implications arising from devolved spending patterns.

← 8. For example, in mountain areas of Nepal the cost of cement and other construction materials is two to three times higher than in Kathmandu or the Terai area in neighbouring India.

← 9. This failure to normalise index variables is surprisingly common, and has been seen in the past in both Mongolia and Nepal (where corrections were subsequently made), and presently in Myanmar.

← 10. There appears to be a widespread tendency for central government officials to aim for similar-sized grant allocations to SNGs – regardless of size or need – perhaps on the grounds that this seems fairer or is more defensible than allocations that vary greatly in size, regardless of the per capita rationale for such variations. For example, in the past, both village and district grants in Nepal were of equal size. When an allocation formula was first introduced for district grants, the “fixed element” was expressly set at such a high level that the variance of total district allocations still remained quite limited.

← 11. This occasional manipulation emerged from a case study of Union Parishad grant allocations within selected Upazilas commissioned by the United Nations Capital Development Fund in 2002.

← 12. Discussions with local authorities in Mongolia in receipt of shared revenues suggests that they are used simply for general investment expenditures, not to address problems arising from mining; the expectation is that such mitigation measures are the responsibility of central government.

← 13. The issues around natural resource revenue sharing are discussed in detail in Bauer et al. (2016[9]).

← 14. This is one reason why it has proved impossible so far to reform fiscal transfers in Indonesia. Allegedly, one reason why reform of own-revenue powers and revenue-sharing arrangements in the Lao PDR was so difficult is that provincial governors sit on the Party Central Committee and are immensely powerful. There are probably similar structural-political reasons why such reforms were also so hard in other one-party countries like China and Viet Nam.

← 15. For example, in Morocco, a 10% increase in UCGs was shown to be associated with a 6.9% increase in SNG own-revenue collection; in the Philippines, a 10% increase in IRA grants with a 3.4-3.9% increase in local fiscal effort; and in Indonesia, a 10% in DAU grants was associated with a 1.2% increase in SNG revenues.

← 16. The strength of the incentive to spend accordingly will be dependent on the perceived likelihood that the central government will indeed monitor use of the transfers, and exert sanctions in case of non-compliance. In this regard, there is considerable variation between countries in the degree of supervisory control and sanction.

← 17. In general, the budget timetables in socialist/transition countries where SNG budgets are approved inside national budgets, they all appear to provide little time for SNG authorities to carry out key steps in the process.

← 18. In Myanmar, township administrations (which receive no advance budget ceiling information) typically submit around five to ten times the number of capital investment proposals than are later approved for funding. It must be understood that there is always an in-built reluctance by communities, local authorities or sector departments to make priorities, unless they are compelled to do so. Making priorities means favouring some proposals but also discarding other proposals; this can be very difficult both technically and politically, particularly where consensual norms prevail.

← 19. These are conditional grants provided under the national flagship “Education for All” programme Sarva Shiksha Abhiyan (SSA).

← 20. These are systems operated under the central ministry of finance, and where SNG authorities must request payments to be made to vendors from their subaccount at the local treasury office. Such systems are quite common in South East Asia. They stand in contrast to systems whereby SNGs manage their own bank accounts, which is more common in South Asia.

← 21. . These programmes were supported variously by the United Nations Capital Development Fund and the World Bank.

← 22. Such multi-sectoral PBG initiatives have been introduced in several Asian countries: Bangladesh, Bhutan, India, Mongolia and Nepal. Indeed, the 14th Central Finance Commission in India has recommended the national rollout of a PBG mechanism, whereby 10% of UCGs in rural areas and 20% in urban areas will be allocated on a performance basis, informed by earlier experiences in West Bengal and Kerala.

← 23. “Outcome” measures were used in India for health transfers but were found to be inequitable, and so changed to “output” performance measures.

← 24. The methodology for determining the size of the PBGs to be received by SNGs in some donor-supported programmes is extremely complex – to the point that the link between performance and reward is probably obscured to all but the programme staff who make the computations.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at