Fiscal Decentralization and Public Subnational Financial Management in Peru
  • 1 0000000404811396 Monetary Fund

Contributor Notes

Author(s) E-Mail Address:;

There is increasing interest in fiscal decentralization in Peru as a mechanism to generate more involved decision-making at the subnational level. This is tempered with a continuing emphasis on overall fiscal stability. However, considerable work needs to be undertaken to define more clearly expenditure responsibilities and financing mechanisms that increase local accountability. In addition, a more transparent fiscal transfer system is needed, together with clarity in expenditure management at all levels of government. The paper suggests that a substantial work agenda is needed to extend the decentralization process with greater transparency.


There is increasing interest in fiscal decentralization in Peru as a mechanism to generate more involved decision-making at the subnational level. This is tempered with a continuing emphasis on overall fiscal stability. However, considerable work needs to be undertaken to define more clearly expenditure responsibilities and financing mechanisms that increase local accountability. In addition, a more transparent fiscal transfer system is needed, together with clarity in expenditure management at all levels of government. The paper suggests that a substantial work agenda is needed to extend the decentralization process with greater transparency.

I. Introduction

This paper reviews the experience and prospects for fiscal decentralization in Peru in the context of the emphasis on macroeconomic stability and fiscal prudence that has characterized government policy in the recent past. We review the available evidence and literature on various aspects of the decentralization process, including the legal framework, the sequencing of the devolution of responsibilities and financing, and the limitations posed by subnational public financial management capabilities.

A stated goal of the Peruvian decentralization process, which started in 2002, is that it should be gradual, address geographical disparities, and ensure fiscal sustainability. Reflecting this gradualism, Peru remains relatively centralized; indeed, in 2004, local governments carried out 13 percent of total primary expenditures of general government. Based on Peru’s relatively limited experience with decentralization to date, and the paucity of available information in many crucial respects, this paper focuses on an overview of the legal framework and its implementation to date, including the sequencing of spending and the devolution of responsibilities and revenue assignments. We also raise issues for future work.

Section II examines the historical background to the current decentralization debate. Section III places decentralization in the context of the overall fiscal framework. Section IV summarizes the legal framework underlying the current decentralization efforts. Section V reviews the components of the decentralization process—the evolving spending responsibilities, the design of own-source revenues, and the current focus of the pending decentralization agenda mostly on revenue-sharing, including of natural resource revenues. Section V also examines the fiscal transfer system. The role of the public financial management system on the sequencing of decentralization is discussed. Section VI concludes.

II. Background

A. Current Decentralization Process

The current Peruvian decentralization attempt started in 2002, in a context characterized by sizable geographical income disparities. Reflecting disparities in revenue bases, income (expressed as gross domestic product per capita) is distributed unevenly across departments. In turn, income levels have an impact on the departments’ capabilities to respond to differential spending needs of the population. Indeed, there is a clear negative correlation between poverty and income levels across departments (Figure 1). Further, within departments, poverty rates in rural areas are significantly higher than those in urban areas; also, the poverty rates in the selva and sierra regions are higher than in the coast (Table 1).

Figure 1.
Figure 1.

By departments: product per capita and poverty 2002

Citation: IMF Working Papers 2006, 120; 10.5089/9781451863802.001.A001

Source: INEI
Table 1.

By Area: Poverty Rate “Extreme Poverty,” 2002

article image
Source: INEI

Regional disparities are a challenge to the efficiency of any decentralization process. While recognizing that the equalization of income per capita may not be feasible or desirable, a decentralization framework should ensure and facilitate a minimum, if not a reasonably similar, provision of access to public services to all sections of the population (Ahmad and Searle, 2005). Otherwise, the decentralization may result in growing imbalances in both incomes and public services.

B. Previous Approaches to Decentralization

Before 2002, Peruvian administrations oscillated between centralized rule and decentralization. Mayors were popularly elected prior to the period of military rule between 1968 and 1979. The subsequent period of extreme centralization was characterized by increasing disparities between the coast (particularly Lima) and the poorer regions of the sierra and selva. The 1979 constitutional reforms reinstated municipalities. However, the attempts to decentralize functions during the García period were hampered by weak administrative and economic management, and plagued by extensive rent seeking and diversion of resources. As a result, there was little evidence of a reduction in regional disparities or poverty.

Despite the fact that the Fujimori government was fairly centralized, there was recognition of the importance of tailoring government programs to local conditions. Indeed, the subnational administrations were effectively deconcentrated agents of the center, which lacked clearly defined own-responsibilities and sources of financing. Centrally-determined and financed programs, such as FONCODES, were often the major source of activities for local governments. While these were, in principle, designed to reflect local priorities, there were few mechanisms to coordinate at the local or central levels, assess tradeoffs and establish priority spending. The center lacked the full information to make allocations, and there often was a lack of clarity between the spending by specialized agencies like FONCODES and the operations of line ministries and local administrations. Local governments were weak and lacked financing or clear responsibilities. The swing of the pendulum towards centralization, particularly in the latter part of the Fujimori administration, brought with it renewed concerns about rent-seeking and an absence of accountability at all levels of government.

These experiences highlight the importance of transparency and good governance in the design of decentralization or central programs to effectively meet the needs of the poorer sections of the population.

III. Decentralization and Overall Fiscal Constraints

Local governments continue to carry out only a small fraction of general government expenditures in Peru. Following a significant increase during the late 1990s, and a subsequent correction thereafter, execution of general government nonfinancial expenditures by local governments increased from 10 percent in 1999 to about 13 percent in 2004 (see Table 2 and Figure 2).

Figure 2.
Figure 2.

Percentage of general government primary expenditures executed at the local level, 1995-2004

Citation: IMF Working Papers 2006, 120; 10.5089/9781451863802.001.A001

Source: Staff with Central Bank (BCRP) data.
Table 2.

Fiscal Operations of the General Government and Local Governments, 2003

(Millions of soles)

article image
Source: Ministry of Economy and Finance (MEF)

Net of transfers among the central government and rest of the general government, which includes ONP, local governments, ESSALUD, FCR, sociedad de beneficiencia, and reguladores (Osiptel, Osinerg, Ositran, Sunass, Conasev, SBS) and registrales.

Includes S./2,609 million of current transfers from central to local governments.

Includes S./249 million of capital transfers from central to local governments.

Local operations are financed by both local revenues and central government transfers. Local revenues accounted for 43 percent of total local government current resources in 2003, and the rest included shared-revenues, such as for natural resources like the canon (which predate the existing decentralization initiative) and specific program transfers (see Table 3). Further, local governments may borrow in the domestic or international markets subject to the fiscal rules on indebtedness stated below. Also, a central government guarantee is needed for external indebtedness operations of subnational governments (General Law of Public Indebtedness, July 2005). Despite availability of financing, subnational governments may incur in payment arrears (that is “floating debt”) reflecting the earmarking, explained in further detail below.2 The magnitude of subnational “floating debt” is unknown, but is believed to be growing.3

Table 3.

Current Revenue Sources of the Local Governments, 2003

(Millions of soles)

article image
Source: MEF.

includes property tax, and casino taxes.

Municipal fees, fines, and revenues from the sale of goods and services.

Canon and sobrecanon.

Foncomún, Vaso de Leche, and other transfers.

In the medium term, local and regional governments may challenge compliance with the fiscal responsibility legislation. In recent years, as seen in Table 4, the overall primary balance of the local governments has been positive (reflecting their limited capacity to use the sizable canon transfers, which are earmarked to capital spending), thus, helping the consolidated nonfinancial public sector (NFPS) comply with the 1 percent of GDP deficit ceiling stipulated by the Fiscal Responsibility Transparency Law (FRTL). However, as decentralization proceeds and the regional and local governments improve their capacity to prepare and execute investment projects, they are less likely to have an overall fiscal surplus and thus their contribution to the NFPS overall balance is expected to deteriorate.4 In any case, any worsening of the fiscal position at the subnational level is limited by the fiscal responsibility rules set in the FRLT and Fiscal Decentralization Law (FDL). In particular, for each subnational government, the 3-year average primary balance must be positive, and the annual real primary expenditure cannot grow more than 3 percent.

Table 4.

Peru: Fiscal Operations of the NFPS

(Percent of GDP)

article image
Source: BCRP.Note: Regional governments are included together with the central government, since during this period they were deconcentrated organs of the center.

In response to any worsening of the fiscal stance of the subnational governments, the central government as well as the rest of the general government and public entities would have to improve their fiscal stance in order to offset this deterioration to stay within the FRTL ceiling. Further, as subnational governments gain more spending responsibilities, more fiscal flexibility by the central government will be needed (for example, to adjust to unexpected revenue shortfalls during the annual budget execution) to ensure compliance with the overall fiscal deficit ceiling. Moreover, as adjustments at the subnational level become more difficult, so do prospects for coordination to achieve the desired outcomes.

IV. Legal Framework

Since the 2002 constitution amendment, a legal framework for a gradual and fiscally neutral decentralization process was initiated. The amendment mandated political and fiscal decentralization, and established three levels of government (national, regional, and local) instead of two levels (national and local).5 This section summarizes the legal decentralization framework and highlights the key aspects that still need to be regulated.

Following the amendment to the constitution, three organic laws were enacted to guide the process of political and fiscal decentralization. These included: (1) the Organic Law of Regional Governments of 2002 (Law No. 27867); (2) the Organic Law of Municipalities of 2003 (Law No. 27972), which defined the organization and responsibilities for these levels of government; and (3) the Framework Decentralization Law of 2002 (Law No. 27783), which mandated a clear, gradual, and fiscally neutral devolution of expenditures and revenues and established stages for the devolution process. In particular, the devolution of responsibilities was to start with social and infrastructure programs, continue with other sectoral functions, and end with the transfer of education and health functions. This law also provides for passing legislation on the creation of regions (resulting from the merger process of regional governments) and incentives for the merging process.

The amendment of the FRTL in 2003 also placed a number of restrictions on subnational operations in order to ensure macroeconomic stability. These include the following:

  • A central government guarantee is required for contracting of external debt by subnational administrations.

  • Such external debt is only to be used to finance public investment.

  • The debt-to-current revenue ratio and annual debt service-to-current revenue ratio must be below 100 and 25 percent6 respectively, for each subnational government.7

  • The 3-year average primary balance of subnational governments must be positive.

In early 2004, the FDL, (Decree No. 955) was approved with the objective of ensuring a fiscally sound decentralization process. The FDL established a two-stage process for transferring revenues to regional governments. During the first stage, regional governments are to be funded through transfers from the central government, earmarked for certain social programs and infrastructure projects and consistent with the principle of fiscal neutrality. Regional governments that voluntarily merge to constitute larger regions will qualify to enter the second stage. At this stage, regions will receive 50 percent of certain taxes—in particular, indirect taxes (VAT and excises) and the personal income tax—collected in their jurisdictions. Further, regions will get additional transfers from the center if expenditures (excluding wage bill increases) exceed shared-transfers because of external factors (such as a natural disaster or a drop in regional income). Also, regions will get a bonus (earmarked for investment and maintenance of infrastructure) equal to the increase in tax collections above their potential level resulting from efforts to improve tax administration and reduce tax evasion.8 According to the law, a bonus to reward efforts to increase revenue collections may also be implemented in the first-stage. In order to ensure fiscal neutrality, the regulations to the FDL of September 2005 clarified that shared-transfers will not exceed the estimated cost associated to the devolved functions. Moreover, the regulations also clarified that fiscal savings resulting from the efficient provision of public functions by regions may be used for investment and maintenance of infrastructure.

In addition to the debt sustainability rules established in the FRTL, the FDL sets additional ceilings on debt service, debt stock (including floating debt), and expenditure growth. These rules include the following:

  • The non-guaranteed debt-to-current-revenue ratio and the annual non-guaranteed debt service-to-current-revenue ratio must be below 40 percent and 10 percent, respectively.

  • Domestic or external indebtedness operations must be used exclusively to finance investment projects.9

  • There is to be a 3 percent annual limit on real primary expenditure growth.

  • Short-term debt (including floating debt) at end-year cannot exceed one-twelfth of annual current revenues.

Further, the FDL established some reporting provisions for subnational operations. Regional and local governments must provide the central government their medium-term fiscal projections (indicating the planned external and domestic indebtedness operations), which must be consistent with the three-year Multi-annual Macroeconomic Framework published by the Ministry of Economy and Finance. Also, subnational governments must report on their quarterly fiscal performance and describe adjustment measures, if necessary, to comply with their annual targets.10

The FDL law also envisaged the passage of an Accreditation System Law to establish a system to assess if regional and local governments meet minimum capacity standards to qualify for the transfer of functions (including personnel) and the corresponding resources. The Accreditation Law was approved in July 2004, and implementing regulations were issued in November 2004. The Accreditation System Law also regulates the annual procedure for the devolution process. Based on the responsibilities assigned to subnational governments in the Organic Law of Regional Governments, Organic Law of Municipalities and the Framework Decentralization Law, the National Decentralization Council (NDC) elaborates an annual transfer plan, which integrates the annual plans prepared by each line ministry specifying the functions that may be transferred during that year and the criteria that subnational governments need to meet in order to qualify for the transfer of the functions and corresponding resources.11 By end-March, the NDC annual plan is approved by the Council of Ministers, and then, the regional and local governments may request the devolution of functions. The subnational governments’ capacity assessment—which may be performed by the NDC or a non-public agency—is undertaken during August and September, and by end-October, the NDC certifies the subnational governments for the delivery and thus further devolution of additional functions during the next budget cycle.

The Law on Incentives for Integration and Creation of Regions of June 2004 provided incentives for the voluntary merger of regional governments (via referendum) in order to obtain an optimum number of economically viable regions. Referenda were scheduled to take place at end-October 2005, 2009 and 2013. For the 2005 referendum, the NDC had approved technical requests for consolidation of 16 departments into 5 regions.12 The result of the October 30, 2005 referendum was that 15 of the 16 departments rejected the proposals, and consequently, none of the regional government are to merge into regions. The result was a reaction to the manner in which the process had been addressed, a lack of information for voters, politicization, changes in the number of votes required to form a region, legal loopholes (no ruling on what would happen to the resources from the canon and how the regional head was to be elected), and different interpretations of the incentives.

Before the FDL regulations were passed, there were different interpretations of the benefits from merger of regional governments. Indeed, under both the FDL and Law on Incentives, the shared-transfers (i.e., the 50 percent of national revenues collected under their jurisdiction) were interpreted by some as an automatic transfer in the sense of not being linked to the devolution of any function, whether or not this breaches the principle of fiscal neutrality guiding the decentralization process, as required under the Framework Decentralization Law and FDL. The FDL regulations state that the shared-transfers should be matched to the cost of devolved functions (as mentioned above). However, there is an incentive in the FDL regulations: each region created in the 2005 referendum, would receive an annual fixed transfer (S./210 million, equivalent to 0.08 percent of GDP) during four years.

Despite the concerted efforts to shore up the legal underpinnings, the legal decentralization framework is incomplete. Indeed, there is still a need to clarify the spending responsibilities of each tier of government and to issue regulations with sanctions for subnational governments not complying with their fiscal rules or reporting requirements.

Additional steps will be needed to ensure a neutral and gradual impact of the decentralization process. The sections below assess key aspects of the likely measures that may be required: (1) the assignment of own-source revenues, to ensure subnational accountability; (2) clarification of subnational responsibilities, together with significant improvements in subnational public financial management (with a focus on whether there is complete and standardized information on the activities of all levels of government); and (3) the design of the transfer system considering the disparities in the country. The administrative capacity of subnational governments to undertake functions needs to be satisfactory, but this also has to be accompanied by strengthened public financial management capacity—indeed, this is needed at all levels of government. Also, at the central level, there is a need to strengthen the accreditation system, to impose monitoring of the quality of the services delivered by subnational units, and to ensure the consistency of subnational fiscal operations with their fiscal rules and the overall macroeconomic fiscal framework.

V. Assignments and Accountability

In this section, we review the key elements of the decentralization process—starting logically with the functions that subnational administrations are required to carry out, and the financing arrangements for these assignments. However, much of the political debate is over the sharing of resources, with relatively little focus on the spending responsibilities. This opens up the possibility of overall imbalances, as the central government is left with more responsibilities than commensurate with its revenue shares.

A. Expenditure Assignments

The expenditure assignments among levels of government (national, regional and local) are, in principle, regulated by the Framework Decentralization Law. This law specified which functions are exclusively assigned to a particular level of government or are shared among several levels. It also provided for the clarification of the shared responsibilities, either with a new Organic Law of the Executive Power or with sectoral legislation. Although a new Organic Law of the Executive Power is under debate at Congress, it does not address the expenditure assignments among levels of governments to avoid duplications. Instead, because of the lack of consensus, the strategy that is being adopted consists in clarifying responsibilities gradually through the medium-term sectoral transfer plans, elaborated by the Ministries in coordination with the local and regional governments, following the guidelines established by the NDC.

Because expenditure assignments are yet to be clarified, it is difficult to assess if the process of decentralization is based on the principle of subsidiarity, which has driven similar efforts in other parts of the world. There is, in fact, relatively little discussion of responsibilities at the subnational level, including policy, and implementation, particularly the hiring and firing of personnel. The legislation does not, at present, assign to local governments the responsibility for either primary education or basic health care—indeed, these are examples of shared responsibilities among levels of government, with the subnational administrations having relatively little influence or control over these functions. It is noteworthy that under the devolution plan education and health are the last functions to be devolved.

The devolution of functions is proceeding gradually. As mentioned above, each year the NDC prepares an annual transfer plan and elaborates a list of subnational governments complying with the certification criteria to qualify for the transfer of functions and associated resources. The transfer plan approved in 2003 and with effect in 2004 included specific social and infrastructure programs and most of the subnational government were certified for these (see Table 5).13 Starting in 2004, the Accreditation System had to be in place to proceed with the transfer of functions for 2005. The transfer plan approved in May 2004 included functions in the following sectors: agriculture, fishing, tourism, trade, industry, artisans, energy and mining. However, because of delays in passing the implementing regulations of the Accreditation System Law, no further functions were devolved in 2005. Instead, the capacity assessment for the devolution of the sectoral functions outlined in the 2004 transfer plan took place during 2005. For these reasons, not many functions have been devolved since 2003.

Table 5.

Devolution of Functions, 2003

article image
Source: PRODES.

Number of governments.

B. Own-source Revenues

The design of the Peruvian intergovernmental system does not provide for control over rates or bases by local governments. Local governments do not have discretion over revenues, but instead, all tax policy issues are determined by Congress. Hence, it cannot be said that local governments have own-source revenues. Indeed, even the rates of revenues from “local” taxes (such as the property tax) that accrue entirely to the local governments are centrally determined. In this sense, such revenues are closer in concept to shared revenues (with a 100 percent share) than own-source taxes.

Discretion over own-sources of revenues is a necessary condition to engender subnational accountability and responsibility. A key element in establishing an effective hard budget constraint at the subnational level of government is that these jurisdictions should have access to own-sources of revenue to use in case of needed adjustments. Further, additional spending should be financed with a burden on the local electorate, so that spending choices are carried out in a responsible manner. This is an essential ingredient, for example, in the operation of participatory budgeting. If, however, the design of the revenue-sharing or transfer systems is such that the local governments are not induced to efficiently use their own-resources, then clearly accountable decisions cannot be assured. This involves control at the margin over the tax-rate structure, (e.g., within bands) or over the bases for particular taxes.

Different arrangements could help bring some discretion over own-source revenues. The legal framework of a unitary state is not, in principle, inconsistent with “room” for own-source revenues at the local level. For instance, Congress could retain control over the rate structures by enacting a band, say for the real estate tax (predial), with an upper and a lower bound within which the municipal government would set its particular rate give its spending needs. This could accommodate some of the scope for generating own-source revenues at the subnational level, while retaining the judicial responsibility for rate setting by Congress.

Similar arrangements are possible for other taxes, including those covered under the general revenue-sharing agreement for merging regional governments.15 Complex revenue-sharing arrangements lead to possibilities for game-play by levels of government, as each tries to maximize its own take. Further, these arrangements may also pose difficulties for tax administration (SUNAT). For example, the allocation of VAT across regional governments is complex since production may take place in one location and final consumption in a different one.

Thus, it would be important to ensure that:

  • Subnational governments have discretion over own-sources of revenues, for example, through rate bands on particular targets.16

  • Discretion over own-source revenues, as described above, does not necessarily imply that there must be local administration of taxes. Suitable arrangements could involve contracting SUNAT to collect some subnational taxes—this would be facilitated if there are shared bases, such as “piggy-backing” arrangements for some taxes. Each option should be subject to feasibility and ease of administration.

C. Sharing of Natural Resources

A major and increasing source of subnational revenues are the canons (revenue-sharing of natural resources with the central government, particularly for mining and petroleum). The main sources of canon revenues in Peru are shown in Table 6, over the period 2000-2004. The recent sharp increase in canon revenues is due both to the extensive new operations that have been brought on stream as well as the recent upturn in commodity prices.

Table 6.

Evolution of the Canon, 2000–04

(Millions nuevos soles)

article image
Source: MEF.

The canons are regarded as payments for the extraction of non-renewable assets. Canons accrue exclusively to producing local and regional governments (which are mainly those with lower poverty rates as shown in Figure 3). Thus, the distribution of canon revenues may further increase disparities between producing and non-producing areas (see Table 7), although this should address the inequalities in local incomes and services if applied effectively. The same concern applies to mining royalties, which have been effective since January 2005 and also accrue to the regional and local governments in the areas where the mines are located.

Figure 3.
Figure 3.

By departments: poverty rates and canon transfers

Citation: IMF Working Papers 2006, 120; 10.5089/9781451863802.001.A001

Table 7.

Canon Distribution by Departments

(Millions of soles)

article image
Source: INEI.

However, within the producing provinces and regional governments, the canon resources are distributed taking into consideration some of the needs of the local population. Thus, within-region inequality may be addressed in those locations. It is not evident that inequalities have been significantly reduced by this mechanism, and the reasons for this (including the possibly excessive earmarking discussed below) need to be assessed with additional data.

D. Earmarking

Earmarking in Peru is extensive at the subnational level and much of this is for investment. Some shared-revenues are earmarked for investment, such as revenues from the canon, which are fully allocated on an origin basis, but have to be used for capital expenditures. The authorities argue since the canon is generated by a non-renewable resource, it is appropriate to earmark it to infrastructure investment. Other earmarked transfers include, for example, FONCOR (the regional government compensation fund), which is also allocated to investment projects distributed among regional governments taking into account population needs. Other transfers, like Foncomún (the municipality compensation fund), which is distributed according to different criteria including population needs, can be used in part for current expenditures (Table 8).

Table 8.

Earmarking of Resources at the Local Government: Cuzco and Cajamarca, 2005

(Thousand soles)

article image
Source: Municipality of Cuzco, and Municipality of Cajamarca.

Originally approved budget, 2005.

Fiscal operations, first quarter 2005.

Earmarking limits operations of local government. Though the global effects of earmarking need to be assessed, the examples presented below of the impact of earmarking on selected municipalities—Lima and Cajamarca—show that earmarking, combined with the virtual absence of own-source revenues and general-purpose transfers (see below), further restricts the autonomous operations of local governments. In addition, since most revenues are earmarked for investment, and in turn, investment projects need to meet some standards according to the National System of Public Investment (SNIP), earmarking ensures that the additional revenues received by local governments in excess of short-run spending needs are not dissipated (as the spending devolution lags the revenue-shares). Nonetheless, with the weaknesses in monitoring of subnational operations, and greater sophistication of local governments in meeting or circumventing nationally set investment conditions, the level of subnational spending is likely to rise. Consequently, spending shortfalls at the subnational level may no longer contribute to meeting the overall targets of the FRTL.

The cases of Lima and Cajamarca presented in Table 9 show that:

  • The extensive earmarking together with limited capacity to prepare and execute investment projects, leads to significant capital under-spending in relation to budgeted resources. While current spending is fairly close to the budget estimates, earmarked capital spending fell far short of the budgeted amounts by 65 percent in Cajamarca and even lower, by about 45 percent in Lima.

  • Capital spending shortfalls and accumulation of cash balances may be accompanied by the buildup of arrears, or “floating debt” if commitments for current spending exceed available non-earmarked resources.

Table 9.

Budgets and Spending in Lima and Cajamarca, 2004

(Thousand of soles)

article image
Source: Regional Government of Cajamarca; Lima municipality.

The shortfalls in budgeted spending, whether or not due to earmarking, together with arrears, in many cases has led to political problems for the local governments, as the municipal electorates are not willing to accept situations that appear to reflect poor management (in addition to inadequate design) in the decentralization process.

As decentralization proceeds, it is likely that the earmarking will be moderated, but it is equally clear that the spending devolution must be the focus of further efforts, so that there is a balance between expenditure assignments and the revenues available at the subnational level.

E. Managing Subnational Spending

Many of the shortcomings of Peruvian public financial management (PFM) are also present at the subnational level, and often difficulties at the centre are magnified at the lower levels due to the more limited human capital and organizational capacity of the subnational governments. In fact, the decentralization process has added an additional layer of complications that needs to be addressed by the PFM system. In what follows, we discuss some of the basic issues that are relevant to each level of government for enhanced and transparent governance:

  • The fragmented nature of the PFM process in Peru generates difficulties in coordination and implementation of government policies. There are expectations that the government financial information system, SIAF, will provide a solution to the governance problems. However, the SIAF is a tool that will only work if the essential institutional and procedural preconditions have been met.

  • Peru lacks of an adequately designed Treasury Single Account at any level of government. The institutional limitations on subnational governments to spend resources (such as canon) result in the existence of large government cash balances of spending agencies at each level of government. The proliferation of government cash balances increases the Treasury’s borrowing requirements, fragments information and transparency and weakens controls on spending.

  • Other PFM shortcomings are the weaknesses in timeliness, clarity and transparency, and in information flows. These are due to the use of functional and economic budget classifications which deviate from international standards and are also subject to change from year to year. The chart of account used by the information system, SIAF, does not fully reflect the budget classification, and this poses difficulties in tracking spending according to appropriations. Moreover, the accounting is performed manually and outside the SIAF, leading to the possibility of discrepancies in spending data.

  • The spending control mechanisms are disjointed, relatively weak, and lack clarity. The problems with SIAF mean that the basic information on which controls should be based is hard to generate.

Although at present, all levels of government in Peru use the same budget classification, this needs to be revised to bring it closer to international standards for the economic and functional classifications. This should be the basis for the tracking of the spending process. The budget classification should be fully reflected in the chart of accounts in order to ensure full transparency. This would require that the SIAF reflects this chart of accounts and perform automatic accounting to generate timely and accurate information on government operations. Any changes that need to be incorporated into the SIAF should be verified at the national level before rolling out to the municipalities. If municipalities are allowed to have their own financial information management systems, these should reflect the same budget classifications and accounting and reporting frameworks.

For transparency and better cash management, there needs to be comparable generation of information on government bank accounts, through a system of treasury single accounts for each level of government. This will minimize the danger of a build-up of idle cash balances in tandem with the development of arrears or borrowing needs. This process needs to be developed at both the central and subnational government levels.

As discussed below, the full tracking of the stages of the budget process is essential in generating a record of spending and liabilities, including floating debt. The critical issue in the process is the setting of consistent standards for all levels of government. There could be a simplified format for the smallest municipalities with relatively few functions to generate a minimum level of comparable information across all jurisdictions.

Participatory budgets

A participatory budget (PB) system was created at the local level to allow citizens to participate in the process of identifying and prioritizing investment projects based on local needs. The political emphasis on participatory budgets is a welcome development, but in itself it does not impose a hard budget constraint on the local governments. Instead, PB creates expectations and pressures for additional spending driven by the following factors:

  • The timetable set for the discussion and approval of the PB is not aligned with that set for the formulation and approval of the entity’s annual budget. In fact, as currently designed, there are mismatches between these two timetables in the sense that the subnational governments begin to discuss the PB based on data and expenditure ceilings that are incorrect, thus, creating false expectations.

  • There is a lack of clear responsibilities or own-source revenues—thus there is no incentive to prioritize according to the intensity and urgency of the need.

  • The ability to monitor spending and outcomes is constrained by the limitations of the reporting and information system.

  • There are no sanctions for not delivering results.17 There are also no adequate formal channels for audit.

Debt management and reporting

Subnational governments may contract either external or domestic debt subject to the limits on subnational debt established in the fiscal responsibility legislation (FRTL and FDL). A central government guarantee is required for the external debt of local and regional governments (amended FRTL of 2003, and General Public Debt Law of July 2005); therefore, external indebtedness operations by subnational governments are counted against the annual limit on indebtedness operations (including guarantees) that the central government may contract or provide in a given year (this limit is established each year in the Annual Public Indebtedness Law). Subnationals may contract domestic debt (including floating debt), and the guarantee by the central government is not required.

At present, it is difficult to estimate the extent of debt exposure at the municipal level. The disjointed treatment of debt, together with an absence of a centralized risk register (which should also cover floating debt, guarantees as well as contingent liabilities) makes it difficult to assess compliance with fiscal rules. Indeed, contingent debt (for example, originating from trials or judicial decisions) and floating debt lack adequate recording.18,19 The World Bank is undertaking a study of subnational debt, including floating, in order to assess the magnitude of subnational debt stock and identify municipalities that might face debt stock problems, including the municipality of Lima. Preliminary conclusions of this study show that unregistered debt is significant (the ratio of unregistered to registered debt is about 4); debt is highly concentrated (in particular, 40 percent of total debt is explained by 20 municipalities); and, a significant number of municipalities do not comply with FRTL rules.

In order to address the loophole in monitoring subnational debt, the General Public Debt Law mandated the registration of indebtedness operations by all levels of government. The General Public Debt Law requires that regional and local governments register all their indebtedness operations using a debt module in the subnational government financial information management system (SIAF-GL)—such a module is not yet operational. As mentioned above, there is an inordinate reliance on a SIAF to address the problems of effective subnational public financial management. While this is a tool that might eventually help, it cannot address basic flaws in procedures or design. Indeed a poorly defined local SIAF might hinder the improvements in subnational governance and financial management that are needed for sound decentralization. In addition to the need to establish a registry of subnational debt, the definition and recording of floating debt (which is not covered under the definition of an indebtedness operation under the General Public Debt Law) still needs to be addressed.

Looking forward, it will be important to establish full information for all subnational operations, including for subnational debt. There should be very clear reporting requirements for subnational governments that build on the standard budget classification, and tracking of all stages of the budget execution process.

F. Transfer Design

The neutrality of the decentralization process is assumed not to exacerbate geographical disparities. Transfers, either in the first stage or second stage of the decentralization, are to be linked to the devolution of responsibilities (see above). In other words, the magnitude of these transfers is to be calculated when the expenditure assignments are determined. The complexity and difficulties of the arrangements in Peru for fiscal management at all levels of administration may jeopardize the “assumed” neutrality of the transfer system.

Moreover, the range of transfers in Peru increases the risk that the decentralization initiative might exacerbate geographical disparities. Few of the transfers take into account population needs in a comprehensive manner, although different aspects are addressed in different programs. For example, FONCOMUN and FONCOR are distributed among municipalities and regional governments, respectively, based on poverty indicators, demography, unmet basic needs, and capacity to execute investment projects. In principle these transfers should contribute to the reduction of selected aspects of regional inequalities. However, the overall effects of the individual transfers need to be assessed in an integrated framework. Other automatic transfers, such as canons, clearly lead to larger geographical disparities and limit the capacity of the central government to “equalize” access to public services across areas.

In many countries, the establishment of an equalization transfer system is an integral part of the design of an effective decentralization program. Best practices (from Denmark to China) suggest that consideration be given to a mechanism that takes into account the differential costs of the provision of services (particularly important in a country with the difficult topography and linguistic differences of Peru), as well as the capacity to raise own-source revenues. Details of international experiences and models are summarized in Ahmad (1997).

VI. Conclusions

The ongoing decentralization effort in Peru has been based on prudent considerations of fiscal sustainability. However, additional work is needed in order to deepen the decentralization process and strengthen the incentives for effective subnational governance.

A critical issue in Peru will be the proper sequencing of the decentralization effort, in order to ensure that appropriate incentives are established to manage subnational operations effectively and in a transparent manner, together with the emphasis on prudent and sustainable macroeconomic management that has been a notable feature of the efforts of the administration in the recent past. This involves clarifying spending responsibilities, making subnational administrations responsible for the functioning of various sectors, together with the greater reliance on own-source revenues. This has to be accompanied by a set of more transparent and equitable transfer mechanisms to avoid exacerbating regional disparities.

In addition, there has to be a significant strengthening of subnational administrative capability to ensure an efficient provision of decentralized public services, together with the requirement of standardized information on operations on all government operations, including on floating debt and the buildup of commitments.


  • Ahmad, Ehtisham, ed., 1997, Financing Subnational Expenditures, an International Comparison of Grants (Cheltenham, UK: Edward Elgar).

  • Ahmad, Ehtisham, María Albino-War, and Raju Singh, 2006, “Subnational Public Financial Management: Institutions and Macroeconomic Considerations,” in Handbook of Fiscal Federalism, ed. by Ehtisham Ahmad and Giorgio Brosio (Northampton, Massachusetts: Edward Elgar).

    • Search Google Scholar
    • Export Citation
  • Ahmad, Ehtisham and Maria Albino-War, eds., 2006, Managing Subnational Finances (Washington: International Monetary Fund).

  • Ahmad, Ehtisham, and Robert J. Searle, 2006, “On the Implementation of Transfers to Subnational Governments,” in Handbook of Fiscal Federalism, ed. by Ehtisham Ahmad Giorgio Brosio (Northampton, Massachusetts: Edward Elgar).

    • Search Google Scholar
    • Export Citation
  • Casas, C., and G. Yamada, 2005, “Medición de impacto en el nivel de vida de la población del desempeño macroeconómico para el período 2001–04,” Universidad del Pacífico. Documento de trabajo.

    • Search Google Scholar
    • Export Citation
  • Dabán, Teresa, 2004, Peru—“Fiscal Decentralization in Latin America: Lessons for Peru,” in Peru: Selected Issues, IMF Staff Country Report No 04/156 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

We are grateful to María Albino-War, Mark Horton, Alejandro López Mejía, Rossana Polastri, Gilbert Terrier, and Mauricio Villafuerte, for helpful comments. Helpful comments were also received from the Ministry of Economy and Finance and the Central Bank of Peru. All errors are ours. This paper is to appear in Ahmad and Albino-War, eds., 2006.


According to the Fiscal Decentralization Law (FDL), domestic or external borrowing by subnational governments may only be used to finance the execution of investment projects, but not for current expenditures. However, the General Law of Public Indebtedness, while indicating that external borrowing should be solely for investment projects, does not address the use of domestic borrowing.


This is suggested by selected surveys of municipal operations initiated by the World Bank.


The Ministry of Economy and Finance is reported to be providing technical assistance to subnational government to strengthen their public financial management capacity and to prepare investment projects.


There are 1,829 municipalities and 26 regional governments in Peru.


This is more restrictive than the limit of 30 percent imposed by the Organic Law on Municipalities and could cause some confusion.


The definition of current revenues used to compute the ratios includes transfers from other levels of government, but excludes the operating balance from previous years, the financing through domestic and external indebtedness operations, and the revenues earmarked to trust funds (fideicomisos) (Regulations for the FRTL, November 2004).


Lima and Callao do not need to merge to get the additional resources envisaged in the second stage.


All investments are subject to a national system that evaluates feasibility (SNIP).


The FDL establishes that by end-2005, local and regional governments will provide the Ministry of Economy and Finance a fiscal management report, that would report on compliance with the fiscal rules and suggest adjustment measures if necessary to ensure compliance after a two-year period.


In turn, the annual plan of the line ministries is based on a five-year sectoral transfer plan prepared by the ministries in coordination with the regional and local governments.


These departments were: Tumbes, Piura and Lambayeque; Ancash, Huánuco, Lima provincias, Pasco, and Junín; Apurímac and Cuzco; Ica, Ayacucho and Huancavelica; Arequipa, Puno, and Tacna.


Lacking an accreditation system, a simplified provisional system for the certification of subnational governments’ capacity to deliver some social and infrastructure programs was implemented in 2003.


While regional governments may get transfers from the central government, including shares of national taxes, no tax has been created for regional governments yet.


However, it is not sufficient to ensure that these revenue-handles will be used, as the incentives may be offset, for example if there are automatic transfers to meet subnational deficits.


The difficulties with special purpose transfers are legendary, and are well known in relation to specific programs in Peru (including the vaso de leche program) and similar programs in other countries. Central programs with local implementation are difficult to monitor and the incentives to manage these effectively is generally absent.


On a regular basis, the Ministry of Economy and Finance collects information on contingent liabilities related to guarantees and judicial decisions.


In some municipalities, such as Cajamarca, floating debt has been accompanied by a build-up of cash deposits at the same time, due to the earmarking of canons and transfers.

Fiscal Decentralization and Public Subnational Financial Management in Peru
Author: Mr. Ehtisham Ahmad and Ms. Mercedes Garcia-Escribano
  • View in gallery

    By departments: product per capita and poverty 2002

  • View in gallery

    Percentage of general government primary expenditures executed at the local level, 1995-2004

  • View in gallery

    By departments: poverty rates and canon transfers