INTERNATIONAL MONETARY FUND WORLD BANK

FOR OFFICIAL USE ONLY INTERNATIONAL MONETARY FUND WORLD BANK BOLIVIA HEAVILY INDEBTED POOR COUNTRIES INITIATIVE ASSESSMENT August, 2004

List of Abbreviations AAP = Assessment and Action Plan CENCAP = Centro Nacional de Capacitación, Training National Center CFAA = Country Financial Accountability Assessment CGR = Controller General Office COSO = Committee of Sponsoring Organizations CUT = Treasury Single Account CV = Vigilance Committee DGC = Accounting Office DGMP = General Directorate for Public Procurement FAD = IMF Fiscal Affairs Department FNDR = National Regional Development Fund FPS = Fondo de Inversion Productiva y Social, Social Investment Fund GDP = Gross Domestic Product GFSM = IMF Government Finance Statistics Manual HIPC = Heavily Indebted Poor Country IMF = International Monetary Found INPS = National Social Security Institute INTOSAI = International Organization of Supreme Audit Institutions MOF = Ministry of Finance MTEF = Medium-Term Expenditure Framework NGO = NonGovernmental Organization PFM = Public Financial Management PEM = Public Expenditure Management PER = Public Expenditure Review POA = Plan Operativo Anual, Annual Operation Plan PRSP = Poverty Reduction Strategy Paper SAFCO = Financial Administration and Control System SAYCO = Administration and Control Systems SCCS = Accounts Section of the Supreme Court SICOES = Sistema de Información de Compras Estatales, Public Procurement Information System SIGMA = Financial Management Information System SINCOM = Sistema Integrado de Contabilidad Municipal, Municipal Integrated Accounting System SNC = Servicio Nacional de Caminos, National Road Service SNA = Servicio Nacional de Aduanas, Customs Agency SNII = Servicio Nacional de Impuestos Internos, National Tax Service TGN = National Treasury UPF = Fiscal Programming Unit VIPFE = Vice-Ministry of Public Investment and External Financing 2

TABLE OF CONTENTS EXECUTIVE SUMMARY...4 I. ASSESSMENT OF THE CAPACITY TO TRACK POVERTY-REDUCING PUBLIC SPENDING.. 7 A. Assessment Coverage...7 B. Budget Formulation...8 C. Budget Execution...16 D. Budget Reporting...20 E. Public Procurement...25 II. ACTION PLAN TO UPGRADE PUBLIC EXPENDITURE MANAGEMENT...27 Tables 1. Public Expenditure Management AAP Indicators in Bolivia...6 2. HIPC II: Balance Distribution (2003) under the National Dialogue Law...8 3. Government Estimations for the Budget (in percent)....11 4. Budget appropriations and execution by line-ministries...12 5. Budget arrears by selected line-ministries and local Governments...17 6. SIGMA Coverage, 2003...23 3

EXECUTIVE SUMMARY From May 10 th to May 19 th, 2004, an IMF/Bank mission conducted a HIPC assessment and action plan (AAP) on the monitoring of poverty-reducing expenditures. The team comprised Carlos Mollinedo-Trujillo (LCSPE, Team leader), Helio Tollini (IMF/FAD), Julio Velasco (LCCBO). Lurdes Linares (LCCBO) provided advice on the internal and external control and ensured coordination with the CFAA team. Simon Cueva, IMF Resident Representative, joined the mission at the first meeting with the Bolivian Authorities. The mission built up on CFAA, PER and IMF/WHD work and coordinated closely with their respective teams. It held discussions with a wide range of senior and technical Government officials (at national and sub-national levels) as well as with donors. The mission met senior and technical staff in MOF s three Vice-Ministries (Budget, Public Investment and Treasury), the Fiscal Programming Unit, the Accountant General s Office, the General Controller Office, the SIGMA/ILACO project, the Department of Public Procurement, the Prefecture of La Paz, the municipalities of La Paz and El Alto, and the donor community. The mission discussed extensively the draft AAP with the authorities and agreed an action plan with them. The AAP was discussed in a final meeting chaired by Mr. Luis Arnal (Vice- Minister of Public Investment and Foreign Financing, Ministry of Finance), with representatives from all Government units involved. The authorities stressed the importance of taking into account the extremely fragile political and economic situation while designing an action plan for PEM improvements in the short-term. After detailed discussions, agreement was reached on each and all AAP recommendations. Based on this frank and open discussion, it is the mission s opinion that the authorities are likely to agree to publish the AAP following the review process. ASSESSMENT In spite of the difficult political, economic and social situation, the draft assessment indicates that Bolivia s PEM system has shown some improvement. Bolivia achieves the benchmark in 4 out of the 16 public expenditure management (PEM) indicators, compared to 5 out of 15 in the assessment carried out in 2001. The assessment shows that Bolivia has improved on 2 out of the 15 PEM indicators assessed in 2001, mainly due to more specific AAP guidelines related to the budgetary coverage. Despite the benchmark being reached, the coverage of the Central Government remains incomplete: public universities expenditures and revenues are poorly covered. The poverty reducing expenditures are clearly defined, although neither a proper functional classification nor a budget tagging mechanism exists. Improvement has been made in the areas of routine accounting and reconciliation since the progressive implementation of an integrated financial management system (SIGMA). Nowadays, the system covers all the Central Government except for the Judiciary, some decentralized agencies, one prefecture and two municipalities. Despite these improvements, the assessment shows that there are still some weaknesses in the PEM system. In total, Bolivia received lower ratings on 6 indicators compared to 2001. Local 4

Governments, which are responsible for more than 50 percent of total poverty related public expenditure, often tend to have weak PEM systems. Significant shortcomings identified in the assessment include: (1) a systematic over budgeting of current and capital items, both at central and local levels; (2) internal and external audit being only partially effective, resulting in weak oversight and a major absence of regular, timely and audited reports and financial statements; (3) the procurement system suffering from a weak legal framework and weak enforcement of rules. The achievement of actions agreed in the 2001 Action Plan has been hampered by the political instability and the difficult economic and social situation. Frequent changes in Administration combined with social unrest and severe fiscal crisis have delayed and hindered the Government efforts to improve PEM in Bolivia; in February 2003 the announcement of the national budget (in which projected fiscal adjustment measures were implicitly included) made by the Minister of Finance triggered a major social unrest. PEM in Bolivia has shown some progress in spite of these difficulties. Out of nineteen actions agreed in 2001, only 6 have been fully implemented, the implementation of another 6 has been initiated and the implementation of further 7 has not started. Most of the 2001 AAP recommendations that have not been completed have been postponed in light of the current social, economic and political difficult context. Additionally, two new actions have been introduced as the new Government decided that more effort is further needed. Table 3 presents the implementation status of the 2001 Action Plan in detail. ACTION PLAN The mission formulated a new action plan, fully agreed upon with the Government, taking into consideration the present reality and needs of Bolivia. In the budget formulation process, the mission recommended the adoption of realistic macroeconomic assumptions, compatible with the fiscal program; the introduction of a true functional classification; the strengthening of technical expertise in Congress on budget formulation issues; and the assessment of the delicate issue of how to better budget public universities. The mission stressed that authorities should have a strategy to develop a Medium-Term Expenditure Framework (MTEF). A former Government priority, the idea was abandoned due to the recent high level of political uncertainty. The mission underlined the importance of maintaining the existing achievements on PEM, while longer-term initiatives continue to be gradually implemented. The mission stressed the importance of maintaining SIGMA and SIIF systems, as there are concerns among Government officials about their financial costs. The action plan presents a number of specific key recommendations on reporting, internal control, and external audit, in order to strengthen these capacities. Regarding procurement, the mission suggested to the Government to build on the good experience of the recent decree, and to later promulgate a law consistent with international practices. 5

Table 1: Public Expenditure Management AAP Indicators in Bolivia ASSESSMENT Standard Desk 2001 2003-04 Benchmark Assessment Assessment Assessment FORMULATION COMPREHENSIVENESS 1 Fiscal reporting adequately covers the Government Finance Statistics definition of the general government sector A B B A 2 Government activities are not funded through inadequately reported extrabudgetary sources to a significant degree A A B A 3 Budget outturn data are quite close to the original budget B C C C 4 Fiscal reports include grants projected to be provided by donors A A A A CLASSIFICATION 5 Budget expenditures are classified on an administrative, economic, and detailed functional or programmatic basis B C B B 6 Poverty-reducing expenditures are clearly defined A B B B PROJECTION 7 Multi-year expenditure projections are integrated into the budget formulation process A C B B EXECUTION INTERNAL CONTROL 8 There exists a small stock of expenditure arrears, with little accumulation of arrears over the previous year A B B C 9 Internal control is effective A C B B 10 Tracking surveys are in use, or are unnecessary B C B C RECONCILIATION 11 Satisfactory reconciliation of fiscal and banking records is undertaken routinely A B B B REPORTING IN-YEAR REPORTING 12 Internal fiscal repots are received within four weeks of the end of the relevant period B C B C 13 Good-quality classification of poverty reducing spending is reflected in the in-year budget reports A C C C FINAL AUDITED ACCOUNTS 14 Routine transactions are entered into the main accounting system (s) within two months of the end of the fiscal year A A B C 15 An audited record of the financial outturn is presented to the legislature within twelve months of the end of the fiscal year B C B C NEW PROCUREMENT 16 The procurement system supports efficiency and effectiveness in the expenditure of public funds through clear A B and enforceable rules that promote competition, transparency and value for money. TOTAL NUMBER OF BENCHMARKS MET 3 5 4 Notes: Please shade cells in cases where the assessed indicator meets or exceeds the standard benchmark level

I. ASSESSMENT OF THE CAPACITY TO TRACK POVERTY-REDUCING PUBLIC SPENDING A. Assessment Coverage Bolivia s General Government is constituted of a Central Administration, 9 Departments and 314 Municipalities (13 new ones will be created by December 2004), besides other decentralized institutions and the social security sector. Each Department has a Prefecture, which corresponds to a deconcentrated Central Government agency at the regional level. The Central Government appoints the prefecture governors, and their budget is included in the Central Government s budget sent to the National Congress. Municipality majors are elected by direct vote every 5 years. The deconcentrated administrations are closely tied to the Central Government structure. For the education and health sectors, the corresponding ministry takes the decisions on formulation and execution of prefectures budget. The same laws and regulations govern both administration levels, and they use the same budget and accounting classifications as well as control mechanisms. The prefectures receive 20 percent of the significant hydrocarbons tax, specific royalties, and budget transfers from the Central Government. Municipal Governments are autonomous; their budgets are not included in the National Budget. However, the (consolidated) Government public investment program is somewhat disaggregated and reports projects executed by the municipalities. The role of municipal Government financial operations is significant: under the law, 20 percent of national tax collections are distributed to subnational Governments; a tributary coparticipation is allocated from the central budget and automatically transferred by the National Treasury (TGN). Locally elected officials are empowered to define local Governments development and spending programs. The municipalities can levy additional taxes on real estate and vehicle property. Currently, approximately 55 percent of poverty related public spending is undertaken at subnational levels of Government, and about half of the total public investment is channeled through municipalities. The National Dialogue Law (Ley del Diálogo Nacional, N 2235/2001), establishes the allocation of the HIPC II resources for a period of 15 years (Table 1). The municipalities receive the major share (56.5 percent). Additional amounts of poverty-reducing spending executed by sub-national levels are not expected in the near future, due to political difficulties in further decentralization. Hence, a significant share of the poverty reduction effort as well as from the HIPC II resources will remain with the Central Government, at least in the short to medium term. 7

Table 2. HIPC II: Balance Distribution (2003) under the National Dialogue Law CONCEPTS HIPC II Resources Bank charges (commissions, cost of checkbooks, etc) Balance transferred to Central Government National Solidarity Fund for Education National Solidarity Fund for Health National Solidarity Fund for SUMI benefits (mother and child) Balance transferred to Municipalities Resources for productive infrastructure (70 percent) Resources for education services (20 percent) Resources for health services (10 percent) Note: estimates from the European Commission in Bolivia. PERCENTAGE 100.0 percent 1.2 percent 42.3 percent 22.1 percent 10.2 percent Max. 10.0 percent 56.5 percent 39.6 percent 11.3 percent 5.6 percent Unless otherwise advertised, the remainder of this assessment will discuss, under the various indicators, local, regional, and Central Government budget management and its capacity to track poverty related spending. The report s findings on prefectures and municipality are based on meetings with officers from major prefectures, municipalities and their association, as well as on previous information available from the PER and CFAA. B. Budget Formulation Indicator 1: Coverage of the budget or fiscal reporting entity Benchmark 1: The budget covers the Government Finance Statistics Manual s (GFSM) definition of the General Government sector, i.e., the central, regional, and local Governments, and all Government operations, whether funded through the budget or not. General Assessment: (A) - Benchmark is met. Overall, the fiscal data match the definition of the General Government sector set out in the GFSM. The central and local Government budgets are quite comprehensive. The financial operations of all these structures are tracked in the existing accounting systems. The budgetary coverage is comprehensive, fitting closely the consolidated General Government as defined in the IMF s Manual of Government Finance Statistics (GFSM). The Central Government s budget includes the deconcentrated Government agencies, including prefectures, and is submitted for approval to the National Congress. Coverage of expenditures and revenues from the 10 universities is limited to inter- Governmental transfers, classified as recurrent expenditures (even if actually spent on capital items), from the Central Government. Universities own-source revenues, estimated to be equivalent to 15 percent of their total expenditures and to approximately 1 percent of the Central Government s budget, are not registered at all. Moreover, although supposed to send monthly execution information to the Accounting Office (DGC), in practice universities usually provide incomplete and delayed information, claiming they have financial autonomy. 8

Information on the Central Government s recurrent expenditure execution is obtained through DGC, and on capital expenditures execution through the Vice-Ministry of Public Investment and External Financing (VIPFE). In both cases, as the financial management information system (SIGMA) still does not cover most prefectures and universities (Table 2), the Fiscal Programming Unit (UPF), which consolidates the General Government numbers, only receives the information after a 5 months lag. Municipalities have their own budget, submitted to their respective Representative Counsel. The central budget only records municipalities operations in respect to the amount of transfers conducted. As SIGMA has only been implemented in two municipalities (La Paz and Cochabamba, off-line), information on their budget execution is still a problem. The UPF, which consolidates the General Government numbers, follows closely 111 municipalities spending, which represent between 75 to 80 percent of local Governments expenditures, and estimates the rest. Other decentralized agencies, including social security institutes (cajas de salud), register all their operations in the Central Government s budget and are required to periodically submit reports. Yet, expenditure allocations are not effectively monitored (in average there is a one-month lag in the submission of information). After the recent implementation of SIGMA, all transactions of these institutions are registered and the information on their activity is made readily available. At the municipal level, the Ministry of Popular Participation has developed a simplified PFM system (SINCOM) for small and medium sized municipalities. Indicator 2: Degree of spending being funded by inadequately reported extra budgetary sources Benchmark 2: Government activities are not funded to a significant extent through inadequately reported extra budgetary sources (less than 3 percent of total expenditures). General Assessment: (A) - Benchmark is met. Extra budgetary operations do exist, but are not significant. The budget incorporates own revenues generated by hospitals and by embassies since the formulation process. During execution, if collection of these revenues shows to be higher than initially estimated, the budget has to identify them, and authorization from the National Congress is needed before they can be spent. The Central Bank budget is also included within the Central Government s budget. Extra budgetary expenditures are not permitted. However, until the end of 2002, due to the inherent weakness in budget planning, a great number of off-budget transactions were still executed outside of the system. This resulted in an overspending of certain budget and 9

treasury accounts, a practice that is in contradiction with the Budget Law, which prohibits this manner of public expenditure management. While SIGMA does not permit off-budget transactions, by subverting the system through manually recording accounting transactions, institutions are able to justify the expenditures ex post through the issuance of an Executive Decree. At the end of the fiscal year, the Government issues a Supreme Decree that regularizes pending cases. During this process, appropriations are transferred from budget categories with positive balances to those that have negative balances and additional spending carried out by agencies with their own resources are finally registered. In this way, public information on the executed budget does not show overdrafts. In 2002 this off-budget transactions amounted to about 1.2 percent of GDP or 3.8 percent of total public expenditure (approximately US$ 100 million). Starting 2003 off-budget transactions will be reduced significantly because this ex-post regularization practice has been explicitly forbidden and enforced. Moreover, although only for emergency adjustments, debt notes (Notas de Débito), an extraordinary fast payment mechanism, can still be used. This prevalent practice last happened during the political and social crisis of October 2003. These notes enabled the payment of the costs of social unrest, and were subsequently registered into the system. For 2004, the Financial Law has explicitly prohibited overspending through the mechanism described above. Additionally, as discussed in Indicator 1, universities do raise non-registered administrative fees and service charges. According to UPF, the known revenues of the universities outside the General Government budget amounted to about B$ 320 million in 2003, or 1 percent of total Government spending. Indicator 3: Reliability of the budget as a guide to future Benchmark 3: Budget outturn data are very close to the original budget General Assessment: (C) The benchmark is not met. There is a systematic over budgeting of current and capital items, both at central and local levels. Budget reallocations are significant, particularly at the institutional level. The Central Government budget is not a reliable guide overall to the Government s fiscal policy. Although the average execution rate of the Central Government shown on Table 4 is highly acceptable (1.9 percent), the original budget tends to over-estimate revenues and overstate capital expenditures. This is part of the bargaining strategy used by Government entities, and gets worst when the formulation process goes to the National Congress, which usually artificially inflates revenues in order to incorporate new expenditures (especially pork-barrel amendments). To estimate revenues, the Government takes into account two variables, the potential nominal growth of the economy and the 10

efficiency gains of the collecting agencies, i.e. Impuestos Internos and Aduanas. Both are typically overestimated. Table 3 presents the real growth estimates used for the revenue forecasts, compared to actual growth rates, and the actual versus predicted deficits for the period 1999 2002 (inflation and exchange rate have been pretty stable in recent years in Bolivia). Growth estimates have been overestimated, on average, by more than 2 percentage points each year. However, most of the revenue overestimation has been linked to unrealistic predicted efficiency gains in revenue collection, which although they have diminished in the last two years are still high; over the 1999 2001 period they averaged nearly 10 percent per year for Impuestos Internos, and more than 16 percent per year for Aduanas. Interviews with participants in the budgetary process suggest that actual efficiency gains have been closer to 3 percent per year. Back of the envelope calculations, based on the evolution of revenues as a share of GDP, suggest that efficiency has increased by about 2.5 percent per year since 1995. Table 3. Government Estimations for the Budget (in percent) Growth Deficit Predicted Efficiency Gains Year Predicted Actual Predicted Actual SNII SNA 1999 5.2 0.6 3.6 3.7 9.0 17.0 2000 4.0 2.3 3.7 4.2 9.9 21.8 2001 4.5 1.5 3.7 7.0 10.0 10.0 2002 3.0 2.7 5.7 8.6 3.0 3.0 2003 2.9 2.4 6.5 7.9 5.0 5.0 As a result, during the year there is an acute shortage of resources to cover budgeted expenditures. Given the fiscal targets, Treasury manages to meet them through establishing monthly commitment and payment limits to line-ministries, which makes cash rationing an important feature of the system. The uncertainties concerning budget ceilings ensuing from this system produce operative inefficiencies over commitments and short-term planning horizons. Thus the budget is not being used as an effective tool for reflecting development plans and policies, but rather for defining budget envelopes for each agency. In general, the political bargaining power of each agency determines the allocations. The TGN decides reallocations of current expenditures, whereas public investment decisions are competence of VIPFE. The budget is modified once over the year, to formalize the situations in which prior authorization by Congress is required. These situations include additional appropriations for the overall budget, for agencies that have underestimated their revenues, for any increase in payroll, and when reallocating from investments to recurrent expenditures. Other in-year expenditure reallocations within line agencies, called reversion, are authorized and can be done at any time during the year. Under this procedure, the agency changes the information introduced into the system after having used the funds for purposes different from those originally intended. Operations related to projects financed by external loans or grants are often introduced during execution and are reflected in the budget execution statement. 11

The information available on municipal budget realism in Bolivia does not show satisfactory results (Table 4). The analysis of selected municipalities shows that in 2 of the last 3 years the executed budget was more than 3 times bigger than the original one - the other year it was almost 2.5 times bigger. The credibility of municipal budgets largely depends on revenue projections, the quality of which is highly variable. Further allocation of funds for investment projects is somewhat less variable and is based on transfers from cofinancing institutions such as FPS and FNDR as well as sources of external funding. Although there are incomplete fiscal transfers under revenue sharing and HIPC II arrangements, these are less variable than revenue estimations/collections. Table 4. Budget appropriations and execution by line-ministries 2001 2002 2003 Approved Executed % Approved Executed % Approved Executed % Total /1 34933 36937 105.7 37794 38343 101.5 34428 33918 98.5 Central Administration 21146 19249 91.0 23880 22746 95.3 25418 25075 98.7 Treasury 16693 14744 88.3 19012 18002 94.7 20258 20415 100.8 Ministry of Government 735 799 108.7 845 824 97.5 921 867 94.2 Ministry of Education 270 336 124.2 313 297 95.1 332 229 68.9 Ministry of Defense 918 1066 116.1 976 1045 107.0 1018 1178 115.7 Ministry of Finance 309 300 97.0 249 193 77.6 239 193 80.7 Ministry of Health 256 221 86.6 266 241 90.6 361 279 77.2 Ministry of Agriculture 366 256 70.0 430 344 80.0 544 384 70.6 Legislative Branch 200 207 103.5 212 222 104.9 203 214 105.3 Judicial Branch 409 316 77.5 435 333 76.5 402 357 88.7 Rest of Central Adm. 992 1004 101.2 1143 1246 109.0 1140 960 84.2 Deconcentrated agencies /1 2531 3716 146.8 3313 4115 124.2 3191 3043 95.4 Public finance sector /1 1338 960 71.8 1693 908 53.7 995 598 60.0 Social security agencies /1 1169 1180 100.9 1334 1420 106.5 239 261 109.3 Public enterprises /1 2893 2776 96.0 1360 1263 92.8 786 947 120.5 Prefectures /1 4375 4525 103.4 4576 4050 88.5 3052 1678 55.0 Municipalities /1 1481 4531 305.9 1638 3841 234.6 747 2316 309.9 1/ Note: Some decentralized institutions, prefectures and municipalities have not reported their financial statements to DGC; thus their information is not available. 12

Indicator 4: Inclusion of donor funds. Benchmark 4: Budgets and/or fiscal reports at all levels of Government include, without exception, grants expected to be provided by donors, and the capital and current expenditure of all multilateral or bilateral on Government activities. General Assessment: (A) The benchmark is met. All activities financed by donors are disclosed on the budget. In general, all grants and donations are included in the budget and in the actual outturn data. During the formulation of the budget law, there are sometimes lags between the approval of financing by donors and the preparation of the beneficiaries budget. In these cases, the beneficiary agency can increase its appropriations and get national resources as counterpart from the Treasury without requiring congressional approval, simply through budget modification. Whenever a credit/donation benefits several public entities, VIPFE records it and distributes the resources. The Government keeps a good track of external inflows. A standard approach to the handling of donor funds seems to be used in the large majority of cases, even at local level. Procedures seem to be simplified, conditionality from donors reduced, and only exceptionally, in small municipalities, deviations from these procedures occur. Sometimes cooperation agencies try to impose their own procedures for resource disbursement, and grants may not be included in the investment budget. In these cases, almost insignificant in terms of value according to the information obtained, donors and beneficiaries carry out the corresponding operations directly with each other, without a budgetary register. In bigger municipalities, according to our meetings with them and information from Government officers, previous register in the budget is required. Indicator 5: Classification Benchmark 5: Budget expenditures are classified on an administrative, economic, and detailed functional or programmatic basis. General Assessment: (B) - Benchmark is met. The budget classification system includes administrative, economic, and programmatic classification. A functional classification is not available. 13

The Central Government s and the rest of the public sector entities budget operations are classified in conformity with international standards. They comprises administrative, economic sector, source of financing, and programmatic classifications. The institutional and economic classifications seem to be properly defined and implemented. Although current and capital budget are not formulated under the supervision of the same entity, they share the same institutional, economic and programmatic classifications (within different levels of Government). Within the programmatic classification, the specific activities and projects are budgeted. Notwithstanding, under the Financial Administration and Control System (SAFCO) norms, all agencies are required to present annual work programs using the program concept Planes Operativos Anuales (POA) which together form the budget submitted to Congress. Although there is a weak link between programs and outputs, the programmatic classification provides sufficient delineation of activities, enough to enable poverty-reducing spending to be identified. Neither the Central Government nor local Governments use functional classification to relate the uses of funds to the ends sought by the public sector. DGC uses a tentative functional classification to produce ex-post reports on budget execution. However, this functional classification is based on manual calculations amalgamating the institutional classifications and does not represent the coding of individual transactions, as acceptable by international standards. Indicator 6: Identification of poverty-reducing spending Benchmark 6: Poverty-reducing expenditures are clearly identified. General Assessment: (B) The benchmark is not met. Although the usage of HIPC II resources are well documented at both central and local level and the poverty-reducing expenditures are well defined, a proper functional classification or a budget tagging mechanism does not exist. Poverty reducing expenditures are defined in the Poverty Reduction Strategy Paper (PRSP), as follows: (i) the Ministry of Education s budget (spending on universities not included); (ii) the Ministry of Health s budget (spending on veterans pensions not included); (iii) expenditure on basic sanitation; (iv) urban and rural development expenditure; (v) other social spending by prefectures. However, there is no tagging of poverty-related expenditures in the budget, neither is it possible to identify expenditure by function, program or project. The poverty-reducing spending forecasts can not be obtained promptly by type of expenditure through a computer application. The process of estimating poverty-related expenditures is based on data from different sources, with the required aggregations being done manually, through a bridgingmechanism. DGC provides information on the Central Government s recurrent poverty 14

reducing expenditures during budget execution, based mostly on SIGMA s registers. Similarly, VIPFE provides information on public investments execution for the Central Government, singling out poverty reducing expenditures. UPF assesses 111 out of 314 municipalities spending, representing 75 to 80 percent of total expenditures a limited but significant coverage providing estimates on poverty reducing expenditure in the municipalities. UPF is also responsible for the final consolidation of this information. However, this is not done routinely nor produced on a timely basis there is at least a 5 to 6 months lag. HIPC II resources are monitored by placing them into different accounts in the Central Bank. The amount transferred to each municipality is available, and expenditures are monitored by relying on quarterly municipal reports. These resources are also identified in the central and municipalities budgets and budget execution reporting by the source of financing classification. Reports are done on each project that is financed in whole or in part by HIPC II resources. Indicator 7: Integration of medium-term forecasts Benchmark 7: Multiyear expenditure projections are integrated into the budget formulation process. General Assessment: (B) - Benchmark not met. No multiyear expenditure projections are integrated into the budget formulation process, neither at the central nor the local level. There is no Medium-Term Expenditure Framework (MTEF) in Bolivia, neither in the Central Government nor in local Government administrations. Multi-year projections are currently being done for the period 2004-2008 as part of the exercise for producing the medium-term framework for the PRSP. There are no tools to integrate these projections into the preparation of the budget law, to assess the long-term costs of current policies, or to forecast poverty-reducing spending. The Central Government budget formulation process is not based on a realistic macroeconomic framework. Fiscal restrictions are not incorporated; macroeconomic parameters have usually been proven to be excessively optimistic. The National Congress always increases the revenues estimations that come in the draft budget law, without technical support, in order to incorporate new expenditures to regions. To achieve the fiscal results intended for the year, MOF has to impose strict commitment and payment monthly limits to line-ministries during budget execution. The Government intends to improve this arrangement, probably by establishing in an earlier regulation the fiscal target and restrictions for the upcoming budget. Moreover, discrepancies between the national budget and the fiscal program agreed with the IMF are significant. 15

C. Budget Execution Indicator 8: Arrears Benchmark 8: Small stock of expenditure arrears, little accumulation of arrears over the past year. General Assessment: (C) This benchmark is not met. Although at Central level the amount of arrears is reasonable, at local Governments level, especially in the large municipalities, arrears constitutes an important problem. Arrears defined as the difference between accrued and paid expenditures are common at all administration levels. In the Central Government and Prefectures, arrears represented in each of the last three years less than 5 percent of paid expenditures, varying within a range of US$ (+/-) 10 million. At deconcentrated (including social security agencies) and subnational levels, arrears are much more significant, with values usually attaining far above 20 percent of paid expenditures. For the consolidated Government, in average, arrears represent slightly less than 10 percent of paid expenditures (Table 5). This is partly due to limitations in data recollection and processing. Recollection delays (approximately a 2-month lag) affect revenues as well as expenditures. Non recorded revenues are lost for the present administration and assigned to finance following year s expenditures. Floating debt defined as the difference between committed and paid expenditures is a problem at all Government levels. In the Central Administration only, floating debt reached about US $ 100 million (over 1 percent of GDP) in 2000. It is estimated that for the Central Government the current share of the floating debt to the total budget reaches about 2.5 percent. At the municipal level, the degree of severity of the problem varies considerably. Even among the municipalities that do not have arrears to the private sector, several have arrears to the National Regional Development Fund (FNDR) and other parts of Government. Some small municipalities have already reached significant amounts of floating debt. And yet for big municipalities, estimations are even higher: an external audit reports that payments in arrears (floating debt) of the three largest Municipal Governments only La Paz, Cochabamba and Santa Cruz reached Bs. 137 million in 1999. This situation has spawned concern among the Government, which is trying to provide solutions through the PRFs (municipal rationalization program). 16

Table 5. Budget arrears by selected line-ministries and local Governments 2001 2002 2003 Paid Arrears % Paid Arrears % Paid Arrears % Total /1 33984 2950 8.7 34937 3407 9.8 31636 2282 7.2 Central Administration 18792 456 2.4 22085 661 3.0 24286 789 3.2 Treasury 14475 269 1.9 17559 443 2.5 19861 554 2.8 Ministry of Government 755 44 5.8 782 42 5.4 824 44 5.3 Ministry of Education 297 39 13.1 282 15 5.4 221 7 3.4 Ministry of Defense 1066 0 0.0 976 69 7.0 1111 66 6.0 Ministry of Finance 283 16 5.7 183 10 5.6 181 12 6.6 Ministry of Health 213 8 4.0 236 5 2.0 262 17 6.6 Ministry of Agriculture 254 2 0.9 339 5 1.4 378 7 1.7 Legislative Branch 193 14 7.3 205 17 8.2 201 14 6.7 Judicial Branch 290 26 9.1 318 15 4.6 344 13 3.9 Rest of Central Adm. 967 37 3.8 1205 41 3.4 905 55 6.1 Deconcentrated agencies /1 2950 766 26.0 3018 1098 36.4 2667 376 14.1 Public finance sector /1 904 56 6.2 867 41 4.8 554 43 7.8 Social security agencies /1 975 204 20.9 1066 354 33.2 146 116 79.5 Public enterprises /1 2684 91 3.4 1154 109 9.4 902 45 5.0 Prefectures /1 4458 67 1.5 3922 128 3.3 1634 44 2.7 Municipalities /1 3221 1310 40.7 2825 1016 36.0 1447 869 60.1 1/ Note: Some decentralized institutions, prefectures and municipalities have not reported their financial statements to DGC; thus their information is not available. Indicator 9: Effectiveness of the internal control system Benchmark 9: Internal control is effective. General Assessment: (B) This benchmark is not met. Internal control is only partially effective. There is a uniform framework for internal controls, but it is unsatisfactory in terms of compliance, quality control of fiscal and financial reports, as well as of its influence on the improvement of efficiency and effectiveness of expenditures. The Internal Control System has a uniform framework of regulations and procedures. The internal control is managed by the Bolivian Regulations for Internal Control, promulgated by the Controller General Office (CGR). In 2001, CGR updated these regulations with the incorporation of the principles set forth in the guidelines of the Committee of Sponsoring Organizations (COSO). The internal control system covers almost all administrative units. Internal control of the compliance of expenditures with laws and regulations remains insufficient. An adequate internal control environment is lacking and there is a strong 17

resistance against changing the situation. The internal control environment is highly dependent on the understanding that regulations are meant to improve transparency and governance regarding expenditure management. In general, the degree of compliance is low, especially within Governmental institutions, where personnel turnover and instability as well as political intervention are high. Quality control of fiscal and financial reports remains insufficient. According to the SAFCO Law, the role of internal audit is to check the functioning of internal controls ex post. There are approximately 207 internal audit units with more than 700 auditors executing annually between 2,500 and 3,000 internal audits. The internal audit units report directly to the head of each agency, but not to the MOF. This is one factor that seriously hampers the effectiveness of internal audit. On the other hand, the implementation of recommendations is solely dependent on the head of each agency; there is no incentive to improve controls and implement recommendations. CGR periodically supervises and assesses the internal audit units (CGR carried out 45 assessments and 71 follow-up reviews in 2002. In 2003, 37 assessments and 49 follow-up reviews were completed). These assessments have shown that internal audit is weak; high levels of turnover and staff instability, low wages and lack of technical competency are all contributing factors to the ineffectiveness of internal audit. According to CGR reports, in 2001 and 2002, only about 25 percent of the recommendations have been implemented. Monitoring of the efficiency and effectiveness of expenditures is insufficient. The weaknesses observed in the management of public expenditures undermine control activities; it is not possible to assess, due to the high degree of budget variation, whether resources are being used for the intended ends, nor whether the resources are being used effectively. Other important instruments for an effective internal control are information and management systems. The implementation of SIGMA along with the implementation of the treasury single account (CUT) and the standardization of the chart of accounts for Central Government entities contributed to improving the system of internal controls. However, besides a few exceptions, no system wide error rates neither material error rates in routine financial documents are monitored. At the local level, the implementation of the internal control system is just being developed. An adequate legal framework for municipal accounts and their auditing is in place. The law requires municipalities and prefectures to report budget executions to the Central Government. In recent years those reports have improved, even though they are not always timely. Moreover, Bolivia has created an alternative form of social oversight to monitor the implementation of municipal investment projects, the Vigilance Committees (Comités de Vigilancia, CVs). Complaints made by these social vigilance committees can be used to stop transfers for a project that is not being handled correctly. The role and performance of those Committees have improved; however, they still need to be strengthened in terms of technical and institutional capacity. CGR, in coordination with the Ministry of Popular Participation, is now working On the legal instruments to allow CENCAP to provide training to the CVs. 18

While all large and a few medium sized municipalities have established internal audit units, internal audit capacity is generally lacking in small municipalities. According to CGR, only 38 of 314 municipalities have an internal audit unit. As in the Central Government, the implementation of internal audit recommendations is solely dependant on the executive (in this case, the mayor). Indicator 10: Tracking surveys are in use. Benchmark 10: Tracking surveys, where necessary, are in use to supplement internal control, but may not be a regular feature of the public expenditure management (PEM) system. General Assessment: (C) This benchmark is not met. Tracking surveys have not been used and the underlying PEM system cannot reliably track spending. For the time being, concrete actions which seek to estimate the amount of public money that is actually delivered through the budget system to front-line service delivery units are still in their infancy. At present there is no tracking of Central Government expenditure through surveys, although their introduction is planned at municipal level. At the local level, tracking surveys are sporadically conducted through Vigilance Committees. However, those are not qualified for this assignment. In particular, CVs do not audit budgetary execution accounts to verify the reliability of budgetary information. They only ensure that the Mayor is carrying out his previous commitments. This limited role is mainly due to lack of resources for financing their main activities; however, a Social Control Fund (Fondo de Control Social) has been established for the specific purpose of financing the CVs expenses related to their functions. The European Commission is planning to conduct a compliance test exercise, which will mainly focus on verifying compliance with procedures and internal control systems to guarantee that funds are being used for the intended purposes. Indicator 11: Quality of fiscal information. Benchmark 11: Frequent reconciliation of all Government bank accounts (those held at the central bank and the commercial banks) with the Government s accounting records. General Assessment: (B) This benchmark is not met. Although the central Government s bank accounts are reconciled regularly and in a timely manner with the Government s accounting records, the same does not happen in most municipalities. 19

For all Government bank accounts held at the central bank and at commercial banks a daily reconciliation is undertaken. This includes resources automatically transferred to municipalities. Bank accounts are reconciled daily both from balances of the above and below the line fiscal accounts. Executive reconciliation reports are available every day; complete reconciliation reports are available monthly. At the municipal level, reconciliation of all bank accounts including own revenues occurs satisfactorily and timely only in a number of municipalities. Bigger municipalities usually perform monthly reconciliation, which is informed a week later to DGC. In the smaller municipalities, reconciliation is not done in a satisfactorily and timely way; however the total amount transferred to small municipalities accounts is small. D. Budget Reporting Indicator 12: Regularity of internal fiscal reporting Benchmark 12: Fiscal reports of spending units are received by the central agency between two and four weeks after the end of the relevant period. General Assessment: (C) This benchmark is not met. Adequate financial information is not timely available. Monthly and annual public financial statements are prepared in a timely fashion by the Central Agency; however, monthly reports are incomplete and are not published. Monthly information is available only of the Central Government and institutions that manage their budget and accounting via SIGMA (of those institutions, financial reporting is available in real time). Under the national accounting framework, obligatory accounting financial reports include: Balance Sheet, Income Statement, Statement of Cash Flow, Statement of Changes in Equity, Budget reports on Execution of Revenues and Expenditures, and the report on Account Savings for Public Investment Financing. However, in 2002 these financial statements covered only a small percentage of the sub-national administration: only 17 percent of the sub-national Government units were in compliance with reporting requirements. In 2003 compliance has improved. The lack of standardization of accounting and reporting requirements for all levels of Government and the fact that Government accounts remain unconsolidated weaken the ability of the Government to use financial statements to inform of public policy decisions. Annually published financial reports contain only statistical information. The reports are not produced in accordance with recognized accounting standards, nor are prior year results presented, which does not allow for comparability of financial activity of the Government. In practice, DGC usually produces financial statements (unpublished) for the Central Government with a considerable delay after June 30. Consolidated annual financial statements for the public sector are not produced, in part due to the lack of financial reports 20

from municipalities and the decentralized sector. As such, there are considerable gaps in the annual financial statements of the Government, and, if let unresolved, they will not result in the publication of meaningful Governmental financial information. In practice, according to the authorities, now that several agencies have been computerized, this time lag has been reduced to less than 30 days. Based on a sample of reports sent by the Ministries of Education, Health, and Agriculture, a transfer period of between three and ten days has been estimated. Municipal accounting systems are weak; therefore reporting to Central Government is not always accurate. The only timely and reliable information on municipalities is prepared by the UPF which collects data of the 111 largest municipalities. Quarterly information is required to monitor the public sector deficit, but the UPF cannot provide municipal Government budget execution data it is too difficult for the UPF to put together and reconcile this information. Indicator 13: Fiscal reports track poverty-reducing spending. Benchmark 13: Good-quality classification of poverty reducing spending is reflected in the in-year budget reports. General Assessment: (C ) This benchmark is not met. Social expenditures are being monitored, but the monitoring is not entirely based on a functional classification of expenditures. Rough poverty reduction spending items estimates are obtained from the programmatic budget structure to identify poverty objectives and policies. Current information systems are not equipped for providing functional classification of the budgetary execution. SIGMA (implemented essentially at Central Government level) as well as the systems currently used in the majority of municipal Government and prefectures does not enable to obtain a functional classification of expenditures. Ad hoc functional classification reports are projected when requested; these reports are not published. The information on budgetary execution is only displayed and published with an institutional or economic classification. There is no tagging of poverty-related expenditures in the budget both at the national and sub-national levels of Government. Social expenditures for the overall public sector are estimated annually outside of any information system. The UPF monitors social expenditures with annual frequency. To do that, it employs information from DGC on current expenditures of the various institutions of the Central Administration, public investment data of the MOF as well as data about 111 municipalities and various decentralized institutions collected by the UPF itself. Estimates of current social expenditures are available only by institutional classification of expenditures. Estimates of capital social expenditures are available by functional and institutional classification. The estimation of current social expenditures include current 21