June 2006

Document of the World BankR

eport No. 34523-C

V

Cape Verde

The Challenge of Increasing Fiscal Space to M

eet Future Pressures

Report No. 34523-CV

Cape VerdeThe Challenge of Increasing Fiscal Space to Meet Future PressuresPublic Expenditure Review

PREM 4Africa Region

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

INE INFA INPS INTERBASE IT VAT MDGs ME MFA MFAP MH MIT MJIA MTEF NDP NIF NOS1 OECD PETROGAL PIP PRGF PROMOTORA PRSC PSI M E SDR SIGOF SOE SONANGOL STASCO TACV TdC TRANSCOR

WTO US-MCC

National Statistics Institute National Institute o f Financial Support to Agribusiness National Inst i tute for Social Protection Company for Sea Products Commercialization Information Technology Value-Added Tax Millennium Development Goals Ministry o f Education Ministry o f Foreign Affairs Ministry o f Finance and Public Administration Ministry o f Health Ministry o f Infrastructure and Transport Ministry o f Justice and Internal Administration Medium-Term Expenditure Framework National Development Plan Tax Identification Number Operational Nucleus for Information Systems Organization for Economic Cooperation and Development Petroleum Company o f Portugal Public Investment Program Poverty Reduction and Growth Facility Venture Capital Stock Corporation Poverty Reduction Support Credit Policy Support Instrument Financial Administration Reform Task Force Special Drawing Rights Financial Management Integrated System State-Owned Enterprise National Fuel Society o f Angola Shell International Trading and Shipping Company Limited Airlines o f Cape Verde General Audit Office Private Company o f Municipal Transportation U S Millennium Challenge Account World Trade Organization

Vice President: Gobind T. Nankani Country Director: Madani M. Tal l

Sector Director: Sudhir Shetty Sector Manager: Robert R. Blake

Task Team Leader: Manuela Francisco

... 111

Table of Contents

.. ACKNOWLEDGMENTS ....................................................................................................................... vi1

EXECUTIVE SUMMARY ..................................................................................................................... viii 1 . MACROECONOMIC TRENDS AND DEVELOPMENTS ............................................................. 1

A . Background ........................................................................................................................... 1 B . Robust Growth and Low Inflation ........................................................................................... 2 C . Remarkable Improvements in the External Position ................................................................. 4 D . Monetary Policy in Line with the Strengthening o f the PEG ..................................................... 5 E . Towards Fiscal Consolidation ................................................................................................. 7 F . 2006 Budget - A Move Towards More Transparency and Fiscal Discipline .................................. 8

2 . FISCAL PERFORMANCE ................................................................................................................. 9 A . Trends and Compositions o f Revenues .................................................................................... 9 B . Trends and Compositions o f Expenditures ............................................................................. 13 C . D . E .

The Efficiency o f Budget Implementation and the Role o f the Public Investment Program ....... 18 PRSP Priorities and the Budget ............................................................................................ 21 The Administrative Classification and the Ministerial Expenditure Patterns - An Analysis o f the Administrative Expenditure Patterns ........................................................ 23

3 . DEVELOPMENTS IN PUBLIC FINANCE REFORMS ............................................................... 27 A . B . Budget Preparation and Adoption ......................................................................................... 28 C . Budget Execution ................................................................................................................ 31 D . Accounting and Reporting .................................................................................................... 35 E . Finance o f Municipalities ..................................................................................................... 36 F . Cash Management ............................................................................................................... 37 G . Debt Management ............................................................................................................... 38 H . The Auditing System ........................................................................................................... 39 I . Procurement ........................................................................................................................ 41

4 . ANALYSING THE IMPACT OF ELIMINATING PETROLEUM PRODUCT SUBSIDIES IN CAPE VERDE ........................................................................................................................................... 42

A . Relevant Facts About the Petroleum Sector, Electra, and Regulation ....................................... 42 B . The Era Pre-Elimination o f the Subsidies .............................................................................. 44 C . The Magnitude o f the subsidies in 2005 and their Impact on the Budget .................................. 46 D . The Elimination o f the Subsidies on April 27, 2006 ............................................................... 47 E . The Impact o f the Elimination o f the Subsidies ...................................................................... 48 F . Impact o f the Elimination o f the Subsidies on the Budget ....................................................... 53 G . The Impact on Electra Sustainability and Recommendations .................................................. 55

5 . PUBLIC PENSIONS AND HEALTH INSURANCE SCHEME .................................................... 56 A . Background ......................................................................................................................... 56 B . Recent Reforms in the Pension System ................................................................................. 61 C . The Impact ofthe New Law ................................................................................................. 63 D . Options for Further Reforms ................................................................................................ 68 Bibliography ............................................................................................................................... 73

The Legal Framework .......................................................................................................... 27

iv

List of Annexes

Annex 1 . Fiscal Performance Technical No te .............................................................................. 75 Annex 2: Mitigating the Impact o f El iminating Petroleum Product Subsidies Technical No te ... 77 Annex 3: Public Pensions and Heal th Insurance Scheme Technical N o t e ................................... 89 Annex 4: Progress o n the CFAA Act ion Plan .............................................................................. 93

List of Tables

Table 1.1: Table 1.2: Table 1.3: Table 1.4: Table 1.5: Table 2.1 : Table 2.2: Table 2.3: Table 2.4: Table 2.5: Table 2.6: Table 2.7: Table 2.8: Table 2.9:

Basic Macroeconomic Indicators. 2000-2006 .............................................................................. 3 GDP . Aggregate Supply and Demand 2000-2006 (percentage o f GDP) ................................... 3 Balance o f Payments. 2000-2005 (US$ Mill ion) ......................................................................... 5 Selected Monetary Aggregates and Indicators, 2000-2005 (USD Mill ion) ................................. 6 Central Government Fiscal Operations, 2000-2006 (Percentage o f GDP) .................................. 7 Revenues, 2000-2006 (CVE million) ......................................................................................... 11 Revenues, 2000-2006 (Percentage o f GDP) .............................................................................. 11 Revenues, 2000-2006 (Percentage o f Total Revenues) .............................................................. 11 Revenues, 2000-2005 (Executed Revenues as Percentage o f Budgeted Revenues) .................. 12 Expenditures, 2000-2006 (CVE million) ................................................................................... 15 Expenditures, 2000-2006 (Percentage o f GDP) ......................................................................... 15 Expenditures, 12000-2006 (Percentage o f Total Expenditures) .................................................. 16 Overall Balance. 2000-2006 ....................................................................................................... 16 Central Government Transfer to Autonomous Institutes. 2000-2005 (CVE Mill ion) ............... 16

Table 2.10: Comparison o f the Original Budget with the Executed.Budget. 2000-2005 ........... ................ 19 Table 2.1 1 : Public Investment Program (PIP) . Execution and Sources o f Funds. 2000-2005 .................. 19 Table 2.12: Comparison o f Amended and/or Revised Budget with Executed Budget, 2000-2005 ............ 20 Table 2.13: PRSP and the Public Investment Program, 2004-2005 (CVE million) .................................... 23 Table 2.14: Administrative Classification o f Executed Expenditures, 2000-2005 ..................................... 25 Table 2.15: Administrative Classification o f Executed Expenditures, 2000-2005 ..................................... 25 Table 2.16: Administrative Classification o f Executed Expenditures, 2000-2005 .................................... -26 Table 2.17: Administrative Classification o f Executed Expenditures, 2000-2005 ........................ 26 Table 4.1: Freight “Premium” (US$/ton) .................................................................................................... 42 Table 4.2: Electra: Salient Statistics ........................................................................................................... -43 Table 4.3: Official Retail Prices .................................................................................................................. 44 Table 4.4: Petroleum Product Subsidies in 2005 (CVE Mill ion) ............................................................... -46 Table 4.5: Direct Government Subsidies to Electra for Gasoil ................................................................... 47 Table 4.6: Price Increases o f April 27, 2006 ............................................................................................... 47 Table 4.7: Access to Electricity and Water by Decile, 200 1-2002 .............................................................. 51 Table 4.8: Consumption o f Electricity and Water (Users Only) ................................................................. 52 Table 4.9: Change in the Per Capita Expenditure and Poverty Headcount, Before and

After Elimination o f Subsidies ................................................................................................... 53 Table 4.10: Projection o f Subsidies - In the Absence o f the Policy Change (CVE Mill ion) ...................... 54 Table 5.1 : Main Indicators o f Public Pension Systems (2004) ................................................................... 56 Table 5.2: Main Rules o f Public Pension Systems (2004) .......................................................................... 58 Table 5.3: Contribution Rates for Social Insurance (in percentage)’ .......................................................... 58 Table 5.4: Accrual Rates and Maximum Replacement Rates in Civil-Service Pension Schemes

in Selected Countries .................................................................................................................. 59 Table 5.5: Contribution Rates for Social Insurance Programs in Selected African Countries .................... 60 Table 5.6: Health Expenditures, 200 1-2004 (Percentage o f GDP) ............................................................. 61 Table 5.7: INPS: Total and Health Expenditures and Revenues, 1983, 1993, 1998, 2004 and 2005 ......... 61 Table 5.8: Financial Gap o f Reform Options .............................................................................................. 70

V

List of Fimres

Figure 1.1: Consumer Price Index. 2000-2006 ............................................................................................. 4 Figure 1.2: Evolution o f the Coverage o f Imports and Emigrants Deposits ................................................. 5 Figure 4.1: Gasoil Price as FOB ($/bbl versus CVEhiter) ......................................................................... 45 Figure 4.2: Indicators o f Targeting Performance ........................................................................................ 50 Figure 4.3: Targeting Subsidies per Type o f Fuel ....................................................................................... 52 Figure 5.1 : Impact on the Number o f Contributors and Pensioners o f the AP PensionProgram with Integration ................................................................................................................................................... 62 Figure 5.2: Combined Pension Expenditures o f AP and INPS Before and After Integration ..................... 64 Figure 5.3: Impact on the Number o f Contributors and Pensioners in INPS with Integration .................... 64 Figure 5.4: Impact on the Pension Program Balance o f the INPS with Integration .................................... 65 Figure 5.5 : Population o f Civ i l Servants from the Old System ................................................................... 66 Figure 5.6: Health Cost using INPS Parameters - Different Leakage Assumptions ................................... 66 Figure 5.7: Health Cost Using INPS Parameters - Different Leakage Assumptions, and 3 Percent Real Increase in Health Cost ............................................................................................................................... 67 Figure 5.8: Impact o f Integration on the Government Budget .................................................................... 68 Figure 5.9: Impact on the Pension Program Balance o f the AP with Integration ....................................... 68 Figure 5.10: Reform Options o f AP Pension System (Percentage o f GDP) ............................................... 70 Figure 5.1 1 : Effect o f Increasing Contribution Rates to the Current Balance o f the Pension Program in INPS ............................................................................................................................................................ 72

List of Boxes

Box 2.1 : Income. Import and Domestic Consumption taxes ....................................................................... 12 Box 2.2: The Budget Process. the Role o f Corrective Budgets and Extra-budgetary Expenditures ........... 17 Box 4.1 : Analyzing the Distributional Incidence o f Electricity Subsidies .................................................. 49

vi

ACKNOWLEDGMENTS

This Public Expenditure Review Update i s based on the findings o f two missions to Cape Verde, headed by Manuela Francisco (TTL, AFTP4). A preliminary version o f the Concept Note was discussed in October 2005 with the Authorities. Manuel Pinheiro, Director General for Planning and Claudino Semedo, Director General for Budget, were assigned by the Authorities to liaise with the World Bank team and coordinate with the Government team. The main f ield mission was conducted in February 2006. The World Bank worked in close collaboration with the Authorities to collect and analyze data. Other participants in the mission included Dorsati Madani (AFTP4) and Robert Cauneau (consultant). In the subsequent mission in M a y 2006, the TTL, in collaboration with the Authorities, focused o n gathering data for the Energy Chapter. During this mission the TTL benefited from the exchange o f information and discussions with the IMF team who were also o n mission.

The task team leader and main author o f this report i s Manuela Francisco (AFTP4). The peer reviewers are Michael Stevens (consultant), Demetrios Papathanasiou (LCSFE), Lucio Monari (LCSFE), Vladislav Vucetic (SASEI), Paul0 Correa (ECSPF) and Anita Schwarz (LCSFM). The report was prepared under the supervision o f Robert Blake, Sector Manager (AFTP4), who offered conceptual guidance, provided critical analytical advice and ensured quality control and management support. Madani Tal l (Country Director) and Iradj Al ikhani (Country Program Coordinator) supported the overall process and the engagement with the Authorities, providing valuable guidance o n pol icy issues. Background chapters were prepared by Dorsati Madani (Fiscal Performance), Robert Cauneau (Public Finance Management), Peter Meier (Energy chapter) and Montserrat Pallares and Kaizo Beltrgo (Pensions). The report benefited from inputs and discussions with sectoral colleagues, particularly Quentin Wodon, Robert Palacios, Stephane Legros, Samuel O’Brien, Soukeyna Kane, Bourama Diaite, and Helene Grandvoinnet. Michael Stevens, Demetrios Papathanasiou, Lucio Monari, and Anita Schwarz offered useful guidance while the document was being prepared. Iradj Alikhani, Sherri Archondo, Samuel O’Brien, Maitland MacFarlan, N i l s Maehle and Izabela Karpowicz provided u s e h l comments. Al ic ia Br i to (Ministry o f Finance Public Relations) provided excellent support to the missions and arranged al l the contacts with the Authoritities. Also, Eng. Antgo Fortes provided excellence assistance arranging the meetings with the o i l companies and the electricity company. Judite Fernandes (AFTP4) provided valuable assistance in document preparation.

vi i

EXECUTIVE SUMMARY

Main Macroeconomic Developments and Poverty Achievements

1. Cape Verde has been pursuing sound macroeconomic reforms. Starting in the early 1990s Cape Verde launched a far-reaching economic reform program, which, despite problems related to insufficient fiscal management early on, showed substantial progress in the period 1998-99 fol lowing an IMF Stand-by Arrangement to restore financial stability. During the period leading up to the elections o f early 2001, Cape Verde’s fiscal situation deteriorated substantially and economic activity slowed down, but starting in 2001, the Government embraced macroeconomic policies aimed at promoting macroeconomic stability. These included a Staff Monitored Program with the IMF in 2001 and a second agreement in 2002, and a Poverty Reduction and Growth Facility arrangement that ended in July 2005. The Government’s development strategy for the period 2004-2007 was set out in the Poverty Reduction Strategy Paper (PRSP), presented to the Board in 2004.

2. Cape Verde has experienced robust growth over the last years, consistently low inflation rate and a narrowing external current account deficit, under a responsive and consistent exchange rate and monetary policy. GDP growth averaged 5.7 percent over 2000- 2005, allowing for a significant improvement in the average standard o f living, as shown by the increasing GDP per capita, respectively averaging 3.6 percent over the period. Consumer inflation has been consistently below 2 percent a year since 2002, bolstered by the firm monetary pol icy and the exchange rate peg to the Euro. The peg has also fostered a steady inflow of capital, in turn reinforcing the credibility o f the peg. The external current account deficit has narrowed substantially from over 11 percent o f GDP in 2002 to approximately 4 percent in 2005. Monetary pol icy has focused o n strengthening the sustainability o f the exchange rate peg by steadily building up reserves and being responsive to signs o f sustainable trends in international reserves and prices.

3. As a result Cape Verde has reached middle-income status and i s on target to achieve most of the Millennium Development Goals (MDGs) before 2015. If GDP continues to grow at the same robust pace, poverty may be reduced to half o f i ts 1990 value by 2015. Some of the M D G s have already been reached, while others wil l gradually be reached, as a result of the poverty reduction strategy implementation.

Fiscal developments

4. Cape Verde has experienced substantial fiscal consolidation but further reforms are necessary. Improvements in both tax collection and expenditure control resulted in a reduction o f the fiscal deficit (including grants) from 15.6 percent o f GDP in 2000 to 5.6 percent o f GDP in 2001 and it has remained under control since then. This i s al l the more important, as in the absence o f an independent monetary pol icy under the f ixed exchange rate regime, the fiscal pol icy must bear the burden o f macroeconomic adjustment. In January 2004, further reforms were implemented, with the introduction o f the VAT, whose successful implementation more than compensated for the decline in revenues from trade taxes. Notably, unlike the overspending

... V l l l

associated with past elections, fiscal pol icy seems to have had a firmer stance in the run-up to the 2006 elections and has been consistent with the economic growth and poverty reduction objectives. Nevertheless, further reforms are needed for: (i) rationalization o f tax exemptions and incentives in order to widen the tax base and strengthen the overall revenue performance, and (ii) the streamlining and reduction o f import tar i f fs in order to promote productivity and competitiveness o f the economy.

The 2006 budget signifies a move towards greater transparency and fiscal discipline. The draft budget bill o f law was submitted to the Parliament in M a y 2006 and it is expected to be approved during the f i rs t semester o f 2006. The draft o f the 2006 budget law eliminates al l o i l and gas related subsidies. Furthermore, unlike previous years, the draft bill includes o i l product subsidies up to April 27, 2006, (date they were eliminated - an improvement for fiscal discipline), and provides for o i l subsidies accrued in 2005. The 2006 budget also includes the settlement o f o ld debts and past arrears.

Fiscal Performance

5. Revenue performance has been strong, reflecting the successful introduction of the VAT in 2004 and overall improvements in tax administration. Total revenues as percentage o f GDP increased from 26 percent in 2000 to 34 percent in 2005. The good revenue performance reflects mainly the introduction o f the VAT in 2004, which has contributed since then to 30 percent o f the tax revenues and improvements in tax collection. However, tax revenues can increase fbrther if tax exemptions and incentives are streamlined, as i t i s estimated that the foregone revenues account for 24 percent o f potential revenue. Furthermore, the elimination o f 1 percent o f the collections for the payment o f the collection services to the B C A (Banco Comercial At lh t ico) wil l also contribute to an increase in tax revenues. This i s o f critical importance as the country needs to create fiscal space to meet future pressures. This pressure is even more acute as external assistance in the form o f grants has been erratic.

6. Non-discretionary expenditures are very high and in the near future are expected to continue increasing. The components o f wages and salaries, transfers and subsidies and interest rates payments averaged 85 percent o f the total o f current expenditures in the last three years. These commitments are going to increase in the near future as a result o f the recent law approved on pensions, the recent heavy borrowing, and the new local finances law. In contrast, the recent pol icy decision o f eliminating subsidies will create some fiscal space.

7. The large share of the non-discretionary expenditures i s of more concern, given that the fiscal policy i s the only instrument available to conduct policy, as the exchange rate regime in place i s a peg to the Euro. In the context o f a large share o f non-discretionary expenditures and revenue collection close to i t s potential, the fiscal pol icy i s strongly limited. To allow some flexibility to the fiscal pol icy is critical as it is the only instrument at the disposal o f the pol icy makers to pursue economic pol icy objectives. T o that end, i t i s o f primary importance in the short term to constrain the increase in the wage bill by controlling the number o f c iv i l servants hired, and the wage increases and interest rate payments by reducing domestic borrowing. Also, it i s important to reduce the benefits associated with the contributory pension system, which is very generous by international standards, in order to reduce the transfers burden.

ix

8. I t i s still difficult to assess the alignment of the budget with the PRSP, as there i s only one year of implementation and data are still preliminary. Analysis o f the existing data indicates that the financial programming associated with the priority actions o f the PRSP document i s ambitious. The government achieved i t s budget related indicators on education and health in 2005, but budget allocation was not yet in line with the PRSP action plan.

Developments in Public Finance Reforms

9. summarized according to the fol lowing categories:

Progress regarding the Public Finance Reforms, including some difficulties, can be

Legal framework: As part o f the ongoing review, Cape Verde has embarked on deep reforms including the budget framework law, the budget planning law and the public chart o f accounts, leading to a comprehensive legal framework. These laws should be voted on by the end o f the year 2006 and adopted once approved.

Budget preparation: The implementation o f the MTEF has not been properly conducted undermining the effectiveness o f this instrument. The global MTEF has not yet been used as an instrument that conveys the overall fiscal strategy by disciplining public expenditure planning. Sectoral MTEFs do not observe ceilings regarding the PIP, set by the global MTEF, and as a result the global MTEF is revised accordingly. In part, this i s because o f the project finance approach o f the donors. This back-and-forth process hampers the implementation o f the MTEF and lessens the meaning o f this instrument. The MTEF should become the overcharging framework for public expenditure planning, pol icy making and annual budget execution. To this end, the global MTEF has to work as a top down envelope setting, with hard budget constraints that define line ministries spending plans.

0 Budget execution: The management o f expenditure and o f revenues has shown improvements, but further strengthening is needed. The Government lacks information regarding the execution o f externally financed projects. The MFPA should agree with donors on a framework for donor financed projects to make sure that they are properly recorded in the budget and the DGT should record in i t s books al l the public funds. Furthermore, sector ministries should be given authority to manage their respective budgets provided that financial control i s effective.

0 Accounting and reporting: The budget accounts are prepared in a timely manner, and since 2004 are sent o n a quarterly basis to the Parliament. However, the State general accounts are not sent yearly to the General Audit Office (TdC) and the newly designed public accounting system has not yet been implemented. The delay in the submission o f accounts to the TdC (for instance, the submission o f the 1998 accounts occurred in 2005) reduces the effectiveness o f the TdC assessment.

0 Cash management: The national treasury i s s t i l l managed o n a daily basis and the lack o f a cash f low plan prevents the smooth financing o f expenditures throughout the year, whereby recurrent expenditures crowd-out other expenditures, which eventually

X

become arrears. Dedicated software should be developed to have daily, weekly, monthly and annual accurate treasury plans.

Debt management: There is no adequate information on domestic debt, stock taking exercises are carried out infrequently and there i s not yet an IT link with the Treasury, limiting the ability o f the Treasury to program payments. Budget execution does not reflect the management o f the public debt since it i s not managed through the current accounting system. The DGT should implement a program that manages domestic public debt.

Auditing: The IGF and TdC have insuficient means at their disposal to fulfill their mandate and their recurrent appropriations continue to be limited. Furthermore, the process o f reform at the TdC has not yet begun, despite the draft o f the new organic law o f the TdC having been prepared since the end o f 2004. The laws on reorganization and competency o f the TdC should be finalized and submitted to the Parliament as soon as possible. Adequate fbnding to conduct their working program should be provided.

0 Procurement: Cape Verde does not have a consolidated legal and regulatory framework for al l public sector procurement and the preparation o f a new National Procurement Code i s ongoing.

Petroleum Subsidies

10. Petroleum product subsidies were eliminated. On April 27 the petroleum subsidies were eliminated, which was an important step to prevent the accumulation o f contingent liabilities and quasi-fiscal liabilities. This pol icy measure i s commended as these subsidies were generally badly targeted, considering the population as a whole, with the exception o f kerosene, and had a strong impact on the budget, especially in the current context o f high o i l prices. In 2005 petroleum subsidies amounted to C V E 1.73 bi l l ion (approximately 2 percent o f the GDP). There were four major subsidies: (i) the subsidies payable to the o i l companies on account o f the retail prices for subsidized fuels (butane in small bottles, kerosene, and fueloil); (ii) the equalization payments to the o i l companies o n account o f notionally unsubsidized fuels (mogas, butane in large bottles, gasoil), but for which at times o f 'persistently rising prices (as has occurred over the past 2 years) also require adjustment to cover o i l company losses; (iii) the subsidy paid to Electra to cover the difference between the gasoil purchase price from the o i l companies and C V E 37. Mi te r ; and (iv) the implici t subsidy that i s implied by an electricity tariff that does not provide adequate returns for Electra to meet its long-term investment requirements. The latter i s not included in the estimated C V E 1.7 billion.

11. Despite the increase in tariffs, Electra remains an unviable company. The tariff increases announced on June 1,' 2006, (23 percent on electricity and 13 percent o n water) to compensate for the elimination o f the subsidies will l ikely be insufficient and thus viability o f the company may be at risk. Depending on the level o f wor ld o i l prices, the tar i f f increases required

x i

to offset the fuel price increases are in the range o f 7 to 17.6 percent'. I t is estimated that the elimination o f the subsidies will free up approximately 5 percent o f the resources. However, if reforms are not undertaken to make the energy sector viable, the implici t subsidy in the form o f tariff deficit or recapitalization o f the company will persist and in that case only 2 percent o f the resources will be released.

Pension and HeaIth Reforms

12. Integration of the two pension systems i s a step in the right direction, but further reforms are necessary to move the INPS towards sustainability, which can otherwise potentially be a large contingent liability for the government. The integration o f c iv i l servants into the INPS has been under discussion for a few years and in February 2006 a new integration law was approved to be implemented when the budget wil l be approved in 2006. Even though, in consolidated terms, there i s a gain in moving c iv i l servants to the INPS system, as it provides fewer benefits, the integration worsens the unsustainability o f the INPS. Preliminary estimates indicate that the I N P S system will be in deficit around 2040, as the system matures and the demographic situation improves. N o w is the time to proceed with further reforms in order to avoid a deficit in 2040. Reforms must consider either a raise in the contribution rates or a substantial cut in benefits. Estimations o f the impact o f a number of reforms have been undertaken in this PER. Regarding the AP (Public Administration) there is a possibility o f making the transition less costly by a combination o f a lower accrual rate, later retirement, and price indexation o f benefits. Regarding the INPS, one possibility would be increasing the contribution rates at an earlier point than when reserves are exhausted.

At crude o i l price levels that prevailed in 2005 (around $70/bbl) it i s estimated that the tariff increase required i s 17.6% (in addition to a 10% increase required to bring the tariff to a level that provides an adequate financial return). If o i l prices returned to the levels o f 2004 (at around $50/bbl), the required tariff increase would be significantly lower. Th is i s explained in detail in Section 4: clearly these calculations are further affected by the future evolution o f the $US/Euro/CVE exchange rates.

x i i

Summary of Recommended Reform Measures

12

to 3 years

Within months

months Reform Measures 12

FiscaVMacroeconomic Stability Increase foreign exchange coverage to further strengthen the X X credibility o f the peg Reduce domestic borrowing X X Control mandatory expenditues X X Fiscal Performance Rationalize the currently fragmented system o f tax incentives and exemptions Streamline import tariffs X Progress towards alignment o f the budget towards PRSP objectives S i g n the protocol vith B C A Public Finance Reforms Completion o f the legal Framework Strengthen the relation between MTEF and budget preparation Strengthen the capacity o n MTEF preparation New accounting system implemented X Clearance o f arrears X X Record domestic debt o n the CSDRMS Develop software to have daily, weekly, monthly, and annual treasury plans Law on competences o f the TdC submitted to the Parliament Clearance o f the backlogs o f public accounts by the TdC Energy Secbr Increase electricity and water tariffs further as suggested by ARE model Application o f the adjustment mechanism for petroleum products as defined

electricity Implement measure to alleviate the impact o f the elimination of the subsidy on kerosene Pensions Further progress with pension reform

X

X

X

X X X

X X

X X

X

X

X

X

Implementation o f the adjustment mechanism for water and

X

... X l l l

1. MACROECONOMIC TRENDS AND DEVELOPMENTS

A. BACKGROUND

1.1 Starting in early 1990s Cape Verde launched a far-reaching economic reform program. Afler the f i rst multiparty elections in 1992, Cape Verde pursued the liberalization o f the country’s economy, aiming to reduce the size o f the public sector and promote growth led by private sector exports. The reform agenda focused on privatization, trade liberalization, financial sector restructuring, and tax reform. Until 1997, the liberalization was conducted without sufficient restraint o n the fiscal policy, and as a result large macroeconomic imbalances emerged from the accumulation o f substantial domestic public debt and the depletion o f foreign exchange reserves. In 1998 the authorities successfully adopted a program supported by an IMF Stand-by Arrangement to restore financial stability.’ Substantial progress followed in the period 1998-99, with real GDP growth accelerating, as a result o f tourism development, foreign investments in the export-oriented manufacturing sector and inf lows o f workers’ remittances.

1.2 In 2000, Cape Verde’s fiscal situation deteriorated substantially and economic activity slowed down. During the period leading up to the elections o f early 2001, the structural reform agenda stalled, while the progress in macroeconomic stabilization suffered a turnaround. In particular the fiscal situation deteriorated considerably, as a result o f restructuring costs associated with the privatization o f commercial banks, and the petroleum price subsidies (9 percent o f the GDP). As the net foreign financing available was largely insufficient, the deficit was largely monetized, while domestic and external arrears accumulated. The recourse to domestic bank credit was extensive (8 percent o f GDP) and the accumulation o f arrears reached 7 percent o f GDP. International capital flows, including donor support, fel l sharply and international reserves declined substantially.

1.3 Starting in 2001, in the aftermath of major macroeconomic imbalances, the Government embraced policies aimed at promoting macroeconomic stability. As part o f the recovery plan, the Government agreed on a Staff Monitoring Program with the IMF in 2001, and as a result o f significant progress in stabilizing public finances, regularizing domestic and external arrears, and rebuilding international reserves, the Government signed a second agreement in 2002 - a three-year Poverty Reduction and Growth Facility (PRGF) for SDR 8.64 mi l l ion (about US$11 million). The PRGF arrangement ended in July 2005. Quantitative performance criteria under the PRGF-supported program were met with a substantial margin and most structural objectives were also met. An IMF team held discussions with the authorities in M a y 2006 on a 3-year program that could be supported under the Policy Support Instrument and it i s expected that the proposed program wil l be presented to the IMF Board by end-July 2006. A sound macroeconomic framework, including debt sustainability and adequate foreign exchange cover, and the development and strengthening o f the financial sector could form the pillars o f the proposed program. Cape Verde has recently signed a compact with the Mi l lennium Challenge Corporation, which will provide grant financing o f around U S $ l 1 0 mi l l ion (over 10 percent o f

’ Between February 20, 1998 and March 15, 2000, Cape Verde’s macroeconomic stabilization and structural reforms were supported by a precautionary Stand-By-Arrangement o f SDR 2.5 million.

1

GDP) spread over the next five years. It has also been selected to apply for funding from the U S Millennium Challenge Account (MCC) for 2006 as part o f a new lower middle income category o f eligible countries. The Fitch rating agency recently assigned a rating o f B+ to Cape Verde.

B. ROBUST GROWTH AND LOW INFLATION

1.4 The Government’s development strategy for the period 2004-2007 was set out in the Poverty Reduction Strategy Paper (PRSP), presented to the Board in 2004. Building up on the Interim Poverty Reduction Strategy Paper,3 which in turn was underpinned by the National Development Plan, the PRSP i s part o f a broad strategy underway in Cape Verde since the beginning o f the decade to promote growth and reduce poverty. The strategy focuses o n five key pillars: (i) promoting good governance; (ii) improving competitiveness and private-sector-led growth; (iii) fostering human and capital development; (iv) strengthening social security and solidarity; and (v) improving infrastructure and land use management.

1.5 Cape Verde has experienced robust growth over the last years (GDP growth averaged 5.7 percent over the period 2000-2005). In 2004, the rate o f growth stood at 4.4 percent, 1.1 percentage points below the PRSP projection^.^ The short-lived slow down in 2004 was the result o f a less favorable agricultural year (hit by a locust infestation in one region, as wel l as a drought), a l o w level o f public investment (administrative delays affecting construction growth) and slightly lower tourism as a result o f external developments. The forecast for 2005 looks substantially stronger, with growth in 2005 projected to reach about 6.3 percent, slightly above the rate envisaged in the PRSP. The strong performance in 2005 reflects a higher execution rate o f public investment and an increase in private investment. Estimations for 2006 indicate that the Cape Verdean economy i s expected to experience a s l ight slowdown in economic growth, but s t i l l maintain a GDP growth o f 5.5-6.0 percent. L o w execution o f the public investment program (as the 2006 Budget L a w wil l be voted in June 2006 only) and the pass-through effect o f petroleum product prices are the main reasons for the expected decrease in the rate o f growth. The medium term outlook (2007-08) is optimistic. GDP growth i s expected to stand at 6.5-7.0 percent on average over that period, as envisaged in the PRSP. Growth i s projected to be led by private investment in tourism, public investment in infrastructure, and private consumption.

1.6 The high growth rate has allowed for a significant improvement in the average standard of living, as shown by the increasing GDP per capita, which averaged 3.6 percent over the period 2000-2005. A s a result, Cape Verde has reached middle-income status and i s on target to achieve most o f the Mil lennium Development Goals (MDGs) before 2015. If GDP continues to grow at the same pace, poverty may be reduced to ha l f o f its 1990 value by 2015. Some o f the M D G s have already been reached, while others wil l be gradually reached as a result o f the poverty reduction strategy implementation.

Presented to the Board in April 2002. PRSP projections for real GDP growth rate were as follows: 5.5, 6, 6.5, and 7.0 percent for 2004, 2005, 2006, and 2007,

respectively. For more details, see Annex 1.

2

Table 1.1 : Basic Macroeconomic Indicators, 2000-2006 Indicator 2000 2001 2002 2003 2004 2005 2006l Real GDP growth (annual percentage change) 7.3 6.1 5.3 4.7 4.4 6.3 5.5 Real GDP per capita growth (annual percentage change) 3.8 4.3 3.4 4.3 2.5 3.8 3.5 Overall balance, including grants (percentage o f

Current account balance, excluding official transfers

Current account balance, including official transfers

Reserve coverage (months o f current year imports and NFS) 1.0 1.5 2.0 1.9 2.6 3.4 3.3 Broad money growth (annual percentage change) 12.9 9.8 14.3 8.6 10.5 15.5 15.4

'Preliminary estimates. Source: IMF, Ministry o f Finance and Public Administration and Staff estimates.

GDP) -15.6 -5.6 -2.3 -3.0 1.5 -2.6 -8.9

(percentage o f GDP) -15.3 -13.9 -17.1 -15.5 -12.4 -9.8 -14.0

(percentage o f GDP) -11.2 -10.1 -11.4 -9.5 -6.7 -4.3 -6.5

CPI annual average (annual percentage change) -2.4 3.7 1.9 1.2 -1.9 0.4 6.2

1.7 Cape Verde's i s predominantly a tertiary sector economy with services representing 75 percent of GDP on average over the period 2000-2005. Cape Verde has no known natural resources and its growth strategy is underpinned primarily on the development o f the tourism industry, the offshore financial center and the regional transport hub. The tourism industry represented 11 percent o f the GDP on average over the period 2003-2005 and i s expected to increase in importance due to large investments initiated in 2005: new hotels in Sal, Santiago and Boa Vista, a new international airport in Boa Vista and Praia, and the upgrading o f the airport in Sa0 Vicente. The primary sector does not represent more than 8 percent and industry (light) stands for about 17 percent o f GDP. The export sector has performed above expectations. Despite negative external developments and high o i l prices, exports increased in 2005 by 23 percent in absolute termsS5 Imports are historically very high. Remarkably, however, the trend has been downward after a peak in 2002. Despite the demand for external investment goods, imports in 2005 increased by 9 percent in real terms only.

Table 1.2: GDP - Aggregate Supply and Demand 2000-2006 (percentage of GDP)

Component 2000 2001 2002 2003 2004 2005 2006l

Primary 9.0 8.4 7.6 7.4 7.2 6.8 6.7

Tertiary 74.1 75.6 75.4 76.1 75.7 75.5 75.2

Consumption

Supply

Secondary 16.9 16.0 17.0 16.6 17.2 17.7 18.1

Demand

Government 18.9 17.3 18.4 20.1 19.9 22.9 22.7 Households 86.7 88.2 88.5 86.5 80.2 74.8 73.1

Gross capital formation 30.7 31.7 35.8 31.1 33.9 38.5 40.7 Resource balance

Exports 20.9 18.9 20.9 14.6 15.2 17.0 16.8 Imports -57.3 -56.0 -63.7 -52.1 -50.9 -53.3 -53.3

'Preliminary estimates.

' The good performance o f exports is explained by an increase in the export o f services (17 percent) and goods (55 percent). Among the goods the export o f fresh fish (200 percent) is noteworthy as a result o f the end o f the European Union (EU) decision to block Cape Verdean exports. Regarding the export o f services, services related to tourism and air transportation continue to play the most important role. Among services there i s a remarkable increase in o i l re-exports (72 percent), largely explained by the opening o f the new airport in Praia.

3

Source: IMF. 1.8 Consumer inflation has been consistently below 2 percent a year since 2002, bolstered by a firm monetary policy and a m exchange rate pegged to the Euro. Prices fe l l 1.9 percent in 2004, mainly because o f reductions in the price o f food, beverage, and clothing throughout the year, and the reduction in customs tar i f fs at the beginning o f the year, which have allowed importers to draw on cheaper and more diversified sources o f supplies. These price reductions helped to offset the impact o f higher prices o f o i l imports. In 2005 inflation stood at 0.4 percent. This increase mainly reflected food price developments: inflation was only 1 percent excluding food over this period, and 0.4 percent excluding food, energy, and water.6 The upward trend is expected to continue in 2006-2007, with inflation reaching approximately 6 percent, reflecting the end o f regulated prices in petroleum products and elevated food prices. Fol lowing this adjustment period, prices over the medium term are expected to stabilize at around 2-3 percent, in l ine with the Euro area, the Euro being the anchor currency. Overall, as a result of l o w domestic inflation the real effective exchange rate has been depreciating since early 2003. During 2005, the real effective exchange rate depreciated by approximately 13.5 percent.

Figure 1.1 : Consumer Price Index, 2000-2006

1m.w

170.W I 2000 2w1 2w2 2003 2w4 2ws 2wB

-CPI - - - - - CPI excluding energy and water

Source: IMF Note: Index level 1989=100 Source: IMF

c. REMARKABLE IMPROVEMENTS IN THE EXTERNAL POSITION

1.9 The peg to the Euro, while avoiding exchange rate uncertainty has fostered a steady inflow of capital, namely remittances, emigrants’ savings, and foreign direct investment, which in turn has reinforced the credibility of the peg. The choice o f the fixed exchange rate regime has served Cape Verde well. The appreciation o f the Euro (and hence CVE) against the dollar does not appear to have affected Cape Verde’s external competitiveness, reflecting the country’s relatively stronger links to Europe with respect to trade, tourism, and remittance flows. The anchor currency, the Euro, is also the currency o f i t s main trading partners (76 percent o f exports, 58 percent o f imports, and more than 50 percent o f total remittances). The sharp

The CPI used for the computation o f the figures presented here makes use o f a 1989 basket, which may not provide a fully accurate representation o f actual inflation. During 2006, INE is expected to begin using a new CPI methodology, including an updated basket.

4

increase in o i l prices in 2004 and 2005 could have had a much larger impact if i t had not been partially offset by the appreciation o f the Euro against the US$.'

Figure 1.2: Evolution of the Coverage of Imports and Emigrants Deposits

Gross reserve coverage in months of current year imports

Source: IMF.

1.10 The external current account deficit has narrowed substantially over recent years, falling from over 11 percent of GDP in 2002 to approximately 4 percent of GDP in 2005 (Table 1.1). The improvements in the external position reflect primarily increasing revenues from tourism, sustained remittance flows (representing approximately 20 percent o f the GDP), effective demand management and a more diversified import sources base.

Table 1.3: Balance of Payments, 2000-2005 (US$ Million)

Component 2000 2001 2002 2003 2004 2005 Trade balance -191.6 -194.4 -236.2 -289.2 -332.4 -354.2

Imports -230.1 -23 1.6 -278.1 -341.9 -390.6 -426.0 Exports 38.4 37.1 41.9 52.7 58.2 71.7

Current account balance Excluding official transfers -82.4 -78.0 -106.5 -126.0 -1 14.8 -96.4 Including official transfers -60.5 -56.9 -71.0 -77.2 -62.1 -42.4

Capital and financial account 40.3 77.5 96.7 63.2 94.9 93.9 Overall balance -39.3 15.4 26.9 -7.1 34.9 51.5

Source: IMF.

D. MONETARY POLICY IN LINE WITH THE STRENGTHENING OF THE PEG

1.1 1 Monetary policy has focused on strengthening the sustainability of the exchange rate peg to the Euro by steadily building up reserves. Cape Verde has lost most o f its monetary independence with the adoption o f the peg, having very l i t t le room for discretion (depending exclusively o n capital account controls). Monetary pol icy is, therefore, subordinated to the target level o f international reserves deemed appropriate to support the peg. The B C V has successfully fulfilled its pol icy objective as reserve coverage has been strong, growing from 1 .O to 3.4 months o f current year imports from 2000 to 2005. Furthermore, in 2002 the B C V implemented a new organic law, further strengthening confidence in the exchange rate pol icy

' In 2005, due to the stabilization o f the US$/Euro exchange rate, the impact o f the sharp increase in o i l prices had a more pronounced effect on the imports bill.

5

choice by virtually eliminating scope for deficit monetization. By this law, the government i s required to settle any outstanding balance on i t s overdraft account in the B C V by year-end.'

1.12 Overall, the monetary policy has been responsive to signs of sustainable trends in international reserves and prices. In 2003, as the expansion of the credit to the economy exceeded the programmed, the BCV adopted a tighter policy by raising the reserve requirements by 1 percentage point and started issuing i ts own securities in order to manage liquidity. The monetary tightening in 2003 produced results and in 2004 there was a substantial increase in foreign reserves and a reduction in prices. As a result, starting in December 2004, the B C V adopted a looser monetary policy. The required reserve ratio decreased 1 percentage point in December 2004 and again in February 2005 (to 17 percent) and the standing lending facility rate decreased 1 percentage point (to 7.5 percent). These two measures were followed by a decline in the commercial banks' lending rates and on the T-bills which in October 2005 reached historically l o w levels, 2.5 percent (for a l l types o f maturity).

1.13 Monetary policy has also been consistent with low inflation. Overall, broad money and credit to the economy have been growing at a rate that supports the overall increase in economic activity (approximately 18 percent on average over the period 2001-2005). However, more recently, mostly in 2005, large unsterilized inf lows o f capital have led to an excess o f liquidity and to acceleration in broad money gr~wth.~ Even though the multiplier effect has been small (as commercials banks have been risk adverse and thus lending moderately), the B C V has intervened more actively in 2006, to prevent capital inf lows leading to inflationary pressures and deterioration o f foreign reserves.

Table 1.4: Selected Monetary Aggregates and Indicators, 2000-2005 (US$ Million) Indicator 2000 2001 2002 2003 2004 2005 Net foreign assets 59.6 77.6 96.7 107.0 155.2 218.3

Foreign assets 77.2 93.2 118.1 137.5 190.4 258.5 O/w: foreign reserves 28.0 42.5 71.6 83.6 127.2 177.6

Net domestic assets 288.8 294.1 350.2 475.2 552.7 584.3 Net domestic credit 344.1 357.8 426.0 557.8 645.4 678.5

Net claims on General Government 181.3 175.8 211.6 261.6 286.2 306.3 Credit to the economy 162.8 182.0 214.4 296.2 356.2 372.3

Foreign liabilities -17.6 -15.7 -21.4 -30.5 -35.2 -40.1

Other items (net) -55.3 -63.7 -75.9 -82.7 -92.7 -94.2

Broad money (M2) Narrow money (Ml)

Demand deposits of which emigrant deposits

Quasi-Money Time deposits

of which emigrant deposits

of which emigrant deposits Foreign currency deposits

Emigrant deposits

348.4 371.6 446.9 582.2 707.9 802.6 168.2 168.5 193.1 235.3 275.0 311.8 114.2 114.1 137.9 168.6 198.9 234.7 11.6 14.3 18.2 22.5 29.2 33.4

161.8 186.1 235.9 324.1 402.7 456.6 151.1 176.9 220.7 306.9 386.4 438.1 83.7 102.8 135.0 193.6 242.3 277.5 18.4 17.1 17.9 22.8 30.2 34.2 9.0 9.3 9.3 10.6 11.0 12.6

104.2 126.5 1 6 2 3 226.7 282.5 323.5

Money multiplier (M2NO) 3.1 3.1 3.2 3.3 3.4 3.3 Money velocity 1.6 1.6 1.5 1.5 1.4 1.3 Source: IMF.

The overdraft allowed i s 5 percent o f the revenues o f the previous year. Mostly reserve requirements and interest rates on the deposit and lending facilities.

6

E. TOWARDS FISCAL CONSOLIDATION

1.14 With the fixed exchange rate regime and thus in the absence of an independent monetary policy, fiscal policy bears the burden of macroeconomic adjustment. As Cape Verde i s a small economy with a high degree of openness (70 percent)" an expansionary fiscal pol icy leads invariably to an increase in imports and consequently to a widening o f the external current account, which in turn i s reflected in a decrease in the level o f foreign reserves. Therefore, in spite o f the steady inflows o f external financing, fiscal pol icy has to be in l ine with the monetary policy, which is subordinated to the target level o f international reserves deemed appropriate to support the peg.

1.15 Cape Verde's fiscal position i s today remarkably stronger than in 2000. From a fiscal deficit (including grants) of 15.6 percent of GDP in 2000, the deficit was reduced to 5.6 percent o f GDP in 2001 and has remained under control since then (Table 1.5). The substantial fiscal consolidation i s the result o f improvements in both tax collection and expenditure control. Total expenditures dropped from 41 percent o f GDP in 2000 to 32 percent in 2001. Over the period 2000 to 2005, expenditures stabilized at around 33 percent o f GDP. Total revenues, excluding grants, have increased steadily from 20 percent o f the GDP in 2000 to 25 percent o f GDP in 2005.

Table 1.5: Central Government Fiscal Operations, 2000-2006 (Percentage of GDP)

Indicator 2000 2001 2002 2003 2004l 2005l 2006l Revenues (inc. grants) 25.67 26.76 32.03 28.43 33.49 30.43 32.22 Expenditures 41.30 32.33 34.35 31.42 31.96 33.02 41.20 Overall Balance (excl. grants) -21.72 -11.60 -11.00 -8.13 -9.31 -8.50 -16.12 Overall Balance (inc. grants) -15.63 -5.56 -2.32 -2.99 1.52 -2.59 -8.97 Foreign (net) 4.22 2.38 1.39 1.79 1.56 2.10 3.48 Domestic (net) 11.41 1.23 0.93 1.19 -0.01 0.49 5.50

'Preliminary estimates. Source: IMF, Ministry o f Finance and Public Administration and Staff estimates.

Primary Balance -14.05 -3.64 0.65 -0.48 4.01 -0.33 -8.97

1.16 Even though the public expenditure trend indicates a substantial improvement in the fiscal position overall, the increase in the deficit in 2003 represents a small set back. In 2003, spending increased above the programmed level and was not off-set by a similar increase in revenues. This was reflected in an increase in the interest rates on BT (which registered the first increase since 2001) and in the decline o f foreign reserves. On the other hand, the fiscal surplus stood at 1.5 o f GDP in 2004, which i s attributable to the stronger-than-expected performance o f the VAT and the lower-than-expected wage bill and execution rate o f public investment. In 2005, the fiscal deficit increased as a result o f a r ise in expenditures related to c iv i l servants (implementation o f the medical career, payment o f arrears related to bonus for teachers), and clearance o f arrears. The increase in expenditure, not matched by an equivalent increase in revenues because o f the delay in the disbursement o f budget support grants, coupled with a decrease in the donor inflow, led to a worsening o f the fiscal position by 2.6 percent.

lo Openness measured as the indicator o f [(Exports + Imports)/GDP].

7

1.17 Deficits have been largely financed by foreign financing. With the exception o f the year 2000, when the deficit was largely monetized, the deficit has been largely financed with foreign financing in concessional terms. In 2001 and 2002, foreign financing accounted for 63 percent o f the financing. In 2004 it dropped to 40 percent and in 2005 increased to 80 percent. Data for 2004 and 2005 are, however, st i l l preliminary estimates.

F. 2006 BUDGET - A MOVE TOWARDS MORE TRANSPARENCY AND FISCAL DISCIPLINE

1.18 The 2006 budget draft bill of law was submitted to the Parliament in May 2006 and it i s expected to be approved during the first semester o f 2006.” The budget draft bill o f law was not submitted to and voted o n in the Cape Verdean Parliament in 2005, because that period corresponded to the one immediately before the legislative and presidential elections (January 22 and February 12, 2006, respectively). This was a common practice in the institutional politics o f Cape Verde,” which wil l be replaced once the draft budget law i s approved. The rationale for this practice was to reserve, with democratic legitimacy, the opportunity to review and approve the budget for the f i rst year o f the term that takes place immediately after the elections for the new Parliament. Until the 2006 budget i s approved the 2005 budget will remain in ~ 1 a c e . l ~ Budget execution shall follow the normal ru le o f spend appropriations established in the expenditure charts by twelfths.

1.19 The draft of the 2006 Budget Law includes measures to strengthen fiscal transparency. The 2006 budget eliminates petroleum subsidies, which were largely badly targeted. Also, unlike previous years, the draft o f the 2006 Budget L a w includes petroleum subsidies for 2006 (until April 27, as they were eliminated at that date) and petroleum subsidies accrued in 2005. This is a remarkable improvement compared to the past practice, where subsidies o f one year were carried forward to the next years’ budget as a result o f the practice o f accounting for expenditures only when payment was made (the payment o f the subsidies was made against taxes). That practice seriously understated the current fiscal position, and thus the budget was not providing adequate support for planning and monitoring current year spending priorities against current year resources. Another important measure to improve transparency was the provision for 35 percent o f the arrears in the budget. A cross-debt study l4 that took place in 2004 identified 5.9 bi l l ion o f net debt o f the central government. This study was revised in 2005 (with regard to the central government debt) and a plan was prepared, whereby the government agreed to settle C V E 4.8 bi l l ion (approximately 5.5 percent o f the GDP) through a three-year plan, beginning in 2006. By the end o f 2008 al l recognized arrears wil l be cleared and no new anears will have been created, with the exception o f the tariff deficit to E1e~tra. l~

1.20 The draft budget for 2006 indicates an increase in current expenditures given the inclusion of the petroleum subsidies and clearance of arrears which will be supported by privatization proceeds and land sale. According to the draft budget for 2006 expenditures are

As per article 21, numeral (5) of the Cape Verdean Framework Budget Law (FBL), the new assembly votes on a new budget within 60 days o f taking office. l2 As per article 21 o f the Cape Verdean FBL.

As per Article 21, numeral (3) o f the FBL. l4 I t analyzed the debt between the Central Government, municipalities, autonomous institutes, and para-statal. The study identified 1 1.6 billion in cross debts, of which 5.9 billion are net debt o f the Central Government. l5 The payment o f the tariff deficit i s scheduled for the next 5 years, starting in 2006.

8

expected to stand at 40 percent o f the GDP. This jump reflects mainly the increases in the categories o f transfers and subsidies (which rose by 141 percent), PIP (by 56 percent) and to a less extent salaries and benefits (23 percent)). Transfers and subsidies include the increase in the petroleum subsidies (which rose by 433 percent), pensions and transfers to the municipalities (a new local finances law established a revised formula, increasing the percentage o f government resources transferred from 7 to 10 percent). While capital expenditures wil l materialize according to the availability o f financing, mandatory expenditures such as salaries and benefits and transfers need definitively to be paid. In contrast, revenues are expected to increase by 22 percent only, and the fiscal deficit would enlarge and achieve 9 percent o f the GDP. The draft budget for 2006 indicates that 41 percent o f the deficit wil l be financed with privatization proceeds and land sales. According to the authorities, accounting privatization proceeds and land sale are reported as revenues and not as financing.16 The risk o f this procedure i s that i t may encourage the expansion o f long-term expenditures when privatization proceeds and land sales are punctual “revenues”, in part because Cape Verde has already privatized the majority o f the public companies.

2. FISCAL PERFORMANCE

2.1 The data used to undertake the analysis was obtained from the authorities. They provided the finalized books for the years 2000-03 in February 2006. Data for 2004 and 2005 are st i l l estimates. The methodology used in presenting the data is similar to that o f the IMF.

A. TRENDS AND COMPOSITIONS OF REVENUES

2.2 The share of total revenues to GDP for the period 2000-2005 shows an upward trend, with some declines in 2003 and 2005. Total revenues as percentage o f GDP increased from 26 percent in 2000 to 30 percent in 2005. They averaged 30 percent over the whole period. While the trend was upwards over the al l period, 2003 and 2005 register declines (5 percentage points in 2003 and 3 percentage points in 2005). The worsening o f the revenue performance in both years was largely due to a reduction in external grants and, to a certain extent, a slower growth o f tax revenues, especially in 2003. External grants account, o n average, for 24 percent o f the total revenues, with the minimums o f 18 percent in 2003 and 20 percent in 2004.

2.3 Supported by two comprehensive reforms in indirect taxation introduced in early 2004 (tariff reforms and introduction of the VAT), tax revenue performance has been strong, Tax revenues show a fast rate o f growth as percentage o f GDP from 2000 to 2002 (16 percent to 21 percent) stabilizing after that around 20-21 percent. The strong revenue performance is due to the introduction o f the VAT, which more than off-set the loss in import tariffs and the improvements in tax administration. Import tariff revenues represented o n average 7.8 percent o f the GDP over the period 2000-2003, and dropped to 4.8 percent in 2004. On the other hand, VAT represents o n average 7.2 percent o f the GDP over the period 2004-2005, while tax consumption represented only 2.9 percent o f the GDP in 2003. Tax debt recovery has improved as a result o f the law (approved in 2003), that empowered D G C I to collect coercively through confiscation o f assets and bank accounts. Further improvements in tax revenue

l6 The Authorities have not yet started using the 2001 GFS Manual, but they intent to adopt it very soon.

9

collection are expected, with introduction o f the NIF (Tax Identification Number), planed for 2006, which will allow a better management o f the tax-payers database. Also, i t i s expected that starting in 2006 revenues will increase with the cessation o f the current arrangement with BCA. Under the current arrangement, B C A retains 1 percent o f the total amount o f i t s collection^.'^

2.4 The 2004 reforms in indirect taxation moved the tax structure and its consequent revenue composition towards a more mature fiscal system (Box 2.1). Import taxes appear to have lost their prominence as a major source o f revenue and the less distortionary and more progressive VAT have gained importance. Since its introduction VAT represents o n average 35 percent o f the total o f tax revenues, while since the tariff reform, import tar i f fs contribution to tax revenues dropped from 45 percent to 24 percent. Nevertheless, the overall tax system has moved towards a more regressive one, as direct taxes have lost importance (from 40 to 30 percent o f the total o f tax revenues, 2000-2005) while indirect taxes gained (from 60 to 70 percent o f the total, 2000-2005).

2.5 Direct (income) taxes constituted a steady 6 - 7.5 percent of GDP during 2000-2005 but were below potential because of a wide-spread and unfounded system of tax exemptions. Estimates for 2005 suggest that revenue forgone from tax exemptions was corresponding to approximately 24 percent o f potential revenue. Furthermore, in the case o f corporate taxes, the rationalization o f tax exempts and incentives could reduce the current dependence on tax collection from key tax payers, as I O companies contribute to 80 percent o f the total corporate tax collection, and one contributes to one third o f the total. Companies contribute, on average, to 40 percent of direct tax revenues.

2.6 While current revenues have shown a rising trend, capital revenues (mostly composed of external assistance) have varied annually, highlighting the perennial uncertainty attached to this source of revenue. The external assistance has spanned from a l o w o f 5 percent o f GDP in 2003, to a high o f 11 percent in 2004, and to a l ow 6 percent o f GDP in 2005. With the graduation to middle-income country, Cape Verde’s eligibility for external aid and concession funds will decrease over time, thus it is o f critical importance that the country creates the fiscal space to meet i t s future pressures.

2.7 Total revenues have been consistently overestimated, with the exception of 2002.18 N o n tax-revenues and external grants present the largest deviations (Table 2.4). In 2003 and 2005 the rate o f execution o f external grants was respectively 59 percent and 50 percent. This i s not surprising as government projections o f external assistance, especially for project financing, tend to be optimistic. I t is important, however, to avoid the overestimation o f revenues as that could encourage overspending. Table 2.4 also shows that VAT revenues have been largely underestimated since their introduction.

A protocol between D G C I and four financial intermediaries for tax collection was signed in October 2005, whereby collectors will receive a flat rate for the service C V E 150-250 per collection. BCA, however, sti l l retains 1 percent o f tax collected but DGCI expects to s ign a protocol with B C A in June 2006 whereby B C A will also receive a flat rate.

Recall that 2004 and 2005 are preliminary estimates.

10

Table 2.1: Revenues, 2000-2006 (CVE million) 2000 2001 2002 2003 2004l 2005' 2006'

Total o f current revenues 12,000 14,134 16,766 17,945 18,466 20,806 24,604 Tax revenues 10,576 12,593 15,141 15,145 16,636 18.543 21,099

Direct taxes 3,912 4,436 5,505 5,146 5,394 5,822 6,955 Income tax 3,912 4,436 5,505 5,146 5,394 5,822 6,955

Indirect taxes 6,663 8,157 9,635 9,998 11,242 12,721 9,643 VAT 1,671 1,951 2,545 2,283 5,591 6,551 8,35 1 Taxes internat. transac. 4,208 5,341 6,126 6,755 3,976 4,231 4,501

Non-tax revenues 1,424 1,541 1,625 2,799 1,829 2,262 3,505 Net lending 634 23 1 220 537 23 1 147 Domestic caa. aarticiaation 10 40 256 External gri ts- 3,931 4,190 6,319 4,087 8,946 5,053 7,099

Total revenues 16,657 18,567 23,305 22,610 27,643 26,006 31,612 'Estimates. ' 2006 Budget Draft Bill o f Law. Source: Ministry o f Finance and Public Administration and IMF.

Table 2.2: Revenues, 2000-2006 (Percentage o f GDP) 2000 2001 2002 2003 2004l 2005l 2006'

Total o f current revenues 18.6 20.4 23.0 22.6 22.4 24.3 25.1 Tax revenues 16.4 18.2 20.8 19.0 20.2 21.7 21.5

Direct taxes 6.1 6.4 7.6 6.5 6.5 6.8 7.1 Income tax 6.1 6.4 7.6 6.5 6.5 6.8 7.1

Indirect taxes 10.3 11.8 13.2 12.6 13.6 14.9 9.8 VAT 2.6 2.8 3.5 2.9 6.8 7.7 8.5 Taxes internat. transac. 6.5 7.7 8.4 8.5 4.8 5.0 4.6

Non-tax revenues 2.2 2.2 2.2 3.5 2.2 2.6 3.6

Domestic cap. participation 0.0 0.0 0.0 0.1 0.0 0.0 0.3 Net lending 1 .o 0.3 0.3 0.7 0.3 0.2 0.0

External grants 6.1 6.0 8.7 5.1 10.8 5.9 7.1 Total revenues 25.7 26.8 32.0 28.4 33.5 30.4 32.2 'Estimates. ' 2006 Budget Draft Bill o f Law. Source: Ministry o f Finance and Public Administration and IMF.

Table 2.3: Revenues, 2000-2006 (Percentage of Total Revenues) 2000 2001 2002 2003 2004l 2005l 2006'

Total o f current revenues 72.4 76.1 71.9 79.4 66.8 80.0 77.8 Tax revenues 63.8 67.8 65.0 67.0 60.2 71.3 66.7

Direct taxes 23.6 23.9 23.6 22.8 19.5 22.4 22.0 Income tax 23.6 23.9 23.6 22.8 19.5 22.4 22.0

Indirect taxes 40.2 43.9 41.3 44.2 40.7 48.9 30.5 VAT 10.1 10.5 10.9 10.1 20.2 25.2 26.4 Taxes internat. transac. 25.4 28.8 26.3 29.9 14.4 16.3 14.2

Non-tax revenues 8.6 8.3 7.0 12.4 6.6 8.7 11.1 Net lending 3.8 1.2 0.9 2.4 0.8 0.6 0.0 Domestic cap. participation 0.0 0.1 0.0 0.2 0.0 0.0 0.8 External grants 23.7 22.6 27.1 18.1 32.4 19.4 22.2

Total revenues 100.0 100.0 100.0 100.0 100.0 100.0 100.0 'Estimates. ' 2006 Budget Draft Bill o f Law. Source: Ministry o f Finance and Public Administration and IMF.

11

Table 2.4: Revenues, 2000-2005 (Executed Revenues as Percentage of Budgeted Revenues) 2000 2001 2002 2003 2004’ 2005l

Total of current revenues 89.2 91.9 97.3 92.2 98.5 104.7 Income tax VAT Taxes internat. transac.

Non-tax revenues

97.6 97.5 109.1 87.5 91.7 94.3 91.7 107.9 101.4 262.5 122.3 99.3 104.6 169.1 80.6 114.0

61.5 63.2 54.4 70.0 58.8 44.4 External grants 61.3 64.6 101.5 58.6 86.5 49.1

Total revenues 100.7 88.7 102.5 80.8 94.9 88.3 ‘Estimates.

Source: Ministry o f Finance and Public Administration and IMF.

Recommendations

2.8 The contribution of import taxes to improve macroeconomic efficiency can be enhanced. As a small island economy, Cape Verde would benefit from further trade liberalization to reduce consumption and production costs and offset to some extent the high transportation costs and the diseconomies o f scale.

2.9 Collection from direct taxes (personal and corporate income tax) and VAT could improve substantially by streamlining tax exemptions and incentives. The VAT i s considered as a less regressive tax instrument than the previous consumption tax. However, the current number o f exemptions wil l reduce its effectiveness as a tax instrument (Box 1). Reduction o f tax exemptions i s also very important as a measure to enlarge the very narrow tax base, and to reduce the current dependency on a few number o f companies. I t i s also recommended to unify al l exemptions in the single law covering al l fiscal incentives.

Box 2.1: Income, import and domestic consumption taxes Import tariffs and trade related taxes: The pre-2004 trade taxation system included a large number of tariff bands, more than 100 exemptions and other trade related taxes. Until end 2003, the tariff system st i l l comprised some 64 bands and peaked at 250 percent, cumbersome for a small country to implement. The weighted average tariff rates stood at 31 percent in June 2002, very high in comparison to Madagascar (4.4 percent in 2001) or Senegal (10.5 percent in 2001).

In early 2004, the Government implemented a sweeping tax reform where most taxes and fees that were administered by customs were eliminated, the number of tariff bands were reduced to seven, and the maximum tariff tops at 50 percent. In comparison, Slo Tom6 and Pn‘ncipe has a tariff structure composed of four bands which peaks at 20 percent, pointing to the continuing high tariff bamers in Cape Verde. VAT (Value-Added Tax): The traditional consumption tax (with rates varying between I O and 60 percent) was distortionary as it made manufactured goods disproportionately more expensive relative to the price of inputs. T h i s distortion affected the producers’ investment decision making process. In addition, consumption taxes have a regressive effect inasmuch as the price elasticity of demand on basic goods i s very small and any increase in prices would affect the poor (who save less) as much as the rich.

In early 2004 the consumption tax was replaced by a 15 percent VAT. In 2005 a reduced rate of six percent was introduced on hotels and tourist activities. A tax on special consumption was also introduced for luxury goods (ten percent, except for vehicles older than four years: taxes range from 30,60 and 150 percent based on the age of the vehicle).

The VAT contains a number of exemptions: (a) for goods considered “worthy” such as goods and services for education and health sectors, training, culture, art and sport, basic foods and medicines; @) some public services, such as post and removal of garbage and funerals; (c) Non-Profit Organizations; (d) some imported goods and import regimes; (e) exemptions with objectives to facilitate tax collection or reduce the cost of collecting tax, to avoid taxing sectors where collection would be difficult such as financial operations, insurance, agricultural activities; and (0 exemptions goods and services obtained by an exporter to produce goods for exports, etc. Direct taxes (personal and corporate income tax): Individuals are supposed to declare their incomes. The taxes on such declarations are progressive. The 2004 and 2005 budget laws included five tax rate bands (from 15 percent to 45 percent). If individuals do not declare their income within the allotted period, a flat rate of 20 percent i s estimated for the person. A rate of 30 percent i s applied to all corporations or individuals who identify themselves as “enterprises.”

The update of the taxpayer registry i s in its final stages. All the enterprises are already registered. The identification of the ‘missing’ or new individual taxpayers i s finishing as well. Identification cards have amved but had not been distributed as of early February 2006.

12

B. TRENDS AND COMPOSITIONS OF EXPENDITURES

2.10 The economic classification of the expenditures show a broadly stable pattern of expenditures: over the last five years, an average o f 62 percent o f resources i s spent o n current expenditures and the remaining resources (an average o f 38 percent) o n the public investment program (henceforth PIP).” This broad distribution i s affected by annual variations, which in turn are affected by the unpredictability o f actual PIP implementation, and to a lesser extent by the costs o f privatizing some public enterprises that spilled over from the late 1990s into the f i rst ha l f o f the current decade. In 2000 the extra-budgetary costs o f privatization amounted to 16.6 percent o f total expenditures, tapering down to 0.9 percent and 2.8 percent respectively o f total expenditures in 2002 and 2003.*’

2.11 The pattern of expenditures over 2000-2005 also highlights the continued prominence of non-discretionary expenditure. The three categories (salaries and benefits, social transfers, and interest rate payments) represent o n average 85 percent o f the total o f the current expenditures and 54 percent o f the total o f expenditures. Continued increases in non- discretionary expenditures wil l constrain the fiscal space available to the government to respond to external shocks. Even though in the medium-long run only interest rate payments constitute mandatory expenditures, as subsidies, salaries and transfer expenditures can always be objects o f reforms that decrease their weight in the budget, i t is important to emphasize that in the short- run, within a constrained budget, if the government does not wish to increase the overall deficit, the weight o f any potential budgetary cuts resulting from a negative shock would fa l l onto discretionary expenditures such as goods and services or the investment budget.

2.12 The recent approved law on pensions will increase even more the mandatory expenditures’ share of the budget. With the new law the Government will be transferring contributions to the INPS, while continuing to pay pensions to al l c iv i l servants that have joined the Public Sector before December 31, 2005. The Central Government will transfer monthly to the I N P S 23 percent o f the wage bill o f the new c iv i l servants, plus 8 percent o f the current active c iv i l servants and pensioners. The expected effect in the 2006 budget is an increase o f 2 percent in the current expenditure. As a result, transfer to the I N P S plus pension payments (contributory), wil l represent approximately 5 percent o f total expenditures and mandatory expenditures will represent 85 percent o f the current expenditures in 2006.21

2.13 The share of salaries and benefits to the GDP rose from 9 percent in 2000 to an estimated 12 percent in 2005, an increase of 33 percent. This increase captures the impact o f several back to back salary hikes in the first ha l f o f the decade to improve the standard o f living for c iv i l services. I t also reflects the hiring o f teachers and health professionals. The increase in 2005, more specifically, reflects promotions and the hiring o f police. The government is aware o f the escalation in wages and benefits and, thus, has decided to impose a freeze o n promotions until the PCCS (Career and Salary Plan) i s completed. Furthermore, beginning in August 2006, the new c iv i l servants database (which excludes the ghost servants) wil l be used to process the

l 9 Note that this classification includes the transfer by the government to autonomous funds and services. 2o In 2000, funds were used to pay for BCA, PROMOTORA and CECV. In 2002, several payments were paid for liquidation of TRANSCOR INFA, and CME. *’ Non-contributory pensions are not included in this analysis.

13

payments. Measures to tighten the wage bill are o f particular importance within the constrained budget.

2.14 Within the social transfers and subsidies category the government has succeeded in controlling scholarship costs. Starting in 2002, the authorities made a pol icy choice to reduce the funds provided for full public scholarship and encourage cost-sharing by the families o f eligible students and the students themselves. I t is estimated that the scholarship costs have decreased by almost 50 percent between 2002 and 2005 (from C V E 555 mil l ion in 2002 to an estimated C V E 282 mi l l ion in 2005).

2.15 In contrast, petroleum subsidies have increased in 2005 after an abrupt decline in 2002. In 2005, reflecting the o i l price increase and the absence o f price adjustments, the total o f petroleum subsidies amounted to 7 percent o f total expenditures. This was not included in the 2005 budget but i t is in the 2006 draft law. The pol icy measure o f eliminating subsidies on April 27th will put an end on the impact o f petroleum subsidies in the budget if reforms are taken to allow the viability o f the sector. Should that not occur, implici t subsidies will persist in the form o f recapitalization o f the company or payment o f a tariff deficit as in the past.

2.16 Interest rate payments increased significantly in 2002 and 2003, being twice as much in 2003 than in 2000. Even though debt service has been decreasing steadily since 2003, the recent increase in domestic borrowing may lead to a reversal of the current scenario. Table 2.3 shows that the interest payments on debt increased from 3.8 percent o f total expenditures in 2000 to 8.7 percent in 2003. This represents a sharp 129 percent increase in the share o f debt service payments in total expenditures. Since 2003 interest rate payments have been decreasing, standing at 6.8 percent o f total expenditures in 2005. During the period 2003-2005, interest rate payments averaged 7.5 percent o f total expenditures or 0.63 percent o f the GDP. However, as a result o f the increase in domestic borrowing, interest rate payments can increase in the future. The stock o f domestic debt increased by E C V 3.1 bi l l ion (or 3.6 percent o f the GDP in 2005) between end-December 2004 and end-December 2005, leading to a stock o f domestic debt o f 33 percent in 2005, while i t stood at 20 percent in 1999. I t i s worth noting that this increase reflects to some extent the policy o f issuing domestic debt to clear arrears.

14

Table 2.5: Expenditures, 2000-2006 (CVE million) 2000 2001 2002 2003 2004’ 2005l 2006’

Total current expenditures 17,201 12,249 15,274, 16,819 15,194 17,751 24,123 Total current expenditures (exc.

extraord. exp.) 12,740 12,749 15,049 16,127 15,194 17,648 24,123 Prima y current expenditure 11,720 11,413 12,887 14,129 13,138 15,719 22,301

Goods and services 422 293 454 1,392 1,002 1,267 1,862 Salaries and benefits 6,420 6,339 7,015 9,574 9,384 10,265 12,712 Transfers and subsidies 4,268 4,363 4,694 2,645 2,158 2,725 6,575

Petroleum subsidies 988 1,174 966 137 450 2,400

Interestpayment on debt 1,021 1,336 2,162 1,998 2,056 1,930 1,822

Interest payment int. debt 495 803 1,437 1,478 1,506 1,384 1,300

Restructuring costs 4,434 225 692

Other expenditures 103 Capital expenditures

Public investment program 9,455 9,680 9,719 8,167 11,191 10,468 16,290 Total expenditures 26,656 22,428 24,993 24,986 26,385 28,220 40,413 ‘Estimates ’ 2006 Budget Draft Bill o f Law. Source: Ministry o f Finance and Public Administration and IMF.

Other current expend. 609 418 724 518 593 1,461 1,152

Interest payment ext. debt 526 533 725 517 550 546 522

Extraordinary expenditures 4,461 225 692 103

Social emergency 3

Table 2.6: Expenditures, 2000-2006 (Percentage of GDP) 2000 2001 2002 2003 2004l 2005’ 2006’

Total current expenditures 26.7 18.4 21.0 21.1 18.4 20.8 24.6

extraord. exp.) 19.7 18.4 20.7 20.3 18.4 20.7 24.6 Prima y current expenditure 18.2 16.4 1 7.7 17.8 15.9 18.4 22.7

Goods and services 0.7 0.4 0.6 1.7 1.2 1.5 1.9 Salaries and benefits 9.9 9.1 9.6 12.0 11.4 12.0 13.0 Transfers and subsidies 6.6 6.3 6.5 3.3 2.6 3.2 6.7

Petroleum subsidies 1.5 1.7 1.3 0.2 0.0 0.5 2.4

Interest payment on debt 1.6 1.9 3.0 2.5 2..5 2.3 1.9

Interest payment int. debt 0.8 1.2 2.0 1.9 1.8 1.6 1.3

Total current expenditures (exc.

Other current expend. 0.9 0.6 1 .o 0.7 0.7 1.7 1.2

Interest payment ext. debt 0.8 0.8 1 .o 0.6 0.7 0.6 0.5

Extraordina y expenditures 6.9 0.3 0.9 0.1 Restructuring costs 6.9 0.3 0.9 Social emergency 0.0 Other expenditures 0.1

Capital expenditures ‘Public investment program 14.6 14.0 13.4 10.3 13.6 12.2 16.6

Total expenditures 41.3 32.3 34.4 31.4 32.0 33.0 41.2 ‘Estimates. ’ 2006 Budget Draft Bill o f Law. Source: Ministry o f Finance and Public Administration and IMF.

2.17 There are still questions regarding whether the government faces contingent liability risks with regards to SOEs and Autonomous institutions (funds and services). Among the SOEs, the study o f the liabilities o f T A C V (the airlines o f Cape Verde), ENMOR (ports operator), EMPROFAC (pharmaceutical distributor), and INTERBASE (fish freezing

15

company) would shed light on the implici t risks to the central budget.22 As for the autonomous institutions, they have l imited resources o f their own but there are also recipients o f large transfers from the central budget (Table 2.9).23 Government data suggest an annual transfer o f C V E 2.1 bi l l ion to C V E 2.9 billion. In the long run, such large transfers put in question the autonomous nature o f these institutes as wel l as the feasibility o f such a n arrangement. This relationship i s certainly worth a closer analysis, conditional on the availability o f reliable data.

Table 2.7: Expenditures, 12000-2006 (Percentage of Total Expenditures)

Total current expenditures 64.5 56.8 61.1 67.3 57.6 62.9 59.7

extraord. exp.) 47.8 56.8 60.2 64.5 57.6 62.5 59.7

2000 2001 2002 2003 2004l 2005l 2006'

Total current expenditures (exc.

Primary current expenditure 44.0 50.9 51.6 65.6 49.8 55.7 55.2 Goods and services 1.6 1.3 1.8 5.6 3.8 4.5 4.6 Salaries and benefits 24.1 28.3 28.1 38.3 35.6 36.4 31.5 Transfers and subsidies 16.0 19.5 18.8 10.6 8.2 9.7 16.3

Petroleum subsidies 3.7 5.2 3.9 0.5 1.6 5.9 Other current expend. 2.3 1.9 2.9 2.1 2.2 5.2 2.9

Interest payment ext. debt 2.0 2.5 2.9 2.1 2.1 1.9 3.2 Interest payment int. debt 1.9 3.6 5.8 5.9 5.7 4.9 1.3

Restructuring costs 16.6 0.9 2.8

Other expenditures 0.4

Interest payment on debt 3.8 6.0 8.7 8.0 7.8 6.8 4.5

Extraordinary expenditures 16.7 0.9. 2.8 0.4

Social emergency 0.0

Cauital exuenditures Public investment program 35.5 43.2 38.9 32.7 42.4 37.1 40.3

Total expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 'Estimates.

Source: Ministry o f Finance and Public Administration and IMF. 2006 Budget Draft Bill o f Law.

Table 2.8: Overall Balance, 2000-2006 2000 2001 2002 2003 2004l 2005l 2006'

Total revenues (CVE million) 16,567 18,567 23,306 22,610 27,644 26,006 31,612 Total expenditures (CVE million) 26,656 22,428 24,993 24,986 26,385 28,220 40,413 Nominal GDP (CVE million) 64,539 69,380 72,758 79,527 82,550 85,461 98,100 Overall balance (CVE million) -10,089 -3,861 -1,687 -2,376 1,259 -2,213 -8,801 Overall balance (?! GDP) 15.6 5.6 2.3 3.0 1.5 2.6 9.0 'Estimates.

Source: Ministry o f Finance and Public Administration and IMF. 2006 Budget Draft Bill o f Law.

Table 2.9: Central Government Transfer to Autonomous Institutes, 2000-2005 (CVE Million)

2000 2001 2002 2003 2004l 2005l Total 2,769 2,829 2,897 3,391 2,625 2,474

Of which: IFA's resources 177 270 260 530 82 330 Of which: transfers from central budget 2,276 2,300 2,481 2,860 2,544 2,144

Expenditures 2,646 2,714 2,739 3,220 2,625 2,577 'Estimates. Source: Ministry o f Finance and Public Administration.

22 EMPA (the national commodities trading company) has been liquidated. 23 Government data, multiple years, (map 111-6).

16

Recommendations

2.18 Better control of mandatory expenditures i s a necessity to ensure an adequate fiscal space for the government to deal with potential external shocks. The following measures should be considered:

i. More control of the number o f civi l servants hired; ii. More control over the rate and extent of salaries increase, with possible alternative payment

systems to consider (e.g., productivity bonus payments); iii. Continued control over the payment o f transfers (e.g., scholarship payments); iv. A better understanding o f the budgetary relationship o f the government with the so called

“autonomous institutions and services;” v. Further analysis o f possible financial and continent liability relation between the government

and SOEs (e.g., ENASA, TACV); and vi. Control over domestic debt.

Box 2.2: The budget process, the role of corrective budgets and extra-budgetary expenditures

The Budgetary Process:

The budget typically i s elaborated in the fal l o f the previous year, adopted in January o f the year it i s to be executed. In election years, such as in 2006, the same budget as 2005 i s executed until the new government takes power. This practice i s to discourage any attempt by political parties to “capture” the budgetary process to their benefit in the “pre-election” period. The elections took place in February 2006 and the new government and national assembly, who took power in early March 2006, have 60 days (by law) to approve the 2006 budget.

Typically, the government alters (or “corrects”) the budget over the year as i ts needs to adjust expenditures within the envelope of the budget law, Within each min is t ry , moving funds within each sub-head i s allowed. However, funds can not be moved from current expenditures to investment or vice versa.

Supplementary budgets: In case the overall envelope of the budget needs revision, an amendment to the budget law i s approved by the national assembly. In 2000, a supplementary budget was adopted to accommodate arrears in payments o f salaries and benefits. In 2002, the budget was amended in August o f 2002 to accommodate the creation o f three new ministries (CVE 1.369 bi l l ion increase in the functional budget) and to regularize the arrears on external and internal debt f rom 2001 (CVE 1.339 billion). In 2005, a supplementary budget was again adopted, th is time to increase the expected revenues by some 5.0 percent (thanks to the success o f the new IVA and import taxation and the new resources provided by the US-MCC program). The amendment also included a 6.8 percent increase in current expenditures (to accommodate election costs, hiring o f new teachers and police). The investment budget i s programmed to increase by a slight 1.1 percent to help recapitalize the Electricity company (Electra), public safety, and support creation o f new municipalities (five municipalities were created).

Extra-budgetary expenditures: during 2000-2003, these expenditures seemed to be focused on paying several payments following the privatization efforts o f the late 1990s and the early part o f the current decade. The extra budgetary spending in 2005 was composed o f a series o f small, un-programmed expenditures, such as acquisition o f a generator for electricity, airplane tickets, training and hotel bills.

17

c. THE EFFICIENCY OF BUDGET IMPLEMENTATION AND THE ROLE OF THE PUBLIC INVESTMENT PROGRAM

C.l Budget Implementation Patterns: Programmed, Revised, and Executed

2.19 As it i s typical in many developing countries, the implementation of current expenditures i s stronger than that o f capital expenditures. Current expenditures typically reach over 87 percent o f the init ial ly allocated budget. The large overrun in current expenditures in 2000 is due to the settlement costs o f privatization (16.6 percent o f total expenditures) and payment o f o i l subsidies to o i l distributors (represented in transfers and subsidies and other current expenditure categories). On the other hand, the best ratio o f implementation to programming for capital expenditure i s 84 percent in the year 2000. In 2003 the ratio is particularly l o w (59 percent), which can be traced to l o w mobilization o f external grants and to the difficulties in predicting the degree o f implementation o f the projects. This outcome i s the result o f two features in the budgetary process that have been already discussed above: (i) mandatory expenditures drive the implementation o f current expenditures; and (ii) the variable inf low o f external assistance affects implementation o f the investment program.

2.20 Within the current expenditures category, the ratio o f outlays to programmed expenditures in the subcategories of mandatory expenditure (salaries and benefit, transfers and subsidies and interest payment) i s characteristically high. For instance, in the salaries and benefit subcategory the outlay i s usually over 90 percent o f the originally programmed budget. In the interestpayment subcategory, this ratio surpasses 100 in 2000,2004 and 2005, with the domestic interest payment almost tripling in size over the period (from C V E 494 mi l l ion in 2000 to CVE 1,383 mi l l ion in 2005). The transfer and subsidies subcategory covers transfers to other levels o f government (autonomous entities and municipalities), petroleum price subsidies as wel l as transfers to students studying abroad. As noted earlier, this category has lost some o f its earlier prominence as the government has curbed student scholarships, but s t i l l has a typical ratio o f implementation to programmed budget o f over 85 percent.

2.21 This high implementation ratio in mandatory expenditures i s often secured at the expense of %on-mandatory’ categories such as goods and services. In the category o f goods and services the ratio o f implemented to programmed budget ranged from 40 percent in 2001 to a high o f 85 percent in 2003, dropping to an estimated 73 percent in 2005.

2.22 The government has made an effort to adjust the original budget allocations to changes in expenditure developments through budgetary amendments (2000, 2002 and 2005) and annual revisions. This effort has dual goals: (i) one the one hand the authorities wish to maintain control over the expenditures; and (ii) o n the other hand, they wish to lessen the tension between allocation and outlays. While these adjustments have led to better distribution o f resources across economic categories and improved the ratio o f actual expenditures to amendedrevised budgets, a comparison o f Table 2.10 suggest that they have not changed the conclusions presented above regarding the nature o f expenditures.

18

C.2 Programming and Implementation of the Public Investment Program (PIP)

2.23 Programming of the PIP for the sectors (namely education, health, agriculture and infrastructure) i s undertaken at the sectoral level and i s usually based on the sectoral strategy as well as national development goals agreed to in documents such as the PRSP and the MDGs. Projects are usually prioritized at the sectoral level and then submitted to the MFAP. In the case o f health, the budgetary demands o f running three hospital and clinics are strong determinants o f the sectoral budget, including the PIP. Programming used to be undertaken under the traditional approach where current and investment budgets were developed separately and negotiated with the MFAP. In 2005, in an effort to unify the budgeting process at the sectoral level, the MFAP provided the sectors with ceilings for both overall and current budgets and encouraged jo int (current and investment) budget development.

2.24 Programming of the PIP has been routinely optimistic over the period 2000-2005 (Table 2.10). The ratio o f implementation to programming has not surpassed 84 percent (2000), and has actually fallen as l o w as 59 percent (2003). This variable performance i s largely due to the fact that the execution o f the (PIP) relies heavily o n projections o f external assistance. Table 2.1 1 shows that over the six year period, over 70 percent o f the executed capital budget is sourced internationally. In 2004 this share reached 92 percent. When disbursement o f external funds are delayed or do not materialize, implementation can be held up or even cancelled.

Table 2.10: Comparison of the Original Budget with the Executed Budget, 2000-2005 (Actual Expenditures as Percentage of Budgeted Expenditures)

2000 2001 2002 2003 2004l 2005l

Goods and services 60.3 40.2 61.3 85.0 59.4 73.1 Salaries and benefits 98.1 9.3.2 96.5 93.7 84.1 91.1 Transfers and subsidies 124.9 87.9 1 10.4 80.7 76.6 95.1 Other 140.3 84.0 91.7 86.7 66.5 114.0

Interest payment on debt 100.8 79.4 93.3 92.3 113.2 117.0 Capital expenditures

Public investment program 83.7 74.8 68.5 59.0 80.4 64.6 'Estimates. Source: Ministry of Finance and Public Administration and IMF.

Total current expenditures 142.1 86.9 98.8 91.4 82.7 94.5

Table 2.11: Public Investment Program (PIP) Execution and Sources of Funds, 2000-2005

2000 2001 2002 2003 2004l 2005l PIP (CVE million) 9,455 9,680 9,719 8,167 11,191 10,468

Domestic financing 29.0 19.1 28.5 14.0 7.9 26.3 Of which (in % o f total PIP)

International financing 71.0 81.0 71.5 86.0 92.1 79.7 PIP (% nominal GDP) 14.6 14.0 13.4 10.3 13.6 12.2

International financing 10.4 11.3 9.6 9.1 12.5 9.0

Of which (in % o f total PIP) Domestic financing 4.2 2.7 3.8 1.2 1.1 3.2

'Estimates. Source: Ministry o f Finance and Public Administration and IMF.

19

Table 2.12: Comparison of Amended and/or Revised Budget with Executed Budget, 2000- 2005 (Actual Expenditures as Percentage of Budgeted Expenditures)

2000 2001 2002 2003 2004l 2005l Total current expenditures 126.2 84.6 95.6 91.2 81.8 90.6

Goods and services 68.4 39.9 69.4 84.4 58.0 72.9 Salaries and benefits 95.9 92.9 96.0 92.4 84.0 88.2 Transfers and subsidies 93.3 80.0 92.6 86.5 77.3 96.5 Other 83.7 62.1 91.2 87.2 102.1 99.5

Interest payment on debt 100.0 97.4 100.0 97.1 100.0 100.0 Capital expenditures

'Estimates. Source: Ministry o f Finance and Public Administration and IMF.

Public investment program 74.5 74.0 62.2 58.6 11.4 58.3

2.25 The authorities have managed to use domestic resources to generate some leverage from variable inflow of external assistance funds. First o f all, an array o f minor investment and small maintenance works are undertaken with domestic funds. Also, when an internationally funded project i s a priori ty and i s not very large, domestic funds (from the treasury) can be marshaled in to close the fiscal gap for a short period - Le. until the international funds are disbursed. In this manner the implementation o f the project can move ahead. However, in cases where investment amounts are significant, for instance in infrastructure projects, the projects are postponed, or remain incomplete.

2.26 This variability of the inflow of external assistance i s rooted in two main factors: (i) the nature of the assistance (project vs. budget support); and (ii) delays related to donor or recipient political or policy implementation processes. First o f all, as i s typical in many developing countries, external assistance to Cape Verde i s mostly channeled through project implementation. This type o f assistance i s prone to delays as donor and recipient bureaucracies address their respective procurement, financial management and other national regulatory requirements. I t also spawns project implementation units that often function outside national budgetary processes and make unifying the budget process difficult. Secondly, sometimes providers o f external assistance have to satisfy their own national, political, or pol icy objectives - which are not always aligned with the objectives o f the recipient country, therefore affecting external assistance for the latter.

Recommendations

2.27 Given that recourse to large amounts of domestic credit i s not a sustainable option for the country, the authorities would greatly benefit from four mediating actions to secure a more predictable inflow of external assistance:

i. .. 11.

... 111.

Pursue more vigorously budget support instead o f project support to circumvent the problems associated with implementing the latter; Continue Public Finance Management reforms to establish a transparent and efficient budgetary process to strength donor confidence in the national budgetary process and therefore providing budget support; Go in pursuit o f harmonization o f strategies, procedures and processes among the development partners o f Cape Verde; and

20

iv. Program donor commitments over multiple years in coordination with the newly launched MTEF process.

2.28 strengthen budget preparation and implementation.

In order to increase the rate of execution of the budget the Government should

D. PRSP PRIORITIES AND THE BUDGET

2.29 I t i s important to emphasize that due to the recent adoption o f the full PRSP it is too early to undertake a full assessment o f the government’s implementation efforts and their results. In this exercise, only investment data are analyzed, because disentangling the reasons behind changes in current expenditures i s very difficult given the lack o f qualitative and quantitative data at this stage.

D.l The PRSP Priorities

2.30 The government of Cape Verde adopted its Poverty Reduction Strategy Paper in mid-2004.24 The PRSP outlines five major pol icy pillars: (i) promote good governance, strength efficiency and guarantee equity; (ii) promote competitiveness to improve economic growth and employment creation; (iii) development and improve human capacity; (iv) Develop infrastructure and promote use o f land and protection o f environment; and (v) promote the efficiency and sustainability o f the system o f social protection.

2.31 The PRSP i s part of a broader strategic planning exercise for economic growth and poverty reduction. I ts f ive major pol icy pillars are the same as the same o f the Grand Options o f the plan 2002-2005 and the National Development Plan (NDP). Furthermore, The PRSP appears to provide a link between different sectoral strategies and the NDP. However, while the PRSP provides an overall strategic consistency, the authorities are fully aware that they need to strengthen the coherence o f the PRSP with the number o f national and sectoral programs with respect to their timeframe (that o f the PRSP is for 2004-2007 while the others range from 2002 to 2015) and to adopt a unified planning tool.

D.2 PRSP Goals and the Public Investment Program (PIP)

2.32 To assess the government’s effort to implement the PRSP, an examination of the connection between the five stated goals of the PRSP and the 2005 PIP (programming and implementation) i s undertaken. This i s because the PIP i s a major instrument for economic development and an integral part o f the MTEF process. I t bears repeating that 2005 was the first full year o f the PRSP implementation and the data for 2004 and 2005 are estimates. So any analysis and conclusions will be preliminary.

2.33 Overall, the financial programming associated with the priority actions of the PRSP document i s ambitious. For instance, against a backdrop o f an annual average PIP implementation o f C V E 9.6 bi l l ion over 2000-2004 (CVE 11.2 bi l l ion in 2004), the PRSP program suggests a needed investment budget o f C V E 14.2 bi l l ion for 2005 (an increase o f 27 percent relative to 2004). This programming i s also optimistic, especially in light o f the fact that

24 An interim PRSP was prepared in January 2002.

21

the country is experiencing a shift from bilateral to multilateral external assistance and will be graduating from the LDC group in the near future. Given these financial realities, unless the authorities intensify their efforts to secure more external assistance or alternative sources o f funding, reaching the investment goals outlined in the PRSP will be very difficult.

2.34 Nonetheless, the government had overall success in achieving its budget related indicators of good governance on education and health. Preliminary calculations suggest that the total expenditures on education reached 31 percent and 29 percent o f the national budget in 2004 and 2005 respectively, thus surpassing the PRSP benchmark o f 20 and 22 percent. Preliminary estimations on health expenditures also suggest that benchmarks were exceeded. Whi le PRSP benchmarks for 2004 and 2005 stood at 6.3 and 6.5, respectively, total expenditures in this sector achieved 8.1 percent and 8.2 percent respectively (see table 2.15).

2.35 The government does not yet have a comparable (one-to-one) programming framework between the PRSP action plan and the PIP. To analyze the PIP expenditures more fully in view o f the PRSP objectives, 2004 and 2005 data on projects and programs provided in the PIP execution framework are grouped under five categories associated with the five PRSP pillars.25 The year 2004 is to give some perspective on the 2005 exercise. Investment totals for each o f the five pillars from the PRSP document are included (Table 2.13).

2.36 Table 2.13 provides the results of this exercise. While the government has refocused its investment budget allocation in 2005 compared to 2004, this new focus i s still not fully synchronized with the PRSP action plan. A br ief comparison o f the 2004 and 2005 allocation suggests that the government had clearly planned to re-focus i t s expenditures on pillars I11 (health and education) and IV (infrastructure and transport) in 2005. The planned increases in pillar I11 and IV are 59 percent and 23 percent respectively, and the two pillars were to capture 75 percent o f the total PIP budget for 2005. However, i t is noteworthy that the budget allocation and the PRSP financial programming for 2005 do not match. The budget allocations for pillars 1-111 largely surpass the PRSP programming estimates. On the other hand, Pillars IV and V are 'under-programmed' by over one-third and one-half, respectively.26

25 Recall the PRSP pillars are: (i) promoting good governance, strength efficiency and guarantee equity; (ii) promoting competitiveness to improve economic growth and employment creation; (iii) developing and improving human capacity; (iv) developing infrastructure and promoting use o f land and protection o f environment; and (v) '' For detailed breakdown on the allocation and execution in 2004 and 2005 see Annex 1- Table 4.

romoting the efficiency and sustainability o f the system o f social protection.

22

Table 2.13: PRSP and the Public Investment Program, 2004-2005 (CVE million)

Pillar 2004l 2004l 2005' 2005l 2005l 2005l Alloc YO Exec YO PRSP YO Alloc YO Rev. YO Exec YO

I 2,099 15.1 815 10.1 150 11.0 892 5.5 2,558 14.2 2,073 19.8 I1 2,422 17.4 1,705 21.1 1,536 10.8 2,575 15.9 2,352 13.1 1,237 11.8 I11 3,815 27.4 2,361 29.2 1,996 14.1 6,050 37.3 5,643 31.4 2,532 24.2 IV 4,965 35.7 2,792 34.6 9,573 67.4 6,094 37.6 6,553 36.5 4,109 39.2

budge?

V 552 4.0 386 4.8 943 6.6 595 3.1 816 4.5 509 4.9 Total 19,915 100.0 8,070 100.0 14,198 100.0 16,209 100.0 17,950 100.0 10,469 100.0 'Estimates.

'Poverty Reduction Strategy Paper. Source: Ministry o f Finance and Public Administration and IMF.

2.37 Of further concern i s that the execution of the budgeted PIP i s not only far from its original allocation, but also from its revised allocation and the PRSP allocation, highlighting the need for an in-depth review of the circuit of programming and implementation for the PIP. For instance, the execution rate o f activities in pillar 111 was only 42 percent o f the original allocation, 45 percent o f the revised allocation but was 126 percent o f the PRSP allocation (Table 2.13).

Recommendations

2.38 I t would be beneficial if the authorities would:

0

0

0

Continue the development and the adoption o f the MTEFs (sectoral and national), and ensure proper integration between the MTEF and budget process; Correct the current programmatic dichotomy between the PRSP actions and the PIP so as to better translate the PRSP plans into action; and Enhance efforts to secure multi-year external assistance programs with international partners to ensure better financial predictability for the PIP.

E. THE ADMINISTRATIVE CLASSIFICATION AND THE MINISTERIAL EXPENDITURE PATTERNS - A N ANALYSIS OF THE ADMINISTRATIVE EXPENDITURE PATTERNS

2.39 Table 2.15 highlights a concentration of expenditure in the PRSP priority ministries. In fact, 85 percent o f the expenditures are undertaken by four o f these ministries (Ministry o f Finance and Public Administration, Ministry o f Education and Sports, Ministry o f Justice and Internal Administration and Ministry o f Health).

2.40 The Ministry of Finance and Public Administration i s clearly dominant. In fact, its share o f expenditure has trended up from 35.3 percent o f total current expenditures in 2000 to 45 percent in 2005, an increase o f 27 percent. The significant share o f these expenditures was to deal with restructuring costs related to privatization and to rising debt interest payments.

2.41 The Ministries of Education and health make up an average 25 percent and 8 percent respectively of total current expenditures between 2000 and 2005. This supports the claim by the authorities that growth and poverty reduction have been a national priori ty before the adoption o f the I-PRSP (in 2002) and the PRSP (in 2004). The lion's share o f the sectoral

23

current expenditures are salaries and benefits for the c iv i l service. According to the authorities, over 65 percent o f each sector’s total yearly budget i s dedicated to salaries and benefits. In 2004 and 2005, estimated expenditures in these sectors have increased (Table 2-10), f ir ther highlighting the focus o f the government o n these priori ty sectors.

2.42 Expenditures by the Ministry of Justice and Internal Administration (MJIA) have trended down over the period, with the estimated 2005 expenditures making up less than half of the expenditures in the year 2000. This i s a surprising result as the sector has benefited f rom new hires in recent years and i s considered a priori ty sector in the PRSP. This result may be due to the preliminary nature o f the 2004-2005 data.

2.43 The Ministry of Agriculture and Fisheries (MAF) and the Ministry of Infrastructure and Transport (MIT) capture little o f the current expenditure budget. In fact the share of MAF in total current expenditure has declined from 2.5 percent in 2000 to 1.24 in 2005. MIT expenditures make up a relatively steady one to 1.5 percent o f the total current expenditures. One reason for these surprising results may be that these sectors usually require large investments. The fhds for these large investments are supplied through external resources, which are registered in the investment budget. Another reason is that financial support and technical outreach to the farmers and fishermen and their dependents may not be financially channeled (and registered) through the MAF, so that the sectoral budget data does not provide an accurate representation o f activities on the ground.

2.44 The overall rate of execution to budget allocation highlights a relatively large annual variation, suggesting that better overall programming of the budget and better preparation to address economic shocks i s a must (Table 2.15). This pattern is especially visible at the MFP, which has been responsible for addressing unpredicted expenditures, such as paying out for privatization costs. Also, the Min is t ry o f Health (MH) registers the increased staffing in 2003 with a large peak in expenditure. On the other hand, the Ministr ies o f Education, Infrastructure, and Justice and Internal Administrat ion display relatively stable execution patterns.

24

Table 2.14: Administrative Classification of Executed Expenditures, 2000-2005 (Percentage of Total Expenditures)

2000 2001 2002 2003 2004l 2005' President of the Republic 0.66 0.62 0.48 0.46 0.71 0.68 National Assembly 2.46 2.38 2.25 2.36 3.72 3.30 Head of Government 8.36 1.60 1.81 2.03 2.32 2.02 Supreme Court 0.13 0.15 0.15 General Attorney 0.05 0.15 0.13 Accounting Tribunal 0.17 0.17 0.18 0.15 0.27 0.24 Min. o f Foreign Affairs and Community 4.26 2.53 3.97 3.45 4.28 4.02 Min. o f Finance and Public Administration 35.35 39.00 44.79 42.14 41.12 44.95 Min. o f Justice and Internal Administration 7.74 7.69 6.55 6.35 3.17 3.34 Min. of Defense 3.94 4.05 2.79 2.84 0.28 0.26 Min. of Agriculture and Fisheries 2.49 2.46 2.14 2.49 1.38 1.24 Min. of Education and Sports 24.10 23.62 20.02 22.46 30.68 28.56 Min. of Health 8.95 8.31 7.30 7.10 8.10 8.25 Min. o f Infrastructure and Transport 1.22 1.10 0.96 1.42 1.47 1.23 Min. o f Tourism, Industry, and Commerce 0.30 0.37 0.39 0.38 Min. of Culture 0.47 0.37 0.35 0.62 0.25 Min. of Reform of the State, Public Administration and Local Authority 4.71 5.09 4.93 0.19 0.16 Min. o f Labor and Solidarity 0.90 0.91 0.92 0.95 0.85 Min. o f Economy, Growth, and Competitiveness 0.45 0.36 Total administrative expenditures 100.0 100.0 100.0 100.0 100.0 100.0 Source: Ministry o f Finance and Public Administration and IMF.

Table 2.15: Administrative Classification of Executed Expenditures, 2000-2005 (Percentage of Initial Budget Allocations)

2000 2001 2002 2003 2004' 2005' President of the Republic 106.5 99.4 99.5 97.9 100.0 110.8 National Assembly 100.7 90.0 94.0 94.0 97.1 100.0 Head of Government 213.8 66.6 22.6 94.5 79.7 120.0 Supreme Court 164.4 121.8 135.1 General Attorney 69.6 122.4 127.8 Accounting Tribunal 72.1 88.0 106.2 85.8 99.0 100.0 Min. of Foreign Affairs and Community 76.4 50.1 92.1 87.7 84.0 91.5 Min. of Finance and Public Administration 134.9 85.8 115.1 92.2 88.6 105.0 Min. of Justice and Internal Administration 60.3 90.8 92.3 91.5 86.6 93.4 Min. of Defense 96.5 99.9 93.7 100.6 80.0 90.8 Min. of Agriculture and Fisheries 95.3 120.5 83.8 78.0 85.0 94.4 Min. of Education and Sports 103.0 96.0 96.3 93.4 94.1 102.8 Min. of Health 105.7 92.0 82.9 100.7 84.2 100.2 Min. of Infrastructure and Transport 189.0 85.1 87.2 80.5 82.5 84.8 Min. of Tourism, Industry, and Commerce 78.4 28.7 96.2 Min. of Culture 77.4 91.9 31.0 Min. of Reform of the State, Public Administration and Local Authority 85.0 64.1 80.9 Min. of Labor and Solidarity 85.7 84.7 86.2 Min. of Economy, Growth, and Competitiveness 77.6 79.6 Total administrative expenditures 104.6 87.0 101.0 92.0 89.5 106.6 Public Investment Program 83.7 74.8 62.2 59.0 80.4 64.6

Source: Ministry o f Finance and Public Administration and IMF.

25

Table 2.16: Administrative Classification of Executed Expenditures, 2000-2005 (CVE million)

2000 2001 2002 2003 2004' 2005' President o f the Republic 95 89 85 92 99 113 National Assembly 357 342 400 470 521 539 Head o f Government 1,210 230 322 404 325 330 Supreme Court 25 21 25 General Attorney 11 21 22 Accounting Tribunal 25 25 33 30 37 40 Min. o f Foreign Affairs and Community 616 363 706 686 600 658 Min. o f Finance and Public Administration 5,120 5,603 7,965 8,392 5,759 7,355 Min. o f Justice and Internal Administration 1,121 1,105 1,165 1,265 444 546 Min. o f Defense 570 582 496 566 39 43 Min. o f Agriculture and Fisheries 361 354 381 495 193 202 Min. o f Education and Sports 3,490 3,397 3,559 4,472 4,297 4,672 Min. o f Health 1,297 1,194 1,299 1,414 1,134 1,349 Min. o f Infrastructure and Transport 177 159 170 282 206 202 Min. o f Tourism, Industry, and Commerce 43 53 69 76 Min. o f Culture 68 66 70 86 42 Min. o f Reform o f the State, Public Administration and Local Authority 677 906 982 27 25 Min. o f Labor and Solidarity 129 161 183 133 139 Min. o f Economy, Growth, and

Total administrative expenditures 14,484 14,366 17,783 19,914 14,006 16,360 Public Investment Program 9,455 9,679 9,719 8,167 11,191 10,468 Total expenditures (including miscellaneous) 23,568 23,63 1 27,502 28,083 25,197 26,828 Source: Ministry o f Finance and Public Administration and IMF.

Competitiveness 63 59

Table 2.17: Administrative Classification of Executed Expenditures, 2000-2005 (Percentage of GDP)

2000 2001 2002 2003 2004' 2005' President o f the Republic 0.15 0.13 0.12 0.12 0.12 0.13 National Assembly Head o f Government Supreme Court General Attorney Accounting Tribunal Min. o f Foreign Affairs and Community Min. o f Finance and Public Administration Min. o f Justice and Internal Administration Min. o f Defense Min. of Agriculture and Fisheries Min. o f Education and Sports Min. o f Health Min. o f Infrastructure and Transport Min. o f Tourism, Industry, and Commerce Min. o f Culture Min. o f Reform o f the State, Public Administration and Local Authority Min. o f Labor and Solidarity Min. o f Economy, Growth, and Competitiveness Total administrative expenditures Public Investment Program Total expenditures (including miscellaneous) 36.52 34.06 37.80 35.31 30.52 3 1.39 Source: Ministry o f Finance and Public Administration and IMF.

0.01 0.04 0.86

10.55 1.59 0.71 0.62 5.62 1.78 0.35 0.10 0.09

1.23 0.23

25.04 10.27

0.03 0.05 0.77 8.61 0.64 0.05 0.24 5.47 1.58 0.24

0.05

0.03 0.16

0.07 19.14 12.25

0.04 0.96 7.93 1.74 0.88 0.56 5.41 2.01 0.27 0.07

22.44 14.65

0.04 0.05 0.52 0.97 8.08 10.95 1.59 1.60 0.84 0.68 0.51 0.52 4.89 4.89 1.72 1.79 0.23 0.23 0.08 0.09 0.10 0.09

0.98 1.24 0.19 0.22

20.71 24.44 13.95 13.36

0.55 0.49 0.55 0.59 0.63 0.63 1.88 0.33 0.44 0.51 0.39 0.39

0.03 0.03 0.03 0.03 0.05 0.73 6.98 0.54 0.05 0.23 5.21 1.37 0.25

0.10

0.03 0.16

0.08 16.97 13.56

26

3. DEVELOPMENTS IN PUBLIC FINANCE REFORMS

A. THE LEGAL FRAMEWORK

3.1 The applicable legal texts on public finance are being reviewed and are in the process of completion. When these texts are implemented, Cape Verde will have at its disposal a very comprehensive legal framework to manage its public finances soundly and effectively. A s part o f the ongoing reform, which begun over three years ago with the Financial Administration Reform task force (RAFE), Cape Verde has embarked on deep reforms in i t s legal framework. The legal texts currently under review include: (i) the budget framework law (ii) the budget planning law; and (iii) the public chart o f accounts. All the related applicable texts have been discussed and are in the process o f being approved.

A.1 Budget Framework Law

3.2 A draft of the budget framework law was submitted to the Parliament in April 2006 and i s expected to be approved by the end of the year. This law wil l replace Decree-Laws No. 78N/98, No. 29/2001, and No. 30/2001. The most relevant new elements o f this draft law are: (i) the presentation and vote o f the following year’s budget every year (even if the following year i s an election year); (ii) the formalization o f the authorizing officers and financial comptrollers; (iii) the deconcentrution o f spending to the line ministries; and (iv) the creation o f ex-ante and ex-post internal and external controls.

3.3 The deconcentrution of the spending will make the line ministries truly responsible for the management of their own appropriations, thus improving managerial discretion and lessening the arbitrariness of budget cuts. The creation o f the function o f the financial comptroller and the expansion o f the function o f principal authorizing officer to sector ministries2’ will improve effectiveness and transparency. Furthermore, the new law wil l allow for a programmatic approach and progress on the MTEF and the program budget agenda.

3.4 The authorizing officers and financial comptrollers are the subject of two specific legal texts, which have been drafted, but have not yet been approved. The financial comptrollers wil l be in charge o f ex ante control o f the regularity, legal, economic and efficient nature o f revenues and expenditure. Under the authority o f the General Directorate for Public Accounting (DGCP), they will report to the MFPA o n a quarterly basis.

A.2 Budget Planning Law

3.5 The draft law was submitted to the Parliament in April 2006 and it i s expected to be approved by the end of the year. This law will establish the objectives, principles and framework for budget planning. I t includes the central government, and the regional and municipal levels. The main objectives are to ensure a balanced development o f sectors and regions, to promote a better quality o f l i fe for the population, to express interests o f the various stakeholders and social groups and to have a rational and efficient resource allocation. This economic and social plan, which must comply with the main options o f the National Development Plan, wil l involve the Parliament, the Economic and Social Council and technical

*’ In the past, this function belonged to the Minister o f Finance only.

27

services. Based on results, th is system wil l include a Medium-Tern Economic Framework (MTEF), with the objective o f ensuring a sound adequacy between public expenditure and macro-economic framework and to facilitate the prioritization o f expenditure.

A.3 Chart of Accounts

3.6 A new chart of accounts i s being fmalized and should be implemented in 2006. The main objective i s to have a modem accounting system to record cash transactions, budgetary transactions (revenues and expenditure), the balance o f fixed assets, and public debt. I t will be used by the central government, the municipalities and autonomous institutes. The reporting should improve, to a large extent, the reliability o f budget reporting and the auditing o f public accounts.

A.4 Recommendations

3.7 Regarding the legal Framework, the following recommendations are proposed:

i. In 2006, al l new laws and decrees on public finance management, which are currently being discussed and under preparation, should be finalized;

ii. The role o f financial comptrollers should be limited to the stage o f commitment o f expenditure and, later, to the stage o f order to pay and check previous budget’s commitment in order to avoid redundancy in terms o f accountability between financial comptrollers and authorizing officers;

iii. The field o f competence o f financial comptrollers should include the DGPE, to subject transactions initiated by this directorate, to appropriate controls; and

iv. The new law should integrate in an explicit way the concept o f public accountant, who should be formally appointed and whose functions should be incompatible with those o f the State authorizing officer (to allow proper segregation o f duties).

B. BUDGET PREPARATION AND ADOPTION

3.8 The budget elaboration cycle i s carried out according to a reliable schedule, which in general has been strictly followed. The main phases of budget preparation can be summarized as follows:

i. First phase (until March 31), the MFAP assembles estimations o f revenues (DGCI, DGA) and expenditures (DGO), and then defines the main pol icy strategies;

ii. Second phase (until July 15), MFPA elaborates the call circular; iii. Third phase (until September 15), the council o f ministers approves the draft budget

bill o f law; iv. Fourth phase (until October 15), the government presents the draft budget bill o f law

to the parliament; and v. Fifth phase, the draft budget bill o f law i s voted within 60 days after i ts presentation.

This schedule has, in general, been rigorously followed. The schedule proposed in the draft o f the budget framework law i s overall very similar, with the noteworthy difference o f moving the deadline for the preparation o f the cal l circular to April 15, thus allowing more time for the sectoral ministries to prepare their appropriations accordingly.

28

3.9 A procedure handbook has been drafted on budget preparation and elaboration and i s accessible on the Intranet. In addition, a comprehensive manual i s currently being drafted, for the use by al l departments at the MFPA. Training seminars have been held to improve the use o f the SIGOF system.’* There i s s t i l l scope for capacity building regarding SIGOF, mainly at the sectoral level.

3.10 The consistency across sectoral budget programming and the quality of the overall budget could be further improved with the presentation of a true general policy framework at the beginning of the annual elaboration of the budget cycle. The budget cycle i s initiated by the MFPA with the estimation o f overall revenues and expenditures, analysis o f the domestic economic trend, definition o f the investment and debt pol icy and general fiscal policy. Based on these first estimates and analysis o f the MFPA, the cal l circula?’ is consequently sent to the sector ministries indicating rough estimates o f the overall ceilings and some general policy priorities. The budget preparation exercise would greatly benefit from improvements in analysis and information conveyed in the call circular. For instance, the absence o f information on the execution o f past years leads to new allocations based on cost estimates o f previous years. In practice, the lack o f information leads to the renewal o f the recurrent budget, with some random incremental adjustments. This practice o f incremental adjustment risks creating a gap between the budget appropriations and real needs. Another limitation o f the call circular is that i t i s weakly linked to the macroeconomic framework, which is a consequence o f weaknesses in macro modeling. A macro modeling tool has been developed locally to improve revenue forecasting. At the present, revenue receipts are from the last three fiscal years and GDP growth expectations.

3.1 1 The coordination between the Budget General Directorate (DGO) and the Planning General Directorate (DGP) i s from now on a de facto effective basis, although the new budget law i s not yet in force. Until very recently, the budget preparation was carried out by two different departments, the DGP and the DGO, despite being done o n the basis o f the same classification. This double intervention created a weakness, especially for the evaluation o f the recurrent costs o f the investments and their integration into the budget. The 2006 recurrent and investment budgets are being elaborated together by the DGP, the DGO and the DGT, thus overcoming the past coordination costs. The elaboration o f both budgets on the same track is a significant improvement and it was made easier by the new budget nomenclature (same nomenclature for recurrent and capital t ransact ion~)~~ that allows for the consolidation o f each ministry’s global expenditure (current spending and investment). At the level o f line ministries, the problem is that two units are dealing separately with the budget: the unit for human resources and the GEP (Directorate o f Studies and Planning). These two units are to be merged into one unit named DGPOC, but this has not been implemented because o f a lack o f human resources. DGPOC exists only at the ministries o f Environment and Family and Solidarity, but st i l l operate in a dual budget mode (investment and recurrent). The classification by function has been designed, but it i s not in l ine with the Classification o f the Functions o f Government o f the United Nations (COFOG) and a bridge between the two wil l be necessary.

Integrated information system with a unique database. *’ Memo sent to the sectoral ministries. 30 Published in the Decree-Law No. 53-A/2003 in the Official Bulletin o f November 26,2003.

29

3.12 The principle for consolidation of the Administrative Public Sector i s stated in the draft Budget Law and it has already been put into practice since 2003. The budget o f autonomous institutes is prepared by the institutes themselves, and then included in the state budget, for transfers f rom the State, as wel l as for their own revenues. The harmonization o f the chart o f accounts for a l l public entities will be effective when the new chart is implemented in 2006. There are st i l l some issues regarding the harmonization o f dates for budget preparation as it seems difficult to impose the State budget schedule on the municipalities as the municipal budget preparation cycle i s less complex and thus shorter.

3.13 Some progress has been made in the preparation of the sectoral MTEFs, but this process i s still in its infancy. International technical assistance and training is being provided to the MFPA and sector ministries, in order to help them develop the global and the sectoral MTEFs. Four ministries have started the elaboration o f the MTEFs: Health, Education, Agriculture, Fisheries and Environment, and Social Protection. The Ministries o f Education and Agriculture, Fisheries and Environment, and Social Protection already have a f i rst draft. The MTEF on Health i s behind schedule and that is primarily the result o f lack o f capacity and shortage o f human resources at the sectoral level. Coordination o n the MTEF exercises between DGO, DGP and the sectors has improved but needs to be strengthened. A lack o f reliable information o n the revenues’ forecast and on donor financed projects, prejudices the budget preparation and slows down the implementation o f the MTEF.

3.14 The 2006 budget was linked to the PRSP and to the MTEF. The four ministries that have produced MTEFs prepared the 2006 budget in accordance with their MTEFs. This is an important improvement but i s s t i l l in a very early stage. Regarding the alignment with the PRSP programs and sub-programs, they are now linked to the budget and the SIGOF functionalities allow for their management in a coherent way. The 2006 budget has been prepared in a decentralized way with l ine ministries, directly entering data o n the SIGOF system and it has also been prepared for investment by program and sub-program. Moreover, sectors are shifting progressively to a results-based management approach.

3.15 Overall the implementation of the MTEF has not been properly conducted, undermining the effectiveness of this instrument. The preparation o f the sectoral MTEFs overlapped with the preparation o f the global MTEF and the interaction between the two processes was insufficient. The global MTEF set ceilings which were not respected as truly upper limits and, as a result, the global MTEF was revised to accommodate the revised ceilings. Furthermore, PRSP and MTEF are not in line, as again the preparation started at about the same time, and thus PRSP does not emerge from the MTEF. This complicates the budget preparation and execution as there i s not a clear and consistent directive. In order to use the MTEF as an instrument that defines the overall fiscal strategy, the global MTEF has to work as a top down envelope setting hard budget constraints that define l ine ministries spending plans. Without a hard budget constraint, derived from the global MTEF, trade-offs between sectors/ministries are not faced and policy i s not disciplined by availabilities.

3.16 The foregoing suggests that the Government has yet to develop a coherent vision of how a MTEF should be developed, and the role it can play in supporting the Government’s fiscal strategy, ensuring that allocations to sectors and ministries reflect political priorities, and that projects and programs are adequately and reliably funded through the budget. I t

30

will be especially important for the Government to develop a sound MTEF based on the budget process in the years ahead, to maximize the development value o f public spending, but also to put the country in a better position to manage any future external shocks, given the central role o f the fiscal pol icy in adjustment in Cape Verde.

3.17 The linkage between government departments and donors i s strengthening. A first memorandum o f understanding was agreed upon in February 2005, and signed in April 2005 by the Budget Support Group, consisting o f the World Bank Group, the European Union and the Kingdom o f the Netherlands. A revised version is being prepared and two new donors are joining the support group: the Afr ican Development Bank and Spain. The DGP is the liaison between the various departments and donors.

Recommendations

3.18 To improve budget preparation and adoption:

i. A bridge between the functional classification and the Classification o f the Functions o f Government o f the United Nations (COFOG) should be defined;

ii. The quality o f the circular call should be improved; and iii. The MTEF should become the overarching framework for public expenditure

planning, pol icy making, and annual budget execution. Subsets o f spending such as the PRSP must fit within the MTEF, both in terms o f immediate outlays as wel l as the spending dynamic that PRSP policies initiated. I t i s particularly important, in this regard, to develop reliable estimates o f the outer year costs, both capital and recurrent, that PRSP policies give rise to.

C. BUDGET EXECUTION

C.l Management of Expenditure

Recurrent Expenditure

3.19 follows:

There are six phases in the expenditure stream. Currently, the six phases are as

i. Emission o f a requisition by the concerned ministry; ii. Validation o f the requisition by the DCP (Directorate for Public Accounting). This

operation, which consists o f checking the availability o f the appropriations, i s called “cabimento;”

iii. Return o f a copy o f the “cabimento” to the technical ministry for order o f work or

iv. After execution o f the service, liquidation o f the amounts due by the DCP. The supply;

V. vi

liquidation consists o f calculating the national debt and certifying the effectiveness o f the expenditure, thanks to the document evidence provided by the qualified service o f the line ministry; Emission o f an order to pay by the DCP “cobranqa;” and Payment to be carried out by the Directorate for Payment Services (DSP) o f the Directorate General o f the Treasury (DGT).

31

3.20 The accountability o f line ministries for their spending decisions will be effective once the draft of the budget framework law i s approved. Currently, the system in force i s highly centralized at the MFPA. The sectoral ministries have only the role o f proposing commitment to expenditure. The orders to pay are manually drafted by them, and then recorded in the information system by the MFPA. The decision to commit “cabimento”, as wel l as to pay, i s under the responsibility o f the MFPA. The draft budget framework law when approved wil l change that, and expenditures and revenue operations wil l be executed at the sectoral ministries’ level. In al l ministries there will be an authorizing officer who will commit to the expenditure, liquidate and send the payment order to the DGT. This wil l be made easier, due to the new functionalities in the information system, the IT-based Financial Management System, which has been deployed to al l Ministries, but o n a read-only basis. The law establishing the mandate and function o f financial controllers exists but a regulation needs to be finalized and approved. Training o f a l l participants will also be needed.

3.21 Due to the large payment of arrears, suppliers have recovered their confidence in the State and thus the demand for simultaneous commitment and liquidation i s less frequent. The cash shortage experienced during the beginning o f the decade led to the loss o f suppliers’ trust in the State, as the debts were so high that i t raised doubts that the payment would ever occur. T o manage that risk, suppliers raised their prices or asked for payment before the service was provided. Since the ministries were and sti l l are not informed o f when the payments are made, commitments were programmed without real-time information on the status o f liabilities toward suppliers. Today’s situation is slightly different and the practice o f commit and liquidate at the same time, which i s against the legal texts, has become less common.

3.22 The salary and pension systems will benefit from a census of the staff and improvements in the management. A census o f the staff has taken place. The database o f the C iv i l Service Directorate i s currently being merged with the database o f the SIGOF o f the MFPA, except for three ministries: Education, Health and Justice. This process wil l be completed by the end o f 2006. The new system, which i s particularly u s e h l given the high mobil i ty o f the staff, will allow the management o f the carriers as we l l as the payroll. Equipment has been purchased and delivered to the DGCP.

3.23 Budget execution does not reflect the management of the public debt since it i s not managed through the current accounting system. The estimation o f the debt and related payments is made with extra accounting tools. Currently, the DGT i s unable to provide the amount o f the stock o f arrears concerning the current fiscal year and previous fiscal years. Furthermore, the lack o f a single accounting system prevents the aggregation o f the financial data from al l public institutions, making i t difficult to carry out a consolidation o f accounts and in particular to determine their debt and arrears. This is a source o f concern because the State wil l have to ensure payment o f these liabilities.

C.2 Investment Expenditure

3.24 The procedures for investment expenditures vary according to their source of financing: domestic revenue, counterpart funds or external financing. The modalities are described below:

32

i.

* . 11 *

Procedures for investment expenditures financed from domestic revenues are the same as for recurrent expenditures; and The procedures for expenditures financed by external financing are different and difficult to monitor. The management o f most investment projects i s done within the framework o f project coordination units, fol lowing donors’ own rules o f management, with the DGP taking part only in the monitoring. This partly explains the important delays in closing accounts.

3.25 The government lacks information regarding the execution of externally financed investment projects and the consolidated position of the treasury of the state. The budget framework law in force (and also the draft budget framework) establishes that the BCV, as the treasury o f the state, should centralize the consolidated position o f the state, including a l l the external grants and loans. However, the regulation applicable to the autonomous accounts for projects is often not respected. For example, even though there has been an effort to close accounts, there are s t i l l accounts opened at institutions other than the Central Bank and thus large amounts in project accounts are not captured in the state’s accounts o f executed expenditures. Another consequence i s that i t makes difficult the assessment o f the execution rate o f investments.

3.26 Unlike the past, currently unspent appropriations for capital expenditure can be carried forward from one year to the next. The action o f carrying forward is now possible for investment expenditure and food aid. According to the DGP and the DGO, appropriations unspent during a given fiscal year are systematically carried forward to the following year. Still, these appropriations need to be voted in the fol lowing years’ budget. These transactions are easily managed by the SIGOF.

C.3 Management of Revenues

3.27 Tax administration capacity i s mixed. The General Directorate o f Contribution and Taxes (DGCI) and the General Directorate o f Customs (DGA) are the main units responsible for tax assessments and collection. The D G C I is in charge o f the income tax and despite progress made, i t i s st i l l not able to conduct internal controls and taxpayer audits, especially on remote islands mainly because the internal audit office at the D G C I is not yet operational. This action has received special attention from the C F A A unit. Equipment has been purchased and delivered and three senior inspectors have been assigned. Terms o f reference have been drafted for developing the specific SIGOF functionalities and procedures handbooks have been elaborated. The internal audit office wil l be operational in the next three months. By contrast, the DGA is a wel l structured organization with branches in different regions.

3.28 The taxpayer registry i s currently being consolidated. The former database was not reliable as, for instance, some taxpayers were registered under five accounts. A system with companies liable for VAT and persons liable for income tax represents the starting point o f the new registry. The objective o f this action is to provide a unique identifier, which wil l be used, not only for fiscal matters, but also to comply with some formalities, l i ke opening a bank account or even getting married. This system will allow customs debt and VAT credits to be compared and to offset liabilities and debts o f the same taxpayer. The text stipulating the creation o f the fiscal identifier already exists. This identifier will be the same for the D G C I and the DGA and it

33

will also be used by the notaries. Every taxpayer wil l receive an electronic card with his identification.

3.29 DGCI i s progressing in identifying missing taxpayers. D G C I specialists are now validating and correcting the information that the registry contains. The D G C I is continuing to gather taxpayer data and enter these in the registry from the databases o f customs warehouses, corporations and c iv i l service rolls (chambers o f commerce and municipal councils have been contacted but they have not yet replied). The link between the c iv i l service database and the fiscal registry has been carried out, except for municipalities’ staff, The specific software developed at the N O S I i s wel l advanced and staff has been hired to manage this software. Sti l l , the D G C I lacks technical assistance to manage daily problems and communication with the NOSI.

3.30 The recovery of tax debts owed i s still weak but it i s improving. Specific software dedicated to proceeding against bad payers is currently being developed. Since 2003, with the approval o f a new law that legitimizes the D G C I to proceed against bad tax-payers, namely through the confiscation o f wages, salaries and bank accounts in general, recovery o f tax debts has improved. Furthermore, D G C I has negotiated with companies o n better terms for the repayment o f their debts, and consequently some debts have been recuperated.

3.31 The management of government revenues i s improving. The SIGOF’s Revenue Management Module with VAT h c t i o n a l i t y i s operational. I t allows recording information coming from monthly and quarterly forms, but it i s not yet possible to manage the process o f refunding. Terms o f reference have been drafted for developing income and corporate taxes functionalities in the SIGOF. A consultant was hired to train staff in the finance units (mainly the heads o f units) in order to ensure the assimilation o f the necessary knowledge to manage tools associated with SIGOF.

3.32 Regarding tax revenue forecasting, the DGCI has no specific tools at its disposal. As mentioned above, i t forecasts the progression o f revenues on the basis o f the moving average o f the three previous years and the expected GDP growth. Particularly difficult i s the VAT revenue forecast, due to its complexity and lack o f time series data.

3.33 The collection of revenues will benefit from modernized methods and training actions. A team with representatives from the DGCI, the NOSI, and the C F A A unit held working sessions in August and September 2005 with the objective o f designing a pi lot project to modernize the system in the finance units (to respond to enquiries and collect and manage revenues). The project is in the discussion and approval phase among the entities involved. Once it has been approved and finalized, a request for financing wil l be prepared and submitted to the C F A A unit.

3.34 Agreement between the MFPA and the banks on the new application that provides information by taxpayer and by type of tax collected in such a way that it can be cross- referenced i s almost complete. Nearly al l the new agreements envisaged have been successfully concluded with four additional banks (BIA, CECV, B C N and the Post Office), while there was previously only one seemingly unwritten agreement with BCA. Meanwhile, there i s progress involving the DGT, the DGCI, the NOSI, and the GEP on improving the linkage to the new

34

computer application that forms part o f the Payments and Revenues Systems project. It has however been noticed that, while the name o f the taxpayers concerned i s effectively in the documents, the D G C I must wait until the end o f every month to get this information, which i s provided by the monthly transfer o f revenues from the finance units.

3.35 A protocol between DGCI and four financial intermediaries for tax collection was signed in October 2005, whereby collectors will receive a flat rate for the service. However BCA, which i s the main collector, has not signed the protocol yet. Therefore B C A sti l l collects tax and retains one percent o f tax collected outside any formal agreement. The new protocols signed with the new financial institution are based o n a flat rate o f 150 to 250 C V E per payment. B C A thus continues to operate under more interesting conditions than the other banks that have signed the protocol. The D G C I expects to sign a protocol with B C A in June 2006.

3.36 But the coordination among the DGCI, the DGA, the DGT, and the DGCP in consolidating tax and customs receipts has not been strengthened yet. All information needed exists, but improvements in coordination depend o n political will. The D G C I can access the database o f the DGA, but the DGT cannot. Furthermore, the DGT does not have at i t s disposal information o n the volume and the nature o f revenues. Despite this, revenues are transferred on a daily basis to the treasury account through the functionalities o f the software Sydonia ++. There is not yet a device with the necessary functionalities to allow a f low o f information that could permit proper bank reconciliation between the DGCI, the banks and the DGT.

Recommendations

3.37 recommendations:

Budget execution could be enhanced by the implementation of the following

i. .. 11.

... 111.

iv.

vi. V.

MFPA should agree with donors on a framework for donor financed projects to make sure that they are recorded properly in the budget; DGT should record in its books al l public funds (including projects), while enhancing the control environment in the Central Bank to prepare it for a progressive take-over o f a l l accounts; Sector minist ies should be given authority to manage their respective budget provided that the financial control i s effective; The census o f tax debts owed and processing against bad payers should be finalized; The new accounting system should be implemented as soon as possible; and Continuation o f efforts to update MTEFs and to use these in the 2007 budget preparation and budget monitoring.

D. ACCOUNTING AND REPORTING

3.38 The budget accounts are prepared in a timely manner, and since 2004 are sent on a quarterly basis to the Parliament.31 Information regarding real-time budget execution i s now available internally on the SIGOF and on the MFPA website (on a quarterly basis). case for expenditure o n food aid, which i s well-known and managed, as wel l as

3' By law quarterly accounts are issued 45 days after the end o f the period.

This i s the expenditure

35

financed by the State owned revenues and budget support. Expenditure related to externally financed projects is difficult to report because this expenditure does not fol low the public accounting rules but mainly donors’ procedures. The aggregation o f the projects’ financial information i s difficult during the course o f the year as wel l as at the end o f the year for the compilation o f the general accounts o f the State. This partly contributes to the delay in the closing o f the accounts. A census o f expenditure financed by grants i s ongoing.

3.39 However, the State general accounts (a partial balance sheet of the State) are not sent yearly to the TdC, as required by law - this problem i s being remedied in the context of the proposed PRSC. Good progress was achieved in preparing the State accounts and clearing backlogs. State accounts for 2000-03 were finalized and transmitted to the Parliament in December 2005. The State accounts for 2004 and 2005 are expected to be finalized in June and September 2006, respectively. The accounts o f the State for a budget year, after the approval o f the TdC, should be submitted to parliament before the 3 1 st o f December o f the following year. However, this has not happened given the delays in closing the accounts by the MFPA and the subsequent delay in the audits. Two main reasons explain the delays in closing the accounts: First, the project grants, as discussed before, and, secondly, the lack o f harmonization of accounting procedures for al l public entities. The lack o f a single accounting system complicates the aggregation o f the financial data from a l l public institutions. It is expected that the implementation o f the future state chart o f accounts will remedy this by easing data consolidation.

3.40 A new public accounting system has been designed, but not yet implemented. After a very long period o f discussion that commenced in 1998, the new chart o f accounts was approved by the Council o f Ministers on December 22, 2005, and it is expected to be approved by the parliament at the end o f 2006. Some corrections are taking place and some points are st i l l being discussed. This is the case for the length o f the supplementary period. Appropriate training is now required. NOSI i s currently working o n the development o f the specific functionalities. Implementation is planned tentatively for July 2006, and the definitive implementation for January 2007. As mentioned above this new chart o f accounts wil l be used not only for the State budget, but also for autonomous institutes and municipalities with the difference that only the root o f the accounts will be identical.

Recommendations

3.41 measures:

Accounting and reporting could benefit from the implementation of the following

i. Dedicated software should be developed to produce annual accurate treasury plans o n daily, weekly, monthly basis; and

ii. CS-DRMS should be updated for information related to domestic debt.

E. FINANCE OF MUNICIPALITIES

3.42 Management of municipal finance i s in progress. The strengthening o f municipal financing and tax administration i s ongoing. Three municipalities have been equipped with computers and with software developed by the NOSI (Sal, Mindelo and Praia). All

36

municipalities of Santiago wil l be equipped in the next months. An agreement has been signed between the NOSI, the A N M C V and the Directorate o f Decentralization in order to support this action. However, the authorities have not secured the financing to complete this action.

3.43 The new local finances law established a revised formula, increasing the percentage o f Government resources transferred from 7 to 10 percent, and linking the resources transferred to revenues collected by the Government in the previous year.’* Municipal resources are composed o f central transfers (the formula-based Financial Equilibrium Fund (FEF) and contract programs), local taxes and proceeds o f decentralized cooperation. The FEF is the main source o f resources for most municipalities given the narrow tax base (for most o f the municipalities) and weak tax administration. Following the promulgation o f the Local Finances Law, the FEF for 2006 increased by 60 percent.

3.44 I t i s difficult to monitor municipalities’ sources of financing. Municipal budgets do not follow a standardized format and proceeds of decentralized cooperation or one-off revenue such as the sale of land may represent an important source of financing. In practice, a complete monitoring o f municipalities’ financing i s feasible for the FEF only. A project o f implementation o f a Local Finance and Decentralization Monitoring Centre has not yet begun.

F. CASH MANAGEMENT

3.45 The National Treasury i s still managed on a daily basis. Yearly program payments and quarterly or weekly loans are s t i l l weak for the treasury. Treasury management is the responsibility o f the Resources Programming and Management Directorate (DGPR) and the Payment Service Directorate (DSP) o f the General Directorate o f the Treasury (DGT). The DGPR has been given the task o f developing monthly and annual treasury budgets, with information o n the estimated level o f the revenues provided by the D G C I and the DGA and expenditures f i om the DGO. However, annual programming exists only in theory and instead budget cash planning i s in place. Norms and procedures pertaining to the management o f the treasury account are regulated by Decree-Law No. 29/98.

3.46 The lack of solid cash flow plans prevents the smooth financing of expenditures throughout the year, whereby recurrent expenditures crowd-out other expenditures, which eventually become arrears. Because o f past fiscal slippages, cash management reflects a very short-term vision (sometimes almost a day-by-day management), intended to face the cash needs qualified as urgent, such as mandatory expenditures l i ke wages, transfers, and interest rate payments. This situation i s aggravated by the cash accounting system as opposite to accrued accounting system, whereby expenditure materializes only when the Treasury receives the request for payment “factura.” A s in many cases the request for payment corresponds to an expenditure o f several months (for instance the utilities bills) and in the absence o f a plan that accounts for future paymenucash needs the treasury may face a situation o f the shortfall, thus unable to liquidate the debt. The absence o f a cash f low plan largely explains the build-up o f arrears. The difficulties with cash management are exacerbated by the non-existence o f procedures handbooks. There are no procedure handbooks, nor specific tools for treasury

32 Law 79NV2005 (published in the Official Gazette, No. 36, September 5,2005).

37

management. Furthermore, treasury staff have not received specific training to deal with cash management.

3.47 Some accounts opened in commercial banks have been closed allowing for more effective management and the reduction of financial costs. This situation generated preventable financial costs as the State was forced to borrow to cover its financing needs, whereas some o f these accounts had a large credit position. A decision to close al l those accounts and centralize SPA resources in B C A has been a plan since 1999 but has not been fully implemented yet. Autonomous institutes are authorized to use accounts opened in a commercial bank, but their transactions are, as from now, centralized in an account open at the DGCP.

G. DEBT MANAGEMENT

3.48 The treasury records external public debt has at its disposal a software that provides timely information on stock and debt services. The Treasury, which is the unit that monitors debt, uses the CS-DRMS 2000+ software to manage external debt. With this software, the Treasury i s able to know the stock o f debt and to project debt services. However, analytic exercises have not yet been undertaken and the Treasury relies o n the IMF/WB debt sustainability analysis. The software CS-DRMS was implemented in 1996 at the debt office, but i t only became operational in 2004 due to the lack o f trained staff. This software is linked to the SIGOF. The NOSI interprets and controls the information transferred from the Treasury. This control is based on the sound execution o f the various transactions. The NOSI does backups o f information o n a monthly basis. The Central Bank also monitors external debt.

3.49 There i s no adequate information on domestic debt. Stock taking exercises are carried out infrequently. The only timely and adequate information on domestic debt held by the Treasury regards Government securities. There is not an adequate recording o f Central Government responsibilities to the Municipalities, Autonomous Institutes, other agencies or the private sector. From time to time the treasury undertakes a stock taking exercise o f domestic debt. For instance, in 2004, as part o f the PRSC-I, a study on cross-debt was carried out. Later in 2005 the study was updated when a plan for the settlement o f those debts was prepared. The treasury updated the domestic debt data recently, as part o f the preparation for the PSI program with the IMF. The B C V also monitors Government securities debt.

3.50 The management of domestic debt i s not yet linked through I T with the Treasury, which limits the ability of the Treasury to program payments. Debt service o f the domestic debt (Government securities) is st i l l calculated manually in an excel sheet from time to time, hampering the implementation o f an adequate cash-flow plan. In general, other domestic debt is not remunerated.

3.51 The Treasury i s in the process of implementing a program to manage domestic public debt. The Crown Agency (the software owner) i s developing / customizing the program with the treasury’s new software for debt management. This software will be linked to SIGOF allowing an interface between the debt management and the accounting system.

38

The internal and external debt situation will be adequately apprehended only when managed by the Treasury through the accounting

H. THE AUDITING SYSTEM

H.l Internal Audit: General Inspection for Finances (IGF)

3.52 The legal framework regulating the mandates and activities of the IGF was revamped in 2004 and 2005. The decree-law on the status o f IGF staff, autonomous fund and financial control, was approved on August 2005 (DL No. 5512005). This framework empowers IGF and allows better coordination with TdC. The IGF reports to the Minister o f Finance and Public Administration.

3.53 Resources at the IGF have improved but remain insufficient. Operations are still weakened by the insufficient means at its disposal: qualified personnel, equipment and the authority to fulfill its mandate. Since 2003, six inspectors have been hired and the recruitment o f an inspector specialist in information systems i s in progress (for interface with NOSI). The current staff consists o f 18 inspectors, plus 8 inspectors seconded in other government administrations. Because o f its really broad f ield o f investigation (government central administrations, 30 autonomous institutes, 2 1 municipalities, embassies) IGF’s l imited resources impede the fulfil lment of its mandate. In 2004 IGF performed 32 inspections. In 2005 the work program had planned 35 missions o f which 32 were performed. In 2006, the work program forecasted 40 missions, but as o f M a y 2006 only nine missions have been carried out. Because of the costs, priority has been given to the missions in Praia.

3.54 The recurrent appropriations to the IGF continue to be limited and working conditions are precarious. The l imited number o f inspectors makes it difficult to appropriately cover the universe o f entities supervised by the IGF. This situation i s aggravated by the territorial discontinuity which increases considerably the inspection costs. Also, due to the limited budget, IGF doesn’t have adequate means to perform its work, despite IT equipment (desktops, laptops and printers) being provided to IGF in 2006 by a C F A A implementation project, financed by the Netherlands. Lastly, the unit is located in small rooms in a basement which strongly hinders the functioning o f the IGF.

3.55 Cooperation with the General Audit office (TdC) has become stronger. The new decree-law establishes, inter alia, the basic principles o f cooperation with TdC. Recently a meeting was held between the IGF and the TdC to discuss actions o f harmonization. The objective was to compare their schedules, in order to avoid redundancies, and to make their respective actions complementary. I t was also agreed that IGF reports should become available to the TdC.

H.2 External Audit - General Audit Office (TdC)

3.56 A draft o f a new organic law that will empower the TdC has been prepared since the end of 2004 and awaits discussion and voting. The TdC has investigative powers covering al l

33 Every event related to external debt should be recorded in patrimonial accounts and the domestic debt should be included in transactions o f creditor’s accounts.

39

entities that manage public funds and al l financial or contractual transactions which commit the State or affect public finance management, including ex-ante monitoring o f large public contracts that commit the State (competences established by L a w No. 84/IV/93). Aiming to strengthen the capacity o f the TdC, a new organic Law was drafted in 2004, however, this draft has not been discussed yet and it i s uncertain when discussion and voting will take place. The lack o f a proper legal framework weakens TdC’s responsibilities significantly and has an impact on the Public Finance Management framework.

3.57 The lack of adequate resources significantly hampers TdC’s ability to fulfill its mandate. The TdC faces substantial financial and human resource limitations and the insufficient number o f auditors (13 auditors) along with an inadequate budget, hamper its ability to fulfill i t s mandate. The TdC has administrative but not financial autonomy, whereby i t s budget i s negotiated annually with the Minister o f Finance o n the basis o f the Court’s proposals (nonetheless it controls i t s financial management and has its own bank accounts). The l imited budget (0.25 percent of total expenditures in 2005) limits the capacity o f the TdC to carry out field missions in the context o f high transport costs given the geographic characteristics o f the country. This i s of particular concern as there i s a pol icy o f decentralization and deconcentration o f administrations.

3.58 The submission of accounts to the TdC i s delayed and the submission of the 1998 accounts occurred only in 2005. As the comments o f the TdC on one year accounts come out so many years later, their contribution to the improvement o f public finance management i s small. In theory, the State Accounts o f one year should be submitted to the Parliament, after approval o f the TdC, before December 31 o f the following year.34 However, in practice, accounts have been closed with several years lag, but it i s worth mentioning again that in the context o f the PRSC that situation has improved considerably. The delay in submitting accounts, to the TdC and the delay in auditing the accounts increases the time lag between the end o f a fiscal year and the emission o f the comments to the National Assembly. In 2005, the TdC received the State Accounts of 1998 to 2003. The 1998 State accounts have been examined and the 1999 State accounts are under examination. The TdC i s requesting financing to hire f ive consultants for one year in order to clear backlogs (2000-2003 State Accounts) and thus be able in the future to audit the State Accounts timely fashion.

3.59 The TdC has been gradually (but slowly) involved on the public finance reform carried out by the Ministry of Finance. For example, i t has no information about the new budget nomenclature or the new public accounting system. It has, however, participated in studies related with the CPAR and CFA. Moreover, the TdC has benefited from some training actions under the C F A A initiative.

34 The procedure i s as follows: State accounts should be sent to the Parliament by the MFPA, who then send them to the TdC for examination. T h e TdC send them back to Parliament, with its comments, to be voted on, with or without conditions.

40

Recommendations

3.60 system:

The following recommendations are suggested in order to strengthen the auditing

i. The laws o n reorganization and competence o f the TdC should be finalized and submitted to Parliament as soon as possible;

ii. TdC and IGF should receive appropriate resources to conduct their work program; and

iii. Hiring temporary staff at the TdC should be considered as an option for catching up with the backlog o f public accounts and to put in place a more updated control mechanism.

I. PROCUREMENT

3.61 Cape Verde does not have a consolidated legal and regulatory framework for all public sector procurement. For the acquisition o f goods, equipment and services (including consultancies), there are only some arrangements in the form o f decrees regulating procurement in sectors such as infrastructure (civil works). In spite o f the inexistence o f a National Procurement Code, there do not appear to have been major deficiencies in procurement. This is due mainly to a cultural environment o f good governance and the frequent use o f World Bank procedures, even when the Bank is not involved in the project financing. The main issues identified in procurement processes are: (i) a prior and post procurement review done by TdC or the Inspection General o f Finances (IGF), which relates more to the regularity o f expenditures on the basis o f an approved budget than to procurement procedures; and (ii) the insufficiency o f regulation regarding the Contract Management Delegation, except for AGECABO, which successfully carried out program projects in the social sector and school constructions.

3.62 The preparation of a new National Procurement Code i s ongoing. A consultant financed by both the World Bank and the U S Mil lennium Challenge Account (MCC) has been hired to draft the National Procurement Code. A final draft o f the Code i s expected by September 2006. There i s a central coordination unit (GEF) for the development and implementation o f reforms, with two steering committees (CFAA and CPAR). However, the lack o f coordination between the two committees and the insufficient involvement o f the TdC o n the ongoing reform hamper the process. Furthermore, a procurement text on c iv i l works is currently being prepared without apparent coordination with the preparation o f the National Code. There is the risk that some conflict could emerge as one code may contradict and overlap with the other. To prevent such conflict, either the preparation o f the c iv i l works procurement text should wait for the finalization o f the National Procurement Code or they should be developed by teams working closely together.

41

4. ANALYSING THE IMPACT OF ELIMINATING PETROLEUM PRODUCT SUBSIDIES

A. RELEVANT FACTS ABOUT THE PETROLEUM SECTOR, ELECTRA, AND REGULATION

4.1 Cape Verde has no refinery and petroleum products are imported, stored and distributed by two companies. They are Shell Cape Verde (a wholly owned subsidiary o f Shell International, that has held operations in Cape Verde since the 1920s); and ENACOL (Empresa Nacional de Combustiveis, established in 1979, and privatized in 1996 with the Government o f Cape Verde, PETROGAL (Portugal) and SONANGOL (Angola) - each holding one third o f the shares). Agents operate some o f the retail outlets, but the two companies own all oil-marketing facilities.

4.2 The regulator sets petroleum product prices from time to time and the most recent was on 27 April 2006 (the last previous adjustment was in August 2005). For some fuels (Mogas, gasoil, butane sold in large bottles), the price set by the regulator i s supposed to equal the price based on the international market price (Le., international price + freight + margins + distribution costs).

4.3 The oil companies report import prices on a cost + freight basis, with freight apparently computed as a constant “premium” (Table 4.1). However this system i s not transparent. The o i l companies should make explicit the fob cost and the freight cost, so the Regulator can at least compare the reported fob cost with the Platts market quotes. It i s perhaps true that the small volumes being imported impose significant additional costs, but the cost structure should be more transparent to the Regulat~r.~’

Table 4.1: Freight “Premium” (US$/ton) Product 2002 Shell, 2005 Gasoil 35 28 Gasolina 70 76 Jet A-1 45 31 Fuel 380 60 42

4.4 Moreover, Electra accounts for roughly 41 percent of total oil imports (see energy balance, Table 3 in h e x 2). Electra also accounts for the same share o f total margin revenue for the o i l companies. The extent to which this reflects the actual cost structure i s unclear.

4.5 Prior to 1999, electricity and water production and supply to final consumers in the city o f Praia, Sal, Sao Vicente, and Boavista was the responsibility of Electra, the state- owned utility. In early 1999, in order to prepare the ground for private sector participation, the electricity operation o f Electra was extended to cover all the islands, and the provision o f sanitation services to include Praia and Minde10.~~

’’ Since 2003 ENACOL and Shell import jointly: each year they call for two offers, one for fueloil and the other for gasoil, gasoline, je t A, and LPG. This year STASCO has the fueloil contract, T O T A L the contract for supply o f the other products.

36 In the other islands, water supply and the provision o f sanitation services remain the responsibilities o f the municipalities

42

4.6 Following the consolidation of Electra's operations in December 1999, the new Electra SA became a private company through the sale o f 51 percent o f the share value o f the production assets o f electricity, water, for Euro 45,3 mi l l ion to a joint-venture o f Electricidade de Portugal (EDP), and kguas de Portugal (ADP), the electricity and water utilities, respectively, in Portugal. The Government maintained 34 percent o f the share value, and the municipalities retained 15 percent. The Privatization Agreement provided Electra with the exclusive rights for electricity generation and water production up to December 31, 2003. In the case o f water, the exclusive rights were extended up to December 3 1, 2005. Since January 2000, the Government entered into a 36-year concession agreement with Electra for the management and operations o f electricity and water distribution, and the collection and treatment o f wastewater. The concession also provides exclusive rights for these services to Electra in its areas o f concession throughout the concession period. Table 4.2 summarizes some key statistics o f fuel consumption, fuel prices, and sales.

4.7 In July 2006 the State regains the position of major shareholder. Strategic partners and State signed an agreement whereby the company would be recapitalized in order to restore its original position. Furthermore, the strategic partners sold their part to the State and thus the latter became the major shareholder.

Table 4.2: Electra: Salient Statistics 2000 2001 2002 2003 2004 2005

142,327 164,332 181,001 198,652 218,813 236,057 130,937 152,953 171,247 189,733 210,246 229,132

7,927 6,426 5,660 5,366 6,430 6,450 Consumption Estado [MWhI 3,617 4,539 5,023 4,793 6,839 7,370 Municipalities [MWhI 42 1 879 1,155 7,163 8,759 9,439

51,081 50,705 64,955 69,943 79,073 85,209 2,026 2,114 2,327 2,177 2,508 2,702

30,821 37,153 43,563 45,997 52,344 56,406 Total [MWh] 87,967 95,391 117,022 130,073 149,524 161,126 Electricity sold [CVE million] -404,3 -987,4 -1,215,9 -606,9 -485,6 -620,5 Water sold [CVEmillion] 1,304,6 1,388,8 1,698,2 2,286,6 2,509,6 2,622,3 Fuel consumption [CVE million]

Unit Generation Total [GWhI Of which Diesel [GWhI

Wind [ G M I

Domestic [MWhI Public firms [MWhI Private f i rms [MWI

Gasoil 434,7 463,8 521,3 770,5 790,l 826,8 Fueloil 180 [Million liters1 16.1 17.7 17.6 Fueloil 380 iMillion litersj 21.7 25.4

Fuel prices [Million liters] 9.3 8.5 Gasoil [CVE/liter] 37.1 37.1 37.1 Fueloil 180 [CVEfliter] 29.1 29.1 29.1 Fueloil 380 [CVEfliter] 24.4 24.4 24.4

Source: Annual reports 2003 and 2004; financial data from Business Plan.

4.8 Over the past few years, Electra has had unsatisfactory financial results, mainly because fuel prices skyrocketed and tariffs have not been adjusted accordingly. This was exacerbated by lack o f investment into fueloil generation. The predictable result was a financial deterioration and the resulting depletion o f the company's capital. The company began to accumulate a tariff deficit due to increasing fuel costs and other charges, partially because o f the institutional failure o f the ARM (Multisectoral Regulatory Agency) to adjust tariffs when fue l price increases were above the benchmark (agreed in January 2000). The Regulator (ARM) was never established and, therefore, no adjustment mechanism was put in place (ARE i s h c t i o n i n g

43

since August 2004). This lack o f adjustment gave latter origin to the “tariff deficit,” US$13.2 mi l l ion over the next five years, in installments o f US$2.63 a year. The continuation o f erosion o f the capital in 2003 and 2004 led to a recapitalization (CVE 1.4 billion). Electra’s poor financial results are also explained by a weak billing collection. Bertolini et al. (2006) report that the billing collection rates for public agencies and public companies are 65 percent and 57 percent respectively. A s o f April 2006 the amount o f overdue electricity and water consumption bi l ls was C V E 1.9 billion, the equivalent to US$21 million. Even though in 2006 the Central Administration cleared i t s arrears to ELECTRA, there i s st i l l a significant amount o f arrears from Municipalities and Autonomous Institutes.

B. ELIMINATION OF THE SUBSIDIES

B.1 Four types of subsidies

4.9 subsidies:

Before the elimination of the subsidies on April 27, 2006, there were four types of

i. The subsidies payable to the o i l companies o n account o f the retail prices for subsidized fuels (butane in small bottles, kerosene, and fueloil);

ii. The equalization payments to the o i l companies on account o f notionally unsubsidized fuels (Mogas, butane in large bottles, gasoil), but for which at times o f persistently rising prices (as has occurred over the past 2 years) also require adjustment to cover o i l company losses;

iii. The subsidy paid to Electra to cover the difference between the gasoil purchase price from the o i l companies and C V E 37.1Aiter; and

iv. The implici t subsidy that i s implied by electricity tariffs that do not provide adequate return for Electra.

4.10 Subsidies to the oil companies on account of the subsidized fuels. Illuminating kerosene, butane sold in small bottles (3kg and 6kg) and fuel o i l were subsidized, reflecting a clear pol icy choice. Prices were set at a price lower than the real cost. The o i l companies recover the price differential based on an equalization procedure (Table 4.3) based o n each shipment.

Table 4.3: Official Retail Prices Feb. 03 Aug. 03 Jun. 04 Nov. 04 Aug. 05 Actual Prices Apr. 06

Aug. 05 Not subsidized (retail price based on international price) Gasoline CVE/liter 105 100 110 120 132 132 145 Gasoil CVEAiter 60 57 63 69 75 75 90 Gasoil bunkers CVEAiter 41.5 41.5 41.5 47 54 54 67.4 Butane, large bottles CVEkg 102.00 95.00 108.00 120.00 134.00 134.00 144.5 Subsidized Kerosene CVE/liter 41.00 41 44 45 45 57 74 Fueloil 380 CVEikg 30.20 30.20 30.20 30.20 30.20 37.10 49.4 Fueloil 180 CVEkg 25.30 25.30 25.30 25.30 25.30 30.20 40.7 Butane, 3kg bottle CVEhottle 290.00 270.00 270.00 270.00 290.00 340.00 395 Butane, 6kg bottle CVEhottle 612.00 570.00 570.00 570.00 645.00 720.00 830

Source: Regulator (CIF price evolution and official sales prices 2004-2005.XLS).

44

4.11 The equalization payments to the oil companies on account of notionally unsubsidized fuels, result from the lag in the adjustment of petroleum product prices by the Regulator. Even though the automatic adjustment mechanism has been in place since 2003, prices have not been adjusted in accordance with fluctuations in the import As Table 4.3 shows, between August 2003 and August 2005 only two price adjustments occurred. As since early 2004 international o i l prices have steadily increased (see Figure 4.1) and the regulator has not applied the automatic adjustment mechanism o i l companies have incurred in losses that required e q ~ a l i z a t i o n . ~ ~ The equalization occurs whenever o i l companies' actual costs differ from the regulated price. The difference i s reimbursed by the Government (if the actual price is higher than the regulated price), or repaid to the Government (if the actual price i s less than the regulated price): this equalization i s done for each imported shipment (see Annex 3-Table 1).

Figure 4.1: Gasoil Price as FOB ($/bbl versus CVEAiter) I 1 I 100 ,

j I I !

4.12 The subsidy paid to Electra to cover the difference between the gasoil purchase price from the oil companies and CVE 37.1/liter. Gasoil was sold to Electra at the prevailing retail price &e., at C V E 75/liter from August 2005-April 2006), but a direct Government subsidy to Electra covered the difference between this (market) price and the special subsidized price o f C V E 37.10Aiter. Since this price has remained unchanged since 2002, the burden o f subsidy o n Government to cover the difference has increased sharply as world o i l prices have risen.

4.13 The implicit subsidy, which i s implied by electricity and water tariffs that do not provide adequate returns for Electra to meet its long-term investment requirements, i s basically the difference between prevalent tariffs and tariffs that would ensure the viability o f the company. The recapitalization o f Electra and the tar i f f deficit are the result o f inadequate tariff setting, and thus can be interpreted as forms o f implici t subsidy. According to the ARE

as it was parameterized in 2004 (taking into account a stream o f investments and

37 The adjustment mechanism establishes that variations above 2 percent should be reflected in the retail price. 38 It i s worth noting that in early 2006, the price of gasoil, FOB Mediterranean, stood at around CVE 45/liter. Given a current retail gasoil price as established by the Regulator at CVE 90fliter, half the retail price i s accounted for by oil company margins and freight. 39 The results, based on the ARE Model, are based on World Bank staff estimates rather than ARE estimates. The model i s s t i l l under development, and has yet to be fully validated by ARE and ELECTRA. The main assumptions of the model are as follows: (i) Demand projections and other key technical parameters are per 2005 assumptions;, (ii) geographically disaggregated (including a representation of every generating unit in the system); (iii) criterion for setting the tariff i s zero cumulative cash flow

45

elimination o f the subsidy in 2006 for the existing subsidy for gasoil), a 10 percent increase in tariffs would be required in order to achieve long-term financial viability.

c. THE MAGNITUDE OF THE SUBSIDIES IN 2005 AND THEIR IMPACT ON THE BUDGET

C.l Petroleum products

4.14 Table 4.5 shows the breakdown of subsidies for 2005, based on the presentations of the oil companies. The gasoil subsidy accounts for the largest single component o f the total CVE 1.3 bil l ion subsidy claim. Table 3 in Annex 2 shows in detail the calculation o f the equalizations. Calculations o f the equalization are made for each shipment on the basis o f FIFO.

Table 4.4: Petroleum Product Subsidies in 2005 (CVE Million) Shell ENACOL Total YO

Subsidy to oil companies Gasoil’ Marine gasoil Butane, not subsidized Butane, subsidized Butane2 Fueloil 180 Fueloil 380 MoGas

274 228 18 28 62

154 0 32

170 166 50 58 42 31

501 37 46 3 62 5

154 11 32 2

337 25 108 8 73 5

Kerosene 23 11 34 3 ‘See Table 4.4 for details o f the Shell subsidy claims on gasoil. The breakdown by subsidized and unsubsidized

butane i s not available for ENACOL.

4.15 For sales to Electra there i s the additional difference between the regulated price and the Electra price of CVE 37.9/liter, an additional (and much greater) subsidy of CVE 37.Miter. This additional explicit gasoil subsidy to Electra i s CVE 382 million, bringing the total subsidy on petroleum products to CVE 1.73 billion (US$20.2 million).

(after outlays for investment) over 30 years; (iv) after initial adjustment in tariffs (for 2005), tar i f fs are assumed to be adjusted every year for inflation; (v) cash flow includes cost of investments, which in some years are high, in others low (the profile o f investments i s going down), but with a spike every five years or so due to investments in capacity, mostly for electricity - investments in capacity for water are less frequent; (vi) no more subsidies for gasoil as of 2005; and (vii) no “extraordinary injections of funds” from the budget. The conclusions are therefore based on this particular set of assumptions; alternative scenarios would change the results accordingly.

46

Table 4.5: Direct Government Subsidies to Electra for Gasoil Volume of gasoil consumed at rural central and Gamboa (liters)

Gasoil sold by Shell 4,073,858 4,206,112 8,279,970 Gasoil sold by Enacol 3,899,680 5,448,711 9,348,391 Total 7,973,538 9,654,823 17,628,36 1 Amount of gasoil subsidies to rural centrals and Gamboa (CVE)

Gasoil sold by Shell 77,961,752 124,648,637 202,610,389 Gasoil sold by Enacol 8 1,477,162 98,508,506 179,985,668 Total 159,438,914 223,157,143 382,596,057

Source: Ministry o f Finance and Planning and Ministry o f Economy

1 St Hal f o f 2005 2”d Ha l f o f 2005 Total

1 St Hal f o f 2005 2”d Ha l f o f 2005 Total

4.16 The oil subsidies pertaining to the year 2005 were not included in the budget, however, if they were, total expenditures would be 6 percent higher. If the explicit o i l subsidies were to be included in the budget, the fiscal deficit would stand at 5 percent o f the GDP rather than at 2.6. This is a big increase compared to the subsides for 2004, included in the 2005 budget, which amounted to C V E 450 million, the equivalent o f 0.5 o f GDP and 1.5 percent o f total expenditures. The drafted 2006 budget included C V E 2.4 bi l l ion pertaining to subsidies for 2004 and 2005. This i s estimated to represent 2.4 percent o f the GDP.

D. THE ELIMINATION OF THE SUBSIDIES ON APRIL 27,2006

4.17 The petroleum product price increases of April 27, 2006, are unprecedented (Table 4.7). For unsubsidized fuels, the increases range from 8 percent to 25 percent, and reflect the increase in international prices o f the last six months (since the last adjustment in August 2005). For the previously subsidized fuels (butane in small bottles, kerosene, and fueloil), the increases reflect the Government’s decision to remove these subsidies, resulting in price increases o f 36 percent to 143 percent.

Table 4.6: Price Increases of April 27,2006 Fuel Old price, Aug. 2005 New price, Apr. Price increase (YO)

2006 Unit Subsidized

Butane, 6kg CVEkg 107.5 138.3 29% Kerosene CVEAiter 45.0 74.0 64% Gasoil Electra CVEiliter 37.1 90.0 143% Fueloil 18Ocst CVEAiter 30.2 49.4 64% Fueloil 360cst CVEAiter 25.3 40.7 61% Unsubsidized

Butane, 3kg CVEkg 96.7 131.7 36%

Butane, 12.5kg CVEhottle 1675 1810 8 % Butane, 55kg CVEhottle 7320 7950 9% Butane, bulk CVEkg 134.0 144.6 8% Gasoil bunkers CVEAiter 54.0 67.4 25% Gasoil (non-Electra) CVEAiter 75.0 90.0 20% MoGas CVEAiter 132.0 145.0 10%

4.18 The potential consequences for electricity and water tariffs are no less dramatic. With its heavy dependence on o i l for generation, fuel costs are the single most important determinant o f Electra’s costs. Simulations using the financial model for Electra developed by

47

ARE, the regulator, show a quasi linear relationship between increases in o i l prices and the increase in tariff levels required to meet Electra's revenue requirements. As indicated by the ARE model, before the increase in fuel prices adopted in April 2006, a 10 percent increase in tariffs was required in order to achieve long-term financial viability (At crude o i l prices o f $80/bbl (representative o f the first quarter o f 2006), after factoring the increase in fuel prices which is o f the order o f 60 percent o n average (not taking into account the termination o f the gasoil subsidy which had been already factored in the tariff simulations by ARE), the required increase in tar i f fs would be about 50 percent. While these estimates are dependent o n a number o f assumptions, i t is clear that higher fuel prices will require tariff increases.

4.19 On June 1,2006, ARE announced price increases of 13.3 percent for water and 25.4 percent for electricity, applied across the board on all tariff categories, which was below the tariff suggested by the World Bank staff. However, there are some mitigating factors that could reduce the need to increase tariffs. In 2006 the fuel m i x has shifted somewhat further towards heavy fuel oil, which could reduce the average cost o f production for electricity. Improvements in collection ratios and loss rates would also lower revenue requirements. And o i l prices could decline (as they did during September 2006). Overall, taking into account these uncertainties, the fact remains that the June 1 tariff increase o f 25 percent wil l not suffice to meet Electra's revenue requirements (which are defined in part by the medium term investment plan o f the firm), and a W h e r price increase will be necessary in the near future (except in the event that o i l prices return to much lower levels). At the average crude o i l prices o f 2005 (around $70/bbl), the World Bank Staff estimates o f the ARE model suggest a tariff increase o f 17.6 percent (and a further 10 percent to ensure long-term investment) at the US$50-60/bbl level (Le. a return to the price levels o f 2004), the required increase would be less than 10 percent (though a W h e r 10 percent to bring financial returns to an acceptable level wil l st i l l be needed).40

E. THE IMPACT OF THE ELIMINATION OF SUBSIDIES

E.l The Impact on Poverty

4.20 Given that this report is focused on the national State budget and the national macroeconomic perspective, this analysis is limited to an examination o f the impact on poverty from the national perspective. I t may wel l be that the aggregated results disguise important differences in impacts upon individual islands. A caveat o f this analysis i s that i t is based on the 2001-2002 survey. A new survey (focused o n electricity, gas and wood consumption) is expected to be launched by the end o f December 2006.

Electricity and Water 41

4.21 A key question i s whether, from the point of view of poverty reduction, the decision to terminate fuel subsidies was adequate. The answer to this question depends in part on whether subsidies are wel l targeted (i.e., whether they reach the poor). Given a f ixed budget constraint, public resources can be used to subsidize many different goods, f rom electricity and

40 A further critical assumption i s the Euro/$US exchange rate: with the CVE effectively tied to the Euro, further depreciation o f the dollar (as many expect) would lessen the impact o f o i l price changes quoted in US$. 4' No analysis is carried out for fueloil, as almost the entire amount o f fueloil imported is sold to Electra for electricity generation or water production (desalinization). So fueloil subsidies served mainly to further lower the cost o f electricity.

48

water to education and health. Therefore, to the extent that subsidies for electricity and water (through the subsidies granted to inputs such as fuel for the production o f these goods) are less wel l targeted than many other subsidies, i t could make sense to terminate these subsidies so as to provide resources for other programs designed to benefit the poor.

4.22 A simple targeting indicator can be used to compare the distributional impact or targeting performance of utility subsidies to that o f other subsidies. The targeting indicator SZ i s the share o f subsidy benefits received by the poor divided by the proportion o f the population in poverty (Angel-Urdinola and Wodon, 2006). A value lower (greater) than 1.0 implies that the average subsidy received by a poor persodhousehold i s lower (greater) than the average subsidy received by a persodhousehold representative o f the population as a whole. i2 can easily be computed fkom household surveys with data o n household expenditure or consumption o f utility services provided that information is available o n the tariff structures. In addition, the value o f SZ can be decomposed to assess key factors affecting i t s value (Box 4.1). Results indicate that the value o f i2 for electricity before the change in tariffs o f M a y 2006 was 0.37, and it was 0.19 for water. Thus, utility subsidies were very badly targeted, since the benefits for the poor were much lower than those for the population as a whole. Indeed, the value o f 0.19 for water subsidies means that when a randomly chosen individual in the population as a whole benefited from lCVE in subsidy, the benefit for a poor person was only CVE 0.19.

Box 4.1 : Analyzing the Distributional Incidence of Electricity Subsidies Define by R the share o f subsidy benefits received by the poor divided by the proportion of the population in poverty. A first such factor affecting R i s access to networks in the neighborhood where households live, denoted by A. Typically access among the poor being i s lower than among the overall population (Ap < AH.) Furthermore, poor households having access to the networks may be also less likely than average to use the services because they may not be able to afford to pay for them. If we denote the share of all households who use the service given that they have access to the network as UflA, this would mean that U f l A < UPIA in some cases. The value of U f l A i s equivalent to the take-up rate o f connections among those with potential access. AH * UHIA i s equal to the actual connection rate (share of households who are connected and use water or electricity). The next key parameters i s the share of eligible utility service users (i.e. households with access, a connection, and meters, i f applicable) who are targeted and therefore receive a subsidy. This i s denoted by Tau. When all consumption i s subsidized, all users have access to the subsidy and therefore Tqp = Tqu = I. The share of households receiving the subsidy i s equal to share of al l connected household times the share o f households eligible for subsidy (i.e. AH * U ~ A * Tqu). The share of poor households receiving the subsidy i s Ap * UP~A * TP~u.

A second subsidy design variable i s the rate of subsidization, which reflects the difference between what households pay for their consumption and what i t actually costs to produce such Consumption. Denote the average unit cost of producing and distributing the good by C, which i s assumed to be constant across al l residential consumers. If the average quantity consumed by subsidy recipients i s Q ~ T , and the average expenditure on water or electricity (i.e. the bill paid by subsidy recipients) i s &T, then the average rate of subsidization i s Rqr = 1 - &qT/ ( Q ~ T * C). Generally, consumption subsidies are designed to try to be progressive. JBT tariffs, for instance, are designed so that low levels o f consumption pay a lower price. Th is certainly helps poor households since low welfare i s associated to low levels o f consumption. However, since IBTs are subsidies to consumption, bigger consumer (generally non poor households) benefit as well for a low price charged to a portion o f their overall consumption. The last subsidy design factor, which i s more in the control of the beneficiaries, i s the average quantity o f water or electricity consumed by subsidy recipients (Q~T). The quantity consumed by the beneficiaries determines the total value of the subsidy each will receive. If a household i s eligible for a consumption subsidy, the more they consume consumed, the larger the nominal value o f the subsidy i t receives. The average value o f the subsidy per beneficiary i s calculated as the quantity consumed by the household times the subsidy rate (Le. R ~ T * Q ~ T * C ) . The average subsidy received by poor subsidy recipients would be R ~ ~ T * Q ~ ~ T * C . In general, one would expect that higher welfare i s associated with larger levels of consumption. It can be shown that:

Thus the targeting performance parameter i s the product o f seven ratios: access, uptake, targeting, rate o f subsidization, and quantity consumed. Typically, one would expect that the ratio of the beneficiary incidence parameters (B) to be below one, because the ratio of access rates (A) will in most cases be lower than one (the poor tend to live in areas with lower access rates than the population as a whole), and that the ratio of the usage or “take-up” rates for the service (U) will also be lower than one (when access i s available in a neighborhood or village, the poor are less likely to be connected to the network than the population as a whole). Where coverage i s not universal, this “access handicap” can be expected to work against the targeting of subsidy benefits to the poor. Subsidy design factors need to overcome the access handicap if the subsidy distribution i s to be progressive. To be progressive, the product of the ratios of the subsidy design factor ratios much be larger than 1. The product of the R and Q ratios -- @PIT/ R qT) * (Q~IT/ Q VT) - i s greater than 1 if the average value o f the subsidy received by poor subsidy recipients exceeds the average value accruing to al l recipient households. If less than 1, the opposite i s true. Source: Angel-Urdinola and Wodon (2006).

49

4.23 Clearly, utility subsidies before the tariff increase of June 2006 were less well targeted than most other types of public transfers in Cape Verde. Figure 4.3 displays the value of the targeting performance indicator SI as well as the so-called errors of exclusion for different public programs apart from subsidies on water and electricity. The error o f exclusion provides the share o f poor households who do not benefit from a subsidy. Fol lowing the methodology used by the National Institute o f Statistics in Cape Verde, a household i s considered poor if its per capita consumption falls below the official poverty line (equivalent to $LC 43,249.8 per capita per year). With that poverty line, roughly 37 percent o f the population i s considered as poor (equivalent to 28 percent o f households). The results suggest that public transfers related to primary education, social assistance, minimum pensions, and health are quite pro-poor (i-e., the value o f C! i s greater than one, meaning that the average benefit for a poor person is larger than the average benefit in the population as a whole). However, with the exception o f primary education and health to some extent, the coverage o f some programs i s very limited. In particular, as suggested by the errors o f exclusion, less than 10 percent o f the overall poor benefit from social assistance or minimum pensions. All other public transfers included in Figure 4.2 (secondary education, Public Study Fund, electricity and water subsidies, reform pensions, and tertiary education) display values o f C! lower than one, suggesting that resources are (in proportion) more heavily allocated to non-poor households. A s described before, l ow values of C! are associated to bad targeting performance, But clearly, water and electricity tariffs were among the worst targeted to the poor.

Figure 4.2: Indicators of Targeting Performance

Primary Educabon

Social Assistance

Minimum Pension

Health

Secondary Educatbn

Public Subsidies

Study Fund

Electnclty

Reform Pensbn

Ternary Education

Water

0000 0200 0400 0600 0800 1000 1200 1400 1600 1800

Source: Wodon, Angel-Urdinola and Echevin (2006).

4.24 A key reason for the poor targeting performance of utility subsidies i s that access to electricity and water i s much higher among rich than poor households. According to the nationally representative 2001-02 household survey, access to electricity i s high in Cape Verde,

50

with 59 percent o f households declaring electricity use. For piped water, the proportion is lower, at 32 percent. Yet the proportions with access are lower among the poor, as shown in Table 4.7.

Table 4.7: Access to Electricity and Water by Decile, 2001-2002 ~~

Decile Electricity Water Households not Share of

Share of declaring payment as Households with Households not declaring Households with share of households access payment as share of households

access with access with access 1 0.207 0.444 0.083 0.4578 2 0.358 0.316 0.141 0.2766 3 0.406 0.416 0.158 0.4051 4 0.581 0.250 0.220 0.1818 5 0.523 0.258 0.251 0.2948 6 0.626 0+.235 0.255 0.2549 7 0.692 0.244 0.388 0.1211 8 0.775 0.227 0.429 0.2075 9 0.839 0.186 0.576 0.1649 10 0.913 0.169 0.735 0.1796 All 0.592 0.275 0.324 0.254

Source: Wodon, Angel-Urdinola and Echevin (2006).

4.25 Furthermore, a high share of households declaring access to these services also i s not actually paying bills to Electra. If the non-payment observed in the surveys corresponds to fraud, then i t can be estimated that Electra loses a maximum o f 27 percent o f its residential electricity revenues, and 23 percent o f water reven~es.~’ These are maximum levels because poor households who are more likely to not pay formal bi l ls also tend to consume lower amounts o f electricity and water. In addition, some such households pay fees to neighbors who are formally connected to the network, so that the consumption i s actually paid to Electra by neighbors at prevailing prices.

4.26 The consumption of electricity and water i s lowest among the poor. Table 4.9 presents monthly total household expenditure per capita as wel l as kilowatts-hour o f electricity and cubic meters o f water consumed by expenditure decile. The expenditures for water and electricity are respectively 3 percent and 4 percent o f total expenditure for those who are using the services. At the tariff structure prevailing at the time o f the survey, only households in the f i rs t decile consume on average less than 40 kWh o f electricity per month. By contrast, households in the first seven deciles consume o n average less water than the current minimum “lifeline” o f 6 m3 per month.

4.27 I t follows that very little of the large subsidies for electricity reach the poor, not only because they have low access, but also because those who do have access consume only a fraction of that of the non-poor. Consumption per household in the top income decile i s 7 times higher than in the bottom decile. The same i s true for water: the top income decile consumes five times the quantity than the bottom decile. In short, the large subsidies for electricity are very poorly targeted.

42 This i s based on the Household survey. According to ELECTR4 (Annual Report 2004) the total losses (fkaud and technical) are about 18.5 percent on average..

5 1

Table 4.8: Consumption of Electricity and Water (Users Only) Decile Electricity Water

Expenditure Expenditure Consumption, YO of per CVE kWWmonth per YO of per capita CVE m3/month capita

per capita household expenditure per capita per household expenditure 1 43.88 29.92 3.33% 50.62 2.29 3.61% 2 76.61 48.12 3.56% 74.83 3.60 3.54% 3 133.67 73.23 4.86% 131.03 5.10 4.70% 4 146.17 72.25 4.21% 91.67 3.62 2.63% 5 198.40 89.92 4.64% 121.19 4.02 2.83% 6 245.59 107.36 4.58% 140.52 4.36 2.62% 7 340.05 133.41 5.11% 229.41 6.72 3.42% 8 380.41 136.97 4.37% 251.51 7.15 2.89% 9 502.40 149.56 3.95% 320.95 6.95 2.51% 10 927.14 204.04 2.59% 679.35 11.32 1.85%

Average 299.43 104.478 4.12% 209.1 1 5.51 3.06%

~ ~~

Other fuels for iluninatin

Kerosene

Gas (other bottles)

Gas (bottles 13kg)

Gas (bottles 5kg)

Gaz (pipe)

0 0.2 0.4 0.6 0.8 1 1.2 1.4

E.2 Butane and Kerosene

4.28 None of the subsidies applied to butane were well targeted. The rationale used for subsidizing Butane is because i t i s primarily used for cooking. The rationale for butane subsidies was not just to assist the poor, but also to encourage substitution from dwindling fuel wood resources. This was recommended by the ESMAP (1990)43 and subsequently adopted by the Government. However, the targeting indicator l'2 indicates that Butane subsidies were benefiting more population as a whole than the poor (Figure 4.3).44

Figure 4.3: Targeting Subsidies per Type of Fuel

4.29 The kerosene subsidy i s well targeted, and it represents a very small amount. According to the targeting indicator l'2 subsidies to kerosene are wel l targeted, meaning the benefits for the poor are much larger than for the population as a whole. Kerosene i s consumed by a very small population, believed to be very poor, who use kerosene as the principal source o f lighting. The total consumption was just 660 tons in 2005 (2800 in 1988). Furthermore, in 2005,

43 ESMAPAVorld Bank, Cape Verde Household Energy Strategy Study, Report 110/90, Washington, DC, February 1990. 44 Once the survey o n electricity and gas i s completed, a revision o f th is analysis could be considered given the l ike ly environment impact o f an increase in the price o f gas.

52

the subsidy to kerosene corresponded to only C V E 34 mi l l ion (0.12 percent o f the total revenues and 0.04 percent o f the GDP).

4.30 What emerges from the targeting analysis i s that only the elimination of the kerosene subsidy will have impact on the poor. The eliminating o f the kerosene subsidy will, however, have a significant impact on the poor households that use kerosene for illumination and for cooking (only 1 percent o f the households report45 using kerosene for cooking). International experience suggests that the poorest households tend to adapt to kerosene price increases by purchasing less, increasing their total monthly kerosene expenditure very little (for example, in the Philippines, recent survey results show kerosene expenditures by the poor have remained largely unchanged since 2001 despite price increases in rural areas o f 100 percent). The sharp increase in kerosene price, from C V E 45 to C V E 74 (66 percent increase) wil l therefore mean a deterioration o f the quality o f l i f e for the poorest families in Cape Verde. For the very few families that depend o n kerosene for cooking in Cape Verde, one may expect that many will shift to fuel wood.

4.3 1 Estimated results for the change in the poverty headcount suggest that the impact o f the recent policy measure of eliminating the subsidies will have little effect on poverty. Table 4.9 shows that with the increase on petroleum products as a result o f the elimination o f the subsidies, per capita expenditure wil l be reduced by 0.01 percent. The total change in poverty is estimated to be 0.5 percent.

Table 4.9: Change in the Per Capita Expenditure and Poverty Headcount, Before and After Elimination of Subsidies

Rural Urban Total Per capita expenditure before 76 139,7 198 523,O 145 906,s Per capita expenditure after 75 595,s 196 313,7 144 447,3 Poverty headcount before (%) 51,1% 25,0% 36,7% Poverty headcount after (%) 5 1,2% 25,9% 37,2% Change in poverty (%) 0,1% 0,9% 0,5%

F. IMPACT OF THE ELIMINATION OF THE SUBSIDIES ON THE BUDGET

4.32 Table 4.10 shows our estimate of subsidies in the absence of policy change, estimated at CVE 1.7 billion in 2006 (on a whole year Were these subsidies entirely eliminated (see row 1 l), then the fiscal space that would be created would be approximately 9 percent o f the current expenditure or 2 percent o f the GDP.

4.33 Unless the ARE increase tariffs further, there will remain a substantial liability to the Government to cover Electra’s revenue requirements, with a corresponding reduction of the fiscal space that has in fact been created. Failure to grant Electra adequate tariff

45 Cape Verde Household Survey. 46 I t i s assumed that oil prices will remain stable. In 2004 and 2005 a significant part of the total subsidy was attributable to regulatory lag during persistent upward trends in international oil prices (in 2005, CVE 460 million of the total estimated subsidy of CVE 1.73 billion was attributable to this cause: the 2006-2009 forecast of Table 4.1 1 assumes this at zero); and, Table 4.1 1 also assumes a 5 percent growth in Electra’s fueloil requirement, but assumes gasoil consumption remains constant: there i s likely to be some price elasticity effect so electricity demand growth wil l be less than would be the case at subsidized prices. Gasoil consumption has leveled off recently as the larger generating units have been converted to fueloil.

53

increases to meet its increased revenue requirements will require a recapitalization (with impact o n the budget as the Government is a stakeholder) or originate another issue o f tariff deficit, similar to the situation in the past. As shown in Table 4.10, without this further tariff increase the fiscal space i s reduced to CVE 340 million. In such a case the fiscal space created in 2006 would represent only 2 percent o f the current expenditures or 0.3 percent o f the GDP.

4.34 Given the l imited time and resources available for this review, Electra's finances have not been studied in sufficient depth to permit precise estimates, and to examine the fiscal implications over a wider range o f international price and macroeconomic scenarios. Nevertheless, though subject to some uncertainty, these estimates are believed to be broadly correct, and sufficiently robust to draw broad pol icy conclusions.

The estimates of Table 4.10 require further study.

Table 4.10: Projection of Subsidies - in the Absence of Policy Change (CVE Million)

2005 2006 2007 2008 2009 1 Butane [Shell] 154 162 170 178 187 2 3 =1+2 4 5 =3+4

6 7 8 9 =6+7+8 10 =5+9

Butane [ENACOL](4) Total Butane Kerosene Total, non-ELECTRA

ELECTRA Gasoil, Direct Government subsidy( 1) Gasoil, subsidy to o i l companies (2) Fueloil subsidy to o i l companies (3) total ELECTRA Total subsidy absent policy change

23 24 25 26 28 177 186 195 205 215 34 34 34 34 34

211 220 229 239 249

328 328 328 328 328 285 285 285 285 285 445 467 491 515 541

1058 1080 1104 1128 1154 1269 1300 1333 1367 1403

Projection of Fiscal space created by elimination of subsidy 1 1 =10 With 100% o f the required ELECTRA tari f f increase 1300 1333 1367 1403

13 2006 adjusted for implementation date (5) 758 1333 1367 1403

15 =13-14 Actual fiscal space increase (without further tar i f f increases) 340 597 612 628

(Le. with a total tari f f increase o f 50%)

14 Less amount corresponding to shortfall in tariff increase (6) -418 -736 -754 -774

1) This payment is pa id directly t o ELECTRA (and covers the difference between the actual purchase price (in the last h a l f o f 2005, the difference between CVE 75/litre and CVE 37. Mi t re) (2) This payment i s pa id t o the o i l companies to cover the difference between the regulated price o f gasoil and the actual international price o n ELECTRA's gasoil purchases. (3) T h i s payment i s made to the o i l companies to cover the difference between the regulated price o f fueloi l (at which ELECTRA makes i ts fueloi l purchases), and the international price o f fueloil. (4) assuming ENACOL's share o f subsidized butane sales in small bottles i s the same as that o f Shell. (5) Taking in to account that the ELECTRA price increase was implemented only June 1, so amount shown i s 7/12"' o f the whole year estimate for 2005, CVE 1300. (6) A 50% tar i f f increase wou ld offset the entire increase in fuel price =loo% o f the total subsidy (e.g. o f CVE 1,154 m i l l i on in 2009). Since only a 22.4 percent tar i f f increase has been implemented, a further subsidy increase o f 27.6 wou ld b e required (if we consider the o i l prices o f the first quarter o f 2006 and the long te rm financial sustainability o f the company- taking in consideration the assumptions in the ARE model wh ich require validation). Hence the fiscal space decreases by 0.276D.5 x 1154 = 774 (in 2009).

54

G. THE IMPACT ON ELECTRA SUSTAINABILITY AND RECOMMENDATIONS

4.35 In addition to the tariff adjustment necessary to accommodate the increase in fuel prices (to eliminate the explicit fuel subsidy), a further tariff increase will be required to bring the tariff to a level that will permit adequate self-financing ratios to meet Electra’s investment requirement. The viability o f the energy sector, with al l the potential implications for the rest o f the economy, depends o n ARE actuation, which has the full authority and responsibility over tariff setting.47 As the increase in water and electricity tariffs o n June 1 were below the price increase required, this suggests that ARE may plan to adjust tariffs by stages. If that i s the case it should be ensured that the tar i f f adjustments would be sufficiently high to enable full recovery o f the fuel price increases by the end o f year 2006. What is important, in light o f the exogenous uncertainties, is transparency in the assumptions that are made by ARE that underpin the final tariff level.

4.36 The present tariff structure for Electra’s domestic consumers in the case of both water and electricity i s known as an inverted block structure (IBT): before the tariff increase o f June 2006, consumption at or below 4 0 k W m o n t h was costing CVE 16.5/kWhY consumption above 4 O k W m o n t h C V E 20.5kWh. Angel-Urdinola and Wodon (2006) show this structure to be poorly targeted, because even those consumers who consume more than 4 0 k W m o n t h (namely al l deciles except the bottom decile) benefit from the lower tariff for the first 4OkWh. The fact that al l tariff bands have been increased proportionately does not change fundamentally the poor targeting performance o f the implici t subsidies that continue to be provided, since the tariff increase i s not sufficient to ensure the financial viability o f Electra in the medium to long run.

4.37 One possibility to reduce the cost of the tariff while better protecting the poor would be to shift from an IBT tariff structure to a VDT tariff structure. Under a VDT tariff, only those households who consume less than the threshold for any tariff band benefit from the lower tariff under that band. This would mean that for electricity, for example, only those households who consume less than 40kwh would pay the lower price for that consumption level (in an IBT structure, a l l customers pay the lower rate for their consumption below the lifeline o f 40kwh). Shifting from an IBT to a VDT, and possibly also reducing the level o f the lifeline o f 40 kwh to 30 kwh or 20 kwh (and doing the same for water) might generate enough resources to provide an even lower rate per kwh for the l o w consumers. In fact, the discount offered o n the first block i s relatively small when compared to international practice, where discounts o f as much as 50 percent are given in some cases, though to a much narrower first block: for example, in the Philippines, the consumption threshold that qualifies for a 50 percent lifeline discount is 1 5 k W m o n t h (though at the other extreme there are examples o f thresholds o f as much as 2 0 0 k W m o n t h as in Yemen). In Cape Verde, as shown by Angel-Urdinola and Wodon (2006), shifting from an IBT to a VDT would help substantially to improve the targeting performance indicator Q for electricity, although the gains for water would be lower.

4.38 Another possibility for energy and water sector assistance to the poor i s to improve access: only 20 percent o f the bottom income decile presently have access to water, as compared

47 According to ARE there i s the in-house expertise to carry out the relevant analysis for the price adjustments.

55

to 90 percent o f the top decile (and only 8 percent o f the bottom decile have access to piped water, as compared to 73 percent for the top decile). Subsidizing the electricity and water consumption o f households currently connected to the network i s much less pro-poor than helping households who are not connected to become connected. In addition, it can be shown that access to electricity and water tends to reduce the unit cost o f energy and water for households, and has a range o f other beneficial outcomes (in reducing chi ld malnutrition and mortality, and improving school performance, for example).

5. PUBLIC PENSIONS AND HEALTH INSURANCE SCHEME

A. BACKGROUND

A.l Pensions

5.1 There are two contributory pension schemes for formal sector workers in Cape Verde. The “Administraqgo Phblica” (AP) scheme, which, according to the c iv i l service census, has an estimated 14,600 contributors (Le., police, teachers, and other c iv i l servants), and the “Instituto Nacional de Previdsncia Social” (INPS), which covered approximately 36,400 private sector workers in 2004 (Table 5.1).48 The INPS i s the National Social Security Institute o f Cape Verde created in 1983 to administrate a compulsory social security scheme (not only covering pensions, but health benefits, family allowances and maternity benefits) for dependent workers in the private sector. The institution was designed to be self-financing through social contributions and investments. However, as a National Institute, the State is ultimately responsible for the provision o f benefits and for the proper administration o f the programs.

Table 5.1: Main Indicators of Public Pension Systems (2004)

Indicator INPS AP (Police and Teachers ) AP (others) Contributors 36,350 8,038 6,542 Beneficiaries 4,705 1,452 1,191

Old-age 1,972 1,118 916 Invalidity 1,063 Survivors 1,670 334 275

System dependency ratio (%) 13 18 18 Pension expenditures ( % o f GDP) 0.8 0.7 0.5

Source: INPS and the Ministry o f Finance and Public Administration.

5.2 one also

Together, the AP (for civil servants) and the INPS (for private sector workers) cover quarter of Cape Verde’s labor force and pay pensions to about 7,300 people. There are three non-contributory pension schemes. The “Pensgo Social Minima” (PSM) is a social

assistance scheme that pays benefits to approximately 5,000 elderly individuals that meet income and age criteria established under a 1995 law, the “Pensgo de Solidaridade” i s paid to more than 9,000 former FAIMO workers over age 60 subject to a (vague) income test and an eligibility condition o f having worked for, at least, ten years under the program. The “Pensgo de Estado” i s paid to individuals with relevant contributions to the arts, culture, and the independence movement o f Cape Verde and covers the residual population.

48 Although I N P S has around 36,400 workers registered, less than 30,000 were actually contributing to it in 2004.

56

5.3 The INPS i s a young scheme, both chronologically and demographically. I t began operating in 1983. In 2004, there were 18 contributors for every old-age-pension beneficiary and 8 for every pension recipient (including those on invalidity and survivors’ benefits). In contrast, the mature pension systems o f older countries, such as those in Europe, have only 2 or 3 contributors for every beneficiary.

5.4 Temporarily, the young age o f the INPS i s highly favorable for the scheme’s finances. The pension program i s currently running a surplus and the INPS has accumulated a substantial reserve o f assets, equivalent to over 15 years o f pension expenditures. Nonetheless, pension expenditures under the scheme face 3 pressures, which will have a profound impact in the short- and medium-tern: (i) the natural maturing o f the INPS, which i s just 22 years-old at the moment; (ii) demographic change, such as longer l i fe expectancy, resulting in population aging; and (iii) a generous benefit promise and health coverage.

5.5 The AP i s a more generous pension scheme than the INPS. The AP pension plan has a retirement age for teachers and police o f 55 for both male and female and 60 for other members while the retirement age for the I N P S i s 65 for men and 60 for women. This i s particularly important as about ha l f o f the c iv i l servants in Cape Verde are teachers, which, along with the police, are covered by the AP pension scheme. As for the basic and accrual replacement rates4’ there are no major differences between the two pension schemes, but the maximum replacement rate i s 20 percentage points higher in the AP than in the INPS. Moreover, while for the AP the number o f years considered for the wage base calculations is only one year for both old-age and invalidity retirement, for the I N P S system it i s 5 years for invalidity and 10 years for old-age. Finally, total contribution rates, as a percentage o f wage bill, are 15 percentage points higher in the I N P S than in the AP. These differences make the AP system a more paternalistic one, which provides enrolled individuals more benefits than those granted by the I N P S (Table 5.2 and Table 5.3).

49 For a detailed explanation on pensions’ terminology, see Table 1 in Annex 5.

57

Table 5.2: Main Rules of Public Pension Systems (2004) Indicator Old INPS New INPS AP (Police and AP (others)

Legal Male 65, Female Male 65, Female 55 Male and 60 Male and Retirement age 60 60 Female Female Required length of service for basic

(April 2004) Teachers)

3 3

15 5

10 10 replacement rate:

Old-age Invalidity Survivors 3 3

Basic replacement rate 24.5% 30.0% 3 1 .O% 29.0% Accrual replacement rate ate 1.5% 2.0% 3.1% 2.9% Maximum replacement rate 85% 80% 100% 100% Number of last years for wage base calculations

Old-age 3 10 1 1 Invalidity 3 5 1 1

Indexation of wage base for pension Not indexed Price indexed Not applicable Not applicable calculation Post-retirement Indexation Ad hoc Price indexed Wage growth Wage growth

Source: INPS and the Ministry o f Finance and Public Administration

Table 5.3: Contribution Rates for Social Insurance (in percentage)'

Social insurance programs: INPS AP2 Old age, disability, and survivors 3% from employee and 7% from 7% from employer (gov. budget)

Sickness and maternity

Work injury

employer 4% from employee and 4% from employer 2% (or 6% from employer, depending on the worker)

1% from employer (gov. budget)

Family allowances 3% from employer Total 23% 8 Yo

Notes: 'Contribution rates are indicated as percentage of gross monthly earnings. 'Contribution rates applied for all civil servants until January 2006. Source: INPS and the Ministry o f Finance and Public Administration

5.6 The Cape Verde scheme i s very generous by international standards (Table 5.4). Cape Verde's accrual replacement rate i s higher than those in other countries. Also, the number o f years' service required for a full pension i s lower than that for many other civil-service schemes. The civil-service scheme in Cape Verde pays a pension o f 100 percent o f final salary after 34 (or 35 for some c iv i l servants) years o f service.so

Teachers and police are subject to slightly different rules than the rest of civil servants and they amount to 55 percent the contingent of civil servants. Teachers are eligible to full pension with 32 years of service.

58

Table 5.4: Accrual Rates and Maximum Replacement Rates in Civil-Service Pension Schemes in Selected Countries

Country Accrual rate Maximum replacement rate ( Y O ) (YO)

Cape Verde 3.10 12.90 100 95 Spain -

Luxembourg 1.667 83 Austria 2 80

80 Italy - Portugal 2 80 Belgium 1.667 75 France 2 75 Germany 1.875 75 Sweden 0.3312.17 73 Senegal 2 70 Netherlands 1.75 70 Greece 1.714 69 Mauritius 2 67 United Kingdom 1.25 67 Switzerland - 65 Finland 1.5 60 Denmark 111.5/1.75 57

Source: Palacios and Whitehouse (2006)

5.7 Cape Verde has already a relatively high contribution rate (23 percent) when comparing with the average rate of the region (16.7 percent) (Table 5.5). However, i t is still around the average or even below it when compared with other regions in the world, being the average o f North-Africa and the Middle-East 23.2 percent, 40.0 percent in Eastern and Central Asia, and 32.2 percent in high-income OECD countries.

59

Table 5.5: Contribution Rates for Social Insurance Programs in Selected African Countries,

(Percentage of Gross Wage)

Country For pensions For all programs Employer Employee Total Total

Benin 6.4 3.6 10.0 21.5 Burkina Faso 4.5 4.5 9.0 23.5 Burundi 5.5 3.0 8.5 10.5 Cameroon 4.2 2.8 7.0 16.9 Cape Verde 7.0 3.0 10.0 23.0

Chad 4.0 2.0 6.0 14.5 Congo (Kinshasa) 4.0 4..0 7.0 13.0 Congo (Brazzaville) 3.6 2.4 6.0 18.5 Cote d’Ivoire 2.4 1.6 4.0 13.0 Equatorial Guinea 26.0

Central Af. Rep. 3.0 2.0 5.0 20.0

Ethiopia 6.0 4.0 10.0 10.0 Djibouti 4.0 4.0 8.0 20.0 Gabon 5.0 2.5 7.5 22.6 Gambia 19.0 0.0 19.0 Ghana 12.5 5.0 17.5 Guinea 4.0 2.5 6.5 23.0

Liberia 3.0 3.0 6.0 7.8 1 .o 4.5 14.0 3.6 9.0 19.0 1 .o 3.0 16.0 1.6 4.0 17.0

7.5 3.0 6.0 5.0

10.0 4.8 17.0 33.0 7.0 24.0 26.0 5.0 10.0 0.0 20.0

Togo 3.6 2.4 6.0 20.5 Uganda 10.0 5.0 15.0 Zambia 10.0

Kenya 5.0 5.0 10.0

Madagascar Mali Mauritania Niger Nigeria Rwanda Sgo Tome and Princ. Senegal Sudan Swaziland Tanzania

3.5 5.4 2.0 2.4

3.0

7.2 17.0 5.0

10.0

Zimbabwe 6.0 Source: Social Security Programs throughout the World - SSA (2004).

A.2 Health Insurance Scheme

5.8 Public sector health services are nowadays basically free in Cape Verde, but there has been a shift in public hospitals in order to increase cost recovery by raising basic fees. Currently, a poverty certificate, easily obtained at the municipalities, i s enough to get free access to most services at public hospitals as well as free drugs (from a basic list). However, there are some services that one cannot be exempted o f payment, even with the poverty certificate. The trend is to increase fees and the l ist o f such services. Further patient participation in health expense is expected as recent data from the Demographic and Health Surveys (DHS) confirm that poor health and high out-of-pocket spending are not major contributors to poverty. Thus, the country’s long-term health financing pol icy is to shift to social insurance schemes, which would allow the Government financing o f health services to be more focused on the poor.

60

5.9 Public expenditures on health services in Cape Verde amounted to 4 percent of GDP in 2004. Public expenditures cost are directed by the Ministry o f Health and by the INPS. The INPS reimburses the Ministry o f Health for services rendered to i t s affiliate but supposedly at a lower cost. There has been a dialogue between the INPS and the Government on adjusting the transfers (Table 5.6).

Table 5.6: Health Expenditures, 2001-2004 (Percentage of GDP) Entity 2001 2002 2003 2004 Ministry o f Health 2.04 2.15 1.95 2.79 I N P S 1.21 1.14 1.16 1.16 Private 1.21 1.20 1.15 1.18 Total 4.46 4.50 4.26 5.13

Source: INPS and Ministry o f Health

5.10 After a long period of continuous increase, INPS health expenditures are getting under control. Between 1983 and 2004, health expenditures presented a 46-fold increase (Table 5.7). In 1983 and 2004, health expenditures represented 33 percent and 57 percent o f total expenditures, respectively. The year 2005 was a turning point and health’s share in total public expenditures reduced to 47 percent. Since 1993 the health component has been running a deficit and a cross-subsidy from the pension component was underway. In 2005, there was a definite turn and the health component presented a small surplus, associated with changes in the rules for delivering services, including a co-payment. According to INPS projections, even doubling the value o f the transfers to the Ministry o f Health (to cover real costs), with the new rules there would st i l l be a small surplus in the INPS health component.

Table 5.7: INPS: Total and Health Expenditures and Revenues, 1983,1993,1998,2004 and 2005 (CVE Thousands)

1983 1993 1998 2004 2005 Total Revenues 22 1,263 870,040 1,487,324 2,632,000 2,900,633 Total Expenditures 63,644 488,113 932,740 1,695,386 2,035,393 Health Revenues 48,101 189,139 5 17,330 9 14,206 1,008,841 Health Expenditures 20,801 27 1,464 53 1,994 967,43 1 972,783

Source: INPS

5.11 Health benefits for INPS affiliate are more generous than those received by civil servants. There has been pressure from civ i l servants for integration mainly because o f the differences in health services. This raises concerns with respect to health costs, since the Government alone will cost the transfers to INPS corresponding to pensioners (8 percent levied o n benefits) and the “Taxa Social Unica,” which is currently part o f the Government’s budget.

B. RECENT REFORMS IN THE PENSION SYSTEM

B.l Legal Framework

5.12 The integration of civil servants into the INPS has been discussed for a few years and, in February 2006, a new integration law was approved. In 2001 a framework law established the principle of such integration. Civ i l servants’ union representatives have been pressuring the Government to implement i t because o f the advantages o f INPS in terms o f health

61

coverage, especially pharmaceuticals that are fully reimbursed to INPS members through a card system. However, several obstacles have existed to this integration, the main one being the potential fiscal costs implied by a government contribution as employer. Despite this constraint and the lack o f data (individual records) to analyze in depth and quanti@ the fiscal impact o f integration, a “Decreto-Lei No. 21/2006” was passed to be implemented at the moment when the budget will be approved in 2006. W h i l e the new law grants that al l c iv i l servants and INPS beneficiaries receive the same assistance in terms o f health, the pension systems o f I N P S and AP were not harmonized and the differences between the two systems have remained.

5.13 The new integration law treats active civil servants and pensioners that started working or retired before December 31Sf, 2005, differently from those hired after December 31, 2005. The f i rs t ones are subject to the same o ld rules for pensions, while the second ones will be completely integrated in W S . Regarding health care, the new law establishes that al l c iv i l servants and pensioners are INPS beneficiaries. Figure 5.1 shows that in 2040 there will be no contributors in the AP system, while there will be AP pensioners until 2075. The number o f AP pensioners will peak in 2049 with 8000 pensioners and will decline after that.

Figure 5.1: Impact on the number of contributors and pensioners of the AP pension program with integration

35

pensioners Without

30 contributors Without integration

25

Number of contributors after integration

I 20

pensioners after i 10 l5

5

5.14 will include:

The Government’s responsibilities as an employer in relation to the integration law

i. 23 percent o f the new c iv i l servants covered wage bill, which will be transferred to the I N P S (8 percent for health and 15 percent for the rest o f the programs, including pensions);

ii. 8 percent o f the current c iv i l servants covered wage bill, which will be transferred to the W S (for the health program);

iii. 8 percent o f the current pensioners pension payments (former c iv i l servants), which wil l be transferred to the INPS (for the health program); and

iv. Pension payments (former c iv i l servants, and the current c i v i l servants that will eventually s t i l l retire under the AP program).

62

The pensionhealth program o f AP will disappear when the last current c i v i l servant dies.51

5.15 On non-contributory pension schemes, a Decree passed on March 6th 2006 (“Decreto-Lei No. 24/2006”) creating new benefits that will, over time, replace the three previous systems. Under the new legislation, individuals with income below the extreme poverty threshold (calculated by INE) are eligible for: (i) basic pension if older than 60 years o f age; (ii) invalidity pension if between 18 and 60 o f age and presenting a disability above 75 percent; and (iii) survivors pension if widow(er) of someone receiving i or ii. While the merger and streamlining o f the o ld non-contributory schemes is welcomed, challenges in implementation o f ‘social pension’ schemes can be significant and costs will rise over time as the population ages.52

c. THE IMPACT OF THE NEW LAW

5.16 Since Cape Verde i s a small country, there are economies of scale associated with the integration of the two pension systems. Hence, a single scheme should be cheaper than two separate schemes because: (i) wasteful duplication o f administrative functions can be eliminated; (ii) mobil i ty between the public and private sectors i s inhibited, reducing labor-market flexibility; (iii) equity between public and private workers i s increased; and (iv) transparency in comparing services’ terms and conditions between public and private-sector jobs i s improved.53

5.17 In consolidated terms there i s a gain with the integration of the two pension systems as INPS provides fewer benefits than the AP (Figure 5.2), but the integration aggravates the unsustainability of the INPS. With the integration there will be a savings in total expenditures starting in 2017 o f about 0.1 percent o f the GDP. The magnitude o f the savings increase over time and during the period 2060-2080 it is projected to reach o n average 2 percent o f the GDP. Considering the INPS solely, the integration aggravates the unsustainability o f the system as it adds new contributors and, subsequently, beneficiaries to a system that i s already unbalanced. Despite some recent parametric changes o f the INPS pension the system will be in deficit, because o f a generous benefit formula and a relatively l o w contribution rate (10 percent) and as a result a deficit i s expected to appear around 2030.

’ I Th is i s expected to happen around 2070. 52 See Palacios and Sluchynsky (2006). 53 For a discussion o f the advantages o f integrating c iv i l service and national pension schemes, see Palacios and Whitehouse (2006). ’‘ Earnings have been gradually changed from the best 3 years’ salary during the final 5 years o f participation in the scheme, to 15 years (INPS benefit formula is: Pension = Earnings Base * [30% + (0.015% x Years of Coverage)]. The required 3 years o f membership to receive a pension has been changed to 15 years and the maximum replacement rate from 85 to 80 percent.

63

Figure 5.2: Combined pension expenditures of AP and INPS before and after integration

C.l Preliminary Estimates of the Integration on the INPS System"

Pensions

5.18 Integration of new civil servants will add to the number of INPS contributors for many years. Figure 5.3 shows the impact on the number o f contributors and pensioners in the INPS. In the future, i t has been assumed there w i l l be no expansion o f coverage in the private sector, and an increase in the total number o f public servants equal to the growth o f the p o p ~ l a t i o n . ~ ~ These data are combined to calculate the number o f new entrants to the I N P S that would otherwise joined the public-sector scheme. By 2049, the whole working public sector i s in the I N P S and the absolute number o f additional contributors reaches nearly 24,600, adding 26 percent to the total affiliates o f INPS. At this point, the profiles o f contributors both pre- and post-integration flatten out, reflecting the maturity o f the system.

Figure 5.3: Impact on the Number of Contributors and Pensioners in INPS with Integration

100

40

20

'' The lack o f accurate data adds considerable degree o f uncertainty to the results, however a big picture o f the impact o f the

56 It has been assumed no growth for c iv i l servants until 2008 integration can sti l l be presented.

64

5.19 The impact of integration on the number of INPS pensioners follows a different profile. There i s no impact until 2037 because only new entrants to the public service are covered by INPS and they begin to retire after 2037. However, the addition o f extra pensioners i s then quite rapid. By the end o f the forecast horizon, there are 25,300 extra pensioners, representing an increase o f nearly 30 percent o n the current INPS (Figure 5.3).

5.20 With or without integration, the INPS system would begin a cash-flow deficit by the early 2040s, meaning that the contribution revenues will be insufficient to meet pension expenditures. The pension promises are readily affordable now because the scheme demographic position i s favorable, but the demographic situation will deteriorate rapidly in the next few years and further reforms are o f critical importance. Contribution revenues wil l continue to increase over the medium-term, but the powerful combination o f the maturing o f the system and demographic change mean that benefit expenditures wil l increase quickly, more than doubling relative to gross domestic product (GDP) in a few years.57

5.21 The integration improves the short- and medium-term finances of the system because of the extra contribution revenues, but aggravates the long-term deficit (Figure 5.4). In the short-term the integration improves the finances o f the INPS and postpones the beginning o f the deficit (by 5 years). However, since the value o f the pension promise i s much larger than the value o f the contributions received, the long-term balance o f the INPS pension program is worsened. As shown on Figure 5.4, in 2064 the two curves intersect and after that the INPS expenditures as percentage o f GDP start to grow faster. The increase in deficit due to the integration achieves 1 percent o f the GDP by the end o f the period.

Figure 5.4: Impact on the Pension Program Balance of the INPS with Integration

INPS pension program balance with intregratlon

2.0% -

1.0% -

0 0 % -

B -2.0%

f -3.0%

-4.0% -

-5.0% -

-6.0% -

INPS Pension 1 \\

-7.0% ’ Note: These balances include returns on pension fund reserves and the assumed interest rate was 3.5 percent.

5.22 Under integration, the population involved in transitory coverage of health and pension expenditures, will completely disappear by 2075. Active c iv i l servants will decrease

’’ Projections are shown when comparing status-quo with the integration process.

65

from now on and pensioners wil l reach a maximum by 2040. Police, teachers, and other c iv i l servants wil l fol low a very similar pattern (Figure 5.5).

Figure 5.5: Population of Civil Servants from the Old System

1 OAP II -ACTIVE WORKERS 18

.AP I -ACTIVE WORKERS

OAP II - PENSIONERS (AFTER 2006)

.AP I - PENSIONERS (AFTER 2006)

.AP II -PENSIONERS (BEFORE 2006)

.AP I - PENSIONERS (BEFORE 2006)

C.2 Health

5.23 The impact o f the integration on health costs for the INPS will depend on how many civil servants already profit from the INPS system. There i s a generalized bel ief in Cape Verde that most o f the c iv i l servants already take advantage o f the health system provided by INPS. Figure 5.6 presents estimated costs assuming constant per capita expenditure and different level o f leakage. Full costs (no leakage) amount to 0.73 percent o f GDP and cost recovery from active c iv i l servants contribution ("Taxi Social Unica") will amount to 0.58 percent o f GDP. Costs under the hypothesis o f a leakage o f 21 percent (this percentage o f c iv i l servants are already using INPS health services) will match contribution in 2006.'* In reality, costs wil l depend on sex and age and an aging population wil l imply higher per capita expenditures.

Figure 5.6: Health Cost Using INPS Parameters - Different Leakage Assumptions (Percentage of GDP)

-mu con

-- 30% IS ALREADY COVERED

-10% ISALREADYCOYERED

+ 70% I S ALREADY COVERED

-CS CONTRIBUTION

--TOTAL. COWRIBUTION

2000 2010 2mO 2mD 2ouI 2DyI nrSa 2070 ZUW 209) 21113

58 Note that all projections were made using a constant per capita expenditure.

66

5.24 Costs will also depend on real inflation of health services. In every country, health costs have been presenting a cost increase above inflation. Figure 5.7 presents the same simulations o f Figure 5.6 under the assumption o f a 3 percent real increase in health costs. Note that in both scenarios, figures are quite similar.

Figure 5.7: Health Cost Using INPS Parameters - Different Leakage Assumptions, and 3 Percent Real Increase in Health Cost

(Percentage of GDP)

ow

0 8% -FULL COST

+yI% I S ALREADY COVERED

-50% I S ALREADY COVERED

+- 70.h I s ALREADY COVERED

-CS CONTRIBUTION

-TOTAL CONTRIBUTION

07U

0 I X

0 5%

0 4u

0 3%

0 P I

0 1%

O O X XW 8 1 0 W B 830 1(yo B W W70 W 2ow Plm

C.3 Preliminary Estimates of the Integration on Central Government Budget

5.25 There will be a “transition burden” for the Government, which will have to pay public-sector pensions and INPS contributions at the same time. The Central Government wil l transfer monthly to the INPS, 23 percent o f the wage bill o f the new c iv i l servants, plus 8 percent o f the current active c iv i l servants and pensioners (Figure 5.8). In addition it continues paying the pensions to the pensioners and wil l pay the pensions o f the current c iv i l servants when they retire (retire under the AP program) (Figure 5.9). In turn, it will save in the provision o f health care to the c iv i l servants. Given the lack o f health services data, i t is not possible to estimate the amount, but there i s a strong bel ief in Cape Verde that c iv i l servants were benefiting unduly from the INPS. In sum, the already high current mandatory expenditure will increase. The net effect on the budget in the short run i s essentially the transfers to the INPS,” and that corresponds to an increase o f 2 percent o f the current expenditures (or 0.6 percent o f the GDP). Transfers to the INPS, plus pension payments now represent approximately 5 percent o f the total o f current expenditures.60 Total budget transfers start declining around 2040 because o f the ageing o f the population under the AP system (Figure 5.8 and Figure 5.9).

’’ Actual pensioners will die and current civil servants will retire. 6o Non-contributory pensions are not included in this analysis.

67

Figure 5.8: Impact of Integration on the Government Budget

Total budget transfers before Integration (pension payments)

4.0%

Figure 5.9: Impact on the Pension Program Balance of the AP with Integration

before Intearatlon

after intrepratlon

-1.0% 1 b, / -2.0% -

; -2.5% -

-3.0%

AP pension program balance before Intearatlon

-- . AP pen9ilon program balance after intrepratlon

4.5% J

5.26 Furthermore, if further reforms do not take place the deficit o f the INPS will become a liability to the central government. Demographic and financial projections for the I N P S pension system in the medium- and long-term under current policies are unsustainable, with a deficit appearing in 2037 and growing very fast. In 2027, i t i s estimated to be CVE 61 mi l l ion and five years latter i s estimated to reach CVE 10 billion.

D. OPTIONS FOR FURTHER REFORMS

5.27 In order to 'harmonize' both pension programs @e., INPS and AP) and try to improve their fiscal situations, some reforms could be considered and their fiscal impact has been analyzed.

68

The AP Pension System

i. Reform I : a first reform, which would affect new pensioners from Ap, could be shifting the accrual rate from the current level o f 3.1 percent for teachers and police and 2.9 for the rest o f c iv i l servants to the 2 percent (accrual under INPS). On average, this reduces the replacement rate at retirement by nearly one hal f and, consequently, the net present value o f benefits and the equilibrium contribution rate by the same proportion;

ii. Reform 2: a second reform i s to increase the pension age to 65, the same age as the I N P S (for men). By increasing the number o f years o f coverage under the scheme, the replacement rate at the point o f retirement is 15 percent higher. However, because the benefit will be paid for 5 fewer years, the annuity factor and the net present value o f total pension benefits are smaller. Also, members will contribute for 5 more years and so the net present value o f employee contributions i s higher;

iii. Reform 3: a third reform i s to index pensions in payment to prices rather than wages as presently. Although real wages have held fairly steady in the last few years, this will not be sustained into the long-term - real pay will once again begin to rise. I t means that price indexation is a cheaper pol icy in the long term. This is captured by the difference in the annuity factor, which falls f rom nearly 16.0 percent in the baseline case to 12.4 percent with price indexation o f benefits. Again, this would harmonize the civil-service scheme with the INPS; and

iv. Reform I+2+3: a final reform presented here could be the addition o f a l l these changes -lower accrual rate, later retirement, and price indexation o f benefits. These three changes together would vastly improve the long-term finances o f the civil- service pension scheme.

5.28 Despite the fact that the AP pension system will eventually be closed, and the balance will always be negative, until it becomes zero, when the system disappears, there i s s t i l l a possibility of making this transition less costly for the system (Figure 5.10). When looking at the financial gaps (present values o f the system balances) o f the different reform options for A P (2006-2100), the savings could be as high as CVE 51.09 bi l l ion (or 60.4 percent o f 2004 GDP), when considering the described reform package (Reform 1+2+3).61 If the system i s not reformed, deficit could be as high as CV E15.2 bi l l ion (or 2 percent o f GDP by year 2036), against CVE 8.4 bi l l ion or less than 1 percent o f GDP when a reform package l ike the one described here i s considered (Table 5.8).

6' Assumed nominal discount rate of 5 percent.

69

Figure 5.10: Reform Options of AP Pension System (Percentage of GDP)

$ -1.0% - 4 -1.2% -

-1.4% - -1.6% - -1.8% - -2.0% -

0.0%

-0.4 -0.40%

-0.6%

-AP after Integratlon Reform I Reform2

Ta-le 5.8: Financial Gap o

-Reform 3 -Reform 1+2+3

Reform Options Financing gap (present value of system balance)

In % 2004 GDP In Millions CVE AP after Integration - 106,900.74 -126 Reform I -90,241.09 -107 Reform 2 -99,675.79 -1 18 Reform 3 -70,036.27 -83 Reform 1+2+3 -55,806.26 -66

The INPS Pension Program

5.29 Reforms would also be necessary to avoid the need for a call on the government budget to meet future benefits. Also to ensure any pension promises that are made can be kept and deliver a fair allocation o f resources between different generations. One particularity (by international standards) o f I N P S i s the evidence that the generous method o f calculating disability pensions combined with high rates o f approval o f disability claims result in high spending on disability pensions. Furthermore, there i s a strong incentive to use disability as a substitute for the o ld age pension.

5.30 Under current circumstances and given the fact that the INPS pensions program i s not sustainable in the future, one possibility to improve the situation would be increasing the contribution rates at an earlier point than when reserves are exhausted. This would have two advantages over leaving the contribution rate unchanged until then. First, i t would be equivalent to a partial pre-funding o f future pension promises. The point at which the cash-flow o f the system will go into deficit would be delayed, meaning that a larger reserve would accumulate in the meantime. This would then be used to pay pensions as the system matures and the population ages. Many countries have established or increased the degree o f pre-funding o f pension obligations in this way, with reserves being used to smooth the cost o f paying for

70

pensions. This also smoothes the path o f the contribution rate, meaning that future increases can be smaller or avoided altogether.62

5.31 INPS potential reforms affecting only benefit formulas will not have an impact until after 2040, since they would only affect new entrants. Only when these workers begin to retire wil l expenditure slowly change. The full impact o f such reforms would only occur when the complete stock o f c iv i l servants pensioners i s covered by the new rules. However, increasing contribution rates (from 10 percent to 15 percent) will affect the finances o f the system from the beginning, but st i l l such measures wil l not be enough to save the system. In addition, if the system can not be sustainable, there would be a need to provide a cross-subsidy from general revenues. This would be unfortunate, as Cape Verdeans would have to pay for pensions (through, for example, income taxes and/or value-added tax or excise duties) not just the members o f INPS. The fiscal benefits in the medium- and long-term because o f the integration are also considerable.

5.32 The financial sustainability o f the pension system i s also dependent on the investment performance of the INPS. The international experience in this area suggests that this can be a challenge unless governance arrangements and investment pol icy are wel l designed.63 L i ke many developing countries, Cape Verde has few traded securities in which the pension fund can invest domestically. This leads many public pension funds to invest in government bonds, but this might in turn lead to higher government consumption than would otherwise have been the case, undermining the objective o f prefunding at the national level. Alternatively, the pension fund might invest in assets that are difficult to value and illiquid raising the possibility o f misallocation o f savings. Finally, corporate governance problems could arise since the INPS would be the largest single institutional investor in the country. Lacking independence from Government could lead to conflicts o f interest. Investing abroad would allow for diversification from country specific risks and would avoid many o f the problems inherent in domestic investment. In general, the reform package should include the introduction o f good practices for pension fund governance, accountability and investment policy.64

62 Note that this i s the same as the ‘scaled premium’ approach advocated by the International Labor Office (ILO), which again aims to smooth the path of the contribution rate over time by charging the long-term equilibrium contribution rate from the beginning o f the scheme and using this to build up a reserve. 63 See Iglesias and Palacios (2000) and Holzmann et al. (2004) for a review o f the international experience and emerging best practice in this area.

See Cannichael and Palacios (2003). 64

71

Figure 5.11: Effect of Increasing Contribution Rates to the Current Balance of the Pension Program in INPS

contribution rate

Pension program balance with contribution rate 10%

-10.00% I

72

Bibliography

Angel-Urdinola, Diego and Quentin Wodon (2006). “Do Utility Subsidies Reach the Poor? Framework and Evidence for Cape Verde”, mimeo, World Bank.

Banco Central de Cab0 Verde (2002). Relat6rio do Conselho de Administraqiio 2002.

Banco Central de Cab0 Verde (2003). Relat6rio do Conselho de Administraqiio 2003.

Cannichael, Jeffrey and Robert Palacios (2003). “A Framework for Public Pension Fund Management”, Paper presented at the Conference on Public Pension Fund Management, World Bank, Washington DC.

ESMAPNor ld Bank (1990). Cape Verde Household Energy Strategy Study, Report 110/90.

Holzmann, Robert and Robert Palacios and Asta Zviniene (2004). “Implicit Pension Debt: Issues, Measurement and Scope in International Perspective, SP Discussion Paper No. 0403.

Iglesias, August0 and Robert Palacio (2003). “Managing Public Pension Reserves Part I: Evidence from the International Experience: SP Discussion Paper 3.

International Monetary Fund (2003A). Article I V Consultation, First Review Under the Poverty Reduction and Growth Facility, and Request for Waiver o f Performance Criteria.

International Monetary Fund (2003B). Second Review o f the Poverty Reduction and Growth Facility Arrangement.

International Monetary Fund (2004). Under the Three-year Arrangement Under the Poverty Reduction and Growth Facility, and Request for Waiver o f Performance Criteria.

International Monetary Fund (2005). Staff report to the 2005 Article IV Consultation, Sixth Review Under the Poverty Reduction and Growth Facility, and Request for waiver o f performance criteria.

Ministry o f Finance and Public Administration, State Accounts (2000 to 2003).

Palacios, Robert and Edward Whitehouse (2006). “Civil-Service Pension Scheme around the World”, SP Discussion Paper 602.

Palacios, Robert and Oleksiy Sluchynsky (2006). “Social Pensions Part I: Their Role in the Overall Pension System”, SP Discussion Paper 60 1.

Potter, Barry and Jack Diamond, (1 999). “Guidelines for Public Expenditure Management,” International Monetary Fund.

Quentin Wodon, Diego Angel-Urdinola, and Damien Echevin (2006). “Oil Prices, Utility Tariffs, and Poverty in Cape Verde”, mimeo, World Bank.

Unidade de Acompanhamento Econ6mico, Several Reports.

73

World Bank (2003). Cape Verde - Public Expenditure Review: Improving Public Spending Under a Tight Budget Constraint.

World Bank (2004A). Cape Verde - Development Policy review: Building on a Strong Record to Meet New Challenges.

World Bank (2004B). Cape Verde - Poverty Diagnostic.

74

Annex 1. Fiscal Performance Technical Note

1.1. PRSP Indicators and Benchmarks

Table 1: Key Economic Indicators

Indicator 2004 2005 2006 2007 GDP (real growth rate) 5.5 6.0 6.5 7.0 Inflation 1 .o 2.0 2.0 2.0 Broad money (growth rate) 8.0 8.1 8.6 9.2 Overall balance (CVE billion) -3.7 -7.1 -7.5 -6.8 Overall balance (% GDP) -4.3 -7.6 -7.3 -6.1 Primary current balance (CVE billion) 3.2 3.4 3.8 4.6 Primary current balance (% GDP) 3.7 3.6 3.7 4.1

Reserves in month imports 2.0 2.3 2.5 2.8 Public investment (CVE billion) 9.7 13.5 15.0 15.5

Table 2: Other Economic Indicators

Indicator 2004 2005 2006 2007 Nominal GDP (CVE billion) 87.4 94.4 102.6 112.1 Total revenues (% GDP) Tax revenues (% GDP) Total expenditures (?? GDP) Investment expenditures (% GDP) Percentage of net credit to the economy in net domestic credit Exports of goods and services (% GDP) Net current transfers (% GDP) Net private transfers (% GDP) Imports of goods and services (% GDP) Overall balance of payments (% GDP) Net foreign direct investment (% GDP)

28.1 27.6 27.7 27.7 19.7 19.0 19.0 19.0 32.3 35.2 35.0 33.5 11.1 14.3 14.6 13.8

55.2 57.0 58.7 60.6 33.9 35.0 36.3 38.2 24.2 24.2 24.2 23.7 18.6 18.4 18.3 18.1 66.1 67.5 66.4 65.6 -9.5 -9.6 -7.0 -4.8 11.7 3.9 5.4 5.4

Table 3: Indicators and Benchmarks of Monitoring the Pillar I

Indicator 2004 2005 2006 2007 FORECAST

Per capita GDP growth rate (%) 3.6 4.1 4.5 5.0

Inflation rate (%) 1 .o 2.0 2.0 2.0 Real GDP growth rate (%) 5.5 6.0 6.5 7.0

Budget deficit (YO GDP) -4.3 -7.6 -7.3 -6.1 Other Indicators of Good Governance Education expenditures (% national budget) 20.0 22.0 22.5 23.0 Health Expenditures (% national budget) 6.3 6.5 7.0 7.0

BENCHMARKS

75

W N W b v ) U J O m m w g g w r c

9

U m 6 ._ E m

a! h

a U

5 i e n

P m

.- 2

Annex 2: Debt Sustainability Analysis (DSA) 65 66

Cape Verde’s debt level has increased in recent years. Despite the rising cost of servicing this debt, the country’s external sustainability ratios indicate a low probability of debt distress under the baseline scenario and in the face of plausible shocks. High domestic debt and its implication for $seal space, however, warrants focusing on total public debt in the analysis.

Background

1. The sustainability o f Cape Verde’s external and total public debt i s assessed using the Fund-World Bank debt sustainability framework for l o w income countries. The previous DSA was carried out during the sixth review o f the PRGF and Article IV c o n ~ u l t a t i o n , ~ ~ based on the outstanding stock o f debt at end-2004. The analysis concluded that, under long-term economic scenarios that were in l ine with or more cautious than recent historical averages, Cape Verde’s public debt appeared sustainable although vulnerable to a variety o f possible economic shocks.

2. Total public debt in Cape Verde has been above 80 percent of GDP for most of the last decade!* At end-2005, debt stood at almost 90 percent o f GDP, with interest payments accounting for nearly 10 percent o f domestic revenues (2.2 percent o f GDP).

3. External public debt increased significantly in the 199Os, but has declined over recent years. After peaking at 64 percent o f GDP in 2001, external debt decreased to 54 percent o f GDP at end-2005. Of this, 79 percent was owed to multilateral creditors, 18 percent to bilateral creditors and less than 1 percent to commercial creditors. The World Bank (IDA) was the largest multilateral creditor, followed by the Afr ican Development Fund and BADEA. Portugal, Russia and Germany accounted for over 65 percent o f bilateral debt.

65 This DSA has been prepared joint ly by staff o f the IMF and the World Bank. The analysis i s undertaken based on aggregate external debt data f rom the authorities, which have been revised since the last DSA.

66 This analysis focuses on central government and government guaranteed public debt only, as comprehensive and reliable data on private external debt are not available. Private external debt contracted i s believed to be around US$25 million. The outstanding amount and the t e r n o f repayment o f this debt are not known.

67 Appendix I o f IMF Country Report No. 05/320,09/2005.

Due to its relatively high income per capita, Cape Verde i s not eligible for HIPC assistance nor does it qualify for debt forgiveness under the MDRI. Cape Verde’s eligibility for concessional financing through the PRGF and IDA i s based on the smal l island exception.

77

Table 1. Cape Verde: Net Present Value of Disbursed Debt Outstanding, 2006-1 1 (Millions of US. dollars)

Actual Projected

Multilaterals IDA IMF Others

Official bilaterals Paris Club Non-Paris Club Export credit agencies

Commercial

395.6 218.0 217.1 214.7 210.0 205.2 199.9 213.9 106.4 108.6 110.3 111.3 112.0 112.7

12.7 10.1 10.4 10.1 9.1 7.5 5.2 169.0 101.5 98.1 94.3 89.6 85.8 81.9 93.9 63.7 55.9 46.7 37.8 32.3 26.9 65.3 42.8 35.8 28.6 21.7 18.3 15.0 28.6 17.4 16.9 15.4 13.9 12.3 10.7 4.8 3.5 3.3 2.7 2.2 1.7 1.2

14.4 12.2 9.9 7.5 5.0 2.4 0.0

Sources: Cape Verdean authorities; and staff estimates and projections.

4. Domestic debt has increased sharply since the debt restructuring operation in the late 1990s. This increase has reflected the Government’s assuming the debt o f some large public enterprises as part o f their privatization, fiscal slippages, and, more recently, increased recognition o f domestic arrears. Central government net domestic debt including arrears amounted to 33 percent o f GDP at end-2005. O f this, 30 percent was short term. Commercial banks were the major creditors, holding 55 percent o f total Government debt. The social security hnd INPS held the largest share o f the debt to non- banks (26 percent).

5. The assumptions underlying this DSA are based on the medium-term macroeconomic framework agreed with the authorities during discussions for the PSI and long-term projections agreed between the Fund and the Bank staff. Over the longer term, real GDP growth i s expected to average 5 percent while inflation gradually decreases to around 2 percent annually in l ine with rates in the euro area. Exports are projected to grow by around 9 percent a year on average, below their historical average growth rate. A cautious approach has been taken toward these assumptions; for example, growth could be higher than projected if foreign direct investment continues to grow at a strong pace.

6. Grants are assumed to decrease to 5percent of GDP by 2026, and new borrowing terms may also worsen as a result o f Cape Verde’s graduation from low income country status in 2008. Hence, while the baseline scenario assumes no commercial borrowing in 2006-09, commercial loans are assumed to gradually replace loans from multilateral creditors and the Paris Club in the long run. Thus, at the end o f the projection period (2026), loans from multilaterals account for 30 percent o f total borrowing, whereas commercial borrowing accounts for 40 percent. As a result, the concessionality o f new borrowing decreases from 27 percent in 2006 to 8 percent by 2026.

78

Table 2. Cape Verde: Macroeconomlc Baseline Assumptions, 2006-1 I (Percent, unless otherwise indicated)

Historical Average 2006 2007 2006 2009 2010 2011

(199s-2005)

Rea GDP growth Inflation Exports of G&S (US. dollars terms) Imports of G&S (US. dollars terms) Current Account Balance (percent of GDP) Foreign direct investment (percent of GDP) Grant element of new extemai borrmng Exchange rate (national currency per U.S. dollar, p.a.) Public sector revenue and grants (percent of GDP)

I

6.6 2.6 14.6 6.9 -9.3 3.5

101.3 30.3

5.5 6.0 6.3 6.6 6.6 6.3 6.2 0.2 0.3 2.1 2.1 2.1 3.6 5.0 6.5 6.4 12.6 10.5 9 2 9.4 9.3 6.6 6.3 6.7

-6.6 -9.9 -10.9 -11.0 -9,9 -7.3 2.1 3.4 4.0 4.4 2.0 2.1

27.3 27.5 29.2 31.1 26.6 25.6 93.1 92.9 92.6 92.4 92.3 92.1 34.1 33.6 34.3 35.0 33.6 30.5

7. In the event that highly concessional loans fall short of the baseline assumptions, an alternative scenario assesses the impact of covering the resulting gap by a limited amount of less concessional borrowing, as provided for under the PSI program. To assess the maximum impact o f such borrowing, the alternative scenario assumes that the full amount allowed under the program (US$20million per year) i s provided on commercial rather than highly concessional terms over 2006-09. The alternative scenario assumes, moreover, that in the long run, commercial loans will replace concessional loans at a faster pace than in the baseline, thus accounting for 60 percent o f total borrowing in 2026.

8. previous exercise in 2005. The main additions are as follows:

The DSA incorporates new information that has become available since the

o Government guaranteed domestic debt i s included in the total debt stock at end-2005; o newly recognized domestic arrears and cross-debt with the public administration and

institutes (amounting to CV Esc 5.3 billion) are included in the domestic debt stock; o revised projections on contributions to the social security fund stemming from the

pension reform o f January 1,2006 are incorporated in the fiscal framework; o oil subsidies are eliminated from the projections, following the Government’s 2006

budget decision to this effect; WE0 oi l price projections are updated as o f the winter 2006 baseline. o

External Debt Sustainability

9. Despite the gradual reduction in concessionality of new loans, Cape Verde’s debt stock and flow indicators under the baseline scenario remain below their policy-dependent thresholds throughout the projection period. The NPV o f debt i s projected to increase minimally from 32 to 34 percent o f GDP, and to decrease in terms o f exports (from 91 to 83 percent). Reflecting worsening borrowing terms, however, debt service payments increase from 7 to 11 percent o f exports. In the alternative scenario, the debt service-to-export ratio would increase to 15 percent over the projection horizon.

79

Figure II 1. Cape Verde: Indicators of Public and Publicly Guaranteed External Debt Under Alternative Scenarios, 2006-2026

(Percent) 80 I 1 80

2 0 -

10

NPV of debt-to-GDP ratio 60

Baseline --- Historical scenario

Most extreme stress test - - -

I I I I I I I I I I I I I I I I , , , ,

250

200

150

100

NPV of debt-to-exports ratio

Baseline Historical scenario Most extreme stress test

- - ---

- ____-- - - - - - - - - - -__ - / - _---- -~

50

70

60

50

40

30

20

10

0

-

I 1 1 1 1 , 1 , , , , , , 1 1 1 , 1 I I

250

200

150

100

50

0

Debt service-to-exports ratio

2006 2008 2010 2012 2014 2016 2018 2020 2022 2024 2026

5 -

1 I I # , , , , , , , , , , , , , , , ,

Baseline Historical scenario

- i l5

---- 1 1 0

Most extreme stress test _--- ----- / / e -

lo t l/ 5

2006 2008 2010 2012 2014 2016 2018 2020 2022 2024 2026

Source: Staff projections and simulations.

80

10. Standard bound tests suggest that even in the presence of external shocks, Cape Verde does not appear vulnerable to debt distress. However, a one-time 30percent exchange rate depreciation relative to the baseline would bring the NPV o f debt-to-GDP very close to the threshold o f 50percent and keep it permanently at this level. The debt outlook also appears sensitive to further worsening o f borrowing terms and a decline in non-debt creating flows (implying higher debt flows). Whi le the former would increase the NPV o f debt-to-exports to 126 percent, a decrease in transfers by one standard deviation compared to the historical average would push the debt service-to- export ratio f rom 7.6 percent in 2006 to nearly 12 percent in 2023.

I Table 3. External Debt Burden Indicators I Cape Verde's ratios

Thresholds' 2006 2016 2026

NPV of debt in percent of: GDP 50 32 33 34 Exports 200 91 96 83 Revenues 300 95 119 120

Exports 25 7 9 I 1 Debt service in percent of:

Revenues 35 7 11 15

Sources: MOF; and staff extimates.

' Based on Cape Verde's 2005 classification as a strong performer. 'Including grants.

Fiscal Sustainability

11. The fiscal DSA i s based on the assumption of ongoing fiscal consolidation in the context o f lower grant inflows. The D S A also builds in and extends the PSI program goal o f a sizeable reduction in domestic debt as a share o f GDP. Specifically, domestic debt would decrease to 16percent o f GDP by 2026. Domestic public revenues are projected to be constant at 24 percent o f GDP over the long term: this i s lower than the recent historical average and medium-term projections, reflecting prudent assumptions on the scope for ongoing gains in revenue collection. The resulting overall fiscal deficit in the projection period averages 3.1 percent o f GDP.

12. The baseline scenario depicts a decline in the NPV of debt-to GDP ratio, in line with efforts to decrease the debt burden and create fiscal space for emerging pressures. This ratio decreases to 50 percent in 2026 from 60 percent in 2006. Whi le the NPV o f debt-to-revenues shows a small improvement, debt service increases sharply from 10 to 20 percent o f revenues by 2026-reflecting the worsening o f borrowing terms in the baseline scenario. In the alternative scenario, with a more rapid deterioration in borrowing terms, debt service would rise to 22 percent o f revenues by 2026.

13. Standardized alternative shocks increase the NPV of Cape Verde's public debt and drive debt service to much higher levels. The most extreme test, which

81

assumes that the primary balance stays at its historical average minus 1 standard deviation in 2007-08, increases the NPV o f debt-to-GDP to 73 percent in 2008 and keeps it above the baseline throughout the projection period. The same shock stabilizes the NPV o f debt to-revenues at around 200percent in the long run, while increasing debt service-to-revenues to 23 percent by 2026.

14. Recent simulations by the World Bank indicate that the INPS pension system will move into a cash flow deficit by 2037. While beyond the projection period in the current DSA, this outlook clearly has negative implications for contingent fiscal liabilities and public debt in Cape Verde over the long run. As in other pension systems around the world, early efforts to address these imbalances would lower the eventual burden o f adjustment . Conclusion

15. In summary, the debt sustainability analysis suggests that Cape Verde i s not likely to face debt distress despite the rising debt service burden, given that sustainability indicators are kept below the thresholds over the forecast horizon. The analysis incorporates cautious assumptions o n macroeconomic variables and builds in a considerable worsening o f borrowing terms in the long run. These results are contingent upon the reduction o f currently high debt ratios, highlighting the need for prudent fiscal and debt management in the period ahead, including to ensure that debt service costs do not crowd out high priority, poverty reducing spending.

82

Annex 3: Mitigating the Impact of Eliminating Petroleum Product Subsidies Technical Note

3.1. Glossary on Fuel:

180 cst

360 cst

bbl

bunkers

gasoleo = Gasoil (diesel) illuminating kerosene international bunkers sold at international prices)

Fueloil wi th a [lower] viscosity o f 180 centistokes: in Cape Verde this i s a blend o f mainly 360 cst fkeloil plus gasoil Heavy fkeloil with a [high] viscosity o f 360 centistokes; a grade o f heavy fkeloil traded internationally Unit o f (liquid) volume used for the pricing o f o i l (42 U S gallons)

Fuel supplied to ships and aircraft (jet A for aircraft, and residual and distillate fuels for ships)

Kerosene used for household lighting

Bunkers supplied to ships and aircraft engaged in international transport (and typically

Jet A

LPG

moGas

Kerosene used in aircraft j e t engines. T h i s has very tight specifications, but i s essentially the same as illuminating kerosene A mixture o f butane, propane and other light hydrocarbons derived f i om refining crude oil. A t normal temperature i t i s a gas, but when pressurized it i s a liquid. = Motor gasoline

petroleo = Kerosene

3.2

As i t i s clear from the sample table o f individual equalization entries for gasoil, each row o f which represents the calculations for a particular import shipment, the gasoil subsidy (e.g., in Nov-Dec CVE 16.27Aiter) represents the difference only between the actual market based price (CVE 91.27Aiter), and the regulated price (CVE 75Aiter).

Computation of the Petroleum Subsidies

Table 1: Shell 2005 Claims for Gasoil Subsidy

Consumption Quantity Price Official Differential From To Consumed Structure Price Negative Positive Total

(liter) (CW/iiter) (CVEfiiter) (CVEfiiter) (CVEfiiter) (CVE) 26-Dec-04 09-Feb-05 3 3 16,2 14 70.90 69.00 (1.90) (6,680,806.60) 10-Feb-05

18-Apr-05 26-May-05

1 2-MU-05

20-Jw-05 16-Jul-05

19-Aug-05 26-Aug-05 16-Sep-05 16-Oct-05

1 4-NOV-05 15-Dec-05

1 1 - M ~ - 0 5 17-Apr-05

25-May-05 19-Jun-05 15-Jul-05

18-Aug-05 25-Aug-05 15-Sep-05 15-Oct-05

1 3-NOV-05 14-Dec-05 04-Jan-06

2,880,977 4,400,405 3,717,851 2,957,692 2,804,567 3,795,027

789,691 2,300,948 3,377,249 3,329,828 3,614,376 1,598,855

72.30 67.70 68.20 74.30 73.10 79.00 83.10 85.87 84.77 90.07 91.27 89.37

69.00 69.00 69.00 69.00 69.00 69.00 69.00 75 .OO 75.00 75.00 75.00 75.00

(3.30)

(5.30) (4.10)

( 1 4.1 0) (10.87)

(15.07) (16.27) (14.37)

(1 0.00)

(9.77)

(9,507,225.23) 1.30 5,720,527.13 0.80 2,974,281.17

(15,675,767.26) (1 1,498,725.76) (37,950,267.42) (1 1,134,643.10) (25,011,304.76) (32,995,719.89) (50,180,503.80) (58,791,432.97) (22,975,546.35)

05-Jan-06 83.47 75.00 (8.47) 0.00 39,083,680 (273,707,134.84)

Source: Shell, Crilculo de diferencas auadro resumo ano 2005.~1s.

83

.. (v

I .I

I 0

cr s

w w o w o -* -, z -

m - o \ o w R 3

N O N m 2

g o * 3

m o m P

2

cl 0

5 00, N N

m m v,

3 M -.

0

0

0

0

0

0

g

b

s 0

3

.. c1

0 0

0 0 - 0 0 2- - N

0

* * *

d . 0 0 0 -- - m d . - O F 2 2 , -

~ " 0 0 r 4 d . t -

m o o

- O W - N 0 0 -

2 %

2 0 0

3

Energy Supply Transformation Final consumption

w m - 0 9 9 0 9 % O S

m \ 9 m 2 0 2 m m

0 0 - m m o d d d ??"!

3.3. The Electra Financial Model

In the 2005 version o f the Electra financial model, objectives are analyzed over a 5 year span (2005- 2010), but the value o f the firm i s evaluated over a longer period from 2005 to 2035. The net present value (NPV) o f the firm over the 30 year period is:

(R, -ck +Trk -Tk - Ik) K, NPV=-Ko +E +- k=l ( l+r ) , (1+rIN

where KO i s the init ial value o f capital, R, i s the revenue o f the firm at date k, C, are operation and maintenance costs, Tr, are transfers by the Government, T, are taxes, I, are investments, K, i s the final value o f capital, and r i s the discount rate (equals 10 percent).

Demand growth is suppose to vary at the demographic rate. Investments are thus both “exogenous” (those that are necessary to increase efficiency) and “endogenous” (those that are necessary to increase the capacity o f the firm in order to respond to increasing demand). Instruments used to increase the value o f the firm are: increasing tariffs, diminishing operation costs and/or losses (thus increasing established and productive efficiency o f the firm), increasing transfers from the Government, and extending the network in various islands (~ross-subsidies)~~.

Assumptions for the basic scenario are the following:

i. No change in tariffs; ii. N o transfer in cash from the Government; iii. Diminishing o f operation costs o f 5 percent from 2005 to 2010; iv. Access rates o f electricity and water o f 90 percent in 2010.

Table 5 gives the resulting cash f low during the 30 year period.

69 Costs are indeed different from one island to another but tariffs remain the same.

87

Table 5: Cash Flow of Electra, 2004-2045

Government Change in Capital Net Flow

base ( R - C + T r l - T - I (K) - Tr2 + K)

and Taxes Investments working (TI (I) capita I

(Tr2)** EBITDA payment of tax (R-C) deficit

(Trl)*

Year

2004 2005 2006 2007 2008 2009 2010 201 1 2012 2013 2014 2015 2016 2017 2018 2019 2020 202 1 2022 2023 2024 2025 2026 2027 2028 2029 2030 203 1 2032 2033 2034

568 536 782

1,213 1,686 2,200 2,384 2,534 2,686 2,873 3,030 3,23 1 3,426 3,626 3,896 4,144 4,388 4,714 4,99 1 5,379 5,716 6,121 6,492 6,877 7,285 7,746 8,209 8,699 9,219 9,771

0 401 40 1 40 1 40 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0

143 120 329 346 362 400 453 489 494 543 594 67 1 716 738 829 863 983

1,022 1,123 1,243 1,368 1,513 1,663 1,727 1,892 2,045 2,200

-5.872 -5.872

2035 10,358 0 2,376 1,351 -186 37,301 44,117 *The tariff's deficit o f Electra is subsidized by the Government with a loan o f CVE 1,604 mill ion on a period of 4 years. ** Tr2 = A(Availabi1ity +Short run credit + Other short run debt). Source: Electra's Financial Model of ARE.

4,194 2,534 2,592 2,538 1,916 3,675 2,495 989 625

1,312 4,126 1,119 814 69 1

2,206 3,827 871

3,951 1,327 4,552 5,573 1,997 1,915 1,053 1,152 8,643 1,262 1,243 1,283 1,312

198 -210 I 50 16

-18 -2 1 -52 -49 -22 -3 0 -79 -60 -66 -12 -4 1

-1 12 2

-118 11

-68 -101 -97 -123 -1 19 -24 -230 -149 -160 -172

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

-3;824 -1,387 -1,415 -1,117

35 -1,786 -437 1,236 1,710 1,130 -1,555 1,696 2,129 2,407 1,03 1

2,891

2,9 19

-358

-68

-168 -811 3,101 3,43 1 4,578 4,740

5,450 5,712 6,052 6,43 1

-2,535

88

Annex 4: Public Pensions and Health Insurance Scheme Technical Note

Accrual replacement rate

Defined benefit Funding

Implicit pension debt'

Indexation

Table 1: Pension's Terminology The rate at which pension entitlement i s built up relative to wages (in earning-related schemes) Benefit based on a prescribed formula Accumulation of assets in advance to meet h t u r e pension liabilities Present value of projected liabilities under the pension system and covers those who are already receiving payments from the system as well as those who have any acquired rights at a particular date Increases in benefits by reference to an index, usually of prices, or average eamings

old-Age dependency ratio PAYG (Pay-as-you-go) system

Pensions contribution rate

Pension coverage rate

Length Of service 1 Number of years a worker has contributed to a system

The ratio of older persons to working-age individuals A method o f financing whereby current pension expenditures are paid out of current revenues from contributions (payroll taxes)

Payroll tax as percentage of an individual gross wage for a pensions scheme The number o f workers actively contributing to a publicly mandated contributory retirement scheme, divided by the estimated labor force or by the working-age

against the employee or employer) in order to avoid subsidy from any external source o f

Replacement Rate

gradually grow old and retire, thereby raising the system dependency ratio to the System maturation

used by the international community and fiscal affairs experts when comparing the maturation and fiscal burden of different pension systems. I t i s important to note that this indicator, as present, as it values calculations i s highly sensitive to the discount rate used in the calculation. Hence with the proviso that macroeconomic assumptions are on balance sound and internally consistent, expressing the implicit pension debt as a percentage of the current year's GDP should be a reliable indicator of fiscal burden. Nevertheless, it should s t i l l be used with prudence when making comparisons with other pension systems. Required length o f service (or vesting period) i s the minimum number o f years required to qualify for fill pension benefits. Basic replacement rate i s the pension obtained for such period.

89

Analysis, Data, and Methodology

The numbers presented here were produced using the Wor ld Bank’s PROST model, a pension simulation model that has been used to evaluate pension reforms in over 80 countries. PROST (Version 12) i s a generic PC-based projection model developed by the Social Protection Unit o f the World Bank. It has been used to analyze and project reform options for the pension programs of AP and INPS in Cape Verde, as wel l as alternative reforms under a few scenarios.70

The program i s designed to simulate the behavior o f pension systems and can assess their financial sustainability under different sets o f assumptions over a long time frame. I t allows modeling different pension reform options - from “parametric” reforms o f pay-as-you-go defined-benefit schemes to systemic reforms, such as the introduction o f fully funded defined- contribution andor notional defined-contribution schemes. The program can be adapted to a wide range o f country circumstances and can handle simulations up to 100 years and more. In the case of Cape Verde, only “parametric” reforms o f pay-as-you-go defined-benefit schemes (Le., INPS and AP) have been simulated.

As with any simulation model, the outcome from PROST depends largely o n the nature and quality o f data as wel l as o n the set o f assumptions being used for the simulations. Since PROST has been extensively used to provide quantitative input for pension pol icy discussions, i t s methodology has proven to be sufficiently robust and its flexibility has permitted easy adaptation to specific country circumstances for sensitivity testing and comparisons under a wide range of economic and policy scenarios.

However, i t i s always important to keep in mind that projections o f this nature should not be quoted in absolute terms without proper reference to the underlying assumptions. The purpose o f the sustainability benchmarks and fiscal implications presented here aims at providing a comparison o f the relative magnitude o f the effects o f different pension pol icy measures under various scenarios.

Pensions data availability and quality can be summarized as follows: (i) lack o f centralization o f the three databases o f I N P S (Sa0 Vicente, Santiago, and Sal), which continues to pose problems; (ii) data inconsistency between statistics and IT department in INPS; (iii) AP census appeared to contain a large number o f errors and individual entries and although it was reviewed, there was st i l l a problem in the lack o f date o f birth entries for AP pensioners. The pension department o f the Ministry o f Finance i s only in charge o f making the pension payments, and do not have the needed support for maintaining a complete pensions database. However, for a good analysis, data by age and sex is needed and only a small sample was available. Although providers o f such data have been very cooperative in Cape Verde, they do not keep the right type o f data to be able to do such type o f analysis. This suggests a strong need for a pension administrative project, which can help them collect and organize the types o f data records they need in order to do pol icy analysis in the future.

’’ The PROST model has been developed in VBA and can be used with Windows 95/98/NT operating systems in conjunction with Office 97 or later platforms.

90

The start o f the simulation horizon was 2004 (base year), the year for which the most complete documentation was found for al l o f the variables needed for PROST. The ending year for the simulation horizon was 2100 (end year), a period viewed to be o f adequate duration to demonstrate the emerging trends o f pension expenditures in most pension schemes.

All data and assumptions required by PROST for population projections were provided by the Wor ld Bank’s Population Unit and INE, including init ial population by age and gender, and projections o f age-specific fertility rates, mortality rates, and migration flows. Total aggregated numbers were also provided from the IMF.

Mortal i ty Rates: Age-specific mortality rates (in five-year age groups), including projected improvements in mortality (in five-year intervals), for years 2004-2020 were provided by They were extended to 2100 using the same pattern. These mortality rates were used to calculate the probability o f dying by age cohort (age 0-85) for both men and women every year during the simulation horizon. As a result o f the mortality assumptions, the average l i fe expectancies at various ages are shown in Table 2.72

Table 2: Projection of Li fe Expectancy Changes 2004 2005 2006 2007 2010 2025 2100

Male Life Expectancy At birth 66.3 66.5 66.7 66.8 67.4 70.1 79.1 A t age 20 51.6 51.7 51.7 51.8 52.0 53.2 59.5 At age 60 16.6 16.6 16.7 16.7 16.8 17.5 22.1 At age 65 13.1 13.2 13.2 13.2 13.4 14.0 18.0 Female L i fe Expectancy At birth 71.9 72.0 72.2 72.4 72.9 75.7 84.6 At age 20 56.0 56.9 57.0 57.1 57.3 58.5 64.8 At age 60 19.8 19.8 19.9 19.9 20.1 21.0 26.1 At age 65 15.7 15.8 15.8 15.9 16.0 16.8 21.5

Fertility Rates: Age-specific fertility rates (in five-year age groups), including projected improvements (in five-year intervals), for years 2004-2020 were provided by INE. They were extended to 21 00 projecting the parameters using a relational-Gompertz model.

Macroeconomic assumptions: Macroeconomic assumptions are presented below. Beyond year 2005, the scenario is based on IMF projections. The longer-term trend i s based o n the assumption that the economy will converge to a steady real growth rate o f 4 percent per year by 2026, while average wage growth was adjusted to reflect the GDP per capita growth. The interest rates were assumed to be real o f 3.5 percent. The inflation rate was assumed to be 2.1 percent in the medium- and long-term. These economic parameters are used in conjunction with the assumptions concerning the population and the system demographics to project the future financial conditions o f the pension systems. A 5 percent nominal discount rate was used for present value calculations (Table 3).

’’ INE’s mortality table stopped at age 65 and considered 5-year age groups. Considering that the elderly population was a focus of the simulation, rates were interpolated for individual age and extended to 100 years. ’* Mortality tables used are available upon request.

91

Table 3: Macroeconomic Trends: Baseline Scenario

Indicators 2004 2005 2006 2007 2010 2025 2050 2100 Real GDP growth 4.5 6.3 7.0 6.5 6.5 5.0 4.0 4.0 Real wage growth 2.5 4.3 5.0 4.6 4.5 3.5 3.2 4.2 Inflation rate -1.9 0.4 2.4 2.1 2.1 2.1 2.1 2.1 Nominal discount rate 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0

92

Annex 5: Progress on the CFAA Action Plan

MODULE 1: STRENGTHENING BUDGET PREPARATION AND PROGRAMMING

1. Budget preparation methodology: A draft budget law and a draft budget planning law are being discussed. The new laws aim at taking into account al l the new reform options, especially progress towards the MTEF and program budgeting and clearly formalizing the schedule for budget elaboration.

The recurrent and investment budgets are now elaborated together by the DGP, the DGO, the DGT and the sector ministries. This has been made easier by the new budget nomenclature, which i s the same for recurrent and capital transactions. Ceilings are specified and sent to the sector ministries. These ceilings are, however, not always deemed as truly upper limits, leading to negotiations and revisions o f the whole overarching framework.

The principle for consolidation is being put into practice. Revenues o f external financing, as wel l as the corresponding expenditure, are included in the budget. The budget o f autonomous institutes i s prepared by the institutes themselves and then included in the State budget for transfers f rom the State, as well as for their own revenues. The harmonization o f the chart o f accounts for a l l public entities will be effective when the new chart i s implemented in 2006.

The new budget nomenclature has been implemented. It wil l permit the consolidation o f each ministry's global expenditure - current spending and investment. The classification by function has been designed, but it i s far from the Classification o f the Functions o f Government o f the United Nations (COFOG) and a bridge between both will be necessary.

The linkage between PND, PRSP and MTEF i s s t i l l weak. The macroeconomic framework for the period 2004 to 2007 was elaborated at the same time as the PRSP was being prepared, with the final version o f the PRSP having been approved by the Council o f Ministers in August 2004. The alignment between PRSP and MTEF should be improved. MTEF and PRSP preparation started at about the same time and thus PRSP does not emerge from the MTEF. The link between budget and PRSP should also be improved. Programs and sub-programs are from now o n linked to the budget.

International technical assistance and training is being provided to the Ministry o f Finance and sector ministries, in order to develop a global MTEF and sectoral ones. Three ministries have elaborated preliminary drafts: Education; Agriculture, Fisheries and Environment; and Social Protection. The preparation o f the sectoral MTEFs overlapped with the preparation o f the global MTEF and the interaction between the two processes was insufficient. The global MTEF set ceilings which were not respected as truly upper limits and as a result the global MTEF was revised to accommodate the revised ceilings. This complicates the budget preparation and execution as there i s no clear and consistent directive and it hampers the budget preparation process.

2. Program budget: The macroeconomic framework i s st i l l not taken sufficiently into account. A macroeconomic modeling tool has been developed locally to improve revenue forecasting. But revenues forecasted at the DGA, as wel l as the DGCI, are st i l l based o n the

93

actual receipts from the last three fiscal years. The evolution o f GDP i s also taken into account. The final decision comes from the MFAP, which usually gives a raise o f the init ial forecast.

A procedure handbook has been drafted on budget preparation and elaboration. A comprehensive manual is currently being drafted for the MFAP and departments. Also, training seminars have been held to improve usage o f the SIGOF system.

The criteria for including a project in the budget are linked to the sector strategy. Sometimes, they are included in specific directives o f the Prime Minister. In this case, the department tries to incorporate them as closely as possible to the sector strategy. The National Assembly can also express recommendations, which lead to changes on the nature o f projects.

The linkage between Sectoral Office for Planning and Studies (GEP) and their respective sector structures i s s t i l l weak, but improving. Training seminars have been provided. The most important constraint i s geographic, the link between Praia and the various islands being difficult.

3. Research Oflce (GPE): The implementation o f the GPE at MFAP is ongoing. The C F A A Unit works closely with this new structure, with the objective o f i t s improvement. The C F A A unit has more o f a technical role, while the GPE has more o f an institutional role.

4. Capacity building of the DGP and the sectors: The linkage between concerned Government departments and donors i s strengthening. The final version o f a memorandum o f understanding was agreed upon in February 2005 and was signed in April 2005 by the Budget Assistance Group, consisting o f the Wor ld Bank, the European Union and the Kingdom o f the Netherlands. A new Memorandum o f Understanding for budget support has been discussed and i s expected to be signed as soon as possible by the Government, IDA, the Netherland Embassy and the European Union. The BAD, Spain and Austria participated in discussions and expressed intentions to j o i n the budget support group. The DGP provides a link between the various departments and donors and i s becoming increasingly effective.

5. Program budget database: The SIGOF database contains the programs and sub-programs and i t s fimctionalities allow managing them in a coherent way. The process o f providing sector structures with procedures and instruments for follow-up in the SIGOF i s ongoing.

MODULE 2: STRENGTHENING MANAGEMENT AND OVERSIGHT OF GOVERNMENT REVENUES

6. Consolidation of the taxpayer registry: The taxpayer registry i s currently being consolidated. The objective i s to provide a unique identifier, which will be the same for the D G C I and the DGA. Every taxpayer will receive an electronic card just i fying his identification. Computers have been purchased recently. This system wil l allow the comparison o f customs debt and VAT credits and to offset liabilities and debts o f the same taxpayer. The law stipulating the creation o f the fiscal identifier has been promulgated.

D G C I has almost achieved the identification o f taxpayers missing. The D G C I i s continuing to gather taxpayer data and enter these in the registry from the databases o f customs warehouses,

94

corporations and c iv i l service rolls. However, chambers o f commerce and municipal councils have been contacted, but they have not yet replied. The link between the c iv i l service database and the fiscal registry has been established, but the one with the registry o f municipalities i s s t i l l expected. The specific software developed at the N O S I i s wel l advanced and c iv i l servants have been hired.

7. Recovery of tax debts owed: The recovery o f tax debts owed i s st i l l weak. A specific software is currently being developed, which will be dedicated to proceedings against bad payers. But the tax debts owed are not well-known for al l municipalities, except for Praia since 2001.

8. Internal Audit Ofice: The internal audit office at the D G C I i s not yet operational. This action has received important attention from the C F A A unit. Equipments have been purchased and delivered and three senior inspectors have been assigned. Terms o f reference have been drawn up for developing the specific SIGOF functionalities and procedures handbooks have been elaborated. However, according to the deputy general director, the office i s currently managed by only one person, who has not received any specific training, and i t seems that the internal audit office i s not yet operational.

9. Management of Government revenues: The management o f Government revenues i s currently improving. The SIGOF's Revenue Management Module with VAT functionality is operational. I t allows the recording o f information coming from the monthly and quarterly forms. Terms o f reference have been drawn up for developing IUR PC and IUR PS functionalities in the SIGOF. A consultant has been hired to carry out training in the Finance un i t s and to ensure that the knowledge and tools necessary for utilization o f the SIGOF are assimilated.

The tax withholding system has not yet stabilized. The gap between the amount o f tax paid by the taxpayer and the amount actually owed will be reduced due to the flexible rates o f the withholding tax.

10. Collections and receipts: The collection o f receipts benefit from modernizing methods and training actions. A team made up o f representatives from the DGCI, the N O S I and the C F A A unit held working sessions in August and September 2005 o n the subject o f designing a pi lot project for modernizing the system in the Finance units for attending to enquiries and collecting and managing receipts.

There was an objective to implementing the agreement between the MFAP and the banks on the new computer application, to guarantee that information i s provided broken down by taxpayer and by type o f tax collected. This objective i s almost met. Meanwhile, work i s proceeding o n improving the linkage to the new computer application that forms part o f the Payments and Receipts Systems project.

Coordination among the DGCI, the DGA, the DGT and the DGCP in consolidating tax and customs receipts has not yet strengthened. All information needed for this coordination i s available. However, the D G C I and not the DGT can access the database o f the DGA and the DGT has not at i t s disposal information o n the volume and the nature o f receipts. Despite this,

95

receipts are transferred o n a daily basis to the treasury account. Regarding the devising o f an information f low that will permit proper bank reconciliation between the DGCI, the banks and the DGT, the functionalities necessary have not yet begun to be designed.

MODULE 3: IMPROVING NATIONAL TREASURYAND PUBLIC DEBTMANAGEMENT

11. The National Treasury is still managed on a daily basis. Yearly program payments and quarterly or weekly loans are st i l l weak for the treasury. The treasury staff has not been trained. There are n o procedures handbooks, nor specific tools for treasury management. The staff i s waiting for assistance from the C F A A unit in these matters. A project o f computerization o f the treasury office has been launched. The MFAP is currently looking for financing.

Some accounts open in commercial banks have been closed. Autonomous institutes are authorized to use an account open in a commercial bank, but their transactions are, as from now, centralized in an account open at the DGCP.

12. Public debt: Regarding public debt, only external loans have been recorded. The software CS-DRMS was implemented in 1996 at the debt office. I t became operational 2 years ago, because o f insufficient trained staff. This software i s linked to the SIGOF.

A training seminar has been held recently on the CS-DRMS. But a real need for technical assistance exists to update the information o f i t s database, especially for the domestic debt. Indeed, only external loans have been recorded in the system. The office does not have procedure handbooks yet and there i s a need for new computers.

13. Guarantees and retrocession accords: They are not yet recorded in the database.

MODULE 4: IMPROVNG THE CONDITIONS OF BUDGET EXECUTION

14. Procedures Manuals: Procedure Handbooks have been drafted. The services have n o w at their disposal a handbook for the execution o f current expenditure, as wel l as investment expenditure . 15. Implementing the new public accounts: A new public accounting system has been designed and approved, but not yet implemented. Some corrections are being done and some points are st i l l being discussed. Until now, no training seminars have been held, but a seminar wil l start soon. A presentation to the staff o f the MFAP and to the public will be done. Its formal publication i s planned for next weeks.

A consultant i s currently sought to work o n implementation o f this new system. The defining o f specific functionalities will soon be done by the NOSI. The tests are planned for July 2006, and the definitive implementation for January 2007. This new chart o f accounts will be used not only for the State budget, but also for the autonomous institutes and the municipalities, only the root o f the accounts being identical.

96

16. Integration of al l investment projects in the SIGOF: Although they are included in the budget, externally financed projects are paid directly by the donors, after proposal by the departments. This expenditure i s later recorded in public accounts within the SIGOF, but they do not fol low the public account rules and the gathering o f information remains very tiresome.

17. Elaboration of the General Government Accounts: Public accounts for the years 1998 to 2003 have been finalized and submitted to Parliament. As stipulated by the public rules, the TdC has been requested to examine these accounts and to express its advice o n their management by the Government. Regarding fiscal year 2004, closing o f the accounts by the MFAP has not yet taken place. The budget law stipulates that the final accounts must be presented to the National Assembly before the fol lowing year.

All the provisional quarterly accounts for 2004, including the corresponding provisional statements o f financial operations, have been elaborated and submitted to Parliament, as wel l as the provisional accounts for the first, second and third quarters o f 2005. I t should be noted that the quarterly accounts are issued 45 days after the end o f the related period. A special office has been created to manage this reporting and to carry out the dissemination o f the reports, which are available for consultation o n the public internet.

18. Salary andpension systems: The salary and pension systems wil l benefit f rom a census o f staff and improvements in the management. The census o f staff has taken place. The database o f the C iv i l Service Directorate is currently being merged with the one o f the MFAP, except for 3 Ministries : Education, Health and Justice. The new system will allow the management o f the carriers as wel l as the payroll. Equipment has been purchased and delivered at the DGCP. Regarding pensions, the database i s currently being separated from the one o f active employees.

MODULE 5: IMPROVING THE MANAGEMENT OF MUNICIPAL FINANCES

19. Strengthening the Jinancial and tax administration of the municipalities: The strengthening o f the municipal financial and tax administration i s ongoing. Three municipalities have been equipped with computers and with the software developed by the N O S I (Sal, Mindelo and Praia). All municipalities o f Santiago wil l be equipped during the next months. An agreement has been signed between the NOSI, the A N M C V and the Directorate o f Decentralization in order to support this action. However, the financing has not been totally completed yet.

20. ongoing.

Register of municipal assets: The implementation o f the register o f municipal assets i s

21. roads.

Crossing of data against the DGTR: This is in progress for vehicles and management o f

22. been set up and transactions for regularization have started.

Clearing of cross debts and credits: Studies have taken place. A plan for payment has

97

23, has been prepared.

Consolidation of decentralized cooperation: A draft law for the decentralized cooperation

24. achieve a consensus.

Consolidation of decentralization: A specific forum will be held in 2006, to try to

25. Modernization of human resource management in the municipalities: The legislation establishing staff regulations for municipal Government employees i s currently being discussed and training actions o f municipal staff have started.

26. Consolidation of the Local Finance Act: The local finance act has been revised and the amended version has been adopted and published as Law 79NV2005. Fol lowing this, a seminar has been held on the topic "Good Governance at the Local Level". I t i s very hard to establish permanent mechanisms for monitoring municipalities' sources o f financing and for decentralized cooperation resources.

27. for this Center has not yet started.

Local Finance and Decentralization Monitoring Center: The project o f implementation

MODULE 6: CAPACITY BUILDING OF THE OFFICE OF THE INSPECTOR GENERAL OF FINANCE (IGF)

28. Capacity building of the IGF : Resources at IGF have improved but remain insufficient. Since 2003, 7 inspectors have been hired and an inspector specialist in information systems wil l be hired soon, to ensure the internal management o f computers and the interface with the NOSI. The current staff o f the IGF are 18 inspectors, plus 8 inspectors seconded in other Government administrations. Some training actions have been carried out.

- As with every audit structure, the IGF can fol low budget execution in real time, thanks to the Intranet o f the Government. Numerous procedure handbooks have been drafted on audit actions, for central administrations, Finance un i t s and municipalities. Regarding i t s office-block, the lack o f space in the building strongly hinders the functioning o f the structure. During 2006 IGF bought, with the support o f the C F A A unit, 12 printers, 5 laptops, and 2 computers.

- I t s action plan i s hard to keep. An action plan is elaborated at the beginning o f the year. But this plan is disrupted by too many specific missions, very often qualified as urgent, and by technical assistance actions. Despite this, the number o f audits carried out by IGF is increasing. O f 32 audits in 2004, they increased to 35 in 2005. They are planned to be 40 in 2006 (focus on institutes and on the assessment o f the debt level o f a l l municipalities) and 42 in 2007.

MODULE 7: IMPROVING THE MANAGEMENT OF GOVERNMENTASSETS

29. Inventory of moveable assets: The inventory o f the Government's moveable assets and vehicles started in 2005 in collaboration with the departments o f Infrastructure, Transport and

98

Foreign Affairs. It i s planned to spread the process to the departments o f Justice and Education. Models o f forms have been designed to collect information not only about moveable assets, but also immoveable assets and lands. The necessary training actions have not yet started, because o f lack o f financing.

30. Inventory of immoveable assets: The inventory o f immoveable assets has started very slowly. There is only one technician available for this work. This inventory wil l be difficult, because many buildings have been built without being registered. A technician o f the DGPE has been seconded at the NOS1 to start the designing o f software dedicated to the management o f moveable and immoveable assets and vehicles.

3 1. Management of goods and services procurement: Progress has been made in procedures relate to the management o f the current expenditure. Actions o f control are henceforth more numerous and stricter. An employee visits once a month the various offices to control the level o f consumption. Regarding consumption o f fuel, a new system has started. A contract has been signed with a supplier, who sends a statement to the Government every month o f the purchases done by the State employees. The specific software dedicated to the management o f the collection o f rents, as well as the purchase o f goods, i s in the process o f improving. The DGPE plans to create an interface with the database o f buildings and vehicles, in order to allow coherent control.

32. Capacity building of the DGPE : It i s planned that, in the context o f the reform o f the DGCP, the DGPE will become responsible for the bookkeeping of patrimonial accounts related to their o w n transactions. In the future, al l transactions will be registered in the internal notarial office, including assets purchased or built o n external financing, which are not yet registered. However, regarding the capital revenues, nothing has been done to improve the recording. Lastly, the study o f the establishment o f internal regulations for the DGPE is delayed, because of a lack o f financing and slowness in procedures.

MODULE 8: IMPROVING THE MANAGEMENT, MONITORING AND RENDERLNG OF FOOD AID ACCOUNTS

33. Improving the management and monitoring of food aid : The management mechanisms have been published and reports are presented o n the use o f counterpart funds on a monthly basis.

MODULE 9: STRENGTHENING CAPACITY OF THE GENERAL AUDIT OFFICE

34 Capacity strengthening of the Board of Auditors: The process o f organizational reform at the TdC has not yet started. More than a year ago, a memorandum was sent to the Government, to ask for comments on an organic law, as wel l as a law on the functioning o f the structure. But to date, the Government has not yet answered and the process is s t i l l at its starting point.

Training actions are held. A seminar has been held, on the initiative o f the CFAA unit, and presented by senior auditors from the Portuguese General Audit Office. Regarding procedure handbooks and training for the judges, a project i s currently being designed.

99

MODULE I O : CPAR

35. CPAR: A specific project has been initiated on this objective. A consultant was hired in September 2005. H e has designed an action plan for the period 2006-2008 with the objectif o f writing up a new National Procurement Code, as from March 2006.

A seminar has been set up for the judges o f the TdC o n public markets, financed with the French cooperation and presented by 2 judges from the French General Audit Office. The TdC has also participated in studies related to the CPAFL

CFAA unit’s Mandate

The C F A A Unit i s responsible for the coordination o f MFAP’s actions related to ongoing reforms. The C F A A i s under the responsibility o f the GPE. The C F A A unit i s considered to have a technical role, while the GPE a more institutional role.

A protocol has been signed between the C F A A Unit, N O S I and GPE. Although financed by the Government budget, the N O S I i s looking for more financing, because currently i t i s unable to respond to al l demands. The C F A A aims to reinforce N O S I capacity.

The last steering committee o f the C F A A project was held in November 2005. The C F A A action plan i s very broad and a lack o f financing requires that prioritization takes place. W h i l e some departments regard C F A A as a channel to implement reforms, others feel that the actions o f the C F A A unit make them less responsible for leading the reforms to be pursued. I t i s important to clarify the role o f every player, in order to remove ambiguities and to revitalize ongoing reforms.

Various actions already financed through the CFAA Unit :

- Municipalities have benefited from actions supporting the drafting o f the new municipal budget law, the setting up o f a seminar on the municipal c iv i l service, o n FADEM (Funds supporting the development o f municipalities), and on the specific functionalities o f the SIGOF.

- Regarding the legal texts o n public finance, the C F A A Unit drafted the terms o f reference dedicated to the recruitment o f the consultant in charge o f drafting these texts: the State budget law, budget planning law, decree-law for authorizing officers and financial comptrollers.

- Seminars on the MTEF have been organized for different department staff.

- On the new chart o f accounts, the C F A A unit has provided advices. T w o o f i t s staffs are former employees o f DGCP and specialists in accounting. Some technical points are st i l l being discussed. This i s the case for the supplementary period. I t has been decided to implement the new chart o f account in a progressive way. A training program i s currently being drafted. The corresponding terms o f reference have been written up, in addition to the functionalities o f the SIGOF necessary to manage the accounts.

100

- For the management o f salaries and pension, the CFAA unit has financed the purchase o f computers. A technical assistant has been hired and training has been provided for the use o f the software.

- Regarding the DGCI, the C F A A unit has supported the creation o f the internal audit office. In principle, three staff should be hired in order to strengthen the unique employee currently appointed. For the management o f debt owed by taxpayers, a project has been elaborated to draw up an inventory o f the sums.

- The C F A A unit has provided the TdC with training seminars presented by judges at the Portuguese General Audit Office.

101