Report No. 40130-STP Democratic Republic of São Tomé and Príncipe Country Integrated Fiduciary Assessment (In Four Volumes) Volume II: Public Expenditure Review June 2007 Poverty Reduction and QK Procurement (AFTPC) Economic Management 3 QK Financial Management (AFTFM) Africa Region Africa Region Document of the World Bank -Vice President = Obiageli Katryn Ezekwesili Country Director = Marie Franqoise Marie-Nelly Sector Director = Sudhir Shetty Sector Managers = Yvonne M.Tsikata Lead Economist = HinhDinh Task Team Leader = Dorsati Madani S9o Tom6 andPrincipe PublicExpenditureReview Table of Contents ABBREVIATIONS AND ACRONYMS........................................................................................ PREFACE .................................................................................................................................... I V 1. MACROECONOMICTRENDS AND FISCAL CHALLENGES ............................... 1 Stable Growth but RisingInflation..................................................................................... 1 Patterns inTrade-A Structural External Deficit............................................................... 3 Monetary Policy-The Need for a New Strategy for the Future........................................ 5 Fiscal Policy-The Medium-Term Analysis (2005-10) .................................................................................. Needto Address the Chronic Structural Deficit ................................. 7 10 Long-Term Analysis (2010-15) ....................................................................................... 12 Alternative Scenario ......................................................................................................... Baseline Scenario.............................................................................................................. 12 14 2. FISCAL CHALLENGESOF THE PREPETROLEUMERA- THE NEEDTO ENLARGEFISCAL SPACE ......................................................................................... 20 1. REVENUESOURCESAND TRENDS ..................................................................................... 20 Trends and Composition o f Revenues .............................................................................. Current Revenues.............................................................................................................. 20 22 Grants and ForeignFinancing ofthe Budget.................................................................... 24 2. PUBLIC EXPENDITURE PATTERNSAND TRENDS............................................................... 26 Objective and Scope of Analysis...................................................................................... Economic Composition of Expenditures .......................................................................... 26 30 Administrative Classification ........................................................................................... 34 Coherence between the NPRS and Budget Execution...................................................... 35 43 Notable Deviations ........................................................................................................... The Central Role ofthe PIP.............................................................................................. Budget Execution and Treasury Operations ..................................................................... 45 48 Expenditure Review of the Health Sector......................................................................... 50 3. THE FISCAL RISKS AND REFORMCHALLENGES OF STATE-OWNED ENTERPRISES ............................................................................................................... 56 1. OVERVIEW OF PRIVATIZATION EFFORTS SINCE 1990 REDUCTION OF STATEOWNERSHIP INTHEECONOMY .............................................................................................................. 56 2. THESTATUSOF THE FOURSOEs...................................................................................... 58 58 The Financial Status ofthe Four SOEs............................................................................. Legal Status o f State-Owned Infrastructure Enterprises................................................... The ContinuedFinancial Connection............................................................................... 59 60 3. GOVERNMENT'S EXPOSUREINRELATIONTO SOES................................................ RISK 65 4. EMAE'sIMPLICIT FISCALSPACE, AND POLICY CHOICES....66 SUBSIDIES, GOVERNMENT 5. THENEED A STRENGTHENEDREGULATORYFRAMEWORK THE FOR FOR INFRASTRUCTURE SOEs ................................................................................................... 71 6. RECOMMENDATIONSFORTHE SHORT TERM(0-12 MONTHS) ......................................... 73 i BIBLIOGRAPHY ........................................................................................................................ 80 LIST OF ANNEXES .................................................................................................................... 81 Listof Tables Table 1: Selected Macroeconomic Indicators................................................................................. 2 Table 2: Sgo TomC and Principe-Contributions Table 3: Mid-Term Analysis (2005-10) ....................................................................................... to Real GDP Growth ........................................ 3 11 Table 4: Baseline Scenario (20 10-1 5) .......................................................................................... 13 Table 5: PRSP Financing to 2010................................................................................................. Table 7: Central Government Revenues - As a percentage o f total non-oil revenues (actual) .....21 Table 6: Central Government Revenues - As a percentage o f GDP (actual) ................................ 16 21 Table 8: Main Expenditure Categories......................................................................................... Table 10: Decomposition o f Staff Costs, 2005 .............................................................................. Table 9: Weight o f Mandatory Expenditure inthe Total Current Expenditure (Execution) ........28 30 31 Table 11: Distribution o f Staff Cost Components by Highest Five Sectors, 2005 ........................ 32 Table 12: Current Transfers from the State Budget....................................................................... Table 13: Most Important Spending Sectors (Current Expenditure) ............................................. 33 34 Table 14: Common Expenditure-Structure o f Spending............................................................. 35 Table 16: Total Budget Executedby Clusters o f Ministries.......................................................... Table 15: Pro-poor Expenditures................................................................................................... 40 41 Table 17: Total Budget Executedby NPRS Ministries inpercent o f GDP ................................... 42 Table 18: PIP-Main Table 19: Total Expenditure Deviation and Composition Variance .............................................. Spending Items........................................................................................... 43 48 Table 20: General Epidemiological Indicators, 2000-05 .............................................................. 50 Table 21:Ratio of HealthWorkers to Population, 2005 ............................................................... Table 22: Total Expenditures o f the National Health System, 2001-05 ....................................... 51 52 Table 23: Total Per Capita Expenditures inthe National Health System, 2001-05 ...................... 52 Table 24: Legal and Ownership Status o f the SOEs...................................................................... 57 Table 25: Summary Data for SOEs, 2005 ..................................................................................... 60 Table 26: Profit Sharing from Infrastructure SOEs to the Treasury .............................................. 62 Table 27: Taxes from Infrastructure SOEs to Treasury (Paid andor Owed) ................................ Table 28: Intra-SOE and Government Debt, October 2006 .......................................................... 62 Table 29: Loans from Financial Institutionsto Infrastructure SOEs via Treasury, ....................... 63 64 Table 30: Government's Combined RiskinRelation to Four SOEs (US$ millions) .................... Table 31: EMAE-Subsidies, Table 32: Descriptive Statistics-Access to Electricity, Consumption, and Expenditure ...............67 Tax Breaks, and Hypothetical Tariff Increases to break even ......65 Table 33: Estimates o f the Impact o f an Increase inElectricityTariffs on Poverty, 2000 ............69 69 Table 34: Estimates o f the Value o f R for Public and Private Transfers, 2001............................. 70 Listof Figures 4 Figure 3: Imports o f Goods and Nonfactor Services ....................................................................... Figure 1 Exports o f Goods and Services ........................................................................................ 5 Figure 4: Money, Credit, and Inflation............................................................................................ 6 Figure 5: ExchangeRate.................................................................................................................. 7 Figure 6: Public Expenditure Breakdown........................................................................................ Figure 8: Sgo TomC and Principe: ExpectedMDGOutcomes under Alternative Scenarios.........14 Figure 7: PIP under Baseline and Zero Deficit Assumptions ........................................................ 9 18 11 Listof Boxes Box 1: The NPRS .......................................................................................................................... 36 Box 2: Reforming the State Sector. 1990-2004 ............................................................................ 57 Box 3: The Positives and Negatives o f Multisector Regulation(MSR) ........................................ 72 List of Annexes . Annex 1.2: Framework o f the RMSM-X Poverty Module........................................................... Annex I1: Sao Tome & Principe Macroeconomic Assumptions 2005-2015.............................. 82 85 Annex 11.1: Tax hindrances to setting up business ....................................................................... 88 Annex 11.2: An alternative income tax proposal for the upcoming income tax reform...............89 Annex 11.3: Tax Institutions. tax structure as o f June 2006 and upcoming tax reforms .............-94 Annex 11.4: Local administration.................................................................................................. 97 Annex 11.5: Information constraints inthe management o f public investment ........................... 98 Annex 11.6: Data on sources o f external aid and PIP financing.................................................. Annex 111.1: EMAE's performance ............................................................................................. 100 101 Annex 111.2: Performance o f ENASA, ENAPORT and CORREIOS.......................................... 115 122 Annex 111.4: Poverty aspects o f electricity tariffs increases: ....................................................... Annex 111.3: EMAE-Tables....................................................................................................... 125 Annex 111.5: Assessing the Targeting Performance of Consumption Subsidies .......................... 126 Annex 111.6: Incentive-Based Performance Contracts for SOEs ................................................. 128 Annex 111.7: Stantardized fiscal template ................................................................................... 130 Annex IV: Per Tables ................................................................................................................ 132 ... 111 PREFACE This i s the final report o f the SBo Tome and Principe's public expenditure review (PER), undertaken by the World Bank in close coordination with the government o f ,530 Tom6 and Principe and the African Development Bank (AfDB). The government group consisted o f a core team in the Ministryo f Planning and Finance: Mr. Americo D'Oliveira dos Ramos (Director of Budget), Mme. Joana Varela (Director of Treasury), Mme. Ana Silveira (Sub-Director o f Patrimony), Mr. Felipe Muniz (Director o f Planning), Mme. Ilza Vas (Director o f Customs), Mr. Agostinho Fernandes (Director o f Taxation), and Filipina de Vera Cruz Rocha (Coordinator o f the NPRS-Poverty Observatory). The directors and staff o f the following state-owned enterprises also worked closely with the team: water and electricity (Empresa de Agua e Electricidade, EMAE), port management (Empresa Nacional de Administrag30 dos Portos, ENAPORT), airport management (Empresa Nacional de Aeroportos e Seguranca Aerea, ENASA), and post (Empresa de Correios, CORREIOS). The mission would like to extend its deepest thanks to his Excellency Mr. Tome Vera Cruz, Prime Minister; her Excellency Mme. Maria dos Santos Tebus Torres, Vice Prime Minister and Minister o f Planning and Finance; his Excellency Mr. Manuel de Deus Lima, Minister o f Natural Resources and the Environment; and all directors and staff o f the Ministry o f Planning and Finance as well as the directors and staff o f the state-owned enterprises for their welcome and their open cooperation. The analytical work of this report i s based on a main mission conducted in June 2006 jointly with the AfDB. Preliminary results were discussed in four meetings at the Ministry o f Planning and Finance inNovember 2006. Her Excellency Mme. Maria dos Santos Tebus Torres, Vice Prime Minister and Minister o f Planning and Finance, presided over the general meeting, where the directors o f the ministry were present or represented. The other three meetings presentedthe preliminary results to the directors and staff o f EMAE, ENAPORT,and ENASA. The PER team was led by Dorsati Madani (task team leader, AFTP3) included Natsuko Obayashi (task team leader, AfDB), Rafael Muiioz Moreno (AFTP3), Maria Vagliasindi (regulatory framework, FEU[), Quentin Wodon (AFTPM), Vitor Dinizio (consultant), Nilgun Gokgur (consultant), Alema Siddiky (consultant), and Thilakaratna Ranaweera (consultant), and was supported by Maude Jean-Baptiste. Maria Vagliasindi (FEU) and Demetrios Papathanasiou (ECSSD) reviewed chapter 3 o f the PER. HinhDinh(lead economist, AFTP3) provided the team with valuable suggestions andguidance. iv 1. MACROECONOMIC TRENDSAND FISCAL CHALLENGES' 1.1 This chapter reviews the recent economic performance o f Slo Tom6 and Principe (STP). The first sectionreviews the country's macroeconomic performance and highlightsthe impacts o f fiscal and monetary policies, as well as external factors, on the economy. The second section analyzes medium- and long-term fiscal sustainability with or without the discovery o f petroleum. The third section focuses on the financial realism o f the expenditure objectives o f the National Poverty Reduction Strategy (NPRS) and the prospects o f achieving some o f the Millennium Development Goals (MDGs).The final section provides policy suggestions. 1.2 The government started to revise its national accounts and the gross domestic product (GDP) series with assistance from the World Bank in 2005, following a new census and surveys available at the time. As a consequence, the new GDP series for 2001-06 that were adopted in early 2007 have been adjusted upward by some 50 percent compared to the old estimates. One major discrepancy came from underestimating the contributions o f some sectors, such as services. Even though the gross national income (GNI) per capita in 2005 has been revised to US$740, compared to US$400 inthe previous GDP series, STP remains a low-income country. 1. MACROECONOMIC PERFORMANCE AND GOVERNMENT POLICIES Stable GrowthbutRisingInflation 1.3 STP's real GDP growth has picked up momentum fi-om 4.5 percent between in 2004 to 5.4 percent 2005 and an estimated 7.0 percent 2006 inexpectation o f a possible oil discovery (see table 1). Starting in 1999, the government implemented structural reforms for economic diversification, deregulated prices, and undertook some privatization, all o f which helped to revitalize the economy. Growth performance accelerated in 2005 to 5.4 percent, based on a services (transporthourism) boom and on a construction boom that were mainly driven by foreign direct investment (FDI) in tourism in anticipation of a possible petroleum era. In 2006, election expenditures further fueled the growth to 7 percent. This high rate o f growth, which i s still less than the 8 percent considered to be the minimum growth rate to achieve MDGs (in the Poverty Reduction Strategy Paper, PRSP), i s not likely to be sustained if commercially viable petroleum i s not found and exploited inthe mediumto long term. 1.4 Real GDP growth i s affected by the unpredictable flow o f foreign assistance. Between 2000 and 2005, foreign assistance financed on average 78 percent o f the public investment; development projects, contracted to local construction companies, constitute a large part of this public investment. Consequently, unpredictable flow o f foreign assistance affects the construction subsector, which in2005 represented 12.1percent o f GDP. 1.5 The services sector i s the main motor o f the economy; it accounted for 56 percent o f GDP in2005.' The financial and banking system accounts for an estimated 2.2 percent of GDP, with new banks recently opened in anticipation o f the possible oil prospects. Tourism has become the most dynamic sector and the primary source o f foreign exchange for the country, with 64.5 percent o f total exports in2005 and an estimated 66.2 percent o f total exports in2006. 'Coauthored byNatsuko Obayashi, AfDB, and RafaelMufioz Moreno and Dorsati Madani, World Bank. In2003, this sector employed 52 percent ofthe active population. 1 Table 1: SelectedMacroeconomic Indicators 2000 2001 2002 2003 2004 2005 2006(p) Real sector Real GDP growth (%) 3.O 4.0 11.6 6.8 4.8 5.4 7.0 Nominal GDP (Dbbillions) 592.0 676.1 835.3 929.3 1060.9 1,188.7 1,526.3 GDP (US$ millions) 74 76 92 99 108 114 123 Population 148,000 151,000 154,000 157,080 160,222 163,426 166,695 GDP per capita 400 448 542 592 662 727 916 Monetary sector Average inflation, CPI (%) 12.2 9.5 9.2 9.6 12.8 16.3 23.1 Exchange rate (dobras per US$,period average) 7,978 8,842 9,088 9,348 9,780 10,470 12,457 Terms o f trade 100 125 186 167 136 123 118 Credit to economy (Db billion) 25.9 25.2 39.9 76.3 168.7 239.5 353.5 - Annual growth ("h) 1.4 -2.7 58.3 91.2 121.1 42.0 47.6 Broadmoney (M3) 102 152 190 276 297 433 603 - Annual growth (%) 16.8 36.7 26.9 41.8 7.4 45.9 39.3 Reserves (months o f imports) 3.4 4.0 3.9 4.8 3.4 4.1 4.7 External sector Current account balance (US$ millions) -29.8 -31.2 -29.2 -33.5 -37.8 -44.7 -80.2 - As percentage of GDP -40.1 -40.8 -31.7 -33.7 -34.9 -39.4 -65.5 Exports, goods and services (US$ millions) 16.3 16.1 18.5 20.7 20.2 21.0 23.6 - Annual growth (%) 11.2 -1.1 15.1 11.9 -2.5 4.2 12.1 Imports, goods and services (US$millions) 44.0 45.3 45.6 53.3 57.0 64.3 102.3 - Annual growth (%) -1.7 2.9 0.7 16.9 7.0 12.8 59.0 External debt (US$ millions) 302.4 302.5 295.8 289.2 281.2 272.4 Net present value o f debt outstanding 142 129 134 131 131 130 128 Fiscal indicators, cash basis (YOGDP) Budget revenue and grants 27.3 36.8 29.4 34.5 35.9 81.8 37.4 Non-oil revenue 13.0 13.2 13.6 15.2 16.8 17.4 21.2 Grants 14.3 23.6 15.8 19.3 19.1 17.2 16.2 Oil signature bonuses 0.0 0.0 0.0 0.0 0.0 47.2 0 Budget expenditure 38.7 49.6 39.0 44.7 51.7 44.5 51.3 Overall fiscal balance -11.4 -12.8 -9.6 -10.1 -15.8 37.4 -13.9 Source: IMF 1.6 The export-oriented agricultural sector, mainly cocoa production, has experienced tepid growth and agncultural exports amounted to only one-quarter o f those of the in 2005.3Although agriculture accounted for only 10.5 percent o f GDP in 2005, it is still a livelihood for 30 percent of the active population (2003) and is the major activity o f the informal sector. During2000-05, cocoa production continued to decline in volume, but a 97 percent increase in the international price of cocoa between 2000 and 2002 benefitedproducers as it increased cocoa export value by 3 Cocoa typically accounts for over 95 percent o f goods exports. Land reform, launched in 1991 and supported by the World Bank and the African Development Bank, was not successful partially because the land was distributed under state concessions and usufructuary rights, which limited access to credits and investment in the sector. Some recent international projects are trying to reverse the dependency on cocoa by introducing high-value-added crops such as pepperand vanilla. 2 42 percent between 2000 and 2003. However, starting in 2004, exports o f cocoa were affected by a two-year drought and the softening o f world cocoa prices. 1.7 A tight fiscal and monetary policy contained inflation to around 9.5 percent on average during 2000-03, but inflation began to trend up starting in 2004 because o f expansionary fiscal and monetary policies and external shocks. In2004, accommodating monetary and expansionary fiscal policies led to increased (by 121 percent) credit to the private sector and larger public expenditures in anticipation o f oil bonuses that did not materialize, pushing inflation to 12.8 percent. The inflation rate further accelerated to 16.3 percent in 2005 because o f a 42 percent increase in credit to the economy and soaring international oil prices that were partially passed through to domestic consumers. In2006, several factors contributed to an increase ininflation to an estimated annual 23.lpercent (table 1). Food shortages in early 2006 led to price pressures in the domestic market. Significant international oil price increases and the government's decision to pass through some o f the increases in fuel prices to move closer to world prices contributed at least 2 percent to the inflation rate. The inflow o f capital to finance presidential, legislative, and local elections as well as an expansionary fiscal stance, including some unplanned spending financed by central bank credit, also contributed to the inflation. Renewed expenditure discipline starting in May 2006 helped curb inflation from a 12-month high o f 26 percent in August to 25 percent inDecember. The central bank also revived its efforts to absorb the excess liquidity inthe last quarter o f 2006. Table 2: Silo Tom6 and Principe-Contributions to Real GDP Growth Percentage growth from previous year 2003 2004 2005 Gross domestic product, mp 6.8 4.8 5.4 Primary sector 5.9 1.6 2.7 Agriculture 6.0 1.I 2.5 Secondary sector 9.8 4.4 4.6 Ofwhich: construction 12.0 3.4 4.1 Tertiary sector 5.7 5.7 6.5 Public administration 3.0 3.5 4.5 Source: IMF PatternsinTrade-A StructuralExternalDeficit 1.8 STP's economy i s characterized by a structural external deficit that i s the result o f its narrow and insular domestic market, with high internal and external transaction costs and a limited domestic production and export base. These characteristics hinder development o f economies o f scale and the valorization o f some potential comparative advantages. Exports o f goods and services represented 18.4 percent o f GDP in 2005. STP's traditional main export i s cocoa (90 percent o f exports o f goods), but it has recently developed a tourism sector that represents 64.5 percent o f total export o f goods and services, equivalent to four times cacao revenues in2005. Expenditure on imports i s highand includes diversified products, from food to oil. Imports represented 56.4 percent o f GDP in 2005. The current account deficit (excluding official transfers) averaged 37 percent o f GDP between 2000 and 2005, financed mainly by official aid flows (grants and loans) and FDI.4In2006, the current account deficit increasedto an estimated 65.5 percent o f GDP, which reflected higher oil prices and investment-related imports o f goods and services for explorations in block 1 o f the joint development zone, foreign investment inthe tourism sector, and increased imports o f consumption goods related to electoral expenditures. 1.9 Terms o f trade are extremely sensitive to international price fluctuations, particularly o f cocoa and oil. STP's terms o f trade improved by 86 percent between 2000 and 2003 because o f an increase in the international price o f cocoa. However, the terms o f trade reversed after 2003 because o f a fall in the price o f cocoa and soaring international oil prices. The managed-float exchange rate regime helped contain the impacts-of volatile prices -for cocoa and oil on the external current account deficit. Figure 1 Exports of Goods and Services E]Cocoa 0TravelandTourism Q Other Services Other Goods Sources: STP government; IMF. Around 75 percent o f STP trade is with Europe (2004), with Portugal as the main exporter (57 percent o f total) and the Netherlands as the main importer (55 percent o f total). The intra-regional trade with ECCAS (Economic Community o f Central African States) represents 14 percent of STP trade, with Angola as the main partner. Angola is the destination of 3.7 percent o f STP's exports and the origin of 15.6 percent of STP's imports. 4 Figure2: Imports of Goods andNonfactor Services 701 1 ~ pFood 0Wtroleumproducts Q Investmentgoods Freight and insurance 0Other Goods Other nonfactorservices E]Technicalassistance Sources: STP government; IMF. 1.10 The trade system i s considered open, with no quantitative restrictions. Since 1998, STP has engaged in wide-ranging trade liberalization, such as the liberalization o f current account transactions in 1999, the elimination o f nontariff barriers, and the liberalization o f the exchange rate regime. The obligation to surrender export receipts to the central bank was abolished. The tariff schedule was reformed in 2002 to allow for four bands spanning zero to 20 percent, with some exemptions and surcharge^.^,^ The government continues to implement a customs tariff reform in order to simplify the tariff structure and increase transparency, compliance, and revenues. It i s pursuing gradual elimination o f remaining customs exemptions, removing the central government import tariff exemptions in mid-2006. The nontariff barrier linked to monopoly practices relates to imports, and distribution by state-owned enterprises has been removed progressively with a partial privatization process. The last nontariff barrier protecting the domestic telecommunications company expired at the end o f 2005. Monetary Policy-The Need for a New Strategy for the Future 1.11 Monetary policy has been relatively accommodating in the past few years, shadowing fiscal policy rather than establishing an independent policy approach. Consequently, broadmoney (M3) grew on average 29 percent annually between 2000 and 2005, and credit to the economy grew by an annual average o f 78 percent between 2002 and 2005. Preliminary data suggest that the 2006 monetary policy has been in line with previous years' stance, with M 3 increasing 39 percent and credit to the economy increasing 47.6 percent. In2002, the government shifted some tariffs on agricultural and industrial goods among three tiers of 5 percent for basic goods, 20 percent for luxury goods, 10 percent for most other goods, and 0 percent for a variety o f capital goods, fertilizers, and seeds ifthey are imported by official agencies, as well as other selected goods such as milk and flour. Some imports, most notably oil products, are subjected to significant surtaxes that essentially serve as excise taxes. All taxes on exports were removed. There are still some exceptional restrictions on the export of wood, stone, and sand, for environmental protection purposes. 5 1.12 This accommodating policy-especially in the face of excess liquidity related to the capital inflow to finance elections in the first half o f 2006-fed inflation, as discussed above. In effect, the central bank o f STP did not intervene actively in the market in the first half o f 2006, because it wanted to ensure minimum required net international reserves and because it was concerned about the medium-term costs o f issuingcertificates o f deposits (CDs) to absorb excess liquidity. The central bank's main intervention has been through the regulations on reserve requirements; that is, unifying the reserve requirement at 22 percent (up from 15 percent for the local currency deposits and down from 30 percent for the foreign currency deposits). However, under the country's 2005-07 Poverty Reduction and Growth Facility (PRGF) program, the central bank i s also to conduct monetary policy through indirect monetary instruments, such as foreign exchange auctions7 and the issuance o f CDs to sterilize excess liquidity while the real interest rates are kept positive. Monetary objectives of the program are to keep foreign reserves above three months' import needs, reduce inflation to 10-15 percent in 2007, limit the base money increase to 26 percent, and improve supervision o f commercialbanks. 1.13 A managed-float exchange rate regime has been inplace since 1997, and the payments system has been liberalized. Therefore, exchange rate fluctuations have generally reflected STP's economic situation, terms o f trade, and capital flows, even though there are still some punctual deviations. Since 1999, the spread between the official rate and the parallel rate has been contained to 3 percent. The adoption o f a flexible exchange rate regime helped the country to adjust its economy to exogenous shocks. However, persistent inflation depreciated the dobra annually by 6 percent on average between 2000 and 2005. The depreciation against the dollar was an annual average o f 4.6 percent in 2004 and 18.7 percent in 2005. It declined to an annual average o f 9.6 percent in2006. 50 M o n e y supply ( M 3 ) Credit to the economy 1/ 40 - 30 - 20 - 10- 0 --- 2000 2 01 2002 2003 2004 2005 -10 - -20 Sources: STP government, IMF 'Thisinstrumentwas introducedin late 2004, abandonedin early 2006, andresumedin late 2006. 6 Figure 4: Exchange Rate Dobras per U S dollar 12000 - 10000 - - A A 2000 - 0 , , Sources: STP government; IMF. 1.15 The banks' main activity i s foreign exchange transactions, and private sector access to credit remains limited mainly to short-term credit to established commercial structures. Even though credit to the private sector has expanded from 4.4 percent o f GDP in2000 to 20 percent o f GDP in 2005, it is concentrated in commercial activities. Investment i s deterred because the spread between the credit rate (32.5 percent nominal in 2005) and the deposit rate (12.8 percent) i s one o f the highest in sub-Saharan Afhca.' Furthermore, about 57 percent o f M 3 i s composed o f foreign currency deposits, which signals a lack o f confidence inthe domestic banlung system and the curren~y.~ 1.16 Inview ofthe recentmonetarydevelopments, the authorities should seek answers to three important questions to help them set medium-term monetary policy: (i) can a country with small net international reserves fight off the monetary impacts o f large currency movements such as those experienced in STP in the first half o f 2006; (ii) if so, do the existing policy instruments work for STP; and (iii) ifnot, would adopting aregional currency (like that of the Economic and Monetary Community o f Central Africa (CEMAC), the CFA franc or the euro or dollar be more advantageous to the country's monetary and economic stability. The authorities are discussing possible membership with CEMAC. In-depth analysis o f these questions would shed light on the policy options available to authorities. Fiscal Policy-The Needto Address the ChronicStructuralDeficit 1.17 Si50 Tom6 and Principe faces a structural fiscal deficit mirroring its external structural (non-oil) deficit. While revenues have improved because o f better taxation policy and collection, continued donor support, and a petroleum signature bonus, expenditures have risen markedly between 2000 and 2005. More significantly, authorities may wish to review the quality o f their expenditures since current expenditures have risen rapidly during2000-05, reducing fiscal space, especially as it relates to investment (see chapter 2). 'Cape Verde hada spreadof 9.6 percentin2003, and the average of all African bankswas 2.6 percent in2002 (Emilio Sacerdoti, "Access to BankCredit inSub-SaharanAfrica: Key Issuesand ReformStrategies," IMFIWPIOSII66). IMF, September2006. 7 1.18 Tax revenues have improved markedly with the introduction o f new taxes, an increase in tax rates, elimination o f tax exceptions, and improved tax implementation, including simplification and improvement in customs. Tax revenues represented 15.1 percent o f GDP in 2005 and an estimated 15.6 percent in 2006, compared with 10.0 percent in 2000. In 2003, the authorities introduced an excise tax o f 2 percent on local services , which was increased to 5 percent in 2005, with a new tax hike programmed to 7 percent in 2007. The government has eliminatedtax exemptions for the state on investment products in 2006 and updated the reference prices incustoms in2005 and again in2006 with new prices ineuros. The most successful reform was the simplification o f the customs tariff structure; more than half o f fiscal revenue comes from customs, mainly import duties and taxes on imported goods (see discussion above). Furthemore, subsidies on kerosene were financed with higher taxes on other petroleum products, and the increase in international oil prices was partially passed through to consumers, which has helped the government to manage the recent hike inoil prices. 1.19 Although traditionally more than half o f total public revenue i s from foreign assistance, the payment o f one oil signature bonus in2005 contributed to revenues. The bonus-awarded for signing the production-sharing contract on block l-amounted to US$49.2 million and was disbursed to the National Oil Account in 2005. Production-sharing contracts were also signed for blocks 2 through 4 in 2005, and STP's share o f US$28.6 million will be disbursed in 2007 after repaying US$15 million to Nigeria for a short-term loan. The funds in the National Oil Account have allowed the authorities to allocate to the budget US$14.3 million in 2005 and US$15.6 million in2006, with USS8.0 million expected for the 2007 budget. 1.20 Government expenditure, particularly current expenditure, has increased since 2000 at the expense o f public investment. Government expenditure has increased to 44.5 percent o f GDP in 2005 and an estimated 46.4 percent o f GPD in 2006, and accounted for 49.6 percent o f GDP in 2001 (see figure below). Furthermore, the share o f current expenditures rose from 18 percent o f GDP in 2001 to 26.5 percent o f GDP in 2005 and an estimated 25.4 percent o f GDP in 2006. Most o f the expenditure components have increased, particularly personnel costs (rising from 6.0 percent o f GDP in 2001 to an estimated 8.7 percent o f GDP in 2005) and transfers (rising from 1.9 percent o f GDP in 2001 to reach an estimated 6.6 percent o f GDP in 2006)." Only interest payments have decreased, because o f the interim debt relief granted to STP. Capital expenditure, inturn,has shrunk from 28.7 percent o f GDP in2000 to an estimated 21percent o fGDP in2006. Moreover, the authorities have dedicated an increasing share o f the public investment program to current expenditures: only 41 percent o f capital expenditure i s actual investment in 2005, compared with 63 percent in 2001. When the proportion o f current expenditures i s accounted for, the actual public investment represents only 8.6 percent o f GDP, compared with 18.0 percent o f GDP in2001. 1.21 Although revenues improved, increasingpublic expenditures and the cycle o f weak fiscal discipline followed by fiscal tightening have led to wide swings in the overall fiscal deficit. The overall deficit widened from -9.6 percent o f GDP in 2002 to -15.8 percent in 2004, shrunk to -9.8 percent in 2005 and rose again to an estimated -13.9 percent o f GDP in 2006. Under the auspices o f the Staff-Monitored Program o f the International Monetary Fund (IMF), the deficit was contained to -10.1 percent o f GDP in 2003, but the increase in the monthly minimum wage o f civil servants from US$31 to US$40 (to address social demands anticipating oil bonuses that did not materialize on time) widened the public deficit again in 2004. Authorities reestablished loTransfers i s a category that sometimes incorporates salaries o f decentralized institutions. 8 discipline in 2005, but the confluence o f external shocks and elections-related expenditures affected the 2006 overall fiscal balance. 1.22 To reduce the fiscal deficit and improve the quality o f expenditure-more specifically, to provide some space for investment and pro-poor expenditure, the authorities plan to reduce the domestic primary deficit from an estimated -8.6 percent o f GDP in 2006 to a projected -7.0 percent o f GDP in2007. This new discipline includes restricting current expenditures to essential purchases, limiting salary increases to 20 percent, and increasing the share o f the Heavily Indebted Poor Country (HPIC) Initiative-related social expenditure from 2.3 percent o f GDP in 2006 to 3 percent o f GDP in2007. Figure 5: PublicExpenditureBreakdown Public Expenditure Breakdown % GDP 2000 2001 2002 2003 2004 2005 2006(p) 2007(f) Year Capital expenditure W Personnelcosts Q Goods and services 0 Interests on total debt 0 Transfers H Other HlPC Initiative-relatedsocialexDenditure Source: STP government 2. FISCALSUSTAINABILITY-SENSITIVITY TO POLICY CHOICES AND PETROLEUMPROSPECTS 1.23 We undertake an indicative analysis o f the mediumto long run o f the fiscal sustainability o f the budget given past policy behavior and the longer-term prospects for petroleum.'' The analysis suggests that past fiscal behavior will not be sustainable in the medium term, even with debt relief. The government will need to adopt fiscal adjustments as soon as possible. We also note that even if some petroleumproduction revenues start flowing to the country inthe middle of the 2010s, the country will still need to adopt fiscal discipline for the interim period. Furthermore, "WeuseanRMSM-X(RevisedMinimumStandardModelExtended)model. 9 if the government wishes to achieve the country's NPRS goals, it will need to redirect its expenditures away from current expenditures and toward investment. Medium-TermAnalysis (2005-10) 1.24 The medium-term analysis (2005-10) i s consistent with the IMF-supported PRGF program. Our analysis suggests annual GDP growth o f between 6 percent and 7 percent. Population growth i s expected to remain at 2 percent per year. Monetary policy will stay the course and aim to reduce annual inflation from 23.1 percent in 2006 to 6.5 percent in 2010. Foreign direct investment will peak to US$14 million in2010, and foreign grants will increase to US$22 million in2010.12 1.25 Assuming the same fiscal behavior as in the past few years, if no new oil bonuses are paid in 2007 or 2008, there will be a financing gap o f 4 percent o f GDP in the public budget as early as 2008. To prevent current expenditures from further crowding out the public investment program (PIP), the government should start retrenching current expenditures as soon as possible. Infact, the government isplanningatight control ofits current expendituresandsalary increases inits 2007 budget. Government revenues are expectedto stay at an average of 19percent ofGDP inthe medium term. We assume that the country will be able to generate increasing revenues to keep the pace o f economic growth through higher indirect taxes and improved tax collection once the tax system i s overhauled (in 2007-08 reforms). Total spending relative to GDP would be trimmed to 35 percent in2010. This reduction i s the consequence o f a strong reduction in current expenditures from 26.5 percent in 2005 to 20 percent in 2010, which can be partly explained by the reduction in interest payments following the HIPC completion point in March 2007. We assume that capital expenditure (public investment program) will nominally stay around 15 percent o f GDP up to 2010, showing a constant effort by the country to implement the NPRS in a framework o f limited additional foreign resources. But we account for the fact that some 50 percent o f the funds assigned to capital expenditures (PIP) are used to finance current expenditure (see the next chapter for further discussion on this topi^).'^ In 2005-06, the fiscal deficit was financed with grant contributions from donors and drawings o f some US$30 million from the oil account, associated with the signature bonus from block 1. In2007, the government can expect a small increase in its fiscal space by receiving debt relief after the HIPC completion point. This debt relief i s to be directed to social sectors and pro-poor expenditures, including investment. Although the government will also benefit from the use of some US$8 million from the National Oil Account (from the signature bonus from block l), sum available from this account is the already dwindlingrapidly. Without further petroleum-related revenues, the account cannot be a source to finance further current e~penditure.'~If there are no additional petroleum-related revenues then between 2008 and 2010 the overall fiscal deficit (including grants) could reach 4 percent to 6 percent o f GDP, which the expected signature bonus (from blocks 5 and 6, approximately US$12 million) will not fully finance. 1.26 Even after enhanced debt reliefthrough the Heavily Indebted Poor CountriesMultilateral Debt Relief Initiative (HIPCMDRI), S5o Tom6 and Principe's external financing needs will not I'The efficiency of investment (Incremental Capital-OutputRatio, or ICOR) is to stay at 8 in 2006 and then be kept at 5-6 for the remainderofthe projectionperiod. l3This translates into the fact that this analysis will keep a constant share of public investment of 8 percent of GDP, which shows the effort to implement the PIP. l4Based on the current state of knowledge, between 2007 and 2010 the petroleumrevenue available to the country is the remainder of the first petroleumbonus and the prospect of receiving some US$12 million from the second bonus (blocks 5 4 ) , which is still heldinNigeria andis expectedto be released to STP in early 2007. 10 be met by the current flow o f donor funds and FDIafter 2008. Domestic absorption will remain above GDP because o f the economy's chronic structural deficit. The current account deficit will remain at around 50 percent up to 2010 as imports remain at 57 percent o f GDP and exports at 18 percent o f GDP. Consumption share to GDP will remain constant to 2010, even though current government expenditure as a share o f GDP will decrease because o f the reduction in interest payments after the HIPC completion point.15 Investment as a share o f GDP will increase to 45 percent o f GDP by 2010. While the authorities continue implementing the NPRS, private investment will remain steady throughout the period at around 30 percent o f GDP. Despite the financial inflows related to PIP and private investment, if there are no new oil bonuses and if corrective fiscal measures are not taken in 2007-08, then the overall fiscal deficit (including grants and current expected funds from the oil account) will not be fully financed in2010. Table 3: Mid-Term Analysis (2005-10) 2005 2006 2007 2008 2009 2010 National income and orices (Annualuercentaeechange, unlessotherwise specified) GDPat constant prices i.4 -7.0 6.0 6.0 6.0 6.5 NominalGDP (Db billions) 1,189 1,544 1,941 2,292 2,629 2,982 Average inflation, CPI 16.3 23.1 18.6 11.4 8.2 6.5 Governmentbudget (InpercentofGDP) Totalnon-oilrevenueandgrants 36.1 35.4 154.5 30.9 30.7 30.7 of which: grants 18.7 14.3 135.6 12.1 11.9 11.9 revenues(excludinggrants) 17.4 21.0 19.0 18.8 18.8 18.8 Total expenditure 42.1 49.0 39.9 37.0 34.9 34.7 of which: current expenditure 26.5 32.0 25.5 22.5 20.2 19.7 capitalexpenditure 15.7 16.9 14.4 14.5 14.8 15.0 Overallpublic deficit (excludinggrantsbefore drawingoil revenues) -24.7 -27.9 -20.9 -18.3 -16.2 -15.9 Overallpublic deficit (includinggrants beforedrawingoil revenues) -6.0 -13.6 114.6 -6.2 -4.3 -4.0 Drawingsfromthe oil fund 23.4 12.8 5.9 2.o 4.9 3.8 Overallpublic deficit (includinggrants after drawingoil revenues) 17.4 -0.8 120.5 4.2 0.6 -0.2 National accounts (Inpercent ofGDP) Consumption 99.6 102.0 105.2 105.2 109.6 107.4 Gross investment 37.8 44.5 38.4 38.3 34.9 37.6 Public investment 7.8 9.3 7.9 8.0 8.1 8.2 Privateinvestment 30.0 35.2 30.4 26.8 29.4 Gross domestic savings 0.4 -18.7 -30.5 16.6 -22.8 -21.1 -32.6 Gross nationalsavings -0.9 -23.2 98.5 -30.8 -30.3 -32.5 Public savings -7.5 -16.6 112.1 -8.6 -6.5 -6.2 Privatesavings 6.6 -6.6 -13.6 -22.1 -23.8 -26.3 External sector (Inpercentof GDP, unlessothewise stated) Current account balance(excludingofficial transfers) -38.7 -53.3 -52.1 -50.9 -51.6 -42.5 Current account balance(includingofficial transfers) -18.7 -34.5 -36.0 -35.1 -36.1 -27.0 Exports earnings (US$ millions) 21 23 24 26 28 30 Value of imports(US$ millions) 64 78 80 85 91 98 Exchangerate(dobrasper US dollar, periodaverage) 10,660 12,665 14,301 15,588 16,509 17,582 Realexchangerate (2005 = 100) 100 100 98 97 95 96 Source: STP Government and World Bank l5The country reached the HIPC completion point in March 2007 and qualifies for additional debt relief from the IMF, the International Development Association (IDA), and the African Development Fund under the Multilateral Debt ReliefInitiative(MDRI)from 2007. 11 Long-TermAnalysis (2010-15) 1.27 For the period 2010-15, we consider two scenarios. Inthe baseline scenario, petroleumi s not commercially exploited and oil bonuses are the only resources coming from the oil sector. In the alternative scenario, we consider the most optimistic scenario where oil production i s commercially exploited starting inthe middle o fthe 2 0 1 0 ~ . ' ~ 1.28 Inthe baseline scenario, annual GDP growth is expected to drop from the 6-7 percent range up to and through 2010 back down to 5 percent as expectations o f finding petroleum fade. GDP growth will remain at 8 percent in the alternative (oil) scenario through 2015. The central bank will continue to adopt policies that lead to a slight reduction o f annual inflation from 6.5 percent in 2010 to 6 percent in 2015 in both scenarios. We assume nominal exchange rate depreciation will follow the inflation rate from 2010 onward. This assumption implies that the real exchange rate will appreciate 3 percent between 2010 and 2015. 1.29 We expect imports and exports to grow at the nominal rate o f growth o f the economy in both scenarios after 2010. In the baseline scenario, imports will increase by a factor o f 2.6 and exports will increase by a factor o f 2.4 between 2005 and 2015. The alternative scenario assumes that imports will be 3 times larger and exports 2.8 times larger in 2015 compared with 2005, because most o f the additional resources the country will get from oil will be used to import products. In the baseline scenario, FDI will steadily decrease from its peak in 2010 to US$8 million, to account for fewer investment opportunities; in the oil scenario, FDI will increase to US$20 million to respond to opportunities in anticipation o f oil exploitation. Foreign grants will remain at US$22 million in the baseline scenario to help the country implement its NPRS and reach the MDGs. Inthe alternative oil scenario, on the other hand, donors will rapidly disengage from STP. BaselineScenario 1.30 If oil is not commercially exploited, then maintaining existing fiscal policy will further widen the public deficit to unsustainable levels: the overall fiscal deficit after grants and drawings from the National Oil Account will widen to 2.6 percent o f GDP in2015. Most o f the deficit will go unfinanced after 2010, since oil bonuses will have been usedup and no other oil revenues are expected. 1.31 STP will not be able to finance its current account deficit and will have to restrain its expenditure policies or accumulate a large foreign debt. In a framework o f lower than expected growth, STP will not be able to keep its large current account deficit and will have to retrench imports and consumption, particularly public consumption, to the country's financing means. The current account deficit will increase from 27 percent o f GDP in 2010 to 50 percent o f GDP in 2015, because a 34 percent increase inimports cannot be matched by the more modest increase in exports." Investment will decrease to 29 percent o f GDP, assuming that there are reductions in both private and public investments. Furthermore, FDIand external grants will not be enough to finance the current account deficit, and the gap fill (the difference between imports and exports) will widen from 6 percent o f GDP in2011to 33 percent o f GDP in2015. l6Assumptions for the oil scenario are taken from Segura, A., "Management o f Oil Wealth Under the Permanent IncomeHypothesis: The Case of Si0 Tome andPrincipe," IMF2006. This model assumes that importswill increase (from 58 percent o f GDP in 2010 to 78 percent of GDP in 2015) and that they will not be matched by an increase in exports, which will continue to account for about 18 percent of GDP throughout the period. 12 1.32 Therefore, if no corrective fiscal measures, such as improving domestic revenues and restraining current expenditure, are taken well before 2010, the impact o f the fiscal gap would be felt on the PIP, cutting it by half in 2015 (see figure below). This reduction would imply that NPRSplans andhopes o f achieving the MDGgoals for STP would have to be scaled back. Table 4: Baseline Scenario (2010-15) 2010 2011 2012 2013 2014 2015 National income and prices (Annual Percentage change, unless otherwise specified) GDP at constantpricks 6.5 5.0 5.0 5.0 5.0 5.0 Nominal GDP (Db billions) 2,982 3,319 3,694 4,111 4,576 5,093 Average inflation, CPI 6.5 6.0 6.0 6.0 6.0 6.0 Government budget (Inpercent ofGDP) Total non-oil revenue and grants 30.7 31.4 31.0 30.5 30.1 29.7 ofwhich: grants 11.9 12.6 12.0 11.4 10.9 10.3 revenues (excluding grants) 18.8 18.9 19.0 19.1 19.2 19.4 Total expenditure 34.7 34.0 33.7 33.5 33.4 33.4 of which: current expenditure 19.7 19.8 20.2 20.7 21.2 21.9 capital expenditure 15.0 14.2 13.5 12.8 12.2 11.5 Overall public deficit (excluding grants before drawing oil revenues) -15.9 -15.1 -14.7 -14.4 -14.1 -14.0 Overall public deficit (including grants before drawing oil revenues) -4.0 -2.6 -2.8 -3.o -3.3 -3.7 Drawings from the oil fund 3.8 2.9 2.3 1.8 1.4 1.o Overall public deficit (including grants after drawing oil revenues) -0.2 0.4 -0.5 -1.2 -1.9 -2.6 National accounts (Inpercent ofGDP) Consumption 107.4 121.0 125.0 129.1 133.5 138.1 Gross investment 37.6 29.2 29.2 29.2 29.2 29.2 Public investment 8.2 7.8 7.4 7.0 6.7 6.3 Private investment 29.4 21.4 21.7 22.1 22.5 22.9 Gross domestic savings -32.6 -29.8 -32.7 -35.9 -39.4 -43.2 Gross national savings -32.5 -30.0 -33.2 -36.6 -40.3 -44.6 Public savings -6.2 -6.0 -6.0 -6.1 -6.3 -6.6 private savings -26.3 -24.1 -27.2 -30.5 -34.0 -38.0 External sector (Inpercent of GDP, unlessothewise stated) Current accountbalance (excluding official transfers) -42.5 -46.0 -49.5 -53.2 -57.2 -61.7 Current accountbalance (including official transfers) -27.0 -31.2 -35.5 -39.8 -44.5 -49.6 Exports earnings (US$millions) 30 33 37 41 46 51 Value of imports (US$ millions) 98 109 122 136 151 168 Exchangerate (dobras per US dollar, period average) 17,582 18,637 19,755 20,941 22,197 23,529 Real exchange rate (2005 = 100) 96 96 97 97 98 99 Source: STP government and World Bank 13 Figure6: PIP underBaselineand Zero Deficit Assumptions 25 -- & A - A A A v - v v 20 - us% -+ 10 - 5 0 4 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year -cPIPbaselineScenano-a- PIPtoAccomdateContinuousCurrentExpenditureRlseandZeroDeficit Source: STP government and World Bank Alternative Scenario 1.33 Even if oil production comes on line in the first half o f the 2010s, continuing with past patterns in public expenditure i s not sustainable. The law on managing petroleum revenues does not allow borrowing against future revenues, and donors are expected to discontinue their STP program once oil discovery i s confirmed. These issues would clearly worsen the macroeconomic balance until STP starts to actually receive oil revenues. 1.34 FDIand donor financing will not be enough to finance the external gap during2010-15. Imports to meet domestic demand are expected to double between 2010 and 2015 in anticipation o f larger drawings from the oil fund. Although exports of goods and services are also assumed to double, they will start from a small initial base, which will lead to an increase in the current account deficit. FDI inflow will rise on the expected new domestic demand, but foreign grants will fall as donors reallocate their funds to other countries. As a consequence, STP will have a widening external financing gap until oil revenues are available to finance it. 1.35 The above discussion highlights the importance of adjusting the domestic expenditure to accommodate (i)actual oil revenues (either new bonuses or production) and (ii) revenues from donors. The government will have to scale back expenditure until production revenues start flowing into its coffers. Furthermore, the government will need to refocus its budget on investing if it intends to reach the NPRS and MDGgoals which will restrain current expenditures. Ifthe mix o f expenditures does not change, the needed overall fiscal constraint may fall more heavily on capital expenditure, which would affect the implementation o f the NRPS action plan and the hopes o f reaching MDGtargets. 1.36 Donors will have to remain financially committed to STP in the interim period between oil discovery being confirmed and the country starting to receive petroleumproduction revenues. Ifthe exploitation of oil is confirmed inthe coming years, there will be several years before the country can actually benefit from the first oil productionrevenues. Giventhe financial constraints that the public budget faces, donors should envisage a phaseout strategy to accommodate the progressive inflow o f oil revenues. 14 Conclusions andRecommendations 1.37 Inthe medium term, NPRS implementation is limited by donor commitments, its main financing source. However, the government may improve PIP effectiveness and the possibility o f approaching MDG targets by reducing current expenditure that i s financed from the PIP budget from 50 percent to 33 percent, and also by refocusing the PIP budget on actual NPRS investments. 1.38 Inthe mediumterm, the fiscal deficit will be financed through oil bonuses, HIPC debt relief, and external grants. However, if no new oil bonuses are paid in 2007 or 2008, there will be a financial gap o f 4.2 percent o f GDP as early as 2008. To avoid the PIP'S being further scaled back, the government should start retrenching current expenditure as soon as possible. In this regard, the authorities have already adopted a 2007 macroeconomic framework based on fiscal discipline. 1.39 In the long run, current fiscal policies are not sustainable. If petroleum is not commercially exploited, then STP will need to find other sources o f growth and scale back current expenditure to create fiscal space needed to implement the NPRS and the PIP. If petroleum i s commercially exploited, current policies cannot be sustained until the government starts to receive revenues from petroleum production. A reduction in current expenditure and a carefully programmed phaseout o f foreign grants will be neededto ensure a balancedtransition to the potential oil era. 3. REALISMAND ACTIVEPOLICY MAKING NEEDED THE NPRSARE IN FINANCIALPROGRAMMING 1.40 Following the fiscal sustainability analysis above, it is useful to assess the status o f the NPRS financial programming and achievements. The findings suggest that authorities will need to take immediate action to improve overall allocation o f resources to NPRS implementation as well as allocation among expenditure categories. 1.41 The resource envelope to support poverty reduction and to achieve MDGs has so far fallen short o f the NPRS action plan's financial programming. The action plan for 2003-10, estimated at US$210 million, i s supposed to be financed through external official transfers (loans and grants), HIPC debt relief, and the Treasury (see table 5)'' However, between 2003 and 2006, the financing o f the action plan has fallen short by a total o f US$8.6 million. 1.42 Moreover, during2000-05, the PIP-as a major instrument for the implementation o f the NPRS actionplan-included an increasing share o f current expenditure to the detriment o f capital expenditures and NPRS implementation efforts. Budget execution data show that around 50 percent o f the investment program in 2005 was used to finance current expenditure, compared with 33 percent in 2000. In other words, actual capital investment in the period 2003-06 was merely US$48.2 million (compared with the total PIP budget o f US$96.4 million), further reducing fiscal space to implement NPRS capitalprojects. HIPC interimassistance ends in2006, and we assume that HIPCMDRI will start in2007. 15 1.43 Although donor pledges after 2007 (especially HIPC and MDRI assistance) could make up the overall financing needs o f the NPRS action plan for the 2003-10 period, STP will still have to refocus the PIP on capital expenditures to see expected results. Donor pledges for the period 2007-10 will compensate for the early NPRS implementation shortfall and will bring the overall resource envelope for the 2003-10 action plan to US$214.1 million, in line with the total NPRS cost o f US$210 million (see table 5). However this amount will not be enough if STP does not correct for the highshare o f current expenditure inthe PIP. Even ifthis share i s scaledback to previous values (it was 33 percent in 2000), the actual public investment in the period 2007-10 would be US$77.6 million and the actual implementation o f the 2003-10 NPRS action plan would fall short by a minimumo f US$84 million. Table 5: PRSPFinancingto 2010 US$ millions - - 7007-1( PRSPfinancing needs 210 105 105 Total financing for the PIP 214.1 96.4 117.6 Treasury 26.6 13.5 13.1 External 149.9 70.2 79.7 Loan 19.3 11.6 7.7 Grant (public inv. projects) 130.6 58.6 72.0 HIPC (interim assistance) 12.8 12.8 0.0 C (c- +MDRT) 34.9 0.0 24.9 Source: IMF. 4. DESPITENOTED PROGRESS, MANY MDGSWILLNOTBEREACHEDBY2015 1.44 This section evaluates the impact o f alternative growth and public expenditure scenarios on achieving full literacy and four MDGindicators: universal primary education, life expectancy, infant mortality rate, and poverty headcount. The methodology used assumes that poverty, education, and health are affected by macroeconomic variables (such as inflation, GDP per ca ita, and social expenditures) and structural variables (such as openness and the urbanization rate)!9 A detailed explanation o f the methodology used can be found inAnnex I. 1.45 We assume three alternative scenarios for 2011-15. The baseline scenario assumes that petroleum i s not found,that GDP growth slows to 5 percent per annum, and that education and health expenditures are around 20 percent and 11 percent o f the budget, respectively. The alternative scenario assumes that the beginning o f oil resource expoitation occurs in the first half o f the 2010s, that GDP growth i s 8 percent per annum, and that education and health expenditures are around 21 percent and 12.5 percent o f the budget, respectively. A third scenario also assumes oil exploitation and very high and accelerating GDP growth rates (from 8 percent in 2010 to 14 percent in 2015), coupled with higher education expenditures o f 23.5 percent o f the budget and health expenditures o f 15.5 percent o f the budget.*' 1.46 The simulations show that economic growth and social expenditures can have a relatively high impact on poverty and social conditions, even in the short seven years to 2015. By The analysis is based on a framework developed at the World Bankusingpooled data for a group o f low-income countries and applied to the specific conditions o f Ssio Tom& (See Chen, Ranaweera, and Storozhuk, May 2004, The RMSM-X+P: A Minimal Poverty Modulefor RMSM-X,World Bank, Washington, DC.) The parameters ineach M D G equation are estimated using pooled cross-country time series regressions based on historical data from low-income *'countries. The parameters o f these three scenarios are summarized in Annex I. 16 extension, this analysis highlights the importance o f strengthening o f the NPRS and PIP implementation(noted above) as a prerequisite for medium- to long-term poverty reduction inthe country. 1.47 Given the short time frame to 2015, while SHo Tom6 and Principe will approach the MDGs, the country will not be able to reach many o fthem under any ofthe three scenarios. Only universal primary education and life expectancy seem to be well positioned to reach their MDG targets, because o f their better initial conditions. Neither the infant mortality rate nor the overall poverty headcount will reachtheir MDGtargets, although the simulations show that more growth and additional social sector expenditures would have a positive impact on poverty and infant mortality. 1.48 Poverty headcount will not reach the MDG target o f 20.5 percent in 2015 under any scenario. Currently the poverty headcount level i s about 54 percent, and under the baseline scenario economic growth could steadily reduce the poverty headcount to about 51 percent by 2015. Under the alternative scenario, higher economic growth after 2011, can help accelerate poverty reduction, but the gains would remain marginal because poverty headcounts would decrease to only 49 percent by 2015. Inthe third scenario, which assumes very highGDP growth rates, the poverty headcount would drop to around 42 percent in 2015, a notable improvement compared with the current 54 percent. Nonetheless, this percentage i s still more than double the MDG objective and marginally higher than the level recorded for 1990, the benchmark year for setting MDGs. 1.49 Projections show that SHo Tome and Principe i s well positioned to achieve by 2015 the target for life expectancy o f 63.3 years. Currently, the country shows a life expectancy o f 63 years, and under the baseline scenario it will reach the 2015 goal. Higher growth and social expenditures in scenarios two and three will cut down the time to achieve this goal by one or two years. 1.50 Although the infant mortality rate i s expected to decline progressively, it will remain far above the MDG target o f 20.3 deaths per thousand live births in 2015. Under the baseline scenario, infant mortality rates are projected to decline by 10 percent (from roughly 61 in2005 to about 55) by 2015. The decline in the infant mortality rate in the alternative and third scenarios (highgrowth and public expenditure) is only to 54 percent and 52 percent by 2015, respectively. This highresistance to decline even with additional growth and health expenditures may reflect the need to improve the efficiency o f health system expenditure and targeting (discussed in chapter 2). 1.51 The universal primary education goal (98 percent) i s within the reach o f Si50 Tom6 and Principe under the baseline scenario. Primary school enrollment i s already quite high in SHo Tom6 and Principe; 97 percent o f school-age children were enrolled inprimary schools in 2001, considerable progress from 85 percent primary school enrollment in 1985. 1.52 The achievements in universal primary education will affect the literacy rates as well. The literacy rate has already risen from about 73 percent in 1990 to 83 percent in 2005. By 2015, under the three scenarios the literacy rate i s projected to increase to about 89 percent, 90 percent, and 92 percent, respectively. Approaching full literacy and universal primary education also implies that additional resources will bring marginal increases in results because the large gains will have been captured already. 17 Figure7: Silo Tom6 and Principe: ExpectedMDGOutcomes under Alternative Scenarios Sao Tome: Poverty Headcount Rate +-BaselineAlternativeb-Very Hlg 63.6- 63.5- 63.4- 63.3- y e i d - A+**-A*--/ /- 30 63.1- 204 I 63.0- : : I 62.9- lo] 0 i;:;; 20052006200720082009201020112012201320142015 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year Year Sao Tome: Infant Mortality Rates Sao Tome : Primary Enrolment I+Baseline+ Alternatlvet- Very High ' ::: 62- \----a- 56- 54- 52- 50- 9 6 . 4 4 48- 4 6 3 , , , , , I , , , I 95.51 I I I , , , , , , , 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year Sao Tome: Literacy Rates 1tBarelinetAlternativet-VeryHigh 82- c 80- I Year Sources: STP Goverment and WorldBank 18 Recommendations Issuesto Address Action to Take Review overall expenditure policy to adjust the Take immediate steps to curb overall expenditures, especially budget to economic realities o f the country. current expenditures. Realign expenditure allocation to the development Stop using the PIP budget for current expenditures, refocus on priorities o f the country. public investment inline with the NPRS actionplan. Ensure donor commitment to the country for at Inform donors and partners o f the realities and uncertainties least untilthe 2013 budget year. attached to the nascent petroleum sector. Ensure better coordination and harmonization o f the donors. Issuesto Address Develop the next generationo f the NF'RS. Ensure donor commitment to the country untilthe country starts receiving oil revenues. Maintain macroeconomic discipline and in intrabudgetary allocations in line with the new NPRS. 19 2. FISCAL CHALLENGESOFTHE PREPETROLEUMERA- THE NEEDTO ENLARGEFISCAL SPACE 2.1 This chapter focuses on the two elements that provide fiscal sustainability and space for economic development. The first section reviews the revenue sources and trends during 2000-05. It notes the strength o f current revenue increases, based on import tax reforms and better overall tax collections over the period. It also registers the volatilitt o f nontax revenues, especially international aid, that affects investment and budget execution. 2.2 The second section o f this chapter benchmarks the expenditure patterns and trends over the first half o f this decade, highlightingthe importance o f more disciplined and policy-focused budget management practices. It first presents the scope o f the analysis and data constraints. It then analyzes expenditures in economic classifications and follows with a review o f current expenditures by spending units.N o standard and complete functional classification exists for data covering 2000 to 2006. Nevertheless, we develop three simple alternative methods to undertake this review and note that the coherence between policy orientations and budget allocations is weak, if not absent. We review the low execution rates in both the recurrent and the investment budget, noting the practice o f cash budgeting. We close with a review o f the allocative efficiency and equity o fpublic expenditure inthe health sector. 2.3 An analysis o f the public finance management and procurement systems (country financial accountability assessment, CFAA, and country procurement assessment review, CPAR) i s presented in the accompanying volumes to this public expenditure review (PER). Therefore, this chapter does not include any discussion on the public finance management system- including its weaknesses and the necessary reforms-that the government started reforming in 2005. 1. REVENUESOURCESAND TRENDS** Trends and CompositionofRevenues 2.4 Total revenues have increased almost tenfold between 2000 and 2006, holding the promise that the government could use these funds as a source to increase its fiscal space. This positive trend i s due to increases in total non-oil revenues and the receipt o f one large one-time petroleum bonus in 2005. Another bonus i s expected in 2007.23 The one-time bonus payment i s not expected to continue in the medium term, but it will impact the fiscal space through the oil revenue management law annual budget contributions (their disposition and use were discussed in chapter 1). 2.5 STP has made an important effort duringthe past five years to increase non-oil revenues. The share o f total non-oil revenues to GDP trends upward from 27.8 percent in 2000 to 34.6 percent in2005 and i s predicted to reach 35.6 percent in2006 (see table 6). These percentages are 2' Annex 11.2 includes a suggestion on the proposed income tax reform to improve progressivity and equity o f taxation, to decrease incentives for evading taxes and to increase participation of the informal sector. 22 The data for this analysis were obtained from STP authorities. Data for 2005 and 2006 are estimates. The methodology used in presenting the data is similar to that o f the IMF. We focus on the analysis o f the non-oil revenues and expenditures. Some data inconsistencies are not resolvable. 23 STP received Db 561.5 billion in 2005. The second bonus, to be paid in 2007, i s a total o f US$28.1 million, but only US$13.1 million will reach STP because the country will pay back the short-term loans it took from Nigeria inthe past three years (US$15 million). 20 significantly higher than the average o f HIPC countries24expected to receive relief (20.3 percent of GDP in2005) and Cape Verde (30.4 percent o f GDP). Table 6: Central Government Revenues As a percentage of GDP (actual) - 2000 2001 2002 2003 2004 2005 2 0 0 6 u Totalnon-oilrevenues (inc.grants) - . 27.8 37.8 29.4 34.4 36.8 34.6 35.6 Currentrevenues (tax and nontaxrevenues) 13.5 14.1 13.6 15.1 17.7 17.4 18.7 Tax revenue 10.0 11.6 11.4 12.3 13.9 15.1 15.6 Direct taxes 2.7 3.4 4.0 3.0 3.9 4.6 4.3 Profit tax 1.o 1.5 2.2 1.3 1.7 2.0 1.9 Salary tax 1.8 1.9 1.8 1.7 2.2 2.6 2.4 Indirecttaxes 6.4 7.3 6.5 7.6 8.4 8.9 9.5 Import duty 3.0 2.9 2.8 3.3 3.4 3.4 3.3 Imported goods excise tax 3.3 4.2 3.4 3.4 4.4 4.7 4.7 Local goods excise tax 0.1 0.3 0.2 0.8 0.7 0.8 1.6 Other* 0.9 0.9 0.9 1.7 1.5 1.7 1.7 NontaxRevenue 3.5 2.5 2.2 2.8 3.8 2.3 3.1 Transfer from state-owned enterprises 0.9 1.2 0.7 0.5 0.4 0.3 0.4 Fishing royalties 1.4 0.4 0.2 1.1 0.7 0.0 0.5 Other** 1.2 0.9 1.3 1.2 2.8 2.0 2.2 Grants 23.6 15.8 19.3 19.1 17.2 16.9 Budget support grants __ 14.3 2.6 0.1 0.3 1.9 1.5 0.6 Project grants 17.9 13.2 15.3 13.8 12.8 12.8 HIPC 3.O 3.5 Oilsignaturebonuses ______ _ _ 3.2 __ 2.6 __ 3.8 3.4 __ 47.2 45.6 NominalGDP 592.0 676.1 835.3 929.3 1,060.9 1,188.7 1,526.3 Source.: MinistryofPlanningandFinance, GovernmentFiscalOperations p:preliminarydata for 2 W . ***Other indirecttaxesincludemyriadsmaller taxes suchas sealtax, assistanceseal tax, propertytransfer tax, andurbanpropertytax Other nontaxrevenueincludesrevenues tom fines andfees, othernonspecifiedreceipts,and the 13 percent markupon oil sales Poma Nigerianoilgrant. Table 7: Central Government Revenues As a percentage of total non-oil revenues (actual) - 2000 2001 2002 2003 2004 2005 2006(p] Total non-oil revenues(inc. grants) 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Current revenues(tax and nontaxrevenues) 48.5 37.5 46.2 43.9 48.1 50.4 52.5 Tax revenue 36.0 30.8 38.6 35.7 37.7 43.7 43.7 Directtaxes 9.9 9.1 13.6 8.7 10.7 13.2 12.1 Profit tax 3.5 4.0 7.4 3.8 4.6 5.7 5.3 Salary tax 6.4 5.1 6.1 4.9 6.1 7.5 6.8 Indirect taxes 23.1 19.4 22.0 22.0 22.9 25.7 26.7 Import duty 10.8 7.6 9.7 9.7 9.1 9.7 9.2 Imported goods excise tax 11.7 11.1 11.6 9.9 11.9 13.6 13.1 Local goodsexcise tax 0.4 0.7 0.7 2.4 1.8 2.3 4.4 Other* 3.1 2.3 3.1 5.O 4.2 4.8 4.9 Nontax revenue 12.4 6.7 7.5 8.2 10.4 6.6 8.8 Transfer from state-owned enterprises 3.3 3.2 2.3 1.5 1.o 0.9 1.2 Fishing royalties 4.9 1.1 0.7 3.1 1.9 0.1 1.5 Other** 4.3 2.4 4.5 3.6 7.5 5.6 6.1 Grants 62.5 53.8 56.1 51.9 49.6 47.5 Budget support grants -_51.5 6.8 0.2 0.7 5.2 4.2 1.8 Project grants _- 47.4 44.8 44.4 37.4 36.9 35.9 HIF'C ___ _ 8.4 8.9 10.9 9.3 8.5 9.8 Oil signaturebonuses __ __ __ __ 181.0 164.4 Source.: Ministry of PlanningandFinance,Government FiscalOperations p:preliminarydatafor 2006. ***Otherindirecttaxes includemyriadsmallertaxes, suchas seal tax, assistancesealtax, propmy transfertax, and urbanpropmytax Othernontaxrevenue includesrevenues from fines and fee, other nonspecifiedreceipts,andthe 13 percentmarkupon oil sales fromaNigerianoil grant 24 Except Sudan. 21 CurrentRevenues 2.6 The government has been very successful inincreasing current revenues (tax and nontax revenues). This upward trend i s due to the steady improvement in tax collections. The improved direct and indirect tax collection i s related to a series o f fiscal reforms that started with the reforms of trade taxes in 2000 and that continue today. Annex 11.3 provides further information on the recent tax reforms. Figure8: DirectandIndirectTaxes inPercentageof GDP, 2000-06 160- 14.0- 120- 2wo 2001 2w2 2003 2004 2005 Source: STP Government. 2.7 The collection o f tax revenue over the period 2000-05 has increased from 10 percent o f GDP in 2000 to 15.1 percent o f GDP in 2005, and 2006 estimates continue this positive trend. This is a better performance compared with Niger's share o f tax revenue to GDP (10.4 percent in 2005), although it i s still far from Cape Verde's share (21.7 percent o f GDP). 2.8 The improvement has been especially notable in direct taxes. They comprise profit tax, salary tax, and a minor tax on urban property. Total direct taxes collected have increased; they represented 4.6 percent o f GDP in 2005, compared with 2.7 percent o f GDP in 2000.25 In corporate taxation, collection has quadrupled between 2000 and 2005 (from Db 5.3 billion to Db 29 billion) because o f improved institutional measures to better assess and collect taxes. The trend i s expected to continue in 2006, although the existing 45 percent corporate tax rate i s higher than international standards and encourages tax evasion and growth o f the informal sector (see Annex 11).The government's plan to reduce the corporate tax rate from 45 percent to 25 percent could reduce revenues in the short term, once the reform i s adopted and implemented, but it i s 25Today, direct taxes represent26.2 percent oftotal non-oilrevenues (excludinggrants), comparedwith 20.4 percentin 2000, which is above the average of HIPC countries expected to receive relief (16.7 percent of total revenues, excludinggrants, in 2005) and is similar to Cape Verde (28 percentof total revenues, excludinggrants, in2005). 22 expected to help reduce tax evasion and informality and to provide a more competitive incentive framework for foreign direct 2.9 The tripling o f the salary taxes between 2000 and 2005 i s mainly the result o f back-to- back public sector salary increases in the past several years. These increases have caused the public payroll to grow from 4.9 percent o f GDP in 2000 to 8.7 percent o f GDP in 2005. N o accurate data are available to analyze the pattern o f salary tax payments inthe private sector. 2.10 Indirect taxes are the main domestic source o f tax revenues. Indirect taxes comprise import duties, and excise tax on imported and local goods and services. Indirect taxes today represent 5 1percent o f current revenues (excluding grants), compared with 47.6 percent in 2000. These percentages are well above the average collection rates o f other HIPC countries (39.3 percent o f current revenues, excluding grants), although they are still low compared to Cape Verde, where they represent 61 percent o f current revenues (excluding grants). Indirect domestic taxes have risenbecause the excise tax rate on local goods and services increased from 2 percent to 5 percent in 2005 and i s supposed to increase to 7 percent in 2007. Despite the major tariff reforms o f 2000, import taxes have risen to contribute a steady 3 percent o f GDP during2000-06, a clear result o f better capacity, better valuation, and better collection efforts at the Directorate o f Customs (Annex 11). 2.11 Reducing import tax exemptions could improve indirect tax collection. Up to end-2005, exemptions alone cost Db 9 billion per year, or 9 percent o f the total amount collected by customs. Removing exemptions from state-owned enterprises (SOEs) should be studied as an option to reduce this type o f revenue loss. 2.12 Despite the positive developments on tax revenues, the revenue picture i s not all encouraging, since the share o f nontax revenues intotal current revenues has fallen by one-third. Nontax revenues include transfers from SOEs, fishing royalties, and myriad smaller fees and fines. Transfers from SOEs have fallen from Db 5.4 billion in 2000 to Db 3.7 billion in 2005 for two main reasons. First, between 1992 and 2004 a number o f SOEs were fully or partially privatized, thereby reducing the potential share o f profits that could be transferred to the budget." Second, o f the remaining four fully state-owned enterprises-water and electricity (Empresa de Agua e Electricidade, EMAE), airport authority (Empresa Nacional de Aeroportos e Seguranca ACrea, ENASA), post (Empresa de Correios, CORREIOS) and port authority (Empresa Nacional de AdministraqZo dos Portos, ENAP0RT)-the first three have reportedlosses for the recent past because o f petroleum price shocks, delays in payment by their main clients (including the government), and interenterprise arrears (see chapter 3 for an analysis o f these SOEs). 2.13 The predictability o f the European Union's payments for fishing rights in the waters o f SZo Tom6 and Principe needs to improve. Transfer o f royalties has had long delays in the past; 26No systematic analysis o f this reform and its fiscal consequencesis possible at this time due to lack o f data. Nonetheless, this tax reform could be an attractive feature for perspective foreign investors. Furthermore, ifthe reduction o f the tax rates is accompanied by continued improvements intax collection and by an effort to formalize economic activity(which would enlarge the tax base), then the reforms should permit the government to recoup its ''fiscallosses inthe mediumterm. The tax administration notes that the 10 largest companies constitute the bulk o f corporate tax collection. State-owned enterprises must transfer 80 percent o f their profits to the treasury because the government is the main shareholder (accounted as nontax revenue), with the remaining profits taxed as private firms are (accounted as profit tax). Private enterprises pay only the 45 percent corporate tax. The state kept 51 percent of ENCO (an importer of petroleum products) and 49 percent o f CST (a telecommunications provider). Air SLo Tomk, partially owned by the state, also has registered losses inrecent years and was dissolved in2006 after its sole plane crashed. 23 slippages affect budget execution and the financial gap. A new four-year fishing agreement with royalties o f US$0.6 million per year has been concluded with the European Union and i s pending approval by the European Parliament.29 Figure 9: BudgetedandActual CurrentRevenue Db millions 250,000 7 200,000 150,000 100,000 50,000 0 2wo 2w1 2w2 2W3 2 w 5 Source: STP government 2.14 Overall, the annual collection o f current revenues i s in line with budgeted amounts, highlighting the authorities' ability to project tax revenues accurately (see figure below). Execution actually surpassed projections slightly between 2000 and 2003, while execution fell slightly short o fprojectionsin2004 and 2005. Grants and ForeignFinancingof the Budget 2.15 While tax revenues have risennotably over the past five years, foreign aid continues to be the main source o f revenues for the government's budget. Foreign aid i s mainly bilateral, has some quarterly and yearly variability, and seems to be in decline. Volatility in these external revenues affects the government's ability to smooth treasury shortfalls. This in turn has led the government to ration cash in budget execution or to access credit at the central bank to finance spending. 2.16 Foreign aid has provided around US$17 million per year (US$15 million in grants and US$2 million inloans) to the budget. This amount financed 50 percent o f the total public budget and represented 17.2 percent o f GDP in 2005, which i s a high ratio o f dependence to external resources when compared with the average o f HIPC countries (4.1 percent o f GDP) or other island states such as Cape Verde (5.9 percent o f GDP). Furthermore, STP reached an HIPC decision point in 2000 and benefited from US$14 million in debt relief between 2001 and 2004. 29 The European Commission and SZo Tom6 and Principe have signed a four-year fisheries partnership agreement that will replace the previous agreement, which ended on May 31 2006. The new agreement will last until May 31, 2010, and will come into effect retroactively June 1 [2006 once it is approved by the EU Agriculture and Fisheries Council and the relevant authorities o f SZo Tom6 and Principe. The EUhas also agreed to a flat-rate financial contribution o f 663,000 per year, which includes 552,500 per year for European access to up to 8,500 tons o f tuna from SZo Tom6 and Principe's exclusive economic zone. Any additional catches beyond this overall amount will be paid retroactively. The total also includes a specific additional annual allocation o f110,500 to support SBo Tom6 and Principe's national fisheries management policy. 24 This debt relief has been channeled through the PIP to finance pro-poor projects. In2005, STP also benefitedfrom Paris Club rescheduling. 2.17 Foreign aid mainly comes in the form o f investment projects financed by bilateral Two-thirds o f foreign aid i s financed by bilateral donors, mainly Taiwan, which provides 52 percent o f total foreign aid to STP. The World Bank, which provided 7 percent o f STP's total aid in 2000-05, i s the fifth major donor. In2004 and 2005, Taiwan and the European Union granted budget support on an exceptional basis, representing 8 percent o f total grants to STP in2001-05. 2.18 There are significant shortfalls in annual realization o f grants compared with the budgeted amount, highlightingthe rationale for better donor coordination and harmonization (see figure 10). These shortfalls are correlated with delays in donors' delivery o f promised assistance. This fact, combined with delays inbudget approval and lack o f national implementationcapacity postpone project launches into the second or third quarter of the year and affect completion of investment projects. 2.19 More worrisome i s the trending down o f the relative share o f foreign aidgrants in total non-oil revenues: the share has dropped fi-om 56.1 percent o f total non-oil revenues in 2003 to 49.6 percent in2005, and i s estimated to be 47.5 percent in2006. This trend i s partially replicated inthe share of foreign aid and grants to GDP: in2005, grants constituted 17.2 percent o f GDP, barely superior to 2002 numbers (tables 6 and 7). This appears to be the consequence o f Siio Tome and Principe's entering the prepetroleum era; donors are assuming petroleum money will flow in the near future and are redirecting their funds to other countries. This early "disengagement" i s misguided because the best estimate suggests that petroleum production revenues will not flow before the middle o f the 2010s and because the country has plans to implement an ambitious PRSP. The December 2005 Donors' Round Table, based on the PRSP, presented a budget envelope o f some US$168 million. STP authorities received approximately US$78 million inpledges, little o f which was new money. Figure 10: Budgetedand Actual Grants (Db millions) 800,000 700,000 600,000 500,000 400,000 300,000 200.000 100,000 O 2000 2001 2002 2003 2 w 4 2005 10 OActualI Sources: Ministryof Planning and Finance. 30This topic i s discussedindetailin further below. 25 Issues to address Action to take The tax system, with high profit taxes, Promulgate the investment code, the draft corporate and multiple small taxes, and ad hoc exemption, i s a burden on business development (see Annex 11.3). Customs norms and procedures are too Promulgate the general code o n customs infractions and cumbersome and exemptions are broadly relatedregulations. granted. Eliminate tariff exemptions given to state-owned enterprises. Fiscal evasion and informality is pervasive. Fully implement the new the General Tax Code and Tax Procedural Code. Strengthen the capacity for tax estimation and verification and legal instruments for enforcement at the Directorate o f Taxation. Revenue and grant receipts are seasonal. Approve the budget before the fiscal year ends. Improve coordination and harmonization o f assistance with donors. 2. PUBLICEXPENDITURE PATTERNSAND TRENDS3' Objective and Scope of Analysis 2.20 The objective o f this second section i s to benchmark the expenditure patterns and trends during2001-06 (estimated), highlightingthe importance of more disciplinedbudget management practices and a more focused link betweennational policies, especially the NPRS, and the budget programming and outlay. 2.21 This section first presents the scope of the analysis and data constraints. Then it provides a brief overview o f expenditures, analyzes expenditures in economic classifications, and reviews current expenditures by spending units. N o standard and complete functional classification existed for the period 2001-06 (estimated). Nevertheless, we develop two simple alternative methods to undertake this review and note that the coherence between policy goals (NPRS) and the budget allocations i s weak, if not absent. To understand this disconnect, we review the low execution rates in both the recurrent and the investment budget, and the related cash-budgeting issue as well as the related institutional issues. We close with a review o f the efficiency and equity o f public expenditure inthe health sector (from a concurrent healthPER). 2.22 The scope of the present analysis i s the budget of the central g~vernment,'~which includes the central administration departments and political institutions such as the National The data cover budgetary expenditure over the past six years (2000 to 2005) provided by the existing budgetary classifications (administrative, economic, and functional) and the 2006 budget. 26 Assembly and the Presidency o f the Republic. Therefore, the budget includes neither the operations o f the state-owned enterprises nor the local administrations, except in what concerns the transfer of finds among those entities. 2.23 Duringthe period under analysis, STP used a dual budget system, which included two distinct sets o f budgetary information: (i) recurrent expenditure, managed by the Directorate o f Budget, and (ii) development expenditure (or the public investment program, PIP), managed by the Directorate o f Planning. The data producedby these budget information systems reveal some important shortcomings: dispersion o f information sources, nonstandardized output formats, and presentation o f budget outlays through a profusion o f Excel files with significant data dis~repancies.~~The only common classification to the two budget components i s the administrative classification (see the CFAA for further details on budget preparation, execution, and control). 2.24 Other data constraints have limited the scope o fthe analysis: The analysis o f the budget execution in the PER i s on a cash basis from the Budget Directorate's electronic database.34There are no charts o f accounts. The Treasury does not do a proper accounting o fthe state budget. 0 There was no standard functional classification for the period. 2.25 The magnitude o f directly executed donor projects i s probably underestimated. The data on investment come from the Directorate o f Planning, which i s responsible for programming and updating the PIP. The director notes long delays in execution and reporting by donors. In addition, we register an increasing proportion o f the PIP budget that i s being used for current expenditures. 0 Sectoral data rarely exist and if available are unreliable and often inconsistent. The health PER that was developed in parallel with this PER i s a first sectoral effort at remedying this data void. The results ofthe healthPERare summarizedinthis chapter. 2.26 Although such limitations do not compromise the results and conclusions o f this analysis as a whole, one notes discrepancies in sums from different sources o f information and across different categorizations. All attempts have been made to reconcile these differences. The overview of total expenditure highlights a weak budget discipline, growing current expenditures, and a shrinking investmentbudget. 2.27 The overview o f total expenditure between 2001 and 2005 (table 8) highlights the weak budgetary discipline o f a country that has been living beyond its means in anticipation o f the arrival o f petroleum revenues (see chapter 1). There are large increases in current expenditures and a shrinking development budget. Current expenditures rose steeply over the period 2001-05, 32 According to the definitions in Government Finance Statistics Manual, the budget of the general government is composedof the budgetsofthe centralgovernment, the local government, and social security.Inthe case of STP, there i s no consolidatedbudgetthat includesall ofthose components. 33 Recent progress has been made to set up procedures for compiling budgetary information in accordance with the IMF'sGovernment Finance Statistics Manual. However, the output ofthis initiative is not yet visible. 34 A complementary registrationis also made on a commitment basis (base de liquidapio). The Budget Directorate's Division of Expenditureauthorizes the expenditureson the basis of the appropriations set out in the budget law. This operationis calledliquidup70 andcan roughlybe likenedto a"commitment." 27 evolving from 18 percent o f GDP in 2001 to a peak o f 27.6 percent o f GDP in 2004, and dropping slightly to 26.5 percent o f GDP in 2005, mainly because o f rises in mandatory expenditures. The main subcategories of mandatory expenditures (interest payment on the debt, personnel costs, and transfers) are discussed later inthe 2.28 Duringthe past four years, talk o f the arrival o f petroleum revenues has fanned public expectations, with immediate impact on the public accounts. Such expectations induced a premature overestimation o f public revenues and therefore expenditure; this overestimation was especially visible in 2004. Also, these expectations have fueled social demands for services and salary increases. These demands have often led to social and political instability because different groups have attempted to capture a share o f the perceived petroleum wealth. These attempts in turnhavebeen addressedthroughunprogrammed, spontaneous additionalpublic expenditure with medium- to long-term consequences on budget discipline. For instance: demonstrations by youth in 2003 led to increased scholarships abroad in late 2003 and 2004; in 2005, nominal wages increasedby 43 percent (considering all the staff costs) after a one-week strike by the civil service in2004 ledto the fall ofthe government; andthe like. Table 8: MainExpenditureCategories p d g e t execution) Db billions 2001 2002 2003 2004 2005 2006 (est.) (Prog ) . Expenditure and net lending 335.3 325.6 415.1 548.5 526.7 706.3 Recurrent expenditure 121.0 156.8 163.3 202.5 314.4 387.3 Personnelcosts 40.5 43.6 55.0 75.7 103.2 124.5 Goods and services 18.8 26.7 43.9 95.8 66.2 79.0 Transfers 13.1 40.7 35.9 65.9 83.2 100.3 Intereston debt 24.8 25.9 26.1 30.1 33.7 38.8 Other 24.8 20.0 22.4 25.0 28.1 44.8 Development expenditure 213.4 168.8 231.8 256.0 214.3 321.0 Public investment programme 194.3 140.0 197.9 234.9 186.4 266.0 HIPC-fundedexpenditure 17.2 22.5 33.6 30.1 27.9 55.0 Public service restructuring 1.9 6.3 0.0 0.0 0.0 0.0 Net lending 0.0 0.0 0.3 -9.0 0.0 0.0 Expenditure and net iendlng 49.6 39.0 44.7 51.7 44.5 46.4 Recurrent expenditure 16.0 18.8 10.7 27.6 26.5 25.4 Personnelcosts 6.0 5.2 5.9 7.1 8.7 8.2 Goods and services 2.8 3.2 4.7 9.0 5.6 5.2 Transfers 1.9 4.9 3.9 6.2 7.0 8.6 Intereston debt 3.7 3.1 2.8 2.8 2.8 2.5 Other 3.7 2.4 2.4 2.4 2.4 2.9 Development expenditure 31.6 20.2 24.0 24.1 18.0 21.0 Capitalexpenditure(Public investmentprogram) 28.7 16.8 21.3 22.1 15.7 17.4 HIPC-fundedexpenditure 2.5 2.7 3.6 2.8 2.3 3.6 Public service restructuring 0.3 0.8 0.0 0.0 0.0 0.0 Net lending 0.0 0.0 0.0 -0.8 0.0 0.0 Sources; Ministry o f Planning and Finance data; new GDP numbers from IMF staff, 2006. 35 The personnel costs include those o f the central government (all ministries), the Presidency o f the Republic, the National Assembly, localities and the government o f Principe, embassies, the justice system (courts and the attorney ffneral), and the accounting tribunal. Common expenditures include goods and services, current transfers, and a catchall category "other current expenditures." Goods and services have been dominated by costs o f electricity, water, and telecommunications in the past few years. Current transfers include payments to the Joint Development Agency (JDA) to cover STP's share o f costs for development o f the petroleum sector in2004-05 (included as autonomous services). 28 2.29 The share o f development expenditures intotal public spending shrunkbetween 2000 and 2005 (table 8). While development expenditures captured 31.6 percent o f GDP in2001, this share shrunkto 18percent in2005 andreachedan estimated21percent in2006. Several factors explain this drop: (i)relative stagnation of capital expenditures (as executed in the public investment a program) while the budget envelope and the economy have continued to grow; (ii) a shift in the composition o f the PIP, where current expenditures have captured a rising share to the detriment o f actual development expenditures (figure below); (iii) variable and often-delayed flow of funds from development partners; and (iv) a likelihood that the expectations regarding the inflow o f oil revenues may have already conditioned donors' decisions regarding the volume and scope o f external aid. Also, two public finance management issues affect the implementation of the investment budget: (i) cash budgetingpractices that focus payment on mandatory expenditures in a tight financial environment; and (ii)the country's low absorption and implementation capacities. These two factors are discussed below (and inthe CFAA volume). Figure 11: Composition of Public Expenditure(Budget Execution) as a percentage of total budget 1 (est.) (Prog.) 2000 I 2001 1 2002 1 2003 I 2004 1 2005 1 2006 1 1 oCurrent 0 Development 1 Source: STP Government. 2.30 In this context, HIPC financing has fulfilled an important role in development expenditures because it freed an average o f 2.8 percent o f GDP over 2001-05, and an estimated 3.6 percent o f GDP in 2006, to the budget. These funds have been focused in sectors that affect the poor directly: health, education, and infrastructure. However, a portion o f HIPC funds have been used for current expenditures, such as increases in salaries for school teachers and emergency payments to health workers duringthe cholera outbreaks o f 2005 and 2006. 2.31 The sharp rise in current expenditures and the sharply diminished petroleum prospects clearly signal the necessity to return to a fiscal policy based on non-oil fiscal discipline and long- term sustainability. The authorities will clearly need to exert more discipline on current expenditure to create fiscal space for growth and poverty reductiondespite their success inraising non-oil revenues in the recent past. This expanded fiscal space can then be used in pro-poor development expenditures. 29 Economic Compositionof Expenditures The economiccompositionof expenditureshighlightsthe growthinmandatory expenditures. 2.32 The increase in the current expenditures was driven by mandatory expenditures (personnel costs, transfers, debt service): the share o f mandatory expenditures in total current expenditure rose from 59.0 percent in 2001 to 70.7 percent in 2005, an increase o f 20 percent in five years (table below). Consequently, it can not be easily revised to create the fiscal space to respondto the objectives o f the poverty reduction strategy. Table 9: Weight of Mandatory Expenditureinthe Total Current Expenditure(Execution) 2001 2002 2003 2004 2005 Mandatory expenditure (Db billions) 43.5 56.5 96.9 133.8 227.2 Current expenditure (Db billions) 73.8 94.4 156.5 208.9 321.5 Weight of mandatory expenditure in current expenditure (percent) 59.0 59.9 61.9 64.0 70.7 Source: STP goverment 2.33 The current budget i s elaborated on an incremental basis without much room to target public spending according to policy objectives. The practice o f cash budgeting, which prioritizes mandatory expenditure, reinforces the dominance o f these expenditures. Cash budgeting i s discussed below (also see the CFAA).37 On the other hand, the public projects selected for PIP financing can intheory be chosen according to their adherence to the policy objectives by phasing in prioritized projects and phasing out projects not meeting the required criteria. PIP issues are discussed below (also see the CFAA). 2.34 The personnel costs explain about one-third o f the total current (and mandatory) expenditure over the last three years. In fact, the share o f personnel costs shows a significant increase from 6.0 percent o f GDP in 2001 to 8.7 percent in 2005, a jump o f 45 percent in five years.38This share was supposed to be controlled to an estimated 8.7 percent in the 2006 budget. The tendency for personnel costs to lose weight in overall current expenditure does not necessarily mean a degradation o f the wage position in the distribution o f current expenditure, because the civil service has been receivingback-to-back salary increases over the past few years (see figure below). It i s rather a consequence o f the faster increase o f other expenditure items, notably expenditure on goods and services and interest on debt in 2003, goods and services in 2004, and current transfers in 2005. 2.35 Four ministries (Education,Health, Defense, Finance) and the embassies represented 85 percent o f total personnel costs in 2005 because they employ the largest number o f civil servant^.^' Furthermore, there i s a clear correlation between sectors with high rankings in the 37For instance, a circular dated May 2006, that was sent to the spending ministries by the Ministry of Planningand Finance (MoPF), states the items that should be prioritized inthe budget elaboration (personnel costs and some goods suchas fuel, food, medicines, andcommunications). 38The increasingtrend is the result of back-to-backsalary rises, with a single 43 percent increaseaccorded to the civil service as a whole in 2004-and implementedinmid-2005-that increasedminimumwage from approximatelyUS$30 to approximatelyUS$40a month. 39Comparedto 80 percent in2003. 30 overall scale o f personnel costs and their higher variable r e m ~ n e r a t i o nFor~instance, Finance, . ~ Health, the Courts, and Education depend considerably on the variable component to make up their total remuneration. Figure 12: Compositionof the Total Current Expenditure 100% 80% 60% 0 Other 0 Transfers % E] Interest on debt 61 Goods and service! 40% Personel costs 20% 0% 2001 2002 2003 2004 2005 Source: STP Goverment 2.36 The variable and social contribution components o f personnel costs are not uniformly distributed across the sectors, raising concerns about the fairness in the distribution of these payments and highlightingthe need for further analysis (table 11). Only three sectors (Education, Health, and Planning and Finance) benefit from the discretionary component (variable remuneration); these sectors absorb 97 percent o f the total, with the Ministry of Education constituting almost half o f the total. Similarly, contributions to social security seem to be a privilege o f a few sectors that absorb 84 percent o f the total o f this category: The Ministry of Education and the embassies clearly benefit, and the ministries o fHealth, Defense, and Planning and Finance benefit to a muchlesser extent. Table 10: Decompositionof Staff Costs, 2005 Percentage of total Fixed Variable Social Total remuneration remuneration contributions Mlnlstryof Educatlonand Culture 68.1 29.0 2.9 100.0 Mlnlstryof Health 61.1 36.4 2.5 100.0 courts 67.6 31.3 1.1 100.0 Embassies 92.1 1.3 6.6 100.0 Mlnlstry of Plannlngand Flnance 61.7 37.0 1.4 100.0 Averape 76.8 79.9 3.2 100.0 Sources: Recurrent expenditure (Ministry of Planningand Finance, budget execution), 40Personnel costs can be broken down into three components: a fixed, stable component; a variable, discretionary component; and the social contributions made by the administration as employer. In2005, 77 percent of the total costs correspondedto fixed remuneration, 20 percent to variable remuneration, and 3 percent to social contributions. 31 2.37 The divergence in composition and distribution o f personnel costs raises concerns about the fairness and cost efficiency o f the current personnel remuneration structure. This divergence highlights the need for more in-depth analysis o f the civil service structure and compensation package to answer the following questions: (i) what are the criteriajustifying such differences in salary compositions; and (ii)s the variable component related to simple overtime or work i incentives (productivity bonuses or performance bonuses). The authorities can use such an analysis to improve civil service compensation packages while better understanding and controlling the impact on the budget. Specifically, controling the discretionary component o f salaries will provide more transparent and predictable personnel cost estimates while simplifying management o f salary calculations. Table 11:Distribution of Staff Cost Components by Highest Five Sectors, 2005 Percentageof total Fixed Variable Social Total remuneration remuneration contributions Mlnistry of Education and Culture 28.9 47.5 29.2 32.6 Ministry of Health 12.6 29.0 12.2 15.9 Ministry of Defense and Internal Administration 18.3 0.4 14.8 14.6 Embassies 13.8 0.8 23.5 11.5 Mlnlstry of Planning and Finance 8.4 19.4 4.4 10.5 Subtotal 82.0 97.0 84.2 85.1 Sources: Recurrent expenditure (Ministry o f Planning and Finance, budget execution,. 2.38 The authorities should use the ongoing public financial management (PFM) reforms, the 2005 general census o f the civil servants, and further studies to develop a modem and decentralized system for managing human resources inthe civil service and to create a system for controlling personnel costs.41First, the new registry o f civil servants will help the authorities to obtain an accurate total o f sectoral and overall numbers o f civil servants, removing any ghost civil servants from the payroll. This new registry will most likely reduce the overall personnel costs. Second, the new registry will help them with the plannedanalysis and consequent restructuring of the civil servant qualification and salary structures. 2.39 Current transfers were the second-highest category o f spending in 2005, with 25.9 percent o f current expenditure and 7 percent o f GDP. This spending category has displayed a typical trend up to 2005 and i s basically composed o f transfers to the National Assembly and to individuals. This last item includes the transfers made by the Ministryo f Labor and Solidarity and transfers to individuals seeking health treatment in Portugal, which i s a considerable cost to the budget. In2005, the structure o f current transfers changed significantly for two main reasons: (i) the reclassification o f the scholarships paid` to students abroad from the public investment program (PIP) to current transfers o f the Ministry of Education and (ii) STP's payment o f its share for activities o f the Joint Development Agency (JDA) in commercializing the Joint Development PetroleumZone (held with Nigeria).42 41As reported inGrandes opp5esdoplano-2006, Ministry of Planning and Finance, January 2006. 42Scholarships are now registered as an appropriation o f the Ministry of Education under the heading "other transfers abroad," and payments to the Joint Development Zone are now registered as an appropriation o f "autonomous services." 32 Execution Table 12: Current Transfers fromthe State BudgetDbbillions 2001 2002 2003 2004 2005 Publlc admlnlstration 3.5 8.2 7.8 8.1 46.2 State budget 2.8 4.2 0.4 0.2 0.3 Autonomous servlces 0.0 3.2 4.7 4.6 37.9 Local admlnlstration 0.7 0.0 3.3 3.4 8.0 Region of Prhclpe 0.4 1.0 2.4 2.5 7.0 Dlstrlct counclls 0.3 0.7 1.0 0.9 1.0 Soclalsecurlly 0.0 0.4 0.0 0.0 0.0 Private Instltltutions 0.1 0.0 0.4 0.2 0.2 Individuals 1.7 0.1 3.4 4.3 4.2 Contrlbutlonsto Internationalorganlzatlons 0.4 2.8 0.7 0.8 1.2 Other transfersabroad 1.1 0.5 0.9 0.8 31.4 Total 6.8 12.0 13.3 13.8 83.3 MemoradumItems: Totalbudget 266.1 408.8 471.6 521.9 Recumnt budget 94.4 156.5 208.9 321.5 Source: Ministryof Planningand Finance. 2.40 The reclassification o f scholarships from the PIP to current expenditures (Db 31 billion/US$2.9 million in2005) highlightsits large relative weight inthe total current expenditure and the clear pressure that it places on the domestic public resources. Representing 42 percent o f the total recurrent budget o f the Ministry of Education and 9 percent o f the total public expenditure, it i s worthwhile pondering the cost-benefit balance o f the present policy in light of the priorities set out in the poverty reduction strategy. Both the NF'RS and the education strategy focus on improving primary education as their main objective inthe sector. 2.41 The third-highestcategory o f expenditure, goods and services, constituted 22.8 percent o f the total expenditure and 5.6 percent o f GDP in 2005. Consumption in this category i s highly concentrated: five spendingministries or authorities (Health,Defense, Education, Embassies, and Presidency of the Republic) make up 80 percent o f the total consumption o f goods and services. Furthermore, the internal structure o f this category changed in the recent past: the oil price increases in 2004 and 2005 led to a crowding out o f goods and services in "common expenditures," under which a substantial part of the goods and services expenditure i s classified. With world petroleum prices receding since August 2006, it is expected that this crowding out will ease. Ifthe present policy regarding scholarships remains and civil service salary increases are not restrained, they may lead to sustained crowding out o f goods and services, causing a deterioration o f the provision o f public goods, namely in the priority sectors targeted by the poverty reduction strategy, like education and health. 2.42 The interest payment on the debt has decreased slightly but steadily from 3.7 percent o f GDP in 2001 to an estimated 2.5 percent o f GDP in 2006 as the country benefited from steady economic growth and the interim debt relief granted by the HIPC program.43The government o f Silo Tomi and Principe has used the interim debt relief to fund investments in NF'RS priority sectors (health, education, infrastructure, and the like). In March 2007 STP reached the HIPC completion point, which has led to a substantial and irreversible reduction of the debt to 43 While debt and interest payment are not categorized as part of the mandatory expenditures in the long run, in the medium term (that is, three to five years) they constitute an important claim on the budget. In the long run, renegotiation of debt, debt forgiveness, and other policy measures may change the importance of debt or interest paymentintotal current expenditures. 33 multilateral and most bilateral donors (Paris Club and Portugal). Debt forgiveness will slightly increase the fiscal space available to the authorities for investment inpoverty reduction. 2.43 Key measures to introduce more flexibility in the government's expenditure policies include (i) improving the overall fiscal situation by increasing domestic revenues, to increase the volume of available resources, and exerting tighter control over expenditures; (ii) analyzing and updating civil service rosters and remuneration structure to improve management and control o f personnel costs; (iii) ensuring that after the HIPC completion point, all HIPC debt relief will not be absorbed by mandatory expenditures; and (iv) undertaking any new borrowing in highly concessional terms (grants or grant elements over 60 percent). AdministrativeClassification The administrative classification of current expenditures highlights the concentration of expenditures in a few spending units. 2.44 Current spending (almost 80 percent) i s concentrated in only six administrative budget lines: common expenditure, the Ministries of Education, Health, and Defense, interest on debt, and embassies (see table below). Table 13: MostImportantSpendingSectors(Current Expenditure) Execution Percentageof total 2001 2002 2003 2004 2005 2006 (progr.) Commonexpenditure 13.2 25.1 20.1 25.0 24.4 23.0 Ministryof Educationand Culture 18.1 17.5 14.8 14.4 23.7 20.6 Intereston publicdebt 12.4 6.1 13.1 10.0 10.3 11.5 Ministryof Health 12.8 9.1 11.4 11.4 8.9 9.5 Ministryof Defenseand Internal Administration 5.5 9.0 6.8 7.4 7.0 7.4 Embassies 10.9 9.3 7.3 6.0 5.4 5.3 Total 72.0 16.1 13.5 74.1 19.1 17.2 Total(Db billions) 71.6 91.7 156.5 208.9 321.5 378.2 Source: STP government 2.45 The Ministry of Education exhibited the second highest share inthe recurrent budget with 23.7 percent in 2005because o f the reclassification o f scholarships in this ministry's recurrent budget. The scholarships correspond to 9 percent o f total expenditure. Without them, the Ministry of Education's share is reduced to 14.7 percent, showing a rising stable expenditure pattern over the past three years. 2.46 The Ministry of Health represented 8.9 percent o f the actual current expenditure in2005, lower than the 11.4 percent reached inthe two preceding years. The spending composition o f this ministry includes primary care, hospital care, and cost o f health care abroad for patients with special medical needs (4 percent of total current expenditures). This topic i s discussed in more detail inthe last section o fthis chapter. 2.47 The overall budget discipline would benefit greatly from a much more vigilant approach to the programming and execution o f the common expenditure budget category. This category has been used as a catchall, overrunning its allocated budget every year and growing in size in the overall current and total budget. Common expenditure has represented 20 to 25 percent o f the 34 total current expenditure every year since 2002 (table 13). Furthermore, the procurement methods usedin spending funds from this category needto be modernizedto ensure transparency. 2.48 The composition o f items procured and classified incommon expenditures i s quite varied. The most important item up to 2005 has been the acquisition o f goods and services, in particular electricity, water, and communications. The Ministry o f Planning and Finance (MoPF) centrally contracts and pays for these services, which are consumed across the entire public administration. Until2004 these items representedthe bulk of common expenditures (88 percent of the However, in2005 current transfers emerged as a major expenditure (26 percent o f the total) with the inclusion o f transfers to the JDA.45The budget for 2006 will likely include an even higher share of current transfers (33 percent), again related to a second payment to the JDA, thereby remaining the largest expenditure. 2.49 Generally, the central acquisition o f certain goods and services, though justifiable on the grounds o f a potential higher economy and efficiency, implies a good knowledge o f the real consumption needs across `the spending units.Moreover, it requires an assessment mechanism to ensure a more accurate analysis o f efficiency inthe provision o f goods and services at the level o f the budgetary units.46 Inthe case of STP, this mechanism is not inplace, which could allow the spending units to misreport the real cost o f the activities carried out. Table 14: CommonExpenditure-Structure of Spending Budgetexecution Percentageof total 2001 2002 2003 2004 2005 2006 Personnel costs 25.7 6.8 1.6 0.7 0.5 1.7 Goods and services 27.2 67.0 71.2 88.0 35.8 16.6 Current transfers 15.3 4.0 3.2 1.I 26.5 34.9 Of which: autonomous services 0.0 0.0 0.0 0.0 26.0 33.2 Other currentexpenditure 31.9 22.2 24.1 10.1 11.2 13.5 Total 100.0 100.0 100.0 100.0 100.0 100.0 Total(Db billions) 9.4 23.7 37.4 52.2 703.9 776.6 Source: Ministry of Planning and Finance. Coherence between the NPRS and Budget Execution 2.50 A major policy and coherence issue is that NPRS i s disconnected from national policy instruments such as the Grandes Opq6es do Plano (GOP), the policy document that the government attaches to the annual budget4' The GOP's "options" or goals are not clearly derived 44In2003-05 this large share was partly explainedby spendingon water,energy, andcommunications-costs for these services doubled between 2003 and 2004 and increasedfurther in 2005. A large share of the electricity production in Si50 Tom6 and Principe has a thermal source. Import o f gasoline constitutes the largest expenditure for the water and electricity company (Empresa de Agua e Electricidade, EMAE), which in turn has increased its electricity tariffs severaltimes, especially for the centralgovernment, to try to remain economicallyviable. 45Classifiedunderthe headingautonomousservices. 46 Defined as the relation between the services provided by each spending unit of each ministry (measured by appropriateoutput or outcome indicators) and the costs incurredfor the provision of that service. Apportioning all the spendingcosts byministry would enablefacilitation of accountabilityfrom spendingunits. 47The mainpolicy documentsthat presently inform, or should inform, the budgetpreparationare the NationalPoverty ReductionStrategy (NPRS), the Grandes Opg6es do Plano (GOP), and the Programado Governo. The GOP maps out 35 from the NPRS's goals, and consequently this document fails to assume the crucial role o f intermediating between the poverty reduction strategy and the state budget.48 Moreover, the "policy measures" set out in the GOP are not clearly reflected in the PIP, which should be the main tool to implement the measures with cost implications. 2.51 Similarly, at the sector level, the NPRS does not seem to have been adopted in practice by the ministries as the main policy reference in the preparation of their investment proposals. The main reasons are (i) the lack o f guidance from the Directorate o f Planning and the NPRS Unit (or Poverty Observatory) for the sectors to prepare their investment proposals inline with the poverty reduction strategy; and (ii) lack o f appropriate tools to prepare and select the the investment proposals in line with the NPRS objectives. The current information systems are not Box 1: The NPRS The NPRS was adopted inDecember 2002, and the government started partially implementing it in the 2003 budget. However, the political instability o f the successive years and the consequent difficulty in mobilizing the required external fmancial resources hindered its implementation. The initial document was updated in2004 and endorsed by the boards o f the IMFand the World Bank in 2005. The 2003-10 budget was estimated at US$210 million. A Priority Actions Program (PAP) was developed based on the NPRS for the Donors' Roundtable held inDecember 2005" and has a financial requirement o f US$168.8 thousand (for the period 2006-08); US$70 million was pledged, o f which little was new money. The NPRS lays out an ambitious program of development based on five pillars: (i) reform o f public institutions, capacity building, and good governance; (ii) accelerated and redistributive growth; (iii) increased employment opportunities for the poor and diversification of the economy; (iv) universal access to social services; and (v) adoption o f mechanisms for monitoring and evaluation with feedback mechanisms to update the strategy. The NPRS also considers cross-cutting issues like governance, gender, and the environment. Each pillar i s broken down in a number o f "Programmes," which are supposed to be consistent with the sectoralpolicy objectives. The NPRS i s based on "development" expenditures, disregarding the sustainability needed in programming recurrent expenditures to ensure key public services in the implementation o f that strategy, inparticular inthe sectors o f education and health. Two reasons could explain this practice. First, the poverty reduction strategy is expected to be basically financed with external resources; NPRS i s a convenient vehicle to mobilize such resources. Second, the rigidity o f the recurrent budget, because of the high weight of mandatory expenditures, makes it a poor instrument to shape public expenditure in line with the poverty reduction objectives. Again, this leaves to the Public Investment Program (PIP) the central role, at least in theory, o f a convenient vehicle to mobilize external resournces and to orient the expenditure inaccordance with the NPRS. the policy options, segmented in"options" or pillars, that the government intends to implementin linewith the budget for the next year. Hence the GOP, besides the fact that the STP constitutionrequires it to be submitted yearly to the National Assembly alongwith the budget, is a very usefuldocument for the public and the decisionmakers.The GOP serves as a bridgebetweenthe overall policy targets, specificallythe targets set out by the NPRS, and the budget. The "policy options" laiddown inGOP are supportedby detailedstated "policy measures," which are a mix of actions with financialimplications,decisions, andtasks to be carriedout by the administration. 48 For instance, the GOP for 2006 states as main objectives "improve the degradation of the economic and social situation", "improve the efficiency of investments in the social sector", and "establish the bases for a sustainable growth". However,there is not aclear matchingbetweenthose objectives andtheNPRS's pillars. 49 The information systems lack consistency (development and recurrent expenditure use different classifications)and integration(PIP is linked neitherto the recurrent expenditurenor to the NPRS). The old database used to manage the PIP does not respond to the information needs raised by the NPRS. The existing functional classification could in theory provide an alternative, but it is too aggregated to allow a direct matchingwith the NPRS objectives. Therefore, the "project fiche" filled upbythe sectorsdoesnot containtherequiredinformationfor that purpose. Roundtablewith the InternationalDevelopmentPartnersof STP, Brussels, December 6,2005. 36 2.52 Furthermore, STP's budget was not classified usingthe standard functional nomenclature (Classification o f Functions o f Government, COFOG) for the period 2001-06 (estimated). The functional classification provides an indication o f the relation between the public expenditure pattern and the government's policy goals. This shortcoming substantially limits the scope o f our analysis. Nonetheless, we developed three simple alternative measures to analyze the connection between the NPRS and budget execution. 2.53 The three alternative methods are to (i)aggregate the different sectors/ministries by general function similar to functional nomenclature; (ii) aggregate the data by NPRS priority sectors/ministries; and (iii) use the ad hoc "pro-poor" definition used by the government, encompassing the Ministries o f Education, Health, and Labor and Solidarity.51,52,53 Care should be given in interpretingall three measures since the clear beneficiaries o f these expenditures can not identified. 2.54 The main findings raise concerns about the quality o f expenditures. The coherence between policy goals (NPRS) and budget allocations needs to be strengthened. Aside from the dichotomy between programmed and executed budget discussed earlier, we note two other dichotomies inperformance: (i) inthe level of current versus development expenditures; and (ii) in the differentiated performance between health/education sectors compared with agriculture/infrastcture sectors. There are also sector-specific issues related to policy and budget coherence as well as programmed versus executed budget. These are discussed in the context o f the review o fpublic expenditure inhealth (inthe last section o f this chapter). 2.55 Usingthe first measure (aggregating current expenditures by general function), we note that the share o f the social expenditures (mainly education, health, and labor) inthe total current expenditure has been a steady 30 percent since 2001; that is, before the adoption o f the NPRS, showing the long-term government's commitment to social sectors. The 2005 increase to 36.6 percent i s explained by the reclassification o f the scholarships in the current budget (in the education sector) and the wage increase in the public administration, which has a higher impact on the social sectors (figure below). 2.56 Economic sectors compose a weak but stable share o f the expenditures, while the general administration andpolitical institutions capture a stable 15 percent o fthe total expenditure. Other sectors (defense and internal afairs, embassies, common expenditure, and interest on debt) still form the largest share o f the functional expenditure with 42.9 percent, despite the decreasing trend over the past six years. 5' These general function categories are (i) administration and political institutions (National Assembly, local general administration, Presidency of the Republic, Prime Minister's Office, Principe, Foreign Affairs [excluding embassies], and Justice); (ii)social sectors (media, labor and vocational training, education and culture, youth and sports, health, natural resources and environment, infrastructure, and land); (iii) economic sectors (finance excluding debt, economy, agriculture and fishing, trade, industry, and tourism); and (iv) other sectors (defense, public debt, common expenditure, embassies, other). 52 The NPRS priority sectors are Ministries o f Education and Culture, Health, Agriculture, Justice, and Infrastructure. 53 As there was no standard functional classification, the authorities developed this ad hoc "pro-poor" definition in2004 to address the concern that the macroeconomic program supported by the IMF (Staff-Monitored Program ) should protect pro-poor expenditures while controlling current expenditures. The pro-poor expenditure (recurrent and capital) i s derived from the economic classification in some selected ministries (Education, Health, and Labor and Solidarity). This category was developed in coordination with the IMF, and with input from the World Bank. The list is included in the IMF report StaffReport for the 2005 Article I V Consultation (2005). 37 -r Figure13: CurrentExpenditureby GeneralFunction 50'0M 60'o1 54.0 n 52.0 ,7., " 7 4 / . L 48.4 42.9 40 0 - 3 i F r - t 30.1 '/E 30.0 20.0 10.0 0 0 2000 2001 2002 2003 2004 2005 Source: STP government 2.57 Using the second measure (aggregating the data by NPRS priority sectors/ministries) shows that, overall, the weight o f the NPRS-related sectors captured a sizable share o f total public expenditure (figure 14 and tables 16 and 17) but seems to decrease between 2002 (57.6 percent) and 2005 (44.8 percent). The budget for 2006 plans an increase. Such an evolution i s disappointing given the fact that the year 2002 sets the initial momentum for the strategy for poverty reduction, following the government's approval o f the Poverty Reduction Strategy Paper. The administration-related ministries reinforced their weight in the total public expenditure until 2004, followed by a decline in2005, which seems to continue in2006. 2.58 Tables 16 and 17 highlight that current expenditures for the NPRS priority sectors have increased over the 2002-05 period, with continued increase predicted for 2006. The performance o f the development budget has been more volatile, with a drop in 2002-05 because o f a number o f political, institutional, and practical difficulties related to securing and implementing the PIP. The political constraints are related to the July 2003 coup and multiple government changes that have contributed to delays inreceiving donor funds and implementing projects. The institutional and practical difficulties inimplementingthe PIP are discussed below. 2.59 Tables 16 and 17 also show that there are divergent sector-specific performances (see also figure 14 below). Education andjustice show an overall upward trend, while health i s stable over the period. Agriculture has shrunk dramatically, while infrastructure shows large annual variations. 38 Figure 14: Weight of the NPRS-Related Sectors inTotal PublicExpenditure 25.0 1 20.0 BJustlce 15.0 Education ae 0 Health 10.0 Infrastructure Agriculture 5.0 0.0 Justice Education 0 Health 0 Infrastructuri Source: STP government. 2.60 The education sector reinforces its share in the total expenditure, from 3.5 percent o f GDP in 2002 to 7.1 percent o f GDP in 2005, even though there was a decline to 4.4 percent of GDP in 2004. The budget for 2006 confirms the upward trend o f the education sector, with 17.5 percent of the total expenditure. The decline in 2004 can be attributed to the deep downward trend of investment in the sector in 2003. However, the negative path o f investment has been offset by the current expenditure in the education sector, which shows a very strong and sustainable increase over the period 2001-05. Infact, the steep increase inthe current expenditure on education in 2005 was mainly due to the reclassification o f the scholarships, which used to be registered inthe PIP. 2.61 The health sector, after gaining some weight in 2003 in comparison with 2002, has remained stable ever since (around an average o f 10 percent over 2003-05). The lack o f further growth i s partially associated with the execution capacity at the sectoral level, curtailing development activities. On the other hand, current expenditures, especially salaries, have increased (as discussed earlier in the chapter). See the analysis of expenditures in the health sector, below, 2.62 Both the Ministry o f Agriculture and the Ministry o f Infrastructure show a steady fall in their participationintotal public expenditure. Expenditure inthe former fell from 14.5 percent of total public expenditure in2002 (more than the education or the health sectors) to 3.3 percent o f the total expenditure in 2005, while the latter fell from 23.7 percent to 10.3 percent (figure above). This decline in agriculture and infrastructure occurred even though both are considered to be priority sectors in the NPRS and essential in fostering wide-based economic growth. This decrease i s a matter o f concern given the relevance o f these sectors in poverty alleviation. A majority o f the poor and extremely poor still live inthe rural areas, and infrastructure constitutes a lifeline for many isolatedvillages and families. 2.63 Several factors may explain this steady fall. One explanation i s that the donor-executed projects in these two sectors are not properly recorded in the central budget-execution system because o f weak management and follow-up o f the PIP and weak coordination of donors. 39 Another related factor may be a low rate o f project execution because o f delayed or cancelled donor funds, weak national implementation capacity, and late approval o f the annual budget. However, in the case o f agriculture, it i s notjust the low execution rate o f the public expenditure (both recurrent and development) that explains the disappointingoutlays inthis sector. Infact, the funds allocated to the Ministry of Agriculture show a consistent tendency to diminish over the past few years, thereby reflecting a policy choice that contradicts the priority role given to this sector inthe NPRS. 2.64 Finally, using the third measure, "pro-poor expenditures," based on the economic classification o f public expenditure, confirms the dichotomy between current and development expenditures discussed above. The weight o f pro-poor current expenditures in total current expenditures has increased between 2002 and 2005. The weight o f pro-poor development expenditures in total development expenditures has decreased in 2004 and 2005. For these two years, the decrease in development expenditures has affected the total pro-poor expenditures in total central budget expenditure^.^^ Table 15: Pro-poor Expenditures 2002 2003 2004 2005 Pro-poor expenditure 105.7 153.6 140.0 152.1 Recurmnt expenditure 23.1 39.0 51.4 69.0 Ministryof Educationand Culture 15.2 21.7 28.3 41.9 Ministryof Health 7.9 17.3 23.1 27.2 Ministryof Labor and Solidarity 0.7 1.3 0.9 1.2 Developmentexpenditure 62.6 114.6 88.6 63.1 Weightof total pro-poor expenditurein totalexpenditure(%) 32.5 37.0 25.1 28.8 Weightof recurrentpro-pwr expenditurein fofalrecurrentexpenditure(%) 14.2 21.3 17.6 22.0 Weightof developmentpro-poor expenditurein fotaldevelopmentexpenditure(%) 50.8 49.5 33.4 38.8 Sources: Recurrentexpenditure(Ministryof Planningand Finance,budgetexecution,Excelfiles). Devdopmnt expenditure(Directorateof Planning,Despesasdireccbnadaspara a redugdoda pobreza) Notes: in 2003 and 2004, the valuesof scholarshipshavebeen deductedfromthe developmentexpenditureto ensurecomparabilityacrossthe years. Source: STP government 54 Care should be taken when interpreting these data because they are based on strong assumptions on who are the actual beneficiaries o f the budget expenditure that have been considered as pro-poor spending. The items selected were recommended in the IMF's staff report for 2005, Article IV Consultation, and first revision under the PRGF. The pro- poor development expenditure includes projects that are deemed to have an impact on alleviating poverty in some selected sectors (education, health, labor and solidarity, agriculture and fisheries, rural development, planning and finance, youth and sport, provision o f potable water, and electrification). 40 3 d "o"m" "m' 4bc p P -0 w The CentralRole of the PIP Improving the PIP'S selectivity and execution is crucial given its central role in ensuring coherence between budget implementationand NPRS goals. 2.65 As noted before, development expenditures have declined over the past few years, to the benefit o f current expenditures. Within the PIP, capital expenditure as a share o f PIP has dropped drastically from 80 percent in 2000 to 45 percent in 2005. Highpressure on the current expenditure, particularly in a context o f fiscal restraint under the Staff-Monitored Program and PRGF programs, has led to a crowding out o f capital spending (table 18). This trend i s explained by a steady decrease in the construction subsector and to a lesser degree by a drop in the equipment component, falling from 13 percent o f total PIP in 2002 to 6.6 percent in 2005. Inparallel with the decreasing capital expenditure, the administration category rose from 5.7 percent o f the total in2000 to 35.4 percent in 2005. Administration includes salaries, consumable goods, fuel, water and energy, and transport and communication. The rising share o f this category signals either an increasingly inefficient project management process or the use o f project resources to support the day to day running of the administration which should otherwise be financed through the recurrent budget. Table 18: PIP-Main SpendingItems Execution Percentageof total Expenditure items 2000 2001 2002 2003 2004 2005 Studies,technical assistance, and audit 11.8 10.8 10.6 21.5 11.4 15.5 Training 1.7 2.5 4.2 0.0 6.9 3.8 Administration 5.7 10.5 14.0 17.1 27.4 35.4 Other (credit, goods and services, and contingency) 1.3 1.6 0.0 0.8 0.5 0.0 Capital expenditure 79.5 74.6 71.2 60.6 53.8 45.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 Source: STP government 2.66 The PIP i s highly dependent on external funding. On average, 70 percent o f the total financing o f PIP over the past four years has come from external sources (78 percent o f the total PIP budget in 2005, with a minimumo f 63 percent in 2003).55Bilateral financing contributes an average o f 64 percent o f this external aid. 2.67 An important organizational constraint o f PIP'S management is the absence o f planning departments in the spending ministries. This leaves the MoPF's Directorate o f Planning without a regular interlocutor to discuss the budget implications o f the sector policies, the elaboration o f the public investment program, and its implementati~n.~~Moreover, the fact that there are no sectoral planning directorates does not ensure a stable and continuous participation inthe PIP process over the 1.1 55 As of late 2006, the public investment was still classified on the basis of (i) administrative classification an (similar to the recurrent budget); (ii)a line-item classification (by categories of expenditure, similar to the economic classificationbut not consistent with the one usedby recurrent spending); and (iii)a programclassification, which i s in fact a simple grouping of projects according to homogenoususes. 1.2 '`In practice, such a role is often filled by the sectoral ministers' direct advisers, but they have multiple responsibilities, diverting them froma regular monitoring ofthe sector investment program. 43 successive budget cycles. The Directorates o f Administration and Finance could assume this role, as some o f them already do, even if their traditional vocation has been more oriented to the financial management o f the recurrent budget. The ongoing PFMreforms will address this issue inthe context of the budget unification, which will require a unit, at the sectoral level, assuming the planning, programming, budgeting, and monitoring functions o f an integratedbudget. 2.68 As notedinthe CFAA (volume 4 o fthe Country IntegratedFiduciary Assessment, CIFA) the investment expenditure process needs to be streamlined because currently it i s cumbersome and involves many actors. These actors include the Patrimony Directorate of the Treasury Directorate, the HIPC Commission, the Planning Directorate, the bid opening committee (in the case o f a public contract, the Accounting Tribunal, the Public Works Executing Agency, and sectoral ministries - based on the technical nature o f the expenditure. The CFAA recommends streamlining the investment expenditure process back to the normal pr~cedures.~~ 2.69 At the operational level, selection criteria and monitoring and evaluation systems needto be improved. The existing selection criteria underpinninga project proposal seem well chosen to meet the purpose o f prioritizing the projects that have a stronger impact in the reduction o f poverty, but they can be improved. According to the present guidelines o f the MoPF, these criteria are the following: HIPC-financed projects; ongoing projects; consistency with the policy objectives o f the respective sector, which should be in line with the poverty reduction strategy; assurance o f financing, particularly from external sources; and quality o f the project proposal submitted to MoPF by line rnini~tries.~~ 2.70 The selection criteria can be improved by ensuring that a project in the PIP adheres to the NPRS goals, verifiable through matching the NPRS objectives with the project's. Presently, such an exercise can not easily be made because project identification and formulation do not reflect a clear hierarchy of objectives based on a sound and systematic m e t h o d ~ l o g y . ~ ~ 2.71 Ensuringthat proposed projects are sustainable can improve the selection criteria. Presently, the proponent sectors do not provide information regarding the fkture recurrent costs generated by the projects6' In some cases, it i s not even spelled out which entity will ensure the management o f the future services that will be made available by the project. 2.72 The practical and systematic application o f these selection criteria can be improved by easing the technical constraints and resisting political priorities. The existing multicriteria method can be improved by setting up a clear scoring o f projects according to their merit in relation to the desired policy objectives. The formulation o f projects should follow a systematic and standard methodology, ''Describedonpage 35 o f the CFAA. 58 Such information has been obtained on the basis of oral statements made by the PIP's responsible staff. No written documenthas beenprovidedon this subject. 59An example of the presentdifficulties is given when one attempts to compare the list of actions includedinthe NPRS document and the list of PIP's interventions. In general they do not match, either because of an identification problem (the same interventionscan assume different names in the NPRS and PIP documents) or because they are neglected in the PIP pipeline. Frequentlytheir absence is explained by the lack of guaranteed financing. 6oThe lack of asystematicprogrammingofrecurrent expendituresmight explainwhy the governmentfelt obliged to resort to usingadhoc and suboptimal procedures. Inthis regard, the Ministry of Health(but only this ministry) was allowed to include in this budget an item for recurrent expenditure under the heading despesas diversas (diverse expenditures) This particular solution introduces an obvious distortion in the classificationof the public expenditure, because it does not distinguishthe expenditure accordingto its nature, and therefore it duplicates the same categoriesofexpenditure alreadyincludedinthe right postsofthe economic classification. 44 such as the "logical framework approach."61 Improving the criteria for selection and implementation will help reduce the number o fprojects that are politically drivenrather than chosen on merits. 2.73 A monitoring and evaluation (M&E) system is needed. As o f mid-2006, no functional institutional setup exists that combines the efforts o f the Directorate o f Planning, the Observatory o f Poverty Reduction, and the sector units responsible for the PIP. A short execution report on PIP implementation i s made each year and included in the budget law o f the following year to be submitted to the National Assembly.62 However, such reporting i s limited to the presentation o f the financial tables without a meaningful analysis on the project performance (outputs and outcomes achieved) or on the factors behind the frequent implementation problems. The set o f indicators laid down inthe NPRS document have not yet been adopted by the PIP management. Consequently, it i s not possible to monitor and assess the interventions derived from that strategy and materialized through the PIP. 2.74 A more comprehensive information system to support PIP'S management, including M&E, also should be adopted in line with the ongoing budget unification (as part o f the PFMreforms). The Directorate of Planning intends to upgrade the present limited information system that i s used to collect and process basic financial information for preparation o f the PIP. 2.75 Finally, weak execution o f the development budget requires the authorities and the development partners to take a number o f actions, starting with the following: Moving ahead with the public finance reform and especially reform o f the national procurement law and processes will enhance donor confidence inthe national procedures and facilitate flow of aid. This reform should include streamlining the investment expenditure process back to normal expenditure procedures. Donors' harmonization o f aid and aid practices i s essential in assisting STP to properly implement its budget. Government coordination o f assistance around its NPRS and budget goals i s as critical as donor harmonizationefforts. 0 Over the medium term, the authorities should consider adopting a medium-term expenditure framework to assist them inlonger-termplanning o f expenditures. Notable Deviations 2.76 The notable deviations between the programmed and executed budget highlights the weak allocative efficiency o fthe budget process. 2.77 Overall, data show a systematic overrun o f the current budget with a peak o f 164 percent in 2004 (figure below), decreasing to a more reasonable 103 percent in 2005. By comparison, Cape Verde has managed to keep the execution o f its current budget below 100 percent since 2001 (82.7 percent in 2004 and 94.5 percent in 2005). On the other hand, the development budget i s underperforming systematically, with a higho f 69 percent in2003, dropping back down to 51percent 61 The adoption of such methodology for project formulation was adopted some years ago but recently abandoned. Only the execution report of the public investment for 2003 has been made available to the mission's staff. Hence, we are not sure whether other reports havebeenproduced. 45 in2005, similar to Cape Verde's overall execution performance (59 percent in2003 and 64.6 percent 2005).63 Such divergent performance patterns signal important weaknesses in the budget programming and execution. The factors affecting the current and development budgets have been discussed inprevious sections. Figure 15: Budget Execution Rates 180.0 163.8 160.0 140.0 120.0 :: 100.0 IRecurrent 80.0 ICevelopmenl 60.0 40.0 20.0 0.0 2000 2001 2002 2003 2004 2005 Source: STP government. 2.78 The overruns o f the recurrent budget highlight that hard budget constraints were not respected over the past few years (figure above).64 Three sets o f factors may explain this problem: (i) the role o f the political institutions in the recurrent budget overrun; (ii)social demands for expenditures in anticipation o f petroleum revenues; and (iii)telecoms and electricity costs and payments to the JDA (classified incommon expenditures). 2.79 Generally, the data suggest that spending "entities" or ministries that are part o f the general administration and political institutions group tend to come close or spend beyond their allocated budgets (figure below). This group includes the Presidency o f the Republic and Office o f the Prime Minister, the Court o f Accounts, and the Ministry o f Planning and Finance. Entities like the local administration, General Attorney, embassies, National Assembly, and Ministry o f Defense exhibit execution rates lower but close to 100percent. For some o f these spending entities, the low execution rate o f the current expenditure i s offset by overspending o f the development budget (National Assembly, Judiciary, and Ministryo f Foreign Affairs). 2.80 The ministries related with the economic functions o f the state tend to show very low execution rates both in the recurrent and the development budget (Ministries o f Natural Resources and Environment, Trade, Industry and Tourism, Infrastructure, and Agriculture and Rural Development). In agriculture and infrastructure, this low rate o f execution in current and development budgets may be because some o f the donor-executed projects are not properly recorded 63CapeVerde datafrom2006 PER, page 35, table 2. 64The non-respect o f the programmed budget is even more apparent as these overruns where not systematically and legally introducedvia amendments of the nationalbudgets and presentedto the National Assembly for new approval until 2005. In2005, the governmentchanged its previouspractice andsubmitted an amended("rectificativo") budgetto theNationalAssembly whenit becameapparent that the approvedexpenditureinsome categorieswill bechanged. 46 and followed up and because no programming is included inthe central budget to capture the need for "recurrent" (or maintenance) costs attached to these development projects. Figure 16: ExecutionRates by Spending Unit (Recurrent and DevelopmentExpenditure) averageof2002-05 Source: STP government. 2.81 The priority sectors inthe poverty reduction strategy show divergent behaviors: for instance, the Ministry o f Education shows a satisfactory execution rate for recurrent expenditure (103 percent) but a very weak execution rate for development expenditure (51 percent); the Ministry o f Health shows an acceptable execution rate for recurrent expenditure (87 percent) but a weaker one in development expenditure (72 percent); the Ministry of Agriculture and Rural Development and MinistryofInfrastructure exhibit low executionrates inboththe recurrent anddevelopment budgets. 2.82 As is the case inmany countries, the Ministry of Education i s particularly affected by salary increases given the highnumber of civil servants it employs. Between 2002 and 2005, salaries in the public administration increased at a nominal annual growth rate o f 38.6 percent, much above the nominal growth rate o f the GDP (15 percent duringthe same period). In fact, this ministry accounts for an average o f 30 percent o f the total personnel costs inthe public administration. 2.83 An analysis o f the deviation and the composition variance o f the executed versus programmed budget (table 19) clearly shows that the total budget suffers fkom large deviation and variance: total expenditure deviation ranges around 30 percent, while total expenditure variance i s above 70 percent.65 The first indicator confirms that the development budget i s the main culprit for the largest deviations, with the situation deteriorating over the past three years. Inthe meantime, the aggregate deviations o f the recurrent budget seem to have improved over the past years (fkom 33.1 percent to -4.8 percent in 2005). However, the high level o f the composition variance signals that budget programming i s clearly out of sync with execution and that shifting funds between budget 65The total expenditure deviation measures the ratio o f the sum across all spending units o f the differences between the budget and actual spending, to the total budgeted expenditure. The fact that the differences might cancel each other out justifies a similar indicator, the composition variance, where the differences between the budgeted and the actual spending are taken in absolute terms. 47 lines i s a normal practice. Figure 16 provides a perspective over three years on which spending or ministries contribute to the discrepancy between programmed and executed current budget. Except for the Ministry o f Education, all NPRS priority sectors are seriously affected by this weak budget execution. 2.84 The serious execution problems o f the development budget are further highlightedby the fact that its execution i s systematically lower than foreseen. The highdeviation and composition variance observed in the development budget confirm the severe constraints faced by the public investment management at the programming and execution phases. The total expenditure deviation o f the investment expenditure has increased over the past three years, signaling either a potential deterioration o f the PIP implementation or collecting and recording o f information on project execution. Table 19: TotalExpenditureDeviationand CompositionVariance Percentage I 2002 2003 2004 2005 Totalexpenditure devlatlon Recurrent 33.1 11.7 -5.3 -4.8 Development 37.4 24.6 33.4 52.5 Total budget 36.0 20.5 21.3 32.7 Cornporltlonvariance Recurrent 39.9 64.7 82.9 Development 108.8 75.6 90.2 Total budget 84.9 70.2 71.4 I Source: Ministryof Planniw and Finance. 2.85 Different causes for the unbalanced execution rates o f the recurrent and investment budget require different solutions. In the case o f the recurrent expenditure, the solution i s the realistic programming o f the budget followed by the respect for the budget ceilings imposed by the budget law inthe context o f the macroeconomic framework. The development budget was discussed above. BudgetExecutionandTreasury Operations Cash budgeting affects allocative efficiency by favoring mandatory expenditures. 2.86 including Rwanda and Zambia - there are continuous tensions between receipts and payments in As is typical inmany Least DevelopedCountries (LDCs) that depend on foreign assistance- Treasury operations (see figures below).66 Although these operations are based on a Treasury plan, whose practical implementationdepends on the evolution o f the primary balance, the unpredictability o f revenues, especially those from external sources, affects the management o f the Treasury operations and forces cash budgeting management (see also the accompanying CFAA volume).67 2.87 A major shortcoming, from the government side, is the lack o f a medium-term financial perspective covering the expected revenues and the likely expenditures consistent with a stable and balanced macroeconomic framework. Budget programming i s made within a one-year horizon, which reduces predictability and increases the risk of shortages at the Treasury. The budget's high dependence on volatile external aid compounds the difficulties o f budget management (see also the accompanying CFAA volume). 66 See Dinh, Adugna, andMyers (2002) andIbudegwu(2005). "CFAA,page38,sectiononaccountingmanagementprovidesfurtheranalysisandtechnicalrecommendations. 48 2.88 Unpredictability o f external funding inflows is one o f the main factors behind the Treasury shortages during the budget execution. Despite the fact that some donors already provide a medium- term financial perspective, setting the priority areas and the respective amounts committed, in practice a number o f difficulties hinder a regular disbursement o fthose funds. 2.89 The tension between revenues and expenditures leads to covering the budget gap by using cash budgeting with a negative impact on the programmed expenditures and on the development priorities. Cash budgeting implies the need to prioritize payments. The prioritized expenditure items generally include mandatory expenditures such as salaries, payment o f scholarships, transfers to health treatments abroad, transfers to embassies, and debt service. Even HIPC-financed expenditure i s often delayed to satisfy the cash budgeting priorities. Besides HIPC, the government has attempted to protect pro-poor expenditures (health, education, and social protection) in its budget for 2005 and 2006, but the process i s more ad hoc than systematic. But the expenditures related to the poverty reduction strategy, that is, the interventions laid down in the NPRS document which should be realized through the public investment program are delayed or not implemented because o f cash budgetingpractices. 2.90 To reduce the negative effects o f cash budgeting on budget execution, the authorities could take steps to (i) create a cash management committee and adopt a systematic approach to cash budgeting; (ii) improve predictability o f aid (discussed below); (iii)improve efficiency inthe use of government resources by adopting a more transparent and programmed approach to cash budgeting: first, publish the cash release plan, and then execute the actual release o f each month according to plan; (iv) improve allocation o f resources by linking cash releases more closely to the budget objectives and long-term priorities o f the government; (v) limit the number o f cases where cash i s released to over-committed ministries or directorates; and (vi) publish budget execution data on a quarterly basis, highlightingdeviations. The accompanying CFAA volume provides further technical recommendations. 8o.m h 8 49 Figure18: Treasury Operationsin 2005 120,000 100,000 ___...__.__._. 80,000 -+-revenues ' -+-Exp??d!ure 40,000 20,000 0 Jan-05 Feb-M Mar-05 Apr-05 May05 Jun-08 Jul-05 Aug-05 Sep-05 Oct-05 Nov-05 Dec-05 Source: STP government Expenditure Review of the Health Sector An expenditure review of the health sector highlights that while total spending has increased, the patternof expenditures suggests inefficient and inequitableuse of budgetary resources. 2.92 Total spending (including private and externally sourced spending) in the sector rose from US$9.9 million in2001 to US$14.4 million in2005, a 45 percent increase. Domestic public spending more than doubled from US$2.7 million to US$6.6 million in the same period, highlighting the commitment o f the government to the sector but also the important role o f foreign aid in the financial sustainability o f the sector. Overall, per capita expenditure o f the total national health system has increased from US$72.5 in 2001 to US$96.6 in 2005. The central budget more than doubledits contribution from US$19.9 per capita to US$44.3 per capita. 2.93 The progress inextending the average life expectancy (now an average o f 65.6 years) and reducing infant mortality and morbidity are positive signs duringthe past five years. While results are above African averages, there i s room for improvement, as shown in the highly successful malaria program that reduced mortality o f children under five from 89 to 60 per 10,000 in a matter o f two years. Table 20: GeneralEpidemiologicalIndicators,2000-05 Per 1,000 inhabitants 2001 2002 2003 2004 2005 Gross rate o f birth 35.6 35.2 34.9 34.4 33.9 Gross rate o f mortality 8.5 8.3 8.1 7.8 7.6 Rate o f infantile mortality 82.7 82.4 81.5 80.7 79.8 Effectiverate o fpopulation growth 1.97 1.97 1.97 1.96 1.95 Life expectency at birth 63.8 64.2 64.7 65.1 65.6 Source: National StatisticalInstitute o f STP, 2006. 2.94 Nonetheless, the current health system suffers from a conjuncture o f human, technical, and physical constraints that lead it to respond essentially to the most prevalent transmissible diseases such as malaria, cholera, and acute diarrhea. At the organizational level, a major 50 constraint i s the lack o f a national integrated information system. This impedes an analysis o f inputs, outputs, and outcomes o f the health system and the evaluation o f policy development, implementation, and financing targets. Institutionally, the health system lacks sufficient administrative and technical capacity. Weak human resources are one o f the major constraints in the country's health system. The low salary base does not attract national specialists studying or residing abroad to return to the country. Furthermore, the country does not have the necessary minimummass for technical training inhealthcare.68 2.95 The supply o f health services i s structured in seven sanitary areas (primary care) and one main hospital and does not respond efficiently and equitably to the population's care needs. The patients with more specialized needs (cardiology, neurology, and the like) are transferred to Portugal at great costs to the health sector budget. 2.96 While STP surpasses the World Health Organization norm o f one physician and two nurses per 10,000 inhabitants (except for the district o f Mi-Zochi), there i s clear disparity in the distribution o f health workers between hospital and primary care, between districts, and with Principe (table below) because o f the concentration of resources at the main hospital and inadequate humanresource development and management. Table 21: Ratio of Health Workers to Population, 2005 Per 10,000 inhabitants Doctors Nurses Others Total Institution International STP Total Per Per Per Per Per Per capita No capita capita capita No capita capita Hospitalcare 13 0.91 22 1.54 35 2.44 102 7.12 614 42.88 738 51.54 Primary care 4 0.28 19 1.33 23 1.61 99 6.91 229 15.99 347 24.23 Agua-Grande 1 0.19 5 0.94 6 1.12 19 3.56 48 8.99 72 13.49 Me-Zochi 1 0.27 1 0.27 2 0.55 25 0.69 49 13.44 75 20.57 Cantagalo 0.00 4 2.89 4 2.89 13 9.40 30 21.69 47 33.98 Lobata 1 0.62 1 0.62 2 1.24 11 6.80 34 21.03 46 28.45 Lembb, 1 0.90 5 4.51 6 5.41 9 8.11 31 27.95 45 40.57 Caue 0.00 1 1.69 1 1.69 13 21.98 32 54.11 46 77.78 Principe 0.00 2 3.16 2 3.16 9 14.23 5 7.91 16 25,30 Total health care 17 1.19 41 2.86 58 4.05 201 14.04 843 58.87 1.085 75.78 Admin. and central services 24 1.68 80 5.59 104 7.26 Source: Administrative and Financial Directorate of the Ministry o f Health, Department o f Human Resources, 2006. 2.97 Total expenditures in the health system have grown 45 percent between 2001 and 2005, notably national public expenditures (tables below), highlightingthe increased attention given to the sector by the government. During the past five years, these expenditures have constituted between 13 percent (2001) and 15 percent (2003) o f GDP. In 2005, total health sector expenditures reached 12.6 percent o f GDP. 2.98 The total expenditures are also highly dependent on international partners and the private sector. For instance, in 2005, 45.9 percent o f the total health budget was financed by the central budget (where 18.2 percent o f the total i s for current expenditures and PIP captures 27.7 percent There is a training center inthe city o f SBo TomC, but it has not functioned regularly over the years. 51 o f the Direct interventions from internationalpartners make up 27.7 percent o f funds, and private sector out-of-pocket contributions, especially to purchase medicine, constitute 26.4 percent o f the total expenditures. 2.99 Cost recovery mechanisms are poor or nonexistent. In effect, in SBo Tom6 and Principe they constitute the out-of-pocket expenditures made by the patient at the time o f receiving services. Cost recovery i s plagued by three major problems: (i)outdated price structure o f services and goods; (ii) use o f collected funds for management o f the center; and (iii) o f lack adequate bookkeeping andreporting inmanagement o f these funds. Table 22: TotalExpendituresof the NationalHealthSystem, 2001-05 US$ thousands Expendituresources 2001 2002 2003 2004 2005 Current expenditures - Central Budget 1,007 899 1,902 2,532 2,625 Investment (capital) expenditures -PIP 1,733 2,218 4,007 3,027 3,980 Total STP public sector 2,739 3,117 5,909 5,559 6,605 External funds ChindTaiwan 959 854 1,127 1,041 710 Portugal 2,108 3,759 3,200 3,929 2,029 Brazil 0 0 0 69 69 World Health Organization 1,226 1,176 1,176 1,176 1,176 Total external finds 4,293 5,788 5,503 6,215 3,984 Private expenditures* 2,937 3,134 3,470 3,858 3,799 Total expendituresinthe nationalhealthsystem 9,969 12,039 14,882 15,632 14,388 Source: STP government. * Estimated from 2001 household survey data. We assumed a constant share o f private expenditures in health inthe family's global expenditures (4.8 percent). Table 23: Total Per CapitaExpendituresinthe NationalHealthSystem, 2001-05 US dollars Expenditures 2001 2002 2003 2004 2005 Public sector 19.9 22.3 41.3 38.1 44.3 Central budget (current) 7.3 6.4 13.3 17.3 17.6 PIP 12.6 15.8 28.0 20.7 26.7 External funds 31.2 41.4 38.4 42.5 26.7 Private sector* 21.3 22.4 24.2 26.4 25.5 Total per capita 72.5 86.0 103.9 107.0 96.6 * Source: Slotomean authorities. Estimated from 2001household survey data. 2.100 While overall expenditures in health have increased, there i s weak allocative efficiency: there i s disconnect between the planning and goals o f the government-as outlined in the NPRS and by the MDGs and the amounts budgeted and executed. Generally, funds are not assigned based on policy orientation. The assignation i s defined based on historical budget shares, with the final decision falling to the Ministryo f Planning and Finance, which executes the overall budget. 69Note that a large share of the overall PIP budget also comes from international partners (see PIP section above). 52 The result i s that while the NPRS and the sectoral strategies highlight the need for improved primary care, 61percent o f the current expenditures are used by the central hospital. Primary care captures only 21 percent o f the current expenditures. Also, as in the case o f the overall budget expenditures, some 40 to 50 percent o f the budget dedicated to sectoral investments are instead used for current expenditures. 2.101 The current expenditure in the sectoral budget i s dominated by three economic categories: salaries and benefits (61 percent o f the total, o f which the main hospital captures 55 percent), goods and services, and medical transfers o f patients to Portugal for specialized treatments. Training and administration have small shares inthe overall expenditure envelope. 2.102 There i s also weak economic efficiency and equity because there are significant cross- district differences in allocation o f expenditures per capita. In2005, the district o f CauC had the highest per capita expenditures (US$14.64), while the district o f Agua-Grande (capital) had the lowest (US$2.12). While per capita expenditure has risen over the past few years, this growth has not been equally shared, with wide variations across districts. In 2005, expenditures for the district o f CauC grew by 1 percent while those for the district o f Agua-Grande (capital) grew by 34 percent. Correspondingly, access and use o f health services i s unequal. Use o f the main hospital and urgent care diminishes in reverse proportion to the distance to the town where the patient resides. On the other hand, the districts o f CauC and Lembh where use o f primary care i s the highest, also use less hospitalizationor ambulatory care. 2.103 The authorities need to reform the organization and the structure o f the health delivery system to improve allocative and economic efficiency as well as equity. The recommendations below provide a way forward. Recommendations:Health Short Term Issues to address I Action to take Misalignment between policy goals Align expenditures on primary health care to match the NPRS and actual expenditures. and national strategy. Lack o f systemwide reliable data. Set up an integrated systemwide monitoring and evaluation system to facilitate follow-up (including a financial accounting system). Issues to address Action to take Weak and unequal cost recovery. Adopt the principle o f equity in terms o f cost recovery: update and unifythe copayments across all health care centers. Weak functioning o f the health (i) the HealthCarePlan(carta sanitaria); (ii) Review define system. reference networks between health units providing care; (iii) improve physical access; (iv) introduce a patient triage system; (v) review the process used to decide about transferring patients to Portugal at government expense; (vi) decentralize budget implementation to provide more flexibility and control to the ministry (after adequate training o f the sectoral staff). Dependence on foreign aid. Consider alternatives to the dependence o f the sector on foreign aid, such as an eventual health system funded by the users (such as a national health care system). 53 Recommendations:OverallBudgetManagement and Coherencewith the NPRS Short Term ~ ~~~ ____ Issuesto address Action to take Budget Management: Measures are needed to introduce (i) Improve the overall fiscal situation by increasing domestic flexibility into the current budget revenues to increase the volume o f available resources, and structure and expenditure policies. exerting tighter control over expenditures; (ii)Analyze and updating civil service rosters and remuneration structure to will help obtain an accurate total o f sectoral and overall number of civil servants, removing all ghost civil servants from payroll. This will improve management and control o f personnel costs; (iii) Ensure that post-HIPC Completion Point, the HIPC debt relief funds will not be absorbed by mandatory expenditures; (iv) Undertake any new borrowing in highly concessional terms (grants or grant elements over 60 percent); (v) consider the preliminary work to develop a medium-tern expenditure framework (MTEF) and (vi) improve donor coordination and harmonization o f donors. Common expenditures has been used Assess the content o f the category common expenditures, as a "catchall" budget category and improving the realism o f funds programmed and the transparency shows systematic programming flaws, o f the funds used. which is reflected in the persistent overruns. The expenditure category transfers, Programthe costs associated with the Joint Development Zone which contains payments to the Joint Administration with a medium-term view to reduce Development Zone Administration unpredictability and impact o n the budget. and scholarship payments, needs to be Review the scholarship policy and costs inlight o f the education reviewed. sector strategy. Weak budget execution: Budget ceilings are not respected. The respect for the budget ceilings imposed by the budget law in the context o f the macroeconomic framework i s critical. The budget circular does not clearly The budget circular that the MoPF distributes to the spending state the government's policy units should clearly state the policy priorities for the main sectors priorities and is not considered a and a financial envelope for each spending ministry (including guide in the spending units' budget recurrent and investment expenditure). preparations. Negative effects o f cash budgeting on Set up a cash management committee and adopt a systematic budget execution. approach to cash budgeting. NPRS and PIP: PIP does not assume its key The newly adopted nomenclature includes functional instrumental role in t i e classification can help redefine PIP projects along the implementation o f the poverty programmatic lines o f the NPRS. reduction strategy. A manual on PIP, covering the elaboration and execution phases, should be elaborated to guide the spending units in the programming o f their investments according to the NPRS objectives, and should focus on performance in addition to the financial management. Project selection criteria should be revised a systematic multicriteria method, including and project recurrent costs, should be elaborated to score the project proposals according to project merit. Harmonization and coordination o f Improve harmonization and coordination o f donor funds. donor funds are weak. The NPRS Unit (Poverty Make the NPRS Unit oDerationa1. The NPRS Unit should 54 Issuesto address Action to take Observatory) i s not fully functional assume its monitoring and evaluation role inpractice, through its and Coordination between the direct involvement inthe ongoing budget reform. It should also be Directorates o f Budget and Planning involved inthe PIP preparation. and the NPRS Observatory i s rather Strengthen coordination between the Directorates o f Budget and weak IPlanning and the NPRS Observatory is rather weak. Mediumterm Issuesto address Action to take Budget Management: A high level o f unpredictability in Elaborate a multiyear macromodel that would allow a financial financial flows affects budget resources perspective, to improve predictability. execution and provokes Treasury Start developing a medium-term expenditure framework; choose shortages. two or three ministries to test the framework. The transfers expenditure category: Costs o f the Joint Development Zone Implement medium-termbudgeting, Administration Scholarships paid to students abroad Consider revising the policy on scholarships paid for studying constitute a heavy burden o n the abroad, based on the analysis done (in the fiist year). One central government and sectoral approach would be to encourage cost sharing with students' budget. families. The system for managing human Use the new civil service registry to help with the planned resources in the civil service needs to analysis and consequent restructuring o f the civil service be modernized and a personnel cost qualification and salary structures. control system created." Common expenditures are used in a Create a mechanism, based on the accounting system foreseen in manner that is opaque and centralized. the context o f the ongoing reform o f public finance management, to apportion the public expenditure made common expenditures to specific spending units. This will enable officials to assess costs in the provision o f public services and the analyze its efficiency. Cash budgeting has negative effects 'he authorities could take steps to (i) improve predictability o f aid onbudget execution. bough better coordination and harmonization with donors; (ii) mprove efficiency inthe use o f government resources by adopting more transparent and programmed approach to cash budgeting: irst, publish the cash release plan and then execute the actual elease o f each month according to plan; (iii) improve allocation o f esources by linking cash releases more closely to the budget ibjectives and long-term priorities o f the government; (iv) limit the lumber o f cases where cash i s released to overcommitted ministries ir directorates; and(v) publishbudget execution data ona quarterly iasis, highlighting deviations. NPRS and PIP: PIP does not have a monitoring and Monitoring and evaluation (M&E) o f PIP'S programmes and evaluation system compatible with the projects should be closely linked with M&E o f the poverty NPRS objectives. reduction strategy, involving the NPRS Unit (poverty observatory) and the national statistics office. Weak coherence o f the NPRS and Review and realign expenditure allocations to the development budget. priorities o f the country. 70As reportedin Grandes Opc6esdo Plano-2006, Ministryof PlanningandFinance, January 2006, 55 3. THE FISCALRISKS AND REFORMCHALLENGESOF STATE-OWNED ENTERPRISES 3.1 This chapter reviews the status o f the fully state-owned enterprises (SOEs) and their fiscal impact on the budget and the economy. The first section provides a brief overview o f the privatization efforts since 1990. The second section benchmarks the regulatory and deteriorating financial status o f the four remaining full SOEs-Empresa de Agua e Electricidade, EMAE (water and electricity); Empresa Nacional de Aeroportos e Seguranca Akrea, ENASA (airport security and administration); Empresa Nacional de Administragiio dos Portos, ENAPORT (seaport security and administration); and Empresa de Correios, CORREIOS (post)-all concentrated in the infrastructure/service sector. Section three provides an estimate o f the fiscal impact o f these SOEs on the central budget. Section four focuses on impacts o f the implicit subsidy given to EMAE and the company's hypothetical tariff increases on the fiscal choices o f the authorities and on poverty. Section five focuses on the need to develop an appropriate regulatory framework, and sections six and seven conclude, highlightingpolicy recommendations for the short- and medium-term horizon. Annex 111.1provides an in-depth analysis o f EMAE's situation. Annex 111.2 provides more in-depth analysis o f the financial performance o f ENASA, ENAPORT, and CORREIOS. 1. OVERVIEW OF PRIVATIZATIONEFFORTS SINCE 1990: REDUCTION OF STATE OWNERSHIPIN THE ECONOMY 3.2 Over the past 15 years, the government o f SZo T o m i and Principe has undertaken an extensive privatization in an attempt to fully or partially withdraw from economic sectors.71The government now has four partially state-owned enterprises: 51 percent o f the petroleum importing and processing company (Empresa Nacional de Combustiveis e Oleos, ENCO), 49 percent o f CST Telecom, 40 percent o f Banco International STP (BISTP), and 30 percent o f Air Si50 Tomi. Air Siio Tom6 went banlaupt in2006, after its lone plane crashed. 3.3 The government has retained and i s still the full owner, regulator, and indirectly the manager and operator o f the four infrastructure SOEs: EMAE, ENASA, ENAPORT, and CORREIOS. These enterprises provide essential services to the public sector and to a growing private sector, as well as to a small segment o f residential consumers. In 2005, the enterprises jointly employed a total o f 527 employees, or 15 percent o f an approximate total o f 3,500 employees in the public sector (excluding the military and police force). They generated combined annual revenues close to US$16 million, equal to 14percent o f the country's GDP.72 7' In 1990, the government established a state-owned enterprise reform commission at the Ministry of Planning and Finance (MoPF) which later was transformed into the Privatization Unit. The Privatization Law in 1992 allowed all state-owned enterprises to be converted into joint stock companies with no limits on foreign ownership while instructing the Privatization Unit to privatize or liquidate 15 nonagricultural enterprises. A summary o f the results o f the privatization effort is presentedinbox 2 below. 72 In2005, GDP was US$ll4 million. 56 Table 24: Legal and Ownership Status of the SOEs ENTERPRISE LEGAL STATE PRIVATEFOREIGN PRIVATE LOCAL NAME STATUS OWNERSHIP OWNERSHIP OWNERSHIP ENCO(oil company) S.A.R.L. 51% 40% 9% Local Sonongol (Angola) CST Telecom S.A.R.L. 49% 51% Portugal Telecom Air SHo Tomb S.A.R.L. 30% 65% 5% Workers Air Portugal E M A E (electricity & water) Statutory corporation 100% ENASA (airport security Statutory 100% and administration) corporation ENAPORT (seaport security Statutory 100% and administration) corporation CORRIEOS (postal services) Statutory 100% corporation INAC (civil aviation) Statutory 100% Source; Compiledfrom various sources. Box 2: Reforming the State Sector, 1990-2004 A Privatization Unit,formed to facilitate the privatizationprocess, managedto liquidate several enterprises: the meat-processing enterprise (ENCAR), poultry-breeding enterprise (EMAVE), fishery (EMPESCA), pharmaceuticals distribution company (ENAMED), publishing (EMAG), construction (CONSTRUCTURA), and transport (TRANSCOMAR). It also transferred 100 percent o f government stake in the former beer manufacturer (ROSEMA), the central repair and maintenance shop (ATELIER CENTRAL), andPoussadaBoa Vista Hotelto the private sector. Some SOEs were partially privatized, including CST Telecom, ENCO, BISTP, and Air SBo Tom6 (see also table 24). The government reduced its 20 percent stake in the brick manufacturer and signed a 30-year concession agreement with a foreign operator to manage Hotel Minamar. It reduced its minority share in Air SBo Tom6 from 35 percent to 30 percent bytransferring shares to the workers. The Privatization Unit apparently received an estimated total o f US$4 million as privatization proceeds. After paying for the operating expenses o f the Privatization Unit, the remainder apparently entered the coffers o f the Treasury and the unit was d i ~ m a n t l e d . ~ ~ The privatization o f agricultural enterprises became the responsibility o f Ministry o f Agriculture under a World Bank (IDA) loan o f US$7.3 million, which was executed between 1992 and 2001 with mixed results.74 Source; Compiled from various sources in S i 0 Tom6 andPrincipe. 73The PrivatizationUnitdid not compile any information as to which enterprises were sold, at what price, who the new owners are, or how labor was affected. 74Information providedby the Debt Division of the MoPF. 57 2. THESTATUSOFTHEFOUR SOES Legal Status of State-OwnedInfrastructureEnterprises 3.4 Created as statutory corporations, the SOEs have their own budgets but have limited management autonomy because their shareholding structure i s not well defined by any legal framework or company law.75They are governed by a three-person board, but the board members are not independent o f the government or the SOEs. Therefore, their reporting relationship to their shareholder (government) i s not based on a specific shareholding agreement or a performance contract but i s exclusively controlled by the respective "tutelle" ministry. 3.5 The SOEs' assets are fully owned by the MoPF, while their respective tutelle ministries regulate their tariffs andor pricing structures and develop their overall policies. On this basis, the country follows a dual model o f state ownership, which still represents the most common practice among developed and developing countries. The tutelle ministries are also expected to monitor the SOEs' performance. In practice, this often translates into interference into day-to-day management o f the SOEs. 3.6 The statutory relationship defines a mutual financial relation that has broken down, to the SOEs' detriment. SOEs are supposed to make payments (taxes and transfers) to the central government's budget and are entitled to receive equity injections under the title o f "investment subsidies" from the government's public investment program (PIP). The SOEs have not received any equity injections or investment guarantees from the government since 2002, which has curtailed their ability to undertake necessary investment in the enterprise^.^^ In the past, these equity injections were either made available by concessional loans received by the government and passed on to the SOEs, or not always attached to an earmarked concessional loan. 3.7 Because o f the nature o f its relationship with the SOEs, the government i s the guarantor o f all their loans, regardless o f whether the loans are obtained through the government, highlighting the risk and burden to the central budget. The government has been the explicit guarantor o f the loans borrowed on behalf o f the SOEs from the international financial institutions, and it will have to service the loans when the grace period i s over if the enterprises are not able to service them on time. Also, the government i s ultimately responsible for the SOEs' short- or medium-term loans from domestic financial institutions, even if it i s not asked to provide guarantees at the time o f lending. The latter constitutes the implicit guarantees o f the Treasury. 3.8 The SOEs attempt to achieve commercial goals (breaking even or making profits) while operating with the mandate o f fulfilling the state's social objectives. However, there are no systems in place to compensate them adequately for such social services: the government does not budget the cost o f its social goals and transfer funds to the SOEs to cover below-cost-recovery tariffs and does not provide the investment funds that SOEs received before 2002. These social objectives include achieving income redistribution and regional development and providing the 75 Inthe evolutionaryprocess, fully state-ownedenterprisestake one of three distinct forms, each with its own status and varying degree of autonomy: departmental undertakings, statutory corporations, or joint stock companies with shares owned by the state. In STP, the infrastructureSOEs were created under a special government decree declaring them "statutory corporations." This form has given them more operating flexibility over the former "departmental undertakings" form, which has beencommoninmanydevelopingcountries. 76 Based on an agreement with the IMF's PRGF-supportedprogram. The IMF was concerned about the soft budget constraint andwanted to introducedmore transparencyinto the centralbudgetprocess. 58 essential goods and services at such prices that would not impact the cost o f living o f the majority of the population. This latter goal has translated into below-cost-recovery tariffs for EMAE. 3.9 The authorities recognize the need for reform, and the management teams o f the SOEs, especially the team at EMAE, are interested in corporatization. Corporatization will change the legal status o f the SOEs from "statutory corporations" to "joint stock companies" (Sociedade Anonima, S.A.)registeredunder Company Law with defined equity shares owned 100percent by the state.77 This legal transformation will help the management o f the enterprises to establish autonomy and improve the performance o f the enterprises. However, the successive government changes have slowed the decision-making process. The Financial Status of the Four SOEs 3.10 The four SOEs have a precarious financial situation (table 25 below). EMAE and ENASA have incurred large losses in 2005, while ENAPORT and CORREIOS report a small profit. All four SOEs have large net liabilities, but EMAE i s by far the most indebted. More analysis o f EMAEs' financial, operational, and technical performance i s included in Annex 111.1. Annex 111.2 contains further information on the financial performances o f ENASA, ENAPORT, and CORREIOS. 3.11 The elimination o f government subsidies-which the IMF recommended in 2002 to strengthen transparency and discipline o f the central budget-partially contributed to the SOEs' financial difficulties. Until 2001, fully state-owned enterprises received operating subsidies from the government's recurrent budget. These subsidies ran as highas Db 5,062 billion (US$551,000) in2001.78 77 Intheory, a successful corporatization is based on the concept of commercialization, but the implementation of this concept i s difficult. Commercialization builds on the concept of an expected economic or social return on investment. Commercialization will require the SOEs to define their commercial and noncommercial objectives in a transparent manner and demand compensation from the government according to the specified noncommercial objectives. SOEs' commercial and noncommercial objectives are clearly separated, and the SOEs specify their noncommercial policies and activities and make a case to the government for separatetreatment. They are thencosted each year: the agreed net cost is included in the government's budget as expenditure and in the enterprise budget as revenue. The government accepts that no instruction or directive can be given thereafter to a public enterprise, or to any officer of an enterprise, except inwriting together with an undertaking to meet the cost (UnitedNations 1995, 15). 78 And Db 646 billion (USrS75,OOO) in operating subsidy in 2000. However, the SOEs may have received some arrear payments from the government during 2003 (USrS1.274 million) and 2004 (USrS2.365 million). These payments are likely to have been made from the budget rubric encargos communs (common charges), which also pays for electricity, telecoms, and water for the state. The 2004 total for encargos communs was much larger than previous years because o fthe doubling o f telecoms and electricity bills. 59 Table 25: SummaryDatafor SOEs, 2005 US$ and ratios PKOFl'T/l,OSS STATEMENT EMAE ENASA ENAPORT CORREIOS Sales revenue 5,936,631 883,254 1,822,422 262,712 Number of employees 186 130 158 53 Operatingexpense 6,789,965 1,060,629 1,820,303 228,85 1 - Wage bill 783,583 848,503 525,259 n.a. - Taxes (a) 92,089 -86,848 (b) 22,128 12,728 Operatingincome -850,334 -177,375 2,119 33,861 Financialexpenses, 973,617 n.a. n.a. n.a. depreciation,and taxes Net profit/loss -1,8 12,724 -177,375 2,119 33,861 Profitsharing 0 0 16,720 0 Balance sheet Liabilities 11,320,814 321,589 772,729 1,249,757 - Short-term debt 3,193,398 - Long-term debt 8,127,416 Equity 5,045,840 739,040 632,822 - 100,256 Assets 16,366,654 883,254 1,405,552 1,149,502 Financialratios Profitmargin -3 1(Z, -20(!/0 0% 13% Return on equity -3 694) -24% 0% -34% Returnon assets -11% -2090 0% 3% Liabilitiedequity 224% 44% 122% -1247% Source: Compiled from SOEs' financial reports; net profit/loss i s after taxes, depreciation, and financial and other expenses. Profit margin i s net income over sales.. Note: n.a. = not available (a) Taxes include payroll taxes, consumption tax, and a stamp duty that were available only for 2005. (b) While the MoPF reports the taxes as received, ENASA reports its taxes in2003, 2004, and 2005 as tax arrears, that is, due but not yet paid. 3.12 The continued financial connection between the SOEs and the government affects the financial health o f the SOEs. These transfers from SOEs to the recurrent budget take on the form o f direct transfers (profit sharing and taxes) and indirect transfers (the government and intra-SOE accumulated payment arrears), as discussed below. 3.13 Inthe meantime, the government has not provided bailouts to help SOEs with their deficits or provided equity injections. This partially explains how the SOEs continue to increase their supplier arrears, including those to other SOEs, augmenting the accumulated inter-arrears among themselves, tax arrears to the Treasury, and occasional wage arrears to their employees. The ContinuedFinancialConnection Transfers from SOEsto the recurrent budget 3.14 The financial transfers between the government and the infrastructure SOEs suggest that SOEs actually benefit the government's recurrent budget. The SOEs' deteriorating financial 60 situation i s a simple indication o f their inability to maintain existing levels o f coverage under the ongoing government and state-owned enterprise relationship. 3.15 The transfers from the infrastructure SOEs to the government's recurrent budget follow two paths: the profit sharing arrangement they have with the Treasury, and the four kinds o f taxes they pay as any other enterpri~e.'~ law, the SOEs are to transfer almost all their total profits to By the government. They transfer 80 percent oftheir net profitsto the government's central recurrent budget every year under the item, "Participacao do Estado no lucro das empresas" (participation o f the state in the profit o f the enterprises." They also pay additional profit tax on the remaining 20 percent o f their net profits at a rate o f 30 percent and a further 15 percent in the event their remaining profits exceed Db 12 million. Besides these profit sharing and profit tax payments, they pay a payroll tax (13 percent o f base salary) on behalf o f their employees, a consumption tax o f 5 percent (which has been raised from 2 percent to 5 percent on July 15, 2005; EMAE i s exempt from this tax), and an additional stamp duty o f 6 percent over their sales revenue. The SOEsare exempt from import dutiesagO Direct Transfers: Profit Sharing and Taxes 3.16 Of the four SOEs, EMAE and ENASA are running large deficits. Therefore, the taxes paid (or owed) to the Treasury far exceed the expected profit sharing and profit taxes on their net earnings.81 The profit sharing as transfers from the infrastructure SOEs to the Treasury declined sharply over the last five years (see table 26).82983The decline in profit sharing from the infrastructure SOEs reflects decreased and even negative profitability (see table above and Annexes 111.1 and 111.2 for further discussion). Since some SOEs have accumulated tax arrears, the taxes actually paid by the SOEs to the Treasury increased only slightly between 2000 and 2005 (table 27). ''Profit sharing (in lieu of dividends for the government's equity) and profit taxes are the only revenuesthe Treasury receivesdirectly from its SOEs. The partially owned SOEs are createdas S.A.R.L.and, therefore, they are expectedto distributedividends to all shareholdersif andwhen profitable.The government receivestransfers from partially owned SOEs. The estimated 2006 value was Db 5.2 billion (based on information from the authorities, and the IMF). However, since the Treasury does not distinguish betweendividendsandprofit taxes, we can not providea breakdown o f the share o f dividendspaidby partially owned SOES inrelationto the state's equity share. The amounts of the import taxes foregoneby the Treasuryare not currently available. Some SOEshaveaccumulatedtax arrears to the state andhavenot paidtheir taxes owed to the Treasury. ENAPORTrecently managed to negotiate with the Treasury to transfer only 60 percent of its profits instead of 80 FFrcent,but it still pays the profit tax over the remaining40 percent. The Treasury does not disaggregate the dividendsand the profit taxes it receives from partially or fully state-owned enterprises; instead, it registers only one single amount under the categoryof "comparticipacao do estado no lucro des empresasestatais" in its recurrent budget. The Treasury should separate the total amount entered into "dividends" and "profit taxes" for the partially state-ownedSOEs, and into "profit sharing" and "profit taxes" for the fully state-owned enterprisesinits recurrentbudget. 61 Table 26: Profit Sharingfrom InfrastructureSOEs to the Treasury (Db millionsand US$)84 2000 2001 2002 2003 2004 2005 EMAE (electricity andwater) 0 1,716 0 0 0 ENASA(airport security & management) 0 (60) (60) 0 0 0 ENAPORT (seaport security and manag't) 1,069 1,383 1,072 347 0 201 CORREIOS(postalservices) 0 50 0 0 0 0 TOTAL., Db millions 1,069 3,209 1,132 347 0 201 US$ 124,144 354,000 122,750 36,840 0 16,720 Sources; Compiled from the income statements of the enterprises and the Treasury; there is, however, a discrepancy between the data provided by the Treasury and ENAPORT. Even though ENASA has not made any profit to pay for the profit sharing, the governmentreports that it receivedtransfers from ENASA in2001 and2002 as indicatedabove. Table 27: Taxes fromInfrastructureSOEsto Treasury (Paid and/or Owed) Db millionsand US$ 2000 2001 2002 2003 2004 2005 EMAE (electricity andwater) 373 603 690 575 690 1,107 ENASA(airportsecurity & mangm't) 3 233 309 (280) (657) (1,044) ENAPORT (seaport security & mgmt) 357 666 452 384 397 266 CORREIOS (postalservices) 43 45 119 119 132 153 TOTAL, Db millions 776 1,547 1,570 1,078 1,219 1,526 US$ 90,000 170,694 170,245 114,681 121,900 127,167 Sources; Compiledfrom data providedby the Direction du Budget, MoPF. Taxes include payroll taxes, consumption tax, and a stamp duty that were available only for 2005. While the MoFP reportsthe taxes as received, ENASA reports its taxes in2003, 2004, and2005 as tax arrears, that is, due but not yet paid. Indirect Transfers: Accumulated Payment Arrears Paradoxically, the accumulated payment arrears from the government and among SOEs seems to have helpedthe governmentin two ways, describedbelow. 3.17 As the major consumer o f the services from the infrastructure enterprises, the government's accumulating payment arrears to the SOEs allows it to create a fiscal space in its recurrent budget.85Accumulating chronic arrears allows the Treasury to allocate cash according to its own priorities while delaying its payment obligations to the SOEs. The stock o f accumulated arrears (the government's unpaidbills) to the four enterprises reached roughly US$4 million by the end o f 2005 (close to a fourth o f the SOEs total revenues), adversely affecting the SOEs' cash flow and operations. Uncollected payments reduce the cash flow of the enterprises and worsen their finances because the enterprises become unable to pay their own suppliers, wage bills, taxes, andmaintenance expenses. 3.18 The arrears that are due to infrastructure SOEs by other (partially owned) public enterprises also benefit the government. The government continues to use the services o f all 84The exchange rateusedto convert the localcurrencyto US dollarsthroughoutthe reporti s providedby EMAE and i s indobras: 8,611 (2000); 9,063 (2001); 9,222 (2002); 9,419 (2003); 10,147 (2004); and 12,021 (2005). 85 For instance, EMAE bills were not always paid in full or on time until recently. ENAPORTreports that it has not been paid for its handling charges of the medical equipment and medication received by the Ministry of Health, or handlingof cars for the Ministry of ForeignAffairs and Ministry of Defense. 62 (partially or fully state-owned) enterprises that are only able to "stay inbusiness" by buildingup their own arrears with one another.86 3.19 There are also accumulated inter-arrears among the infrastructure SOEs themselves. ENCO, the oil importer and partially owned state enterprise, is reported to refuse to pay the bills it receives from ENAPORT. EMAE has been suffering from ENASA's accumulated arrears (see table 28 below). CORREIOS, on the other hand, has accumulated arrears to ENASA. 3.20 An indicative calculation o f arrears among the SOEs and with the Treasury as of October 2006 i s presented in table 28 below. The calculations are limited by the quality o f the data provided by the SOEs and it i s not possible to verify them independently. The debts are mostly commercial, usually cleared within three months. For instance, part o f ENCO's Db 18,775 million debt to the government represented tax arrears that were cleared in November 2006. The exception i s the longer-term arrearddebt triangle between EMAE, ENCO, and the government, which emanates from the precarious financial circumstances o f EMAE and the sharp oil price increases in 2004-06. EMAE's debt to ENCO contains both programmed payments from previous debt rescheduled in April 2006-under a triad financial agreement with the government-and more recent arrears accumulated inthe summer and fall o f 2006. Table 28: Intra-SOE and GovernmentDebt, October 2006 Db millions (Oct 06) I Source: SOE financial statements. There have not been any systematic equity injections from the government to SOEs since 2002. 3.21 Up to 2002, the MoPF provided equity injections to the SOEs through its public investment program (table 29). These "investment subsidies" to the SOEs were primarily the transfer o f concessional loans that the government had taken out on behalf of the infrastructure enterprises. The total concessional loans on behalf o f two o f its infrastructure SOEs, EMAE and ENASA, are equivalent to US$22.5 million and constitute 7 percent o f Sgo TomC's debt stock o f approximately US324.5 million as o f end-2005.87 EMAE's loans alone represent 40 percent of 86ENCO fits into this category. While it is apartially privatizedenterprise, its majority owner, the state, still behavesas ifit is the fullowner andreceivesimplicitsubsidiesby continuingto usethe servicesofferedbythe other fully owned SOE, namely that of ENAPORT, free of charge. ENCO, on the other hand, providesan implicit subsidy to EMAE by selling fuel to EMAE at a price that is 35 percent below the domestic market price. Furthermore, it imports fuel every 45 days simply for EMAE's sake; it would import fuel every 75 days if it were not the sole provider of fuel to EMAE. "AccordingtotheDirectiondelaDetteattheMoFP,thispracticestoppedafter2002,leavingtheSOEstorelyon their own very limited funds for investmentandmaintenance. Someof the SOEstried to secure loans directly from the 63 total loans from financial institutions to the Infrastructure SOEs via Treasury with US$9.25 million. European Investment Bank (ED) provided the largest loan for the electricity sector (US$5.28 million). EMAE's water sector has apparently received far less compared to other infrastructure sectors, only US$1 million over the past fifteen years: it received two small loans from Arab Bank o f Economic Development inAfrica, BADEA (US$O.15 million) in addition to a larger loan from Organization o f Oil Producing Countries, OPEC (US$0.86 million). The remaining infrastructure loans o f US$4.4 million were for the expansion o f the Sgo Tom6 Airport. Table 29: LoansfromFinancialInstitutionsto InfrastructureSOEs via Treasury, between1986 and 2006 (US$ millions) SECTOR Original Year Interest Grace Maturity Amount Amount to amount rate period period disbursed be paid Water (EMAE) BADEA 3.20 2000 2.00% 5 25 0.10 0.10 BADEA 1.50 2002 2.00% 6 25 0.06 0.06 OPEC 1.10 1990 3.25% 5 17 0 0 OPEC 1.29 1997 3.OO% 5 17 1.3 0.09 Electricity (EMAE) European 7.00 1995 1.OO% 0 0 5.28 5.28 Investment Bank (EIB) African 4.55 1994 0.75% 10 50 2.52 2.50 Development Fund Subtotal(EMAE) 18.64 9.25 8.80 Airport (ENASA) African 3.83 1992 0.75% 10 50 3.33 3.29 Development Fund African 12.93 1986 0.75% 10 50 12.93 1.12 Development Fund Subtotal(ENASA) 16.76 13.3 4.40 TOTAL 35.40 22.50 13.20 Source; Compiledfrom data receivedfrom Direction de la Dette, Directeurde Tresor, MoPF; euro and SDR amounts were converted to US dollars by the consultant. The utilized amount reported for BADEA loan was taken from EMAE's own documentsas US$59,090,while the Cabinet de laDette reported a lower amount of US$35,000. 3.22 Only 63 percent o f these loans from the international financial institutions to two infrastructure sectors have been disbursed. The remainder either has been cancelled or has not yet been disbursed partially because o f low national absorption capacity or problems with donors' disbursement procedures. Of the amount disbursed, the Treasury has already paid back some 40 percent, or US$9.3 million. The outstanding debt i s around 58.7 percent, or US$13.2 million,88 market ,but these still required government backing, which did not materializebecause it would translate into the government's contractingfurther liability/debt andbecauseit went againstthe government's agreement with the IMF. Half of this amount, USS6.9 million from ADF, could be forgiven once the country reaches the HIPC completion point. For ADF, all public and publicly guaranteed external loans disbursed and outstanding before end-2004will be cancelled under MDRI. Those loans remaining after MDRI and disbursed before end-2005 (essentially all 2005 disbursements) will receive HIPC debt relief as part of topping up. However, this forgiveness will depend on discussions between authoritiesand the ADF because World Bank records did not list this credit as part of the HIPC completionpoint debt envelope. 64 Loans from the African Development Fund (ADF), BADEA, and EIB are under discussion for potential debt relief.89 3. GOVERNMENT'S EXPOSUREINRELATIONTO SOES RISK 3.23 Section three provides a measure o f government risk exposure regarding the deteriorating operations o f its infrastructure enterprises and an estimate o f the fiscal impact o f these SOEs on the central budget. 3.24 The liabilities o f the SOEs are especially worrisome in light o f their zero or negative profitability and return on equity ratios combined with declining equity in their balance sheets (see Annexes 111.1and 111.2 for details on all four SOEs). Therefore, as the owner o f these SOEs and the explicit and implicit guarantor o f their liabilities, the government's risk exposure i s the sumofdirect and any future contingent liabilities: 0 direct liabilities that these enterprises have accumulated over the years to their suppliers (domestic or international), their short-term (domestic) and long-term (international) debt to financial institutions guaranteed explicitly and implicitly by the government, and the needed equity injections to recapitalize them; 0 contingent liabilities if EMAE enters into a power purchasing agreement with an independent power producer and fails to honor the contract, the government will have a contingent liability. 3.25 The calculation of the full risk exposure o f the government was not possible because o f lack o f data. The direct liabilities were accounted for and by end-2005 summed to US$13.66 million (see table 30 below), an increase o f 159 percent in five years. The government needs to decide how to address this large liability by setting aside reserves or negotiating an understanding with lenders. Table 30: Government'sCombinedRiskinRelationto Four SOEs (US$ millions) 2000 2001 2002 2003 2004 2005 Share as of 2005 (Yo) Total 5.28 6.38 7.22 10.05 11.14 13.66 100.00 liabilities --EMAE 4.09 5.29 5.72 7.86 8.54 11.32 82.69 --ENASA 0.18 0.19 0.24 0.25 0.29 0.32 2.35 --ENAPORT 0.46 0.21 0.41 0.64 1.05 0.77 5.65 --CORREIOS 0.55 0.70 0.84 0.1,3 1.25 1.25 9.15 Source; Complied from all liabilities reportedinAnnexes I-IV. 3.26 The proper calculation o f equity injections from the government to recapitalize the SOEs requires more in-depth study by the government and EMAE. 89The only EIB loan outstanding as of end 2005 is the loanto Power Rehabilitation(loan 71001). The creditor decides whether this loan will receive debt relief. However, according to the creditor's information (confirmedby authorities), the interimrelief that has beenprovided already covers the EU/EIB share of HIPC relief (provided in other loans that were outstanding), so for this loan, we would expectthat only asmallportion would be cancelled, as part of the topping up process. 65 3.27 Also, the government should be aware o f the financial and regulatory risks in EMAE's signingpower purchasing agreements. The government has to calculate the contingent liabilities attached to EMAE's purchasing agreement and should follow the best practice o f setting aside reserves to cover this liability. These reserve funds can be invested to maximize the value o f assets.90Also, on the regulatory front, the government needs to address the current weakness in regulatory supervision inthe electricity sector (see section 5). 3.28 However, putting aside a reserve fund to pay for the total liabilities o f these SOEs (including the equity injection to recapitalize the SOEs) will not put an end to the ongoing problems o f its infrastructure SOEs. Enterprise and institutional reforms in this subsector are required. Recommendations for short- and medium-term reforms are discussed in sections 6 and 7 below. 4. EMAE'sIMPLICITGOVERNMENT SUBSIDIES, FISCALSPACE, AND POLICY CHOICES 3.29 As the most important SOE interms o fbothits potential impact on the central budget and its central role inaffecting economic development, EMAEis worthy o f a more in-depthfiscal and operational analysis, which i s provided inAnnex 111.1. 3.30 This section estimates the implicit subsidy and tax break that EMAE receives. It also looks at a hypothetical increase intariffs that may be necessary for EMAEto break even (with or without subsidies or tax breaks), analyzing the distributional impact o f these tariff increases on the poor compared with other public expenditure choices. Finally, this section reviews the policy choices for the government. Implicit Subsidies and TaxBreaks Benejiting EMAE 3.31 The implicit subsidies that EMAE receives are related to the price it pays for its fuel to ENCO. Calculations based on the structure o f diesel prices (Annex 111.1 table) confirm that compared with the CIF (cost, insurance, and freight) import price, ENCO sold fuel to EMAE at a price o f 37 percent (or approximately Db 1,500 per liter) below the world market price. In October 2006, this subsidy rose to Db 2,457 per liter (but still 37 percent below the world market price) because o f the sharp increase inworld prices." 3.32 The total value o f the implicit subsidy in2005 i s estimated to be as high as Db 15 billion (US$1.43 million). The total implicit subsidy received from ENCO in 2006 i s estimated to be above Db 28 billion or US$ 2.26 million (table 3l).92'93 The government might decide to hold its reserves offshore in a foreign currency or domestically in the domestic currency. Ifthe guarantees are denominated in US dollars, the government should consider investingthe reserve fund indollar assetsandpossiblykeepingthe reserve offshore to circumventconvertibility risk issues. 9'The estimate is supportedby statements by ENCO that it currently sells fuel to EMAE at a price that is 37 percent below the world marketpriceand, furthermore, it importsfuel every 45 days simply for EMAE's sake insteadof every 75 days if itwere not the sole provider of fuel for EMAE. 92The implicit subsidy i s estimated at Db 2,457 (as of October 2006). The price paidby EMAE at Db 7,000 per liter was only half the amount chargedto end users. Three major factors account for the differencebetween the pricepaid by EMAE and the price paidby end users. First, the fuel usedby EMAE for electricity generation is not taxed, which leadsto a savings of 42 percent of the CIF price, or Db 2,773 per liter. Second, the domestic retail costs for EMAE's fuel are lower than for end users of diesel, at Db 477 dobras per, because there is no road maintenancetax to pay, as well as no retailer's margin. That is, the distribution costs for EMAE fuel is only Db 477 per liter, which corresponds 66 3.33 Inaddition, EMAE does not have to pay the 5 percent import duty on diesel, nor a 37 percent tax on diesel fuel paid by other end users. The value per liter o f these taxes was equal to Db 2,714 per liter inJune 2006 and Db 2,773 per liter inOctober 2006. Taking a value o f 2,750 dobras per liter for the year as a whole, the tax break provided to EMAE customers, given that electricity sales are not taxed, was equivalent to Db 33 billion, or US$2.64 million. In2005, the estimates are lower because o f the lower CIF import price, but the tax break i s still substantial at US1.85 million. Total implicit subsidies and tax breaks provided to EMAE for 2005 are estimated at US$5.1 million, and at US$3.25 million when only the implicit subsidies are included(table 31). Table 31: EMAE-Subsidies, Tax Breaks, and Hypothetica Tariff Increases to break even 2005 2006 Exchange rate (dobras per U S dollar) 10.700 12,500 Diesel consumptionby EMAE(liters) 10,646,383 12,005,469 ENCO implicit subsidyper liter (dobras) 1,44 1 2,350 ENCO total implicit subsidy(dobras) 15,341,437,903 28,212,852,878 ENCO total implicit subsidy (US$) 1,433,779 2,257,028 EMAEtax break per liter (dobras) 1,861 2,750 EMAEtotal tax breakper liter (dobras) 19,812,918,763 33,O 15,040,602 EMAEtotal tax breakper liter (US$) 1,851,675 2,641,203 Operating loss o f EMAE (dobras) 1,812,724 n.a Total shortfall including tax break (US$) 5,098,178 n.a Total shortfall excluding tax break (US$) 3,246,503 n.a Total revenues (US$) 5,936,631 n.a Estimated share o f electricity revenues (% o f total) 75.0% n.a Electricity revenues (US$) 4,450,258 n.a Hypothetical increase in electricity tariff incl. ENCO subsidy and tax break e!) 114.6% n.a Hypothetical increase inelectricity tariff excl. tax break (%) 73.0% n.a Hypothetical increase in electricity tariff excl. ENCO subsidy and tax break (%) 40.7% n.a Source: Wodon et al. 2007. Note: NA =not available To break even, EMAE tariffs will have to increase substantially. 3.34 If the authorities decided to increase tariffs to help EMAE break even financially, the hypothetical tariff increases would be substantial. This indicative analysis assumes no new investment, no subsidies, and no tax breaks. In 2005, the average increase in electricity tariffs needed for EMAE to cover its operating losses i s 40.7 percent (assuming it continues to receive tax relief and subsidies from ENCO). The tariff increase would have to be 73 percent to cover operating losses while ENCO implicit subsidies are terminated. The tariffs would increase further only to the retail distribution fees, as opposedto Db 1,521 per liter for end users. Third, ENCO is apparently making a loss on the fuel it sells to EMAE, while it is making a profit on the fuel sold to end users. The "extra profit or loss" versus the normal 8 percent marginthat ENCO should be able to obtain was Db 2,457 per liter in October 2006 (see Annex I11for calculations). Calculationsbased on the structure of diesel prices (Annex 111.1) confirm that compared with the CIF import price of Db 6,602 per liter inOctober 2006, ENCO sells fuel to EMAE at a price of 37 percent (or Db 2,457 per liter) below the world market price. The figure for June 2006 was fairly similar at Db 2,258 per liter. In 2004 and 2005, when import prices were lower, the "extra loss" was smaller and more on the order of Db 1,500 per liter. 93 At an average exchangerate of about Db 12,500 per USdollar in2006. 67 to 114.6 percent if the tax breaks that EMAE receives are terminated.94 For 2006 we expect the tariffs to be even higher because both subsidies and tax break were higher compared with 2005.95 This minimum tariff increase might not by itself lead to any improvement in service quality, unless incentives for improving performance are introduced, nor would it provide resources for expanding the network, since it would only enable EMAE to break even by covering its current expenses, including depreciation and taxes. 3.35 If in addition to increasing tariffs there was some rebalancing of the rates between different categories o f customers, so that public administration entities would not face the same increase as other categories (because the price paid by public administrations today i s much higher than for the other categories o f customers), the increase to be faced by residential customers would be even larger. Poverty Aspects of Hypothetical Electricity Tariff Increases 3.36 Inthis section, we turn to the issue of the distributionalproperties of existing subsidies and tax breaks for EMAE, and to the potential impact o f a tariff increase on the poor. The analysis i s based on the E C V M (EnquCte sur les Conditions de Vie des MCnages) a survey o f households' living conditions. The data and basic statistics as well as methodology used are provided inAnnex 111.4. 3.37 An increase in tariffs could.have a substantial impact on poverty among households connected to the network and especially for the poorest households with a connection (table 32). Overall, only 42 percent of households have access to electricity. As expected, access to electricity, both at the nei hborhood level and at the household level, i s lower among the poorer deciles o f the population.9 t 3.38 Total expenditure for electricity i s nonnegligible for those with a c ~ n n e c t i o nIn~the first . ~ decile, for example, the average expenditure on electricity for those with a connection i s Db 26,579 per month. This can be compared with total household spending o f Db 352,374 in the lowest decile. Thus, for the poorest households with a connection, spending on electricity absorbs approximately 5.3 percent o f total pend ding.^' The share o f total spending allocated to electricity increases to 7.5 percent in the seventh decile, before receding and reaching to 3.8 percent in the top decile. Overall, these are rather high shares o f total spending allocated to one single 94In this case, the average increase in electricity tariffs that would have been needed in 2005 in order for EMAE to break even without subsidies from ENCO and if taxes were applied to EMAE's diesel consumption would have been 114.6percent. 95We assume that 75 percent o f EMAE's total revenues came from sales o f electricity; 75 percent is the proportion observed in 2005 according to data from EMAE. The excersise can not be duplicated for 2006 because we do not have the operating loss o f EMAE for 2006. 96Each decile accounts for 10 percent o f the population, and the deciles are ranked according to the level o f per capita consumption o f households. Inthe first four deciles o f the population, the average consumption is at or below 100kWh ger month. Consumption reaches 200 kWh per month only inthe top decile. Even though some connections are available in the neighborhoods where households live in close to 85 percent o f cases (this means that only one in two households actually connects to the network where the network i s available in the area where the household lives). Access to electricity, both at the neighborhood level and at the household level, i s lower among the poorer deciles o f the population. Consumption levels are often low. In the first four deciles o f the population, the average consumption i s at or below 100 kWh per month. Consumption reaches 200 kWh per month only inthe top decile. 98This i s for simplicity an approximate share based on the mean values reported intable 32 (a ratio o f means is not the same as the mean o f ratios that would have been computed for each household, but differences are typically small). 68 infrastructure service. Indeed, it i s often assumed that for services such as electricity or water, a bill that represents more than 5 percent of a household's total consumption would be unaffordable. In very poor countries such as SZo Tome and Principe, even the threshold o f 5 percent itself may be considered high, given the cuts that households would have to make elsewhere to allocate resources to electricity consumption. Table 32: DescriptiveStatistics-Access to Electricity, Consumption, and Expenditure 2000-01 Decile ofper Total Household Share of Average kWh Access to Access to capita household expenditurefor spending consumedper electricity electricity expenditure expenditure electricity allocatedto monthper in the at the household (1) (2) electricity household household's level Db Db (2)/(1) neighborhood 1 497,966 26,579 5.3% 44.185 0.551 0.074 2 700,701 36,418 5.2% 56.936 0.735 0.214 3 927,652 61,033 6.6% 87.093 0.771 0.273 4 1,152,034 79,510 6.9% 106.8 0.857 0.374 5 1,304,874 86,784 6.7% 115.579 0.880 0.360 6 1,357,096 90,931 6.7% 121.337 0.924 0.520 7 1 319,432 113,962 7.5% 138.873 0.914 0.528 8 1,972,822 130,057 6.6% 158.817 0.920 0.580 9 1,996,876 124,417 6.2% 151.08 0.947 0.624 10 5,209,4 12 196,b59 3.8% 209.662 0.971 0.652 All 2,056,657 113,442 5.5% 138.194 0.847 0.420 3.39 We consider three potential tariff increases that could have been applied proportionately to all customers in 2005 to enable EMAEto break even): 41 percent (to cover EMAE's operating losses in 2005), 73 percent (to cover operating losses and the implicit subsidies from ENCO), and 115percent (to add the taxes paidby end users on diesel fuels). Share of householdsinPoverty Shareof PopulationinPoverty Households Households Full sample with electricity Fullsample with electricity Baseline 45.5% 25.6% 53.1% 33.7% Increase intariff of 41% 46.2% 27.4% 54.1% 35.8% Increase intariff of 73% 46.9% 29.1% 54.8% 37.3% Increase intariff of 115% 48.4% 32.5% 56.5% 40.9% 3.40 The increase in poverty that would result from a large increase in electricity tariffs i s relatively large across the board and may not be a socially acceptable option in the short to medium term. For example, at the national level, the share o f the households inpoverty increases by 1.4percentage points when the tariff increase is on the order o f 73 percent. Among households connected to the network, the increase i s larger, at 3.5 percentage points, but from a lower base given that households connected to the network tend to be less poor than households without access. The relatively large impact on poverty stems from the fact that the connection rate to electricity i s relatively highby African standards (approximately US$l50), and that the share o f total consumption allocated to electricity i s also relatively high, at above 5 percent as discussed 69 earlier. Adopting well-targeted subsidies on electricity consumption rather than connection would have a better poverty outcome for the country (see Annex 111.1for further analysis). 3.41 For poverty reduction, the authorities should investigate whether there are better outlays for the funds used for the implicit subsidies provided to EMAE for electricity. As a simple comparison, we note that the total cost to the authorities o f the shortfall in electricity revenues i s estimated to be as large as US$5.1 million.99 This amount compares to a budget for recurrent expenditure for education o fUS$7.1 million, and recurrent spending for health o f US$2.7 million in2005. 3.42 Usingthe data from the 2000-01 survey, we can assess which households use a range o f other services subsidized by the government and thereby compute the share o f subsidies (public transfers) received by poor households divided by the share o f the poor in the population (the value o f ~2intable 34). We assume that the unit cost o f provision for various types o fbenefits in- kind (such as education received by children inpublic schools or the usage o f publicly provided health care) i s constant for all those who benefit from these transfers.'" By comparison, the share of the implicit subsidies inthe tariff structure receivedbypoor households divided by the share o f the poor in the population (R) took a value o f 0.486 in 2000 under the existing volume- differentiated tariff (VDT) structure and with the lifeline at 100 kwh for the first bracket o f consumption(see Annex I11for detailed data and methodology)."' 3.43 The benefits inkind (primary education, secondary education, and health centers) appear to be better targeted to the poor than electricity subsidies, since the values o f R are larger for these transfers than for electricity subsidies (table 34). The difference intargeting performance i s largest for primary schooling, followed by health services taken as a whole. It can also be seen that the public (as well as private) cash transfers that can be identified in the survey are also on average better targeted than electricity subsidies. Table 34: Estimatesof the Value of Rfor Publicand PrivateTransfers,2001 I Share of benefitsfor the poor Estimateof Primary school 0.57 1.082 Secondary school 0.29 0.541 Public transfers 0.31 0.593 Private transfers 0.40 0.759 Healthcenters 0.40 0.754 Perfectly targeted transfer 1.oo 1.887 99Subsidies through losses by EMAE will have to be repaid later, or through implicit subsidies from ENCO and tax breaks. looThis assumptionmay overestimatethe benefitsof public transfers for poorerhouseholds if those households live in poorer areas that benefit from lower-quality and lower-cost services. Still, the analysis that is permitted by this assumptionis usefulfor comparisonpurposeand inthe absence ofmoredetaileddata. lo' For the simulations, we use the tariff structure that prevailed at the time, namely the December 2000 tariffs presented in Annex I11tables. Ifan invertedblock tariff structure had been appliedinstead of a volume-differentiated tariff (VDT), Rwould have taken avalue of0.417. The value ofRdoes not change significantlyifthe lifelinevalue for the first bracketof consumption is modified. 70 The above analysis suggests two complementarypolicyoptions: i. Newinvestmentsintheproductionanddistributionofelectricitywouldhelpdecrease the cost o f production, alleviating EMAE's financial difficulties and reducing the need for potentially large tariff increases. The government i s seeking private investments inelectricity productionto stabilize supply inthe short runand to expand coverage inthe medium to long run(see Annex 111.1). ii. Thegovernmentmaywishtoreviewthetargetingandeffectivenessofthesubsidies in its current pro-poor programs in reaching desired populations and reducing poverty. It should also analyze policy choices in providing implicit subsidies to EMAEcompared with fulfilling other social goals (such as provision o f social tariff rates for electricity, budgeted payments for electricity connection, education, and health and other social protection services). 5. THENEED FORA STRENGTHENED REGULATORYFRAMEWORK FORTHE INFRASTRUCTURE SOES'~~ 3.44 Inview of the government's interest inattracting private sector investment inelectricity and water production and in port and airport development, a strengthened regulatory framework i s o f crucial importance. 3-45 The current regulatory authority was created inJuly 2006, as a follow-up to the July 2004 promulgation o f the telecoms regulatory law and the August 2005 adoption o f the decree law defining the nature, structure, role, and responsibilities o f the General Regulatory Authority o f the Democratic Republic o f Sgo Tom6 and Principe. The authority i s a multisectoral body that regulates the telecoms,water, and electricity sectors. It has a minimal expert staff, trained over the past few years by international partners, including the World Bank and Portugal.'03 More hiring i s planned in 2007, with hrther need for training. As o f late 2006, the legal, institutional, and operational framework finally allowed the authority to start its regulatory work, starting with issuing two telecoms licenses in January 2007. It i s also working on a plan to better plan the distribution o f telephone numbers at the national level plan and a law to regulate radio frequency. The authorities have yet to start the development o f the regulatory framework for the water and electricity sector. 3.46 A number o f alternative regulatory frameworks exist but in the case of SHo Tom6 and Principe the choice has been already made: a multisector "umbrella" regulator i s one o f the best options, despite potential implementationchallenges. We define multisector regulation as a single regulatory agency that has responsibility for sectors such as telecom, energy, water, and transportation. With further economic growth, the need for more specialized regulation may arise, but this first stage allows STP to develop not only the industries but also the expertise needed for more focused regulation. Several countries-such as Botswana, Senegal, Mauritania, Niger, Jamaica, Costa Rica, Panama, and Latvia-have chosen to have a multisector regulatory agency rather than specialized bodies. Opting for multisector agencies seems particularly adequate for small economies in which the lack o f shlled staff and financial resources constitute a binding constraint and in which the stage o f development o f industries i s such that expanding production lo' BasedonVagliasindi(2006a, 2006b). lo3 Currently, the authorityhas a president, alegalexpert, an economdfiscal expert, and an engineer/computerexpert, along with some support staff. 71 and investment i s more an issue than promoting competition or dealing with sophisticated access issues. Opting for multisector agencies also skirts the duplication costs associated with setting up several sector-specific agencies. 3.47 Nonetheless, it might be hard to meet the criteria required by effective regulation and more importantly to establish and maintain credible infrastructure regulatory agencies for a small low-income country for a number o f reasons.lo4 The first, and most obvious, reason i s the shortage o f qualified regulatory staff. Since regulation depends on the separation o f regulatory and policy powers, this staff shortage can make the establishment o f separate regulatory agencies a notional rather than a genuine exercise. Second, the lack o f reliable law courts and/or the capacity o f the commercial law courts can also be a major impediment to independent regulation. Finally, the political economy and the "institutional endowment" o f the country might not be consistent with the development o f commercialized utilities and separate regulatory institutions. This final argument is particularly strong when infrastructure utilities have not yet been commercialized, so that the separation o f ownership from regulation might not be a genuine one, unless the government first undertakes serious governance reforms o f the state-owned enterprises (see Annex 111.3). 3.48 While the multisector regulator i s the best option for STP, the authorities will need to be aware o f positive as well as negative aspects o f a multisector regulator. These are summarized in the table below. Box 3: The Positives and Negatives of Multisector Regulation (MSR) Positive aspects Negative aspects Q Economiesof Scaleand Scope:MSR increases the 0 Economiesof ScaleandScope: MSRcreates economies of scale and scope inthe development greater complexity inestablishing the legal of the regulatory agency and provides more framework, potential delays inthe reform process, effective means of dealing with converging sectors and difficulties inmerging the agency and and bundledprovision o f services. achieving consensus. 0 RegulatoryCapture:MSRreduces the riskof 9 RegulatoryCapture:MSRincreasestheriskof industry and political capture by one sector, capture by a dominant incumbent because of multisector responsibility. operator/ministry. 9 RegulatoryConsistencyacrossSectors:MSR 0 RegulatoryConsistencyacrossSectors:Failure encourages transfers ofregulatory know-how by the regulator in one sector provokes cascades to between regulators responsible for different other sectors and dilution o f sector-specific infrastructure sectors. technical expertise. 3.49 To strengthen the regulatory framework of the existing regulatory body, the government should consider the following actions: 0 Speed up the full operationalization o f the regulatory authority to remove the conflict o f interest between operation and regulation. The introduction o f incentive-based regulatory rules to improve SOE performance i s advisable. IO4There is now an extensive literature o f the key criteria, including independence, accountability, and transparency. For a thorough review o f these criteria see Brown et al. (2006) and Stern and Holder (1999). For empirical evidence see Vagliasindi (2004,2006) . 72 Encourage the informal exchanges o f information and pooling o f resources between regulators o f different countries. This practice seems to be developing in a very interesting and potentially productive way in Southern Africa for both telecommunications and electricity. It i s also the route being followed in the European Union. A particular advantage i s that informal pooling can be market driven, both responding to and helping encourage trade, integrating markets and networks, and increasing the scope for competition. 0 Develop a decree law to operationalize the regulation o f EMAE's power purchasing agreements (discussed below), accompanied by appropriate capacity building for sectoral supervision. Inthe medium to long term, consider extending regulatory oversight over ENASA and ENAPORT, especially because the government i s interested in private sector investment inbothport and airport subsectors. 6. RECOMMENDATIONSFORTHE SHORT TERM(0-12 MONTHS) Measure the full direct and indirect subsidies between the government and the SOEs to fully understandthe financialstatusof the SOEs andthe exposureof the centralbudget Measure the sum o f any direct (cash equity injections from the government, donor grants, and asset transfers) and indirect (equity support, favorable financing terns, and favorable fiscal terms such as exemptions from import duties and others) subsidies to the SOEs. Have a clear assessment o f the contingent liabilities that the government i s accepting on behalf o f EMAE if the entreprise later enters into a contract with an independent power producer for electricity production. Adopt a financial policy in terms o f setting funds aside to cover the financial exposure. Adopt a standardized fiscal template (see Annex 111.7). Assess the policy choices regarding the implicit subsidies to EMAE (diesel prices and tax breaks) versus developingtargeted pro-poor social programs. 0 Undertake a full assessment o f the efficiency and targeting o f current pro- poor programs. 0 Analyze policy choices in providing implicit subsidies to EMAE versus social goals. Address the financial difficulties o f the SOEs by undertaking a medium-term reform program that would include the following for the short (12-month horizon): SettleAccumulatedArrears andEstablishPaymentDiscipline 0 The infrastructure enterprises need to settle all their accumulated arrears with the government as well as their inter-arrears with one another. The clearing o f arrears and inter-arrears i s the first step to improve the enterprises' cash flow. Once the government clears and settles all its payment arrears with EMAE, ENASA, 73 ENAPORT, and CORREIOS, the enterprises in turn would be able to clear the supplier, tax, and all other arrears with one another. 0 The government can devise it own methodology o f settling arrears, but it needs to establish a strict payment discipline using international experience. In the early 199Os, for example, Senegal and Ghana used the following steps to clear the SOEs' arrears with the state as well s the inter-arrears among the SOEs: First, the accumulated arrears o f the state and the inter-arrears among the enterprises were calculated. Second, all parties concerned were given an opportunity to agree or disagree with the amounts declared. Third, net arrears and inter-arrears were agreed upon. Fourth, a payment schedule was determined. Fifth, a contract was drafted for every paying party to adhere to ina given time period. 0 The payment discipline should be extended and communicated to the private sector (domestic and international) as well as to the international public or semipublic entities operating in STlo Tome and Principe 0 Adopt as a policy that no government agency, entity, or ministrywould be allowed to accumulate arrears in the future. One measure that was tried in Francophone countries inthe 1990swas the introduction o f estimated amounts for expected use o f service by each ministry into their recurrent expenditures. The estimated amounts at least allowed each ministry or government entity to reserve funding for its use o f services from the infrastructure enterprises. Even though the ministries were conservative in their estimates, they could at least pay some o f their bills from the programmed expenditures and avoid accumulating large arrears. Consider Establishinga State-OwnedEnterpriseUnit 0 The government does not have a policy o f state-owned enterprise management, or any mechanism that i s capable and reliable for the task o f managing the performance o f SOEs (their operational and investment expenditures), as well as the relations between the government and SOEs, in a systematic and coordinated ma~mer."~ Without comprehensive, easily accessible information about the SOEs, a development strategy for each, and well-qualified personnel acting as advocates for the state-owned infrastructure sector, each successive government has faced large financial losses and inadequate service. Each has been forced to take reactive, piecemeal, ad hoc actions. The government i s interested in opening the sector to private investment. It will need an entity that could systematically study and evaluate the options and proposals that the authorities face. The SOE Unitshould 0 not be intended as an additional layer o f supervision but as an entity workmg closely with the infrastructure enterprises, the MoPF, and the "tutelle" ministries. It also needs to act as an intermediary regarding the SOEs' operational and investment lo' This recommendationis in line with the absence of any monitoringand evaluationsystem and procedures for the government'spublicinvestmentprogram. 74 needs, aligning them with those o f SHo Tome and Principe and preparing master plans for each infrastructure sector and the enterprise; have the capacity and expertise to take full stock o f SOEs' financial positions (accompanied by private forensic audits): SOEs' financial reports need to be reviewed indetail; help clarify, settle, and monitor all accumulated arrears with the government and inter-arrears among SOEs and enforce a strict payment discipline (see recommendationabove). 7. RECOMMENDATIONSFORTHE MEDIUM TERM(THREE-YEAR HORIZON) Corporatize and Commercialize the SOEs The four SOEs, especially EMAE, are interested in corporatization. The proposed institutional reform has to be designed to strengthen the enterprises' autonomy as a distinct legal identity; the reform should separate the SOEs from their "technical tutelle" ministry and their "financial tutelle" ministry (Ministry o f Planning and Finance). Ifagreed, all SOEs would be incorporated under the same laws that govern private corporations. 0 The nascent regulatory body o f the country will need to be strengthened and, if necessary, the coverage o f the regulatory law adjusted to ensure proper supervision of all the infrastructure corporations. Currently, the law includes telecoms, water, and electricity. Corporathation does the following: 0 Provides SOEs with company status, clarifies their shareholding, and gives them more managerial independence. Their reporting relationship to their shareholder (government) would be in terms o f a shareholder agreement and performance contract, and not interms o f day-to-day management decisions. Promises to free the concerned SOEs from the long tradition o f the cabinet delegating absolute control to the "tutelle" ministry. This freedom would also require the introduction a robust performance evaluation to provide incentives or delegate autonomy to SOEs' managers. 0 Encourages commercialization, which means that the government needs to either allow tariffs to cover full costs, or fill the gap between the cost o f service and the tariff revenue with explicit and transparent subsidies. 0 Improves corporate governance rules and standards. These improvements, which can help considerably to reduce political interference, include the introduction o f (i)a structured and skilled nomination process (by establishing a database o f qualified candidates) for boards o f directors; (ii)a training and induction process on board responsibility to enhance members' professionalism; and (iii) independent directors. Consider AdoptingIncentive-BasedPerformanceContracts 0 The four SOEs could benefit from incentive-based performance contracts, but careful analysis and development o f the contracts is critical to their success. For more information on these contracts, see Annex 111.7. 75 0 The overriding principle for preparing performance contracts remains that they are based on simplicity so that their content can be easy to monitor and to evaluate. 0 The performance contract system must be coupled with long-term planning and a reform program (namely, the PRSP) consistent with the priorities established by the government through sector ministries. Expand the Roleofthe SOE Unit 3.50 The role o f the SOE Unit could be expanded to take on additional responsibilities. These responsibilities could include one or more o fthe following activities. 0 Inconsultation with all stakeholders, develop andproposereformsthat will improve financial relations between the government and the enterprises based on assessment results. The infrastructure enterprises have been operating within a system o f explicit and implicit subsidies; once defined, these subsidies can be lifted to allow the enterprises to operate in a commercial manner combined with cost-recovery tariffs (withdirect subsidies for services providedto low-income consumers). 0 Review, in consultation with all stakeholders, options for better price setting in the absence o f competition. Presently, the tariff setting i s the responsibility o f the relevant "tutelle" ministry. However, the subsidy structure between block tariffs needs to be reviewed, especially inthe case o f electricity and water and, ifnecessary, ithas to beredesigned. 0 Receive the unsolicited proposals that STP i s often subject to and secure technical expertise to review them; prepare competitive international tenders when necessary; assist in selecting the right private operator; and help the regulatory body to monitor the operator's performance and compliance with the contractual terms. 0 In consultation with all stakeholders, help to plan for how to increase service coverage, quality, and affordability, especially for electricity and water. Review, in consultation with all stakeholders, options for private sector participation. These options could range from simple contracting arrangements with the private sector for the design, building, financing, and operation o f greenfield facilities; rehabilitation, operation, and transfer o f existing facilities (EMAE); concession contracts for the container terminal (ENAPORT); or management contracts (EMAE and ENASA) . Help the existing regulatory body o f the country to design, implement, and monitor incentive-based performance contracts between the government and the SOEs. Prepare, in consultation with all stakeholders, the master plans for infrastructure sectors and their respective SOEs. All infrastructure enterprises complain about the lack o f strategic master plans articulating their respective sector's needs in the long run and their role fulfilling those needs. For example, EMAE and its "tutelle" ministry needs to explore alternative primary sources of energy for electricity production, including hydropower sources; the government also needs to develop a long-term energy strategy by working closely with the Ministry o f Natural Resources 76 and Environment, the MoPF, and the international donor community.'06 The particular needs o f the island o f Principe also have to be taken into full consideration while developing an energy strategy for the whole country.'o7 Assess net subsidies and potential financial savings from reforms. The SOE Unit can create a database to calculate the net subsidies from the government to the infrastructure SOEs and from the SOEs to the government; direct subsidies in the form o f cash equity injections, donor grants, and asset transfers; and indirect subsidies in the form o f equity support, favorable financing terms, and favorable fiscal terms from the government to the enterprises. Analyze the Possibilityof SeparatingWater fromElectricity Functions On the positive side, EMAEmight consider reviewing the social and economic costs and benefits o f separating the water and electricity supply while creating two separate enterprises out o f EMAE. The provision o f water supply could be converted into a Water and Sanitation Authority, taking advantage o f the the international community's priority for the water sector so that it can have direct access to the HIPC Fund. Once established, the Water Authority could develop a strategy to expand the limited coverage and quality o f the country's piped water.lo8 The Water Authority might also include the sanitation facilities, and it can develop its long-term strategy by working with the Ministry o f Natural Resources and Environment and the MinistryofHealth. The PhnomPenhWater Supply Authority is anexample. In 1998, itbecame a financially viable public enterpriseunder commerciallaw and at the same time established a revolving fund to finance domestic connections to help the poor to connect to the network. By 2004, the water utility's connections increased almost ninefold. Onthe negative side, since STP is a small island, creating another separate institution may further burdenthe human and financial capacity o f the country. Train Staff and Build Capacity The enterprise managers and engineershechnicians can all benefit from training or from having an ongoing "twinning arrangement" with another state-owned utility in the region or in Portugal or Brazil. There are only a few examples o f such twinning arrangements, since they have been studied at great length and benefits have not been determined. SZo Tom6 already has a limited internal capacity not only inall layers o f government but also within its infrastructure enterprises. Even if corporatized, without private equity partners these enterprises would still need international expertise and management know-how to improve all levels o f efficiency and profitability. The international donor community might be willing to fund such The energy strategy has to outline plans for expanding energy access in the short term and how to increase energy capacity in order to foster economic growth in the medium to long term. It also should consider ways to increase energy security and decrease dependence on imported fuels, ensure energy efficiency, decrease the cost o f electricity and fuels, assess technological change, and consider local environmental effects and ways to protect environmental assets. The objectives o f the master plans should also be inline with the policy options laid down inthe PRSP. lo'Sa'o Tome`and Principe Advisory Project: Report on Energy Infrastructure, Earth Institute, Columbia University, 2004. lo*Debt Relief International specifically talks about the government's access to the HIPC Fund for all investments related to water supply and sanitation or for maintenance expenses. 77 training programs, particularly if linked with twinning arrangements. For example, the "operations and maintenance" contracts often provided by the international electricity energy companies have managed to transfer know-how to the utilities in developing countries. The training component either was part o f the original operations and maintenance contract or was paid for by the bilateral German development agency to the German company, STEAG Inc. Analyze and discuss the options to facilitate public infrastructure investments. One possibility i s excluding public infrastructure investments from the fiscal targets that developing countries must meet to qualify for financial assistance.109,110 Another option i s to allow a larger percentage o f primary deficit with a clear commitment that the funds are used indomestically financed investment for EMAE.Inanalyzingthese suggestions, their impacts on the fiscal and financial health o f the budget and the economy, the economic viability o f the investment project in the long run, and the sustainability o f the country's public debt needto be taken into account. 0 Develop its project evaluation and management capacity to prepare and propose priority public investment projects in line with infrastructure sector strategies (which were available as o f June 2006). Recommendations: Short Term Issues to address Action to take Accumulated arrears need to be settled and Calculate, schedule, and clear arrears and payments payment discipline adopted. among SOEs and between the SOEs and the government. Clear arrears with private companies and international entities. Adopt a policy to forbid buildup o f arrears within public entities. Include estimated amounts for each ministry's expected use o f services in their annual recurrent budgets. Adopt a standardized fiscal template (see Annex 111.7). A State-Owned Enterprise Unit should be The SOE Unit: established. Should not be intended as an additional layer o f supervision but as an entity working closely with the infrastructure enterprises, the MoPF, and the "tutelle" ministries. Should have the capacity to take full stock o f SOEs' financial positions (accompanied by private forensic audits). Should help to clarify, settle, and monitor all accumulated arrears with the government and inter- arrears among SOEs and to enforce a strict payment discipline. Fullassessment ofdirect andindirect subsidies Help the MoPF to measure the sum of any direct (cash from the government to the SOEs and equity injections from the government, donor grants, government contingent liabilities does not exist. and asset transfers) and indirect (equity support, favorable financing terms, and favorable fiscal terms Hemming and Ter-Minassian (2004, 30). `lo All infrastructure SOEmanagersrepeatedlyexpressedthe needfor investments. 78 Short Term Issuesto address Action to take such as exemptions fiom import duties and others) subsidies to the SOEs. Assess the contingent liability related to EMAE's signing a contract with an independent power producer for electricity production. Adopt a financial policy to set finds aside to cover the central budget financial exposure. Policy choices regarding the implicit subsidies to Review efficiency and targeting o f current pro-poor EMAE and targeting o f pro-poor expenditures programs. and social programs should be analyzed indepth. Analyze policy choices inproviding implicit subsidies to EMAEcompared to social goals. MediumTerm hree-year horizon) Issuesto address Action to take The SOEs should be corporatized and Develop sector-specific strategies. commercialized. The law should be developed and approved by authorities. The nascent regulatory body o f the country will need to be strengthened and if necessary, the coverage o f the regulatory law adjusted to ensure proper supervision o f all the infrastructure corporations. Consider adopting an incentive-based Use performance contracts that are simple so that their performance contract. content can be easy to monitor and to evaluate. Combine the performance contract system with long- term planning and a reformprogram consistent with the priorities established by the government through sector ministries and the PRSP. Ifa well-trained andcapable SOEUnitis is inplace, it can monitor performace contracts as a provisional solution to improve the administration and effectiveness o f its SOEs. Train staff and buildcapacity. Adopt "twinning arrangements" with another state- owned utilities in the region or in Portugal or Brazil for training purposes. The regulatory body would also benefit from such training. Consider the role o f the SOE Unit and expanding A number o f activities are suggested in the text. The itto assume more responsibility. new responsibilities will depend on the decision o f the authorities. Consider making room for public infrastructure Analyze and discuss with international partners ideas investments . to facilitate investments in infrastructure. One possibility i s to exclude public infrastructure investments fiom fiscal targets; another is to dedicate an extra percentage o f domestically financed investment for EMAE infrastructure (production and distribution). Analyse the possibility o f separating water from Have international experts conduct an analysis. electricity functions. 79 BIBLIOGRAPHY Brown, Ashley; Jon Stem; Bernard Tenenbaum and Defne Gencer (2006) Handbook for Evaluating Infrastructure Regulatory Systems, World Bank, Washington, DC. Dionizio, Vitor. August 2006. Background chapter on revenues and expenditure patterns. "Public Expenditure Review". World Bank. Dinh, H. T., A. Adugna, and B. Myers. November 2002. "The Impact of Cash Budgets on Poverty Reduction in Zambia." Africa Region Working Paper Series Number 39, World Bank, Washington, DC. Eifert, B., and A. Gelb. 2005. "Improving the Dynamics of Aid, Towards a More Predictable Budget Support." WPS #3732, World Bank, Washington, DC. Ibudegwu, C. April 2005. "The Medium-TermExpenditure Framework." Africa RegionWorking Paper SeriesNumber 82. InternationalMonetary Fund. 2007. Staff report for the 2006 Article IV consultation, thirdreview under the three-year arrangement under the PovertyReductionand GrowthFacility, Washington, DC. Gokgiir, N. July 2001. "Privatizing Infrastructure in Egypt." Report prepared for USAID MissiodEgyptand Carana Corporation. Gogkiir, N. August 2006. Background chapter on state-owned enterprises. "State Owned Infrastructure Entreprises". World Bank. Hemming, R., and T. Ter-Minassian. December 2004. "Possible New Approaches to Fiscal Accounting." Finance and Development. Stem, Jon and Stuart Holder (1999) "Regulatory Governance: Criteria for Assessing the Performance of Regulatory Systems. An Application to Infrastructure Industries in the DevelopingCountries of Asia", Utilities Policy, 8 (l), 35-50. UnitedNations. 1995. Guidelines on Performance Contracting: A Practitioner's Manual. United Nations Department for Development Support and Management Services. Vagliasindi, M. 2006a. "Infrastructure Regulatory Issues for Small and Low-Income Developing Countries." Backgroundpolicy notes, World Bank, Washington, DC. Vagliasindi, M. 2006b. "Performance Evaluation and Monitoring for Infrastructure Service Providers." World Bank, Washington, DC. Wodon, et al. January 2007. Electricy Tariffs and the Poor in Sgo Tom6 and Principe. Draft, World Bank. World Bank. 2006. "Public Expenditure Management and Financial Accountability in Niger." Washington, DC. World Bank. November 2003. "Zambia Public Expenditure Management and Financial Accountability Review." Washington, DC. 80 LIST OFANNEXES Annex I.: 1 Sao Tome & PrincipeMacroeconomic Assumptions 2005-1015 Annex 1.2: Framework o f the RMSM-X Poverty Module Annex 11.1: Tax Hindrances to setting up Business Annex 11.2: An Alternative Income Tax Proposal for the Upcoming Income Tax Reform Annex 11.3: Tax Institutions, Tax Structure as o f June 2006 Annex 11.4: Local Administration Annex 11.5: Information Constraints inthe Management o f Public Investment Annex 11.6: Data on sources o f external aid and PIP financing Annex 111.1: EMAE'sPerformance Annex 111.2: Performance o f ENASA, ENAPORT and CORREIO S Annex 111.3 EMAE-Tables Annex 111.4: Poverty Aspects o f Electricity Tariffs Increases: Data and Methodology Annex 111.5: Assessing the Targeting Performance o f Consumption Subsidies: A Framework Annex 111.6: Incentive-Based Performance Contracts for SOEs Annex 111.7: Standardized fiscal template Annex IV: PERTables 81 ANNEX 1.1: SAOTOME& PRINCIPE MACROECONOMIC ASSUMPTIONS2005-2015 IfSTPstartsexportingoil, GDPwillincreasebeyondtheabsorptioncapacityofthe country. To minimize "Dutch disease" most o f oil income will not be readily available to the country but rather deposited in an Oil Fund and disbursed based on the fiscal rule set at the ORML and the needs o f the country. Under the rules o f the oil fund, the government will withdraw progressively according to the absorptive capacity o f the economy and allocate the drawings to development objectives, especially social pro-poor goals. However, the commercial production, if any i s not expected before 2012. Inthe mean time, the country's sole oil related revenues are the oil bonuses paid on concession rights. So far, only the contract signed on block 1 has been disbursed in 2005, amounting to 49.2 million. Blocks 2 to 4 were awarded and STP share amounts 26.6 million US$ that, after repaying $ U S 15 Mto Nigeria for a credit to finance the JDA budget will be disbursed to the country in2007. Inorderto getabetterideaofthe actualimpactontheeconomy ofthe oilerawe setupa model o f the domestic economy. It does not consider expenditures or revenues from foreign companies directly onto the oil sector, without local production content, or oil revenues deposited inthe oil fund. More precisely we definemacroeconomic variables as Where I* i s the non-oil sector related investment and I"s foreign investment directly i channeled to the oil sector (for set up o f the oil industry) undertaken by international companies inforeign currency. Where M" are imports for the oil sector, undertaken directly by the oil companies in foreign currency. Therefore only M*,the imports that are incorporated inthe domestic economy, are accounted. We can assume that any import that does not imply a currency exchange and i s directedto the oil sector i s not accounted. Where X* are exports o f non oil products and X" are exports o f oil done directly by oil companies to international markets in dollars which do not revert directly to the country but are rather deposited into the oil fund. Even though the oil fund i s compounded by foreign reserves they are not accounted in our definition o f reserves as they are not readily usable by the monetary authority. The only foreign currencies to sterilize are the drawings from the oil fund. * G+Drawings from Oil Fund = G* is current expenditure that takes into account yearly drawings from the oil hnd for the government budget. Accordingly the value o f the oil fund i s not accrued to the GDP which can be expressed as: Y* = C G* I*X* - M* + + + This Non oil values are calculated assuming the following assumptions for the oil variables previously described: 82 Table 1: Assumptionson OilMacroeconomicVariables (CurrentPrices,BillionDobras) Assumptions on Oil Macroeconomic variables, Current prices, Billion Dobras 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Oil Investment, I" 0 187 212 262 310 365 4,569 4,403 1,541 4,574 1,609 3,209 Exportso f Oil, X' 0 0 0 0 0 0 0 3,846 23,927 31,011 32,414 28,415 Imports for the Oil Sector, M" -- Services 0 187 212 262 310 365 3,641 4,003 1,973 4,630 2,095 3,371 Goods 0 131 148 183 217 255 2,165 2,478 953 2,659 1,000 1,879 0 56 63 79 93 109 1,476 1,525 1,020 1,972 1,095 1,491 Oil fund outlays 279 198 352 124 108 93 78 64 209 478 1,190 1,207 - FromBonussignatures 279 198 352 124 108 93 78 64 51 41 32 26 ~ FromOil Production 0 0 0 0 0 0 0 0 158 437 1,158 1,181 Source: IMF 83 9\49 N W F o ! Y f 9 \ 4 ? \ 4 m w m m w - omomm- - :?g m w m Y t Y 9 ' 4 o q Y m - - 0 o m - N m m - - m w w ANNEX 1.2: FRAMEWORKOF THE M S M - X POVERTY MODULE The poverty analysis inthis document i s based on the RMSM-X+P fkamework developed at the World Bank."' The RMSM-X+P consists o f a RMSM-X macroeconomic model with additional sub models for poverty and social indicators. The linkages among the models facilitate the analysis o f the impact o f various macroeconomic shocks on a selected set o f key social indicators. Poverty analysis i s performed using a poverty equation (which i s estimated using pooled data for a group o f low-income countries) that links the incidence o f poverty to inflation, the literacy rate, real GDP per capita, the degree o f trade openness and income inequality. The effects o f various macroeconomic shocks on education and health are analyzed with the help of separate equations for education and health. This annex presents the poverty model followed by the equations for education and health sectors. The PovertyEquation A number o f macroeconomic and structural factors are assumed to influence the incidence o f poverty (measured by the headcount index) in an economy. Among these factors, increases in the adult literacy rate, the level o f per capita real GDP, the level o f openness to international trade are postulated to reduce poverty while increases in the inflation rate and the degree o f income inequality are assumed to increase poverty levels. Equation 1 below shows the poverty incidence equation while the parameters used for the analysis are shown inthe table at the end ofthis annex: (1) Poverty Incidenceit = PO+PI Inflationit +p2 Literacy Rateit+ p3 logker capita realGDPit) +p4 Opennesit +p5 Giniit The EducationEquation. The adult literacy rate i s taken to be a proxy for the general level o f education in a country. As inthe case o f the incidence o f poverty, the level o f general education in a country i s influenced by macroeconomic and structural factors. The factors that have been explicitly identified in the current analysis are the level o f per capita real GDP, the urban share o f total population (i.e urbanizationrate) and per capita public expenditure on education. The education equation i s given below. (2) log(Literacy RateiJ = + Po PI log(Averageper capita Public Education ExpenditureiJ + pzlogber capita real GDPiJ +/& log(Urbanization RateiJ The HealthEquation. Infant mortality rate i s postulated taken as a measure o f the general level o f health o f the population o f a country. Infant mortality i s affected by per capita real GDP, per capita public health expenditure and the adult literacyrate as shown below: (3) log(Infant Mortality RateiJ = Po+PIlogber capita Public Health ExpenditureiJ+ PZ logker capita real GDPiJ +p3log(Literacy RateiJ `I1See, Derek H.C. Chen, Thilak Ranaweeraand Andrei Storozhuk. The RMSM-X+P: A Minimal PovertyModule for the RMSM-X. World BankPolicyResearchWorking PaperSeries, No 3304, May 2004 85 Parameter Estimates(t-ratiosinparenthesis) The parameters ineach of the equations shown above were estimated with pooled cross- country time series regressions based on historical data from low-income countries. Estimates of the poverty equation are based on pooled cross-section of 52 observations including 24 low- income countries for the 1980-98 period. The sample data for the education regressions included 229 observations for 32 countries covering the period 1983-98. The healthregression i s based on 174 observations for 43 countries covering the period 1990-99.'l2 Estimated equations as shown inChen et. a1(2004) are reproduced below: (I) Poverty Incidence = 49.6446 + 0.1883Inflation 0.3176Literacy - Rate -2.544 logbercapitarealGDP)-0.7585Openness+0.0046OpennessSquared (2.8894) (3.1373) (-4.8665) (-0.9524) (-5.2597) (5.5421) + 0.9534 Gini113 (9.6837) R2adj.=0.9276 ; SE=15.6843 ; F-stal~109.9705 (2) Log of Literacy Rate = 0.0229log (Ave.per cap Edu Exp current +previous 3 periods) (3,9909) + 0.088 logher capita real GDP) + 0.8455log (Urbanization Rate) (2.0641) (I4.5031) R2adj .=0.9917 SE=0.0414 F-stat=13650 (3) Log of Infant Mortality Rate = -0.0218 log (per cap Health Expenditure) (-3.7675) -0.1I31 logbercapitarealGDP)-0.6004log(Literacy Rate) (-I.9545) (-6.9598) R2adj.=0.9876 ; SE=0.0694 ; F-staF6939.2 Simulations Poverty simulations reported in chapter one were conducted talung the estimated parameters from the above equations as given. The only exception was the estimate of the elasticity of poverty with respect to GDP per capita. The estimated elasticity was modified to take into account the specific situation of Sao Tome. The modification of the elasticity of the poverty index with respect to growth is based on estimates of this elasticity as shown in a number of international studies.'14 In particular, a recent study by Besley and Burgess (2003) has estimated an elasticity of 0.75 (with a standard "* For more details, please see Chen et. a1(2004) The growthelasticitywas modifiedto fit Sao Tome experience. See, for example, RavallionM. (2001), Growth, Inequalityand Poverty: Looking BeyondAverages, World Development, Vo1.29, pp. 1803-15;BesleyandBurgess(2003). Halving Global Poverty, Journal o f Economic Perspectives,Vo1.7, pp. 3-22. 86 deviation o f 0.25) using data for less developed countries and 0.49 for sub-Saharan Africa (witha standard deviation o f 0.23). Ravallion (2001) estimated the poverty elasticity to be around 2.0 ina sample o f developing countries. It would appear that a range o f 0.2 to 1.2 may be applicable to most developing countries. Higher elasticities are reported in the literature, especially when expenditure data are usedrather than GDP per capita data as the explanatory variable. The graph below shows the sensitivity o f the elasticity o f poverty with respect to GDP growth. The poverty impacts were calculated for four different parameter estimates for the elasticity, holding all the other parameters constant. The results show relatively robust poverty indices notwithstanding changes in the elasticity. Only a high elasticity yields a relatively large impact on poverty. Sao Tome: Sensitivity Analysis of Poverty Elasticity I --e E=0.25 -4- E=0.5 -f- E=2.0 +e=1.0 I -8 53.0 58 CI 52.0 - 51.0 Q) & 50.0 49.0 1 I 1 I , I I I I I I I 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Years 87 ANNEX 11.1: TAX HINDRANCESTO SETTINGUPBUSINESS There are a large number o f inefficient taxes that hinders formality and business set up. Companies in STP spend 1008 hours per year to prepare, file and pay the corporate income tax compared to the average 394 hours in Sub-Saharan Africa. Not only time to comply tax i s long but the time requiredto open a business is around 192 days compared to 64 days inSub-Saharan Africa (Doing Business 2006). Some taxes that have a negative effect on doing business are: Highcorporate tax rates: discussed above inbox 2. MinimumTax: Companies are supposed to file a "Minimum Tax" during the first semester o f the year even though it has limited collection capacity (between 2$ and 61$ at most per company, deductible when filing final tax records). Urban Property Tax: It charges 15 percent tax rate on any urban property but have several correction factors depending on the year o f property registrationthat reduces this amount. Legislation inthe field is cumbersome and difficult to apply as there is not a reliable cadastre. Furthermore, companies set up inrural zones are excluded o f thistax which may impact onbusiness location. Licenses and Registrations Fee (Alvaras): Among other minor fees and administration, small enterprises must register and pay "Licenses" and "Registrations" (Alvaras) when they start up. The tax depends on their economic activity and ranges between 11$ for locksmiths to 45$ for soap production activities. Companies must then pay yearly 20 percent o f the initial tax. To overcome some o f these problems the authorities are ready to approve new corporate and investment codes that reduce profit tax rate to 25-35 percent, and simplify tax compliance procedures to encourage business creation. The new investment code will provide equal treatment for domestic and foreign investors operating in STP and will rationalize preferential tax regimes inorder to eliminate tax exemptions to enlarge the tax base. The government i s also ready to approve new guidelines on urbanproperty taxation, and new legislation on inheritance but some o f the multiple taxes and fees that impact negatively on business set up and tax compliance will remainunaffected. 88 ANNEX 11.2: AN ALTERNATIVE INCOME TAX PROPOSALFORTHE UPCOMING INCOME TAX REFORM Current Income Tax andGovernment Reform Proposal The current salary tax i s a flat 13 percent and has led to the collection o f Dbs 30 Billion in 2005 (compared to Dbs 10.5 B in ZOOO), representing 7.5 percent of total non oil public revenues. However there are many deductions and exemptions that make the tax less effective and possibly less fair"'. Also, the tax base i s reduced due to a large informal private sector and limited administrative capacity to enforce tax payment. STP Government has drafted a salary tax bill (see table 1) as part o f its upcoming broad reform of its tax system. The reform aims to: 0 Increase income revenue to keep pace with needed social expenditure. The billplans an increase o f the 13 percent tax rate for those earning salaries above Dbs 36 M. Expandthe tax base. The bill eliminates most o f the exemptions and incorporates salary complements (per diem, and the like) to calculate the tax base.'16 Also, by simplifyingthe payment procedure the government wants to encourage the application o f the salary tax to the informal sector. Introduce some progressivitvby charging higher tax rates on higher incomes and exempt earners with income less than Dbs 1.8 M Analysis o f the current and proposed income tax structure provides us insight into the impact o f the proposed tax reform, highlightinga source o f inequality in taxation, and suggests a possible correction. The analysis i s partial intwo ways. First o f all, we rely only on civil service payroll data for 2005 and do not include any form o f private income, either earned outside the public function or by employees or self-employers working in the private sector'". Secondly, payments o f extra hours after budget approval (a third o f the labor cost in Ministries o f Education, Health and Economy) and per diems are not analyzed as there i s not clear record of their distributionamongworkers' 18. The government proposal suggests that while tax i s easy to calculate and collect, which i s essential given the low capacity o f STP tax administration, its unclear about whether some high bracket income earners, i.e. diplomats, will remain exempted. ' I 9 'I5There i s a D b 5,000 monthly deductible and the earned annual income up to Db 480,000 i s exempted. There are also exemptions for salary supplements up to 10percent o f the fixed monthly salary, expenses allowances, per diem, representation allowances and certain moving expenses up to the limits set for government employees. Income o f clergy from the exercise o ftheir spiritual functions and foreign personnel o f diplomatic and consular missions or inthe service o f international or foreign organizations are also exempted. 'I6It clearly includes bonus, per-diems, benefits, awards, as well as in-kind and foreign income, and the way to value them. Many o f these complements are actually paid by foreign financed projects and remain unclear whether they appropriately withhold the tax at source. 'I7State data i s not available for this expanded exercise. An informal and very rough estimate puts private employment at 25- 50 percent o f total employment with an income distribution skewed towards higher income than the public sector. 'I8 Per diems are mainly financed off budget and through externally funded projects. 'I9The tax bill clearly specifies the schedule and tax amount to pay depending on the monthly salary and states that the employer has to withhold at source. Only those who change income brackets during the year have to file a tax reconciliation which should bring the role o f tax administration to a minimum. 89 The tax bill achieves the goal o f raising tax revenues from civil service salaries by Dbs 670 M (+6 percent) compare to the current tax system (Table 2. 1). The draft bill reduces the number o f exemptions and applies a tax rate above 13 percent to gross salary above Dbs 36 M. This increases the effective tax rate from 13 to 14percent. However, the tax structure presents flaws in terms o f equity, as similar incomes are not taxed similarly, lower income taxpayers face steep tax increases and marginal taxation i s not taken into account The bill states that annual net salary cannot be inferior to Db 1.5 Mafter taxes. Italso statesthat the entire gross salary is taxed at the tax rate o fthe income bracket. This implies that a move up from a lower income bracket to an upper one raises the effective tax rate for all earned salary. Consequently increases in salaries close to the upper limit o f the bracket are discouraged, which i s against the principle o f tax neutrality (tax structure should not heavily affect taxpayer economic decisions). To tackle this flaw the bill states that lowest net salary in one bracket cannot be inferior to the highest net salary o f the income bracket preceding it, which implies that similar taxpayers receive different tax treatment. The figure 2.3 shows how effective tax rates increase abruptly for income earners indifferent income brackets. Figure 1: Tax Rate of Government Proposal Tax Rate of Government Proposal 30% 25% 2.1w.000 74.634.146 123,000,WO 150.000.OO0 Gross Annual Salary Also, the government's proposed income brackets can be more progressive as the tax bill does not appropriately address the need to broaden the tax base and impacts negatively on progressivity. Some major income earners, particularly diplomats, remain exempt. Only 225 public employees (3.5 percent o f total public employees) earn more than Db 36 M and would pay under the new tax bill a tax rate higher than 13 percent. Actually, the highest income bracket would consist in 31 employees, merely 0.5 percent o f total public employees (mainly diplomats posted abroad). Also the lack o f progressivity in the lower scale o f the tax structure does not provide incentives to the self-employed or employees earning more than Db 2.1 M (the large majority o f the population) to register their activities12'. IZoThe move from zero percent tax bracket to 13 percent tax bracket for those on the verge o f Db 2.1 Mdiscourages themfrom complying with incometax inthe absence of other incentives (exemptions, benefits associated to formality 90 Two alternative Proposals for the Income Tax Reform Inthe first alternative, we retainthe general principles, tax rates and income intervals of the government proposal, but introduces the marginal tax rates' thus the tax structure smoothes. Figure 2: Alternative proposal: progressive tax structure Progressive Tax Structure 1 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 1,5(m,OOO 36,000,000 72,000,000 120,000,00 150,000, GrossAnnual Salary Based on public service pay roll, and using the government suggested income brackets, the income tax with marginal tax rates decreases tax revenues by Dbs 862 M (-7.7 percent) (Table 2-1). This i s mainly explained by the general exemption o f the first Dbs 1.5 M earned by the taxpayers, bringingthe effective tax rate to 12 percent compared to 13 percent inthe current tax system. However, it this alternative proposal reduces the disincentives for lack o f compliance for low income earners and to reporting salary increases. Taxpayers inthe low income brackets have a more progressive increase in the tax rate which reduces the disincentive to tax compliance. Also, there i s no abrupt move between tax brackets to discourage salary increases. The second scenario suggests that to improve tax collection and ensure better progressivity, it would be necessary to redefine income brackets and increase tax rates. To bring salary tax collection to figures close to the government draft bill all tax rates should be increased one percent (table 2-XXX) `*l. The highest marginal tax rate i s still below other Afiican Countries (Botswana's i s at 25 percent; Ghana and Nigeria are at 25 percent; Kenya, Tanzania, Uganda and Zambia are at 30 percent). This new proposal increases income tax on public employees by Dbs 528 M, (+4.7 percent) compared to the current systembut shares the burden o f the tax reformina fairer way than the government draft bill. The effective tax rate for employees earning less than Db 16 M (77 percent o f total public employees) would be less than the current 13 percent, leaving them better o f f than inthe current tax system. Employees earning more than and the like). Onthe other hand, the costs o f remaining inthe informal sector is not as high since if taxpayer decides not to comply voluntarily with tax payments, the authorities have few administrative means to enforce payment. 12'The highest marginal tax rate in STP, 21 percent, would still be below than in other African Countries (i.e. 25 percent inBotswana, Ghana, Nigeria; 30 percent inKenya, Tanzania, Uganda, Zambia) 91 Db 16 M (23 percent o f total public employees) will pay at a higher effective rate bearing the overall tax increase. Conclusions andrecommendations: The analysis above suggests the following conclusions and recommendations: The draft o f the income tax bill addresses administrative issues properly and tries to expand the tax base by incorporating salary complements (i.e., per diems, and the like). It is important to ensure that: 0 All income earners, i.e. diplomats, pay taxes 0 Foreign fundedprojects withhold taxes at source The draft billneeds to consider a progressive tax system with marginal tax rates: 0 The draft bill does not account equity and progressivity concerns which eventually disincentives formalization and certain salary increase. Progressivity, particularly for low income earners, would be improved by including a minimumincomeexemptionas suggested bythe draft. However, it is important to note that marginaltaxation i s essential to progressivity o f taxes, This would also make it easier to broaden the tax base andreduce the incentive to remain inthe informal sector. 0 Income brackets and tax rates o f the draft bill should be redefinedto distribute the tax burdeno fthe reformina more equitable way. 92 ANNEX11.3: TAX INSTITUTIONS,TAX STRUCTUREAS OFJUNE2006 AND UPCOMINGTAX REFORMS Tax Institutions Tax Institutions are hampered by low capacity and old legal structure. The Tax Directorate lacks the human resources to properly administer complex tax regulations and the legal authority (backed by appropriate laws and regulations) to enforce tax collection.'22 The tax base has remained narrow due to a large informal economy and the lack o f capacity to identify enterprises, assess their tax load and make them pay taxes legally. As most o f the economy remains informal registered information, particularly in profit tax, i s scarce and inaccurate. Consequently the directorate concentrates on tax compliance supervision with very limited tax auditing'23. Emphasis on revenue collection places unfair tax burden on the few registered companies due to the fact that some tax rates have been high(for instance the corporate rate o f 45 percent), while not remedingtax evasionby many. Government plans to further strengthen the Tax Directorate and tax collection. The government already introduced a single taxpayer identification number for all revenue agencies but due to lack of computer capacity, information stemming from it is not complete and has not being appropriately e ~ p l o i t e d ' ~Recently, the Tax Directorate started implementing a resolution ~ . from the Ministry o f Finance to collect arrears dating back from 2003, particularly excise taxes from domesticallyproduced beer. Also National Authorities plan to approve a new tax procedural code that will strengthentax directorate ability to enforce tax compliance Current structureof the tax system Indirect taxes: Tariffs and trade taxes: In2000 the customs tax was overhauled. Export taxes were removed. The import tariff structure for imported goods was reduced to three tax brackets, 5 . percent on the value o f basic goods, 20 percent on luxury goods and 10 percent on the remaining import goods. Surtaxes apply to petroleum products (between 57.9 percent and 317 percent), alcoholic beverages (between 25 percent and 55 percent), tobacco (55 percent) and used vehicles (0 percent to 20 percent). For tax collection purposes, imports were originally valued at c.i.f. price with several goods paying a reference price set by the government. The price was set in Dobras and seldom adjusted which, due to dobra depreciation, progressively reduced import values and therefore the tariff collected. The government updated the Dobra based reference price in2005 and in2006 set the new import values to euros to avoid hrther erosions o f the tariff collection base due to the depreciation o f the Dobra. Broadly granted exemptions on import duties have cost custom 9 percent o f its collection in recent years. Exemptions are granted to imports o f capital goods for the private sector or state owned enterprises, several donor projects, churches, embassies and NGOs related to development projects and all imports for the government. In the first half of 2006, the 55 employees in2005 The number o f audited private firms is around 10 firms a year Notably lack of cross data information between customs, tax directorate and treasury among others. 94 government eliminated exemptions on imports o f the Central Government imports (but not for state owned enterprises). Tax on petroleum goods has been simplified, and includes 5 percent customs, 4 percent tax allocated to road maintenance fund and ENCO' handling and distribution cost. Subsidies were eliminated and nowadays the full cost passes to consumers, which in 2005 represented 171 percent o f gasoline cost. In 2003 the authorities introduced an excise tax on local services (mainly utilities and hotels) o f 2 percent that was increased to 5 percent in 2005 with a new tax hike to 7 percent programmedin2007. In2005, the tax was also expanded to cover domestic beer production. This tax has been very successful and has increased local excise tax to 2.3 percent o f total revenues in 2005 compared tojust 0.4 percent in2000. Direct taxes: Corporate/profit tax collection has increased but tax rates are high, providing a disincentive to formalize economic activity contributing to a narrow tax base: Private firms with profits o f less than 12 million Dobras pay 30 percent intax which become 45 percent beyond that amount'25.These tax rates are very high when compared to other African Countries (Tanzania, Uganda and Zimbabwe 30 percent, South Africa 29 percent or Botswana 15 percent). Few companies file taxes and the informal sector i s wide spread. Only firms with more than 30 million Dobras ($2800) in sales must keep proper accounting books'26.The remaining companies either are taxed based on information provided by the company to the Tax Directorate or the tax i s estimated by the Directorate itself based on the company sector. In 2001 there were 2781 companies filing profit taxes with around 200 companies in the first category. To ease the tax burden, the authorities have used the investment code to grant ad-hoc temporally tax breaks to certain companies Salary tax: uunlike many other African countries, Sao Tome has an individual flat income tax that i s levied on all domestically earned income. The tax i s withheld at source and deposited inthe treasury the first eight days o f the month after the income was earned. Upcomingtax reforms Tax reforms are unfinished: Up to end 2006, many o f tax laws and codes still dated back to colonial dates, with a patchwork o f amendments that made them cumbersome, inefficient and full of exemptions. There are also a large number of different taxes with different tax structures that are difficult for taxpayers to understand, allowing discretionary powers to tax enforcers. The 2000 initiative was a reform process to modernize the tax system in line with other African Countries, trying to foster tax collection, broad tax base and simplify tax compliance to eliminate growth hindrances related to the tax system. However due to political uncertainty that started in 2001 and continued until after the elections o f August 2006, refoms related to other taxes - notably profit, corporate, personal income and customs related taxes (new customs organic law and a new general code on customs infractions) -slowed down. While draft laws lZ5 Only 50 percent ofprofits fromAgriculturalcompanies are taxed They were around200 companiesthat fell inthis group in2000 95 were prepared, the reforms were not fully discussed or adopted by the previous National Assembly. S i 0 Tom6 and Principe i s updating its legislation to bring it in line with international standards. Several draft tax codes are with the National Assembly for approval between end-2006 and end-March 2007, and implementation during2007 and early 2008. The new Investment Code (codigo de investimento) establishes equal treatment between foreign and domestic investors, guarantees unrestricted repatriation o f capital gains and profits, and sets the general conditions for tax exemption regimes, which must be stipulated inthe tax legislation. Itwas beingdiscussed by the National Assembly inearly 2007. The General Tax Code and Tax Procedural Code provide the framework under which tax obligations and procedures are defined (codigo geral tributario, codigo de process0 e procedimento tributario, codigo de sisa, codigo de imposto sobre o conumso). The codes modernize tax administrationprocedures and create specialized judicial institutions for resolving appeals and enforcing collections. The latter i s critical, as under current legislation, the tax authorities have only a few instruments to fight tax evasion. Effectiveness is expected within 90 days o f approval o f the codes. The General Tax Code and Tax Procedural Code were approved by the National Assembly inDecember 2006. The draft corporate and personal income tax codes (codigo de imposto sobre rendimento de pessoas colectivas, codigo de imposto sobre rendimento de pessoas singulares) seek to harmonize taxation between enterprises and individuals, improve the rate structure, and strengthen tax assessment. The drafts have benefited from technical assistance from LEG and advice from FAD. A key issue for corporate taxation i s to lower the tax rate (currently at 40- 45 percent) in line with international levels. The draft personal income tax code brings all forms o f income under one tax (not just salary, as i s the case today) and introduces progressivity. The choice o f tax rates and their impact on revenue are being studied, in consultation with Fund staff. Bothcodes are expected to be voted in2007 and become effective in2008. Two more draft direct tax laws (for inheritance tax and urban real estate tax) aim at enhancing efficiency o f collection (codigo de imposto sobre sucessao e doacoes, codigo de contribuicaopredial urbana). As o f early 2007 these codes were being discussed in the second commission o f the National Assembly. The objective o f the Customs Code and related regulations i s to unify today's disperse regulations into a single body and to provide customs with adequate tools for assessing import values. In this respect, customs procedures will be updated and alternative valuation methods considered, for their gradual application in line with strengthening capacity o f customs administration. The Customs Status was approved by the Cabinet and signed by the President o f the Republic in early 2007 and i s to be published. The Customs Code and Code for Customs Infractions are to be discussed by the Cabinet inthe first halfo f 2007. 96 ANNEX 11.4: LOCAL ADMINISTRATION Local administration i s represented in the central budget through the transfers made to the island o f Principe (Regizo do Prin~ipe)'~'and Cdmaras Distritais (District Councils). The spending level and structure o fthese administrations i s however not known for no comprehensive and timely reporting i s made. Direct transfers to the local administration do not represent a significant share o f the central budget (3.3 percent o f the total recurrent expenditure in 2005) signaling a high level o f centralization in the public expenditure management. There was a significant increase in the amount of transfers in absolute terms between 2002 and 2005, but even though its relative weigh in the total expenditure has deteriorated after the maximum o f 4.2 percent reached in 2003. Nevertheless, the decreasing share in the total recurrent expenditure has been offset, over the last years, by an increasing participation in the development expenditure. The execution rate o f the transfers to the local administration fluctuated widely over the last years, signaling that the local administration i s strongly affected by the treasury shortages. It consequently creates a situation of poor budgetary predictability for the local administration. Table 36: Transfersto the Local Administration Total local administration Programmedtotal expenditure 4.1 6.5 7.3 21.6 33.4 Executedtotal expenditure 2.4 6.5 7.6 13.9 Executionrate (%) 58.9 100.0 104.2 64.2 Ratioto executed total recurrent expenditure(%) 2.6 4.2 3.6 3.3 2.7 Ratioto executed total expenditure(%) 0.9 1.6 1.6 2.7 3.5 Source:Ministry of Planning and Finance Data on local administration is presented differently inthe state budget. Regiono fPrincipe i s registered under the heading o f 'Encargos Gerais' whereas the transfers to the districts are classified as 'Cfimaras Distritais' 97 ANNEX 11.5: INFORMATIONCONSTRAINTSINTHEMANAGEMENT OF PUBLIC INVESTMENT The management o f the Public Investment Program lacks a proper information system. In fact, the system i s not supported by an adequate database, the classification standards are poor as they do not ensure the consolidation o f the recurrent and development expenditure. . Moreover, it should be highlightedthat the present budget classifications do not support a proper articulation between the NPRS, the recurrent budget and the PIP. The information provided in the project fiches i s not enough to ensure the financial sustainability o f the projects, let alone the possibility o f estimating the future recurrent expenditure generated by those projects. PIP data i s presently managed though a database built in 1996, which is conditioned by many constraints: i)programming i s not user friendly; ii)it does not export files to spreadsheets to ease further data processing; iii)Classifications used in the database do not allow an integrationor linkage with the recurrent expenditure. A new database to support the management o f the Public Investment has been built, but its implementation has been withheld on the expectation of the evolution o f the budget reform. However, whatever this development, the new database should be adopted as a management tool of PIP, in close relation to the NPRS strategy. To this end, it should cover all the phases o f the budget cycle (programming, implementation, monitoring, evaluation) and accommodate a number o f still missingelements like the assessment o f future recurrent costs. A first analysis o f the information content o f the new base, as far as it could be assessed through the information content o f the new `project fiches', shows some progress in relation to the previous one. For instance, from now on it i s possible to classify the development expenditure according to the functional and economic classifications, though there i s duplication o frequested information (i.e. some items specifically related with investment are requested on the basis o f specific classifications, though they could be obtained by simply using the economic classification with a proper breakdown to accommodate the needs o f the public investment). Public investment data i s organized according to the following classifications: `Administrative', similar to the one used inthe recurrent budget; `Sectors', which i s an aggregation o f the administrative classification after the reclassification o f some expenditures items, with the purpose o f reclassifying the investment expenditure by `function'l28. This classification was the main one until 2002, being replaced by the `administrative classification' (by ministries and other spending units),but itremaineduntilnow, so both classifications coexist; `Programmatic', which i s inpractice a mere aggregation o fprojects by homogenous areas; There i s also a sort o f `line item classification' that i s similar to the `economic classification' but with a different structure. It refers to spending categories that are relevant and specific o fprojects (like technical assistance, studies, etc.); It shouldbe recalledthat only the recurrent expenditureis classifiedinthe budget accordingto the `Functional Classification'. This leaves a substantialamountofthepublic expenditureout ofthat classification. 98 Finally, there i s an ad hoc classification o f projects according to their `pro-poor' orientation. Despite, all these classifications the link with the recurrent budget can only be ensured through the `administrative' classification. The lack o f a programmatic classification does not enable an accurate matching between policy and project objectives. PIP i s in fact constituted by a list o f actions ranging from large projects (like for instance the `Modernization o f the international airport o f S. TomC', for 2 million dollars), to very small actions (like a simple study o f 3 thousand dollars relative to the `New working timetable for the public administration'). The existent `Program' classification used in the PIP reports does not characterize real programs, being solely an ad hoc way o f grouping the public investment in some meaningful way, according to homogeneous areas o f intervention. The adoption o f the NPRS did not contribute thus far to change the situation. The programmatic classification underlying the NPRS has not been adopted by the PIP management as mainreference for investment programming and monitoring. A major problem typically affecting the `development budget' in other countries is the difficulty in distinguishing the current and capital expenditures. Actually, the `economic classification' is not appliedto the investment expenditure o f STP, which does not enable to make that distinction. However, such exercise i s made by the Institute ofStatistics with the purpose o f estimating the Gross Fixed Capital Formation for the National Accounts, but it seems that there i s no feedback o f this exercise to the Planning Directorate. In the context o f the present PER an estimate o f the capital content o f the development expenditure has been made, on the basis o f the investment data by `expenditure item'. However, the robustness o f such estimate could not be verified because the original data provided inthe project fiches was not assessed. 99 ANNEX 11.6: DATA SOURCES OFEXTERNAL AID AND PIP FINANCING ON Figure3: Sources of externalaid, 2002-2005 1 ' O ' O 51.1 50.0 40.0 S 30.0 20.0 12.8 10.0 I ... 5.3 5.2 5.1 0.0 Taiwan European United BADFAD World Bank France Fbrtugal Other Other Union Flations Bilaterab Witilaterak Table 4: Financingof the PublicInvestmentProgramme in % of total investment 2002 2003 2004 2005 2002.05 Budget Executed Budget Executed Budget Executed Budget Executed Budget Executed Financing 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Foreing 83.34 78.81 85.68 63.38 59.51 87.4s 81.23 78.11 76.51 70.73 Grants 74.78 66.58 69.52 56.51 45.74 53.44 67.32 71.20 63.17 60.59 Loans 8.56 12.22 16.16 6.87 13.77 14.04 13.91 6.91 13.34 10.13 Domeotlc 16.66 21.19 14.32 36.62 40.4s 32.51 ia.77 21.89 23.4s 29.27 Counterpartfunds 0.00 0.61 0.00 0.00 0.00 0.00 3.61 0.00 0.97 0.12 Budget 4.51 7.29 4.76 25.54 30.94 17.68 7.09 9.70 12.84 16.26 HlPC fund 12.15 13.29 9.56 11.08 9.55 14.84 8.07 12.19 9.68 12.89 Sources:Minstryd Plannlnp and FIM~M, Gowrnmem Fisc3 Opraimns 100 ANNEX 111.1:EMAE'S PERFORMANCE This section reviews the financial and operational performance o f EMAE, reviews its tariff structure, and the cost o f the implicit subsidy it receives. The section also reviews the amount o f tariff increases that would help EMAEbreak even and the impact o f such increases on the poor. Financial Performance EMAE'sannual revenues have increased over the last six years, reaching close to US$ 6 million (one-tenth o f the GDP) by 2005. Revenues from water alone make up between 25 to 30 percent o f its gross revenues (see table below). Nonetheless, the enterprise has increasingly experienced financial difficulties and i s now heavily indebted to its suppliers and to international and domestic banks. Its liabilities constitute the bulk o f the of SOE liabilities for which Government i s responsible. Table 5: Summary Information on EMAE'sProfit and Loss Statement and Balance Sheet((US$) ProfitJLoss 2000 2001 2002 2003 2004 2005 Statement Sales Revenue 2,527,985 3,542.782 5,337.981 5,552.875 5,682,122 5,936,631 Cost of Sales 2;798;616 3,697,440- 3;947;708 5; 134;567 5;494;484 6,789,965 -Wage Bill 432,990 557,884 799,281 803,692 799,558 783,583 -OperatingExpenses 2,273,957 3,036,888 3,061,478 4,219,508 4,555,296 5,873,O 18 -Other Expenses 68,248 72,580 86,949 111,367 139,630 130,364 Operating Income -270,634 -154,658 1,390,273 388,308 187,638 -850,334 Financial Expenses, 59,812 -356,096 1,030,639 514,450 512,959 973,6 17 Depreciationand Taxes Ket PrufitlLoss -376,696 -1 66,260 !I 1,057 -235,06? -333,713 -1,8 11,724 Balunce Sheet Liabilities 4,092,534 5,292,894 5,723,040 7,864,512 8,541,554 11,320,814 -Short termdebt 628,836 432,783 666,193 1,775,363 2,742,554 3,193,398 -Long-termdebt 3,465,698 4,860,111 5,056,847 6,089,149 5,799,000 8,127,416 Equity 8,53 1,767 10,889,45 1 11,820,212 10,5 80,037 9,369,920 5,045,840 Assets 12,624,301 16,182,345 17,543,251 18,444,549 17,9 11,474 16,366,654 Finnnciul Ratios Profit Margin -13% -5% 2% -6% -6% -3 1Yo Return on Equity -4% -2% 1YO -3% -4% -36% Return on Assets -3% -1% 1% -2% -11% LiabilitiedEquity 48% 49% 48% 74%" .-2% 91% 224% Source: Compiled from EMAE's financial reports; net profitlloss i s after taxes, depreciation and financial and other expenses. Profit margin i s net income over sales. EMAEfinancial difficulties spring from increased cost o f its operating expenses (mainly fuel), a higher wage bill, weak collection rates and increased financial expenses to domestic banks. The enterprise has been runningnet deficits, negative profitability (profit margin as well as return on equity and return on assets) over the last six years with the exception of one single year in2002. 101 Figure 4: EMAE's Profit Margin as Net ProfitLoss over Sales 5% 0% -5% -10% -I 5% -20% -25% -30% -35% 2000 2001 2002 2003 2004 2005 Source: Compiled from financial data intable 6. Fuel costs constitute the bulk, three-fourths, o f EMAE's operating expenses. Since only 10percent o f EMAE's electricity is produced by its hydro-electric plant, 90 percent, comes from its thermal plant operating exclusively on imported fuel. The utility has become increasingly vulnerable to the increase in imported oil prices in recent years. The market price o f oil has increased from 6,000 dobras (gasoline) and 5,000 dobras (petroleum) per liter in 2000 to 12,500 dobras (gasoline) and 11,000 dobras (petroleum), including a 15 percent fuel tax, in 2005. Even though EMAEpays 35 percent below this price to its major supplier, ENCO, its total fuel costs have tripled within the last six years fromUS$ 1,681,583 in2000 to US$4,381,998 in2005. While EMAE's wage bill also increasedby 80 percent from US$432,990 in2000 to US$ 783,583 in 2005, the perception within the government that EMAE's main problem i s overstaffing or a large labor force i s misguided. The increase in salary i s due to increase in number and average salaries o f the employees. The 37 percent reduction in the number o f employees - implemented as part o f the 2001and 2002 reforms - has been almost fully reversed (see section below). Also, the average annual wages per employee more than doubled from US$ 1,468 in2000 to US$ 3,378 in2005129130. However, the total wage bill has gone from 17 percent o f total revenues in 2000 to 13 percent o f total revenues in 2005. Fuel costs and operational inefficiencies(discussed below) are the main problems. Inaddition to salaries, EMAE's employees seemto collect additional company benefits while receiving interest-free loans when needed in addition to a lower electricity and water tariff--none of which i s calculated in monetary terms or included inthe wage bill. 102 After paying for the high fuel cost and the wage bill, EMAE has not been able to spend adequately on muchneeded maintenance or investment since 2000. In2005 EMAE managed to double its maintenance budget and to invest for the first time in six years (table 35). The utility's engineers maintain that a further increase in maintenance expenses will reduce EMAE's fuel consumption by half, generating additional savings. Their analysis shows that an additional 600,000 Euro (US$ 750,000) for maintaining the thermal plant will generate savings in fuel consumption equivalent of 1,250,000 Euro (US$ 1,500,000).'31 The same report also argues that an additional spending o f 350,000 Euro (US$400,000) for further rehabilitation of the Hydro-electric plant Contador would save an additional 1,080,000 Euro (US$ 1,350,000). Table 7: EMAE's MaintenanceExpensesand Investments(2000-2005)(US$) 2000 2001 2002 2003 2004 2005 Maintenance 62,759 116,613 187,088 774,790 356,460 600,699 Investments 0 0 0 0 0 1,411,500 Source: Compiled from EMAE's financial reports. Weak collections compounds the financial difficulties related to the high operating costs and low maintenance and investment budget and point to the need for a systematic upgrading and reform o f the collection system (Table 36). The stock o f customers' payment arrears (which constitute the bulk of the receivables) increased almost five-fold between 2000 and 2005, with a 40 percent increase between 2004 and 2005. One-half o f these arrears belonged to the public sector (Government and other state-owned enterprises-partially or fully owned) and the rest to the private sector. In late 2005 and early 2006 EMAE's management attempted to clear the arrears with usingtwo approaches. On the public administration front, the management initiated two agreements for settling the government's arrears and its own vis-a-vis its sole supplier o f gasoline, Empresa Nacional de Combustible e Oleas (ENCO). On the private sector front, the management introduced a limited pre-paid system and institutedpenalties for theft and fraudulent behavior in an attempt to improve the collection efficiency. These two efforts while useful will need to be incorporated into a systematic reform and investment program to ensure their sustainability. Table 9: Stock of CustomerArrears inEMAE's Total Receivables(US$) 2000 2001 2002 2003 2004 2005 Customer Arrears 581,507 615,920 1,266,109 1,921,627 1,849,247 2,4453 15 ----ofpercent which public sector 290,753 307,960 633,055 960,814 924,624 1,222,658 share 50% 62% 50% 50% 50% 50% Lendingto Employees 628 3,303 4,381 3,666 10,188 9,544 Receivablesfrom state 66,393 184,685 46,810 54,423 35,562 33,288 Bills to be collected 347,529 226,600 688,043 741,018 708,489 0 Other receivable^'^^ 340,907 20,666 67,866 1,989,972 1,702,203 1,031,996 TOTAL 1,336,965 1,051,174 2,073,209 2,791,001 2,641,354 3,520,144 Source: Calculated from the information in EMAE's financial reports; the public sector arrears, the responsibility of the Government, are roughly estimated. A detailed breakdown o f customer arrears under the receivables inthe balance sheet i s not available at this point to conduct a more detailed study. 13' FaustoNetto, EMAE's Former DirectorofElectiricity Operations,"Prezuijos CausadosAo Sistema de Produgaoe Distributicaode Electricidadeem S. Tome e Principe," June 8, 2006, p. 4 i s not clear what these "other receivables" are andthey needto be investigatedas part o f the proposedforensic audit exerciseproposedinthe recommendations.(See 6.2. State-OwnedEnterprise Unit) 103 While EMAE suffers from unpaidbills, it accumulated its own liabilities inturn between 2000 and 2005, creating a major problem both for EMAE'smanagement and for the Government as its sole owner (figure 8). Its short term liabilities are made o f its accumulated supplier arrears and its short term debt to domestic banks. By the end o f 2005, the utility's accumulated supplier arrears (combined in dobras and in foreign currency) were equivalent o f US$ 2,112,554. The bulk o f these supplier arrears, US$ 1.5 million, were owed to ENCO, its lone gasoline supplier which EMAE managed to settle in March 2006 while settling simultaneously the government's arrears. The rest i s the short-term debt which the last management o f EMAE obtained from domestic banks. For instance, one US$600,000 loan i s from the Bank International o f Sao Tome (BISTP), carries an interest rate o f 18 percent and i s for three months.'33 EMAE`s increased medium and long term liabilities from US$ 4.1 million in 2000 to U S $ 11.3 million in 2005 has serious implications (table 37). Its debt ratio (liabilities over equity) jumped from 48 percent in 2000, 2001, 2002, to 74 percent in 2003, 91 percent in 2004 and to an unacceptable 224 percent in 2005. The long term debt i s responsible for this level o f high indebtedness as it jumped from 41 percent in 2002 to 161 percent in 2005. Under these circumstances, the debt coverage ratio (EMAE's ability to service its debt) i s naturally negative since the net income has remained negative. The aggressive financing strategy in 2005 may have putthe enterprise infar worse financial shape. EMAE's increasedliability has increasedthe fiscal exposure o f the central budget as the Government i s the guarantor (explicit or implicit) o f all SOE loans. It not expected to pay back the concessional loans it receives from the international institutions via the Treasury as they will be forgiven under the HIPC completion point / MDRIprocess. It has regularly been servicing its loan to the European Investment Bank (EIB). The loan i s under discussion for potential debt relief for its remaining loan o f US$4,652,527 at the HIPC completion point. This international loan i s an explicitly guaranteedby the Government. Besides the EIB loan, the other two medium term loans are from domestic banks and are implicitly guaranteed by the government: US$ 1,300,000 from AFRILAND First Bank at an interest rate o f 11 percent and US$ 668,000 from COSTP with an interest rate o f 18percent.'34 `33This information i s gathered from EMAE, "Relatio e Contas," Exercisede 2000-2005, andthe interviewsconducted with EMAE's Director GeneralinJune 2006. Apparently,BISTP, apartially state-ownedbank, agreed to lendfunds to EMAE knowing that the Government will shoulder the financial responsibility if EMAE could not service its debt. EMAE also appears to have benefitedfrom a lower interest rate of only 18% than the on-goingmarket rate of 23 or even 24%. This is similar to the 35% reduction in price paid by EMAE to ENCO, another partially state-owned enterprise, all needto betaken into account ifandwhen there is an attempt to calculate the net subsidies. `34It is not clear if loanswere reportedcorrectlyinfinancial statements. 104 Figure 4: EMAE'sIncreasing Liabilities $10,000,000 < - " - - - - - . - - -I- ^I_ $8,000,000 - - $6,000,000 ,_.- __ - - - $4,000,000 $2,000,000 $0 2000 2001 2002 2003 2004 2005 ~ Source: Compiled from financial data analyzedintable 6. OperationalPerformance EMAE's financial problems are closely related to its weak operational performance (technical and commercial performance). EMAE can increase revenue (the volume o f water and electricitypaid for) by improving operational efficiency even without changing its tariff structure. In the water sector, the combined technical and commercial efficiency (ratio of water paid to water extracted) i s less than 50 percent, which i s low in comparison to other countries This combined efficiency indicator is influenced by a low technical efficiency (ratio of water distributed over water extracted), which reached 57 percent in2005 (up from 45 percent in2000). A 57 percent technical efficiency is still weak by comparison to other counties. The increase in technical efficiency between 2000 and 2005 was achieved as the water supply operations o f the utility increased the amount o f water extracted and distributed by approximately 60 percent. The enterprise now extracts 10.6 million cubic meters and distributes 8.4 million cubic meter o f water. On the other hand, the amount o f water billed over water distributed (water accounted for) increased from 58 percent in 2000 to 72 percent in 2005 and inversely the unaccounted for water (UFW) appears to have declined from 44 percent in 2004 to 28 percent in2005. 105 Table11: EMAE'sOperationalPerformancefor Water (in cubic meters measuringquantities) WA1'EH 2000 2001 2002 2003 2004 2005 Extracted 6,708,473 7,892,322 8,752,552 9,298,827 9,751,858 10,630,347 Treated 3,057,600 4,555,200 5,3 19,464 6,852,072 7,454,235 8,790,3 10 Distributed 5,232,609 5,424,625 5,751,734 6,881,131 7,265,134 8,395,817 Billed 3,032,000 3,235,000 3,558,500 3,830,250 4,102,000 6,035,000 Paid 2,767,000 2,777,000 3,054,700 3,359,400 3,692,000 4,861,000 Ratios Distributed/Extracted= 45% 41% 41% 41% 42% 57% TechnicalEfficiency Billed/Distributed= 58% 60% 62% 56% 56% 72% Water AccountedFor 1-Water Accounted 42% 40% 38% 42% 44% 28% For=UnaccountedFor Water (UFW) Paidmilled= 91% 86% 86% 88% 90% 81% CommercialEfficiency PnirUE~~truct~d=Lbinbiiiled 41%I 35% 15"' 3 i o 36%) 38%) 46%) .I Effficiency Source: Data provided by the Directorate o f Water, EMAE; for the missing data the consultant chose a method o f approximation. EMAE's commercial (or collection) efficiency for water (ratio o f water paid to water billed) has also improved slightly up until 2004 but declined significantly in 2005. While the exact cause o f this abrupt decline could not be identified from available information, it may have emanated from the management correcting one or more o f the factors common in utilities worldwide that lead to highrate o f unaccounted for water (UFW): 1) inaccurate estimate o f the amount o f water pumped due to not metering all water at the intake source or by usingraw water or finished water meters that are inaccurate or improperly installed; (2) inaccurate customer meters; (3) bookkeeping errors, (4) non-metered uses such as water used inthe treatment process, and (5) water leaks.'35 Figure 7: WATER: Technical,Commercial ai 1CombinedEfficiency I 100% 1 80% +Technical Efficiency 60% --.I--. Commercial 40% Efficiency I 20% Cornbined Technical and Commercial 0Yo 2000 2001 2002 2003 2004 2005 ~ Source: Compiled fiom data provided by Directorate o f Water, EMAE as reported inAnnex 5 '35 Itwill be useful to undertakeastudy to explorehow EMAEmanagedimprove operationalperformance and what needsto be done to improveit even further. 106 EMAEmightconsider further technical improvements via installingbulkmeters and leak repairs as well as improvements in collection efficiency via reaching un-billed subscribers, identifyingillegal connections, follow up disconnections and meter replacement. It can increase capacity and water treated, distributed, billed and paid by upgrading facilities in every phase o f the distribution process up untilit collects the payment for the services rendered. In the electricity sector, authorities need to take immediate attention as the combined technical and commercial efficiency (electricity paid for over electricity produced), has declined from 55 percent in 2000 to only 51 percent in 2005. This decline is due to the falling performance o f the commercial efficiency (the ratio o f electricity paid for over electricity billed) from 91 percent in2000 to 75 percent in2005 (table 38 and figure 9). The commercial efficiency has been partly affected by the payment arrears from the Government as well as arrears from all other customers including other infrastructure SOEs. While EMAE's management claims to be improving the collection efficiency for electricity, the 2005 numbers did not yet supporting this turnaround. Table 13: EMAE'sOperational PerformanceFor Electricity (inkWhmeasuring;quantities) ELECTRICITY 2000 2001 2002 2003 2004 2005 Produced 26,050 26,467 31,208 35,889 37,243 41,453 Distributed 25,000 24,692 25,207 32,865 35,496 39,471 Billed 15,782 17,162 22,505 23,867 25,228 28,224 Paid 14,406 15,103 19,129 19,571 20,182 21,168 Ratios Distributed/Produced=Technical 61 Yo 65% 72% 67% 68% 68% Efficiency Paid/Billed=CommercialEfficiency 91% 88% 85% 82% 80% 75% Paid/Produced=ConibinedEfficicicy 5 5 % 57Y" 61'%, 55%) 54%) 5 I"& Source: Data provided by the Directorate of Electricity, EMAE; for missing data the consultant chose a method o f approximation While technical efficiency (electricity distributed compared to electricity produced) has improved slightly from 61 percent to 68 percent in the past five years, it still has much room to improve as technical losses should not be more than 10 percent. Also, the recent technical improvement has not been enough to stop the decline o f the combined efficiency. The slight improvements have covered a major rehabilitation o f the hydro-electric plant, Contad01-l~~and the reduction o f the transmission and distribution losses seem to have been reduced. The biggest gains to technical efficiency are likely to come from improving maintenance, rehabilitating and upgradingthe existing electricity transformation and distribution network. 136 The ratio between hydro-electric and thermal production was halved from 26% in2000 to 13% in2005 as aresult o f rehabilitating. The conclusion o f the rehabilitation efforts will likely to reverse this trend inthe near future. 107 Figure9: ELECTRICITY:Technical,Commercialand CombinedEfficiency loooh 80% 60% 40% 20O/O Source: Compiled from dataprovided by Directorate o f Electricity, EMAE Consumers Over the last fifteen years, EMAE's coverage has increased substantially, but will needto expand firther for the public. EMAE's water customers doubled from 4,000 in 1991 to 8000 in 2005 and the total number o f electricity customers reached some 21,700 in200513'. Less than one third o f these electricity connections are the residential users, while two-thirds are for the private sector and Government (central government, ministries, public agencies and entities). This translates into a rough estimate that only 40% o f the Sao Tome's population receives electricity and water. This coverage i s low compared to the regional averages: Dakar i s 55 percent electrified, Abidjan i s 60 percent electrified and in Cape Verde, 59 percent o f the households declare electricity use13*. Many o f the urban and rural poor have no access to electricity or piped water. The lack o f new investments into EMAE i s the major barrier to increasing coverage o f both electricity and water output and services. (See Table 11 and Annex 3 containing EMAE tables). Cost o f access to and use and reliability o f electricity supply can be improved notably. At present, households pay for a connection fee and the Volume differentiated bloc tariff based on their usage per lulowatt hour (see annex 3 for details on bloc tariffs and their evolution between 2000 and April 2006)13'. Many o f the poor households can not afford the connection fee to the grid which cost around 1.5 million dobras (US$ 125). In 2006, the government launched a program that allows connection fees to be paid over several months, thus reducing this barrier for the poor. 13'Approximately7,000 residentialcustomers(or 28,000 people) nowreceive bothelectricityandwater andget one single bill while only 1,000 customers are billed only for water. '" 13'Cape Verde data for 2001-2002, datafrom Wodon, Angel-UrdinolaandEchevin(2006). Volume differentiated bloc tariffs provide lower prices to clients at low levels of consumption only to those householdswho consume less than agiventhreshold,. This thresholdis often calledthe "lifeline" levelof consumption that shouldbe affordableto all. It can be shown that VDTs tend to bebetter targetedthan IBTs, but inmost cases only marginally so. VDTs have one weakness: they imply a discontinuity in terms of the amount of the bill for the customer. For example, assume that the lifeline is set at 100kwh, andthat alower priceper kwh applies to households with consumption levelsbelowthat lifeline, while ahigher price appliesto all the consumption of householdswhile the consumptionlevel is above the lifeline. This means that ahouseholdconsuming 101 kwhper monthwill have a much higher bill than the householdconsumingonly 99 kwh, while both householdmay be equally poor. In addition, some householdsmight move from one priceto another dependingoftheir monthlyconsumptionandthe billing system. 108 Average electricity tariffs o f U S 15 cents (2005) are higher than average electricity tariffs per kwhwas U S 9 cents inCote d'Ivoire (2003), U S 8.6 cents inUganda (2002), U S 3.1 cents in South Afiica (2002) and U S 14 cents in Senegal (20O4).l4O Service has been unreliable since 2004 as production can not meet demand. Frequent blackouts, brownouts, and load shedding affect both residential and commercial consumers. Inresponse, many private businesses have set uptheir ownelectricity generators to ensure reliable service. Between 2000 and 2006, nominal tariffs for electricity have increased substantially and at different rates across consumption brackets, but in real terms, the average price per Kwh for residential customers has remainedflat over the years, even though the operating costs for EMAE have increased faster than inflation due to the increase in world oil prices. For residential customers, the increase has also been differentiated by consumption bracket, with a sharper rise for the lowest consumption where usually concentrating the poor consumers: a 122 percent increase for the lowest bracket (so that the price per Kwh was 2.22 times higher in 2006 than in 2000), 75 percent for the second bracket, and 50 percent for the third bracket. While these increases are large, they are on average in line with the inflation over the period, whose cumulative impact has been such that on average consumer prices at the end o f 2005 were 94 percent higher than in2000. Inreal terms, the changes intariff rates across the non-residential consumers has also been differentiated, with the public administrationtaking brunto f the increase: the price per Kwh for public administration entities i s much higher than for any other type o f customer, and has increased faster than other tariffs, since in April 2006, the price was 3.07 times higher than in December 2000. Prices paid by state-owned firms have more than doubled over the same period (ratio o f 2.18 over time), while prices for commercial and industrial customers have increasedby a much lower 39 percent, and remain much lower than the prices paid by the state administration and SOEs (annex 111). Table 14: EMAE'sAverage Electricityand Water Tariffs (dobras and US$) 2000 2001 2002 2003 2004 2005 ELECTRICITY: 1126 1298 1342 1490 1634 1805 KWHindobras KWHinUS$ 0.13 0.14 0.15 0.16 0.16 0.15 WATER: 1149 1831 2205 2448 2780 2818 M3indobras M3inUS$ 0.13 0.20 0.24 0.26 0.27 0.23 Source: Compiled by the informationprovided by EMAE. Incontrastto electricity, the average tariffs for pipedwater is only US 23 cents per cubic meter which i s reasonably low compared to regional averages--US 84 cents in Senegal (2003), U S 58 cents in Uganda (2004) and U S 41 cents in Mozambique (2003) (see annex XXX for details on water price^).'^' Nonetheless, between 2000 and 2005, the cost o f one cubic meter increased differentially across consumption brackets, with the highest rise affecting the lowest 140Data on cross-countycomparisonsfromthe South African Developmentthrough Electricity (SAD-ELEC), 2002; SenegaleseandCote d'Ivoire datafromthe two cases studies conductedon the privatizationimpactassessmentof SENELEC and Cote d"1voire Electricity, for the World Bankby the BostonInstitute for InternationalDevelopment (BIDE), 2006. Nilgiin Gokgiir and Leroy Jones, "Privatization ImpactAssessment of MozambiqueWater," case study preparedfor the World Bank, fundedby the NorwegianTrust Fund(NTF), 2006. 109 bracket (multiplied by 3.60 consumption below 10 m3'.For the upper consumption bracket the price was multiplied by 3 (consumption above 10 m3). For commercial customers, the price o f water has been multiplied by 2.52 over the period under review, and for industrial customers, the price has been multipliedby 1.77 (annex 111). Also, coverage remains limitedas some 60 percent o f the population presently relies on untreated water. Targetingof the electricityconsumptionand connectionsubsidies The existing electricity consumption subsidies are not well targeted to the poor. As mentioned earlier, the share o f the implicit subsidies in the tariff structure received by poor households divided by the share o f the poor in the population (a), i s 0.486 in 2000.'42 This means that if a randomly chosen person in the population as a whole can expect to receive 100 dobras in subsidies, the likely subsidy to be received by a poor person i s only 48.6 dobras. Nonetheless, the share o f the subsidy received by the poor at the national level i s o f an order o f magnitude similar to that obtained in other countries such as Cape Verde, urban Rwanda, Cameroon, and Guinea. The above estimates of are based on data from the 2000-2001 household survey, and the tariff structure o f December 2000. The advantage o f using these data i s that they are coherent, because the observed consumption levels correspond to the tariff structure actually in place at the time o f the survey. The tariff structure has however changed since December 2000, as shown in Annex 111. Yet the average cost per kwh for residential customers has remained essentially flat since 2000 inreal terms. Given the flat average level o f the tariff over time inreal term, it i s probably reasonable to assume that the consumption in kwh o f connected households has not changed in major way. Using the tariff structures post-2000, and assuming stable levels o f consumption for households over time, the above decomposition was implemented with the new tariff structures, and the results didnot change inany substantial way'43. The poor targeting performance o f the electricity consumption subsidy (a)i s driven by the comparatively low connection rates to electricity amongthe poor as opposed to the population as a whole (low access rates in the neighborhoods where the poor live in comparison to the overall access rate and low take-up rates where there i s potential access among the poor in comparison to the overall take-up rate)'44. Finally, the low value o f i s also due to higher consumption levels among the population connected as a whole as compared to the consumption amongthe poor benefiting from connections and the lower tariffs. 14* Under the existing VDT tariff structure and with the lifeline at 100Kwh for the first bracket of consumption(see Annex V for detailed data and methodo1ogy)For the simulations, we use the tariff structure that prevailed at the time, namely the December 2000 tariffs presented in Annex 111tables. Ifan Inverted Block Tariff(IBT) structure had been applied instead o f a VDT, R would have taken a value o f 0.417. The value o f l2 does not change significantly ifthe lifeline value for the first bracket o f consumption is modified. '43 Remember that the value o f the parameters A and Udo not change between 2000 and 2006 according to these simulations, since we do not have data on connection patterns after the survey; the fact that the values of R change very little with the new tariff structure may be due inpart to the fact that we need to assume due to data constraint no change inconnection patterns as well. It couldbe that there hasbeen a slightincrease inconnection rates over time, inwhich case the value o f R m a y increase a bit. But this may be compensated by the fact that the increase inthe cost per kWh has been slightly higher at the lower brackets o f the cost structure than at higher brackets, which would tend to reduce the value ofl2 slightly. 144 With an InvertedBlock Tariff(IBT), becauseall householdswho are connected to the networkbenefit from the lower tariff rates for the lower consumption blocks, the targeting parameters Tpand TNare both equal to one. With the VDT, the parameters are only marginallybelow one, becausethe great majority o fresidential customers consume less than300 Kwh. 110 An alternative to benefit the poor is to provide connection instead o f consumption subsidies. Connection subsidies may help the poor to connect to the network, because many households today cannot afford the cost o f connection to the grid which cost around 1.5 million dobras (US$ 125). To compare the two subsidies, three alternative connection scenarios are de~eloped'~~: first, the households who receive the connection subsidies are selected randomly from the population without a connection today, which from a measurement point o f view would be the same as giving a connection subsidy to all the population not connected today (scenario 1 in Figure9). Second, the beneficiaries for the connection subsidy are selected from the population with access but not usage or take-up, which is the scenario 2 in Figure 10 (this means that to benefit from access, households must not use the service but live in an area where the service exists). Third, the beneficiaries for the connection subsidy have the same distributional characteristics than the households connected today, for example because o f a complete lack o f targeting kin the connection subsidy design and a low access rate preventing trickle down to the poor from the increase inconnection rates made feasible thanks to the connection subsidy. Connection subsidies in all three scenarios benefit the poor more compared to consumption subsidies, with scenario 1 and 2 being clearly pro-poor. Connection subsidies have the potential to be better targeted than consumption subsidies, and compare to results from Latin America (Figure 1l)14? However, note that inpractice this will depend on how the mechanism for providing eligibility for connection subsidies would be im~lemented'~'. Figure 13: Connectionvs. consumptionSubsidies in Sao Tome and Principe 0 0 2 0 4 O B 0 8 1 1 2 1 4 0m.g. Source; Authors using STP's 2000/2001 survey. 14'The methodological details for considering those different scenarios are explained inAngel-Urdinola and Wodon (2005) 146by Estache et al. (2002) . 147 While connection subsidies have potential, they need to be implemented well to ensure good targeting and limit costs. Intheir study on social water connections in Abdijan and Dakar, Lauria and Hopkins (2004) explain how social connections were financed through a Water Development Fund paid for through a surcharge on water tariffs. Poor targeting resulted in 90 percent of residential connections in Abidjan being eligible for the subsidy. In addition the program suffered from distorted incentives as flat fees paid for each social connection to private operators were an incentive for them to maximize the number o f subsidized connections while at the same time seeking for these social connections richer households likely to consume more and located closer to the pipes. 111 Attempts at reform over the last five years There were several attempts to improve EMAE's financial, operational and commercial status. All efforts took, however, the form o f a piecemeal crisis interventionand were affected by frequent changes inmanagement and governments. Therefore, they could not eliminate EMAE's financial, operational and commercial problems for the long-term. InSeptember 2001 the Government appointed a three person team of Saotomeans for a period of two years under a management contract to turn around the then failing enterprise. The management contract included a management fee but lacked any specific performance incentives. 148The three person team achieved a net profit by the end o f 2002 mainly by reducing the size o f the labor force by 37 percent (from 295 employees in 2000 to 187 in 2002). The change in government ended the management contract in 2003 and three other management changes occurred between 2003 and June 2006.14' Before the management change o f June 2006, some long standing suggestions o f previous management teams were implemented: i. Forinstanceanewbillingsystemwasinitiatedtoallowcustomerstwobillingscycle before cutting service for nonpayment. If their service was cut-off, the customer were obliged to pay 50 percent o f the amount hehhe owed plus the re-installationfee. The customer was given a maximumo f six months to completely pay the amount. 11. The cycle o f payment arrears between EMAEsmajor supplier, ENCO, and its client, Government was tackled with a new agreement. Non-payment o f its services by the Government was preventing EMAEto pay its supplier, ENCO, which intwn was not able to pay its taxes to the Government. InMarch 2006, EMAE cleared it arrears o f close to 18 billion dobras (US$ 1.5 million) to ENCO while Government paid 10 billion dobras (US$ 830,000) o f its own arrears to EMAE. A Memorandum o f Understanding (MoU) was signed by all three parties to prevent future arrears. ... 111.Finally, the long delayed efforts to launch the pre-payment technology was undertaken on a test case basis. It's implementationwas slowed by social resistance and the poor quality o f electricity supply. Management has decided to delay installationo f full system untilelectricity supply becomes more reliable. Pending Private Entry into Generation The authorities, guided by the NPRS and their MDG goals, are aware o f the need for investment to expand coverage and improve the quality o f water and electricity services. They are inthe process of opening electricity productionto the private sector, and diversifying the source o f their energy. Overall, the authorities would like to reach a production capacity of 30 Megawatts over the medium term. InM a y 2007 the authorities signed an understanding with a I4*Themanagementfee alonewas consideredto act as a monetaryincentive for the membersofthe teamto performits verybest.. Eachindividualmember ofthe threepersonteamwas reportedto havebeenpaidapproximately around US%2,500 per month, a salary somewhat above the regular salariesreceivedby EMAE'sprevioussenior management staff. 14'A new managementtook over in 2003. EMAEhad another change intop management in March2005 which only lasteduntilJune 2006. 112 private investor for nine mini-hydroelectric units to come on line over 2007-2008. The investment i s supposed to total some US5 dollars and provide a total o f 8-9 Megawatts. Also, in April 2006 the government signed an investment agreement (also known as an Independent Power Producer (IPP)). We are negotiating a thermal electricity production Purchasing Power Agreement (PPA) with another private in~estor'~'.PPA i s to provide 6 Megawatts o f electricity intwo years withthe possibility of at total of 15 Megawatts inthe mediumterm. The investment i s to cost US$25 millions. If the Purchasing Power Agreement (PPA) were to be signed, the Government, as the owner o f EMAE and guarantor o f the PPA, agrees to its potential contingent liabilities on the central budget. The government will therefore need to be aware that there are several concerns about the manner inwhich private entry into electricity generation has taken place: EMAEdoes not have a strategy and development vision to guide it inits investment and growth decisions. The process o f choosing a company was not done in competitive and transparent biddingapproach, butrather throughun-solicitedproposals by different investors. The private investor has changed its original design and i s now proposing a thermo plant not dependent on fuel but on coal, with possibly serious environmental implications for EMAE and the country in terms o f supply and pollution issues. The 2004 economic impact and environmental analysis is therefore outdated as it focused on light fuel compared to the current coal fuelled thermal units. The financial terms proposedby private investor do not seem favorable to the country as the PPA contains both a commercial risk and a foreign exchange risk for EMAE and indirectly for the Government. As o f October 2006, the private investor was asking for Euro 11 cents, close to the current EMAE average tariff which i s Euro 11.5 cents (US 15 cents or 1805 dobras). Also, as the dobra has depreciated notably in the past two years, there is concern that the exchange rate risks are high in agreeing to pay for the electricity production in Euro (even if the final payment will be made inlocal currency). The proposed construction cost o f US$ 25 million for a 20 MW plant. This amount seems higher than comparable IPPs already built and operating in West Africa. For example in 2000 and 2001, a similar size IPP with a production capacity o f 20 MW inMalicost Euro 9.2 million; 13 MW inBurlunaFaso for Euro 8.5 million; and 18.9 MW inGambia for Euro 11.28million. 0 EMAE is not yet in a position technically to distribute the additional production through its existing transmission and distribution network. This i s one reason why the authorities are more interested in ensuring a phased increase in production capacity and electricity purchase, using the interim period (2-3 years) to bring in large clients on board and improve the transmission and distribution network. 0 Also, EMAE needs to improve bill collection before agreeing to purchase large amounts o f electricity (the original PPA noted purchasing a minimumo f 15 MW and paying US$ 7.8 million per year, the current values are being negotiated). This i s I 5 OThe Treasury signed the IPP agreement, an agreement between the investor and the Government with the tax incentives awardedto any other investor. Treasury receivedabonus ofUS$13,000 upon signingthe contract. 113 especially worrisome in light o f the on-going practice o f accumulated payment arrears, especially those o f the Government EMAE has already established a Review Commission to study the impact o f the PPA if the Government proceeds with the already signed IPP agreement. The Review Commission requested technical assistance fiom Electricit6 de France (EdF) to explore the financial and environmental impact, and develop a strategy to alter the PPA and making it more favorable to EMAE's as well as country's fiture needs. The authorities may wish to consider a more thorough analytical (technical, environmental and financial) approach and a strategic view to the sectoral developments in general (production, distribution, collection) and to private entry into the country's electricity and water sectors before finalizing any IPP with private investors. 114 ANNEX111.2: PERFORMANCEOF ENASA,ENAPORT AND CORREIOS This annex provides a brief review o f the performance o f each SOE-- financial and operational to the extent o f available data ENASA (Airports Security and Administration) After EMAE, ENASA i s another SOE with financial difficulties as it tries to manage the country airport's security, operate and maintain its commercial activities on behalf o f the Government. The Sao Tome airport was originally constructed in 1948 and passenger terminal, technical building an air traffic control tower and parking facilities were added only in 1992. Most recent investments were funded by multilateral and bilateral organizations. The Government o f Taiwan was among the key grant providers to improve the airport services. The airport has a small (only 750 square meter) passenger terminal with limited passenger activity depending on a few airlines such as TAP Air Portugal, Air Luxor, TAAG Angola Airlines, and Air Service Gabon which provide service to Lisbon, Libreville, Luanda and Principe. International services to Lisbon, Portugal i s once a week whereas the continental services are on a few times a week. Table15 Summary Informationon ENASA's ProfitbossandBalance Sheet (US%) Profit/Loss Statement 2000 2001 2002 2003 2004 2005 Sales Revenue 712,263 1,132,291 1,219,345 1,475,752 1,163,932 883,254 OperatingExpense 547,168 1,364,393 1,340,869 1,627,355 1,425,680 1,060,629 Of which wage bill 437,734 1,0915,5 14 1,072,695 1,140,544 Net Profit/Loss 165,095 -232-102 -121,524 -1,301,884 15 1.603 -26 1.749 -848,503 177.375 Bnlnnce Slieet Liabilities 179,216 185,831 243,861 250,573 288,402 321,589 Equity 1,949,898 1,694,901 1,527,717 1,376,782 1,137,278 739,040 Assets 1,963,084 1,740,350 1,650,054 1,475,752 1,163,932 883,254 Finrinciril Ratios ProfitMargin 23% 12% -10% -10% -22% -20% Return on Equity 9% 8Yo -8% -11% -23% -24% Return on Assets 8Yo -13% -7% -10% -22% -20% LiabilitiesiEquity 9% 11% 16% 18% 25% 44% Source: Compiled from various financial data received by ENASA; ENASA does not have any long term debt; its liabilities are its short-term debt to its suppliers. ENASA operates under its own "tutelle" Ministry o f Public Works and Infrastructure and it employs a total o f 130 workers and 113 (86 percent) o f which are permanently employed and the rest (14 percent) work on contractual basis. ENASA's annual revenues o f close to US$ 1 million come from landing fees, runaway lighting, passenger charges, cargo, aircraft parking, and handling charges. Landing fees and the passenger charges are the top revenue sources. Besides, it derives additional commercial revenue from renting the facilities, auto rental services, advertising, car parhng and refueling services provided to airlines. This stream o f revenue seems to be in early stages o f development but it has the potential to grow as the airport i s expanded. Similar to EMAE, ENASA too suffers from un-paid bills o f the Government and in turn it can not pay its own suppliers and has accumulated supplier arrears to EMAEfor the use of its electricity and water. In2005, the wage bill constituted the bulk, 80 percent (US$ 848.503) o f its operating expenses; the remaining 20 percent (US$ 183,254) was spent on security, cleaning, maintenance, communications, energy and water (services provided by EMAE). ENASA has not 115 been able to invest to upgrade its facilities. It has been running deficits over the last few years and unable to transfer any profit sharing to the Treasury besides paying its taxes. Its profit margin declined from 23 percent in 2000 to 12 percent in 2001 and remained negative between 2002 and 2005. ENASA has been eating up its equity and accumulating short term liabilities to its suppliers which as doubled over the last six years. Its equity decreased by half over the same period. (See Annex 2) Figure 14: ENASA's ProfitMargin as Net ProfitLossover Sales 2000 2001 2002 2003 2004 2005 ._ - __ ~ Source: Compiled from financial data analyzed in Annex 2. ENASA's management i s concerned about the enterprise's ability to clear its own increased liabilities (short-term only). These liabilities are the accumulated arrears to its suppliers (EMAE, CST Telecom, ASECNA, and SAT) and tax arrears to the Government. By 2005. EMAE's combined supplier and tax arrears reached 3,859,068,000 dobras (USS 321,589), one- third of its sales revenue. Half of this, 1,986,579,306 dobras (USS 165,000), constituted its supplier arrears to EMAE alone. ENASA IS not able to pay the social secunty payments on behalf o f its employees either."' Figure 15: ENASA's IncreasingLiabilities 2000 2001 2002 2003 2004 2005 1 Short-term 1 Source: Compiled from financial data inAnnex 2. ENASAhas a longterm operationscontract with ASECNA of Senegalfor airport traffic security butit hasnot been able to pay its bills. 116 Like EMAE, ENASA suffers from delinquent customer payments. By 2005, its unpaid bills from all its customers reached 2,396,021,804 dobras (US$ 200,000), a quarter o f sales revenue, almost half o f this or 1,128,566,087 dobras (US$ 94,000) was alone the responsibility o f Air Sao Tome (Air STP), partially state-owned and bankrupt national airline.15' ENASA is waiting for Government's approval for corporatization to be transformed into S.A. However, the airport security has to be separated from ENASA while the latter becomes a separate corporate entity. Next step would to design a private management contract and tender it competitively and also seek other sub-contracting opportunities to farm out to private sector (domestic or international) and improve operations. ENAPORT (Seaports Security andAdministration) ENAPORTi s incharge of administering the two ports (Sa0 Tome Port and PrincipePort) with 158 employees and generates annual revenues close to US$ 2 million. While both EMAE and ENASA are running large deficits, ENAPORT (as well as CORREIOS) generates a small net profit and it i s apparently able to pay its profit sharing obligations and taxes to the Treasury. It appears that it has recently receivedin2005 an "investment subsidy" (equity injection) equivalent o f US$ 237,844 from the PIP in order to maintain operations, it complains of not having the financial resources to maintain its obsolete and ailing equipment and it desperately needs new investments into its facilities. Table 16: Summary Informationon ENAPORT's ProfitLossStatement and BalanceSheet (US$) ProfithossStatement 2000 2001 2002 2003 2004 2005 Sales Revenue 1,350,776 1,229,512 ,329,225 2,233,701 2,157,666 ,822,422 OperatingExpense 1,335,362 2,415,739 ,198,386 2,186,163 2,070,240 ,820,303 ofwhich the wage bill 237,307 376,188 395,361 439,380 511,949 525,259 Net ~r~~~~~~~~~~ 15,414 171.635 130.S3S 37.538 57,426 2,139 Bukince Sheet Liabilities 456,861 207,412 414,229 644,582 1,053,190 772,729 Equity 247,390 574,867 343,113 445,239 281,771 632,822 Assets 704,25 1 782,279 757,342 1,089,822 1,334,961 1,405,552 Fin(inriu1Rcitios ProfitMargin 1% 14% 10% 21% 4% 0% Returnon Equity 6% 30% 38% 11% 31% 0% Returnon Assets 2% 22% 17% 4% 7% 0% Liabilitiesmquity 185% 36% 121% 145% 374% 122% Source: Information received from ENAPORT's financial statements; the balance sheet for 2003 was problematic and the consultant constructed her own approximation to calculate "owners' equity." In2005, ENAPORT has apparently received an equity injection, the only one among all four SOEs. 152Informationprovideduponrequestfrom ENASA to the consultant, "Divida dos principaisclients da ENASA e dividascornprincipaisfomecedores." 117 Figure 16: ENAPORT's Profit Margin as Net Profit/Loss over Sales 2000 2001 2002 2003 2004 2005 1 ProfitMargin 1 Source: Compiled from financial data analyzed inAnnex 3 ENAPORT too suffers from its inability to pay for its accumulated liabilities which are only short term. However, they have been increasing over the last six years and its liabilities over equity ratio has remained quite high over the same period ranging from 185 percent in 2000 to 145 percent in 2003 and to 122 percent in 2005. ENAPORT's operations continue to subsidize not only EMAEbut also indirectly another partially s4ate-owned enterprise, ENCO and therefore the state. As discussed earlier, ENCO has refused paying its bills and ENAPORT has stopped billing ENCO. ENAPORT has no other choice but to revert to supplier arrears itself thus accumulating short-term debt to its own creditors. Figure 17: ENAPORT'sIncreasing Liabilities 2000 2001 2002 2003 2004 2005 I L Source: Complied from analyzed financial data inAnnex 3. ENAPORT's management is eager to change the enterprise's legal status into SA. creating a Port Authority owning and managing all the assets of the port for the state and 118 operating commercially. It hopes to bring in a private operator through a concession agreement to manage its container terminal. The latter has become increasingly important as the container traffic i s growing rapidly. Even though the fees are not changed during the last few years, the container revenues constitute the bulk o f ENAPORT's sales. Once the Port Authority i s created, the Government can then incorporate the container terminal as a wholly owned subsidiary and seek to award a long term license to a private operator (domestic or international or both) to operate the container terminal. As the Port Authority becomes a financially autonomous public corporation, the container terminal can be set up as a legally private limited company but under public ownership (Government owning 100 percent of its equity). The private operator or the concessionaire can propose a management policy of ensuring the effective deployment o f workers and machines; increasing dockside equipment; providing labor with better working conditions, higher wages, and improved benefits and incentives and generally treating labor as the most important factor in the organization; improving work systems (including changing and improving operational procedures relating to ship and box transport); developing direct-user relations, enhancing employees' skills, and restructuring organization. Once established, the new Port Authority needs to review the successful international examples such as the concessioning the Kelong Container Terminal in Malaysia and Container Terminal in Dar es Salaam in Tanzania, both are among the most successful international examples in terms o f significant efficiency gains and welfare benefits to workers, consumers, government and the enterprise itself. CORREIOS(Postal Services) CORREOIS i s a small state-owned enterprise with only 53 employees and few post office branches, and it has the capacity to handle limited items per year. Its current annual revenues o f US$262, 712 hardly cover its operating costs generating a very modest profit. Even though it has its own budgets, it operates almost more like a government department with no separation between owner, operator and regulator. However, the enterprise has been operating on negative equity. Its "investment subsidies" or the equity injections it receivedin2001 are already depleted. Its liabilities that are only short-term have been increasing constantly and reached US$ 1,249,757 (five fold o f its annual revenues) and far above those o f ENAPORT and ENASA. Table 17: Summary Information on CORREIOS'SProfitLoss Statement and Balance Sheet (US$) ProfWLoss Statement 2000 2001 2002 2003 2004 2005 Sales Revenue 563,227 277,605 269,652 305,112 359,874 262,712 Operating Expense 417,918 263,147 350,109 490,394 323,610 228,851 Net ~ ' ~ ~ ~ ~ / ~145,?0`) o ~ s L4,JSS -so.457 1x5,2s2 36.264 33,851 ~ Balance Sheet Liabilities 547,916 699,351 839,936 1,293,083 1,254,641 ,249,757 Equity -6,645 -185,075 -314,248 57,150 8,083 -100,256 Assets 541,270 514,275 525,688 1,350,233 1,262,723 ,149,502 Finunciul Ratio5 Profit Margin 26% 5% -30% -60% 10% 13% Return on Equity -2187% -8% 26% -324% 449% -34% Return on Assets 27% 3% -15% -14% 3% 3yo Liabilitiesmquity -8245% -378% -267% 2263% 15522% -1247% Source: Information received from CORREIOS'S financial statements; the balance sheets for all six years were difficult to dissect; CORREIOS received "investment subsidies" from the Treasury in2001, 2002 and 2003 as equity injections and the company still has negative equity. 119 Like ENAPORT, CORREIOS too has a forward looking and pretty entrepreneurial management. It has been well aware o f the international trends and change in the demand and technology for postal services to encompass differing forms o f communication, transportation and delivery support, and financial and other related services. It follows the on-going reforms and liberalization o f markets and globalization on exchanges o f goods and services, and the increasing role o f private participation. The new management i s now seehng commercial freedom and a clear market orientation. It has entered a long relationship with CORREIOS Portugal to create a mixed enterprise. The inaction on the part o f the Government in 2003 interruptedthe negotiations and now the international private partner seems not interested. Figurel8: CORREIOS'SProfitMargin as Net ProfitLossover Sales 2000 2001 2002 2003 2004 2005 pGGi&K Source: Compiled from financial data analyzed inAnnex 4. Figure 19: CORREIOS'SIncreasingLiabilities 2000 2001 2002 2003 2004 2005 1 Short-term - - . _ Source: Compiled from financial data analyzed in Annex 4. CORREIOS i s also seeking a new legal status allowing for a high degree o f commercial freedom and flexibility, precise definition o f universal service obligations and clear distinctions and accountabilities between the government, regulator and operator. The Government needs to recognize the benefits to be gained from a reformed postal system. Most and foremost, the postal system needs to improve postal operations and efficiency through expanded use o f information technology and expand access to the information infrastructure through the post office. It also 120 needs to consider the introduction of new services as the population i s likely to demand in the future: 0 Billpayment services (money order andpostal savings); 0 Pension and other benefits service; 0 Informationrepository andmarket/data analysis services. Even if it can not find a private investor, the management can enter into a twinning arrangement with another modernized postal system in Brazil and Portugal to improve and modernize services. 121 0 8 0 5 0 8 z 0 N N V m w v - 0 8 z 0 D h P rA g z a m 9 c ! 2 s 2 2 a C E 1C -i c r c h $ 8 8 a"-5 m m 0 0 b b '",v! a $6 3 $6 3 d W 3 W i5m k U9mtg e8 I c, U9 a f .I .I ? c, e i f c, 8 E 0 9 .I 0 c, 0 B1 N z+ L v) 31 9 c, 0 22 8Um Bm k '13 m e m tg I I c, m 0 c, v1 B - v) Q 0 N U zi P a LI 0 30B e a.. 0 p c.l Q) ANNEX111.4:POVERTYASPECTS OF ELECTRICITYTARIFFSINCREASES: Data and Methodology Inthis section, we turn to the issue of the distributionalproperties of existing subsidies and tax breaks for EMAE, and to the potential impact o f a tariff increase on the poor. In section 3.1, we start by providing basic statistics on access rates to electricity among residential customers, and consumption patterns. Section 3.2 then presents a framework for assessing the distributional properties o f existing subsidies. The empirical results obtained on the targeting performance o f the subsidies are presented in Section 3.3. Section 3.4 discusses the potential targeting performance o f connection subsidies. Finally in Section 3.5 we discuss the potential impact on poverty o f an increase in tariffs, and also compares the distribution o f spending on electricity subsidies and tax breaks with that o f other services provided by the government o f Sao Tome and Principe, with a focus on education and health. Comparisons are also provided in that section between government spending for education and health and the total implicit subsidies and tax breaks provided to EMAE as well as the losses incurred by the utility, which will all ultimately be paid by the state either right now or in the future when existing losses are transferred into short or long term debts. Data and Basic Statistics The analysis in this section i s based on data from the ECVM survey (Enquete sur les Conditions de Vie des Mknages) conducted in 2000-01, which estimated that 54 percent o f the population was poor at that time and 15 percent lived inextreme poverty, with standards o f living being especially low inrural areas (particularly in the north o f SBo Tom6 island and the island o f Principe). These estimates of poverty are based on a relative moderate poverty line equal to two thirds o f mean per capita expenditure or 2,638,618 dobras per year (about US$294/year or 0.81 US$/day). The estimated total number o f residential customers i s about 7,000, representing approximately 40% of the population. The estimates obtained with the ECVM are very close to that proportion, as shown intable 8. The table indicates that 42% o f the households have access to electricity, even though some connections are available in the neighborhoods where households live in close to 85% o f cases (this means that only one in two households actually connects to the network where the network i s available inthe area where the household lives). As expected, access to electricity, both at the neighborhood level and at the household level, i s lower among the poorer deciles o f the population (each decile accounts for ten percent o f the population, and the deciles are ranked according to the level o f per capita consumption o f households). 125 ANNEX 111.5: ASSESSING THE TARGETINGPERFORMANCE OF CONSUMPTION SUBSIDIES A FRAMEWORK'53 Define 0,as the share o f the implicit subsidies inthe tariff structure receivedby poor households divided by the share o f the poor inthe population. Inmathematical terms, where S denotes the nominal implicit subsidies for the poor (with subscript P and for the population as a whole with subscript H), P denotes the number o f poor households and H represents the total number o f households. IfR takes a value o f one, this implies that the subsidy i s roughly neutral from a distribution and poverty point o f view, with the share o f benefits going to the poor equal to their population share. A value above (below) one implies that the program i s somewhat pro-poor (not pro-poor) since the poor benefit fkom a large (smaller) share o f the total benefits than their populationn share. Inpractice, manypoor householdsinthe populationare not subsidyrecipients andthusthe value SZ i s usually lower than one. As discussed in Komives et al. (2005) who provide analysis with data for many different counties, this occurs due to several reasons. For example, access to networks may not reach poorer areas. If we denote access to networks by A, this means that in many cases AP< AH. Also, poor households, albeit having access to the networks, may be less likely the population on average to use the services due to lack o f affordability. Ifwe denote the share o f all households who have access to the service intheir neighborhood and use the service connections among those with potential access, AH * as UHlA, this would meanthat UHlA UPIAmost cases. While < in UHiA represents the take-up rate o f UHIA represents the actual household connection rate, with the same relationships for the poor (subscript P). the next key parameter i s share o f eligible utility service users (i.e. households with access, a connection, and meters if applicable) who are targeted and therefore receive a subsidy. This i s denoted as THlU. when all consumptioni s subsidized, all users have access to the subsidy and therefore THlp= THlU = 1, The share o f households receiving the subsidy was equal to share o f all connected household times the share o f households eligible for subsidy (i.e. AH * UHlA * T HIU). Similarly, the share o f poor households receiving the subsidy i s AP UplA T * * We will use below the variable B to capture this beneficiary incidence, so that: B P = * * TPIU A second important variable for assessing the targeting performance of subsidies is the rate o f subsidization or the difference between what households pay per kwho f electricity versus what it actually costs to produce, transmit, and distribute it. Denote the average unit cost o f producing, transmitting and distributingthe good by C. Then the average rate o f subsidization i s R = 1 EH~T / (QHiT * - C), with QHlT being the average quantity consumed by subsidy recipients and EHlT being their average expenditure on electricity. Again, these parameters can be estimated for the poor as a group (subscript P instead o f H). The average subsidy benefit per household receiving '53 This section essentially replicates part of the materialprovided in Angel-Urdinola and Wodon (2006a, 2006b). 126 (and per poor household) inthe population as a whole (and among the poor) can then be written as The benefit targeting performance indicator 0which again represents the share o f the benefits o f the subsidy obtained by the poor divided by the share o f the poor inthe population i s equal to: Thus five ratios determine the value o f the overall performance parameter f2 : access, uptake, targeting, rate o f subsidization, and quantity consumed. The ratio o f access rates (A) will inmost cases be lower than one simplybecause the poor tend to live in areas without access to electricity. The usage or "take-up" ratio (U) will also be lower than one if the cost of connecting to the network is high for the poor, or if they live further away from the grid even when there i s access intheir neighbourhood. This means that the product of the R and Q ratios must be greater than one for the subsidy to be progressive, but as we will see in the case of Rwanda (and as has been observed elsewhere), this i s rarely the case. 127 ANNEX 111.6: INCENTIVE-BASED PERFORMANCECONTRACTSFOR SOES As a tool for reforming and evaluating the performance o f SOEs, PCs were designed and implemented in several countries ranging from South Korea, to India and to France and later in West and East Africa in the 1980s between the governments and the SOEs' public managers. They were often monitored by a third party or an oversight agency. While South Korean (signaling system) and Indian PCs (Memorandum o f Understanding or MoU) were based on performance information, performance evaluation and performance incentive system, the French style contract plan" lacked all those elements and they were not incentive- based and therefore it " was difficult to evaluate the performance o f the public managers and get results. Those French style PCs were therefore not successful while their well-designed counterparts in South Korea and India led to significant performance improvements. Unfortunately South Korean and Indian PCs did not make it to Sub-Sahara, Africa. To a large extent not-incentive based PCs emulating the French style contracts were introduced and implemented in few selected countries in Sub-Saharan Africa. Senegalese PCs were reviewed widely. 154 Since the outcomes were less than satisfactory, the international community dismissed them as not a reliable enterprise reform measure. However, the recent experiments with privatizing the state-owned infrastructure enterprises did not always lead to remarkable performance improvements either. It mightbe timely for the African governments including Sao Tome and Principe who are not ready to privatize their SOEs to re-consider the PCs. Nonetheless, the international community needs to show some willingness to put in the similar kind o f technical assistance and training to drafting and implementing PCs as it did into privatization in Sub-Saharan countries inthe past. The optimal use o f performance contracts can be made if they are part o f a reform package which advocates mechanisms to evaluate SOE performance (fully and partially state- owned). A management improvement system through incentive-based performance contracts does then needto have three elements: Performance evaluation system translates national goals into explicit and quantifiable entrepreneurial targets captured in performance criteria which are fair to the manager as well as being simple and easily monitorable; each performance criterion has to have a criterion value and targets. Performance information system monitors and measures the actual performance o f an SOE using the accurate information needed to evaluate the management performance; Performance incentive system directly links the well being o f the managers and employees o f and SOE to national welfare through a system o f bonuses, either monetary or non-monetary. Managers are thus either rewarded or penalized for their performance, based on the contract goals and criteria and calculated composite score, at the end o f each year o f the contract period ranging from 3 to five years.'55 `54Nilgun Gbkgur, "Performance Contract Experience o f SENELEC and SONATEL in Senegal," Teaching Case prepared for the former Harvard Institute for International Development (HIID), Harvard University, 1990. Is'NilgunGokgur, `Performance Contracts: A Tool for Reformingand Evaluatingthe Performance o fPublic Enterprise," paper delivered at FirstEvaluation Seminar for Central America, Panama & the Dominican Rewblic by the Inter-American Develoument Bank, San Jose, Costa Rica, 1994. 128 It is important to note that the performance contract system must be coupled with long- term planning, which means that every enterprise has to develop medium and long-term plans that are consistent with the priorities established by the government through sector ministries, namely the PRSP. The overriding principle for preparing a performance contracts remains that they are based on simplicity so that their content can be easy to monitor and to evaluate. 129 ANNEX111.7: STANTARDIZED FISCALTEMPLATE Questionnaire on Enterprise Governance Questions Allowed Responses - 1) What percentage of the enterprise i s ownedby Percentage(sum=lOO%) a) Local private sector b) Foreignprivate sector c) Employees .I 6a d) Central government t L e) Municipal government 0 f ) Others 3 2) What i s the current legal status o fthe Uncorporatized state owned enterprise=l ; Corporatized state enterprise? owned enterprise=2; Limitedliability shareowned company=3; Publicly trade corporation =4; None o f the above=5 3) Ifthe enterprise i s corporatized, inwhat year Year didcorporatizationtake place? 4) Does the utility have a Board o f Directors? Yes=l; No=Os 5) How many members does it have? Number 6) How were the Boardmembers Appointed by the supervisory authoripl; Selected by a group o f selectedappointed? shareholders=2, Electedby all shareholders=3; Other (specify)=4 7) Howmany o f the Board's membersare drawn Number from each o fthe following groups? a) Management b) Employees c) Supervisory Authority d) Central GovernmentAuthorities e) Local Government Authorities f ) Foreign owners/partners g) Banks h)Private sector experts independent fromthe state and management: i)Others(specify) 8) Overall, who has the most decisive influence Management=l ;Board=2; on the following activities/decisions o f the Government=3; Other =4 enterprise? a) Hiring o f workers b) Laying offo fworkers c) Setting wagesibonuses d) How much to produce e) Who to sell to f ) How to set prices g) Approval o f individual investmentprojects h) Onreceipt ofpublic subsidyandits use 130 9) Are any management or performance contracts made between your fmand ownershit, entities? I Yes=l; No=Os 10) D o you have performance based incentive systems inwhich payments and promotion o f managers are determined by their performance? Yes=l; No=Os 11) Ifyes, when performance basedincentive systems have been introduced? Year 12) Are there penalties for poor performance o f managers? Yes=l; No=Os 13) Ifyes, when penalties for poor performance o f managers have been introduced? Year 14) I s your Derformance monitored on a regular basis? Yes=l: No=Os 15) What i s the monitoring period? Monthly= 1;Quarterly=3 ;Semi- annually=3; Annually-4; Other=5 16) D o you have external auditors? Yes=l; No=Os Specialized state audit firms=l; 17) Who is authorized to audit your firm? Private sector auditors selected by you=2 Other=3 18) Are your annual reports o nperformance available to the public? Yes=l; No=Os 19) Does your fmfollow the International Financial Reporting Standards (IFRS)? Yes=l; No=Os 20) Has the enterprise hadan independent audit o f its accounts? Yes=l; No=Os 21) Ifso, when was the first year that such an independentaudit was carried out? Year 22) Is the audit Dublic? Yes=l: No=Os 23) I s the enterprise required to earn a rate o f return? I Yes=l; No=Os 24) I s the enterprise required to pay dividends? Yes=l; No=Os 8 25) I s the enterprise exempt from any form o f taxation (e.g. VAT)? 9 Yes=l; No=Os tz O=below the market rate, l=equal to vi 26) At which rates does the enterprise have access to debt? the market rate, 2=above the market 27) I s the enterprise remunerated for non-commercial activities by the state? Yes=l; No=Os 29) Has the enterprise ever been granted a state (centralhegionalllocal) guarantee? Yes=l; No=Os 30) I s the enterprise at liberty to lay-off workers if needed? Yes=l; No=Os Comparable to the public sector=l; 30) H o w would you characterize the wageshalaries paid by the enterprise?? Comparable to the private sector=2; Somewhere inbetweenuublic and private sector levels=3 31) Has the enterprise contracted out significant activities to the private sector under service contracts? a) Billing and collection b)Humanresource I Yes=l; No=Os c)IT services d) Others (Specify) 131 ANNEXIV: PERTABLES Table 1: Sao Tome & Principe Gross Domestic Product (Current price- Millions of Dobras) 2001 2002 2003 2004 2005 2006 Primary Sector: 133,060 164,397 192,831 238,430 201,940 Agriculture 90,927 112,341 134,437 169,623 124,825 Fisheries 42,133 52,055 58,394 68,807 77,115 Secondary sector: 113,971 140,812 162,876 222,509 246,690 Construction 62,784 77,570 90,235 129,813 143,581 Utilities 9,754 12,052 13,844 15,472 21,141 Other 41,432 51,190 58,797 77,225 81,967 Tretiary sector: 375,613 464,072 504,317 530,710 669,704 Commerce 155,492 192,111 210,307 207,418 236,747 Hotel and Restaurant 9,726 12,017 14,977 16,456 23,911 Transport 91,938 113,590 120,875 128,418 208,224 Financial institutions 15,809 19,533 21,346 24,101 26,281 Public administration 44,448 54,915 56,193 60,804 60,225 Other services 58,200.34 71,906.97 80,619.24 93,513.36 114,315.50 Gross Domestic Product 676,071 835,291 929,287 1,060,891 1,188,718 1,526,338 Source: IMF Table 2: Sao Tome & Principe Gross Domestic Product (Constant price- Millions ofDobras) 2001 2002 2003 2004 2005 2006 Primary Sector: 133,060.0 148,475.8 157,246.5 159,700.9 163,999.2 Agriculture 90,927.3 101,461.8 107,549.5 108,732.5 111,450.8 Fisheries 42,133 47,014 49,697 50,968 52,548 Secondary sector: 113,971 127,175 139,702 145,855 152,578 Construction 62,784 70,058 78,465 81,133 84,459 Utilities 9,754 10,885 12,038 12,893 13,731 Other 41,432 46,232 49,198 51,829 54,387 Tretiary sector: 375,613 419,130 447,126 476,197 509,583 Commerce 155,492 173,506 182,876 193,848 205,867 Hotel and Restaurant 9,726 10,853 13,023 13,714 15,771 Transport 91,938 102,589 107,924 113,644 118,985 Financial institutions 15,809 17,64 1 19,405 20,958 23,053 Public adrmnistration 44,448 49,597 51,085 52,873 55,252 Other services 58200.3 64943.2 72813.3 81160.5 90655.3 Gross Domestic Product 676,071 754,398 805,828 844,700 890,147 952,457 Source: IMF 132 Table 3: Summary of CentralGovernmentOperations (Billions of Dobras) 2001 2002 2003 2004 2005 2006 Act Act Act Act Act Est Revenue(including Oil Bonuses) 248.8 245.6 321.0 381.2 973.3 570.0 Revenue(includinggrants) 248.8 245.6 321.0 381.2 411.8 570.0 Tax revenue 77.2 94.9 114.2 147.3 180.0 267.0 Direct taxes 23.8 34.1 33.0 44.5 54.3 Indirect taxes 49.5 54.1 70.3 89.2 105.8 Other tax revenue 3.9 6.8 10.9 13.6 19.8 Non-tax revenue 12.0 18.4 27.4 31.3 27.4 57.0 Grants 159.5 132.2 179.4 202.6 204.4 246.0 Project andbudget support related 138.1 110.4 144.5 166.4 169.3 199.0 HIPC 21.4 21.8 35.0 36.1 35.1 47.0 Oil Signature Bonuses -- _ _ -- -- 561.5 0.0 Expenditureand net lending 335.3 325.6 415.1 548.5 528.7 783.0 Current expenditure 121.9 156.8 183.3 292.5 314.4 441.0 Personel costs 40.5 43.6 55.0 75.7 103.2 133.0 Goods and services 18.8 26.7 43.9 95.8 66.2 91.0 Transfers 13.1 40.7 35.9 65.9 83.2 108.0 Interest o n debt 24.8 25.9 26.1 30.1 33.7 52.0 Other 24.8 20.0 22.4 25.0 28.1 57.0 Capital expenditure (Public 194.3 140.0 197.9 234.9 186.4 307.0 investmentprogramme) HIPCfunded expenditure 17.2 22.5 33.6 30.1 27.9 35.0 Public ServiceRestructuring 1.9 6.3 0.0 0.0 0.0 0.0 Net lending 0.0 0.0 0.3 -9.0 0.0 0.0 Overallfiscal balance(commitment -86.5 -80.0 -94.1 -167.3 444.5 -213;O basis) Changes inarrears (net; reduction -) 3.8 48.8 28.3 41.0 -134.2 21.0 Overallfiscalbalance (cashbasis) -82.7 -31.1 -65.8 -126.3 310.4 -192.0 Financing 82.7 31.1 65.8 126.3 -310.4 193.0 External (net) 31.9 3.6 24.7 64.4 -80.8 31.0 Domestic 35.6 -1.8 16.0 27.0 -8.6 162.0 OilReserve FundFlows (net) 0.0 0.0 0.0 0.0 -263.2 0.0 Change inarrears (principal) 15.2 29.3 25.1 34.9 -101.9 0.0 Paris Clubrescheduling and debt 0.0 0.0 0.0 0.0 144.1 0.0 relief Sources: Ministryof Planningand Finance, Government Financial Operations Note: 2006 data i s from IMF Annex 133 Table 4: Summary of CentralGovernment Revenue (In billionsof Dobras) 2000 2001 2002 2003 2004 2005 2006 Act Act Act Act Act Act Est Revenue(including grants) 164.6 255.2 245.7 319.9 390.2 973.3 570.0 Non-oil Revenues 164.6 255.2 245.7 319.9 390.2 411.8 570.0 Tax Revenue 59.2 78.6 94.9 114.2 147.3 180.0 267.0 Directtaxes 16.3 23.1 33.3 27.9 41.8 54.3 Profit Tax 5.7 10.2 18.2 12.1 18.0 23.7 Salary Tax 10.5 13.0 15.1 15.7 23.8 30.7 Indirecttaxes 37.8 49.5 54.1 70.3 89.2 105.8 Import Duty 17.9 19.4 23.8 30.9 35.7 40.1 Imported Goods excise tax 19.3 28.3 28.5 31.8 46.4 56.2 Local Goods excise tax 0.7 1.8 1.8 7.6 7.2 9.5 Other 5.1 6.0 7.5 16.0 16.3 19.8 Non-tax Revenue 20.5 17.1 18.4 26.2 40.3 27.4 57.0 Transfer form enterprises 5.4 8.2 5.7 4.9 3.8 3.7 FishingRoyalties 8.1 2.7 1.6 9.8 7.2 0.4 Other 7.0 6.2 11.1 5.2 19.0 13.6 OilRevenue -- -- -- 6.4 10.4 9.6 Grants 84.9 159.6 132.3 179.5 202.5 204.4 246.0 budget support grants 17.3 0.5 2.4 20.4 17.5 2.0 Project grants 120.9 110 142.1 146 151.8 197.0 HIPC 21.4 21.8 35.0 36.1 35.1 47.0 Oil SignatureBonuses _ _ _ _ -- -- -- 561.5 0.0 Sources: Ministryo f Planning and Finance, Government FinancialOperations Note : 2006 data is from IMF 134 0 0 b 0 0 d N - r ooriowvidri - - N v) m 3 x o\ b 8 N N O O O W 0 0 0 0 0 9 8 2 - 0 : 8 ? " ? P O - 0 0 - 0 2 8 `3? 9 f 0 0 - 2 8 2 - 0 W 2 0 8 "9099 5 - 0 0 0 c? 0 8 Table 7: Government Expenditure by quasi-functional classification Budget execution (Billion Dobras) Functions 2000 2001 2002 2003 2004 2005 (Program) General Administration and political 8.7 11.3 14.5 24.0 31.2 45.9 institutions National Assembly 1.5 2.0 2.5 4.7 6.3 12.2 Local administration 0.0 1.1 1.3 2.2 2.7 2.6 Presidency o f the Republic 1.5 1.1 2.1 2.4 2.4 3.0 Prime Minister Office 0.7 0.9 0.8 1.5 2.8 3.0 Principe 1.6 2.0 2.4 4.3 4.9 5.7 Foreign Affairs (exc. Embassies) 0.4 1.1 1.5 1.9 2.0 3.4 Justice 3.0 3.1 3.8 7.0 10.1 16.1 Social Sectors 16.1 25.4 27.9 47.5 61.1 111.7 Media 0.5 2.4 0.4 1.2 1.8 2.6 Labour and Vocacional Training 0.0 0.4 1.3 2.1 2.0 3.1 Education and Culture 8.4 12.8 16.0 23.2 30.0 71.6 Youth and Sports 0.0 0.0 0.8 1.1 1.1 1.6 Health 6.6 9.0 8.4 17.9 23.8 28.9 Natural Resources and Environment 0.6 0.8 0.9 2.0 1.2 2.0 Infrastructures and Land 0.0 0.0 0.0 0.0 1.2 1.9 Economic Sectors 3.9 6.0 7.7 14.9 20.6 24.0 Finance (exc. Debt) 3.0 3.7 4.4 7.0 10.5 9.5 Economy 0.9 1.1 1.6 4.0 5.0 7.2 Agriculure and Fishing 0.0 0.6 1.1 1.9 2.4 3.4 Trade, industryand tourism 0.0 0.5 0.5 2.0 2.7 3.8 Other sectors 33.5 45.0 44.1 73.8 101.0 131.0 Defense 3.5 6.5 7.0 10.6 15.4 21.5 Public Debt 7.6 10.0 5.6 20.4 20.9 37.0 Common Expenditure 13.5 20.6 23.1 31.4 52.2 56.0 Embassies 5.5 7.8 8.5 11.4 12.5 16.6 Other 3.3 0.1 0.0 0.0 0.0 0.0 Total 62.1 87.6 94.1 160.3 214.0 312.7 Source: Ministry o f Planning and Finance; 'Orqamento Geral do Estado' Note: Values are on a commitment basis, or 'base de liquidaqlo' 138 Table 8: Government Recurrent Expenditure by economic classification (Billion Dobras) Economic classification 2000 2001 2002 2003 2004 2005 Personel costs 28.8 36.6 40.6 54.7 75.9 108.2 Goods and services 15.3 9.9 24.9 48.3 77.8 73.1 Interest on debt 10.9 8.9 5.6 20.4 21.0 32.6 Transfers 7.3 6.8 8.2 13.3 13.8 83.3 Other 11.5 9.8 13.4 19.6 20.4 24.1 Total 73.8 71.9 92.5 156.3 208.9 321.2 Sources: Ministry o f Planning and Finance, Government Fiscal Operations Note: Transfer includes JDA and scholarship funds Table 9: Current Tansfer from the State Budget (Billions of Dobras) 2000 2001 2002 2003 2004 2005 Public Administration 3.5 8.2 7.8 8.1 46.2 State budget 2.8 4.2 0.4 0.2 0.3 Autonomous services 0.0 3.2 4.1 4.6 31.9 Local Administration 0.7 0.0 3.3 3.4 8.0 Region of Principe 0.4 1 .o 2.4 2.5 7.0 District councils 0.3 0.7 1.o 0.9 1.o Social Security 0.0 0.4 0.0 0.0 0.0 Private Instititutions 0.1 0.0 0.4 0.2 0.2 Individuals 1.7 0.1 3.4 4.3 4.2 Contributions to international organizations 0.4 2.8 0.7 0.6 1.2 Other transfers abroad 1.1 0.5 0.9 0.6 31.4 Total 11.0 21.1 24.4 25.3 137.5 Source: Direction of Planning, Ministryof PlanningandFinance 139 Table 10: PublicInvestmentProgramby mainsectors (MilUS%) 2000 2001 2002 2003 2004 2005 Public administration 1.6 2.9 3.6 6.6 11 6.0 Agriculture 2.6 2.3 3.8 1.8 0.9 1.4 Water and sewage 3.1 1.7 1.9 3.7 1.9 0.9 Education 1.3 4.0 1.5 3.8 1.8 0.9 Energy 2.6 1.6 0.7 0.8 0.2 0.0 Housing 0.6 0.3 0.0 1.2 0.4 0.3 Fisheries 0.8 0.6 0.3 0.4 0.1 0.0 Health 0.3 1.7 2.2 4.0 3.9 3.2 Transport and telecommunications 4.7 4.7 4.1 3.9 3.4 3.8 Other 1.1 0.5 0.3 0.6 4.4 2.3 Total 18.7 20.4 18.5 26.8 27.9 18.7 Source: Directionof Planning, Ministryo f Planning and Finance 140 WB21847 C:\Documentsand Settings\WB21847My Documents\SAO TOME\STP PEMFARPEMFAR- FMAL\STP CIFA VolumeII.6.5.Final2.doc 06/08/2008 5:34:00 PM 1