Haiti - Public expenditure review (Vol. 1 of 2) : Main text

Report No. 61.13-HA Haiti Public Expenditure Review (In Two Volumes) Volume i: Main Text September 29, 1986 Latin America and the Caribbean Regional Office FOR OFFICIAL USE ONLY Document of the World Bank This report has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALES Currency Unit = Gourde (G) G 1.00 = $0.20 US$1.00 = G 5.00 The Gourde has been pegged to the U.S. Dollar since 1919 at this rate. WEIGIITS AND) MEASURES Metric System FISCAL YEAR October 1 - September 30 This report uses FY for fiscal year. Thus FY86 refers to the period from October 1, 1985 to September 30, 1986. ABBREVIATIONS AAN Autorite Aeroportuaire Nationale (National Airport Authority) AOPS Association des Oeuvres Privees de Sante (Association of Private Health Organizations) APN Autorite Portuaire Nationale (National Port Authority) BCA Bureau de Credit Agricole (Agricultural Credit Agency) BCI Banque de Credit Immobilier (private mortgage bank) BNDAI Banque Nationale de Developpement Agricole et Industriel (National Agricultural and Industrial Development Bank) BRE Banque de la Republique d'Haiti (Central Bank) CAHEP Centrale Autonome d'Eau Potable (Port-au-Prince water company) CGCE Caisse Cent-ale de Cooperatior. Economique (French aid agency) CIDA Canadian International Development Agency CONATRA Compagnie Nationale de Transport (Port-au-Prince public passenger transport company) DDT Departement de Transport (MTPTC) DEL Direct Exchange Line (telecommunications) DHFN Divisior. d'Hygiene Familiale et Nutrition (MSPP) (Division of Family Hygiene and Nutrition) DRE Directior. de Ressources Energetiques (MMRE) (Directorate of Energy Resources) EdH Electricite d'Haiti ENAOL Entreprise Nationale des Oleagineux (vegetable oil company) EPI Extended Program of Immunization (MSPP) EPPLS Entreprise Publique des Logements Sociaux (public housing agency) IDB Inter-American Development Bank I1fSSD International Drinking Water Supply and Sanitation Decade IPN Institut Pedagogique National (MEN) (National Pedagogical Institute)- ITU International Telecommunications Union (of the United Nations)  FOR OMCIAL USE ONLY SIMOPSIS This report reviews Haiti's public expenditure in the context of economic stagnation, the need for major structural reform, and a proposal that the country adopt a 3-year rolling investment program. It discusses the level and control of public expenditure as a whole (both recurrent and investment, and including public sector employment), the role of aid donors, and nine individual sectors: agriculture, industry, electric power, transport, telecommunications, the urban sector, water supply, education and health. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  HAITI: PUBLIC UXPENDITUU REVIEW TABLM OF CONTENTS VOLUME I: MAIN TEXT SYNOPSIS TABLE OF CONTENTS CHAP!ER I: THE ECONOMY ..............*..................1 Recent Developments..................... ......... - Economic Projections .............** 9 *9*99...9999949. , 3 The Need for Structural Reform.3 Resources for Public Expenditure ...........................4 CHAPTER II:PUBLIC EXPENDITURE.... Composition of Public Expenditure. 99999997 Past Treasury Expenditure: The Role of Extrabudgetary Spending ........................999999999999999999.9...... , .8 Recurrent Expenditure: Miuistries and Offices. 9 Other Identified Purposes......................o.......13 Outstanding Obligations to Domestic Suppliers. ........13 Debt Service .13 Unidentifed 1 ....... Evaded Recurrent Expenditure: The Role of NGOs.*...........*15 Investment Expendituree...........*999999999999949994999 }15 The Return on Public Investment...........................19 General Recommendations................................ 21 Recommended Public Expenditure Program, FY86-89. .........22 CHAPTER III: AGRICULTUR 8........... The Sector and its Planning..........................e.o.28 The Poor Return on Agricultural Investment ...............30 A. Composition of the Investment Budgetd.............. 30 B. Ministry Operating and Management Problemsems**4**. 30 C. Complementary Policies and Organization........... 32 Investment Program and Financing Plan, FY86-89**..........35 Recommendation s ..s............... 9999e9999*o*e*****o*37 This report is based on the findings of a mission to Haiti in October-November 1985 led by Nicholas Burnett and consisting of Peter Gyamf i (transport), Robert Lacey (public industrial enterprises, telecommunications), Adrian Lambertini (power), Julio Linares and Carlo Rietveld (water supply), William McGreevey (health), Maryvonne Plessis-Fraissard (urban sector), Antoine Schwartz (education), Tom Zalla (consultant, agriculture) and John Gallup (research assistant). Jacomina de Regt (education) and Noel King (energy) also contributed to the report. A draft was discussed with the Haitian authorities and the major aid donors during July and August 1986.  CHAPTER IV: PUBLIC INDUSTRIAL ENTERPRISES .......O....................41 I. Infrastructure Needs of Private Exporters.................41 Infrastructure and Industrial Parks....................41 Recommendations ......... <4 ,4,., * , 0442 II. Public Industrial Enterprises.............................42 Financial and Operational Aspects. ....... ...........,42 Debt.o. .*.o.oo.....o o.......oo...o.o.o..o...o.oo 0.44.44 Balance of Payments I mpact 45 Recent Government Actions: ENAOL and USND ............46 Options: La Minoterie, Ciment d'Haiti and USN .......... 47 Recommendations............................................... 47 CHAPTER V: ELECTRIC POWER,................... 4 50 The Energy Sector Context ..... ... ,, 0 The Power Sector and its Planning.........................51 Electricity Demand 4..52 Tariffs and Operating Costs........................4*. 54 Expansion Plan, FY86-95.* *4040444444444400400004* .0.4.4.54 Investment Program and Financing Plan, FY86-89.***o#......55 Recommendations...e...... e.. 4444444404404456 CHAPTER VI:TRANSPORT.... 58 Mhe Sector and its Planning..............,................58 Recurrent Spending. * 4444444444444444 44.4.444.59 Investment Program and Financing Plan, FY86-89. ..........61 Recommendations. .......... .................... .4.0.0.0.62 CHAPTER VII: TELECOMMUNICATIONS ......67 The Sector and its Planning 67 Financial Aspectso..o.............4.4.4.44.o.4.o0.o... .*469 Investment Program and Financing Plan, FY86-89.**.o.....o.70 Recommendations.e............ ..............o . 44444.0444472 CHAPTER VIII: THE URBAN SECTOR. o..e.......o.o..........o44440444404474 The Sector and its Planninge..o... ....oo .o...o..oooo0.44.74 Recurrent Spending. .....o..oo............o.o..o..... 77 Investment Program and Financing Plan, FY86-89. ...oo..oo.78 Recommendations .......o........o.o..o.......... 00.04..4478 CHAPTER IX: WATER SUPPLY............ ............... . o....... . o82 The Sector and its Planning...... 4*44*0404o**00*o 444 82 Financial Aspects .......e...o.......o..o....o......... 83 Investment Program and Financing Plan, FY86-89o9o. oo.....84 Recommendations. 4.44.4o.e..o o86 CHAPTER X: EDUCATION. 4044 00044*044444440444404040444440440487 The Sector and its Planningo.... ......... .44.44.4.4.4o. *87 Past Expenditure. 4...o.444@..404004004444044.4444.400 ,0l91 Expenditure Program and Financing Plan, FY86-89o...o.....94 Recommendations... ............o.............o... e...o..96 CHAPTER XI: HEALTH...... 04440400044444 444044440044 44444040444444103 The Sector and its Planringo...........................o.103 Past Expenditure 107 Expenditure Program and Financing 'lan, FY86-89a*o.......109 Recommendations. , 440400444111  VOLUME IT: APPENDICES APPENDIX 1: ECONO01IC PROJECTIONS APPENDIX II: STATISTICAL APPENDIX Index of Tables Tables APPENDIX III: PROJECT-BY-PROJECT DATA Table of GContents Explanatory Note Project-by-Project Data LIST OF TEXT TABLES (VOLUME I) Page No. CHAPTER I: 1.1 Public Sector Operations, FY81-85......,.........,.........2 1.2 Public Enterprises Operat-ng Surpluses and Transfers to the Government, FY84-85 ................................ 5 CHAPTER II: 11.1 Budget, FY86 ........................ . 8 II.2 Treasury Expenditure, FY81-85, and Budget, FY86o............9 II.3 Recurrent Expenditure on Ministries and Offices, FY82-85, and Budget, PY86 ............ . 10 II.4 Actual and Budgeted Public Employment, FY79-82.............11 1}.5 Personnel Share in Ministry Spending, FY82-85.............12 II.6 Government Debt Service, FY86, and Projected, FY87-89.*,.l.14 I1.7 Expenditure of 71 NGOs, FY82-85.............m.............15 II.8 Financing of Public Investment, FY84-85, and Budget, FY86..16 11.9 Planned, Budgeted and Actual Public Investment, FY81-84....18 Il.10 Sectoral Allocation of Public Investment According to Financing Source, Average FY84-85. . ...................... .21 I1.11 Recommended Recurrent Expenditure of Development Ministries, FY86-89 ................. 23 11.12 Recommended Public Investment Program and Financing, FY86-89. ***9#9S*9*#99+**.**@O@99.*O^*99*.9* .*v...,.e....24 I1.13 Recommended Treasury Expenditure, FY86-89, and Budget, FY86s ...........- **@--@sv42  CHAPTER III: 111.1 Public Agricultural Investment Budgets FY86..... ............29 111.2 Public Agricultural Investment Expenditure, Actual versus Budgeted, Y28 .......,- 3 111.3 Composition of Agricultural Investment Budget, FY86...*.....36 111.4 Financing of MARNDR Budget, PY85-86 os.ee.e..................37 111.5 Recommended MARNDR Expenditure and Financing, FY86-89, and Actual Budget, FY86....................................39 CHAPTER IV: IV.1 Public Industrial Enterprises: Summary Annual Average Financial Results, FY83-85 .........o.............o..........................43 IV.2 Public Industrial Enterprises: Comparison between Unit Production Costs and CIF Import Prices, FY85.....o*.4o...44 IV.3 Economic Impact of Public Industrial Enterprises, FY82-85...46 IV.4 Public Industrial Enterprises: Illustrative Investment Program, FY86-89 ............... ........... ........... 48 CHAPTER V: V.1 8dH: Investment Program, FY86-89 ...........................55 V.2 EdH: Financing Plan, FY86-89................................56 CHAPTER VI: VI.1 Preliminary Transport Investment Program, FY86-91.......... .61 VI.2 Recommended Transport Investment Program, FY86-89...........65 VI.3 Recommended Transport Expenditure, FY86-89..................65 CHAPTER VII: VII.1 TELECO: Income Statement, FY81-85, and Budget, FY86.........69 VII.2 TELECO: Investment Program and Financing, FY86-89 ........... 70 VII.3 TELECO: Budgeted and Actual Investment, FY84-86.............71 CHAPTER VIII:- VIII.1 Programmed and Executed Urban Investments, FY81-85..........76 VIII.2 Urban Sector Recurrent Budget, FY86.........................77 VIII.3 Proposed Urban Sector Investments, FY86-88..................78 VIII.4 Recommended Urban Sector Investment Program, FY86-89........80 CHAPTER TX: IX.1 Water Investment Program and Financing, FY86-89.............85 CAWTER X: X.1 The Private Sector's Role in Education, FY8S................87 X.2 Public Recurrent Expenditure on Education by Level, FY85....93 X.3 Public Education Investment and Financing, FY86-91..........95 X.4 Recommended Education Recurrent Expenditure, FY86-96 ........ 99 X.5 Recommended Educatio,n Investment Program, FY86-96..........100 X.6 Recommended Education Expenditure and Financing, FY86-89... 101  CHAPTFR XI: XL.1 Total Health ExpendliVe, FY81-8$..........................107 X1.2 Public Health Expenditure, FY80-85, and Budget, FY86.......109 XI.3 MISPP: Sources and Uses of Investment Fundes, FY86, Budget and Likely ...........l XI.4 MSPP: Recowmended Expenditure Program and Financing FY86-9O....... ............ 0**00000* 113  -i- S31RY i. Recent political changes provide Haiti with a unique opportunity to adopt policies to initiate sustained economic growth and begin to alleviate the desperate poverty of the average Haitian. The relentless pressure of population on limited natural resources can be relieved only by increased output, coupled with family planning and improved human capital. Major pricing and structural reforms are needed in the productive sectors, and public expenditure requires overhauling. Haiti: Policy Proposals for Growth (World Bank report 5601-HA, June 10, 1985) recomnends major policy changes for agriculture and industry; this report compl_ments them with proposals for the level and allocation of public expenditure, including thiat of the public enterprises.-/ Public expenditure has averaged 22 percent of GDP in the last five years, of which 54 percent has been recurrent spending by the Treasury and 46 percent capital expenditure by the entire public sector. ii. Haiti's economy has stagnated for six years; per capita incomes are still nine percent below those of 1980. Aid alone cannot overcome this malaise. The only route to sustained growth is that of major policy reform, coupled with a very modest increase in aid. With reform, a sustained real rate of economic growth of about 3-4 percent a year is feasible for the rest of the 1980s, a!nd would lay the basis for faster expansion in the 1990s. iii. 'fIhe needed pricing, structural and expenditure reforms are interlinke4. Thus, for example, continued public investment in irrigation rehabilitation could not be effective without first setting agricultural prices appropriately and establishing systems to recover irrigation operating costs and to carry out maintenance. iv. A balanced budget for the consolidated public sector, financed only by tax and customs revenues, earnings of public enterprises and concessional aid, is an essential element of future expenditure. Financial control has been inadequate, including of the public enterprises. In the recent past, unproductive, ad hoc extrabudgetary spending, e.g., on an uneconomic new sugar mill and on military jet aircraft, created fiscal deficits averaging almost ten percent of GDP, that were financed by concessional aid (67 percent), Central Bank credit to the public sector (28 percent) and external commercial borrowing (5 percent). Haiti is too poor to borrow commercially and, until recently, the excess liquidity from money creation spilled over into the balance of payments and threatened the parity of the Gourde. v. Public expenditure must be adjusted to increase the returns to existing private and public investments. This means an adequate level of recurrent expenditure by development-oriented ministries, especially agriculture, public works, education and health. Modest increases for 1/ Policy changes are now underway in all these at:eas. This report uses as its organizing framework the original FY86 budges; subsequent major changes are reflected where possible in the text, but not in the numerical material, which generally refers to late 1985. Updating the data would not, however, in any way change this report's principal conclusions.  Hii- education and health were implemented in mid-FY86. Although too much public expenditure is on salaries, the four development-oriented ministries all have large staffs and suffer from relatively low salary scales compared to other ministries and, in particular, from Inadequate non-salary operating funds. Agricultural extension workers cannot get to the field, schools do not have books, public health clinics are without drugs. A real increase of about 25 percent above the FY86 budget in the recurrent budgets of these four ministries is called for, from about G 300 million to about G 400 million, almost all for non-salary operating funds. Major improvements in their internal efficiency are also necessary. vi. Partly because some ald donors have increasingly ehanneuled their aid away from the public sector, Non-Government Organizations have become very important in several sectors, notably agriculture, education -nd health; their increasing expenditures have, in a sense, permitted ttie Government to avoid some recurrent expenditure. The Government must improve its coordination with NGOs, especially in agriculture and education; that in health is relatively effective. vii. Unless these recommended shifts are made in recurrent expenditure and coordination with NGOs is improved, the returns on public investment cannot be increased. The recurrent budget is the responsibility of the Ministry of Economy apd Finance; the investment budget has been that of the Ministry of Planning.'/ Past returns on puv"lic investment have been low, because complementary pricing and structural reforms have not been made; because much of the investment budget has not been used for fixed investment but rather for salaries or diverted to non-development purposes; because of inadequate complementary recurrent expenditure; and because of poor project selection. There has been a chaotically complex system, nominally based on a Five Year Plan, of planning, budgeting, monitoring and disbursing public investment, which provided numerous opportunities to divert funds. Projects financed wholly from Government resources or by foreign commercial borrowing, like the public sugar mill and vegetable oil factory, were not selected using economic criteria. The Government has no project analysis capacity and has accepted donor-offered projects uncritically. Donors have not evaluated projects adequately; less than five percent of current projects have had an economic rate of return calculated. Almost none have had their recurrent cost implications assessed. Except in the electric power sector, aid coordination has not been effective; there are particular problems in agriculture, transport and education. It is striking how many donors are involved in so many sectors. viii. The ctmbersome planning and budgeting system should be replaced with something more flexible. This report recommends a three year rolling public investment program, and suggests investments within this framework. Experience elsewhere has shown that effective budgeting and finaneial control is almost impossible without consolidating responsibility for both recurrent and investment spending. An effective system of financial control 2/ The Ministry of Planning was closed in August 1986. A new National Promotion and Public Service Commission was established.  -ii-.. is needed. Project analysis capacity must be developed. Aid coordination must be improved and the Government may wish to consider establishing one point within the administration responsible for all contacts with donors and also limiting the number Of donors involved in individual sectors. Some of these measures will take time and require technical assistance. There is no reason, however, that effective systems should not be in place by the early 1990s. ix. Public investment should concentrate on consolidating and rehabilitating existing investments. Investment expernditure could be held roughly constant through FY89 at about G 840 million per year in FY86 prices (about eight percent of GDP), with increases compared to the past for agriculture, health, education and water supply, and decreases for industry, transport and non-development purposes. This level of public investment would represent a modest real increase over FY85 but a significant decrease compared to the inflated expectations of external aid contained in the FY86 budget. The returns on public investment should increase, however, as much more of it will represent actual fixed investment. This level of public investment would pose little financing difficulty. Treasury contributions would not be needed above the G 136 million level budgeted for the current FY86; the bulk of public enterprises' investment would be by the electricity and telecommunications companies, which are in satisfactory financial shape to generate necessary internal cash for investment and have appropriate investment programs. Average annual concessional aid disbursements of about US$110 million would be required, about half of which is already committed. X. The financing issue for public expenditure concerns the recurrent budget. The increases recommended for the four development-oriented ministries can be financed through the elimiuation of waste and expenditure that would not stand scrutiny. Despite an official figure of 32,400, there are almost 60,000 public servants, far more than are needed to run the Government effectively. In particular, there are many non-existent and non-performing employees on the public payroll. An initial cut in public employment at the ton-development ministries is recommended; this would probably cover only nonexistent employees. In addition, unidentified recurrent expenditure, especially 'Special Obligations" and "sans justification" budget lines, represents about 17 percent of total Treasury outlays. These have already been reduced and the Government intends to end them in the FY87 budget. Assuming, probably generously, that 25 percent of these expenditures as budgeted in FY86 may, in fact, be reasonable, the elimination of the other 75 percent would provide adequate savings to finance the necessary increases in recurrent spending for development. xi. Further potential savings could undoubtedly be identified through an audit of all recent public expenditure, to determine where diversions occurred, and through an analysis of public employment needs and current staffing. With these analyses in hand, schemes could be developed to reduce the size of the public service over time through attrition, early retirement, redundancy bonuses, and so on.  -iv- xii. Indeed, so extensive is the waste in the current budget that the modest shifts and cuts recommended above, before any major reductions in public employment and before the elimination of all dubious expenditures, could by the late 1980s result In a current budget surplus of some G 200 million per year, if revenues had remained at their budgeted FY86 level. In fact, revenues fell from 12 to 10 percent of GDP in FY86, following tax reductions; a budget surplus therefore no longer seems so possible. Nevertheless, resources freed by eliminating waste could be used to ircrease the salaries of those who remain in the public service, especially in the development-oriented ministries, like teachers and medical personnel. xiii. This report contains detailed recommendations for all developmetit-related sectors which cannot be briefly summarized. The paragraphs below therefore only highlight the principal recommendations for each sector; the individual chapters must be consulted for detailed analysis and recommendations. As noted, their effective implementetion will require improved aid coordination and appropriate technical assistance. xiv. Agriculture. As noted, expenditure reforms cannot produce results in agriculture if they are not accompanied by pricing and structural changes. The most important -xpenditure shift should be to increase non-salary operating funds to about 40 percent of the total agricultural recurrent budget. This should be complemented by major managerial improvements to consolidate all agricultural projects under the Ministry of Agriculture, Natural Resources and Rural Development; improve work discipline; reallocate staff from Port-au-Prince to the rural areas; and streamline disbursement procedures. Projects are rarely fully implemented, in part because of poor design and also because of weak implementation procedures. Except for extensions of existing projects, almost no new projects should be started through FY89; the focus should be on getting existing ones to work. The investment priority should be irrigation rehabilitation, provided appropriate complementary operating policies are adopted. Rural development projects need a more specific production focus, and those that are ir trouble should be reassessed. Soil conservatior. projects should be seen as experiments to find a workable anti-erosion strategy for the 19909 and should not be expected to produce effective results in the short term. A nroposed project to create regional milK processing plants does not appear justified. Major improvements are needed in agricultural credit. Further detailed work is necessary and urgent to identify more precisely than the mission was able those projects that should be eliminated or reshaped. Despite the general recommendation to avoid new projects, two were identified: a private sector input supply project and an extension education and diffusion project. Coordination and the sharing of experience among the Government, NGOs, and aid donors should be improved. xv. Industry. Five public industrial enterprises resulted in economic losses during FY82-85 equiva ent to almost four percent of GDP; the Haitian economy has not grown by even three percent in any year since FY80. The vegetable oil factory wrs 'losed in July 1986 and the market for oil  freed in Aprilo One sugar mill was also closed; the second should follow suit. The cement factory should cease producing clinker and instead crush and bag imported clinker, requiring investments in port facilities, and the cement market should be freed* The flour mill should be subjected to competition by removing its monopoly on wheat and flour imports, and freeing the.market. There is no need for further public industrial parks as the private sector is meeting demand; Ideas for a private industrial Free Zone near the port of Port-au-Prince need to be thought through but should be encouraged. xvi. Electric Power. The Government should curb-the excessive growth of public sector power consumption and settle its arrears to Electricite d'Haiti. The company, while generally well run, should increase its operating efficiency, especially the number of customers per employee, as its tariffs are alteady among the highest in the hemisphere.. Its investment program, centering ca the construction of the Artibonite 4C hydro project, is reasonable and well-balanced, and should not prove difficult to finance. The program does not, however, deal with the problem of the siltation of the reservoir behind the existing Peligre dam, although a CCCE-financed study of alternatives is underway. Raising the height of the dam would flood a small area of the neighboring Dominican Republic and discussions between the two countries should begin to assess its political feasibility. The interconnection of Cap Haitien, Gonaives and St. Marc to the Port-au-Prince system should not be made until the late 1990s, as in the system expansion plan. xvii. Transport. The Ministry of Public Works, Transport and Communications' excess staff should be trimmed over four years, with annual reductions of about five percent. The Government's list of projects for the transport sector is too ambitious and should be scaled back. Road priorities should be to complete the ongoing bridge and culvert program, to reconstruct the bridges at Fer-a-Cheval and Guayamouc, to construct and rehabilitate about 120 km of secondary roads per year, to upgrade the Mirebalais-Pont Sonde road, and to rehabilitate sections of the Port-au-Prince to Mirebalais road. Port and maritime priorities are port ma'lntenance equipment, minor works to complete the port at Cap Haitien, rehabilitation of that at Jeremie, a radio communications system, an inspection slipway and a training center. No new cabotage port rehabilitation is appropriate in light of the disappointing results of recent experience. Minor civil aviation works proposed are reasonable, except for the proposed airport modernization at Cap Haitien, which has no economic or financial justification, and the national air navigation system, which requires economic and financial study to determine its feasibility. Transport sector aid coordination meetings should be revived. xviii. Telecommunications. Management of the public company, TELECO, has improved recently. It is now being audited. However, the company should raise domestic tariffs to at least cover marginal costs, contain current costs (especially by halving wage costs per line installed), diversify its sources of external finance, and adoot international competitive bidding for its equipment procurement, which has been costlier than necessary because of  -vi- the use of tied aid. It is to be hoped that the master plan currently being developed will build on recent organizational improvements and encourage more In-depth analysis of prospective projects than in the past. TELECO's investment program, consisting principally of introducing digital exchanges to the Port-au-Prince telephone network and improving the telex service, is in general appropriate and meets immediate needs. The company should not need any further external financing through FY89. The Government should not use TELECO's cash surpluses to finance other activities, e.g., in the past, it was used to subsidize the operating costs of the nesrer public sugar mill. xix. The Urban Sector. Port-au-Prince urgently needs an urban development plan. Its absence greatly complicates the major urban Investment, a drainage program, and also the planning of water supply investments. Drainage is a high priority because of the city's location on a swamp beneath steep, eroding hillsides. Further investments, however, should await the urban plan, the establishment of a system of tariffs for urban solid waste collection, the setting up and funding of a maintenance program for the drains already and to be built, the resolution of the relative roles of the municipality and the Ministry of Public Works, Transport and Communications, and a reassessment of the project's phasing. The future of the Croix-des-Boissales market project should be resolved. More attention should be paid to urban transport, minor works and traffic management to relieve congestion that impedes workers access to industrial assembly plants and the plants' access to the port and airport. The public bus company CONATRA cannot compete with the efficient, private "tap-taps" and should be closed. Technical assistance should concentrate on preparing an urban plan, improving municipal finances and implementing a cadaster. xx. Water Supply. The investment program implicit in the plans of the three sector agencies is well balanced and justified, although the lack of an urban plan makes this judgement tentative for Port-au-Prince. A drastic improvement is necessary in the management, operations, finances and project execution capacity of CAMEP, the Port-au-Prince water supply agency, including particularly a halving of its staff per connection. If this is done, then a project proposed for Port-au-Prince will be feasible. Investments outside Port-au-Prince should not be held up while similar actions are taken by the relevant agencies, SNEP and POCHEP, although steps must be initiated to start to recover at least some operating costs. xxi. Education. Appropriate, achievable objectives are to complete the reform of basic education introduced since 1982 and to provide universal access to primary education for all six-year olds by the year 2000. The bulk of primary education is in the private sector and this will remain so. The public sector should concentrate on improving quality and access throughout the educational system. The expansion of basic education should be slower than in the past, but efficiency improvements should compensate. There should be a three percent per annum expansion of public basic education but no increase at any other level. Cost recovery should be introduced at the university and public vocational and technical schools. Agricultural technical training needs attention. Many efficiency improvements are needed in basic education, including salary increases for  -vii- qualified teachers under the new Teacher Charter, bonuses for those teaching in remote rural areas, and the use of double shifts in urban classrooms. The question of minimum teacher qualifications needs to be reviewed. Parental contributions to their children's education should be r#dtced not by lowering fees but by providing free textbooks at the primary level. The internal efficiency of the recurrent budget for education should be further improved and its overall level raised. The use of double shifts in urban classrooms would permit a reduced investment budget, but one that is operationally sustainable. No new aid commitments are needed through FY89 but donor coordination needs improving, and some minor reshaping of overlapping projects is necessary. Donors financing low priority higher education-projects should consider alternative uses for their funds. No new teacher training college is necessary at the higher level and there is as yet no case for public investment in the private Institut Roi Henri Christophe. xxii. Health. As in education, account should be taken of the effective role played by institutions outside the public sector. Public resonrces should be shifted away from Port-au-Prince toward rural areas and away from curative services toward primary and preventive health care, family planning and targeted nutrition. The proposed reduction in staffing in Port-au-Prince is appropriate, but the large ir-crease intended in rural areas is excessive. Where possible, NGOs should be utilized to provide rural care; coordination between the public sector and NGOs is effective. Employment at the Ministry of Public Health and Population could usefully be frozen and the size of the medical school class at the university halved. Dispensaries and health centers should be used more intensively and cost recovery should be further tried for curative services. Even with these efficiency improvements, recurrent spending should increase gradually. Donor support has been declining but should be increased if these reforms are successfully introduced by the Government. Investment projects are broadly appropriate, but many experts question the emphasis on the malaria program. The public sector's capacity to deliver rural health services must be strengthened before this project can be expanded. The new maternity hospital proposed for Port-au-Prince should not proceed. The decline in family planning service delivery must be reversed and support for nutrition programs should quadruple.  CHfPTUR I THE ECNONWi Recent Develo_ments 1.01 Recent Government charges in Haiti provide a unique opportunity to. reform economic policies and launch the economy, which has stagnated since 1980, onto a path of sustained growth and development. Haiti: Policy Proposals for Growthl/ described developments through FY84 and discussed their main causes: (a) growing population pressure on limited natural resources and inappropriate policies in agriculture; (b) inappropriate incentives for private industry and uneconomic public industrial enterprises; and (c) inadequate public savings and an inappropriate allocation of public expenditure, both recurrent and investment. 1.02 More recent developments reinforce the findings of Haiti: Policy Proposals for Growth. At US$377 in FY85, per capita income was still nine percent below its FY80 level. Agriculture has continued to stagnate. The import substitution industries have exhausted the growth potential that protection afforded them in the 1970s. Only the export assembly industry has continued to flourish during the first half of the 1980s, despite the U.S. recession. Recently, however, it too has shown signs of faltering, because of the shakeout in the U.S. computer industry and polltical events in Haiti since November 1985, the latter culminating in a change of government in early February 1986. Several export assembly firms have closed their Haitian operations, either in response to difficulties with the customs authorities or in reaction to political upheaval.. Problems in the assembly industry follow the loss of three traditional export markets in recent years: the bauxite mine closed in 1982; tourism has suffered severely since the AIDS disease came to be associated with Haiti; and the threat of a U.S. ban on mango imports treated with the pesticide EDB. 1.03 The negative impact of these developments on the balance of payments has been compounded by a public sector deficit averaging 9.6 percent of GDP in FY81-85, financed by external aid (67 percent), central bank credit (28 percent) and foreign commercial borrowing (5 percent). The public sector deficit has shown a pattern of control followed by relaxation, falling from almost 13 percent of GDP in FY81 to nine percent in FY83, then rising again to ten percent in FY84 and falling to six percent in 1/ World Bank Report No. 5601-HA, June 1985.  -2- FY85 (Table 1,1). Recourse to central bank credit to the public sector reached about 3.5 percent of GDP in both FY81 and FY84; by FY85, it fell to about two percent. Public savings have followed the same pattern. Since February 1986, however, public finances have been controlled, the budget balanced and there has been no Central Bank credit to the public sector. Table 1.1: PUBLIC SECTOR OPERATIONS, FY81-85 (percent of GDP) FY81 FY&2 FY83 FY84 FY85 Receipts 10.4 12.1 12.6 12.9 13.8 Treasury revenue 9.1 10. I 10.3 100 11,2 Transfers from public enterprises - 1.0 0.9 1.0 0.7 Retained surplus of public ents. 1.3 1.0 1.4 1.9 19 Expenditure 23.4 21.4 21.9 23,1 20.2 Treasury current expenditure 11.3 11.2 11.6 12.4 12.6 Public sector capital expenditure 12.1 10.2 10.3 10.7 7.6 Public sector deficit 13.0 9.3 9.3 10.2 6.4 External concessional aid 6.2 6.5 7.4 6.8 5.2 Grants 3.2 3.9 3.4 3.4 2.6 Loans (net) 3.0 2.6 4.0 3.4 2.6 Public sector deficit after aid 6.8 2.7 1.9 3.4 1.1 External commercial financing (net) T.T -U.T T.7 -07 -8 Domestic financing (net) 3.4 3.3 1.2 3.5 1.9 Memo item: Public savings -0.9 0.9 1.0 0.6 1.1 Source: Statistical Appendix Table- 1.11. 1.04 The public sector deficit ri recent years spilled over into the balance of payments. Much of the financing of both in FY81 was foreign commercial lending for the construction of a new public sector sugar mill, the Usine Sucriere Nationale de Darbonne (see Chapter IV). In general, however, the balance of payments deficit, which averaged 4.4 percent of GDP ir. FY81-85, was f4nanced by concessional external aid, borrowing from the IMF, reserve losses -nd arrears. As noted, there was a considerable improvement in the public sector deficit in FY85 and the payments deficit declined. However, Haiti's fiscal and balance of payments situation is so precarious that no deficit not financed by concessional external aid can be afforded. Excess liquidity resulted in the emergence of a parallel foreigr. exchange market, the discount on the Gourde varying between 8 and 25 percent since FY84, giving rise to trade disruptions, notably suspensions in fuel supplies. With renewed fiscal discipline, the discount had dropped to about 5 percent by August 1986. Commercial borrowing must also be repaid and new borrowing is difficult and unwise, given Haiti's limited creditworthiness.  -3- Net aid disbursements declined since FY83, and net foreign reserves at the end of FY85 were deeply negative at minus US$67 million, having fallen in each of the last five years, and have not improved since. By January 1986, there were payments arrears for oilseeds and to the IMF. Gross reserves are essentially nonexistent. Public enterprises were ofter required to go to the parallel market to raise foreign exchange. Aid lhas Increased since February and payments arrears are being eliminated. Econoic Projections 1.05 Despite the gravity of its economic and financial circumstances, the Haitian economy can start to grow again. The mission updated the -economic projections of Haiti: Policy Propoaals for Growth, and confirmed that report's conclusions (Appendix I). 1.06 Three scenarios were compared. Same would be a continuation of past policies and aid levels; Aid a continuation of past policies but with a significant increase in aid; Reform a combination of policy reform and increased aid. All three scenarios assume that the present negative level of international reserves cannot be allowed to continue and must rise to zero no later than FY90. 1.07 The results demonstrate dramatically the reed for policy reform and underline that increased aid alone will not be enough to set the Haitian economy onto a path of sustained growth. The Same scenario would result in zero economic growth, or continued stagnation and falling per capita incomes. The Aid scenario would also fail to even maintain per capita incomes, the economy growing at about 1.5 percent per annum while the population grows at 1.8 percent. Only the Reform scenario would result in sustained growth, though moderate at about 3.5 percent per annum. Thus aid will be a necessary but not a sufficient condition for economic growth; policy reform is more important. The Need for Structural Reform 1.08 The population problem cannot be exaggerated. Though Haiti's population growth rate is under two percent, the increasing pressure on natural resources has become severe. A recent study by FAQ compared the carrying capacity of national agricultural systems with current and future demand for food as it relates to population size and growth. The study found that food demand in the 1970s was already exceeding agricultural system capacity by 75 percent. Even with probable improvements in agricultural technology, Haiti's population is projected to grow so that it will exceed capacity by 87 percent in the year 2000. This, coupled with recent economic stagnation, points to the need for effective family planning, to increase agricultural output, and to increase exports to purchase the food that will have to continue to be imported. 1.09 Sustained economic growth cannot occur without major structural change. The broad policy changes necessary for agriculture, industry and public revenues are discussed in Haiti: Policy Proposals for Growth. The benefits of the growth of 1975-80 have been channelled into inappropriate,  -4- unproductive and inefficient activities, including corruption, and led to economic stagnation. The opportunity of a new Government can now be seized to permit this to be changed. Fundamentally, the economy must be shifted toward exporting; there is no other option for a small economy like that of Haiti. Other measures concerning the allocation and control of public expenditure, including the public enterprises, form the basis of Chapters II-XI of this report. 1.10 In agriculture, incentives must be set correctly e.g. the coffee export tax should be further reduced, other export taxes should be abolished, farmer prices for cotton should be raised to the equivalent of import parity level, rents should be raised on State lands, large landholdings should be taxed, and irrigation charges should be increased. In addition, the operations of the Ministry of Agriculture, Natural Resources and Rural Development (MARNDR) should be drastically overhauled, improvements are necessary in marketing, and an agricultural information base is urgently needed. (Other measures affecting public expenditure on agriculture are discussed in detail in Chapter III of this report) 1.11 In industry, efficiency should be improved, principally through abolishing import licenses, quotas and price controls, and reducing tariffs, and also through administrative improvements at the Customs and for all procedures concerning investment, exports and export-related transactions. Steps are riow being taken to do this. (Other measures needed for industry--the timely provision of economic infrastructure--are discussed in Chapters IV-IX). Inefficient public industrial enterprises should be closed, dive&ced or made economically efficient (Chapter IV). 1.12 Over time, the national tax effort needs to be increased and the tax system improved, largely by continuing the trend toward raising revenues from internal taxes on value added and income rather than from excises and taxes on external trade, especially exports, and by improving collections. A growing economy would, of course, make this much easier. The tax base should be enlarged and rates might then be reduced, thereby helping to overcome taxpayer resistance. Tax collection must continue to improve. 1.13 In the absence, however, of any indication that these measures are likely to be adopted, this report adopts a conservative no-growth hypothesis for the next several years when considering public expenditure. To the extent that this proves incorrect, the provision of public services and the volume of public investment should be able to increase. Resources for Public ExpendIture 1.14 If borrowing from the Central Bank is ended. the resources available for public expenditure would consist of the current revenues of the Treasury enhanced by transfers from those public enterprises that are financially profitable, the retained operating surpluses of public enterprises, and concessional aid grants and loans. As noted, foreign commercial borrowing is currently impossible for Haiti.  1.15 Tax revenues inereased from nine percent of GDP in FY81 to over 11 percent in FY85 (Table 1.1) but were reduced to about 10 percent in FY86. New increases cannot be expected until the economy starts again to grow significantly. Modest improvements are possible, however, especially from the value added tax, receipts from whi'.h have grown from 1.1 to 1.8 percent of GDP from FY83-85, from improved income tax collections, and from import tariffs replacing quotas. A major increase in the tax effort would not, however, be appropriate in the immediate future as the private sector needs resources to grow. 1.16 In addition to tax revenues, an increasing share of public sector resources has come from the operating surplus of the public enterprises, which averaged 2.8 percent of GDP in FY84-85 (Table I.1). Three enterprises account for over 80 percent of this surplus: EdH (the electricity company), TELECO (telecommunications) and La Minoterie (the flour mill) (Table 1.2). Indeed, the Government has since FY82 supplemented its tax revenues with direct transfers from TELECO and La Minoterie equivalent to one percent of GDP in FY82-84 and 0.7 percent in FY85. All public enterprises have been subject to taxation since the beginning of FY86. The transfers are thus taking place in a different manner. Chapter IV proposes that the Minoterie should lose its flour monopoly and that imports should be permitted, subject to a tariff set to retain Government flour revenues at their present level. Whether this proposal is adopted or the present situation continues, the net impact on the fisc will be approximately the same. TELECO, however, should not have its surplus transferred to the Government as it has considerable investments to finance in the next several years and its tariffs for international telephone calls are already prohibitively high (Chapter VII). Table I.2: PUBLIC ENTERPRISES OPERATING SURPLUSES NDT'rRANSFERS TO THE GOVERNMENT, FY84-85 (G million at current prices) FY84 FY85 Operating Transfer Operating Transfer Surplus to Government Surplus to Government APN (port) 19 - 28 - CAMEP (water supply) 1 -3 1 -3 Ciment d'Haiti 17 - 22 - EdH (electricity) 45 - 57 - Minoterie (flour mill) 65 46 69 61 TELECO (telecommunications) 112 67 86 35 USND (sugar mill) -2 -19 -4 -20 Total 257 91 258 73 Source: Statistical Appendix Table 2.6 1.17 Taking recent tax revenue and public enterprise operating surplus trends into account, resources other than foreign aid available to the Government through the 1980s are assumed at a constant 12 percent of GDP. This compares very unfavorably with the average of 16.7 percent for low  -6- income countries with per capita GDP below US$400 but is nonetheless realistic. In the 1990s, once the economy is growing steadily, steps can be taken to increase the tax effort further. 1.18 In addition, public enterprises are assumed to have a net operating surplus after transfers and taxes sufficient to meet the detailed internal cash generation requirements of their recommended investment programs proposed elsewhere in this report. These would peak at G 177 million in FY87 (expressed in FY86 prices) or 1.6 percent of a stagnating GDP i.e. below that of FY84 and FY85. This is realistic only if the recommendations in Chapter IV to reform the operating environment of public industrial enterprises are followed.  -7- CIAPTER II PUBLIC EXPINDITURE 2.01 Structural reform is needed not only in policies affecting the productive sector but in the composition of public expenditure. This report does not examine all public spending but concentrates on that which is most important for development: the recurrent and investment expenditure of the Government and the investments of the public enterprises. This chapter surveys these components in the context of the FY86 budget, discusses the impact of public spending since FY80, notes changes since February 1986, and makes recommendations for the future. Detailed sector-by-sector recommendations are made in Chapters III-XI. Composition of Public Expenditure 2.02 The FY86 budget represented a major step forward by consolidating all public expenditure in the budget and by identifying explicitly for the first time many items that were previously either excluded from the budget or not identified. Despite this, much budgeted expenditure was not explicitly identified. The presentation here supplements the published budget where appropriate. 2.03 Table ILI, sets out the main elements in public spending, which was budgeted to be split roughly equally between recurrent and investment expenditure. Sixty percfent of recurrent expenditure was for operating expenses of ministries and offices, including the armed forces; eight percent for operating expenses of other public institutions; four percent for payments to domestic stuppliers for purchases made in previous years; and 13 percent for debt service. By July 1986, debt service had risen to 17 percent of the total. Fully 17 percent of recurrent expenditure, however, was for unidentified purposes. About two-thirds of investment expenditure is by the Government proper and one-third by the public enterprises. 2.04 The budgetary process does not permit the precise allocation of investment expenditures to individual ministries, agencies and public enterprises that characterises recurrent expenditure. This reflects the relative roles of the Ministries of the Economy and Finance (MEF) and Planning (MP). The former is responsille for the recurrent budget and several public enterprises; the latter / for the investment budget of the Government--but not the public enterprises--within a global investment envelope determined by MEF. This budgetary overlap continues into disbursements with the MP having to pass on all investment expendftures of ministries before authorizing MEF to disburse. Further confusion is caused by MP classifying all investment expenditures according to sector and MEF simply adopting this classification and including MP tables in the consolidated budget; the data are never linked to actual spending ministries. 1/ The Ministry of Planning was closed in August 1986. A new National Promotion and Public Service Comdissi-on was established.  -8- Table 1I.1: BUDGET, FY86 G million Percent Recurrent expenditure 1,196 52 Ministries and offices 712 31 Other identified recurrent expendIture 73 3 Civil and military pensions 29 1 Local Government 17 1 Banque Nationale de Developpement Agricole 12 1 et Industriel International Organizations 7 - Outstanding obligations to domestic suppliers 49 2 Debt service 156 7 Amortization of external commercial debt 22 1 Amortization of external concessional debt 20 1 Interest on external debt (i.ainly commercial) 19 1 Interest to IMF 35 2 Interest on Government domestic bonds 60 3 Unidentified recurrent expenditure 204 9 Special Obligations 179 8 "Acreditifs' 25 1 Investment expenditure 1,113 48 General government ?36 32 Public enterprises 377 16 Total expenditure 2,309 100 Source: MEF, 3udget, FY86. 2.05 The recurrent budget, whioh accounts for current outlays to run the Government and also development projects, is financed out of Treasury revenues and transfers from the public enterprises. The investment budget is largely financed by external grants and concessional aid. In FY86, total expenditure by the Treasury was budgeted at G 1,332 million, G 1,196 million for recurrent spending and G 136 million for investment. Total aid to the public sector was forecast at G 718 million, 65 percent of investment expenditure and 31 percent of total expenditure, (of course, this excludes aid to NGOs and to the private sector). In general, and FY86 was no exception, anticipated aid is always overstated in Haiti's Budget. Past Treagury Expenditure: The Role of EXtr8bugetEry Spending 2.06 For FY81-85, extrabudgetary expenditure (that not explicitly programmed in the budget) took on major significance. Table II.2  -9- presents Treasury recurrent and investment expenditure since FY81 (as noted, the FY86 budget breaks new ground and its detailed categories cannot be applied to the past). The role of extrabudgetary spending can be clearly seen; on average in FY81-85 it amounted to 22 percent of that anticipated. Except for FY81, extrabudgetary expenditure was largely for recurrent items. Some of these were proper recurrent purposes, such as debt service, which were not included in the budget belore FY86. Most recurrent extrabudgetary expenditure was ad hoc and unproductive, however. Quantitative estimates are not available but it is known to include military jet aircraft, outlays for the FY85 referendum, and a presidential yacht. Extrabudgetary capital expenditure has been both investment and financial. The former was represented largely by the new public sugar mill and inventory accumulation at certain public enterprises and agencies. linancial capital expenditure was in the form of the purchase of stock in Ciment d'Haiti. There has been no extrabudgetary spending since February 1986. Table 11.2: TREASURY EXPENDITURE, FY8185, AND BUDGET, FY86 (G million at current prices) FY81 FY82 FY83 FY84 FY85 FY86 Budget BUDGETARY 877 962 901 1,005 1,117 1,332 Ministries and Offices 538 567 540 679 702 712 Other recurrent 204 251 266 230 288 484 Investment 135 144 95 96 127 136 EXTRABUDGETARY 314 41 163 301 273 - Recurrent 84 12 14 3 T; Y - Investment 230 29 10 55 - Financial capital - - 10 32 - - TOTAL 1,190 1,003 1,065 1,306 1,390 1,332 Sources: Statistical Appendix Tables 2.3 and 2.5. 2.07 Unproductive extrabudgetary spending has been the principal cause of Haiti's public sector deficit, and hence its balance of payments deficit. The elimination since February of all extrabudgetary expenditure is thus very significant. When all anticipated expenditures are consolidated in the budget, total budgeted Treasury expenditure in FY86 appears to increase by a massive 19 percent from FY85. Compared to actual expenditure, however, it falls by four percent in nominal terms. However, a system of strict financial controls, including the public enterprises, has yet to be established. Audits have been started of many of the public enterprises. Recurrent Expenditure: Ministries and Offices 2.08 Recurrent expenditure can be categorized in both functional and economic terms (Table I1.3). Functionally, expenditure on ministries and offices has accounted for 57-68 percent of total recurrent spending,  -10- averaging 58 percent, and was scheduled for 60 percent in FY86. It is split about equally among General Public Administration (Justice, Interior and National Defense, Armed Forces, Information and Public Relations, Presidency, Foreign Affairs, Legislative Chamber and Religion); Economic Services (Agriculture, Public Works, Transport and Comnuications, Commerce and Industry, Economy and Finance, Planning} Mines and Energy Resources and Superior Court of Accounts); and Social Services (Education, State University, Health, Social Affairs, and Youth and Sports). However, it' has recently emerged that a significant portion of the budget of the Ministry of Economy and Finance was used for the Armed Forces and for Foreign Affairs; this has now been transferred to the budgets of these ministries. The key expenditure ministries from a development perspective are Agriculture (MARNDR), Public Works, Transport and Cowmunications (MTPTC), Education (MEN) and Health (MSPP). Table II.3: RECURRENT EXPENDITURE ON MINISTRIES AND OFFICES, FY82-85, AND BUDGET, FY86 (G million at current prices) FY82 FY83 FY84 FY85 FY86 Budget Total 567 540 679 691 712 Functional Classification General Public Administration 184 177 235 249 248 Economic Services 182 175 238 223 246 Social Services 201 188 206 220 219 Economic Classification Personnel a/ 400 407 491 528 463 'Sans Justification" .. .. .. 35 21 Other a! 167 133 188 128 228 a/ Armed Forces personnel costs are excluded from Personnel in FY86 and included in Other. Sources: MEF; Statistical Appendix Tables 2.7-2.8. 2.09 Government working hours are from 8:00 a.m. to 2:00 p.m. Monday through Friday, much less than the eight-hour day normal in the private sector. 2.10 An economic categorizstion shows that personnel costs represent the major component of spetiding by ministries. They are budgeted at 65 percent of the total in FY86, down from the 71-77 percent range of FY82-85 only because the Armed Forces' expenditure (including salaries) is now included within the 'Other" category. Budgetary data show that in FY84-86 there were about 32,400 civil servants, excluding military personnel and employees of public enterprises, twice the FY71 level and nine percent more than in FY82. Official numbers have grown at 4.8 percent per annum since FY71,  -11- though the rate since FY80 has been a lower 2.4 percent. Whatever the rate, it has put relentless pressure on the recurrent budget. Public employment has been frozen through September 1986. 2.11 The official number of public employees is 32,400. The actual number, however, may be as high as 57,000. There is no current register of civil servants, at either the level of the Public Service Commission or the individual ministries. Yet the basis for establishing one exists: all public employees are now paid with computer-generated checks. Recently, however, the Public Service Commission compared official numbers of public employees with the results of a Ministry of Planning investigation of actual numbers for FY79-82 and found there were over 75 percent more civil servants than previously thought (Table I1.4). No measures have been taken since FY82 to lead to any change in the ratio of actual:official numbers. In In addition, many officials have been paid both salaries and 'expenses"; this practice will be eliminated in FY87. Table II.4: ACTUAL AND BUDGETED PUBLIC EMPLOYMENT, FY79-82 (thousands) FY79 FY80 FY81 FY82 Actual 41.2 46.3 50.8 54.9 Budgeted 22.4 27.9 29.3 29.6 Ratio 1.83 1.66 1.73 1.86 Source: Public Service Commission. 2.12 The difference in numberrn is explained partially by the use of the investment budget to finance recurrent costs, including employees (see below). According to the Director-General of the Public Service Commission, it is also the result of the widespread existence of unproductive employees, frequently reflecting the use of one budgeted salary to pay two or more employees--because of their political connections--who are not expected to perform on the job, or ineeed, to appear at the ministry. Other corrupt practices include the payment of salaries to people who do not exist at all, the funds in fact going to those politically favored. The mission did not examine all ministries but found evidence of these phenomena at MARNDR, MTPTC, MEN and 1SPP (Chapters III, VI, X and XI). Active steps have been taken since February 1986 to begin to deal with these problems. 2.13 In October, 1985, the Director-General of the Public Service Commission estimated that the current functions of the Haitian Government require less than half the approximate current total number of civil servants. This may not, however, take full account of health and education needs. 2.14 Although the public enterprises do not figure in the recurrent budget, their staffing also reflects the broad pattern of public employment. The five public industrial enterprises employed over 1,800 people but cotld also have operated with less than half (Chapter IV). The electricity company has one of the lowest ratios of customers per employee  -12- in the region (Chapter V). The telecommunications company has five times the number of staff needed per 1,000 direct exchange lines (Chapter VII). The Port-au-Prince water supply company has over twice as many staff per 1,000 connections as the average for Latin America and the Caribbean (Chapter IX). 2.15 While personnel costs account for 75 percent of recurrent expenditure by ministries, they represent a greater proportion at three of the four key development ministries (Table II.5). MARNDR, MEN and MSPP are shackled because they have completely inadequate funds for necessary non-personnel operating costs, including, for example, transport for agricultural extension workers and school inspectors, maintenance of health facilities and schools, teaching materials, and drugs and medical supplies (Chapters III, X and XI). Non-personnel recurrent spending at MSPP has recently been increased, however. Yet apparent average annual salaries are much lower at these ministries than at others: G 12,700 at MARNDR, G 7,300 at MEN and G 11,400 at MSPP in FY85 compared to a Government-wide average of G 14,300, excluding the Armed Forces, and a staggering G 25,300 for those in the General Public Administration. A clear pattern therefore emerges of lower relative salaries and inadequate operating funds at large development-oriented ministries and high salaries and a higher proportion of recurrent expenditure available for operating funds in the General Public Administration. The apparatus of the State has been favored over development. Despite the establishment of a Public Service Commission and enactment of a Public Service Law in FY82, there is no Government-wide salary scale or standard employment procedure. The impact of these two factors and the existence of many non-functioning employees has been detrimental to morale among well-motivated staff at development ministries. Recurrent expenditure on health and education was increased in mid-FY86, however. Table II.5: PERSONNEL SHARE IN MINISTRY SPENDING, FY82-85 (percent) FY82 FY83 FY84 FY85 Agriculture 88 96 93 95 Public Works, Transport & Communications 71 78 74 78 Education 92 93 90 91 Health 74 87 87 87 General Public Administracion 64 68 64 66 All Ministries 71 75 72 77 Source: Statistical Appendix Table 2.8. 2.16 Several ministries and Government agencies have made some improvements in their staffing in recent years. MARNDR had begun to reduce unproductive staff in FY85 and planned a ten percent cut for FY86. During FY85, MEN attempted to clean payrolls of salaries for nonexistent staff; it identified 400 fictitious posts and cut them in the FY86 budget. MSPP planned in FY86 to transfer staff out of Port-au-Prince to the rural-areas  -13- where health services are more urgently needed. The housing agency EPPLS halved its staff and quadrupled its productivity between FY83 and FY85, although it experleneed some reversals in FY86. Most ministries have also been hindered by their inability to transfer even small items from one budget line to another, i.e., typically from salaries to materials, withv'ut the express permission of the Ministry of Finance for each transfer within each budgetary unit at the individual ministry. 2.17 Civil servants are entitled to apply for a psnsion after reaching the age of 55 and after 25 years' service. They did not until 1986 necessarily receive one; this was at the discretion of the incumbent Minister when they applied but is now automatic. After 25 years, the pension is 50 percent of salary, rising to 75 percent after 30 years. Teachers, doctors and nurses were not subject to the age minimum and could apply for 100 percent pensions after 25 years service; this is no longer the case. Total pension provisions in the FY86 budget amounted to only G 24 million. Civil servants also have seven percent of their salary deducted for pensions which are apparently paid out of current income; this would have yielded G 37 million in FY85. The total pension bill is thus apparently about G 65 million. Recurrent Expenditure: Other Identified Purposes 2,18 The FY86 budget is useful in that it identifies for the first time some recurrent spending other than for ministries and offices. These include G 29 million for pensions, as noted; G 17 million in transfers to local governments which themselves raise almost no revenues; and G 12 million to pay the salaries of employees of the National Agricultural and Industrial Development Bank (BNDAI). The agricultural operations of this credit bank are shown to be highly inefficient in Chapter III; the same is true of its industrial ones and BNDAI urgently needs rehabilitation. There is no case for an income-generating development bank having all of its staff costs paid by the Treasury. Other identified recurrent expenditure is very minor. Recurrent Expenditure: Outstanding Obligations to Dometic Suplier. 2.19 About G 49 million in the FY86 budget represents cumulative purchases by the Government that have not been paid for. It is confusingly included within amortization in the budget document. Once it is spent in FY86, there will be no reason for further expenditure in this category, beyond any normal fiscal year-to-fiscal year short-term float. Recurrent Expenditure: Debt Service 2.20 Debt service by the Treasury consists of amortization of and interest on foreign commercial and concessional loans, interest to the IMF, and interest on Government bonds held by the Central Bank. In the past, debt service was contained within extrabudgetary spending but was in no way an improper expenditure. The Government is now trying to improve its handling of debt service. 2.21 Assuming that all new debt incurred by the Government is on concessional terms, service on which will be minor in the immediate  -14- future, debt service through the 1980s can be approximately projected on the basis of existing debt (Table 11.6). In real terms, debt service by FY89 should be only two-thirds its FY86 level as IMF and commercial debt service is reduced. Table 11.6: GOVERNMENT DEBT SERVICE, FY86, AND PROJECTED, FY87-89 (G million at current prices) FY86 FY87 FY88 FY89 AmortIzation 42 20 20 20 Foreign commercial debt 27 -- - Foreign coneessional debt 20 20 20 20 Interest 114 115 110 110 Government donestic bonds 60 - IMF 35 30 20 15 Other external debt 19 20 20 20 Total debt service 156 135 130 130 Total debt service (FY86 prices) 156 127 113 105 Sources: MEF; IMF; mission estimates. Recurrent Ezjpenditure: Unidentified 2.22 In the FY86 budget, G 204 million, nine percent of proposed total expenditure and 15 percent of Treasury expenditure, was unidentified, G 179 million labelled as "Special Obligations' and G 25 million as 'acreditifs." This was discretionary expenditure, G 43 million greater than the total public investment of G 136 million to be financed by the Treasury. While all governments have discretionary funds, their level has been excessive in Haiti in recent years and has largely contributed to the public sector deficits. Some of this unidentified expenditure in the past has been for purposes such as debt service and pensions which are now explicit in the FY86 budget. Most, however, may have been for improper purposes that would not stand scrutiny, and can be cut. Precisely how much of the G 179 million budgeted for FY86 was for legitimate purposes cannot be determined. An audit is necessary of past expenditure to determine how much should be continued. Since April 1986, these budget categories have been drastically reduced. 2.23 Unidentified expenditure also exists within the Ministries and Offices category of the budget. Most departments and units have a "Sans Justification" budget line. This accounted for five percent of their spending in FY85 and was budgeted at three percent in FY86. It accounted for 47 percent of the G 5.6 million budget of the Administrative and Accounting Directorate of the Ministry of the Presidency (abolished in February 1986), 65 percent of that of the G 8.9 million budget of the Administrative Affairs Directorate-of the Ministry of Information and Public Relations, and for lesser percentages of other ministries. It was not possible for the mission to examine the allocation of this and other  -15- discretionary spending; some is reasonable, however, e.g., for foreign embassy expenses under the Ministry of Foreign Affairs. Again, an audit is necessary to determine what should be continued. Starting in FY87, the Government intends to identify all spending by specific purpose. 2.24 Taken together, these two unidentified categories accounted for over G 225 million of the FY86 Treasury budget, or 17 percent of total Treasury spending. Evaded Recurrent lxpenditure: The Role of N60. 2.25 The Government has, in a sense, avoided some development-oriented recurrent expenditure because of the large and growing role of Non-Governmental Organizations (NGOs) in Haiti. There are at least 250 and probably over 300 NGOs; information on their activities is hard to assemble. However, the Ministry of Planning recently surveyed 71 NGOs including the largest and most important. Their total annual spending averaged G 133 million in FY82-85, the bulk for community development (Table 11.7). However, NGO spending on education and health almost tripled over the period. Most NGO development-as opposed to missionary or humanitarian--activities have been financed by USAID and CIDA, which increasingly channelled their bilateral aid away from the Government during the first half of the 1980s. (NGO activities and expenditure in agriculture, education and health are discussed in Chapters III, X and XI). Agriculture and community development spending by NGOs in FY85 amounted to 66 percent of the public recurrent budget for agriculture, that on education to 50 percent, and that on health to 45 percent. Table II.7: EXPENDITURE OF 71 NGOs, FY82-85 TG million at current prices) FY82 FY83 FY84 FY85 Agriculture 6.6 8.6 7.9 8.3 Education 16.4 41.8 39.2 45.8 Health 11.9 10.1 29.7 39.0 Subtotal 34.9 60.5 76.8 93.1 Community Development 94.6 a/ 59.4 48.9 54.3 Cottage Industry 1.6 1.1 2.5 2.8 Total 131.1 121.0 128.2 150.2 a/ Probably unusually high because of post-Hurricane Allen relief efforts. Source: Ministry of Planning. Investmet Expediture 2.26 Public investment is, in theory, made in the context of a five-year planning process. The FY82-86 plan was abandoned in FY84 because broad economic developments (Chapter I) had made it unrealistic and was  -16- replaced with a special two-year plan for FY85-86. A new five-year plan for FY87-91 was under preparation in early 1986 by a cumbersome system of interlocking sectoral and general commissions examining the desired shape of the country in the year 2000 and beyond. The broad objectives of the current two-year plan are to: (a) increase production, employment and exports in agriculture; (b) develop an integrated industrial sector based initiaily on the growth of the export assembly subsector and eventually on backward linkages to industries producing inputs from both domestic and imported raw materials; (c) improve the quality of human resources, through strengthening basic education and technical training; (d) promote decentralization away from Port-au-Prince; and (e) correct the public finance and balance of payments deficits and move gradually toward financial stability. No significant departures from these objectives are anticipated for the next plan. The first three objectives affect the sectoral allocation of public investment (agriculture, infrastructure for industry, and education), the fourth its spatial distribution, and the last its level -(financeable without resort to expanding credit to the public sector). 2.27 Ithere is, however, little relationship between these objectives, public policies and the list of projects in the plan. The projects are not necessarily inconsistent with the objectives; rather they have been selected using the totally different criterion of cataloguing the projects that donors are financing or intend to finance. Public investment is financed either domestically (by the Treasury and by public enterprises' internal cash generation) or by external aid donors. In recent years, donors have financed about 63 percent of total public investment, two-thirds of it in agriculture, transport, energy and health (Tables II.8 and 11.10). It is also striking how many donors are involved in each sector, particularly agriculture, education and transport. Table 11.8: FINANCING OF PUBLIC INVESTMENT, FY84-85, AND BUDGET, FY86 (percent) FY84 FY85 FY86 Budget Domestic resources 34 41 26 Treasury 9 15 12 Public enterprises 26 26 13 External aid 66 59 74 Public investment (G million) 1,067 793 1,113 Source: Statistical Appendix Table 2.12. 2.28 These data are approximate. While it is hard to determine the allocation of much recurrent expenditure, at least the level is known. This is not the case for investment spending, except for that of the Treasury. There are major discrepancies among budgeted and actual sector allocation data coverage by the Ministry of Planning stemming largely from the  -17- inconsistent inclusion of public enterprise investments; between disbursement data reported by the Ministry of Planning and by the Ministry of Finance (Statistical Appendix Table 2.17); data submitted to the mission on a project-by-project basis by both the Government and donors (Appendix III); overall public sector capital expenditure reported to the IMF (Statistical Appendix Table 2.3); donor grant disbursements supplied by the Ministry of Planning (Statistical Appendix Table 2.18); the same data as reported to the IMF and the World Bank (Statistical Appendix Table 2.19); and data on individual sectors supplied variously by the sector ministry and the Ministry of Planning. This applies even to budgetary data. In the education sector, for instance, MEN records show total budgeted investment expenditure at G 177 million in FY82-84 and realized expenditure at G 105 million; MP, by contrast, show budgeted spending at G 181 million and executed at G 152 million. 2.29 These inconsistencies result from a chaotically complex planning, budgeting, monitoring and disbursement system. So complex is it that it cannot be examined in detail in this report. In brief, however, the Ministry of Planning was in theory responsible for working with planning units in sectoral ministries to draw up five-year development plans which are translated into annual investment budgets. In fact, there was no effective planning in either sectoral ministries or the Planning Ministry. The latter, indeed, concentrated in its five-year plans on defining long-term physical objectives without taking any account of financial and economic constraints. It did not even link these objectives to the proposed project list. As donors finance the bulk of all public investment, the projects in the five-year plans tended to be no more than a compilation of ongoing and likely donor-financed projects together with certain proposals for projects to be financed entirely from domestic resources. There is no project evaluation at all, unless donors carry it out. The annual investment budget was drawn up by the Ministry of Planning within an overall envelope for Treasury investment spending established by the Ministry of Finance but taking no account of the likely impact of this envelope for counterpart funds on donor disbursements. Hence, annual investment programs were always smaller than those foreseen in the five-year plan and actual investments were always less than those budgeted even on an annual basis (Table I1.9). Disbursement procedures were extraordinarily complicated, requiring authorizations within sector ministries, and by the Ministry of Planning before the Ministry of Finance may release funds. In practice, even full authorizations did not necessarily result in full disbursements. Monitoring of projects and expenditures was a function of both sectoral ministries and thie Ministry of Planning; in fact, this was only performed for Treasury-financed expenditure. To give another example from education, MEN records showed disbursements on two IDA-financed projects at the end of FY85 at G 1.6 million whereas IDA records showed them at G 11.1 million. 2.30 This complex and obfuscating system is not just inconvenient and impractical. A context in which channels of information are ineffective or nonexistent--and controls are numerous but uncoordinated and unrelated to the objectives of the investment program-provided numerous opportunities to divert resources away from their original purpose. Indeed, it was probably designed to work in this manner. The present Government intends to reform the system.  -18- Table 11.9: PLANNED, BUDGETED AND ACTUAL PUBLIC INVESTMENT, FY81-84 a/ (G million) Domestic Foreign Total Financing b/ Financing Five-year plan provision 1,260 2,640 3,900 Revised annual provisions 1,091 2,087 3,178 Actual investment 772 1,459 2,232 Ratios (x) Actual to revised 71 70 70 Actual to 5-year plan 61 55 57 a/ Excludes USND sugar mill and purchase of shares in Ciment d'Haiti, neither of which were included in plans or budgets. b/ "Domestic financing" here follows Government practice and includes USAID PL-480 counterpart funds which are properly foreign financing; these account for about 30 percent of the domestic financing. Sources: REF; MP; mission estimates. 2.31 The mission has done its best to reconcile conflicting data on the investment program. Nonetheless major problems remain. In FY84-85 domestic financing was 67 percent concentrated on energy, telecommunications, and industry (Table II.10). 'Other' sectors-those not examined in detail by the mission because they appeared to have little linkage to development-- represented a further 11 percent. A public enterprise's internal cash generation is, in general, used only for investment in that same enterprise. Treasury funds, however, may be used for any sector and their distribution perhaps gives the best indication of the actual priorities of the Government: 'other," urban, agriculture and transport account for 76 percent. "Other' is largely composed of regional and special activities, much of which is probably disguised recurrent spending. It is very clear that the priority said to be attached to agriculture and education is not reflected in practice in the investment program; this is apparent also from the inadequate recurrent budgets of these two sectors. Indeed, when the first four years of the current plan period are compared with the FY77-81 plan, agriculture's share of public investment has fallen from 17.3 to 15.6 percent and education's from 7.6 to 4.4 percent. 2.32 The mission was able to assemble reasonably consistent sectoral allocation data for FY84-85. The results, in Table 11.10, exclude the huge emphasis the public sector has placed on industry, through construction of the USND sugar mill in FY81-83, and through the acquisition of bankrupt or faltering private enterprises through the purchase of shares in Ciment d'Haiti and ENAOL and acting as receiver for the USN sugar mill (Chapter IV).  -19- Table 11.10; SECTORAL ALLOCATION OF PUBLIC INVESTMENT ACCORDING TO FINANCING SOURCE, AVERAGE FY84-85 (percent) Total Treasury Total Domestic External Aid Agriculture 17 14 14 24 Industry 7 5 13 4 Energy 21 1 34 13 Transport 13 11 7 17- Telecomminications 9 - 20 2 Urban 8 17 5 9 Water 2 5 3 2 Education 5 4 1 7 Health 9 7 2 13 Other 11 34 11 11 Average (G million) 930 108 345 585 Source: Statistical Appendix Tables 2.13-2.16. 2.33 The FY86 budget appeared to make some progress. The total number of projects was cut from about 300 in FY85 to 110, the bulk of those eliminated being "projets d'appui" which had no visible outputs and only served to provide employment. Staff cuts did not follow this step, however, implying that employees were absorbed by other projects. Even the 110 projects turn out to consist of 145 because of the lumping of several old projects, especially in agriculture and energy, under new titles. Projects for which the Ministry of Planning was authorizing disbursements in October 1985 numbered 166. In mid-FY86 the Government cut Treasury development spending from about G 11 million per month to G 3 million. It will rise again in FY87, however. The Return on Public Investmoet 2.34 In general, public investment has not been efficient in Haiti. Chapters III-XI discuss past investment in each sector. Had public investment in agriculture since FY76 yielded a rate of return of 12 percent, which is the approximate opportunity cost of capital in Haiti, agricultural GDP would have been seven percent higher in FY85 and thus overall GDP perhaps two percent-higher. More important, efficient public investment in agriculture would probably have raised the returns on private agricultural investment. The resource loss resulting from the protected monopoly status of five major public industrial enterprises in FY82-85 was equivalent to nearly four percent of GDP. Action taken in July 1986 to close or divest two of these, the ENAOL vegetable oil factory and the USND sugar mill, will reduce their resource loss. Returns are harder to quantify for other sectors but the pattern of inefficient investment is repeated across the board, with the exceptions of the power sector and most roads.  -20- 2.35 These poor results are all the more astonishing when one considers the number of external aid agencies involved in Haiti, the staff resources they have devoted to project planning, and public investment's approximate ten percent share of GDP in FY81-85. 2.36 Why has public investment yielded such low returns? First, as discussed in Chapter I, necessary structural reforms have not been made to set appropriate incentives for private producers in agriculture and industry. This affects the return on public investment directly. An irrigation scheme cannot sustain improved crop yields if there is no system for allocating water or paying fo- maintenance. In addition, public industrial enterprises enjoyed a protected monopoly status that enabled them to operate so inefficiently that they have become a drag rather than contributing to increased output and income. ENAOL lost its vegetable oil monopoly in April 1986. 2.37 Second, much so-called "investment' financed from domestic resources is not in fact investment. Investment funds are used for salaries, for non-investment purposes, and diverted. Some 50-60 percent of the Treasury contribution to investment is composed of salaries, hence the high proportion of Treasury investment funds used for "other" sectors. This is "rational" in the sense that recurrent budgets are inadequate. The same is true of a lesser proportion of donor-financed investment, especially for USAID PL-480 counterpart funds. Some donors are now stopping the payment of such salaries, usually in the form of salary supplements, but this may simply shift the burden onto the Treasury. Other non-investment purposes have included financing the operating costs of the supermarket of the now-disbanded Volunteers for National Security (G I million in FY85). Diversions from the investment budget cannot be quantified but are confirmed by overwhelming anecdotal evidence from donors. 2.38 Third, project selection has been poor. This is not just a poor fit between selected projects and broad economic objectives but also the selection of individual projects and project alternatives. The Government has no capacity to appraise projects and has tended to accept projects proposed by donors without assessing their economic impact. Projects financed solely by the Government or with external commercial loans, notably the USND sugar mill and most salary-paying projects, have been selected for reasons having nothing to do with economic returns. Very few donors have appraised projects they have financed, as evidenced by replies to questionnaires summarized in Appendix IV. In terms of numbers, less than five percent of all projects have had an economic rate of return estimated. 2.39 Fourth, necessary recurrent expenditure to yield acceptable returns on investments has not been provided at either the project or the sector level. At the project level, neither the Government nor donors could provide any information to the mission (Appendix III) on the likely recurrent expenditure impacts of investment projects, the major exception being the health sector. While donors have routinely criticized the Government at aid meetings for its failu' i to assess these costs, they have done nothing to estimate them even for * 'ects they finance. Failure to provide necessary non-personnel operatir.. inds was discussed above under recurrent expenditure. In some sectors, it also included salaries themselves; schools have no qualified teachers because salaries are too low  21- even to attract potential primary teachers to attend free teache. training Normal Schools. If physical construction--irrigation schemes or schools--cannot be operated, it cannot yield returns. General Recon_endations 2.40 As discussed in Chapter I, reforms in public expenditure should form one part of a major program of structural reform designed to set the Haitian economy onto a path of sustained growth. Such a program has now been initiated. 2.41 Within this context, four of the five broad economic objectives of the current plan are entirely appropriate. Agricultural and industrial growth, especially for export, must be achieved. Education, especially the expansion of literacy and specific technical training, is essential for sustained development. Financial stabilization must occur if reserves are to be rebuilt and the economy is not to operate on the edge of potential disaster. Since February, it has begun. 2.42 One objective is only partially appropriate, however: decentralization. Over time, there is no question that a more even spatial distribution of the benefits of economic growth is called for; Port-au-Prince is currently heavily favored (Chapter VIII). First, however, there must be economic growth. Industrial growth, at least in the next four to five years, must come by expanding the already industrialized area, i.e., Port-au-Prince. It already has much of the necessary infrastructure and efficiency dictates that it will be easier to supplement existing facilities than start new ones. In practical project terms, little change is necessary in the investment program as a result of this recommendation: there should be no new pub'ic industrial parks outside Port-au-Prince (Chapter IV) and there is at present no case for major extensions to the airport at Cap Haitien (Chapter VI). By contrast to industrial location, there should be an increased decentralization of social sector expenditure; rural education and health have been badly neglected (Chapters X and XI). 2.43 Decentralization did, hoaever, occupy much of the time of the Ministry of Planning, although this work was not based on any analysis, e.g., of differentiated economic returns or social impacts from investments in different parts of the country. A legal framework being established in early 1986 to create decentralized planning directorates in four regions and three subregions should not proceed further. 2.44 What should be the objectives of public expenditure in light of the four appropriate broad economic objectives? First, its overall level must be such that it can be financed now and in the future with available resources and without resort to destabilizing expansion of credit to the public sector or foreign commercial borrowing. Extrabudgetary expenditure must therefore stop; this has been the case since February 1986. Sustaining this will require an improvement in the Ministry of Finance's financial control, including of public enterprises, requiring some reorganization and external technical assistance. Eliminating extrabudgetary expenditure does not mean that budgetary allocations should not change during a fiscal year but rather that explicit increases in one sector should be offset by corresponding cuts elsewhere or by increased revenues.  -22- 2.45 Second, public expenditure should concentrate in the Immediate future on increasing the returns to existing investment, both private and public. This means, In particular, providing adequate recurrent funding for agricultural extension and education; completing priority investments whose execution has been delayed, especially in agriculture; and providing the necessary infrastructure for agriculture and industry. The Government may also wish to extend official working hours to increase daily productivity. 2.46 fhird, the process of planning, budgeting, disbursing and monitoring public expenditure should be overhauled to enable the elimination of waste and corruption, proper project selection and the preparation of a rational, efficient and coordinated public investment program. The present complex system should be greatly simplified. Major institutional changes beyond the scope of this report are clearly needed, very probably along the lines of devolving planning, project selection and monitoring to the sector ministries while leaving the Ministry of Finance in charge of all budgeting and disbursement anid introducing forward budgeting and medium-term financial planning. 2.47 Planning in Haiti requires drastic revision; the only effective function of the former Ministry of Planning was to authorize disbursements, which should be a Ministry of Finance responsibility. There is clearly, however, a need for a small unit concerned with medium-term development, the broad lines of the economy, the overall shape of public investment, the performance of sectoral ministries and public enterprises, and the like. World Bank experience elsewhere, especially in Africa, shows that satisfactory budgeting can only be achieved when the same Minister is responsible for both recurrent and capital expenditures. 2.48 The five-year plan cycle is inappropriate to the modern and rapidly changing world. It could be replaced, as it has in many countries, with a rolling three-year public investment program. Following this recommendation, the rest of this report concentrates on the current FY86 and the next three years, FY87-89. Project selection must be improved by the introduction of routine project appraisal (or at least, appraisal of projects above a certain size) and by full consideration of the recurrent cost implications of investment projects. 2.49 Coordination among Government and donors also needs drastic improvement. Coordination has been effective in power but not elsewhere. Sectors with coordination problems are agriculture, education and transport (Chapters III, X and VI). Beyond these specific sectors, however, individual donors may wish to consider limiting their involvement to fewer sectors or, alternatively, increasing cofinancing with different donors taking the lead in different sectors. The development of a proper public investment program could help enormously in this regard. A central point should be established within the Government for donor coordination and contact. 2.50 All these institutional changes will require time and appropriate technical assistance if they are to be implemented. There is no reason that virtually all of them, however, could not be in place by the early 1990s.  -23- lePublic Ex iture f gram, n86-89 2.51 Chapters III-XI make detailed public investment proposals for the different sectors in the context of the above general recommendations. For the first time, recurrent expenditure requirements are also explicitly determined for agriculture, public works, education and health, the four developments within each. This section aggregates the recommendations for both recurrent and investment expenditure of the rest of the report. It shows that they are feasible without any real increase in Treasury expenditure and within the projected external aid envelope of even the Low case economic scenario projected in Chapter I. This is largely because identifiable areas of waste, corruption and inefficiency in the present budget were so significant. Further steps are now underway that should permit the elimination of further waste in the future, after appropriate audits and studies. 2.52 Table II.11 summarizes the recommended recurrent expenditure levels for the develkpment-oriented ministries, largely derived on the basis of increasing non-personnel funds essential for effective operations. With the exception of MSPP, the recommended levels Include recurrent expenditure previously concealed within the investment budget; MSPP data did not permit this to be done at this stage. A sizeable increase from the FY86 budget is called for. Even these increases, however, assume that NGOs will continue to provide development services at least at their FY85 level. The recommended levels include potential efficiencies within each ministry (discussed in the relevant chapter) but do not assume any change in the relative structure of public service salaries, except for those anticipated for teachers under the Teacher Charter (Chapter X). Table II.11: RECOMMENDED RECURRENT EXPENDITURE OF DEVELOPMENT MINISTRIES, FY86-89 (G million at FY86 prices) FY86 FY86 FY87 FY88 FY89 Budget MARNDR 37 85 85 85 85 MTPTC 77 82 96 84 78 MEN 105 a/ 115 119 119 125 MSPP 90 85 90 95 100 TOTAL 309 367 390 383 388 a/ Includes the State University, UEH. Sources: MEF, FY86; mission recommendations. 2.53 If these recurrent funds are made available and pricing and structural reforms are made (Chapter I), existing and new public and private investments should start to yield higher returns. If they do not, there will be little point in new public investment, at least in agriculture, education and health.  -24- 2.54 Table 11.12 sets cut the recommended public investment program, on the assumption that the recurrent expenditure recommendations are followed. Like them, public inv-stment would remain roughly constant in real terms through FY89, above its actual level in FY85 but below the typically inflated expectations about external aid disbursements that characterize the FY86 budget. Further real increases are not feasible until recurrent expenditures and day-to-day operationo of relevant ministries have improved. The recommended levels do, in fact, include some real increase in fixed investment, since they exclude formerly disguised recurrent expenditure. Table II.12: RECOMMENDED PUBLIC INVESTMENT PROGRAM AND FINANCING, FY86-89 (G million at FY86 prices) FY86 FY87 FY88 FY89 FY86-89 G million Percent SECTORAL TOTAL 800 806 821 877 3,304 100 Agriculture 155 155 155 155 620 19 Industry 25 50 15 - 90 3 Power 192 136 155 161 644 19 Transport 88 109 120 148 465 14 Telecommunications 61 50 50 50 211 6 Urban 67 82 94 102 345 10 Water 40 53 50 67 210 6 Education 72 42 42 42 198 6 Health 70 100 110 121 401 12 Other 30 30 30 30 121 4 FTNANCING TOTAL 800 806 821 877 3,304 100 Domestic 251 300 286 259 1,096 33 Treasury 133 123 125 129 510 16 Public Enterprises 118 177 161 130 586 18 External Aid 550 506 535 618 2,209 67 Existing 543 259 206 174 1,18i 36 New 7 247 329 444 1,027 31 Sources: Statistical Appendix Tables 2.23-2.26. 2.55 Compared to the recent past, the major shifts in sectoral allocations are to increase outlays in agriculture, health, education and water supply and decrease those for industry, transport and "other." All these shifts are consistent with the plan's objectives, with the apparent exception of increases for health and water supply and the decrease for industry. The former, however, is largely a matter of accounting, because it has not proved possible to net recurrent expenditure out of the health  -25- sector's investment budget. Hence, the increased recurrent budget for MSPP in Table 11.11 is much lower in proportional terms than those for agriculture and education. The decrease for industry is because the record of public industrial enterprises in Haiti has been so poor that all future industrial investment should be left to the private sector. 2.56 The recommended level of public investment would permit private investment to increase considerably, especially under the Reform case scenario of Chapter 1. By FY89, for example, the private sector share of total investment could rise to 58 percent, up from 44 percent in FY85. 2.57 The recommended public investment program should pose few financing difficulties. The Treasury contribution in each year from FY86-89 never exceeds the G 136 million budgeted for FY86. Internal cash generation from public enterprises is two-thirds for EdH (power) and TELECO (telecommunications), both companies that should have little difficulty generating the required operating surpluses (Chapters V and VII). Aid disbursements will be required of about G 2,200 million (US$440 million), of which about G 1,200 million (US$240 million) has already been committed. New commitments of about US$200 million are therefore called for. Average annual aid disbursements of about US$110 million compare favorably to likely total aid even under the Same case scenario of Chapter I. This involved about US$190 million per year by FY89, including over US$110 million for the fisc even after aid for NGOs and for credit has been excluded (This comparison also indicates Haiti's need for aid to support NGOs and the balance of payments and not simply to finance public investments). 2.58 The financing issue therefore becomes one of where the resources can be found to finance an increased recurrent budget for the four development-oriented ministries. So extensive was the waste in the current budget, however, that this poses little problem. The question becomes, rather, what to do with the extra resources that would be released by eliminating waste after FY86. 2.59 *able II.13 illustrates this. Ministry expenditure is that recommended above for the four development ministries plus existing expenditure for other ministries in FY86. Other ministries are cut by the assumed laying off of non-performing and nonexistent staff aad by eliminating the "sans justification" category. This ten percent cut is very conservative; no ministry that was examined by the mission had as few as ten percent of its staff in these categories. The effect of these changes will be to shift expenditure toward development and reduce the relative share of the General Public Administration. Of course, reducing staff is politically very difficult; nonetheless, a start must be made. The subsidy to BNDAI is cut out of the other identified recurre. category from FY87 onward. Obligations to domestic suppliers are assumed to be paid in full in FY86 and then not to reappear. Debt service is on the basis of existing debt and as projected in Table 11.6. Unidentified recurrent spending is reduced to 25 percent of its FY86 budget level on the assumption that the bulk of it is inappropriate and unwarranted; assuming that half has already been spent in FY86, this 75 percent cut is only applied to the remaining half for FY86. Investment expenditure is as recommended above.  -26- Table II.13: RECOMMENDED TREASURY EXPENDITURE,IV FY8649, AND BUDGET, FY86 (G million at FY86 prices) FY86 FY86 FY87 FY88 FY89 Budget Ministries and offices MARNDR, NTPTC, MEN & MSPP 309 367 390 383 388 Other 403 403 363 363 363 Other identified recurrent 98 98 85 85 85 Outstanding obligations to domestic suppliers 49 49 - - - Debt service 156 156 127 113 105 Unidentified recurrent 179 112 45 45 45 Investment 136 133 123 125 129 Total 1,330 1,318 1,123 1,104 1,105 a/ From domestic sources. Sources: Tables 11.1, 11.2, II.ll and 11.12. 2.60 These modest shifts and cuts in expenditure mean that the FY86 recommended program could have been achieved within likely resource availability, while those for FY87-89 will yield a G 200 million annual current surplus for the Treasury if it continues to receive only the revenues it anticipates in FY86. It could be more if donors start to finance recurrent spending on an explicit basis, although this does not seem necessary. It will be more if there is any economic growth and revenues rise; indeed, it could increase to about G 400 million by FY89 in the Reform case scenario of Chapter I. The G 200 million annual surplus is so large that even if some of the proposed cuts prove impractical--and this seems unlikely--there will be no difficulty in financing public expenditure necessary for development. They depend critically also on their being no extrabudgetary spending. 2.61 In practice, furthei elimlnation of waste from the recurrent budget is undoubtedly possible. Two stucies may highlight these further economies. First, an audit should be made of the actual allocation of recent recurrent expenditure to determine where the abuses lay (and hence savings found) and also to investigate unexplained categories like "accreditifs." Second, an estimate should be made of the public employment necessary to run the Government, a comprehensive list of current employees should be drawn up, the two should be compared, and appropriate cuts made, using programs to encourage employees to leave the civil service. The appropriate size of the civil service should be determined on a more  -27- reliable basis than current estimates by the Public Service Commission staff; these estimates do indicatet however, that the scope for reductions is huge. Reductions may not necessarily save resources in the first few years--because of the cost of incentives to resign and retire and probably of an increased pension bill--but should lead to important savings over time. A proper pension scheme should also be set up. 2.62 The precise mechanisms of reducing public employment remain to be determined. They could include: (a) development of a wage and salary policy, following the personnel audit; (b) appropriate policy measures, e.g., attrition, early retirement, termination bonuses, etc.; (c) eliminating rigidity In staff rules and regulations; and (4) developing alternative employment opportunities. 2,63 Resources freed by eliminating waste could be used to increase the salaries of those who remain in the civil service, and especially to redress the balance between the General Public Administration and the development-oriented ministries. Teachers and medical personnel could benefit; indeed educational expenditure proposed in Chapter X could perhaps be increased before FY89 if savings can be made relatively soon. 2.64 The institutional reform and hence technical assistance needs of the program proposed above would be very significant, but nonetheless must be met. Priority areas should be (a) redressing the information chaos; (b) instituting financial control; (c) improving the process of budgeting and investment planning, especially forward budgeting and medium-term financial planning; (d) coordinating donors and resisting donor pressure for inappropriate projects; and (e) controlling public enterprises.  -28- GRAPTIfR III kGRICULTURE The Sector and Its Planning 3.01 Haitian agriculture has stagnated for at least ten years and probably much longer. In the last decade, per capita agricultural GDP has declined at 1.5 percent per annum from G 363 to G 312 in PY76 prices. Rural per capita incomes hover around a very low level subsistence equilibrium. This equilibrium can only oe broken through a series of far-reaching structural reforms and improved pricing policies, combined with more efficient public and private investment and improved public services to farmers.1/ 3.02 The principal ministry responsible for agriculture is that of Agriculture, Natural Resources and Rural Development (MARNDR); other ministries playing important roles include both Community Development and Plan (MP), each of which runs regional development projects with agricultural components; Commerce and Industry, which is responsible for export promotion; Economy and Finance (MEF), which oversees public agro-industries; and Public Works, Transport and Comumnications, which builds and maintains rural roads. Excluding agro-industries and rural roads (which are discussed in Chapters IV and VI, respectively), Table III.1 shows that some 21 percent of the public agricultural investment budget is handled by ministries other than MARNDR. 3.03 In addition to this public investment budget, which is 90 percent financed by donors when U.S. PL-480 aid is taken into account, U.S. and Canadian bilateral aid to agriculture has increasingly been channeled through Non-Government Organizations (NGOs), which now spend at least G 30 million per year, G 23 million alone coming from USAID. For the most part, these NGOs and their budgets are neither controlled nor coordinated by either MARNDR or MP, though MARNDR recently indicated that it wished to play a stronger role toward NGOs involved in river basin management and soil conservation. / Many of the structural and pricing measures are described in World Bank, Haiti: Agricultural Sector Study, Report No. 5375-HA; June 1985, and summarized in Chapter III of World Bank, Haiti: Policy Proposals for Growth, Report No. 5601-HA, June 1985.  -29- Tabe .II-: PUBLIC AGRICULTURAL INVESTMENT BUDGET, FY86 (G million) Source of Funding Internal Externalu/ Total MARNDR 21 185 206 Other Ministries 23 32 55 Total 44 217 261 a/ Includes G 42 million of U.S. PL-480 aid and G 1 million of similar commodity aid from other eountvies. Source: Statistical Appendix Table 3.2. 3.04 Public sector and foreign aid-financed NGO efforts over the last decade to raise agricultural production and rural incomes have had little noticeable effect because the relatively low priority given to agriculture has led to a low level of effective public expenditure in the sector and to waste and inefficiency in the expenditure actually made. Haitian government--as opposed to external aid--resources, both budgeted and actually spent, for agricultural investment declined every year but one between FY81-85. Investment was highly inefficient. Had MARNDR's investment budget since FY75 yielded an average return of 12 percent, for example, real agricultural GDP would have been over seven percent higher in FY85 than in FY76, rather than at the same level. With only a six percent return on investment, output would still have been five percent higher; these figures exclude agricultural investments made by other ministries. Since mid-1984, agriculture has, however, begun to receive higher priority in government policy, in terms of both attention and budget. 3.05 Agricultural objectives of the current FY85-86 plan include a shift toward exports, a two percent sector growth rate for FY86, increased employment, increased rural purchasing power and a reduced food deficit. Actions proposed in the plan to achieve these objectives are an improvement in the administration of public development projects; irrigation rehabilitation and extension, especially in the Artibonite valley; re-establishment of the pig population; fiscalization of revenues such as irrigation taxes and fees; expansion of export crops through a reduction in the coffee export tax and the abolition of the export tax on essential oils; reforestation and soil conservation programs to reduce erosion, protect irrigation schemes from siltation, and to generate employment; better distribution of fertilizer and improved seeds to accelerate production of food crops; and improved data-gathering ir the sector. Many of these measures are to be implemented through irtegrated rural development projects. The objectives are reasonable, but might above all have included raising productivity, which is particularly impeded by population pressure on the land. 3,06 Formal planning in agriculture is essentially in the hands of MARNDR's Unite de Programmation (UP), which is also responsible for project evaluation and monitoring. UP has done most of the preparatory work in  -30- MARNDR for the next FY87-91 five-year plan. UP does not, however, systematically evaluate proposed projects and has only examined an average of six per year in the last five years; does not examine the recurrent cost implications of proposed projects; does not have good information on donor-financed project expenditures; and does not monitor at all projects financed entirely by donors or administered by other ministries. Indeed the generally poor quality of available project data in Haiti is most marked in the agricultural sector. Tbe Poor Return on Public Agricultural Investment 3.07 The record of public investment in agriculture has been poor. In part, this has been due to sociological, anthropological, historical and political factors. However, it also reflects three factors more amenable to change: (a) the composition of what is called investment; (b) severe operating and management problems in relevant ministries, above all at MARNDR; and (c) a chronic failure to adopt complementary policy and organizational measures, including the funding of adequate recurrent costs, essential to make investmeats work. These also help to explain why so many projects seem never to be completed but continue on from year to year and indeed from five-year plan to plan. Each is now discussed in turn. A. Coposition of the Investment Budget 3.08 Some 35-50 percent of MARNDR's internally-financed investment expenditure is, in fact, used to pay the salaries of staff who should be contained within the recurrent budget. A further significant but unknown fraction of the investment budget is allocated to non-salary recurrent items to compensate for their virtual non-existence within the recurrent budget (see B below). In addition, donors have used portions of their contributions to MARNDR's investment budget to pay salary supplements and other costs which are more properly recurrent. Thus the proportion of the investment budget that is actually invested is probably no more than half. 3.09 In addition, much of the agricultural investment budget (Table 111.3) goes to rural development and natural resource conservation projects. The former have tended to lack a specific production focus and include many socially desirable components like rural education and water supply that make a contribution to productivity only in the very long term. The latter concentrate on anti-erosion measures on the hillsides which, while important, are unlikely to contribute to major productivity improvements. B. MItistry Operating and Nanageeent Problems 3.10 1. Inadequate Operating Funds. Funds budgeted and spent by MARNDR for operating support--fuel for vehicles, supplies, etc.--are severely inadequate, representing less than five percent of the recurrent budget. The rest goes to salaries. The operating cost proportion of the recurrent  -31- budget is a little higher-about 20 percent-for semi-autonomous organizations under MARNDR, including several regional rural development agencies. Be it five or 20 percent, however, it is much too low for a ministry implementing projects and delivering services to rural Haiti, Forty percent would be reasonable. 3.11 Ministries do not have the discretionary authority under present budgetary procedures to shift funds among categories within the recurrent budget; unspent funds within a line item have to be returned to the Treasury and cannot be used for other purposes. There is thus no incentive to reduce staff in order to increase operating funds. This problem affects all ministries but is particularly acute for MARNDR, which has circumvented it by using the investment budget for operating support. Even this has been difficult, however, because 35-50 percent of this budget also goes to salaries. Indeed, taking the two budgets together, over 85 percent of total recurrent expenditure is on salaries. 3.12 2. Redundant Employment, Incorrect Allocation of Staff, and Lax Work Discipline. MARNDRfs budgeted staff positions increased by five percent in FY85, following several years at around 2,200, reportedly because the then Minister recruited some new staff and increased the ministry's staff outside Port-au-Prince. In November, 1985, the Director of Administration was, however, preparing to reduce total staff by about 10 percent, with the cuts intended to fall on the significant number of professional staff who either do not even attend the ministry or do no work while there. These staff undoubtedly represent more than ten percent of the total and are protected by political connections. In addition, there are apparently considerable numbers of phantom employees. The mission was unable to pursue this very far but it did observe that UP had an employee roster of 26 persons against a budgeLed 39. 3.13 There is no question but that too high a proportion of MARNDR's staff is located in Port-au-Prince. Except for those working within the integrated rural development projects, MARNDR has almost no functioning extension agents in a country that is 80 percent rural. This helps explain the poor returns to private as well as public agricultural investment. 3.14 3. Cumbersome Disbursement Procedures. The general disbursement problems that affect Haiti's public expenditure are discussed in Chapter II. They are aggravated for agriculture because of the weakness of the relevant MARNDR staff and the large number of donors involved in the sector, each with different procedures. Table 111.2 shows that actual Table II.2: PUBLIC AGRICULTURAL INVESTMENT EXPENDITURE, ACTUAL VERSUS BUDGETED, FY82-85 (percent) Source of Finance FY82 FY83 FY84 FY85 Internal and PL-480 67 85 86 84 External 30 65 47 Total 39 70 56 Source: Statistical Appendix Table 3.6.  -32- internally-financed expenditures have averaged 85 percent of those budgeted in the last several years; this is an improvement on the recent past and reflects the greater priority agriculture now receives. Donor-financed expenditures have varied much more, averaging 47 percent. Combined agricultural investment expenditures from all sources have therefore averaged only 55 percent of those budgeted. 3.15 The apparently lower disbursement rate of donor-financed expenditures reflects several very different factors. First, agriculture is no exception to the general pattern of the five-year and annual plans' consistent over-estimation of likely donor expenditures. Second, within agriculture, there have been overly optimistic projections of available donor support for high priority but unfinanced projects. Third, some donors have reduced disbursements pari passu with the Government's own counterpart shortfall. Fourth, the Government has frequently failed to meet conditions that have to be fulfilled before donors can disburse. Fifth, donor procedures for recruitment, project approval, commitment and disbursement of funds all vary and put a particularly heavy burden on the agricultural sector because it involves so many donors but has a relatively weak public staff. 3.16 Rapid Turnover of High-Level Personnel. On average, MARNDR has had a new Minister every year over the last decade. Other key personnel also turn over rapidly. This makes managers unwilling to take action against non-performing staff because the necessary political support could evaporate overnight. Frequent turnover has also affected program continuity: no Minister has had a chance to develop and implement a strategy or even to learn from experience. C. CD Ple teDM Policies and Organization 3.17 Agricultural projects can be grouped into seven principal program areas: integrated rural development, irrigation, credit, applied research, conservation of natural resources, crop production and extension, and livestock. This section discusses the complementary policies and organization needed but generally lacking for each. In addition, and of greatest relative importance, are the inappropriate price incentives that farmers face. 3.18 1. Integrated Rural Development. The largest share of the agricultural investment budget goes to externally-financed integrated rural development projects. Financing of such projects is adequate and their disbursement levels are the highest in the sector. Despite this, the projects are severely affected by the general implementation problems that afflict MARNDR; those which enjoy greater organizational autonomy are the least hampered. In addition, the recurrent expenditure implications of integrated rural development projects are very large as they provide types of services that were not readily available before the projects. Apart from charges for services provided, there are no obvious ways in which the projects can recover these recurrent costs, much less capital ones. They will have to be met, therefore, by increasing MARNDR's recurrent budget or by NGOs financed by either MARNDR or external donors.  -33- 3.19 The integrated rural development projects have not led to major increases in output, except in the Artibonite valley. This is partly due to the implementation and recurrent cost problems discussed above but also reflects inadequate levels of technology. More needs to be done to provide a properly integrated set o production services to farmers. Pilot applied research to identify suitable technologies is usually essential before full-scale projects are launched. There is considerable scope for sharing research results-and implementation experience--among projects, which currently tend to operate each on their own, and for improving the links between research and extension. 3.20 2. Irrigation. Irrigation investments are usually included within integrated rural development and rightly emphasize rehabilitation over new construction. Their inclusion in rural development, however, tends to deflect attention from irrigation's key role as the major investment necessary to increase production and productivity. In addition, projects have so far failed to address the interrelated reasons that irrigation systems are not maintained in the first place. Chief among these are insecurity of land tenure, the absence of water users' associations, inadequate operations and maintenance funding, and poor water utilization. 3.21 Haiti should enact and enforce an up-to-date system of land and water use laws and regulations, to include provisions for full cost pricing of irrigation services provided by MARNDR and retention of irrigation fees by the perimeter collecting them. This would force a more efficient allocation of water and land and increase agricultural production. Higher fees would also generate revenues essential to finance system operations and maintenance. The concept of water users associations, legal entities with the power to withold water to members who fail to pay fees, could be expanded. The associations could also be used as mechanisms for distributing inputs, delivering extension services and credit and conducting applied research. Other mechanisms may also be effective. 3.22 Once these problems have been resolved, irrigation projects could be concentrated on certain watersheds with joint planning among the Government and all involved donors of both irrigation and watershed management. 3.23 3. Credit. The two agencies extending agA cultural credit, the Bureau de Credit Agricole (BCA) and Banque Nationa.e de Developpement Agricole et Industriel (BNDAI), are in poor shape. BCA's operating costs are G 29 per G 100 it lends and its repayment rate is under 80 percent, with the result that it costs the Government and donors G 7.5 million per year to lend G 20 million. The position of BNDAI appears even worse. 3.24 This resource drain is principally the result of the agencies' inability to dismiss non-performing staff and of poor repayment rates by public sector employees and others with political connections. The Government must fire poor employees and restore order to the administration of contract law as it applies to credit. The credit agencies must then vigorously prosecute those defaulting without justifiable reason. This is a much more important step than establishing a new credit institution which would likely be plagued with the same problems as those now in existence.  -34- 3.25 4. Applied Research. Applied research is essential in Haiti with its many micro-climates. Substantial variations in rainfall and altitude prevent the transfer of crop varieties among ecosystems and regions. Research is all the more Important if the integrated rural development projects are to make an impact on production. The current USAID-financed Agricultural Production Research Project Is working well, but needs strengthening to diffuse successful approaches more quickly and to reduce duplication In experiments. 3.26 5. Conservation of Natural Resources. Severe erosion on the hillsides is a major problem for hill farming; for agriculture on the plains, where irrigation systems silt up and often suffer severe damage during heavy rains; and for hydroelectric power, where the useful capacity of the reservoir behind the Peligre dam is declining (see Chapter V). 3.27 Despite the seriousness of erosion, MARNDR has not developed a replicable or cost-effective mechanism for implementing soil conservation projects. Promising pilot programs are being tried in various places but few, if any, are attractive to farmers unless the latter are subsidized or even paid directly to carry out conservation works. Urgently needed is a replicable strategy integrating soil conservation and reforestation into hill farming that provides direct economic benefit to farmers and is linked to watershed management. Such a strategy is clearly much easier to advocate than to design; it is, nonetheless, perhaps the top pre-investment priority for agriculture and probably for the entire economy if successful soil conservation investments are to be made in the 1990s. Current projects should experiment in order to try to develop such a strategy and should frequently exchange experience. 3.28 Further problems afflicting soil conservation projects are absentee ownership and poor management of State lands; both inhibit coordinated action. Tenants must have adequate security of tenure if incentives for improved land use are to be effective, if regulations to conserve the soil are to be enforced, and, of course, if production is not to continue to be lost as a result of the poor utilization of good land. Implementation would also be easier if MARNDR concentrated its soil conservation efforts on designated watersheds or water conservation districts in conjunction with irrigation rehabilitation (see 2 above). 3.29 6. Crop Production and Extension. The crop production projects provide extension services outside the integrated rural development project areas. These projects-which should properly be part of the recurrent budget-are not well funded because of disagreements among MEF, MARNDR ant: USAID (the external donor) on how they should operate. Extension services are hampered by limited operating funds, inadequate in-service training of agents, poor supervision and weak management. In addition to taking steps to overcome these problems, MARNDR cannot afford to ignore the contribution that NGOs could provide, and might consider offering training to NGO staff on a fee basis alongside its own agents; this would help cover the recurrent costs of its new middle level training institutions. 3.30 7. Livestock. MARNDR's IDB-financed pig restocking project has begun very slowly because of difficulties in providing crucially important  -35- veterinary support. CIDA and FAC Intend to finance similar projects using stock considered more appropriate for small farms. In addition, a USAID project implemented exclusively through NGOs has already led to a population of about 10,000 pigs. Beyond veterinary services, the pig restocking projects have no recurrent cost implications. Investm t Proam and Eianin Plan, VY86-69 3.31 Planning is not sufficiently advanced by either the Government or donors to present a comprehensive investment program for FY86-89. Candidate projects can, however. be identified from these sources: (a) MARNDR projects in the FY86 annual plan with a total project cost of G 1,360 million; (b) other ministries' agriculture projects in the FY86 annual plan; (c) unfinished projects in the FY85 annual plan with a total project cost of G 440 million, which were not listed in the FY86 annual plan but will probably be restarted during the next few years; and (d) donor proposals for future projects with total donor financing of about G 750 million but no estimate of Haitian counterpart requirements.2/ 3.32 These various sources indicate that the future project mix will be much like the present one, indeed most of the investment program will consist of completing ongoing projects. The program will continue to favor exports. It will emphasize rehabilitation and expansion of irrigation systems. It will continue to include investments in rural development, soil conservation and reforestation, pig repopulation, credit expansion, and distribution of seeds, fertilizers and other inputs. The next five-year plan will continue to emphasize food crop production and the spread of new technologies and will therefore imply more extension, extension education and close cooperation between the public sector and NGOs. It is clear that there is little prioritizing of projects and expenditures. 3.33 Judging from the new projects under consideration, donors will continue to have the same individual priorities. USAID intends to work closely with NGOs in agro-forestry and hillside farming/soil conservation. French aid will be for ongoing integrated rural development projects. IDB will continue its interest in the Artibonite valley. IDA will follow up its experimental forestry project with a larger tree crops project. The bulk 2/ See Statistical Appendix Tables 3.09-3.12.  -36- of all new external resources will go into irrigation within rural development projects. 3.34 SO numerous are agricultural projects and so poor is data collection by both the Government and most donors that it is not possible to determine how much has already been spent on projects under implementation, nor do estimates exist of expenditures necessary to complete these projects. No financing plan as such therefore exists, except for the FY86 budget (Table 111.3). Unfortunately, the data do not permit irrigation to be separated out from rural development. There is no reason to expect the shares to vary significantly for FY87-89. Table III.3: COMPOSITION OF AGRICULTURAL INVESTMENT BUDGET, FY86 (G million) Internal a/ Donors Total Share (X) Integrated Rural Developaent 15.3 74.4 89.7 34 (includes irrigation) Credit 13.7 16.5 30.2 12 Applied Research 3.5 7.7 11.2 4 Conservation of Natural Resources 25.8 16.7 42.5 16 Crop Production and Extension 8.5 22.5 31.0 12 Livestock 6.0 6.0 12.0 5 Other (mainly studies) 14.1 30.2 44.3 17 Total 86.8 173.9 260.7 100 a/ Includes G 43 million of U.S. PL-480 and similar donor aid. Source: Statistical Appendix Table 3.2. 3.35 The total budget for agriculture for FY86 appears to show a sharp rise in both the recurrent budget and the internally-financed investment budget, from a combined total of G 72 million in FY85 to G 102 million (Table III.4). In fact, it reflects an increase of 133 percent in USAID PL-480 funds, which the Governmer.t classifies as internal resources but should more properly be classified as externally-financed. Moreover, the apparent decline in the externally-financed share of the investment budget is largely due to a corresponding reduction in non-PL-480 assistance from USAID. The total agricultural budget is in fact lower in FY86 than in FY85, and agriculture's share of the combined recurrent and investment budgets for all sectors has not changed at all from the 13 percent level that it reached in FY85.  -37- Table III.4: FINANCING OF MARNDR BUDGET, FY85-86 (G million at current prices) FY85 FY86 Recurrent Budget: Treasury 33 38 Development Budget: Treasury 18 21 (Internal) U.S. PL-480 18 42 Other commodity aid 3 1 Development Budget: Donors 185 142 (External) Total Agriculture: 257 244 Treasury 51 59 Donors 206 185 Sources: MEF, Budgets, FY85 and FY86; MP, Annual Plans, FY85 and FY86. Reconendations 3.36 Key agricultural policy reforms to help ensure that investments are attractive and yield satisfactory returns include a framework of appropriate prices; the establishment of a codified and enforced system of land and water use laws, including irrigation charges to cover full costs; the proper administration of contract law as it relates to agricultural credit; and improvements in land tenure. Irrigation charges should help increase revenues, and provide a small part of the extra Treasury resources necessary for agriculture. Other revenue-raising measures within the agricultural sector include raising rents on State lands to market levels and taxing large landholdings. These policy and revenue measures will not be easy to carry out; they are, however, just as essential as the shifts in public expenditure recommended below. Without them, public investment -n agriculture will continue to yield low returns. 3.37 Implementation obstacles must be overcome. This implies managerial improvement at MARNDR and a major increase in its operating funds. The two must go together, however. 3.38 Managerial improvements at MARNDR should include improved work discipline and disbursement procedures, and a longer tenure for Ministers and other key personnel. MARNDR should promote a wider exchange of experience among projects, NGOs and donors. Much duplication at present arises from ignorance of what others are doing. Donors clearly have an obligation here to inform themselves and MARNDR of each others' activities  -38- and technical results. In addition, MARNDR must cooperate more with NGOs, as it has begun to do in the soil conservation and reforestation areas. Haiti's agricultural needs are so great that the Government cannot afford not to utilize the services of any organization willing to help, particularly when it does not represent a cost to the fisc. A receptive attitude by MARNDR, which seems now to be emerging, is very important if such cooperation is to be successful. 3.39 MARNDR's operating funds should be greatly increased, whether they are in the recurrent or the investment budget, as should local counterpart funding in order to take full advantage of investment funds available from donors. Very few new projects should be undertaken before FY90 in order to allow time to implement properly projects currently under execution and to adopt appropriate new policies for the agricultural sector. This would not reduce the flow of aid if disbursements were to increase. A more detailed assessment is needed of existing projects to eliminate those that are not effective. 3.40 Many of the projects in the FY86 agricultural investment program and those likely to be in that for FY87-89 fit well with sector priorities. There are three exceptions however. First, rural development projects remain too complicated with an inadequate focus on production. Second, there is a proposed project for a series of regional milk processing plants that do not appear to have an obvious justification. Third, soil conservation projects do not yet have an effective strategy and should be seen more as experiments to develop one than as a solution to Haiti's severe erosion problem. Major investments could then follow in the 1990s. 3.41 - There are too many agricultural projects and too little available project data to prepare a detailed recommended investment program and financing plan for FY86-89. Rather, the level of annual public expenditure in agriculture should be roughly as suggested in Table III.5, which also presents the FY86 budget for comparison. Total expenditure is recommended at G 960 million in constant FY86 prices, of which G 540 million (56 per- cent) would come from the Treasury and the remaining 44 percent, or G 420 million (US$84 million), from external aid donors. It is not possible to estimate the foreign exchange component of the recommended expenditure, but it is unlikely to be above the 44 percent to be financed by donors. These proposals amount to a major agricultural budget rationalization and may well require considerable technical assistance to implement. 3.42 Table 111.5 puts all recurrent expenditure in the recurrent budget; this is for logical convenience but need not be implemented in practice if, for instance, it proves easier for some donors to finance recurrent costs by having them labeled as investments. The recommended expenditure levels do involve a major increase In Treasury funds allocated to agriculture from G 59 to G 135 million a year; budgeted external aid would fall by over 40 percent, but the resulting G 105 million (US$21 million) per year should actually be disbursed and would therefore be well above actual aid levels for recent years which have averaged around G 70 million (US$14 million).  -39- Table 111.5: RECOMMENDED MARNDR EXPENDITURE AND FINANCING, FY86-89, AND ACTUAL BUDGET, FY86 (G million at FY86 prices) Recommended Annual Expenditure Budget FY86-89 FY86 Applications 240 244 Recurrent 85 38 (Personnel) (50) (35) Investment 155 206 (Personnel) (-) (22) Sources 240 244 Treasury 135 59 External Aid 105 185 Sources: Mission estimates; MARNDR. 3.43 Adequate data are not available to estimate how much of the G 420 (US$84 million) of external aid required for FY86-89 has already been committed. (For the purposes of the overall analysis in Chapter 1I, it is assumed at US$52 million, or the entire requirement for FY86 and 50 percent of that for FY87-89; this implies new commitments of US$32 million). 3,44 The recommended spending above also excludes that by ministries other than MARNDR. If this is assumed to continue at the levels of the past, then the above estimates should be increased by 25 percent to cover the entire sector. The Government should consolidate all agricultural spending under one ministry. It makes little sense, for example, to have the Ministry of Planning implement a rural development project as it does now. 3.45 The recommended recurrent expenditure of G 85 million per year should include at least G 35 million of operating funds. In addition, 15-20 percent of MARNDR's staff should be laid off-those who are not performing--and those remaining should be awarded selective salary increases on the basis of performance and willingness to be transferred from Port-au-Prince to the field. This should reduce the total salary bill from G 57 to G 50 million, or by about 14 percent. 3.46 The recommended investment level above is approximately G 50 million less than that budgeted for FY86. About half of this is purely accounting, representing the transfer of operating costs to the recurrent budget. The other half will have to come from cuts in some projects planned for FY86. Future detailed work is urgently needed here. Cuts could come  -40- from a reduction in credit project funding if the credit agencies' internal administration costs and loan losses can be reduced. Subject to future analysts, some rural development projects might also be eliminated. 3.47 For FY87-89, priority areas should be irrigation rehabilitation, applied research, Input supply, and conservation. Few new projects should be started. The principal exceptions would be continuations of existing ones, which are often labelled 'new' by donors but are not really so. Rural development, however, must concentrate on irrigation and not become encumbered with too many social components. In addition, two new projects are needed. First, there should be a private sector input supply project to increase the availability of inputs in rural areas through existing merchants. Second, it is essential to have an exteitsion education and diffusion project to provide in-service training for agents at all levels, ideally including those of NGOs. These two project ideas have not been costed but should be included within the overall sector expenditure proposed for FY87-89. Coupled with more effective operation of MARNDR services, these two projects should sharply increase the rate of return on agricultural investment.  -41- CRHAPTR IV PUBLIC INDUSTRIAL ENTERPRISES 4.01 Haitian industry has three subsectors: export assembly, private sector import substitution and the public industrial enterpri3es. As discussed in Chapter I and in Haiti: Policy Proposals for Growth, industrial expansion can come only from exporting, the promotion of which requires eliminating the anti-export bias of the current incentive system. Public expenditure affects the sector in two principal ways: through the public provision of infrastructure for exporters and through the industries owned by the state. I. Infrastructure Needs of Prilvate Exporters Infrastructure and Industrial Parks 4.02 Private exporters need regular electric power, transport facilities, telecommunications, and water supply. These are largely treated in Chapters V-IX which deal with individual infrastructure sectors. Particular current problems mentioned by exporters in Port-au-Prince are, in rough order of frequency, poor or unavailable telex services, unreliable power supplies and inadequate water supply. Investments to deal with the first two are already underway; those to provide water are being prepared but will take longer. 4.03 An important aspect of industry's infrastructure which is not treated in the individual infrastructure chapters is its spatial location. At present, the bulk of industrial activity is concentrated in the Port-au-Prince metropolitan area, where there is a public industrial park currently undergoing expansion (SONAPI), a private industrial park (SODECOSA) and plans for two further private parks. In addition, there are some ideas to establish an industrial free zone at or near the port of Port-au-Prince. Outside Port-au-Prince, there is a small private industrial park-an extension of SODECOSA-at Gonaives; an IDS-financed pre-feasibility study is also being prepared for two possible public industrial parks at Cap Haitien and Les Cayes and for a free zone in Port-au-Prince. 4.04 Government planning documents have for some years emphasized regional development and called for policies to encourage industry to locate outside Port-au-Prince. At present, there is, however, no evidence that private industry wishes to establish itself elsewhere. Port-au-Prince has a good port and airport and other necessary infrastructure. Moreover, the private sector is satisfactorily meeting the demand for industrial parks. In the future, decentralization may well be called for. First, however, growth must be re-established.  -42- 4.05 Beyond the current expansion of SONAPI, there is no need for further public sector involvement in industrial parks in FY86-89, especially outside the Port-au-Prince metropolitan area. Indeed, the Government may wish to consider divesting the profitable public enterprise SONAPI. Ideas for an industrial free zone need to be thought through, including whether it should be public or private, and necessary studies wust be conducted before any possible investments are made. The existing industrial parks are virtually free zones. Again, no public investments seem likely in FY86-89. II. Public Industrial Enterprises 4.06 There were in early 1986 five major and several minor public industrial enterprises. This section is confined to the five principal ones: (a) Entreprise Nationale des Oleagineux (ENAOL), a vegetable oil mill; (b) La Minoterie d'Haiti, a flour mill; (c) Ciment d'Haiti, a cement plant; (d) Usine Sucriere du Nord (USN), a dilapidated sugar mill acquired from the private sector; and (e) Usine Sucriere Nationale de Darbonne (USND), a new sugar mill. The first four were initially established as private or mixed companies which were eventually taken over by the State. USND was built directly by the State as a public investment. 4.07 Through 1985, each enjoyed a special place in the economy. ENAOL had a production monopoly in the crushing of vegetable oil seeds and a trading monopoly in both semi-refined oil and the export of the meal by-product. La Minoterie had a monopoly in wheat and flour imports and in the production of wheat flour. Ciment d'Haiti had a de facto monopoly in cement production and trade. The prices of all the enterprises' products were controlled by the Government. Until March 1985, they were exempt from taxes. Financia and Oprto al sets 4.08 Through 1985, all the enterprises charged prices for their domestically sold outpit considerably in excess of border prices. Despite this, their financial performance was mixed (Table IV.1). Both sugar mills  -43- and the vegetable oil mill incurred heavy losses, particularly when depreciation and financial charges are taken into account. The flour mill and cement factories were profitable; indeed, were it not for La Minoterie's very large profits, the other four enterprises as a whole would have shown an annual average loss of some G 38 million after depreciation and financial charges. Table IV. 1: PUBLIC INDUSTRIAL ENTERPRISES: SUMMARY ANNUAL AVERAGE FINANCIAL RESULTS, FY83-85 (G million at current priees) Enterprise Operating Operating i rsting Financial Overall Revenues Expenditures Profit Charges & Profit Depreciation ENAOL 148 180 19 25 (7) La Minoterie 294 228 67 3 64 Ciment d'Haiti III 82 29 22 7 USN a/ 18 20 (2) 4 (6) USND 20 27 (7) 25 (32) Total 641 536 105 79 26 a/ FY84 and 85 only. Source: Mission calculations, based on information from enterprises. 4.09 The profitability of two of these enterprises should not, however, be taken as an indication of efficiency. Rather, it reflects the fact that the controlled prices of flour and cement were sufficiently above border prices that La Minoterie and Ciment d'Haiti made profits despite their inefficiencies. Vegetable oils and sugar were also sold at substantially above world prices but this was insufficient to compensate financially for the inefficiencies of ENAOL, USND and USN. Table IV.2 indicates that production costs were well above the CIF prices of competing imports in each case. 4.10 The causes of the low level of efficiency vary from enterprise to enterprise. However, all shared gross overstaffing and high overhead costs. All the enterprises, whose total employment exceeds 1,800, could have operated with less than half their employees, in some cases with less than a third. At La Minoterie, ENAOL and Ciment d'Haiti, administration and other fixed costs were also substantially above those which more competitive operations could sustain. ENAOL and the two sugar mills suffered from low capacity utilisation, the former because of low domestic demand and the latter because of lack of raw material. Those enterprises which imported raw materials--ENAOL, La Minoterie and, to a lesser extent, Ciment d'Haiti-paid CIF prices significantly above those they should have, according to known world prices and transport costs. This was frequently due to excessive concentration on one source of supply, be it a country or a company. It may also have been related to poor payment records: both ENAOL and La Minoterie were in early 1986 in arrears to grain suppliers. In some  -44- ca3es, costs were high because of antiquated or poorly maintained equipment which led to frequent interruptions to operations, in turn increasing fuel consumption substantially because of constant restarting. Table IV.2: PUBLIC INDUSTRIAL ENTERPRISES: COMPARISON BETN UNIDfT PRODUCTION COSTS AND CIF IMPORT PRICES, FY85 Company Product Unit CIF Production Ratio Price Cost a/ ENAOL Semi-refined US$ per 732 1,520 2.1 soybean oil metric ton La Minoterie Superior US$ per 12 19 b 1.5 flour 100 lb bag Ciment d'Haiti Bagged cement US$ per 50 88 1.8 metric ton USN Raw sugar US cents 15 / 33 2.2 per lb USND Raw sugar US cents 15 cl 87 5.8 per lb a/ Including depreciation and financial charges, but excluding indirect taxes. b/ An audit of La Minoterie being carried out in July-August 1986 may throw more light on its production cost; the figure here is the mission's best estimate on the basis of information available to it. c/ Long-term projected world price. Source: Mission estimates. Debt 4.11 Several enterprises were heavily in debt (especially ENAOL and the sugar mills), the service on which added considerably to their costs, especially as the bulk of the debt was short term. Further, as noted, arrears represented a serious problem for at least two enterprises. ENAOL's relatively expensive short term obligations were increased while its long term loans were paid off. ENAOL's outstanding debt of US$17.4 million as of September 30, 1985 had been reduced by only US$0.4 million by December 31. Just over half was due to the Junta Nacional de Granos, an Argentine government soybean marketing organization, and was over six months in arrears. 4.12 An audit now being made of La Minoterie may clarify its debt situation. As best the mission could determine, and excluding liabilities to the central bank (BRH) and other local banks, La Minoterie had an outstanding total debt of US$15 million on September 30, 1985; preliminary Information indicates it had been reduced to about US$11 million by December  -45- 31. Of this, US$7 million was due to the Government of Canada, which took over a debt due to a Canadian company, Maple Leaf, which previously had a management contract with La Minoterie; some US$2 million was over four months in arrears. 4.13 USND had an outstanding debt of US$46 million as of December 31, 1985, practically all related to the construction of the sugar mill at Darbonne. Of this total, US$38 million was owed to a consortium of European banks and was guaranteed by the Italian Government (the mill equipment was largely supplied by an Italian firm). USN's total debt of about US$6 million was owed in its entirety to the Banque Nationale de Credit, a Haitian public sector commercial-bank. 4.14 Ciment d'Haiti has paid off the greater part of its outstanding debt, both short term obligations to commercial banks and a long term loan from the previous foreign shareholder. Balance of aymnts lpact 4,15 One of the principal stated justifications for the establishment, expansion or continuation of the public industrial enterprises is that they save foreign exchange by producing goods which the country would otherwise have to import. In fact, partly because of their high costs and inefficiencies, the three import-substituting enterprises--ENAOL, La Minoterie and Ciment d'Haiti--were heavy net users of foreign exchange. Their net foreign exchange loss was about US$30 million per year, or 9 percent of Haiti's total exports of goods and services. 4.16 Since commencing operations in FY82, ENAOL spent about US$140 million in foreign exchange through FY85 for imported raw materials, spare parts, fuel, technical assistance and for debt servicing. This was equivalent to US$34 million per year, or about 11 percent of exports. Direct foreign exchange earnings from ENAOL's exports of soybean meal averaged US$11 million per year and the company's crushing operations produced crude soybean oil which would have had an avarage CIF cost of some US$8 million per year. ENAOL, therefore, had a negative balance of payments impact of about US$16 million annually, equivalent to the measures needed to create perhaps 6,000 new jobs each year. 4.17 La Minoterie's balance of payments impact was also heavily negative. Direct foreign exchange requirements were about US$45 million per year, or 14 percent of exports. About US$2 million was earned from the export of wheat chaff for animal feed, and about US$31 million worth (at CIF prices) of flour and chaff was produced for the local market. The net foreign exchange cost of the mill's operations was therefore about US$12 million per year, equivalent to the creation of perhaps 4,500 jobs each year. 4.18 Even Ciment d'Haiti, which uses local raw materials, was a net drain on foreign exchange resources, principally because of its very high fuel consumption. The company's total foreign exchange cost of producing a metric ton of bagged cement was about US$52, whereas bagged cement could  -46- likaly have been imported for about US$50 per metric ton. This implies a net annual foreign exchange loss of about US$0.5 million. 4.19 The public industrial enterprises had a very substantial net negative impact on Haiti's economy. Table IV.3 attempts to quantify this cost from FY82-85 by adding together outlays for investment and the real resource impact of the enterprises' current operations. The investments Include the capital cost of works directly financed by Haiti, either from internally generated resources or by borrowing, plus the cost to Haiti of compensating previous-foreign shareholderso As an approximation of the real resource cost of current operations, the consumer subsidy has been measured as the difference between the enterprises' costs of production and the amount which Haitian consumers would have paid for the CIF cost of imported substitutes, had the enterprises not existed. Since the economic rate of return on all the enterprises is negative, the investment cost is also a total loss. It can be seen that the total economic cost of the public industrial enterprises in FY82-85 was of the order of G 1.5 billion (US$300 million), nearly four percent of total GDP over the four-year period. Had the investment been made elsewhere with a return of 12 percent, industrial GDP would be G 380 million greater in FY90, rather than continuing to lose almost G 400 million per year in economic terms. Table IV.3: ECONOMIC IMPACT OF PUBLIC INDUSTRIAL ENTERPRISES, FY82-85 (G million in current prices) Enterprise Investment Consumer Subsidy Total ZNAOL 91 340 431 La Minoterie 25 363 388 Ciment d'Haiti 70 189 259 USN 33 57 89 USND 289 92 381 Total 506 1040 1546 Source: Mission calculations. Recent Government Actions: EEAOL and USND 4.20 Recognizing the problems at ENAOL and USND, the Government has since April 1986 taken several very significant actions, which, if sustained will greatly reduce the fiscal and economic losses generated by the public industrial enterprises. First, it freed previously restricted imports of semi-refined and refined vegetable oil, subjecting them to a liberally interpreted quota and to a 20 percent ad valorem tariff. Retail prices have as a result dropped dramatically from $1.60 to $1.10 per liter. Second, the Government closed the ENAOL plant on July 31, assuming its debt and offering generous severance pay to its 200 workers but also offering them the option of running the factory as a cooperative. By mid-August, about 100 workers had elected severance pay. Third, the Government has closed the  -47- USND sugar plant, is offering severance pay and compensation to affected mill workers and cane growers, and will provide infrastructure and technical assistance to help the latter attempt to convert from sugar cane to other crops. The physical disposition of the assets of both ENAOL and USND remains to be determined; at present, they remain in the hands of the State. Optioln: La Nlnoterie, Ciment dqfilti and USE 4.21 To inform policy decisions to prevent or minimize further losses, the mission carried out an economic analysis of future options for the three remaining enterprises. Each analysis centers on the option of continuing with existing operations or expanding them after undertaking the requisite investments. Extremely favorable assumptions are made from the viewpoint of continued or expanded operations. All investments and losses incurred through FY85 are assumed written off. It is assumed that international levels of efficiency would be obtained with modest investments at La Minoterie, Ciment d'Haiti and USN. At USN, it is also assumed that supplies of sugar cane increase dramatically to enable the mill to achieve a more reasonable capacity utilization. All inputs and outputs are valued at long term international prices. The analysis of La Minoterie may have to be refined once detailed audit results are available, but these are unlikely to change the conclusions significantly. 4.22 Specific results for each enterprise are as follows: (a) La Minoterie. Physical yields cannot be improved significantly above their current unsatisfactory levels without investing in a new flour mill; equipment at the existing mill is obsolete. At international prices, however, a new mill would have a heavily negative net present value, even assuming considerable improvements in operating efficiency. (b) Ciment d'Haiti. Continued production of clinker is uneconomic; switching to the crushing and bagging of imported clinker would be marginally justified, under favorable assumptions and a discount rate of 12 percent. (c) USN. Even under highly favorable assumptions concerning cane supply, the international value of sugar, and efficiency improvements, continued operation would result in negative net present values at the mill and is therefore economically unjustified. Iecoe_dations 4.23 ENAOL should remain closed and its physical assets should be disposed of. The annual public investments of about G 5 million designed to stimulate local production of oilseed substitutes for imported soybeans to supply ENAOL should cease. The Haitian vegetable oil market should remain open. ENAOL's debt, now assumed by the State, should be renegotiated.  -48- 4.24 Similarly, USND should remain closed and its physical assets should also be disposed of. Its debt, also assumed by the State, should be renegotiated. 4.25 So as to assure competitiveness under any kind of ownership, no special privileges should be made available to cooperatives or other private groups that may operate either of the former ENAOL and USND plants. This includes the granting of monopoly status, restrictions on imports, and public funds, credit and debt guarantees. 4.26 La Minoterie's monopoly on imported wheat and wheat flour should cease and imports of all food grains and of flour should be freely permitted with tariffs to protect local agricultural production and provide fiscal revenues. Price controls should be removed. La Minoterie should respond to this by becoming more efficient, through cost-cutting measures. The State should aot invest in a new flour mill under current market conditions. Investments in improving the efficiency of the existing mill may be necessary to enable it o confront increased competition, but no further investments should precede the market reforms. Importing flour rather than wheat may require additional investments in storage facilities. This possibility should be carefully analyzed together with the organizational improvements required for successful flour storage and distribution. 4.27 Ciment d'Haiti should cease clinker production and crush and bag imported clinker. The cement market should be liberated and the company's trade and production monopolies lifted. It would seem that new power generating equipment is needed, but its size should be determined by the requirement to crush clinker. Additional investments which may be required in port and storage facilities for imported clinker should be identified and subjected to a feasibility study. 4.28 On economic grounds, the USN sugar mill should be closed. It is economically unviable even if the opportunity cost of sugar cane is reduced to the G 30-40 (US$6-8) per ton paid by the guildives (small distilleries producing potable alcohol) rather than the G 65 (US$13) per ton paid by the centrifugal mills (reflecting the argument that the land and labor dedicated to cane production in Haiti has no more profitable alternative use). On social grounds, h>,wever, this recomendation would need to be implemented in the context of the identification of alternatives to cane production. 4.29 Very few investments will therefore be necessary for the public industrial enterprises in FY86-89. The only two which may need to invest are La Minoterie and Ciment d'Haiti. A rough investment program might look like that presented in Table IV.4. It must be emphasized, however, that these numbers are purely illustrative. 4.30 Given recent experience with public industrial enterprises in Haiti, there is no reasor, for the Government to establish or take over any new enterprises. In particular, should gold or lignite mining prove economically feasible, they should be left to the private sector to develop (Chapter V also discusses lignite briefly).  -49- Table IV.4: PUBLIC INDUSTRIAL ENTERPRISES: ILLUSTRATIVE INVESTMENT PROGRAM, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 La Minoterie 5 35 40 Ciment d'Haiti 20 30 50 USN _ Total 25 65 90 Source: Statistical Appendix Table 4.1. 4.31 In addition to La Minoterie and Cinent d'Haiti, there are at least five small public industrial enterprises (Flore d'Haiti, Industrie Marbriere Haitienne, Port Dauphin, Produits Metalliques, and Societe de Nutrition Animale) and perhaps more, several owned via the State holding company Societe d'Equipment Nationale (SEN), that the Government may wish to consider divesting.  -50. CHAPTR v ELECSTIC ?OSBR TV ame=Y Sector Context 5.01 The mission did not examine public spending on energy other than electricity in any detail, principally because no major investments are planned for FY86-89. Before turning to electric power, however, it is useful to set the subsector in its full sectoral context and also to make a few recommendations abotlt sector planning that have implications for investments in the 1990s*./ 5.02 In calorific terms, wood is Haiti's most heavily used fuel. It accounted for 74 percent of total energy supply in 1982, the rest of which came from imported petroleum products (14 percent), hydropower (5 percent), bagasse (5 percent) and imported coal (2 percent). Indigenous energy resources are limited. Forest resources are being depleted for use in household cooking, commerce and industry. This has led to serious erosion problems which pose a grave threat to agricultural productivity. In addition, erosion-induced siltation impedes irrigation and has reduce. the useful volume of Haiti's main hydroelectric scheme, Peligre. Undeveloped hydropower potential is only 90 MW, compared to the current installed capacity of 173 MW. No petroleum deposits have been identified, although there is hydrocarbon potential. Under present conditions, it is not economic to mine the country's largest lignite deposit (Maissade) for power generation. 5.03 Unsatisfied energy demand may contain Haiti's economic growth. In particular, a reliable supply of power is essential to the growth of the export assembly industry. Key objectives for the energy sector as a whole are to meet overall demand while balancing wood supply and demand on a sustainable and ecologically iound basis, minimizing the oil import bill, and developing the power system on a least-cost basis. Achieving these goals will require: increasing the economic efficiency of energy use, for which there is considerable scope; developing indigenous energy supplies to the extent consistent with efficient resource allocation; ensuring that oil is imported on a least-cost basis; and analyzing the possibilities of importing more economic substitute fuels. 5.04 Three institutions have responsibility for energy. The Ministry of Agriculture, Natural Resources and Rural Development (MARNDR) deals with forestry and agriculture (see Chapter 3); Electricite d'Haiti with electricity (discussed below); and the Ministry of Mines and Energy Resources (MNRE) with substitute fuels (including indigenous fossil fuels) and energy efficiency. In addition, the Ministry of Commerce, 11 A more complete analysis of Haiti's energy situation is contained in UNDP/World Bank reports Haiti: Issues and Options in the Energy Sector (3672-HA, June 1982) and Haiti: Energy Assessment Status Report (041/85, August 1985).  -51- which oversees the petroleum supply industry, is addressing the issue of the least-cost procurement of oil. The Ministry of Planning plays no effective role in the sector. Given such a fragmentation of responsibilities, coordination is vital to ensuring a coherent and balanced approach to the sector as a whole. Efforts in this direction need to be vigorously pursued. 5.05 Within the MMRE, there is scope for a small unit such as the existing Directorate of Energy Resources (DRE) to strengthen its role in: (a) advising the Government on overall sector planning, and (b) exploring energy options (i.e., fuel substitution and increased efficiency) not covered by other institutions. The DRE's pre-investment work in the next several years will have implications for investments in the 1990s. Consequently, it should focus its work program so as to play a catalytic role in channelling investments into priority areas. 5.06 The DRE is currently placing a great deal of emphasis on investigating the exploitation of Maissade lignite as a woodfuels substitute. Consideration should be given to developing this deposit only once it has been established that producing a suitable fuel from it is technically and economically feasible. Over the next few years, the DRE should renew its emphasis on: (a) building up energy balance demand data; (b) promoting energy demand management among the major industrial consumerd of oil, power and wood; (c) targeting the guildives (small alcohol distillers) to reduce their wood consumption and perhaps replace it with bagasse; (d) establishing a self-sustaining system for producing and marketing charcoal stoves in Port-au-Prince; and (e) carrying out in conjunction with MARNDR the proposed regional analyses of the woodfuels situation. 5.07 With respect to petroleum exploration, given its high risk and the shortage of investment funds, the Government should ensure that its costs are borne by international oil companies. This is being dealt with under a MMRE/IDB program to formulate a strategy to promote oil exploration. The Power Sector and its nlanniDg 5.08 EdH has 173 MW of installed capacity; there is also an estimated 40 MW of captive plant. As of October 1985, EdH had 160 MW to serve the Port-au-Prince area's demand of 68 MW. Despite apparently large reserves, EdH has had difficulty meeting dry season demand in the last two years. Low rainfall and the effects of siltation have limited the availability of Peligre's 47 Kt; 5.09 There no detailed records of captive plant capacity. The largest plants are those of Ciment d'Haiti (14.5 MW), the Minoterie (5.6 MW), and the four sugar mills (USND 5.2 MW, HASCO 2.5 MW, USN and Dessalines); the balance belong to isolated users and to EdH customers taking precautions against outages. 5.10 The main constraint to expansion is the high investment cost of hydro schemes. Variations in the regime of the rivers make reservoirs necessary; large reservoirs are very costly because of the terrain and the lack of roads.  -52- 5.11 EdH is the sole entity responsible for power supply. The Minister of Public Works, Transport and Communications chairs the Board of Directors on which sit the Ministers of Finance, Commerce and Industry, and Platming, the Governor of the central bank, and EdH's General Manager. Policy directives are set by the President; the Board approves major investment plans, tariffs, borrowings and broad policies. The General Manager is appointed by the President and has reasonable autonomy. 5.12 EdiH's current organization, based on a 1981 study by foreign consultants, is adequate. Improvements in technical departments have been more modest than those in administration, partly because of staffing constraints. Indeed, EdR's dependence on foreign consultants will have to continue several years until a number of young Haitian professionals have acquired enough training and experience to take over technical and managerial responsibility. Despite manpower limitations, EdH's project implementation capability is reasonably good. Financial management is sound; EdH has maintained an average return of 6 percent on revalued assets except under adverse weather and technical conditions in FY83. 5.13 Power sector planning is based on a system expansion program prepared by EdH, with the assistance of foreign consultants, following the 'Long-Term Pre-Investment Study of the Power Sector", drawn up in 1978 and periodically updated since. This study, which emphasized the expansion of generation in Port-au-Prince and other major urban areas, has been endorsed by the Government and the main external aid agencies supporting the power sector. The latest expansion program was reviewed in May 1985 by an independent Board of Consultants. The aid agencies involved in the sector (CCCE, CIDA, IDA, IDB, KfW, UNDP and USAID) will submit their views to the Technical Coordination Commission during 1986. This commission was established in 1980 and has since sa-Asfactorily coordinated technical and financial aid to the power sector. 5.14 The prospective growth of EdH will require strengthening the budgetary control of projects under execution. Foreign consultants are assisting EdH address this and other management issues. Electricity Demand 5.15 Electrification is limited. In 1985 only 10 percent of the population was connected to the system, mainly in Port-au-Prince where 45 percent of the population had electric service compared to only 3 percent elsewhere. Industry accounts for 45 percent of EdH's sales, followed by residential customers (38 percent), public sector institutions (12 percent) and commerce (5 percent). Total sales in FY85 increased by 6 percent on FY84, slightly below the FY80-85 average of 6.3 percent. Port-au-Prince, with 87 percent of sales, continued to dominate the market.  -53- 5.16 The growth of private sector demand in Port-au-Prince has followed that of the economy, averaging 16 percent per year in FY75-80 and 5 percent since, consistent with the economic stagnation that has gripped Haiti since FY80. Public demand, by contrast, grew at 6 percent per year in FY75-80 but has since increased to a staggering 14 percent in FY80-85. This distressing growth in total public consumption has occurred despite a 6.4 percent decline in the street lighting component even though coverage has been expanded-achieved through technical measures such as wider spacing and the use of high efficiency lamps; other public consumption has risen by 23 percent. In addition, Government payments for electricity are 6-12 months in arrears. 5.17 As is characteristic of most developing electricity markets, the rate of consumption in Haiti is led by the rate of connections. Active users in Port-au-Prince increased by 6.2 percent per annum between FY80 and FY85 while consumption per user rose by only 0.2 percent per annum. Consumption grew more in the provinces, reflecting investments in distribution and the conversion of unnaccounted-for consumption into sales. 5.18 At about 27 percent of generation, losses are still a major problem in Port-au-Prince, and are split evenly between technical losses and stealing. They represent about G 18 million per year. A loss reduction program is under way to renew about 75 percent of the distribution network by FY87 and reduce technical losses and theft to about 17 percent of gross generation by FY89. Once this targst is achieved, EdH will have to continue its efforts to reduce lossas even further to more acceptable levels. Theft has increased very substantially in the first quarter of 1986. Losses are no longer a problem in the provinces, where network renovation and legal enforcement resulted in their reduction from 26 percent of gross generation in FY78 to 15 percent in FY84. 5.19 Two electricity consumption scenarios were assumed for evaluation of Ed.'s investment programs for FY86-95, reflecting uncertainty about the growth prospects of the economy in general and industry ii particular. The long gestation periods of power projects mean that the evaluation period must necessarily be longer than the investment horizon. The minimum scenario assumes that total sales will increase at 6 percent per annum and the high scenario assumes 8 percent; the latter is consistent with the rate assumed by EdH for expansion planning purposes. It would require only that part of the reduction in losses in Port-au-Prince be added to the minimum scenario. 5.20 Under either scenario, sales growth in the provinces is assumed to slow down significantly, as the expansion effort in Cap Haitien, Gonaives and Les Cayes during FY80-85 is now complete and further growth is expected to continue at the rate of normal network development. The minimum scenario is based on already moderate plans for expanding the distribution network and consumption growth below 6 percent per annum is unlikely. However, growth above 8 percent per annum eannot be ruled out unless public consumption is brought under tight control.  -54- Tariffs and Operating Costs 5.21 The tariff structure is based on the underlying structure of marginal supply costs for different user categories, in line with the recommendations of a 1983 review. Tariffs are adjusted every year to reflect the system's operating conditions. 5.22 ahe average tariff per kWh reached US$0.146 in FY85, one of the highest in the western hemisphere. In FY85 the rate of return on fixed assets was estimated at 5.3 percent and EdH was able to meet 39 percent of its capital requirements from internal cash generation. Operating costs have, however, risen as surplus staff have been added, the total rising from 1,130 in FY82 to 1,430 in FY85. The number of customers per employee has dropped from 65 to 60 and compares very unfavorably with other countries in the region, e.g., Brazil (146) and Ecuador (90). 5.23 If the loss reduction target (para 5.18) Is achieved, average electricity tariffs could be reduced about 6 percent from current levels. Expansion Plan, FY86-95 5.24 Following the 1983 system expansion study and its 1985 review by an independent Board of Consultants, EdH has updated the generation expansion program for FY86-95. In contrast to the Board of Consultants' projection of an annual energy sales growth rate of 10.9 percent in Port-au-Prince alone, EdH assumes growth of about 8 percent per year for the whole country, in line with the assumption that part of the reductions in system losses will be converted to sales. 5.25 For Port-au-Prince, the FY86-95 plan includes the development of the Artibonite 4C hydro project for commissioning in FY92, and the first investments on a coal fired steam power plant for operation in FY96, with initial and final capacities of about 35 and 210 MW respectively. These two developments will serve Port-au-Prince for the next 20 years. Development of the remaining hydro project (La Chapelle, with a capacity of about 60 MW) is still doubtful in view of its very high cost (US$4,000 per KW compared to US$3,000 for Artibonite 4C) and the social problems associated with the displacement of 6,000 people and the loss of 4,000 ha of agricultural land. Both Artibonite 4C and the steam plant are included in the Board of Consultants least cost expansion program. EdH's revised generation expansion program excludes the installation of additional diesel-electric power plants in Port-au-Prince (20 MW in FY86-95) recemmended-by previous programs. However, if energy sales grow faster than assumed, or if Artibonite 4C is not ready in FY92, new diesel-electric capacity will be required. 5.26 For the provinces, dBR plans to continue moderate expansion, based on the construction of new small hydro projects (an increase of 8 MW by FY95) and diesel-electric plants (an increase of 18 MW). 5.27 The siltation of the Peligre dam is advancing rapidly; by the second decade of next century, no regulation of the Artibonite river for power generation, irrigation and flood control will be possible. The  -55- Board of Consultants recommended raising the height of the Peligre dam as a posesible way to increase the life of the reservoir behind, but EdH's expansion program does not include any funds for this. The CCCE is currently financing a study of what would be needed. 5.28 Nonetheless, EdH's overall expansion plan is reasonable. (Statistical Appendix Table 5.1 summarizes the system in 19X5 once the program is completed). The construction of Artibonite 4C would allow EdH to operate the upstream Peligre plant more flexibly without detrimeLt to Irrigation water requirements. During the dry season, water regulation by Artibonite 4C would allow Peligre to meet up to 35 MV of peak demand compared to its current maximm of 15 MW. The one weakness in the plan is the exclusion of measures to arrest the siltation of the Peligre dam; these cannot be costed until the above study is completed. Iyes~t Profram ad Lianclmg lan. T8689 5.29 Table V.1 summarizes EdR's investment program for FY86-89. The program shows an adequate balance among investments for generation and distribution and between Port-au-Prince and the provinces. Planned Investments in the provinces are modest but enough to meet a power macket growth of 8 percent per year, which would increase from 3 to 6 percent the proportion of the population outside Port-au-Prince with access to electricity. CIDA is financing a Master Rural Electrification Study seheduled to start in 1986. At present, EdH plans only minor investments in rural areas, including line extension near small towns and small hydro plants. Table V.1: EdH: INVESTMENT PROGRAM, FY86-89 (G million at FY86 prices) Port-au-Prince 504 Generation and Transmission: Artibonite 4C (partial) m Distribution and General Plant 283 Provinces 72 Generation and Transmission 'T Distribution and General Plant 21 Total Direct Investment 576 Interest during Construction 42 Working Capital 26 Total Investment aud Working Capital 644 Sources: EdE; Statistical Appendix Table 5.3. 5.30 EdE's FY86-89 investment program, including interest during construction and working capital requirements, amounts to G 644 million in constant FY86 prices, of which 89 percent corresponds to direct investment. Ongoing works represent G 278 million (43 percent) and  -56- include: (a) renovation of the distribution system, transmission lines and substations in Port-au-Prince; (b) construction of a new head office; (c) Artibonite 4C design; and (d) routine investments. Future works represent G 298 million (46 percent) and include (a) partial construction of Artibonite 4C and (b) distribution and general investments in Port-au-Prince. The foreign component is estimated at G 464 million (US$93 million) in FY86 prices, or 75 percent of the total. 5.31 EJH's net lnterral cash generation would finance G 314 million of the program (49 percent of the total); the Government and consumer contAibutions would fund G 23 million (3 percent); and the remaining G 307 million wouLd be covered by foreign aid, of which G 165 million (26 percent) corresponds to existing loans and grants and G 142 million (22 percent) to future ones. Past experience indicates that EdH has good prospects of securing the incremental foreign currency financing. Table V.2: EdH: FINANCING PLAN, FY86-89 (G million at PY86 prices) FY86 FY87-89 FY86-89 Applications 192 452 644 Investment MT WI7T Variation in Working Capital -9 35 26 Sources 192 452 644 EdH Internal Cash Generation 31T 252 TI Borrowing 124 183 307 Existing IDA Loans (73) (48) (122) Other Existing Loans (44) - (44) Future Loans (7) (135) (142) Consumer Contributions 2 6 8 Government 4 11 15 Source: Statistical Appendix Table 5.3. 5.32 In addition to meeting the local costs of the program and the rate of return levels agreed with foreign donors, EdH could meet about one third of the foreign costs of the program, provided it has access to foreign exchange. However, limitations imposed by the central bank on foreign exchange purchases of spare parts and even fuel have affected its operating efficiency in the past. lecomundations 5.33 Despite staffing limitations and other constraints, EdH is satisfactorily managing the growth of Raitils power sector. The company has demonstrated good project implementation capacity and the ability for sound financial management, but must improve budgetary control of projects under execution. EdH must aggressively reduce losses and pare its excess staff, however.  -57- 5.34 The Government should (a) curb the growth of electricity consumption by public sector institutions; (b) settle permanently the payment of public sector arrears and keep all its electricity bills current; and (c) lift foreign exchange restrictions on EdH. 5.35 The least cost expansion program was reviewed by World Bank staff in July 1986, following the drop in world petroleum prices, and Artibonite 4C remains preferable to a thermal alternative. The FY86-89 investment program is reasonable and well balanced, both in light of its planned construction undertakings as well as its financial viability. Indeed total direct investment at G 576 million in FY86 prices over four years will be slightly less than the G 592 million in FY86 prices invested during the preceding four. 5.36 The one weakness in the investment program, as in the expansion plan, is the neglect of the siltation of the Peligre dam which poses an increasing threat to the system's reliability and may derate the country's hydro potential unless it is addressed in a timely fashion. There is a sense in which this is a much broader problem than one for the power sector; interministerial attention needs to be paid to the entire question of erosion, siltation and watershed management. id! cannot artord, however, to assume that massive, long-gestating and hard-to-finance watershed management projects will provide it with a solution. The CCCE-financed feasibility study of raising the height of the dam is thus essential and action must be taken once it is completed. Raising the dam would flood a small area of the Dominican Republic and so discussions should begin with Haiti's neighbor to assess its political feasibility. In addition, and for the much longer term, both countries might gain from interconnecting their power systems. 5.37 - New foreign aid loans and grants of G 142 million (US$28 million) are needed for FY86-89; a higher commitment will probably be necessary to finance the entire cost of the Artibonite 4C project, construction of which will continue through FY91. 5.38 Some donors have recently been asked to consider the possibility of financing a US$50 million interconnection cf Cap Haitien, Gonaives and St. Marc to the Port-au-Prince system, for operation in 1990, to help absorb important captive loads along the coast. This large investment is now scheduled for the late 1990s in the system expansion program and should not proceed until then.  -58- CHAT VI- TRANSPOIT Mme Sector and its Plnnming 6.01 The blueprint for investient in transport is the National Transport Study (NTS), carried out by foreign consultants in 1975-77, which recommended as priorities: (a) improved road maintenance; (b) selected upgrading and rehabilitation of primary and secondary roads and drainage structures; and (c) rehabilitation of selected coastal shipping ports, to permit deferring costly road development. 6.02 These priorities have been closely reflected in subsequent Investments. Significant achievements include: (a) strengthening SEPPRN, the national road maintenance agency, to cover acceptable maintenance of some 2,655 km compared to 1,450 km in 1979; (b) reconstruction and rehabilitation of about 473 km of the paved road system, including Port-au-Prince to Cap Haitien (250 km), Port-au-Prince to Lee Cayes (180 kIm) and Grand-Goave to Jacmel (43 km); (c) construction or improvement of about 1,600 km of secondary and tertiary roads; Cd) reconstruction or rehabilitation of some 70 bridges and culverts; (e) improvement and reconstruction of the two international ports at Port-au-Prince anl Cap Haitien and rehabilitation of ten coastal shipping ports; and (f) resurfacing and extending the runway of the Port-au-Prince international airport. These accomplishments have largely met the principal objectives of the NTS. 6.03 The returns on most of the roads appear good and improved maintenance is assuring their sustainability. The Cap Haitien port b'.s not, however, yielded the-expected benefit of actively serving its hinterland and tourist shlpe; traffic has averaged only two ships per week. Given its relatively short distance from the more developed port at Port-au-Prince, the commercial and government center, it is unlikely that the Cap Haitien port will ever assume the significant role once projected for it. In retrospect, the investment cannot be justified economically. Investments in coastal shipping ports have not fared much better; unsuitable pricing and the weakness of the National Maritime Services Agency (SEMANAH) have not encouraged their full use.  -59- 6.04 The Ministry of Public Works, Transport and Communications (MTPTC), through its Department of Transport (DDT), is responsible for overall sector management, including policy formulation and investment planning and coordination. DDT is also responsible for the road subsector. MTPTC also has jurisdiction over the National Airport Authority, the National Civil Aviation Office and SEMANAR. MEF is responsible for the National Port Authority (APN) and influences transport investments through its control of annual budgetary allocations for both capital and recurrent expenditures. As for other sectors, the Ministry of Planning (MP) is theoretically responsible for defining the national development strategy and for reviewing, consolidating and coordinating the investment proposals of the various ministries and agencies. 6.05 In this institutional framework, transport investment planning in the public sector is expected to start with MP's definition of the national development strategy. DDT should then derive from it a transport investment strategy from which the various modal agencies should develop individual investment proposals for subsequent coordination and consolidation by DDT into a national transport plan. The latter, in-turn, should be reviewed by MP for consistency with national objectives and for financial feesibility within national plan financial ceilings set by MEF. The entire process should be interactive both among DDT and the modal agencies and among DDT, MP and MEF. 6.06 In practice, however, MP has not provided sufficient leadership. Transport planning is therefore reduced to a compilation of the perceived needs of each modal agency. Also, DDT has so far made no serious attempt to ensure the screening and coordination of these agencies' proposals into a coherent plan. Moreover, DDT does not systematically review and update the transport plan each year to reflect changing needs and funds availability and as a basis for budget requests. To some extent, this results fron DDT's lack of authority in actual practice over the financially autonomous implementing agencies. It also stems from the shortage of experienced transport planners and economists in DDT. 6.07 Donors active in the sector include CCCE (airports), EEC (roads), FAC (roads), IDA (roads, ports), IDB (roads, ports), KfW (ports) and USAID (roads); OPEC is considering an involvement in roads through cofinancing with IDB. 6.08 There are no public sector carriers, except for CONATRA, the Port-au-Prince bus company (Chapter VIII). Recurrent Spending 6.09 No accurate statistics on past recurrent spending on transport are available. Information from MEF, which, however, differs from that from other sources, indicates that some G 285 million were spent in FY83-85, slightly more than the G 275 million on capital investments. 6.10 Annual recurrent budget proposals for APN, AAN and OFNAC are drawn up by the agencies themselves as they are financially autonomous. They are now included in the consolidated budget. MTPTC's recurrent funding proposals, however, have to be-reviewed by MEF which determines the  -60- final budgetary allocations. MTPTC's proposals are aggregated from those of Its various departments with little serious review of their justification. As with other ministries, the bulk of MTPTC's recurrent budget covers saliaries, wages and allowances. The remaining components of the recurrent budget proposal are typically inflated with the expectation that they will be automatically reduced by MEF. This long-standing practice prevents realism in the budgetary allocations and invariably results in lower allocations for materials and equipment than are needed. 6.11 Control of public transport expenditures by MTPTC is lax. First, the budget allows a rather sizeable amount to be used at the discretion of the minister of MTPTC without the need for justification. The exact amount is unknown, but is generally believed to be high, especially for MTPTC which is a large public sector employer. Second, unlike most ministries (Chapter II), MTPTC is not held accountable for the justification of expenditures against individual budgetary items or individual entities; funds are considered totally fungible. It is therefore impossible for NEF to properly account for the use of funds against individual items or entities in MTPTC. For example, it was not uncommon for the mission to receive three separate expenditure figures from MEF, MTPTC and the individual entity involved for the same items in the budget. A strong need exists to: (a) base budgetary proposals on agreed work programs; (b) account for expenditures against individual budgetary items; and (c) ensure that there is consistency among the accounts kept by MEF, MTPTC and individual entities within the ministry. 6.12 The FY86 recurrent budget for the entire sector, including MTPTC, AAN, APN, OFNAC AND SEMANAR, totals about G 88 million. The bulk is allocated to SEPPRN for road maintenance (G 40 million) and to APN (G 32 million). SEPPRN was transferred from the development to the recurrent budget in FY84; in general adequate funds have been provided in the past for road maintenance, about G 36 million per year; the FY86 allocation shows that this will continue and is justifiable. That for the financially autonomous APN enables it to operate at a profit with tariff levels which fall within the middle range of those charged at other Caribbean ports. 6.13 MTPTC is overstaffed. Two-thirds of its budgeted recurrent expenditures are for salaries, and this excludes the funds the minister may spend without justification. Less information is available for the autonomous agencies. Since disaggregated data for salaries and wages, materials, etc. do not exist, it is difficult to make any detailed assessment of the manpower budget. It is, however, quite evident that there is a high level of overstaffing in the sector at both professional and non-professional levels. Professional staff numbers at MTPTC have increased disproportionately to increases in work volume as each new minister comes in with a not insignificant number of professional employees, often to satisfy political constituencies, without terminating those from previous ones. The extremely high turnover of ministers--four in 1985 alone--has made the problem acute. 6.14 At the non-professional level, many staff, originally employed for fixed terms to satisfy peak demands, have remained on the payroll even though in some cases the demand for their services has ceased to exist. The most recent former minister of MTPTC estimated that the entire ministry's  -61- manpower could be trimmed by up to 20 percent if these staff could be laid off. 6.15 The excess staff are not necessarily financed by the recurrent budget. Within MTPTC, the recurrent budget and the Government contribution to the development budget are considered fungible. A large number of staff performing recurrent operations are in fact paid out of the development budget, even after the transfer of SEPPRN. During the FY86 budget review, MEF attempted to reduce redundant staff within MTPTC by eliminating all wages, salaries and allowances associated with programs no longer funded under the development budget. MLPTC estimated that laying off all staff thereby affected could reduce its payroll by some 53 percent. Perhaps 30 percent of these could be redeployed to more justifiable recurrent tasks or useful development jobs. Some 20-25 percent could still be terminated with no reduction irn MTPTC's output and savings of some G 24 million per year. Action to lay off the redundant staff has, however, got nowhere because of continuing frequent ministerial changes and because of the MTPTC's intepretation of the Minister of Finance's communique of October 18, 1985 to mean that no public employee could be dismissed (Chapter I). Investumt Program and Finecin Plan, 7Y8-89 6.16 As noted above, no serious integrated planning is carried out for the transport sector. There is thus no official program to present for FY86-89 or any other period. What does exist is the investment budget for FY86 and a list of projects under discussion in DDT for the next five-year (FY87-91) Plan (Table VI.1). No financing plan exists beyond FY86 although certain major aid loans and grants are anticipated; hence Table VI.1 does not show financing. Table VI.1: PRELIMINARY TRANSPORT INVESTMENT PROGRAM, FY86-91 (G million at FY86 prices) FY86 FY87-91 FY86-91 Roads - budgeted 67 1188 1255 - unbudgeted 17 - 17 Ports and maritime 9 239 248 Civil aviation 15 101 116 Total 107 1527 1634 Sources: Statistical Appendix Tahles 6.1 and 6.2. 6.17 FY86. G 91 million is budgeted to be spent on transport investments in FY86, a very slight real increase over recent years. The program consists of nine projects, covering roads (73 percent), ports (10 percent) and airports (17 percent). 6.18 Two of the road projects are continuations of ongoing activities: secondary road construction (G 32 million) and the rehabilitation of the northern section of the Port-au-Prince to Cap Haitien Route Nationale #1  -62- (G 22 million). The latter is underfunded, however. Also proposed are studies for bridges and culverts (G 3 million), construction of the bridge at Guinodee (G 4 million) and the improvement of the road from Milot to the Citadelle (G 2 million). No economic feasibility study has been carried out of the Milot-Citadelle road, the principal benefit from which would be improved access to Haiti's major tourist attraction. Missing from the budget are funds to start bridge and culvert reconstruction. 6.19 Port investments include minor works to complete the port at Cap Haitien (G 3 million), port management improvements (G 1 million) and the improvement of the cabotage ports at Ile de la Tortue and Ile a Vache (G 6 million), which provide the two islands' only link with the outside world. A new investment proposed for FY86 is the modernization of the airport at Cap Haitien. 6.20 Some 81 percent of the budgeted investment program of G 91 million would be financed by donors with either grants or concessional loans. These funds are all committed. About 82 percent of the cost of the two cabotage ports would be financed by KfW. local AAN funds would be used for the Cap Haitien airport. Significant omissions from the budget are G 10 million for the early start of the bridge rehabilitation program under the proposed IDA Transport VII project and a shortfall of G 7 million for completion of ongoing works on Route Nationale #1. This means that the total expected investment program for FY86 is in fact G 107 million,' of which G 73 million (68 percent) will be financed by external aid and G 34 million (32 percent) by the Haitian public sector. 6.21 FY87-91. Haiti's preliminary project list for FY87-91 would cost about G 1,527 million in FY86 prices. The roads component would be the largest at about G 1,188 million (78 percent), including upgrading about 430 km of primary gravel roads to paved standards, rehabilitating about 170 km of existing paved roads, constructing or improving some 1,100 km of secondary and tertiary roads, constructing or rehabilitating selected drainage structures, and carrying out feasibility and engineering studies for some 480 km of primary roads. Investments in the ports and maritime subsector of G 239 million (16 percent) would involve expanding installations at Port-au-Prince including additional cargo handling and maintenance equipment, rehabilitating five further cabotage ports, additional facilities at Cap Haitien to make it fully functional, a maritime training center, a marine radio communications system and a slipway for boat inspection and repair. The civil aviation component would cost G 101 million (7 percent) to continue the modernization of the Cap Haitien airport, to expand the arrival and departure halls at the Port-au-Prince airport to provide a national air navigation system and to purchase five trucks and emergency generators. Iscouneudations 6.22 External donors participate in financing almost every transport project and so their preferences heavily influence investment selection. Up  -63- to 1983, the Government met annually with the donors as a group to discuss and coordinate assistance to the sector. The Government is interested In reviving these meetings and should do so, following the excellent precedent established in the electric power sector (Chapter V). Aid coordination meetings could help to avoid low priority investments, ensure a good balance between capital and recurrent expenditure, and coordinate institutional development and policy changes in the sector. 6.23 Recurrent spending to maintain existing investments is projected to average about G 100 million per year, a real increase of about 7 percent on the recent past. This could be substantially reduced if MTPTC's excess staff could be trimmed. A program to do this over the four years FY86-91 would involve annual staff reductions of only 5 or 6 percent with annual expenditure savings of about G 6 million. 6.24 The principal role of the transport sector for the next decade should be to support agriculture and the export industries. Investments in the next few years should therefore concentrate on completing high priority ongoing and rehabilitation projects and on consolidating gains in institutional strengthening. 6.25 The FY86 budget and supplementary commitment plus the preliminary FY87-91 project list would cover practically the entire transport infrastructure needs of Haiti for the next two decades, at an estimated cost of G 1,634 million. No attempt has yet been made by the Government to prioritize the investments or constrain their timing by financial or implementation capacity. More than 60 percent of them would be premature based on their low economic returns. 6.26 The highest priority road investments are: (a) completion of the bridge and culvert program; (b) reconstruction of the bridges at Fer-a-Cheval and Guayamouc (Hinche); (c) construction, improvement and rehabilitation of about 600 km of secondary and tertiary roads i.e. about 120 km per year instead of the proposed 1100 km (220 km per year); (d) upgrading the Mirebalais-Pont Sonde road; (e) rehabilitation of sections of the Port-au-Prince to Mirebalais road; and (f) studies of about 200 km of roads. Feasibility studies have shown all these investments to be justified economically; they are all directly linked with the provision of access to agricultural areas and the main consuming or export centers. The Milot-Citadelle road should not be constructed until a feasibility study Is completed. In addition, G 17 million should be added to the FY86 budget  -64- for Route Rationale #1 rehabilitation and for bridge and culvert reconstruction. The recommended investments would cost about G 597 million at FY86 prices over FY87-91, of which about 76 percent or G 454 million (US$91 million) would be in foreign exchange. 6.27 Port and maritime priorities are: (a) port maintenance equipment, especially at Port-au-Prince; (b) minor works to complete facilities at Cap Haitien; (c) rehabilitation of the-port at Jeremie; (d) a radio communication system for SElANAH; (e) a slipway for vessel inspection and repair; and (f) a training center for SEMANAH. There are also long-term proposals to expand the Port-au-Prince port and facilities, which depend on the Government's plans to create an industrial free zone (Chapter IV); any needed investments are unlikely before FY89. 6.28 The proposed rehabilitation of additional cabotage ports should be postponed. Studies found rates of return of 7-8 percent, well below the opportunity cost of capital. More importantly, experience with the ten ports already improved has shown that they have not yielded the benefits anticipated because of pricing problems and SEMANAR's weakness as an institution. No more cabotage ports should be built until the completed ones have been improved by strengthening SEMANAH's management and introducing more appropriate pricing; the only exceptions are Ile de la Tortue and Ile a Vache which, though not economically justified, do represent the only communication between the two islands and the mainland and will be 82 percent financed by external aid grants. It would be preferable if donors would fund the entire investment. 6.29 Including these two cabotage ports, the priority port and maritime investments would cost about G 74 million over FY86-91, of which about 79 percent or G 58 million (US$12 million) would be foreign exchange. 6.30 With two exceptions, the projects proposed in civil aviation are relatively minor and justifiable. The exceptions are the airport at Cap Haitien and the national air navigation system. Modest though it is, the modernization of the Cap Haitien airport has no economic or financial justification; at most, one or two commercial flights a week would be expected. It would not be able to recover its operating costs, much less the initial investment, and would therefore become a continuous fiscal drain. The air navigation system also requires more in-depth economic and financial study. Unless such study shows the system to be feasible, civil aviation investments should be limited to G 8 million over FY86-91. 6.31 In the event any private Haitian airline should experience financial difficulties, it should not be subsidized, taken over or have its debt guaranteed by the public sector.  -65- 6.32 A revised list of transport investments along the lines described above would cost about G 683 million in FY86-91. Table VI.2 shows the investments of G 465 million recommended for FY86-89 and their financing. About 80 percent or G 372 million (US$74 million) would be in foreign exchange. Some G 200 million of the entire FY86-91 program would be left for execution in FY90-91. Table VI.2: RECOMMENDED TRANSPORT INVESTMENT PROGRAM, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 Applications 88 377 465 Roads 77 399 Ports and Maritime 9 53 62 Civil Aviation 2 6 8 Sources 88 378 466 tiernal Cash Generation (AAN, APN) -T -Zr 75 Government 26 60 86 Existing External Aid 58 17 75 New External Aid - 279 279 Source: Statistical Appendix Table 6.4. 6.33 In terms of cost, this recommended program is only 42 percent of the list under consideration in DDT. It would still involve a 14 percent annual real increase in transport investment, although the Government's contribution would remain constant at about G 20 million per year. Even this reduced program may be beyond the sector's implementation capacity; historically, only 60-70 percent of planned transport investments have actually been executed. A smaller real program than recommended above is therefore very possible. Table VI.3: RECOMMENDED TRANSPORT EXPENDITURE, FY86-89 (G million at FY86 prices) FY86 FY87 FY88 FY89 Capital Expenditure 88 109 121 148 Internal Cash Generation and Government 7 28Z External Aid 58 82 92 122 Recurrent Expenditure 82 96 84 78 Total Expenditure 170 205 205 236 Iinternal Cash Generation and Government 112 i23112 1}54 External Aid 58 82 92 122 Sources: DDT; Statistical Appendix Table 6.4; mission estimates.  -66- 6.34 These recommended capital and recurrent expenditures are combined in Table VI.3. The public sector's expected contribution to total transport spending would be only slightly higher in real terms than in the past.  -67- CHAPTER VII TILHcoHNIJICATIONS Tb. Sector and its Plapalm 7.01 The responsibility for the provision, operation and maintenance of public two-way telecommunications, both domestic and international, telephone and telex, is the monopoly of Telecommunications d'Haiti (TELECO). Ninety-six percent of its shares are held by BRH in the Government's name; the remainder are privately owned. The Minister of Finance chairs its Board; the Minister of Public Works, Transport and Cowmunications is a member. TELECO appears to be appropriately organized for efficient management; its planning, operations and staff training have improved considerably in recent years, partly thanks to technical assistance from the International Telecommunications Union (ITU). The only aid donor currently involved in financing investments in the sector is the French CCCE. Japan is interested in future financing but has as yet made no commitment. 7.02 During 1985, TELECO's new management improved organizational efficiency and planning, reduced debt and strengthened the company's finances. TELECO's Planning Department, created in 1984, is completing a telecommunications development master plan, with the assistance of ITU experts. The plan, to be ready in early 1986, will analyze fully the demand for services in Haiti as a whole and will seek to satisfy already evident demand in the Port-au-Prince area while at the same time improving provincial and rural coverage. Digital technology is currently being introduced in parts of Port-au-Prince; a strategy to extend it throughout the country will likely be proposed, together with required quality improvements. Recommendations will also be made about tariff policy and a management information system. 7.03 Operational, technical and service quality improvements since 1983 include: (a) agreement with EdH to use electricity poles for telephone lines for G 75 per pole per year, replacing exclusive telephone poles; (b) improvements in the outdoor plant and in line maintenance which have reduced the average annual number of faults per Direct Exchange Line (DEL) from 1.9 to 1.3; (c) taking over of the operation and maintenance of the satellite earth station and the international telex exchange; (d) the drawing up and successful application of a new employment and recruitment code which clearly specifies the qualifications for and responsibilities of each grade and position.  -68- 7.04 Yet major problems remain: (a) Service Goverage. Despite recent and ongoing expansions, telephone density is still very low and access is unsatisfactory outside the Port-au-Prince area. There were approximately 32,000 DELs in October 1985, i.e. 0.6 telephones per 100 inhabitants compared to averages of 5.5 for Latiai America and 0.8 for Africa. Five central exchanges with 24,000 DELs serve the Port-au-Prince area. There are also exchanges at Les Cayes, Cap Haitien and Gonaives and very old ones at Port de Paix and Jereiie. Communications elsewhere must frequently rely on two-way radios. About one third of the urban demand for telephone subscriber services and one half that for telex services remain unsatisfied. (b) Tariffs. There is over-reliance on income from international services while domestic tariffs are far below costs. (c) Staffing. TELECO is overstaffed with approximately 1,500 employees, or 50 per 1,000 DELs in service. In the United States and Western Europe, the modern industry standard is about 10 per 1,000 DELs; as a developing country Haiti should aim for no more than 30. Cd) Procurement. Although the new management has placed more emphasis on competitive bidding, there is still over-reliance on tied aid for most new investments. (e) Relations with Government. The central Government exercises negligible control over policy planning or implementation. MEF has not yet used the authority that is now offered by the inclusion of TELECO within the consolidated budget for the first time in FY86. KTPTC and MP play no practical role in determining policy or investment levels. Given TELECO's current relatively competent management, this is not at present a major cause for concern, but it does make the system potentially vulnerable to abuse. The company's accounts are not subject to any external audit. While investment budgeting procedures are adequate, and planning has improved considerably, the Government makes virtually no intervention in TELECO's investment cycle procedures. However, while the Govern-ent hardly regulates TELECO, it uses it as a source of funds and foreign exchange in ways which limit its freedom of action to implement improvements. TELECO and the Minoterie are the two public enterprises which make significant profits, and TELECO earns foreign exchange. Its funds have been used in the past, for example, to meet both the current and some investment costs of the uneconomic USND sugar mill (Chapter IV).  -69- Financial Aspects 7.05 Analysis of TELECO's finances must bear in mind that its accounts are not subject to external audit. With these reservations, recent financial performance has been strong with operating ratios averaging about 50 percent since FY81. The return on average net fixed assets has averaged about 35 percent. (TELECO's balance sheet at the end of FY84 is in Statistical Appendix Table 7.2.) Table VII.1: TELECO: IT.OME STATEMENT, FY81-85, AND BUDGET, FY86 (G million at current prices) FY81 FY82 FY83 FY84 FY85 FY86 Est. Bud-et Operating Revenues 82 123 151 196 185 169 Operating Expenses 30 48 68 60 81a/ 768/ Current Surplus 53 75 83 137 104 93 Interest 11 16 9 9 9 3 Depreciation 15 20 21 27 .. .. Net Income 27 39 54 102 95 90 Operating Ratio (percent) 54 56 58 44 44 46 Return on Net Assets (percent) 25 34 36 46 38 44 a/ Includes depreciation in FY85 and FY86. Source: Statistical Appendix Table 7.1. 7.06 Nearly 80 percent of total revenues come from international traffic. Of international revenues, 85 percent are earned from overseas calls placed to Haiti (including collect calls from Haiti), 90 percent of them from the United States. TELECO's tariffs for overseas calls from Haiti are extremely high by international standards. Calls to Europe cost a flat G 25 (US$5) per minute and those to the United States average about G 10 (US$2) per minute. TELECO is reluctant to reduce its international tariffs; it prefers people abroad to call Haiti instead and so earn it and hence the Government foreign exchange. Revenues peaked in FY84. Lower revenues in FY85 and budgeted for FY86 reflect a fall of about G 20 million in income after increased fraud led to rcatrictions on credit card calls originating in the United States. 7,07 Revenue from domestic traffic has also been increasing. The number of DEL. has risen from 11,500 in 1978 to 32,000 today, an annual growth rate of 16 percent. Domestic tariffs were increased by 20 percent in February 1984 and currently consist of a flat monthly charge of G 75, irrespective of the number or duration of calls. These domestic tariffs do not cover costs and there is substantial cross-subsidization from  -70- international revenues. The tariff study to be completed shortly with ITU assistance is expected to recommend the gradual introduction of tariffs reflecting the number and length of calls. 7.08 It is probable that there is a large suppressed demand for domebtic andt particularly, for international services, which are normally highly price elastic. Quality improvements, such as the introduction of direct overseas dialling, would further stimulate demand. However, management fears that direct dialling facilities might overload the system, encourage fraud and reduce foreign exchange receipts. 7.09 Annual cash outlays have risen from G 40 million in FY81 to G 76 million in FY85. The key element in these increasing expenditures has been the wage bill which has increased dramatically from G 10 million in FY81 to a budgeted G 54 million in FY86, an average annual rate of growth of 42 percent in nominal terms and 30 percent in real terms. Part of this was due to necessary restructuring of salaries and grades in FY84-when the wage bill rose by 71 percent-to recruit and retain skilled personnel; part is also due to overstaffing. Investmnt Program and Finant Plan. FT8-9 7.10 An estimate of the current investment program is shown in Table VII.2; detailed investment plans have not been drawn up, pending completion of the financial section of the master plan. Total investments are estimated at G 211 million in FY86 prices, G 61 million in FY86 and about G 50 million each year in FY87-89. About 75 percent would be in foreign exchange. 7.11 The centerpiece of the current program is a two-phase modernization of the Port-au-Prince telephone network through the introduction of digital electronic exchanges. This will almost double the number of subscribers to about 60,000 DELs in phase one and increase them by a further 15,000 in phase two. Table VII.2: TELECO: INVEST ENT PROGRAM AND FINANCING, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 Applications 61 150 211 Port-au-Prince telephone network 50 T-3 180 Telex 3 6 9 Other 8 14 22 Sources 61 150 211 Internal Cash Generation Tr Tff TN Concessional Aid (CCCE) 40 45 85 Sources: TELECO; mission estimates. 7.12 Over half this demand is already clearly identified through DEL requests and it may be anticipated that the new capacity will be swiftly  -71- absorbed. Also, in phase two, existing equipment in Port-au-Prince will be dismantled and reassembled in a number of provincial cities. It will be compatible with the new digital equipment in the capital and so will greatly tuprc-ve cowmunication between Port-au-Prince and the interior of the country as well as between provincial cities. Provincial lines would increase from the current 8,000 to about 23,000. Phase one started in FT85 and was scheduled for completion in early FY86. Phase two Is due to begin In mid-FY86 and will be completed by about mid-FY88. 7.13 The total cost of the expansion is estimated at about G 175-200 million in FY86 prices. About 65 percent (US$25 million) would be in foreign exchange. Detailed costs have not yet been determined for phase two. CCCE is financing about half the total program, and thus about five sixths of the foreign costs, with FF 105 million in phase one and perhaps FF 55 million in phase two. TELECO's own contribution is expected to be roughly the same, and will therefore also involve financing about one sixth of the foreign costs. No central Government funds are involved. CCCE aid is tied and only French firms may bid to supply the equipment. The total project cost (including all transmission, distribution, buildings, vehicles, etc.) of about US$1,000 per new line installed is reasonable, although World Bank experience in other countries shows that full international competitive bidding can reduce equipment costs by as much as 30 percent. This is particularly so in an industry at the leading edge of technological change; both demand and supply are Increasing at unprecedented rates and international competition is especially intense. 7.14 TELECO also plans to install a new national and international telex center and replace the existing stock of 500 machines on hire, many of which were installed more than 20 years ago. It appears that the foreign firm which operated the international telex service before January 1985 invested little in maintaining quality. The new center will quadruple traffic capacity and remove many of the current sources of consumer complaints. Congestion constraints cause some difficulty for current users but, more importantly, restrict access for new users. Over the next three years capacity will expand to 800 machines. Data processing equipment will also be acquired for the new management information system. These investments will cost a total of about G 15 million in FY86-88, of which about 90 percent (US$2.7 million) will be foreign exchange, and will be financed entirely from TELECO's own resources. Table VII.3: TELECO: BUDGETED AND ACTUAL INVESTMENT, FY84-86 (G million at FY86 prices) FY84 FY85 FY86 Budget 46 81 61 Actual 15 50 Implementation Rate (percent) 32 61 Source: Statistical Appendix Table 7.3.  -72- 7.15 Table V11.3 indicates that the rate of implementation of investment was very low in FY84 and that, although FY85 registered a substantial improvement, there is still considerable scope for eliminating slippage or for more realistic budgeting. Early indications are, however, that the FY86 budget will be spent within the first half of the year. ec ndations 7.16 So far, TELECO has financed almost all its relatively modest investments from internally generated funds, but the costs of its share of the current modernization program will exact a heavy toll on its financial reserves. Moreover, in order to continue to modernize and expand the network, especially in the provinces and rural areas, it is likely that investment costs per new line installed will} increase sharply. According to World Bank experience elsewhere, costs of US$3,000 per new line are not uncommon in remote rural areas in small developing countries. 7.17 It is therefore important that TELECO continues to generate strong current surpluses. This will require three complementary policies. First, domestic tariffs need to be raised and linked to the length and duration of calls to cover marginal costs; international tariffs are already very high and if anything should be reduced. Second, the Government should cease using TELECO's financial resources for investments or current expenditures for extraneous uneconomic projects. Third, TELECO must contain its own current expenditures. The investment program should reduce maintenance expenditures considerably in Port-au-Prince and also permit substantially greater labor productivity. As a priority, wage costs per line installed should be reduced by at least one half as capacity is increased by a total of 60,000 new lines. Surplus labor must be eliminated. 7.18 TELECO should not need any further external financing through FY89. In the future, however, TELECO would do well to diversify its external sources of funds as much as possible and not rely so much on one country of supply or on tied aid. TELECO could realize savings through regular international competitive bidding procedures, particularly if it resorts to price and terms bidding as is now increasingly practiced in the telecommunications sector. 7.19 The company's accounts should be externally audited on an annual basis by a recognized auditing firm. This had begun by July 1986. 7.20 TELECO's investment program answers immediate needs by satisfying already apparent demand and implementing service quality improvements. While it is nevertheiess unfortunate that major ongoing investments and those for the next two years were decided upon before the master plan was ready, all other future investment requirements should be identified and analyzed in the plan. It is to be hoped that the master plan both builds on recent organizational improvements and encourages more in-depth analysis of prospective projects.  -73- 7.21 A full economic and financial analysis of major investments should be included as part of TELECO's regular planning procedures. The new automatic telephone system for Port-au-Prince was subjected to a thorough technical analysis but there was no assessment of its economic feasibility. It is almost certain that suppressed demand in Port-au-Prince and the dramatic increase in services elsewhere will result in this project having a high social and economic rate of return. However, this may not be the case with future investments.  -74- CIAPTUE ViII THE URBAN SECTOI The Sector and Its Plan1lag 8.01 With less than 30 percent of its population living in towns, Haiti is one of the least urbanized countries in the Western Hemisphere. Three different categories of cities can be distinguished with very different planning, investment and management needs: the capital, the three major secondary eities and 22 smaller urban centers. 8*02 Port-au-Prince is the center of Government administration and of economic activity. Its metropolitan area accounts for 85 percent of industrial output, 60 percent of commerce, and 70 percent of real estate, housing and government services and administration. It has about one million inhabitants and is growing at over 5 percent per year. The problems of rapid and unplanned growth are compounded by the heavy concentration of low income people, high residential densities, traffic congestion and scarce basic services. 8.03 The three secondary cities of Cap Haitien (7C,000), Gonaives (47,000) and Les Cayes (32,000) are growing much more slowly than the capital. Their principal problems are the provision of shelter and basic services. These cities also lack the capacity for basic planning and administration and suffer from a lack of attention and failure to coordinate among relevant ministries. 8.04 The 22 smaller urban centers are widely scattered and have a total population of about 350,000. They suffer from the same problems as the three secondary cities, but poor transport and telecommunications links mean that they are virtually ignored by the central administration. 8.05 Municipal governments are not elected and come under the tutelage of the Central Government. Their responsibilities are restricted to the administration of a few urban services, such as collection and disposal of garbage. Their staffs are insufficient and inadequately trained. Their financial positions are precarious; more than half of their revenues come from taxes earmarked for the municipalities and collected by the Ministry of Finance, such as the property tax and the license tax on enterprises. The remainder consists essentially of grants from the Central Government. In real terms, revenues from local taxes have been falling in recent years and tax collections are inefficient. 8.06 Public responsibility for planning and managing the urban sector is dispersed and overlapping. The primary agency is the Urban Directorate of MTPTC with three divisions: (a) the Urban Planning Division is responsible for all urban and rural plans and regulations and for technical specifications and designs for all types of construction  -75- in both the public and the private sector. Its routine work overloads it to the point it cannot provide meaningful planning advice. (b) the Urban Works Division is supposed only to execute infrastructure works. In practice, it also constructs public buildings and executes works that are formally the responsibility of the municipalities. (c) the Maintenance Division maintains urban infrastructure. However it is so inadequately staffed and equipped that SEPPRN, which is supposed only to maintain inter-urban roads, is often called on to carry out urgent or priority urban rehabilitation works, especially in Port-au-Prince. MTPTC is also responsible for CONATRA, the public bus company in Port-au-Prince, and CAMEP, the capital's water supply company (Chapter IX). 8.07 The second major ministry involved in the urban sector is the Ministry of the Interior which has responsibility for the 29 municipalities and pays their staff. It also contains the National Cadaster Office, which has not yet done any work in urban areas. 8.08 The Metropolitan Government ("communaute urbaine") of Port-au-Prince falls under the Ministry of the Interior and is a legal umbrella for four municipalities, having no staff or coordination procedures of its own. The major municipality of the four, that of Port-au-Prince, has no financial or administrative capacity to discharge its responsibilities for garbage collection and disposal. -In addition its administration is weak, and there is no clear division of responsibility among it, the other three municipalities, the Metropolitan Government and the various relevant central Government ministries regarding urban planning, the enforcement of land use legislation, investment and maintenance. The result is serious inefficiency and waste. 8.09 The autonomous public housing agency EPPLS reports to the Ministry of Social Affairs. Its role is to prepare programs and implement social housing projects for low income people, and to ensure economic rates of return and full cost recovery through both rentals and sales. Other ministries involved in early 1986 in housing included the Ministry of the Presidency, which developed two projects totalling 400 units without any cost recovery, and the Ministry of Agriculture which played a negligible role. 8.10 These dispersed and uncoordinated urban responsibilities mean that there is no explicit Government strategy for the urban sector as a whole. Chapter II notes the Ministry of Planning's irrelevant effort to promote regional planning. There is no urban master plan for Port-au-Prince. There is little or no planning of urban transport needs. There is no housing policy. Drainage is being addressed through major urban investments  -76- without necessary complementary actions to protect the surrounding watershed, the erosion of which is silting up the new drain structures. 8.11 Urban expenditure considered in this chapter includes that on drainage, housing,1markets and transport; water supply is discussed in Chapter IX. In reent years, as Table VIII,1 demonstrates, some two-thirds of planned investment was actually made; two-thirds of this was for drainage in Port-au-Prince. About 48 percent of the total was in foreign exchange. Table VIII.1: PROGRAMMED AND EXECUTED URBAN INVESTMENTS, FY81-85 (G million at current prices) Programmed Executed Drainage (Port-au-Prince) 241 198 Housing 71 53 Markets 65 5 Transport 46 22 Technical Assistance 10 10 Other 5 5 Total 438 292 Source: Statistical Appendix Table 8.1. 8.12 Drainage investments in Port-au-Prince have been financed principally by IDB. However these investments have been jeopardized by inadequate phasing, a lack of coordination with other infrastructure investments, and inadequate solid waste collection. A major program of maintenance is already under way for drainage works that were only recently completed but were not routinely maintained. Some rehabilitation is also needed because of the lack of maintenance. Maintenance costs have as yet no assured financing. Uncollected trash clogs streets and drains, rendering any drainage system inoperative. Trash collection is therefore essential to permit drainage investments to yield returns. There is still no system for recovering even operations and maintenance costs for th3 new fleet of garbage trucks that is intended to increase collection coverage from 40 to 90 percent. The clear definition of respective responsibilities between MTPTC and the Municipality of Port-au-Prince is crucial to an adequate return on drainage investments, as it is for the proper functioning of the urban system. 8.13 EPPLS has since 1982 developed five housing projects, producing about 1,100 new units per year and rehabilitating a substantial number more, with financing from IDA under its market project and from the Federal Republic of Germany. EPPLS' effectiveness has suffered from the recent political events and may take time to be restored.  -77- Recurrent Spendlog 8.l/i No useful data on past recurrent spending are available. The FY86 budget amounts to about G 18.4 million, or some US$2.5 per urban inhabitant (Table VIII.2). Personnel costs represent almost 60 percent of the total. In faet, the share of personnel is almost certainly higher as G 6.0 million of the G 7.8 million of non-personnel costs represents a transfer from the Ministry of the Interior to the Municipality of Port-au-Prince, much of which is very probably used for sal4ries. By low, Port-au-Prince should have its own budget; in practice, it does not. Table VIII.2: URBAN SECTOR RECURRENT BUDGET, FY86 (G million) Ministry/Agency at Personnel Other Total MTPTC 6.9 0.8 7.7 EPPLS 1.4 0.2 1.6 Interior 2.3 6.8 9.1 Total 10.6 7.8 18.4 a! Excludes the Ministry of the Presidency for lack of data. Source: Statistical Appendix Table 8.2. 8.15 The incremental annual recurrent cost of maintaining the FY81-85 transport and drainage investments is about G 8.0 million, assuming an average annual maintenance cost equivalent to 4 percent of the initial investment. Total maintenance expenditure planned by MTPTC for FY86 in fact amounts to only G 4.0 million, and this is for all maintenance and urban works, not just those in Port-au-Prince. It is also much more than has ever been successfully implemented in the past. 8.16 G 1.3 million of the MTPTC budget is for CONATRA, the public bus company for Port-au-Prince. This company's fleet of vehicles has declined from 80 on its creation in FY80 to only 15 today; it has only an insignificant role in the active, competitive passenger transport market in Port-au-Prince which is very adequately handled by private sector jitneys known as "tap-taps." 8.17 As discussed in Chapter VI, M?TPTC's staff is excessive. Investm8nt Program and Financing Plan, FY86-89 8.18 So dominant is Port-au-Prince and so overwhelming are its needs that this chapter does not further consider other urban centers. The intentions of the various agencies and donors involved in the urban sector are summarized in Table VIII.3. None has concrete plans for PY89 and those  -78- for FY87 and FY88 are not clearly separated, though most agencies plan to spend about half their two year total in each year. (The data exclude investments by the ministries of Interior and Presidency for lack of available information). 8.19 Total investments planned in December 1985 for the three years amount to G 270 million at FY86 prices, of which about 48 percent or G 130 million (US$26 million) would be in foreign exchange. Foreign costs would all be covered by external financing of G 206 million. Table VIII.3: PROPOSED URBAN SECTOR INVESTMENTS, FY86-88 (G million at FY86 prices) FY86 FY87-88 Applications 149 121 Drainage 56 54 Housing 27 20 markets 34 36 Transport 24 Technical Assistance and Training 8 12 Sources 149 121 Government 53 12 Concessional Aid (Loans and Grants) 97 109 Source: Statistical Appendix Table 8.3. 8.20 The Port-au-Prince drainage project will continue to represent the bulk of sector investment; it is the only project for which new external financing is firmly anticipated before FY90 - a new IDB loan of US$45 million. In addition a drainage rehabilitation program of about G 30 million is being implemented. No transport investments are planned after FY86; those in FY86 itself are, in fact, studies and not physical construction. The technical assistance and training pipeline is growing rapidly and will represent 9 percent of the total after FY86. 8.21 Not only is there no clear division of responsibility among seceor agencies, there is no urban development strategy or plan. A compreher.aive plan for the Port-au-Prince metropolitan area should be elaborated to guide future investments, including those for water supply (Chapter IX). It should include in particular land use, and also reforestation studies for the watersheds surrounding the city. Corresponding enforcement procedures should be defined and applied. If the plan is not prepared and used to influence future investments in the 1990s, rapid growth in some areas of the metropolis and the de facto and uncoordinated plans implicit in the ongoing drainage and proposed water supply projects (Chapter IX), will result in major areas being inadequately or not at all serviced, and in increased strain on existing infrastructure.  -79- 8.22 The FY86-89 investment program is inadequate In terms of implementation capacity, subsectoral balance and provision for recurrent cost financing. In addition, the future of the IVAfinanced Croix-des Bossales market is in doubt. 8.23 Implementation Capacity. Although planned overall sector spending is modest on a per capita basis, that for the one year FY86 is, even in nominal terms, 250 percent of the annual average In FY81-85. In addition, the proposed Treasury share of 30 percent in total investment is unrealistic in view of recent experience - it has averaged about 19 percent. Indeed, the very large program proposed for FY86 itself reflects the accumulated backlog of investments planned but not executed since FY82. Both Government and donors are over-optimistic in assuming that the backlog can be completely made up in one year. Much more reasonable would be to spread the entire proposed FY86-88 program over the four years FY86-89 with a modest expansion each year. 8.24 Subsectoral Balance. Drainage is a high priority area for Port-au-Prince. The city is sited on and below a steep hill subject to heavy rains and soil erosion with much of its areas built on a swamp, subject to flooding. It is expanding in the swamp area. Discharge of stormwater to the sea is difficult. Nonetheless, no further drainage investment should be made until certain steps have been taken to ensure that it will function once it is constructed. The Government should draw up the comprehensive metropolitan plan (para. 8.21); arrange for tariffs for solid waste collection to cover at least operating costs; and establish a system for regular maintenance of the drains; and define clearly the responsibi- lities of the Municipality and MTPTC. In addition, the phasing of the drainage investment program has been arbitrary, defined by availability of US$50 million amounts, without reference to physical work, and should be reassessed. The current phasing has resulted in serious inconsistencies e.g. works not being carried out in contiguous areas but in unconnected places. The drainage investment program for FY86 and FY87 should therefore be cut back and that for FY88 and FY89 expanded, to permit these prerequisites to be taken before FY88. 8.25 The future of the stalled project to construct the Croix- des-Bossales market should be determined. 8.26 Urban transport has received insufficient attention in recent years and should now have top priority. This does not, however, necessarily mean a large investment program. What is critical for relieving traffic congestion and hence saving petroleum imports, easing workers' travel from their homes to their workplaces, and speeding the delivery of raw materials and finished products between the airport, port and the assembly industry is a series of minor works and a traffic management program. Such a program could be put together relatively rapidly. Aid donors have not so far shown significant interest in Haiti's urban transport subsector and should be encouraged to do so. 8.27 The scale of the rapidly growing technical assistance and training pipeline shiould be reassessed. Top priority should go to improving  -80-. municipal finances and implementing a cadaster that German aid is financing. 8.28 Recurrent Cost Financing. The proposed FY86-88 investment program would further increase recurrent cost needs by about G 5.0 million per year. Clearly maintenance capacity, programming and financing should be assessed in detail before further investments are made, especially those in drainage as noted above. 8.29 CONATRA should be closed. In particular, no new buses should be purchased for it. 8.30 Taking into account the above considerations, the investment program for FY86-89 might be revised as illustrated in Table VIII,4. About 48 percent would again be foreign costs. Table VYII.4: RECOMMENDED URBAN SECTOR INVESTMENT PROGRAM, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 Applications 67 278 345 Drainage (Port-au-Prince) 0 97 107 Housing 20 60 80 Markets 17 51 68 Transport 15 45 60 Urban Plan - 10 10 Technical Assistance 5 15 20 Sources 67 278 345 Government 1 33 66 Existing/planned aid - d--inage(IDB) 8 78 86 Existing aid - other 46 74 120 New aid - 73 73 Source: Statistical Appendix Table 8.4. This program assumes, however, that the Croix-des Bossales market continues as programmed; this now seems unlikely. Even this investment program may prove too ambitious, requiring a 15 percent annual real increase in actual expenditures Disbursements from new aid commitments of about G 73 million (US$17 million) are called for in FY87-89, perhaps implying total commitments to run through a longer period of US$25 million, all for the urban plan, for housing and for urban transport. If local financing of housing can be improved, external aid requirements could be lower. The critical component of the program is the preparation of an urban plan; a donor to finance this is urgently needed. 8.31 In addition to the above capital spending, recurrent spending for the urban sector will have to rise from its current level of about G 4  -81- million to at least 0 12 million in FY87 and perhaps G 18 million by FY89. Most or all of this could come from cuts in inessential parts of MTPTC's budget (excess staffing; CONATRA) and from effective recurrent nost recovery for solid waste collection and drainage services.  -82- VA2 SUPY Tb Seeter Iad Its Pl.dmX 9.01 Water and sanitation service levels in Haiti are extremely low. About 40 percent of the urban and 5 percent of the rural population have access to public piped water of uneven quality through house connections and publie standpipes. The remainder obtain water from vendors and from streams, rain and springs. There are no sever systems; less than 40 percent of the population have safe excreta disposal by latrines or septic tanks. Water and sanitation-related diseasee are among the principal causes of illness and death. Information on sanitaion is hard to obtain, and this chapter concentrates on water supply. 9.02 Partly because of the density of the population, the inadequacy of services in Port-au-Prince is critical. About 17 percent of the metropolitan population have direct access to water through 28,000 house connections and about 28 percent through standpipes. Limited sources of supply and a high level of losses result in extensive rationing for all classes of consumer. Low income groups suffer the most. Higher income people get around these problems by constructing costly private water tanks and buying water from trucks at G 7-20 per cubic meter for nondrinking purposes and indus*trially produced bottled water for drinking. No hard data are available on consumption or prices but it may safely ise said that average consumption is less than sae third that in cities of comparable size in the region and is high cost. 9.03 HTPTC is responsible for setting water supply and sewerage policies and for coordinating projects. The Minister of NTPTC chairs two public enterprises which are responsible for the provision of water supply: Centrale Autonome d'Esu Potable (CAMP) in Port-au-Prince and Service National d'Kau Potable (SNEP) everywhere else. Both are inefficient, overstaffed and poorly mnaged. Theoretically autonomous, they are heavily influenced in their tariff and personnel policies by the Central Government, which has failed to support professional management. 9.04 CAMEP has a history of adequate senior management, but is poorly operated. Its excessive, untrained and poorly disciplined staff of 600 provides poor service and maintenance. With 20 staff per 1,000 connections, CAMEP compares unfavorably with the 5-10 found elsewhere in Latin America and the Caribbean. A good indicator of CAMEP's Inefficiency is the high level of unaccounted-for water, estimated to reach a staggering 60 percent, mainly because of large system losses and numerous illegal connections. Management's efforts to improve the organization have been hindered by its inability to hire qualified and fire unqualified and undisciplined staff.  -83- 9.05 SNEP supplies some 450,000 people in 230 towns, providing free water through standpipes to two-thirds of its served population and heavily subsidized unmetered water through house connections to the remaining third. Although It also has 20 staff per 1,000 connections, it provides virtually no maintenance or operational assistance for the systems for which it is responsible. Its management has no planning or operational perspective, beyond a firm commitment to subsidized water. Its poor management and project execution capacities have led aid donors, like USAID and a Belgian NGO, to execute investment projects directly before turning them over to SNEP to operate. 9.06 The Ministry of Health (MSPP) has no legal responsibility for water supply but is executing water projects under its IDB-financed Postes Communautaires d'Hygiene et d'Eau Potable (POCHEP) program in small towns that are part of SNEP's responsibility. POCHEP is only an investment oversight unit, which lacks operational capabilities. Given SNEP's inefficiency, the IDB's insistence on a separate project execution unit is understandable. In addition, the Bureau d'Assainissement (BA) is responsible for sanitation. German aid is financing technical assistance to boot SNEP and BA aimed at institutional strengthening. 9.07 None of these sector institutions has an adequate planning or execution capability. Nor can they finance even part of their investments. Indeed, the sector is unable to finance its own operations and maintenance, much less investments. Lack of cost recovery has compounded institutional weaknesses. The result has been inadequate maintenance, insufficient investment, reliance on central Government subs'.dies to cover operating costs and dependence on foreign aid donors for almost all investment financing. 9.08 In FY80-85, annual investments averaged G 22 million in current terms, fluctuating between G 5 and G 34 million. This was not much more than that necessary to replace the existing capital stock (G 12 million of revalued depreciation) and maintain service levels (G 5 million) to accommodate population growth. Principal donors have been IDB (G 40 million in FY80-85), KfW (G 20 million) and IDA (G 11 million). Government contributions for investment have been G 40 million, on top of G 10 million to cover operating costs. Detailed income and funds flow statements for the sector as a whole, CAMEP, SNEP and POCHEP are included in Statistical Appendix Tables 9.1-9.4. 9.09 Despite the water agencies' planning and execution problems, recent investments were justified. Aid financing has led to donors determining priorities; foreign technicians have been largely responsible for sucessful physical execution. Bilateral donors and IDB have financed small (mainly gravity) systems in over 100 small communities that previously had no safe water. IDA financed SNEP's expansion of systems in seven provincial towns. 9.10 As noted, both CAMEP and SNEP fail to cover recurrent costs.. CAMEP has had a net annual operating loss of between G 1 and 3 million since FY80 and SNEP between G 2 and 3 million.  -84- 9.11 To overcome CAMEP's institutional and financial weakness, the Government has considered leasing the operations and entrusting the execution of investments of CAMEP to a foreign water supply company. The Government, however, now considers that conditions for that arrangement appear unfavorable in the near future. A management company could well Improve the operating efficiency and financial performance of CAMEP. An alternate route could be drastic Government actions to raise tariffs, support professional management and enable CAMEP to benefit from external management expertise, under a CCCE-financed project (Emergency Program). The Government's and CAMEP's performance in this program will largely determine CAMEP's ability to carry out further investments and the Government's strategy to pursue sector objectives. 9.12 SNEP could improve its institutional capabilities and also its financial position, but this would be a very difficult task involving a major change of attitude by its management and customers. Most of its small town systems are gravity-fed, however, and cost very little to operate and maintain. No significant change is anticipated in the immediate future. Invest_nt Prorm and Financing Plan, 1Y86-89 9.13 The Government has established International Drinking Water Supply and Sanitation Decade (IDWSSD) targets for 1990: (a) increase urban water servi^e coverage from 40 to 80 percent of the population; (b) increase rural water service coverage from 5 to 50 percent; (c) provide sewerage service to 40 percent of the population of Port-au-Prince and six medium size towns; (d) provide "adequate" sanitation to the remaining urban and rural population. Achieving these goals would require annual investments over five times larger than those achieved in recent years and is clearly unattainable. 9.14 Indeed such investment plans .as do exist reflect the impracticality of the IDWSSD goals. Although the Government has no structured plan for the sector as a whole, it does have material to draw upon. For Port-au-Prince, there is available a specific investment and financing plan to upgrade water supply services during FY85-95. The obstacle to industrial and commercial development that water supply poses in Port-au-Prince has led the Government to assign a high priority to investments in the metropolitan area. In addition, there is an investment program for the provincial towns and the rural areas for FY86-87 and a partial indicative program for FY88-89. These plans can be combined as presented in Table IXJI. Total investments in FY86-89 would be G 201 million in FY86 prices. No detailed estimates are available but experience suggests that about 60 percent would be in foreign exchange.  -85- 9.15 In line with Government priorities, 48 percent of planned investments are for CAMEP in the Port-au-Prince metropolitan area, including: (a) urgently needed technical assistance, engineering studies, construction of wells and purchase of equipment, to be financed mainly by the CCCE Emergency Program; and (b) the first phase of a G 300 million project to improve CAMEP's water services and institutional capacity, which IDA is preparing. This ten year project would increase those served with house connections from 17 to 25 percent of the metropolitan population and meet the demand of the most of the rest of the population through a system of commercially operated public standpipes, raising total service coverage to about 90 percent of the population. Assuming CAMEP is well managed, internal cash generation should finance some 20 percent of its investment program. Table IX.1: WATER INVESTMENT PROGRIA AND FINANCING, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 Applications 38 162 201 Port-au-Prince (CAMEP) 0w T Small towns (POCREP) 7 24 31 Small towns and rural areas (SNEP) 17 55 72 Changes in working capital - 4 4 Sources 38 162 201 Internal cash generation 2 28 -ff Existing aid loans and grants 31 78 109 New aid loans and grants - 63 63 Government contribution 4 3 7 Source: Statistical Appendix Table 9.1. 9.16 The other half of the investment program (52 percent) outside Port-au-Prince would include: (a) water systems in small towns to be constructed under the IDB-financed POCHEP program; and (b) water systems in small towns and also rural wells with handpumps to be built by SNEP, to be financed by grants from donors. No data are readily available on the service level improvements that would result. Internal cash generation in small communities is nil and Government funding of these small town and -ural systems is insignificant; these investments will thus depend almost entirely on donor financing.  -86- 9.17 No major sewerage or sanitation investments are planned, although the Port-au-Prince drainage project (Chapter VIII) will Inevitably carry sewage, though this is not its primary purpose. Drainage must precede sewerage in any case. 9.18 The proposed investment program would be ficanced 54 percent by existing external aid loans and grants, 31 percent by new aid, 11 percent by internal cash generation from a better operated CAMEP and 4 percent by Government contributions. New aid of G 63 million (US$13 million) would be needed, including G 46 million (US$9 million) of loans and G 17 million (US$4 million) of grants. Reco=endations 9.19 The above investment program implicit in the plans of CAMEP, SNEP and POCHEP is well balanced and justified. It is technically and financially feasible, although it does involve a real rate of expansion of investment of 30 percent per year; significant slippage is therefore possible. New aid commitments of US$13 million for FY86-89 are needed; this seems reasonable, though a larger commitment may in fact be called for to cover the entire CAMEP program through FY92. 9.20 No major sanitation investments are planned, although _here are undoubtedly components within community and integrated rural development programs. This approach is sensible because experience has shown that the construction of latrines etc. is only successful when it is an integrated part of other development. 9.21 As discussed in Chapter VIII, a major impediment to urban -investments, including water and drainage, in Port-au-Prince is the lack of an urban development plan. The preparation of the proposed CAMEP project has essentially led to the drawing up of such a plan, but only from the perspective of water supply. Before further investments are undertaken in the 1990s, an urban master plan is essential. 9.22 The feasibility of the Port-au-Prince component of the program--and thus the 30 percent real increase-depends on a drastic improvement in CAMEP's management, operations, finances and project execution capacity. Such an improvement can only result from drastic financial measures and management upgrading in CAMEP, or the hiring of a foreign water supply management company. CAMEP needs to raise tariffs to most consumers, restructure its staff, reduce their number per 1,000 connections by perhaps 50 percent, and hire and train qualified local personnel who might in the future take over the administration not only of a stronger CAMEP but of the entire water sector. Improved financial and administrative performance are assumed in the financial and investment projections used in this chapter; it remains to be seen if CAMEP will turn around its performance. 9.23 The strengthening of SNEP and POCHEP requires a radical revision of financial and institutional policies for small communities, inc3id1vg the introduction of at least partial cost recovery. While this does not appear very likely at present, it should be encouraged. POCHEP is considering introducing tariffs that would cover operating costs. Costs should also be  -87- cut by reducing the ratio of staff to connections, particularly as little maintenance Is necessary for the gravity systems that form the bulk of SUEP's responsibility. Inadequate financial performance is not, however, a reason to delay proposed Investments wnich are essential to maintaining current rural service levels. Poor project Implementation capacity means that donors will continue to execute these projects directly. The Government should, however, address the issues of at least recurrent cost recovery, management and project implementation capacity in the next few years in order to prepare for further investments in the 1990.. If it does not, the IDWSSD goals will not be met by the turn of the century, much less 1990.  -88- CHAPTER X RWUCATON the Sector and its Plannlig 10.01 Haiti's education indicators show It to be more like a low income African country than either its Caribbean neighbors or Asian low income countries. The adult literacy rate is around 35 percent, reflecting both a low primary school enrollment rate of 69 percent and a very inefficient education system. Total enrollments rose from 22 percent of the 6-24 year age group in 1971 to 38 percent in 1982, with a dramatic increase in rural enrollments from 15 to 32 percent. About one percent of the Haitian population has had some higher, nine percent some secondary, and 26 percent some primary education. Three-quarters of all enrollments are in primary education, of which almost 60 percent are in private institutions. 10.02 Indeed, a striking feature of the Haitian education system is the dominant role of the private sector, which includes schools operated for profit and by religious and other NGOs. The private sector has absorbed the bulk of the rising demand for education: 79 percent of the increase in primary enrollments since FY79, 93 percent of that at the secondary level, and virtually all that at vocational, technical and higher institutions. It now accounts for two-thirds of primary and over three-quarters of secondary and vocational and technical education. Only the higher level is an exception because the State University of Haiti (UEH) is public. Table X.1: THE PRIVATE SECTOR'S ROLE IN EDUCATION, FY84 (percent) Student Qualified Level Enrollment Schools Teachers Teachers Primary 58 69 67 34 Secondary 85 92 86 of Vocational/Technical 70 79 77 Hlgher 21 .. .0 Total 62 .. .. .. a/ Teachers with at least 10 years of basic education. Sources: Statistical Appendix Tables 10.3-10.7.  -89- 10.03 The Ministry of Education (MEN) has since 1978 been responsible for national educational policy and also, with certain exceptions, for all public educational expenditure. The chief exception is UEH, autonomous since FY86, which is funded directly from the general recurrent budget, except for the Faculties of Agronomy and Veterinary Medicine (MARNDR recurrent budget), Medicine and Pharmacy, and Dentistry (MSPP). The Ministry of Social Affairs operates six vocational training centers. 10.04 MEN's Planning Department, created in the late 1970s'with technical assistance from CIDA and UNESCO, is responsible for the preparation and mronitoring of the five-year and annual plans, the programming and evaluation of the basic education reform now underway (see below), the coordination of external aid, and the collection of statistics. Originally a relatively- effective unit, the department lost political support in the last three years, its functions were limited and the quality of its work deteriorated. The gathering of statistics is the only function it now carries out effectively; it does no substantive expenditure planning, evaluation or monitoring. 10.05 Primary Education. Primary enrollments have accelerated since the early 1970s, growing at 3.3 percent per annum in FY58-70, 5.9 percent in FY70-79 and 8.2 percent since, the recent rates well above population growth of 1.8 percent. Since FY79, private school enrollments have increased at 12 percent per year, three times the four percent rate at public s4hools, the latter being concentrated in urban areas. 10.06 The structure of primary education is very complex, according to whether it is urban or rural, public or private, traditional or "reform." Seventy percent of primary classrooms still use the traditional curricula and teaching methods, consisting of a preparatory year followed by one cycle of six years, and taught throughout in French (Haiti's official language that is spoken by only about 20 percent of its people). This traditional system has been characterized by large numbers of overage students, high repetition and dropout rates, and low quality instruction. On average, approximately 12-13 student years of instruction are required to produce a primary school graduate, ranging from about eight in the best private urban schools to about 15 in rural for-profit ones. These figures probably overestimate the primary system's internal efficiency, howeve , especially in rural areas, because an unknown but significant proportion of dropouts later enroll again and are hence more strictly repeaters, costing the system more. 10.07 The reform system, in which about a third of all primary children are now enrolled, has been introduced gradually since 1982 in an effort to overcome the major drawbacks of the traditional system. Intended to cover all schools by 1991, it consists of two cycles of primary education (4 and 3 years), followed by one three-year cycle of basic secondary education. Functional literacy is to be achieved by the end of the first primary cycle. The restructuring of the cycle is accompanied by new curricula; the use of the universally-understood Creole as the language of instruction; changed teaching methods (e.g. a teacher staying with a student cohort for two years, discovery learning rather than memorization, grouping of students by ability level); automatic promotion from grades 1 to 2 and 3 to 4; and improved school supervision and inspection. Reform schools are heavily  -90- concentrated in the public sector. No data are readily available on the rural:urban breakdown of reform schools. 10.08 In general for primary education, however, rural areas are much worse off than urban ones. They contain about 80 percent of the population but only 46 percent of primary enrollments, and an even lower 41 percent of teachers. This imbalance is even more pronounced in the public sector, with only 44 percent of enrollments and 39 percent of teachers I rural areas. As a result, the gross rural enrollment rate is well below the urban one, and student:teacher ratios are relatively high, especially in public schools where they reach almost 60:1. Qualified teachers are in short supply everywhere, but again more so in rural areas: only 28 percent of all rural primary teachers--and-only 17 percent of those at rural private schools-have themselves had at least 10 years of basic education, compared to 42 percent overall. Graduates of primary teacher train..ng schools (Normal Schools) are even scarcer, and practically all work in urban areas. 10.09 Teacher quality is thus a major problem for primary education, especially in private and rural schools. It is hard even to get potential teachers to enroll and then remain in the ten Normal Schools; graduates averaged barely 250 a year in FY82-84, only 50 percent of the Government's target. Salaries have been too low to attract appropriate teachers. The average salary of all employees of the MEN was G 7,310 per year in FY84 compared to an average for all public employees of G 14,250. The average private primary teacher earns G 315 per month. Until recently, the monthly income of a trained public primary school teacher was G 500. Neither private nor public teachers are provided any housing, expenses or incentives to teach in rural areas. In order to consolidate the basic education reform, a 1984 Teacher Charter provides in principle for salary scales and career paths for all public education staff, taking into account qualifications and experience. It has so far only been applied to primary teachers. It raises the monthly salary for a trained primary teacher with four years of experience to G 640 (rising to G 840 after 20 years) and cuts the salaries of unqualified teachers back from G 500 to G 400. 10.10 The implementation of the Teacher Charter since FY85 has, however, been flawed by resource constraints, ambiguities in its application, and the lack of-a reliable personnel administration system. There are indications that, rather than act as a mechanism to upgrade teacher quality, the Charter is instead being used in a period of severe resource constraints to minimize the burden of teacher salaries. From FY85 to FY86 there has been no increase in average primary teacher remuneration. Existing staff are not being upgraded (only 2,400 teachers have received an increase and 600 identified as meriting one have not received it because of personnel file anomalies), and almost all newly hired teachers Appear to be inadequately trained and thus paid only the minimum G 400 per month. 10.11 Secondary Education. Secondary enrollment growth has been higher than primary, though starting from a much lower base, at 11 percent a year since FY79, with private schools (12%) again outstripping public ones (5%). Of the 316 establishments, only 26 are public. The seven-year secondary  -91- cycle is broken into a four-year ' wer cycle and a three-year upper one. It takes about 5.5 student years to produce a graduate of the lower cycle and almost nine for the upper one. Student:teacher ratios and students per classrbom are rather higher than in primary schools. Teacher quality is again a major problem, with 43 percent of private and only 18 percent of public secondary teachers having graduated from either teacher training college or university. Relatively low salaries are again a basic cause. Most secondary schools are urban, almost all in the Port-au-Prince metropolitan area. Secondary education--even at the few public schools--is beyond most Haitians' ability to pay. 10.12 Vocational and Technical Education. The lack of centralized information and systematic data collection make it very difficult to obtain a comprehensive view of vocational and technical education. There arz about 500 vocational and technical schools, 20 percent of them public, the bulk in the Port-au-Prince area. Eighty percent of both private and public institutions are girls' home economics centers which account for two-thirds of enrollments. The private sector's growth has not, however, been as rapid as in general secondary education, because of the investments required. Although considerable on-the-job training takes place in industry, the assembly industry requires relatively few trained employees, except for certain specialties like mechanics. Technical and vocational education for agriculture have received very little attention to date; the focus has been on university training to the detriment of middle and lower level agricultural technical training. 10.13 The principal issues in vocational and technical education, mostly interrelated, pertain to structural problems, coordination and standardization of programmes, quality of training, adaptation to manpower requirements and financial practices. Combined, they result in a low efficiency end low quality system which does not respond to labor market requirements and is relatively costly. 10.14 Higher Education. The 4,500 students at the State University of Haiti (UEH) represent 80 percent of all those in higher education. Private institutions are a relatively new, though rapidly growing, sub-sector. About 90 percent of all secondary school graduates pursue higher education, about 60 percent in Haiti and the rest abroad. Tuition at UEH is free, but access is limited by the number of places available. Over a third of all students are enrolled in the Faculties of Law and Economics, Ethnology and Human Sciences where internal efficiency is disastrous: only two percent of their students graduate. The Faculties of Medicine and Pharmacy, Dentistry, Sciences, and Agronomy and Veterinary Science are more efficient, with about 80 percent of all students graduating. Access to private institutions is limited by students' ability to pay. The emigration of trained Haitians considerably reduces the rate of return on public resources allocated to higher education. 10.15 Government Oblectives. The Government has two general long-term objectives for education: (a) universal literacy by 2000; and (b) consolidation of the basic education reform, including full introduction of the Teacher Charter.  --92- In addition, it has two specific short-term objectives: (a) achievement of a gross primary enrollment rate of 83 percent by FY86; and (b) complete introduction of the basic education reform by FY92. 10.16 The 83 percent primary enrollment target has already been met thanks to a rapid growth of private primary enrollments. 10.17 The introduction of the reform is progressing steadily but - accumulated delays mean it cannot be completed before FY93, or one year later than the target. Disagreements among experts concerning the use of Creole in the second primary cycle have slowed the preparation of new curricula and syllabi. Results in reading and writing Creole have been disappointing, primarily because teachers as well as children have had to learn to read and write what had only been a spoken language, but also because of the policy that parents should bear the coat of texts and other teaching materials-only 25 percent of rural children possess the new books, for instarnce. In addition, the use of Creole has generated strong resistance among powerful urban groups who have interpreted it as a banishment of French, seen as essential to a successful career. In fact, however, French as a second language seems to be taught very effectively in reform schools after the first cycle, and resistance is lessening as the impact of the reform materials on students' cognitive levels becomes apparent. 10.18 Grouping children according to ability level has so far proved impossible. Logistical problems have delayed the introduction of automatic grade promotion; new advisors created specifically to implement the reform have been integrated into the regular inspection corps which is, however, still too small and too short on logistical support to function properly. Financial constraints in the public sector have impeded the introduction of the Teacher Charter, the recruitment of qualified teachers, and the upgrading of unqualified teachers. MEN is attempting to introduce the reform into private schools with a combination of parental pressure and specific incentives, teacher training, teaching materials, and scholarships. Since FY85, it has required private schools participating in the school feeding program to introduce the reform. MEN also intends to reintroduce a national examination for grade 7 school-leavers. Past Kxpeuditure 10.19 Total education expenditure--public and private, recurrent and investment--in FY83, the most recent year for which comprehensive data are available, amounted to about G 500 million, or over 6 percent of GDP. About 20 percent came from the Treasury; 30 percent from external aid and from local and foreign charities (official aid from donors plus flows from religious organizations and NGOs); and 50 percent from parents. Fifty-five percent of thxe parental contribution was for primary education, and 41 percent for secondary. Combined public recurrent and investment expenditure on education represented about G 120 million, or 9 percent of total public  -93- expenditure, equivalent to about 1.5 percent of GDP. This is considerably below the 2.8 perceut found in sub-Saharan Africa. However, total education expenditure when the private sector is included is very probably higher in Baiti than in Africa. 10.20 Parental contributions range from 12-100 percent of costs, according to level and type of school. By definition, they amount to 100 percent at private for-profit schools. In the public sector, they average 25 percent at the primary-level, about 34 percent at the secondary, three percent at vocational and technical schools and five percent at U1M. Rural schools run by NGOQ require the smallest parental contributim foz rprimary education of 12 percent. Parental contributions consist of fees, books and materials, uniforms and, except for primary schools, transportation. Fees in primary education range from G 12.50 per year in public schools to G 900 in high quality urban private schools. ?arents trying to enroll their children in public schools are, however, sometimes required to pay more than the standard fee. Books and other materials for primary schools average about G 50 per year. Uniforms add another G 50. Parental contributions to primary education rise through the cycle so that they are three times the grade I level by grade 3, a key explanation of the dropping out of poorer children in the higher primary grades. 10.21 Parental contributions are a major k rden on disposable household incomes. It is nonetheless striking that Haitian parents are very willing to pay for their children's education. Willingness is not, however, the same as the ability to pay. The rural and urban poor with household incomes below G 2,500 who make up two-thirds of the population are barely able to afford even the cost of enrollment in private schools run by NGOs; many of these require low fees. The cost of parental contributions makes all education above the primary level completely out of the question for the bulk of the population. 10.22 The bulk of public expenditures is financed out of general revenues and external aid. Such cost recovery as exists in the system is very limited and extremely inefficient e.g. a payroll tax of one percent that is supposed to be used for technical and vocational education; and fees at public schools, that do not appear to end up in the general revenues of the Governmer.t. 10.23 Public recurrent expenditure has increased at about four percent per year in real terms since FY79, or roughly the same rate as public sector enrollments. Table X.2 shows that 62 percent of public recurrent expenditure is for primary education. Almost 90 percent of recurrent spending is on personnel, of which about ten percent is estimated to be paid to nonexistent "phantom" employees, including support staff, administrators and teachers. Only about ten percent of the entire recurrent budget is therefore left for non-salary operating funds. About a third of this ten percent consists of the central administration and adult literacy programs, anrd a third of the remaining seven percent is used for rental costs of buildings and other transfers to the private sector. This leaves less than G 5 million to cover materials at all levels of the public education system, or about G 12 per student. The effect of this chronic underfunding of operating cos4s is evident throughout public educational institutions: inadequate or nonexistent school maintenance, immobile and hence inefcective rural school inspectors, etc. Basic material needs like books and supplies  -94- are, in fact, met by contributions from parents or from external aid donors (in the latter case, hidden within the investment budget). 10.24 Ninety percent of public recurrent expenditure on education is accounted for by the MEN budget. MEN overspent Its recurrent budget by G 4.2 million in FY85, largely as a result of implementing the Teacher Charter without making adequate provision for its financing. Actual overspending for primary education was G 5.5 million, or twice a G 2.6 million supplemental budget allocation that was made to accommodate the Charter. A substantial saving was made, however, on non-salary operating expenditures which were underspent by ten percent. A familiar pattern was thus repeated: excess spending on salaries has a negative impact on the funding of operating expenses with often disastrous consequences for the effectiveness of the public education system. Table K.2: PUBLIC RECURRENT EXPENDITURE ON EDUCATION BY LEVEL, FY85 (percent) Distribution Share of Personnel Central administration 6 68 Primary a/ 69 94 Secondary 10 96 Vocational/Technical 9 81 Higher (UEH) 13 90 Adult literacy 5 92 Total 100 89 Expenditure (G million) 106 95 a/ Including teacher training and inspection. Source: Statistical Appendix Table 10.11. 10.25 Public investment expenditure has been 80 percent-financed by external aid. The major donor has been IDA, which has typically financed over half the investment budget. Other donors have included CIDA, FAC, UNDP, UNESCO and other UN agencies. 10.26 Public Invesetment has declined steadily in real terms since FY79. During FY82-84, tile overall plan called for expenditures of G 290 million, budgets amounted to G 168 million (58 percent), and actual spending was only G 105 million (36 percent). Worse still, over 37 percent of investment expenditure was, in fact, recurrent expenditure listed under two budget lines: institutional support (Programme de Renforceent Institutionnel) and unplanned (Programme Hors-Plan)* The latter in particular was used to channel current transfers to UEH. Chronic underfunding of the recurrent budget has led to the use of the Treasury component of the investment budget to meet recurrent costs, in turn reducing that available as counterpart to external aid for genuine investment. As a result, physical investment was only 57 percent of the total.  -95- 10.27 The underfunding of investment and misallocation of recurrent outlays is at the root of the steady decline of public relative to private education. Very little real itwvestment has been completed: only 377 primary classrooms were built in FY82-83 compared to a targeted 1,000; none of the 400 existing primary classrooms have been provided planned equipment; 200 planned housing units for primary school staff have not been constructed; and no planned vocational and technical construction took place. This underfunding of the investment budget in turn results from the underfunding of the recurrent-budget. Expenditure Program and Fiusnciug Plan, FY"6-89 10.28 There is no public expenditure program as such for the next few years. However, a broad picture can be drawn from an examination of the Government's objectives for education, the recurrent and Investment budgets for FY86, Government ideas for new projects, and the pipeline of donor-financed projects. 10.29 The FY86 recurrent budget for MEN represents a 4.8 percent nominal increase over actual FY85 expenditure i.e., z.o increase in real terms. Salaries are to increase by four percent overall, and by five percent for primary education, including both teachers and support staff. The average percentage salary increase for primary teachers will, however, be lower than that for secondary teachers or administrative professioaals. This confirms that the Teacher Charter is not being used to upgrade the quality of primary teachers but rather all new hiring is of unqualified and hence lower-paid staff. 10.30 In one important regard, however, the FY86 budget represents a significant improvement over the past. MEN, like all ministries, has been plagued with excess unproductive and even phantom employees, especially among the central administrative and support staff. Support staff have been cat by 30 percent in FY86, while the total staff has risen by only seven percent. There appears to have been a real attempt to prune the payroll, and to use reductions in support staff to upgrade the professional staff, both teachers and administrators, although particularly the latter. In addition, non-salary personnel expenses (allowances and representation) have been cut by over a third, and the budget for external personnel services has been virtually abolished. 10.31 The FY86 investment budget continues the trend of FY85 spending (Table X.3). There has been a further dismantling of the institutional support program, part of which is now integrated into the recurrent budget, and a greater emphasis on construction and equipment through a reduced number of projects. Against this, however, is the continued failure to understand the implications for external aid disbursements of arbitrary cuts in local counterpart funding; the cuts from those originally planned imply that a 64 percent reduction in Treasury expenditure from G 14 to G 5 million would provoke only a 34 percn-t decrease in external aid disbursements from G 62 to G 41 million; this is simply not realistic.  -96- Table X.3: PUBLIC EDUCATION INVESENT AND FINANCING, FY86-91 (G million at curtent prices) FY86 FY87-91 Applications 72 264 MEN 57 193 Primary and Secondary 46 144 Vocational and Technical 7 30 Non-formal 4 19 Higher 15 70 UEH 10 Xi Other 5 29 Sources 72 264 Treasury 9 18 External aid 63 246 Sources: Statistical Appendix Tables 10.17 and 10.18. 10.32 Table X.3 also indicates likely investment in FY87-91, the period of the next five-year plan. Some G 264 million is to be invested by the public sector, 93 percent of it financed by external aid. In addition, external donors intend to finance 86 per_ent of identified private sector education investment of G 81 million. The bulk of this would be USAID support to private primary and secondary education. USAID and IDB would, together with IDA, account for over 70 percent of aid to MEN and over 60 percent of that to the sector as a whole. All three projects would concentrate on basic, particularly primary, education. The FY87-91 program involves a major shift away from construction and equipment provision toward support activities requiring relatively less counterpart funding (technical assistance, training, logistical and institutional support, etc.). Thus total Treasury funds required for the entire five years would be only twice those budgeted for the one year of FY86. 10.33 On the basis of a performance contract, the USAID project will provide a portion of existing primary schools in rural and depressed urban areas with a package of pedagogical and material resources over a continuous 4-5 year period. The IDB project will finance the construction of public primary schools, the provision of free textbooks for public and private reform school grades 1-4, extensions of Normal Schools for teacher training, and in-service teacher training for the first primary cycle. IDA would assist the Government to consolidate the reform, finance part of the Teacher Charter, and provide logistical support to the school inspector corps. CIDA and FAC are also likely to remain heavily involved in the sector, the former concentrating on higher and vocational and technical education, and the latter 50 percent on basic and 25 percent each on higher and on vocational and technical education.  -97- 10.34 The predominant role of the private sector means that public expenditure must be considered in the context of the education sector as a whole. At all levels, the education system is characterized by low quality instruction and inefficiency, reflecting run-down facilities, a shortage of qualified teachers and the low status of the profession, outmoded curricula and teaching methods, and a lack of necessary materials. Despite this, analysis--on the basis of a computer model developed to project recurrent expenditure on primary education--shows that the objectives of completing the reform and achieving access to universal enrollment (as opposed to universal literacy) by 2000 can be met, provided fundamental policy changes are introduced. 10.35 The major policy modifications necessary are: (a) a slower expansion of both public and private sectors than in the recent past; (b) cost recovery at UEH, the public university, and vocational and technical schools; (c) efficiency and quality improvements throughout the system, especially in the public sector; (d) reduced parental contributions for primary education at both public and private schools; (e) an increased recurrent budget for education, concentrating on the primary level; and (f) an investment budget within the sector's absorptive capacity, again concentrating on primary education. 10.36 Slower expansion. Universal literacy (the Government's objective) could eventually be a^hieved by providing universal access to basic education for all six year old children by the year 2000. This would require a relatively modest three percent per year expansion of new primary grade intake, well below the ten percent average for 1978-82. The same three percent rate could be applied to both private and public sectors because expansion in the private sector is expected to slow down. Many households already spend the maximum they can afford on education, NGOs are unlikely to able to increase the external resources they receive, and necessary quality improvements will result in increased costs. The public sector then would have to supply the remainder of the new student places required to reach the objective of universal access to basic education. The costs of this public sector expansion seem feasible. 10.37 The Government should rethink policies regarding teacher recruitment, qualifications and remuneration. At the present pay scale under the Teacher Charter, fully subsidized Normal Schools cannot attract qualified students, and graduates leave the profession as soon as possible. Although increasing the pay scale might attract people into the profession and the subsidization of Normal Schools could be eliminated, a better option might be to reevaluate the level of teacher qualifications needed for grades  -98- I to 4. By lowering the qualifications required, subsidies for pre-service training could be avoided, and present pay scales would be adequate to maintain teachers in the profeasion. 10.38 Public secondary, vocational and technical enrollments and facilities should not be expanded through FY89, although quality and efficiency improvements including increased teacher salaries and non-salary operating funds are essential. If expansion does occur, it cannot be financed by the fisc and will have to be met entirely from fees; this essentially limits secondary expansion to the private sector. High emigration rates and the low current demand for skilled labor make the social returns to these education facilities very low. This is not to say that certain specific skill training may not be called for, and a comprehensive study of vocational and technical education needs is necessary to identify them. In addition, the Government may wish to consider divestiture of all or some public vocational and technical training establishments, probably excluding those providing agricultural training. By the 1990s, some expansion of secondary education will undoubtedly be required if there is sustained economic growth and to absorb some of those then graduating from primary school. 10.39 UEH should not expand. Indeed, there is a strong case that those faculties which are inefficient and do not contribute to national development should be closed or divested to the private sector. The medical school class could well be halved (Chapter XI). The Faculty of Agronomy and Veterinary Science should concentrate more on secondary level training at the agricultural schools it runs than on higher level education. 10.40 Cost recovery at UEH. UEH represents 0.5 percent of all sector eurollments but 13 percent of the recurrent budget. Cost recovery through fees should be introduced in all faculties. Needy students enrolling in faculties contributing to development could be subsidized through scholarships. The introduction of full cost recovery-with some provision for scholarships and some recurrent funding for the Faculty of Agronomy and Veterinary Science conditioned on its use for its secondary agricultural training schools-should reduce the university's recurrent budget from G 12 to G 5 million in FY86 prices, split among the UEH autonomous budget (20 percent), MARNDR (40 percent) and MSPP (40 percent). Cost recovery could also be introduced at vocational technical training institutes. 10.41 Improved efficiency and quality in education could be assisted by: (a) ensuring that all classrooms have teachers, even if unqualified, and that these teachers have orientation and training; and assigning non-teaching teachers to other posts, so that the full complement of tea_hing posts is filled with teaching staff (this could be facilitated by the computerization of salary and personnel records currently underway); (b) hiring of all graduates of existing teacher training schools by the public sector, minimizing the recruitment of unqualified teachers;  -99- (c) reviewing teacher qualifications before expanding the Normal Schools; no expansion would probably then be necessary, although one of the exising public ones badly needs rehabilitation; (d) upgrading 350 unqualified public primary teachers per year through FY91 with in-service training; (e) gradually reducing average classes in rural public primary schools from the present 52 to 45 students by FY96; (f) the increasing use of double shifts in urban primary classrooms, with separate teachers for each shift; (g) full application of the Teacher Charter; (h) a financial bonus for experienced teachers whose lack of formal qualifications excludes them from increases under the Charter but who are satisfactorily teaching reform classes (i) another bonus for teachers in remote rural areas; (j) the provision of free or heavily subsidized textbooks and teacher guides for all primary first cycle students and teachers in both public and private reform schools; (k) an increased allocation for non-salary operating funds; (1) a strengthened iuspector corps with adequate non-salary operating funds to inspect all schools, including especially private rural ones; and (a) the introduction as soon as possible of national testing to monitor quality Improvements and to signal parents which private schools are the more efficient. 10.42 Reduced parental contribution for primary education. Fees at public primary schools should not be changed. The financial burden on parents would be alleviated, however, by the proposed free provision of primary books. It could also be eased, where appropriate, by abandoning compulsory uniforms. 10.43 Increased recurrent budget. The efficiency and cost recovery policies advocated above would still not be enough to finance all expenditure necessary to move toward access to universal basic education by 2000. An increased recurrent budget is therefore necessary (Table 1.4). Through FY96, there would be an average annual increase of 5.6 percent, but heavily concentrated in the 1990s; the recurrent budget would grow only by 2.8 percent through FY89. Primary education's share in the total would rise from 69 percent in FY86 to 75 percent in FY89. The share of personnel expenses within total primary expenditure would fall from 87 to 69 percent by FY96 as more adequate operating funds were provided, but only to 85 percent by FY89. This is less than desirable but unavoidable if teachers' salaries are also to be raised. A more rapid shift toward non-salary  -100- operating funds would only be possible if donors were willing to finance other recurrent funds for the next few years. Table X.4: RECOMMENDED EDUCATION RECURRENT EXPENDITURE, FY86-96 (G million at FY86 prices) Year MEN UEH Total Primary (% personnel) Other Total FY86 79 (87%) 24 102 12 115 FY87 86 (87%) 24 110 9 119 FY88 88 (88%) 26 114 5 119 FY89 94- (85%) 26 120 5 125 FY96 159 (69%) 35 194 5 199 Source: Statistical Appendix Table 10.20. 10.44 About 30 percent of the incremental recurrent budget in FY86-96 would be for recurrent expenditures now classified in the investment budget, i-ncluding scholarships, in-service teacher training, transport for school inspectors, and textbooks. Except for scholarships, these items are all financed by external aid donors. One donor, IDA, would also finance about 14 percent of the FY86-91 incremental teacher salaries generated by the Teacher Charter. This external financing amounts to only about G 6 million per year in FY86-89. Of the total incremental FY86-96 recurrent expenditure, about 30 percent would be for primary teachers and school directors' salaries, 20 percent for school books and teachers' guides, 20 percent for maintenance, and 10 percent for in-service teacher training. 10.45 The vocatioi!al and technical tr4tning subsector requires further review, including of cost recovery mechanisms. Even without such a review, however, it is clear that more public agricultural technical training is necessary and that the use of puBlic resources for home economics courses is questionable. 10.46 Investment expenditure needs to be reduced to an operationally sustainable level commensurate with meeting sector objectives. This can be done because of the proposed use of double shifts in urban primary classrooms. Expenditure could be about G 125 million over FY87-89, 97 percent financed with existing donor commitments, on an annual basis less than 60 percent of the FY86 program (Table X.5). Construction should amount to about G 44 million and support programs to some G 81 million. The foreign exchange component has not been determined but will be much less than the 93 percent of the FY86-89 program to be financed by donors.  -101- Table X.5: RECOMMENDED EDUCATION INVESTMENT PROGRAM, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 Budget Recommended Applications 72 125 197 MEN 57 98 155 Primary and Secondary 46 78 124 Vo-cational and Technical 7 20 27 Non-formal 4 4 Higher 15 27 42 UEH -R -t Other 5 - 5 Sources 72 125 197 Treasury 9 4 13 Existing external aid 63 121 184 Sources: Table X.4; Statistical Appendix Table 10.22. 10.47 Two projects included in the Governmentts list of likely projects should not proceed. There is no need to construct a new teacher training college at the higher education level as appropriate lodging has been found. Support to the Institut Roi Henri Christophe, a private higher education institution, should be examined in the context of a policy for the possible subsidization of selected private sector institutions in lieu of expanding public ones. Such a policy should be developed first. 10.48 To avoid overlap and inconsistencies with other donors and sector policies, the Government could consider reshaping the IDB-financed project. No new expansion of Normal Schools for primary teacher training is needed until the existing ones are fully utilized and the teacher qualification question is resolved; this project component could therefore be delayed until the 1990s. Funds so released might be reallocated to (1) the rehabilitation and construction of rural primary schools beyond the 300 already to be financed under the existing IDA and IDB projects; and (2) the possible reconstruction of the existing Normal School at Damien which is in appalling condition. The computer model projection probably underestimates the number of new rural public primary schools needed, as it assemes that many rural children will continue to walk many hours a day to urban schools. In addition, some 20 percent of existing primary schools are in poor condition and need rehabilitation. Some minor reshaping of the IDA project is also necessary to avoid overlap e.g. IDA can increase its management restruc-turing component and eliminate training of grade 1-4 teachers, the latter now being financed by IDB.  -102- 10.49 No new aid commitments are necessary in PY86-89 unless donors are willing to finance a greater portion of recurrent costs* Disbursements from existing commitments of G 184 million (US$37 million) at FY86 prices will be adequate, provided the above reshaping occurs. 10.50 Some G 57 million of proposed donor-financed projects for FY87-91 has been excluded from the recommended investment program, mainly higher (G 23 million) and non-formal (G 33 million) education. This is primarily aid from CIDA and FAC. The Government should influence donors to finance only development-oriented higher education within the context of increased privatization. The funds for literacy could remain, as long as there are no counterpart requirements. These are low priority areas and should be left to the private sector. The Government cannot afford to use its own resources for them. 10.51 The recommended recurrent and investment expenditure programs and financing are combined in Table X.6. To properly prepare and monitor these and future expenditures, MEN urgently needs to strengthen its Planning Department and improve its monitoring processes. It must also improve its salary administration and, as noted, set up a better coordination mechanism among the donors and the ministry. Table X.6: RECOMMENDED EDUCATION EXPENDITURE AIND FINANCING, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-89 APPLICATIONS 186 488 674 Recurrent 114 363 477 MEN 102 344 446 UBH 12 19 31 Investment 72 125 197 MEN 57 98 155 UEH 10 27 37 Other 5 - 5 SOURCES 186 488 674 Treasury 117 361 484 Existing external aid 69 127 190 Sources: Tables X. 4 and X.5. 10.52 Greatly improved coordination is needed among the Government and donors financing basic education. Already, overlaps and inconsistencies have become apparent among the IDA, IDB and USAID projects; as noted, some reshaping of existing projects is necessary.  -103- CHAP= xi hALTS- The Sector and ite ?lanning 11.01 HaitiIs 5.3 million people suffer from poor health and nutritional conditions, more like other low income countries than Haiti's Caribbean neighbors. Life expectancy at birth is only 53 years; the infant mortality rate was 120 per 1,000 live births in 1984, and over 25 percent of children suffer from second or third degree malnutrition. The crude birth (36 per 1,000) and total fertility rates (4.6 births per woman) are comparable to those in African countries at similar income levels. Over one quarter of Haitian mothers did not breastfeed their infants during the most recent birth Interval. Ninety percent of deaths among 1-4 1ear old children are associated with malnutrition and diarrheal diseases*_/ 11.02 Haiti has 516 health facilities, 90 of them hospitals and health centers with beds, the majority public. The occupancy rate, especially In the public facilities, is typically belcw 50 percent. The Ministry of Public Health and Population (MSPP) employs about 810 physicians, over half in Port-au-Prince. Most private physicians are also in Port-au-Prince. Outside the capital, there are 1.4 physicians per 10,000 people. Over 20 percent of all public employees work for MSPP, which is the second largest employer in the country after the Ministry of Education. 11.03 Over 200 private health establishments, many run by externally- funded voluntary organizations, operate almost half the health facilities. Most are located in isolated rural areas, many serving the poor and needy. There is fairly effective cooperation between them and the Government, especially since the organization o; the Association of Private Health Institutions (AOPS) in 1982. Some 28 AOPS affiliates will offer primary health care to about 500,000 rural people in the next few years, consistent with NSPP policies antd norms and with MSPP paying the salaries of some employees. 11.04 Population Status and Policies. The major demographic problem facing Haiti has been clear since the early part of this century: the relentless pressure of population on limited natural resources. This is likely to be exacerbated in the future for two reasons. First, improved rural primary health care delivery can dramatically reduce infant and child mortality very quickly, leading to more rapid population growth. Second, the distribution of modern methods of contraception has not expanded rapidly enough to meet potential demand, especially in rural areas. 11.05 Today, the population is growing at about 1.8 percent per annum. Port-au-Prince, with an official population of 750,000 and perhaps as many 1/ In addition to field work during the mission, this chapter draws on previous work summarized in World Bank Report No. 5699-CRG, Haiti: Situation Note on the Population, Health, and Nutrition Sectors, May  -104- as 1,000,000 people, is at least ten times larger than Cap Haitien, the second city. Over 15 percent of persons-born Haitian, perhaps one million people, now live abroad. The total population in Haiti will grow to 7 million by 2000 and perhaps 13 million by 2050.2/ With more rapid fertility and mortality decline, the population in 2050 could be held to 10 million. The population-resource balance in rural Haiti will grow increasingly unfavorable unless effective measures are taken to slow population growth. There is widespread knowledge of moderri contraceptive techniques, but less than 5 percent of married woman use contraceptives, according to recent surveys, though official figures show higher rates around 11 percent. Be it 5 or 11 percent, however, there is a wide gap between need and practice. 11.06 The outline of a population policy is contained in the FY81-86 five-year plan and is repeated in more recent annual plans. MSPP's "New Orientation", a health policy drawn up in 1982, sets as goals for the year 2000 a crude birth rate of 20 per 1,000; a fertility rate of 3 children per woman; a crude death rate of 8 per 1,000; an infant mortality rate of 50 per 1,000; and a life expectancy of 65 years. MSPP recently established a Directorate of Population to work with other key miaistries, especially MARNDR and MP, concerned with population dynamics and nutrition. 11.07 Health Status and Policies. The death rate has fallen from above 16 per 1,000 in the early 1970s to about 13 today. Only a fraction of people who need medical treatment receive it. Preschool children are sick about half the time, 80 percent with diarrhea. Malaria is the most frequently reported cause of adult morbidity, followed by pneumonia, bronchitis, diarrheas, and malnutrition. Typhoid, syphilis, intestinal parasites, and respiratory tuberculosis are diagnosed about half as often as malnutrition. Other common causes of ill health are anemia, infections of the ear, eye or skin, and complications of childbirth. 11.08 Service statistics calculated from incomplete data give a picture of very inadequate health coverage. Only 40 percent of women eligible for prenatal services receive them. Family planning service utilization is constrained by an inadequate supply system. Pediatric services are provided to only one quarter of the 0-5 age group. Vaccinations only protect a fraction of the population under 5 years old: five percent are immunized against diphtheria, pertussis and tetanus, three percent against polio, two percent against tuberculosis, and 0.1 percent against measles. In 1978, the average number of patient-contacts in a MSPP dispensary was nine per day, compared to 26 in private dispensaries. The dispensaries in the Petit Gofve region dealt in 1979 with an average of 46 clients per day: those in the Jacmel rogion with only five. An informal survey by USAID staff in 1985 found similar low levels of facility use. In the North region in 1979, there were 0.69 patient visits per person; in Jacmel there were 0.10. Much greater efficiencies of operation are possible just by bringing laggard regional programs up to the standard of the better-run subsystems. 11.09 The problem of inefficiency is not confined to a shortage of services; the extent of underutilization of available services is extraordinary. The bed occupancy rate of 36.5 is due not to a 2/ World Bank, World Development Report, 1984, p. 192.  -105- shortage of clients, who stand in line for hours waiting for attention, but rather to a shortage of staff, equipment and supplies. Most physicians are not prepared to work in ill-equipped hospitals in a deplorable state of repair (most were built during the U.S. Occupation between 1915 and 1930). A few dedicated staff persevere in spite of difficult conditions, but their ability to carry out necessary medical procedures is severely hampered by the absence of functioning equipment and shortages of supplies, and by quite unsanitary surroundings. Under these circumstances, staff morale is often low, and supervision is infrequent. People who need care have learned that they may walk miles to a clinic, only to find it closed. If the staff are present, they typically do not have the drugs which their patients need, so patients face a longer journey to a commercial pharmacy in town, to buy drugs at a price they cannot afford. The private dispensaries tend to be better supplied and equipped, and this is reflected in the service statistics. 11.10 The New Orientation emphasizes primary health care and identifies six priorities for MSPP action: (1) diarrheal disease control; (2) immunization against communicable diseases; (3) tuberculosis control; (4) improvement of nutritional status; (5) maternal/child care and family planning; and (6) malaria and endemic disease control. Each priority will receive special attention for a specific period. The first, the diarrheal disease control program, launched nationwide in July 1983, led to widespread knowledge of oral rehydration, considerable use of commercially available packets of oral rehydration salts, and a reduction in diarrhea cases presented for treatment at health facilities. The second priority, the expanded program of immunization (EPI), will be the focus of special attention for two years from October 1985. Again, the need is great; as noted, only five percent of children under the age of five are vaccinated against diphtheria, pertussis and tetanus, for exainple. 11.11 Public policy in favor of primary health care was not adequately reflected in MSPP's recurrent budget before FY86. Forty percent of MSPP staff positions were assigned to hospitals and medical schools, and over half of all public sector health workers to the Port-au-Prince metropolitan area. Including administration, almost sixty percent of the FY85 recurrent budget was assigned to spending in Port-au-Prince which contains only about 20 percent of the population. 11.12 Nutrition Status and Policies. Malnutrition is one of Haiti's most serious problems: about 30 percent of rural children and 48 percent of urban ones are anemic. Between 24 and 57 percent of children aged 0-6 had second- or third-degree malnutrition by Gomez' scale of weight for age in four sample surveys during the 1960s and 1970s. More recent data from the national Nutrition Survey indicate that only one-quarter of children-under 5 years old are in the normal range (90 percent of the reference median) of weigbt for age. Rural children of weaning age (18-24 months) are frequently not fed enough to keep them from being undernourished; they are not given special attention, and do not compete successfully at the communal bowl. An episode of illness (e.g., diarrhea) often proves fatal because the child's resistance is lowered by undernourishment, and feeding practices hinder recovery (e.g., withholding liquids, or not persevering to persuade a marasmic child to accept food). Feeding practices embodied in the use of a  -106- bottle instead of breast-feeding, pose an increasing threat to the health of children in the pre-weaning age-group as more and more rural people move to Port-au-Prince and adopt this urban practice. 11.13 The causes of malnutrition include poverty and food shortages, droughts and floods which destroy crops, erosion of arable land, some traditional practices of feeding children (for example, one meal per day), and some "modern' infant feeding practices. The use of bottle-feeding instead of breastfeeding has been increasing throughout Haiti; 96 percent of urban mothers and 59 percent of rural mothers had used a bottle at some time to feed their youngest child. This is a staggeringly high figure when it is recalled that only two percent of rural Haitians have access to safe water; that fuel (for boiling water) is scarce; and that adequate infant formula is too expensive for most Haitian families. In rural areas, weaning age (18 to 20 months) is much higher than in the cities, where half as many mothers now breastfeed during the child's second year as in the rural areas. The strong association between bottle feeding, frequent diarrhea and severe malnutrition has been documented. 11.14 The fundamental constraint to improving nutritional status is the shortage of food. Families do not waste the food they have; a long-term solution to Haiti's problem of undernourishment must involve overall income and productivity growth. An additional problem is dealing with the significant amounts of food aid Haiti receives. The management problems are enormous and improvements in nutritional status is not always evident. 11.15 A 1975 survey found prevalence of blindness due to Vitamin A deficiency was 8.1/1,000 in the North and 1.2/1,000 in the South. The rate in the North was alarmingly high, compared with the WHO standard. In response to these findings, a nationwide program to distribute Vitamin A capsules to ill and malnourished children and nursing mothers began in 1976. A 1979 survey evaluated this program and found the prevalence rate for children under seven with keratomalacia had fallen to 0.9 in the North, i.e., one-tenth of the 1975 rate. The lower prevalence was not entirely due to the program, since the North was experiencing a famine in 1975. However, effective monitoring and timely intervention, as in the Vitamin A program, can prevent some severe consequences of nutritional deficiencis. 11.16 Closely targeted intervention can be effective. In Cite Simone, a Port-au-Prince slum typical of low income urban areas, three percent of children are severely malnourished. However, this small, high risk group consituted 10 percent of all patients at an oral rehydration center, 35 percent of all hospital admissions for children under five, and a staggering 65 percent of all deaths among the under-five population. The risk of death for this small group of severely malnourished children is 25 times higher than for the much larger group with only second degree malnutrition.3/ These findings suggest that targeted nutrition interventions should be introduced at an acceptable cost as a component of maternal and child health services. 3/ Reginald Bouloe et al., 'Cite Simone, an urban slum of Haiti, the malnutrition challenge," mimeo, Cite Simone Medical Complex, p. 7.  -107- 11.17 There is no one mlnistry with responsibility for nutrition-those involved include MSPP, MARNDR and MP-and there is no Government nutrition policy. Immunization, growth monitoring, oral rehydration, family planning, and the promotion of breastfeeding are central components of the primary health care system in the pilot programs developed so far. They offer some hope of ameliorating nutrition problems and represent an advance over earlier programs that were prohibitively expensive per child served. Past Expenditure 11.18 Given Haiti's low income level and poor health and nutrition status, total public and private expenditure (both recurrent and investment) has been surprisingly high at G 81-116 (US$16-23) per capita in the 1980. (Table XI.1). No other low income country among ten surveyed by the World Bank spends as wmch; Sri Lanka is closest but enjoys a substantially better health status.4/ The high Haitian level is probably due principally to inefficiency but also reflects Haiti's location close to high income North America, which pulls up the wages of internationally mobile professionals. Table XI.1: TOTAL HEALTH EXPENDITURE, FY81-85 dmillion at current prices) Source FY81 FY83 FY85 Public expenditure 160 215 230 NGOs 30 80 120 Direct payments to private providers 215 25 265 Total 405 550 615 Total per capita 81 108 116 Source: Statistical Appendix Table 11.3. 11.19 Individuals pay directly to private providers for about 40 percent of their health care. Private voluntary organizations provide about 20 percent of sector resources in cooperation with the Government and external aid donors. Health, like other sectors, has recently seen an increase in aid being channelled through such Non-Governmental Organizations (NGOs); the programs of 71 NGOs covered by a Government survey tripled between FY83-85. 11.20 The remaining 40 percent of sector resources comes from public expenditure, split roughly equally between internal and external financing. "Internal" financing, however, is itself about 50 percent financed with U.S. PL-480 counterpart funds. Thus the Treasury provides only about one quarter of total public expenditure on health. Real public health spending has declined during the 19809 from its FY80 level of G 51 per capita (in FY86 prices) to G 43 in FY85. See David de Perranti, Paying for Health Services in Developin; Countries: An Overview, World Bank Staff Working Paper 721, 1985, p. 99.  -108- 11421 Most private health spending is for curative services; so, unfortunately, is the bulk of public spending. Indeed probably less than five percent of total health spending is for prevention. This low share is the major cause of sector inefficiency.- 11.22 The principal external donor is USAID, through both its PL-480 and direct aid; UN agencies, IDB, Japan, the Federal Republic of Germany and France play a secondary role. These donors have helped to strengthen MSPP services, supported the malaria program, and aided the maternal and child health program. Total external assistance in FY85 was over G 100 million when PL-480 funds are included, up from about G 70 million in FY80. 11.23 Total public health expenditure represented eight percent of Central Government expenditure in FY85. This is far more than the three percent average for all low income economies, and even the five percent average for middle income ones. It is so high largely because donors give such a high priority to these services. 11.24 It is essential to analyze public expenditure on health in terms of both the recurrent and the ilvestment budget together for two reasons. First, the btlk of all expenditure, whichever budget it is in, is for services and thus is properly recurrent, not investment spending. Expenditures on personnel and materials by MSPP's Division of Family lygiene and Nutrition (DHFN), its National Service Against Endemic Diseases (SNEM), and other projects are counted as part of the investment budget even though they are clearly recurrent items. Second, new investments create the need for future recurrent costs. 11.25 Over 85 percent of the formal recurrent budget is used for wages and salaries, thus limiting complementary operating funds for services such as drugs, medical tests, essential transport arnd maintenance. A 1982 report estimated that recurrent spending would have to grow dramatically to operate rural health delivery, malaria control, health center maintenance, and maternal and child care projects then under implementation: by 22 percen in FY83, by 15 percent in FY84, and by 10 percent in each of FY85 and FY86._f In the event, recurrent expenditure financed by the recurrent budget fell by 25 percent in real terms between FY82-85 and is budgeted to fall further in FY86 (Table Xl.2). As a result, the Government has not been able to assume responsibility for the recurrent costs of any of the projects which came on stream in the early 1980.. To pay for them fully, it would have had to raise its recurrent budget by nearly 50 percent in FY86. 11.26 The gap between projected necessary recurrent spending and actual outlays has been partially made up in three ways. First, some donors have continued support for the malaria program and for some7salaries and salary supplements in, for example, the maternal and child health program. Second, there have been substantial cuts in real salaries among all MSPP staff and 5/ Peter Cross, "Financial Analysis and Financial Projections prepared for DSPP," mimeo, Management Sciences for Health, 1982. Progress in implementing the report was reviewed by C. Leighton, 'Issues and Recommeidations for Addressing the Recurrent Cost Problems of the Miuistry of Health, Government of Haiti," USAID and Birch & Davis Associates, 1985.  (.. -o109- especially those who no longer receive supplements. Third, some health activities have been shifted to private sector NGOs. Despite these factors, ..he quality of service delivered has undoubtedly been Impaired, especially rural primary health care and family planning, as necessary non-salary operating resources like transport and drugs have been cut very severely. Table XT.2: PUBLIC HEALTH-EXPENDITURE, FY80-85, AND BUDGEs,, FY86 (G million at FY86 prices) Year Recurrent Budget Investment Budget Total FY80 100 93 193 FY81 95 67 162 FY82 131 107 238 FY83 105 109 214 FY84 102 100 202 FY85 97 134 231 FY86 Budget 90 116 206 Source: Statistical Appendix Table 11.6. 11.27 Partially as a result of the first factor above, and also reflecting donors' priority for the health sector, the investment bidget has remained roughly constant around G 100 million during the 1980s. It is, however, budgeted to fall by some G 18 million from FY85 to FY86. The investment budget depends critically on external donors for its size, content and project mix. ftpenditure Program and Financing Plan, 1Y86-89 11.28 MSPP's recurrent budget for FY86 takes a bold step by cutting about 900 staff positions in the Port-au-Prince metropolitar. area and increasing those in the regions by about 1,000. It must be noted, however, that these are budgeted staff positions; actual staff numbers may well be different (Chapter II). 11.29 MSPP investment expenditures for FY86 were originally budgeted at G 116 million but later reduced to G 102 million. However, even this level is inconsistent with anticipated donor disbursements. USAID anticipates disbursement difficulties with the rural health delivery system project that is budgeted to absorb about 55 percent of PL-480 funds; UNFPA is suffering general budget constraints that may well lead to cuts; and IDB plans for hospital and health center construction and re-equipment are delayed at least until FY87. Rather than an expenditure of G 102 million, of which about G 80 million would have come from donora when PL-480 is included, a more likely level would be G 70 million, of which perhaps G 55 million would be donor-financed. This assumes that even Treasury disbursements will be well below the G 22 million budgeted, largely because of the shortfall of donor assistance. The total likely expenditure of G 70 million would be lower in real terms than in any year since FY81 (Table XI.2).  -1 10- 11.30 Regionalization and strengtheuing rural services have both been supported by USAID and IDB with the objective of extending low cost primary health care to rural Haiti. There has been progress wlth oral rehydration therapy, growth monitoring, and some distribution of family planning services. However, the training of community volunteers has fallen short of objectives; donors have essentially suspended disbursements for FY86 until thls and other problems are resolved. In addition, construction of new facilities to have been financed by IDE has been delayed until MSPP and IDB can work out a plan which will. avoid massive recurrent cost implications for the Governmeut. Table XI.3: MSPP: SOURCES AND USES OF INVESTMENT FUNDS, FY86, BUDGET AND LIKELY (G million at FY86 prices) Budget Llkely Applications 102 70 Regionalize services 25 13 Strengthen rural services 10 3 NCR/Family planning 25 16 Nutrition 1 1 Endemic disease control (malaria) 34 34 Other 6 3 Sources 1Wa 70 Treasury 22 15 Donors 80 55 USAID PL-480 18 17 USAID direct 19 17 -DB 21 4 Others 22 17 Sources: Statistical Appendix Table 11.7; mission estimates. 11.31 Maternal and child care sad family planning operated from the mid-1970s through 1983 as a vertical program supported by USAID and UNFPA. However, its effectiveness has waned since it began to be integrated into MSPP's regular programs and there has been a decline in family planning service delivery. 11.32 The largest program and major beneficiary of external aid has beer and continues to be the malaria eradication program. 11.33 The largest item in the budget to be financed entirely from Treasury resources is a maternity hospital to be built in Port-au-Prince. Budgeted at G 10 million, this would require 10 percent of the investment budget and use nearly half the Treasury allocation of G 22 million to the health sector.  -1}1- 11.34 Because of the country's extreme poverty, rapid population growth, poor health and undernutrition characterize the current situation. A major policy option is to shift the balance between budgets for population, health and nutrition. In 1982, only about 16 percent of the MSPP aggregated operating and development budgets were devoted to population programs broadly conceived; the share was even less in 1985. The nutrition program accounts for only one to three percent of the MSPP budget. Population and nutrition programs are significantly underfunded and should be expanded relative to health activities. 11.35 - To achieve a greater concentration of resources on population and nutrition programs it will be necessary to achieve greater efficiencies in the use of funds committed to health programs, both in operating expenses financed through the operating budget and in investments financed through the development budget. Significant cost savings can be identified: efficient use of existing hospital resources, reduction of referrals beyond the primary-care system, more intensive use of dispensaries and health centers, more complete shift of ministry operating budgets to finance primary rural care and, possibly, devolution of tertiary care to private providers if costs continue to be prohibitively high. 11.36 The most urgent problems are the low levels of coverage and efficiency in providing basic services, resulting largely from an emphasis on cure rather than prevention. In population, the annual cost of G 750 (US$150) per modern family planning user is unnecessarily high; coverage is low and unmet need is six times greater than the number of current users. In health, curative care is costly, facilities are understaffed and underequipped and hence underused; most people who need primary and preventive services cannot obtain them. In nutrition, curative approaches have proven too costly, preventive programs have not advanced beyond the pilot stage, and responsibility is dispersed among too many agencies. 11.37 Equity is also a major concern because of the significant rural-urban and regional disparities in service availability. Too much is spent in Port-au-Prince; too little benefits rural Haiti. Neither low coverage nor inequity can be corrected without greater efficiency. MSPP has rightly recognized--with its New Orientation, its constructive cooperation with NGOs, and its reassignment of staff positions in the FY86 recurrent budget-that the extension of coverage within current budgetary restrictions requires giving up some urban hospital services as well as achieving greater efficiencies. 11.38 MSPP should Implement planned staff reductions in the Port-au-Prince area and strengthen rural and regional services. The proposed reduction of about 900 staff positions in Port-au-Prince is laudable and could save about G 9.5 million, or some 12 percent of the MSPP salary budget. By July 1986, however, it had not taken place. 11.39 Further cuts could be achieved by reducing the number of medical students. Fewer than half the current graduating class of 100 (already down  -112- from 250 in the 1970s) are likely to find work in Haiti6/ and a further reduction to 50, with commensurate cuts in the teaching staff, would save on subsidies to the medical school and its students plus about G 0.5 million per year in direct staff costs. The medical school was built for only 35 students and its quality would improve if student numbers were reduced. 11.40 The proposed increase of more than 1,000 in regional staff is excessive. There is no guarantee that those transferred will either move from Port-au-Prince or have the skills necessary to deliver primary health serviees in rural areas. These services can be delivered effectively and at lower cost by volunteers, auxiliaries and nurses than by trained physicians who are needed only for referrals at higher levels of service. Since many persons no longer needed in Port-au-Prince may refuse placement in the regions, total MSPP employment could fall by 500 persons, resulting in an actual saving of about G 5 million, or five percent of the recurrent budget. More effective management of resources in the recurrent budget, which is not at all financed by aid, will be an important sign of progress in the sector. 11.41 It would be useful to freeze employment in MSPP to increase the efficiency of its staff. A solution may lie in further staff cuts and readjustments as advocated above, accompanied by contracting out the delivery of more public services to the private sector and especially to NGOs. MSPP's record of cooperation with NGOs has been good but more can be done. 11.42 The shift of public health resources toward primary and preventive care should continue; where possible, NGOs should be mobilized to deliver rural services. 11.43 Efficiency could also be increased by the more intensive use of dispensaries and health centers. The devolution of some tertiary care to private providers, and some forms of private or social insurance to finance health care, should be considered. Cost recovery on the basis of user charges for curative services is being tried and should be expanded. Recent progress made by MSPP in analyzing how resources are allocated and in improving health status statistics should be consolidated and continued, in order to provide a basis for increasing productivity throughout the public health system. 11.44 The new maternity hospital for Port-au-Prince proposed for FY86 should not proceed, given excess capacity in existing facilities and its inconsistency with the priorities of the New Orientation. This would save G 10 million. 11.45 If these recommended adjustments to the recurrent and investment budgets for FY86 are made, total spending could be held to G 155 million, the lowest level since the 1970s. While a substantial share of the burden of health care has and will continue to shift to NGOs, continued public 6/ R. Bicknell et al., Paying for the Public's Health: Haiti, 1984-1994, Health Policy Institute, Boston University, 1985.  -113- expenditure at the G 155 million level is inadequate. MSPP will have to increase its recurrent budget gradually, in addition to improving efficiency; an annual increase of G 5 million in FY86 prices, or approximately six percent, should be enough for FY87-89. This assumes, however, th;,t donors continue to pay some recurrent costs through their aid to the "investment" budget. 11.46 If MSPP is successful in making the adjustment it has proposed for FY86, reducing staff in Port-au-Prince and building up its capacity to deliver rural health services, donors may wish to step up their support from FY87 onward. Table XI.4 shows a recommended flow of funds for public expenditure on populatior., health and nutrition in FY86-89. It assumes that the FY86 period of adjustment will be successful and that the investment budget will expand from FY87 on. Whether this expansion can occur depends fundamentally or. two factors: (1) improved efficiency in service delivery by MSPP; and (2) the quality of individual projects proposed. Table XI.4: MSPP: RECOMMENDED EXPENDITURE PROGRAM AND FINANCING, FY86-89 (G million at FY86 prices) FY86 FY87-89 FY86-8T APPLICATIONS 155 615 770 Recurrent budget 85 285 370 Investment budget 70 330 400 Regionalize services 13 82 94 Strengthen rural services 3 39 42 MCH/Family planning 16 72 87 Nutritiorn 1 9 10 Endemic disease control (malaria) 34 108 142 Other 3 21 24 SOURCES 155 615 770 Treasury 95 330 425 Recurrent budget 85 285 370 Investment budget 10 a/ 45 55 Donors b/ 60 285 345 USAID: PL-480 17 63 80 : direct 17 78 95 IDB c/ 4 67 71 Others 22 77 99 a/ Assumes Port-au-Prince maternity hospital does not proceed. bl Aid committed only for FY86 and 10-20 percent of FY87-89 requirements. c/ Assumes MSPP and IDB develop a satisfactory program (see text). Sources: Tables XI.3, 11.8, and 11.9.  -114- 11.47 The regionalizatton and strengthening rural services projects are valuable but effective design, management and execution have so far proven elusive. If MSPP can improve its capacity to deliver rural health services during FY86, higher funding levels can be resumed in FY87-89. 11.48 As resources shift away from curative care, it should be possible to provide more support to population and nutrition programs which have not received adequate budgets in the past. 11.49 The maternal and child care and family planning program has run into difficulties but can be resuscitated. The decline in family planning service delivery must be reversed. There is a large and growing unmet need to provide contraceptives as an integral part of primary health care. Some combination of vertical and horizontal service delivery may be essential to assure than emphasis on maternal and child health is not diluted when it is combined with other health actions at the regional level. Population control programs, as well as other programs of preventive health care, need continued nurturing and strong support from the central management of MSPP and technical backstopping from DHFN. The introduction of incentive pay could be considered. MSPP's neu Directorate of Population needs strengthening. 11.5& Less than two percent (if sector resources have been allocated to the pressing problem of nutrition in rscent years and this will continue in FY86. Support should quadruple ir. the future provided that current rehydration therapy programs can first be expanded at low cost per beneficiary. Nutrition interveations should also be highly targeted on those most in need. 11.51 Several experts question whether so much emphasis needs to be placed on the malaria program, wxhich is at p:-sent the largest recipient of external aid. Past evaluation studies should be reviewed to assess whether a program of control--rather thim eradication--might be more effective in light of other sector needs. 11.52 The investment program also provides small amounts for epidemiological surveillance and urban health and sanitation. These amounts are appropriate and should be camplemented with some Treasury resources to support the maintenance and re-equipment of the University Hospital. 11.53 Ir. constant FY86 prices, the total recommended expenditure program for FY86-89 is G 770 million, of which G 425 million (55 percent) should come from the Treasury. External aid donors should finance the remaining 45 percent, or G 345 million (US$69 million). No new aid commitments are needed for FY86 but 80-90 percent of the recommended FY87-89 program will need new commitments i.ev, G 250 million (US$50 million). The foreign exchange component of the program is not known but is unlikely to be greater than the 45 percent to be financed by external donors. Donors will have to continue to finance what are in fact recurrent costs even though they are in the investment budget.  -115- 11.54 The Government will have to concentrate its financial resources on the recurrent budget which is recommended to rise by G 5 million in real terms each year. No increase ir. Treasury contributions to the investment budget is therefore realistic and it is recommended to remain constant at G 15 million in FY86 prices. Total Government funding of the health sector should grow from G 95 million ir. FY86 to G-115 million in FY89, a seven percent real annual increase. 11.55 As noted, aid commitatents have not been made for most of the investment program recommended in Table XI.4, except for FY86. Moreover, there are serious questions about certain donors being able to finance the assistance implied in Table XI.4. USAID has in recent years shifted its financing toward N