Urban finances in Bogota, Colombia

引 This collection of essays was prepared under the auspices of World Bank Research Project No. 671-47 ("Bogota City Study"). At the ti-ine of the preparation, Messrs. Greytak and Wolff were consultants to the World Bank. Ms. Fawcett was research assistant and Mr. Linn., a Senior Economist in the World Bank's Urban and Regional Division, Development Policy Staff. WPS/BR-001covs/2.8.85 ABSTRACT This report presents a collection of four studies on the Urban Finances of Bogota, Colombia. All four studies emphasize the importance of the institutional and political dimensions and constraints of urban govern- ment. Chapter 1 provides an overview of urban finance in Bogota, including the structure of urban government, trends in expenditures and revenues of city government, and issues in urban taxation, user charges, revenue sharing, and budgeting and investment planning. Chapter 2 analyses budgeting, investment planning and financial control, with particular emphasis on the institutional and political aspects of urban financial planning in Bogota. Chapter 3 reviews alternative methods of urban fiscal forecasting and assesses the fore- casting practices employed by city agencies in Bogota. Chapter 4 analyses the system of intergovernmental fiscal relations and revenue sharing in Colombia as it affects in particular the financial status and service provision of local authorities in Bogota. Special attention is given to the organization and financing of education and health services. WPS/BR-001covs/2.8.85 FOREWORD This collection of four studies summarizes the research on the urban public finances of Bogota, Colombia, which was carried out under the auspices of World Bank Research Project No. 671-47 ("Bogota City Study"). This research is complementary to other components of the City Study, which deal * with a broad range of urban development, planning and policy issues in the areas of housing, land, transport, employment and industrial location. The work reported on in this volume extends beyond the more common type of case studies on urban finances in developing countries, in that it emphasizes the importance of the institutional and political dimensions and constraints of urban government which permeate many of the urban finance issues in the cities of the developing world. In particular, the papers deal primarily, and in some detail, with issues that are often neglected or treated only in passing, namely urban government budgeting and fiscal forecasting, financial planning and control, and intergovernmental fiscal relations. The purpose of this volume is therefore not only to provide a description and analysis of conditions and practices prevailing as of 1980 in these important areas of urban finance and public administration in Bogota, but also to serve as an example of how these issues may be approached in other cities. Chapter 1 provides an overview of urban finance trends and issues in Bogota in comparison with other cities in Colombia and in other developing countries. Chapter 2 analyses the budgeting, investment planning and finan- cial control process and procedures in Bogota. Chapter 3 reviews and assesses the fiscal forecasting methods employed by the major urban government entities in Bogota in relation to the experience with the practices employed in U.S. WPS/BR-001covs/2.8.85 cities. Chapter 4 provides an analysis of the system of intergovernmental fiscal relations and revenue sharing in Colombia, with special attention to how it affects the financial situation of urban government in Bogota. In addition to the work reported in this volume, a series of comprehensive urban finance statistics for 1961-1980 was collected for the local public agencies of Bogota under the auspices of the City Study. An updated version of this series, including 1981 data, was published by the District Controller's Office in december 1982. 1/ The authors wish to express their deep gratitude to the many public officials in Bogota which have made this research possible by providing the data, information, and many of the elements of analysis reported on in the four studies contained in this volume. In addition, the authors gratefully acknowledge the assistance and comments of their colleagues involved in the City Study, in particular Gregory Ingram and Jose Fernando Pineda. Richard Bird also contributed many helpful suggestions and comments, particularly in the preparation of Chapter 4. All errors of fact or interpretation remain, of course, the sole responsibility of the authors. 1/Contralaria de Bogota D.E., "Finances Publicas Consolidadas del Distrito Especial de Bogota, 1961-1981", Boletin Estadistico, Vol. 2, No. 5, December 1982. Table of Contents Chapter 1. ADMINISTRATION AND FINANCING OF PUBLIC SERVICES IN BOGOTA: SUMMARY OR FINDING by Johannes F. Linn.................................. 1 Introduction .......... 1 Responsibilities and Structure of Urban 1.... 1 Government in Bogota ................................... .... 1 Trends and Composition of Expenditures and Revenue of District Agencies ........ ........................ 3 District Taxes .......... 18 User Fees and Development Charges ......................... 28 Revenue Sharing ......... 34 Budgeting and Investment Planning in Bogota ................ 38 CONCLUSIONS ......................a ...... 40 Chapter 2. BUDGETING AND INVESTMENT PLANNING IN BOGOTA, COLOMBIA by J. H. Wolff ............................. ......... 43 Introduction and Summary ..................... 43 The Bogota Budgetary System and Format ...................... 48 Budgeting: The Organization ...................... ...... ..51 Budgeting: The Process *** * ..............****.................. 61 Credit Raising and Investment Planning: The Organization ....................................****** 74 Credit Raising and Planning: The Process *****... . 78 Budgeting, Planning and Politics: Political Culture versus Technical Rationality ....................... .... 86 Recommendations .......... 99 Chapter 3. REVENUE AND EXPENDITURE FORECASTING IN BOGOTA by David Greytak ...................................... 105 Introduction.... .....105 Urban Government Revenue and Expenditure Forecasts: The Background....... .. .........106 General Purposes .................. 110 Common Practices..........113 Medium-term Forcasting in Bogota..........123 Bogota Special District........ ...........123 Empresa de Energia Electrica de Bogota (EEEB)........... 135 Empresa de Acueducto y Acantarillado de Bogota (EAAB)... 148 Empresa de Telefonos de Bogota (ETB).............. 158 Summary Assessment of Forecasting Practices in Bogota... 161 Chapter 4. THE REVENUE SHARING SYSTEM IN COLOMBIA: A CASE STUDY OF BOGOTA by Caroline S. Fawcett Introduction Revenue Sharing ...... 168 District Consolidated Budget ....... 170 Central Administration Budget ....... 176 Shared Taxes and Grants ...........* .....................180 Adjusting for F.E.R. and S.S.B. Transfers ....*188 Intergovenmental Transfers to the District .... * o............192 The Situado Fiscal .................................... 192 Shared Sales Tax ...........o....................... 199 Comparative Analysis of Situado Fiscal and Sales Tax Sharing Program ........................ 204 Beer Tax Sharing ....................... . ....... 249 Tobacco Tax Sharing .............................. . 255 Liquor Tax Sharing ...... ........................... 256 Revenue Performance of Beer, Tobacco and Liquor Taxes .. 257 Additional Tax Sharing Arrangements .................. 264 Grants ........................................ ..273 Impact on District Revenue and Expenditure ..... ...... 288 Introduction .................................. 288 The Workings of S.S.B....... ......................... 289 The Workings of F.E.R...... ............................ 306 The Impact of National Transfers on S.S.B. and F.E.R. Expenditures ...................... 321 Conclusions ....................................... ........ 328 WPS/BR-001/2.12.85 Chapter 1. ADMINISTRATION AND FINANCING OF PUBLIC SERVICES IN BOGOTA: SUMMARY OR FINDINGS by Johannes F. Linn I. INTRODUCTION This chapter summarizes the major findings on the administration and financing of public services in the Special District of Bogota, Colombia, re- sulting from a number of studies carried out under the auspices of World Bank Research Project 671-47 ("Bogota City Study"), and based on research pre- viously carried out by the author under World Bank Research project, 670-70 ("Administration and Financing of Urban Governments in Developing Countries"). It is organized as follows: Section II discusses the responsi- bilities and structure of urban government in Bogota. Section III reviews the trends and composition of expenditures and revenues of urban government agen- cies in Bogota. Section IV looks at taxes levied by the District Government, while Section V considers user charges and development contributions for urban services. Section VI deals with intergovernment transfers and relations in Bogota. Section VI assesses the budgeting and fiscal planning practices by local agencies in Bogota. Reference is made at the end of the chapter to background documents which provide documentation, detailed analysis and further elaboration of the major points discussed in this chapter. II. RESPONSIBILITIES AND STRUCTURE OF URBAN GOVERNMENT IN BOGOTA Compared with the responsibilities and structure of urban governments in the other cities of the developing world the following observations may be made regarding urban government in the Special District of Bogota: WPS/BR-001/2.12.85 - 2 - 1. The District Government is responsible for a larger number of urban services than is commonly found for local governments in other LDC cities. Public housing is, however, not a major local responsibility in Bogota, but is in the hands of a national-level agency, the Instituto de Credito Territorial (ICT). 2. The administrative structure of the District Government is charac- terized by a high degree of decentralization, but in contrast to cities in the United States this administrative fragmentation is along functional lines rather than along geographic boundaries separating autonomous jurisdictions within the metropolitan area. The extent of functional fragmentation in Bogota is unusually high, when compared with other cities in Colombia and in other developing countries. 1/ 3. Some minor problems in the planning and management of urban services iv the metropolitan area of Bogota have occurred in selected areas of the city where metropolitan growth has spilled over from the SpeciJL District in to adjacent municipalities. However, these problems have not yet reached the degree of severity of geographic fragmentation experienced in some other large cities in developing countries, such as Mexico City and Calcutta. In the past, much of the potential problem has been avoided by annexation of selected municipalities; at present, some of the more serious problems are avoided by 1/ The District Government is comprised of the District Administration under the direct authority of the Mayor and a number of autonomous, or "decentralized" agencies in charge of particular services, including the Water Company (EAAB), the Power Company (EEES), the Telephone Company (EEEB), the enterprise in charge of garbage collection, street cleaning, markets and cemeteries (EDIS), the Urban Development Institute (IDU), the District Bus Company (EDTU), the District Littery, University, and Orchestra, the District Housing Agency (CVP) and the District Social Security Agency (CPS), to mention but the most important. WPS/BR-001/2.12.85 - 3 - extension of service perimenter of the major public utilities in Bogota which have begun to provide services to adjacent municipalities (especially for water and electric power). Nonetheless, interjurisdictional frictions may be expected to grow as Bogota continues to expand and may well include increased problems in the areas of competition between jurisdictions, refusal by rural municipalities to accept the dumping of District refuse, refusal and suburban municipalities to cooperate with the District in zoning and planning, etc. 4. In various areas of urban serNice provision District agencies have to adhere to national government policies and control, as in the case of public education and health services, in the setting of urban service charges and of many local tax rates, and when incurring foreign or local debts. In contrast to many other cities in developing countries, however, the District Government has extensive de facto authority over the urban planning functions and faces virtually no intervention by national authorities in the preparation and implementation of its annual budgets. III. TRENDS AND COMPOSITION OF EXPENDITURES AND REVENUES OF DISTRICT AGENCIES A series of expenditures and revenue data of all District agencies was put together for the "City Study" in order to permit an analysis of the trends and composition of consolidated District expenditures and revenues for the last two decades'. The following conclusions are based on these financial WPS/BR-001/2.12.85 - 4 - data. 1/ Where possible, comparisons are drawn with the financial trends of the local government of Cali, the third largest city in Colombia. Consolidated District Expenditures Between 1961 and 1979 consolidated District expenditure grew by a fac.tor of 37 (Figure 1.1), when taken La current prices. However, when expressed in constant prices and on a per capita basis, consolidated expendi- tures did not exhibit a strong trend, but fluctuated quite widely around an average level of about Col. $2,100 (in constant 1976 prices) (Figure 1.2). These fluctuations appear to reflect largely the changing rate of-inflation: during periods of high inflation District revenues did not manage to grow in tandem with the general price level, thus resulting in a slow down or contrac- tion of real per capita expenditures. The same conclusion also applies to the expenditure trends of the municipal government in Cali, which were almost parallel to the fluctuations of consolidated District expenditures in Bogota (Figure 1.3). Maybe more surprising is the finding that per capita spending by local authorities in Bogota and Cali were virtually at equal levels between 1966 and 1976, despite the fact that the local government in Cali had fewer tax instruments at this disposal than the District Government in Bogota, and despite the fact that the economic-- and thus taxable--base in Cali was below that of Bogota when measure in terms of per capita disposable income. 1/ In 1979 only the following agencies were included in the expenditure and revenue data: District Administration, EAAB, EBB, ETB, EDIS, IDU and CPS. In 1978 these seven agencies accounted for 90 percent of consolidated District expenditures. Furthermore, for 1979 District Administration expenditures reflect budget estimates, rather than closed accounts. -5- Figure 1.1 Total Consolidated Expenditure (1961-79) (Real expenditure In 1976 prices) 2.0 2,0 1.0?- -1.0 0.5 -0.5 '61 '63 '65 '67 '69 Yea '71 '73 '75 '77 '9 Year Legend: Nominal Expenditure World Bank-26162 . Real Expenditure -6- Figure 1.2 Total Consolidated Expenditure Per Capita (1961-79) (In 1976 prices) 4,000- 4,000 3,000- -3,000 2.000?.- - 2,000 1,000,- -1,000 0. 1 + 0 61 '63 '65 '67 '69 '71 '73 '75 '77 '79 Year t tged ..... Nominal Uxpenditume Real Expendilture World Bank-26163 Figure 1.3 Bogota and Cali: Consolidated Total Expenditure Per Capita (in 1976 prices) 3,000- 2.500- F 2,50 2,000- -2,000 1,500 - 1,500 1,000 J-1,000 500- -500 0 0 '61 '63 '65 '67 '69 '71 '73 '75 Year Legend: U - -Call World Bank-26164 . -Bogota WPS/BR-001/2.12.85 - 8 - As regards the structure of District expenditure, the following observations may be made: First, the breakdown between recurrent and capital spending fluctuated extensively over the last two decades, but did not show a clear trend, and the level of per capita spending in constant prices did not show a rising or falling trend over the same period (Table 1.1). Second, the decentralized agencies contributed an overwhelming share of total consolidated District expenditure. Especially the three public utility companies respon- sible for water, power and telephone services weighted very heavily, contributing over half of total consolidated expenditures (Figure-1.4), and generally more than 70 percent capital spending (Figure 1.5). Third, when considering the functional breakdown of consolidated District spending the following findings appear from Table 1.2: General administrative expenditures generally accounted for less than 10 percent of total spending and on balance it appears to have accounted for a slightly higher share in the 1970s than it did during the 1960s. Debt service fluctuated between 14 and 20 percent of total spending, but in recent years remained comfortably below the 20 percent level for all District agencies combined. 1/ Education spending saw a steady increase as a percent of total District 5 spending until 1976, when all educa- tional expenditure responsibility was vested in the Regional Education Fund (FER) which is under the joint control of the National Ministry of Education and the District Secretariat of Education. As a result, education spending by the District dropped drastically after 1976. Public health services, although 1/ This is of some relevance, since each individual agency is prohibited by the statute to exceed a debt service-revenue ratio of 20 percent. Individual District agencies (such as EAAB) have reached this ceiling, while others, such as the Distr.ct Administration as yet are substantially below it. WPS/BR-01TI-1/2.12.85 9 Table 1.1: Bogota: Consolidated Capital Expenditures (Per Capita) (in 1976 prices) 1 2 3 4 5 6 7 8 9 10 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1. General Administration 0.0 0.0 3.0 a - 1.4 0.0 1.4 5.3 40.2 24.1 2. Police and prisons 0.0 0.0 1.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3. Fire Protection 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4. Roads and Bridges 48.6 38.5 24.2 56.6 50.1 73.3 235.3 323.6 201.4 117.0 - 5. Parks and Recreation 0.0 0.6 0.1 0.3 0.3 0.4 17.9 27.2 21.3 11.0 6. Education 4,1 1.1 16.2 3.0 7.1 11.3 21.3 4.2 35.2 .37.0 7. Health 0.0 0.1 0.2 1.2 1.5 4.5 1.9 0.9 1.8 0.8 8. Transportation 0.0 2.1 0.5 0.0 0.0 0.0 0.0 68.9 0.0 32.3 9. Welfare 0.0 0.6 6.3 0.9 1.7 0.0 0.0 0.9 6.2 15.9 10. Water and Sewerage 96.4 109.1 126.4 110.6 93.6 110.4 153.6 240.9 463.8 402.7 11. Other Sanitation 0.0 0.0 42.8 71.0 8.6 3.3 1.2 - 6.1 4.9 12.3 12. Electricity 572.3 693.1 725.9 518.9 354.5 166,9 105.4 202.7 261.4 157.6 13. Telephones 53.5 111.3 75.7 46.0 57.2 109.3 228.8 89.5 126.2 141.2 14. Housing 15.1 13.1 3.2 23.0 17.8 20.6 4.0 18.2 16.2 12.9 15. District Social Security 1.4 14.0 6.7 4.3 4.9 12.8 6.4 4.6 11.5 14.2 16. Debt Service 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17. Trans. To Outside Dist. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 18. Total Expenditures 791.4 983.7 1032.4 837.3 598.7 512.8 777.3 993.1 1190.0 983.0 1 2 3 4 5 6 7 8 9 1971 1972 1973 1974 1975 1976 1977 1978 1979 1. General Administration 12.3 9.9 12.2 20.7 10.8 12.4 2.1 3.3 9.0 2. Police and prisons 0.0 0.0 0.0 1.5 0.0 0.0 0.0 0.0 1.8 3. Fire Protection 0.0 0.0 0.0 0.2 2.5 2.4 0.0 0.1 0.0 4. Roads and Bridges 110.1 76.0 92.6 173.2 98.8 132.1 97.7 69.7 60.1 5. Parks and Recreation 6.5 4.5 4.1 4.0 6.9 8.8 3.3 2.3 4.3 6. Education 37.6 11.6 4.2 9.8 4.5 5.0 0.5 1.9 6.4 7. Health 1.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8. Transportation 33.1 0.5 0.0 19.3 99.0 6.2 0.0 0.1 1.5 9. Welfare 48.3 51.2 40.6 40.0 8.7 18.6 16.1 10.8 7.4 10. Water and Sewerage 277.3 370.4 225.2 180.9 335.1 293.5 223.3 222.1 151.7 11. Other Sanitation 27.2 7.5 4.2 21.2 2.0 0.1 0.0 0.0 1.5 12. Electricity 201.9 136.4 236.1 99.9 196.6 131.7 88.1 120.7 57.5 13. Telephones 171.6 217.9 135.5 258.4 199.1 191.5 136.2 182.7 191.6 14. Housing 15.1 6.5 4.2 27.7 39.1 45.9 36.1 41.8 0.0 15. District Social Security 6.7 1.2 20.1 10.0 0.8 0.2 0.4 0.3 0.2 16. Debt Service 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17. Trans. to Outside Dist. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 18. Total Expenditures 949.0 893.6 779.0 866.9 1003.9 848.5 603.9 655.8 492.9 - 10 - Figure 1.4 Composition of Total Expenditures Per Capita (in percent) 80- 80 60l 60 04 4-40- -40 C 20- -20 '61 '63 '65 '67 '69 '71 '73 '75 '77 '79 Year Legend: - EAAB, EEEB, ETB S - District Administration World Bank-26165 - Other Decentralized Agencies 긔 WHIBR-OM-2/2.12.85 12 Table 1.2; Bogota: Structure of Consolidated Total Expenditures (Percentage Distribution) 1 2 3 4 5 6 7 8 9 10 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1. General Administration 6.4 5.1 6.0 5.5 5.8 5.8 4.7 6.3 9.4 8.5 2. Police and prisons 0.9 0.8 0.9 0.9 1.0 1.0 0.8 0.7 o.6 0.5 3. Fire Protection 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.2 4. Roads and Bridges 4.4 3.4 2.7 4.7 4.9 6,4 14.6 17.0 9.9 7.1 5. Parks and Recreation 0.2 0.2 0.2 0.2 0.3 0.3 1.0 1.4 0.9 Q.5 6. Education 6.4 5.9 6.0 6.5 7.6 8.8 7.5 5.7 6.3 10.6 7. Health 2.1 1.7 1.5 1.8 1.9 2.1 1.5 1.4 1.3 1.4 8. Transportation 4.3 3.2 3.0 3.0 3.1 2.8 2.2 5.2 2.0 3.7 9. Welfare 1.6 1.8 2.1 1.3 0.8 0.9 0.7 0.7 1.0 1.3 10. Water and Sewerage 9.1 9.1 9.7 10.7 11.1 13.2 13.9 16.8 23.6 22.5 11. Other Sanitation 2.4 2.2 4.1 6.2 3.2 3.1 2.6 2.7 2.5 3.3 12. Electricity 34.8 34.2 37.5 32.9 27.7 18.3 13.2 15.1 15.8 12.4 13. Telephones 5.9 8.2 6.4 5.3 6.9 10.8 15.9 7.4 8.7 10.0 14. Housing 2.0 1.5 0.9 2.4 2.1 2.6 0.8 1.4 1.0 0.9 15. District Social Security 3.3 4.9 4.1 4.4 5.0 4.0 4.6 3.1 2.9 3.2 16. Debt Service 15.5 17.4 14.8 13.9 18.6 19.7 15.7 15.2 14.0 13.9 17. Trans. To Outside Dist. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 18. Total Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1 2 3 4 5 6 7 8 9 1971 1972 1973 1974 1975 1976 .1977 1978 1979 1. General Administration 7.3 8.6 6.5 10.9 6.0 6.7 9.2 7.9 6.4 2. Police and prisons 0.5 O.R 0.5 0.5 0.5 0.4 0.5 0.4 0.5 3. Fire Protection 0.2 0.2 0.2 0.1 0.2 0.3 0.2 0.2 0.2 4. Roads and Bridges 6.1 4.5 5,2 8.1 5.1 7.0 6.7 4.5 4.9 5. Parks and Recreation 0.3 0.5 0.8 0.7 0.8 0.8 0.5 0.6 0.4 6. Education 11.2 11.5 10.7 9.8 10.7 10.5 5.1 1.8 1.2 7. Health 1.6 1.5 1.5 1.4 1.2 1.2 1.3 1.1 1.3 8. Transportation 3.4 2.0 1.7 2.2 5.0 1.9 1.6 1.8 0.6 9. Welfare 2.6 2.7 2.6 2.4 1.4 1.8 2.0 1.7 1.7 10. Water and Sewerage 16.0 19.7 16.6 13.9 19.0 18.0 17.5 15.7 17.0 11. Other Sanitation 3.8 3.6 3.4 3.9 2.8 3.0 3.4 2.9 3.3 12. Electricity 14.1 12.7 16.2 10.7 15.7 15.3 16.4 23.6 24.3 13. Telephones 10.4 12.2 9.9 14.1 11.5 12.4 12.4 15.6 19.3 14, Housing 1.0 0.6 0.5 1.2 1.6 2.1 2.1 2.1 0-.0 15. District Social Security 3.7 3.6 3.9 3.8 3.s 3.7 3.9 4.3 1.6 16. Debt Service 17.7 15.8 19.8 16.4 14.8 14.8 17.0 16.0 17.3 17. Trans. to Outside Dist. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 18. Total Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 WPS/BR-001/2.12.85 - 13 - similarly under joint national-District administration, never were provided from District budgetary resources, expect for minor expenditure items. More important, Table 2.2 reflects the relatively small and erratic share of spend- ing on roads and bridges and on transportation and transit services in Bogota. This reflects the fact that mass transit is largely operated by private bus companies and that transport infrastructure investments have tended to be heavily bunched in certain years, in particular between 1967 and 1971 and between 1974 and 1976. Other cities, such as Mexico City, which are heavily engaged in the provision of mass transit services have fodnd that this involvement tends to act as a major drain on urban finance resources. Similarly, in line with the minimal responsibility of District agencies in public housing, a minimal percentage of District spending is devoted to this category, again in contrast with cities where local authorities have major responsibility in this area (e.g., Zambian cities and towns). Finally, utility services constitute the major source of spending in Bogota in func- tional terms, as was noted previously when discussing the breakdown of District spending by agency. Consolidated District Revenues Total consolidated revenues (including borrowing) of the District Government moved very closely in tandem with consolidated expenditures, due to the fact that none of the District agencies carry major reserve funds from which they can draw or into which they can deposit funds to support sustained deficits or to absorb surpluses. As regards the structure of revenues, tax revenues declined relative to other sources during the last two decades. In 1961, taxes contributed 27.0 percent to total consolidated revenue, reaching a minimum share of only 12.4 WPS/BR-001/2.12.85 - 14 - percent in 1974, and rising thereafter to 16 percent in 1979 (Table 1.3). In terms of constant 1976 prices, per capita tax revenues dropped from Col. $482 in 1961 to Col. $381 in 1979 (Table 1.4). Revenues on account of service charges and developement contributions ("self-financed revenues") grew in relative terms from 45 percent in 1961 to 60 percent in 1979, as well as in terms of real per capita collections (Tables 1.3 and 1.4). Revenues from borrowing fluctuated widely over the years, contri- buting generally between 20 and 30 percent of consolidated Distridt revenues, but dropping to 14 percent in 1979 (Table 1.3). Almost all borrowing was carried out by the decentralized agencies. Revenues from national grants and shared taxes amounted to between 8 and 15 percent of consolidated District revenues, a low percentage in compari- son with the share of transfers and shared taxes in the total revenues of urban governments elsewhere in the developing world (see Tables 1.3 and 1.5). The dramatic decline of revenues on account of grants between 1975 and 1979--in terms of constant 1976 prices per capita revenues in this category declined from Col. $393 to Col. $210 during this period--was due to the nationalization of education services in 1976, and the resulting cut in transfers to the District Administrative budget. 1/ In sum, the revenue structure of the District Government changed quite drastically during the last two decades, with a reduction in the reliance on tax revenues and transfers, and a significantly increased reliance 1/ The transfers were channeled instead directly to the FER, as discussed further below. WPS/BR-01T1-3/2.13.85 - 15 - Table 1.3: Bogota: Structure of Consolidated Revenues (Percentage Distribution) 1 2 3 4 5 6 7 8 9 10 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 i. Tax Revenues 27.0 22.5 18.4 18.8 19.9 18.3 16.2 14.2 15.1 13.8 2. General Non-tax Revenues 1.4 1.3 1.0 1.2 1.4 0.5 0.5 0.5 0.4 0.3 3. Self-Financed Revenues 44.4 41.4 39.5 41.0 45.9 49.5 44.5 46.5 54.1 53.3 4. Shared Taxes 8.0 6.4 9.1 11.2 10.1 9.6 7.9 6.2 7.2 9.3 5. Grants 0.9 2.4 3.6 3.4 2.8 3.6 2.3 3.5 1.8 4.0 6. Borrowing 18.3 26.0 28.3 24.4 19.8 18.4 28.5 29.1 21.4 19.3 7. Total Revenues 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1 2 3 4 5 6 7 8 9 1971 1972 1973 1974 1975 1976 1977 1978 1979 1. Tax Revenues 13.0 13.5 14.7 12.4 12.8 13.5 13.5 13.5 16.0 2. General Non-tax Revenues 0.4 0.3 0.2 0.4 0.3 0.3 0.3 0.4 0.9 3. Self-Financed Revenues 52.6 51.7 50.7 49.3 44.5 53.9 60.0 60.9 60.3 4. Shared Taxes 10.1 9.6 8.6 8.8 9.3 7.9 7.9 7.8 8.8 5. Grants 3.8 3.3 4.4 4.1 4.8 4.5 0.9 1.2 0.0 6. Borrowing 20.0 21.7 21.4 25.0 28.2 19.9 17.3 16.1 14.0 7. Total Revenues 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 - 16 - Table 1.4: Bogota: Consolidated Per Capita Revenues (in 1976 prices) 1 2 3 4 5 6 7 8 9 10 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1. Tax Revenues 482.4 460.7 373.3 338.3 354.1 288.2 325.8 324.6 358.1 348.6 2. General Non-tax Revenues 25.2 26.3 21.2 21., 25.5 8.6 10.4 12.1 9.8 8.1 3. Self-Financed Revenues 792.6 847.2 803.5 740.3 815.8 778.3 895.1 1061.4 1285.9 1348.1 4. Shared Taxes 142.4 131.5 185.7 201.9 179.3 150.6 159.7 141.5 171.1 1235.2 5. Grants 16.6 48.1 73.4 61.7 50.3 57.1 46.0 79.3 43.7 102.4 6. Borrowing 327.4 530.6 575.5 *440.2 352.9 289.4 572.5 665.5 509.8 488.7 7. Total Revenues 1786.7 2044.4 2032.7 1803.9 1777.9 1572.2 20J9.5 2284.5 2378.4 2531.1 1 2 3 4 5 6 7 8 9 1971 1972 1973 1974 1975 1976 1977 1978 1979 1. Tax Revenues 364.3 357.4 382.2 325.7 357.4 350.5 326.3 345.7 380.7 2. General Non-tax Revenues 12.6 8.3 5.6 10.8 8.9 6.8 7.3 10.4 21.3 3. Self-Financed Revenues 1474.7 1373.8 1317.0 1300.3 1241.7 1395.9 1451.0 1560.3 1432.6 4. Shared Taxes 284.7 255.5 224.0 233.2 259.7 205.5 191.6 199.1 210.3 5. Grants 107.2 86.5 113.5 108.3 133.5 115.3 22.4 31.5 0.0 6. Borrowing 562.2 575.6 557.7 658.8 787.8 516.2 419.0 413.4 332.8 7. Total Revenues 2805.6 2657.2 2600.2 2637.0 2788.9 2590.1 2417.7 2520.5 2377.7 WPS/BR-O1T1-5/2.13.85 - 17 - Table 1.5: Financing of Local Public Expenditure In Selected Cities: Percentage Distribution According To Type Of Revenues Total Local Local Self-financing Other Local Total External Grants and Net City Year Revenue Taxes Services Revenue Revenue Shared Taxes Borrowing_1/ Total Francistown (Botswana) 1972 102.9 46.8 56.1 -- -2.9 1.9 -4.8 100.0 Mexico City (Mexico) 1968 101.9 70.9 5.2 25.8 -1.9 8.9 -10.8 100.0 Le Paz (Bolivia) 1975 97.0 61.9 3.6 31.5 3.0 9.0 -6.0 100.0 Tunis (Tunisia) 1972 93.9 36.8 7.1 50.0 6.1 0.7 5.4 100.0 Kitwe (Zambia) 1975 92.7 35.0 53.1 4.6 7.3 2.2 5.1 100.0 Valencia (Venezuela) 1968 90.8 44.8 13.4 32.6 9.2 9.2 -- 100.0 Lubumbashi (Zaire) 1972 90.5 72.8 -- 17.7 9.5 9.5 -- 100.0 Rio de Janeiro (Brazil) 2/ 1967 88.4 74.5 7.2 6.7 11.6 1.7 9.9 100.0 Ahmadabab (India) 1970/1 86.3 38.6 41.8 5.9 13.7 4.2 9.5 100.0 Bombay (India) 1970/1 84.6 37.9 38.7 8.0 15.4 1.0 14.4 100.0 Karachi (Pakistan) 1974/5 84.1 67.6 2.2 14.3 15.9 2.8 13.1 100.0 Seoul (Korea) 1971 80.0 30.3 36.3 13.4 19.9 15.8 4.1 100.0 Jakarta (Indonesia) 1972/3 78.8 40.6 15.2 23.0 21.1 21.1 -- 100.0 (81.9) (43.7) (15.2) (23.0) (18.1) (18.1) (--) (100.0) Lusaka (Zambia) 1972 78.2 39.3 36.9 2.0 21.8 6.0 15.8 100.0 Cali (Colombia) 1974 74.4 15.6 57.5 1.3 25.7 2.8 22.9 100.0 Calcutta Corp. (India) 1974/5 73.8 64.4 -- 9.4 26.2 19.4 6.8 100.0 Cartagona (Colombia) 1972 70.4 23.3 43.3 3.8 29.6 12.8 16.8 100.0 Mbuji-Mayi (Zaire) 1971 70.2 66.5 -- 2.7 29.8 29.8 -- 100.0 Manila 4/ (Philippines) 1970 70.0 55.0 10.0 5.0 30.0 30.0 -- 100.0 Bukaru (Zaire) 1971 69.9 67.4 - 2.5 30.1 30.1 100.0 Madras (India) 1975/6 69.2 54.5 3.7 11.0 30.8 25.1 5.7 100.0 Bogota 3/ (Colombia) 1972 62.5 13.7 48.5 0.3 37.5 14.0 23.5 100.0 (72.4) (23.6) (48.5) (0.3) (27.6) (4.1) (23.5) 100.0 Tehran (Iran) 1974 46.9 42.8 -- 4.1 53.1 45.2 7.9 100.0 Kingston (Jamaica) 1971/2 30.1 23.9 2.7 3.4 69.9 67.2 2.7 100.0 Kinshasa (Zaire) 1971 26.9 25.4 -- 1.5 73.1 73.1 -- 100.0 Median 78.8 42.8 7.2 6.7 21.1 9.5 5.1 100.0 - (Average) (76.6) (46.0) (19.3) (11.2) (23.4) (17.7) (5.7) 100.0 1/ Net borrowing consists of loan financing minus net changes in financial assets or revenues. 2/ Due to exclusion of autonomous agencies the contribution of self-financing service revenue and that of all locally raised revenues are probably understated, while the remaining entries are overstated. 3/ Figures not in brackets include shared taxes under grants; figures in brackets include shared taxes under local taxes. 4/ Total revenues are used instead of total expenditures. Source: Johannes F. Linn, "Urban Finances in Developing Countries." In Urban Government Finances: Emerging Trends, ed., Roy W. Bahl, Beverly Hills: Sage Publications, 1980. WPS/BR-001/2.12.85 - 18 - on self-financed revenues, especially service fees and development charges. Graphically, this development is summarized in Figure 2.6. IV. DISTRICT TAXES Special attention has been given to an analysis of taxes levied by local authorities in the Special District of Bogota, in particular the most important among these, the property tax. The Property Tax Traditionally, the property tax has been the most importgnt tax of the District, although during 1979 the collection on account of the industry and commerce tax for the first time in many years exceeded those of the property tax (in 1980, the property tax was again the most important). The fact that property tax is commonly the most important of local taxes, not only in Colombia, but also in other developing (and developed) countries (Table 1.6), explains the considerable attention whi-h has been given to this tax in the context of the "City Study". In particular, the analysis has focussed on the revenue performance of the property tax in Bogota and its incidence. Revenue Performance. Between 1960 and 1979 the receipts from the property tax declined when measured in per capita terms and in constant prices (Table 1.7). Even though there was a strong surge in property tax collections in current prices between 1978 and 1980, the per capita revenues in 1980 did not exceed the average per capita collections for the years 1974 to 1976 when measured in constant prices (Table 1.7). At the same time the relative importance of the property tax declined dramatically: In 1961, 40 percent of the revenues of the District Administration were contributed by the property tax, while in 1979 its share amounted to only 22 percent. In fact, the - 19 - Figure 1.6 Composition of Total Revenues (in percent) 70 -70 60- 60 50- -50 40- -40 C 30- -30 20 -20 0 '61 '63 '65 '67 '69 '71 '73 '75 '77 '79 Year Legend: - Self-financed Revenues Ap....--m& - Shored Taxes 0 - Cr6ditos Borrowing - Grants World Bank-26167 - Tax & Non-tax Revenues WPS/BR-OITI-6/2.13.85 Table 1.6: Percentage Distribution of Local Tax Revenues According To Source In Selected Cities Local Taxes as Industry Percent of Property General and Motor All Total Local Property Transfer Income Sales Beer Gasoline Entertainment Commerce Vehicle Gambling Other City Year Expenditure Tax Tax Tax Tax Octrol Tax Tax Tax Tax Tax Tax Taxes Total Managua (Nicaragua) 1974 84.3 -- -- -- 69.4 -- -- -- 2.3 21.1 3.1 - 4.1 100.0 Mexico City (11exico) 1968 70.9 33.5 2.8 - -- - - 1.1 2.6 44.2 -- - 15.8 100.0 Valencia (Venezuela) 1968 44.8 21.4 - - -- - - -- -- 66.7 11.8 - - 100.0 Bogota (Colombia) 1972 13.7 58.4 - 1/ - -- - - 1/ 1.8 7.0 18.2 5.1 9.5 100.0 Call (Colombia) 1974 23.2 54.0 - - -- -- - -- 6.1 27.8 4.3 - 7.8 100.0 Cartagona (Colombia) 1972 23.3 61.2 - - -- -- - - 4.4 12.2 2.1 5.8 14.2 100.0 La Paz (Bolivia) 1975 61.9 5.2 - -- -- -- 7.1 -- 1.5 73.8 -- -- 12.4 100.0 Manila (Philippines) 1970 55.0 61.9 - -- -- -- - 2.2 -- 32.1 - - 3.8 100.0 (Group Average) (37.0) Karachi (Pakistan) 1974/5 67.6 46.0 - - -- 49.9 -- -- - 3.0 - 1.0 100.0 Ahmedabad (India) 1971/2 38.6 43.0 - - -- 52.0 - - - 2.0 -- 3.0 100.0 Bombay (India) 1971/2 37.9 55.6 -- - - 37.7 - -- 0.3 -- 3.7 -- 2.7 100.0 Calcutta Corp. (India) 1974/5 64.4 64.8 -- - -- 27.1 21 -- - - -- - -- 8.2 100.0 (Group Average) (41.7) Francistown (Botawana) 1972 48.8 12.9 -- 61.1 -- - - -- -- -- - - 26.0 100.0 Luaska (Zambia) 1972 39.3 74.6 - 25.4 -- -- -- -- -- -- - -- 100.0 Ndola (Zambia) 1972 n.a. 75.6 - 24.4 -- - -- -- - -- -- - -- 100.0 KLtwe (Zambia) 1972 n.a. 80.0 - 20.0 -- - -- -- - -- - - - 100.0 O Kinshasa (Zaire) 1971 25.4 -- - 14.4 -- - 62.5 - - -- - - 23.1 100.0 Bukaru (Zaire) 1971 67.4 - -- 3.7 -- - 87.0 - -- -- - -- 9.3 100.0 Mbuji-Mayi (Zaire) 1971 66.5 -- -- 62.7 -- -- -- - 37.3 100.0 Daegu (Korea) 1976 n.a. 49.5 21.2 9.1 -- -- -- - 10.4 -- 5.4 - 3.5 100.0 Gwangju (Korea) 1976 n.a. 50.3 23.1 13.2 - - -- -- 6.6 -- 4.1 -- 2.9 100.0 Daejeen (Korea) 1976 n.a. 51.0 2! 1 9.7 - - -- - 10.7 -- 5.5 - 3.0 100.0 Jeonju (Korea) 1976 n.a. 52.0 2i.4 8.9 -- - - - 7.5 -- 4.9 -- 2.1 100.0 (Group Average) (23.0) Seoul (Korea) 1971 30.3 20.6 34.8 -- -- - -- - 16.4 -- 22.2 - 6.0 100.0 Madras (India) 1975/6 54.5 68.9 5.1 - -- - - -- 16.0 -- - - 10.0 100.0 Tehran (Iran) 1974 42.8 55.3 -- - -- -- -- -- 9.1 -- 10.1 -- 25.6 100.0 Tunis (Tunisia) 1973 36.8 82.6 12.8 - -- -- -- -- 4.6 -- -- -- 100.0 Jakarta (Indonesia) 1972/3 43.7 -- 1/ -- - -- - - - 16.9 -- 50.2 26.9 6.0 100.0 Lagos (Nigeria) 1962/3 50.9 100.01 - -- -- -- -- - - - - -- -- - 100.0 Kingston (Jamaica) 1972 23.9 100.0 -- - - -- -- -- -- - - - - 100.0 Rio de Janeiro (Brazil) 1967 84.4 3.9 1.0 - 89.2 -- - -- 1 -- - -- - - 5.9 100.0 Overall Median 51.0 (Overall Average) (44.6) 1/ Share tax receipts not shown under local tax revenue. 2/ Share of Octrol receipts collected by the Calcutta Metropolitan Development Authority. Sourre; Linn, "Urban Financeg." Op.Cit. WPS/BR-O1T1-7/2.13.85 - 21 - Table 1.7: Bogota: Per Capita Revenues of the District Administration (in 1976 prices) 1 2 3 4 5 6 7 8 9 10 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1. LOCAL TAX REVENUE 347.8 456.2 434.0 349.7 315.2 328.8 257.4 290.4 280.1 303.1 2. Property & Complement 192.7 258.0 259.7 209.5 192.7 198.9 163.8 189.4 172.0 187.1 3. Industry & Commerce 41.6 111.0 94.0 79.4 65.2 70.1 44.6 50.7 48.6 51.9 4. Public Entertainment 0.0 1.8 0.4 0.0 0.0 0.0 0.1 0.0 0.0 10.9 5. Advertisment 0.0 0.0 0.5 0.5 0.4 0.4 0.5 1.5 4.0 4.4 6. Vehicles 37.3 44.7 38.9 25.9 21.2 25.4 20.1 18.7 21.3 16.5 7. Urban Construction 11.6 10.0 11.9 10.7 14.2 12.3 9.7 9.9 15.1 14.2 8. Public Road Use 0.6 0.9 0.8 0.8 0.7 0.6 0.5 2.3 1.9 1.4 9. Raffles 0.8 0.8 0.9 0.7 0.9 0.9 0.9 0.9 0.4 1.6 10. Betting 48.5 11.1 10.0 9.2 8.6 9.5 8.4 8.6 8.9 7.5 11. Gasoline ($0.04/Gal) 14.8 17.9 17.0 13.1 11.4 10.7 8.9 8.3 -8.0 7.6 12. Foreign Cigarettes 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 13. Others 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 14. LOCAL NON-TAX REVENUE 22.3 25.2 26.3 21.2 21.5 25.5 8.6 10.4 12.1 9.8 15. Sanitation Licenses 1.7 1.6 1.1 0.8 0.8 0.8 0.7 0.6 0.5 0.4 16. Nomenclatura Urbana 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 17. Laboratory Services 0.9 0.7 0.5 0.4 0.3 0.3 0.2 0.3 0.2 0.2 18. Interest Receipts 3.1 10.0 13.6 11.1 12.5 11.7 0.0 0.0 0.0 0.0 19. Fines 7.9 6.4 6.1 6.1 5.9 5.9 5.9 7.2 9.3 7.2 20. Confiscated Land Sales 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.0 0.2 0.1 21. Rental Receipts 2.6 1.9 1.1 0.5 0.3 0.4 0.1 0.1 0.1 0.0 22. Real Estate Sales 0.0 0.0 0.1 0.2 0.2 5.1 0.6 0.4 0.0 0.0 23. Others 6.0 4.5 3.8 1.8 1.5 1.3 1.1 1.8 1.7 1.8 24. SHARED TAXES 105.6 142.4 131.5 185.7 201.9 179.3 150.6 159.7 141.5 171.1 25. Sales Tax 0.0 0.0 0.0 0.0 0.0 0.0 11.1 0.7 0.0 6.6 26. Beer Tax 84.7 94.1 82.2 145.4 139.6 123.5 116.1 122.8 113.2 133.3 27. Gasoline ($0.06/Gal) 19.3 26.9 25.4 19.6 17.1 16.1 12.6 12.4 11.9 11.2 28. Other 1.6 21.4 23.8 20.7 45.3 39.7 10.8 14.1 13.4 20.0 29. Tobacco Tax 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.6 2.9 0.0 30. GRANTS 25.1 10.9 32.0 45.7 42.8 38.0 43.0 31.8 40.8 34.5 31. For Teachers Salaries 4.6 5.9 29.8 38.2 42.0 37.2 42.3 31.6 26.5 28.5 32. For Regional Ed. Fund 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 33. For School Cafeterias 0.3 0.4 0.3 0.2 0.2 0.3 0.3 0.2 0.2 0.2 34. For Health Services 18.9 3,4 0.8 6.5 0.0 0.0 0.0 0.0 0.0 0.9 35. For Social Welfare 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 36. For Econ. Dev. Works 0.0 0.0 - 0.0 0.0 0.0 0.0 0.0 0.0 14.0 4.9 37. From Decent. Agencies 1.3 1.2 1.0 0.8 0.6 0.6 0.4 0.0 0.0 0.0 38. CAPITAL RESOURCES (DEBT) 75.0 5.1 4.5 0.0 0.0 0.0 0.0 168.9 186.5 4.2 39. Loans 75.0 5.1 4.5 -0.0 0.0 0.0 0.0 58.0 11.3 4.2 40. Bond Issues 0.0 0.0 0.0 0.0 0.0 0.0 0.0 110.9 175.2 0.0 41. TOTAL REVENUES 575.8 639.7 628.2 602.3 581.5 571.6 459.7 661.2 661.0 522.7 WPS/BR-01TI-7b/2.13.85 - 22 - Table 1.7: Bogota: Per Capita Revenues of the District Administration (cont'd) (in 1976 prices) 1 2 3 4 5 6 7 8 9 10 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1. LOCAL TAX REVENUE 314.7 298.1 276.4 276.7 240.1 281.1 269.7 260.7 279.6 318.3 362.8 2. Property & Complement 188.8 169.4 166.8 172.6 144.0 158.5 141.5 108.7 122.5 121.6 146.0 3. Industry & Commerce 59.3 56.4 51.9 58.3 51.9 83.7 88.2 95.9 101.2 130.4 130.3 4. Public Entertainment 10.6 11.1 8.4 6.9 5.9 6.2 5.7 6.2 6.1 6.3 5.5 5. Advertisment 4.7 2.3 0.8 1.0 0.4 0.9 0.5 0.6 0.4 0.4 0.3 6. Vehicles 18.8 20.0 14.7 10.1 11.7 9.0 8.8 16.3 19.8 29.8 25.0 7. Urban Construction 15.2 12.5 10.3 10.6 9.9 8.6 9.2 13.8 12.4 13.3 31.2 8. Public Road Use 1.5 1.0 1.1 0.8 1.1 0.7 0.7 3.4 3.9 4.0 9.3 9. Raffles 1.1 1.0 1.1 0.5 0.6 1.5 1.1 2.1 2.0 1.5 1.3 10. Betting 7.6 9.0 10.7 9.1 10.2 8.5 10.5 8.1 6.4 4.3 3.1 11. Gasoline ($0.04/Gal) 7.0 7.1 5.8 5.3 4.2 3.5 3.0 2.5 -2.1 1.6 1.3 12. Foreign Cigarettes 0.0 8.5 4.8 1.5 0.1 0.0 0.5 3.1 2.8 5.0 9.3 13. Others 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 14. LOCAL NON-TAX REVENUE 8.1 12.6 8.3 5.6 10.8 8.9 6.8 7.3 10.4 21.3 43.8 15. Sanitation Licenses 0.5 0.6 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 16. Nomenclatura Urbana 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17. Laboratory Services 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 18. Interest Receipts 0.0 0.2 0.0 0.0 0.1 1.8 1.3 6.2 4.3 10.8 39.8 19. Fines 1.6 1.7 4.2 3.9 5.2 5.4 3.4 0.0 4.5 9.4 0.0 20. Confiscated Land Sales 0.1 0.2 0.1 0.1 0.0 0.0 0.0 0.1 0.0 0.0 0.0 21. Rental Receipts 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.0 22. Real Estate Sales 0.0 0.2 0.0 0.0 0.0 0.0 0.1 0.1 0.1 0.0 2.4 23. Others 5.7 9.6 3.8 1.6 5.4 1.5 2.0 1.0 1.5 1.1 1.7 24. SHARED TAXES 235.2 284.7 255.5 224.0 233.2 259.7 205.. 191.6 199.1 210.3 222.6 25. Sales Tax 13.9 31.1 36.4 21.5 37.7 62.6 31.0 17.6 28.8 16.1 23.3 26. Beer Tax 186.0 204.6 177.2 150.2 148.8 157.9 138.1 138.6 138.1 161.0 185.7 27. Gasoline ($0.06/Gal) 10.0 10,1 7.9 6.4 6.1 5.7 4.5 3.6 3.1 2.1 2.1 28. Other 22.3 22.1 19.1 20.7 23.7 16.1 17.0 16.6 18.1 19.6 21.0 29. Tobacco Tax 3.0 16.9 14.9 25.3 16.8 17.4 14.9 15.3 11.0 11.5 11.0 30. GRANTS 73.3 80.0 69.1 86.1 82.0 117.8 97.4 0.0 0.0 0.0 1.8 31. For Teachers Salaries 0.0 0.0 0.0 0.0 0.0 0,0 0.0 0.0 0.0 0.0 0.0 32. For Regional Ed. Fund 71.9 78.5 69.0 86.1 81.5 113.8 97.4 0.0 0.0 0.0 0.0 33. For School Cafeterias 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 34. For Health Services 1.0 1.5 0.1 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 35. For Social Welfare 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 36. For Econ. Dev. Works 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 37. From Decent. Agencies 0.0 0.0 0.0 0.0 0.5 3.0 0.0 0.0 0.0 0.0 0.0 38. CAPITAL RESOURCES (DEBT) 15.9 22.4 3.1 13.4 39.1 91.0 39.0 7.6 7.6 8.8 286.0 39. Loans 15.9 17.7 3.1 13.4 39.1 91.0 39.0 7.6 7.6 8.7 216.7 40. Bond Issues 0.0 4.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 69.4 41. TOTAL REVENUES 647.2 697.8 612.4 606.0 605.2 758.4 618.4 467.3 496.7 558.6 917.1 WPS/BR-001/2.12.85 - 23 - decline in real per capita receipts of District taxes in the last two decades is almost entirely due to the decline in real per capita property tax receipts. The effective rate of property taxation, as measured by the ratio of property tax collections to estimated market value of real estate in Bogota, has always been low when compared with the effective property tax rates com- monly encountered in the United States, where it is generally greater than 1 percent. In Bogota, it has been estimated that in 1972 the effective property tax rate in Bogota did not exceed 0.42 percent, and may well have-been sub- stantially lower. It is also of interest to observe that in 1976 per capita revenues from property taxation in Bogota (Col. $147) where almost 20 percent below those in Cali (Col. $176), again confirming the relatively low effective property tax effort in Bogota, even when compared with another Colombian city. For 1972 the factors explaining the poor property revenue performance were carefully analyzed: A low statutory tax rate (0.84 percent for most properties) and a low collection rate (only 68 percent of property tax liabil- ities were actually collected) were the major culprits, with tax exemptions and poor assessment contributing to a lesser extent. According to preliminary analysis, the collection rate improved between 1972 and 1979, reaching 84 percent during the latter year, while the percentage of exempt properties declined. The reduction in real per capita property tax collections in recent years accordingly was caused by a slow growth in the tax base, i.e., the assessed value of District properties. Considering this tax base, the nominal assessed value of properties listed in the computerized files of the District Computer Agency (SISE) tripled between 1974 and 1979, while the number of properties doubled during WPS/BR-001/2.12.85 - 24 - the same period (Table 1.8). However, if one allows for inflation and in- crease in population, the real per capita property value declined during these years (Table 1.8). Since the "City Study" has elsewhere estimated that the real value of urban land in Bogota increased during the same period and more- over a positive income elasticity of demand for housing has been estimated, one must conclude that between April 1974 and February 1979, the assessment ratio (as measured by the ratio of assessed to market value of properties) declined in Bogota. A separate statistical analysis was carried out to investigate the relationship between market value of land and assessed value (as shown in the land value map of the District Cadastral Office) for a sample of properties in 1976. The results indicate that the assessed land values on average are above market value. When analyzing the ratio of assessed value to market value by location, it is found, moreover, that in the Center the assessment ratio is substantially greater than unity, while it is less than one at the periphery. This probably reflects the fact that land values have been in- creasing at the least rapid rate in the center and most rapidly at the periphery, but that efforts to reassess property values in recent years have generally concentrated on the central areas of the city. One additional conclusion emerges from the high average assessment ratio of land as shown on the Cadastral land value map: To the extent that properties have been severely underassessed in Bogota this is not a result of poor land assessment techniques. Rather, it is likely to be the result of the failure to (a) incorporate land values as estimated for the land value map into the computerized property value files of SISE used for tax billing; (b) to value adequately the improvements on land (buildings in particular); and WPS/BR-001/2.12.85 - 25 - Table 1.8: BOGOTA: CADASTRAL DATA 1974 (April) 1979 (February) Number of properties 224,520 421,488 Number of people per property 13.2 8.7 Property value (in millions of current pesos) 35,600 100,966 Ptoperty value (in millions of pesos at 96,038 100,966 constant 1979 prices) Per capita property value (in constant 1979 pesos) 32,333 -30,295 Average property value (in current pesos) 158,560 263,272 Average property value (in constant 1979 pesos) 427,748 263,272 Source: Calculated from data provided by SISE. WPS/BR-001/2.12.85 - 26 - (c) to update property files to reflect changes in land use (subdivision, change from residential to commercial/industrial use) or improvements in public infrastructure. The District City Council in 1980/81 debated far- reaching reform proposals to improve the city's Cadastral system. The most important changes which those proposals should lead to are (a) increased efforts and resources to update and maintain the property registration and valuation system, and (b) integration of various duplicate archival systems for property registration and valuation data in Bogota, which currently lead to extensive problems in tax and user charge collection. The Incidence of the Property Tax. In popular and academic discussions of property tax reform, the assumption is often made that the property tax is regressive. In fact, however, the "City Study" concludes that the incidence of an increase in the property tax on a nation-wide basis is progressive, while a property tax increase in the Special District of Bogota, relative to the average level of property taxation in Colombia is likely to h.ve an approximately neutral distributive impact. These conclusions are based on an analysis of the property incidence in Colombia, which includes an in-depth case study of Cali and which allows for possible biases in property tax administration, as introduced for example by differential assessments, exemptions, or tax collection. It is concluded, moreover, that the total or partial exemption of properties with low value, as has been applied in Bogota and Cali, results in a more progressive property tax than in the absence of such exemptions without resulting in a significant revenue loss. 1/ 1/ This exemption practice is comparable to the "life-line" tariffs for public utilities, which are further discussed below. WPS/BR-001/2.12.85 - 27 - Furthermore, property value assessment practices were found to be biased in favor of lower income areas where land tends to be underassessed and against high-income areas, where land values tend to have been overassessed. Other District Taxes In addition to the property tax, the District relies on a number of other taxes, among the most important the industry and commerce tax. Also of special note are the motor vehicles tax and an urban construction tax. The Industry and Commerce Tax. Until 1974 this tax showed a declining revenue trend when measured in per capita terms and in Eonstant prices. Since the fiscal reform of the District in 1974, when this tax was converted from a complex set of specific taxes on commerce and industry to a relatively simple tax on gross sales or business activity, the industry and commerce tax has been transformed into one of the most buoyant District taxes (Table 1.7). This growth is due to the fact that the tax is now ad valorem and thus grows automatically with the growth in economic activity and with inflation. Moreover, in recent years, a special effort has been made to reduce tax evasion. The Motor Vehicle Tax. Until 1976 the real per capita revenues on account of this tax declined, but almost tripled between 1976 and 1980. Nevertheless, they remained below the level achieved in 1965 (Table 1.7), despite the fact that the number of automobiles during that period has increased quite rapidly. In some other large cities of the Third World, particularly in Jakarta (Indonesia), motor vehicle taxes have been a very important and buoyant revenue source, and have contributed to an equitable and efficient financing of urban services. Efforts probably could and should be WPS/BR-001/2.12.85 - 28 - made in Bogota to strengthen the administration of this term and to increase the effective rates of motor vehicle taxation. The Urban Construction Tax. The Urban Construction Tax (Impuesto de Delineacion Urbana) experienced a rapid increase in revenue in 1980 due to higher rates and greater collection efforts. The tax is levied on new construction carried out with formal permits as obtained from the District Planning Department. Compared to the general property tax, the construction tax is undesirable for the following reasons: First, it provides a disincen- tive for the investment in urban construction and improvement. SJch a disincentive has, in fact, been always the reason which economists have given in favor of a pure site value tax, as compared with a tax on land and improve- ments. Second, the tax is likely to act as a disincentive to obtaining building permits, contributing to the extensive illegal and unregistered building activity in Bogota, which in turn results in added difficulties for Cadastral authorities, who, to a large extent, have to rely on building permits to signal when improvements are being made on a property. V. USER FEES AND DEVELOPMENT CHARGES Revenues from user fees and development charges ("self-financed service revenues") have contributed the most important means of public resource mobilization for the financing of urban services in the Special District of Bogota. Between 1961 and 1971 real per capita revenues from this source doubled, but between 1971 and 1979 they showed no longer a rising trend (Table 1.4). In analyzing these trends, it is useful to consider first the revenues derived from water, power and telephone services, and then the financing of road infrastructure investment through the valorization system. WPS/BR-001/2.12.85' - 29 - Water Supply, Electric Power, and Telephone Services Financial Considerations. All three services have been operating on a financially self-sufficient basis without receiving, or paying transfers from or to any other public agency. However, since for all three services, unit costs per connection and per quantity of service increased, the three agencies providing these services (EAAB, EEEB, and ETB), encountered financial difficulties at various times wien their service charges were not regularly adjusted to keep up with rising real costs as well as inflation. -This diffi- culty was particularly troublesome for the Water and Sewerage Company (EAAB), which at various times had to finance operating deficits by borrowing. Automatic monthly increases of small percentages are now the rule for water and sewerage charges and for electricity charges, providing at least a partial solution to the financial problem. Efficiency Considerations. Public utility service charges have generally been set in Bogota without reference to the marginal costs of pro- viding these services. In recent years this apparently has led to consumption charges for water above marginal costs, while the reverse has been the case for electricity. Estimates have put average incremental cost (AIC) of water in Bogota in 1978 at about US$0.06 per cubic meter of water while consumption charges (i.e., charges per cubic meter of water consumed) amounted up to 360 percent of AIC for residential users and up to 470 percent for industrial and commercial users. For electricity, average marginal cost in Bogota in 1979 were almost twice the average tariff. Thus, consumption of water was unduly restricted, while consumption of electricity was unduly encouraged. WPS/BR-001/2.12.85 - 30 - In addition, the structure of marginal costs has not been adequately reflected in utility charges in Bogota. For the case of electricity, high- voltage industrial and commercial consumers have been charged with higher rates than residential consumers, although relative marginal service costs are just the reverse. For water supply, a supplementary pumping fee has been applied in areas of the city where pumping is required; however, this fee apparently has been set without reference to the actual differential in marginal cost between water service with the without supplementary pumping. To some extent, financial self-sufficiency of the public-utility agencies accounts for the failure to set prices according to marginal cost. However, the financial goals in principle could more appropriately be pursued by setting the consumption charges according to marginal cost, while setting fixed monthly charges so as to satisfy the financial requirements. In the case of water this would mean that fixed monthly charges would be set above the monthly customer costs (metering, billing, etc.), while the reverse would be done for electricity. Alternatively, it would be possible--again, in principle--to transfer financial surpluses generated by marginal cost charges for electricity to the water company to cover the financial deficit resulting from marginal cost charges for water supply. Transfers of this type have in fact taken place in some recent years in Cali, where the public utility services are provided by a single agency. In Bogota, this solution in prac- tice would be difficult to implement given the strict separation of public utility accounts. Equity Considerations. Public utility pricing in Bogota (as throughout Colombia) has been quite heavily influenced by income distribution considerations. Water and telephone charges are progressively linked to WPS/BR-001/2.12.85 - 31 - property value, while residential electricity charges increase in blocks with increased consumption per connection. Thus high income consumers who tend to live on higher value properties and consume relatively more electricity than poor consumers have tended to cross-subsidize the low income consumers. Overall, estimates of the distributive impact of these cross-subsidies is not very substantial in equalizing incomes; however, it does provide some relief from user charges to the low income groups. Special consideration needs to be given to the distributive impact of connection fees, which can be quite substantial. To the extent that those households which are not connected to the service system tend to be among the lowest income groups, subsidization should focus on lowering the costs of connection, rather than on cross-subsidies between existing consumers. To some extent this has been done in Bogota by linking connection fees to property value, and, since late 1980, low value properties have been entirely exempted from parts of the connection fee for water. Valorization The valorization system has been designed to recoup the cost of investments in road infrastructure from the owners of properties in the vicinity of the project. It has also been applied to finance the secondary and tertiary sewerage collector network in Bogota. The method of distributing the costs among property owners is quite complex and relies basically on a number of actors designed to reflect the relative benefit received by each property from the project (e.g., distance from the work, shape and size prop- erty, topography, etc.) as well as ability to pay of the population living in the area. WPS/BR-001/2.12.85 - 32 - In Bogota, the valorization system has been extensively utilized in past years, especially between 1967 and 1971. In 1968, expenditures on valoriation works accounted for 16 percent of all consolidated District expenditures, while in 1969 revenues from valorization contributions amounted to more than two-thirds of property tax revenue (Table 1.9). During those years, the valorization system assisted substantially in financing a rapid expansion in the road network of Bogota, cortributing extensively to the good road access of most neighborhoods in Bogota, including those in the poorer areas of the city. I During the 1970s, however, a dramatic decline occurred in the use of the valorization system. In 1976, the Urban Development Institute (IDU), whose primary function lies in the execution of valorization-financed projects, stitl contributed 8.6 percent to total consolidated District expenditure; in 1979, its share had dropped to 3 percent. Revenues on account of valorization in 1971 represented about 5 percent of the consolidated current revenues of the District; in 1979, they amounted to only 2.5 percent. Among the principal reasons for this decline were the following: (a)absence of integrated plans of urban service expansion which could have been financed by the valorization system; (b)legal norms which prohibited the concurrent coverage of one property by valorization charges resulting from two or more independent projects; (c)1ow interest charges and low penalty charges for term payments and arrears, respectively; (d)substantial cost overruns due to delays in project executioin, frequently caused by difficulties in land acquisition for construction; WPS/BR-001/2.12.85 - 33 - Table 1.9: VALORIZATION ACTIVITY IN BOGOTA, 1959-74 Valorization Valorization Total per Capita Expenditure as Contributions as Valorization Percentage of Percentage of Expenditures in Local Authority Property Tax Year Prices (Pesos) Spending (%) Revenues (%) 1959 16.4 n.a. n.a. 1960 14.8 n.a. n.a. 1961 17.1 n.a. 5.9 1962 10.8 n.a. 19.5 1963 12.0 3.2 41.9 1964 14.1 5.8 20.7 1965 12.4 5.5 21.6 - 1966 13.3 6.6 29.7 1967 36.4 14.3 59.2 1968 49.1 15.9 60.2 1969 35.3 10.3 66.9 (68.0) 1970 21.7 6.8 62.0 (75.4) 1971 17.8 5.1 44.6 (57.6) 1972 15.2 4.2 31.0 (38.8) 1973 9.8 n.a. 22.0 (27.5) 1974 16.9 n.a. 18.2 (34.1) Average 1961-1966 13.3 5.3 23.2 1967-1970 35.6 11.8 62.1 (65.7) 1971-1974 14.9 4.7 29.0 (39.5) Source: Doebele, Grimes and Linn, op.cit. WPS/BR-001/2.12.85 - 34 - and a failure to pass these cost overruns on to project beneficiaries by recalculating valorization charges; (e) failure of non-exempt public agencies to pay their valorization debts; and (f) termination in 1976 of all transfers from and tax-sharing with the District Administration, which had been intended originally as a safety cushion for IDU to provide it with a financial buffer against liquidity losses and failure to cover all costs, especially in poorer neighborhoods. Efforts were underway in 1980/81 to begin to deal with many of these difficulties. To what extent they can be remedied by legal or administrative reform remains to be seen. As of 1981, it appeared that the valorization system which commanded considerable attention by the city's top administrators in the 1960s, but lost much of it during the 1970s, was still not given the priority it deserves as an importance local revenue instrument. VI. REVENUE SHARING There exist, broadly speaking, two types of revenue sharing in Bogota: first, there are those shared taxes and transfers which enter the budget of the District Administration, such as the shared beer tax and sales tax; and second, there are those national transfers which support the Regional Education Fund for Bogota (FER) and the Regional Health Service of Bogota (SSB), agencies jointly administered by their respective national ministries, and by the District government. WPS/BR-001/2.12.85 - 35 - Revenue Sharing in the District Administration The revenues received on account of the shared been tax are the most important single item of current revenues received by the District Administration; beer tax revenues exceed revenues from this property tax and from the industry and commerce tax (Table 7 above). In terms of its incidence, this tax is probably more regressive than the other main District taxes, but it is relatively simple to administer and historically had experienced quite a high buoyancy (though this revenue performance was not maintained during the 1970s, when real per capita revenues actualty declined somewhat). Shared sales tax revenues, which had shown a good growth tendency in real per capita terms until 1976, thereafter diminished in terms of revenue directly accruing to the District Administration, because of the nationaliza- tion of education which resulted in the direct transfer of a substantial part of the shared sales tax revenues to FER. The remaining shared sales tax revenues were subject to quite large and unpredictable fluctuations, caused to a large extent by the failure of the national government to effect the trans- fers of funds in a regular and timely fashion. The nationalization of education also resulted in the elimination of the transfers from the so-called "Situado Fiscal" (best translated as "general revenue sharing"). These funds had been destined to the payment of teachers' salaries, and as of 1981 were also paid directly to FER. Figure 1.7 shows graphically the development of transfers and shared taxes in Bogota between 1969 and 1979. The broken and dotted lines after 1976 indicate the develop- ments of receipts accruing directly to the District Administration on account of the "situado fiscal" and the shared sales tax, respectively. - 36 - Figure 1.7 Per Capita District Revenues from Transfers and Shared Taxes, Adjusted for the Transfers Received by F.E.R. and S.S.B. (in 1976 prices) 250- 250 200- -200 150 - -150 0 100 - -100 50- 1-50 0 0 '69 '70 '71 '72 '73 '74 '75 '76 '77 '78 '79 Year Legend: -BeerTax -OWL - Situado Fiscal World Bank-26168 - Sales Tax WPS/BR-001/2.12.85 - 37 - The Financing of Education and Health Although the nationalization of education brought substantial changes in the financial arrangements of this service in Bogota, as mentioned above, the District Government as of 1981 retained important responsibilities in the education sector, as well as the administration of health services. The financing of these services therefore has remained of considerable importance for the District. The national transfers and shared taxes for health and education services, in particular the shared sales tax and the "siduado fiscal" grew rapidly during the last decade, despite considerable fluctuations (Figure 1.7). In 1976 the transfers on account of the "situado fiscal" exceeded even the revenues derived by the District from its shared beer tax. The rapid growth of "situado fiscal" revenues is based on the fact that the transfers are calculated as a percentage of Central government tax revenues, which have been growing more rapidly in recent years in Colombia than subnational taxes. The formula which is used to distribute the revenue sharing pool of "siduado fiscal" between the various subnational jurisdictions in Colombia has given Bogota a share that has been Lower than its share in the Colombian population, a share which moreover has been declining over recent years. This bias, to a certain extent has been reduced by supplementary transfers during the fiscal year which were not determined by the basic dis- tribution formula. The net effect, however, remains one of (partial) equalization of subnational government resources through this revenue sharing arrangement, as was its original intent. Transfers to FER can be law only be applied towards teachers' salaries. The financing of school equipment, teaching materials, text books, WPS/BR-001/2.12.85 - 38 - etc., has to rely on extremely limited resources, and non-teachers' salary ex- penses therefore appear to have suffered in recent years in Bogota. In the health sector, in contrast, no similar restrictions have been placed on the use of transfer funds and as a result more balanced expenditure pattern be- tween personnel and support services appear to have occurred. VII. BUDGETING AND INVESTMENT PLANNING IN BOGOTA Budgeting and investment planning in Bogota proceed in a highly decentralized manner, reflecting directly the high degree of admiffistrative decentralization. In the absence of general urban planning and of systematic methods for economic evaluation of investment projects and programs, the coordinating agencies exert only a very limited influence over the budgeting and programming activities of the many separate District agencies. The preparation of the budget for the District Administration is largely in the hands and under the control of the executive branch under the direction of the Major and his staff, rather than a prerogative of the City Council which has only quite limited scope for modification. Budget execu- tion, moreover, is marked by frequent changes and additions to the initial budget during this fiscal year, thus further eroding the influence of the Council as well as the importance of the budget as a fiscal planning device. In the execution of its budget, the District Administration has also been hampered by complex disbursement methods, which have led to serious delays in payments to contractors and in underspending of the District's investment budget. The fiscal control and audit, which is in the hands of the District Controller, suffers from excessive complexity and serious other shortcom- WPS/BR-001/2.12.85 - 39 - ings. Rather than employing selective, random ex post audits, the Controller's office employs a large staff to carry out complete ex ante and ex post audits, implying that a representative of the controller is involved in every financial transaction as it takes place. This dilutes the responsibil- ity of the financial officers of the executive branch and implicates the auditors directly in the transactions, making completely independent ex post audits an impossibility, quite apart from the fact that it requires many more staff resources than selective ex post auditing would require. What is more, no performance audits appear to be carried out to ascertain whethEr programs have actually been carried out in accordance with their original design and have achieved their intended objectives. No systematic city-wide investment programming, capital budgeting, or integrated area development--covering more than one service at a time in a particular neighborhood or area--appear to be in use. This has resulted in poor timing or sequencing of investments by separate service agencies, has resulted in serious engineering planning errors, 1/ and in difficulties with cost recovery through the valorization system. What is more, no systematic effort eve seems to have been made to begin to evaluate the relative needs of different public service sectors for further investments. 1/ For example, one overpass, which was designed and constructed by IDU, had to be completely redesigned while construction was already under way, because it was discovered during construction that a major set of supporting columns would have been placed right on top of a water main, had the original design been followed. WPS/BR-001/2.12.85 - 40 - VIII. CONCLUSIONS The ur.ban government of Bogota is best characterized as one that has interfered relatively little with the private market forces which have given the major impetus to the development of the city, both in terms of the growth of its economic base, as well as in terms of the land use patterns in the -city. Zoning regulations and building codes appear to have been largely honored in the breach. Illegal "pirate" developments (but not squatting on private land, so-called "invasions") have generally led, rather than followed the provision of public infrastructure, which--compared with many other cities of similar size in the Third World--has been quite successful. Most households have electric power, potable water and sanitary sewerage facilities in their homes and have good access to public transportation, which in turn is relatively cheap and efficient. With all its apparent difficulties the decentralized service provision in Bogota may well account for this particular pattern of public urban management. The diffusion of responsibility and authority has made a forceful application of regulatory control very difficult, if not indeed impossible. The existence of financially and administratively autonomous agencies quite likely contributed to a substantial degree of self-financing of infrastructure services and thus the relatively extensive provision of services throughout the city, including the extensive road network. What is more, in the absence of a financially and administratively strong central authority in Bogota, private mass transport was never seriously in danger of being taken over by the local government and thus continued to expand its service along with the growth of the city. Finally, given the relatively WPS/BR-001/2.12.85 - 41 - strong tradition of independent local government in Colombian cities, the national government--which itself is characterized by a highly decentralized administrative structure--has never intervened directly on any large scale in the provision of urban services or the management of urban growth. All in all, Bogota probably has not been badly served by its particular brand of urban government. There are, of course, areas where the analysis carried out for the "City Study" suggest considerable room for improvement. Among these is the improvement of the city's local tax structure, in particular, of the property tax and the motor vehicl-e tax. Improvements might also be made in the pricing of public utility services, particularly water supply and electric power. Reviving and improving the administration of the valorization system could contribute to a more effective provision of urban infrastructure. Social service provision, particularly in the public health and education fields, could strengthened by a more flexible financing scheme and a more systematic evaluation of the needs of the poorer neighbor hoods. Finally, there is room for improvement in the financial man- agement practices in the city--budgeting, accounting, forecasting, investment planning and auditing--even while maintaining its highly decentralized urban management structure. The remaining papers in this volume address some of these issues in some detail, while others have been addressed elsewhere as in- dicated by the references listed below. WPS/BR-001/2.12.85 - 42 - References W. Doebele, 0. Grimes and J. Linn, "Participation of Beneficiaries in Financing Urban Services: Valorization Charges in Bogota, Colombia." Land Economics, Vol. 55, 1979. Caroline Fawcett, "The Revenue Sharing System in Colombia: A Case Study of Bogota, D.E." Chapter 4, below. Jeffrey D. Lewis, Alberto Hernandez G., Marco Tulio Ruiz, S., "Finanzas Publicas de la Administraction Central en Bogota, Colombia: Un Resumen Estadistico." City Study Intermediate Paper No. 25, May 1979. Johannes F. Linn, "Urban Finances in Developing Countries." In Urban Government Finance: Emerging Trends, ed. Roy W. Bahl, Beverly Hills: Sage Publications, 1980. , "Public Utilities in Metropolitan Bogota: Organization, Service Levels, and Financing". Urban and Regional Report No. 78-9. World Bank, 1976. , "Estimation of Water Supply Costs in Cali, Colombia". Urban and Regional Report No. 76-14. World Bank, 1976. , "The Incidence of Urban Property Taxation in Colombia". In The Taxation of Urban Property in Less Developed Countries. Edited by Roy W. Bahl, Madison: University of Wisconsin Press, 1979. , "Property Taxation in Bogota, Colombia: An Analysis of Poor Revenue Performance". Public Finance Quarterly, Vol. 8, October 1980. M. Wilhelm Wagner, "Vacant Lots, Land Prices, and Assessment Practices: The Case of Bogota, Colombia". City Study Project Paper No. 17. Urban and Regional Report No. 81-14. World Bank; 1981. Wolff, J. H., "Budgeting and Investment Planning in Bogota, Colombia." Chapter 2, below. WRA/BR-001-2a/3.15.85 - 43 - Chapter 2. BUDGETING AND INVESTMENT PLANNING IN BOGOTA, COLOMBIA by J. H. Wolff* I. Introduction and Summary The Government of Bogota, capital of the Republic of Colombia and of the Department of Cundinamarca differs somewhat from that of other departments and municipalities in Colombia: it is a Special District (Distrito Especial, usually abbreviated D.E.) rather than a municipality forming part of a depart- ment. The Mayoz of Bogota (Alcalde Mayor) is directly nominated (and removed) by the President, without intervention of any governor. 1/ In 1954 the former Municipality of Bogota (which had been given some special taxation rights but had remained part of the Department of Cundinamarca) and several adjacent municipalities were joined together as the Special District of Bogota. Except for its "independence" from the Department of Cundinamarca (particularly the direct nomination of the Mayor by the President), the internal adminstrative The author wishes to express his sincere gratitude to the several current or former public officials of the District, of the national administrative agencies, and of decentralized public enterprises that offered their time and energy for interviews carried out in Bogota in March and April 1980. Richard F. Bird, Caroline Fawcett, Johannes F. Linn and Jose Fernando Pineda all have carefully read the draft of this paper and provided'a number of most valuable criticisms, suggestions, and additions. These definitely helped clarify my ideas and avoid several mistakes. Finally, Mary Ann Heraud ably processed the manuscript. As usual, any errors entirely rest in the responsibility of the author. 1/ The fact that the mayor has no fixed term of office represents one of the most severe impediments to any long-term policy and planning. Over the last 70 years, the average term of office of Bogota's mayors has been slightly more than a year. WRA/BR-001-2a/3.15.85 - 44 - structure remained basically that of a municipality. 1/ Only in 1972 did the city govGrnment attempt to decentralize its administration by establishing the 16 English Alcaldias Menores, which are directly under the control of the Alcaldias Menores. These offices were politicized from the very beginning, however, and not given sufficient tasks and administrative means. The purpose of this paper is to describe and evaluate the budgetary process and investment planning at the level of the City Government of Bogota. After a short outline of the system (Section II), the bufeaucratic and political network controlling the budget is described (Section III). After a description of the budgetary process itself (Section IV), two sections on public credits and investment investment planning (V, VI) follow. Budgeting in Colombia is a rather formal process of compiling pro- jected income and expenditure on the basis of prior legislation. It is therefore important to examine the basic decisions, the most important of which are investments at the District level (hence including the decentralized agencies). Since the city is also under close supervision of the national government, it is unavoidable that several national authorities are also in- volved in these decisions. The principal findings of the study are that budgetary and investment planning above the level of the single agency does not exist (Section VII); that this is not merely a technical defect, but is intimately related to the political structure of the country (Section VIII); and that reform measures must take these political factors as given, thus 1/ The basic legal source for the City's current organization and prerogatives is national decree 3133/1968, one of the results of the constitutional and administrative reform of that year. The majority of national legal sources may be found in Ortega Torres (1976). From time to time, proposals for changing the legal status of Bogota are discussed. WRA/BR-001-2a/3.15.85 - 45 - being content with relatively small modifications in existing practices rather than fundamental changes (Section X). An important limitation of .his study is the fact that it concentrates solely on the interactions between agencies (with the exception of the administrative units in charge of the budget). The agencies themselves, however, are treated "black boxes." Budgetary and investment procedures are thus described only after the various administrative units and public utility agencies present their requests. The remainder of-this section summaries the main findings of this paper. Bogota's budgetary system is almost pure cash, i.e., income and expenditure are accounted when actual movements of funds occur. The bureau- cratic network responsible for the budget includes a central administration as well as several public utility corporations, commercial firms, special funds, and decentralized units, (decentralizadas) 1/ most of which possess administrative economy. These entities generate their own revenues and there- fore operate on separ&te budgets, the total of which far exceeds the central budget. The central units of and coordination to the politically much weaker than the more important of decentralized agencies. The District Treasury and the Comptroller's Office act as agents of the City Council which, conse- quently, has some influence also in the executive realm. However, the Mayor does not depend on the political confidence of the Council. For international loans, the higher authorities of the national government (Finance Ministry and National Planning Department) come into play. 1/ in this paper, the decentralized city agencies are often referred to by their Spanish name, "decentralizadas." WRA/BR-001-2a/3.15.85 - 46 - The Directorate of the Budget (an agency of the city's Finance Secretariat) actually prepares the central budget of current expenditures; the District Planning Department formulates the budget for investment expenditures, based on the initiatives of the decentralizadas. Approval by the Planning Council, a mixed executive-legislative body, and by the City Council is basically a formality. The City Council directly determines some minor expenditures (auxilios) that are used by individual councilors to obtain or augment political support. The Budget Directorate also oversees the actual expenditures during the fiscal year. Additions and amendments to the original budget are frequent. Investments are often characterized by marked underspending. Payments out of the Treasury are extremely complicated, and are therefore often delayed for months. The auditing system usually requires more than one control on the same procedure and considerable paperwork. Political patronage by City Councillors has resulted in featherbedding in the Comptroller's Office and among the (independent) officials in financial control of the Water, Telephone and Electricity Companies. Auditing activities therefore concentrate on numerical legal aspects of the budget, neglecting political and administrative controls. Loans and credits, traditionally vital for the financing of public investments, are subject to control mechanisms at both the District and national levels. Although the involvement of planning authorities in these investment decisions is ostensibly to guarantee conformity with overall development goals, comprehensive planning does not exist. Credit decisions are thus made on the basis of fiscal criteria. Duplication of control also WRA/BR-001--2a/3.15.85 - 47 - occurs at the Finance Ministry. A number of other organizations and councils must approve, more or less formally, every credit raised by an organization in the Special District of Bogota. Much delay therefore occurs before a credit can effectively be contracted. Bogota's budgeting process is seriously fragmented, much like political power in the District. Streamlined budgeting and comprehensive planning neither exist nor would be feasible in Colombia, where the political system generally avoids making irreversible decisions. "Muddling-through" is preferred to technical "ideal" or even "rational" long-range decisions on development policy in this rather volatile society. The lack of coherent and continuous social policy is also reinforced by the rapid turnover of govern- ment personnel. Incrementalism is thus the order of the day, and is unavoidable unless a basic change of the present political system occurs. Given Bogota's political realities, recommendations for improving procedures include the following reforms: exclusion of the National Planning Department from the process; abolition of one of the two authorizations required of the Finance Ministry to engage in credit negotiations and to sign contracts; and major reorganization of personnel and procedures in the Comptroller's Office. In addition, a comprehensive analysis of intergovern- mental fiscal relations and of the revenue structure of the District should be undertaken. International lending organizations, in any case, must treat the political constraints as given and concentrate on the economic and social quality of individual projects and program. WRA/BR-001-2a/3.15.85 - 48 - II. THE BOGOTA BUDGETARY SYSTEM AND FORMAT Unlike the national budget that adheres to a modified accrual system ("sistema de ejercicio"), the Special District's system is almost a pure cash- flow budget. 1/ Only actual revenue and actual expenditures during a fiscal year (i.e., the calendar year) are counted; appropriations not spent by December 31 do not simply accrue but must be explicitly included in next year's budget; expenditures for bills received but not paid during a fiscal year, as well as investment expenditures, are included as special items (vigencias expiradas) in the subsequent budget. 2/ The decentralized agencies of the District government also operate under this type system. A pure cash budget, of course, lends itself to manipulation. By delaying payments for services received, an artificial surplus can be created when there is actually a deficit. This is a serious temptation given that a number of prescriptions prohibit deficits, that the term "surplus" has positive political associations, and that the mayors and the heads of the decentralized agencies change so frequently that often the successors must bear the consequences. Furthermore, for business ventures such as the District's major public utility firms (telephone, water and sewerage, and electricity), a pure cash flow budget is insufficient to allow thorough 1/ Arts. 31, 40 C.F.; Art. 66 Decreto 3133/1968. 2/ More precisely, the Comptrollers Office (Contraloria) books only real expenditures (gastos efectivos), whereas the Budget Directorate includes reserves (Contractual obligations for payments); actual payments plus reserves are therefore considered actual expenditures. WRA/BR-001-2a/3.15.85 - 49 - economic analysis. No wonder, then, that these firms use additional budgets for internal purposes. 1/ Although the District budget documents contain a number of summary tables at the central level 2/ (and, in less detail, at the decentralized level 3/) indicating the economic and functional distribution of expenditures, the main criterion for budgetary classifications is the prevailing administrative structure of the District government. Consequently, there are two important sets of budgets: the central budget (comprising tlie "central" administration including the District Council, and its dependencies); and the budgets of the decentralized agencies of the District 4/ and of some separate financial funds ("fondos rotatorios"). On the income side, ordinary revenue (ingresos corrientes) is distinguished from capital revenues (recursos de capital). The first comprises tax and non-tax income such as interest receipts, fines, shared taxes, and transfers; the second type consists largely if credit income. One major category of current expenditures includes salaries, general administrative costs, expenditures for supplies, and transfers (appearing only in the budget of the Finance and Education Secretariats); the second category includes investment expenditures. 1/ The American consulting firm Peat, Marwick, Mitchell and Co. is actually elaborating a new budgeting system for the Electricity Company (EEEB). 2/ Presupuesto Central 1980: 147 ff. 3/ Presupuesto Descentral 1980: 269 ff. 4/ Transfers from the national government giving Bogota a certain share of sales and consumer taxes paid in the city. Technically, they must be considered taxes as well, although not levied by the District itself. WRA/BR-001-2a/3.15.85 - 50 - In the past, taxes accounted for something more than half the total revenue of the Central District Administration; between 1975 and 1978, however, capital income has varied from 12 percent to 1.6 percent of total revenues. Shared taxes 1/ have continued to increase in importance, accounting for more than one-third of all revenue of the central administration. 2/ While current expenditures usually account for more than 95 percent of all expenditures, 3/ the recent Mayor, Hernando Duran Dussan, has made great efforts to increase the share of investment expendiftures. 4/ The decentralized agencies thus make the bulk of investment expenditures: 38.8 percent of their combined outlays in 1977 were for investments. 5/ It can safely be assumed that this percentage has not varied much subsequently. 6/ 1/ Transfers from the national government giving Bogota a certain share of sales and consumer taxes paid in the city. Technically, they must be considered taxes as well, although not levied by the District itself. 2/ Source for these figures: Castano 1979: 88. 3/ Source: Informe Financiero 1977: 12. 4/ Projected figures for 1979 and 1980 were 15 and 20 percent respectively (Presupuesto Central 1980: 11). These cannot be compared with the above mentioned figures, however, since a wide gap usually exists between budgeted and actual expenditures. 5/ Informe Financiero 1977: the percentage has declined somewhat since 1974 (44 percent); see Castano 1979: 122; for absolute figures corrected for inflation, see Linn 1980: 11 ff. based on Lewis 1979. 6/ For 1980, something more than 50 percent of combined expenditures for the decentralized agencies is to be spent for investments (Presupuesto Descentral 1980: 270); however, under-spending is the usual result. WRA/BR-001-2a/3.15.85 - 51 - Earmarked taxes or other ordinary earmarked revenues, at the central level, no longer exist 1/ with the exception of taxes establishing by previous laws and District ordinances (acuerdos) for specific development programs. Since there is no consolidated budget, 2/ such a prescription appears somewhat theoretical: in effect, the revenue of the Descentralizadas are earmarked for special purposes, namely, to finance the particular activities of that agency. The same holds true, mutatis mutandis, for the principle of the unit of accounts ("unidad de caja"). 3/ III. BUDGETING: THE ORGANIZATION The Bureaucratic Network of Bogota's Central Adw-nistration The mayor acts as head of Bogota's public administration executing the ordinances (acuerdos) adopted by the City Council, he also acts as the President's direct agent and is head of the police (the administration concerning public order and security) in the Distrito Especial. 4/ The mayor's cabinet consists of five secretaries, i.e., Interior Administration (Gobierno); Finance (Hacienda); Education (Educacion); Health (Salud); and Public Works (Obras Publicas), a structure resembling that of government at all administrative levels in Colombia. The secretaries are often political 1/ Art. 37 C.F. All revenues enter a common fund; all expenditures are paid out of this. 2/ For a discussion, see Informe de Tesoreria 1979a: 10 ff. 3/ Art. 38 C.F. 4/ Alcaldia Mayor 1972: 40. The big units of the central administration of the district did not undergo any significant change since 1968. WRA/BR-001-2a/3.15.85 - 52 - appointments, although they are not drawn from the City Council, which in turn has no legel influence upon their nomination. Corresponding to the Administrative Departments (departamentos administrativos) at the national level, four departments at the District level provide technical assistance only, i.e., the Departments for Planning, Transit and Transport, Community Development, and Social Welfare. In theory, non- political experts (career civil servants) who are not part of the mayor's cabinet should head these agencies; in the highly politicized Colombian context, however, no difference exists between the office of secretary and that of an administrative department head. That these positions are political posts may be unavoidable because of the tasks entrusted to the departamentos administrativos: Contrary to popular opinion, transport or planning matters are not technical questions (alone), but are deeply rooted in the political domain. 1/ Three other administrative units represent somewhat special cases: the Office of the Legal Representative of the City ("Personero"), the Office of the Treasurer ("Tesorero"); and the Office of the Supreme Auditor (Comptroller) of the City ("Contralor"). While all three offices depend directly on the city council, two are at least functionally part of the central administration. Finally, while the budget of the city Council is also included in the central budget, the Council must be regarded as a political rather than an administrative institution. 1/ Concerning planning, see Wolff, 1977; for Bogota, Wolff, 1980. WRA/BR-001-2a/3.15.85 - 53 - Between 1970 and 1976, education accounted for about 45 percent of all central expenditures 1/ dropping sharply to 19 percent in 1978 and about 14 percent in the 1980 budget following the nationalization of education in 1975/6 (by Law 43/1975 and Decree 102/1976). The Secretariat of Finance (including all transfer payments and the debt service) was responsible for 47.7 percent of the central budget in 1978 and 52 percent in 1980. The next largest share of expenditures is by the Secretariat of Public Works--about 11 percent in 1980--whereas the other 12 units of the central administration account for well below 10 percent each. The Decentralized Agencies As in the case of the national government, several bureaucratic units at the District level (Entidades decentralizadas) are independent legal entities, enjoy administrative autonomy 2/ and a number of other members 1/ Prior to 1975, the national government transferred funds to the central administration earmarked for teachers' salaries; the sharp drop thus is more of a technical than a real nature. 2/ This representation of the City Council in the different juntas directivas provoked one of the most serious controversies between Mayor Gaitan Mahecha and the city council in 1977. "The council as co-administrator" (or as controlling body) was used to justify all interference in the day- to-day administration although this expression does not appear in the constitution, the Codigo de Regimen Politico y Municipal, or in decree 3133/1968. On September 8, 1977, the City Council voted an acuerdo modifying the composition of the juntas directivas of the water and elec- tricity companies cancelling the membership of the semi-private economic associations and augmenting the number of councilors to be on par with the members of the administration (El Tiempo, Sept. 17, 1977). Gaitan Mahecha objected to the decision as unconstitutional and inappropriate. Finally, the Council of State, supreme court for administrative matters, ruled that participation of the councilors on the board of directors of public cor- porations was unconstitutional (for details, see Concejo 1978: 8 ff.). Since mid-1978, however, practice has simply continued: the subsequent Duran Dussan, has explicitly accepted the participation of councilors and their alternates in the meetings of the juntas (El Tiempo, October 26, 1978). Evidently, some legal loophole had been found to circumvent the WRA/BR-001-2a/3.15.85 - 54 - varying from one case to the next, e.g., high ranking members of the city administration, representatives of national banks in the housing sector, or representatives of the private economy. In addition, the Comptroller (Contralor) of the District also attends the meetings of the juntas with voice but without voting power. Of the 17 decentralized agencies, 1/ the following dispose of the most important budgets in 1980: . the Electricity Company (Col$ 14.1 billion, nearly three-times the central budget); . the Telephone Company (Col$ 6.4 billion); . the Water and Sewerage Company (Col$ 5.6 billion); . the Institute for Urban Development (Instituto de Desarrollo Urbano, IDU; Col$ 1.2 billion); . the Company for Urban Services (Empresa Distrital de Servicios, EDIS; Col$ 0.9 billion); . the District's Social Security Fund (Caja de Prevision Social; Col$ 0.8 billion); and . the Lottery of Bogota (Col$ 0.6 billion). The budget of the Electricity Company alone accounts for nearly 45 percent of the combined budgets of the Descentralizadas, the Telephone Company for 20 percent, and the Water Company for nearly 18 percent. The three largest public utility companies thus dispose of 83 percent of the total judgment of the Council of State. 1/ The proliferation of the decentralized agencies (at both the national and the District level) is at least in part due to the policy of international lending agencies to require occasionally the creation of new infrastruc- ture to administer an important loan and to serve as counterpart organiza- tion. WRA/BR-001-2a/3.15.85 - 55 - budget of the Descentralizadas. If the institute for Urban Development is included, this percentage amounts to 87 percent. The Budget Administration, the Mixed Bodies, and the Comptroller's Office The two major administrative units in the central administration that prepare and execute the budget are the (Secretaria de Hacienda) and the Department of Planning Departamento Administrativo de Planeacion Distrital, responsible for budgeting current and investment expenditures, respectively. The Secretariat of Finance comprises four directorates, i.e., for taxes; property assessment (catastro); 1/ internal administration; and budget. 2/ Preparation and execution of the budget occur at the Directorate of the Budget. Besides two divisions in charge of accounting, the budgetary matters of the District 3/ and of the administration of the District's public debt, are entrusted to two divisions in the Dicectorate of the Budget: The Financial and Fiscal division is essentially in charge of preparing the income side of the budget and of compiling the various budgets of the decentralized agencies; the Execution and Control Division prepares the expenditure side of the budget and supervises budget execution. The first agency has 13 staff 1/ The land-register office has been reorganized by decree 1082/1978. For details, Informe Financiero 1978/9: 55 ff. 2/ The Junta de Hacienda is included in its budget, it cannot be regarded and administrative part of the Secretariat. See below. 3/ According to Art. 212 C.F., the Budget Directorate is in charge of the District's accounting; according to Art. 539, however, the Comptroller's Office has to deliver the general balance ("Cuenta General") of the budget and the treasury, and report on the corresponding state of budget execution. WRA/BR-001-2a/3.15.85 - 56 - members (nine of which are professionals), and the second one has 22 (21 of which are professionals). As stated, the District Planning Office (DAPD) is organized as a Departamento Administrativo, and thus directly linked to the mayor. It is subdivided into five units: one for its internal administration; and one each for Neighborhood Upgrading, Studies and Investigations, Urban Development, and Fiscal and Budgetary Analysis. 1/ In the present context, this last unit is the one of most interest. It comprises the Division of Financial Analysis, in charge of fiscal studies in the widest sense of the term (financial situation of the District; requirements for the financing of proposed investments, studies of tariffs for public set '.ces) 2/; and the Budget Programming and Control Division which compiles the investment budget of the central admin- istration, studies the demands for the investments of the decentralized agencies; and controls the execution of investments duri-ag the fiscal year. 3/ The Treasury (Tesoreria) 4/ is the central revenue collection and management authority of the District. With five divisions employing over 400 officials, its task should be regarded as highly technical rather than political. Historically, however, the head of this authority (Tesorero) is directly elected by the City Council for a period o,: : years (with 1/ Statues underlying today's organization are Acuerdo 1/1975 (of the City Council), regulated by decree 816 of the same year. 2/ Art. Decree 816/1975. 3/ Art. 6 Decree 816/1975. 4/ On administrative structure and functions of the different officials see Estructura 1977: passim. Basic norms in Art. 33 Decreto Ley 3133/1968. WRA/BR-001-2a/3.15.85 - 57 - succeeding terms permitted). 1/ This insticutional arrangement has in the past often resulted in serious friction between the mayor and the tesorero, two officials who must collaborate closely. 2/ Since the Treasury is charged only with the revenue side of 'he budget, its role here is of little importance. The Comptroller's Office (Contraloria de Bogota), like the Treasury, is directly responsible to the City Council. Following the British model, the legislative branch, as holder of the "power of the purse", has an-auxiliary organization to control the execution of the execution branch's authority for the use of public funds, i.e., the budget. 3/ Hence, the council elects the comptroller for two year terms; since a reform in 1979, no reelection is allowed. In addition to several staff units and committees, there are five divisions in the Comptroller's Office: Division of Ex-ante and Physical 1/ This arrangement goes back to the Castilian tradition: at the level of the municipalities, finances were in the hand of the City Council. The Colombian administrative reform of the year 1968 tried in vain to make the Tesoreria) part of the (Secretaria de Hacienda), hence to consider the treasurer as agent of the "mayor" (Art. 22 Decreto Ley 3133/1968). In 1969, the Supreme Court, in 1969, declared void this prescription. 2/ More specifically, a narrow collaboration with the Computer Centre of the District (SISE), the Directorates for Taxes and for Land-Register and the "Division Financiera y Tributaria" is vital for an orderly procedure of tax collection and budgeting. 3/ Internationally, two main types of audit organizations exist: The English (legal) system, where a bureaucratic unit is instituted as an auxiliary organization of the parliament, and the French (jurisdictional) system, where an independent court of accounts supervises public finances (Camargo 1969, 14 ff.). In Latin America, the first organizational type prevails; Colombian auditing at all levels (national, departmental, municipal, including the District of Bogota) is to be subsumed in a somewhat extreme form under the first group. WRA/BR-001-2a/3.15.85 - 58 - Control; Division of the Permanent Representatives of the Controller in the various district agencies and administrative units called Auditores; division of Ad-hoc Control by Visitadores; Division of Ex-post Control; and Division for the Special Control of the Administration of Foreign Credits. The "Controller" is responsible for the auditing the budgets of the central administration as well as of the districts decentralized agencies with the exception of the three big public utility companies. The Revisores Fiscales, nominated by the mayor and elected by the City Council for a period of two years, control the budgets of the utilites. 1/ They are in charge of separate offices composed of about 50 staff members each for the Water Company, the Electricity Company, and the Telephone Company, so that, in total, an additional number of about 150 persons are in charge of fiscal control. Finally, several committees bring together members of the executive and the legislative bodies at the District level. Information on the composition, functions and actual importance of these committees or councils to the budgetary process is summarized in Table 2.1. 1/ The existence of Revisores Fiscales who are not members of the Comptroller's Office is a result of the historical development of the three public service companies. Originally, they were private companies; after the take-over by local authorities, the Bank of the Republic (the issuing bank) was to administer a trust, to assure continuing effi- ciency. The City Council proceeded to elect the auditores out of a the three names suggested by the mayor. - 59 - Table 2.1: MIXED COMMITTEES IMPORTANT FOR BOGOTA'S BUDGETARY PROCESS Name Composition Functions Actual Importance Junta Distrital Alcalde Mayor (Chairman) Consultative body for all tax Most important for the grant- de Hacienda Secretario de Hacienda; matters; grants tax exemp- ing of tax exemptions and the Personero; Tesorero; three tions; decides on tax appeals decision of disputes between Councillors with their alter- (raclamacion y recurso de tax payers and the tax nates. The Contralor has apelacion) of individual tax administration 1/ voice but no vote. (Art, payers. 26 C.F.) Comite Financiero Still in discussion; in Investment plan for the entire Obviously not yet operative Distrital principle DAPD plus District; unifying of informa- representatives of the tion and accounts; revision of decentralized agencies. the Fiscal Code 2/ Junta de Alcalde Mayor (Chairman); one Important function in two More Important for investment Planeacion secretary or head of admini- fields: urban planning of decisions than for urban plan- Distrital strative department (Art. 24 Bogota (Decreto-Ley ning; more important than the Acuerdo 7/1977); four 3133/1968); defining the main other committees mentioned. councillors and their alter- lines of the investment budget nates; Personero; Director of of the district (Art. 75 C.F.) IDU; Secretaries of Finance and approving any amendments and Public Works. With voice to it (Art. 238 C.F.). but no vote: director of DAPD; Contralor; Secretary of the Junta. 3/ Junta Asesora y de Alcdlde Mayor as Chairman; all Authorizing the mayor to enter Normally formal approval of Contratos the secretaries and directors into contracts up to Col$ 3 contracts and changes of the of administrative departments; million; to authorize all con- budget during the fiscal year. three councillors and their tracts; to authorize changes alternates; Personero; with of the budget during the fis- voicer but no vote: cal year; to authorize the Contralor, Tesorero, and conditions of invitations to Secretary General of the bid; to approve the zones muo- Personeria (Art. 314 C.F.) ject to valorization charges (Art. 316 C.F.). Comite Tecnico Four councillors, technical Coordination of public Only 5-6 meetings per year, Coordinador de representatives of the public services by first discussing functions only with respect to Servicios Publicos services corporations, proposed investments before particular plans Direccion del Catastro, IDU, submitting them to the DAPD Jefe Unidad de Barrios, Representative of Public Works Secretariate of the D.A. de Accion Comunal 11 Details: Informe Financiero 1978/9: - 45 ff. tInterview Material collected by Johannes F. Linn in 1979. 3/Until 1977, the three professional societies of architects, planners, and engineers were represented in the Junta. They often assumed a critical attitude towards technically unsound planning proposals. This led probably to their exclusion from the commission. WRA/BR-001-2a/3.15.85 - 60 - The City Council Composed of 20 members and 20 alternates (suplentes) elected in the spring of every second year, 1/ the City Council represents the supreme law- giving authority at the level of the District. It is entrusted to pass District ordinances (acuerdos), one of the most important of which is the annual budget. The distinction between legislative and executive powers, with the latter represented by the local administration, is obscured by the fact that the Rouncil regards itself as co-administrator with the Mayor. This finds its expression, for example, in the fact that the Council elects the legal representative of the town, the Treasurer, and Comptroller as well as the Revisores Fiscales for the three major public service companies. For handling budgetary matters, the City Council oversees: 2/ . the "Comision Segunda" (for Finances and Budget), composed of eight councilors (and eight alternates) elected by the Council according to the political composition of that body; . the General Committee,-in its composition identical to the plenum; and finally, . the plenum itself. 3/ 1/ For more than a century, the Colombian party system has shown conflict and cooperation between the two Liberals and the Conservatives, affiliations which do not reflect any ideological orientation; instead, within the parties factions centered around important politicians constantly emerge and disappear. Details in Woff (forthcoming). 2/ According to Arts. 94 ff. of the Reglamento del Concejo. 3/ Theoretically, the Planning Committee of the Council could have an important influence upon budgetary matters; in practice, if it is operating at all, it concentrates; upon broad urban development, on public investments and their fiscal consequences. The Public Works Committee has not been operational during the last few years. - 61 - IV. BUDGETING: THE PROCESS Preparation and Presentation of the Proposed Budget The budget process starts with an estimate of expected revenue by the Financial and Fiscal Division (Division Financiera y Tributaria); expected revenue thus determines 4xpenditures, not vice versa. It is noteworthy that the cash system allows a very quick accounting of revenue; 1/ estimated revenue for the next budget can thus be based upon actual figures for the first half of the year. 2/ The estimation process entails simple extra- polations, semi-logarithmic extrapolations, and direct estimation of the tax base and, hence, of the resulting tax. While recommending budget estimates to the City Council, the Financial and Fiscal Division also provides various alternatives and their justifications. The Budget Committee discussions with the councilors also include the Financial Secretary, the Director of the Budget and the Head of the Fiscal and Tax Division. 3/ Often, though not always, the councilors are 1/ Note that a manipulation of the revenue side which would correspond to the manipulation of the expenditure side is not possible. The Treasurer reports daily on actual income, together with percentage of the budget obtained and cumulative figures. 2/ See, e.g., Presupuesto Central 1980: 35 ff. This situation differs radically from the national budetary process where estimates usually start in October of the second year before the budget year in question. 3/ As stipulated in Art. 103 C.F. WRA/BR-001-2b/03.15.85 - 62 - inclined to adopt the estimates of higher revenue levels and permitting higher expenditures. 1/ The initiative of the financial bureaucracy results in a provisional distribution of expenditures that forms ceiling figures for later discus- sions. This procedure definitely is politically more feasible than to start with the demands of the administrative unit. The unusual procedure for provisional4y allocating expenditures is by simple increments. Year by year, current expenditures and gastos-generales rise by 15-25 percent, most of which is due to increases in the remuneration of public officials, following the high inflation rate in Colombia. The strength of the trade unions is also a substantial factor. In 1980, an in- crease of 22 percent was included in the budget, and 26-28 percent was finally agreed upon, partly in response to union pressure. 2/ Investment expenditures are determined as a sort of residual, i.e., total revenue less current expenditures, debt service, etc. It should be noted, however, that this does not apply to the credit-financed investments, the majority of which are made by the decentralized agencies of the District. Initiatives in budgetary matters occur within the decentralized units, each having their own department. The central organizations of the 1/ The contention is that no matter how reasonable the revenue estimates are, there is a tendency for politicians to opt for higher expenditures to be financed by a higher revenue estimate. The fact that additions to the budget are much less frequent than transfers from one title to the next, however, seems to indicate a realistic policy of estimation. 2/ At the witional level, care must be taken not to insert too high a figure in the budget that the trade unions would regard as a lower limit. On the district level, this consideration does not seem to be of any importance: the trade unions present their demands, regardless of the budget. WRA/BR-001-2b/03.15.85 - 63 District only come into play after the Junta Directiva of the individual units has agreed to the budget; their influence usually is limited to the approval of transfers out of the central budget. Upon receiving the budget of ordinary income and current expenditures, the Directorate of the Budget only analyzes such questions as: Does the budget conform to the newly established budgetary classification? Do proposed transfers out of the central budget and transfers among decentralized budgets correspond to appropriate statutory requirements. The District Planning Office studies all investment expen-ditures, including direct investnents of the central administration and of the decentralized agencies and the indirect investments made by the decentralized agencies but financed out of the central budget by transfers. The Division de Programacion X Control Presupuestal prepares a statement for submission to the Junta de Planeacion. 1/ The Junta then formally approves the budget. A special case that deserves brief mention here relates to the nationalization of the educational sector in 1975 and 1976. Currently education is financed at both the national and district levels. Direct negotiations between the District and the National Ministry of Education (which in turn has to negotiate with the Ministry of Finance) are therefore necessary. Payments are made through the Regional Educational Fund for Bogota (Fondo Educativo Regional, FER), whose board of directors includes the mayor, representatives of the Ministry of Education, the District Secretary of Education and the Secretary of Finance, members of the ICCE (the national 1/ The DAPD is regarded as technical secretariat of this body. More specifically, the Director of the Unidad Administrativa acts in person as Secretary of the Junta, preparing the meetings and the agenda, maintaining the records, etc.). WRA/BR-001-2b/03.15.85 - 64 - agency for school construction), of ICETEX (the agency in charge of scholar- ships and educational credit), an auditor, and a representative of the teachers. While the mayor has the authority to execute the board's actions, the Secretary of Education of the Special District actually carries out the decisions of the Junta. The procedure thus is extremely complicated; transfers from the national government are only definite after the national budgetary law is enacted so that changes the district education budget become necessary during budget execution. Discussion and Approval of the Budget After the mayor formally approves the proposed central budget and the Junta de Planeacion and various boards of directors accept the budgets of the decentralized units, the Secretary of Finance presents the two budget docu- ments to the City Council. This stage must be completed before November 1st of each year. For every ordinance (acuerdo), authorization requires two readings on different days: one in the General Committee, and one in the Plenum. 1/ One member of the Budget Committee called ponente (often its President 2/) is charged with delivering a report on the proposed budget, to suggest changes and to recommend its adoption or rejection by the Council. For each budget proposal of the decentralized agencies, the same procedure is followed. Final decision must be taken before December 10. Approval by the General Committee and the Plenum (identical in composition) is usually a mere 1/ Art. 83 Reglamento del Concejo. The text follows various articles of this code. 2/ In Spring of 1980, identical to the President of the City Council. WRA/BR-001-2b/03.15.85 - 65 - formality because of time constraints. The report of the Budget Committee is presented shortly before the deadline so that the City Council has no other alternative than to approve the budget by a very formal consent where the votes are not even counted. 1/ The City Council's approval of the budget is not the last step of its formal acceptance. The Mayor must first approve it, and then the city government (and the boards of directors in the case of the different descentralizadas) must translate it into an executive orders. To-that effect, the Mayor has to promulgate a decree for a precedure called "liquidacion," and the boards of directors of the descentralizadas have to take a corresponding decision. A number of aspects of the Council's budget approval process are noteworthy. Contrary to a procedure reported to occur in the United States, 2/ agency heads and councilors infrequently meet directly to challenge the budgetary figures that the Mayor has accepted. Two reasons for this are clear. Given the weakness of a professional civil service, more or less open revolt against the mayor or the Secretary of Finance would quickly result in the dismissal of the rebellious official. In addition, the autonomy of the City Council is in no way comparable to that of the North American Congress. Thus, to increase any of the estimates of expected tax revenue or of the budgetary surplus available for expenditure requires the prior, written 1/ Should the Concejo not come to a decision in time, the proposal as present by the administration is automatically in force (Art. 106 C.F.). A similar prescriptionn exists for the national budget. 2/ Wildavsky, 1974: 88f. WRA/BR-001-2b/03.15.85 - 66 - consent of the Secretary of Finance; 1/ the same holds true for any modification, transfer, or inclusion of new expenditures in the budget. 2/ The Council must, in fact,.accept a number of expenditure items (debt service, contractual obligations, planned public works, public security, education, public health). 3/ Although the influence of the city Council in budgetary matters must therefore be considered small, individual councilors do demon- strate particular preferences; their influence, however, is usually exerted before presentation of the budget proposal to avoid open confronta7tion with the District's Secretary of Finance. A very special Colombian system of parliamentary grants (auxilios parlamentarios) should also be mentioned. These funds are not found in the budget proposal that the Secretary of Finance presents but are added after- wards by the City Council. In 1980, parliamentary grants arrounded to Col$ 70 million, or about 1.5 percent of the central budget. After the Council has fixed the aggregate figure, the funds are distributed among the individual councilors who determine their use, e.g., for minor public's works, scholar- ships, donations to neighborhood committees, etc. The official justification for the grants is that the councilors, as representatives of the electorate, can use the funds to alleviate needs of the population that lie outside the auspices of the ordinary bureaucracy. Even if this were true, it is question- able why the auxilios do not appear in the ordinary budget but are distributed 1/ No. 3 Art. 101 C.F. 2/ No. 5 idem. 3/ No. 8 idem. WRA/BR-001-2b/03.15.85 -67- directly by the councilors. 1/ The explanation, of course, lies in the political realm, discussed below. Budget Execution The Fiscal Code of the District of Bogota mentions two systems of expenditure control, one referring to contractual obligations, the other to effective payments. Only the latter system, called control de ordenacion de gastos, or simply control de gastos, is used at the central as well as at the decentralized level. Every month, the various administrative units present their financial requirements--i.e., for investments financed by central administration units or by decentralized units with transfers from the central budget--for the following month to the Division of Budget Execution and Control and to the Planning Department of the District. 2/ The established schedule for investments, the availability of revenues, the time elapsed since the beginning of the fiscal year and before budget execution are the criteria for proposing the acuerdo 3/ de ordenacion de gastos for signature by the Secretary of Finance and the Mayor. The document then becomes the legally binding upper limit for actual expenditures in the next month. The decentralized agencies employ a similar procedure. The Board of Directors of the Water and Sewerage Company decides four times a year on the acuerdo. Only current expenditures, however, are included; for investments, a 1/ The administrative and constitutional reform of 1968 prohibited the Council to vote auxilios (No. 2, Art. 14 Decree-Law 3133), but the Supreme Court voided this clause in 1969. At the national level, several attempts have also accurred to abolish the auxilios, also in vain. 2/ Should they not do so, the Budget Directorate may unilaterally fix an amount; Art. 145 C.F. 3/ The expression is misleading; the City Council is not involved. WRA/BR-001-2b/03.15.85 - 68 special account (reserva) for the whole year constitutes the ceiling for effective payments (giros). Special allocation rules apply to avoid too many expenditures in the first months of the year. Changes of the budget during the fiscal year are of two types: additions (creditos adicionales), which augment the total amount of the budget; and transfer (traslados) which more funds from one expenditure title to another. While traslados occur frequently (about 150 cases a year), there are seldom additions to the budget. For transfers, the bureaucratic machinery for transferring funds from one title to another prevails over any authority of the City Council: 1/ After the Comptroller has certified that the funds are available and after the approval of the Junta Asesora y de Contratos, 2/ the Mayor may simply decree that the money been shifted. In the case of transfers within the investment budget both at the central and at the decentralized levels, the District Planning Department must consent. This prescription also applies to investments by the decentralized agencies that are not financed out of transfers of the central budget. To increase the total amount of the budget or to create a completely new expenditure title, the City Council must formally approve the change. This may be given by the 1/ Note that para. I, Art. 182 C.F. which speaks of the necessary approval of the Budget Committee of the City Council has been declared void by sentence of the Administrative Court of Cundinamarca in 1978. Art. 184 C.F. prescribing the approval of the Council itself during periods of sessions (four times a year, for a period of one month each), according to information obtained in the Budget Directorate, is not observed. 2/ Art. 316 C.F. Note that in this body the representatives of the administration form the majority. WRA/BR-001-2b/03.15.85 - 69 - Budget Committee between the sessions 1/, during these by the plenum itself. The Mayor's decree concludes the procedure. 2/ It is interesting to note that the budget does not have to be modified during its execution by an additional budget enacted in the same form as the basic budget the usual procedure in industrialized countries. More than 100 single decrees represent that many ad hoc decisions. It is thus questionable whether systematic considerations of fiscal policy--if they ever influenced the initial shape of the budget--operate during the budget's execution. 3/ The bureaucratic procedures leading to a payment by the Treasury are numerous and therefore onerous. Many authorities are involved: the initiating unit, the comptroller's Office, the Secretariat of Finance and the Treasury. Three sets of controls exist and the only official entitled to order a Treasury payment is the Secretary of Finance himself (or the legal representatives for the units). 4/ For contract partners of the District, considerable delays in payments of cou'rse results; this entails problems not only for the private party, but also for the District which must pay interest on the arrears. 1/ In that case, total additions may not normally exceed 25 percent of the initial budget; Art. 189 C.F. 2/ It may be noted that, according to Art. 28 Decree 3133/68, decrees modifying the original budget need the prior consent of the Junta Asesora y de Contratos. In the practice, however, this prescription has little relevance. 3/ Caiden/Wildavsky 1974: 66ff. call this phenomenon "the disappearing budget." 4/ Art. 117 C.F. WRA/BR-001-2b/03.15.85 - 70 - Strange as it may seem in a developing country--poor by definition and with tremendous needs for urban public services--underspending is the rule in Colombia. In 1977, actual central investment 1/ were only 22.9 percent of budgeted investments 2/, and those of the decentralized authorities were only 65.8 percent. 3/ In 1978, central investments rose to 35.7 percent, 4/ while the Departmento Administrativo de Bienestar Social spent only 2.8 percent of its investment budget. 5/ In 1979, decentralized agency investments were only 53.7 percent of the total. 6/ These figures include investments committed during past years (vigencias expiradas) (they normally come close to 100 percent). In addition, the percentages refer not to the initial budget, but to the budget including all subsequent modifications. If it seems likely that planned investments will not be made during this year, the expenditure title may thus be transferred to another use; this, of course automatically increases the percentage of planned investments actually realized, and vice 1/ Of course, the definition of "investment" is open to some discussion. We cannot enter into details, however. 2/ Informe Financiero 1977: 65. 3/ idem: 107. 4/ Informe Financiero 1978: 27. A simple comparison of percentages is difficult, however. See the following text. 5/ idem: 33. 6/ Source: idem. - Historical statistics on budget execution to be found in Revista de Hacienda y Planeacion, Vol. 1 No. 1 (May 1978). The central administration usually has spent only about two percent of its budget for - investments, the decentralized agencies around 40 percent. In other words, in the period from 1973-1976 the central administration has only implemented one percent of the investments of the district (idem: 59). The four big public service corporations (plus the IDU) accounted for about 90 percent of all investments of the decentralized agencies. WRA/BR-001-2b/03.15.85 - 71 - versa, if the investment budget has been raised during the fiscal year. Thus, in 1977 central budget investments amounted to 28.3 percent of initial appro- priations, 1/ decentralized investments, however, drop to 59.9 percent. 2/ The District Social Security Fund, which reported a proud 100 percent invest- ment execution, actually spent only 0.17 percent of the initial amount ear- marked for that purpose; 3/ the rest of the funds were simply transferred to another use. Auditing Types of budgetary oversight commonly include: 4/ political control, i.e., the mechanism that ensures that public expenditures meet politically set priorities, usually embodied in a popularly elected group that holds the "power of purse" (the Concejo in the Bogota case); administrative control, i.e., the mechanism to ensure that the discretion which the bureaucracy exercises after the acceptance of the budget law accords with sound principles of fiscal management; and numerical-legal control, i.e., the mechanism that ensures that revenues and expenditures legal constraints the most important of which, of course, concerns the budget itself, the tax prescriptions and the various rules on contracts with the fisc. Due to its close relationship to budgeting in general, political control of Bogota's budget is discussed in more detail later in this paper. Since both bureaucratic units charged with preparing the budget and supervising its execution (the Planning Department and the Secretariat of 1/ Calculated following dates on Informe Financiero 1977: 65. 2/ Following dates idem: 107. 3/ Following dates in idem: 149. 4/ For Colombia, at the national level, cf. Wolff 1977b. WRA/BR-001-2b/03.15.85 - 72 - Finance) lack the necessary administrative capabilities, we may simply state that administrative control is nonexistent. We therefore concentrate here on the system of numerical-legal control exercised by the Comptroller's Office of the District and by the Revisores Fiscales and their collaborators in the three major public service companies. Unlike the system employed in most industrialized countries, the basic principle of Colombian numerical-legal control is not selective, i.e. based on Random Sampling; 1/ officials of the Comptroller's Office (or the revisores) must verify all movement of public funds. Second, all controls are carried out twice: once before and once after the expenditure. Colombian law also specifies a third type of control to determine the physical existence of funds and material and to compare it to the book of accounts. If misappropriations of funds are suspected, the Division de Visitadores of the Comptroller's Office carries out a special investiga- tion. 2/ The Section of Supervision and Evaluation attempts to audit each bureaucratic unit once or twice a year; the Section for Investigation, however, performs in-depth inquires after a denunciation has been made. The division also oversees the permanent auditores delegados in the different administrative units under the Comptroller's Office. In contrast to national auditing where irregularities are only detected by ex-post facto investiga- tions, Bogota benefits from these ongoing controls. Since officials in the 1/ This is not exceptional for Latin America. Comparable are control types - in Costa Rica, Cuba, El Salvador, the Dominican Republic, and Uruguay. For details, see Camargo 1969. Control types are identical at the national level; Wolff 1977b. 2/ See Arts. 562 ff C.F. WRA/BR-001-2b/03.15.85 - 73 Colombian bureaucracy change constantly, detection of irregularities after a great deal of time is rather pointless. It should be recalled that the Comptroller is represented without voting power in the majority of the Juntas Directivas of the District's decentralized agencies. Despite this legal constraint, he can nonetheless impede plans proposed by the decentralized agencies. In addition, the Comptroller has in recent years entered the realm of other types of control and even the field of general administratiTe tasks. The Division Financiera-Operativa makes proposals for administrative reform in concrete cases; if, during the budgetary process, some person or organizations do not comply with requirements (e.g., if payment are not made), the Comptroller intervenes. The Office's annual reports (Informes) 1/ have also criticized cases of inefficient budgeting (especially under-spending), with some consequences in for personnel. A similar observation can be made about the Revisores Fiscales, at least in the case of public service corporation. Finally, it is noteworthy that the Comptroller's Office is financed essentially by transfers of 1.5 percent of the budget of current expenditures of the decentralized agencies. 1/ Although a general clause ("intervenir en todo el proceso presupuestal" 2/) would permit the Comptroller to intervene at every step of the budgetary process, it has refrained from doing so. 3/ 1/ The seemingly clear-cut distribution of tasks is not always strictly observed. Thus, by resolution 10/1972 of the Comptroller, auditing in the Personnel Division of the District was entrusted to the Division de Control Previo of the Contraloria. 2/ Art. 539 C.F. 3/ The Comptroller is also entrusted with special additional tasks in the field of adminstration of assets of the district, but the procedures have WRA/BR-001-2b/03.15.85 - 74 - V. CRDIT RAISING AND INVESTMENT PLANNING: THE ORGANIZATION A very high percentage of public investments in the special District has traditionally been financed out of loans and credits. Between 1971 and 1976, credits have financed between 56.1 and 93.8 percent of investments. 1/ Of total credit income, only 5.9 percent would come from national sources. For the Central Administration of the District, the maximum percentage of investments financed by foreign credits for a single budget year was 31.1 percent between 1969 and 1976, with the average being only 18.3 percent. 2/ For the decentralized agencies, some 50% of investment expenditures were financed by credits. In Colombia a basic principle for raising credits by any public body, at the national as well as the District Level, is that no entity is entirely free to engage in a loan contract. In every case, controls come into play. Credit Control Authorities At the District level, the control mechanisms are identical to the bureaucratic units engaged in budget preparation and execution (with the partial exception of the Budget Directorate); i.e., the District Planning Office, District Planning Council, Coordinating Committee of Public Services, and the City Council. The Budget Directorate concentrates on technical aspects only and accepts the recommendations of the agencies. little relevance here. 1/ Hacienda y Planeacion, Vsl 1, No. 1 (May 1978), Part 2: 60 f. 2/ Castano 1979: 121. WRA/BR-001-2b/03.15.85 - 75 - Raising credits by any public authority--with the exception of very small credits below 1 million Col$, about US$22,000--requires the consent of the National Government. Thus, even after the District authorities have approved the credit, the National Government must carry out its evaluation "from the ground up." Therefore, although the National Government does not officially interfere in Bogota's budgeting (e.g., by giving approval or disapproval to the budget accepted by the City Council), it obviously influences the basic decisions underlying budgetary figures, especially investment choices. The National Planning Department (Departamento Nacional de Planeacion, DNP) is a staff unit at the national level, 1/ representing the central bureaucracy on questions of economic and social planning and is the technical secretariat of the National Council of Economic and Social Policy (Concejo Nacional de Political Economica y Social, CONPES). DNP comprises seven units (unidades) 2/ in charge of research and planning for a particular policy area (e.g., human resources, industrial and agrarian studies) or a cross-section (e.g., the Public Investment Unit). The unit in charge of budgetary questions, called Unidad de Inversiones Publicas, has responsibility for the investment side of the (national central) budget; within this unit are three divisions, i.e., for external financing, budget programming and control, and financial analysis. 1/ The concept hence corresponds to the basic idea of the organization and planning of the district. Once again, however, planning is not the apolitical task this name implies; the heads of DNP have long been politicians or at least persons with political inclinations. 2/ Presidencia 1978: 389ff. WRA/BR-001-2b/03.15.85 - 76 - The first division is especially important in the present context because it assesses the economic viability of projects to be financed by foreign credits. The Plannitg Department is thus involved in determining public investments to assure conformity with overall development goals, including that of balance of payments stability. The Finance Ministry (Ministerio de Hacienda y Credito Publico), in addition to several staff units, includes a number of general directorates; the Direccion General de Credito Publico is responsible for collecting credit resources and controlling the credits raised by the different bureaucracies at the central, the central, the decentralized, and the territorial levels. 1/ This directorate has five divisions, the most important of which are the Division de Credito Publico Interno and the Division de Credito Publico Externo (in charge of studies and recommendations on economic, fiscal and financial aspects of proposed national, and foreign credits, respectively by any public authority). In addition to these major authorities, several units and committees o .;he central government participate in the evaluation of credit requests. The previously mentioned CONPES is a sort of cabinet committee, in the European sense, composed of the President, the Ministers of Finance, Economic Development, Agriculture, Public Works, the Head of the DNP, the Manager of the Central Bank, and--a very Colombian characteristic--the Manager of the Coffee Growers' Association (Fedecafe). Since 1968, the Monetary Council, the Council of Foreign Commerce, and the Council of Customs Policy each have a permanent member in the CONPES; other public organizations can be invited 1/ Art. 1 (national) (Decree 076/1976. WRA/BR-001-2b/03.15.85 - 77 - ad hoc to send representatives. 1/ Legally, CONPES is the higi%ast decisionmak- ing body on questions of economic and social development policy. More specifically, it approves national plan proposals for presentation to the Congress. 2/ Connected to the Finance Ministry, the Monetary Council is a unit "for consulting, coordination and decision-making on specialized matters." 3/ The Finance Minister acts as chairman, with participation by the Ministers of Economic Dev-lopment and Agriculture, the Head of the National Planning Department, the Director of the Banco de la Republica and of the Institute of Foreign trade, the economic Secretary of the Presidency, and, two consultants without voting power. 4/ Since major development projects may influence price levels and the country's foreign exchange situation, it is understandable that the Monetary Council has a say in these matters. The Central Bank (Banco de la Republica) must register credit contracts involving repayments in foreign currency 5/ so that the necessary permits may be issued at the proper time. 1/ Rivera-Ortiz 1976: 67, 71. 2/ This is legal fiction, however; the Plan Committees of both housesnever actually get formed, so that the first reading (corresponding to the discussion in the committee) cannot take place. Consequently, the president has used his prerogative to promulgate the plans by decree with the binding force of laws. It may be added that the CONPES is only involved in the case of foreign, but not of national credits. 3/ Art. 3 Decree 080/1976. 4/ Decree-Laws 2206/1963 (Art. 1); 3171/1968 (Art. 1); 151/1976 (Art. 9). 2206/1963 (Art. 2). 5/ Law 441/1976 (Estatuto Cambiario). WRA/BR-001-2b/03.15.85 - 78 - Because a national government guarantee is often required, domestic credits must have approval of the Council of Ministers, a group composed of the President and 13 ministers. 1/ Finally, an Interparliamentary Commission--four members of Congress, two designated by Committee No. 3 of the Senate, the other two by the Corresponding committee in the House of Representatives--act as consultants to the government 2/ on matters relating to foreign credits or to guarantee credits contracted by other public agencies. VI. CREDIT RAISING AND INVESTMENT PLANNING: THE PROCESSES The process of credit raising and investment planning both in Bogota and nationally includes at least three steps: (a) the initiative; (b) control procedures at the level of the Special District; (c) control procedures at the level of the National Government, as well as negotiations with the credit grantors. 1/ To issue prescriptions on the organization of the public administrationis one of the prerogatives of the Congress. The number of 13 ministers is a heritage of the Frente Nacional, the "big" coalition- of the Liberal and Conservative parties from 1958 to 1974 (but continuing, in many aspects, thereafter) aimed at ending political violence and enabling a transition to unrestricted political competition. Six ministers were conservatives, six liberals, and one a non-political professional military. 2/ Art. 8 Law 123/1959. It can be added that, according to Art. 8 Law 3/1972 an equal number of congressman, elected by the same committees, participate with voice and vote in the deliberations of the CONPES if credit-financed investment plans are discussed. Their role is to insure fair distribution to various regions of the country. WRA/BR-001-2b/03.15.85 - 79 - Initiative Initiative for particular investment projects usually originates with the local unit that intends to borrow funds, i.e., the Secretariat of the Central District Government, or, more frequently, of the decentralized agencies. 1/ Although no data are available to suggest why particular investment or credits are proposed, it is nonetheless evident that each agency works on the basis of investment plans dating-back several years. This is true for water and sewerage investments; for the production and distribution of electrical energy; for the construction of roads; and, presumably for the telephone network. After the internal decision making process of an agency has ended and its manager has approved the proposal, 2/ the board of directors, and occasionally the Coordinating Committee for Public Services, evaluate the proposed credit; if they approve, the project is then submitted to the District Planning Department, Division of Financial Analysis. Control by District Authorities Upon receipt of an agency's borrowing proposal, the Division of Financial Analysis prepares a review paper which is then presented to the Planning Council for formal approval, a procedure not unlike the preparation 1/ There are exceptions, however, to this rule. For example, the two-level street intersections under construction in 1980 were originally proposed by the Institute of Urban Development, but the mayor transferred responsi- bility to the Secretariat of Public Works. If the Bogota subway is ever be built, it will be at the initiative of mayor rather than any manager of a descentralizada. 2/ This decision making process includes technical and financial feasibility studies. A special decentralized unit supervised by the National Planning Department has been created to finance these studies. WRA/BR-001-2b/03.15.85 - 80 - of the budget proposal. The justification for the planning authority's involvement is ostensibly that it assures that the proposed investment/credit conforms with the general development plans of the District and of particular programs. In practice, however, the Planning Department judges the investment according several fiscal rules aimed at guaranteeing proper financial management and avoiding heavy indebtedness. These standards include the following: Total debt service may not exceed 20 percent of ordinary income (or 10 percent of total income) unless income of the new investment w11 "automatically" produce the exceeding amount; 1/ the initiating agency must also demonstrate satiifactory financial and investmen management for the preceding years. 2/ In the case of direct investments by units in the central administration, revenue must be sufficient to support the debt service. The ,-Ianning Department thus fails to evaluate the technical feasibility and economic value of the proposed project. Since the Planning Department functions as a sort of technical secretariat of the Planning Council, the Council normally accepts its recommetdations. Not infrequently, however, the manager of the agency, usually accompanied by his financial and technical directors, will explain and defend the project. After the Planning Council's approval (a2d the approval of the Secretariat of Finance if a transfer from the central Ludget is necessary), the City Council provides formal consent in the case of credit 1/ A typical Colombian loophole: the seemingly strict prescription is mostly invalidated by an exception or general clause. In this case, no future income can be assumed "automatic" since it is always based on estimates. 2/ According to Art. 6 Decree 1050/1955, a Certificate of the Contralor or Visitador is to be included in the record. WRA/BR-001-2b/03.15.85 - 81 - contracted by the central administration. 1/ As representative of the District, the mayor then signs the file and presents it to the national authorities, the third level of review. The formal process is actually based upon extensive informal coordi- nation and consultation. A widely.accepted social norm in the Colombian bureaucracy assures that conflicts among officials of more or less-equal rank are avoided. The Planning Department therefore usually approves proposals as presented. Control by National Authorities For both national and foreign credits, approval of the Finance Minister is require; if the proposed credit exceeds Col$5 million, the Head of the National Planning Department must also provide formal consent. The Mayor thus presents the formal demand among with numerous documents to justify the 1/ Art. 420 C.F. Authorization is necessary for every foreign credit and for all national credits of more than 3 million Col$, about US$67,000. WRA/BR-001-2b/03.15.85 - 82 - project, in two copies to the Finance Minister and an additional copy to the DNP. 1/ The next steps of the decisionmaking procedure differ slightly, depending on the origin of the proposed credit. If a domestic credit is proposed, negotiations with the lending bank have already occurred and a provisional agreement achieved. The project contract thus forms part of the 1/ Most of these supporting documents are enumerated in Decrees 1059/1955 and 150/1976. According to Normas y Procedimientos 1979: 1-12, these are: 0 the written request itself; a the authorization by the political or directing body; 0 a feasibility study concentrating on economic and financial aspects; 0 an "investment plan," for financing the project and the time schedule for using the proposed credit; & a scheme for the proposed debt service; 0 a draft of the proposed credit contract, sometimes requiring several other certified documents; 0 a certificate of the Contralor that the unit in question has the financial capacity for the new indebtedness; * the guarantees, requring several certificates, stating among other things that the agency is free from other charges; the consent of the Mayor, in the case of the decentralized agencies of the District; *4 * a projection of income and expenditure for a-period not less than five years; * the budgets for the current and the preceding year; * the balance of the preceding year's budget; * in the case of decentralized units, the last two balances plus income and expenditure for the two preceding years; * account of the state of existing debt and its projection; if a unit is asking permission for a credit for the first time, its statutes; * if material is to be acquired abroad, a certificate of INCOMEX that there exists no national equivalent; * in the case of investments in the telephone network, an evaluation of a body called National Coordinating Council for the Telephone Sector. WRA/BR-001-2b/03.15 .85 - 83 - file, which the Finance Minister is authorized 1/ to sign after the agreementof the National Planning Department has been given. In the case of foreign credits, 2/ however, the procedure is much more complicated. First, official negotiations with any credit institutions cannot begin without the Finance Minister's prior authorization (also based in the agreement of the Planning Department). This resolution sets the terms that the unit in question can engage in, i.e., maximum interest, grace period, duration of credit, etc. Since such terms can only be fixed on the basis of an intimate knowledge of international capital markets, the involvement of the Public Credit Directorate of the Ministry is obviously important. As mentioned above in the description of the bureaucratic network, several additional units participate in approving the proposed investment, including evaluations by the Council of Ministers and the Monetary Council. A supporting statement by the CONPES is also required, although this body always agrees to the decisions of the National Planning Department. Finally, the Comision Interparlamentaria de Credito Publico must be informed. After this, a second ministerial resolution authorizes the entity in question to sign the credit contract and, having heard once more the Council of Ministers, the President then gives his consent. 1/ By a so-called "resolucion ministerial." Thereafter, approval by the Council of Ministers is required (Art. 2 Decree 2449/1975). 2/ If the credit is above 10 million pesos or US$500,000; Art. 114 Decree-Law 1670/1975 as modified by Art. 2 Decree 2449/1975; see also Decree 150/1976. To assure conformity with the prescriptions, the Public Credit Directorate of the Finance Ministry has even published a model contract (Minuta Modelo 1976). WRA/BR-001-2b/03.15.85 - 84 - The complications of this procedure easily explain why it usually takes five to six months until the contract can be signed. 1/ The evaluation by the National Planning Department is primarily a financial analysis rather than a technical one. From the available information, the examination is apparently based on data transmitted by the interested unit; the DNP rarely initiates its own investigation. One can conclude once more that the review process provides rather implicit control, e.g., by checking the coherence of the proposal's arguments. Again, however, the justification for including the DNP among the supervision bodies is that it should guarantee the conformity of a particular project with development plans and programs. It can be added that during project execution, the two planning departments (DAPD and DNP) receive progress reports. If foreign credits are involved, control by these two departments is much stricter. Unlike the earlier requirement under which the National Congress had to consent to every foreign credit, a Congressional lump sum authorization (which is raised from time to time) is given by law. The current limit is US$2 billion. 2/ If the credit-giving bank is a Colombian institution, there are other small modifications in the approval procedure. A final formality must be mentioned. Like every contract relating to the District's fiscal activities, credit contracts must be registered and 1/ If an invitation to bid is necessary, another higherly complicated process - will follow. One interviewee reported that in one case, even after the award to his firm, one year and a half passed before the contract was finally signed. 2/ Law 631/1978. WRA/BR-001-2b/03.15.85 - 85 - countersigned by the auditing authority, be it the Comptroller in the case of the central administration or the auditor (revisor fiscal) of a decentralized agency. 1/ As at the first stage of review (the District level), details of the legal prescriptions are in practice frequently not observed. It was therefore determined that a single ministerial permit was not sufficient to control the procedure; since the minister, for political reasons, often could not withold his approval despite unfavorable terms, another ministerial resolution was introduced. 2/ This requirement, however, is often not observed since informal discussions with the bank start well before the official authorization is given. Interested private firms even participate in preparing the invitation to bid where they will be one of the participants. Informal discussion also occurs among the supervisory bodies; when a request is actually presented, the review meeting is only intended to formalize an agreement that has already been reached. Objections to a borrowing contract are therefore infrequent. 1/ Art. 279 C.F. 2/ By Decree 2832/1966. WRA/BR-001-2b/03.15.85 - 86 - VII. BUDGETING, PLANNING AND POLITICS: POLITICAL CULTURE VERSUS TECHNICAL RATIONALITY Control and Coordination at the District Level As at the national level, decisionmaking--and therefore political power--is highly fragmented in Bogota. The fragmentation is clearly manifested in the budgetary process. A budget for Bogota as a whole does not exist; each of the decentralized agencies instead decides on its own budget, several of which are considerably larger than the central administrative budget. The important budgetary entities are also functionally independent: Apart from small amounts, the important decentralized firms neither -eceive nor pay any funds to the central budget. 1/ The multiplicity of budgets in itself necessarily does not mean that they follow divergent or incoherent policy lines; functional fragmentation could in theory be associated with efficient coordination at the central level of the District. In this context, however, coordination and control are much weaker than the forces of independence among the different units. Control and coordination, of course, aim to influence the behavior of some agencies (e.g., the decentralized public service agencies) by another (e.g. the District Planning Department). 2/ Defending their own autonomy, however, is one of the 1/ As noted by Linn 1976: 146. This is net to say that functional frag- mentation-may not have its merits such as a higher professionalization and a possible shielding from political pressures of all sorts (idem: 143 ff., as confirmed by interview material). The point is, however, that a clear and coherent policy of the city government is extremely difficult or impossible. 2/ See Etzioni 1975: passim, and Thompson 1967: Chapter 10. WRA/BR-001-2b/03.15.85 - 87 - fundamental reactions of all bureaucratic entities. Thus, the means of control become the decisive factor. According to Etzioni, coercion, economic incentives, and normative values are the three major types of control; a com- bination of the three would be necessary to influence the behaviour of complex organizations. In the case of Bogota, though, the supervisory bodies lack most of these means of control. Given the administrative independence of the decentralized agencies, coercion would only be an option if individual officials were guilty of gross misbehavior; it is not a method, however, that can be used to direct the policy of an organization. 1/ Although their formal consent is necessary, the supervisory agencies also can not dispose of economic goods, especially funds. Finally, normative values would amount to legal prescriptions; as we have observed, though, formal rules determine only the procedures to be followed and the documents required. The only real prescriptions that exist--the constraint on maximum indebtedness included in the fiscal code (which may easily be circumvented), the Finance Minister's resolution on the terms of a loan contract--attempt to avoid economic losses rather than offer any guidance for policy. Obviously, the City Council, exerts some control over budgeting decision. It theoretically holds the "power of the purse" by having authority to approve the central budget as well as the acuerdos for the decentralized budgets. In addition, individual council members are also on the boards of directors of the decentralized agencies, thereby insuring that agency budgets 1/ The hiring and firing of managers may also be regarded a special case of coercion. It is questionable, however, whether this method can actually influence organizational behavior. In addition, the more hiring and firing occurs at the top, the more important and powerful the permanent day-to- day administration. WRA/BR-001-2b/03.15.85 - 88 - reflect the popular will. This model of representation assumes that the City Council develops a clear policy aboutthe major budgetary decisions made by the decentralized agencies, and that the councilors use their votes to influence decisions accordingly. Neither assumption, however, is supported empirically. The City Council is elected every two years, and meets for four sessions of one month's duration each year. Although some continuity exists, the composition of the group varies sufficiently to make its long-term orientation very difficult to predict. Moreover, a political entity such as a city council is not likely to consider basic policy orientation in the abstract but rather vis-a-vis concrete issues: e.g., should a major arterial road be constructed favoring high-income residents, or should improvements be made to small neighborhood streets in the barrios? Should the water and sewerage lines be extended to undeveloped land, or should investments concentrate in providing services to squatter settlements? This is all the more true in Colombia where politics is more "the art of muddling through" 1/ than decisionmaking on the basis of rational analysis of problems. It thus follows that the City Council normally does not take a clear position on the many decisions to be taken by the boards of directors of the decentralized agencies; that its representatives on these boards do not truly represent the council at all. It would be incorrect, however, to conclude that the councilors necessarily have little influence over the boards of directors. Insofar as these boards are important, councilors may play an important role depending of their individual personality and political connec- tions. Their power to impede a plan, however, is greater than their ability 1/ Braybrooke/Lindblom 1967. WRA/BR-001-2b/03.15.85 - 89 - to make a positive contribution to the business policies of the various corporations. The second domain of control lies in the political leadership of the District, exerted through the mayor's chairmanship of the boards of directors, and through the representation of public officials on those boards. 1/ Very little information exists on the power relationship between the permanent bureaucracy of the semi-independent authorities and their boards of directors. At the national level, an investigation of the Higher School for Public Administration 2/ demonstrated that the established bureaucracy clearly dominates over councils that meet only occasionally and must make decision on the basis of information transmitted by the manager. Interviews among Special District officials suggest the same conclusions; one intecviewee even mentioned a "delegation" of tasks to the bureaucracy. As in budgetary matters, the control of the business policy of the decentralized units is therefore rather loose. 3/ Finally, there is the control at the higher levels of the national government. 1/ Another control body could be added: The Council of Ta-iffs of Public Services (Junta de Tarifas de Servicios Publicos) is a oody connected to the National Planning Department that must agree to proposed changes in the tariffs for services rendered by public or semi-public enterprises. Normally this agreement is automatic. Interestingly enough, the Electricity and Water Companies of the Special District are held by international agreements to charge tariffs sufficient to avoid operational deficits. 2/ Escuela Superior 1973. 3/ Another argument to support this contention is the rapid change in the offices of the mayor and the managers, a change. Influence upon decisions is certainly a function, of permanence in one's office, among other things. WRA/BR-001-2b/03.15.85 - 90 - National authorities do not, however, exert sufficient political (and, to a somewhat smaller degree, legal) authority to direct the policy of the decentralized and centralized authorities of the District. 1/ They nearly always find themselves in the position of reacting to the initiatives of others, usually the management of the decentralized corporations. 2/ Control criteria at the national level are usually financial in nature. In other words, there are no possibilities for structuring investment decisions on the basis of a comprehensive development plan. For various political reasons, such a plan cannot exist, 3/ at least in the sense of a document that guides the actions of public authorities in the field of development. Relative Political Weight of the Bureaucratic and Political Units Generally speaking, the executive units exercise more political power than the legislative City Council. The full council only approves or disapproves the overall budget; it cannot modify single expenditure titles. 4/ 1/ The fact that the Mayor of Bogota is directly appointed by the President of the Republic should not be considered a counter-argument. The mayor may or may not belong to the national power elite; he may also find himself in a web of antagonistic political forces that make simple adherence to superior orders difficult. Finally,.most important, Colombian presidents usually do not operate out of a clear and coherent line of policy that can be transmitted to subordinator. 2/ It should be noted again that, in contradiction to legal prescriptions, contact with creditors usually occurs before an official authorization is given. 3/ Wolff, Planung in Kolumbien. Moglichkeiten und Crenzen rationaler Entwicklungspolitik in einer klientelistischen politischen Kultur, manuscript in print. 4/ Budgetary law, in fact, demonstrates a marked distrust of the City Council. WRA/BR-001-2b/03.15.85 - 91 - Even the Council's Budget Commission (and its various working groups that analyze the decentralized budgets) normally approves the budget proposals as proposed by the executive offices. It is therefore not surprising that in recent years, only three to five percent of the total budget has been modified as a result of the City Council's deliberations. Presumably, the slight influence that the City Council exerts is not only a result of the "natural" power of the executive, the lack of technical support available to the Council, and the fact that councilors usually are not fulltime politicians, but it is also a result of the councilors' own lack of interest. 1/ Two exceptions, however, must be mentioned. The City Council is clearly interested in the current expenditures of the central budget, especially in the biggest item: the salaries of city employees. 2/ In addition, the councilors pay attention to parliamentary grants (auxilios parlamentarios) to a degree out of proportion to their size (small, even by Colombian standards). The political importance of the auxilios is discussed in more detail below. Budgeting, Interest Groups and the Voter The direct influence of interest groups (gremios), neighborhood committees, or of social forces in General upon the budgetary process is not clear; decisionmaking seems to occur almost exclusively within the 1/ Additional data would be needed to confirm this presumption conclusively. It follows from interview material and from the general similarity of attitudes of district politicians and of national politicians, the latter definitely lacking interest in budgetary questions. 2/ No consistent wage structure exists: Every decentralized company separately discusses the salaries with the different trade unions. WRA/BR-001-2b/03.15.85 - 92 - administrative-political machinery. 1/ The only exception where budgetary expenditures directly relate to the demands of the electorate is the distribution of auxilios. Auxilios may also be viewed as a direct link between potential votes and the budgetary output of the political system. For politicians, budget decisions may be a process in which funds are invested to gain political support. Obviously, this political function is independent of the auxilios' legal purposes. One small exception to the observation that budget decisions are independent of community preferences is the Fondo de Accion Comunal, 2/ a special fund supported by contributions from the Neighbothood Committees. The city government is required by law to carry out small public works projects if the community collects a certain percentage of the costs in advance. "Special Purpose" vs. "General Purpose" Units The fragmentation of the budgetary process and the resulting weakness of coordinating and supervisory bodies is manifested in yet another way. "General purpose" units, especially the planning units, have much less power than "special units" such as the big public service agencies. The planning departments, both at the District and at the national level, do not exercise the necessary control over budgeting decisions; even their technical expertise is somewhat useless since traditional planning criteria (e.g. calculating and comparing marginal rates of return for different investments) are required only occasionally. 1/ Two qualifications are in order. First, interests may be internalized by different bureaucratic units so that direct pressures by interest groups may be unnecessary. Second, individual councillors "patronize" organiza- tions and interests, with the same result. 2/ Administered by the "Departamento Administrativo de Accion Comunal." WRA/BR-001-2b/03.15.85 - 93 - -Auditing The dependence of the Comptroller's Office on the City Council implies various assumptions: first, that the legislative body in practice is able to affect the size and composition of the budget; second, that it is concerned how the budget is actually implemented. As should be evident, the first assumption is highly questionable: The influence of the City Council upon the budget and upon actual expenditures is small. Moreover, the Council normally does not discuss the results of budget's actual execution; indeed the Council's approval of the final accounts of budget implementation, one year after its presentation, is assumed by law 1/ if the City Council has not taken any decision. Clientelism Since there is little justification for the direct election (and, hence, dependence) of the Contralor (auditor) by the City Council, the political reason for this procedure is all the more important. From the fact that 527 persons (190 of which are professionals) work under the direct supervision of the Comptroller's Office 2/ clearly suggests that it is essentially a patronage institution. The larger the institution is, the easier patronage appears; the larger it is, the easier the famous 1/ Art. 211 C.F. 2/ The Comptroller's Office of the Federal Republic of Germany has a staff numbering in the same order of magnitude, while the budgets under its supervision have a volume of about 538:1 relative to that of Bogota. WRA/BR-001-2b/03.15.85 - 94 - "milimetria" should be, i.e., the balancing of the different party wings in relation to the number of their votes. 1/ Colombia thus has essentially a "clientelist" political structure. The so-called "parties" (more accurately, "political camps") do not assume .ideological positions in the European sense. To win followers and active members, they must therefore dispense material gratifications. An obvious means for this is patronage in the civil service in general, and in the Controlaria in particular. 2/ Here the City Council has direct aCcess since the Contralor acts as the Council's agent. A change in Comptroller does not, however, provoke a change in all of the staff members of the Contraloria. In general, though, changes of personnel are frequent in the Colombian bureaucracy. The City Councilors thus exert considerable influence upon the nominations by way of political recommendations. The exposure of the Comptroller to those more or less open political pressures, of course, results from the fact that the Colombian civil service system has no basis in individual merit; and even if it had, the Comptroller's Office represents a typical "puesto de confianza." The City Council is thus 1/ This is a result of the Frente Nacional (1958-1974), the power-sharing between Liberals and Conservatives to bring the Rojas Pinilla dictatorship to an end and to prevent a revival of the "violencia." Representacion equitativa in public service is a right conceded to the respective minority party. According to interview information, patronage by councilors among low-ranking jobs in the decentralizadas is also common; technical positions are apparently not involved. 2/ A professional civil service in the District "Contraloria" therefore does not exist; see Art. 555 C.F. ([Corresponde al contraLor] "nombrar y remover libremente los funcionarios de la Contraloria.") WRA/BR-001-2b/03.15.85 - 95 - more likely to elect a person with some political background rather than a "tecnico" unwilling to play the councilors' political game. In addition, the two-year term of office is so short that the Comptroller normally aspires to re-election; to be re-elected, however, he must maintain good working relations with the City Council, in particular by accepting political recommendations. To protect the Comptroller from these traditional pressures, it would be preferable to make the office a life-time (or at least a very long term) appointment or to prohibit re-election. The first solution has the advantage of insuring that the Comptroller has the technical competence for the position and reduces the possibilities that the candidate will engage in all sorts of promises in exchange for election. These technical advantages are, however, political disadvantanges; it is therefore not surprising that a reform in 1979 opted for the second alternative. Although it is certainly too early for a final evaluation, it is likely that the technical problems created by the political dependence of the Contralor on the City Council will not change drastically. Outdated Control Procedures As for the other expenditure control institutions in Colombia, auditing procedures seem to be based on the assumptions that double (or multiple) controls are more effective than a single one; and a complete (100 percent) control is more effective than a selective or random procedure. If a single control is effective, however, it is questionable whether a second procedure enhances it in any way; or, vice versa, if a first control is demonstrably inefficient, it is doubtful that a second (inefficient in itself) adds anything to overall efficiency. WRA/BR-001-2b/03.15.85 - 96 - Misappropriation of public funds can normally be deterred by making the risk of detection so high that officials in charge of funds are discouraged from illegal operations, even if the particular operation is not subject to a control procedure. This goal can be achieved, among other means, by thorough control of randomly selected cases. The Colombian control procedure, however, assumes that misappropriation cannot occur if auditors are involved. This would be a reasonable concept if the actual auditing did not cause serious delays; if the control system were not so superficial, at least relative to one based on random selection; and if the auditors had no incentives to collude with the agency officials they are supposed to oversee. A radical change of control procedures therefore is necessary. Political implications, of course, require careful consideration given the amount of work auditing creates, the large number of officials involved in the procedure, and the inherent patronage possibilities of the system. It is thus clear that politicians must be offered incentives to reduce the client-based network. The most general rule of politics in Colombia is that clear-cut and final policies are to be avoided. Colombian politicians resist making harsh and irreversible decisions that necessarily foster some groups and discriminate against others. This attitude, of course, is in direct opposition to the goals of rational decisionmaking in the technical sense, i.e., to optimize or maximize benefits as so many planning models attempt to do. The political advantage of the current Colombian approach, however, is WRA/8R-001-2b/03.15.85 - 97 - the possibility at all times to react upon new demands, to solve newly created problems, and to avoid the social tensions that gave rise to the "violencia." The budgetary process also reflects this general decisionmaking pattern. No central authority may dispose of the total expenditures by agencies of the Special District. The District's financial resources are therefore not an effective instrument for implementing political priorities. For example, a mayor who wants to improve the water systems in the barrios and to finance the project with a surplus from the telephone company would find it impossible to do so; the Telephone and Water Corporations have budgets that are strictly separated. The opposite case may also occur: an agency may make expenditures simply because of the excessive availability of credit funds, even while there is a scarcity of funds in other parts of the District administration. A balancing of surpluses and deficits is thus extremely difficult, an expression of the desire to avoid clear decisions for important entities such as the Special District. The budgets proposed by the city Government and authorized by the City Council normally do not contain the last word an actual expenditures. Some expenditure titles are artificially inflated relative to requirements so that transfers to other titles are possible; as noted above, such transferencias and traslados are much more frequent in Colombia than inindustrialized nations. Exaggerating somewhat, the budget acuerdo is not only a set of norms but a point of departure for discussions on what really is to be done. This may also be the case, to some extent, for investments involving international financing. Another characteristic of the fragmented decisionmaking system is the continual rotation of officials and important politicians, resulting in WRA/BR-001-2b/03.15.85 - 98 constant reorientation of policies. Each new mayor, new councilor, and new manager of a decentralized firm has his own ideas about priorities; all too often, earlier plans are discontinued and even projects that are underway may come to an abrupt end (obras inconclusas). Nearly complete investments (e.g., the Hospital Distrital del Norte) are not put into operation for lack of a very small amount of funds, and the responsible officials (also in the national administration) simply are not interested in the matter. If clear and long-term policy guidelines do not influence budgetary decisions, old fashioned incrementalism becomes all the more important. Year by year, expenditures rise at least in nominal terms; as a result of the largely unplanned development of the city, new needs arise that the bureaucracy reacts to. Incrementalism, according to Wildavsky, possesses the important advantage of allowing the government to avoid statements of principle: discussions and evaluations center only around the value of a single proposed project, thus ignoring such questions as the relative merit of investments in road construction vs. education. Antagonisms that exist in every society, but especially in a developing country such as Colombia, are thereby impeded from growing into violence. Generally, the ability of Colombia's current political system to devise and implement a vigorous and concise development policy is rather limited. Actions of the nation and the district constitute but a minor part of the factors influencing Bogota's growth. While the fragmentation of political and administrative power weakens the impact of public actions, the system at least has the advantage that planning mistakes generally do not reach gigantic proportions. WRA/BR-001-2b/03.15.85 - 99 - To summarize, then, it is clear that (at least in the short run) 1/ Colombia's political exigencies and the demands of efficient budgetary analysis and planning are very different. On the one hand, the goal of maintaining the political system and avoiding social upheaval is paramount; on the other, achieving technical and economic optimization is desired. In this sense, delaying or reversing decisions is perfectly rational behavior. VIII. RECOMMENDATIONS Simplication of Control Procedures As long as comprehensive development planning for Bogota and for the nation does not exist, the participation of the District and the National Planning Department in designing the investment budget and in raising credit seems rather pointless. As we have shown, since budgetary decisions are based primarily on the financial condition of the individual agency seeking credit, finance departments are a more appropriate supervisory body than planning departments. In any case, the duplication of review of borrowing requests at the District and at the national level could be avoided by abolishing this procedure at the District level, since at the national level there exists a better system of evaluation and control. It may be noted in passing that the District Planning Office (DAPD) does in fact occasionally introduce urban and environmental criteria into decisionmaking process. Even if the DAPD does not participate in approving 1/ In the long run, it may well be that survival of the political system depends on its ability to meet technical criteria of efficiency. This, however, is beyond the scope of this paper. WRA/BR-001-2b/03.15.85 - 100 loans, however, it could still maintain environmental standards through its control of building permits. Another procedure that needs simplifying is the Finance Ministry's authorization to engage in negotiations with lenders and to sign the resulting contract. Like the double controls described above, the first control and authorization for investments (the Finance Ministry's determination of the terms of planned credits) is never observed; a second control step is therefore required. It would be preferable to retain only one authorization (i.e., the one based upon an examination of the proposal of the credit contract) and to use all possible administrative means to assure compliance. This should not be too difficult to enforce because the National President must sign the contracts; he could thus refuse to approve them if the credit terms were unfavorable. Consolidation of the District Budget If political guidelines were to determine total public expenditures, a more coherent process of preparing, passing, and implementing a consolidated District budget, including the budgets of the decentralized agencies, would have to be devised. Two observations, however, suggest that this idea is not feasible. First, a political center that determines the District's overall policy consistently over a number of years does not exist. The chances of introducing such dramatic changes into the political culture are dim. Budget consolidation and coherent budgetary procedures are not values per se; they would have to serve a purpose (more rational planning and decisionmaking), and if this does not occur as expected, a mere change of procedures is pointless. In addition, a radical reform of this nature would entail far- . reaching legal amendments that run counter to the spirit of the decentralized WRAIBR-001-2b/03.15.85 - 101 - agencies (i.e., reducing the administrative autonomy) and also to established vested interests. Not only would it be nearly impossible to overcome resistance to the changes, but it is also not altogether clear that it would be desirable to do so. Reform of the Comptroller's Office Radical reform of the Contraloria seems imperative. The office should be made independent of the City Council, with the Comptroller's term of office drastically extended. Auditing procedures should be simplified, using one thorough control system based upon random samples. Organizational changes should include: removing permanent representatives of the Contralor from the administrative departments; while at the same time drastically reducing their number; complete independence for the individual official when searching for irregularities; extension of the Comptroller's responsibility to include administrative budget control; and presentation of an annual report that cites financial irregularities, for discussion by the City Council. District politicians would no doubt object to such drastic changes in the Comptroller's Office, simply because of its value as a patronage institution. The only means to overcome this resistance 1/ would be to offer some type of compensation the City Council members. An increase in the auxilios would be one (though not ideal) solution; a partial financing of political parties and their factions by public funds would be another. Policy of International Lending Agencies 1/ It should be recalled that many nationally important politicians have been members of the City Council. To recommend reform by order of the State would overlook the support which City Council could mobilize also at the national level, if they wish to oppose the reform. WRA/BR-001-2b/03.15.85 - 102 - Given the actual situation, the policy of international lending agencies should concentrate on the social, as well as economic, value of individual projects they are asked to finance. To this end, good working relations with the planning units within the particular borrowing institutions seem essential; such relationships are more important than with the City or National Planning Departments, which concentrate on purely financial c:iteria for approving loans. WRA/BR-001-2b/03.15.85 - 103 - REFERENCES FOR CHAPTER 2 Alcaldia Mayor 1972. Bogota, Alcaldia Mayor, Departamento Administrativo Distrital, Desconcentracion Administrativa: Alcaldias Menores, Bogota. Banco de la Republica 1976. (Colombia) Banco de la Republica, Departamento de Investigaciones Economicas, Division Jurídica, Compilacion de Disposiciones Organicas de las instituciones financieras, Vol. II: Junta Monetaria, Bogota. Braybrook, David and Charles E. Lindblom, A Strategy of Decísion. Polícy Evaluation as A Social Process, New York. 1967. Caiden, Naomi and Aaron Wildavsky, Planning and Budgetíng in Poor Countríes, New York. 1974. Camargo, Pedro Pablo, El Control Fiscal en los Estados Americanos y Mexico, Mexico. 1969. Castano, Yesid. "Las Finanzas del Distrito Especial," in: Revista de la Camara de Comercio de Bogota, pp. 95-177. 1969. Concejo de Bogota, D.E<, "La partícipacion de los honorables concejales en las juntas directivas de empreses e institutos descentralizados del Distrito Especial de Bogota," Bogota. 1978. Escuela Superior de Adminístracion Publica, "La toma de decisiones en la administracion publica colombiana. Aspectos de las relaciones entre los gerentes, directores o presidentes de los organismos publicos descentralizados nacionales y las respectivas juntas o concejos directivos," mimeo. Bogota. 1973. Bogota Distrito Especial (Servio Tulio Ruiz, Tesorero de Bogota, D.E.; Jorge Dussan Abella, Secretario General), Estructura Organica y Manual de WRA/BR-001-2b/03.15.85 - 104 - Funciones de la Tesoreria Distrital de Bogota (Resoluciones 1 y 2 de 1977), n.p. (Bogota) n.d. (1977). Amitai Etzioni, A Comparative Analysis of Complex Organizations, New York (2nd ed.). 1975. Contraloria de Bogota, Distrito Especial (Enrique Low Murtra, Contralor de Bogota, D.E.), Informe Financiero 1977, Bogota 1978. Contraloria de Bogota, Distrito Especial (Rosa Natale Rodriguez de Primero, Contralora), Informe Financiero 1978, Bogota 1979. Bogota Distrito Especial, Secretaria de Haciendaa (Jorge Guzman Moreno, Secretario de Hacienda), Informe Financiero 1978-1979, Bogota. WRA/BR-001-2b/03.15.85 WRA/BR-001-3/3.15.85 - 105 - Chapter 3 REVENUE AND EXPENDITURE FORECASTING IN BOGOTA by David Greytak I. INTRODUCTION * In Bogota the principal government unit is that of the Special District of Bogota (BSD). There are also a number of independent or quasi independent entities which perform public functions. Primary among these are five decentralized agencies (empresas decentralizadas) which are responsible for the major public utility services, i.e., Electricity (EEEB), Water and Sewerage (EAAB), Telephones (ETB), Public Transportation (EDTU) and Sanitation (EDS). In addition there are a number of special purpose offices, institutes and revolving funds. In a sense, all these entities engage in revenue and expenditures forecasting. However, for the most, their efforts are limited to the estimation required for the preparation of the annual budget. For others, in particular BSD, EAAB, EEEB, and ETB, intermediate-term forecasts of revenues and expenditures have been prepared on a more or less regular basis. The purpose here is to review the intermediate-term forecasts of revenues and expenditures have been prepared on a more or less regular basis. The purpose here is to review the intermediate-term (3-8 years) forecasting methods and procedures employed by these entities. As will be seen, there are both substantial differences and similarities in the approaches to forecasting. The remainder of this paper consists of two sections. Section II reviews the state of the art in municipal fiscal forecasting. It draws * All references are located at the end of this chapter. WRA/BR-001-3/3.15.85 - 106 - heavily on the experience of the United States in such endeavors as it is there that municipal forecasting has had its broadest application. 1/ Section III reviews and assesses the forecasting efforts undertaken by the government of the Bogota Special District and by the major decentralized public works agencies. II. URBAN GOVERNMENT REVENUE AND EXPENDITURE FORECASTS: THE BACKGROUND Evaluation of public sector expenditures and revenues ha!r recently emerged as an important element in the fiscal planning of local governments in many areas through the world. In large part the driving force underlying the growing interest in fiscal forecasting is the need to inform various officials, department heads, and project planners of the budgetary implication of policy decisions and the budget constraints that may confront them in formulating policies in the future. The most immediate and the most common reason for developing expenditure and revenue estimates is to evaluate the short-term budgetary implications of anticipated service levels and revenue for the next fiscal year. It is a process regularly undertaken as part of budget preparation by local governments worldwide. However, as annual budgets generally have a flavor of committment and accountability, they usually require that the revenue and expenditure estimates be those most precisely related to the policies and conditions expected to prevail during the budget year. Thus, the process is essentially one of balancing the various departmental and agency 1/ Department of Housing and Urban Development, "Multi-Year Revenue and Expenditure Forecasting" (Washington, D.C.: Office of Policy Development and Research, 1979). WRA/BR-001-3/3.15.85 - 107 - expenditure obligations and requests against the expected revenue and borrowing capacities. This is not to say that in budget preparation the revenue and expenditure implications of alternative policies such as employment cutbacks or tax rate increases are igr :s. Rather the consideratior. of such alternatives is most often undertaken as an exercise in the final stage of budget preparation as an attempt to balance overall budget requests with anticipated revenues. In this regard, the governmental institutid-ns in Bogota (i.e., BSD, EEB, etc.) are similar to their counterparts in other areas of the world. Budgetary numbers, however, do not represent true forecasts. Rather they represent the obligations and commitments which have been selected as a result of considering objectives and goals and the ability to achieve them in the context of the economic and political circumstances which are expected to evolve during the budgetary period. 1/ Alternatively, forecasting is an attempt to depict some set of outcomes which will likely accompany some set or sats of circumstances which may or may not be expected to evolve during the future. Forecasts and the predictions they contain are primarily designed as 1/ In most countries, there are provisions for altering local government budgets via adjustments and supplements as conditions require during the course of the budget years. In some areas, local governments are afforded a great deal of flexibility in this regard. For example, in the Philippines, local government effectively operate as if they were on a monthly budgetary cycle. For a discussion see, "Strengthening the Fiscal Performance of Philippine Local Governments: Analysis and Recommendations", Part 1, (Roy Bahl, David Greytak, Larry Schroeder, Kenneth Hubbell and Benjamin Diokno), Research Report, Local Revenue. Administration Project, The Maxwell School, (Syracuse, NY: Syracuse University, 1981). WRA/BR-001-3/3.15.85 - 108 - an aid to informed decisionmaking. 1/ This is not to say that forecasting may not be an integral part of the budgetary process, for it may be, indeed it should be. But the value of forecasting lies in its ability to explore the outcomes of a variety of conditions, some of which may be anticipated, but the more important of which, like the revenue impact of changes in various tariff rates, may be purely conjectural. Of particular import is the recognition that fiscal projections or forecasts are not a description of what a government or agency will be doing at some future date. That is,-forecasting is a means of exploring the implications of various possibilities. Some of these may be strong commitments, e.g., wage contracts. Others may be more tentative, e.g., increases in service levels. Perhaps some of the more important may be hypothetical, e.g., the rate at which population and/or income may increase, inflation rates etc. The value of forecasting is that it provides a means of exploring what might be expected to follow from some specific changes in some given set of circumstances or conditions. While there is little doubt that interest in local government expenditure and revenue forecasting is on the rise, the extent to which it actually is practiced has not been documented. Even in the U.S. where local 1/ In general usage, the distinction between the terms forecast and predictions is akin to the difference between foretelling and predicting. The former implies accuracy and certainty while the latter connotes some degree of uncertainty, usually the existence of a specific set of circumstance. It is the latter connotion, estimates based on the assumed realization of a specific set or sets of conditions which is appropriate to the issues at hand. Generally, the literature does not distinguish between the two terms or where it does, it does so in such a subtle manner that the general convention is to use the terms interchangeably. In what follows, this convention will be followed--the terms will be considered synonomously--both taken in the sense of conditional estimates. WRA/BR-001-3/3.15.85 - 109 - government forecasting, perhaps, has its longest history, there exists no complete inventory of forecasting practices and procedures. In fact, the most comprehensive information on fiscal forecasting practices by local governments in the U.S. has been derived from a relatively small sample of large cities and counties. 1/ All 28 of the responding governments considered themselves to be involved in forecasting, however, most focused on a few components of their revenues and/or expenditures with the majority constraining their efforts to the comparatively short time horizon of budgetary cycle-s of one to three years. One would suppose that if this is the situation of the larger local governments in a country where there is ready access to the requisite talent, experience and equipment, then comprehensive intermediate term forecasting by local governments generally, and more particularly in developing countries, is not a common practice. If this is the case, then intermediate forecasting in Bogota may be more advanced than in many other areas as a result of its recently developed forecasts by the BSD, the major public agencies in Bogota, (including EEEB, EAAB and ETB) engage in intermediate term forecasting on a more or less regular basis. As to the state of the practice, the available information about the techniques used in forecasting is again scarce. If the experience of the abovementioned local governments in the U.S. is reflective of what is common, then most local governments do not take advantage of even the more elementary of recently developed forecasting techniques. Rather, multi-year forecasting has been accomplished by methods which essentially constitute simple 1/ "A Process for Revenue and Expenditure Forecasting in Local Government," Management, Finance and Personnel Task Force, The Urban Consortium, Unpublished mimeo, February 1977. WRA/BR-001-3/3.15.85 - 110 - extrapolations of the recent past. Even in this regard the variety of methods, models and procedures which have been employed is such that there cannot be said to be a general approach to Local government forecasting. To an extent, forecasting practices in Bogota can be similarly characterized. There are instances, in particular the BSD, where the forecasting techniques are quite rudimentary. But in others, in particular some elements of EEEB and EAAB forecasting, the underlying practices and techniques are complicated and fairly sophisticated. Such differences are not surprising for the forecasting techniques and procedures employed have been selected and tailored by the different institutions independently and for their specific purposes. Whether this is a result of the latters' compliance with the requirements of inter- national lenders is an issue that need not be dealt with here. The result is a diversity of approaches to forecasting whose variety precludes the charac- terization of forecasting in Bogota as of a general or common approach. General Purposes 1/ Traditionally, intermediate term revenue and expenditure forecasting has been most commonly used for two purposes: fiscal gap and impact analysis. In general, the objectives of both types of analysis are to explore potential trends in the growth of revenues and expenditures. The basic difference between the two types of analysis is one of focus for, in Bogota, as well as elsewhere, the same procedures are often used for each 1/ This review draws in large part upon Roy Bahl, Larry Schroeder, "Forecasting Local Government Budgets," Occasional Paper No. 38, Metropolitan Studies Program, The Maxwell School, (Syracuse, NY: Syracuse University, 1979), p. 14. WRA/BR-001-3/3.15.85 - Ill - purpose. 1/ Indeed, it is often the case that the elements of the two analyses are carried out simultaneously. That is, intermediate term forecasts generally involve prediction of the impact of specific policy changes in the context of alternative assumptions about the macroeconomic factors which govern the growth in revenues and expenditures. A more recent development is the use of forecasting techniques to consider the sensitivity of expenditures and revenues to their determinants, particularly, those which can be influenced by policy. Fiscal gap analysis is concerned with the budgetary implications of various scenarios of behavior in an area's macroeconomy with particular concern focused on those economic variables beyond the control of Local governments. It generally involves the estimation of current account expenditures and revenues based on alternative assumptions about such factors as growth in population, production, and inflation. The objective is to compare alternative futures and to determine the likelihood that revenues will at some future point differ from expenditure requirements. This type of analysis allows policymakers to consider the magnitude of revenue and expenditure adjustments which may be possible in the case of a surplus or necessary in the case of a shortfall. In the U.S., it is often employed as a means of determining whether adjustments in tax rates will be required in the near future. Another objective frequently served by gap analysis is related to capital planning. By projecting anticipated revenue and expenditure patterns, the amounts of capital which can be feasibly financed for public investment can be evaluated. 1/ These will be discussed below. WRA/BR-001-3/3.15.85 - 112 - It is to the service of this end, i.e., determination of debt carrying capacity that the BSD has recently introduced its multi-year forecasting. This too, has been an objective of the multi-year forecasting of the major decentralized agencies in Bogota. In particular, EEB, ETB and EAAB forecasting has been an integral part of of their investment planning process. The most common use of impact analysis is to consider the fiscal implications of the various possible discretionary policy decisions which may be under consideration. Such information can be a valuable input in the decisionmaking process. A common use for impact analysis in the U.S. and by the decentralized agencies in Bogota as well is the projection of future debt service costs required to complete new capital projects. Indeed, this is the principle purpose served by ETB, EEB and EAAB forecasting. Beyond this, these institutions, unlike BSD, have developed projection methods which accommodate the analysis of a variety of other impacts. In particular, EAAB and EEB forecasts generally are built upon a series of assumptions about such things as service levels, tariff rates, inflation rates, debt financing costs and depreciation rates. These analyses serve the purpose of providing forecasts of the revenues and costs which could be expected under the assumed circumstance. Perhaps because they have been primarily used in conjunction with capital financing programs, forecasting efforts in Bogota has not been generally directed to what is perhaps one of the more valuable uses. That is, sensitivity analysis. The purpose of sensitivity analysis is to identify the response of costs and/or revenues to specific factors, be they policy decisions, e.g., wage contracts, or changes in underlying economic conditions, e.g., inflation rates. The objectives of sensitivity analysis are best served WRA/BR-oo1-3/3.15.85 - 113 - when the principal determinants of revenues and expenditures are incorporated in forecasting models in a way which identifies the impacts that a change in some factors may have both directly and indirectly on revenues or expenditures as the case may be. For example, a negotiated wage rate increase will directly increase wage and salary costs but its effect on aggregate labor costs may be much greater if the levels of other types of costs, e.g., pension and retirement benefits are tied to wages and salaries. To an extent, EEB and EAAB have been concerned with the sensitivity of their revenues and expenditures to a variety of factors, e.g., levels of services, inflation, tariff rates, wage costs. However, their forecasting has been undertaken as an element of investment planning; it has been viewed only incidentally as a tool of ongoing analysis underlying the management process. Common Practices The reliability of the forecasts, as well as the method employed to obtain them, depends in large part upon the quantity and quality of data and policy plans available to the forecaster. Ideally the forecasts are derived from a model which formulates the appropriate behavioral relationships among the relevant determining factors or variables. In addition, once the behavioral relationships have been formulated in the form of a model and the model estimated, impact and fiscal gap analyses require the use of projections of future policy alternatives and economic conditions. Thus, the first requirement for most projection methods is an adequate data base with which to work. In practice, the availability of data and tesources to properly forecast, and the size of the local government unit have been the overriding determinants of the methods used in projecting local government revenues and WRA/BR-001-3/3.15.85 - 114 - expenditures. As a general rule, the larger the entity, the more likely it is that it will have access to the facilities required for the more sophisticated statistical methods. Also, as larger cities usually have more diverse revenue sources and expenditure obligations, the more important is it that the individual components of each should be systematically projected in order to avoid mistakes in the gap and impact calculation. 1/ Trend Extrapolation. The simplest of the calculating techniques for making budget projections is trend extrapolation of historical data series where revenues and expenditures are functions only of time and it is assumed that the components of the budget grow at a constant rate. This may not be an unreasonable assumption in the absence of any major changes in policy or unexpected economic fluctuations. However, with time trend extrapolations it is not possible to determine the budgetary implications of other changing economic conditions (e.g., changes in the rate of inflation), 2/ and the important effects of policy changes cannot be evaluated. 3/ 1/ When there is an insufficient data base to draw upon or when the resources required for the use of more sophisticated technique are lacking, forecasts often represent simply the best guss of an informed official. This may be the most cost-effective technique for smaller local governments and/or for some short-term budgetary purposes. Still, informed judgments based on the knowledge and intuitive feel of the forecaster are most likely to be on or close to target for shorter term budget projections. The longer run allows room for the significant changes which erode the knowledge and experience which make for expertise. As a result, the reliability of the ad hoc forecasts usually decreases as the forecast period lengthens (and as a jurisdiction grows), thus giving rise to the need for a more systematic technique. 2/ Bahl and Schroeder, "Forecasting Local Government Budgets," p. 14. 3/ Ibid. WRA/BR-001-3/3.15.85 - 115 - In practice, some local governments in the U.S. adjust their time trend forecasts to reflect anticipated changes in major policy variables, e.g., wage rates, service levels and/or economic conditions, e.g., inflation rates. However, even the large local governments have been faulted for simply extrapolating on the basis of current wage and tax rates, and recent inflation and population growth. 1/ The problem with trend extrapolation is that it provides no indication of dependence or causality between economic and policy variables and budgetary elements. The influence of such factors can be evaluated but to do so requires more elaborate (e.g., behaviorally based) models and more extensive data bases. Be that as it may, the severe data limitations (e.g., data availability and reliability) which exist in many developing countries may preclude the use of more sophisticated projection methods. Thus at times, some combination of expert "guessing" and trend extrapolation is the best available technique. Among the larger governmental units in the U.S., trend extrapolation is frequently used to project the smaller categories with more sophisticated techniques applied to the larger, more important categories. In Bogota, trend extrapolation plays a major role in the forecasting of two of the major units. In the case of the BSD, intermediate term forecasts are based on straightforward trend extrapolation of aggregate own source revenue. The simple nature of the extrapolation and the use of a single extrapolation as the basis of the major categories of revenues and expenditures greatly qualifies the usefulness of their forecasts. 1/ David J. Ott, Attiat F. Ott, James A. Maxwell and J. Richard Aronson, State-Local Finances in the Last Half of the 1970s (Washington, D.C.: American Enterprise Institute for Public Policy Research, 1975), p. 12. WRA/BR-001-3/3.15.85 - 116 - Alternatively, EEB has developed a complicated trend extrapolation method to forecast its sales by type of consumer as a basis of its revenue and expenditure forecasts. Such a disaggregate approach is certainly more appropriate than relying on a limited aggregate forecast as does the BSD. However, aLhough disaggregation does allow for the incorporation in the forecasts of behavorial differences among consumer classes, the basic approach is still one of trend extrapolation and thus suffers from the abovementioned limitations. It should be noted however, that EEB extrapolation of consumption is but one element, albeit the major, of its forecasts. It is combined with assumptions about elements of costs and tariff rates to generate revenue and expenditure forecasts. The Major Determinants Approach. Under the major determinants projection technique, other variables besides time are taken to determine expenditures and revenues. Specifically, a stable relationship is established between expenditure and revenue levels and some exogenous variables such as local population and/or income. Two determinant relations often used in expenditure projections are the per capita multiplier method and the service standard method. 1/ Under the per capita multiplier method it is assumed that changes in future expenditures are equal to current per capita costs multiplied by the future change in population. Alternatively, given a policy decision on future service levels, projected expenditures under the service standard method are calculated by multiplying the desired increase in population being serviced times the labor and nonlabor expenses required to attain that service level. 1/ "A Process for Revenue and Expenditure Forecasting in Local Government," Management, Finance and Personnel Task Force, The Urban Consortium, (Unpublished mimeo, February 1977). WRA/BR-001-3/3.15.85 - 117 - This method is often used in developed as well as developing countries where local governments are attempting to expand service levels to accommodate increased population in its jurisdiction. Of these two methods the per capita multiplier approach is the simplest and not so frequently used any longer. Indeed, it does not appear to be a part of the forecasters tool kit in Bogota. Perhaps because their concern with estimating investment costs for future undertakings, some of the smaller organizations, e.g., ETB and Caja de Vivienda Popular, employ something akin to the service standards methods. However, they do not attempt to use this or other techniques for multi-year forecasting in the systematic manner as does the BSD or the major decentralized agencies. Recently an expanded version of the major determinants approach has been designed primarily for the dual purposes of sensitivity analysis and forecasting. Its focus is on the cost of providing services but it disaggregates those costs not only by categories of services, usually identified with a particular government function or department, but also by principal factors determining costs. This approach identifies the contributions of such things as changes in employment levels (by detailed classification), wage rates, inflation, material inputs to expenditures. By partialing out the detailed bases, it provides a means of evaluating the expenditure implications of alternative assumptions about future values of the cost determining factors. This particular approach has been termed the public employment approach primarily because such large proportions of most local WRA/BR-001-3/3.15.85 - 118 - government costs are related to employment levels, either directly (wages, pensions and fringe benefits) or indirectly (supplies and materials). 1/ It is readily apparent that this approach is particularly appropriate for intermediate term impact analyses, with the planners developing alternative policies with different labor and nonlabor input requirements. Unlike straightforward trend analyses, it provides a valuable tool for policy analysis, as it allows the analysts to evaluate the expenditure implications of a variety of policy options including changes in employment levels, wage rates, fringe benefit packages, as well as those related to the general economic climate, e.g., rates of inflation. This approach to local government forecasting is relatively recent and this has had limited application. 2/ A variant of the determinants approach, which has a substantial history in revenue forecasting, is elasticity-based projections, with percent changes in revenues (both taxes and user charges) being related to percent changes in some base, most frequently, income. 3/ In practice, the elasticity approach generally has been applied in a rather mechanistic way to translate income growth into revenue increases. The problem with this approach is twofold. First, it assumes that revenue bases are equally affected by all 1/ Robert W. Buchell and David Listokin, The Fiscal Impact Handbook, (New Brunswick, New Jersey: The Center for Urban Policy Research, 1978), pp.5- 6. 2/ Roy Bahl, "Forecasting Urban Government Expenditures," (with Richard Gustely), Proceedings of the Sixty-Seventh Annual Conference of the National Tax Association, 1974. For early application see, Claudia Scott, Forecasting Local Government Spending, (Washington, D.C.: The Urban Institute, 1972). 3/ See, for example, Charles Mansfield, "The Elasticity and Buoyancy of a Tax System: A Method Applied to Paraguay," in Intergovernmental Monetary Fund Staff Papers (Washington, D.C.: July 1972), pp. 426-428. WRA/BR-001-3/3.15.85 - 119 - income changes whether they be purely nominal or real and whether they are due to growth in population or to higher average incomes. Second, it assumes that the factors governing the revenue response to changes in tax bases, e.g., tax rates, administrative efficiency, do not change during the period of analysis. For the purpose of forecasting, these assumptions may not be problematic if they are realistic, but that is doubtful. Moreover, the elasticity approach as usually employed, does not differentiate among the various factors determining the effective rates and growth in the-revenue base and thus is of limited use to policy analysis. In fact, revenue estimates based on elasticities alone are less common than they once were. More frequently, elasticities are now used primarily as a means of classifying the responsiveness of taxes and tax structure. Econometric Models. In all of the above forecasting techniques the forecaster has essentially played a passive role in that he has searched for estimates to explain the relationships between a set of economic and demographic variables. In the econometric approaches discussed below, the forecaster first identifies the relevant explanatory variables for the various revenue and expenditure categories and then develops a functional form that best explains the underlying behavioral relationships. 1/ There are three important benefits of econometric based projections: (1) having based a projection on a theoretical model, statistical inference can be applied to test for levels of significance of the 1/ For an in-depth discussion of the econometric formulation of projections and the statistical .-fficulties associated with it, see C. Kurt Zorn, (Ph.D. Dissertation, Syracuse University, 1981). WRA/BR-001-3/3.15.85 - 120 - coefficients (i.e., t-tests) to establish firm relationships between the dependent variable and the independent variables, (2) the resulting projection equation can be used to conduct impact analyses of various discretionary policy alternatives, and, (3) to the extent that such variables are included, it is possible to evaluate the fiscal gap implications of changes in socioeconomic variables such as population and income. Most large cities in the United States use some form of the independent time series-regression method for projecting revenues, while expenditures are usually projected by some variation of the determinant method discussed above. 1/ Therefore, in discussing econometric approaches to forecasting, examples are confined to the revenue side of the budget. It is significant to note that econometric approaches have been of growing popularity in the local government forecasting in the U.S. Econometric techniques are also used by the BSD, EEEB, and EEAB. However, in Bogota, their use is primarily as an estimating method in time trend analysis, usually of just one of the revenue determining factors as is the case of electricity and water sales by EEEB and EAAB. The more general approaches used in the U.S. is to develop explanatory or independent variables for the regression equations or some combination of factors thought to be the principal determinants, e.g., personal income, local population, the revenue base, the revenue tax rate, a price index, and dummy variables for discretionary policy changes. The equations are estimated in a time series manner and are usually in a linear or 1/ Bahl and Schroeder, "Forecasting Local Government Budgets," p. 51. WRA/BR-001-3/3.15.85 - 121 - log-linear form, with the decision as to the correct functional form often dictated by the equation yielding the highest R2 (the best fit criterion). 1/ Although economic theory should be the final judge both about which independent variables are to be included and the appropriate functional form of the equation, for several reasons data availability usually constrains the forecaster in his choice of independent variables. The lack of a time series of adequte length can restrict the number of independent variables due to degrees of freedom difficulties. In such cases, it is common for the forecaster to include in the estimating model only the variables which explain a large portion of the variation in the dependent variable. In addition, often the data for the explanatory variables are simply not available at the local level. Under such conditions two alternatives are common. A substitute or proxy variable can be used in place of an unavailable one, e.g., the common use of a general price index when the desired measure is a more specific measure of prices (e.g., sales value of property). Alternatively on the assumption that a local variable is functionally linked to a regional or national counterpart the latter may stand as a substitute for the unavailable local data. For example, assuming that the local business cycle parallels national economic trends, a measure of local personal income could be obtained as a linear function of national personal income. Or, as is not uncommon, gross national product may be used in place of its local equivalent when a local income measure is not available. Finally, in order to utilize the 1/ For a discussion and application of a statistical method of determining the appropriate functional form of elasticity estimating equations, see David Greytak, "Functional Form in State Income Tax Elasticity Estima- tion," National Tax Journal Vol. XXXII, No. 2, June 1979, (with Jerry Thursby). WRA/BR-001-3/3.15.85 - 122 - equation for projecting future revenues there should be reliable projections of the dependent variables. If these are not available on the local level, they can be obtained by linking the local variable in some manner to a national variable that is regularly projected. Alternatively, taking into account recent policy decisions which could affect the future values, projections of independent variables may be based on extrapolation from past observations. 1/ Although there are a variety of other problems, those of the availability of the requisite data for estimating the model and for establishing the future values of the independent variables are of the first order. 2/ They confront forecasters even in areas like the U.S. where data are plentiful. Statistical accuracy as well as evaluation of the model generally require that the number of observations exceed the number of independent factors included in the model to an extent, i.e., time series of at least 20 to 30 observations, that is difficult to accommodate, even in the U.S. For many of the factors underlying public sector revenues and expenditures in Bogota, e.g., income, employment, characteristics of various population groups, the unavailability of data all but precludes the use of broad based econometric modeling like that becoming popular in the U.S. In addition, other problems arise when the available data base does not have the length or depth necessary, consistent with efficient estimation of the 1/ Bahl and Schroeder, 'Forecasting Local Government Budgets,' p. 20. 2/ C. Kurt Zorn, "Local Econometric Multi-Year Revenue Forecasting: Problems and Solutions," (Ph.D. Dissertation, Syracuse University, 1981). WRA/BR-001-3/3.15.85 - 123 - behavioral model. 1/ In particular, when the estimating equation includes only a few of the explanatory variables, long-term forecasts become problematic for as structural relationships between variables evolve over time other socioeconomic variables may become more relevant. III. MEDIUM-TERM FORECASTING IN BOGOTA Bogota Special District The methods and procedure used by the BSD for medium-term forecasting represent an extention of those employed in their preparation of their annual budget. Thus, a brief discussion of the budgetary process is appropriate prior to the consideration of forecasting methods and procedures. 2/ BSD Budgeting. In the construction of its annual budget, the Bogota Special District confronts a constraint which is imposed on local governments generally. That is, its anticipated revenues must equal or exceed its planned -expenditures for current operations and debt retirement. Unlike many local governments, or at least unlike those in the United States, the BSD is additionally constrained in that it is not at liberty to alter the rates of any of its major taxes in the event of a projected revenue shortfall. As will become apparent, this constraint has greatly influenced the procedures employed by the BSD in developing its forecasts. 1/ In addition, statistical problems can arise such as omitted relevant variables leading to autocorrelation and, high correlations among the independent variables (i.e., multicollinearity) causing t-tests to be suspect. 2/ For a detailed consideration of the budgetary and planning process in Bogota, see Chapter 2 above. WRA/BR-001-3/3.15.85 - 124 - The first step in the preparation of the budget is the estimation of anticipated current accounts revenues. To accomplish this, revenues are estimated separately for each of the 19 major categories (rubros) of tax, shared tax and non-tax revenues. While there is generally conformity in approach, the estimates which ultimately enter the calculation of the total anticipated revenues for budgetary purposes may be the product of one of a number of estimation models. For some, but not all sources, revenue estimation involves the application of three different types of revenue estimation procedures. This first, and most common, is a time series extrapolation of past revenues. The second, and least common, termed the economic base study, forecasts revenues as an extrapolation of the tax base. The third is referred to as the base legal (legal base). It projects a flat 10 percent increase in collections. It follows from a legal restriction which limits the revenue growth for budget purposes to 10 percent in the absence of a more sophisticated procedure justifying a forecast for larger increments. 1/ The base legal method and the straight trend extrapolations are applied to each of the revenue categories while the use of economic study methods is selectively applied depending on the availability of base data. In general, the time series extrapolations involve the regression of revenue receipts on a time variable using two different equation forms. The first is a simple linear regression of collections on a time variable which has the value of unity in the base year and is incremented by one 1/ In actuality the base legal limits the estimate of revenue growth to 10 percent of the revenue collected during the first half of the current year and the second half of the previous year. WRA/BR-001-3/3.15.85 - 125 - chronologically (i.e., a counting variable). The second equation involves the same data with a log transformation of the dependent variable. Reliance on these relatively straightforward techniques is dictated by the data series which covers the comparatively short period since the last major alteration in the District's tax structure, i.e., 1974 to the current year. 1/ The choice of which method or equation form to be used in the projection of revenues for budget purposes essentially is determined by the goodness of fit criteria (i.e., high R2). In the case of unsatisfactory fits regression results, generally defined as a low R2 (although in some cases it is a matter of low revenue estimates), the base legal provides the revenue estimates. Once having selected a set of estimates for current accounts revenues, the revenue side of the budget is completed by the addition of capital resources, i.e., loans and the surplus or deficit from the previous years. The next step of the budgetary process is the assignment of revenues among the various dependencies of the district. This process is carried out under the direction of a council consisting of the District Mayor, the Budget Director and the Secretary of Finance. The essence of the allocation is a comparison of current expenditures and the amounts requested by the various agencies, funds, etc. While each entity has the responsibility of preparing and providing its own budgetary request, it is the council which ultimately determines the allocation among the various departmental agencies. To be sure 1/ It should be noted that there are techniques which can be used to adjust tax data to a comparable base by normalizing the impact of discretionary changes in the tax system. However, these are designed to deal primarily with changes in rates rather than major alterations in the basic structure of the tax. WRA/BR-001-3/3.15.85 - 126 - these allocations and the budget estimates or requests underlying them are guided by recent or expected outcomes of collective bargaining, debt retirement obligations and plans for improvement in the quantity and quality of services. These, or similar procedures are not uncommon among local governments. worldwide. They are noted here because they form the foundation of the long- term forecasting procedures employed by the District. BSD Forecasting. Forecasting beyond the immediate budget year by-the Bogota Special District government is a relatively recent undertaking initiated primarily for the purpose of determining potential debt carrying capacity. The intermediate forecasts are derived from a method which involves procedures whose nature is even more restricted than those +nderlying the annual budget projection. The most recent forecast, dated 1980, are based on a rather short data series running from 1973 to 1981 with a forecast through 1986. While this would imply a forecast period of five years, data for 1980 are based on estimates as of May 1980, while 1981 data are those contained in the preliminary budget for 1981. Thus, since all post-1979 data are estimates, the forecast period could be said to be of seven years. Whether the forecasting period is taken as five or seven years, the fact is that they are based on a relatively short series of actual data, i.e., 1974-1979. In essence, the BSD intermediate forecasts produce revenues and expenditure estimates for a number of years exceeding that of the data upon which they are based. The limited number of basic observations, of course, at least in a statistical sense, greatly limits BSD's ability to allow for the variety of WRA/BR-001-3/3.15.85 - 127 - factors which likely determine the pattern of revenue growth. Of course, as the experience of each additional year is catalogued, this limitation becomes less restrictive. But still, something much more than five to ten years of observations is desirable, at least for purposes of statistical evaluation of the forecasting model. In comparison, even the more sophisticated forecasting undertaken by local governments in the U.S. has been based on limited series observations which span perhaps 10 to 15 years on the average. 1/ The implicat-ion is that with continued data collection, the BSD will approach what may be termed the state of the practices, at least in terms of revenue (and expenditure) components of the data base underlying some of the more sophisticated forecasting models currently used by local governments in the U.S. BSD Forecasts: Revenues: As is the case of budget estimation, the first step in the forecasting procedure is the generation of revenue estimates. Whereas in the budgetary exercise total current account revenues are calculated as the sum of the projected amounts for each revenue source, the only current account revenue figure reported for the forecasts is total revenues (Table 3.1). The forecasted values of total current account revenues are generated by trend extrapolation using a semi-logarithmic equation, InR = 13.53376896 + .26145077 (X), with R and X representing revenues and time, respectively. The coefficient of determination (.0032) is quite high and can be taken to indicate that the equation conforms closely to the historic performance of total revenues. 1/ Bahl and Schroeder, "Forecasting Local Government Budgets." WRA/BR-001-3/3.15.85 - 128 - TABLE 3.1 SPECIAL DISTRICTS OF BOGOTA /a REVENUES AND EXPENDITURES 1980 and 1985 (thousands of Pesos) 1980 1985 I. Current Revenues /b 5,007,248 17,386,041 1) Ordinary Revenue 4,991,048 17,386,041 2) Capital Contributions 16,200 --- II. Current Expenditures 4,147,898 8,820,002 1) Operating Expenses /c 2,184,704 5,215,812 2) Transfers /d 1,522,983 3,129,487 3) Financial Costs (Public Debt) 440,211 474,703 III. Current Savings 859,950 8,566,039 IV. Financing (Net Credit) 2,280,467 (931,309) 1) Resources from Credit 2,445,859 /e --- 2) Amortization 165,392 931,309 V. Available for Investment 3,140,417 7,634,730 1) Investment 2,681,192 /e 7,634,730 VI. Surplus or Deficit on Investment 459,225 a/ Revenue estimated to the 31st of December based upon the collections up to May 31. b/ The estimation for 1981 was carried out based upon the Provisional Budget for the current fiscal year 1981. c/ The estimations for 1980 and following years are 30 percent of ordinary revenue. d/ This classification was estimated for 1981 and the following years and represents 15 percent of the ordinary revenue. e/ Budget 1980. SOURCE: Finance and Taxation Division of the Secretary of District Finance, Bogota Special District. WRA/BR-001-3/3.15.85 - 129 - Perhaps it is for this reason that this particular equation was chosen as the forecasting model. Indeed, it is common in local government forecasting everywhere that among the possible equations which could be used for generating forecasts, that which is to be actually used is selected on the basis of a high coefficient of determinC-tion (or correlation). But this should not be the sole criterion, for each equation form has imbedded an assumption about the nature of the relation between the variables. Semi- logarithi4,c equations like that estimated for the BSD revenue forecasts imply not only that revenues will grow over time but that they will grow at increasing rates. 1/ In fact, the estimated equation implies that four years into the forecast period, 1985, the annual rate of revenue growth will have increased by 50 percent, i.e., from 200 percent in 1981 to 300 percent in 19,85. Moreover, the estimated equation implies that in 1989 revenues would grow by 400 percent. It could be argued that in those periods when substantial population growth is accompanied by rapid inflation, annual revenue increases may be 1/ The revenue time relation implied in the BSD forecasting equation is (13.5338 + .26145X) which implies the growth in revenues over time is increasing, i.e., dR = (.26154)e (13.5338 + 2145X) at an increasing rate dX dR2 = (.26145)2e (13.5338 = 26145X). dX Substituting the revenue forecasting equation in the surplus equation given above, current accounts surplus are projected as S = .52R = (52) (.26145)e 13.5335 + .26145X) which implies that over time the growth in surplus dS = .26145X, increase S WRA/BR-001-3/3.15.85 - 130 - quite high. However, extrapolating past growth exponentially into the future can be quite hazardous for even a slight moderation in growth of the factors which have generated past revenue increments, e.g., population, or price, would result in over-estimation errors which would compound themselves over time. BSD Forecasts: Expenditures. The expenditure forecasts (Table 3.1) of the BSD lists two major types of spending, current and capital. Included in the first are current expenditures which consist of operations cost, Eransfers, public debt and savings on current accounts, i.e., categories II and III. The second is related to the financial cost of new investments IV, investment spending V and overall surplus VI. Of the three categories of current costs, two, operations and transfer expenditures account for the vast majority, i.e., about 90 percent. The third category, the costs of servicing public debts differs from the others in that the terms of existing debt, the larger part, are known and thus need not be forecasts. Thus, in relation to current costs, BSD forecasting efforts are focused on the projection of operations and transfer costs. Projected spending in the investment categories are derived from investment plans, but are constrained to total no more, and in fact are projected to equal savings on current account. Thus, the new investments spending projection of the BSD forecasts are determined by the projected current accounts surplus, i.e., the difference between projected revenues and projected current costs. The procedures by which the BSD forecasts current operations and transfer expenditures are similar. That is, the projected amounts are calculated as fixed proportions of current revenues, 30 percent in the case of WRA/BR-001-3/3.15.85 - 131 - operating expenditures and 15 percent for transfers. Savings on current account are then projected as the difference between projected revenues and projected operations and transfer expenditures plus the costs of financing public debt. While it would be tempting to consider this surplus as a forecast of the fiscal gap on current account, such a reference would not be consistent with the general use of the term. Generally, estimates of fiscal gaps, such as current accounts surpluses or deficits, are derived from independent forecasts of revenues and expenditures. In the case at hand, however, it could be argued, and quite appropriately so, that given the existing and committed borrowing, the magnitude of the current account surplus is determined by the revenue forecast. This follows from the fact that the current operations and transfer cost componenLs are forecast as fixed proportions of revenue, while public debt service costs are given. As service of the public debt is slated to grow rather slowly in comparison to revenues, the BSD forecast procedures generate an increase in current accounts surplus that is quite dramatic. In fact, the reported forecasts project a current surplus which amounts to 49 percent of forecasted revenues in 1985. By comparison, current accounts surplus in 1980 amounted to just 17 percent of current revenues. Although not explicitly specified, the BSD forecasting procedures project the current account surplus to increase at increasing rates over time WRA/BR-001-3/3.15.85 - 132 - such that it will approach an amount approximately 52 percent of current revenues. 11 Forecasting current account surpluses as a proportion of revenues may appear to be a dubious practice but it would seem to be consistent with budgetary practices. As was indicated above, the uses of the District's revenues is determined by the District Council in line with their commitments and priorities. Thus, subject to the servicing of debt, the council can determine the level of forecasted current accounts savings by setting the proportion of forecasted revenues to be allocated to the operating and transfer components of current spending. But still, this is not a common procedure, for forecasted surpluses (or deficits) are usually based on the difference between revenues and expenditures which are independently forecast. Be that as it may, the next element of the forecasts is one which is generally a part of fiscal gap analysis. That is, from the estimated current accounts savings, debt amortizations and other obligations are subtracted while available capital funds are added. The result is the amount of funds available for investments or to be retained as a general budget surplus. Indeed it is the determination of the amounts likely to be available for 1/ This can be seen as follows. Letting S stand for surplus, R for revenues, E for expenditures and a for debt service costs are assumed to equal respectively 30 and 18 percent or together 48 percent of revenues. S R - E E .48R + a S R - .48R - a S .52R - a which approaches 52 percent as a is forecast to increase on very little. WRA/BR-001-3/3.15.85 - 133 - future investment which is the basic reason for BSD initiation of intermediate forecasts. The most recent forecasts are based on an overall budget balance. That is, all available capital funds are invested each year. Quite appropriately the attendant interest and amortization cost of new investments appear as expenditures in the forecasts for subsequent years. 1/ BSD Forecasts: Summary. The procedures employed for a medium range forecasting by the Bogota Special District represent a combinationi of time trend extrapolation plus the addition of estimated (not forecasted) investment costs. In a strict sense, statistical elements (i.e., the use of regression techniques) are limited to trend extrapolation forecasts of current operating revenues. However, as the major elements of current expenditures are assumed to remain a fixed proportion of total revenues, in essence they too are forecast by a time trend extrapolation. Investment and debts obligation thus are not truly forecast. The principal use of the forecasts is the estimation of the levels of funds and current accounts savings available for investment purposes. The costs associated with the financing aspect of new investment are incorporated into the forecasts. Thus, it can be said that underlying the forecasts should be a consideration of the impact of new investments on the interest and amortization in and on expenditure budgets. A shortcoming of the BSD forecasting procedure is that the impact of new investments on the operating budget is not considered. In essence, it is 1/ It is worth noting that the forecasts indicate annual increases such that the Level of funds invested in 1985 is three and one half that in 1981. WRA/BR-001-3/3.15.85 - 134 - assumed that the operating and maintenance of the facilities to be financed will not affect the District's current operating expenditure during the forecast period or if they do, they can be accommodated out of the 30 percent of current revenues which are allocated to current operations in the forecasts. Given that the purpose of these forecasts are related to the determination of the availability of funds for investment and debt carrying capacity, they can be seriously faulted. This follows from the faict that there is no independent estimate major categories of expenditures. Rather, it is assumed that operating and transfer expenditures grow at the same rates as revenues. This may be consistent with budgeting procedures which require a balance between expenditures and revenues. However, it precludes virtually any analysis of any policy options related either to the composition of services provided or to rules and/or regulations governing employment and renumeration levels. Simi arly, the BSD forecasting procedures are capable of shedding absolutely no light of socio-demographic or economic changes, e.g., inflation, on its cost of operation. In large part the shortcomings of current revenue forecasting procedures can be attributed to the relatively short data series available on the revenue side. As time passes and additional observations become available, the strength of this limitation will decline. Thus, there will be the possibility of introducing more sophisticated economic techniques in the revenue forecasting. The procedures employed in the forecasts of current accounts expenditures preclude even the most rudimentary of analysis. Given the fairly extensive and detailed budgetary data which is available it would seem that WRA/BR-001-3/3.15.85 - 135 - the constraint so limiting on the revenue side of the ledger would be much less serious on the expenditure side. Indeed, even with a data series of no greater length than that used for the revenue forecasting, fairly detailed and extensive expenditure analysis would appear to be feasible. This is not to fault the District government for a lack of effort or sophistication. Indeed in revenue estimation for annual budgets, it has developed a variety of models and thus appears willing to experiment and upgrade its efforts. There is no apparent indication that BSD forecasting efforts have not been sufficient for the primary purpose to which they are directed, i.e., estimation of revenues for the purpose of establishing annual budgets. But the extention of these to multi-year revenue forecasting can be faulted on a variety of accounts. In particular, the time trend extrapolation of aggregate revenues precludes any determination of the relative impact of policy (e.g., tax and collection rates) and other factors, (e.g., inflation, population growth) or revenues. The same can be said about expenditure forecasts. Without some form of independent forecasting procedures, little can be accomplished by way of evaluating the determinants of future spending levels. Empresa de Energia Electrica de Bogota (EEEB) The EEEB has been involved in intermediate term forecasting since 1975. The forecasts are produced for the purpose of providing financial projections in large part to determine the financial viability of existing and potential additions to capacity. In addition to projecting the financial implications of new investments, the forecasting procedures have been developed and have been used ina way which allows systematic evaluation of some types of policy action, particularly those related to tariff rates. WRA/BR-001-3/3.15.85 - 136 - Although they are both tied to projected sales, the major elements of revenues and expenditures are projected in ways which facilitate analysis of a variety of possibilities concerning price levels, employment and renumeration policies as well as variation in the financial requirements of investment activities. However, except for the two noted exceptions, variations in tariff and investment terms, EEEB forecasts have generally been applied to the broad range of impact and sensitivity analysis which they could accommodate. The driving force of EEEB forecasting is a demand model which generates sales projection. 1/ As the structural and operational details of the model are available elsewhere only its salient aspects will be considered here. 2/ It was initiated in 1975 with monthly energy consumption for the period 1968-75 and is continually updated. The essence of the demand model is a set of time series extrapolations of energy consumption for five types of consumers (residential, commercial, industrial, block, and official and public lighting) subclassified by rate class, i.e., level of consumption. For each subclass within each consumer type a best fit logarithmic regression equation of energy consumption on time is selected for use in forecasting consumption over a five year period. For each sub-class of consumer, the best fit equation (highest R2) is selected from six equations which differ by the weights applied to the most recent observation. The six best fit equations are the basis for the time 1/ For a review of the broader issues in energy demand forecasting see IBRD, Energy, Water, and Telecommunications Department, Energy Demand Forecasting: A Critical Review of Current Approaches (February 1977). 2/ Modelo Corporativo, Calculos De Ingressos per venta de Energia. Empresa de Energia Electrica de Bogota, D.E. (undated). WRA/BR-001-3/3.15.85 - 137 - trend extrapolation (forecasts) of consumption by each sub-class of consumer which sum to total consumption. In actuality, two sets of forecasts are generated by these procedures, one based on monthly and the other on annual data. Those based on annual data compose the actual forecasts of consumption. The forecasted monthly consumption is employed primarily as a check on the forecasted annual values. The shortcoming of this approach lies in the inflexibility of time trend extrapolation. If the factors which have governed consumption in the past continue to exert their influence as they have in the past, then time trend extrapolations will have the desired degree of accuracy. However, consumers can be expected to alter their consumption in response to a variety of factors such as climatic conditions, the growth in their income, and changes in composition. More importantly, consumption levels will undoubtedly be affected by the costs of service which, for accuracy of time trend forecasts, implies one of two possibilities. Historic changes in tariff rate and consumer response to them must persist through the forecast period. Alternatively, if they do not, then the effects of other factors influencing consumption must counterbalance them. This, of course, is the shortcoming of all time trend extrapolation. They simply provide no means of evaluating the impacts of the basic factors, e.g., tariff rates, family income or consumption. EEEB Forecasts: Revenues. Projected sales are one basic component of EEEB's revenue forecasts. The product of tariff rates and consumption provide the basis for the forecasts of sales for the three major categories of operating income: sales revenues, connection charges and other incomes. WRA/BR-001-3/3.15.85 - 138 - Forecasts of sales revenues are calculated as the product of forecasts of consumption and tariff rates. The most recent forecasts (Table 3.2, February 1980), for example, incorporate an across-the-board 12 percent increase in tariffs in July 1980 and a 2.2 percent increase in average tariffs monthly through 1985. However, as it is structured, the forecasting procedure could be used for estimation of the revenues which would be produced by any of a variety of tariff rates. On this account, the forecasting procedures are well suited for gap and/or impact analysis. Just as EEEB procedures allow for the estimation of consumption by each of the classes within the major categories of consumer, so too do they allow for the calculations of the income derived from each class of consumer. However, as their forecasts are of the total quantity on a year-by-year basis, they employ an average tariff rate in their forecasts of sales revenues. This rate is calculated as a weighted average of the tariff rates assumed to be applicable to each class of consumer each month during the forecast period. The weights used in this calculation are the forecasted quantities of consumption for each class of consumer. Thus, based on the structure of tariffs assumed to be in effect, an average rate is estimated for each year. The product of this average rate and projected total consumption yields the estimate of total sales revenues. An alternative less cumbersome method for estimating annual gross sales revenues would be to simply sum the monthly revenues from class of consumers which could be directly forecast as the product of the assumed tariff rate and projected monthly consumption. In fact, this more direct procedure is equivalent to the weighted average method used by EEEB at least WRA/BR-001-3/3.15.85 - 139 - Table 3.2: EMPRESA DE ENERGIA DE BOGOTA (ELECTRIC COMPANY) REVENUES AND EXPENDITURES ACTUAL 1978 AND PROJECTED 1985 (millions of pesos) 1978 1985 Operating Revenues Sale of Energy 3,020 43,556 Other 388 6,098 Total Operating Revenues 3,408 49,654 Operating Expenses Generation 320 1,233 Purchase (I.S.A.), 728 5,~723 Transmission 100 209 Distribution 184 717 General and Administration 330 1,792 Public Lighting 128 1,164 Contributions (I.S.A.) 115 0 Depreciation 693 5,070 Total Operating Expenses 2,598 15,908 Net Income 810 33,746 Total Investment (1979) 1,146 8,952 SOURCE: EEEB, "Proyecciones Financieras," (February 1980). WRA/BR-001-3/3.15.85 - 140 - in so far as forecasts of annual sales revenues should be identical. 1/ However, the more direct estimation would have the advantage of directly indicating the revenue dependence of EEEB on class of consumer and the revenue impact which may arise from inaccuracies in the forecasted consumption of any class of consumer. In addition to sales revenues, EEEB identifies two other categories of operating income in their forecasts. They are connection charges and other incomes. There are projected to amount to a constant 15 percent of sales income with connection charges accounting for 85' percent and other incomes 15 percent of the combined total. Forecasting these values in this way implicitly assume that the levels of connection charges and other incomes are governed by the levels of sales and tariff rates. While these may be the primary determinants of some components of other incomes, this forecasting procedure would seem to have particular limitations in regards to connection charges. In particular, if there is a relation between the number of connections and average tariff rates, common sense would seem to dictate that it be negative. That is, if anything, the continual monthly increase in tariffs which EEEB has scheduled in their forecasts should, if anything, reduce the demand for new 1/ Let i stand for class of consumer m for month, E for consumption and t for the tariff rate. Total annual sales is the sum of month consumption by each class of consumer, i.e., S = E Cim. The weighter average tariff rate is T = E. T. C. /EC. . Revenues forecasted as the product of the two annual values which by substitution is TS = (Z. t. C. /ZC. ) ZC. imimm 1 i m im which upon cancellation yield Z. t. C. the sum of monthly sales revenues from all class of dS = .26145X, increase. S WRA/BR-001-3/3.15.85 - 141 - connections. In fact, the assumed 15 percent of sales revenues for these component represents a decrease from the 18.5 percent estimated in 1979. However, there is no indication that EEEB foresees this decline as a result of reduced demand in response to higher tariffs. Rather, it appears they perceive this decline simply because tariff rates are slated to increase while connection charges are not. Alternatively, the number of connections may vary with total consumption in the manner implied by the forecasting procedure if the number of new consumers increase at the same rates as they have in the past and if new consumers consume average amounts. But the direction of causality should be the reverse, i.e., new connections to consumption, of what EEEb's procedure implies. A more appropriate procedure would be 'to base the forecasts of connection revenues on independent forecasts of new connections. Finally, the other elements of revenues, e.g., interest and dividends, are based on current contractual arrangements and are generally known and need not be forecasted. In general, the EEEB revenue forecasting procedures are well designed to evaluate the impact of tariff policies on revenues and budget balance (i.e., fiscal gap analysis). However, the basic time trends extrapolation methodology precludes the broader analysis of the factors which determine energy consumption and thus sales revenues. EEEB Forecast: Expenditures: Just as was the case with the revenue forecasts, EEEB forecasts of operations expenditures are based on their projections of energy consumption. For the most part the major elements of costs of supplying energy demand are estimated in constant prices (1977) and then inflated according to a scale of inflation rates. WRA/BR-001-3/3.15.85 - 142 - EEEB forecasts separately the costs of hydro and thermal generation. The amount of power to be obtained from the two sources is based on the assumed availability of water from the Chingaza and Mesitas operations, the number of accounts, and existing contracts for energy purchases and historic levels of various measures of average costs. The costs of hydroelectric power are forecasted on the basis of past unit costs of capacity (see Table 3.3). The costs of thermal power consists of two parts, the cost of coal and other costs. Purchase costs of coal are projected on the basis of existing contracts and historic increase in coal prices. Other thermal costs are estimated in constant prices by an equation relating them to the total quantity supplied. Other major current cost items are forecasted on the basis of kilometers of transmission line (transmission costs), number of accounts (distribution and transmission costs) and as a proportion of operating expenditure (general and administrative expenditures). A shortcoming of EEEB forecasts, or at least their reporting of them, is their failure to identify procedures underlying their projections of these basic factors, number of accounts, km of line, which are the determinants of the related expenditures. In general, the coefficients or multipliers used in the estimation of these costs are based on historic relationships and are stated in constant prices. Thus as with other elements of expenditure, forecasts are in constant prices which are then adjusted for inflation. The other major elements of current costs, e.g., depreciation, are based on knowh accounting schedules and need not be forecasted in a formal sense, while the capital cost elements of the budgets are estimated on the basis of recent experience. As structured, the procedures underlying the cost forecasting produce what in essence is a major determinants type of model. Thus, they lend WRA/BR-001-3/3.15.85 - 143 - Table 3.3: UNIT COST ESTIMATING VALUES UNDERLYING EEEB EXPENDITURE FORECASTS, 1980 Bases Value Hydro Generation Average of real cost C$212.8 per MW per M.U., 1970-77 1977 constant dollars Thermal Generation Coal existing contracts Other costs cost = 27.141 + .035 GWH 1977 constant dollars Transmission Average of real cost C$53.91 Per km of line, 1972-77 1977 constant dollars Distribution Average of real cost C$203.8 per subscriber 1970-77 Billing and Average of real cost C$122.9 Collection per s-scriber 1973-77 General and Constant proportion of 24.14% Administration generation, transmission and billing and collection cost SOURCE: EEEB "Proyecciones Financieras", February 1980. WRA/BR-001-3/3.15.85 - 144 - themselves quite readily to both fiscal gaps and impact analysis. Noteworthy is the fact that forecasts are stated in constant prices, providing a means of estimating the impact of inflation on costs. There are shortcomings with this approach. The first is that the average unit values used to translate the various quantity measures into costs, e.g., real cost per unit of electricity, per km of line, per sub- scriber, are constant. Whether in fact they are is an empirical question. However, if they were then the unit values of any one year would be su:*fi- cient. But in fact the unit values of the factors which are used are averages based on various periods depending on the category of costs. This alone would seem to imply that there has been some measure of variation in these cost factors over time. It would seem that the interest of forecast accuracy would be served by developing procedures which account for changes in real unit costs and apply them in conjunction with projected quantities, e.g., consumption, km of lines, etc., to obtain the expenditure forecasts. The forecasts of general and administration expenditures is a case in point. Assuming that they remain in fixed proportion to other operating costs (or equivalently that they row at the same rate) implies that they are set by two types of determinants, the levels of service activities (e.g., total consumption, km of lines) and the unit costs of them. In regards to the first, it may well be the case that the amounts of general and administrative activities required is related to the levels of operations, but that too is a question which could be answered by a look at the facts, in this case the growth in general and administration expenditures relative to other operating expenditures. But even if there is stability in the amounts of general and administrative services required, there is no a priori reason to suppose that WRA/BR-001-3/3.15.85 - 145 - their unit costs, e.g., wage rates, material costs, etc., should be related to generation distribution and transmission costs in any particular fashion let alone in the way of a fixed relation which would seem to be implied by the current procedures. But this too is reflective of the general procedure of assuming a fix unit costs of the various operations. A preferable approach would be to develop more elaborate means of considering unit costs in terms of employee and other input requirements and their unit costs. The shortcominTg of the procedures is that employment and personnel costs are not considered separately. Thus, it is not possible to directly evaluate the effects of changes in employment and employee compensation policies or changes in the prices of other inputs on expenditures, i.e., impact analysis, or on the budget balance, i.e., fiscal gap analysis. In this regard it is interesting to note that EEEB has in another context projected their employment level to increase at an average annual rate of about 6 percent per year. 1/ However, it does not appear that the expenditure implications of these increases have been directly incorporated in their forecasts. Rather, it appears that it is assumed that the costs of such employment increases will be accommodated within the expenditure increases forecasted for the various categories of operating expenses. In a similar vein, EEEB includes in their forecast report a steady growth in the number of subscribers but they provide no indication of the considerations which underly these or the projection of transmission lines. While it can be argued that these are discretionary in the sense that they are 1/ Cite IBRD appraisal project. WRA/BR-001-3/3.15.85 - 146 - determined and controlled by planned expansions in the system, their relation to the projections of consumption and sales is nowhere spelled out. From all appearances, projected growth in sales, distribution lines and number of accounts and employment as well are determined independently rather than simultaneously as should be the case for consistency in expenditure (as well as revenue) forecasting. EEEB Summary. The basic driving force underlying the EEEB revenues and expenditure forecasting are a detailed series of time trend extrapolations of energy consumption. 1/ As these forecasts are in terms of the quantity rather than the value of energy consumed, they lend themselves quite readily to calculating the impact on revenues of changes in the prices charged consumers assuming that consumption itself is not responsive to changes in tariff rates. The logic of time trend extrapolation is that the rate structure and any alterations in it which have occurred during the basic observation period will be repeated during the forecast period. If this is not the case then it would be expected that consumption would be different as consumers adjust to different tariff rates. To ignore the possibility of a consumption response to tariffs is tantamount to assuming that consumption is determined by and large by factors other than tariffs. Indeed it appears that EEEB energy consumption forecasts are based on the premise that other factors such as basic household and production requirements rather than cost of electricity determine usage. 1/ For a detailed review of the general practices employed in energy demand forecasting see, IBRD, Energy Demand Forecasting. WRA/BR-001-3/3.15.85 - 147 - On the expenditure side, the forecasting methodology is of determinants type designed to forecast the effects of changes in energy supply. In this regard the forecasting methodology lends itself to both a fiscal gap and an impact analysis. Regretfully, the procedures are wanting on at least two counts. The first is that in many instances they are based on categories of expenditures, (e.g., administration), rather than objects (number of employees) expenditures. This precludes the analysis of such things as changes in personnel policies (i.e., employment levels and renumeration rates) and in the quantity and values of many types of materials and supplies. Second, operating costs associated with the maintenance and operation of new facilities do not enter the forecasts in any explicit way. Rather it is assumed that in per unit constant dollar terms, these costs will remain constant at levels calculated at historical average. Whether constant dollar operating and maintenance costs are related to energy generated by new facilities just as they have been in existing facilities is a matter of conjecture, and one which greatly limits their usefulness, for constant returns to scale is a characteristic not often found for electricity generation. However, it is an assumption embedded in EEEB current forecasting practices. Finally, it should be noted that while the EEEB data base and basic procedures lend themselves to both fiscal gap and impact analysis, their use has been limited. Primarily they have been employd to determine the budgetary implications of capacity expansions via the effect on energy supply, i.e., fiscal gap analysis. But there does exist the potential for the explicit analysis of some other possibilities, e.g., alterntive tariff structures, different patterns of inflation, etc. WRA/BR-001-3/3.15.85 - 148 - Empresa de Acueducto y Alcantarillado de Bogota (EAAB) The most recent set of EAAB budget forecasts were produced in 1980 (Table 4). The general approach is similar to that of EEEB (and to that of ETB as will be noted below). That is, the principal purposes of EAAB forecasting are related to its investment activities. However, they have had application in what is essentially a limited type of impact analysis. On the revenue side, the implications of alternative tariff rates have been considered. On the expenditure side, the cost increases due to irTflation have been calculated. EAAB's supply forecasts are grounded in a data set originating in 1971. However, other than establishing existing levels of consumption and supply, these data are used for the purpose of calculation of various average unit cost measures. The driving force behind EAAB's forecasts are projections of number of accounts, sales, and production. Each of these is projected to increase at a constant annual rate during the forecast period. Considering that the number of new accounts and the volume of production can in large measure be determined as a matter of policy, increases at constant annual rates could be the outcome of company policy. However, the growth in water sales is another matter, for, in addition to the number of consumers it depends on the amounts they consume. But this too is assumed to increase at a constant annual rate. The forecasts produced in 1980 indicate that they are based on the assumption that each year following 1973 up to 1985 population increases at a 4.5 percent rate. Certainly it is unlikely that population would increase at a constant rate over a 12 year period but such is the outcome of time trend extrapolations which so frequently underly population projections. It appears that the projected equal annual average rate of WRA/BR-001-3/3.15.85 - 149 - population growth is what underlies the constant annual average increases in sales and production and indeed it is so stated in the case of the number of new connections. This could be problematic for whenever some elements of a forecast are tied directly to another projection then the accuracy of the forecasts depend on the validity of the projected factors. Errors in the latter produce errors in the former, which will likely be compounded over time. Whether the population growth in Bogota will increase by an amount equivalent to an annual average rate of growth of 4.5 percent cannot at this point be known. Recent experience would seem to indicate that actual population increases between 1971 and 1980 fell below the 1.341 million which is built into EAAB's forecasts. If the population projection through the forecasts period exceed actual population growth then EAAB's revenue and expenditure forecasts will also be high. In fact, EAAB forecasts that the number of new accounts will increase at a greater rate (5.75 percent per year) than population. The implication is of course that the percent of household and business connected to the water system will increase year by year. Given that EAAB does not reach the whole of the population--in 1980 their objective was to reach 80 percent of the urban population--, it is conceivable that the number of accounts could increase even if the population growth falls short of that predicted. 1/ The principal result would be a more rapid achievement of service level objectives and somewhat less need for capacity expansion. 1/ The World Bank, Colombia, Third Water Supply Sewerage and Drainage Project, Staff Appraisal Report, (May 2, 1979). WRA/BR-001-3/3.15.85 - 150 - Indeed, EAAB appears to believe that the water needs will increase less than proportionate to population. On the basis of past experience, EAAB project water sales to increase by 3.8 percent per year. The fact that the projected increases in sales and number of accounts remain constant throughout the forecast period implies that on the average purchases by new accounts will be the same. That is to say that water sales are in fact forecast in terms of a fixed multiple of the number of new accounts. As the number of new accounts is forecast as a fixed multiple of population, EAAB sales are in actuality forecast as a fixed multiple of forecasted population increase. 1/ This still implies that over the years average service usage will decline. But since connections are projected to increase at a lower rate than population, average consumption levels are projected to decline. This would follow if the extension of services were primarily directed to lower income smaller consumerss and/or if it is anticipated that tariff increases would depress consumption. As the water sales projection are derived from trend extrapolation based in a period when tariff rates did increase, some of the projected decline in average consumption may be reflective of higher prices. 2/ However, there is no clear indication that consumer response to 1/ That is, if the annual percentage rates of consumption, AC/C, new accounts AN/N and population are consistent the ratio of two of them is also a constant. Thus, if AC/C - AN/N = k and AN/N + AP/C = n, and AP/P = R, the ratio of the constant annual percentage rates of growth in consumption and accounts = R/n = a and of accounts to population = N/C = b(N/P) AP and as a = k/n = AC/C i AN/N, AC = a(C/N)AN and substituting for AN yields AC a(C/N) b(N/P)AP = a(C/P)AP. Since the base period values of C, N and P are all known, AC as well as AN is forecast as a fixed - multiple of AP, population increase. 2/ Tariffs were substantially increased between 1976 and 1978 with monthly increases of 1 percent, initiated in May of 1978. WRA/BR-001-3/3.15.85 - 151 - tariff increases explicitly affect the projections and this failure may affect forecasts accuracy if the pattern of tariff rates change is as planned, different from those which have historically occurred. Rather, it appears to be the case that EAAB sales forecasts are based on project plans to expand service primarily to lower income, smaller consumers. 1/ EAAB Forecasts: Revenues. The projections of water consumption revenues from water sales (Table 3.4) are forecasted by applying the existing tariff schedules to projected sales. 2/ As the current schedules include a bi- monthly rate increase of 2 percent, the forecasts indicate a revenue increase at a 12.62 percent compound annual rate. As has been indicated, these tariff increments differ from those of the recent past and thus it would be expected that they may alter the trend of average water consumption. However, as has also been indicated, water consumption is forecast as a fixed multiple of population, and thus assumed to be unaffected by tariff rates. The second major component of the forecasts is the revenues from connecting fees. These are based on projections of the number of accou-rts by class of consumer and the legislated fees. The projected number of new accounts in each class is determined by allocating the projected number of new accounts to the various classes of consumers on the basis of the percent 1/ See World Bank, Colombia, Third Water Supply Project, Staff Appraisal Report, (May 2, 1979). 2/ Empresa de Acueducto y Alcantarillado de Bogota, D.C., Commentarios a Los Proyecciones Financieras, Junio 25-1980. WRA/BR-001-3/3.15.85 - 152 - distribution of new accounts in the recent past. 1/ The essence of this procedure is to project on the basis of historical experience the distribution of new consumers in the various consumer classes and then to apply this to independently determined population based projections of total new accounts. In essence, EAAB constrains the aggregate forecasts for the separate consuming classes to equal the growth projected on the basis of a forecast of total population and an assumed fixed rate between new accounts and new population. In this regard, EAAB differs from EEEB whose forecasts of total sales or usage is calculated as the sum of separate forecasts of usage by consumers in each consumer. Which of these approaches produces the more accurate forecasts will be known only at the end of the forecast period. However, as the numbers of actual and potential consumers, as well as the tariffs they are charged, differ by class of consumer, it might be expected that the number of new accounts in the various class might add up to a number quite dif.'erent for the independently projected number of new accounts. In any case, given the estimates of new accounts in each class of consumer, the revenues are calculated in terms of the new tariff rates. For supply fees, this amounts tu 4 percent of catastral value while license fees are given as C$37.14 and C$68.77 pesos depending on consumer class. 2/ To 1/ There are four classes of consumers, residential, indusstrial and commercial, preferential and government. Tariff rates for each class of consumer depends on catastral value including a monthly fee and an increasing marginal rate structure based on water consumption. 2/ Although these fees are set in relation to catastral values, the initial levy is generally set in terms of the cost of construction. The intent is to reclassify property on the basis of catastral values and/or as these values are changed. However, it appears that although catastral values may be altered, property are seldom reclassified. WRA/BR-001-3/3.15.85 - 153 - Table 3.4: EMPRESA DE ACUEDUCTO Y ALCANTARILLADO DE BOGOTA (WATER AND SEWERAGE) REVENUES AND EXPENDITURES ACTUAL 1978 AND PROJECTED 1985 (millions of pesos) 1978 1985 Revenues Sale of Water 1,263 3,460 Licenses and Supplies 350 1,275 Other Income 107 .401 Total Revenues 1,719 5,134 Expenses "Energia de Bombeo" 59 - 201 Chemical Products 45 620 Personnel Operating 103 698 General Operating 11 138 Maintenance 62 310 Total Operating Expenses 279 1,967 Personnel-Administrative 94 725 General Administration 74 814 Administrative Maintenance 35 398 Total Administrative Expenses 203 1,938 Paid Social Benefits 221 1,716 Unpaid Social Benefits 91 1,003 Total Expenses 794 6,623 Net Income Before Depreciation 925 (1,487) Depreciation 95 5,017 Net Income Before Interest 829 (6,504) Total Investment 1,188 14,629 Source: EAAB WRA/BR-001-3/3.15.85 - 154 - obtain the forecast of aggregate revenue, an estimate of other revenues, projected to increase at 21 percent per year, is added to the forecasted water sales and connection fees revenue. In the forecast writeup the factors generating this rate of growth are not identified. In general, the EAAB procedures of forecasting revenues are somewhat limited. The fact is that the projected population growth is the primary force leading to increase consumption and number of accounts. That these principal elements of revenus are considered fixed in relation to population over the forecast period is a particular shortcoming for any inaccuracy in the population forecasts is directly translated into the revenue forecasts. There is a possible inconsistency in the forecasting procedure. That is, sales revenue are forecasts of projected aggregate sales and an average tariff rate. Alternatively, to estimate connection fees, EAAB provides an estimate of the number of new consumers by class of consumers. This would seem to be the first step toward a projection of usage by consumer class. Given that different tariffs are applicable to different classes of consumers, usage and revenue projections by consumer class would seem more desirable than simple projection of aggregate values particularly given the apparent policy of directing system extention to lower income households. EAAB Forecasts: Expenditures: In general, any category of expenditures can be considered in terms of the cost per uni! of the good, service or activity and the quality produced or purchased. This is the approach used by EAAB to forecast but two elements of their expenditures: pumping cost and the costs of chemical products. Unit costs of these two expenditure elements are based on historical costs projected through the forecast period. The volume of production is projected to increase at a constant 3.3 annual percentage WRA/BR-001-3/3.15.85 - 155 - rate. 1/ Thus, the procedure underlying these costs are similar to those employed in the forecast of revenues, i.e., unit volumes are projected to increase at constant annual rates. The projections of other expenditure categories are accomplished in similar fashion, i.e., increase at constant annual percentage rates. 2/ However, in these cases the forecasts are not built up from separate estimates of unit costs and level of operation or quantities purchased. Rather it is the gross expenditures of the remaining categories of operating expenditures are forecasts (Table 3.5). Although projected annual increments in the range of 30 to 46 percent may seem substantial, they must be considered in light of EAAB's anticipated rate of inflation, i.e., 35 percent. That is, with the exception of nonpersonal costs of general administration, where a 1 percent annual increment is allowed for "normal growth of the enterprise", all categories of expenditures are projected at rates equal to or less than the rate of inflation. That is to say that EAAB forecasts of operating cost are based on an empirical assumption of essentially no increase in real (constant dollar) costs between 1980 and 1985. Indeed a decline in real personal and social obligation costs is forecasted. In conjunction with the projected increases in production, number of accounts and sales, constant real costs of operation implies that per unit of production and/or sales, the levels of operations 1/ The production and sales volume differ by water loss in transmission as well as by theft. Historically these have been quite high. They are estimated to decline slightly from roughly 25 to 23 percent of volume. 2/ Social obligations not paid are assumed to be estimated as 40 percent of the social obligation reserve. It should be noted that the pension fund is estimated to increase by 10 to 20 percent as determined by contracted salary increments. WRA/BR-001-3/3.15.85 - 156 - Table 3.5: PROJECTED ANNUAL AVERAGE INCREAS-S OPERATION, MAINTENANCE AND ADMINISTRATION 1980 - 1985, EAAB Projected Annual Percentage Increase Expenditures 1980-85 Personnel Operations 30 Administration 30 Other (General) Costs Operations and Maintenance 35 General Administration 36 Social Obligation to be Paid 35 Unpaid Social Obligations (alternate years) 48 Source: EAAB WRA/BR-001-3/3.15.85 - 157 - inputs, i.e., labor supplies and materials must be declining. While water supply is often characterized as a decreasing (real) cost industry, that pertains to the whole of cost, fixed as well as operations, and usually it is the high fixed costs which lead to decreasing costs. The functions of general administration, operations and maintenance are generally not considered to be of the declining cost variety. But given that in the aggregate personnel expenditures are forecasts to grow at a rate slower than anticipated inflation, it would appear that EAAB anticipates a reduction in employment levels and/or a reduction in the real cost (wages) of employees. Whether such declining real cost can realistically be expected depends on the outcome of wage policies and labor negotiations. That it is an issue here results from the fact that EAAB cost estimates are not built upon projected employee and other input requirements. EAAB's financial forecasts are completed by adding in contractual and other calculated amounts, (e.g., depreciation, interest) to obtain the total expenditure. EAAB Forecasts: Summary. Although EAAB forecasts are initiated by projected changes in the volume of water production and sales, they are not tailored in a fashion which would facilitate anything like a broad based analysis of the variety of factors which likely affect the pattern of growth in their revenues and expenditures. There appear to be elements of inconsistencies of growth in both their revenue and expenditure forecasting procedures. This is regretful, and perhaps avoidable, for there does exist a rather lengthy and detailed data base which could facilitate the development of even some of the more WRA/BR-001-3/3.15.85 - 158 - sophisticated forecasting procedures. 1/ More importantly is that the cost projection imply no (or little) increase in real cost of labor; any increase in employment levels, real wage rates or other elements of labor cost will increase EAAB's need for operating funds. In general, the expenditure estimating procedures employed by EAAB should come as no surprise as their forecasting procedure has not been designed for these specific purposes. Rather, like EEEB and th BSD, expenditure forecasting has been allied with analysis of capital investment projects. But even in this regard, the failure to explicitly account on the impact of capital projects on current operating costs is a shortcoming. Empresa de Telefonos de Bogota (ETB) As with the other entities considered here, ETB initiates its forecasts with projection of demand. In large part the projections are based on the historical experience of the company in that accumulated requests for installations provide a guide to changes in demand. In addition, historic experience in regard to the number of calls and average duration provides the basis for the projections of line usage. However, within ETB it is generally believed that request for new installations reflects less than half of the actual demand. As the number of requests for installation during the 1970s exceeded the number of line installed, there is generally a perception of excess demand. 1/ For an example of a recent use of the EAAB data base for a rather sophisticated analysis see "Colombia, Third Bogota Water Supply, Sewerage and Drainage Project," Staff Appraisal Report, Latin America Projects Department (Washington, D.C.: The World Bank, 1979) especially Section VII. WRA/BR-001-3/3.15.85 - 159 - Given the perception of excess demand, ETB forecasts are dependent on its plans for system expansion to a much greater extent than the other entities considered here. The principal factor governing the most recent forecasts is the objective of providing the additional facilities and equipment required to accomodate its pending excess demand as well as the anticipated new demand. As this objective is stated in terms of telephone density, per capita telephone lines, the level of system operations and revenue forecasts are guided by anticipated population growth. ETB Forecasts: Revenues. Estimates of current excess demand and the increase in demand attendant to projected population growth combined with the plan to service 98 percent of demand by 1986 generate both the number of new installation and the total number of lines in service. System usage is then estimated based on the number of calls per line, their average duration which are derived from company data on telephone usage in the recent past. Revenues (Table 3.6) are then estimated by the application of the schedule of charges which currently call for an average 2.5 percent increase per month through December 1981. ETB forecasting procedures as they are currently structured and applied provide a means of evaluating the revenue impact of population growth on service levels and of changes in usage charges. In this regard they represent a form of deterministic forecasting. 1/ ETB Forecasting: Expenditures. Just as the planned level of service to be provided is the driving force behind ETB revenue forecasting, so too it is the major factor governing its expenditure forecasts (Table 3.6). In general 1/ ETB, Investment Plan, 1979-86. WRA/BR-001-3/3.15.85 - 160 - Table 4.6: EMPRESA DE TELEFONES DE BOGOTA (TELEPHONES)- REVENUES AND EXPENDITURES ACTUAL 1979 AND PROJECTED 1985 (millions of pesos) Actual Projected 1979 1985 Total Income 1771.2 6021.4 Expenses Operating 311.5 1392.4 Administration 347.5 1622.4 Social Payments 123.7 846.4 Services 26.2 178.2 Various 13.2 23.2 Sub-Total 822.1 3999.7 District Participation 38.1 177.3 Other Non-Operating 28.6 131.2 Financing 118.1 480.2 Depreciation 150.0 461.4 Total expenses 1156.9 5249.8 Net Income 614.3 771.6 Total Investment* 1820.0 * From the ETB's (unpublished) Investment Plan 1979-1986. Source: Unpublished Worksheets of the E.T.B. dated 6-19-79 entitled "Proyeccion de Ingresos Totales" and "Proyeccion de Gastos Por Rubro." WRA/BR-001-3/3.15.85 - 161 - operating costs are estimated in terms of the unit costs which have characterized ETB operations in the recent past. These, however, are in some cases, modified or adjusted to reflect anticipated (or estimated) unit cost differences which may be anticipated as a result of the installation of new equipment. The financial costs of investment activities is a known factor, at least in regards to existing commitments, and thus does not require forecasting. Cost of new undertakings is another matter. In this case, ETB forecasts are based on planned investments. In part these are to be financed out of anticipated operating surplus although the majority of these expenditures are expected to be financed through credit. These new and anticipated financial charges are incorporated in the budget forecasts. ETB Forecasts: Summary. The basic methods underlying ETB forecasts are of the variety often employed in impact (or service standards) analysis. They are generally designed to determine the budgetary impact of investments required to meet planned objectives. The levels of activities to be required by the realization of these objectives is determined by anticipated growth in the area's population and by existing levels of excess demands. The budgetary implications of planned levels of operation is in large part determined by the company's recent experience in terms of system usage and unit costs. Summary Assessment of Forecasting Practices in Bogota The forecasting methods and procedures employed by the four public entities considered herein represents a variety of techniques and procedures implemented with varying degrees of sophistication. The purposes for which they have been undertaken represents a blend of riscal gap and impact WRA/BR-001-3/3.15.85 - 162 - analysis. Most generally, the efforts are directed towards the determination of the extent to which fiscal gaps may be anticipated, i.e., a forecasted surplus on current accounts. This information provides one of the bases for investment planning, for it in part, determines debt carrying capacity. Thus it can be said that one object of these entities' forecasting is to determine the impact of anticipated current accounts operations in the capital budget. Moreover it is also the case that the financial costs of investments are incorporated in the forecasts. This provides a means of determining whether gaps between overall revenues and expenditures are likely to materialize. Although it is not the case with the BSD, forecasting procedures are designed so as to accomodate some measure of impact analysis, particularly in regards to revenues as they may be affected by population changes, tariff adjustments, and service standards. However, it appears that comparative analysis, i.e., a consideration of the fiscal implication of differing assumptions concerning the determinants of revenues and expenditures, has not been an objective of forecasting. This is as true of expenditure forecasting as it is of revenues. In fact, in many instances the procedures underlying the expenditure forecasts may obscure many of the elements which could be useful for policy analysis. This is the result of building forecasts upon aggregate categories of types of expenditures, e.g., administration and current operation rather than objects of expenditure, e.g., employee and materials. Even in those cases where important material costs are identified explicitly in the forecast, e.g., cost of water in the case of EAAB, other costs are not identified by object. However, it must be recognized that the overriding purpose of forecasting has not been to facilitate a comparison of budgetary impacts. WRA/BR-001-3/3.15.85 - 163 - Rather, forecasting has primarily been undertaken for purposes of determining the impact on the budget of a narrowly defined set of circumstances, e.g., service standards level of population, and tariff structures and to some extente, inflation rates, which may characterize the future. Given the above, it should not be surprising that to a large extnet, forecasting has been of the major determinants form. That is, based upon some identified relation between costs or revenues and some variable, the forecasts are structured to answer questions such as what will be the impact on revenues or expenditures of some specified change in the variable? In some cases, the forecasting techniques are designed in a manner which would accomodate considerations for a variety of factors. Most common are population, price levels, and service levels or standards. Other important determinants such as wage rates, employment and level of income have not received similar treatment, since the objectives of forecasting have not generally included impact analysis. To a great extent, particularly in the cases of BSD, trend extra- population has been a major element in forecasting. Use of this procedure may be problematic for it will yield accurate forecasts only if the future replicates the past. Some of the severity of this assumption can be alleviated if the extrapolations are based on the historic behavior of subclasses rather than aggregates. But still, there remains the presumption that the factors governing historic outcome will extend with similar influence throughout the forecast period. As was indicated previously, the determination of behavioral relations can best be determined by statistical modeling. While statistical WRA/BR-001-3/3.15.85 - 164 - models have been applied as an element of some forecasting, i.e., EEEB, its uses even then are generally to facilitate trend extrapolation. The intent here is not criticized for the lack of behavioral modeling, for there is no guarantee that it can produce forecasts which are any more accurate than those generated by simpler means. In fact, much of the data required for behavioral modeling is not readily available. Thus it is likely that its use will be limited to those. However, as both trend extrapolations and the determinant based forecasts presume a stable relation between a number of variables, behavorial modeling could provide a useful check on the validity of these assumptions. Thus it should be applied whenever possible. It is in this regard, as an adjunct to other forecasting methods, that statistical modeling may be most useful to forecasters. - 165 - REFERENCES FOR CHAPTER 3 Benjamin Ardila Duarte (Tesorero) and Angela Robayo Bello (Secretaria General), Informe de Tesoreria, Primer Semestre de 1979, Bogota. Lewis, Jeffrey D., "Consolidated Public Finance in Bogota and Cali, Colombia," Washington, D.C. The World Bank (mimeo.). 1979. Linn, Johannes F., "Public Utilities in Metropolitan Bogota: Organization, Service Levels, and Financing." Washington, D.C., The World Bank (mimeo.). 1976. Linn, Johannes F., "Summary of Preliminary Findings and Outline for a Monograph Entitled Metropolitan Government in Bogota: Options for Efficient and Equitable Financing of Urban Growth," Washington, D.C., The World Bank (mimeo.). 1980. Ministerio de Hacienda y Credito Publico, Direccion General de Credito Publico, Division Legal, "Minuta Modelo para los contratos de emprestito externo garantizado por la Nacion," Bogota (mimeo.). 1970. Ministerio de Hacienda y Credito Publico, Direccion General de Credit Publico, Division de Credito Interno, "Normas y Procedimientos para la Contratacion de Credito Publico Interno," mimeo. Bogota. 1979. Torres, Jorge Ortega (comp.). Codigo de regimen politico y municial (Ley 4a de 1913 y normal legales complementarias), Quinta edicion actualizada, Bogota. 1976. Paredes, Luis Ricardo, "Colombia's Urban Legal Framework," Bogota (mimeo.). 1979. WRA/BR-001-2b/03.15.85 - 166 - Pineda, Jose Fernando, "Sistemas de Ejecucion y Estrategias Regional de Desarrollo en Colombia," in: Revista Camara de Comercio de Bogota, Vol. 9, nr. 36, Sept. 1979, p. 153-168. Bogota Distrito Especial, Alcaldía Mayor, Secretaria de Hacienda, Direccion Distrital de Presupuesto, Division Financiera y Tributaria, Proyectos de Presupuesto de las Entidades Descentralizadas y los Fondos Rotatorios de Bogota, D. E., Vigencia Fiscal de 1980, Bogota 1979. Bogota, Distrito Especial, Secretaria de Hacienda, Díreccion de Presupuesto, Proyecto de Presupuesto de Bogota, Bogota 1979. Presidencia de la Republica, Secretaria de Administracion Publica, Manual de organízacion de la rama ejecutiva del poder publíc, Bogota. 1978. Reveiz, Edgar, Jaime Millan, Gabriel Murillo, Charles Richter, Elizabeth Ungar, Poder e Informacion. El proceso decisorio en tres casos de politica regional y urbanaen Colombia, Bogota: CEDE, Universidad de los Andes. 1977. Rivera-Ortiz, Angel Israel. "The Politics of Development Planning in Colombia," Ph.D. Díssertation, State University of New York, Buffalo. 1976. Thompson, James D., Organizations in Action: Social Science bases of Administrative Theory, New York. 1967. Wildavsky, Aaron, The Politics of the Budgetary Process, 2nd edítion, Boston 1974. Wolff, J.H., Planning in Entwicklungslandern. Eine Bílanz aus Politik- und Verwaltungswissenschaftlícher Sicht, Berlín (Duncker und Humbolt). 1977(a). WRA/BR-001-2b/03.15.85 - 167 - Wolff, J.H., Reflexiones sobre el control fiscal en Colombia, in: Universitas (Bogota), No. 53, Dec. 1977, pp. 33-51. 1977(b). Wolff, J.H., La politica del proceso presupuestasl en Colombia. Consider- aciones preliminares, in: Universitas (Bogota), Nr. 54, 19-34. 1978. Wolff, J.H., Physical Planning in Bogota, Colombia, in Jahrbuch fur Regionalwissenschaft, Vol. 1, pp. 124-141. 1980. Wolff, J.H., Bureaucractic Politics and Administration: The Colombian Case (forthcoming). WRA/BR-001-2b/03.15.85 WPS/8R-001-4a/3.20.85 - 168 - Chapter 4 THE REVENUE SHARING SYSTEM IN COLOMBIA: A CASE STUDY OF BOGOTA by Caroline S. Fawcett I. INTRODUCING REVENUE SHARING Revenue sharing is the transfer of tax revenue from one level of government to another. 1/ The system of revenue sharing as it affects the District of Bogota cannot be so simply defined; it encompasses several budget processes, various disbursement procedures, and a wide range of policy and political initiatives at both the District and national levels of government. The purpose of this paper is to describe and analyze the revenue sharing system of the District. Given the complexity of the system, such description and analysis is no modest task. The limitations of the paper are two. Firstly, the national revenue sharing system is described only as it relates to the District system. Secondly, there is no comprehensive analysis of the District's tax contribution to the revenue base which feeds revenue sharing programs. The focus of this paper is to examine the many factors which enter into the transfer of revenues to the District, their importance to District revenue, and their impact on District expenditures. Section I of this paper briefly reviews the overall trends and patterns of transfer flows as received by the District over the last two 1/ Revenue sharing in the context of the United States refers to the distribution of national direct tax revenues to local governments. For the purposes of this paper, a more general definition is adopted which includes revenues from direct, indirect, and shared tax sources. WPS/BR-001-4a/3.20.85 - 169 - decades. Section II describes the various revenue sharing arrangements, be they shared taxes or grants, as they affect the District. The type of transfer, the collection, disbursement and earmarking of the revenue, and the revenue performance are considered for each transfer instrument. The District transfer receipts are then compared to the nation-wide allocation of transfers, and also the inter-departmental distribution of funds. In the case where specific policy objectives underlie the revenue sharing process, as in the situado fiscal and the shared sales tax program, the objectives of the programs will be discussed in light of the authorization and distribution of transfers. Section III concentrates on the impact of the main revenue sharing programs at the District level. The flow of transfer revenues to the District Administration, the Local Health Agency (S.S.B. - Servicio Seccional de Bogota), and the Regional Education Fund (F.E.R. - Fondo Educativo Regional) is discussed. 1/ The workings of the two agencies which receive a large portion of District transfers are scrutinized, allowing us to understand the importance of the transfer revenue in the respective budget processes, the disbursement procedures, and the administration of these two agencies. In addition, the impact of the various revenue sharing arrangements on expenditure level and composition is studied. Section IV summarizes the main conclusions of this chapter. 1/ The local government of the District consists of the "District Administration" headed by the Mayor and a number of autonomous local service agencies. The S.S.B. and the F.E.R. are agencies jointly - administered by the local and national government. Their administration and financing are closely examined in Part Three below. WPS/BR-001-4a/3.20.85 - 170 - The first half of the paper largely approaches revenue sharing from the "top down", from the national to the District level. The second half of the paper views revenue sharing as a circular process, from the local level to the national level, and then back down to the local level. These two approaches emphasize the intragovernmental dynamics of revenue sharing arrangements in the District. Revenue sharing is certainly not a self- contained process, rather it touches all levels of government policy, administration and management. When examining the role of revenue sharing in the District, three budgets and closed accounts must be distinguished: the District consolidated budget and account; the Central Administration budget and account; and the budgets and accounts of F.E.R. and S.S.B. The consolidated budget includes the revenues and expenditures of the Central Administration and the local decentralized agencies. 1/ The budgets of the F.E.R. and S.S.B. are subject to national approval and are not entered into the District consolidated budget. District Consolidated Budget As can be seen in Figure 4.1 and Table 4.1, the shared taxes and grants have played a minor role in the consolidated budget (on average 8.8% and 2.9% respectively for the years 1961 to 1978). The self-financed revenue from the decentralized public utilities agencies have contributed around 49.6% on average; and their contribution has been increasing substantially from 1975-1979. Borrowing represented 21.6% on average during the same time 1/ For greater explanation of the District Budget and accounting framework, see Chapter 2 above. - 171 - Figure 4.1 Composition of Real Per Capita Total Revenues 80- -80 60 - -60 0 40- -40 I 20-. -20 0 0 '61 63 '65 '67 '69 '71 '73 '75 '77 '79 Year Legend: - Seff-financed Revenues - Shared Taxes -Borrowing 0 - Grants [ . 0 Tax & Renewal Non-tax Revenues World Bank-26169 WPS/BR-001-4a/3.20.85 - 172 - Table 4.1 Annual Average Compound Growth Rates and Percentage Distribution of District Consolidated Revenue (1961-1978) Average Percentage In Current In Constant Real Distribution of Prices 1976 Prices Per Capita Total Revenues Tax and Non-Tax 20.1* 4.3* -1.1 17.1% Self-financing 25.6* 9.0* 3.3* 49.6% Shared Tax 24.2* 7.84* 2.2* 8.8% Grants 25.0* 9.03* 3.5 2.9% Borrowing 21.6* 5.6* .09 21.6% * Significantly different from zero at 95% confidence level. Source: Fawcett and Lewis (1981). WPS/BR-001-4a/3.20.85 - 173 - period, yet its contribution has been declining in the last several years. General local tax and non-tax revenues has also declined in the last years. Its contribution for 1961-1978 averaged at 17.1%. Of all five revenues sources, shared taxes and grants have made the smallest constribution to total District consolidated revenue. The performance of shared taxes and grants must not be judged solely on its average contribution to consolidated revenues. In fact, the growth rates of the shared revenue, as well as the elasticity of the revenue to District income and population, presents a much different perspective for shared taxes and grants. The average annual compound growth rates at constant 1976 prices for the years 1961-1978 are 2.2% per capita for shared taxes and 3.5% for grants. 1/ This contrasts with the negative real growth rate per capita of general local taxes at -1.1% and the low growth rate of borrowing at +.09%. Self-financing, the healthiest of all District revenue sources, has an annual average compound growth rate of 3.3% (Table 4.1). 2/ The population and income elasticities of transfer revenue measure the responsiveness of transfer revenue to increases in District income and population, and are calculated for the years 1961-1978. 3/ 1/ The revenue performance of grants in recent years has declined sharply due to reforms in revenue sharing arrangements, as will be discussed later in Part One. This explains that the real per capita growth rate of grants is not significantly different from zero. 2/ Throughout this paper the consumer price index for Bogota (workers) is used in deflating financial data. The price index is based on Revista del Banco de la Republica de Colombia, various issues. 3/ Double log equations (n R = a + In b Pop and In R = a + b In Y) where R represents District revenue in real 1976 prices, Y equals real District income, and Pop is District population. WPS/BR-001-4a/3.20.85 - 174 - The population elasticities of shared tax and grant revenue is greater than one (Table 4.2), showing that for each additional percent of population in the District, there has been an increase of 1.3 percent of real shared tax revenues and 1.4 percent of real grant revenues. As Table 5.2 points out, shared taxes and grants are not significantly different from unity at a 95% confidence level. This compares with a population elasticity of self-financed revenue of 1.4 which is significantly different from unity at a 95% confidence level. The general tax and non-tax revenue is inelastic at .70, so that for an increase of ten percent in population, there has only occurred an increase of 7 percent in general tax revenue. The income elasticities show similar results, although they are lower than the population elasticities due to growth in real per capita income. The revenues of shared taxes, grants, and self-financing services have experienced a unitary income elasticity. 1/ Thus an increase of real income by one percent in the District has been associated with an increase of real revenue by one percent for each of these revenue sources. The unitary income elasticity of these three revenue sources is in marked contrast to that of general tax and non-tax revenue which is inelastic to income at about 0.5 and significantly below unity at a 95% level of significance. For each increase of real income by ten percent, there has only a five percent increase in general tax and non-tax revenues. In sum, one can conclude that shared taxes, grants, and self-financed revenus have kept up with, if not a little ahead of increases in population and income in the District. However, it must be noted 1/ The regression coefficients are not significantly different from unity at the 95% confidence level. WPS/BR-001-4a/3.20.85 - 175 - Table 4.2 Responsiveness of Consolidated Revenues to Changes in District Population and Incorm (1961-1978) Consolidated Revenue Type Population Elasticity Income Elasticity Tax and Non-Tax .7033" .518* Self-Financing 1.4112* 1.066 Shared Tax 1.264 .956 Grants 1.389 1.009 * SIgnificantly different from unity at 95% confidence level. Source: Fawcett and Lewis (1981). WPS/BR-001-4a/3.20.85 - 176 - that the responsiveness of grant revenue declined from 1975-1977. This drop will be explained below. General tax and non-tax revenus have not managed to keep pace with the increases of population and income. Central Administration Budget The above discussion has analyzed the performance of the various revenue sources in the context of the consolidated District budget. When distinguishing between revenues from the Central Administration and that of the Decentralized Agencies, the revenue composition takes on a different form. Shared tax revenue is received only by the Central Administration budget. Self-financed revenues are only found in the Decentralized Agencies budget. Grant revenue flows to both Central Administration and Decentralized Agencies. Grants are directed to only certain decentralized agencies, such as the C.V.P., E.D.T.U., the District Orchestra and University. However, total grant receipts have been only a small fraction, approximately 5%, of total decentralized agencies' revenues. Grant revenues have been of greater importance in the Central Administration budget, representing an average 7.8% of total revenues from 1961-1979 1/ (Table 4.3). Shared taxes are the second largest of Central Administration budget, averaging 33.2% for the years 1961- 1979. General tax revenue represents the largest revenue source at 51.7% for 1961-1979, and non-tax revenue (i.e., interest fees and fines) is the smallest at 2.3%. Borrowing on average has contributed 4.9% of District Central Administration revenue; however, this contribution has varied considerably over the years 1961-1979. 1/ The decline of grant receipt by the Central Administration after 1976 will be discussed later in this section. WPS/BR-001-4a/3.20.85 - 177 - Table 4.3 Annual Average Compound Growth Rates and Percentage Distribution of Central Administration Revenue (1961-1979) 1/ Average Percent Average Annual Growth Rate Distribution of In Current In Constant Real 1961-1979 Prices 1976 Prices Per Capita Total Revenues Taxes 18.1* 3.55* -1.94* 51.7% Non-tax Revenue 14.3* .15* -5.4* 2.3% Shared Taxes 23.3* 8.09* 2.4* 33.2% Grants 2/ 32.4* 17.8* 11.2* 7.8% Borrowing 31 7.7* -8.13 -13.7 4.9% * Significantly different from zero at 95% confidence level. 1/ Time period is calculated at 1961-1979 unless otherwise specified. 2/ Growth rates calculated for 1961-1976. 3/ Growth rates calculated for 1967-1979. Source: Fawcett and Lewis (1981). WPS/BR-001-4a/3.20.85 - 178 - Figure 4.2 graphically presents the contribution of the various revenue sources as a percentage of Central Administration revenues. The contribution of local direct taxes has been declining during this period while the shared tax revenue has been increasing. To some extent, this can be explained by the fact that selected local taxes were redefined into shared taxes under several fiscal reforms. The share of grants steadily increased until 1976 at which time the F.E.R. receipt was redirected, by-passing the Central Administration alLogether. Non-tax revenue has slightly declined during the 1961-1979 period. Annual average growth rates per capita also reflect this changing composition of Central Administration revenues (Table 4.3). For the tim' period of 1961-1979, the shared taxes performed best with a 2.4% growth rate in real per capita revenues. In contrast, local tax revenue and non-tax revenue declined in real per capita terms, at rates of -1.9% and -5.4% respectively. Rates of growth of real per capita revenues from grants and borrowing were calculated only for the years 1961-1976 and 1067-1979, respectively, because for some of the other years no revenues were received. 1/ The annual average compound growth of real per capita revenues from grants was exceptionally high at 11.2%. While the growth rates of these two sources are not comparable to those calculated for years 1961-1979, it can be concluded that the rate of growth in grants received by the Central 1/ The semi-logarithmic regression equation used to estimate trend growth rates cannot be applied for years when observations are equal to zero. - 179 - Figure 4.2 Composition of Central Administration Current Revenues (real figures in 1976 prices) 70 70 60- 60 500 50 - -5 40- -40 30- 30 20- -20 10 - - 10 o1 1 1 Io '69 '70 '71 '72 '73 '74 '75 '76 '77 '78 '79 Year Legend: - Taxes 0 O - - Situado Fiscal & Grants - Non-tax Revenues - Shared Taxes World Bank-26170 WPS/BR-001-4a/3.20.85 - 180 - Administration has been impressive, while its reliance on borrowing a balance has declined. 1/ The population and income elasticities calculated for these Central Administration revenue sources have comparable results (Table 4.4). The income and population elasticities of taxes, non-tax revenue, and shared taxes were calculated using the double log equation discussed earlier in the paper. The local direct tax is inelastic with respect to population and income at .52 and .37 respectively (Table 4.4). The elasticity of real per capita non-tax revenue is with respect to real per capita income is negative, as is the population elasticity of real non-tax revenue. Not surprisingly, shared tax revenues are responsive to income and population, and with elasticities not significantly different from unity. Elasticities of grants and borrowing are calculated for shorter time periods, 1967-1976 and 1067-1978 respectively. Grants are highly elastic to population and income (3.5 and 2.0, respectively). Real borrowing, similar to non-tax revenues, is negatively related to real per capita income and population (Table 4.4). Shared Taxes and Grants Shared taxes and grants became increasingly important sources of revenue for the Central Administration during 1961-1976. Shared taxes have consisted of. the beer consumption tax, the shared sales tax, the gasoline tax, 1/ However, the decline in borrowing by the Central Administration was revised in 1980, when this agency borrowed approximately two billion- Colombian pesos, amounting to almost 25% of total District Administration. WPS/BR-001-4a/3.20.85 - 181 - Table 4.4 Responsiveness of Central Administration Revenue to District Population and Income Changes Central Administration Population Income Revenue Type 1/ Elasticity Elasticity Taxes .5159* .374* (1961-1978) Non-Tax Revenues -.1338* -.1874* (1961-1978) Shared Taxes 1.264 .956 (1961-1978) Grants 3.465* 2.011* (1967-1976) Borrowing -1.325 -.8007 (1967-1978) * Significantly different from unity at 95% confidence level. 1/ Figures in parentheses are time periods for which the elasticities were calculated. Source: Fawcett and Lewis (1981). WPS/BR-001-4a/3.20.85 - 182 - and the tobacco tax. The main grant transfer during this period was the national transfer to pay for teacherst salaries, and, as of 1969, the situado fiscal (Table 4.5). The beer tax revenue trend of Figure 4.3 is a direct reflection of various changes in the tax rate and distribution procedure. The slight decrease of beer tax revenue from 1964-1968 can be explained by the restructuring of the beer tax, from a local departmental tax to a shared national and departmental tax. The dramatic increase in beer tax revenue in 1970-1971 represents the change in distribution procedure. Prior to 1970, the District government received a percentage of the Department of Cundinamarca's receipt; since 1970, the District government receives its revenue independent of the Department's distribution. The sudden drop of beer tax revenue in 1972 is a consequence of the tax rate being reduced from a 60% rate to a 48% rate. The contribution of beer consumption tax revenue for 1976 is 22.3% of total Central Administration revenue. The next largest transfer revenue share in the Central Administration is that of the grant to pay for teachers' salaries. The name of this transfer has changed throughout the years. Since 1969, the national revenue sharing program, the situado fiscal for education, has earmarked funds for teacher salary expenditure. This is included in the F.E.R. receipt as listed in Table 4.6. This transfer has been channeled directly to the F.E.R. since 1977, and for that reason receipts on account of situado fiscal dropped to zero in the District Administration budget as of 1977. Also, the dramatic increase of funds from 1969 to 1970 is largely attributable to the total national set WPS/BR-001-4a/3.20.85 - 183 - Table 4.5 Annual Average Compound Growth Rates and of Shared Taxes and Grants of Central Administration Percentage Distribution Real of 1976 C.A. Revenues 1/ Current Real Per Capita Total Revenue Beer Tax 23.15* 7.9* 2.3* 22.3* (1961-1978) Gasoline Tax 6.34* -7.3* 13.2* .7% (1961-1978) Shared Sales Tax 31.8 11.2 6.03 5.0% (1969-1979) Tobacco Tax 31.5 10.1 4.8 2.4% (1970-1979) Situado Fiscal 34.8 18.06 13.2 15.7% and Grants 1/ Figures in parenthesis are periods for which growth rates were computed. 2/ Given that a number of these shared tax and grant arrangements began in different years after 1961, an average historical percentage distri- bution must adjust for these various effective time periods. For that reason, only a percentage distribution for year 1976 is presented as an indicator of the revenue contribution of each of these sources. Source: Fawcett and Lewis (1981). WPS/BR-001-4a/3.20.85 - 184 - Table 4.6 Responsiveness of Shared Taxes and Grants to District Population and Income Changes Type of Shared Population Income Tax/Grant 1/ Elasticity Elasticity Shared Sales 3.920* 1.664 (1969-1978) Beer 1.200 .9154 (1961-1978) - Gasoline -.8486* -.7443* (1961-1978) Tobacco 3.519 1.408 (1970-1978) Situado Fiscal/Grants 2.35* 2.168* (1961-1976) * Significantly different from unity at 95% confidence level. 1/ Figures in parentheses are time periods for which elasticities were calculated. Source: Fawcett and Lewis (1981). - 185 - Figure 4.3 Composition of Central Administration Transfers 250- 250 200 CL150- -150 100- -100 0 Å '61 '63 '65 '67 '69 '71 '73 75 '77 '79 Year Legend: . - Shared Sales Tax . . - Tobacco Tax - Beer Tax - Situado Fiscal/Grants - Gaolln TaxWorld Bank-?6171 WPS/BR-001-4a/3.20.85 - 186 - aside for the situado fiscal for education as compared to the grant program in operation from 1961 to 1968. In 1976 this transfer program contributed 15.7% to total Central Administration revenue. The annual average compound growth rate calculated only for years 1961-76 using real revenue per capita is an extremely high 13.0%. The minor tax sharing arrangements are the shared sales tax, the gasoline tax, and the tobacco tax. 1/ Of the three, the shared sales tax revenue is the largest, making a 1976 revenue contribution of 5.0%-to total Central Administrtion revenue. Similar to the situado fiscal, the real growth per capita is high at 6.0%. The rate is calculated using only years 1969- 1979, and thus, it is not directly comparable to that of the beer tax or situado fiscal. The sales tax program began in 1965 with revenues reaching a peak in 1972 and 1975. The decline in revenues after 1975 reflects the change in the distribution procedure, as the result of which one-half of the District receipt of shared sales tax revenue was re-directed to the Ministry of Education. The gasoline shared tax, commonly known as the ECOPETROL tax (referring to the taxation on the Nationalized Petroleum Corporation - ECOPETROL), contributed .7% to the Central Administration during 1976. The real growth rate of per capita gasoline tax revenue, calculated for 1961-1979, has been negative at -13.2%. The steady decline of the tax revenue reflects the fixed rate of gasoline tax. The ECOPETROL tax rate of Col.$.06 per gallon has been in effect since 1959. 1/ Yet smaller tax-sharing arrangements and grants, such as the registration tax and the national transfers to decentralized agencies, will be discussed later in the paper. WPS/BR-001-4a/3.20.85 - 187 - The tobacco tax based on an ad valorem rate has performed much better than the gasoline tax. The contribution of the tobacco tax revenue in 1976 was 2.4%. During 1961-1979 the per capita receipts increased in real terms. The increase of revenues from 1968-1971 largely stems from the restructuring of the tax from a fixed rate schedule to an ad valorem schedule in 1966, and the increase of the ad valorem rate from 8% to 10% in 1971. The real growth rate of per capita tobacco tax revenue for 1970-1979 was 4.8%. In testing the responsiveness of these revenues to District population and income, the results are similar to the growth rate of the various transfers. 1/ (Table 4.6) All of the transfers are elastic with respect to population, except for the gasoline tax revenue. Shared sales tax revenue was most responsive to changes in population; with every additional percent of population in the District, there has been an increase of 3.9 percent of real shared sales tax revenue. Real shared tobacco tax revenue also has an extremely high elasticity of 3.5. The population elasticity of the situado fiscal revenues (1961-1976) is high; for each additional percent of population, real revenues from situado fiscal increased by 2.4 percent. This compares to real beer tax revenues, increasing by 1.2 percent for every percentage increase in population, The gasoline tax revenue, given its fixed rate structure, is inelastic with respect to population, such that with increasing population, we find decreasing amounts of real gasoline tax revenue of -.847. The beer tax and tobacco tax elasticities are not significantly different from unity. 1/ As can be seen by Table 4.6, the elasticities were calculated for varying years, depending on the beginning year of the revenue sharing arrangement. WPS/BR-001-4a/3.20.85 - 188 The income elasticities of the transfers do not vary as widely as do their population elasticities. Real beer tax revenue is not significantly different from unity at a 95% confidence level as are the real shared sales tax and tobacco tax revenues. The real revenue from grant and situado fiscal transfers is most responsive to real income at a 2.2 ratio elasticity. For every additional percent of real income, there is an increase of real situado fiscal of 2.2 percent. This contrasts with the responsiveness of gasoline tax revenue which has a negative income elasticity (-.74). Adjusting for F.E.R. and S.S.B. Transfers The growth rates and population and income elasticities have indicated the growing importance of the shares sales tax and situado fiscal revenue transfers during 1961-1976. However, as of 1977, their contribution to consolidated revenues and District Administration revenues dramatically declined. As mentioned, the distribution procedures of these shared revenues have been revised, by-passing the Central District Administration. The F.E.R. and S.S.B. directly received the transfer revenues of the situado fiscal. One-half of the shared sales receipt is channeled to the F.E.R. where it pays for secondary education expenditure, known as the "nationalization" of education. Although these transfers no longer show up in District budgets and accounts, it can be argued that the F.E.R. and the S.S.B. are to some extent under the control of the local authorities (see Section III below) and their revenues therefore ought to be accounted for in considering consolidated District revenues. To adjust for the transfer receipt, the 1976-1979 situado fiscal to S.S.B. and F.E.R. is added to the 1961-1976 Central Adminsitration situado fiscal transfer. Additionally, the shared sales tax receipts to the F.E.R. is added to the Central Administration sales tax receipt. Figure 4.4 - 189 - Figure 4.4 District Receipt Adjusting for F.E.R. and S.S.B. Receipt 250- -250 200- -200 a 150- - 150 915 S100- -100 50- 50 UI 0 10 '69 '70 '71 '72 '73 '74 '75 '76 '77 '78 '79 Year Legend: - Beer Tax Revenues C - Shared Sales Tax Adjusted - Situado Fiscal Adjusted - CA Shared Sales Tax Receipt - CA Situado Fiscal Receipt World Bank-26172 WPS/BR-001-4a/3.20.85 - 190 - illustrates the transfer revenue trends adjusting for the F.E.R. and S.S.B. receipt. The steady increase of situado fiscal revenue is most noteworthy. The real growth rate per capita of the situado fiscal with this adjustment (1969-1979) is 15.3%. The shared tax revenue under this adjustment is less consistent, rising quickly from 1976 to 1978 and then radically dropping in 1979. This decline represents the decrease in real revenue of both the Central Administration and the F.E.R. receipts. To emphasize the significance of national transfers to Di9trict finance, Table 4.7 presents the revenues and expenditures of the District Administration, District Decentralized Agencies, F.E.R., S.S.B., and the public hospitals. 1/ Although the data on the S.S.B. and the hospitals are incomplete for the 1978-1979 period, several observations can be made from the table. Of the District revenue and expenditure which approximates $11 billion real pesos in 1979, 11% of the revenue (or expenditure) is from the F.E.R. budget. S.S.B. represents a minimal 2% of this total consolidated budget. Based on the 1978 estimates, F.E.R. contributes around 11%. The hospitals' budget is approximately 6% of the total amount. District revenue trends clearly point to the increasing importance of transfer revenue in the District. Accompanying these transfers are a wide range of interrelationships between the different levels of District and national government. The rest of this paper is devoted to understanding the transfer processes, its ties to local and national government, and its impact on the District. 1/ S.S.B. and F.E.R. are almost entirely supported by national transfers. The hospitals receive approximately half of their operating revenues from national transfers. See Part Three for a more detailed breakdown of revenus and expenditures. WPS/BR-001-4a/3.20.85 - 191 - Table 4.7 Consolidated Revenue and Expenditure Budget of District Administration, Decentralized Agencies, F.E.R., S.S.B., and the Hospitals, 1978-1979 (in thousands of constant 1976 pesos) 1978 1979 Revenue Expenditure Revenue Expenditure District Administration 1.744.853 1.401.570 2.045.114 1.761.705 Decentralized Agencies 7.219.159 6.623.477 6.947.572 7.681.971 F.E.R. 1.155.848 1.161.994 1.286.414 1.239.329 S.S.B. n.a. n.a. 265.085 179.651 Hospitals 680.137 451.174 n.a. n.a. Source: Fawcett and Lewis (1981), F.E.R. (1980), and S.S.B. (1980). WPS/BR-001-4b/3.20.85 - 192 - II. INTERGOVERNMENTAL TRANSFERS TO THE DISTRICT In Section I various revenue sharing arrangements were identified. Section II examines each transfer arrangement in detail. To the extent possible, comparisons are made between the transfer programs. Especially the situado fiscal and the shared sales tax program have many similarities. The shared taxes on "vice", namely liquor, beer, and cigarette taxes, also are comparable. Tha remaining tax sharing arrangements and grant programs cannot be so easily grouped or compared. The Situado Fiscal The situado fiscal is a revenue-sharing program which transfers earmarked funds from the national government to the sub-national government for primary education and local health expenditures throughout the 33 1/ Colombian departments (provinces), territories and the Special District of Bogota. Because Bogota is the capital district of Colombia, combining in essence municipal and departmental characteristics, national funds flow directly to its local government, by-passing the Department of Cundi namarca 2/ (see Bird, 1975, p. 39-40). Base Formula. The origins of the situado fiscal are found in the national tax reform of 1968. Under Laws 33/68 and 39/69, Congress widened the national government tax base by nationalizing sales, registration, and beer taxes. Moreover, Congress assumed the responsibility to re-distribute tax revenues. As part of its tax distribution efforts, Congress authorized tax sharing 1/ There are 33 sub-national governments receiving situado fiscal funds as of 1979 (Republic de Colombia, 1979). 2/ This is the department surrounding the Special District of Bogota. WPS/BR-001-4b/3.20.85 - 193 - arrangements, one of which was a revenue sharing program earmarking funds from total national revenues. This program has become more commonly known as the situado fiscal. While the legislative directive of 1968 fails t specify the exact set-aside amount, Law 33/68 does establish a standard formula to allocate funds to the departments, territories, and the Special District: 30 percent of the total to be distributed in equal shares and 70 percent of the total to be divided in accordance with population using 1964 census statistics. 1/ The need to elaborate the 1969 tax sharing programs led to the enactment of the situado fiscal of 1971. Under Law 46/71, the formula for earmarking funds to the situado fiscal is as follows: SF = (%) * [IC-(PIV+PIVL+PIR)] where SF is the amount set aside for the situado fiscal grant; (%) is the percentage authorized by Congress which has been 15 percent since 1975; IC is the national current revenues; PIV is that share of the sales tax turned over to the sub-national governments; PIVL is the sales tax on departmental liquors; and PIR is the share of the national income tax received by a few departments. The deductions (PIV+PIVL+PIR) are nothing more than amounts already transferred from the national government to the department and District under other tax sharing arrangements (Bird, 1975, p. 39). Under the implementing Decree 1064/72, the percentage (%) set-aside was established at 1/ Strictly speaking, the law specifies that the officially ratified population figures of the last census are to be used. However, since to the present data the 1973 census figures have not been official released, the 1964 population figures still apply. WPS/BR-001-4b/3.20.85 - 194 - 13 percent (1973), at 14 percent (1974), and at 15 percent (1975), with future increases tied to revenue growth. 1/ Budget Process. The authorization of situado fiscal funds is provided through the national budget process. 2/ The Ministry of Finance, responsible for preparing the national budget, sets aside a certain amount of national revenue based on the above formula for the situado fiscal program. These funds are earmarked for primary education and health care expenditures according to the guidelines of Law 46/71, Article 2: "situado fiscal funds will be-distributed - as follows, 74% to primary education and 26% to local health care." In the national budget, the situado fiscal is entered under operational expenditures as follows: 1) funds for primary education under the Ministry of Education; 2) funds for local health care under the Ministry of Public Health; and 3) primary education and local health care costs of the National Territories under the Ministry of Finance. 3/ As part of the national budget preparation, the Ministries of Education and Health must submit their own appropriate budgets. The estimated - transfers from the situado fiscal are included in their expenditure budgets. In the Ministry of Education, situado fiscal funds are entered under 'primary 1/ Under this decree, the (%) was to increase by a maximum of 2 percent if current revenues increased by more than 15 percent annually as compared to the average of the paft three years" revenue. However, this criterion has not been applied. The (%) has remained at 15 percent since 1975. In any case, no more than 25 percent of current revenues can be appropriated for the situado fiscal under present law (Ruiz Betancourt, 1978, p. 50; Bird, 1975, p. 39). 2/ Decree 1064/72. 3/ The budget of the national territories is the responsibility of the Ministry of Finance, and for purposes of this paper will not be discussed. WPS/BR-001-4b/3.20.85 - 195 - education' expenditures to be transferred to the decentralized adminsitrative units, the Regional Educational Funds (Fondo Educational Regional, F.E.R.). In the Ministry of Public Health, the situado fiscal is not entered under a specific program expenditure, but is under the general transfer of Ministry funds to Local Health Agencies, (Servicios Seccionales de Salud, or S.S.S.) for the various departments, territories, and Special District. Distribution of Funds. The distribution of situado fiscal funds to the local S.S.S. and F.E.R. is still based on the 1968 formula: 70% based o-n population and 30% on equal shares using *1964 population census figures. Although this 70/30 formula is the only authorized criterion to be used, other factors enter into this distribution process through the supplementary budget appropriations. The initial budget outlay for the departments, territories, and Special District follows the 70/30 formula; the budget supplementals, or "modificaciones", do not require such approval. The importance of these supplements, authorized only for educational expenditures, will be discussed later in this paper. The existence of the supplementals, however, illustrates that the 70/30 distribution formula is not strictly adhered to, at least in the Ministry of Education. The disbursement of situado fiscal funds as with other national transfers to the S.S.S. and F.E.R. in the departments, territories, and Special District follows procedures stipulated in contracts signed by the respective Ministries and governors, or in the case of Bogota, the Mayor. These contracts vary among jurisdictions on such issues as payment schedule, types of local expenditure to be covered by the transfer payment, and the extent of control exerted by the Ministries over the disbursement process. According to District officials, as a general rule, the F.E.R. and S.S.S. of WPS/BR-001-4b/3.20.85 - 196 - large and complex educational and health systens located in the urban centers of Bogota, Cali, and Medellin have less binding contractual agreements with the Ministries. The autonomy of the District's F.E.R. and S.S.S. is certainly a case in point, and will be discussed later in this paper. Figure 4.5 summarizes the disbursement channels of the situado fiscal. Appropriated in the national budget (Level 1), situado fiscal funds represent 15% of current revenues. The distribution formula of 70/30 is applied (Level 2). Bogota receives 7.75% based on the population share and 3.03% based on the equal share. Once assigning the appropriate departmental/District share, the funds are then budgeted under the operational expenditures of the Ministries of Education and Health (Level 3). Through these Ministries the funds are transferred to decentralized administrative units, F.E.R. and S.S.S. 1/ (Level 4) in accordance with the calculations made at Level 2. Underlying Objectives. In understanding the development of the situado fiscal, it must be remembered that the national government of Colombia has long been involved in financing local educational and health-related services. Since the 1950s, national grants have paid for primary teachers' salaries and contributed to local hospitals. Education grants were based on teachers' salaries and numbers of students in the Department. Health care was funded through the Social Security Agency (ICSS) and the lotteries. Recipients of these funds were primarily urban for a number of reasons: (1) the highest teachers' salaries and greatest number of students enrolled 1/ From 1969-1975, the District Administration received the situado fiscal transfer from the Ministry. Since that period, the District F.E.R. receives its funds directly from the Ministry of Education. - 197 - Figure 4.5 Distribution of Situado Fiscal to District, D.E. Total National Current Revenues (net other tax-sharing transfers) 85% Revenues 15% Set Aside for General for Situado Level i Budget Expenditure Fiscal 70% 30% Level 2 7.75% 3.03% (Bogota) (Bogota) 74% 26% Ministry of Ministry of Level 3 - Education Health F.E.R. S.S.S. Level 4 I I r;dimary Local Health Educaflon Cost Services Source: Ruiz Betancourt, 1978. World Bank--26173 WPS/BR-001-4b/3.20.85 - 198 - were in urban areas; (2) the lotteries and ICSS originated in urban areas. However, the regional and income inequities springing from the grants programs were conflicting with the government's stated objectives for social and economic development in Colombia (Musgrave, 1971, p. 154; Netzer, 1967, p.57). The situado fiscal was introduced to eliminate the regional bias of the grants system in Colombia, as well as to equalize the opportunity to health and educational services throughout Colombia. The double objectives of the program, regional equity and minimal health and educational se-rvices provision, are the basis for the disstribution formula. 1. Regional equity calls for an equal distribution of funds to all regions, regardless of need. The formula reflects this objective by: (a) 30 percent of the funds being distributed based on equal shares, giving additional income to less densely populated, rural regions. (b) the use of 1964 population statistics which over-represent present rural population, thus favoring rural regions. 2. Minimum standards for services in health and education requires funding of elementary education and basic health care, The formula accounts for this objective by: (a) the earmarking of national funds, 74 percent to primary education and 26 percent to basic health services. (b) the increase of percentage revenue appropriated for the situado fiscal from 6.5 percent (1970) to 15 percent (1975), ensuring greater public spending on health and education expenditures. From this neat and tidy description of the situado fiscal, one might conclude that the program has re-directed national revenues to lower income, WPS/BR-001-4b/3.20.85 - 199 - lesser populated and rural regions. In analyzing the performance of the situado fiscal, this generalization cannot be easily substantiated. However, let us first look at the "sister" of the situado fiscal -- the sales tax sharing program -- and then proceed to a comparative analysis of the two programs. Shared Sales Tax. The sales tax is a national ad valorem tax on non-essential goods. Established in 1965 during a "state of economic emergency", the sales tax was changed from a specific tax on manufactured items and impo7ts to a general tax on all goods and services. Exemptions are limited to food, drugs and other basic items. From the period of 1965-74, increased sales tax rates have been established by Congress; and consequently, the sales tax revenue, as of 1979, contributes approximately 15% to total national revenue (DNP, 1979). Base Formula. The sales tax sharing program, established in 1968 under the same legislation of the situado fiscal (Law 33/63), distributes sales tax revenue to the departments and the District, a total of 23 sub-national gov- ernments. The national territories do not participate in this program. 1/ As with the situado fiscal authorization, the national sales tax sharing program falls under the national budget. Revenue from the national sales tax flows into the national account. Of the total sales tax revenue collected, 30% is 1/ Although Law 10/79 stated that the territories shall participate in the sales tax sharing program, the Supreme Court declared the decision . "unconstitutional". Only the 1980 shared sales will be re-distributed to the Territories, and not thereafter. WPS/BR-001-4b/3.20.85 - 200 - set aside in the budget for the sales tax sharing program. 1/ The remaining 70% of sales tax revenue is devoted to general national budget expenditures. Following the formula guidelines of the situado fiscal the distribution of sales tax revenue to the 23 departments and District is 70% based on population using 1964 census statistics, and 30% based on equal shares. Figure 4.6 illustrates the formula application. Level 1 represents the 30% set aside of total national sales tax revenue. Level 2 shows the distribution of these set-aside funds to the 23 sub-national governments. Earmarking of Funds. After applying the 70/30 formula for departmental distribution the set-aside funds are then earmarked for specific expenditures within each department and the District. Under Law 43/75, the distribution guidelines are (Level 3 of Figure 4.6): * 13% to the Ministry of Education to be passed on to the respective FER to finance secondary education; * 10% to the Ministry of Health to be transferred to the departments and District to finance the public employee pension funds (Caja Prevision Social, C.P.S.); * 77% to the departments for distribution to municipalities; in the case of Bogota, these funds are collected by the District Administration treasury for the Central Administration budget. Since 1976, however, 50% of the last share, i.e., the allocation to the departments (and District) for municipal revenue sharing is actually channeled 1/ The percentage of sales tax revenue allocated to the program has increased through the years: 10% in 1969, 20% in 1970, and 30% since 1971. - 201 - Figure 4.6 Distribution of Sales Tax Revenue to District, D.E. (1975-1980) Total National Sales Tax Revenues (100%) Level 1 70% Revenue 30% Set Aside for General for Shared Budget Expenditure Sales Tax Level 2 70% Based on 30% Dividend 1964 Population As Equal Shares 1 I 7.74% 4.35% (Bogota) (Bogota) Level 3 13.1% 76.9% 10% Ministry of Distict/ CP.S. Education Departmental (through Ministry Allocation of Health) Level 4 38.45% Earmarked for Nationalization (50% of 76.9%) 38.45% District District Receipt C.P.S. (50% of 76.9) District Dstict F.E.R. Treasury Source: Caceres. 1978 World Bank-26174 WPS/BR-001-4b/3.20.85 - 202 - to the Ministry of Education to support secondary education (Level 4). The Ministry of Education then transfers the 50% to the departmental and District F.E.R. to cover secondary education (Caceres, 1978, p. 129-30). Budget Process. Under these formula guidelines, the sales tax sharing programs are accounted for in the national budget under: (1) secondary education program expenditures in the Ministry of Education; (2) direct transfers to departmental District C.P.S. through the Ministry of Health; and (3) direct monthly transfers to the departments and the District through the Ministry of Finance. Like all national budget expenditures, these transfers which are paid out in monthly quotas must have the monthly approval of various top-level officials (i.e., National Controller, and Representatives of the Minister of Finance and of the National Bank). The transfers to the Ministry of Education are part of the overall program expenditures of that agency. Just as the situado fiscal is tied to the financing of primary education and local health care, so the shared sales tax program supports the financing of secondary education or, in Colombian budgetary terms, the "nationalization of education". Nationalization of Educationt In 1975, Congress nationalized education in Colombia by shifting the responsibility for teachers' salaries entirely to the national government. The already-existing situado fiscal paid for primary teachers' salaries, leaving only secondary education to be assumed by the district government. The shared sales tax was selected as the mechanism to pay for this nationalization. As stated in the 1975 formula, 50% of the departmental and District receipts flow to the Ministry of Education to cover secondary education expenditures, primarily that of salaries. WPS/BR-001-4b/3.20.85 - 203 - Under the 1975 nationalization, all municipalities, departments, and the Special District were to phase out their financing of secondary education over a five-year period with the Ministry assuming total financial responsibility by the end of 1980. Once the Ministry of Education achieved full responsibility for financing education in 1980, the distribution formula was to be changed: of the total shared sales tax 10% was to go to the departmental and District C.P.S., and 90% directly to the municipalities and District. There is much confusion over this 90% flow to the municipalities. In the legislation (Paragraph 3 of Law 43/75) there is no reference to shared sales tax continuing to support secondary education expenditure. The shared sales tax was authorized to sustain the nationalization only during the phase- out. The extension of the authorization or a permanent authorization was being discussed in Congress in 1980/81. 1/ Should, in fact, the shared sales tax continue to support the 'nationalization', the following procedure would probably be implemented: one-half of the 90% designated for the municipalities and the District would flow through the Ministry of Education to the departamental and District F.E.R. (Caceres, 1978, p. 129-30). There are many similarities between the sales tax sharing program and the situado fiscal which include: * funds authorized through national budget; * distribution formula of 70/30 using 1964 statistics; * expenditures earmarked for local education; * growth dependent on national revenue. 1/ Congressional action in November, 1980 extended Paragraph 3 of Law 43/75. The extension is in effect only until Congress reconvenes in 1981. WPS/BR-001-4b/3.20.85 - 204 - Given the shared characteristics of the two programs, it is possible to make a comparison of the programs' performance levels, and this is the task of the next section. Comparative Analysis of Situado Fiscal and Sales Tax Sharing Program. In assessing the performance of the two revenue sharing programs, let us first look more closely at their twin objectives: balanced regional development and equalizing opportunities in health and educational services. Government policy to bring about balanced regional growth7is part of Colombia's overall economic development policy. The five major regions of Colombia--Atlantic, Eastern, Bogota, Central, and Pacific--differ signifcantly in per capita output and income levels, and the fiscal capacities of the lower level of governments in these regions reflect these differences (Berry and Urrutia, 1976). The present formula of 70/30 using 1964 census statistics is aimed at re-directing national revenue from higher income to lower income regions. The direction of national resources towards lower income regions was thought to spur productivity and income growth in these regions, thus promoting regional income equity. The second objective, that of equalizing access to education and health, is a recurring theme of Colombian development policy. Education and health services are considered to be basic to meeting a minimum standard of life in Colombia and essential elements for economic development. Increased attention has been given to both quality and quantity of services in the lower income regions. The nationalization of education and health and their accompanying financing measures--the situado fiscal and the shared sales tax program--aim to equalize both opportunity to and quality of services in the five regions. WPS/BR-001-4b/3.20.85 - 205 - Putting aside the political rhetoric of the above policies, it may seem strange that redistribution of income was not a stated objective of the programs. Nowhere in the legislation of either program is there a reference to income redistribution. Yet, is not the consequence of tiansfers from high income to low income regions redI t.-butive? In fact, income radistribution takes place through a number of means. The obvious is the transfer from regions with higher per capita income to regions with lower per capita incomes, or regional redistribution. 1/ Secondly, public expendit-ures on specific education and health services (e.g., primary education and hospitals outpatient centers) benefit lower-income brackets more than higher-income brackets, redistributing income within the region (Selowsky, 1979; Berry and Urrutia, 1976). On the basis of the available evidence, there is considerable room for optimism that increased equity in access to education and health services contributes to improved vertical income equity. The underlying premise of the two programs is that national transfers can bring about regional equity and increased health and educational basic service levels. To test this premise would be an arduous task. A myriad of economic and social factors influence regional equity as well as the provision of health and educational services. The scope of analysis in this section is therefore more limited. It focuses on the factors within the revenue sharing process, by examining the distribution formulas and various other factors which determine the actual allocation of funds. An understanding of the impact of these factors provide 1/ The extent of redistribution depends on "the regional differences in. average income and the [income] distribution among poorer and richer regions"; see Berry and Urrutia, 1976, p. 136. WPS/BR-001-4b/3.20.85 - 206 - insight into the workings and effectiveness of the two revenue sharing programs. Sources of National Revenues Contributing to Set-Aside Amounts. As explained earlier, the set-aside formula for the situado fiscal is established at 15% of total current revenues minus other tax-sharing revenues. As can be seen in Table 4.8 the 15% set-aside applied only for years 1975-1977. In 1973, the set-aside percentage was 13%, in 1974 it was 14%. During 1973-77 current revenue of the national government-increased at an annual average compound growth rate of 29%. 1/ This growth figure compares to the situado fiscal growth of 34% during the same period. The higher growth rate of situado fiscal is explained by the increasing percentage of set-aside revenues. Four main tax revenue sources contribute to national current revenue: income tax, customs tariffs, national sales tax, and national gasoline taxes. During the years 1970-1979, the average contribution of these sources to current revenues was (Table 4.9): * 40.6% from income taxes; * 13.7% from custom tariffs and fees; * 17.6% from national sales tax; * 5.4% from gasoline taxes. Income taxes were by far the most significant national revenue source, with national sales tax and custom tariffs secondary in importance. Gasoline tax revenue contributes only a relatively small amount of revenue. 1/ All growth rates are calculated as average annual compound growth in current prices unless otherwise specified. WPS/BR-001-4b/3.20.85 - 207 - Table 4.8 Total Current Revenues, Percentage of National Government Set Aside and Authorization of Situado Fiscal Total Current Situado Fiscal Revenue Percent Appropriated 1973 23.948 13.0% 2.261.7 1974 32.615 14.0% 2.685.0 1975 42.630 15.0% 4.204.5 1976 55.591 15.0% 5.547.1 1977 65.157 15.0% 6.926.0 Source: Republica de Colombia, (1973) (1977). WPS/BR-001-4b/3.20.85 - 208 - The growth rates of the four sources are presented in Table 4.9. For the years 1973-77 the annual growth rate of the sales tax was highest at 42% in current prices. The other three sources showed lower, but still substantial rates of growth during the same period: the gasoline tax grew at 28% per annum, the custom tariff at 25%, and the income tax at 25%. Revenue performance must consider both the percentage contribution to total current revenue and the growth rate of the tax revenue. To gauge the full impact of the tax source on total revenue, the change of revenue from each source is calculated as a percentage of the change in total revenue over a period of time. 1/ Using 1970-79 budget data, the total current revenues increased around Col.$104 million while income tax revenues increased around Col.$28 million; or income tax revenue increase is 27% of current revenue increase. When comparing the other tax revenue increases to current revenue increases, sales tax is 25%, customs is 14%, and gasoline is 6%. From this simple test, it can be concluded that income tax increases best explain the growth of total current revenue--the revenue base of the situado fiscal program. Sales tax also contributes a significant amount to current revenues. As discussed earlier, 30% of sales tax revenue is set-aside for the shared sales tax program, and thus does not enter into the pool from which situado fiscal revenues are distributed. The sales tax sharing program is entirely dependent on sales tax revenue. With the high current growth of sales tax revenue at 42.1% per annum (1973-1977), it is surprising to find that the shared sales tax (30% of total 1/ Given the limited amount of data (5 observations) no multiple regression could be run to statistically determine the significance of these four revenu sources to total current revenue, or the situado fiscal program. WPS/BR-001-4b/3.20.85 - 209 - Table 4.9 Annual Average Growth of Major Revenue Sources and Percentage Contribution of Total Current Revenues (1970-1979) Annual Average Average Percentage of Revenue Source Growth Rate Total Current Revenues Total Current 28.9% 100.0% Income Tax 25.0% 40.6% (Personal and Corporate) Tariffs and Custom Fees 25.0% 13.7% Sales Tax 42.1% 17.6% Gasoline Taxes 28.4% 5.4% Source: Republica de Colombia (1970) (1979) WPS/BR-001-4b/3.20.85 - 210 - sales tax revenue) has grown only at 28% annually during the same years. To test if the shared sales tax program received its entitled 30% of sales tax revenue, the following regression equation is estimated using 1973-1977 data: SS = a + al ST where SS is shared sales tax receipt and ST is total national sales tax revenue. The regression coefficient a1 was estimated to equal 0.11, which suggests that the shared sales tax program is not receiving its full 30%, but only 11% of the marginal increase in sales tax revenue. 1/ The situado fiscal receipt, supposedly 15% of total current revenue, can be tested in like fashion; however, the 15% set-aside must be adjusted for "other shared tax arrangement". Subtracting these other shared tax revenues results in decreasing the pool of funds for the situado fiscal. (One such arrangement is the shared sales tax program, but since the shared sales tax revenue is only 3% of total current revenues, the effect of this subtraction is insignificant.) To test if the situado fiscal program is receiving its 15% of authorized funds 1/ the following equation is calculated using 1973-1977 data: SF = a + al (CUR - SS) with SF being situado fiscal receipt, CUR being total national current revenue, and SS being shared sales tax revenue which represents "other shared tax arrangements". The results show that the situado fiscal is receiving 1/ The regression coefficient .11 is, however, not significantly different from 30%, at the 95% confidence level. 1/ Given that the percentage was 13% in 1(713, and 14% in 1974, we would expect the 1973-1977 data to show-the mean percentage for these years at 14.4%. WPS/BR-001-4b/3.20.85 - 211 - 12.4% of a marginal increase in total current revenue, almost its full legally authroized amount. However, this coefficient is significantly different from 15% at the 95% confidence level. The findings of the above analysis can be summarized as follows: The growth of situado fiscal [at 34%] has been greater than that of its revenue base, total current revenues, (at 29%]. * Income tax revenue appears to be the most significant factor explaining the growth of current revenue, the base for ca-culating the situado fiscal. * The situado fiscal program is receiving 12.4% of the marginal increase in current revenues; this compares to the 15% authorized percentage. * The growth of shared sles tax program is substantially below that of its revenue base, the national sales tax. * The shared sales tax program is receiving only 11% of the marginal increase in sales tax revenue; this compares to the 30% authorized set-aside percentage. Factors Determining Inter-departmental Appropriations of Situado Fiscal and Shared Sales Tax Programs. A number of factors must be considered when examining the appropriations of situado fiscal and shared sales tax. First, there is the 70/30 distribution formula based on the 1964 census. What regions receive the most revenue per capita? How do the variables of population and income enter into the formula? Does the formula allocate funds based on "need", as measured by population size? Or alternatively, does the formula actually redistribute income? These questions examine the appropriateness of the formula under the two objectives of equity (e.g., WPS/BR-001-4b/3.20.85 - 212 - regional income redistribution) and need (e.g., health and educational service provision). Secondly, the supplemental appropriation of the situado fiscal for education acts as an additional factor in the authorization process. To what degree do the supplementals redirect revenues? What political, financial, or economic factors determine these supplementals? Let us now turn to some of these questions, to gain greater insight into the workings of revenue-sharing in Colombia. a. The Formula used in the Initial Appropriation Let us first look at the distribution of revenues under the formula of 70/30 using 1964 population statistics. The initial appropriations of the situado fiscal and the shared sales tax program strictly follows the formula criteria (authorization figures are shown for 1975 in Columns 3 of Table 4.10 and 4.11.) Alongside these authorized amounts are population and GDP per capita 1/ figures for each department and the District. For purposes of our discussion, GDP per capita is taken as a baseline for measuring equalization of income, and population is taken as a measure of "need". Although the relationship between these two variables, population and CDP per capita is positive, as illustrated in Figure 4.7, there are several departments which have low populations yet mid- to high-income per capita levels, such as Quindio, Risaraldra, and especially Guajira, where the increase of CDP has been dramatic in recent years. 1/ CDP is here defined as gross regional product for each department and the District, respectively. WPS/BR-001-4b/3.20.85 - 213 - Table 4.10 Departmental/District Authorization of Situado Fiscal 1975 Initial Total Department/ Appropriation Supplemental Appropriation District GDP per Capita Population Per Capita Per Capita Per Capita (Thousands of (Thousands) (Thousands of (Thousands of (Thousands of Pesos Pesos) Pesos) Pesos) Antioquia 16,053.4 3,488.0 122.3 9.4 131.7 Atlantico 21,556.9 1,078.2 138.7 11.6 150.3 Bogota 29,535.0 3,011.4 100.9 17.3 118.2 Bolivar 18,042.5 956.1 152.7 0.6 153.3 Boyaca 12,573.6 1,153.9 166.6 37.5 204.1 Caldas 12,584.0 855.4 174.2 18.4 192.7 Cauca 8,107.7 786.3 168.2 3.6 171.8 Cesar 23,856.4 399.8 194.6 5.1 199.8 Cordoba 14,313.7 767.7 168.0 8.6 176.6 Cundinamarca 16,334.8 1,301.0 164.0 9.3 213.4 Choco 4,146.5 232.9 281.0 55.6 336.6 Guajira 18,887.4 209.1 286.7 10.1 296.8 Huila 12,296.3 557.2 183.6 29.8 213.3 Magdelena 14,228.6 623.5 192.3 11.1 203.4 Meta 22,591.1 270.4 232.9 20.3 253.2 Narino 8,751.4 892.2 165.5 29.2 194.7 N. de Santender 11,507.5 755.7 164.5 9.1 173.5 Quindio 12,310.7 344.5 246.6 11.0 257.7 Risaralda 15,923.2 528.9 199.6 5.2 204.9 Santender 16,246.7 1,385.2 148.0 24.7 173.7 Sucre 15,875.7 379.2 227.0 9.7 236.7 Tolima 15,336.0 1,091.2 155.0 16.7 171.7 Valle 20,508.4 2,390.0 129.4 1.7 131.1 Source: DNP (1978) WPS/BR-001-4b/3.20.85 - 214 - Table 4.11 Departmental/District Authorization of Shared Sales Tax 1975 Authorization Department/District GDP per Capita Population Per Capita (Thousands of Pesos) (Thousands) (Thousands of Pesos) Antioquia 16,053.4 3,488.0 77.3 Atlantico 21,556.9 1,078.2 92.9 Bogota 29,535.0 3,011.4 64.6 Bolivar 18,042.5 956.1 102.3 Boyaca 12,573.6 1,153.9 109.5 Caldas 12,584.0 855.4 116.5 Cauca 8,107.7 786.3 11-3.8 Cesar 23,856.4 399.8 140.2 Cordoba 14,313.7 767.7 113.8 Cundinamarca 16,334.8 1,301.0 106.9 Choco 4,146.5 232.9 208.0 Guajira 18,887.4 209.1 215.8 Huila 12,296.3 557.2 127.4 Magdalena 14,228.6 623.5 131.4 Meta 22,591.1 270.4 173.0 Narino 8,751.4 892.2 110.9 N. de Santender 11,507.5 755.7 112.0 Quindio 12,310.7 344.5 175.2 Risaralda 15,923.2 528.9 138.3 Santender 16,246.7 1,385.2 97.7 Sucre 15,875.7 379.2 161.0 Tolima 15,336.0 1,091.2 102.7 Valle 20,508.4 2,390.0 82.9 Source: DNP (1978) - 215 - Figure 4.7 GDP 1975 By Population (Thousands of Pesos Per Capita) 30 Bogota 25 - * Cauca Meta O Valle 20- * Guajira .2 o a. 15 0 Antioqula E 0 _ Quindio@ 0 0 10 0 10 * Naino 0 5 * Choco n I I I II 0 500 1,000 1.500 2.000 2.500 3,000 3,500 Population of Departments Cuentas Reglonales (1977), DNP (1978) Word Bank-26175 WPS/BR-001-4b/3.20.85 - 216 - Given that the distribution amounts under the situado fiscal and shared sales tax differ, it is necessary to analyze each individually. To test if the initially authorized amount of the situado fiscal is distributed on a "1need" basis, the initial per capita authorizations for the 23 departments and the District are regressed on population for years 1973, 1975, 1977: IA/PC = a0 + al Pop where IA/PC is initial authorization per capita of situado fiscal, and Pop is the population of each department and the District. The results of the regression are the following: al Year (t statistic) 1973 -2.8 x 10-5 (3.50) 1975 -4.2 x 10-5 (-5.44) 1977 -6.4 x 10-5 (5.53) In all cases, the initial appropriation per capita is negatively related to population, and significantly so. Thas, departments with larger populations, such as the Antioquia and Valle, and the District of Bogota, receive lesser amounts of initially allocated funds per capita than the less-populated departments. Clearly, the formula is not based on "need". To test if the distribution formula is "re-distributive", the initial appropriation of situado fiscal per capita for 1973, 1975, and 1977 are regressed with GDP per capita as the indepartment variable: AI + GDP P- a + C WPS/BR-001-4b/3.20.85 - 217 - with the following results: a, Year (T-statistic) 1973 -4.764 x 10-3 (-2.23) 1975 -2.835 x 10-3 (-1.63) 1977 -0.947 x 10-3 (-.72) which show a negative direction of the relationship, such that departments with greater GDP per capita receive less situado fiscal funds as allocated by the formula. However, only in 1973 is the regression coefficient significant at a 5% confidence level. Since that time, the negative relationship between per capita initial allocation and per capita income has become increasingly insignificant. Figure 4.8 graphs this relationship. Several departments-- Guajira, Meta, --lead there lationship to break down. These departments with relatively high GDP per capita receive a much larger share of situado fiscal funds than other departments at a comparable GDP per capita level. The formula, to the extent that it is represented within the initial authorization of situado fiscal funds, has been redistributive in the past, yet appears to have no significant redistribution effects as of 1977. 1/ 1/ Income elasticity for L" initial appropriation was also tested. Given that the results bore the same findings as the above, they will not be presented in this paper. - 218 - Figure 4.8 Initial Appropriation 1975 By GDP (Thousands of Pesos Per Capita) 0.30 * Guajlra * Choco 0.25 * Quindio 0 Meta 0 c 0 10.20 0 0 o o C0 O0 0 00 O O Norino 0.15 0 00 O Valle 0.10 Bogota &l 0 5. 10 15 20 25 30 35 GDP of Departments Cuentos Regionales (1977). DNP (1978) World Bank-26176 WPS/BR-001-4b/3.20.85 - 219 - Under similar analysis, the shared sales tax appropriations have much the same results. In testing the shared sales tax appropriation to the objective of "need", the initial appropriations per capita for each department and the District are regressed on population for 1973, 1975, 1977: SS S a + a1 Pop PC o0 The results of the test show a statistically significant negative relationship between the two variables of shared sales tax receipt per capita and population. a1 Year (T-statistic) 1973 -1.1 x 10-5 (-5.19) 1975 -3.4 x 10-5 (-5.24) 1977 -4.9 x 10-5 (-3.53) To test the redistributive effect of the shared sales tax, the initial appropriations per capita are regressed by GDP per capita for 1973, 1975, and 1977: SS CDP m a0 + a PC o 1, PC The statistical findings are: a1 Year (T-statistic) 1973 -.00105 (-1.38) 1975 -.002020 (-1.38) 1977 +.00140 (+1.12) WPS/BR-001-4b/3.20.85 - 220 - The regression lines based on the regression coefficients a1 are negative for years 1973 and 1975, illustrating a redistributive effect of the program, albeit insignificant at 5% confidence level. The a1 coefficient for 1977 is positive, indicating just the opposite, that the appropriation was not redistributive. In all cases, the a1 coefficients are extremely small, and more importantly, are insignificant. The regression of the initial appropriation per capita of the sales tax on per capita GDP of each department and the District shows the same bias toward low populated departments regardless of its GDP level, as was the case for situado fiscal (Figure 4.9). Once again, the District of Bogota, and the departments of Valle, Antioquia, and Atlantico receive the smallest transfer per capita. Based on the above findings, the following conclusions can be drawn given that the initial appropriations of the two revenue-sharing programs accurately reflect the allocation of funds based on the 70/30 formula using 1964 statistics: * The formula clearly does not allocate funds on "need" in either the situado fiscal or shared sales tax program, where need is defined in terms of departmental and District population size. * The formula appeared to be significantly "redistributive" in 1973; however, it has failed to achieve this objective in 1975 and 1977. - 221 - Figure 4.9 Appropriated Shared SalesTax 1975 By GDP (Thousands of Pesos Per Capita) 0.25 Choco 0 Guajira 0.20 QuIndio 0 0 Meta - fn0 0.15 0 Cauca 0 o O O 0.10 O 0 Valle 0 Bogota 0.05 0 I I I I I 0 5 10 15 20 25 30 35 GDP of Departments Cuentas Reglonales (1975). DNP (1978) World Bank-26177 WPS/BR-001-4b/3.20.85 - 222 - * The extent to which the formula can redistribute revenue is closely tied to the relationship of GDP per capita and population of each department/District. When there is a strong inverse relationship (e.g., high GDP per capita and low population) the formula-cannot achieve its redistributive objective. b. Factors Influencing Supplemental Appropriation Only the educational component of the situado fiscal has supplemental appropriations, 1/ funds which are added to its budget during the course of the fiscal year (Table 4.10). Under the supplemental, the 70/30 formula is not used, rather other factors come into play. Let us first identify the jurisdictions which receive supplemental appropriations and then proceed to analyze what factors may influence the appropriation of these supplementals. Under the regression analysis of supplementals to GDP per capita, the low populated departments of Choco, Meta, and Cauca again receive the greater amount of supplementals for their GDP per capita levels (Figure 4.10). More interestingly, the observations of Bogota and the department of Cundinamarca lie well above the regression line. The District of Bogota with the highest GDP per capita and greatest population of any sub- national government, receives a greater share of supplemental funds than 1/ The available information only notes supplementals in 1973, 1975, and 1977. It is not clear from the data source if there were no supplementals in 1974 and 1976, or if they were not reported. - 223 - Figure 4.10 Supplemental 1975 By GDP (Thousands of Pesos Per Capita) 0.07 0.06 - * Choco 0.05 - * Cundinamarca 8 0.04 - 0 Cauca E 0.03 - O E CD .Meta 0.02 - Bogota 0.01 -Quindloe O 0 O Guajira Narnoe 0 0 J 1 n A eValle 0 5 10 15 20 25 30 35 GDP of Departments Cuentas Regionales (1977). DNP (1978) World Bank-26178 WPS/BR-001-4b/3.20.85 - 224 - other departments with similar GDP levels. The regression coefficients relating per capita suppementals to per capita GDP are negative, implying that the supplementals are redistributive. al Year (T-statistic) 1973 -3.087 x 10-3 (-2.56) 1975 -1.002 x 10-3 (-1.89) 1977 -0.440 x 10-3 (-1.06) However, the coefficients are small and insignificant at a 1% confidence level. (Only the 1973 coefficient with a t-statistic of -2.56 is significant at a 5% confidence level). Having identified which jurisdictions benefit especially from the supplemental appropriation, it is necessary to ask what they have in common. Since the supplemental appropriation supports primary education expenditures, one might expect that demand for primary education may be a factor which influences the supplementals. Under this assumption, supplementals would be responsive to demand for educational funds. To measure this demand, the number of school age children per capita of each department and the District was calculated (DNP, 1979). In testing the influence of educational demand, the supplemental appropriations per capita of the 23 departments and the District are regressed on three variables for 1973, 1975, and 1977: SUP GDP SCHAGE PC- a + a P+ a2 Pop + a3 SC PC o 1 PC 2 3 PC WPS/BR-001-4b/3.20.85 - 225 - where SUP/PC is supplemental appropriation per capita, CDP is GDP per capita, POP is population, and SCHAGE/PC is number of school age per capita. In all three regressions, the results of all three variables are insignificant at a 10% confidence level. 1/ The above analysis thus provides little insight into the factors which determine the-supplemental appropriation. Another factor which may explain the supplemental is the political nature of the appropriation process. Political "noise" made during the fiscal year from the various departments and the District F.E.R.s may be an important determinant of the supplemental appropriation. Measuring the political pull of each department and the District is clearly difficult. Yet, just as important as the "noise", is the reception of the "noise" at the Ministries of Education and Finance. As a proxy variable of the special consideration (i.e., the supplementals) given to each department and the District, the proportion of liberal votes in the Presidential election of 1974 were taken for each jurisdiction. In running a correlation of these votes to 1975 supplementals per capita, correlation for 1975 is extremely low and insignificant at .0829. The above analysis does not allow us to distinguish any specific factor which determines the supplemental appropriation. Nevertheless, it is likely that a combination of factors including educational service demand and political considerations, enter into the appropriation process. 1/ Given the insignificance of the regression coefficients, the results are not shown here. WPS/BR-001-4b/3.20.85 - 226 - Another factor which may explain the supplemental is the targetting of funds to specific expenditure items and its effect on the appropriation process. Given the complexity of the process, the legitimacy of the above proposition will be examined later in this paper. c. The End Result -- The Total Appropriation The total appropriation is the sum of the initial and supplemental appropriations. As seen in Table 11-3, the supplementals are a small percentage of the total situado fiscal appropriation. For that reason, the total appropriation largely reflects the initial appropriation in the situado fiscal. Following the earlier analysis, the two objectives of need and equity are tested for in the total appropriation of situado fiscal. Under the need objective, the total appropritions per capita (TOT/PC) are regressed by population (POP) for years 1973, 1975, and 1977: TOT - a + a, POP PC o1 The following are the regression findings: al Year (T-statistic) 1973 -3.97 x 10-5 (-3.17) 1975 -4.40 x 10-5 (-4.74) 1977 -6.51 x 10-5 (-4.32) In all three years, the a1 coefficients are significantly negative. These results reconfirm our earlier findings that departments with lesser population receive greater amounts of situado fiscal per capita. WPS/BR-001-4b/3.20.85 - 227 - Similarly, when testing the equity objective, the total appropriations per capita for 1973, 1975, and 1977 are regressed by GDP per capita (GDP/PC): TOT GDP = + ai~ PC o0 PC As expected, the results of this test are similar to those in the earlier analysis of initial appropriation by GDP per capita. a1 Year (T-statistic) 1973 -7.851 x 10-3 (-2.44) 1975 -3.837 x 10-3 (-2.05) 1977 -0.507 x 10 -3 (-.33) However, two distinctions can be made regarding the two regression findings. First, the coefficients a1 are significant at a confidence level of 5% for both 1973 and 1975. Under the initial appropriation analysis, only the 1973 coefficient is significant. Secondly, the effect of the supplementals can be seen in the total appropriation (Figure 4.11). Particularly the observations of Bogota and Cundinamarca are clearly above the observations of the initial appropriation. Much of the above analysis has emphasized the normative objectives of equity and need which are the underpinnings of the distribution formula. As we have seen, the formula allocations have not been based on need. Departments with lesser population receive greater per capita amounts of appropriated situado fiscal or shared sales tax funds. The data has weakly suggested that the formula meets the equity - 228 - Figure 4.11 Total Appropriation 1975 By GDP (Thousands of Pesos Per Capita) 0.35 0 Choco 0.30 * Guajiro c 0.25 - * Qulndlo a Metr 0. a0 cV 2o a0ogt S0.20 - O 1-0-Norino 0 O O0 O 000 0.15 - O O O 0 Vallee Bogota 0,10 I I I1 0 5 10 15 20 25 30 35 GDP of Departments Cuentas Regionales (1977), DNP (1978). Wodd Bank-26179 WPS/BR-001-4b/3.20.85 - 229 - objective, by channeling greater amounts of transfers per capita to jurisdictions with lower GDP per capita levels. However, the single regression analysis of the two variables (e.g., population and GDP per capita) does not hold the one variable constant while testing for the second. To clearly distinguish between these two variables, and their significance in the distribution of situado fiscal and the shared sales tax, multiple regression analyses are necessary. Furthermore, to distinguish between "population need" and "educational need", two variables, population per department (POP) and number of school age children (SCHACE) per capita of each department, are included in the multiple regression equations tested for years 1973, 1975, and 1977: (Situado Fiscal) TOT a + a + a POP + aSCHAGE PC 0 PC 2 3 PC SS GDP SCHAGE (Shared Sales Tax) a + a + a POP + a C PC o 1 PCP2 3 PC Table 4.12 presents the results of the multiple regression analysis of the situado fiscal. The population variable remains inversely related to total appropriated situado fiscal per capita, and is significant for all three years tested. The equity variable, that of GDP per capita, is also negatively related to the total appropriation per capita yet this relationship is insignificant. Surprisingly, holding population constant the total appropriation per capita and "educational need" have a positive relationship which is significant in 1973 and 1977. These new results suggest that the situado fiscal may well achieve the need objective, if one defines need as school age children per capita. Also, based on these multiple regression findings, the "equity" WPS/BR-001-4b/3.20.85 - 230 - Table 4.12 Multiple Regression Analysis of Total Appropriation and Shared Sales Tax per Capita by Population, GDP per Capita, and Educational Need Variables TOT/PC = a + a1 GDP/PC + a2 POP + a3 SCHAGE/PC Year a a a R2 1973 -.005306 -.0000279* +.10213* .415 1975 -.002006 -.0000426* +.13415 .607 1977 -.012029 -.0000527* +.44909* .664 and SS/PC = a + a1 GDP/PC + a2 POP + a3 SCHAGE/PC 1973 +.000185 -.0000231* +.000055 .58 1975 +.000158 -.0000625* +.000138 .58 1977 +.00227* -.00001722* +.000662* * Significance to a 90% confidence level Source: Fawcett calculation based on DNP (1978) data. WPS/BR-001-4b/3.20.85 - 231 - objective seems to be insignificant in allocation of situado fiscal funds. If any general statement can be made regarding the appropriation and its formula (which largely determines the various allocations), it is that inter-departmental authorization tends to allocate revenues to the lower populated departments regardless of GDP per capita levels. Under similar multiple regression analysis, the shared sales tax program has similar results. Population is negatively related, and significantly so at the 90% confidence level in 1973 and 1975, and at the 99% confidence level in 1977. The educational demand variable -- number of school age children per capita in each department -- is positive, reconfirming our findings in the situado fiscal. Most curiously, the equity variable -- GDP per capita -- is positively related to shared sales tax transfer per capita. In the simple regressiosn analysis, this relationship is negative. However, in both tests, the coefficients a1 are highly insignificant (except in the 1977 multiple regression analysis where it is significant to a 99% confidence level.) Based on these findings, it must be emphasized that the shared sales tax program clearly does not allocate transfer revenues under the "equity" objective.' Educational need does seem to be a factor in the distribution of funds. Moreover, the allocation, similar to the situado fiscal, redistributes revenues to lesser populated departments and not according to lower GDP per capita levels. Actual Distribution of the Funds by Department/District. Under both the situado fiscal and the sales tax sharing program, the actual distribution of WPS/BR-001-4b/3.20.85 - 232 - funds is much less than the authorized total appropriation (Tables 4.13 and 4.14). 1/ The gap between authorized and distributed amounts is most striking in the case of the shared sales tax program (Figure 4.12). As illustrated, there is an increasing gap in authorized versus distributed amounts, as well as a decline in the actual distribution of shared sales tax funds during 1975- 1976. Under the situado fiscal for education, the DNP data only reports a gap in the authorized and distributed amounts in 1973 and 1976. (DNP, 1978). In 1973 the departments with the greatest gaps relative to its total appropriation are: Choco, Cauca, and Cundinamarca. (See Table 4.15). In 1976, a uniform (and larger) percentage shortfall was reported for all departments with actual allocations accounting for 94.3% of the authorized amounts. The situado fiscal for health expenditures has an even greater gap between authorized and distributed amounts. Table 4.16 presents this gap for 1976. The impact of the shortfall of revenues in the situado fiscal and sales tax sharing program is probably most serious for local government planning. Before turning to these consequences, an explanation of the gap between authorized and distributed amounts is in order. Under the appropriation and authorization process, responsibility for the situado fiscal for health and education falls under the Ministries of Health and Education respectively. The shared sales tax budget is the responsibility of the Ministry of Finance. It could be expected that the 1/ The authorized amount not distributed is recorded under 'reservas' of the national accounts. However, there is no evidence if the 'reservas' are distributed during the next fiscal year. WPS/BR-001-4b/3.20.85 - 233 - Table 4.13 Authorization and Actual Distribution Cap of Shared Sales Tax Revenue, 1973-1977 (in Colombian pesos) Year Authorized Actual 1973 691,999 690,343 1974 840,686 825,024 1975 2,367,003 2,161,203 1976 2,479,708 1,776,202 1977 2,468,688 1,630,296 Source: DNP (1978) Table 4.14 Authorization and Actual Distribution Gap of Situado Fiscal for Education, 1973-1976 (in Colombian pesos) Year Authorized Actual 1973 2,261,700 2,244,235 1976 5,547,100 5,343,092 Source: DNP (1978) - 234 - Figure 4.12 Authorized vs Distributed National Sales Tax (Thousands of Pesos) 2.500 / . - -- - --------- --- - Auhrzto 1 Authimztion Versus Actual 2,000 Receipt Gap Decline in Authorized 0 Authorization 1,500 Distributed 1,000 500 '73 '74 '75 '76 '77 Year DNP (1978) World Bank-26180 WPS/BR-001-4b/3.20.85 - 235 - Table 4.15 Authorized vis-a-vis Actual Allocations of 1973 Education Situado Fiscal by Departments and District (Thousands of Current Pesos) Gain/ Percent of Department/District Authorized Actual Shortfall Authorized Antioquia 225.0 224.6 -.44 99.8% Atlantico 79.0 78.6 -.38 99.5% Bogota, D.E. 171.2 171.5 -.28 99.8% Bolivar 78.2 78.0 -.22 99.7% Boyaca 126.7 125.8 -.81 99.4% Caldas 80.4 80.3 -.09 99.9% Cauca 93.7 92.4 -1.35 98.6% Cesar 45.4 45.1 -.28 99.4% Cordoba 62.9 62.7 -.19 99.7% Cundinamarca 133.6 132.6 -1.06 99.2% Choco 74.8 72.7 -2.10 97.2% Guajira 39.9 39.5 -.41 99.0% Huila 63.1 62.4 -.72 99.0% Magdalena 77.1 76.9 -.53 99.3% Meta 42.8 42.3 -.51 98.8% Narino 102.4 101.8 -.64 99.4% N. de Santender 67.9 68.1 -.18 100.2% Quindio 49.7 49.2 -.45 99.1% Risaralda 58.6 58.4 -.18 99.7% Santender 113.2 112.8 -.40 99.6% Sucre 49.5 49.4 -.19 99.6% Tolima 109.0 108.4 -.56 99.5% Valle 168.9 168.4 -.44 99.7% Source: DNP (1978) WPS/BR-001-4b/3.20.85 - 236 - Table 4.16 Authorized vis-a-vis Actual Allocations of 1976 Health Situado Fiscal by Departments and District (Thousands of Current Pesos) Gain/ Percent of Department/District Authorized Actual Shortfall Authorized Antioquia 175.7 152.3 23.4 86.7 Atlantico 61.6 58.6 3.0 95.1 Bogota, D.E. 125.2 118.9 6.3 95.0 Bolivar 60.2 57.2 3.0 - 95.0 Boyaca 79.2 68.6 10.6 86.7 Caldas 61.4 58.3 3.1 95.0 Cauca 54.5 51.8 2.7 95.0 Cesar 32.1 27.8 4.3 86.6 Cordoba 53.1 46.5 6.6 87.6 Cundinamarca 87.9 76.2 11.7 86.7 Choco 27.0 23.4 3.6 86.7 Guajira 24.7 21.4 3.3 86.6 Huila 42.1 40.0 2.1 95.0 Magdalena 49.4 46.9 2.5 94.9 Meta 25.9 22.5 3.4 86.9 Narino 60.8 52.7 8.1 86.7 N. de Santender 50.0 47.3 2.7 94.6 Quindio 35.0 30.3 4.7 86.6 Risaralda 43.5 41.3 2.2 94.9 Santender 80.1 69.4 10.7 86.6 Sucre 35.5 30.7 4.8 86.5 Tolima 69.7 66.2 3.5 95.0 Valle 127.5 121.1 6.4 95.0 Source: DNP (1978) WPS/BR-001-4b/3.20.85 - 237 - Ministries of Health and Education are most intent on receiving the authorized monies to provide the services as programmed in their budget. The Education and Health ministers not only report to the Ministry of Finance and the Controller, but act as the main spokepersons for their respective services, which are administered at the local level by the respective F.E.R. and S.S.S. On the other hand, the Ministry of Finance is not intimately tied to the workings of each local government's budget operation. Should the actual sales tax receipt fo the municipalities not equal the authorized amount, there is no channel to present these local grievances to the Ministry. This contrasts with the Ministry of Education where teachers' strikes are a very real threat. This suggests that the lack of linkage between the Ministry of Finance and the local government and the fact that the Ministry of Finance does not have a mandate of direct service provision account for the sizeable gap in the authorized versus distributed amounts of shared sales tax. The District Receipt of the Situado Fiscal and Shared Sales Tax. Throughout the above discussion, the receipt of funds under the situado fiscal and the shared sales tax program to the District has been compared to that of the other departments. Our findings have shown that: a) The 70/30 formula using 1964 population statistics does not favor distribution to the District. Greater amounts of revenues per capita flow to lesser populated departments than to the District with its relatively large population; b) The supplemental appropriation of the situado fiscal tends to benefit the District; c) The gap between authorized and distributed funds decreases the receipt of funds in the District more than for other jurisdictions. WPS/BR-001-4b/3.20.85 - 238 - Since 1975, the District Administration has only received the shared sales tax directed to municipalities and that receipt is only 50% of that previously allocated. From 1969-75, the District received both the situado fiscal and shared sales tax receipt. 1/ Figure 4.13 reflects this change. The 1976 and 1977 receipts of sales tax revenues decreased significantly as compared with 1975. The 1977 receipt is almost one-half of the 1975 receipt. This partly explains the minimal contribution of the sales tax to District finances. The situado fiscal receipt of 1977 drops to zero, reflecting the fact that these transfers are now directed to the District F.E.R. and S.S.B. To adjust the Central Administration reven+s for the transfers, the F.E.R. and S.S.B. receipts from the situado fiscal is added for the years 1976 through 1979. During the period 1970-1979 the annual average growth rate of the real situado fiscal per capita is then calculated as 15.3%. Even with the high growth of the two programs, the District feels shortchanged. 2/ a. The Distribution Formula The distribution formula is the main exampale of the "anti-Bogota" tendency of revenue-sharing. To calculate how much the District actually loses under the revenue-sharing arrangements, it would be necessary to determine both how much it gives and how much it receives. The problem in calculating the District revenue which 1/ The C.P.S. and S.S.S. receipts for the District are outside of the District Administration budget, and for purposes of comparison will not be included in the following discussion. 2/ Castano (1978), p. 81, 82, 84. Much of the criticism deals with the contribution of the District to the national government for the situado fiscal program, as reflected by national tax collections in Bogota. - 239 - Figure 4.13 Percent Sales Tax and Situado Fiscal Receipts (of District Administration Revenue 1969-1977) Sales & Situado Fiscal7 25 - 151 10 Situado Fiscal o Sales Tax '69 '70 '71 '72 '73 '74 '75 '76 '77 Year Lewis (1976) World Bank-26181 WPS/BR-001-4b/3.20.85 - 240 - supports the situado fiscal and shared sales tax program is that of income tax and sales tax reporting. Headquarters for many industries are located in the District. These companies report their income earned or sales made from headquarters. However, neither the income nor the sales may actually have originated i*n the District. Thus, the tax incidence does not fall on the District, but rather on another region of Colombia, undiscernible from the sales tax or income tax data. Because of this complication, this paper will not attempt to calculate how much Bogota contributes to the revenue- sharing programs. Rather, it looks at the effects of the formula and other national factors on the District receipt. The District receipt of situado fiscal is presented in Table 4.17, and of shared sales tax funds in Table 4.18. The situado fiscal actual receipt has been less than the authorized amount from 1973 to 1977. The shared sales tax receipt has been Less for every year, except 1974. 1/ For purposes of comparison, the formula has been hypothetically changed in Table 4.19. Two alternative formulas applied are: 1) distribution based on 100% of current population according to DANE estimates; and 2) distribution based 100% on equal shares amongst the departments, District, and territories. Under- standably, the amount of funds received under the present formula is less than the formula using only population statistics., 1/ It may be that the 1974 receipt included back payments from 1973. The data does not allow for such distribution. WPS/BR-001-4b/3.20.85 - 241 - Table 4.17 Authorization vs. Actual Distribution Gap of District Situado Fiscal Receipt to F.E.R. (Thousands of pesos) Percent of Year Authorized Actual Receipt Authorized 1973 171,542.2 142,426.0 83.0 1974 212,598.6 166,475.9 78.3 1975 355,982.9 297,803.3 83.6 1976 446,557.6 314,596.4 70.4 1977 584,062.1 546,992.5* 93.6 * Estimated as average of 1976 and 1978 receipt. Table 4.18 Authorization vs. Actual Distribution Gap of District Shared Sales Tax to District Treasury (Thousands of pesos) Percent of Year Authorized Actual Receipt Authorized - 1973 56,691 35,571.0 62.7 1974 68,815 77,138.5 112.1 1975 194,576 163,899.8 84.2 1976 181,510 100,082.0 55.1 1977 96,860 74,414.0 76.8 1/ Amounts include C.P.S. receipt as well as District receipt. Sources: Lewis (1979), DNP (1978) WPS/BR-001-4b/3.20.85 - 242 - Table 4.19 Distribution of Situado Fiscal Varying Base Formula in the District Total Situado District Allocation Years Total Population Fiscal Appropriation Based on Formula 1/ % of Total (thousands) (thousands of pesos) (thousands of pesos) 1973 2,280.4 2,293.7 77.978 7% 1974 2,344.6 2,747.0 102.586 7% 1975 2,376.1 4,303.4 146.556 - 7% - 1976 2,440.3 5,377.3 204.069 7% 1977 2,501.0 7,632.1 578.286 7% Alternati7e 1 = Using Only Alternative 2 = Using Only Current Population Equal Shares Base (Thousands of Pesos) % of Total (Thousands of Pesos) % of Total 1973 278.5 12.1% 69.5 3% 1974 338.3 12.3% 83.2 3% 1975 545.4 12.7% 130.4 3% 1976 692.0 12.9% 162.9 3% 1977 920.8 13.1% 213.1 3% 1/ Using 70/30 formula and 1964 census statistics Source: Lewis (1979) WPS/BR-001-4b/3.20.85 - 243 - From the calculations, the District would benefit most from a formula based totally on yearly DANE population estimates. Under such a formula, the District's share of situado fiscal would have increased through the years, 12.1% in 1973, 12.6% in 1975, 13.1% in 1977. The increases in the share reflect the increases in population of Bogota relative to all of Colombia. These increases contrast with the fixed share established under the present formula, 7.0%. Since population is held constant at the 1964 census level, there is no change in the share (%) received by the District, nor for all other subnational governments throughout 1970-79 under the present formula. The receipt of shared sales tax would have similar results. The District's share of sales tax would increase through years by percentages equal to those above, 12.1% in 1973, 12.6% in 1975, and 13.1% in 1977. The fixed share established under the present, 7.6%, is greater than that of the situado fiscal. This difference is due to the fact that sales tax revenue is distributed to only 23 subnational governments. From these findings, it appears that the population base used in the calculations and the proportion of funds allocated according to population and equal shares respectively in the formula (e.g., 70/30, or alternatively, say, 80/20) are of importance in determining the District transfer receipts. To look more systematically at these changes, the findings of Table 4.19 are plotted on a graph (Figure 4.14) for 1975. The "percentage of formula based on population" refers to the 70% of the present formula based solely on population. When varying this percentage from 0 to - 244 - Figure 4.14 Distribution of Situado Fiscal for 1975 to District, D.E. (Using Different Formulas - Thousands of Pesos) 600 Based on Current Population 500 0-'0 0 100 400 Based on Population/73 Actual t 300 Distribution 200 Based on Population/64 100liilf 0 20 40 60 80 100 Percentage of Formula Based on Population DNP (1978) Wodd Bank-26182 WPS/BR-001-4b/3.20.85 - 245 - 100%, the District receives increasing amounts of situado fiscal funds. This increase of District receipt is also tied to the population census used in the calculations. Under the 1973 census and current DANE population estimates, the District receipts are larger than under the 1964 census population figures. Both factors, i.e., distribution formula and population base, play an important role in determining the transfer revenues accruing to the District; moreover, the largeu the share of transfers allocated according to population (rather than in equal shares), the greater is the impact of switching from one population base to another. b. The Gap Between Authorized and Distributed Funds In our earlier analysis, the gap of authorized and distributed amounts under the shared sales tax program was much greater than the gap under the situado fiscal. The difference was explained by the budget channels of the revenues; under the situado fiscal the Ministries of Health and Education are pressured to allocate the full transfer amount; under the shared sales tax program, little pressure can be exerted on the Ministry of Finance. The District gap between authorized and distributed confirms the national findings. (Tables 4.17 and 4.18) The situado fiscal distribution of funds to the District is below the authorized amount (Figure 4.15). The shared sales tax receipt in the District has been consistently (except in 1974) much smaller than the amount authorized (Figure 4.16). Alongside this gap in distributed funds, the shared sales tax receipt is considerably delayed. In July of 1980, it was - 246 - Figure 4.15 Authorzed vs Distributed District Receipt of Situado Fiscal (Thousands of Pesos) 600 500 - 400 - Authorized o 300- 200 --- --- Distibuted 100 '73 '74 '75 '76 '77 Year DNP (1978) World Bank-26183 - 247 - Figure 4.16 Authorized vs Distributed District Receipt of Sales Tax (Thousands of Pesos) 2oo 150 -% 100 C 011 // / // '73 '74 '75 '76 '77 Year DNP (1978) Wodd Bank-26184 WPS/BR-001-4b/3.20.85 - 248 - reported by a District official that the District treasury had not received its transfers since December, 1979. The lateness of the payment, as well as the gap between authorized and distributed amounts, makes the shared sales tax an unreliable source of revenue for the District Administration. The delays of payment of the situado fiscal are not as serious, and will be discussed in Section III of this paper. From the above discussion, it is clear that the national government and the District Administration have different, and often conflicting, perspectives on the role of revenue sharing. The distribution formula's objective of regional redistribution, although consistent with the national policy of balanced regional growth, is perceived by the District as being unfair. The shortfall in the District receipt may be advantageous from the national perspective when national fiscal resources are tight; however, this shortfall jeopardizes the planning and implementation of the District budget. Thus, to discuss the objectives of revenue sharing, clear distinctions must be drawn between the national vis-a-vis District objectives. Under the national objective of regional equity and health and educational service provision, the distribution of situado fiscal and shared sales tax revenues do not totally.achieve this dual objective, nor do the programs completely fail. Unfortunately, lack of data makes impossible a comparison between situado fiscal or shared sales tax program and previous grant programs, and thus rules out an assessment of the impact of revenue sharing on regional equity and public service provision. The District objective of revenue sharing is to promptly receive a share of total transfers that is equal at least to its (growing) share in WPS/BR-001-4b/3.20.85 - 249 - total population. For that reason, the District favors the formula based solely on population. More critical to the District, however, is the shortfall of the national receipt under both the situado fiscal and shared sales tax programs. The consequences of this gap may include budget deficits and inadequate levels of health and educational expenditures. In Section III, we will examine these consequences of revenue sharing in the District. Beer Tax Sharing. The beer tax is the most important source of sales tax revenue in the District Administration, contributing in 1976, 22.3% to total District Adminsitration revenue. As early as 1928, the District received revenue from the taxation of beer. Since those early days, the tax has changed considerably: * The tax, originally levied at a fixed rate, is now an ad valorem tax. * The tax revenue, once under the exclusive domain of the departments, is now shared between the national government, the departments and the District. * The District's receipt, before administered by the Department of Cundinamarca, is currently the responsibility of the District Administration. These reforms are intimately tied to the workings of the beer tax sharing program. The Rate of Tax. The beer tax began as a local District tax during the 1920s. In 1928 (District Ordinance 19/28) the District levied a one centavo tax (Col$ .01) on a bottle of beer (390-400 grams). This rate was successively increased in 1930, 1954, and 1963. The principal objective*of the increased tax rate was to raise additional revenue for the departments and WPS/BR-001-4b/3.20.85 - 250 - the District, because under the fixed rate structure revenues did not increase in pace with income. 1/ The fixed rate schedule was cast aside in 1966 (Decree 1665/66), and the ad valorem rate structure was adopted. The rate ?stablished in 1966 was 60%; this was lowered to 48% in 1971 (Decree 61/71). The 48% rate remains the rate as of 1980, and includes both a national tax (8%) and a departmental tax (40%). In implementing the tax, the rate is applied on the producer price. Each brewery is required to pay the beer tax; and in turn, the tax is included in the wholesale price of the beer and passed on to the consumers. 2/ Collection and Distribution of Revenues. Throughout the years, there has been a legislative volley in gaining control over the collection and distribution of beer taxes. Firstly, the national government authorized the right of the departments to tax beer (Law 19/28). The overriding legislative power of the Congress remains intact, reminding the sub-national governments "that which has been given, can also be taken away". Nonetheless, the national gc,;rnment did not exercise its control over beer taxation from 1930-1960. During the 1960s, the nature of the tax changed significantly. First, the departmental tax sharing arrangement, as of 1966, has distributed revenues based on consumption of beer (see Figdre 4.17). The national government assigns consumption percentages to each of the departments and the District. The breweries throughout Colombia send the set percentage of 1/ This low revenue buoyance of a fixed rate tax is well documented in public finance literature. 2/ This is not to say that the consumer bears the entire incidence. As learned in elementary microeconomics, both producer and consumer may bear the burden of the tax. - 251 - Figure 4.17 Collection and Distribution of Beer Tax Revenue Beer Producer Beer Sold Pays National ito Consumers & Local Tax Includes Tax in Price Per Bottle District/ National Beer Tax Departmental of 8% Directly Tax of 40% Remitted to Local Distributed Based S.S.S on Consumption Other 23 Departments District Treasury Receipt of Remaining Receipt of 20-25% 70-75% of Total of Total Source: Based on Interviews with District Administration officials World Bank-26185 WPS/BR-001-4b/3.20.85 - 252 - collected revenue to each departamental and District government. The purpose of distributing revenue based on consumption is to equalize the incidence of the beer tax. Furthermore, the 1960s reforms included the 8% national ad valorem tax -in 1966. The revenues from the 8% tax are earmarked to the S.S.S. located in the same jurisdiction as the brewery (Decree 161/71). In the case of Bogota, the revenues flow directly from the producer to the district S.S.B. The national beer tax distributes revenues based on production and does not have a redistributive mechanism. Additionally, the national beer tax revenues do not flow through the national budget as do other national taxes. However, the Ministry of Health includes the revenues in their planning budgets for S.S.S. At the same time that Congress established the national beer tax, it also created several mechanisms to control the pricing of beer, the collection of beer taxes, and the distribution of beer tax revenues. In Decree 2338/66, reporting requirements were established. The private producer has to submit information on the production costs of the beer to support the brewery price. In 1969, these data requirements were further elaborated to include: (1) level of production and total costs; (2) location of distribution; (3) business receiving goods; (4) roads used in d ,.tribution; and (5) collection of the tax. The information must be sent to the respective treasuries every 15 days. The purpose of the reporting requirements is to make the beer industry accountable to the natin-zal government. Under the requirements, the national government monitors the collection of beer taxes. Through the Administration of National Taxes of the Ministry of Finance, the national WPS/BR-001-4b/3.20.85 - 253 - government exercises the control to ascertain the correct payment of taxes by the producers in terms of amount and timing. The pricing and reporting requirements are enforced by the National Superintendent on Prices. Non- compliance with the national regulations leads to loss of the beer production license for a period of five years. This national information is most important, however, for its use by the National Tax Administration when determining the beer consumption by departments and District. Since the 40% departmental share of beer tax revenues is distributed based on "consumption of beer", the department and District governments report consumption levels in their area to the National Tax Administration. Comparing these reported levels with the distribution data, the National Tax Administration assigns consumption percentages to each department and the District (Table 4.20). The producers are then instructed to send the 40% tax share directly to the treasuries of the department, District, and territories in proportion to their corresponding consumption levels. The District benefits greatly from the structure of distribution. With a large urban population, the District receives almost 25% of total beer tax revenues. And since 1970, the District receipt is forwarded directly to the treasury of the District Administration, by-passing the Department of Cundinamarca's treasury altogether. Once the District treasury receives the beer tax receipt, there is no earmarking of the revenue within the District. 1/ However, it is generally understood that departmental and District beer tax revenues are to support education (Decree 3258/68). 1/ The District Administration works under a consolidated account system, "Caja de Unidad", with no revenues being tied to specific expenditures. WPS/BR-001-4b/3.20.85 - 254 - Table 4.20 Beer Consumption by Departments and District - 1978 Antioquia 10.4 Atlantico 5.4 Bogota 22.1 Bolivar 2.0 Boyaca 7.8 Caldas 2.8 Cauca .9 Cesar 1.3 Cordoba 1.2 - Cundinamarca 11.6 Choco .8 Cuajira .7 Huila 1.6 Magdalena 1.5 Meta 2.5 Narino 1.1 N. de Santender 3.8 Quindio 1.5 Risaralda 1.8 Santender 6.4 Sucre .5 Tolima 5.0 Valle 7.2 100.0% Source: Controlari de Bogota (1979) WPS/BR-001-4b/3.20.85 - 255 - National beer tax revenues of course directly support health services as mentioned above. The reforms of the beer tax sharing program have been considerable. Its revenue performance can largely be attributed to these changes. But before assessing the merits of the reforms, let us review two close relatives of the beer tax, that of the tobacco tax and liquor tax. Tobacco Tax Sharing The national tobacco tax 1/ plays a much lesser role in District financing than the beer tax, contributing on average only 1.4% to District Administration revenue in recent years. The weak performance of the tax is mainly attributable to its underlying structure. The Rate of Tax. The tobacco tax was established in 1944 (Departmental Ordinance 24144) as a departmental tax with the then-municipality of Bogota receiving 20% of the Department of Cundinamarca's tobacco tax revenues. The tax followed a fixed rate schedule per pack of cigarettes. The producer was responsible to pay the tax to the department treasury. In 1966 the National Congress established an additional national tobacco tax at an ad valorem rate of 8%; the rate was increased to 10% in 1971. Distribution of Revenue. The national tobacco tax revenue is distributed according to consumption of tobacco. However, within the Department of Cundinamarca, the District's participation is limited to 20% of the revenues collected-both in the District and the department. The District, contributing a large percentage of total departmental tobacco tax revenues because of a relatively large share of population in the District, feels this small 20% 1/ The national tobacco tax should not be confused with the foreign cigarette tax which is a local, non-shared tax in the District. WPS/BR-001-4b/3.20.85 - 256 - share is unfair. The District revenue share paid by the producer is sent directly to the District Treasury. Liquor Tax Sharing. It may seem strange to discuss liquor tax sharing as the District Administration receives no revenues from this source. 1/ This "non- receipt" of liquor tax revenues is an especially sore point in the District. To understand why the District Administration receives no liquor tax revenues, it is necessary to understand the structure of the liquor industry in Colombia. For better or for worse, the liquor industry in Colombia is a departmental public monopoly with each department having its own distillery. The potential costs of this monopoly are well known -- inefficient production, low quality of liquor products, high liquor prices, and high department liquor taxes (Musgrave, 1971). The liquor enterprise of each department is comprised of producers and distributors who establish the liquor prices and pay the tax on the liquors. The tax is levied on the wholesale price of the liquor which includes the producer and distribution costs. Our understanding of the tax on liquor products is confounded by the fact that there are a number of departmental liquor and wine taxes on: liquor from the local producer, liquor from other departments, foreign liquor and wine, and domestic wine. All of these taxes accrue only to the departments. Over and above these taxes, there is the national sales tax on liquor which flows to a special national liquor account. The revenues of this account are included in the national budget. Similar to the situado fiscal appropriation 1/ To be distinguished from the District Administration is the decentralized agency of S.S.S., which receives a minimal contribution of the national sales tax on liquor. This will be discussed later. WPS/BR-001-4b/3.20.85 - 257 - for health, the national liquor sales tax revenues are earmarked for the S.S.S. and are budgeted under "transfers to the S.S.S." in the Ministry of Health's expenditure budget. The District Administration's participation in liquor taxation is zero for one simple reason: it does not have a distillery. Prior to 1970, the District liquor tax revenue was under the authority of the Department of Cundinamarca. It received a 9.7% share of all the Department's revenue on the various department liquor taxes. Since 1970, the District receipt of shared taxes such as shared beer tax, has gone directly to the District treasury, by- passing the Department of Cundinamarca. At the same time, the District lost its departmental receipt of liquor tax revenues. The District population without a distillery, continues to buy liquors from the distillery of Cundinamarca, yet receives no compensation for paying the tax on these liquors. Revenue Performance of Beer, Tobacco, and Liquor Taxes. The taxation of vice has long been a legitimate revenue source in Colombia. The "do-good" objective of vice taxation attempts to limit the consumption of "vice" by raising the price of the good. And, as with the beer tax, the revenue raised goes to support "good" causes, such as health expenditures. The soundness of this traditional argument may be questioned. In doing so, it is necessary to look at how much "good" does it do (e.g., revenue performance). Revenue Performance. The beer tax clearly outshines the tobacco and liquor taxes in contributing revenues to the District. As illustrated earlier, the beer tax once contributing around 15% to District Administration revenue, now contributes over 40%. The real annual growth rate of beer tax during the WPS/BR-001-4b/3.20.85 - 258 - period 1960-79 was 9.2%. This compares with the weak performance of the tobacco tax and the liquor tax. When comparing the District revenue receipt to the total national revenues generated from the three revenue sources, the same conclusion can be made. The District receipt of total national beer tax revenues for 1978 was around 22%; for tobacco tax revenues the local share was 2%; for liquor taxes revenues, it is 0% (Table 4.21). Given the importance of the District beer tax receipt let us look more closely at the reforms and its effects on revenue performance. A measure of the impact of the beer tax reforms is the revenue arrears on collecting beer taxes. It could be hypothesized that with the increased national controls of 1966, the arrears should be smaller. When examining the revenue arrears from 1969 to 1977 (Table 4.22), we can see that arrears have increased, not decreased in current prices. In 1970 the amount increased substantially, and has now levelled at around 50,000,000 pesos per year (1977). The revenue arrears from beer tax have been decreasing in percentage terms compared to total revenue arrears of the District Administration (Table 4.23). Revenue arrears in the same years for property tax and the industry and commerce tax are a larger percentage of their respective tax collections, on average, 20.1% and 8.6%, with beer tax at 4.7%. Even more impressive, the arrears for "previous years" under the consumption tax have declined from 4.3% (1961) to 0.1% (1977). The reforms of the 1966 legislation and its reporting requirements apparently brought a decline in arrears. However, the 1970 separation of Department and District beer tax revenue WPS/BR-001-4b/3.20.85 - 259 - Table 4.21 District Receipt of Beer and Tobacco Tax Revenues as (%) of National (1973-1977) Beer Tobacco 1973 25.1% 4.6% 1974 24.3% 3.0% 1975 24.5% 2.9% 1976 22.1% 2.6% 1977 21.3% 2.6% Source: Controlari Distrital (1973)-(1977) Table 4.22 Beer Tax Arrears 1969-1977 (Current Prices) 1969 23.935.2 1970 51.416.2 1971 60.519.5 1972 48.452.0 1973 48.198.1 1974 51.320.1 1975 52.062.0 1976 44.216.0 1977 46.833.9 Source: Controlari Distritat (1973)-(1977) WPS/BR-001-4b/3.20.85 - 260 - Table 4.23 Structure of Revenue Arrears (1969-1977) (Percentage Distribution) Industry and Beer Tax Property Tax Commerce Tax 1969 2.8 17.8 8.8 1970 5.5 18.6 7.4 1971 6.1 21.5 6.2 1972 6.4 21.7 7.8 1973 5.9 18.9 7.9 1974 4.7 18.9 -<.2 1975 3.2 18.3 4.9 1976 3.0 22.1 -4.1 1977 3.7 21.0 7.8 Source: Lewis (1979) WPS/BR-001-4b/3.20.85 - 261 - seemed to have an opposite effect, given the increase in arrears of beer tax revenue since 1970. It must be noted that the arrears percentage may be associated with the amount of total revenue, rather than result of the collection performance of the taxes.- To test the hypothesis, the following equation was regressed: Ln (arrearsl) = a + b Ln (Tot Rev1) for the three District taxes, the property, industry/commerce, and beer consumption. The results of the regression (Table 4.24) show: 1) beer tax arrears increased less than in proportion with total beer tax revenue (+.835992). 2) induustry/commerce arrears increase at a lesser rate than its total tax revenue (+.778837). Its rate is the lowest of the three taxes. 3) property tax arrears increase at a greater rate than total property tax revenues (+1.28049). Clearly, the collection of beer and industry/commerce taxes has improved when compared to property tax collection. To measure the magnitude of this change, the ratio of revenue to arrears is calculated for each of the three taxes (Table 4.25). The revenue-arrear ratios for the beer tax is very high and increased substantially over time (1960-77) from 4.324 (1960) to 10.090 (1976), pointing to the success of the reporting reforms. This contrasts with the industry/commerce tax which increased from 1.647 (1960) to 4.672 (1976). As expected from the regression findings, the revenue-arrear ratio of property tax declines from 2.387 (1960) to 1.383 (1977). The evidence of the sucessful beer tax performance contrasts sharply with that of the tobacco and liquor performances. The limited 20% participation of the District under the tobacco tax, and the virtual exclusion WPS/BR-001-4b/3.20.85 - 262 - Table 4.24 Regression of Arrears to Total Revenue Ln (Arrearsi) = ao + al Ln (Tot Revi) 1960-1977 Tax Base al R2 (T-STAT) Beer .<36 .9609 (19.836) Industry and .779 .9011 Commerce (12.0734) - Property 1.281 .9511 (17.769) Source: Lewis (1979) WPS/BR-001-4b/3.20.85 - 263 - Table 4.25 Revenue Arrear Ratio of Property, Industry, and Beer Tax Revenues (1960-1977) Pro IC Beer 1960 2.387 1.647 1961 3.829 2.014 4.825 1962 3.529 2.495 8.676 1963 3.521 2.320 6.655 1964 3.358 2.421 6.759 1965 2.896 2.227 7.223 1966 2.437 1.425 7.255 1967 2.722 1.408 9.760 1968 1.725 1.140 9.739 1969 2.749 1.541 12.594 1970 2.587 2.030 8.663 1971 2.018 2.347 8.579 1972 2.706 2.344 9.834 1973 3.197 2.576 8.874 1974 2.067 1.729 8.613 1975 1.664 3.293 9.392 1976 1.403 4.672 10.090 1977 1.383 3.272 9.966 Source: Lewis (1979) WPS/BR-001-4b/3.20.85 - 264 - of the District in the liquor tax explain why the District does not benefit more than se tax sharing arrangements. Additional Tax Sharing Arrangements. Two additional tax sharing arrangements are the registration tax (or property transfer tax) and the Ecopetrol gasoline consumption tax. 1/ The contribution of these taxes to District Administration reveune is minimal as illustrated in Table 4.26. The registration tax on average contributed 2% to District Adminstration revenues between 1964 and 1977. The contribution of Ecopetrol revenues to District Administration revenu was .07%. To explain the minor role of these two taxes, several factors must be examined: - the tax rate - the District's share of national and/or departmental distribution - the market determinants of the tax revenue, namely value of property transactions and gasoline consumption. A careful study of the specifics of the registration and gasoline taxes points to the importance of the three factors. Registration Tax Sharing. The tax on the transfer of property is more commonly known as the registration tax. As early as 1896 the national government authorized taxing property and other public/private transactions. Since that time various reforms have increased the tax rate and established both a national and departmental tax. 1/ The entertainment tax revenue is also shared: All public events are taxed at a 30% ad valorem rate, with 20% of this tax flowing to the District treasury, and 10% earmarked for the National Sports Administration. No figures are available on the District's share of grants received by the National Sports Administration. WPS/BR-001-4b/3.20.85 - 265 - Table 4.26 District Administration Revenues from Registration Tax and Ecopetrol 1964-77 (thousands of current pesos) Year Registration Tax Ecopetrol Receipt Total Sales Tax Total Revenue Revenue Revenue 1964 9.194.0 6.566.0 77.632.0 223.546.0 1965 7.597.0 6.989.0 77.794.0 247.997.0 1966 5.880.0 6.842.0 81.896.0 249.941.0 1967 8.895.0 7.857.0 100.843.0 417.520.0 1968 9.555.0 8.521.0 100.977.0 471.730.0 1969 16.331.0 9.183.0 139.852.0 427.236.0 1970 20.635.0 9.248.0 217.248.0 597.814.0 1971 23.657.0 10.780.6 305.206.2 748.212.9 1972 24.372.0 10.019.9 326.072.5 781.579.1 1973 34.241.0 10.580.8 370.425.5 1.001.837.0 1974 48.496.0 12.416.2 476.496.5 1.236.705.2 1975 42.187.0 14.825.9 679.597.9 1.984.449.2 1976 55.053 14.552.2 663.873.6 1.997.573.8 1977 70.307 15.056.0 809.653.0 1.974.238.0 Source: Lewis (1979) WPS/BR-001-4b/3.20.85 - 266 - a) The Tax Rate The Departmental registration tax was first established in 1913 as a fixed rate tax at Col$10.00 for each transaction (Law 51/13). In 1920 this fixed rate tax was changed to an ad valorem tax of Col$0.20 on each Col$100.90 pesos of property value. The ad valorem rate has been increased several times since that period (Decree 2838/54). The most recent rate change effective of 1963 (Law 24/63) established the rate at Col$1.00 for every Col$100.00 pesos of property value. Accompanying the increase in tax rate, various fees have been placed on property transactions, over and above the basic tax rate, such as: The fixed fees for additional service, unlike the basic ad valorem rate, is an inelastic revenue source, and the base ad valorem rate is a more important source of tax revenue than that of the service fees. The national registration tax must be distinguished from the departmental tax. It originated as an "emergency tax" to raise revenues for the national government for six months. Under Law 78/30, fifty percent surcharge was levied on the department's registration tax revenue to be channeled to the National Ministry of Finance. In 1969 the surcharge rate was revised to include only 10% of the registration tax revenue. No national fees are charged. The registration tax is collected by notaries or registrars who forward the tax revenues to respective treasuries. The distribution formulas used in allocating the tax revenues under the departmental tax and national surcharge are extremely different. b) The Distribution of Registration Tax Revenue. The District's receipt of registration tax revenue from the Department of Cundinamarca and WPS/BR-001-4b/3.20.85 - 267 - the national government is determined by the differing distribution guidelines. To a large degree, these guidelines be they for the national or departmental tax, reflect the inherent interest of the specific governmental unit. The national registration tax revenue is returned to the departments and District (Law 33/68). The distribution of the funds is authorized by the National Budget Office in accordance with guidelines set by the National Budget Office in accordance with guidelines set by the National Controller of the Republic. The distribution of revenues is based on the percentage contribution of each department and the District. The District receives the same percentage of funds that it contributes; the District's receipt is not tied to that of the Department of Cundinamarca. Responding to the erosion of tax authority, the Departments carefully guard their revenues from registration taxes. This is clearly illustrated in the case of the Department of Cundinamarca. The District contributes some 75% to total registration tax revenue in the Department; it receives only 30% of these revenues from the Department. The 30% distribution to the District has been cited as a "great injustice" (Chamber of Commerce, 1978). The District receipt, however, is channeled directlyto the District treasury, by-passing the Department's treasury. The administration of the registration tax works as follows. The notaries and registrars collect the registration tax at the time of the property transfer, and send 30% of the tax to the District Treasury. Both the collection office of the Department of WPS/BR-001-4b/3.20.85 - 268 - Cundinamarca and the national government monitor the District transfers. Although the Department of Cundinamarca receives 70% of the departmental registration tax revenues, the importance of the registration tax contribution to the Department is.minimal. 1/ The only earmarking of funds, as noted in the legislation (Law 24/63) is that any additional revenues due to the increased tax rates of 1963 are to be targeted to social assistance programs. c) Other Determinants of Registration Tax Revenues. One important determinant of the amount of revenue received by the registration tax is the property value of the transfer, and the national tax calculated as a percentage of this initial rate, the value of the property transaction plays an important role in determining the registration tax revenues. Table 4.27 lists the value of property sales in the District from 1964-75. The annual growth rate during 1964-75, of property sales value in current prices was 19.5% during the entire period, and 7.0% in constant prices. The growth rate of registration tax revenue was a little higher at 20.7% in current prices and 8.0% in constant prices during the same period. Table 4.27 also presents the effective rate of registration tax revenue to value of total property sales 2/ during the years 1964-75. There is no clear pattern of the rate changes with the 1/ The most important revenue sources for the Department in 1977 were: liquor tax at 41%, beer tax at 26%, and cigarette tax at 25% (Caceres Bolanos, 1978, p. 131). 2/ Sales value is equated to market value for purposes of our discussion. WPS/BR-001-4b/3.20.85 - 269 - Table 4.27 Property Value of Sales and Registration Tax Revenue (thousands of Pesos) (1964-75) Value of Sales Registration Ratio Tax Revenue 1964 1.046.9 9.194 .00878 1965 1.174.8 7.597 .00646- 1966 1.620.0 5.880 .00362 1967 1.777.2 8.895 .00500 1968 2.070.1 9.555 .00462 1969 2.315.9 16.331 .00705 1970 2.366.441 20.635 .00872 1971 6.741.136 23.657 .00351 1972 5.411.177 24.372 .00450 1973 3.958.76 34.241 .00865 1974 7.726.516 48.496 .00628 1975 7.962.532 42.187 .00523 Source: Lewis (1979) and DANE (1974-1975) WPS/BR-001-4b/3.20.85 - 270 - average rate equal to .00603. While the registration tax revenue is a very small percent of total property value, when testing the relationship of registrtion tax revenue to sales value of property, there is a strong correlation (0.89) between the two variables. Given this high correlation, it can be surmised that sales value of property is another important determinant of revenues generated by the registration tax. Ecopetrol Gasoline Tax. Authorized under Law 15/59, the Ecopetrol gasoline tax, levied at the production level, requires the state oil enterprise Ecopetrol, to pay to department and District a fee of Col$0.06 for each gallon of gasoline sold. The revenues received from the tax are distributed based on the gasoline consumption in the departments and District. The District revenues from the Ecopetrol tax are very small: on average they amounted to 1.96% of District Administration revenues for 1964-75. The annual average growth rates of the revenues for the period 1960-77 were +7.1% in current prices, and -5.0% in constant 1976 prices. This poor revenue performance reflects the low and fixed rate of tax, the District's share of revenue, and the market determinants influencing the revenue distribution. a) The Tax Rate The rate of taxation, Col$0.06 per gallon of gasoline, was established in the original legislation. This rate has not changed in twenty years. A fixed rate tax, i.e., a fee on a specific quantity, is an inelastic revenue source. Gasoline tax revenues do not proportionately increase with increases in the relative price of gasoline. Moreover, the tax rate has not been changed to reflect WPS/BR-001-4b/3.20.85 - 271 - inflation. The erosion of the effective gasoline tax rate results from this fixed rate of tax. When examining the other types of gasoline taxes, such as the local fixed rate tax of Col$0.04 per gallon, 1/ the national ad valorem tax of 10%, and the highway Trust Fund Tax, it is clear that there is considerable inconsistency in the gasoline tax structure. The two taxes which directly contribute to the District Administration treasury are the Col$0.04 per gallon distribution tax and the Ecopetrol tax, both being fixed rate taxes with an extremely poor revenue performance (Musgrave, 1971). b) The Distribution of Ecopetrol Revenues The distribution of Ecopetrol tax revenue is based on consumption of ganoline. Ecopetrol is required to forward the tax revenue to the respective treasuries; in the case of the District, Ecopetrol forwards the tax revenue to the District Administration treasury, completely by-passing the Department of Cundinamarca. Consumption is calculated by Ecopetrol, based on reports submitted by the departments and District on the consumption of gasoline in each area. Table 4.28 presents gasoline consumption in the different departments and District for 1964-74. Throughout this period, the District has consumed approximately 25% of all gasoline in Colombia. 1/ There seems to be some confusion as to whether the gasoline tax of Col$0.04 is a shared tax. Only in the 1980 budget is it considered to be one. The tax is paid by the distributors, ConMobil and Texaco, based on the distribution of gasoline. For purposes of this paper, it will be considered a direct tax paid to the District Administration. Table 4.28 Gasoline Consumption (1964-74) by Departments/District 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 Antioquia 62390 64864 69163 72126 75051 82344 90453 94466 98729 1055776 113137 Atlantico 26203 28485 31135 30592 31939 34604 38109 42105 56494 59477 64948 Bolivar 20594 21569 22906 17748 16752 18142 20215 23060 23567 25159 27729 Boyaca 12971 13939 15107 16119 17965 19495 20723 21169 21376 22970 23424 Caldas 29657 32134 30287 16079 15473 16574 17816 18347 18924 20255 21302 Cauca 6478 6924 7258 6384 6412 7067 7143 7709 8953 11698, 11930 Cordoba 7385 8290 8052 8655 8808 10018 11326 11954 13500 15609 16567 Cundinamarca 24881 25383 25302 23467 24737 87188 29333 29923 34780 36028 36692 Bogota, D.E. 109441 116479 126225 131420 143683 155011 167606 176927 184415 197201 221084 Choco 363 251 243 187 497 566 659 330 485 602 659 Guajira 480 720 339 1082 1366 1563 1468 1937 2365 2616 3965 Huila 7502 7881 8184 8402 9150 9918 10584 11067 12101 13270 14292 Magdalena 13939 16301 17893 17515 8308 9164 9623 10221 10941 11113 12040 Meta 4461 5074 6390 7006 7880 7934 8034 8112 9495 10424 10576 Narino 7926 9178 9916 10970 12569 12680 13728 12461 10366 10985 13569 N. de Santander 6592 8195 9718 9039 8382 10307 9123 8927 8689 7699 7593 Santander 24812 26937 27677 28635 30438 31181 52149 38508 39814 43594 47703 Tolima 20608 22042 23838 23195 26169 27553 30688 31449 30440 3368 34435 Valle 63800 67172 73107 74567 81555 83602 92500 97999 101681 110615 115367 Cesar 244 9843 10885 11490 10565 10056 11641 13475 Quindio 4196 8046 8570 9251 9879 10961 11335 12811 12881 Risaralda 8422 10443 11096 12199 12685 13520 15126 15996 Sucre 3468 4665 4926 5430 6397 7494 7774 8564 T. Nacionales 1323 2182 2402 2626 2972 3505 3661 4998 8919 9962 8960 Total 451806 484000 519348 525994 563627 604574 673939 692296 738438 795573 856890 Source: ECOPETROL (1975). WPS/BR-001-4b/3.20.85 - 273 - The annual average growth rate of gasoline consumption in 1964-74 has been 7.0% in the District, keeping -ace with the national rate of 6.5%. Information is not available on the total Ecopetrol revenues, making it impossible to calculate tax ratios of District receipt to total tax revenues. However, a simple correlation of District gasoline consumption and Ecopetrol revenue receipts produces a very high correlation coefficient of .9774, suggesting that gasoline consumption is indeed the criterion used to calculate the distribution of Ecopetrol tax revenues. c) Other Determinants of Ecopetrol Tax Revenues The consumption of gasoline as measured by the number of gallons consumed, results from vrious market forces as well as government regulations (i.e., gasoline subsidy program) and other taxes (i.e., 10% ad valorem tax). The gasoline industry is carefully regulated by the Colombian government. Any discussion of changes in Ecopetrol gasoline tax will likely take into consideration the overall impact of this tax and others on the gasoline industry. Grants. So far the discussion in this section has focused on the shared taxes of the District of Bogota, National grants in housing, education, health, and transportation have been reformed to a much lesser degree, and play a lesser role in the financing of District expenditures. As seen earlier national grants on average contributed 3% to total consolidated District revenues compared to shared tax contribution of 8.7% during the period 1966-77 (Table 1, Section I). Furthermore, national grants have grown in real terms at an annual rate only of 9.8% (1961-1976), compared to shared taxes real growth of 15.8%. WPS/BR-001-4b/3.20.85 - 274 - One critical difference between grants and shared taxes is that grant revenues do not flow directly to the District treasury. Rather, national grants are earmarked for specific expenditures and are channeled to the decentralized agencies, outside the District Administration budgetary process. All of the grants, with the exception of the national bus subsidy received by EDTU, are lump-sum grants received for construction or maintenance of facilities, and include transfers for schools, health facilities, and housing. Due to the specific nature of the trants and the lumpiness of the grant flows into the District, grant administration does not involve the complexities of the shared tax programs. The disbursement of grant funds in the District is relatively straightforward. EDTU Bus Subsidy. The national government subsidizes the bus industry through block grants issued by the Transportation Finance Corporation (Corporacion de Finanzas Transportivas, C.F.T.). The C.F.T., an autonomous credit agency sponsored by the national government, distributes grant revenues to bus companies, both private and public, based on the number of buses and age distribution of buses. EDTU (Empress Distrital de Transported Urbanos), the public bus company of the District, received only a minimal share, approximately 5% of the total given to District operators subsidy. 1/ However, this contribution represented 23.6% of the total budget of EDTU in 1977 (Table 4.29). To assess the importance of this subsidy, both its past performance and the factors which determine its present amount must be considered. 1/ This figure is an approximation based number of buses in service (Table 11-22). Table 4.29 Empresa Distrital de Transporte Urbano (EDTU) Revenues, 1961-76 (thousand Pesos - current prices) 1 2 3 4 5 6 7 8 1961 1962 1963 1964 1965 1966 1967 1968 1. Total Current Revenues 7293.0 7026.0 12588.0 13742.0 11800.0 10543.0 18739.0 25308.0 2. Operation 7218.0 6943.0 12492.0 13652.0 11700.0 10405.0 12640.0 19892.0 3. Other Revenues 75.0 83.0 96.0 90.0 100.0 138.0 99.0 218.0 4. Nat'l. Gov't. Subsidy 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5. Trans. from District 0.0 0.0 0.0 0.0 0.0 0.0 6000.0 5198.0 6. Total Capital Revenues 0.0 0.0 0.0 0.0 0.0 3816.0 12335.0 1888.0 7. Total Revenues 7293.0 7026.0 12588.0 13742.0 11800.0 14359.0 31074.0 27196.0 1 2 3 4 5 6 7 8 1969 1970 1971 1972 1973 1974 1975 1976 1. Total Current Revenues 31886.0 45082.0 42435.0 41091.0 45964.0 42506.0 91141.0 97923.0 2. Operation 30400.0 34971.0 41611.0 37921.0 33864.0 33428.0 48137.0 52064.0 3. Other Revenues 486.0 2571.0 824.0 609.0 3471.0 668.0 7732.0 1176.0 4. Nat'l. Gov't. Subsidy 0.0 5642.0 0.0 2561.0 8629.0 8410.0 8143.0 26425.0 5. Trahs. from District 1000.0 1898.0 0.0 0.0 0.0 0.0 27129.0 18258.0 6. Total Capital Revenues 9189.0 35546.0 118509.0 11054.0 45240.0 136974.0 259155.0 32492.0 7. Total Revenues 41075.0 80628.0 160944.0 52145.0 91204.0 179480.0 350296.0 130415.0 LA Lewis (1979). WPS/BR-001-4b/3.20.85 - 276 - a. Past Performance The bus subsidy began in 1959 when the national government established a subsidy of 700 pesos per month for each vehicle in service (Law 15/59). During 1959-68 these subsidies as well as the price of fares, were raised because of increasing operational costs. Year Fare Subsidy (Monthly) Per Bus 1959 $.25 $ 700.00 1962 $.25 800.00 1966 $.30 1,500.00 The original objective of the 1960s subsidy program was to maintain low fares. In 1975 the subsidy program was re-directed. The prime objective of the new system is to stimulate investment in bus transportation. Subsidies are granted on the basis of the age of the vehicle with newer buses receiving a larger subsidy than the older buses. The following is the differential schedule used (Cifuentes, 1979). Model Monthly Subsidy (pesos) Per Bus 1959 earlier $ 750 1960-64 2,500 1965-69 4,000 1970-73 5,000 1974 and on 17,500 WPS/BR-001-4b/3.20.85 - 277 - With this built-in bias toward newer buses, the system encourages bus companies to purchase new vehicles. As of 1977. the program was extended to cover Minibuses, but eliminated its subsidies to gasoline-powered vehicles registered after February 11, 1977. Also, it appears that trolley buses do not receive subsidies (Linn, 1976a). To assess the performance of the subsidy program, information is needed on the receipts by EDTU and the private bus companies. Unfortunately, there is no detailed data on private bus companies' receipts. The 1970 EDTU receipt is found in the Informe Financiero 1970 (Table 4.29). In the years 1972-1977 EDTU receipts are well documented in EDTU data. The EDTU receipt in 1975 was pesos 8,143; in 1976 it was pesos 26,425. Clearly the age distribution formula, rather than allocation based on the number of buses, favored EDTU to the extent possible. Let us look at the two factors, number of buses and age distribution of buses, which determines the present performance of the subsidy. b. Factors Determining EDTU Subsidy EDTU received a larger amount of funds under the new subsidy program for two reasons: firstly, EDTU has only 2-5% of total District buses in service; and secondly, EDTU has a high percentage of new buses compared to old buses. The number of EDTU buses in service has grown from-104 in 1964 to 233 in 1977, representing a growth of 13.05%. However, when comparing this EDTU number of buses to total District buses, EDTU has an average 4.29% of District buses from 1964 to 1972 (Table 4.30). This percentage has remained unchanged with EDTU percentage of District buses in Service at 4.44% in 1977. The second factor which must be considered is the age distribution of buses. Table 4.30 presents the age of vehicles between the total private and WPS/BR-001-4b/3.20.85 - 278 - Table 4.30 District Buses in Service (1964-72) EDTU EDTU Total Total % 1964 104 2196 4.74 1965 85 2263 3.76 1966 77 2490 3.09 - 1967 77 2679 2.87 1968 120 2941 4.08 1969 165- 3165 5.21 1970 192 3436 5.59 1971 195 3893 5.01 1972 184 4350 4.23 1977 233 5247 4.44 Source: Cifuentes (1979) WPS/BR-001-4b/3.20.85 - 279 - public bus companies in the District in 1977. EDTU owns new buses purchased in 1968 and 1975. In grouping the ages of these vehicles into year periods of 1955-1962, 1963-70, and 1971-77 period, (Table 4.31) EDTU has an extremely high concentration of buses in the 1971-77 period (70%). This contrasts with the District total number of buses built during 1971-77 of 44.9%. Although in 1973 EDTU made considerable invesstments in new buses, this did not lead to an efficient transit operation. (DADP, 1980). Maintenance costs have increased sizeably. No new investment in buses has taken place since 1973. Although the national subsidies received by EDTU have increased, EDTU has become less financially viable. This gloomy financial picture of EDTU leads to greater pressure to dissolve the public transit company, a proposition that has been discussed for many years. CVP Construction Grants. The Caja de Vivienda Popular (C.V.P.) is an autonomous local agency responsible for public housing and other community facilities in the District. Established in 1942, the activities of C.V.P. include: - constructing public housing, health facilities, schools and markets; providing home financing to low-income families; - stimulating private investment in housing and coordinating all housing activities in the District. Transfers from the national government, primarily the Ministry of Education and the national decentralized housing agency, I.C.T., (Instituto de Credito Territorial) support C.V.P. activities. Table 4.32 presents the national government's revenue contribution to C.V.P. During the years 1960-77 the WPS/BR-001-4b/3.20.85 - 280 - Table 4.31 Distribution of Buses in District 1977 Year of Age Distribution Age Distribution Purchase of All Buses of EDTU Buses 1955 25 1956 27 1957 1 1958 7 1959 40 1960 236 - 1961 299 1962 280 1963 111 1964 9 1965 428 1966 122 1967 206 1968 581 70 1969 159 1970 358 1971 407 1972 93 1973 203 1974 174 1975 797 163 1976 459 1977 225 Total Number 5,247 233 of Buses Source: Cifuentes (1979) WPS/BR-001-4b/3.20.85 - 281 - contribution on average represented 10.5% of C.V.P. total revenue receipts. The national transfers in the form of investment grants were earmarked for specific construction projects. Of the total national transfers in 1961-72, the largest share, 83%, is from the I.C.T. and is targeted to joint ICT-CVP housing projects. The I.C.T. is an autonomous agency attached to the Ministry of Economic Development and provides housing loans and grants to low-income families. Grants are issued to specific public housing projects, and usually in combination with loan financing arrangements. The grant transfers to the C.V.P. have been quite minimal. Joint ICT-CVP projects have focused on the "normas minimas" housing settlement, a program to improve sub-normal settlement. ICT investment in CVP projects comprises grant and loan financing. With the information available, (Stevenson, 1979) it is impossible to ascertain the amounts of grants as against loan. For that reason, the following is a general discussion of ICT-CVP projects. Of the many joint activities of the ICT and CVP, six projects have been most important: El Araujo, Las Garces Naves, Las Guacamayas, Las Lomas, La Manuelita, and El Alfonso Lopez P. Each of these projects has been co- financed by ICT and CVP. The first joint project, that of El Araujo, began as a slum improvement project on private initiative. ICT and CVP joined forces in 1964, and supplemented the project's private revenues until 1966. The second project of Las Garces Naves differs from El Araujo. It began in 1970 completely-under ICT initiative as a planned community with infrastructure provided (i.e., water, electricity, sewerage). ICT paid for 91% of the construction costs of Las Carces Naves. Table 4.32 Caja de Vivienda Popular (CVP) Revenues, 1961-76 (thousand Pesos - Current Prices) 1 2 3 4 5 6 7 8 1961 1962 1963 1964 1965 1966 1967 1968 I Total Current Revenues 7421.0 7738.0 8716.0 17709.0 15407.0 21923.0 24027.0 19051.0 2 General Revenues 5681.0 3376.0 4411.0 4534.0 4054.0 10319.0 6244.0 4105.0 3 Amortization Payment 2687.0 903.0 1106.0 1146.0 1124.0 1586.0 957.0 1031.0 4 Other 2994.0 2473.0 3305.0 3388.0 2930.0 8733.0 5287.0 3074.0 5 Transfers 1740.0 4362.0 4305.0 13175.0 11353.0 11604.0 17783.0 14946.0 6 From National Govt 0.0 100.0 1820.0 3760.0 1161.0 400.0 355.0 400.0 7 From District Govt 1740.0 4262.0 2485.0 9415.0 9938.0 10859.0 17399.0 14333.0 8 From Dist. Agencies 0.0 0.0 0.0 0.0 254.0 345.0 29.0 213.0 9 From EAAB 0.0 0.0 0.0 0.0 114.0 159.0 0.0 0.0 10 From EEEB 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 11 From LOTERIA 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 12 From CPS 0.0 0.0 0.0 0.0 140.0 186.0 29.0 213.0 13 Total Capital Revenues 240.0 0.0 0.0 0.0 0.0 500.0 0.0 2578.0 14 Total Revenues 7661.0 7738.0 8716.0 17709.0 15407.0 22423.0 24027.0 21629.0 1 2 3 4 5 6 7 8 1969 1970 1971 1972 1973 1974 1975 1976 I 1 Total Current Revenues 22035.0 30354.0 32881.0 24049.0 23082.0 39943.0 51379.0 61649.0 00 2 General Revenues 3762.0 3414.0 24639.0 17189.0 15458.0 20148.0 28251.0 47560.0 3 Amortization Payment 1109.0 1183.0 810.0 2188.0 0.0 0.0 14747.0 35502.0 4 Other 2653.0 2231.0 23829.0 15001.0 0.0 0.0 13504.0 12058.0 5 Transfers 18273.0 26940.0 8242.0 6860.0 7624.0 19795.0 23128.0 14089.0 6 From National Govt 0.0 14067.0 4032.0 4183.0 4093.0 7000.0 9500.0 2614.0 7 From District Govt 16112.0 12586.0 4210.0 2677.0 3531.0 12795.0 12966.0 11475.0 8 From Dist.Agencies 2161.0 287.0 0.0 0.0 0.0 0.0 662.0 0.0 9 From EAAB 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10 From EEEB 208.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 11 From LOTERIA 1953.0 287.0 0.0 0.0 0.0 0.0 0.0 0.0 12 From CPS 0.0 0.0 0.0 0.0 0.0 0.0 662.0 0.0 13 Total Capital Revenues 0.0 0.0 0.0 0.0 0.0 13881.0 35133.0 50226.0 14 Total Revenues 22035.0 30354.0 32881.0 24049.0 23082.0 53824.0 86512.0 111875.0 Lewis (1979) WPS/BR-001-4b/3.20.85 - 283 - During this same time period of 1970-74, three other projects were co-financed by ICT and CVP, that of Las Guacamanyas, La Manuelita, and Las Lomas. Since 1956, the CVP had been active in renovating and legalizing these three squatter developments. Under the Plan of the PIDUZOB (Programa Integrado de Desarrollo Urbano de la Zona Oriental de Bogota), these three projects, part of the "normas minimas" program, provided public services (water, electricity, and sewerage) to the communities, as well as schools and health centers. Funding for the projects has been shared by IDB (Inter- American Development Bank), IDU, ICT, and CVP. While these projects may be an administrative nightmare to implement, given the number of project contributors, it certainly reflects the concern of the local, national, and international communities about housing improvement in Bogota. Finally, the project Alfonso Lopez P. is a middle income housing project for District of Bogota employees. CVP has a statutory responsibility of assisting District employees with housing, and the Alfonso Lopez project resulted from this CVP responsibility. The direct ICT investment in the District, including grants and loans, has been high in relation to total national ICT investment (Table 4.33), on average 28.1% during 1960-74. 1/ Annual average growth rate of ICT investment in the District is 6.6% in real terms, compared to the annual average growth rate of total ICT investment of 6.4% in real terms. Furthermore, the District has received the largest share of ICT investment 1/ Table 4.33 shows investments in all of Cundinamarca, including Bogota. However, most of this investment is likely to have taken place within the boundaries of the District. WPS/BR-001-4b/3.20.85 - 284 - Table 4.33 Achievements and Investments by the Land Credit Institute, ICT, in Bogota and the Country Years No. of units Investment, current prices Cundina- Nation % in Cundina- Nation % in marca (1) Cundinamarca marca (1) Cundinamarca 1960 1001 10,284 9.8 19.8 168.0 11.8 1961 1607 18,793 8.5 17.6 107.5 16.3 1962 10660 31,922 33.3 79.4 219.4 36.4 1963 2499 12,367 20.3 68.0 290.9 23.5 1964 2136 8,463 25.1 60.2 159.2 27.6 1965 648 2,339 27.0 67.0 129.2 51.6 1966 1666 9,069 18.4 57.9 194.7 29.8 1967 4367 12,082 36.2 114.4 291.9 39.2 1968 3492 12,148 28.7 193.9 396.0 49.0 1969 3720 10,519 35.4 94.6 399.6 23.6 1970 3544 11,404 31.1 172.0 648.3 26.6 1971 5018 13,997 36.0 192.5 751.1 25.6 1972 6356 20,190 31.4 241.0 1,031.6 23.4 1973 8446 23,887 35.6 196.2 1,131.6 17.3 1974 2484 18,926 13.0 303.3 1,502.3 20.0 1975 X 20,926 - X 1,698.6 1976 X 22,282 X 2,076.0 DATA: Inversion y construcciones del ICT, 1942-1974, CENAC, 1965 (1) Bogota, D.E., represents more than 95% of Cundinamarca X No data available Source: Stephenson (1979). WPS/BR-001-4b/3.20.85 - 285 - compared to other subnational governments (DNP). Under the ICT minimum needs housing program, the District received 8% of the national grant funds; in the ICT Basic solution housing projects, the District received 8.1% of the national grant funds. These percentages are higher when looking at total investment (grants and loan ICT financing) with the District receiving 8.7% of total basic housing funds. However, when compared to the share of the District in total population of Colombia, these percentages of ICT investment in Bogota are quite low. The future of the ICT grants system is uncertain at this time. Many of the projects of the 1960s and 1970s are now completed. The need to integrate the ICT, the National Fund for Military Housing, and the Federation of Cooperative Housing is most apparent. Also, there is considerable pressure throughout the government to encourage private housing investment rather than public housing developments. Due to these factors, housing grants are becoming fewer, not only to the District of Bogota, but for all of Colombia. Community Action Grants. National community action ("accion comunal") grants support a small number of investments and program expenditures in the Distsrict. The 600 community action groups which exist in the District under the Community Action Administration receive grant monies for construction projects (i.e., water system) and education programs (i.e., recreation activities). National funds budgeted under the Ministry of Interior flow through the Community Action Administration to specific projects in the groups. Theca monies support only administrative costs within the program budget. Direct investment costs must be paid by either the community or a specific ministry, such as the Ministry of Public Works or Ministry of Education. The need to WPS/BR-001-4b/3.20.85 - 286 - coordinate the grant monies from the different ministries, as well as to seek community support for the project may be deterrents to the project's success. Not surprising, this grant program's contribution to the District revenue is insignificant. Additional Grant Programs. In addition to the CVP, EDTU, and Community Action groups, three other agencies receive grant transfers: the University, the Orchestra, and the District Institute for Child Protection (IDIPRON). For each of these agencies, the national giants are the primary revenue source. In 1976, the national grant monies contributed to 42.6% of the University revenues, 97.4% of the Orchestra revenues, and 91.2% (Lewis, 1979) of the IDIPRON revenues. The grant system under the three agencies is most simple. The grants for the University and Orchestra are budgeted under the Ministry of Education. IDIPRON receives grants from various national sources (i.e., the Ministry of Education, the Colombian Institute on Family Planning). As discussed in Section I above, the monies from these grants represent an extremely small share of total District revenues. The above discussion has not mentioned the grants received by the F.E.R. and S.S.S. in the District. A number of grants accompany the tax share receipts to these two decentralized units of the Ministries of Education and Health. Since these grants are closely tied to the budget process of these agencies, they will be discussed in the following sections on expenditures. The grants system in the District decentralized agencies is much less complex than the situado fiscal and shared tax arrangements. No formulas, no double layers of collection and distribution exist under the grant system. The direct national transfers to various decentralized agencies in the District, usually targeted for specific project expenditures, best describes WPS/BR-001-4b/3.20.85 - 287 - this grant system. At the same time, it is difficult to pinpoint a policy focus of the grants programs. Only the CVP and EDTU has had any policy directives, that of sites and service ("normas minimas") housing and bus investment incentives. There is a striking .difference in the administrative and policy sophistication between the situado fiscal and the IDIPRON grant program. The extent to which this paper deals with the various transfer arrangements largely reflect this level of sophistication. Part Two of this paper has described and analyzed the different tax-sharing and grant programs in the District. For each of these District transfers, there is a District expenditure. Section III focuses on the impact of these national transfers in the District, concentrating on two agencies, FER and SSS. WPS/BR-001-4c/3.19.85 - 288 - III. IMPACT ON DISTRICT REVENUE AND EXPENDITURE Introduction The District receipt ofr shared tax and grant revenues has been discussed in the eirlier sections of this paper. In summary, the following points can be made: . the bias of most revenue sharing arrangements against the District, . the gap between authorized and actual amounts distributed to the District; . the delay of payments, especially the shared sales tax revenue. The greatest impact of these shared tax and grant revenues is on the decentralized agencies of the F.E.R. and the S.S.B. (Servicio Seccional de Bogota). For both of these agencies, this national transfer revenue represents over 90% of their budget. Section III will analyze the effect of revenue sharing on these two agencies. 1/ There are several limitations in such an analysis. Most importantly, the findings are not generalizeable to other departments in Colombia. The workings of the decentralized agencies of F.E.R. and S.S.S. differ greatly between departments, territories, and District. To be sure, the agencies in the District and the Departments of Valle and Antioquia are quite similar, yet differences do exist between the agencies. Secondly, the dynamics involved in providing health and educational services are extremely complex in the District, the largest city in Colombia. In addition to its own population, the District is believed by 1/ Look to Part One for a discussion of the impact of revenue sharing on the District Administration. WPS/BR-001-4c/3.19.85 - 289 - local authorities to provide services to rural population who do not have adequate services in their own region. There are many factors which enter into the demand and supply of health and educational services in the District, (i.e. migration rates, location of services). For purposes of this paper., it must be acknowledged that the social service sector in the District is not a closed system and interjurisdictional externalities are of key importance. 1/ The Workings of the S.S.B. (Servicio Seccional de Bogota) The S.S.S., the decentralized administrative unit of the Tinsitry of Health, overseas health expenditures in the District. Since 1973 when the S.S.S. was first established under the nationalization of health services in Colombia, all national grants and earmarked taxes have flowed through its treasury. The District public hospitals report directly to the S.S.B., and the local Secretariat of Health coordinate activities with the S.S.B. Several questions arise regarding the integration of health services of the S.S.B., the Secretariat, and the hospitals. To what extent does the S.S.B. or the Secretariat or Health control health policy and administration in the District? What control mechanism are available to the Ministry? A first step in understanding the interaction of local and national health agencies is to study the S.S.B. budget. 2/ The Budget Process. The S.S.B. budget is only one of many budgets relating to District health expenditures. To a large degree, the S.S.B. coordinates the 1/ Discussion of types of externalities can be found in Oates, (1972). 2/ Unless otherwise stated, the information on budgeting, financial and administrative procedures was obtained from interviews with FER and SSB (District) officials, as well as descriptive information in national and local budgets. WPS/BR-001-4c/3.19.85 - 290 - various budgets, those of the fifteen public hospitals and the Secretariat of Health. The Social Security hospitals, the Military hospitals, and the private hospitals are outside of S.S.S. budgetary influence. The public hospitals submit their operational budgets to the S.S.B. for approval. Included in these budgets are revenues from national transfers and service fees, and expenditures of personnel, administration, general expenses, and transfers (Figure 4.18). Although the hospital budgets are not consolidated into the S.S.S. budget, the national transfers received by the hospital are included under "revenues received by the national government" and "transfer expenditure to hospitals" in the S.S.B. budget (see Table 4.34). Based on the expenditure data, the hospitals receive about 58% of the S.S.B. budget. The remaining operational expenditures of the S.S.B. support the sixty health centers in the District, as well as special health programs such as the Maternal and Infant Health program. National transfer revenue fund the S.S.B. operational costs. Also, service fees charged to patients in the Health Centers goes into the S.S.B. treasury. The budget of the Secretariat of Health includes only personnel salaries for various local programs. This budget is submitted to the District Administration for approval, as do all District Secretariats. However, the main expenditure of the budget, approximately 65% of the total, is for medical personnel; this expenditure finances salary expenses in programs of the S.S.B. to local and regional needs. Also, the amount of personnel expenditure of the Secretariat is double that of the S.S.B. The overlap of personnel between the - 291 - Figure 4.18 Budget Process of S.S.B. MINISTRY OF FINANCE * Prepares National Budget DECENTRAUZED AGENCIES (11 TOTAL) MINISTRY OF HEALTH 1* Approves S.S.B. Budget * Budget: Revenue from National Budget Transfer to S.S.B. CajaPrevsion National CAJA eiNA Hospital Service Fund DISTRICT ADMINISTRATION .Submits Budget to D.A S.S.B. BUDGET: SECRETARIAT OF HEALTH / *Revenues from National Transfer Budget /. & S.S.B. Service Fees C Adm Sasnditures for Health Centers SpeciWl Programs, & Hosprtal Transfers Submits Budget to /S.S.B. for Approval HOSPITALS BUDGET. * Revenues from Transfer & Service Fees * Current Expenditures World Bank-26186 WPS/BR-001-4c/3.19.85 - 292 - Table 4.34: S.S.B. BUDGETED REVENUE AND EXPENDITURE 1979 (thousands of pesos) As percentage of total Revenues Col$,000 Revenues/expenditure 1. Transfers Situado Fiscal 226,445 25.2% National Liquor tax 92,364 10.3% National Budget earmarked for Hospitals 112,800 12.5% Special Programs Pre- and Post-Natal care 3,200 0.4% Tuberculosis 1,081 0.1% Disease Control 1,600 0.2% Immunization 1,673 0.2% National Rural Health 17,000 2.0% National Nutrition 3,200 0.3% Drug Control 1,000 0.1% Shared Beer tax 129,000 14.4% Lottery tax 51,246 5.7% Horseracing tax 35,000 3.8% TOTAL 675,609 75.2% 2. Service Fees 28,100 3.1% 3. Other Revenue Account Balance 64,226 7.1% Late payments of transfer Situado Fiscal 14,632 1.6% Pre- and Post Natal 580 0.1% Tuberculosis 83 0.0%* Disease Control 2,068 0.2% Nutrition 500 0.1% Shared Beer tax 11,674 1.3% Lottery 2,279 0.2% Horseracing 8,197 0.9% National Budget to Hospitals 67,550 7.5% TOTAL 107,563 12.0% Interest from Loans 55 0.0%* Computerization Grant Monies Total Revenues 898.417 100.0% Expenditures 1. Personnel Services 61,531 7.8% 2. Ldministsration 54,110 5.7% 3. General Expenses 11,350 1.4% 4. Transfers Social Security 4,190 0.5% Institute for Family Planning 2,461 0.3% Social Welfare Administration 2,000 0.3% Hospitals 451,682 57.5% Other Institutions 35,132 4.5% Back payment of transfer 67,550 8.6% 5. Investsment 104,108 13.3% Total Expenditures 785,114 100.0% Source: Miniscry of Health Projected budget figures - 1979 Percentage is smaller than 1/100th. WPS/BR-001-4c/3.19.85 - 293 - Secretariat of Health and the S.S.B. suggest the highly integrated nature of the two organizations. This point will be further elaborated in our discussion of administrative authority of the S.S.B. Once the S.S.B. budget has been prepared, it is submitted to the Ministry of Health. 1/ There are a number of players in the budgeting of funds for the S.S.B. at the national level, that of the decentralized agaencies (i.e. National Hospital Funds), the Ministry of Health, and the Ministry of Finance. The Ministry of Health acts as the coordinator of these various budgets. The decentralized agencies, such as CAJANAL (the national pension program) and the National Hospital Fund, directly report to the Ministry of Health. Their budgets are submitted to the Minsitry of Health, and in turn all the national transfers earmarked to the decentralized agencies flow through the Minsitry and then unto the decentralized agencies. The central expenditure budget of the Ministry of Health includes the transfers to the S.S.B. (and the 22 departmental S.S.S.) and to the decentralized agencies. The transfers to the decentralized agencies are blocks grant to specific institutions with no set criteria for distribution of funds. This contrasts with the transfers to the S.S.B., such as the situado fiscal and the national liquor tax revenues, in which the distribution of funds is determined by an established formula. Other funds budgeted for the 1/ As the above discussion emphasized, the S.S.B. budget preparation is highly interactive with the District Secretariat budget and the public hospital budgets. The data available does not allow for distribution between the S.S.B. submitted budget, and the revised Ministry of Health's budget for the S.S.B. Information is based on interviews with both Ministry and District request. Secondly, the actual allocation of - Ministry to the S.S.B. funds may be less than the Ministry's budgeted amounts. WPS/BR-001-4c/3.19.85 - 294 - S.S.B. are under special national programs--the National Rural Health Plan, the Tuberculosis Campaign, and the Drug Control program. The S.S.B. is assigned as a set amount of these special block grant funds. The Ministry of Health submits its central budget to the National Budget Administration, an offshoot of the Ministry of Finance. There is considerable interaction between this office, the other Ministries and the National Comptroller. The double objective of the office of National Budget Adminsitration is that of balancing the budget and of achieving nalional policy goals. Health officials have not been pleased with the juggling act. (S.S.B., 1980). Health expenditures have decreased as a percentage of the national budget from 1974 - 1979 (Table 4.35), indicating that: 1) health expenditure is not a critical national policy objective; and/or 2) within the internal workings of the Ministry of Health there are certain obstacles to exerting control over the national budget process. This issue will be discussed later in the context of the comparative analysis of the health and education sectors. Suffice it to say, the Ministry of Finance and the Comptroller hold a steady grip on the initiatives of the Ministry of Health. The approval of the national budget clears the way for the disbursement of national transfers to the Ministry of Health. The flow of funds from the national treasury to the Ministry and to the local S.S.S. is no automatic process. Let us look more carefully at the different channels in which the transfers must flow before arriving at the local S.S.B. WPS/BR-G01-4c/3.19.85 - 295 - Table 4.35: NATIONAL BUDGETED APPROPRIATION FOR VARIOUS SECTORS 1974, 1979 (thousands of Pesos) Budget Budget 1974 (%) 1979 (%) Health 2.681.079 (13.1) 7.444.183 (12.6) Education 5.665.765 (40.0) 21.373.566 (36.3) Civil Aviation 132.995 (0.9) 742.339 (1.3) National Planning 78.246 (0.6) 1.544.145 (2.6) National Security 2.393.195 (16.9) 9.723.116 (16.5) National Defense 1.075.824 (7.6) 6.723.116 (11.4) National Police 117.690 (.8) 540.646 (.9) Work and Social Security 774.846 (5.5) 3.573.819 (6.1) National Resources 97.070 (.7) 1.865.871 (3.2) Comptroller 206.797 (1.5) 1.550.398 (2.6) Jurisdictional Bench 798.665 (5.6) 3.228.114 (5.5) Public Works 143.731 (1.0) 585.271 (1.0) TOTAL 14.165.903 58.894.584 100% Source: S.S.B., 1980 WPS/BR-001-4c/3.19.85 - 296 - Financial Flow of Transfers. The Ministry of Finance only approves the disbursement of funds; it does not issue the money to the Ministry of Health. The Comptroller of the Nation, in conjunction with the Bank of the Republic, issues the monthly disbursements to the Ministries (Figure 4.19). The national transfers to the Ministry of Health, such as the situado fiscal and national grant: are part of this process. The transfers to the decentralized agencies flow through the Ministry to the decentralized agencies. In the transfers earmarked to the S.S.B. and to specific hospitals, the Ministry of Health acts as a clearinghouse for the disbursement of funds. The extent to which the Ministry actually changes the destination of funds is difficult to discern from the available data. The gap in the actual distribution of funds to its appropriation, as seen in the situado fiscal, might reflect Ministry of Health action, 1/ but more probably stems from the control of the Comptroller and/or Ministry of Finance over the disbursement process. The Ministry of Health and the National Hospital Fund forward the earmarked revenues to the S.S.B. The other decentralized agencies, such as CAJANAL, forwards its local transfer funds to regional representatives. The grant revenues for special programs remain in the S.S.B. treasury, and pay for specific program expenditure. The shared tax revenues, that of the situado fiscal, the Liquor tax, the beer tax, and the lottery, tax receipts, cover general expenditures of the S.S.B., and pay for the operation of the health care centers, its general administration, and its maintenance. Table 4.36 1/ The most powerful check of the Ministry of Health is its ability to - withhold funds from the S.S.B. should the local agency not be carrying out the Ministry directive. Table 4.36 Percentage Distribut-on of 1980 Budgeted Revenue to S.S.B. and District Hospitals Revenue Source S.S.B. Lorencita Misericordia Roosevelt San Juan Samaritana Santa Clara San Ignacio Total - Percentage of Total Revenue - Service Fees 0.0* 5.9 0.9 0.7 1.0 1.9 0.8 6.7 17.9 Situado Fisca! 17.5 17.5 Liquor Tax 4.3 4.3 National Hospital Budget 5.9 0.0 0.7 0.1 3.8 0.8 2.9 0.0 14.2 Pre and Post Natal Program 0.2 0.2 Tubercu!-ois 0.0 0.0 Disease Control 0.2 0.2 National Rural Health 2.0 2.0 -4 Other National Transfers 0.2 0.2 Beer Tax Receipt 10.7 10.7 Lottery Receipt 2.3 0.0 0.0 0.0 1.1 0.2 0.0 0.0 3.6 S.S.B. Transfer n.a. 2.8 3.3 0.7 9.7 3.0 0.3 2.3 22.1 Other 4.6 0.6 0.1 0.6 1.2 0.0 0.0 0.0 7.1 Total- 47.9 9.3 5.0 2.1 16.8 5.9 4.0 9.0 100.0 * 0.0 is any percentage less than one-hundredth of a percent. Source: S.S.B., 1980. WPS/BR-001-4td3.16.85 - 298 - Figure 4.19 Financial Flows of S.S.B. Budget Comptroller Approved * Issues Monthly Payment MINISTRY OF FINANCE by Congress *uh Approval o * Prepares Budget for * Ministry of Health Congressional Approval * Ministry of Finance * Representative - Bank of Republic DECENTRALIZED T AGENCIES NATIONAL LIQUOR ACCOUNT MINISTRY OF HEALTH * National Liquor Tax Transfers Include: * Situado Fiscal * Earmarked Taxes * National Grants * Liquor Taxes National Hospital CAJANAL Liquor Earmarked Earmarked Fund Producers Transfers Grantsto to S.S,B. S.S.B. Regional S.S.B. SECRETARIAT OF HEALTH Representative Transfers Flow. * District 8udget Pays *General S.S.B. Account Adrninistrcative Salaries * Directly to Hospitals Every 15 Days Beer Producers \ . (8% Tax) Lottery 7 Revenues Administration Revenues Flow to HOSPITALS S.S.B. Transfers Horseracing National Ta Hospital Fund National Budget Funds S.S.B. Funds World Bank-26187 WPS/BR-001-4c/3.19.85 - 299 - presents a detailed summary of the destination of national transfers and other revenues within the S.S.B. and the District hospitals. Alongside the S.S.B., the Secretariat of Health receives District Administration revenues, with one- half of this expenditure supporting S.S.B. activities in the District. Although the hospital grants flow through the S.S.B., these grants are immediately sent to the hospitals upon S.S.B. receipt. The national treasury checks, are made payable to the hospitals, and not tile S.S.B. The soures of transfer revenues of the S.S.B. are both national and local in origin. The local sources of revenue include the lottery and horseracing tax revenues which are sent directly to the S.S.B., the first directly, the second indirectly. The 8% beer tax revenue is sent from the producer to the S.S.B.; the national liquor tax revenues, collected by the distillery and forwarded unto the National Liquor Account, is disbursed with the other Ministry of Health transfers to the S.S.B. The above discussion of the financial flows of the S.S.B. and its corresponding budget process has pointed to the interaction among organizations at the local, national, and intergovernmental levels. There is no one organization which dictates to the others; rather, the budget and financial processes depend on the art of compromise between organizations. Even though the national government determines the "purse" of the S.S.R., once the revenue transfers are received by the S.S.B. and the hospitals, the national government has little control over how the local agency/hospital spends the funds (i.e. personnel, administration, maintenance). As one Ministry of Health official says, "the Ministry has its hands tied, once the transfers are distributed to the local S.S.S." Neither the hospital nor the S.S.B. are required to spend transfer revenues on WPS/BR-001-4c/3.19.85 - 300 - specific expenditure items--salaries, supplies, etc. Given this leeway, the S.S.B. and the hospitals control their own administration. Decision-making largely rests with the local S.S.B. in regards to administration and management, and this will be the topic of the next section. Administration and Management. The administrative organization is as complicated as the budget and transfer processes. The S.S.B. coordinator has the title of "chief" of the S.S.B., and is selected by the Ministry of Health to administer national policy and integrate this policy with local- initiatives. The administrative power of the Coordinator is diluted by the structure of decision-making and in particular the role of the Regional Health Council. As Figure 4.20 illustrates, there are a number of officials which enter into the decision-making process of the S.S.B., the District Mayor, the District Secretary of Health, the District Secretary of Finance, the District Lottery Directory, and the Representative of the I.C.S.S., as well as the S.S.B. Coordinator. The Mayor of Bogota acts as the president of the Regional Health Council. The main decision-makers of the Council are primarily local officials. The Mayor directly appoints the Secretariate of Health and Finance as well as the Lottery Director. The I.C.S.S. Representative and S.S.B. Coordinator are the only members not directly under the influence of the Mayor's office. In short, the District government tends to control the administration and management of the S.S.B., with the Mayor and his Secretariat of Health acting as the key decision-makers. The S.S.B. Coordinator, the spokesperson for the Ministry of Health's policy, appears to play a secondary role in administration and management. - 301 - Figure 4.20 Administrative Authority of S.S.B. REGIONAL HEALTH COUNCIL MAYOR MAYOR SELECTS MINISTRY Selects S.S.B. I.C.S.S. Lottery Sec. of Mayor DISTRICT OF HEALTH S.S.B. Chief REP Director Health Counci ADMINISTRATION Chief ADMINISTERS/ S.S.B. 7/ I/ S.S.B. SECRETARIAT OF HEALTHI * S.S.B. Chief Reports I *Under Authority of Mayor to Ministry & Mayor *Memnber of Council *Under Authonity of CouncilI I WodBn-68 WPS/BR-001-4c/3.19.85 - 302 - The decision-making structure of the Regional Health Council is characteristic of Colombian organization: 1) the need to resolve conflict before any action can be taken and 2) the lack of accountability in the decision-making process. Moreover, while the Council is held responsible for the operation of S.S.B., the Council cannot control the actions of the national. health officials, nor of the hospitals of the District which remain detached from the S.S.B. operation. Problems Confronting the S.S.B. The above discussion has identified a number of flaws in the budget, financial, and decision-making processes: the various budgets needing coordination, the delay and/or non-payment of transfers to the S.S.B., the cumbersoie decision-making structure of the S.S.B. All act as obstacles to effective health policy and administration in the District. Moreover, these factors are mutually reenforcing, further exacerbating the situation. For example, the delayed transfers are accounted for in the next year's budget under "Revenue" (see Table 4.34). These overdue transfer payments are often not paid by the Ministry, and the S.S.B. finds itself short of revenue on account of the arrears from the current as well as the previous years. After a ntmber of years, the problem of unpaid transfers snowballs to the point where it is difficult to assess the financial and budgetary status of the S.S.B. The complications arising from the budget and financial processes reduce the amount of transfer money to the S.S.B. The consequence is that the S.S.B. receives a smaller share of national funds than it did in the past (Table 4.37). Administration of S.S.B. health program and hospital care program and hospital care cannot be effective without national revenues which contribute to 94% of the S.S.S. recurrent budget and 61.8% of the hospital WPS/BR-001-4c/3.19.85 - 303 - Table 4.37: S.S.B. PARTICIPATION IN THE SITUADO FISCAL AND TOTAL NATIONAL BUDGET 1974-1979 (Colombian Pesos) S.S.B. S.S.B. Receipt National Situado Fiscal Situado Fiscal as % As % of Year Bud- t Health Receipt of National Situado Fiscal Budget Health .1974 30.303.190 697.080 55.285 0.0018 0.08 1975 33.854.088 679.722 66.453 0.0020 0.10 1976 53.386.577 1.320.248 125.202 0.0023 0.09 1977 62,748.192 1.377.944 138.609 0.0022 0.10 1978 86.580.619 1.673.080 175.578 0.0020 0.10 1979 108.256.514 1.935.487 226.445 0.0019 0.10 Source: S.S.B., 1980 WPS/BR-001-4c/3.19.85 - 304 - budget (S.S.B. 1980). The District Administration has not been able to make up for the deficiencies of the national transfers. Its contribution has increased from 2.3% in 1973 to 3.9% in 1977. At the same time that the health service sector has received a smaller share of funds from the national government, the health policy has called for extending services to local health centers. Under the decentralized planning of the Ministry, the local health center is to provide the initial health treatment. The second level of treatment is at a general hospital, and the third level is at a specialized hospital. The objectives of the local health center located in the actual neighborhood of its patients are two: preventive medicine and outreach. The goal of the program is to treat the patient at the neighborhood level, eliminating as far as possible the second- and third-levels of treatment. According to interviews with District officials, given the shortage of revenues in the S.S.B., the local health centers all said not to have been adequately manned or supplied. 1/ Due to the insufficient resources of the centers, service levels have declined; and accompanying the decline is the deterioration of the health of the patient who frequents the local health centers. As the patient's health deteriorates to a critical level, the patient is forced to go to the hospital, often under emergency circumstances. The need for emergency care directly 1/ The performance level of the sixty Centers of Health in the District varies tremendously. Several centers have five to seven doctors, others have only one doctor part-time and few supplies. The information on the adequacy of the health centers is based on interviews with S.S.B. and hospital officials, academics, and reports made by Accion Communal leaders. WPS/BR-001-4c/3.19.85 - 305 - results from the poor services offered at the local health agency. The hospitals forced to provide care to the patients are now in a serious financial bind. The decentralization of health services did not decrease the demand for hospital services, rather it re-defined the type of demand, that to emergency service care. Furthermore, national funds have been re-directed to the local health centers, and the hospitals are receiving less national funds than four years ago. The conflict between the policy directives and the financial and budget realities of the health care sector in the District is a delicate topic of conversation with government officials. The inaction of the governments, both local and national, to resolve the situation can best be explained by the decision-making structure of the S.S.B. As we have seen, there is no one organization, not even the Regional Health Council, that can reconcile the policy versus financial conflict. The policy mechanism of the Ministry cannot even muster the power to increase its percentage of the national budget. The Mayor and his Regional Health Council control the administration of services, but not the 'purse' to provide these services. This analysis of the workings of the S.S.B. and District hospitals has concentrated on the general issues of concern. Unfortunately, this method only allows generalities, such as the need for accountability in the decision- making structure, for a consolidated budget in the S.S.B. and for increasing health expenditures at the local health centers. No specific statements can be made in regards to the type of transfer program, and its impact on expenditures from the above analysis. Before developing such a framework, letus study the workings of the F.E.R., and then proceed to a comparative WPS/BR-001-4c/3.19.85 - 306 - analysis of the impact of national transfers on the expenditures of S.S.B. and F.E.R. The Workings of F.E.R. The operation of the F.E.R. is very similar to that of the S.S.B. The purpose of the decentralized unit of the Ministry of Education is to encourage educational services to be responsive to local needs, concurrently with national policy objectives of universal primary education. The District F.E.R. adminsiters primary and secondary public education, and several special programs for pre-schoolers, adults, and the general public. Since its establishment in 1969, the District F.E.R. has received increased responsibility; and in 1976, it began to receive the national transfers directly from the Ministry of Education, by-passing the District Treasury. Let us now look at the budget, financial, and decision-making channels which control educational policies and administration in the District. The Budget Process. The District F.E.R. budget is a much more consolidated document than that of the S.S.B. The budget is structured around 'expenditure programs', such as primary, secondary and special programs (see Table 4.38). Teachers and adminsitrative salaries, supplies, and maintenance are included under each 'expenditure program'. The budget for each school falls under a number of expenditure items (i.e. teacher's salaries, supplies) within a given program (i.e. primary education). Primary and secondary education expenditures represent 88% of total expenditures. While the consolidation efforts of the budget are commendable, the massiveness and general unaccessibility of the d cument makes it less operational than that of the S.S.B. WPS/BR-001-4c/3.19.85 - 307 Table 4.38: 1979 Budgeted F.E.R. Revenues and Expenditures and as Percent of Total (Thousands of Pesos) Colombian As Percentage Revenues $P,000 of Total National Transfers Situado Fiscal 1.038.056 43.5 Nationalization (shared sales tax revenues) 308.165 12.9 Additional Salaries 126.315 5.3 Cooperative Schools 76.855 3.2 Technical Training (INEM/ITAS) 130.763 5.5 Building Maintenance 346.738 14.5 Additional Grants 6.443 0.3 Adult Education 4.450 0.2 Teacher Training 1.194 0.0* Total Transfers 2.038.979 85.4 District Transfer 323.000 13.5 Education Fees 24.830 1.1 Total Revenues 2.386.803 100% General Administration 22.779 1.0 Preschool Education 4.422 0.2 Basic Primary Education 1.106.709 48.1 Secondary Education 440.582 19.2 (Nationalized) Post Secondary Education 24,601 1.1 (Non-University) Special Education 133 0.0 Nonformal 13,455 0.6 Training 9.266 0.4 Secondary Education (Tehcnical) 477.202 20.7 Cooperative Schools 74.730 3.2 Additional Salaries 126.315 5.5 Total Expenditures 2.300.194 100% * 0.0 indicates that the percentage is less than 1/100 th of a percent. Source: FER, 1980. WPS/BR-001-4c/3.19.85 - 308 - The revenues of the budget include national and District transfers, service and matriculation fees (see Table 4.39). The national transfers consist of the situado fiscal, the shared sales tax revenues earmarked for secondary education, and a variety of grants for specific programs, namely adult education, teacher in-service training, plant maintenance, and technical education. These national transfers represent 85.4% of F.E.R.'s total. The District government supports the F.E.R. in two ways: 1) direct transfers to the F.E.R. to pay for teacher's salaries; and 2) payment of Secretary of Education administrative salaries. As in the Secretary of Health, the Secretary of Education works in conjunction with the decentralized unit, and submits its own budget to the District Administration (Figure 4.21). The District F.E.R. submits its budget to the Ministry of Education for approval although the Ministry uses this tentative budget as a guide in planning its own budget. In 1978-1979 the Ministry revised the initial buget figures significantly. The central budget of the Ministry of Education earmarks all of its transfer expenditures to the specific F.E.R. In the Ministry budget, these transfers are listed, but are not disaggregated into specific expenditure items (i.e. teacher's salaries). The expenditure programs are then detailed as specific transfers to the respective F.E.R. The decentralized agencies of the Ministry of Education, twenty-seven agencies in all, are also part of the national administration budget. ICETEX (Institute for Educational Grants), ICCE (Institute for School Construction), and SENA (National Apprenticeship Service) submit their tentative budgets to the Ministry of Education. Transfers under these decentralized agencies are WPS/BR-001-4c/3.19.85 - 309 - Table 4.39: Ceiling Limits on the Nationalization of Secondary Education District/Departments Secondary Education Expenditure, 1975 Bogota, D.E. 92.239.100 Antioquia 190.102.120 Atlantico 51.274.795 Bolivar 67.580.00 Boyaca 69.818.903 Caldas 99.935.413 Cauca 26.047.426 Cesar 11.753.600 Cordoba 32.821.384 Cundinamarca 174.785.067 Choco 4.190.420 Guajira 7.929.384 Huila 20.464.575 Haqdalena 13.129.645 Meta 19.530.330 Narlno 37.205.218 Norte de Santander 36.260.180 Quindio 37.656.724 Risaralda 48.241.978 Santander 73.927.595 Sucre 21.721.653 Tolima 48.137.000 Valle 135.342.939 TOTAL 1.320.095.449 Source: Azuero de Munoz (1980) - 310 - Figure 4.21 Budget Process of F.E.R. MINISTRY OF FINANCE DECENTRAUZED Prepares National Budget AGENCIES (27 AGENCIES) SENA CMINISTRY OF EDUCATION - Approves FER Budget - Central & Decentralized Budgets Consolidated - Transfers Under 'Program Expenditure' Submits Budget to Includes District Transfer MEN for Approval DISTRICT ADMINISTRATION in Revenue Budget MNfrApoa * Payment of Secretariat * Direct Transfer FER SECRETARIAT OF FE.R. Budget Submits Budget EDUCATION / * Revenues from District to D.A & National Transfers * Administrative / * Consolidated Program Salaries I Expenditures / SCHOOL BUDGET INCLUDED IN PROGRAM EXPENDITURE World Bank-26189 WPS/BR-001-4c/3.19.85 - 311 - listed by project, such as specific schools to be constructed, scholarship to certain schools, or funds to a specific university. The central budget of the Ministry of Education is then presented to the Office of National Budget Adminsitration of the Ministry of Finance. As mentioned earlier, the preparation of the national budget entails the specific ministries, officials from the Budget Administration and Ministry of Finance, and the Comptroller's office and Bank of the Republic. The relative priorities given to these programs of the various offices in the budget preparation can be seen by the amount of funds budgeted under education, as seen in Table 4.35. Education receives a large percentage of national funds, 38% an average of 1974 and 1979. Although this percentage has decreased in the past years, it is still much higher than that of health (12.6%). Clearly, education is recognized as a vital policy concern by the National Budget Administration, and the educational budget reflects the muscle of the Minsitry of Education in the budget process. Financial Flow of Transfers. While the Ministry of Education exerts considerable influence over the budget process, it does not seem to control the disbursement of funds to the same extent. As discussed earlier, the disbursement process under the wing of the National Comptroller works in the following manner: the Ministries of Education and Finance, the Comptroller's office, and the representative of Bank of the Republic must sign off on each monthly issuance of funds to the Ministry of Education (see Figure 4.22). Once the payment is cleared by these various offices, the funds flow through the Ministry of Education to their earmarked destinations. The transfers earmarked in the budget preparation include: situado fiscal, shared - 312 - Figure 4.22 Financial Flows of F.E.R. Comptroller Issues Monthly Payment with Approval of: CONGRESS Ministry of Education APPROVES BUDGET Ministry of Health Ministry of Finance Representative - Bank of Republic MINISTRY OF EDUCATION Transfers include: Situado Fscal: Shared Sales Tax Additional Salades/Grants National Savings Funds SENA ICETEX ICCE Earmarked DISTRICT ADMINISTRATION District Transfers Transfer to F.E.R. (Salaries) SECRTARIT OFF.E.R. SECRETARIAT OF Funds Flow to FER Treasury " EDUCATION Teachers Salaries REGIONAL REPRESENTATIVES Administrative Salaries Special Programs Administration/Supplies SCHOOL OPERATION World Bank-26190 WPS/BR-001-4c/3.19.85 - 313 - sales tax, additional salary payments, 1/ plant maintenance, and technical education. The grants of the decentralized agencies have also been targetted to specific projects. The difference between the distribution of these transfers is that all Central Ministry transfers flow to their regional representatives. In spite of Article 90 of Decree 102/1970 which ordered that all educational revenues from the Ministry--with the exception of ICCE and the university grants--be directed to the F.E.R. treasury, the decentralized agencies continue to flow transfer funds to the regional headquarters. Once again the legislative intent of the budget process is in contrast to the actual workings of the budget process. The only interaction of these headquarters with the District F.E.R. is in the Regional Education Council. The F.E.R. receipt of the national transfers flow into its treasury, and are used to pay for teacher's salaries, special programs, and administration. The expenditure composition of the different transfers is dictated by the contractual agreement of 1974 and 1976. These contracts differ with each F.E.R., and are the legal mechanisms of the Ministry to control F.E.R. expenditures. The composition of the District F.E.R. is roughly 9% for salaries (administrative and.teachers) and 3% for general expenditures and transfers to social security. The payment of these transfers is made on a monthly basis. In discussing the disbursement process with the F.E.R. coordinator, it seems that the Ministry of Education does not send one total monthly payment, rather issues several payments for specific expenditure 1/ These payments are for teacher's salaries not covered by the Nationalization (e.g. shared sales tax). WPS/BR-001-4c/3.19.85 - 314 - items, namely primary teachers salaries, secondary teacher salaries, administrative costs. In addition to the national transfers, the District Administration forwards revenues to the F.E.R. These direct transfers support salaries. The District Administration also pays for the administrative salaries of the Secretariat of Education. Salary checks are sent to Secretariat employees from the District Treasury. Another revenue source is the service fee which represents a small percentage of F.E.R. recurrent revenues (1.1%); these fees pay for general operational expenses, such as books, pencils, materials, and other supplies. Administrative Authority. The underlying objective of establishing decentralized adminsitrative units (F.E.R.) was to create an organization which would follow the policy mandates of the Ministry, yet at the same time be flexible enough to meet local needs. Like its sister agency, the S.S.B., the F.E.R. administration is dominated by the Mayor's office. And in fact, the Ministry of Education appears to have less direct input into local administration than in the case of the Ministry of Health in the S.S.B. The representative of the Ministry of Education is not the "chief" of F.E.R., rather a F.E.R. coordinator is selected by the Regional Educational Council to administer F.E.R. The Regional Educational Council, headed by the Mayor, has the following members: the District Secretaries of Education and Finance, and one representative from the teacher's association, ICETEX, ICCE, and the Ministry of Education (see Figure 4.23). Given that the Mayor only appoints two out of the six members, one would expect that District and Ministry interests to be balanced. This is not the case. The Mayor and his Secretary of Education are - 315 - Figure 4.23 Administrative Authority of F.E.R. MINISTRY OF EDUCATION Selects REGIONAL EDUCATION COUNCIL MAYOR Ministry Represent. Selects Sec. Ministry ICETEX Sec. of of Finance Rep. Rep. Finance MAYOR DISTRICT ADMINISTRATION ICCE Teacher Sec. of Rep. Rep. Educ. Selects FER Coordinator Select Sec.cretary of Selects Sec Education ofEducation Double Role F.E.R. C a SECRETARIAT OF EDUCATION * Coordinator Reports to *UdrAtoiyo ao Mayor/Secretary of Education * Under Authority of Mayor * Under Authority of Council * Member of Council World Bank-26191 WPS/BR-001-4c/3.19.85 - 316 - largely responsible for F.E.R. administration. The role of the representative of the Ministry is to ensure that the F.E.R. follows the contractual obligations of the F.E.R. to the Ministry. The other representatives act as coordinators of activities between their appropriate organization and the F.E.R. The autonomy of the District F.E.R. from the Ministry is most pronounced. This autonomy can be explained in part by the financial role of the District in supporting educational services. The District also has a greater willingness to pay for educational services given this level of F.E.R. autonomy from the national ministry. As a general observation, the F.E.R. administration seems intimately tied to the political strings of the District Mayor's office. This contrasts with the S.S.B. adminsitration where the Regional Health Council, also headed by the Mayor, seems to lack consensus and control. The decision-making power of the Regional Education Council clearly rests with the Mayor and the Secretary. And it is believed that the Mayor can solve both financial and administrative problems. The influence of the Mayor over Ministry-related matters (i.e., financial) may stem from the fact that Hernan Duran Dussan -- the mayor from 1978-1981 -- was previously the Minister of Education during the presidency of Lopez Michelson. The national prestige of Duran may explain the high District profile in educational issues, and as such may be a transitory phenomenon. The Problems Confronting the F.E.R. A number of problems present themselves in the operation of the F.E.R.: the delayed transfer payments, the cumbersome consolidated budget, and to a certain extent, the political overtones of- F.E.R. administration. To a large degree, these factors overlap. Similar to WPS/BR-001-4c/3.19.85 - 317 - the budget process of the S.S.B., the unpaid transfers of fiscal year 1 are budgeted in the following year's budget. The transfers arrears are often never paid to the F.E.R., and thus budgets fall short of necessary revenues to implement the educational programs, on account of non-payment of current transfers obligations as well as arrears. The District Administration and the Mayor's office become involved in the disbursement process when the F.E.R. does not receive the expected transfers. The transfer payments to support secondary teacher's salaries and F.E.R. administrative staff are especially slow. To resolve the situation, the Mayor can either pressure the Ministry, and/or contribute additional District funds to pay for these expenditures. This haphazard process of authorizing funds seems to follow the principle of 'robbing Peter to pay Paul'. The shortage of revenues cannot be blamed entirely on the budget and financial processes; so as to be discussed below, the structure of financing primary and secondary education, namely the situado fiscal and the nationalization of 1975, has proven to be problematic. Also, the restructuring of teacher's salaries in 1979 has placed a heavy financial burden on the F.E.R. The District F.E.R. receipt of the situado fiscal and shared sales tax revenue are largely determined by the 70/30 formula: 30% distributed equally and 70% distributed based on 1964 population statistics. However, according to the District F.E.R. accounts for 1978-1979, (F.E.R.,1978-79) supplemental authorization are issued for all of the revenue sources -- the situado fiscal, the nationalization, and all other specific grant programs. These supplementals increased the F.E.R. revenue budget by 27% in 1979. -The consistent use of supplementals and additional salary payment point out that WPS/BR-001-4c/3.19.85 - 318 - the formulas alone are not adequate to determine the educational funding needs of the F.E.R. 1/ Secondly, under the nationalization of secondary education, the phase out of District financing of secondary education has proved to be unpracticable. Under the plan, District financing of secondary education was to be phased out at a rate of 20% per year 1976-1980. While the nationalization established Ministry responsibility in paying for secondary education, the amount of financial responsibility of the Ministry has a ceiling limit. 2/ The limits established for each department/District (Table 4.39) are equal to estimated 1975 secondary education expenditures. No provisions are included for increasing the amount for additional teachers. 3/ Only the salaries of the number of teachers based on 1975 expenditures will be covered. The inflexible nature of the nationalization act has led to inadequate financing of secondary education in the District. The insufficient level of national financing is well illustrated in Table 4.40 where the District contribution has certainly not been phased out. The Ministry's response to this demand for additional revenues is the "additional salaries" to the departments/District, and it appears that this transfer, along with the supplementals, are a respnse to the political pressure and to the actual need of the various F.E.R.s. 1/ The impact on expenditure level will be discussed further below. 2/ There is an escape clause in Ley 43/75 which state that "any sum which exceeds this guaranteed amount will be the responsibility of the respective department/District" Art 1, Par 4. 3/ Additional teachers can be hired by the F.E.R. but need the approval of the Ministry. WPS/BR-001-4c/3.19.85 - - 319 - Table 4.40: Authorized Versus Actual Phase-Out of District Financing of Secondary Education 1/ (1975-1980) Base Phase- Rate of Authorized Actual Difference Year Out Amount Phase Out Financing Financing (A)-(B) ('000 Pesos) (Percentage) ('000 Pesos) ('000 Pesos) ('000 Pesos) 1975 92,239 100% 92,239 131,033 1/ 38,794 1976 92,239 80% 73,791 138,423 2/ 64,632 1977 92,239 601 55,343 141,850 86,157 1978 92,239 40% 36,860 140,800 103,940 1979 92,239 20% 18,448 140,800 122,312 1980 92,239 0% 0.0 140,800 140,800 1/ District Budget Data 2/ Estimated on 1977-1980 Data in District Budget Source: Calculated using F.E.R. 1980 data and District budget data (1975)- (1980). WPS/BR-001-4c/3.19.85 - 320 - On top of all these troubles, the wage structure of teacher's salaries has been reorganized. The Ley 2277 of 1979 revised the traditional structure of limiting primary school teacher's salaries to only four grade status. Under the new system, all teachers, of primary and secondary are eligible for 8 grade classifications (General 1-4 and Specialty 1-4). Primary and secondary teachers receive their status based on education level, teaching experience, and number of years in service. This restructuring aims at improving the quality of primary school teachers. Primary education has traditionally employed teachers with the lowest skills, namely a high school education and a certificate from a Teacher's College; and for that reason, primary education teachers were eligible for only 4 grade salary steps. This new wage structure emphasized curriculum development techniques and skills which contribute to quality teaching. The costs involved in improving teaching in the primary sector, are extremely high. It is predicted that the restructing will cost roughly Col$1,500 million (Munoz,1980). During an interview in mid-1980, the F.E.R. coordinator estimated teachers' salaries to go up by 45% in the District for school 1980-1981. As of the end of 1980, the restructuring had not been completed. The F.E.R. coordinator stated that until the District F.E.R. receives the transfers ,from the Ministry to pay these higher salaries, the F.E.R. will not follow the guidelines of the restructuring. Once again we see the conflict between policy and financial structures. The overly ambitious policies fail because of the unpractical financing structures. Interestingly, the politicalization of the process -- the supplementals and the "additional salaries" -- enters when revenues are inadequate. The above discussion has concentrated on the actual workings of WPS/BR-001-4c/3.19.85 - 321 - the F.E.R. and S.S.B.; it has been largely descriptive, rather than analytical. The following section offers an analysis of the impact of the revenue transfers of the F.E.R. and S.S.B. on their respective expenditure levels. The Impact of National Transfers on S.S.B, and F.E.R. Expenditures The preceding paragraphs have described the workings of the two decentralized entities, the S.S.B. and the F.E.R., highlighting problems confronting them. The similarities of S.S.B. and F.E.R. are striking. Both rely on national transfers as their primary source of revenue; both are governed by a Regional Council. It could be postulated that the transfers would have comparable effects on the expenditures of the two agencies. To analyze these effects, it is necessary to stud_ the composition of revenue in each agency, and the percentage of the transfer to total revenue and expenditure. In our earlier discussion of financial flows to the F.E.R. and S.S.B. transfer revenue was identified as the main source of revenue. In examining the composition of revenue in the F.E.R., the situado fiscal contributes 44%, the shared sales tax 9.5%, the "additional salaries" 5.6%, the District Administiation 15.2% and the plant maintenance grant 15.1%. Four of the five transfers, representing 74%, are earmarked primarily for teacher's salaries. The 1978-79 F.E.R. revenue and expenditures data are presented in Table 4.41. Based on these figures, the total appropriation of F.E.R. expenditures does not correspond to the total revenue appropriation. The actual revenue receipt is approximately equal to the total appropriation. However, the 1978-1979 total expenditures are lower than the amounts specified in either the initial or the total expenditure appropriated. The F.E.R.- officials stated that the shortfall in actual expenditure is due to the WPS/BR-001-4c/3.19.85 - 322 - Table 4.41: F.E.R. Initial and Total Appropriation and Actual Receipt of Revenue and Level of Expenditures 1978-1979 (Thousands of Pesos) % of % of 1978 (Original) 1979 (Original) Revenuea Initial Appropriation 1.375.812 (100%) 1.698.421 (100%) Total Appropriation 1.750.727 (127%) 2.390.247 (141%) Actual Receipt 1.733.772 (126%) 2.387.584 (141%) Arrears Initial Appropriation 867.755 117.410 Actual Receipt 000 64.427 Expenditure Initial Appropriation 1.772.960 (100%) 2.582.143 (100%) Total Appropriation 2.148.225 (121%) 2.495.448 (97%) Actual Expenditure 1.742.991 (98%) 2.300.195 (89%) a Revenue figures do not include the arrears from previous years. Source: F.E.R., 1980. AR WPS/BR-001-4c/3.19.85 - 323 - delayed payments and non-payments of the previous arrears. In 1978, the 900 million pesos arrear was never paid, and in 1979, only one-half of the 117 million peso arrear was paid. The S.S.B. transfers have also decreased. While the national transfers represent 95% of total revenues of the S.S.B. (see Table 4.34), on average in 1979-1980, 60% of the transfer revenue received by the S.S.B. flowed directly to the hospitals. When netting out the hospital transfers from the S.S.B. revenues, there is a sizeable decrease in transfer revenues supporting direct expenditures of the S.S.B. (see Table 4.42). The contribution of transfer revenue (net of hospital transfers) is around 88% of total revenues. Service fees have taken on new importance, bringing in 12% of total revenues. Information on actual S.S.B. expenditures is extremely limited. Budgeted expenditures were significantly less than budgeted revenues in 1979-1980. Hospital revenue and expenditure data is much more detailed than that of the S.S.B. (see Table 4.43). For years 1974-1978, the revenue transfers have been declining in real peso terms and as a percentage of total revenues. The contribution of service fees has increased in real terms during this period. The level of expenditure has been rel.atively stable from 1975 to 1978. In fact, there has been only a small decline in expenditure level in constant terms during this period. Based on the above revenue and expenditure data, several conclusions can be drawn: 0 National transfer play a critical role in financing F.E.R., S.S.B., and hospital expenditure; however, their contribution to the total budgets of these agencies has declined in recent years. WPS/BR-001-4c/3.19.85 - 324 - Table 4.42: S.S.B. Budgeted Revenue and Expenditures - 1979-80 (Thousands of Current Pesos) 1979 (% of Total) 1980 (% of Total) Total transfer 837.631 788.000 (Net Hospital Transfer) (376.964) (554.364) S.S.B. Revenue 460.667 (93.6%) 223.636 (81.7%) Service Fees collected 31.331 (6.4%) 50.020 (18.3%) Total Revenue 491.998 (100%) 273.565 (100%) Expenditure S.S.B. Expenditure 333.432 156.149 (Net Hospital Transfers) Source: Ministry of Health, 1979-80. WPS/BR-001-4c/3.19.85 - 325 - Table 4.43: Hospital Revenues and Expenditures (1974-1980) (thousands of constant 1976 Pesos) Revenues (% of Service Fees Year Transfers Total) and Other (%) Total (%) 1974 246,555 (70) 351,638 (30) 598,193 (100) 1975 342,496 (75) 458,118 (25) 800,614 (100) 1976 239,143 (46) 524,990 (54) 764,133 (100) 1977 331,546 (60) 549,954 (40) 881,500 (100) 1978 233,077 (52) 447,060 (48) 680,137 (100) Expenditures Year Salaries (%) Administrative (%) Other General (%) Total (%) 1975 245,695 (60) 106,357 (26) 60,188 (14) 412,240 (100) 1976 271,465 (62) 123,193 (28) 45,318 (10) 439,976 (100) 1977 275,698 (63) 123,993 (28) 40,019 ( 9) 439,710 (100) 1978 222,397 (61) 102,927 (28) 125,850 (28) 451,174 (100) Source: S.S.B., 1980 WPS/BR-001-4c/3.19.85 - - 326 - 0 F.E.R.'s shortfall in actual expenditures is due to the past revenue arrears which are added into the current revenue budget but are never received during the fiscal year. o S.S.B.'s budgeted expenditure is significantly less than its budgeted revenue in 1979-1980; service fees are increasingly becoming an important source of revenue. o Hospital expenditures have been relatively stable in current prices during 1975-1978; transfer revenue has declined in constant terms during this period. o Hospitals appear to be less financially dependent on national transfers with service fees contributing an increasing percentage to total revenues. To examine more carefully the impact of revenues on expenditures, the type of expenditure that national transfer support can be analyzed. National transfers primarily support recurrent expenditures. Under all three budgets, recurrent costs are listed as salaries, administration, and general operational expenses. During 1979, teacher salaries made up 96.8% of the F.E.R. budget. Administration and operational expenses make up only 3.2% of total recurrent expenditures. (MEN, 1980). It is generally considered that these two expenditure items are loosely tied together. An increase in the number of teachers to serve a greater number of students usually requires additional operational materials (i.e. books, supplies) and administration. The composition of health expenditures differs significantly. In S.S.B., approximately 35% of expenditures go to operational expendiures (i.e. supplies, medicines). Both administrative and technical salaries account for WPS/BR-001-4c/3.1.85 - 327 - 64% of S.S.S. expenditures. It must be emphasized that the types of expenditures in the health sector are closely tied. For efficient treatment of the patient, there must be both a salary expenditure (i.e. doctor or nurse) and a general expenditure (e.g. medicine). Although the S.S.B. and hospital expenditure data does not allow for exact comparisons by typs of expenditure, it is generally thought that education expenditure is primarily of one type -- teacher salaries; and health expenditures are of several types -- general, salaries, administrative. To what degree do national transfer revenues encourage these expenditure patterns? In the case of the F.E.R., all national transfers are earmarked to specific expenditures. Some 60% of the transfer funds are earmarked to teacher salaries. The rest of the transfer funds go into special programs. Only an extremely small amount of the situado fiscal (0.3% of total revenue) and special grant funds (.9% of total revnue) pay for general operational expenditures and supplies. Service fee revenues, only 1.3% of total revenue support general operational expenditures. The composition of educational expenditures reflect the earmarking, 98.5% to teacher's salaries and 1.5% to administration. National transfers are not earmarked in the S.S.B. nor the hospitals. Transfers support adminsitrative and technical staff, administrative supplies, and general maintenance. To a large extent, the unearmarked transfers reflect the integrated structure of expenditures in the health sector. The effective provision of services in the health sector demands that the recurrent expenditures be tied together. The importance of earmarking is that it adds accountability to the the transfer process; in essence, it creates a pressure on the national WPS/BR-001-4c/3.19.85 - 328 - government to pay for specific expenditures. For example, the Ministry of Education is responsive to the demand for supplemental appropriations to the F.E.R. The Ministry of Education is well aware of the potential threat of a teachers strike, should the salaries not be paid (or the transfer payment not received by the local F.E.R.). With the unearmarked transfer revenue of the S.S.B. and the hospitals, the S.S.B. has been ineffective in pressuring the Ministry. No one factor, be it doctor salaries or medical supplies will be exclusively affected by the non-payment of transfer revenue. No strikes, lawsuits, or bureaucratic nightmares threaten the Ministry of Health. The conclusion to be drawn from this analysis is that expenditure and financing structures are extremely interwoven with the services provided by the health and education sectors. Moreover, these financing and expenditure structures largely determine the responsiveness of the national government in paying the transfer revenues. Education with a high expenditure percentage devoted to teachers' salaries has most of its transfer revenue earmarked to this specific expenditure item. The political clout of the teacher's union and the direct link between transfer revenue and specific expenditure allows F.E.R. to have significant bargaining power with the Ministry of Education. In contrast, the S.S.B. transfer process does not provide a mechanism for the S.S.B. to apply pressure on the Ministry of Health. IV. CONCLUSION Revenue sharing in the District is clearly part of the policy and political processes, how revenue sharing is supposed to work (i.e., the policy process) and how revenues sharing actually does work (i.e., the political process). The policies are oriented towards economic growth and service WPS/BR-001-4c/3.19.85 - 329 - provision. The formulas--the guidelines for revenue sharing--aim to ensure conformity of revenue sharing with economic and social goals. The political process is concerned primarily with conflict management. The trade-off between policy objectives and political realities explains the shortfalls of the revenue sharing system. Any conclusions to be drawn with regard to the system must recognize the interaction of these two processes. The conclusions drawn from our findings have several limitations. Much of the paper has dealt with descriptive material relevant to the District. Only in the relative comparisons of the District participation to other department presented in Part Two can generalizations be made on the national impact of revenue sharing. Secondly, the data used for comparisons, especially those of the District F.E.R. and S.S.B., are extremely limited. For that reason, the data here only supplemented our discussion of the budget and disbursement processes, and do not document trends of revenues and expenditures. Thirdly, the transfers to the S.S.B. and F.E.R. come in any shapes and sizes. The conclusions of this paper will focus on the larger revenue sharing programs, primarily the situado fiscal. The first conclusion to be drawn from our findings is that the budget process primarily reflects the policy concerns and that the disbursement of funds is entangled in the political sphere. The gap between authorized funds to actual receipt illustrates the wide schism of the budget and disbursement processes. This is not to say that the budget process itself is apolitical. The formula guidelines are imposed when preparing the initial budget; however, the supplementals and "additional salaries" certainly suggest that politics are at work in the budget modification. One means of closing the gap is to simplify the disbursement process, WPS/BR-001-4c/3.19.85 - 330 - creating a special account for shared tax revenue and by-passing the national butdget. The revenues would be distributed from this special account based on the distribution formula. The simplification of the collection and disbursement might certainly eliminate some of the politicizing of the budget and distribution of revenues. However, the delay in the transfers of the shared beer tax to the S.S.B., a shared tax which is not within the national budget, certainly does not leave one overly optimistic that a special shared tax account will necessarily close between authorized and actual amounts. This conclusion only begs the question of the financial viability of the shared taxes. The authorized amounts of shared tax revenues do not equal the percentages prescribed in their respective legislative mandates. Furthermore, not one department nor the District receive their full amount of authorized funds. This suggests that the legislated percentages are high in relation to the national budget constraint. This paper does not examine this question, yet emphasizes that it is most pertinent to the effective administration of the shared tax programs. Another conclusion reflects our findings on the distribution formula of the situado fiscal and the shared sales tax program: the formula of 70/30 using 1964 census statistics does not achieve the national policy objective of equity. The stated objectives of the situado fiscal are the equalization of regional income and opportunity to health and educational services. While these objectives are not contradictory, the issue of equity take a back seat to educational service provision. The differences between national and local policies are even more striking. The national government's goal of regional equalization is interpreted in the District as an anti-Bogota policy. The District, now WPS/BR-001-4c/3.19.85 - 331 - receiving less national revenues per capita, suffers the consequences of this redistribution program. This study does not allow for any general statement on the price that Bogota pays under the revenue sharing arrangements for two reasons: 1) no calculations have been made on the District's revenue contribution to tax sharing; and 2) the externalities of the social goods (i.e. health and educational services) have not been measured. We can only conclude that the District perceives revenue sharing as a national policy not in its interesc. It seems only fair to judge the appropriateness of the formula by the objectives as stated in the situado fiscal. As we have seen earlier, the formula is slightly redistributive, yet responsive to educational "need". However, the formula is becoming less responsive to income, as seen in the insignificance of the relationship between situado fiscal and income in 1977. Clearly, the formula must be updated to reflect the changing composition of population and income in Colombia. Also, the actual income levels of the departments should be measured into the formula, and the equal share component be discarded. This would prevent departments with high income, but low population (such as Guajira) from receiving a relatively large percentage of transEers per capita given its income level. A second issue tied to the effectiveness of the formula is that of the supplementals. The impact of the supplementals is not particularly important to the total distribution of the situado fiscal; yet their presence shows the breakdown of the authorized guidelines. The supplementals could be institutionalized into the revenue sharing process under a matching grant program, over and above the formula based distribution. The institutionalizing of the supplemental might control the political pressure on WPS/BR-001-4c/3.19.85 - 332 - the situado fiscal. As one final note, no magic formulas or quick cures have been offered in our discussion of the formula guidelines. If anything, the paper has emphasized the number of factors which enter into the distribution process, and the formula being only one of many considerations in the distribution of situado fiscal and shared sales tax. The third conclusion based on our findings is that the earmarking of revenues to specific expenditure items politicizes the revenue sharing process. Under traditional grant analysis, earmarking is an accepted mechanism to depoliticize the grant system. As we have seen in the case of the educational expenditures, the earmarked transfers to cover teachers salaries puts pressure on the F.E.R. for additional revenues to cover increased operational and administrative costs. Tied to this issue is the type of earmarked expenditure. As mentioned, teachers are unionized, and each holds a contract with the F.E.R. Teacher strikes are a threat to the F.E.R., the Ministry, and the entire national government. If the full amount of transfer is not received and the teachers are not paid, action can be taken. Under this system, earmarking creates both a pressure for additional revenues to pay for operational and administrative costs, and also acts as an incentive to close the gap between the authorized and actual receipts. It also eliminates local control and shifts the problem up to the national government distributed amounts. Earmarking brings the bitter with the sweet. It ushers political influences into the revenue sharing process; yet it makes the Ministry accountable to the F.E.R. for the transfer payment. Even if the earmarking is stopped, it is unclear if the influence of the teachers would be eliminated. Given that the expenditure on teachers salaries is around 95% of total expenditures, transfer revenue probably would support teachers salaries -- WPS/BR-001-4c/3.19.85 - 333 - whether it be earmarked or unearmarked. To a large extent, the expenditure composition of the educational sector opens the door to this political pulling. This contrasts with the health sector where there is no one expenditure item and no earmarking of revenues to specific expenditures. Because of this, the S.S.B. can apply little pressure to the Ministry. The gap between authorized and actual distribution remains larger and no additional transfers are made. The lack of accountability within this revenue sharing process illustrates the inappropriateness of the formula and earmarking to the health sector. These guidelines should be more attuned to the expenditure composition of the health sector, and not simply the same formula used in the education sector. Much of this paper has been devoted to a detailed description of the District revenue sharing arrangements. We have identified many levels of government as well as officials within each level as the decision-makers in the revenue sharing process. There is no single chain of command dictating within the process. Even the autonomous administration of F.E.R. or S.S.B. do not hold such power. By adopting this aggregate framework of analysis, it is difficult to make any single recommendation -- for each conclusion is intimately tied to the next. For example, earmarking of funds leads to closing the gap between authorized and actual receipt, yet many lead to politicizing within the allocation process. As one closing comment, the interrelationships of the many factors must be emphasized; and once accepting these factors, the need to balance the budgetary, fiscal, and political controls between the different levels of government takes on new importance in th planning of revenue sharing arrangements. WPS/BR-001-4d/3.20.85 - 334 - REFERENCES FOR CHAPTER 4 Berry, Albert and Miguel Urrutia (1976), Income Distribution in Colombia, New Haven and London, Yale University Press. Berry, Albert and Ronald Soligo (eds) (1980), Economic Policy and Income Distribution in Colombia, Westview Press, Boulder, Colorado. Bird, Richard M., (1975), "Intergovernmental Fiscal Relations in a Developing Country: The Case of Cali, Colombia". World Bank, Urban and Regional Economics Division; mimeo. Boulding, K.E. (1973), Transfers in an Urbanized Economy: Theories and Effects of the Grants Economy, Wadsworth Publishing, Inc., Belmont, California. Bourguignon, Francois (1979), The Role of Education in the Urban Labor-Market during the Process of Development: The Case of Colombia. Draft. Caceres Bolanos, Robert (1978), "La situacion financiera de Cundinamarca". Revista Camara de Comercio de Bogota, No. 30-31: 123-133. Martin Carnoy, ed. (1979) Can Educational Policy Equalize Income Distribution in Latin America? Saxon House and ILO. Hants., England. Chamber of Commerce (1978). Bogota Estructura y Principales Servicios Publicos. Benjamin Villegas and Asociados. Cifuentes, Jorge Ignacio (1979). Urban Transportation in Bogota, World Bank: Washington, D.C., mimeographed paper. Controlaria Distrital. (1975)-(1980). "Budget Data of District Administration". Bogota, Colombia. Coombs, P.H., Hallak, J. (1972). Managing Educational Costs, New York: London: Oxford University Press. DANE (1974)-(1975). Anuario Estadistico. Bogota, Colombia. DANE (1979), Cuentas Regionales de Colombia, 1960-75, Bogota, Colombia. DANE (1978). "Situado Fiscal de Education 1973-78". UDS-DE-02. DAPD (1980). "EDTU Situacion Actual y Alternativas de Solucion". Bogota, Colombia. DNP (1979), "Estadisticas Basicas para el Estudio de la Education Primaria 1964-1977", Bogota. WPS/BR-001-4d/3.20.85 - 335 - Fawcett, Caroline and Jeffrey Lewis (1981). "Revised Consolidated Public Finance in Bogota". World Bank, Urban and Regional Economics Division. Mimeo. F.E.R. (1980), Initial and Total Appropriation Budgets of District F.E.R. - 1978-1979, Actual Revenue and Expenditure Data (Closed Accounts) 1978- 1979, and Projected Budget Data, 1980. Jallade, Jean-Pierre, (1974). Public Expenditures on Education and Income Distribution in Colombia. World Bank Staff Occasional Papers No. 18. Johns Hopkins University Press. Lewin Figueroa, Alfredo, (1978). "Regimen fiscal de los municipios y departamentos en Colombia". Revista Camara de Comercio de Bogota, No. 30- 31: 51-62. Lewis, Jeffrey D. (1979). "Consolidated Public Finance in Bogota and Cali, Colombia". World Bank, Urban and Regional Economics Division. Mimeo. Lewis, Jeffrey D., Alberto Hernandez G., Marco Tulio Ruiz S. (1979). "Finanzas Publicas de la Administracion Central en Bogota, Colombia: Un Resumen Estadistico". World Bank, Urban and Regional Economics Division, mimeo. Linn, Johannes F. (1976a). "Public Transportation and Housing in Bogota: Organization, Service Levels, and Financing". World Bank, Urban and Regional Economics Division. Mimeo. Linn, Johannes F. (1976b). "Education and Health Services in Metropolitan Bogota, Colombia: Organization, Service Levels, and Financing". World Bank, Urban and Regional Economics Division. Mimeo. MEN (1978), "Educacion Para el Desarrollo Rural de Colombia". Programa Colombia-PNUD-UNESCO. Proyecto Col. 76/003. MEN (1980) "Financiamiento de la Educacion". Oficina Sectorial de Planeacion Educativa (OSPE) Ministerio de Educacion Nacional, Bogota, Enero de 1980. Ministry of Health (1979) Budget of S.S.B.,1979. Munoz, Azuero de, Beatriz. (1980). Normas Aplicables, Ministerio de Educacion. Murillo, Gabriel C., Elisabeth Ungar B. (1978), Politica, Vivienda Popular y el Proceso de Toma de Decisiones en Colombia. Departamento de Ciencia Politica. Universidad de Los Andes, Bogota. Musgrave, Richard A. and Malcolm Gilles, (1971). Fiscal Reform for Colombia. The Law School of Harvard University. WPS/BR-001-4d/3.20.85 - 336 - Netzer, Dick (1967). "Some Aspects of Lc-eal Government Finances in Colombia". Economic Development Report No. 51. Development Advisory Group, Harvard University. Oates, William (1972). Fiscal Federalism. Harcourt Brau Jovanovich, Inc. Perry, Guillermo (1979). "Estructura del Impuesto a las Ventas" en La Estructura Fiscal Colombiana, edt. de Arango Fonnegra, Sebastian, Jaime Bueno Miranda, y Florangela Gomez de Arango. Universidad Javeriana, p. 275-284. Republica de Colombia (1970)-(1979). Ley de Presupuesto (1970)-(1977). Bogota, Colombia. Richardson, Harry (1959). Regional Economics, Praeger Publishers, New York, Washington. Ruiz Betancourt, Maria Cristina (1978). El Situado Fiscal. Investigacion de Grado para Optar al Titulo de Economista. Universidad de Gran Colombia. Selowsky, Marcelo (1979). Who Benefits from Government Expenditures. The Johns Hopkins University, Baltimore, Md. S.S.B. (1980). Estudio del Financiamiento de los Hospitales Adscritos y Vinculados, 1974-1979. Stephenson, Rafael (1978). Housing Programs and Policies in Bogota: An Historical and Descriptive Analysis. World Bank, Washington, D.C., mimeographed paper. Unidad de Desarollo Regional y Urbano. "La Inversion Publica Nacional y Su Impacto En el Desarollo Regional (1970-1975)". DNP-UDRU-DER 38.