Colombia - Reforming the decentralization law : incentives for an effective delivery of services (Vol. 2 of 2) : Annexes

Report No. 15298-CO Colombia Reforming the Decentralization Law: Incentives for an Effective Delivery of Services (In Two Volumes) Volume II: Annexes April 29,1996 Countrv Opertions Countly Department III Latin America and the Caribbean Region Document of the World Bank   Decentralization in the Education Sector ANNEX 1: DECENTRALIZATION IN THE EDUCATION SECTOR A. INTRODUCTION The good management of a decentralized education system requires of the various actors that they have clearly defined duties, the authority and means to perform such duties, as well as incentives to do it effectively (World Bank (1996)). The regulating framework in place today comes primarily from Law 60/93 (Ley de Competencias y Recursos), Law 11 5/94 (Ley (;eneral de Educaci6n), the Estatuto Docente of 1979, and various other regulatory decrees. The great challenge facing Colombia is the implementation of these regulations. This becomes a strenuous task, due to the number of norms and their com- plexity, which in most cases generate confusion, doubt or uncertainty, making it difficult to transfer responsibilities. To some extent, this confusion reflects the presence of certain basic conflicts on the management of the education system. Different actors understand the regulatory setting according to their own perceptions. The major barrier to an effec- tive management of the sector is the fragmentation of duties among the different govern- ment levels, which result in a scheme where no individual is fully accountable, and there- fore. no one is responsiblforor the results (World Bank (1996)). In our study, we found barriers to good management, causing three types of problems. The first problem, and probably the most common one, occurs when those who are sup- posed to perform certain duties, lack the tools to carry them out. For instance, Law 115 confers a special role to the school in the education management process. However, this has no control over any kind of financial resource, nor does it have any instruments to in- fluence the decisions relating to the appropriation of the educational budget to the local level, which in the end, will determine their ability to develop their activities and programs. A second type of problem emerges when those who are performing a particular duty, can- not find incentives or encouragement to carry it out effectively. A case in point, the geo- graphical distribution of both national and nationalized teachers would not seem to re- spond to any criteria of efficiency or equity, causing great disparities in the student/teacher relationship between departments and municipios. Neither the Fondos Educativos Re- gionales (FER) nor the departments-the two levels invested with the relevant author- ity-face the right incentives to rationalize the personnel distribution, since it is up to the nation and the municipios to pay for the cost of such inefficiencies. A third type of problem arises as a result of the inconsistencies in the designation of duties to the different institutions, which create-voids and/or a conflicting overlapping. For ex- ample, the three levels of government are supposed to prepare plans for the education sector, in addition to the educational project prepared by the schools. Nonetheless, there are no mechanisms, provided either in the regulations or in real practice, through which such plans are coordinated so that the job carried out by a given level is not dislodged by a higher level - 1-  I)ecentralization in the Education Sector B. THE FUNDING OF EDUCATION IN COLOMBIA In Colombia, the funding sources of basic education include: automatic transfers, sectoral transfers, and own resources from the territorial entities. Of these sources, automatic transfers (SNituado Fiscal and Participaciones Municipales) are the most important in terms of quantity. According to Law 60, a minimum of 60% of the Situado and a 30% of the Participaciones must be directed to education. Sectoral transfers include national and departmental co-financing funds, the most important of which is the Social Investment Fund (FIS). Finally, their own resources come from several departmental and municipal sources, their amounts varying significantly among entities. No information source con- solidated at the national level is available on own resources and their allocation by sectors, to allow a global vision of how education is funded. A rough estimate suggests that the Situado and other national sources make up two thirds of the education expenditures, while the rest is made up of departmental and municipal resources (including the Partici- paciones). The Situado represents a share of the current national income (which in 1996 will increase to 24.5%), aimed at covering current educational and health costs. As to the educational sector, funds are managed by the FERs, institutions deconcentrated from MEN, basically charged with the payroll of national and nationalized teachers. In 1994, the amount of the Situado for education was 2% of GDP and, undoubtedly, became the most important source of funds for the educational sector. In the case of Antioquia, for example, the Situado was 56.6% of the total expenditure on education in this department (including de- partmental and municipal resources) in 1994.1 In the case of Huila, for example, the Situado represents 68% of the expenditure for teachers (the remainder coming from the department and the municipios), and probably somewhat less than the total expenditure on education. The distribution of the Situado funds among departments is strongly influenced by inertial trends, since their main purpose is the payment of teachers' salaries, the regional distribu- tion of which depends on historic decisions and events. According to Law 60, each de- partment receives at least an equivalent amount, at constant prices, to the amount ex- pended in 1993.2 Only the balance of the Situado is distributed following the formula es- tablished in that Law. Such balance was positive in 1994, but negative in 1995 and 1996. With this scheme in place, the national level has been forced to cover the FER deficits (i.e., the difference between the Situado and the actual costs) for several departments and districts. The Situado distribution is not equitable as far as the distribution of the same amount of financial funds per student served, among departments. For 1994, the Situado directed to education was $160,000 per student-in public schools, but with enormous variations among departments. Even excluding departments such as Guainia, Vaupes and Vichada, IData for Antioquia are taken from its development plan. 2 Additionally, each department or district receives an aliquot part (equivalent to 15% of the Situado), which is incorporated into the calculation of the minimum Situado. - 2 -  Decentralization in the Education Sector which receive amounts four or five times greater than the average, we find some signifi- cant differences. For instance, the Situado per student in Boyaca is twice that of Bolivar ($214,000 vs. $107,000). These differences are basically the result of two factors. For one, the apportioned part of 15% (the alicuota), involving very dissimilar sums per stu- dent, depending on the population of the institution. On the other hand, the minimum amount of the Situado, which depends on decisions not so transparent from the past. Currently, equity in the intermunicipal distribution of the Situado depends exclusively on the geographical distribution of national and nationalized teachers. There are clear indica- tions that such distribution is extremely inequitable (See Section C). In the case of Huila, for example, the Situado per student fluctuates from a minimum of around $100,000 to a maximum of nearly $300,000. The municipios of Oporapa and Tesalia, for instance, have the same number of students enrolled, but the latter one receives 2.5 more than the for- mer. The differences in the Situado per student are not compensating for the differences in municipal or departmental expenditures. In fact, those municipios with a high Situado per student tend to show a high level of expenditure per student (Graph No. 1). Quantitatively, national co-financing funds are much less important. In 1994, the total budget execution of FIS for projects in the educational sector did not even reach 0.1% of GDP (one third the intended amount). The national level establishes indicative shares for departments (equivalent to 70% of the budgeted amount), using criteria of poverty (NBI) and sectoral criteria Counterpart shares fluctuate between 10% and 50%, depending on each municipio. The approval of projects funded by FIS is made at the department level for small projects, and at the national level for all the rest of the projects. In practice, the process of submitting proposals and approving projects has become so complex that it has greatly limited the total extent of those transfers. This helps explain the low quantitative importance of this funding source for education. Considering the financial resources available for education, the department is the weakest level. In 1994, the total public expenditure of departments was only 16.8% of the national total. There is no information available on what per cent of this departmental expenditure was directed to education. The department of Antioquia, for example, allocated nearly 17% of its total budget to education in 1994, a 21.6% of the total educational expendi- tures for this department. This percentage would likely be less in other departments, con- sidering the fiscal capacity in Antioquia. The departmental authority to assign education funds is limited to its own resources, given that the Situado is managed by FER. In most departments, these resources are relatively low, compared to those for the nation or the municipios. Once it becomes certified, the department will have the authority to assign the Situado funds, including the decision on what percentage is allocated to education and how it will be distributed among municipios (following the criteria previously set forth by the departmental assembly). In practice though, the salary costs of the existing teaching staff paid via FER will continue to deter- mine the use of the Situado. In principle, however, a certified department will have the authority to reassign teaching personnel and to redistribute, consequently, the Situado among municipios -3 -  Decentralization in the Education Sector The current and future increase of municipal funds (the result of the increase of Partici- paciones) suggests that the municipal level will play a significant, growing role in the funding of education. In fact, only the municipios has a significant discretionality in the use of funds for the sector. It is estimated that, if the assignment of minimum percentages is achieved even without using own revenues, municipios will allot to education an amount which will grow from 0.3% of the GDP in 1994 to 0.8% in 1998, and close to 1% beginning the next century.3 No figures are available that show the total amount of expenditures in education (including own funds) of the municipios as a whole. In the case of Antioquia, (which, as we saw, assigns a high percentage of the departmental funds to the sector), municipal expenditures in education are 21% of the total. This percentage will grow strongly in the years to come with the amount of transfers, and is likely to be higher in other departments having a fiscal capacity significantly lower than that of Antioquia. The distribution of funds among local schools and zones depends almost exclusively on the transparency of the process of assigning the local budget. Currently, there are only a few formal levels where the educational community is participating in such decisions. The JUMEs are only beginning to be formed, and have no formal power in this issue. Funds at the school level are scarce. Primary schools have only voluntary contributions from the community, through Parents Associations and a few other resources generated by school restaurants. Schools also have some funds provided by the Fondos de Servicios Docentes, coming from annual tuitions, and fees related with special services. There are no clear ref- erences to cases where part of the local budget for education is channeled directly to schools with the participation of the School Boards. C. THE HUMAN RESOURCE The management of human resources in the education sector has experienced repeated changes through the years. Ongoing incidents of centralization and decentralization have left a complicated, difficult to manage, education picture as a legacy. The nationalization of educational services in 1975 charged the national level with the financial, administrative and political responsibility, but it left the appointment of teachers in the hands of gover- nors. The decentralization set out by Law 29 of 19894bestows upon mayors the duty of managing both national and nationalized teachers, without modifying the teaching staff or its operating mechanisms. The departmental quality of this staff was reconfirmed by Law 60, which nonetheless, charges the muinicipio with the management of education. It is no surprise, therefore, that we find confusion today on the nature of regulations about accountability in the management of human resources in the sector. Most local govern- 3Transfers were 1.7% of GDP in 1994. This year, the minimum of 30% allocated for education was ap- plied only to 50% of the total amount. For 1998, transfers are estimated to be 2.6% of GDP. See DNP (I 995c). 4Law 29 of 1989 assigned mayors the duty of appointing, transferring, removing, controlling and, in gen- eral, managing the teaching and administrative staff of educational institutions of the national level . ac- cording to existing regulations of the Fstatuto Docente and to the availability of open jobs in the staff. - 4 -  Decentralization in the Education Sector ments accepted the administrative responsibilities established by Law 29, and have per- ceived changes occurred over the past few years as part of a process of educational 'departmentalization'. At the departmental level, difficulties associated with the manage- ment of human resources discourage the assumption of responsibilities provided by Law 60. FECODE, the teachers' union, considers the so-called 'departmentalization' an alter- native to a much repudiated 'municipalization'. Meanwhile, the central level not only continues to fund most of the teachers but also continues, in fact, to assume the political and administrative costs involved. At least five types of teachers coexist today, depending on the institution they were ap- pointed by, and that which pays them: (i) national: appointed by the nation and paid for with funds from the Situado; (ii) nationalized: originally appointed by the departments and paid for with funds from the Situado; (iii) departmental: appointed and paid for by the de- partment; (iv) municipal: appointed and paid for by the municipio, and (v) others, includ- ing teachers on a per-hour basis, educational solutions, etc. Authorities at the national level do not have the information that allows them to determine the number, nature and distribution of teachers throughout the national territory. The number of teachers is esti- mated at approximately 250,000, over two thirds of whom are paid for with funds from the Situado. Some 50,000 teachers are being paid for with local funds, co-financed in many cases by the nation and the department. The basic regulations for the management of teaching staff are found in the Estatuto Do- cente. Starting with Law 60 and Law 115, teachers nationwide are subject to this legisla- tion. The teacher is classified in an escalaf6n, and is promoted within it based on the number of years served and on courses certified by the authorities, regardless of who their employer is (be it public or private, at any level). Promotions are not linked to the teacher's performance, or to any incentive policy the system might adopt. The same regulations apply to school management. Being a principal or director does not come as the result of a process which evaluates a teacher's performance and skills, but is rather a further echelon in the educational career. The disciplinary regime is basically ruled by the Estatuto Docente. The processes begin at the institution or educational nucleus level, then moving up to the municipal level. It is up to the Junta de Escalaf6n Departamental (JED), and ultimately to the national level, to enact the corresponding regulations. Regarding the management of human resources, Laws 60 and 115 issued a new regime, which in fact, would apply upon the department certification. Its basic components are: The teaching staff paid for with funds from the Situado is departmental. This means that the department (or district) regulates the contests, appoints the teachers, and as- signs a municipal teaching staff according to the needs in coverage, quality and effi- ciency. Transfers of teachers between municipios require a previous concept from the JUDE. -5 -  I)ecentralization in the Education Sector * The municipio manages education.5 This means the authority to transfer teachers within the municipio (following a previous concept by JUJME), to approve, provide incentives, and grant licenses, among other decisions. The municipio may hire teach- ers using its own resources, in which case it has the joint administrative authority over the human resources. * Municipios with a population over 100,000 may become certified and assume all the duties pertaining to a department. This requires the authorization of MIEN, as well as the approval of the departmental assembly. Additionally, municipios with populations less than 100,000 may receive the authority to appoint teachers by delegation from 6 the department _ This scheme starts with a situation where the distribution of teachers, determined by his- toric patterns, is extremely inequitable and inefficient. Thus, certain departments have too many teachers for their current needs in coverage. Furthermore, the distribution of teach- ers paid for by FER in municipios within one department, does not seem to respond to the criteria of efficiency or needs. The magnitude of the differences between munici'pios may be viewed in Graph No. 2 for Cundinamarca. The rationalization in the distribution of human resources is the major, and probably the most complicated, issue in the process of decentralizing education. In the current situa- tion, the national level faces a paradoxical predicament: whereas it funds most of the teacher salaries, it has only few tools to control growing costs. The salary level is deter- mined in a bilateral negotiation (in which there is no participation of departments or mu- nicipio.s), where the teachers' union would seem to carry the critical leverage. The level and composition of employment is totally out of control by the national level, which has a very limited capacity to control absenteeism, promotions, transfers, and other key vari- ables. The national level, therefore, must virtually assume bills it has to pay. Nonetheless, neither the departments nor the mniicipios have the authority to change the level or dis- tribution of teaching staffs paid for by FERs- D. MAIN CONTRADICTIONS It is not easy to understand the operation of a system as complex as the Colombian educa- tion sector. Regulations are complex, and reality is extremely diverse. In general, four basic contradictions may be identified in the way the educational system works. First: The municipal level plays a central role in the management of the sector, but has a very limited authority in managing the human resources which, undoubtedly, is the main input in the educational process. This is a basic principle established by Lawv 60 (article 2). Lawv 115 (article 153) provides that the ad- ministration involves. among other things, the appointment of, removal of, transferring, sanctioning, in- centivating. giving licenscs and permits to teachers, managers and administrative personnel. 6This may occur in at least two X avs The department may assign this responsibility, or the <nunicipio may request it. once certain conditions established by Law 60 have been met. -6 -  Decentralization in the Education Sector From the viewpoint of regulations, the municipio is responsible for administering educa- tion. From a practical viewpoint, it is the best positioned governmental level to do it: it is the level closer to the community, and the only one having financial resources for educa- tion which may be assigned with a certain discretionary margin among the different activi- ties (Teachers, textbooks, infrastructure, etc.). Over 70% of the municipios have assumed the responsibility of the educational sector ac- cording to Law 29. There is no study indicating, in a systematic fashion, how the as- sumption of such responsibilities has influenced the management of education. In general, the perception is that in those municipios which assumed education, mayors somehow feel compelled to pay attention to a sector they are responsible for. In contrast, in those mu- nicipios not assuming education, a mayor with no particular motivation to educational matters would seem to have a good justification to provide only a marginal care to the problems in the sector. Nevertheless, municipios have a very limited authority to manage the human resource. The vast majority of teachers are part of a staff managed by what is FER today, and which will be managed by the departments in the future. The geographical distribution of such staff does not respond to criteria of equity and efficiency, but to historic patterns. The same is true of the distribution of resources through the Situado. The power a municipio has to determine the number and quality of teachers is limited to a great extent to the fi- nancial capacity they may have to hire teachers using its own funds. The authority they have could be curtailed even more, should the certification of departments be interpreted in practice as a 'centralization at the departmental level'. An efficient administration of education requires municipal authorities to be able to define and adjust the distribution of the teaching staff within the municipio. Currently, mu- nicipios which adopted Law 29 have this power not only for those teachers paid for with their own funds, but for those paid for with funds from the Situado. Even though, in our opinion, the regulatory framework is clear in that the munmcipio has the authority to transfer teachers among schools, given a previous concept by the JUME, just the opposite interpretation has developed among certain groups. Without this power, the supposed managerial role of the municipio would become merely nominal. The municipal authority to manage the daily activities of education, is very restricted. Promotions of teachers are not associated to their performance, and becoming a principal is not the result of a process of performance and skills evaluations, but basically the effect of criteria such as time spent in the service. The application of disciplinary processes, which is conflictive per se, depends to a great extent on a departmental level (JED), which lacks the incentives to implement it effectively. From the viewpoint of the mu- nicipio, disciplinary processes represent.high costs which overwhelm the possible benefits they may provide. It seems to be greatly associated with bottlenecks at the JED level. The composition of JEDs is such that, should the commitment of the departmental ad- ministrations exist, they would have the ability to enforce current regulations. In general, however, a disinterest on the part of the authorities may be observed which, broadened by the low administrative capacity of JEDs, results in very poor management. This is not at all surprising. From the viewpoint of the departmental authorities, costs related to 'discipline enforcement' is high in terms of conflicts with the teaching staff The benefits - 7 -  Decentralization in the Education Sector are not quite obvious: the departmental level does not profit from the resulting 'economic benefits' coming from an improved discipline, for instance, the benefits of less absenteeism (appropriated by both the FER and the national authorities). Second: Regulations emphasize the role of the department in issues such as planning and human resources, but in practice, this is the governmental level with the lowest discretionary degree in the management of financial resources. With only a few dis- cretionary funds, the department has a low capacity to develop educational policies. The implementation of the new decentralized management scheme cannot be devel- oped until the departments become certified, a process made difficult by reserva- tions both at the national and departmental levels. Law 60 assigned the department a critical role in the management of education, as the ba- sic planning level which manages the teaching staff transferred to it by the nation along with the Situado funds to help fund it. But, relatively, the departmental level is extremely weak financially. Most of the Situado funds are committed to the payroll. Consequently, managing the Situado involves very little discretionary power. Compared to municipios, the department may count solely on its own limited resources to develop educational policies. At present, the certification of the department7 by MEN is one of the major pitfalls in helping implement the new regulatory scheme. Without it, few incentives exist to produce the rationalization of the teaching staff, a critical aspect to improve both the efficiency and equity in the system. The advancement of this process would seem to be influenced both by the fears at the national level and the hesitancy shown by departments. The result is a scheme in which the national level continues to exercise highly relevant duties through FERs, without having the effective capacity to change the way the system is really work- ing. Municipios, on their part, face the uncertainty of what kind of consequences for their capacity to manage education the mentioned certification will bring about. The first condition for an effective management is that departments have the capacity and incentives to adjust the total number of teachers to levels suiting the real coverage needs. Depending on each case, this will require to increase or decrease staff reductions. Finan- cially, most of the departments will rely on the Situado for the payroll. In certain cases, though, they will start with a situation where the Situado is insufficient to cover their cur- rent payroll. For most departments, this generates doubts on the benefits the certification may bring about, as well on their financial capacity to achieve the objectives of coverage expansion, which are likely to come mostly from municipal funds. Besides, those depart- ments with a surplus of teachers in relation to their enrollment, will find it difficult to re- duce employment under a very restrictive legislation. Such adjustments should probably be effected in the long term, based on attrition and coverage expansion. A second condition is that the distribution of teachers among municipios be equitable and flexible, so that it may respond to the major demographic changes observed in Colombia 7 The certification process provided in Law 60, is regulated by Decree 2886/1994, which establishes eight conditions that must be satisfied for the MEN to certify the department. Recently, MEN certified a first group of departments. -8 -  Decentralization in the Education Sector Under the current regime, the department has no incentives to rationalize the inter- municipal distribution of teachers, since these are paid for by the national level which, in fact, is forced to cover wage costs. Even if it had the motivation, the department has no full power to adjust the teachers' distribution, since FER and the municipios exercise a certain control over it. Once they are certified, and if the nation makes the principle of fund allocation binding only based on a formula, departments will have greater incentives to develop a better staff distribution: given their budget, the only means to guarantee cov- erage is through inter-municipal transfers. Third: The school is seen as the basic educational unit developing its own Institu- tional Educational Project (PEI) and with a high level of autonomy, but it is not as- signed either the financial funds or the true authority to manage teaching or admin- istrative human resources. Law 115 sought to strengthen the role of school as a semi-autonomous unit charged with the definition of pedagogical policies, the identification of needs, as well as action planning and management. PEI is a key element in that process. Basically, the educational com- munity (teachers, staff, parents, and students) defines the pedagogical objectives taking the guides generated by MNEN as a reference, and based on these, establishes their needs for human resources, materials, processes, etc. However, establishment lack the financial resources to implement those decisions. There are no formal mechanisms which schools may use to directly access municipal resources or co-financing funds for this purpose. Thus, there is a risk for the PEI to become a purely theoretical exercise. It is remarkable to see the lack of formal mechanisms which ensure that the appropriation of financial resources in the educational sector at the local level, re- sponds to the needs and priorities emerging from the PEIs. Insufficient statistical infor- mation on education at all levels, restricts even more the possibility to develop adequate plans. Law 1 15 gives the school a central role in the management of human resources. That role, however, contends with the limited authority principals really have. From the formal point of view, principals have no authority in areas such as the teacher selection, appointment and transfer. Regulations give them the authority in issues such as evaluation and control, discipline, pedagogical orientation and task assignment. In practice, however, their power would seem to depend basically on their persuasion ability, since the way the Estatuto Do- cente is applied, it is extremely restrictive of their capacity to enforce their formal author- ity. In addition, they do not have any formal tools allowing them to provide incentives to teachers, since they bear no influence in determining wages or promotions. Thus far, there are no formal mechanisms to evaluate teachers' performance. Ultimately, the only means a principal has to reinforce his/her administrative capacity is to 'leverage' in informal alli- ances with parents, the head of the educational nucleus, the secretary of education, or the mayor. In many cases, the principal himself has relatively few incentives to become committed to a good management of the school. Basically, his appointment follows a process preestab- lished by the escalaf6n, and his permanence in the job is not subject to evaluation of his - 9-  Decentralizationi in the Education Sector leadership performance. Why, then, push initiatives forward that could meet head to head with the school teachers and management staff Law 115 grants the School Board an important position in school governance. In prac- tice, few boards have been formed. Even though the motivation of the educational com- munity may lead to many of these to be formed, their survival will depend critically on the real power they may exercise. This Law enables them to 'approve the income and ex- pense budget of their own resources', which gives them substantial authority. However, since there are no significant sources of funds, this authority may disintegrate in time. Fourth: Although decentralized management is emphasized, few mechanisms of control and accountability are in place on the use of resources and results. Community participation is not reinforced by means of controlling levels in the assignment and use of financial resources. The inadequacy of information systems at all levels hinders both decision-making and result evaluation. The good performance of a decentralized system depends primarily on the control exer- cised by the citizens and other government levels over the use of financial resources and educational results, in issues such as coverage and quality. This control requires the avail- ability of information that may allow the result evaluation and decision-making. The lack of information on resources and results is very obvious, and may be observed at all levels. The national level finds it difficult to know for certain how many teachers there are nationwide, and how much municipios and departments are spending on education. Departments, supposed to develop educational plans for their regions, often lack even the information on the number and status of their own schools. The situation is not much different at the municipal level. Overall, it is very difficult for any citizen today to know the efficacy of his authorities in the use of public resources, since there is no systematic evaluation of the performance of teachers or students. Law 115 assigns a critical role to participating levels such as Juntas de Educaci6n Mu- nicipales (JUMis) and .Iuntas de Educaci6n Departamentales (JUDEs), as well as Foros Educativos. However, these levels only act as advisors. For example, the JUME, the major participating level in the sector at the municipal level, has no formal influence on the way resources of the municipal budget are assigned to education. This not only limits the ability the community has to influence such process, but also discourages the very forma- tion of Juntas. E. MAIN RECOMMENDATIONS The following are the main recommendations for changes at the three governmental levels, in order to improve the management of the educational system. The role played by MEN. Within the framework of a decentralized educational system, MEN plays a key role in facilitating management by the territorial entities, and in develop- ing a national policy through regulations and criteria. The Ministry is not an efficient en- tity in playing the role of a manager. In order for the decentralized management to de- velop, it becomes necessary for the process of certification of territorial entities to be ac- - 10-  Decentralization in the Educatiotn Sector celerated. The effective transfer of responsibilities is a necessary condition for MEN to be able to perform its new role as guide, leader and promoter. At the national level, doubts and fears associated with the transfer of responsibilities arise partly due to a lack of confidence in the capacity of departmental management. This im- plies a very high cost, since it leaves the educational system undefined. This cost is probably much higher than the alleged risks of decentralizing. Whereas, in many cases, departments truly have a low capacity, it is unlikely that this capacity may develop in ab- stract, but on the contrary, only if they really discharge the duties assigned to them by law. In many cases, the limiting factor will not be the timidity of the central level, but fears at the departmental level. The national government will have to use a combination of incen- tives and persuasion to neutralize such fears. In certain cases, the departmental reluctance is associated with the financial costs expected from certification, since under existing conditions, departments would run into a deficit. In such cases, it must be considered to what extent it is justified to negotiate a transition scheme, rendering these costs less bur- densome. For those departments having a surplus of teachers, it will be necessary to de- velop a practical mechanism to make the transfers to other jurisdictions. Furthermore, the certified departments could be rewarded, by offering them a privileged place in the design of educational policies at the national level, or a preference in the access to co-financing resources, as provided by Law 60. We believe it is essential that the certification process should extend to municipios of populations over 100,000. These will require applications by municipal councils and ap- proval by departmental assemblies. The most controversial issue may be the approval of the distribution criteria on the Situado. However, once an agreement is reached on the matter (a subject which has to be resolved even without the municipal certification), we believe it will be of mutual interest to both the department and the municipio, to obtain the approval of the municipal autonomy. The national level may collaborate by making the certification easy, and by promoting the arrangements between the department and the miunicipio. The incentive program for teachers and schools, approved this year by MEN8, is a very innovative and potentially powerful means to motivate the best performance among teach- ers and quality improvement at the school level. Its true value goes beyond its direct im- pact, and it will depend on the demonstrative effect it may have for the communities and municipal authorities. This is why its development and implementation are so important in the near future, in a way that it may involve the community and local authorities. We be- lieve this to be an area of priority action for MEN. Promoting the use of information supported by adequate systems is a key element to this approach, both for the management by territorial institutions and schools, as well as by the nation. It is our belief that the improvement in educational information (on the use of fi- nancial, human and physical resources, and on the performance of teachers, students and schools) is an area where the national government must lead, assisting and promoting ini- Sec Ministerio de Educaci6n Nacional (1995). 1 1-  Decentralization in the Education Sector tiatives at the other levels. A reform of the current statistical educational scheme requires the schools to feel they are committed to quality work. MEN should consider the intro- duction of changes in the information gathered, by assuring this information is useful to schools for their daily operation, and in such a way that it may be aggregated to allow the higher administrative units to gain a descriptive view of what is going on in their territo- ries. It would also be necessary to assign the municipios the duty of summarizing the in- formation generated by the schools. The department-municipio articulation. The excessive fragmentation of authority is likely to be the highest risk in the implementation of the new administration, leading to a scenario where no one feels responsible for the management of the sector. Consequently, the major challenge is to obtain coherence in the management of human resources. If the municipio lacks the tools required to manage the human resource, and the certification process is interpreted as the loss of responsibility by the local authorities, the doors will open for an extremely inefficient management scheme to come in. The management of important financial resources coming from the Participaciones could cause serious distor- tions in the assignment of resources, inasmuch as the municipio does not have the author- ity before its citizens to either manage the human resource or the responsibility of the edu- cational management. The department is too distant yet from the schools to expect acquiring the management of human resources. Only the mnunicipio will be in a condition to do it. This implies that the municipios should, undoubtedly, have the administrative authority bestowed upon them by law. (Including aspects such as transfers between schools, discipline, etc.) Of course, in the case of teachers paid for with own resources, such authority must include their appointment and selection. On this subject, we find today that quite a confusion pre- vails as to what exactly the departmental certification involves for municipios which have already assumed education according to Law 29. MEN should consider organizing a clarification campaign directed to departments and municipios. Of most concern is the climate of discouragement which might be generated among the municipal authorities, who consider the certification as a loss of responsibility. We believe the commitment of the mayor and the municipal council is critical for the new administration to create an ade- quate management of educational services. Such commitment will be difficult to maintain if the municipal level does not clearly understand it still has the basic responsibility of managing education. Another essential point in the department-municipio articulation has to do with the proc- ess of rationalizing the departmental teaching staff and their distribution among mu- nicipios. The great risk is that this process may be guided by not so transparent criteria, leading to inequity, inefficiencies and conflicts. This process is directly associated to the assignment of the Situado among municipios, which the certified departments should do. Such assignment should be done on the basis of logical, objective criteria (such as the population in school age and its density, as provided by Law 60). The strict application of such criteria will force a rationalization process of the teaching staff, which might involve transfers and adjustments between municipios. The emphasis in the rational, transparent distribution of the Situado among municipios will have the additional benefit of facilitating the administrative and planning action by the municipal authorities, who will thus have a - 12 -  Decentralization in the Education Sector clear picture of the balance they will have to establish among the several sources financing their teaching staff An efficient administration of the human resource will not be possible, unless the discipli- nary processes do not become more agile and responsible. Even though there are some disagreements, it would seem as though the Estatuto Docente shows a margin of impor- tance not used because of an apparent lack of motivation and a poor capacity of JEDs and the National Board. As long as the members of such boards do not assume their full re- sponsibility, this will be a permanent bottleneck for local and school authorities. This re- quires a strong commitment on the part of governors and other authorities at the depart- ment level, as well as a more active participation by the municipal authorities. In the area of educational planning and teaching training, Law 11 5 omitted the role of the municipio, which still is responsible for the key financial decisions relating these issues. This unbalance must be corrected recognizing, however, the need of a regional agreement between the municipios and their department. The balance of authority at the local level. At the regulatory level, there is an adequate emphasis on the community participation and the role of the educational community as the manager of the educational process. In practice, however, this may be limited by the lack of access to resources by the educational community. The role of this latter must be strengthened, clarifying their responsibilities and giving them the instruments to implement them. The creation of JUMEs and JUDEs will be slow as far as the idea of belonging to them is not made attractive to the different actors. This is why the third component of our proposal is to strengthen both their role and the authority law has given the community. The creation of PEIs is a key element in the new regulations. Beyond the practical diffi- culties which may be encountered at the school level to prepare the PEIs, the main prob- lem we have identified is the inexistence of concrete mechanisms, through which these can influence key decisions at the municipal level. Law 115 does not establish an important role for the municipal plan, and is unclear as to how the PEls must associate with it. In this sense, we believe it necessary to establish a closer connection among PEIs, the mu- nicipal plan, and the assignment of resources. This may be obtained: (i) by giving the JUMEs an important role in the development of the municipal educational plan, taking the PEIs in consideration, and (ii) by bringing the JUMEs closer to the process of preparation of the municipal educational budget. These two steps should contribute to make the deci- sions concerning municipal investments in the educational sector more transparent, and respond to the needs demonstrated at the educational community level. The School Boards are a major level in the school administration. We believe it necessary to strengthen them by assigning them higher authority in managing financial resources. In principle, the municipios should channel part of their funds for education to maintenance and provision of the Fondos de Servicios Docentes administered by the School Boards. This will have the double benefit of strengthening this key level and of probably creating the opportunities for schools to be in a better maintenance position. The limitation of op- eration of such Fondos at the school level limits their impact as instruments to manage re- sources by the educational community. In this sense, it would be convenient to expand - 13 -  Decentralization in the Education Sector their applicability to schools. The FIS may promote these arrangements by offering better conditions to municipios, in order to channel their local counterparts through these funds. It is also necessary to strengthen the role of principals in their role of school managers, particularly in the issue of human resource. (discipline, task assignment, etc.) This in- volves a greater authority, but also a greater responsibility to the community. To this end, it would be recommendable to consider some changes in the selection processes for the principal, giving a greater influence to the educational community (for example, selection made by the School Board, and for a fixed term) While these mechanisms exist, the prin- cipal should hold a higher authority to exert the disciplinary duties at the school level. - 14-  sluepn)s 00&Y L OOOZ L 0000 1 0009 0009 000& OOOZ 0 .1. ~~~~~~~0 ,, I . ,, .. .. . .. . . .. ...1 .... .. ,,.1 ... ....... i . , ,O .... ''' . '.: '': . ,' ' , ''' ...... '''' ' .. ' ":''''' '',,,.''''...,,'.. . .'': ..... .~~~~~~~~~~~~~~~~~~~~~~~~~~~. .. ' -''-...... ... :' .':.-.'-...... .:.'. '' :::'. ... '' '' ' ' '' .... . ''' ... -'' '' - - ... ' .'........... .....-: ... . -:'......''' , ', , ^ . s '~~~~~~~.... .'. . . ... . ..... l ua pni si open} !s .: . ', . + . .. . . . . . -- - . . . . - . . ., . . . . . . . . "~~~~~~~.. ... .. .. ... . ... ... .. .. . ... ... .. . ..... .. '~~~~~~~~~~~~~~~~.... .. .... . ,,.C.. . OOOO LE f . . . . . .. . . .. . . . ~~~~~~~~~~~~~~~~~~~. . . .. . . .. . . . ... ''' ,''^..00000 ...... .. .... . ' . ....'.' ,, .. 00000 . .... . . .000008 . . . ... .........0 . 4. %*,~~~~~ .... ..... juapnIsfeJn1l!puadx3; . ..: ..0.... 0001' u'....., IuGpnllS/op2nl!S . . ........... 0.. 4 . ..~~~~~~~~~~~~~~~~~~~~~~~~~~~... ... so!dp*iunwl ueaM4aq SGOUOJGJa!G :eI!nH 1 qdewg  Graph 2 Cundinamarca: Differences between municipios 70.0- .. ... .-- .......... .. .: .4 ; - .....:0;:: . . E :: E,,,E... . . . . . E . . .. . . . .,,,. E. .. . . . . . . . . . .. . . . .. :::: E:ER ':iC iEE.E} . ...EE: ;....... : ... ....... . ... . . . . . . .. .. . . . . . . . . . . . . E iE:ER . . : : . .E :. i E.ERE::: : E :E E :: : i l. . i E - i i : i , , : iR iE:E:...... ... ... ..... . ... ... e SEEi. :. ! ... E.. . ::: .:. i.... ............ ...... :S... . . . ...'. . : .....:E ~~. . . . . . . . . . . . . .. . . .-. . . . . . . . . . . . . E :ERE~~.. .. R .. .. . : :.: ..: . ::: . .:: . . .: E E... I, E:: ... ...... .. .... .. ; i EEE .: iS . .... . . .. . . ..o ...; ...... .0 .0 E . ...E ... ; ..!E .... .ffE.0 ..: .....: . . . ! . . ... .... .... ... .. .ER . ..E :0;E. .EE.fEE...Wi E !0 i ES0! E:RS . ..... . ... ......... . . ..... ..... . . .. i E .. .. . .. O .. . . ... .. . . .. .. . .: : .. .. .. ....... o~~~~~~~~........ 200 .. 0 ..... ..... l.00. 00 40 ... . .. ................udent . ~~~~~~~~ ~ ~ ~ ~ .. .....16.....-... . . ..  Rejorms in the Health Sector ANNEX Il: REFORMS IN THE HEALTH SECTOR A. INTRODUCTION. The issues concerning the health sector in Colombia are intimately linked to a radical reform of how health care is to be financed, as a result of Law 100 of 1993. One critical factor to be considered because of its impact on the design of intergovernmental transfers is that everything about Law 60 is territorial-money flows from one level of government to another corresponding to the responsibilities of a governing unit. In contrast, the POS mandated by Law 100 (see below) is individual and non-territorial, this is equally true for the full POS, to be delivered under the Regimen Contributivo, and the initially half-size POS-S or plan for the poor, to be delivered under the Regimen Sub- sidiado. The entities which are supposed to receive the funds for financing POS and POS- S are simply insurers, and while they may concentrate their membership in a particular geographic area they are non-territorial and can enroll members, and make payments for medical services, anywhere in the country. A municipio-or a consortium of small municipios, or in principle even a depart- ment-may organize its own, publicly-owned EPS under Law 100, but that EPS is required to be legally and administratively distinct from the government that organizes it, so while the owner of the EPS may be territorial, the EPS itself is not. Similarly, the fact that the value of the capitation payment may vary from one part of the country to another, to take account of differences in the cost of providing the services composing the POS or in the frequency with which those services may need to be provided, is only a geographic distinction, not a territorial one. Differences in the ownership of an EPS do not matter for this distinction. The shift from a territorial to an individual basis, however, applies only to the POS (and POS-S), not to every kind of health expenditure mandated by Law 100. The plan for spending on public health measures-the PAB-cannot be administered through EPS because some of its components are provided to the public or to the environment generally rather than to identifiable individuals. The PAB therefore continues to be territorial. This annex summarizes those aspects of Law 100 and its implementation that are more relevant to the system of intergovernmental relations established by Law 60. B. THE BASICS ABOUT LAW 100 Law 100, enacted in December 1993, mandates radical changes in the way health services are organized, financed and delivered. If implemented, the proposed reform would greatly expand coverage and reduce or eliminate differences in access because of income, enlarge private insurance and medical care, and emphasize preventive services in publicly-provided care and financing. The start-up date for the new system was January 1, 1995, with implementation of all components to be phased in over a six-year period. Law 100 establishes seven basic principles to guide the delivery of health services to all Colombians: - 17 -  Reforms in the Health Sector * Separation of financing from service delivery and freedom of choice. The institutions responsible for insuring beneficiaries and handling the administrative and financial as- pects of coverage (Empresas Promotoras de Salud, or EPS and Empresas Solidarias de Sal/ud or ESS) are separated from those that will actually deliver services (Intstituciones Prestadoras de Servicios or IPS). Consumers are free to select the in- suring organization to which they affiliate. EPS are required to offer their members a choice of at least two IPS. * Universal coverage. The new system seeks universal health insurance coverage through its provisions for obligatory coverage of workers, coverage of dependents, subsidized services for the poor, and the fostering of competition for beneficiaries among multiple service provider organizations, both public and private. By expanding health insurance coverage to all in individuals regardless of occupation and income, and by establishing a market comprised of multiple insuring organizations and health care providers, Law 100 enables the poor to obtain health services on a subsidized basis. * Comprehensive basic health package. Beneficiaries will be entitled to a basic package of health services (Plan Obligatorio de Salud or POS) without exclusions due to pre- existing conditions. In the initial implementation phase, however, the service content of the POS would vary by type of beneficiary (i.e. the poor and the unemployed would be guaranteed a subsidized package with fewer services, called POS-S, than the one to be made available to those beneficiaries who pay health insurance contributions). Complementary plans providing additional services will be available to those who wish to purchase them. * Solidarity and cross-subsidization. The system has two schemes. The first, a contribu- tory scheme (Regimen Contributivo), will be financed by obligatory social insurance contributions equaling 12% of wages paid by workers (4%) and their employers (8%). The self-employed are required to contribute 12% of their income. The second, a subsidized scheme (Regimen Subsidiado) for the poor will be financed through general government revenues at the national, departmental and municipal levels and I percent- age point of the obligatory social insurance contributions that will serve as a cross- subsidy (to be channeled to the special Solidarity and Guarantee Fund or Fondo de Solidaridady Garantia, which will help finance the POS-S for the poor). Under the current system, a share of transfers to municipalities and departments must be dedi- cated to finance these subsidies. * Cost control mechanisms. The system makes use of prospective fixed per capita payments to the insuring organization for each beneficiary which vary according to the person's age and sex. The EPS will contract with providers to deliver care, and pay them on any basis they chose (including capitation for a set of services, DRG or fees- for-service). * Incorporating private providers into a single system comprised of both public and private insuring and service delivery organizations. The new system seeks to integrate the private sector into the unified system by offering private groups the opportunity to form EPS and ESS and receive the fixed contribution per affiliate, and to provide complementary health plans which will be financed outside the obligatory salary con- tributions and which cover improved technology, amenities, hospital ancillary services, - 18 -  Reforms in the Health Sector etc. By increasing competition among service providers for contracts with the insuring organizations, the system also intends to reduce the public sector's monopolistic posi- tion as provider of health services to low-income persons and foster greater opportu- nities for private sector service providers. The mandatory and free provision of a basic public health program. In addition to the provision of curative health services, the law contemplates a public health program (PAB) covering aspects such as sanitation, health education, and vector control to be financed by the national government. Providing the PAB is regarded as a responsibil- ity of both departments and municipios, for the sound reasons that some public health activities may have to be conducted over an area larger than a municipio, and that some municipios may lack the capacity to provide those services unless they get help from the departments. C. IMPLEMENTATION OF THE REFORM Despite the great detail contained in Law 100 and the numerous companion decrees prepared in 1994 for its implementation, it is clear that much remains to be worked out to effectively operationalize the system. The difficult question is how to manage the transition from a regime of financing supply directly in public facilities, to one of financing demand directly and making facilities, public or private, get their funds indirectly by selling services to EPS. The law itself established a transition period of two years (ending in January 1996) during which public hospitals and clinics are protected from change and guaranteed the same budget, in real terms, as before. That provision could be used to buy time to get ready for the change to demand financing, but it neither specifies how to make the change, nor gives any particular incentives to prepare for it. The large number of decrees to spell out details of the law, issued in the summer of 1994, also did not adequately specify the transition. The start-up implementation of the health reform has been slow in part due to initial doubts and concerns by the Ministry of Health (MOH) during the first year of the new administration. For example, the so-called Transition Decree (No. 2491) of Novem- ber 1994 created a new institutional entity, the transitory public EPS, still without helping promote the change to a demand-driven system. It also fomented the registration of the poor as beneficiaries of those public EPS, without really giving them a choice of insurers, and thereby may have done two other things which are inimical to the reform: reinforced the role of mayors as being responsible for people's health care, when according to the law they are supposed to be responsible only for people's affiliation to the level of insurance to which they are entitled; and by emphasizing the distribution of 'carnets' (carnetizaci6n) to beneficiaries without really changing anything else in the system, gave people the impression they were already enrolled according to the law, which probably makes it harder to affiliate them with competitive EPS and go through registration a second time. Nevertheless, over the last few months, progress has been made in defining the transition and in setting up the subsidized regime. While the transition involves many complex issues, the following are perhaps the crucial ones from the point of view of inter- governmental relations. - 19 -  Reforms in the Health Sector * Implanting the POS-S (that is, subsidies to the demand side) requires a clear definition of who qualifies (coverage), for what services (the amount of the subsidy) and the way they are enrolled and the money is transferred to providers. In order for people to be classified as subsidy recipients, it is necessary for them to register with a local entity (municipio or department) and get a carnet identifying them as eligible for the basic package of services. The challenge is how to manage simultaneously the processes of registration (carnetizaci6n) and affiliation (i.e. sign-up of the population with EPS) of the subsidized recipients in order for them to receive the basic package of services. Currently, the identification of beneficiaries is done through a system (the SISBEN) that classifies individuals as poor based on a variety of indicators. The process is slow because of the difficulties of implementing this system. In addition, there are doubts about the adequacy of the targeting methodology. An alternative is to move in the direction of a system in which individuals self-select themselves into the subsidized regime. This option would involve some type of rationing mechanisms for the services not covered under the subsidized program. The actual enrollment of beneficiaries, regardless of who qualifies, will require an active involvement on the part of municipal authorities acting as agents of the national government. To accelerate this process, it would be important to offer incentives to municipios to enroll beneficiaries. Under the current system, however, municipios must also pay for these subsidies out of the transfers they receive from the national government (15% of the Participaciones Municipales). In fact, there is very little justification for this complex way of transferring resources to an EPS. A direct transfer to the EPS (with the municipio simply reporting the names and addresses of the identified beneficiaries) would make the process of enrollment more transparent and probably faster. * How are public hospitals prepared to operate under the system of demand financing? What do they need in the way of better management, information, cost systems and ability to bill for services ? To get ready for the transition, hospitals and clinics have to learn how to keep track of costs and bill for services-and they are unlikely to do that, without the perceived need to learn because their revenues are going to depend on it. However, it should be clear that the establishment of a cost system and the design and pilot testing of new resource allocation systems are complex tasks that require time and specialized knowl- edge. For example, the design and pilot testing of prospective payment systems for hospitals and municipal-run primary care facilities in Chile was carried out over a four- year period. How can public hospitals be protected against a catastrophic decline in revenues, even if it is only temporary ? How will public hospitals be paid during the transition period? How will they be paid for services not covered under the standard package? As shown in OECD countries and elsewhere, traditional fee-for-service systems tend to encourage more consultations and diagnostic tests, overprescription of drugs, higher surgical rates, and higher costs. The use of this reimbursement modality is in contradiction with Law 100, which mandates the use of prospective payment mechanisms that offer a vast array of incentives for providers to increase efficiency in their medical - 20 -  Reforms in the Health Sector practice, since they must absorb any additional cost if they exceed the fixed amount allocated for a defined package of interventions per person or for treating an average diagnosis. This type of arrangement may be required also for those services to be pro- vided by public hospitals but not covered by the standard health care packages. What is needed is some combination of protection of revenues-that is certainly not a bad idea in itself-with an incentive to get ready for that protection to end after a fixed time. And the interval of protection and preparation probably should not be uniform, as required by the measures now in place, but should start when a public hospital meets certain requirements and then end six or twelve months later. That way the whole system would not have to go through the transition at the same speed. Whether or not there were explicit pilot facilities or municipios, the effect would be to let the transition happen earlier in some places than others, and allow for the laggards to learn-and take heart-from the pioneers' experience. Something like this, with explicit pilot experiences, seems to have worked in Chile public hospitals were given six months to learn the new system of payments, during which they were guaranteed the larger of the revenues under the new or the old system. ARer that they knew they would have to survive with the new system. Of course the Chilean reform is much less radical than that contemplated by Law 100, because there are no EPS in the scheme, and public hospitals have to contract only with the Health Service to which they belong. Even so, some such explicit, staggered period of transition is needed. In summary, it is time to recognize transition as a process with certain require- ments, to let it happen at different speeds in different institutions or places, and to ensure that there are incentives for actually getting it underway. Freezes during which nothing happens and after which everything is as uncertain as before, do not help at all. - 21 -  Local Resource Mobilization ANNEX III: LOCAL RESOURCE MOBILIZATION A. INTRODUCTION A recurrent theme in Colombia's fiscal history has been concern with the effort, or lack of effort, demonstrated by departmental and muncipal governments in mobilizing local fiscal resources. This theme was, for example, emphasized in varying degrees by two major reviews of fiscal decentralization undertaken in 1981 and 1992.' More recently, the effect on local fiscal effort of the increased transfers to municipal governments resulting from Law 60 of 1993 has given rise to much concern and to a series of relatively inconclusive empirical studies in Colombia.2 This annex is intended as a contribution to this discussion by presenting an analysis of departmental and municipal finances from 1988 to 1994, using as consistent and comparable data base as can be compiled for the purpose. The next section briefly describes the data base that has been utilized. One clear conclusion emerging from this study is that the national government must devote more resources and effort to establishing a reliable data base on subnational finance if its decentralization policies are to be based on objective knowledge. At the present time, it is not possible for anyone in Colombia to have a good idea of what is going on at the subnational level: this is not good news in a country in which so many important public sector activities are now carried on by subnational governments. The third section, supported by an extensive statistical appendix, presents the major conclusions that emerge from these data. Contrary to what many seem to believe, there is little or no evidence of either generalized fiscal 'laziness' or wasteful expansion of local bureaucracy. On the contrary, Colombian local governments have generally expanded their own-source current revenues at approximately the same rate as their current expenditures, with the increased resources made available by larger national transfers, like the increased borrowing in some cities, going largely to finance expanded investment expenditures. As discussed in the main report, there are clearly a number of serious problems with Colombia's present system of intergovernmental finance. It is possible that at some future date undue decreases in local efforts at resource mobilization may properly be counted among those problems: but up until now this has not been the case. The fourth section comments briefly on both the relative uselessness of the concept of 'fiscal effort' in the design of intergovernmental transfers and on the limited extent to which subnational governments in Colombia can affect through their own actions and choices the amount of revenue at their disposal. When coupled with the extensive earmarking discussed in Annex IV-which means that subnational expenditure choices are also heavily constrained-the limited revenue discretion available to departmental and ISee Bird (1984), World Bank (1987) and World Bank (1989), and Wiesner (1992). 2See Wiesner (1994), Steiner and Correa (1994), Ferreira and Valenzuela (1993), Ahmad et al. (1995), Fainboim, Acosta and Cadena (1994), Vargas et al. (1994), Junguito, Melo and Misas (1995), Maldonado (1995), Sanchez and Gutierrez (1994, 1995), CGR (1995), and DNP (1995e). - 22 -  Local Resource Mobilization municipal governments suggests that the heavy emphasis on local 'fiscal effort' in much recent Colombian literature seems misconceived. The major problems in intergovernmen- tal finance in Colombia arise at, and must be resolved at, the national, not the local, level. B. DATA ON SUBNATIONAL FINANCE IN COLOMBIA Virtually every study of subnational finance in Colombia laments the poor state of the data. In particular, the reliability of the data for years since 1991-the period of most interest in this report-has been severely questioned by some, often with good reason. The core data used in the present study are based on data on municipal finance collected for 1988-1991 by the Banco de la Republica (BOR) and for 1992-1994 by the Departa- mento Nacional de Planeaci6n (DNP). These two distinct data bases were made available in machine-readable form by the DNP, and the present report is based on a special data base assembled by combining these two sources, as discussed below. Although both sources provided data with a common municipal and departmental code and employed similar relational database software, the DNP format for the later years is considerably more disaggregated than the BOR format used for the earlier years. To create a unified database, it was therefore first necessary to aggregate the 1992-94 data to match the 1988-91 format. Indeed, since a different format was used in 1992 than in later years, two different reclassification exercises were necessary. In theory, this process may seem relatively straightforward, but in practice a number of problems had to be overcome, in some instances rather arbitrarily. In the first place, over 100 of the smaller municipios do not appear in the DNP data base at all.3 Consistent data could be obtained for all seven years covered in this report only for 864 municipios. Athough the municipios included in the unified database accounted for 92.1 percent of the total population in 1994, and those excluded were probably for the most part relatively poor, the totals reported here consequently understate municipal revenues and expenditures by a small amount. This understatement should, however, be relatively consistent over time and should therefore not affect the trends reported. More importantly, as a result of the change in data bases over the period a number of critical variables were redefined and in some cases reclassified. In particular, the treatment of royalties, transfers, capital transfers, and investment differs, in some cases very substantially, from year to year. The result is that what is essentially the same item could conceivably be classified under three different headings in 1991, 1992, and 1993. To mention only one instance, the original data for 1992 and (with some variations) later years, counted as 'investment' items that were considered to fall under the concept of 'inversion social' as specified in Law 60, that is, such items as payments to teachers and health workers that had previously been classified as 'gasbos corrientes'. Such problems were resolved to the extent possible by reclassifying the data for 1992 and later years into the less disaggregated classification system used for 1991 and earlier years. Even after cleaning the data set as best as possible some hard to understand year-to-year fluctuations It is not clear why payments should be made to municipios that do not report the minimal required financial data. This seems to be the simplest and most easily enforceable, monitoring technique available. - 23 -  Local Resource Mobilization remain for particular departments and municipios. When the governments affected are relatively large such changes, which may sometimes reflect real policy changes (e.g. the increased property taxes in Bogota in 1993 and 1994) but which may also sometimes simply be reporting or classification errors, may obviously affect the totals reported here. For such reasons, caution must be used in interpreting the data on which this note is based. Despite the understatement noted above, and some likely classification and reporting errors (perhaps particularly in 1992, the first year of the new system), the municipal data base that has been assembled should in principle be sufficiently clean and consistent to be used for analytical purposes, as in the next section of this annex.4 An additional cautionary note needs to be sounded at the beginning, however. The financial information discussed here covers only municipal and departmental 'central' governments, that is, decentralized agencies and enterprises are excluded except to the extent transactions with them are reflected in government budgets. In later analysis, some correction is made for the resulting understatement of the level of subnational activity by taking into account (estimated) expenditures and revenues from the Situado Fiscal and the cofinancing funds, but there is no doubt that the picture of subnational financial activity sketched here remains incomplete in some important respects. Nonetheless, despite the many caveats we and others may have on the data used here, on the whole they appear to be clearer and more comparable and consistent than the data upon which some other recent studies have been based. In addition to the core data on municipal finance, additional data on other aspects of subnational finance have been obtained from a variety of sources, as indicated in the notes to the tables-DNP, DANE, Controlaria General de la Republica, Ecopetrol, and Instituto Geografico Agustin Codazzi. The comparability of these data is not always very clear, but these were the best figures available at the time of this study. Some specific comments regarding these various sources are noted where appropriate in the next section. Finally, where presented, 'real' data are shown in 1988 pesos, with expenditure figures deflated by DANE's government services price index and revenue figures deflated by the CPI. While other deflation methods may be used, this particular deflation procedure is intended on one hand to provide a better idea of the purchasing power of the resources over which government secures control and on the other to give an idea of the extent to which the private sector gives up purchasing power. Because these two indexes may move differently over time--and indeed did so over this period, with the CPI rising less rapidly than the government index before 1990 and more rapidly thereafter-the trends shown in the 'real' tables are not the same as those shown in the tables in nominal values. c. TRENDS IN SUBNATIONAL FINANCE Municipal governments. Tables 1- 4, present the basic data on municipal revenues and expenditures over the period 1988 to 1994, in current pesos, constant 1988 prices, real Comparisons of the data used here with that appearing in sources as Wiesner (1992), Ahmad et al. (1995), and DNP (1995c) reveals some differences, but none of great magnitude that cannot be explained by some of the classification problems (e.g. with respect to 'investment') mentioned in the text. - 24 -  Local Resource Mobilization pesos per capita, and as a percent of GDP. The format of the tables and some of the ways in which certain items are classified is different than is common in Colombia. Some of these changes have been made to overcome data problems, but most are intended to bring out more clearly the aspects of recent trends that are important for the present discussion. The first category shown in these tables, called 'own-source revenues', includes all tax and current nontax revenue received by municipal governments while excluding transfers from other levels of government, borrowing, and other capital income.5 Owing to data classification problems, it was not possible to separate 'tax' and 'nontax revenues' consistently over the period, other than for the two major taxes (see also later discussion on certain other taxes) so they have been lumped together in row A2 of the table. It should be noted, however, that royalties are also included in this figure, on the grounds that they constitute an 'own' revenue, rather than treating them as an additional 'national transfer' as is often done in Colombia. The rationale for this treatment is that most royalties (over 60 percent for Ecopetrol in 1994, for instance) go to the relatively few departments and municipios in which production takes place, for such reasons as to compensate for environmental costs or simply because it is perceived to be equitable or politically necessary for a disproportion- ate share of the returns to the nation to go to those areas (a similar, though more tenuous, argument might be made for the smaller share going to fluvial and coastal ports since 1994). From this perspective, royalties are like other fee or license revenue and should be treated similarly. It may be true, as some have argued, that a rapid increase in an 'unearned' income like this is as likely to lead to unwise expenditures in a Colombian department or municipio as in a Gulf state. Since the relative importance of royalty income actually fell over the period, the result of the treatment followed here is to lower the rate of increase in current revenues that would otherwise be shown in these tables. As a percent of GDP, local own-source revenues rose from just over I percent in 1988 to just over 2 percent in 1994, with the increase in the property tax (and the catch-all 'other' category which, as just noted, includes declining royalties) being particularly marked. On the expenditure side, wages and social benefits constitute the item shown as 'personnel', and apart from debt service, all other current expenditure items are aggre- gated in row B3, again to facilitate comparability over the period. Row C, 'own source deficit' is simply the difference between current own-source revenues, as defined here, and current expenditures. One of the most striking features of these figures, which is quite different from the conclusions some commentators seem to have drawn about trends in recent years, is that local own-source revenues have expanded more quickly than current expenditures, with the result that there has been virtually no change in the current account Some smaller municipios may include sales proceeds from municipal waterworks in nontax revenues (and operating costs in current expenditures). Most large ones operate such activities in enterprise form and hence only record net transfers (positive or negative) in the central government budget. Only the surplus (deficit) of sales receipts over operating expenses truly constitutes a revenue (expenditure) comparable to other budgetary items. Although there is no way to tell by how much the municipal revenue and expenditure data may be inflated for this reason, it seems unlikely to be large. - 25 -  Local Resource Mobilization deficit of the municipal sector over the period as a whole (see e.g. Table 2 in real terms and Table 3 in terms of percent of GDP). Apart from a jump in 1992 (when the data are shakiest owing the change in collection agency and in the classification system), these data offer no support for either of two common hypotheses in Colombia: there is no evidence of slackening fiscal effort and no evidence of excessive expansion of 'burocracia' financed by 'free' national transfers. As a percent of GDP, local expenditures on personnel did indeed rise from 0.6 to 1.0 percent over the period, but this increase, which took place mainly after 1992, was much less than that in own current revenues and was, moreover, undoubtedly influenced by the earmark- ing of a significant portion of the increased national transfers to 'social investment', a term which was defined to include payments to teachers and health workers.6 As Row D shows, even when such payments are excluded, the single most striking feature of Colombian municipal finance in recent years is the very large increase in investment expenditure, particularly after 1992.7 From 1988 to 1994, for example, while central government capital expenditure remained constant at an estimated 2.4 percent of GDP, local capital expenditure more than doubled, to 1.8 percent of GDP. The result of this substantial increase of municipal government investment, of course, was a marked increase in the overall deficit of the municipal sector, from 0.8 percent of GDP in 1988, to 1.9 percent in 1994, as shown in Row E of Table 3.8 How was this deficit financed? As shown in section F of the tables, the main source of finance was, as might be expected, the increased national transfers resulting from Law 60.9 As a percent of GDP, transfers doubled over the period, from 0.7 to 1.4 percent, As noted elsewhere, the data presented here include such current expenditures (which are, in accordance with Law 60, often classified as 'investment' in Colombian data) in expenditures on personnel. 7The data for post- 1991 years are somewhat inflated because they include payments for certain education expenditures (becas and dotaci6n) that should arguably be included in current expenditure both for analytical reasons and for comparability with the earlier data. 8It is interesting to note also that, while the local current account balance (as defined here) rose from a deficit of -0.2 (-0.6 in 1992) to one of -0.1 percent of GDP from 1990 to 1994, the central government's current account balance fell from a surplus of 1.3 (2.5 in 1992) to one of 0.6 over the same period. These data suggest that it is far from clear that the behavior of the municipal sector contributed to the stabilization problems of the public sector over this period. Note that no distinction is made here between 'current' and capital transfers, although it is common to do so in Colombian data: for example, national transfers to municipios (participaciones) are often divided into the 80 percent for inversion social (a concept which, as noted elsewhere, includes both investment and current expenditures) which is labeled a capital transfer and the balance (libre asignaci6n) which is labeled a current transfer (even though much of it also seems to finance investment). From the point of view of local resource mobilization, these two components are clearly more similar than different. Incidentally, although provision is made in the data base to show municipal receipts of cofinancing transfers (all classified as 'capital transfers'; although in fact many finance current expenditures) virtually no municipios record such transfers in their budgets, no doubt in part because the current cofinancing procedures are so cumbersome that no one is ever sure what they may get from this source, or when they may get it. - 26 -  Local Resource Mobilization accounting for 74 percent of the increased finance required for investment purposes, with the balance coming for the most part from increased borrowing.'