22 The Experience of Bogota

Marc Robinson
Published Date:
October 2007
  • ShareShare
Show Summary Details
Israel Fainboim Yaker

During the decade 1993-2003, the city of Bogota made significant progress in many areas. Civic culture strengthened. The coverage of core social and public utility services expanded and their quality rose, improving poverty indices based on the coverage of these services. Crime rates declined. The existing transport system is being replaced by a faster and more efficient system based on articulated buses circulating on exclusive lanes. Schools and libraries were built with high-quality infrastructure in the poorest areas of the city. Public spaces were reclaimed and recreational areas expanded with the construction of parks.

This progress has been impressive especially because it occurred in a context of rapid population growth, which created pressure to provide infrastructure and public utilities in the city. Population growth was greater in Bogota than in other Colombian cities because of the internal migration of people displaced by violence or attracted by the services offered.

These developments have been made sustainable by the strategy of financing the city’s expenditure needs through an increasing tax effort and by ensuring efficient and transparent use of public funds. Improvements in the provision of services resulted from a process of modernization of the public sector that involved the restructuring of some agencies (particularly the electricity company) and the introduction of a number of market-based governance elements (particularly concession contracts). Within this process, and to improve public management, increase transparency and accountability, and strengthen the city’s finances still further, in 2000 the city began to develop a new budgetary system based on a performance budget,1 linked to a medium-term expenditure framework (MTEF) and supported by an activity-based costing system.

The recent progress and the fact that Bogota has pioneered the development of these new budgetary instruments in Colombia and in Latin America can be attributed largely to the political and institutional changes made in the second half of the 1980s and the early 1990s, which included the introduction of democratic elections for Mayors, making it possible for mayorships to be won by candidates who were independent of the traditional parties and untainted by the cronyism practices of the past, and the approval of an Organic Statute for the city, which abolished the co-administration of its entities by the city council and the latter’s participation in procurement decisions and treasury management, thereby substantially reducing corruption.

The experience of Bogota with performance budgeting is relevant for at least three reasons. First, the deepening of decentralization and the increasing transfer of competencies and resources in Colombia over the last decade, as occurred in many other countries, mean that it is the local authorities rather than the national government that have the responsibility for the provision of the most important social and public services, and therefore it is they that should be held accountable for these services and not the national government; and a performance budget is the best instrument that can be used by them for accountability purposes. Second, the experience of Bogota shows that even the cities in developing countries can implement performance budgets without undue difficulties. Finally, the planning and budgeting institutions in Bogota are very similar to those of many Latin American countries, therefore the changes made can be used as a reference for other countries in the region.

This chapter describes the process of implementing a performance budget in Bogota. It begins by explaining the reasons and motivations for introducing this instrument in the city. It goes on to describe the activities carried out to implement this tool, and identifies the obstacles encountered and presents the results obtained. The last section provides some conclusions and recommendations, suggesting, among other things, possible ways of ensuring the permanency of the instrument. This latter point is important because although the city made considerable progress in implementing a performance budget between 2000 and 2003, the use of this instrument became a mere formality and for all practical purposes came to a halt in 2004 following the change in government—a risk faced by all reforms that are not embedded in the country’s legal framework. Before describing the implementation process, basic information on the city is provided to give the reader at least a general idea of the setting in which the process has unfolded.

Bogota’s background

Bogota is Colombia’s biggest city, with a population of 7 million, which makes it larger than at least six Latin American countries; its share of Colombia’s population is 18 percent. It had a per capita income in 2004 ($ 2,700) 30 percent larger than that for the nation ($ 1,900). It is “the country’s major financial center, the largest industrial city and the center of culture” (Gilbert and Davila, 2002, p. 29).

The city’s administration comprises 64 institutions and 53,000 civil servants. It has the responsibility for the provision of the most important social and urban services (water, sanitation, electricity, sewerage, transport, road maintenance and construction, education, health, housing, social assistance, street lighting, and so on) and competes in the provision of telecommunications services.

As it is also a capital district, it receives both departmental and municipal revenues. Its main revenue sources are the industry and commerce tax (a business tax levied on gross revenues), the property tax, and the gasoline surcharge, which between them account for over 90 percent of the city’s tax revenues. The city’s budget for 2006 amounts to roughly $2.4 billion.

Political context

Up until 1988, the capital was different from other cities in Colombia in that the Mayor was appointed by the President; in the other cities, he/she was appointed by the Governor, who in turn was appointed by the President of the Republic (Gilbert and Davila, 2002, pp. 32-3). The President usually appointed Mayors who shared his political affiliation. With the constitutional reform of 1986, all the Mayors are now democratically elected; the first of these elections took place in 1988. The system of democratically elected Mayors has resulted in the election of candidates that are independent of the traditional parties. Of the seven Mayors elected since that year, only two have been from the President’s party (Gilbert and Davila, 2002, p. 33). This has led to periodic conflicts with the national government and to the city’s dependence “on the goodwill of the president and the finance minister when it came to finding money for major projects” (Gilbert and Davila, 2002, p. 45).

The Mayor’s term of office was initially set at two years, but the constitutional reform of 1991 increased it to three and a new reform in 2004 further raised it to four—the same term as the President. The benefit is that it increases the authority of the Mayor and makes longer-term planning possible. The legislation allows the Mayor to be reelected, but not immediately, which gives rise to changes every four years in the management positions of institutions, limits policy continuity, and prohibits timely rewards to Mayors who have had a successful administration. Since the introduction of democratic elections, one of the Mayors (Antanas Mockus) has been re-elected.

The election of independent Mayors has also resulted in conflicts with the city council. The council has been democratically elected since 1974 (Gilbert and Davila, 2002, p. 33) and its term is the same as that of the Mayor. The legislation establishes that the number of council members must increase in proportion with the population, as a result of which the number has grown from 20 in 1972 to 42 in 2004. The council is very fragmented politically, which makes it very difficult to have dialogue to approve projects submitted by the administration, which must be sanctioned by practically every council member.

Although the council has always been dominated by traditional parties, and predominantly by the Liberal Party, the councils have no particular party line and are composed of various factions which, in many cases, are connected to individual council members; over time, the number of organizations represented on the council has increased, particularly religious organizations and movements with short political lifespans.

According to Gilbert and Davila,

given the fragmented nature of party politics in Colombia, and particularly the lack of party discipline, it is in the interest of many councilors to give the mayor a hard time. Few councilors can be relied upon to support the mayor; practically every vote has to be won. How that vote is to be won is a real problem insofar as it has frequently required the granting of favors. The mayor is expected to help secure public contracts, to direct public works projects to selected barrios, and to find jobs to councilors’ friends and allies. (2002, pp. 43-4)

This practice has caused serious conflicts between the executive and the council, which were particularly acute during the administrations of Mayor Antanas Mockus, who was opposed to any type of agreement that involved trading favors between the administration and the council for the approval of draft arrangements.

Up until 1993 the council members also sat on the boards of enterprises and district institutions, and participated in decision-making on hiring and cash management in the city. The Organic Statute of the city (Decree 1421 of 1993), issued by the central government in compliance with and obligation established in the 1991 Constitution, changed that situation and established a clear separation of functions and powers between the district executive and the council. The statute abolished the co-administration of its entities by the city council and the latter’s participation in procurement decisions and treasury management, thereby substantially reducing opportunities for corruption.2 It also initially afforded the city greater administrative, tax, and budgetary autonomy than other municipalities in Colombia and granted the executive the power to change the internal structure of institutions without needing approval from the council.

The relationship between the city administration and the national Congress has not been good. Several city administrations have had to make enormous efforts to have members of Congress defend their interests. This is partly due to the fact that their interests are under-represented in the Congress. Although the city has become a separate electoral area since 1991, allowing it to send its own representative to Congress, the number of representatives that it can elect is very small in proportion to its population. The majority of the Congress is opposed to the interests of the capital, using anti-centralist arguments, despite the fact that many senators (elected at the national level) come to office predominantly on the basis of votés from the city.

In terms of intergovernmental transfers, Bogota has received per capita (or per poor inhabitant) much less than other cities. It was only in 1996 that the Congress accepted the results of the 1993 population census, which improved the population’s situation with respect to transfers.

The difficult relationship between the National Executive, Congress, and the council has motivated the trend to have the city increase its self-financing capacity and to cease being dependent on decisions from those agencies to finance its plans and programs.

Despite Mayors’ short terms in office and the impossibility of immediate reelection, one of the factors that explain the improvements in the quality of life in the city in the 1990s and 2000s is the maintenance of a certain degree of continuity in policies and long-term targets across administrations. According to Gilbert and Davila (2002), the quality of city government “improved during the 1990s as a succession of mayors built on the achievements of their predecessors and made genuine improvements to the quality of urban management.” This has been because Mayors have tended to agree on the appropriate solutions to the challenges facing the city, and also because they have agreed that these responses must be medium term.

Planning, budgeting, and evaluation cycle at the sub-national level

Performance budgeting was not developed in Bogota in a vacuum, but drew on processes and instruments already established in the city, by integrating and strengthening those processes most useful for performance-based management and budgeting. These processes and instruments, which are largely the same as those used by the national government, include a Development Plan, a Multi-year Financial Plan, a Multi-year Investment Program (MIP), and an annual investments work plan (POAI), together with an action plan and the development of strategic plans by individual institutions. One major difference with the national government is a requirement that a document containing the government program must be prepared.

As at the central government level, the district budget is essentially based on the city’s Development Plan (PD). Unlike the national approach, the PD must be based not only on the multi-year financial plan but also on the Government Program. The Government Program is the document that each of the mayoral candidates presents to the public as a political platform for the elections. The program has to be prepared by identifying the city’s main needs and proposed solutions. The elected candidate must set out in the Development Plan the objectives and instruments proposed in the Government Program. When the citizens vote for a candidate, they vote for a Government Program and have the right, granted by the Constitution and the law, to demand that the Mayor-elect acquit this program.3

The Constitution incorporated an additional mechanism of democratic political accountability, the recall of mandate in the event of non-fulfillment of the Government Program,4 which strengthens the traditional tool of elections and provides incentives to an administration to produce results. Although a recall is hard to achieve because the requirements are very demanding,5 its mere existence forces administrations to prepare realistic Government Programs (and as a consequence, realistic development plans) that they can fulfill.

The link between the Development Plan and the agencies’ responsibilities with regard to specific targets is provided by strategic plans that define the entity’s mission and vision, and establish its long-term objectives. Their preparation is not compulsory in Colombia, but they have long been prepared in most public entities and are a useful tool for public managers. The strategic plans are the basic instrument of the entities for preparing their annual budgets.

The budget process in Bogota (see Box 22.1) is a two-stage and centralized one in which the agency responsible for finance—the Finance Secretariat—prepares the Annual Financial Plan, determines the aggregate expenditure ceilings for each of the agencies, and submits the budget proposal to the council. It is difficult for the council to amend the draft budget, because any proposal to amend the budget, or any Proyecto de Acuerdo (regulations issued by the local council) that involves additional spending, requires prior authorization by the government for its approval. The executing units also have very little leeway to change their budget during its execution.

Box 22.1.The budgetary process in Bogota

The budgetary process begins with the Mayor defining the budget policy guidelines, which are notified to line agencies by April 1 of each year. These guidelines are established after consultation with the District Economic and Fiscal Policy Council (CONFIS) and define the objectives, targets, and parameters to be considered in drafting the budget for the next fiscal year.

The Budget Directorate of the Finance Secretariat (SH) prepares central government revenue forecasts for the following fiscal year, in coordination with the Tax, Treasury, and Economic Research Directorates of the SH; and the forecasts for public establishments to generate their own revenue are revised. The expenditure initially forecast is the result of working group discussions between the Finance Secretariat, the District Planning Administrative Department (DAPD) and the various line agencies, in a forum created for that purpose called a “functional committee.”

On the basis of these exercises, the SH adjusts the Financial Plan (PF) and prepares an annual version of the plan, which it presents to CONFIS for consideration. Based on the annual PF, the SH defines and reports overall expenditure quotas for each entity, for functional expenses and investment. Each entity prepares an Annual Operational Investment Plan (POAI) based on the assigned quotas and the working group discussions and submits it to the DAPD so that the projects can be recorded in the Programs and Projects Databank, and produces and submits its preliminary draft budgets to the SH and the DAPD.

The preliminary drafts are finally discussed with the SH and the DAPD in the “functional committee” and, when an agreement is reached, the preliminary draft POAI is prepared, followed by the preliminary draft budget, which must be approved by CONFIS. The latter may amend the drafts to provide a final draft budget, which is sent to the city council. Since 2000, the performance budget has been included as an annex to this document. In 2003, the draft also included information on contingent liabilities and medium-term expenditure framework exercises for various sectors.

The draft must be submitted to the city council at the beginning of November for review and approval. The council has until December 31 to approve it. The council may only increase aggregate projected revenue and/or expenditure with the agreement of the Finance Secretary. It is also restricted in its power to reduce or delete individual items in the expenditure budget. If the council does not forward the budget by December 31, the government draft prevails, including any amendments approved during council debates. If the draft is rejected, the current budget prevails. If the council approves the budget but the Mayor objects to it on the grounds that it is illegal or unconstitutional, the government must refer the budget to the administrative tribunal of Cundinamarca for judgment.

Budget appropriations may be changed by the administration during the budget execution stage by means of budget transfers, additional credits or reductions, postponements or cancellations of previously approved appropriations. Budget transfers within an aggregate set aside for investment, operations or debt service may only be executed with the prior endorsement of the Budget Directorate (and the DAPD if the transfer affects investment items). In the event that the transfer involves a change in the aggregate amounts initially authorized, it must be submitted to the council for approval. Increases in appropriations must also be approved by the council.

Motivation for the move to performance budgeting

The performance budget (PB) was introduced in the city as an initiative of the administration and not as the result of any legal obligation to do so, and the motivation for its adoption arose from the government and not from the legislature. There is no legal obligation in Colombia to adopt performance budgeting at the national or local levels, unlike, for example, the legal requirement supporting the medium-term fiscal framework. In fact the performance budget in Bogota has been presented to the city council alongside the traditional budget project, and the latter is still the document that the council is legally required to approve.

Basically there were three reasons for Bogota to move to a performance budget: the first was to increase the transparency and efficiency of the use of public funds, and to hold the government more accountable for its results. Prior to 2000, very few public entities produced periodic reports on the outcome of its programs, but new programs and new expenditures continued to be executed without considering how the existing programs had performed. Inefficiency and waste of resources were rife, and widespread political corruption and cronyism had caused politics to lose credibility.

The aim was to restore confidence in the government and increase its legitimacy through transparent and efficient management. The administration of Mayor Mockus (2001-03), during which the PB was consolidated, considered public resources to be sacred; the public officials held them in trust, had to answer for them, report on their use, and use them efficiently Accountability and honesty in government were central to his program;6 and one of the main objectives of his administration was to improve probity in the use of resources. It was thought that the PB, which uses language that is accessible to the public, would facilitate accountability and promote public debate and participation in discussions on the budget and in monitoring and control of its execution.

The second motive for implementing a performance budget in the city was to reorient the public debate on the budget, by curtailing the endless discussions on line-items between the administration and the municipal council and refocusing it on the outcomes and outputs it was proposed to obtain with the budget. This sought to eliminate or at least restrict the “micro-management” intentions by the council, through lengthy discussions on investments, general expenses, wages, and contracts of each public entity. The belief by the administration that the presentation of the budget frames the public debate was a key factor underlying this decision.

Similarly, the PB was intended to discourage ad hoc budget reform proposals that would favor the interests of individuals or special interest groups by subjecting all expenditure proposals to the test of how they contribute to achieving any public objective established in the government program and in the PD.

Finally, the third factor that accounts for the introduction of the PB in the city was the involvement of former staff and consultants of the national government who played a key role in diagnosing existing problems in the budget process in Bogota, and implementing reforms. This group of technocrats, which had a shared vision of the reforms needed in the budget process, found that Bogota was fertile ground for applying their knowledge, since the matter was not moving forward at the national level.

Information base for performance budgeting

Information systems

In 1999 work began on development of an integrated financial management system in the city developed in-house with the support of a small group of external experts. The project was developed in modules, and in 2003 the key modules (budget, accounting, and treasury) were already in place.

Alongside this, in 2001 a performance information system was designed, that complemented the budget information module (PREDIS) of the financial management information system. This system records information on outputs and outcomes and their costs. Special care was taken in its design to ensure that expenditure data prepared on the basis of an economic classification more or less consistent with the Government Finance Statistics Manual (GFSM) 1986 would be used to produce information by objectives and programs. As no functional classification of expenditure has been implemented in Bogota, the accounts that were structured by institutional and economic classification had to be reorganized by objective.

At the same time, a number of information systems were being developed and used in the sector, which produced periodic information on outputs and outcomes to feed the performance information system developed by the Finance Secretariat.7

Selection of performance indicators

The selection of indicators by public entities for the performance budget was based fundamentally on the indicators of the Development Plan. The Development Plan of the Mockus administration (2001-03) included 81 specific objectives, of which:

  • 41 corresponded to outcomes or impacts by individual entities, including citizens’ perception of those benefits among them

  • 25 involved the provision of goods and services (outputs) by individual entities

  • 15 involved the provision of outputs by more than one agency.8

The task of selecting the indicators was facilitated by the fact that the Development Plan largely followed the recommendation that “key priorities should be measurable, achievable, outcome-oriented, and have defined time horizons.”

Although the indicators selected by the entities for their performance budgets covered 60 percent of the plan’s targets (48 indicators included in the PD), their measurement required a much larger number of indicators, because they were defined in aggregate form in the Development Plan.9

The main reasons why the selection of outputs predominated in Bogota were as follows:

  • For some areas it was very difficult to define the outcome indicators. A clear example of this problem arose in connection with the Citizenship Culture priority, whose outcomes are very difficult (or impossible) to quantify.

  • Some outcomes take more time to achieve than the three-year term of office. It was practically impossible to commit to the achievement of them in such a short time span. Output indicators can be used to check targets annually.

Efforts were made to specify relatively few outcomes and outputs for each entity, although this objective has not been fully achieved. The performance budget programmed for 2004 included 102 outcomes (objectives) and 230 outputs, which amounted to an average of 14 indicators per entity. Four to five indicators per program, defined for the most important programs, would have been sufficient.

The indicators were selected on the basis of characteristics widely recognized in the literature as appropriate (Hatry, 1999; Government of Alberta, 2000), such as meaningfulness, reliability, timeliness, and comprehensiveness.

Baselines and benchmarks

Baselines were established for 2000, and annual targets were set for the period 2001-04. Benchmarks often depended on the sector and program in question. In education, the results of academic tests were compared between public and private schools. Also, the cost of providing basic education services for a child was compared with the cost resulting from a concession contract. In the case of health services, benchmarking was performed by comparing the financial results of the 11 public hospitals as well as some non-financial performance indicators. In the case of security (or crime rate) indicators such as the homicide rate, comparisons were made with the country’s other large capital cities and with the nation as a whole.

Measurement of indicators

Indicators were measured almost entirely by the agencies. In some areas surveys were used to obtain information on public perceptions of the effects or to quantify them; this happened particularly with regard to the Citizenship Culture and Admirable Public Management priorities. The quality of the indicators used and of their measurement improved steadily during the government term. The agencies provided regular feedback with comments on the data recorded, thereby making improvements possible. However, there was no external auditing of the quality of them.

Adoption of ABC cost accounting for measuring efficiency and effectiveness

There are two ways in which a government can implement cost accounting (Kelly and Rivenbark, 2003):

  • develop an independent costing system, and reconcile it regularly with the existing integrated financial information system (IFMS) to ensure its accuracy. The advantage of this alternative is its flexibility in identifying the costs of service functions separately. Its disadvantage is that it takes time and uses resources to keep the system

  • adopt program-based costing, with accounts structured around programs. This makes it possible to monitor direct costs permanently with the IFMS. The total cost of each program can be calculated by distributing support services data from the IFMS among programs, using a standard methodology that includes indirect costs and depreciation of the capital assets of the respective program.

Following an investigation of existing costing models in national public sector bodies, and also in other countries, it was decided to adopt the first strategy and develop an activity-based costing (ABC) system; and to implement it initially among secretariats and decentralized bodies in 20 district entities, for later extension to the remaining entities. It is worth mentioning that there is no regulation requiring public bodies, either at the national or at the local governments’ level, to produce cost accounting information.

The ABC system was chosen, among other reasons, because it was considered as the method that best allocates indirect costs (overhead) among products and because it provides data for cost management, making it possible to identify inefficient areas and activities. ABC provides a clearer picture of the total cost of a process than other costing methods.

Also, standard accounting does not allow for reporting on the costs of processes that traverse the boundaries of an organization’s various departments and therefore it has shortcomings for decision-making. Standard accounting allows for costs to be determined only in terms of vertical organizational structures, whereas ABC permits cross-cutting costing procedures. This is because in the first case the information is structured for the collection of data on transactions by departments.

It is worth noting that the resistance faced for adopting ABC by the private sector, because it requires great detail and much more work than other costing methods, has diminished over time as ever-more efficient information technologies have become available. It should be mentioned that Bogota has an accrual accounting system, which facilitated the implementation of ABC, because ABC basically translates information from financial statements into cost data (Cokins, 2000).

As a result of actions taken in the selected pilot entities,10 the costing model was defined for 16 entities, structured as follows:

  • Definition of resources. Accounts were analyzed and resources defined on the basis thereof. Standardization of resources and their allocation to activities allowed for a cross-cutting analysis of their use.

  • Graphical model of activities. This made it possible to group together the different tasks undertaken by the public entities to achieve the objectives and fulfill the mission of each organization. Roughly 2,500 activities were defined, and specifications of the support activities of the different entities were standardized, thereby allowing for analysis not only at the entity level, but also at the sectoral and district levels.

  • Cost objects. The outputs, goods, and services to be costed were defined.

  • Inductors. Inductors, or criteria for distributing resources to activities, were denned.

The ABC system was put into operation in 20 entities by December 2003. Between 2001 and 2003 a simplified costing method was used, which implied the definition of “cost centers” for allocating costs to products. No comparison of the accuracy of the two methods was made, but the cost of implementing the simplified methodology was much lower.

Development of a performance evaluation methodology11

Although performance evaluation is the last stage in the cycle of public management, it must provide feedback for the budget planning, budget programming, and public policy execution processes, and provide information for reporting. Performance evaluation can yield the following benefits, among others:

  • at the level of organizational units, it can enhance operating capacity (improving internal processes) by identifying the causes of the failure to achieve targets

  • at the sectoral level, it can improve coordination among the agencies responsible for achieving shared objectives and defining public policy.

In the final months of 2003, a methodology for evaluating the results of the term was developed. The methodology was designed to determine not only whether or not the institutions achieved their intended objectives, but also to establish whether the outputs and objectives were well defined and coherent. Consequently, this was based on integrating the three levels of evaluation.

At the first level, the aim was to evaluate the strategic planning capacity of the institutions based on the consistency of the institutional mission with the organizational objectives defined by the institution and the consistency of those objectives with the outputs delivered. For that purpose:

  • a consistency indicator was developed that identifies the number of objectives that did not contribute to the fulfillment of the mission, or whose contribution is unclear

  • efforts were also made to determine whether the objectives are well defined, that is, whether they correspond to a benefit that the institution hopes to provide to a third party

  • an index of the relevance of each indicator to the objective being measured was also constructed. Its purpose was to establish whether the index measures progress toward the objective rather than what the institution is doing to achieve such progress, that is, if it adequately covers in its definition the benefit produced. The indices for individual indicators were consolidated into a summary index of the relevance of each institution’s performance indicators as a whole

  • similar indicators were established for outputs, with the purpose of: (1) determining whether they are well defined, that is, whether they correspond to a good or service that the institution hopes to provide to a third party; (2) whether they adequately measure at the individual level the quantity, quality, and/or timeliness of delivery of goods and services; (3) whether they do so as a group, in order to identify whether there are outputs whose delivery is not measures, or whether it is not clear what is being measured; and (4) whether the output is consistent with the objective.

Based on a simple average of these indicators, an overall rating of strategic planning by institutions was established.

At the second level, the intention was to evaluate to what extent the delivery of outputs (goods and services) had been achieved. For this purpose:

  • an indicator of effective provision of outputs which compares planned targets with actual values was defined

  • an indicator of changes in efficiency in the provision of outputs was also defined. It relates goods and services delivered to resources transferred and compares the performance of prior fiscal years. There was no proposal first to measure efficiency per se since the methodologies developed for rigorously obtaining information on output costs have not yet been applied.

At the third level, the aim was also to evaluate success in impacting the sectoral level (that is, achieving the objectives identified), by means of an analysis of the change in the institution’s effectiveness. To do so, based on the evaluator’s selection of the most important sectoral objectives, weighted effectiveness is compared using the weight of outputs as regards the achievement of objectives with the effectiveness of the weighted objectives based on the importance of their indicators.

The indicators being developed are easy to understand, and although they are not completely sound conceptually and reflect some of the subjectivity in their design, they are relevant and low cost in obtaining information and estimates. To quantify more rigorous indicators would have meant a much higher cost for information-gathering and estimation.

The evaluations draw on the information reported by the institutions for the performance budget. The ratings generated automatically through the performance evaluation sub-module designed for that purpose, which is connected to the budget module (PREDIS) of the Bogota financial information system. To be useful to the managers, the ratings must be accompanied by a qualitative analysis that seeks to identify the reasons for poor management performance.

To make the process more credible, two evaluations were to be produced: one by the institution itself (self-evaluation) and another by an external entity, more specifically the Finance Secretariat. The financial information system included an option for to self-evaluate, in addition to receiving the automatic rating it generates.

Along with the development of evaluation tools by the Finance Secretariat, the DAPD produced performance evaluations of the Development Plan using its own information system, the Development Plan Monitoring System. This is a reflection not only of the existence of two different evaluation processes in Bogota, but also of planning and budgeting.

Implementation steps

Before describing some key implementation steps of the performance budget it should be underscored that its introduction in Bogota received political support from the highest level of the administration, both from the Mayor and from the directors and managers of the various entities.

The District Finance Secretariat was in charge of implementing this instrument. The administration opted for a quicker full-scale implementation of the PB, due to the limited time available; a pilot or incremental approach increases the overall length of time required for implementation, and the Mayor’s term in office was only three years.

Institutional diagnostic study

The implementation process began with a diagnostic study of existing institutional arrangements, and particularly of the rules of the budgetary process in force in the city and their impact on fiscal sustainability, strategic prioritization of expenditure, and efficient resource use.

The diagnostic study was performed as part of the District Institutional Reform Mission, conducted by a team of consultants of the Finance Secretariat, the main objective of which was to provide a comprehensive overview of governance institutions in the city and propose a set of short- and long-term reforms to improve it.12 This made it possible to focus the project from the outset on addressing the problems identified and overcoming potential obstacles to implementation of the desired budgeting system.

Identification of best practices

In designing the performance budget, studies were made of performance-based budgetary systems and governance models in Australia, New Zealand, the United Kingdom, Sweden, Canada, the United States, and some North American cities that use (or have used) performance budgets, such as Sunnyvale (California), Coral Springs (Florida), Austin (Texas), and Indianapolis (Indiana).

External technical assistance

The design of the instruments also benefited from technical assistance provided by some international experts. Malcolm Holmes, co-author of the World Bank’s Public Expenditure Handbook, provided technical assistance in the design of the performance budget and, in conjunction with Joris van Dijk,13 assisted the administration in developing the medium-term expenditure framework. Design of the ABC costing system received assistance and training from Paul Sharman.14

Creation of a special unit to design and implement new budget instruments

To coordinate and direct the process, a special unit composed of consultants was set up outside the Budget Directorate, attached directly to the Office of the Secretary of Finance, with responsibility for definition of an action plan and the design and operational testing of the PB, and training of staff. A consulting team reporting directly to the Accounting Director Office of the city (an office that depends on the Finance Secretariat) was also established to design the costing system.

Possibly one of the errors made in the process was not to integrate these units into the structure of the Finance Secretariat; the fact that it was composed of consultants generated some resistance from the staff, which hindered their assimilation of the system. Nonetheless, if this unit had been incorporated into the structure of the Secretariat, it would have been unable to hire qualified staff owing to salary restrictions.


Part of the team of consultants was assigned to provide training and technical assistance to the entities. Each consultant was responsible for providing training and technical assistance to roughly six entities. Presentations were made, and workshops and meetings were held with staff from all entities, to disseminate the new concepts and encourage the development of planning exercises within them, thereby laying the foundations for annual programming of the performance budget.

Interagency coordination

As many outcomes were the responsibility of several agencies, achieving them required a coordinated effort. This was particularly important in the areas of mobility, public space, and social development. For that purpose, the administration organized coordination committees for each strategy of the Development Plan (the plan had nine strategies) and for each sector, chaired by the person with greatest responsibility for the outcome in question.

The Mayor held regular meetings with the objectives coordinator and the sectoral coordinator, “which made it possible for the actions to be seen from the two perspectives. These two different perspectives enhanced monitoring and the number of opportunities for coordination” (Mockus, 2004, p. 55). This was also important in a context in which public institutions are not properly grouped and ranked on the basis of sectoral structures that allow the public to clearly identify those responsible for performance.15


Coverage of the performance budget

The performance budget was implemented as an information structure in all secretariats, administrative departments, and public establishments (that is, a total of 37 entities). Systemized information on costs, outputs, and outcomes for these entities is available for the period 2000-03.16 Progress was expected to be made in 2004 and 2005 in the coverage of public enterprises, hospitals, and local district administrations. As mentioned above, ABC costing was put into operation in 20 entities in 2003. Although the performance budget was not implemented in all the city’s entities, its adoption was always planned as a gradual process that would take a number of years to consolidate.

Monitoring and evaluation of outcomes and outputs

Information is published annually on performance and output in an annex to the budget documents. As mentioned earlier, only in 2003 was a performance evaluation methodology developed, and it cannot be considered a full program evaluation (for example, cost-benefit or cost-effectiveness evaluations),17 although it was more than a mere analysis of indicators.

As indicated, this methodology includes a review of consistency of objectives with outputs; an analysis linking outputs an objectives to offer some information on expenditure effectiveness; a comparison between results planned and results obtained (efficacy); and a comparison between results obtained with a given budget in the respective year with those obtained in previous years, to offer some information on expenditure efficiency. The methodology has been applied in the environment, education, health, social welfare, and culture sectors, with the results reported (see Polania, 2004a; Moreno, 2004).

Use of performance information by government and council

Since year 2002, during the budget preparation stage allocation decisions were supported with performance information, resulting in some modifications, albeit minor, to resource allocations. Outcomes and outputs began to be analyzed and achievements used in the “functional committeees” as an input for feedback on decision-making. These committees are the mechanism through which the Finance Secretariat discusses with the corresponding directors each entity’s expenditure quotas in the budget and its distribution between the different programs. The Secretariat’s sectoral specialist, the director of the special unit for developing the performance budget and the directors of the public entity attended these meetings. Usually the Finance Secretary attended the meetings for those entities that have large budgets.

The most important programs of each institution were analyzed with some degree of detail in this scenario. If the institution could show that the failure to achieve intended outcomes was due to inadequate budget funding, more appropriations were sought for the program. However, if the sectoral specialist from the Finance Secretariat showed that the outcomes were poor due to improper program management, the resources per unit of output would not be changed and the discussion would focus on ways of improving program implementation. As the Finance Secretariat began to produce approximate information on the unit costs of outputs since 2000, it was not easy for the institutions to prove that they needed additional resources per unit of output to produce a predetermined level of goods and services.

Since 2000, performance budget documents began to be delivered annually to the council in addition to the required budget documents. These documents contained information on the mission, vision, strategies, and priorities, together with a list of outcomes and other outputs; information on achievements in the previous year (without comparing them against the targets of that year); projected targets for the current year and those programmed for the following one; and information on investment and operating costs for the current and following year, whenever that was easy to determine. However, the document did not identify or provide information on uncontrollable external factors that could affect outcomes.

The external audit entity (Contraloria) and the council used the information provided to debate budgetary issues and to exercise political oversight of commitments undertaken and the outputs and outcomes achieved by the city administration.18 Since 2000, the budget discussion with the council has been increasingly centered on outputs and outcomes and less on inputs. This change in the way the budget is discussed was a very important step forward, since it was the first time in the history of Colombia that an institutional arrangement provided a mechanism whereby the inhabitants of the city and their political representatives could access information to evaluate mayoral performance.

Although the council continued to appropriate resources by agency and expenditure item, rather than by outputs or outcome targets, the information contained in the performance budget allowed for a more informed discussion of the budget and resulted in a number of minor changes in appropriations.

It should also be mentioned that in 2003 the council approved Agreement 67, which requires the administration to publish information on performance indicators. The Decision 67 initiative arose partly by contagion resulting from the use of performance budgeting by the administration, and partly also as a result of the influence of the Bogota Como Vamos project (described below). The indicators selected in the decision were developed with the unit responsible for the performance budget project in the city, although the council added a number of its own.19

The other key component of a performance budget is the flexibility to be given to managers in administering programs, in exchange for assuming responsibility for their outcomes. This doesn’t exist in the city, because managers are free to reallocate resources only between “investment programs” and within them, provided the entities did not weaken their output (or, in some cases, outcome) commitments and conditioned to an authorization by the Finance Secretariat (or the DAPD in the case of capital expenditures). Furthermore, the city council approves the budget for the institutions and large items (total per institutions for investment, operations, and debt service),20 which gives the administration greater flexibility in reallocating expenditure. The proposal for this, which was not implemented, was that as public managers produced better outcomes, the Finance Secretariat would allow them more flexibility in the use of resources.

Impacts on effectiveness and efficiency of service delivery

Although government provision of goods and services became considerably more effective and efficient between 2001 and 2003, it is hard to attribute the improvements to the use of performance budgeting. This is due to the fact that performance budgeting was introduced along with a large number of reforms aimed at improving governance and the provision of services to citizens, which jointly can be classified as “market-based governance” mechanisms. These included the liquidation of a number of public agencies that were no longer providing useful services to citizens; partial privatization of electric power and telecommunications firms; restructuring of entities through radical staff downsizing; and the introduction of concession mechanisms for the provision of many services.21 In several cases these mechanisms included the adoption of performance contracts (water supply park maintenance, education services, transport).

Effects on transparency and accountability

Results data (outputs and outcomes) were collected in many cases on a monthly basis; reports to citizens were provided once a year and annual accountability meetings were held in 2002 and 2003 (reports were also made to the council in budget documents, which are public). In 2002 and 2003, annual reports on the results of PB management were produced, and two exercises of citizenry accountability were performed on the results obtained. A standard alternative was thus offered for reporting the administration’s performance to the council and the citizens. Formerly, each entity presented individual outcome reports, which usually entailed a numerical-legal analysis of public expenditure, without a major benchmark in terms of outputs delivered and the benefits for the city.

Citizens were able to ascertain the approximate cost of most goods and services provided by the administration. Moreover, in the discussion of priorities, these costs have become units of measurement for the existing trade-offs, such as: how many school places would be sacrificed if it were decided to construct a pedestrian crossing?

Box 22.2.The Bogota Cotno Vamos project: an important accountability mechanism

The Bogota Como Vamos project has been promoted since 1998 by Casa Editorial El Tiempo (the country’s largest circulation daily newspaper), Fundación Corona (a nongovernmental organization), and the Bogota Chamber of Commerce, associated with the strongest private sector professional association in the city. It aims to measure changes in the levels of quality of life and well-being in Bogota, to provide citizens with basic information to evaluate governance, and to promote efficiency and transparency in district administration. The project evaluates the fulfillment of each administration’s Development Plan, analyzing changes in people’s access to core services and the level of citizen satisfaction through indicators of the coverage and quality of services, and citizens’ perception towards them.

There are 12 areas of evaluation: education, health, highway mobility, public space, public finances, environment, governance, security, public utilities, housing, citizen responsibility, and economic development. Indicators for each area were defined on the basis of discussions with the administration, and are measured annually. Most of the information is provided by the city. Each administration can include new indicators in the evaluation, but cannot remove the existing ones. On a complementary basis, the project gathers monthly data on 75 variables, in a process known as the Bogotanometro.

Also, to gauge citizens’ perception on a wide variety of issues, it conducts an annual survey, interviewing no less that 1,500 citizens. The survey asks about: the climate of opinion (general climate, economic climate, food situation); sectoral aspects (health, education, childhood issues, domestic services, housing, and so on); general aspects (citizen responsibility, mobility, environment); and government aspects (public administration, and so on). Each evaluation is debated by public officials, and the results are disseminated through a variety of communications media.

It should also be stressed that Bogota is subject to external monitoring of the outcomes of the Development Plan, undertaken by the Bogota How Are We Doing (Bogota Como Vamos) project (see Box 22.2). This has made a significant positive contribution to improving the quality of public debate, providing annual data on the outcomes achieved by the development plan, and discussing them publicly. The information collected by the project comes from the public agencies, and is complemented with survey data.


Although the district has made a great effort to implement a performance budget, much hard work is still needed for its consolidation. The performance budget adopted by the city worked well as an information structure and therefore as an instrument for political control and citizen’s participation. However, its role as a tool for expenditure allocation was more limited, as was the use of the performance information as a managerial tool.

The budget cycle must go beyond a list of performance commitments annexed to the budget. The process of allocating resources must be based on how enforceable and sustainable such commitments are. The budget now includes information on performance, yet decisions are not being systematically taken on that basis, nor are the medium-term implications of the various programs being systematically considered, despite the fact that this activity is a key underpinning of the performance budget.

The following additional steps are desirable to create the conditions for consolidating the performance orientation of the public resource management cycle:

Step 1. It is advisable for Bogota to institutionalize a high-level (cabinet) forum for open, collegial procedures that would evaluate the various policy options and have them compete for resources, within the framework of medium-term restrictions. This forum would be in charge of ensuring that the policies are enforced and feedback is given on their design (or for developing alternative policies), based on performance. Sectoral expenditure directors and the authorities responsible for discipline and guidance in government management should have seats in the forum to make the process of resource allocation an open discussion, which would legitimize the policies to be developed within the administration, and would help reconcile top-down restrictions (political priorities and availability of resources in the medium term) with bottom-up restrictions (resource needs in the short and medium term by the different sectors of government to maintain current policies or introduce new ones).

Step 2. Another point that merits attention in the future is the importance of broadening the vision of the budget to a multi-year format. To achieve this, there must be continuity in the implementation of the medium-term expenditure framework, not only as a tool for information on future spending, but as a policymaking process that requires agreements among public managers on the direction of policies and, consequently, expenditure trends.

To the extent that Bogota still does not regularly use comprehensive expenditure categories at the policy and program level, it is not clear how institutions are linked to develop them and whether policy performance is enforceable. The preparation of Sectoral Strategic Plans as building blocks of the medium-term expenditure framework will make it possible to consolidate credible medium-term contracts to allocate public resources to strategic priorities and enforce the respective outcomes. With these plans, objectives common to several entities could be coordinated on the basis of these policies and inter-institutional work encouraged.

The sectoral focus would provide a more strategic approach to the budget, creating a better environment for its formulation. It will also reduce pressure on public spending, since the formulation of sectoral plans would increase the possibility of transferring resources within the sector, promoting their internal reallocation. In a context of tight resources restrictions, additional amounts for new policies will have to be sought within the sector by promoting program savings. The Sectoral Plans and the medium-term expenditure framework should gradually replace the Development Plan as the medium-term reference framework for annual budget decisions.

Step 3. The annual budget process should include: (a) a definition and/or review of sectoral policies, with a view to reaching an agreement at the highest level on the amounts to be allocated to each policy; (b) an initial definition of quotas, based on which meetings would be held (task forces) focusing on analysis and establishing performance commitments; and (c) a performance analysis as a reference for final discussions on budget distribution.

Step 4. The United Nations functional classification of expenditure (COFOG) should be adopted as a bridge between the budget information by line-item and budget performance data. This classification should have the force of law within the administration.22

Step 5. One key institutional aspect is the importance of creating the organizational conditions necessary for formally implementing the new budget rules within the political/administrative structure of the district. On this point, the problems with the current organizational structure are well known, namely too many institutions operating within an inflexible institutional structure, with no connection to sectoral structures to which one can delegate and demand accountability for performance. As a result, in each administration, as mentioned earlier, informal coordination mechanisms arise that do not have sustainability over time.

Step 6. The new budget instruments (medium-term expenditure framework and performance budget) should be given a firm legal basis.23 This would guarantee its permanency across changes of government.24

Step 7. Line agencies should be given more autonomy over the management of their resources in exchange for accountability for outcomes. Ex-ante control imposed by the District Finance Secretariat and the District Planning Administrative Department on internal transfers of large investment and operating items should therefore be lifted, provided that these entities do not dilute their commitment to outcomes. This increased flexibility could be granted gradually and selectively to those institutions that demonstrate better-defined objectives and outputs and greater commitment to outcomes with the resources available.

Step 8. Finally, without changing the civil service system, the use of the performance information as a managerial tool will have limited impact. The civil service system seriously restricts the introduction of individual incentives. Also, civil servants are generally unskilled and hold their posts through political recommendation. The related legislation, which is national, is at a stalemate and staff have been unable to be promoted on the basis of performance. Other incentives to achieve results relate to public recognition and they were used considerably by Mayor Mockus’s administration.


It should be noted that the new budget model was called a “results-oriented budget” not a “performance budget,” but the latter designation will be used in this chapter to refer to the new instrument the city has adopted.

Also, the 1991 Constitution prohibited the system of transfers or “auxilios,” whereby a small portion of the budget is distributed among the council members (or, at the national level, among legislators) for them to use the resources supposedly for projects in their communities.

Article 259 of the 1991 Constitution introduced the concept of “programmatic vote” in elections for Mayors and Governors (regulated by the Law 131 of 1994), which means that voters electing candidates to such posts mandate the elected candidate to implement the Government Program that he or she presented when registering as a candidate.

Articles 40 and 103 of the Political Constitution.

This mechanism can be invoked when at least one year of the term of office has elapsed, following a call for popular vote by at least 40 percent of the total votés obtained by the elected candidate. Recall can occur if it is approved by an absolute majority of citizens participating in the vote, provided the number of votés is at least 55 percent of those registered when the candidate was elected.

His popularity came from the fact that he did not represent any of the parties and from his record of honesty.

One of these is the SUIVD (Unified System of Information on Violence and Delinquency) (<>), operated by the Government Secretariat and drawing data from the Metropolitan Police and the Forensic Medicine Institute. The other is the information system operated by the District Culture and Tourism Institute.

Some analysts have argued that an organization should use both output and outcome indicators, and even input indicators, on the grounds that the various users of performance information (politically elected officials, managers, citizens, legislators) require different indicators.

For example, one of the targets was “to achieve useful inoculation coverages (95 percent for polio and measles, and 85 percent for diphtheria, tetanus, whooping cough, haemophylus influenza, hepatitis B, rubella, and tuberculosis).” As a specific measurement was required for each type of vaccination, this target alone required nine indicators.

These were public entities from the central and decentralized sectors of the Capital District, including an industrial and commercial enterprise (Metrovivienda) and a state social enterprise.

This section is based on Polania (2004b).

The Mission studies were supported by a loan provided by the Inter-American Development Bank (IADB).

Consultant in macroeconomics and public finance for the Division of Macroeconomic and Sector Policies of ECORYS, Netherlands.

President, Focused Management Inc.

The Mockus administration proposed to the city council on several occasions that they adapt the city’s institutional structure, grouping it into sectors and establishing a hierarchy of relationships between the institutions in each sector, but the council never approved the proposals.

This can be consulted directly on the internet at the Finance Secretariat website (<>).

Note that “outcome measures can identify trends and developments that need attention, but a full program evaluation is needed to establish whether the cause of poor performance…has anything to do with its financing” (Joyce, 1999).

The external audit entities in Colombia, as in most other countries, are required to examine not only the legality of operations but also the economy, efficiency, and effectiveness of the provision of goods and services.

Currently (2006), although the information requirements of the Agreement are fulfilled, its presentation has become a mere formality.

It should be noted that the existing flexibility diminishes at the national level because Congress has the power to approve the budget up to the investment program level.

Among the services concessioned to the private sector were: education, street cleaning and refuse collection, mass transportation (Transmilenio), the client services activities related to water and sewerage services, management of the Tibitoc water treatment plant, the sanitary landfill, parking lots, cemeteries, gas, lotteries/betting, street lighting, issuing vehicle number plates and administering mechanical checks on buses and trucks, citizen information services and payment of utilities’ bills (CADEs), and park maintenance.

It could be part of the Budget Settlement Decree approved by the Mayor upon completion of the budget discussion in the council.

This requires changes in the national legislation.

This is a necessary but not a sufficient condition for a successful adoption of a performance budget. It is the use that the different stakeholders (legislature, citizens, civil servants) give to these instruments that will determine its consolidation as a governance tool.


    CokinsG.2000Overcoming the Obstacles to Implementing Activity Based Costing,Bank Accounting and Finance Vol. 14(1).

    GilbertA. and J.Davila2002Bogotá: Progress Within a Hostile Environment,” in Capital City Politics in Latin America: Democratization and Empowermented. D.J.Myers and H.A.Dietz (Boulder: Lynne Riener Publishers).

    Government of Alberta2000Results-Oriented Government. A Guide to Strategic Planning and Performance Measurement in the Public Sector<>, accessedJuly2006.

    HatryH.P.1999What Type of Performance Information Should be Tracked,” in Quicker Better Cheaper? Managing Performance in American Governmented. D.W.Forsythe (New York: Nelson A. Rockefeller Institute of Government) ch. 2.

    JoyceP.G.1999Performance-Based Budgeting,” in Handbook of Government Budgetinged. R.Meyers (San Francisco: Jossey-Bass) ch. 24.

    KellyJ.M. and W.C.Rivenbark2003Performance Budgeting for State and Local Governments (Armonk, NY: M.E. Sharpe).

    MockusA.2004Modernizacion del presupuesto, evaluatión y calidad del gasto publico,” in ¿Por qué evaluar el gasto publico? Experiencias internacionales y el caso Colombiano (Bogotá: BID, World Bank, and DNP).

    MorenoF.2004Aplicación de la propuesta para evaluar la gestión pública. Sectores: educatión, cultura, bienestar social y salud.Mimeo Secretaría de Hacienda de BogotáJanuary.

    PolaniaD.F.2004aEvaluatión de la Gestión pública en el sector medio ambiente.Mimeo Secretaría de Hacienda de BogotáJanuary.

    PolaniaD.F.2004bLineamientos y propuesta procedimental para la evaluatión de la Gestión publica en el Distrito Capital.Mimeo Secretaría de Hacienda de BogotáaJanuary.

    Other Resources Citing This Publication