Design and Implementation of Spending Reviews in Colombia: A Case Study and Recommendations (Complete note) 1 World Bank Group Governance Global Practice March 2018 1 This note was prepared by Tomás Martín, Pablo Andres Guzmán, Joanna Watkins and Pedro Arizti. Content I. Country Background ............................................................................................................... 4 II. Conceptual Framework for Spending Reviews in Colombia .................................................. 5 A. Principles, levels of analysis, and guiding questions ............................................................. 5 B. Political leadership for the public spending reviews............................................................ 10 C. Objectives of the public spending reviews........................................................................... 11 D. Scope of the reviews and features of Colombia’s budgeting system................................... 11 E. Stakeholder roles and responsibilities .................................................................................. 12 III. Implementation of the Spending Reviews in 2016 ............................................................... 14 A. Phases One and Two ............................................................................................................ 15 B. Phase Three .......................................................................................................................... 16 C. Phase Four ............................................................................................................................ 16 D. Phases Five and Six ............................................................................................................. 18 III. Results, lessons learned and recommendations ..................................................................... 18 A. Results and lessons learned .................................................................................................. 18 B. Recommendations ................................................................................................................ 21 References ..................................................................................................................................... 23 Abbreviations and Acronyms COP Colombian peso STIP Territorial Management and Public Investment Subdirection (Subdirección Territorial y de Inversión Pública) DIFP Direction of Public Investment and Finances (Dirección de Inversiones y Finanzas Públicas) DSEPP Direction for Monitoring and Evaluation of Public Policies (Dirección de Seguimiento y Evaluación de Políticas Públicas) GDP Gross domestic product IMF International Monetary Fund MoF Ministry of Finance NPD National Planning Department OECD Organization for Economic Cooperation and Development WBG World Bank Group 2 Design and Implementation of Spending Reviews in Colombia: A Case Study and Recommendations 2 1. Spending reviews offer a valuable tool to identify structural savings, while promoting the efficiency and effectiveness of public investment. Several OECD member countries have implemented such reviews since the last decade and particularly in the aftermath of the 2008 financial crisis. The tight fiscal and expenditure constraints resulting from the crisis led many countries to resort to spending reviews as a pivotal tool in its fiscal consolidation efforts. Since the pioneering implementation of spending reviews in Canada and the United Kingdom during the 90s (Stacey, 2015), Australia, France, and Ireland, among other OECD countries, have carried out similar works. Likewise, the World Bank Group (WBG) and IMF have recently carried out pilot spending reviews in Croatia (Madzarevic-Sujster and others, 2015; and Olden and others, 2015) and Poland (Holda and Skrok, 2015). In the case of Colombia, spending reviews were recommended by the 2013 OECD Public Governance Review as a mechanism to improve the linkage between strategy and budgeting as well as a prioritization tool for the allocation of fiscal space. 2. Spending reviews also provide a useful mechanism to improve the quality of public expenditure. On the one hand, they contribute to making explicit the alignment of expenditure with the government policy priorities, thus increasing the efficiency of resource allocation. On the other hand, they examine the efficiency (or lack of) in the processes of service delivery of goods and services to the citizens (operational efficiency). Spending reviews can provide valuable inputs to carry out planned budget cuts without negatively affecting the government’s capacity to deliver strategic programs. As such, they are a useful and necessary tool to improve the efficiency and effectiveness of public expenditure in a pragmatic and concrete way. 3. In this context, Colombia’s National Planning Department (NPD) requested WBG support to develop and implement a spending review methodology. The methodology had the objective of identifying potential savings resulting from efficiency gains by reallocating budget from low-priority or low-performing public investment initiatives toward high-performing projects, or carrying out focalized cuts depending on the fiscal situation. The WBG support also sought to generate institutional capacity at the National Planning Department and sector-level institutions by training staff in the implementation of the methodology so that it could be replicated in future exercises. 4. As a result of the technical assistance project, spending reviews were carried out in two pilot sectors. The reviews produced budget reallocation recommendations for the 2017 national budget. The reviews analyzed the social inclusion and agriculture sectors, and included an assessment of the alignment of public investment projects with the government objectives and the achievement of financial progress milestones and results. 2This document is based on the final deliverables of a Reimbursable Advisory Services technical assistance provided by the World Bank to the Colombian Government’s National Planning Department in 2016. The Bank team of senior sector experts was led by Pedro Arizti and Joanna Watkins (Governance Global Practice), and comprised Tomás Martín, Genni Cavalli, León Montealegre, and Santiago de la Cadena. Juan José Perfetti (Research Associate from Fedesarrollo) and Roberto Angulo (independent consultant) also participated in the spending reviews. 3 I. Country Background 5. In 2015, Colombia´s fiscal position required the implementation of measures to strengthen the management of public resources due to the sudden drop in fiscal revenues. Increasingly low oil and commodities prices on international markets, as well as expectations of a reduction in oil production, led to a significant decrease of exports that in turn affected the central government revenues. According to data from the Ministry of Finance (MoF), dividends from the state-owned, oil-producing company, ECOPETROL, which in 2014 accounted for 15.6 percent of total government revenues — or nearly 2.6 percent of GDP — dropped by over 50 percent in 2015. Nevertheless, the country maintained a positive economic growth trend based on a prudent macroeconomic and fiscal policy. However, to keep this trend in 2016, the Government of Colombia sought to develop tools that could review and improve the effectiveness of public expenditure in the context of scarce resources. 6. In recent years budget appropriations had experienced sustained growth, mainly driven by increases in public investment. Overall, government expenditure rose from 16 percent of GDP in 2011 to 19 percent in 2015, whereas public investment as share of total Government expenses increased from 13 to 21 percent during the same period. Meanwhile, debt service and recurrent expenditure remained stable. Public investment was thus the key driver behind the pressure on government finances, forcing the government to seek optimization measures that could enhance the fiscal space and restructure spending in an effective and efficient way. 7. The increase of government expenditure, accompanied by a drop of revenue, led to an increase of the fiscal deficit. The fiscal deficit, which in 2014 amounted to 2 percent of GDP, increased to 3 percent in 2015 and reached 4 percent in 2016. The deficit increase was mainly driven by the downturn in oil prices, increasing the gap with the structural deficit established by the fiscal rule. 8. While the Government began a series of ambitious structural reforms dating back to 2014 to cope with the fiscal consequences of oil price cycles, they lacked a strategic approach regarding the rationalization of public expenditure. The reforms implemented ranged from budget cuts to a tax reform which increased non-oil tax revenue by nearly 0.6 percent of GDP, whereas the implementation of aggressive budget cuts led to savings of approximately 0.8 percent of GDP in 2014 and 1 percent in 2015. In this context, and in line with a presidential mandate to curb public expenditure, a government-wide budget cut amounting to 0.7 percent of GDP was established for 2016. However, the implementation of this directive entailed a significant redistribution of resources across sectors and entities, with the risk of losing funding for strategic and high-economic, social-value programs. 9. Though it is crucial for Colombia to implement immediate measures to mitigate the effects of a downturn, such measures should also consider the effectiveness and performance of public investment projects. The decision to carry-out budget cuts or redistribute resources within Government needs to be underpinned by an analytical framework of public investment expenditure that provides insights regarding both the efficiency and effectiveness in the use of resources vis-à-vis the achievement of government priorities, and identifies potential savings 4 derived from postponing or eliminating non-priority expenses. Such a framework would allow the government to ensure that expenditure programs with high economic and social value are shielded from the downturn and continue contributing to the achievement of national development objectives. II. Conceptual Framework for Spending Reviews in Colombia 10. Overall, spending reviews must be sufficiently flexible to adjust to the specific features of each sector and to changing economic and political circumstances. The establishment of overly rigid or standardized processes runs the risk of being perceived as disconnected from the country or sector context and thus regarded as a purely formal step that does not add value to manage public expenditure. This may lead to a loss of interest by public agencies in the process and can affect their degree of participation and commitment. A. Principles, levels of analysis, and guiding questions 11. The following basic principles guided the spending review process in the case of Colombia: 3 a) The review process must have high-level political support. This support is key for the success of the review. In the case of Colombia, the process should ideally be backed by the President, the Minister of Finance and the NPD Director. They should provide a clear mandate so that line ministries are aware of the review objectives and issue the necessary instructions to their staff to carry out the review. b) Clear savings or reallocation targets should be established to guide the exercise. Without an explicit savings or reallocation target, the linkage between the review process and the desired effect is lost considering that it would not be possible to assess if the exercise has served its purpose. Such targets are critical to guide the work of the teams in the field and need to be credible to provide the necessary compliance incentives. The process by which the targets are defined can vary depending on the country, but the targets must be consistent with a medium-term expenditure framework. c) Milestones for the spending review process must be aligned with the budget calendar. In the absence of clear milestones and deadlines, the spending review might extend its work indefinitely and become irrelevant as an input for the budget formulation process. d) The proposals resulting from the spending review must be jointly discussed with the Ministry of Finance and the National Planning Department, as well as the relevant sector ministries or agencies. The spending reviews should be carried out as partially decentralized exercises to maximize ownership and participation from line ministries in the 3 The principles and questions outlined in this section were discussed with NPD and were proposed by the WBG team based on a literature review and international experiences (Vandierendonck, 2014; Robinson, 2013; Madzarevic-Sujster and others, 2015; Holda and Skrok, 2015; Olden and others, 2015; Stacey, 2015: HM Treasury, 2010 and 2015; German Federal Ministry of Finance, 2016). 5 identification of savings. This also simplifies the access to information and facilitates implementation of the recommendations later. e) A procedure to prepare savings proposals at the center must be defined to solve controversies or cases when sector-level review teams do not complete their work on time. This procedure should set guidelines for both National Planning Department and Ministry of Finance to centrally develop or complete savings proposals in case the review teams at the sector level do not complete the work within the given deadlines or do not achieve the targets set at the start of the exercise. The threat of developing unilateral proposals at the center, together with the setting of credible strict deadlines, provides incentives for the sectors to actively participate in the review process. f) Reallocation proposals should only be presented either by the sector review teams based on the methodological framework or through a pre-defined centralized procedure in case the deadlines or targets are not met. The sector-level review teams are key actors throughout the process and have the formal mandate to prepare savings and reallocation proposals. This implies that recommendations from stakeholders external to the review teams — and aside from the National Planning Department and Ministry of Finance — need to be first considered and agreed to by the review teams in order to be formally included in the reallocation proposals. g) Final decisions regarding savings proposals must be considered by a high-level political committee. Spending review processes entail a political dimension that needs to be considered to ensure sustainability. In this regard, the members of the proposed committee should have power and control over budget and expenditure decisions. In the case of Colombia, this committee should comprise the President, Minister of Finance, and NPD Director. Nevertheless, the spending review framework can provide strong technical inputs to serve as the basis for this discussion. In addition, by following a structured process to carry out the reviews, this process can also increase transparency and scrutiny over expenditure decisions. h) Spending reviews should be complemented by a comprehensive analysis of programs and policies in the medium term. While a rapid assessment focused on the identification of efficiency gains in the allocation of resources for the coming fiscal year is useful, there are limitations which should be acknowledged. A deeper analysis regarding the performance and impact evaluation of programs or government policies over time lie outside the scope of spending reviews and require the implementation of additional instruments in the medium term. 12. Setting the guiding questions for the review process facilitates the orientation of the team throughout the process. In line with OECD practice, these questions usually seek to assess the degree of alignment of given policies or programs with the government priorities, as well as the need of government intervention to achieve the program or policy objectives, vis-à-vis private sector initiatives. 6 13. In the case of Colombia, the guiding questions focused on assessing the alignment of programs with the government priorities at the time and on program efficiency in the use of resources. Table 1 presents the guiding questions used for the spending reviews in 2016. Some of these questions were explicitly included in a framework for the identification of savings, whereas some were expanded to obtain greater detail in the analysis. These questions established an analytical framework for the spending reviews and were operationalized at three analytical levels: (a) strategic; (b) alignment with program priorities and performance assessment; and (c) operational. Table 1. Guiding questions for Colombia´s Spending Reviews (2016) Alignment with government priorities • Review the Presidential goals, the National Is the program aligned with the government policy Development Plan, and the Strategic Sector and priorities? Institutional Plans, etc. Performance • Review achievement of related goals from the National Development Plan To what extent has the program achieved • Review the achievement of goals of related government policy priorities? public investment projects • If available, review program impact evaluations Efficiency • Analyze the program objectives, target Is there overlap with other programs? population, and beneficiaries • Collect user feedback Is there space to improve the delivery of services • Review program costs or do it at lower costs? • Carry out user satisfaction surveys • Analyze administrative procedures to identify Is there space to merge programs and share simplification opportunities services or simplify administrative arrangements? • Identify opportunities to pool and share resources • Carry out market studies to establish the Is it possible to outsource or involve a non-state feasibility of outsourcing delivery of specific service provider at a lower cost? program services and activities at a lower cost (1) Strategic level 14. This level focuses on sector priorities and objectives and reviews the suitability of existing policies required to achieve them. Overall, the analysis reviews whether the strategic objectives and priorities of a given sector can be achieved with the current policies or if they need to be adjusted. The identification of such gaps will provide inputs to develop resource reallocation proposals that could generate potential savings by means of restructuring. Moreover, it intends to shed light onto the need of potential structural reforms in case significant gaps are found. This analysis is carried out based on international experiences and impact evaluations (if available). 7 15. A macro level analysis should ideally be done at the start of the spending review process, looking at current policies and the sector organizational structure, using senior technical experts. This is a key lesson learned from the Colombian experience where this step was implemented after the spending review teams had completed their work. This left little space to include the resulting recommendations into the savings and reallocation proposals. Furthermore, the political environment must be considered as the magnitude of some of the recommended measures require reaching the highest level of political commitments, including participation of sector ministries and vice ministries, especially when they envision structural reforms. 16. Restructuring sector policies may increase both the effectiveness and efficiency of government programs; however, implementation takes time and requires considerable political support. The strategic analysis of existing policies vis-à-vis the declared reform and priorities as part of a spending review process can open the door to explore a potential institutional overhaul to consolidate potential savings identified as a result of the analysis. In this respect, the construction of political support and commitment is key to effectively generate savings proposals and reap the benefits of the proposals. (2) Alignment with priorities and performance assessment at the program level 17. This level focuses on assessing the allocative efficiency of public expenditure by assessing program performance and their alignment with policy priorities at the strategic level. The analysis is based on the assessment performed at the strategic level discussed above, which outlines the key priority policy. A given program is deemed to be efficient if spending is aligned with priority policies and is well performing in terms of results achieved. The resulting proposals aim to increase the alignment of the budget with the achievement of government priorities. 18. Policy priorities must be clearly outlined and explicitly defined to carry out the alignment analysis at the program level. Policy priorities are considered as an exogenous variable for assessing the alignment of programs. In the case of Colombia, key sources of information in this regard are the National Development Plan as well as the sectorial and institutional plans. 19. Moreover, it is crucial to establish a clear link between programs and the associated budget. These links are the focus of analysis and must enable tracking of the allocation of resources. Without them, it will not be feasible to carry out a meaningful spending review. 20. The analysis also requires performance monitoring data (if available) to assess the achievement of results. While it is still possible to carry out a spending review with limited data on performance (as was the case in the Canada and the United Kingdom reviews), such reviews will provide little information regarding the effectiveness of public expenditure as their focus on program alignment only captures efficiency in the allocation of the budget. 21. When considering both efficiency and effectiveness in the allocation of public resources, programs can be classified into four distinct categories: (a) high priority and high performing; (b) high priority but low performing; (c) low priority but high performing; and (d) low priority and low performing. This characterization makes it possible to quickly identify programs that could either benefit from additional resources in case their performance is associated 8 with budget restrictions or the ones that could be scaled back to produce savings or closed. Figure 1 depicts the features of the classifications. Figure 1. Performance and Priority Matrix Performance High Low Priority High-priority and low- High-priority and high- performing programs. performing programs. Requires improvements to High Candidate to receive additional strengthen performance. budget if required. Potential candidates to receive additional budget. Low-priority and high- Low-priority and low- performing programs. performing programs. Low Potential to provide resources Requires restructuring. Primary for budget reallocations toward candidates to provide resources high-priority programs. for budget reallocations. (3) Operational level 22. This dimension focuses on improving the operational efficiency of public expenditure by identifying mechanisms that would allow programs to maintain their output production and objectives at a lower cost. This can be achieved through identification of potential economies of scale in the production of goods and services or simply by cutting back redundant spending. The guiding questions to carry out the assessment include, among others: what are considerations regarding the feasibility of sharing administrative services to lower costs, and what is the degree to which a program overlaps with another initiative and duplicates work? Table 2 summarizes the strengths and weaknesses of the three analytical dimensions described thus far. Table 2. Comparison chart of weaknesses and strengths of the three analytical levels Level Strengths Weaknesses • Savings are obtained in the long run • Generates long-term savings • Requires high-level and constant • Helps to solve sector structure or policy political support Strategic alignment challenges • Recommendations can be costly and • Enhances impact of cost-effectiveness difficult to implement measures • High aversion to change Alignment with • Generates savings over short and medium • Requires strong clear linkage between priorities and term programs and budget performance • Improves the alignment of expenditure • Requires quality program assessment with policy priorities performance information • Generate rapid savings • Easier to implement recommendations • Savings can be modest Operational • Establishes good practices in terms of • Savings tend to be one-time only value-for money and cost-effectiveness in actions sector agencies 9 23. The three analytical dimensions discussed in this section are parts of an integral spending review process and as such are not mutually exclusive. The emphasis on one or the other depends on country-specific context and characteristics. For example, in context where political support might be difficult to consolidate for complex reforms, but rapid actions are needed, more emphasis should be made on the operational level rather than on proposals focused at the strategic or alignment and performance levels. 24. The implementation of this framework can contribute to significantly increase the quality of public expenditure by reallocating spending across programs based on policy priorities and program performance. Programs with little alignment to the government policy priorities and ineffective in terms of performance generate low-quality expenditure. These resources could arguably be reallocated to high-priority programs targeting pressing concerns, thus increasing efficiency in the use of public resources. B. Political leadership for the public spending reviews 25. In Colombia, the initiative to carry out spending reviews came from the National Planning Department, which provided the necessary political support. Specifically, the NPD’s Territorial and Public Investment Subdirection (Subdirección Territorial y de Inversión Pública, STIP) led the implementation of the exercise and acted as liaison and coordinator with the sector agencies involved. As part of the implementation process the National Planning Department selected the pilot sectors and STIP disseminated the methodological approach at the sector level as well as the objectives of the exercise and agreed on saving targets for the public investment budget. 26. Colombia’s National Planning Department is responsible for the management of the public investment budget and the development of the National Development Plan, which covers a four-year period. While ministries of finance are often the lead actor of expenditures reviews in other countries, in Colombia the legal authority over the management of the budget is split between the Ministry of Finance and National Planning Department, which oversees progress in the implementation of public investment projects and the monitoring of results, centralizing performance information of government programs. In addition, the National Planning Department is in charge of establishing guidelines for preparing public investment projects and approving or rejecting any budget modifications of public investment projects presented by line ministries or other government agencies. However, final decisions over the consolidated budget are made by the Ministry of Finance, which is in charge of managing current expenditure and operating budget, which is the lion’s share of government spending. This separation of duties created a challenge for the implementation of the spending reviews. 27. As the institution leading the process, the National Planning Department coordinated with the ministries that comprise the agriculture and social inclusion sectors to obtain their support for implementing the spending review. The NPD Director sent letters to the Agriculture Minister and to the Social Prosperity Department (which like National Planning Department has ministerial rank) requesting their support as sector heads, as well as the collaboration of their staff for the spending review. In addition, STIP liaised with the planning offices of these institutions to explain in detail the methodology of the review and the human resources and time needed for implementation. 10 C. Objectives of the public spending reviews 28. The main objective of the spending review exercise in Colombia was to improve the efficiency of public investment spending. Overall, spending reviews can serve different purposes. The consolidation of fiscal sustainability is perhaps the most evident one. As was previously discussed, this can be achieved through reductions in the overall level of public expenditure either by means of reallocating resources from low-priority and low-performing programs to high-priority and high-performing ones or by improving government operating efficiency. In the case of Colombia, the spending review process also included high-level strategic considerations regarding the alignment of programs with the government policy priorities. 29. The agriculture and social inclusion sectors were selected as pilots by the National Planning Department due to their strategic importance for the National Development Plan and given the potential negative impact of expected across-the-board reductions in the 2017 budget. In this regard, the review sought to provide inputs to implement budget reductions in a strategic way. However, the schedule for the spending reviews was not aligned with the calendar for budget formulation. This implied that the spending review process could not be completed by the time the formal budget allocations discussions took place, and thus its recommendations were not considered in the 2017 Budget Law. Nevertheless, the National Planning Department was set to guide the exercise with focus on allocation of a share of the public investment budget for the selected sectors. 30. The National Planning Department established a goal to generate savings amounting to 10 percent of the unrestricted public investment budget of the target sectors. The unrestricted budget comprised the available public investment resources and did not include commitments made beyond the current fiscal year as well as resources from international cooperation. The spending reviews were expected to generate savings and budget reallocation proposals to optimize public investment expenditure. D. Scope of the reviews and features of Colombia’s budgeting system 31. In Colombia, the spending reviews were focused on the public investment portion of the national budget. International experiences are usually focused on operational spending, and only on rare occasions intervene on the budget of on-going or shovel ready public investment projects. In the case of Colombia, however, the definition of public investment is rather broad. As per the current classification, public investment does not only include investment on physical capital but also transfers and other operational expenses related to investment on both social and human capital comprising, for example, the salaries of teachers or medical personnel. In 2010, only 45 percent of the overall public investment budget funded capital expenditures as defined by the classification of the 2014 IMF Government Finance Statistics Manual. This broad definition of public investment was made to consider the fact that the investment on social and human capital are as important as investment made on physical assets. 32. Public investment projects were, in turn, the main unit of analysis for the spending reviews. Given the broad definition of public investment in Colombia, most projects include operational expenses and few contain pure capital investment as defined by IMF. However, each project has a performance framework, including indicators to monitor the achievement of activities 11 as well as results and impact goals. In this regard, public investment projects in Colombia function in effect as budget programs by linking expenditure with performance measures. 4 33. The broad definition of public investment allowed the review to assess not only pure capital investment expenditure but also the allocation of current expenditure within projects. However, operational expenses were only considered in the extent to which they were clearly linked to the achievement of specific products. 34. The spending reviews were focused on the agriculture and social inclusion sectors using information from 2015 and 2016. However, in Colombia, the sector classification reflects the government administrative organizational structure. To address this limitation, the reviews included all entities and agencies that were formally mapped under those sectors. In the case of the agriculture sector, the reviews included an assessment of the expenditures made by the Ministry of Agriculture and Rural Development, Rural Development Agency, National Land Agency, Rural Agricultural Planning Unit, National Agriculture and Fishing Authority, and Land Restitution Unit. Regarding the social inclusion sector, the review comprised expenses made by the Social Prosperity Department, Colombian Institute of Family Welfare, Victims Support and Integral Reparation Unit, and National Center of Historic Memory. E. Stakeholder roles and responsibilities 35. The implementation of spending reviews can follow different approaches. Based on experiences from OECD countries, there are three main options to implement the reviews. First, a bottom-up review by which the sectors are responsible for preparing their own savings proposals based on MoF guidelines. Second, a top-down review process implies that the savings and reallocation proposals, while centrally prepared by the Ministry of Finance, have limited sector participation. In contrast, a third joint-review approach implies that the proposals are prepared by teams with participation from both the Ministry of Finance and the sectors. 36. For the Colombian case, a joint-review approach was recommended, with participation from the National Planning Department, Ministry of Finance and sector entities. This approach allowed the teams to have more profound and up-to-date knowledge of the policies and projects from each entity while retaining the overall perspective of national and fiscal priorities. In addition, by participating in the joint process, the ownership of the reallocation proposals is increased, which is crucial for their effective implementation. 37. In terms of the governance structure for the process, a Steering Committee and review teams were established. Their specific tasks were as follows: a) Steering Committee. Overall, the Committee’s objective was to coordinate the spending review process (including administrative aspects), define the savings goals and participating institutions. The Steering Committee was also charged with approving guidelines and templates for the implementation of the methodology and monitoring the implementation of the proposals. The Committee was composed of staff from the National 4 In 2015, a World Bank team carried out an assessment of Colombia´s Public Investment Management practices. As part of its recommendations, the study suggested the development and implementation of differentiated procedures and metrics for strictly pure public investment projects, and those that include current expenditure. See World Bank (2015) for additional reference. 12 Planning Department and its STIP, as well as representatives from Direction of Public Investment and Finances (Dirección de Inversiones y Finanzas Publicas, DIFP), Direction for Monitoring and Evaluation of Public Policies (Dirección de Seguimiento y Evaluación De Políticas Públicas, DSEPP), and the technical departments of relevant sectors. b) Review teams. The core task of the teams was to identify potential savings and prepare budget reallocation proposals to be discussed by the Steering Committee. The teams comprised staff from the National Planning Department and from the sectors and entities subject to review. Figure 2. Organization framework for the 2016 spending reviews 38. The WBG team developed the methodological tools for the review as part of its technical assistance to the National Planning Department. The support included consolidating and constructing the necessary databases that were the main input for the spending review as well as providing guidance to the review teams for the implementation of the methodology. The WBG team comprised senior international experts with extensive knowledge of the sectors and provided guidance to the review teams and analytical insights to the National Planning Department and the sectors involved. 13 III. Implementation of the Spending Reviews in 2016 39. The spending reviews were to be implemented in several phases based on the Colombian institutional setting as well as international experiences. Table 3 summarizes the implementation phases and the processes they entailed. The World Bank Group provided support during phases 1 and 2 as well as for selected processes in phases 3 and 4. Table 3. Summary of Spending Review Phases and Processes Phases Processes • Obtain high-level political mandate to perform the spending review Phase 1. Establishing a • Establish Steering Committee mandate and a • Define the review scope, including its overall targets sector targets, guiding questions, framework for the schedule, and organizational arrangements spending review • Prepare terms of reference for the review teams and guidelines for each dimension of analysis • Define the analysis unit and necessary information. Phase 2. Gathering of • Staff the review teams. information and database • Gather and consolidate data on single database. organization • Preliminary assessment of alignment with priorities and performance. • Carry out the strategic-level analysis. Phase 3. Explore and • Draft recommendations. implement strategic • Explore high-level political support for the implementation of strategic-level analysis* recommendations. • Design an implementation plan for strategic-level recommendations. • Carry out joint review of the alignment of priorities and performance assessment. Phase 4. Develop savings • Identify potential savings. proposals • Discussion of alternatives and drafting of savings proposals. • Each entity discusses and approves proposals that do not imply a breach in the achievement of an entity’s target involve a single entity. • The steering committee discusses and approves proposals that imply changes in the achievement of an entity’s target and involve more than one entity. Phase 5. Decision-making • If necessary, discuss and approve centrally generated savings proposals prepared by the National Planning Department or the Ministry of Finance. • If necessary, the central level identifies additional savings by following the spending review methodology. • Implement savings proposals during the budget programming stage or through budget Phase 6. Proposal modifications and decrees to postpone appropriation of resources. implementation • Monitor the implementation of savings proposals. * Phase 3 can be implemented in parallel with phase 2 or upon completion of phase 4. 14 A. Phases One and Two 40. As mentioned earlier, the National Planning Department provided the political support required and led the implementation of the spending reviews. Preliminary work began in early 2016 with the establishment of a Steering Committee and the drafting of terms of reference. The Committee had representatives from STIP, as well as from the MoF Investment and Public Finances Directorate and the NPD Public Policy Monitoring and Evaluation Directorate. Among its other tasks, the Steering Committee defined the scope of the review process, schedule, targets, and sectors that participated in the review. The WBG team prepared the terms of reference to guide the work of the review teams. With final approval by the Steering Committee, the terms of reference outlined the review objectives and methodology. 41. The next steps involved defining a unit of analysis for the review and consolidation of information in a single database. As mentioned earlier, public investment projects were selected as the main units of analysis given the possibility of establishing a link with performance results indicators and the budget at this level of granularity. However, some large projects had to be broken down into smaller units tied to specific and clearly identifiable results indicators. Once the unit of analysis was identified, the teams needed to gather information on the achievement of both performance indicators and historical budget-execution rates. This process proved difficult since the National Planning Department did not have a tool to produce consolidated reports linking both budget execution and performance information. The information was scattered across seven different databases. Furthermore, some projects were more recent and did not yet have performance measures. 42. A preliminary assessment of the alignment with policy priorities and performance was then carried out based on methodology designed with WBG support. The methodology sought to associate each public investment project with a performance and priority score in a straightforward and transparent way. As a first step, the team reviewed the sector-related objectives stated in the National Development Plan and those of the sector strategic plans, and then compared them with the project objectives. A score was then assigned based on the fit of the project objective with the government goals and policy priorities. With this information a priority ranking sorted the projects from high to low alignment. Similarly, the team reviewed the achievement rate of project objectives 5 and financial execution rates. These rates were then used to construct a project performance ranking that sorted the projects from high to low performance. However, since new projects had no performance information, they were only considered in the priority ranking. A preliminary assessment using this methodology was implemented by the World Bank with a team of local consultants with sector expertise. The results were intended to serve as a guiding input for the review teams and were validated with both the National Planning Department and the participating sector entities. The simplicity of the methodology contrasted with the usual NPD approach, which tends to favor the use of complex statistical methods that are often unclear for non-specialized personnel in the sector entities. 5 The methodology for preparing public investment projects in Colombia establishes that project objectives should be associated with products, defined as the goods and services provided to the citizens upon completion. Although several of the reviewed projects complied with these guidelines, some projects contained objectives associated with the completion of activities, and in some cases with expected impacts (the results of delivering goods and services). In this regard, objectives associated with impacts must be linked to high-level budget programs rather than individual projects. 15 43. Implementation of phases one and two took approximately six months to be completed due to unforeseen delays in the definition of the framework for the spending reviews. The activities of phases one and two were finalized in July 2016 with an important lag in the schedule. While these phases were expected to be completed over a maximum three-month period to initiate the fieldwork of the review teams in March or April, significant delays were encountered during consolidation of the information. In this regard, coordination within NPD directorates that were part of the Steering Committee proved to be challenging, particularly with regard to a joint definition and agreement on the scope of work and approach of the spending reviews. B. Phase Three 44. This phase was focused on the analysis of the information at a strategic level and could thus occur at different times throughout the review process. In the case of Colombia, a team of high-level sector experts reviewed the sector strategic objectives and priorities to assess if they could be achieved with current policies in place or if adjustments were required. Such adjustments can generate savings; the experts were instructed to analyze any gaps in the alignment of policies with objectives in search of potential savings that could be achieved in the medium and long term. This implied that the resulting policy reform proposals were not neutral in a fiscal and budgetary sense and were to lead to increases of public expenditure over time. This phase began in late August 2016. However, phase three could be implemented either in parallel to phase two or upon completion of phase four. In addition, the experts participated in some of the meetings of the review teams to inform their proposals, which were also validated by the National Planning Directorate and the participating sector entities. C. Phase Four 45. Phase four covered the development of savings proposals by the field-based review teams. To guide the review on the sectors, terms of reference were prepared and approved by the National Planning Directorate. Implementation of this phase began in August 2016 with the formal establishment of review teams for the agriculture and social inclusion sectors. The review teams comprised staff from STIP, the Public Policy Monitoring and Evaluation Directorate, the Rural and Social Development Directorates, and the institutions subject to review. The first task of the teams was to review the terms of reference for the exercise and agree on a timeline for review implementation. During this phase, a WBG team acted as moderator and facilitator. 46. A key task was the review of the project portfolio of each entity and preliminary assessment in terms of alignment with priorities and level of performance. During this phase, the teams validated the information of the project portfolio and identified projects that because of their nature and size needed to be disaggregated in smaller components for the analysis. The review of the results from the preliminary assessment was agile and required two work sessions since the methodology required explanation beforehand. Figure 3 depicts the social inclusion projects grouped by their performance and priority scores (priority is understood as and alignment with government priorities). 16 Figure 3. Project performance and priority assessment in the social protection sector (low to high) PRIORITY PERFORMANCE (low to high) Source: Authors, based on scores from the preliminary assessment agreed with the review teams. 47. While the scores and graphics from the preliminary assessment did provide insight in terms of alignment with priorities and performance, they should not be the only criteria taken under consideration for immediate decisions over budget allocations. The scores allowed the identification of potential candidates for budget reallocations but needed to be complemented with additional criteria. For example, the preliminary assessment found in cases of low-priority projects connected with support processes (such as maintenance of information systems) and budget cannot be reduced without compromising the entity operations. On the other hand, there might be high-priority and well-performing projects that include inefficient expenditure items. 48. Savings proposals were to be developed once the review teams reached a consensus regarding the results of the priority and performance assessments. Potential savings were identified with a guide prepared with WBG support. This guide contained instructions for the review teams and an analytical framework to assess the expenditure efficiency of budget items at an operational level. The guide recommended that teams begin with an examination of low-priority projects to identify potential savings opportunities and then assess the expenditure operational efficiency of each project regardless of their priority and performance. 49. Savings identified by the entities were discussed within the review teams for consideration, using the same analytical framework to assess budget efficiency at the operational level. To be included in the final proposals, the alternatives required concurrence from all team members. The proposals were then summarized using standard templates requiring the 17 teams to identify the project and describe rationale for the modification, as well as the amount that could be reallocated (saved). In addition to this information, the review teams were also required to clearly state if the proposal would have a potential impact on the achievement of the project objectives. The proposals were revised and approved by the teams during two separate work sessions and finalized in November 2016. 50. However, the savings proposals did not achieve the saving targets set at the start of the exercise. There was no follow-up action by the National Planning Department or Ministry of Finance to identify additional cuts as was originally designed. This was a result mainly from the misalignment of the budget preparation calendar and the implementation timeline for the spending review, and the fact that sizeable cuts had been already defined in the 2017 Budget Law. Thus, the government had no need to enact additional budget reductions. D. Phases Five and Six 51. These phases foresaw the enactment of the savings recommendation and their implementation. However, given the challenges discussed above, they have yet to be executed. Overall, by the end of 2016, it was agreed that the heads of the sector entities and the National Planning Department would make a final decision on recommendations in early 2017 and would then agree on implementation plans with participation of the Ministry of Finance. III. Results, lessons learned and recommendations 52. Overall, the results achieved give reasons for optimism and suggest that there is an opportunity to institutionalize spending reviews in Colombia. The positive first experience and the fiscal outlook favor the implementation of initiatives to improve the prioritization of public spending. In this regard, by the end of 2016, the National Planning Department, with the technical advice from Fedesarrollo, a local think-tank, was exploring the possibility of implementing spending reviews in five additional sectors to provide inputs for the 2018 budget. A. Results and lessons learned 53. Upon completion of the spending review exercise, the teams encountered specific reallocation opportunities that could be implemented during FY2017. Nevertheless, neither sector achieved the 10 percent savings target established by the National Planning Department. In the case of the social inclusion sector, the target amounted to approximately COP 280 billion (US$93.3 million 6), whereas the savings proposals accounted for COP127 billion (US$42.3 million), nearly 4.5 percent of the total unrestricted public investment budget. Likewise, in the agriculture sector, the target amounted to COP 64 billion (US$21.3 million) while the savings proposals accounted for COP 28 billion (US$9.3 million), approximately 4.4 percent of the unrestricted budget. Among the entities that identified savings that surpassed the target as share of its flexible public investment budget were the Victims Support and Integral Reparation Unit from the social inclusion sector; and, from the agriculture sector, Rural Development Agency, National Agriculture and Fishing Authority, and the Land Restitution Unit. All other entities identified savings short of the 10 percent target. 6 The exchange rate used was USD 1 = COP 3,000.71, which was the exchange rate reported by the Bank of the Republic (Colombia’s central bank) for December 31 of 2016. 18 54. A key reason for missing the target related to limited high-level political support provided to the exercise. As was mentioned in the previous sections, letters requesting support and signed by the NPD Director were submitted to the Minister of Agriculture and to the Director of the Social Prosperity Department. While both conveyed their willingness to comply and notified their staff accordingly, they did not stress the importance of direct participation of staff from the sector entities in the review team, thus leading to limited involvement. Furthermore, despite coordination meetings that took place between the National Planning Department and the Ministry of Finance, the limited involvement of the Ministry did not generate a credible incentive for review teams. The MoF staff that participated in the coordination meetings did not receive clear instructions or a mandate from the Minister to implement the review. This lack of political support, combined with the fact that the 2017 budget reductions were defined before the review process took place and no real threat of unilateral action for not meeting savings targets existed, eroded the main incentive for sector participation. 55. The strategic-level analysis and proposals were deemed useful by the sector review teams, however, high-level discussion of results to explore the feasibility of implementing the proposals did not occur. This was a significant challenge since decisions involving reforms needed to be approved by high-level authorities before implementation. The discussions did not happen partly because the strategic-level proposals were presented to the sector teams late during the review exercise, and at that stage the savings proposals had already been prepared. In addition, the sector ministers had limited interest and incentives to discuss the proposals with the senior experts. 56. Involving the Ministry of Finance as an active partner in the design, implementation, and coordination of spending reviews was a necessary condition for the success and consolidation of the process. If the Ministry of Finance had not joined the process in a protagonist role together with the National Planning Department, it would have been difficult to ensure the effective implementation of saving proposals. Moreover, current expenditure, which accounts for a key share of the budget, could be left outside the scope of the reviews. 57. Likewise, defining budget reductions prior to the spending review left little margin and incentives to identify additional savings. While the work of the review teams began in September 2016, the Ministry of Finance had already announced the 2017 budget cuts to each entity in June, along with a substantial reduction of the investment budget. In the case of the agriculture sector, the reduction amounted to nearly 12 percent and 30 percent in the case of the social inclusion sector compared to 2016 budget. 58. The anticipated budget reductions for 2017 could have served as a useful guide to the process, if the timeline for the review had been aligned with the budgeting calendar. This would have implied that the review began its work a couple of months earlier so that results became available in May in time for discussion and implementation of the savings proposals as part of the notifications issued in June. 59. The structure of the budget information in Colombia also limited the scope of the spending reviews. Overall, multiple challenges were encountered during the process in terms of 19 quality of the information. A methodology for the definition of results-based budget programs was designed in 2015 to better associate policy goals with public expenditure. This methodology comprised a framework to consolidate public investment projects and link the products produced at the project level with the program results. Nevertheless, the methodology was only implemented for public investment budget, and thus it is not feasible to link recurrent and investment expenditures with policy goals and performance measures. Moreover, the current programs are defined by high-level policy goals and are associated with the budget needed for them at a highly aggregated level. This feature makes it unfeasible to implement spending reviews given the broad nature of programs as analytical units. For this reason, one of the main recommendations to the National Planning Department was to consider the implementation of sub-programs that would provide an intermediate level between the high-level budget programs and the individual projects. Nevertheless, while projects do operate as budget programs and cover a considerable share of the budget, their structure is not homogeneous and the quality of the information provided at this level is lacking. Finally, the National Planning Department did not have a support tool to generate consolidated information reports at the project level, combining both performance and budget data. This made consolidation a highly time-consuming, manual task. 60. Although implementation of the spending reviews in two sectors proved challenging, it laid the foundation for institutionalizing this process in the future. The United Kingdom, the Netherlands, and other countries have required several years to refine their spending review methodologies and develop the necessary technical capacity to lead the process. Considering this, the Colombian experience was a valuable lesson to rolling out spending reviews in other sectors. 61. A crucial first step toward institutionalizing spending reviews is adapting the design and process of the methodology to country-specific context. Special attention was given during the design of the methodology to adapt it to the Colombian budgeting context. In this regard, the implementation of the spending reviews in two representative sectors led to the identification of concrete savings proposals and provided feedback to refine the methods used. In addition, while both the National Planning Department and the Ministry of Finance had tools providing them information for budget allocation negotiations with sector entities, information was highly asymmetrical in terms of power and knowledge. Both the Ministry of Finance and the National Planning Directorate have more power over the budget than sector entities; however, they lack detailed knowledge of sector policies, priorities, and expenditure allocation. By establishing a joint approach to spending reviews, this process contributed to generating trust and reducing information gaps. 62. The spending reviews established a precedent for joint work between the National Planning Department and the sector entities, in terms of information sharing and the achievement of concrete results. The methodological framework for the spending reviews explicitly acknowledges sector knowledge over their policies, priorities, and budget. This represented a positive shift in National Planning Department´s usual work dynamic where collaborative work between public sector institutions in Colombia is not common. The methodological approach taken by the spending reviews was able to cover information gaps and facilitate the analysis. These advantages were acknowledged by NPD and sector staff that participated as team members during the review process and considered the methodology to be easy to use, with potential to improve the quality of Colombia´s public expenditure. 20 63. Although the savings targets were not met, the results achieved in both sectors were significant. The savings proposals identified were substantial and were validated by the National Planning Department and the sector entities. This was a breakthrough considering that this was the first time sector spending reviews had been implemented in Colombia and that political support and alignment with the budget calendar was not optimal. Furthermore, the spending review exercise proved its potential to improve the quality of public expenditure through the identification of inefficient and low-priority expenditure. The identification of potential savings opens the door for enactment of measures to increase resource allocation efficiency. The next big step in this regard will be deciding to implement the proposals during the Colombian budgetary calendar. 64. Finally, it is important to highlight that the strategic analysis conducted as part of the spending review allowed identifying concrete proposals for strategic policy reorientation in both sectors, with the potential of generating savings over the medium term. In the case of the agriculture sector, reform recommendations involved subsidies reduction and increases in the production of public goods; whereas in the social inclusion sector, a key proposal was modifying the conditional transfer program, Familias en Acción, to mobilize program resources from urban to rural areas where the impact of the program could be greater. The high-level policy recommendations could have a profound effect over the budget allocations, completing the results of the spending review analysis at an operational level. However, there are currently no concrete plans to implement these proposals, and it is not envisioned that they will be enacted given the limited participation of government officials with decision-making power, and the focus on short- term savings. B. Recommendations 65. The most important condition for the institutionalization of spending reviews in Colombia is obtaining high-level political support. A high-level political committee should support the second wave of reviews at a level above the Steering Committee and the review teams. Comprised of the President, NPD Director, and Minister of Finance, the proposed committee should provide the mandate to carry out the review and decide on potential reforms formulated at the strategic analysis level that imply potential savings or that could affect more than one sector. In addition, the committee should decide the sectors subject to review and the relevant savings targets. 66. The implementation schedule for the spending reviews should be aligned with the budgetary calendar and should provide sufficient time for discussion of savings proposals and their inclusion in the formal budget allocation process. To achieve this, it is necessary to review the budget preparation calendar to include the expenditure as part of the process. In several countries, spending reviews are completed in time to feed into the negotiations over budget allocation between the sectors and the Ministry of Finance. 67. Savings proposals should be delivered in time for drafting the Budget Law or by the time the medium-term expenditure framework is updated. Depending on the available time, the Steering Committee can prepare an action plan and agree with the sector team on the milestones 21 for the review. As a reference, the spending review process in two sectors took approximately 12 months to be completed. 68. Some modifications to the budget structure could be considered in order to ease the implementation of the spending reviews. Specifically, consideration should be given to the development of budget programs that encapsulate both capital and current expenditure and the instruction of classifications as stated in the IMF 2014 Government Finance Statistics Manual. These measures would improve the quality of the information available for spending reviews. 69. A budget program classification should be developed not only for the public investment budget, but also for recurrent expenditures. The Government may consider the possibility of developing sub-programs to better associate policy goals with public expenditure. Thus far, the implementation of budget programs has been led by the National Planning Department with no participation from the Ministry of Finance. Increased coordination between both entities is crucial for the success and sustainability of the spending review process and the implementation of recommendations to improve the quality of public expenditure. 22 References HM Treasury (2010) The Spending Review Framework. HM Treasury (2015) A country that lives within its means. Spending Review 2015. Holda, Paulina and Emilia Skrok (2015) “Spending Review Pilot: Low-Income Family Support in Poland.” World Bank. Madzarevic-Sujster, Sanga, Theo Thomas y Suzana Petrovic (2015) “Croatia: Spending Review Pilot Process and Lessons Learnt”. Macroeconomic & Fiscal Management Practice Notes 10. World Bank. German Federal Ministry of Finance (2016) “Improving the German Federal Budget: Spending Reviews.” OECD (2013) Colombia: Implementing Good Governance. OECD Public Governance Reviews, OECD Publishing. Olden, Brian, Jason Harris, Amanda Sayegh, Duncand Last y Chris Uregian (2015) “Republic of Slovenia. Establishing a Spending Review Process”. Fiscal Affairs Department, International Monetary Fund. Robinson Marc (2013) “Spending Reviews”. Annual Meeting of OECD Senior Budget Officials. OECD, Working Party of Senior Budget Officials. Stacey, Bruce (2015) “Expenditure Reviews: The Canadian Experience.” Presentation on the Public Spending Quality Seminar organized by the World Bank and the National Planning Department in Bogota on September 2015. Vandierendonck, Caroline (2014) “Public Spending Reviews: Design, Conduct, Implementation.” Economic Papers 525. European Commission. World Bank (2015) “Modelo de gestión y control para la inversión pública en Colombia.” Part of the Reimbursable Advisory Agreement signed between the World Bank and the National Planning Department. 23