1
Draft
7-5-00
GGuess
PUBLIC FINANCIAL MANAGEMENT REFORM IN HONDURAS:
A CASE STUDY
By Bill Dorotinsky and Yasuhiko Matsuda
Abstract
The Sistema de Administracion Financiera Integrado (SIAFI) is a relatively small system composed of four core modules installed in the ministry of finance (SEFIN) and four other central ministries of the GOH. Its emphasis has been accounting, reporting and control of budget execution. SIAFI was developed as part of the new Modernization of the State program and was viewed both as a mechanism for strengthening existing financial management systems and as a catalyst for more profound institutional and systemic changes in governance, such as fiscal decentralization. Since very little had been accomplished in the state reform program at the point of SIAFI initiation, subsequent changes both in PEM and wider state reform can be attributable directly or indirectly to SIAFI. SIAFI’s implementation began in 1997 and was interrupted by Hurricane Mitch in 1998.
Nevertheless, substantial systems training and module implementation has occurred. Three of four modules are operative at the four ministries and a network is in place for exchange of information between them and SEFIN. Further progress has been impeded by: laws and practices inconsistent with wider budget discretion at lower administrative levels, by institutional fragmentation which has weakened expenditure control at the center (e.g. the treasury department of SEFIN, Tesoreria General de la Republica or TGR), and by the culture of centralization that has worked against deregulation of administrative process.
I. Introduction
Documented experiences with the design and installation of integrated financial management systems (IFMSs) in Latin America have accumulated over the past 10 years (Asselin, 1999a; Bartel, 1996; Guess and Jutkowitz, 1999). The major research question has been whether IFMSs have contributed to improved behavioral and technical aspects of fiscal management. Can SIAFI-type systems break through the barriers imposed on policy-making and management execution of programs by the “top-down financial control model” (Reid, 1994) of Latin America?
In all IFMS programs, the expected sequence of improvements or operating premise assumes that: properly designed and installed IFMSs (1) will generate improved data (e.g. contingent liabilities, accrued obligations), (2) which permit technically qualified decision-makers to develop analytic information (e.g. medium-term maintenance requirements; tighter fiscal policy; service performance requirements; re-engineering of internal process needs), (3) which will permit improved allocation of resources consistent with national policy priorities (e.g. health and education), which will then allow (4) improved efficiency and effectiveness of program operations. The question is whether this virtuous sequence of events has happened, and what bureaucratic and other factors account for success or failure?
This is an evaluation or case study of the Honduran public financial management reforms that were initiated in 1996 as part of a broader program of public sector modernization. It should be noted that none of the above sources mentions the Honduran reforms, in part, because they are so recent, but possibly also from the perception that not much progress has been made here. If this is the perception, it can be argued that it is not true. The core of these reforms has been an IFMS known as SIAFI.
Based on the TOR, the study will address four questions: First, what was the context of the public financial management reforms and how did the reforms appear on the policy agenda? Second, how successfully was SIAFI designed and implemented? Third, what has been the broader impact of SIAFI on public sector reform generally and PEM systems and behavior in particular? Fourth, what factors, conditions or determinants have contributed to the success or failure of SIAFI? In order to answer these questions, we used data generated from the attached survey (Appendix A) and secondary sources.
II. Summary of SIAFI Progress
SIAFI progress to date in five major areas can be summarized as follows:
*Design: First, SIAFI was fitted into an ongoing modernization of the state program. SIAFI was expected to be the catalyst for the effort that stressed re-engineering of state bureaucracy. The SIAFI design phase assessed user needs. It is clear that process re-engineering did not take place in design prior to purchase of software and implementation of SIAFI. Needs assessments were performed. But units such as the accounting department of SIAFI (Contaduria), despite its planned leadership role, were largely ignored when it became evident that it lacked sufficient technical capacity to perform necessary functions such as property management and inventory control. More importantly, vertical and horizontal integration of SIAFI with existing administrative units and information systems was not thoroughly reviewed before implementation.
*Data Generation: Second, data generation via SIAFI’s modules is preceding more or less as planned for SEFIN and three of four pilot ministries. SIAFI began with the budget module to develop counterpart support. It was concluded early on that focus on all four modules would diffuse counterpart support and weaken implementation. The modules now provide data in real time to decision-makers in SEFIN and to a lesser extent in the sectoral ministries. The budget execution module is quite advanced; there is much greater access to information on budget execution; the treasury module (beyond a narrow emphasis on payments) has not been developed as yet. The website is under construction and presents the Honduran government (GOH) as a professionally transparent operation. Any institutional impacts in deconcentration and decentralization of authority or responsibility for data use have been less apparent.
*Data Use: Third, legal restrictions still impede ministry flexibility and authority to manage budgets. New SIAFI data has been used for narrower improvements in documents processing. For example, the velocity of payments approvals has been increased. But pre-existing centralized pay order approval procedures in SED and other ministries vitiate SIAFI. How permitting and licensing processes have been streamlined to eliminate approval steps is less clear. External audit has moved from a transactional to a sampling basis which has reduced expenditure approval time. All modules except public credit are operative.
*Budget Framework/Document: Despite the relatively sophisticated website, the annual budget document still is very badly put together. In violating almost every major norm or standard (e.g. Government Finance Officer’s Association of the U.S. and Canada or GFOA) pertaining to the proper organization of budget documents, it details transparently neither inputs nor outputs of the GOH. The budget process is not built around an MTEF or multi-year expenditure planning exercise and does not produce a modern performance-based budget. The budget is the central incentive system of government. The budget process provides the singular opportunity to re-evaluate macroeconomic plans and results, program allocations and operations. Absence of a document that permits technical questions on input-output relations to be asked (a) weakens the chance for a rational assessment of expenditures, and (b) virtually eliminates staff incentives to generate data for expenditure analysis and control via systems such as SIAFI.
*Financial Management/State Reform: SIAFI has led to substantive improvements in processing payments and performing other basic GOH transactions. Wider improvements in state reform are being discussed but have not occurred to date. Much of the wider improvements depend on changes in law that have been opposed because of their major structural and political power implications, e.g. shifting management control of the unified account to the treasury from central bank.
III. Context of the Reform
Two Honduran background features in which SIAFI was designed and approved should be noted: (1) macroeconomic and political, and (2) microeconomic- financial. In 1995-1996, the Reina Administration faced a severe economic crisis derived from major electricity shortages that inhibited existing business activity and diminished tax receipts. On the expenditure side, rigidities due primarily to the fixed costs of public investments (debt service payments reached 31% of exports) and inefficient regulations prevented major expenditure cuts and discouraged private investment. Growing fiscal deficits in this period prevented the Administration from meeting its campaign promises to increase social sector expenditures. In this political-economic context, the Administration reached agreement with IMF on an economic reform program that required unpopular austerity measures to ration electricity, increase electricity tariffs, expansion of the sales tax base and elimination of many tax exemptions.
Faced with an inability to pull macroeconomic levers in the short-term to meet the growing political-economic crisis, the Reina administration focused in 1995-1996 on microeconomic problems. These included: poor quality social services that were both inefficient (high unit cost) and ineffective (lack of coverage); late and poor-quality budgeting and accounting data; lack of programmed maintenance for capital projects; and antiquated regulations that produced budgets inconsistent with policy priorities. In the context of political-fiscal problems at all three budgetary levels (discipline, strategic planning, and operations), the Administration embarked on a public sector reform program, of which the largest component was a World Bank-financed integrated financial management system (SIAFI). Modernization of the state preceded and led to the SIAFI effort which was viewed as a catalyst. In short, the financial management reform program appeared on the agenda in 1995 from the push of a local fiscal economic crisis and the pull of an IMF austerity program.
One should also ask the question, why was an IFMS selected rather than a MTEF as are found more often in Africa and elsewhere? Given the complex flow of technical systems ideas in international development, it cannot be definitively said that, for example: “the World Bank came up with this one and told the GOH to do it”. Nor can it be concluded that the cause was singularly, for example, that SEFIN learned of IFMS achievements in Bolivia, Argentina and elsewhere over the past 15 years and asked the Bank to produce a similar system in Honduras. Nor can we explain definitively why MTEFs are included in financial management reforms only at later stages in Latin America, rather than as part of the packages of initial reforms as in Africa. Physical project planning and programming has been a part of Latin American policy-making since the 1960s. But macroeconomic planning linked to budgeting has not been. In short, SIAFI’s appearance on the GOH agenda would have to be a combination of both flows of information.
IV. Evaluation of SIAFI Design and Implementation
A. Fiscal Crisis before SIAFI
As indicated in the previous section, in 1995-1996, problems could be identified at all three levels of budgeting that moved the GOH and the Bank to design a project of financial management reform. At Level 1, fiscal deficits in this period were the highest in Central America (10.5% of GDP in 1993 and 7.5% of GDP in 1994). The bulk of the deficit could be attributed to high and inefficient levels of public investment. Unsustainable numbers of projects were planned and financed by the budget that also increased accumulated debt to 101% of GDP. Planned investment budgets were regularly exceeded by the four main public enterprises (ENEE--electricity, ENP--ports, HONDUTEL—telecommunications, and SANAA—water and sewerage). Via direct transfers, the annual budget had to cover investment debt service (ENEE and SANAA) and public enterprise payrolls (SANAA). Financial and economic analyses of many projects were questionable. Political management of tariffs (consumer subsidies) decreased revenues and led to operational inefficiencies that contributed to deficits.
At Level 2, strategic planning and sectoral allocations were not coordinated. Inefficiencies in current and capital expenditures for example in education led to high unit costs and low service coverage. Sectoral policy goals and objectives could not be derived from expenditure patterns or results. In most cases, there were no sectoral strategies to begin with. This was largely a problem of political/ legal intrusions in a budget process (e.g. the dictamen requirement, an extra bureaucratic-legal “analytic” step that doesn’t exist in most other Latin American countries) that itself lacked basic information and technical capabilities. Financial management information needed for sectoral planning and budgeting was delayed (public enterprises delayed reporting for up to a year in some cases). Without timely and accurate accounting and reporting information on actual expenditures by departments, program costs, and year-to-date project expenditures, budgets were often made during execution instead of the formulation phase. Managers could not manage budgets properly. In short, because of changes during the year, congressional approvals and departmental requests bore little relation to actual expenditures.
At Level 3, operations and service performance were weak in most areas. Road planning and maintenance was the one exception. Health and education facilities maintenance and program operations have been weak. GOH regulations restricted private participation and competition in these areas which prevented improvements. The exception was road maintenance, which was contracted-out and for which performance was improving.
What emerges from this brief review is a catalogue of major weaknesses in budgetary processes and data/information systems facing the administration in 1995-1996. As described next, the Bank response via SIAFI focused largely on accounting and control systems and to a lesser extent on reforms in planning/budgeting such as MTEF/MYEP frameworks, or linkage of physical performance and fiscal data to expenditure planning and evaluation. The budgetary emphasis was on improved access to data during execution.
B. The GOH Informational Response to the Fiscal Crisis: SIAFI
Based on the assessment of fiscal management problems performed, the Bank developed a project with the following five objectives: (1) develop/implement the SIAFI computer system (hardware and software), (2) develop/implement the investment programming (and monitoring) computer system (SISPU), (3) develop methods and procedures to support basic financial management functions, e.g. budget performance indicators and project evaluation criteria, (4) develop a strategy to reorganize the institutional framework for financial management and investment programming, e.g. financial devolution and reduced/streamlined bureaucracy, and (5) train the technical staff in financial management units in GOH. In the view of some, SIAFI was narrowly conceived as an accounting information system for financial transactions. Only in the process of implementation did expectations arise that it was something broader, an integrated system of public management (Albano, 2000:4).
These are typical objectives for IFMS projects and the only questions would pertain to the specifics of design for users and proper implementation. It should be noted that the Bank uses other models in other regions. In Thailand, for example, budget modernization is the major goal and component parts of the “Budget Modernization” project: such as financial management and funds control include IFMS development. In that project as well as many others outside of Latin America, the budget function is viewed as the central incentive system of government into which support systems, e.g. accounting and procurement, feed data for decision-making. In Honduras, as in many IFMS-specific projects, the budget function is included as simply one of the financial management functions that need to be strengthened.
By contrast in Thailand, the Bank is encouraging technical gap-filling in areas such as: multi-year budget planning, procurement management, activity costing, performance reporting, asset management, and internal audit, prior to financial devolution. There is no such strategy in Honduras. In Thailand, each agency (ministry) signs an MOU with the BOB (like SEFIN) detailing compliance with checklists and specified thresholds in each area. The devolution will act as an incentive to meet standards; compliance with Bank standards will lead to more flexible, better-planned and cost-controlled budgets. Whether improved financial management decisions are facilitated to a greater extent by “budget-centered” or “system-centered” projects is an empirical question that to date has not been examined. As noted, the basis for this substantive design difference between the two regions by donors is not clear. While the comparative design of financial management projects is not specifically the theme of this case study, it can be said that the results from several projects in both regions are not dramatically different in resolving impact problems at all three levels of budgeting. This issue should be explored in another study.
C. Process of SIAFI Project Design
The needs assessment process for this project really occurred in parts over a 12-year period from 1985-1997. While the Bank’s formal needs assessment was much shorter (1995-1996), information had been generated on financial management needs since 1985. In 1985, Price Waterhouse did an initial study financed by USAID, which recommended an IFMS. A substantial portion of the assessment examined problems and needs in external audit (Contraloria General-CG). Following a request for assistance by the CG, USAID provided a $0.5 million contract to strengthen the CG as part of its new RFMIP I project (1988-90). The primary focus of this assistance from Price Waterhouse was accounting and auditing systems. A major product of these efforts was a new Organic Contraloria Law (Decreto 224-93); other financial management issues were not really addressed at this stage.