In the initial stages of the transition to performance-oriented budgeting, reforms in public expenditure management have largely focused on what may be termed the “macro” aspects of PEM reform—namely, putting in place new systems for budget preparation such as introduction of MTBFs and better-designed and more policy-relevant program structures. The broad thrust of this macro approach, which was described in the previous section, is to emphasize reforms within the central agencies of the government. Often, however, the viability of this new approach depends on supporting changes at the “micro” level of the budget system—that is, within the line ministries and other agencies of government that are required to modify their internal management procedures to successfully operate this new model of budget management.
There is increasing recognition within emerging economies that OECD-type budget systems must be based on a solid platform of financial management within such government institutions, which has led to a greater emphasis on the micro basis of broader budget system reforms. At the same time, however, many of these countries have been forced to acknowledge that this aspect of their PEM systems is problematic. Three areas related to internal management control systems have generally been found weak: internal audit, management information systems, and costing systems. This section explores some of the problems faced by emerging economies in each of these areas and reviews the generally recommended solutions to these problems. First, by reviewing the nature of internal control systems, this section stresses their importance not only for an effective PEM system but also more widely for ensuring good governance in the more devolved performance-based budget management model. Second, by reviewing international standards for government auditing and accounting, this section emphasizes that internal audit is a central component of internal financial controls aimed at protecting the government’s financial interests. Third, by emphasizing the need to create suitable management information systems to facilitate and anchor reforms at the micro level, this section emphasizes the need for providing the performance information required to ensure accountability at that level. Fourth, by examining in more depth the required costing arrangements, this section stresses the challenges faced in generating adequately precise and meaningful cost information on government programs that is required for comparison with output information in order to judge performance.
Internal Control Systems
Internal management control systems encompass a range of management tools aimed at various broad objectives: first and foremost, to ensure compliance with laws and regulations; second, to ensure the reliability of financial data and reports; and, third, to facilitate the efficiency and effectiveness of government operations. In this way, a sound internal control framework is designed to assure the public that these operations attain some basic fiduciary standards: in guarding against the misuse, and inefficient use, of financial and human resources; safeguarding assets; achieving budgetary objectives as set out in government policies and spending plans; countering fraud and error; and maintaining satisfactory accounting records to enable the organization to produce timely and reliable financial and management reports.63 As such, internal controls can be regarded as one of the foundations of good governance and the first line of defense against improprieties. They also provide the public with “reasonable assurance”64 that if improprieties do occur, they will be made transparent and appropriately addressed.65
In many countries, the heritage of traditional budget systems is a weak internal management system. Under traditional systems, detailed spending priorities tend to be set at the central level, rather than by the operational units, and therefore the latter have little incentive to develop their management systems in order to judge the efficiency with which inputs generated outputs or delivered needed services. Instead, allocations are driven by rules—or budget norms—at a detailed budget line level. Good management at the operational level is directed at preserving or increasing historical levels of funding in each budget line, rather than at making these funds more productive in terms of final outputs or outcomes. Managers have no incentive to spend their funds better. Indeed, being more cost-efficient would most likely result in future reductions in funding. In this environment, there is little incentive to develop internal management accounting, reporting, and controls. Adopting a more performance-oriented approach requires changing this set of incentives, and hence increases the pressure to strengthen internal line agency management systems.
The range of management competencies and procedures required for good PEM at the agency level is quite wide, including the following, among others:
planning budgets, setting their priorities, and reformulating them in the required detail for management control purposes;
putting in place a financial management system that enables appropriate categories of costs to be clearly identified, accounted for, and reported on;
establishing internal procurement procedures to ensure that purchasing is directed to the most economical source, meeting acceptable standards of quality and timeliness;
putting in place an accounting system to allow correct recording of transactions, and generating required management reports ensuring control of, and transparency in, the agency’s operations;
developing specific management procedures for important items of spending, such as capital assets and human resources;
strengthening the agency’s internal control mechanisms to prevent any new financial freedom, brought about by the dismantling of central controls, from resulting in less effective use of public funds. These internal controls—based on a strong internal audit function—should also be viewed as performing a watchdog role within agencies to ensure the effectiveness of the other internal management systems.
Obviously, the way that internal PEM is organized and effected varies considerably among countries and ultimately reflects the country’s predominating budget management philosophy. For emerging economies transforming their budget management and basic institutions to accommodate a more performance-oriented approach, the strengthening of internal control systems should be emphasized as one of the basic preconditions for the success of this reform process. Three areas are highlighted here, partly reflecting their importance but also reflecting their previous neglect: internal audit, management information systems, and costing systems.
Internal Audit
International Standards
In moving a budget system toward OECD standards, the role of internal audit (IA) will generally need to be better defined. Recently, there has been progress in reaching a consensus on the audit standards that governments should meet. Specifically, both the International Organization of Supreme Audit Institutions (INTOSAI)66 and the Institute of Internal Auditors (IIA) have issued standards to guide the auditing and accounting professions. While these are not compulsory, they are generally regarded as “best practices,” and so it is expected that as countries develop their own public sector auditing standards, they will try to keep them consistent with these international standards.
The standards adopt a broad view of the role of IA, placing it as a more central element of PEM rather than as a narrow function of compliance or financial regularity. The emphasis is on IA as a management tool and as an integral part of both management controls and information and communications processes. From this perspective, the purpose of IA is to review, appraise, and report to budget managers on the soundness and adequacy of internal controls (e.g., safeguarding assets, ensuring reliable records, promoting operational efficiency, monitoring adherence to policies and directives).67 The IA is thus seen as performing a watchdog role to ensure the effectiveness of internal management controls.
The standards stress four aspects of IA, which unfortunately appear hard for many countries to attain, including many emerging economies:
Independence to make objective judgments: This implies that the auditor will have no direct management responsibility for what is being audited, will be free to choose any transaction/topic for audit, and will be allowed access to all necessary information to come to an informed judgment. Unfortunately, in many countries, systemic governance problems often imply real difficulty in ensuring the auditor’s independence.
Professional proficiency: This assumes an appropriate audit methodology, technical competence, and sufficient levels of resourcing for the IA function. In many countries, skilled auditors are in short supply, professional proficiency is very low, and/or the government’s pay scales are insufficient to attract or retain suitable staff. These factors often represent an important constraint on attempts to strengthen the IA function.
Wide scope for IA: The scope of the IA described in these international standards is based on the broader view that IA is a tool of management, helping to close the loop in the agency’s PEM management cycle by ensuring the efficient and effective use of resources.68 This, in turn, assumes a mechanism for follow-up and action on audit reports. In many parts of the world, including in many emerging economies, IA continues to be defined rather narrowly—focusing on financial compliance and regularity, rather than on broader management issues. Moreover, governance problems and a lack of professional competence also constrain the internal auditor to this narrower role and hinder his or her ability to generate timely and relevant reports.
Effective management of the IA function: In many countries, management of IA is poor—poor work practices, lack of planning and personnel management, and little support from the external audit agency. Additionally, management is constrained by the institutional arrangements for IA, which often compromise the role of the auditor as an aid to internal management.
IA in the OECD Countries
Even among OECD countries, organization of the IA function varies greatly, for example, from centralized to decentralized approaches. In the centralized model, the MoF not only plays a key role in budgeting and allocating funds to line ministries, but also directly intervenes in ex ante controls, placing its own staff in the line ministries. In this environment, the IA is concentrated in a specific organization performing certain control functions, traditionally a centralized ex ante financial control organization, an inspectorate general, or a treasury external audit service. In the decentralized approach, each line ministry takes full responsibility for spending its own budget and for ensuring appropriate checks and safeguards. In this environment, the IA is focused on the ministry’s overall system of organization, controls, rules, procedures, and regulations set up to ensure the most economic, efficient, and effective use of resources. To do this, the IA control system includes a range of ex ante controls, systems, performance, and IT audits. Other models appear to mix internal and external audit functions. For example, in Germany, IA is not part of a government agency’s control system, but can be viewed as a component of external audit.69
Despite the variety of IA approaches among OECD countries, there are some common general principles. First, IA is viewed as a central component of internal financial controls aimed at protecting the government’s financial interests. The important concept is the internality of this executive function, distinguishing it from external audit. Second, although IA activities include traditional compliance and regularity operations, they can be defined quite widely to include substantive tests and systems, performance, and IT audits.70 Third, to function effectively, the internal auditor must be functionally separated from the day-today management of an organization (otherwise the accountability of designated managers will be diluted), but at the same time, the auditor must have input to top management to ensure that his or her findings and recommendations result in corrective action. Fourth, internationally recognized auditing standards should be upheld.
Developing the IA Function in Emerging Economies
In developing the IA function, the most important step for emerging economies is to determine the role of the IA in the country’s budget management system. Establishing such an overall framework involves two basic design issues:71
whether to adopt a control or a management orientation—a question of the objectives to be pursued by the IA function; and
the degree to which IA is centralized—a question of the organization of the IA function.
Once these overarching issues are resolved, there are further issues related to how this will be implemented at the agency level:
the relationship with external audit—a question of responsibilities and coordination; and
how to restructure work practices in agencies—a question of operational effectiveness.
These are addressed in turn.
Deciding the Objectives of IA
There are a variety of interpretations of the role of IA. At one extreme, there is the centralized view of IA as a support function, assisting the MoF in monitoring ministry and department compliance with MoF financial regulations, instructions, and accounting procedures. The emphasis is on compliance and control. Alternatively, under a decentralized approach, IA is viewed as assisting budget managers in the effective discharge of their responsibilities by providing feedback on their use of public resources through the submission of reports and, when justified, the provision of recommendations for corrective action. The emphasis is on efficiency and effectiveness in the use of resources and the delivery of services. Between these two approaches exist many variants, including the use of a central inspectorate to set standards and to assist decentralized IA units in the line ministries with delegated responsibilities.
The overall design of the IA function should be geared to the specific priorities of the country. For countries with governance problems, the foremost objective should be to ensure compliance with financial laws and regulations. For those emerging economies facing a high degree of fiscal stress, the need to attain macroeconomic objectives should be paramount. For those countries that can ensure compliance with the law and have reached a fair degree of macroeconomic stability, more attention can be paid to ensuring the efficiency and effectiveness of resource use, as currently emphasized in the OECD countries.
Deciding the Degree of Centralization
A fundamental design issue is the degree of centralization in the organization of the IA function. The centralized approach has often been viewed as better from a capacity-building viewpoint, because it is argued that this approach—
allows easier maintenance and better development of the proficiency of internal auditors. With a scarcity of skilled manpower, a decentralized approach often implies the diversion of IA staff to other duties that will reduce the proficiency of the staff. However, if the MoF develops a special cadre, it will be able to concentrate scarce auditing resources and so maintain proficiency, ensure the auditors’ specialization, and develop centralized standards and training programs for them.
fosters greater independence. The audit should be conducted with adequate independence. The centralized option is better in this regard, it is argued, since the IA is managed by the MoF outside the direct control of line ministry managers. However, the necessity of independence is in direct conflict with the necessity for the MoF’s close cooperation with other departments for budget management.
There are also some disadvantages to centralization, namely that it—
weakens accountability of the line ministry management. It could be argued that the prime responsibility for internal control should rest with, and be “owned” by, the line ministry management. However, the centralized option divides responsibility between the line ministry management and the MoF, obscuring the ownership (or accountability) of this control mechanism. The line ministry’s management may be only too happy to consider the responsibility for internal control as belonging to the MoF.
is of limited effectiveness because of weak transparency. In many countries, the flow of information to external officials (including internal auditors from the MoF), is typically limited and untimely, constraining the effectiveness of the IA.
fails to foster close cooperation with other departments. Close cooperation among departments is essential for efficient IAs.72 However, the centralized approach does not promote such cooperation—the internal auditor is viewed as the “spy” of the MoF, rather than as a member of the line ministry management team.
In weighing these two options—a centralized or a decentralized design for the IA—there are considerations that suggest the answer will be country specific. First, for many countries the danger is very real that in an entirely centralized approach, the MoF will assume responsibility for the rectitude of financial management in budget institutions, undermining the basic accountability of budget managers.73 Second, in some countries, the risk of political interference with routine budget management is high, so that the budget manager’s accountability is undermined from above, and therefore a centralized system is justified. Third, where the administrative capacity to perform IA functions is low, with regard to the recruitment and retention of competent staff, a centralized system controlled by the MoF is also recommended. Given the time it takes to establish a professional corps of internal auditors, including in many emerging economies, this is a most relevant consideration.
This presents a dilemma. Taking due account of the above considerations, a centralized approach for emerging economies is often recommended, at least initially, as the most prudent approach, although this runs counter to the basic decentralized institutional model that underlies many OECD IA systems that support performance budgeting. As argued previously, although the decentralized budget management model, focused on performance, is the system that emerging economies will eventually move toward, until budget management capacity has reached an adequate threshold, in the interim, a more centralized IA system may be required.
Strengthening IA at the Agency Level
Ensure the independence of IA An important objective in restructuring the IA function is to give some assurance of its independence from day-to-day management and hence of greater objectivity in its evaluations. Obviously, the degree of IA independence is not the same as for external audit, which reports to parliament. Rather, the IIA (1999) defines IA independence as follows: “Internal auditors are independent when they can carry out their work freely and objectively. Independence permits internal auditors to render the impartial and unbiased judgments essential to the proper conduct of audits. It is achieved through organizational status and objectivity.” (p. 11)
Ideally, the internal auditor should be responsible to the minister or the chief executive of the ministry or agency. In a decentralized model, the internal auditor will report directly to this top official. In the centralized approach, having the centralized audit office reporting directly to the minister of finance helps ensure the independence of the IA.74 The internal auditor is responsible to the head of the ministry/agency and is part of that agency’s staff and part of the chief executive’s management team. However, care must be taken not to infringe the cardinal rule of audit: an auditor should not audit him- or herself. Typically, this is avoided by several institutional mechanisms:
A clear and agreed definition of the internal auditors’ tasks: This is a way of clarifying the place of the IA function in the work of the budget institution, dispelling ambiguities and avoiding disputes.
Establishment of line ministry audit committees: These should be formed from the top management of the institution and technical experts in the accounting and budget fields. The aim is to act as a steering committee for the work of the IA, identifying problems as well as the corrective or preventive action. Not only does this act to strengthen the role of IA within the agency in enforcing financial discipline, but it also creates some distance between the agency’s regular operations and IA evaluations.
Create a central audit committee in MoF: Similarly, to enforce this distance from day-to-day management and to offer some external support, it is also recommended that a central audit committee in the MoF review the findings of the IA units and pursue remedial actions. Another possible mechanism is to have outside professionals conduct an independent external review of IA practices every two or three years, thereby countering any tendency for agency managers to interfere with the IA.
Clear demarcation of responsibilities in relation to external audit: In some ways, this can be addressed by a clear and well-documented definition of the duties of internal auditors. At the same time, the relationship between the two functions should be recognized as symbiotic—it is important for the IA that there be a strong external audit, and vice versa. The external audit should use the work of the IA, and the IA should be guided by the findings of the external audit.
Restructure work practices A strategic decision to be taken in many emerging economies is where best to deploy scarce audit skills. There are ways to economize on the use of these scarce IA resources:
Prioritize to extend the scope of IA: Improved work practices can often offer significant savings—say by moving away from extensive pre-audit of vouchers to a sampling approach—as can improved management of the audit function through focusing on priority areas and key weaknesses. One area that can typically benefit from IA review is the evaluation of internal controls. An important function of the IA should be to examine and evaluate the adequacy and effectiveness of internal controls in existing systems, as well as in any new systems before these are introduced. This clearly implies that the entire system of internal control has to be reviewed for each ministry, department, or agency as well as by function. This needs to be emphasized because, with strong internal controls, the system will automatically have its own checks and balances to minimize the possibility of errors, irregularities, and fraudulent manipulations.
Create special teams: In examining the IA of transitional economies, it is not unusual to discover many functions that are either not being performed or for which coverage is superficial because of inadequate staff, lack of specialized skills, etc. Often the most productive use of limited IA staff is in central teams earmarked for conducting special audits in government agencies with the assistance of IA staff already stationed there.
Better formulate work plans: Existing operational standards for IA require that the internal auditor adequately plan, control, and record his or her work. Such planning should be done not only for individual audit assignments, but also for varying time periods such as a quarter, a year, or even longer periods.
While the above approaches can make IA more effective, undoubtedly this requires proper resourcing, which may be difficult. However, the returns could be substantial insofar as a sound IA function is considered a precondition for introducing a more devolved performance budgeting model.
Financial Management Information Systems
The term financial management information system (FMIS) usually refers to comprehensive computerization of PEM processes. At its most general, an FMIS includes budget formulation, budget execution, and accounting with the help of a fully integrated system for financial management of the line ministries and other spending agencies. The full system also should be securely integrated and in communication with other relevant information systems.75
Although it is no panacea, an FMIS provides substantial benefits. First, improved recording and processing of government financial transactions allows prompt and efficient access to reliable financial data. This supports enhanced transparency and accountability of the executive branch to the legislature, the general public, and other external stakeholders. Second, an FMIS strengthens financial controls, facilitating a full and updated picture of commitments and expenditures on a continuous basis. Once a commitment is made, the system should be able to trace all the stages of the transaction, including budget release, commitment, purchase, payment request, reconciliation of bank statements, and accounting of the expenditure. Third, more comprehensive financial information on current and past performance assists budgetary control and improves economic forecasting, planning, and budgeting. As a consequence, establishment of an FMIS has become an important benchmark for a country’s budget reform agenda, often regarded as a precondition for achieving effective management of budgetary resources and as a requirement for introducing performance budgeting procedures.
As the name implies, there are, and should be, three guiding characteristics for an FMIS:
The FMIS is a management tool. When developing an FMIS, it is important that it cater to the needs of managers—not just in the central agencies but also in the line agencies. Moreover, it should support the management of change. It must be viewed as an integral part of budget system reform—and hence should be designed to meet not only present budget system requirements, but also those that are likely to arise as parallel budget reforms are implemented elsewhere.
The FMIS should provide a wide range of financial information. As a tool of management, it should provide the information required for decision making. It should be anchored in the government accounting system and be designed to perform all necessary accounting functions as well as to generate custom reports for internal and external use. However, this does not mean that it should concentrate exclusively on financial information. Managers will require other, nonfinancial information, including for example, personnel information (i.e., numbers of employees, their grades within the organizational structure, and their rates of remuneration). Performance information will be important to managers, such as the identification of program objectives or outcomes, the types of goods and services produced, as well as indicators by which to judge programs’ efficiency and effectiveness.
The FMIS is a system. Its role is to connect, accumulate, process, and then provide information to all parties in the budget process on a continuous basis. All participants therefore need access to the system and need to be able to derive the specific information they require to carry out their different functions. The converse is also true: if the FMIS does not provide the required information—that is, does not have the correct functionality—it will not be used and will cease to fulfill its central function as a system. Box 25 broadly describes the attributes usually required of an FMIS.
Existing information systems will need to be modified to accommodate a performance information system, the need for which was outlined in the previous section. Generally, more modern government FMISs require a progressive shift from a focus on annual, cash-based budgets toward a more medium-term, financial planning approach, based on a clearer policy framework in terms of performance and focused on the sustainability of the associated resource implications. Also, as discussed more fully in Section VII, institutional PEM arrangements are an integral part of this new performance orientation and will most likely need to be modified. For example, they must accommodate the trend toward decentralization, devolution, and the adoption of competitive contracting arrangements for the delivery of public sector services. Therefore, any FMIS in this new environment needs to be sufficiently flexible to address the financial management needs of a variety of organizational service delivery arrangements.
The Attributes of a Well-Designed FMIS
A well-designed financial management information system will—
be modular and capable of progressive upgrading to cater to future needs.
offer a common platform and user interface to the stakeholders in different agencies responsible for financial management to allow them to add to and access the information database (in its absence, each agency will have the incentive to develop “its own” FMIS to meet its needs).
maintain a historical database of budget and expenditure plans; transaction data at the highest level of detail; cash flows and bank account operations including checks issued, cancelled, and paid; cash balances; and floats.
have dedicated modules to handle forward estimates of revenues, expenditures (prepared by the agencies), and resulting cash flows on a monthly, rolling, short-term (one- to three-month), and longer-term (three-month to year-end) basis.
have built-in analytical tools to offer trend analysis of various elements of fiscal operations to permit a forward look at the emerging events bearing on the fiscal stance.
compile formal government accounts from the database of the agencies’ authorizations and cash allocations, primary revenue and expenditure transactions, and of treasury operations, avoiding the need to duplicate data entry for accounting purposes.
enable real-time reconciliation of parallel but related streams of transaction data—at the agency level, to reconcile checks issued with those paid by the banks; at the treasury level, to reconcile receipts from banks with the checks paid by taxpayers; and at the agency level, to reconcile cash balances reflected in the agency ledgers with the cash balances in the banks.
mechanize all possible routine tasks at the central and spending agencies—generating various forms/authorizations, checks, outputting hard copies of key registers and statements, and the like.
be flexible enough to provide user-defined management information from the database, aggregated at the desired level of detail.
The FMIS must also accommodate and facilitate a number of other reforms, which have already been described and will be discussed more fully in subsequent sections. Among these are: the transition to modified cash/accrual accounting and eventual progression toward full accrual accounting for at least some parts of government; the phased introduction of performance budgeting with clear definition of inputs, outputs, and expected outcomes; introduction of competitive markets (purchaser/provider splits); decentralization (delegation of authority and responsibility for resource management); more responsiveness to beneficiary/client requirements and more inclusive and participatory approaches; and improved approaches to strategic management and human resource management. Needless to say, the new performance budgeting orientation adds new tasks and requires more information, focusing on: improved analytical and policy formulation capacity; technical capacity to move toward more performance- or results-oriented planning, budgeting, and accounting systems; and improved control, monitoring, and performance review of the government’s budget (Box 26).
Changes in Information Needs to Accommodate Performance Budgeting
Traditional financial management needs:
institutional or expenditure heads, as detailed in the budget approved by parliament
release of funds against expenditure heads by detailed line item
recording of expenditures and reconciliation with bank statements
periodic financial statements on cash-based accounting
recording of debt and other liabilities
New performance-related tasks added to these information requirements:
programs and activities as detailed in the budget approved by parliament
allocation of funds based upon particular, desired results described in physical and financial terms
budget documents that describe the expected outcomes and performance measures to monitor achievement
responsibility for achievement identified and physical targets tracked
broader-based performance reports that document physical as well as financial measures of performance
data required to administer a set of performance-based rewards and sanctions
All such tasks require the use and development of improved IT and communication facilities to meet the needs for improved analysis and efficient information processing. The backbone of this effort will continue to be accounting and financial information systems that enable the production of complete and timely periodic financial information on revenues, expenditures, and financing, disaggregated by organizational responsibility, nature of account (expenditure, revenue, asset, liability), activity, project, and sector or economic category, as appropriate. The system should incorporate recognized accounting standards and support enhanced levels of public accountability.
Development of an FMIS
Introducing an FMIS has proved to be a challenge in every environment.76 Based on the international experience of FMIS projects, a number of implementation issues should be stressed to ensure success:
Identify the owner(s) of the system: The importance of this phase of the project may not be inherently obvious. An owner, or set of owners, needs to be identified because there will be many critical issues that arise during the course of the project that require quick and decisive resolution. It is recommended that the ownership role not be spread too thinly, for instance across a panel of users, since this may compromise the required decision-making process and its legitimacy.
Select a dedicated project team and steering committee: Many FMIS developments have failed, or produced unsatisfactory results, because insufficient investment was placed in the team that managed the project. Using a part-time team is not recommended, because such an approach invariably produces an FMIS that is late or fails to adequately satisfy user requirements. The team needs to be separated from day-to-day work responsibilities, but should not be too isolated from the ultimate users of the system, who can provide valuable feedback on its development. It is advisable that the project be driven by the users on the team rather than the IT members of the team.
Establish a high-level steering committee: A steering committee is needed to provide strategic guidance throughout the life of the project and to assist the owners in making crucial decisions that significantly affect the direction and final character of the system. Major users of the system should be represented on this steering committee, which should address such critical issues as: specifying user requirements; decisions on in-house or contractor development and the issue of locking in contractors; the use of standard software packages; and the extent of customization. These issues will be discussed more fully below.
Clearly specify the development milestones: To ensure that the system is developed on time and that any contractor payments are only made as parts of the system are successfully delivered, it is essential to clearly define milestones. It may be difficult to do this during the formulation stage of the project, but the key stages should be listed with details on the delivery components to be agreed at a later date. No contractor payments should be made until the contractor can justify the completion of a milestone, with late delivery subject to penalty clauses, with partial delivery subject to discounts, and with both specified in the contract. Successful completion of a milestone should be supported by appropriate testing and audit reports.
Produce system and user documentation: Throughout the development of the FMIS, system and user documentation must be produced on a regular basis. There is a risk in large system development projects that required documentation is left to the final stages of the project, but regular user documentation assists in the development of training modules and thereby allows the essential task of training the ultimate users to commence while the system is being developed. This not only provides a pool of knowledgeable users to test the final system before it comes on-line, but also provides valuable feedback on the system modules as they are produced. Regular system documentation is essential to avoid reliance on a single contractor, or group of contractors, for maintenance of the system.
Provide appropriate training: It is recommended that the training function be handled in-house. The risk of outsourcing this function is that the training provided may be superficial and not at the appropriate level for those users who will be using the system extensively, the so-called power users. Different types of users should be identified, and not all users should receive the same level of training. Power users should be trained first and should have access to follow-up training throughout the development process. Other users should be able to indicate the level of training they require and should receive such training as close to the roll-out of the system as possible to reduce the need for follow-up training.
Negotiate an agreement for long-term maintenance and integration of enhancements: The issue of contractor lock-in also extends to the maintenance of the system after its implementation. It is essential that the long-term maintenance agreement be subject to a full tender process, as was the original contract. Obviously, the developer or developers would have a comparative advantage in bidding on the maintenance agreement, but a full tender serves the interests of securing a competitive price and high level of service.
Conduct an independent review of system development: It is recommended that a full, independent review of the FMIS be conducted within six months of the system’s introduction. This should be in addition to the audit reports on the completion of each module of the system. The review should initially be provided to the nominated owners of the system, and then to a panel of users for review and comment.
Costing Systems
The costing methodology for performance-based budgeting has often been neglected. That is, the new emphasis on outputs or outcomes does not remove the need for input identification, which remains the most widely used means for annual budgeting and accounting. For the program structure to be implementable, there must be adequate mechanisms to fully cost programs, so that the outputs of programs can be related to their budgetary costs, and then ultimately to their benefits in order to judge program performance. Costs can be allocated on many different criteria.77 From a program budgeting viewpoint, the more pertinent criteria are the direct or indirect association with the program—i.e., based on the principle of assignment. Costs also can be assigned to different cost objects: operating units, cost centers, outputs, and projects. Whatever criteria are used, they should be clearly specified in a government-wide costing system that not only can produce accurate cost information on programs, but can also assign those costs correctly and is neither too complex nor too resource-intensive to apply (Box 27). Because these characteristics can be hard to satisfy simultaneously, important trade-offs exist.
To be accurate, the cost allocation must avoid two important types of distortion. The first is a flow distortion, when the cost bases used to describe program activities do not accurately capture their consumption of resources. The solution is to devise more accurate assignment mechanisms or to decompose program activities into more homogenous categories. Both solutions, however, involve increased complexity and higher costs. The second type of distortion is a stock distortion, when the costs assigned to program activity outputs do not reflect the actual quantity of resources consumed. Simple costing systems often assume that all resources aggregated in an operating unit are consumed proportionately by all outputs that are produced by the unit. If this is not the case, then quantity distortions occur. The solution is to assign costs by more appropriate operations and to allow each operating unit to use multiple assignment mechanisms.
Attributes of a Costing System
Relevant: Provide reasonably accurate information in a form relevant for decision making.
Clearly assigned: Offer a clear assignment of cost to each government activity, avoiding the risk of cross-subsidization of activities.
Cost-effective: The expense of managing the costing system must not be onerous in terms of resources (including staff time).
Cost-allocation systems are designed to provide cost information along various dimensions, such as organizational units, cost centers, outputs, or projects, which are described as the cost objects of the system. The choice of cost objects affects the administrative complexity and expense of operating the cost-allocation system. There is an obvious trade-off between the degree of detail by cost object and the resources required to operate the system. There are dangers at both extremes: providing a lot of cost information to decision makers is expensive, but an overly simple cost-allocation system may be of poor quality, may increase the risk of cross-subsidization, and may give misleading information to decision makers. Unfortunately, the increasing complexity of government operations has added considerably to the difficulties of cost allocation.78
For the purposes of program budgeting, the manner in which resources are consumed by programs, activities, outputs, or cost centers determines whether they are allocated as direct or indirect costs. To increase the accuracy and relevance of total program or output costs and, at the same time, to reduce arbitrary cost allocations, it is necessary to categorize as many of the resources as possible as direct costs.79 Of course, it may not be possible to directly allocate some costs to cost objects when they are common to more than one object. However, such indirect costs may still be apportioned to cost objects on the basis of consumption, i.e., on the extent to which they contribute to, or are incurred because of, the cost object.80 Much depends on how closely the costing system follows an existing organizational structure. If each operating unit or cost center is contributing to only one output (program or activity), the need for allocation in the second stage will be minimized. However, if costs accumulated in a cost center must be allocated to two or more outputs, or if some units within an organization provide services to other units, then a more complicated two-stage allocation (or step-down) process is required. Under this method, the costs accumulated in cost pools are then allocated down to other cost pools or cost objects, using a variety of allocation bases, until all costs have been allocated to final cost objects.81 The process is shown schematically in Box 28.
Basic Steps of a Program-Based Costing System
Break down the program into the activities that produce the program’s objective, service, or product, and identify the work units undertaking each activity.
Identify all resources used and their associated costs in terms of the basic work units delivering program outputs/outcomes.
Categorize and measure direct costs—those arising from the operations of a work unit or the costs incurred on behalf of the unit through the operations of other work units—and those government-wide or department-wide indirect costs, i.e., overheads.
Assign or allocate all direct and indirect costs to specific activities, using an agreed costing methodology.
A variant, more in line with the output budgeting approach, is activity-based costing (ABC). In this approach, programs are broken down into their basic activities—the steps or sequences of processes that convert inputs into outputs. As a first step, this approach also starts with apportioning an organization’s expenses to a set of cost pools, but then ABC compiles cost information on all activities performed in an organization, including activities that do not vary with output volume. ABC is oriented to allocating overheads with a primary focus on identifying the full costs of providing government services.82 Organizations that are most likely to benefit from ABC are those with high overhead costs and widely diverse operating activities or service lines—i.e., most government agencies. This approach is typically most applicable at the operational level on a service-by-service basis, enabling service managers to better understand the true costs of an activity. In this way, ABC enables costs to be better related to performance measures and allows managers to budget more confidently for target activity levels and to break down the budgetary impact of each level of activity.83
There is common agreement that the uniform measurement of the full costs of end products or activities across all government programs is logically required in order to judge performance from a policy perspective. There are many practical problems to overcome, however. First, there is the difficult problem of indirect costs, or overheads. The activities of any agency will include certain common staff functions such as personnel management, accounting, and general management, which are difficult to break down and apportion to specific activities. Indeed, some would argue that the usefulness of the program structure could be substantially diminished in certain government sectors by the presence of such indirect costs. Second, the measurement of program costs is further complicated by the existence of joint products and activities by the same basic unit. In the private sector, such cost measurement is the special domain of managerial or cost accountants geared to servicing internal management decision making. In the government sector, these skills are typically underdeveloped and often nonexistent. Last, but not least, it is often argued that capturing full costs ideally requires an accrual system of accounting, as used in the private sector, where the whole rationale is to ascertain the true cost of “outputs”—i.e., through an output-oriented accounting system. The case for moving to accrual accounting in government, at least in principle, is a strong one because it enables the full costs of output production to be ascertained each year and compared among years. However, in practice, the case may be less strong, as will be discussed more fully in Section VI.
In the first stages of implementing performance budgeting, it may be prudent to adopt systems that do not require complex product costing. As indicated, one of the important principles in designing costing systems is their cost-effectiveness (see Box 27). It should be recognized that the allocation of indirect/overhead costs could involve potentially high costs, and for this reason alone costing structures should not be overly ambitious. For example, “full” cost attribution for programs or even outputs, that is, distribution of all indirect and capital costs, may not be required in all cases. If the manager is determining his potential activity level within a given budget constraint, then variable and marginal costs are more relevant, since full costing likely contains elements that would not change if the level of output changes.84 Of course, if the decision making affects numerous budgets and programs, then full costing becomes more relevant. This brings into operation the first principle in designing costing systems, that it be relevant (see Box 27). Based on this principle, it can be argued that more sophisticated costing methodologies need to be deployed selectively rather than indiscriminately throughout the budget sector, and that for countries with more limited financial and human resources, costing systems, at least initially, should serve modest objectives.
Assessing the Performance of the Budget System
A central message of this section is that before a performance management framework is introduced, there should be some assurances that it will rest on a solid basis of public expenditure management. The question then arises how to judge whether a PEM system is robust enough to accommodate the previously discussed changes required for the introduction of performance-oriented budget management. Answering this question in turn requires an assessment of the overall performance of the PEM system.
From this perspective, it is important to note some recent work directed at measuring the performance of the PEM system as a whole.85 This effort has arisen from the fact that donors and multilateral development agencies increasingly direct aid to countries through general budget support rather than through projects. Given that expenditure is fungible, these donors and development agencies need to look at government spending in a more comprehensive way and, for fiduciary purposes, to examine the performance of PEM systems as a whole. Work is ongoing to replace the multiple diagnostic tools used by bilateral and multilateral agencies with common tools for measuring the performance of budget systems. This effort has already resulted in the development of a standardized assessment framework, the Public Financial Management (PFM) Performance Measurement Framework, which is designed to measure the performance of countries at all levels of PEM development from both a strict fiduciary risk perspective and a development risk perspective.
Critical Objectives of PEM Systems Measured through a Standardized Assessment
Budget realism: Budget is realistic and implemented as intended in a predictable manner.
Comprehensive, policy-based budget: Budget captures relevant fiscal transactions and is prepared with due regard to government policy.
Fiscal management: Aggregate fiscal position and risk are monitored and managed.
Information: Adequate fiscal, revenue, and expenditure records and information are produced, maintained, and disseminated to meet decision making, control management, and reporting purposes.
Control: Arrangements are in place for the exercise of control and stewardship in the use of public funds.
Accountability and transparency: Arrangements for external transparency and scrutiny of public finances are operating.
The PFM Performance Measurement Framework was developed by the World Bank with the assistance of its development partners, including the IMF, and is aimed at streamlining and condensing the previous diagnostic tools.86 It takes six critical objectives of a well-functioning PEM system (Box 29) and derives a small number of high-level indicators to assess performance against these objectives.87 Each indicator is measured against a four-point ordinal scale (A–D), and guidance notes have been developed on what level of performance would meet each score for each indicator. This has then been applied in pilot countries for testing.
The result is a standardized assessment tool by which countries’ budget systems are assessed against benchmarks based on international good practices in a variety of different dimensions. Even though the tool is a synthesis of assessments generated through existing instruments, the objective in the medium term is to replace the existing instruments and to harmonize recommendations for improving and strengthening PEM systems.
The ultimate objective is to make aid more effective. However, this approach also has potential as a guide to budget system reform. The power of the approach is its recognition of the PEM system as a system in which different phases of the process are interrelated. This integrated approach to performance assessment is beneficial in identifying key weaknesses that cause problems elsewhere in the system, and this assists in prioritizing remedial actions, which can be especially useful for capacity-constrained public administrations. The emphasis has been placed on proactively using this tool for capacity development, first by diagnosing weaknesses in the PEM system and then by developing a strategy and action plan jointly with the donor community and the government to correct these weaknesses. Using this approach should make it possible to strengthen PEM systems to a minimum level that accommodates performance management reforms.
Concluding Remarks
A central message of this section is that the wide scope of budget system reforms required at the micro level to accommodate a performance-oriented approach stands as a warning not to underestimate either the magnitude of the task or the time that this reform may require. These new management procedures need to be effected throughout all levels of the government sector. Unlike macro reforms, which have been focused on a few central agencies, the micro reforms potentially affect a large number of institutions and hence a much larger number of people. Accordingly, the scale of the task raises important issues related to administrative capacity—for example, the need to develop internal auditing capacity when there is a general shortage of accountants in the economy and especially in the government sector, or the lack of specialized computerization skills, which is a significant constraint on development of modern accounting and information systems.
Another important dimension of administrative capacity is the preparedness of agency managers to adopt change. Certainly, the budgeting and financial management methods used in the past are likely to require modification to serve the new performance-oriented system. Hence those who advanced to management positions in the old system may not be the best suited for the new. Given the evident skills required to implement the procedures described above, at a minimum one would have to recognize the need for an extensive and necessarily lengthy training process. One also might have need for a more basic restructuring of incentives in support of budget system reforms. Specifically, larger-scale administrative reforms may be required that allow government pay scales to offer adequate incentives to attract and retain specialized skills. Discussion of this topic is beyond the scope of this study.
This section has also emphasized the connection between the macro and micro levels of reform, which raises the issue of how to ensure a balanced approach. When reforming budget systems, there may be no alternative to adopting a top-down approach and introducing the macro framework first; at the same time, these macro reforms should not get out of gear with the capacity of the system to implement them. There is a concern that this has been the case in the past and that there have been instances in which the centralized institutions, by moving too far too fast, have seen reforms flounder and stall due to the lack of management capacity at the agency level.
These considerations argue for coupling budget system reforms with management improvement at the agency level and underline that the lack of basic management skills in government, especially at the micro level, can be a binding constraint on the speed at which budget systems can be reformed. Traditional budgeting procedures required administrators, not managers. In the old systems, controls were centralized and imposed from the top down. Reforms envisaged at the micro level, as discussed above, require a more bottom-up focus and depend on adequate management capacity at lower levels to be successful. Not only does this take time to create, but the well-documented bureaucratic resistance to change will only be overcome with time (and special skills). This implies a second order of management skills—the need for central agencies when introducing budget system reforms to invest in developing a change-management strategy and in change-management skills. This issue will be explored further in Section VIII.
See Allan and Tommasi (2001), which defines internal control as “the organization, policies and procedures used to help ensure that government programs achieve their intended results; that the resources used to deliver these programs are consistent with the stated aims and objectives of the organizations concerned; that programs are protected from waste, fraud and mismanagement; and that reliable and timely information is obtained, maintained, reported and used for decision-making.” (p. 260)
This concept is important, since there can be no absolute guarantee against wrongdoing or honest error. Rather, a control system should be designed to reduce that risk to a “reasonable” level compatible with the cost of implementing the control system.
See IMF (2001b, especially pp. 56ff).
INTOSAI (1995); Internal Auditing Standards Board (2001). See also OECD (1999).
The IIA defines internal audit as “an independent, objective, assurance and consulting activity designed to add value and improve an organization’s operations. It helps an organization accomplish its objectives by bringing a systematic, disciplined approach to evaluate and improve effectiveness of risk management, control, and governance processes.” (IIA, 1999)
See Gray, Jenkins, and Segsworth (1993). The external audit institution does this for the government as a whole.
While the IA cadre operates within agencies, they are subject to technical and professional guidance, as well as supervision, by the German supreme audit institution, the Federal Court of Audit. They report only to the Federal Court of Audit and perform a “pre-audit” role rather than a traditional IA role. See discussion in Diamond (2004).
“IA is the total sphere of activities of ex post verification by an organization (located within the organization to be audited but independent of the management functions of that organization) of whether management and control systems comply with budget specifications, objectives, rules and standards and, more generally, to the principles of sound financial management” (Larsson and Madsen, 1999, p. 5).
These are discussed more fully in Diamond (2002b).
Although at the same time, operating at arm’s length secures some independence from day-to-day operations.
The European Commission recently decentralized responsibility for spending to avoid weakening accountabilities of Commission managers. That is, “the explicit prior approval of a separate financial control service has been a major factor in relieving Commission managers of a sense of personal responsibility for the operations they authorize … while doing little or nothing to prevent serious irregularities. …” OECD (2001, p. 272)
Of course, it is necessary for the MoF to maintain its own IA and be subject to the close scrutiny of the external audit institution.
Because of the integration requirement, the FMIS is commonly characterized as an Integrated Financial Management Information System (IFMIS).
For some specific issues arising in developing countries, see Diamond and Khemani (2005).
For example, by their nature (variable or fixed); production function (output or nonoutput); decision making (investment, sunk, opportunity cost); accountability (controllable or noncontrollable).
As pointed out by Gearhart (1999), “… technology investments and the expenses associated with increased overheads, increased services, complex processes, and changing roles of government have significantly complicated the cost profile of government organizations” (p. 13).
There are many ways to do this, e.g., time recording systems, coding actual transactions, and judgment by management.
Where costs cannot be directly assigned to cost objects, agencies must identify appropriate bases for assigning them, reflecting in some way the consumption relationship between the costs and cost objects.
For example, the allocation process could begin with depreciation of buildings and fixtures. These costs would be distributed across all remaining cost pools. The amount to be distributed from the next pool (building maintenance) would now include not only its own costs but also the amount allocated to it from the previous step. This total would then be allocated to all the remaining pools, and so forth. (Thompson, 1998, p. 387)
In this way, it supports what is often termed “activity-based management” which focuses on the management of an activity to continuously improve its efficiency and effectiveness. For a case study of its application, see Abrahams and Reavely (1998).
Although the ABC approach is not used by higher levels to set overall budgets, it can lead to greater knowledge about unit costs and can provide a basis for setting benchmarks for government activities.
This also applies to joint costs, i.e., costs that contribute to multiple products but do not vary with the changes in the production level of only one of those products.
The Public Expenditure and Financial Accountability (PEFA) Program is a joint program of the World Bank; European Commission; IMF; development agencies from France, Norway, Switzerland, and the United Kingdom; and the Strategic Partnership with Africa (SPA). The four-year program was established in December 2001, and its secretariat is based in the World Bank.
While the approach concentrates on a standard set of indicators considered relevant to all country circumstances, it is recognized a country may add further indicators to meet specific needs.
