Performance Budgeting: Past Initiatives Offer Insights for GPRA Implementation (Letter Report, 03/27/97, GAO/AIMD-97-46). Pursuant to a legislative requirement, GAO reviewed the implementation of the Government Performance and Results Act of 1993 (GPRA), focusing on key design elements and approaches of GPRA as compared with those of past initiatives which also sought to link resources with results, a concept generally termed performance budgeting. GAO noted that: (1) in its overall structure, focus, and approach, GPRA incorporates critical lessons learned from previous efforts, but many of the same issues encountered in previous initiatives remain and will likely pose significant challenges if GPRA is to achieve its aim of better linking resource decisions to performance levels; (2) where past efforts failed to link executive branch performance planning and measurement with congressional resource allocation processes, GPRA requires explicit consultation between the executive and legislative branches on agency strategic plans; (3) past initiatives' experiences suggest that efforts to link resources with results must begin in the planning phase with some fundamental understanding about program goals; (4) where past initiatives devised unique performance information formats often unconnected to the structures used in congressional budget presentations, GPRA requires agencies to plan and measure performance using the "program activities" listed in their budget submissions; (5) where past initiatives were generally unprepared for the difficulties associated with measuring the outcomes of federal programs and often retreated to simple output or workload measures, GPRA states a preference for outcome measurement while recognizing the need to develop a range of measures; (6) GAO's discussions with selected legislative staff and agency officials revealed fundamental differences in perspectives and expectations that are often a necessary consequence of the system of separated powers; (7) past initiatives often foundered because no mechanism existed to reconcile or even to address these legitimate, but at times competing, views; (8) GPRA, through required consultations and formal, public documents, is intended to encourage an explicit and periodic exchange of views between the branches; (9) GPRA differs from prior initiatives in two important respects; (10) past performance budgeting initiatives were typically implemented governmentwide within a single annual budget cycle; (11) GPRA, in contrast, defines a multiyear and iterative governmentwide implementation process that incorporates pilot tests and formal evaluations of key concepts; (12) GPRA will face an operating environment unknown to its predecessors: persistent efforts to constrain spending; (13) past initiatives demonstrate that performance budgeting * --------------------------- Indexing Terms ----------------------------- REPORTNUM: AIMD-97-46 TITLE: Performance Budgeting: Past Initiatives Offer Insights for GPRA Implementation DATE: 03/27/97 SUBJECT: Planning programming budgeting Congressional/executive relations Budget decision units Zero-base budgeting Budget cuts Reporting requirements Strategic planning IDENTIFIER: Planning, Programming, and Budgeting System ****************************************************************** ** This file contains an ASCII representation of the text of a ** ** GAO report. Delineations within the text indicating chapter ** ** titles, headings, and bullets are preserved. Major ** ** divisions and subdivisions of the text, such as Chapters, ** ** Sections, and Appendixes, are identified by double and ** ** single lines. The numbers on the right end of these lines ** ** indicate the position of each of the subsections in the ** ** document outline. These numbers do NOT correspond with the ** ** page numbers of the printed product. ** ** ** ** No attempt has been made to display graphic images, although ** ** figure captions are reproduced. Tables are included, but ** ** may not resemble those in the printed version. ** ** ** ** Please see the PDF (Portable Document Format) file, when ** ** available, for a complete electronic file of the printed ** ** document's contents. ** ** ** ** A printed copy of this report may be obtained from the GAO ** ** Document Distribution Center. For further details, please ** ** send an e-mail message to: ** ** ** **** ** ** ** with the message 'info' in the body. ** ****************************************************************** Cover ================================================================ COVER Report to Congressional Committees March 1997 PERFORMANCE BUDGETING - PAST INITIATIVES OFFER INSIGHTS FOR GPRA IMPLEMENTATION GAO/AIMD-97-46 Performance Budgeting (935220) Abbreviations =============================================================== ABBREV BAPA - Budget and Accounting Procedures Act BOB - Bureau of the Budget DOD - Department of Defense GDP - gross domestic product GPRA - Government Performance and Results Act MBO - Management by Objectives NAPA - National Academy of Public Administration OMB - Office of Management and Budget PFP - Program and Financial Plan PM - Program Memoranda PPBS - Planning-Programming-Budgeting-System ZBB - Zero-Base Budgeting Letter =============================================================== LETTER B-275095 March 27, 1997 The Honorable Fred Thompson Chairman The Honorable John Glenn Ranking Minority Member Committee on Governmental Affairs United States Senate The Honorable Dan Burton Chairman The Honorable Henry A. Waxman Ranking Minority Member Committee on Government Reform and Oversight House of Representatives Since 1950, the federal government has attempted several governmentwide initiatives designed to better align spending decisions with expected performance--what is often commonly referred to as "performance budgeting."\1 Consensus exists that all of these efforts, whether launched by the legislative or executive branch, failed to shift the focus of the federal budget process from its longstanding concentration on the items of government spending to the results of its programs. In 1993, the Congress enacted the Government Performance and Results Act (GPRA) to improve the effectiveness, efficiency, and accountability of federal programs by having agencies focus their management practices on program results. Through better information on the effectiveness of federal programs and spending, GPRA seeks to help federal managers improve program performance; it also seeks to make performance information available for congressional policy-making, spending decisions, and program oversight. With regard to spending decisions, GPRA aims for a closer and clearer linkage between resources and results. In this sense GPRA can be seen as the most recent event in a now almost 50-year cycle of federal government efforts to improve public sector performance and to link resource allocations to performance expectations. GPRA mandates that GAO review the implementation of the Act's many requirements and comment on the prospects for compliance by federal agencies as governmentwide implementation begins in 1997. This report is one component of GAO's response to that statutory mandate.\2 Specifically, this report compares and contrasts the key design elements and approaches of GPRA with those of past initiatives which also sought to link resources with results, a concept generally termed performance budgeting.\3 A principal hypothesis of our work was that understanding past initiatives can aid the Congress in anticipating future implementation challenges for GPRA. In addition to an extensive literature review of past initiatives and GPRA, we convened panels of agency officials and legislative staff involved in GPRA implementation, as well as academic and other experts familiar with GPRA and some of the prior initiatives. Panelists were asked to comment on a set of challenges we identified for GPRA implementation from our review. Throughout this report, we refer to these panelists by their affiliation with a particular branch of government or as "experts" due to their background in budgeting and public administration. Although not necessarily complete or generalizable, the views expressed by the panelists cover a broad range of perspectives reflecting multiple congressional committee jurisdictions and a wide range of executive departments and agencies. Lastly, we discussed this report with senior officials from the Office of Management and Budget (OMB). They proposed several technical adjustments, which we have incorporated as appropriate. -------------------- \1 In this report, we use the term "performance budgeting" to refer generally to the process of linking expected results to budget levels, but not to any particular approach. As discussed in the body of this report, both the concept and techniques of performance budgeting have evolved considerably since 1950. \2 For additional discussion of GPRA implementation issues, see Executive Guide: Effectively Implementing the Government Performance and Results Act (GAO/GGD-96-118, June 1996); Managing for Results: Achieving GPRA's Objectives Requires Strong Congressional Role (GAO/T-GGD-96-79, Mar. 6, 1996); GPRA Performance Reports (GAO/GGD-96-66R, Feb. 14, 1996); and Managing for Results: Status of the Government Performance and Results Act (GAO/T-GGD/AIMD-95-193, June 27, 1995). \3 As discussed in the body of this report, GPRA requires performance budgeting pilots that present the anticipated levels of outputs and outcomes that would result from varying spending levels. RESULTS IN BRIEF ------------------------------------------------------------ Letter :1 In its overall structure, focus, and approach, GPRA incorporates critical lessons learned from previous efforts. Nevertheless, many of the same issues encountered in previous initiatives remain and will likely pose significant challenges if GPRA is to achieve its aim of better linking resource decisions to performance levels. -- Where past efforts failed to link executive branch performance planning and measurement with congressional resource allocation processes, GPRA requires explicit consultation between the executive and legislative branches on agency strategic plans. Past initiatives' experiences suggest that efforts to link resources with results must begin in the planning phase with some fundamental understanding about program goals. The challenge for those implementing GPRA will be to ensure that consultations are substantive and address the sometimes conflicting and competing goals of federal programs and the differing expectations of participants. -- Where past initiatives devised unique performance information formats often unconnected to the structures used in congressional budget presentations, GPRA requires agencies to plan and measure performance using the "program activities" listed in their budget submissions.\4 However, program activity structures vary throughout the federal government, and the extent to which current structures can support both GPRA performance planning needs and congressional budget decision-making is also likely to vary. -- Where past initiatives were generally unprepared for the difficulties associated with measuring the outcomes of federal programs and often retreated to simple output or workload measures, GPRA states a preference for outcome measurement while recognizing the need to develop a range of measures, including output and nonquantitative measures.\5 Focusing on outcomes shifts the definition of accountability from the traditional focus on inputs, processes, and projects to a perspective centered on the results of federal programs. However, the difficulties associated with selecting appropriate measures and establishing relationships between activities and results will continue to make it difficult in many cases to judge whether changes in funding levels will affect the outcomes of federal programs. Our discussions with selected legislative staff and agency officials revealed fundamental differences in perspective and expectations that are often a necessary consequence of our system of separated powers. For example, legislative staff concentrated on their oversight role and stressed near-term program performance, consistency over time in information presentations, and accountability. Conversely, executive agency officials stressed long-term goals, adaptability to changing needs, and flexibility in execution. Past initiatives often foundered because no mechanism existed to reconcile or even to address these legitimate but at times competing views. GPRA, through required consultations and formal, public documents, is intended to encourage an explicit and periodic exchange of views between the branches; nevertheless the inherent challenges posed by our system of checks and balances will inevitably and appropriately remain. GPRA differs from prior initiatives in two important respects. First, past performance budgeting initiatives were typically implemented governmentwide within a single annual budget cycle. GPRA, in contrast, defines a multiyear and iterative governmentwide implementation process that incorporates pilot tests and formal evaluations of key concepts. In this manner, GPRA increases the potential for integration of planning, budgeting, and performance measurement while guarding against the unreasonably high expectations that plagued earlier initiatives. Second, GPRA will face an operating environment unknown to its predecessors: persistent efforts to constrain spending. This restrictive budgetary climate can create an imperative for linking performance information to resource decisions but will likely intensify existing differences in expectations between executive and legislative branches. Past initiatives demonstrate that performance budgeting is an evolving concept that cannot be viewed in simple mechanistic terms. The process of budgeting is inherently an exercise in political choice--allocating scarce resources among competing needs and priorities--in which performance information can be one, but not the only, factor underlying decisions. GPRA is based on the premise that budget decisions should be more clearly informed by expectations about program performance. Ultimately this goal of linking resources with results implies both risks and rewards. The risk lies in expecting too much too soon--for example, that discrete outcomes can be associated with specific resource commitments, or that performance information can quickly provide solutions to today's budgetary pressures. But rewards exist as well. GPRA holds the potential to more explicitly infuse performance information into budgetary deliberations, thereby changing the terms of the debate from simple inputs to expected and actual results. -------------------- \4 The term "program activity" refers to the listings of projects and activities in the Appendix portion of the Budget of the United States Government. Program activity structures are intended to provide a meaningful representation of the operations financed by a specific budget account. \5 GPRA defines outcome measures as an assessment of the results of a program activity compared to its intended purpose. Output measures, conversely, refer to the tabulation, calculation, or recording of activity or effort, such as checks processed or students enrolled. FEDERAL INITIATIVES HAVE TAKEN VARYING APPROACHES TO PERFORMANCE BUDGETING ------------------------------------------------------------ Letter :2 At the federal level, interest in performance budgeting has led to numerous initiatives since World War II, including four that were governmentwide in scope: (1) reforms flowing from the first Hoover Commission in its efforts to downsize the post-World War II government, (2) Planning-Programming-Budgeting-System (PPBS) begun in 1965 by President Johnson, (3) Management by Objectives (MBO) initiated in 1973 by President Nixon, and (4) Zero-Base Budgeting (ZBB) initiated in 1977 by President Carter. Each of these efforts established unique procedures for linking resources with results. The following discussion briefly summarizes and relates each of these initiatives; appendixes II through V provide additional background information. First championed in 1949 by the Hoover Commission, a federal "performance budget" was intended to shift the focus away from the inputs of government to its functions, activities, costs, and accomplishments. Rather than emphasizing items of expenditure--for example, salaries, rent, and supplies--a performance budget was to describe the expected outputs resulting from a specific function or activity--for example, weapons, training, insurance claims, construction projects, or research activities. Consistent with the Commission's recommendations, the Congress enacted the Budget and Accounting Procedures Act of 1950 (BAPA), which, among other things, required the President to present in his budget submission to the Congress the "functions and activities" of the government, ultimately institutionalized as a new budget presentation: "obligations by activities." These presentations were intended to describe the major programs, projects, or activities associated with each federal budget request--in a sense, the "performance budget" of a government which at that time was primarily involved in directly providing specific goods and services.\6 Workload and unit cost information began to appear in the President's Budget, associated with the "obligations by activities" presentations, providing a means of publicly reporting the outputs of federal spending. The Planning-Programming-Budgeting-System (PPBS), mandated governmentwide by President Johnson in 1965, assumed that different levels and types of performance could be arrayed, quantified and analyzed to make the best budgetary decisions. In essence, PPBS introduced a decision-making framework to the executive branch budget formulation process by presenting and analyzing choices among long-term policy objectives and alternative ways of achieving them. Multiyear planning was to be based on an agency's "program structure," which was to provide a coherent statement of a national need, an agency's directive to fill that need, and the activities planned to meet it. Performance was generally defined as agency outputs, with an agency's program structure linking outputs to long-term objectives. Systems analysis and other sophisticated analytical tools were an intrinsic part of PPBS, with measurement seen as an essential means to better understand federal outputs, benefits, and costs. Management by Objectives, which was primarily a federal management improvement initiative, ultimately sought to link agencies' stated objectives to their budget requests. Initiated by President Nixon in 1973, MBO put in place a process to hold agency managers responsible for achieving agreed-upon outputs and outcomes. Agency heads would be accountable for achieving presidential objectives of national importance; managers within an agency would be held accountable for objectives set jointly by supervisors and subordinates. Performance was primarily defined as agency outputs and processes, but efforts were also made to define performance as the results of federal spending--what would today be called "outcomes." Zero-Base Budgeting (ZBB) was an executive branch budget formulation process introduced into the federal government in 1977 by President Carter. Its main focus was on optimizing accomplishments available at alternative budgetary levels. Under ZBB agencies were expected to set priorities based on the program results that could be achieved at alternative spending levels, one of which was to be below current funding. In developing budget proposals, these alternatives were to be ranked against each other sequentially from the lowest level organizations up through the department and without reference to a past budgetary base. In concept, ZBB sought a clear and precise link between budgetary resources and program results. -------------------- \6 The Hoover Commission's recommendation for a performance budget was first made a statutory requirement applicable to the Department of Defense through the National Security Act Amendments of 1949. President Truman's fiscal year 1951 budget, released in January of 1950, was subsequently hailed as the "first performance budget," as it applied the Commission's concepts governmentwide. However, BAPA, which was enacted in the fall of 1950 and effectively institutionalized the President's 1951 budget presentation, did not contain the phrase "performance budget." During final debate, some members argued that the term was redundant with the phrase "functions and activities" and could be restrictive of future budget presentations. THE LEGACY OF PAST INITIATIVES: THE EVOLUTION OF PERFORMANCE BUDGETING ---------------------------------------------------------- Letter :2.1 Past initiatives, although generally perceived as having fallen far short of stated goals, contributed to the evolution of performance-based measurement and budgeting in the federal government. Many concepts first introduced by these initiatives became absorbed in the federal government and persisted long after their origins in PPBS, MBO, or ZBB had been forgotten. -- Hoover Commission reforms ultimately led to permanent changes in the President's budget presentations and a greater inclusion of performance information in the narrative summaries associated with each budget account. The "obligations by activities" presentations established in response to the Commission's performance budgeting recommendations continue today, although they are now referred to as "obligations by program activity" or, more informally, "program activities." -- PPBS and MBO fostered exploration of difficult performance measurement issues, ultimately demonstrating the inherent limitations of analysis in a political environment and the often complex and uncertain relationship between federal activities, outputs, and outcomes. -- ZBB illustrated the usefulness of defining and presenting alternative funding levels and expanded participation of program managers in the budget process. When viewed collectively the past initiatives suggest two common themes. First, any effort to link plans and budgets--that is, to link the responsibility of the executive to define strategies and approaches with the legislative "power of the purse"--must explicitly involve both branches of our government. PPBS and ZBB faltered in large part because they intentionally attempted to develop performance plans and measures in isolation from congressional oversight and resource allocation processes. Since goals, objectives, and activities were not jointly discussed and agreed upon, there was no consensus on what performance should be, how to measure it, or how to integrate performance information with resource decisions. Second, the concept of performance budgeting has and will likely continue to evolve. Past initiatives illustrate a progression from the straight-forward, efficiency notion implicit in the Hoover Commission recommendations, through the increasingly complex and mechanistic processes of PPBS and ZBB. Budgeting is the process of making choices, and all of these initiatives sought to improve the rationality of budget choices by focusing on the results of activities--however those results might be defined. This history suggests that no single definition of performance budgeting encompasses the range of past and present needs and interests of federal decisionmakers. One commentator has summarized this reality as follows. "To a student of politics and of legislative bodies, it [performance budgeting] means . . . a presentation and review of budget requests in such a manner as to emphasize issues and make possible more effective choices. To a top administrator, it . . . also [means] greater flexibility and discretion in his operations, plus better control and accountability with regard to his subordinates. Down the line of an agency, it may mean a single source for funds, an enlargement of authority, flexibility, and responsibility in the use of funds. . . . To the accountant, it means accrual accounting, cost accounting, segregation of capital from operating accounts, working capital funds, and many other techniques."\7 In other words, the multiplicity of definitions reflects the differences in the roles various participants play in the budget process. And, given the complexity and enormity of the federal budget process, performance budgeting at the federal level will need to encompass a variety of perspectives in its efforts to link resources with results.\8 -------------------- \7 Frederick C. Mosher, Program Budgeting: Theory and Practice with Particular Reference to the U.S. Department of the Army (New York: American Book-Stratford Press, 1954), pp. 80-81. \8 This observation is not limited to the federal government. See Performance Budgeting: State Experiences and Implications for the Federal Government (GAO/AFMD-93-41, February 17, 1993); Using Performance Measures in the Federal Budget Process, prepared by the Congressional Budget Office, July 1993; Joint Staff Report: Performance Budgeting, prepared by the Joint Budget Committee and Office of State Planning and Budgeting, State of Colorado, September 20, 1995; and Budgeting for Results: Perspectives on Public Expenditure Management, prepared by the Organisation for Economic Co-operation and Development, 1995. THE FUTURE OF PERFORMANCE BUDGETING: THE GOVERNMENT PERFORMANCE AND RESULTS ACT ---------------------------------------------------------- Letter :2.2 As the current federal initiative seeking to link resources to results, GPRA seeks to involve all participants, directly ties plans and measures to budget presentations, and centers attention on outcome performance measurement. GPRA requires all federal agencies to set strategic goals in consultation with the Congress and key stakeholders; develop plans for program activities; measure performance; and annually report to the President and the Congress on the degree to which goals were met. Appendix VI contains additional information on GPRA's purposes and requirements. GPRA can be seen as melding the best features of its predecessors. Its required connection to budget presentations harkens back to BAPA; its interest in performance measurement and cross-agency comparisons reflects PPBS; and its concern with outcomes and outputs emulates MBO. In performance budgeting terms, GPRA avoids the mechanistic approaches of previous efforts, notably PPBS and ZBB. The Senate committee report on GPRA\9 emphasized that although "this Act contains no provision authorizing or implementing a performance budget," it was imperative that the "Congress develop a clear understanding of what it is getting in the way of results from each dollar spent." Recognizing that "it is unclear how best to present [performance] information and what the results will be," GPRA requires pilot projects to develop alternative forms of performance budgets. As one observer has noted, "there is no magic bullet that will replace budget judgement and budget policies with science."\10 Past initiatives demonstrate that any link between performance information and resource allocation decisions is unlikely to be straightforward. The implicit presumptions of PPBS and ZBB--that systematic analysis of options could substitute for political judgment--ultimately proved unsustainable. GPRA recognizes that decisionmakers, rather than budget systems, must provide judgments needed within a public sector context. That is, in a political process, performance information can be one, but not the only factor in budgetary choice; performance information can change the terms of debate, but not necessarily the ultimate decision. Finally, GPRA should be seen as part of a series of critical managerial and financial reform efforts currently underway in the federal government that share common goals of better management and accountability for results. For example, the Chief Financial Officers Act and efforts by the Federal Accounting Standards Advisory Board seek to increase public confidence in government through improved financial reporting. These efforts will, among other things, help achieve improved cost accounting and reliability of data, essential steps in accurately matching resources to program performance. -------------------- \9 S. Rep. No. 103-58 (1993). \10 John Mikesell, Fiscal Administration, 4th ed. (Belmont, CA: Wadsworth, 1995), p. 190. KEY DESIGN ELEMENTS OF GPRA INCORPORATE LESSONS FROM THE PAST, BUT IMPLEMENTATION CHALLENGES REMAIN ------------------------------------------------------------ Letter :3 In its structure, focus, and approach, GPRA incorporates important lessons from past federal performance budgeting initiatives. For example, -- by requiring consultation between the executive and legislative branches on overall agency goals and missions, GPRA addresses past failures to link planning and goal setting processes with the congressional budget process; -- by requiring use of program activities in agency budget requests as the basis for performance planning and measurement, GPRA enhances prospects for effective links with the budget; and -- by emphasizing a range of performance measures that strive toward but do not initially demand outcomes, GPRA provides a realistic framework for the expectations and capabilities of performance measurement in the federal environment. Nevertheless, many of the challenges which confronted earlier efforts remain unresolved and will likely affect early GPRA implementation efforts. Agency officials, legislative staff, and other experts we met with recognized these continuing concerns and emphasized the need to adjust expectations as new approaches and capabilities are developed and tried. GPRA EMPHASIZES FORMAL STRATEGIC PLANNING INCORPORATING STAKEHOLDER CONSULTATION ---------------------------------------------------------- Letter :3.1 Where most past initiatives did not link performance information developed within the executive branch with congressional processes, GPRA provides that agencies must consult with cognizant congressional committees, and other stakeholders, as strategic planning efforts progress. This requirement is GPRA's most fundamental change and perhaps its most significant challenge, because any effort to link resources and results must encompass some fundamental understanding of the goals of a particular program. However, discussions between agencies and the Congress on strategic planning are likely to underscore the competing and conflicting goals of many federal programs as well as the sometimes different expectations among the various stakeholders in the legislative and executive branches. In addition, the federal government's increasing reliance on third parties--principally, state and local governments and contractors--further complicates efforts to reach consensus on program goals. And, significantly, executive branch officials and legislative staff we spoke with seemed to approach strategic planning consultations with very different expectations. For the most part, past initiatives defined planning processes as internal agency activities, with limited external visibility and virtually no external involvement. Not surprisingly, where initiatives were in practice confined within the executive branch, legislative oversight and budget decision-making were ultimately unaffected; the Congress resorted to traditional information sources, which agencies quickly reemphasized. For example, in PPBS, executive agencies did not provide the Congress with information on alternative program choices or even on the basis for decisions to pursue a particular program course, often despite requests from the Congress.\11 During MBO, presidential objectives approved by the administration were made public, but congressional involvement in determining these objectives was not sought. Although some ZBB decision packages were made available to the Congress, differences in format and the voluminous amount of paperwork limited congressional interest and discouraged use.\12 GPRA's premise of joint legislative and executive involvement in strategic planning is new. GPRA requires a formal document based in part on consultations with the Congress and other interested stakeholders. The Senate report on GPRA indicates that strategic plans are to be the basic foundation for a recurring process of goal-setting and performance measurement tied to the agency's program activities and that goals must be clear and precise in order to maintain a consistent direction. The Senate report on GPRA recognizes that shifts in political philosophy may alter priorities and means of achieving objectives but assumes that legislatively determined missions and goals would remain largely unchanged from year to year. GPRA strategic planning was viewed as fundamentally different from previous efforts, requiring that agency missions and goals be connected to day-to-day operations. Past governmentwide initiatives suggest that achieving GPRA's strategic planning consultation goals will be difficult, particularly given the changes in emphasis and approach established by GPRA. For example, reaching a reasonable level of consensus on clear and precise strategic goals will almost certainly encounter political hurdles. Competing and/or ambiguous goals in many federal programs are often a by-product of the process of consensus building; strategic planning which is seen as merely rekindling old conflicts may not be well-received within the political process. Furthermore, the federal government's continued and, in recent years expanded, reliance on state and local governments and other third parties to deliver federally funded services--some of the stakeholders that would likely be part of the consultation process--adds extra complications to the prospect of reaching consensus. For example, applying PPBS to programs requiring participation by federal, state, and local governments was seen as a major implementation problem. Discussions with legislative and executive branch staff confirmed the above concerns and also suggested that these officials may be approaching strategic planning from fundamentally differing perspectives. Agency officials viewed strategic plans as a potentially useful means to a dialogue with congressional committees but were skeptical that consensus on strategic goals could be reached, especially given the often conflicting views among an agency's multiple congressional stakeholders. Some noted that achieving consensus may result in rhetorical rather than substantive plans and doubted the capacity of such plans to inform congressional decision-making. Legislative staff characterized some of the early strategic plans as lacking in substance and requisite detail. One staff member expressed concern that agency strategic plans would be used to present political agendas and justifications for the status quo, rather than real assessments of need and value provided by specific program activities. Another legislative staff suggested that the broader focus of the strategic planning process could hinder traditional congressional oversight and control processes. Some experts we contacted suggested that the expectations for strategic planning must be lowered, particularly for the initial attempts at congressional consultation. Specifically, they urged agencies and the Congress to seek a "reasonable degree" of consensus on draft strategic plans and allow several iterations to refine plans and demarcate lines of conflict and agreement. In the opinion of these experts, establishing an ongoing dialogue between the branches will be more important than seeking immediate consensus. -------------------- \11 In a congressional hearing discussing the availability of PPBS information for the Congress, the Bureau of the Budget (BOB), the predecessor to OMB, took the position that information used to develop the budget was internal to the executive branch. However, BOB stated that budget requests and legislative justifications should incorporate evaluation data and cost estimates that arose from PPBS analysis. \12 Some ZBB pilots were congressionally directed prior to governmentwide implementation, but results were similar to the Carter initiative: congressional use was hindered by large volumes of information in unfamiliar formats. GPRA PERFORMANCE PLANNING AND MEASUREMENT REQUIRES DIRECT LINKAGE WITH THE BUDGET ---------------------------------------------------------- Letter :3.2 Where past initiatives tended to devise unique structures to capture performance information that ultimately proved difficult to link to congressional budget presentations, GPRA requires agencies to plan and measure performance using the same structures which form the basis for the agency's budget request: program activities. This critical design element of GPRA aims at assuring a simple, straightforward link among plans, budgets, and performance information and the related congressional oversight and resource allocation processes. However, the suitability of agencies' current program activity structures for GPRA purposes is likely to vary widely and require modification or the use of crosswalks. Discussions with agency officials and legislative staff suggest that both are well aware of potential challenges in implementing this GPRA requirement but, again, tend to view the need for and benefits of adjustments to program activity structures from very different perspectives. As discussed previously, the "program structures" (PPBS) and "decision units" (ZBB) of earlier performance budgeting initiatives were not intended, at least initially, to explicitly connect to either an agency's organizational structure or congressional budget justifications.\13 Attempts to crosswalk PPBS program structures to budget presentations proved unduly cumbersome, and subsequent efforts to align these structures with the federal budget were ultimately unsuccessful. Similarly, under ZBB, crosswalks were needed between decision units and budget structures, and decision unit consolidations obscured the analysis of alternative spending levels and performance that was ZBB's presumed hallmark. Congressional interest in both initiatives quickly waned as plans and performance information could not be directly linked to familiar oversight and budget structures. In the end, structural incompatibilities meant that resources were not linked to the new results information. GPRA's required use of program activities appearing in the President's Budget as the basis for performance planning and measurement is intended to establish the direct budgetary link absent in earlier initiatives. But this goal is dependent on the capacity of the current program activity structures to meet GPRA's needs. That is, where the success of earlier initiatives hinged on the extent that unique planning structures could link to congressional processes, current program activities structures useful to congressional resource allocation processes must prove their suitability for planning and measurement purposes. Subject to clearance by OMB\14 and generally resulting from negotiations between agencies and appropriations subcommittees, program activity structures differ from agency to agency and, within an agency, from budget account to budget account. Program activities, like budget accounts, may represent programmatic, process, organizational, or other orientations\15 and, similarly, their suitability for GPRA planning and measurement purposes will also vary. For example, during ZBB, some agencies used their program activities as the basis for consolidated decision units; one agency that did so found that the process orientation of its program activities (e.g., regulatory development) rendered ZBB rankings meaningless. Under GPRA, when program activity structures present challenges to performance planning and measurement objectives, agencies have options. GPRA allows agencies to consolidate, aggregate, or disaggregate program activity structures for performance planning purposes, where needed. This approach would of course require subsequent crosswalks, but presumably not as burdensome as those of prior initiatives. Agencies may also attempt to renegotiate program activities with their appropriations subcommittees and OMB. Program activities, however, serve specific functions and may prove resistant to frequent or substantial change. For example, program activities (1) provide a relatively consistent structure for OMB and the Congress, allowing comparison of current spending to estimates of future needs, and (2) often form the basic unit of congressional oversight for determining reprogramming thresholds.\16 Agency officials we spoke with confirmed the varying suitability of their program activity structures for GPRA purposes. One agency successfully worked through the performance planning process using its existing program activities; another agency found it necessary to devise a separate planning structure and then link back to program activities using a crosswalk. This second agency had a program activity structure that reflected its organizational units--a structure useful for traditional accountability purposes but less useful for outcome planning. Still other agencies separated performance planning from program activity structures, believing it necessary to first establish appropriate program goals, objectives, and measures before considering the link to the budget. These agencies planned to rely on GPRA's provision to aggregate, disaggregate, or consolidate program activities. Our discussions with agency officials and legislative staff highlighted a potential tension on the use of program activities as a basis for agencies' performance planning and measurement. Some agency officials saw program activity structures as secondary to strategic planning; thus, where current program activity structures proved unsuitable for planning purposes, these officials viewed change in the program activity structure as inevitable and appropriate. Legislative staff generally viewed these structures as fundamental to congressional oversight of agency activities; thus, change was viewed with apprehension and concern. Legislative staff were generally comfortable with existing structures and questioned whether changes would frustrate congressional oversight. Agency officials generally saw a need to be flexible in using program activities as a planning mechanism, and considered it likely and desirable to change program activity structures to better align with GPRA goals and objectives; however, they noted that changes could prove difficult and time-consuming to negotiate with the Congress. In addition, agency officials were not convinced that changes to program activities would necessarily achieve GPRA's purposes, particularly when competing or unclear goals existed or when agency goals and objectives were likely to change over time. The experts we met with generally agreed that the program activity requirement of GPRA would likely constitute a significant implementation challenge. One expert expressed the tension between legislative and executive branch officials as a difference in the purpose and role of the program activity structure. Congressional interests emphasize oversight and control, thus necessitating detail and continuity in the structure. Agencies, however, use program activities for managerial purposes, thus seeking less detail in favor of more flexibility. Another expert noted that GPRA does not define how to aggregate, disaggregate, or consolidate program activity structures. -------------------- \13 In fact, under PPBS, budget presentations were ultimately expected to conform to the new program structures. PPBS guidance noted that over time it "may be necessary and desirable for the program by activity portion . . . to be brought into line with the program structure developed." \14 OMB Circular A-11, "Preparation and Submission of Budget Estimates," requires that an agency's program activities must be useful for the analysis and evaluation of budget estimates, be related to the administrative operations of the agency, and have accounting support. \15 For a discussion of this point, see Budget Account Structure: A Descriptive Overview (GAO/AIMD-95-179, September 18, 1995). \16 Reprogramming is the shifting of funds within an appropriation to purposes different from those contemplated at the time the appropriation was requested and provided. Several appropriations subcommittees use program activity structures to establish reprogramming thresholds. If an agency needs to shift funds from one activity to another above the defined threshold, it is expected to notify the subcommittee. GPRA PERFORMANCE REPORTING EMPHASIZES OUTCOMES ---------------------------------------------------------- Letter :3.3 GPRA performance reporting allows agencies to use a range of performance measures but contains a specific emphasis on outcomes--the actual results of a program activity compared to its intended purpose. Past initiatives struggled with a variety of approaches, ultimately finding it more practical to measure agency processes and outputs than outcomes. Agency officials implementing GPRA affirmed the value of outcome measurement and were also exploring alternative approaches due to the inherent challenge of outcome measurement in a federal environment marked by entitlement programs and other programs performed by nonfederal actors. Legislative staff questioned the validity and usefulness of outcome data in decision-making and perceived a potential for loss of needed detail. Taken together, the views of executive and legislative officials suggest GPRA will be challenged to identify performance measures that are both outcome-based and useful for traditional accountability purposes.\17 Past initiatives struggled with performance reporting. Taken together, their experiences reflect a slow refinement of the objectives and awareness of the difficulties of performance measurement within the federal government. Efforts spurred by the Hoover Commission centered on identifying the activities to be performed and their costs, most commonly described as unit cost and workload analysis. Under PPBS, the purposes and uses of analysis were expanded to include a decision-making component; hence, not only were outputs and their costs analyzed, but PPBS expected that such analysis could define the most urgent national goals and determine the most effective and efficient means of reaching these goals. But agencies which attempted to gather this performance-oriented data found the process to be far more difficult than expected, and officials reported that several years would be required to develop the information and collection systems envisioned by PPBS. Agency officials reporting on their experiences under PPBS also noted situations where it was difficult to relate programs to a stated outcome or to separate out other influences that might affect ultimate outcomes. For example, the Upward Bound program was designed to increase skills and motivation for low-income high school children. However, PPBS officials had no way to isolate the program's effect from other environmental influences which might also have contributed to the success or failure of different program participants. While subsequent initiatives reduced their expectations regarding the use of performance measurement and analysis, they continued to encounter similar difficulties. Under MBO, in contrast to PPBS experiences, presidential objectives and related agency programs were to be determined, and then followed by discussion of appropriate measures. This approach recognized that some presidential objectives, such as achieving cooperation with other countries or successfully negotiating international economic treaties, did not lend themselves to scientific analysis. ZBB decision packages were expected to include the outputs or accomplishments expected from a program. However, these performance measures were very quickly overwhelmed by the need to present decision packages within budget deadlines. ZBB allowed the use of proxy measures of performance and even indicated that decision packages were expected to be ranked with or without the benefit of performance information. In fact, a subsequent analysis of ZBB efforts found that fewer than half of the decision packages examined had quantifiable accomplishments, workload, or unit cost information. While acknowledging the inherent difficulties of performance measurement, GPRA requires that agencies establish performance indicators to be used in measuring the relevant outputs and outcomes of each program activity. The Senate report on GPRA indicates that sponsors understood the importance of measurement to any performance-based initiative and that outcomes are the most desirable performance indicator. However, GPRA also accepts that measurable outcomes may not always be possible--that causal links between federal efforts and desired outcomes may never be established--and encourages that a range of related indicators, such as quantity, quality, timeliness, and cost be developed and used to approximate outcomes. Executive officials we spoke with were strongly supportive of performance measurement, including outcome measurement, but raised concerns about the use of this information, particularly as a vehicle for congressional oversight. These officials saw value in defining outcomes for planning purposes and were also testing various approaches, including identifying intermediate performance measurements,\18 using multiple measures to reflect different stakeholders' interest, and applying nonquantitative measures, due to the difficulties inherent in outcome measurement. But executive officials were concerned that in today's federal environment, full or ultimate program outcome was typically not under the control of a single federal agency, complicating responsibility determinations and resulting resource allocation decisions. In some cases, outcomes can only be achieved over many years; in other cases, federal activities are but one, and often a small, component of total public and private sector interventions in a given program area; and in still other cases, intended results cut across the activities of several agencies. In each of these cases, any individual agency outcome measurement is often incomplete and therefore of limited value to budgetary decisions. Moreover, the increasing role of state and local governments as well as of other third parties as the delivery agents for federally financed activities means that in achieving many federal outcomes, the efforts of nonfederal actors--and their objectives and concerns--were critical factors in performance measurement. Lastly, the predominance of entitlement spending, in which federal actions are typically a function of statutory eligibility determinations, further clouds the ability to hold agencies accountable for outcomes by shifting attention from broad goals (e.g., assuring a certain standard of living) to specific processes (e.g., ensuring correct and timely payments to individuals). Legislative staff also expressed concerns regarding the use of outcome measurement for oversight purposes, but principally in terms of the completeness, validity, and reliability of the data for decision-making. In particular, legislative staff were reluctant to have outcome information substitute for the more detailed information they customarily receive, indicating that such a substitution could lead to less, rather than better informed legislative decision-making. One official described an agency's strategic plan as outcome-based, but with little discussion of the activities planned to meet the established agency goals; others expressed frustration that an agency's goals defined in its GPRA plans can be very different from those negotiated in congressional oversight and resource allocation processes. Finally, legislative staff also expressed strong interest in congressional involvement in measurement questions. Although concerned about the added burden for congressional staff, legislative staff felt that the Congress should take an active interest in what is measured and how it is measured. GPRA performance information, augmented by audited financial data, was seen as most useful for the Congress, but the staff emphasized that the quality of this information would need to be greatly improved. Experts we spoke with encouraged agencies to identify a range of measures and indicated that this approach is particularly useful for programs with multiple or conflicting goals. Nonquantitative measures were also cited as important for activities such as research and development, and one expert urged the use of multiyear measures where goals could not be realistically achieved in a single year.\19 One expert cautioned against agencies identifying outcomes too quickly, indicating that such a practice risked rhetoric over measurement and would not be useful in holding agencies to a level of performance. Similarly social indicators--poverty rates or mortality statistics--should only be used where it is evident that federal actions have the capacity to affect the indicator. -------------------- \17 For a discussion of related issues within foreign and state governments, see Managing for Results: Experiences Abroad Suggest Insights for Federal Management Reforms (GAO/GGD-95-120, May 2, 1995) and Managing for Results: State Experiences Provide Insights for Federal Management Reforms (GAO/GGD-95-22, Dec. 21, 1994). \18 Intermediate outcomes are outcomes that occur between outputs (delivery of products or services) and the achievement of the ultimate purposes of a program (reducing pollution and improving health, for example). Intermediate outcomes might include client satisfaction, actions taken by other levels of government, or actions by those in the private sector. \19 See, for example, Managing for Results: Key Steps and Challenges in Implementing GPRA in Science Agencies (GAO/T-GGD/RCED-96-214, July 10, 1996). GPRA DEFINES A PHASED, ITERATIVE IMPLEMENTATION PROCESS THAT ENHANCES PROSPECTS OF INTEGRATION WITH THE BUDGET ------------------------------------------------------------ Letter :4 Unlike past initiatives, GPRA's implementation design enhances prospects for a fuller integration of performance information with budgeting. As noted above past initiatives were generally attempted within a single annual budget cycle and tended to lack processes for addressing implementation problems. In contrast GPRA posits a multiyear, iterative implementation process, built around periodic publicly available products, that will allow agencies and the Congress the opportunity to refine performance planning, measurement, and reporting, and to modify, as needed, current budget processes and presentations. Past initiatives tended to take an "instant implementation" approach that limited their capacity to address challenges as they arose. At their outset, these initiatives generally gave agencies little time for complex implementation tasks. For example, PPBS gave agencies 10 weeks to develop requisite program structures--a task which the Department of Defense, the originator of PPBS, took 10 years to accomplish. ZBB similarly imposed numerous changes to executive branch budget formulation processes within a single budget cycle, with guidance agencies believed was inadequate on key requirements. Given this abbreviated implementation process and the fact that cost estimates and decision packages developed under these initiatives were not routinely made available to the Congress, it is not surprising that congressional budget decision-making was unaffected. In contrast, GPRA defines a 7-year implementation time frame, from initial pilots to first governmentwide performance reports, and incorporates feedback mechanisms such as required evaluations of key concepts before governmentwide implementation.\20 Once key requirements have been phased in, successive iterations of agencies' strategic plans, performance plans, and performance reports will allow opportunities for needed refinements. In addition, GPRA's products, which will be part of the public record, are to be made available routinely to the Congress in time to allow for the information to be integrated with congressional budget and oversight processes. For example, the Senate report on GPRA states that its plans and reports can give the Congress the ability to identify where planned resources do not appear adequate to achieve intended results and then to make realignments as appropriate. GPRA's implementation approach also provides for 2-year pilot projects of alternative performance budget approaches in at least five agencies. During the second year of these pilots (fiscal year 1999), performance-based budget presentations for each of the designated agencies are to be included in the President's Budget submission to the Congress. The pilots' aim is to test possible approaches and develop capabilities toward realizing the potential of performance budgeting, and to present varying levels of performance, including outcome-related performance, resulting from different budgeted amounts. GPRA also requires OMB to evaluate the results of the pilots by March 31, 2001, and assess whether legislation requiring performance budgets should be proposed. The Senate report on GPRA said that the performance budgeting pilots are to begin "only after agencies had sufficient experience in preparing strategic and performance plans, and several years of collecting performance data." In this context, and recognizing the importance of concentrating on governmentwide GPRA implementation in 1998, OMB indicates that these pilots will be delayed for at least a year. As envisioned under GPRA, performance budgeting will require the ability to calculate the effects on performance of marginal changes in cost and funding. According to OMB, very few agencies currently have this capability, and the delay will give time for its development. -------------------- \20 OMB called for changes in the quality and quantity of performance information in agency budget submissions earlier than required by GPRA. For further discussion, see Office of Management and Budget: Changes Resulting from the OMB 2000 Reorganization (GAO/GGD/AIMD-96-50, Dec. 29, 1995). GPRA WILL FACE UNIQUE CONFLICTS ARISING FROM BUDGETARY PRESSURES ------------------------------------------------------------ Letter :5 In one critical dimension, GPRA will face an environment unknown to previous performance budgeting initiatives: sustained, real declines in discretionary spending.\21 Past efforts faced budget-related tensions, but nothing comparable to that which will likely form the initial operating environment for GPRA. Both implementation challenges and opportunities will likely arise from different expectations regarding the appropriate role for GPRA within this period of declining resources. To executive officials we spoke with, performance information was seen as essential to justify and improve current program performance; to legislative staff, performance information was expected to prove valuable as a government downsizing tool. As GPRA is implemented governmentwide, total discretionary spending is projected to decline in real terms, continuing the pattern of the last 6 years. This constitutes a unique implementation environment when compared to past initiatives. Hoover Commission recommendations were implemented as the federal government shifted from a wartime bureaucracy; PPBS faced the competing spending tensions of the Vietnam war and an ambitious social agenda; ZBB was instituted as federal deficits reached then post-war highs and the economy experienced unusually high inflation. While all of these concerns affected consideration and passage of the budget, federal spending during each of these initiatives generally continued to experience real increases, particularly for discretionary spending. Budgetary constraints will likely raise implementation issues for both agencies and the Congress. Experts we spoke with noted that in implementing GPRA all participants will need to build capacity to develop and use performance information. For agencies, this will mean acquiring necessary resources and skilled personnel, and developing the management leadership needed to sustain a performance-based organization. Similarly, the Congress will need to expand its capacity to actively participate in strategic planning, effectively communicate results-based expectations, and manage its use of performance information provided by agencies. Generally, executive officials did not see resource availability as a significant concern; they tended to view GPRA as the "right thing to do" and believed that needed resources would be found. However, they were concerned about the potential burden of expanded performance measurement requirements, noting that GPRA's requirements could be especially onerous if, as some expected, they were layered on top of existing information requirements. In our discussions with executive officials and legislative staff, both agreed that declining budgets provided new incentives to use performance information as a key input to decision-making, but each had differing expectations as to how this should be done. In effect, each had differing views on what constituted appropriate and effective "use." Executive officials believed that GPRA can be used to more effectively present budgetary requirements in performance-based terms, for example to the Congress. In addition, they noted that GPRA can be useful within the executive branch to identify ways to streamline operations and to make necessary budget reductions; its principal value was internal and management oriented, stemming from its ability to clarify missions and performance expectations. However, they also noted that current budgetary pressures and apprehension about use of GPRA information could increase levels of defensiveness among agency staff. Legislative officials agreed that GPRA should aid in presenting budgetary requirements in performance-based terms. They saw GPRA as encouraging both agencies and the Congress to revisit current functions and activities in relation to their articulated mission and to identify poorly performing or overlapping program activities. Legislative staff added that, given the difficult budget choices facing the nation, terminating programs based on GPRA performance information was a far more defensible practice than instituting across-the-board reductions in all spending--all too often the only other alternative. These staff expressed concern that as agencies and the Congress search for ways to reduce federal spending, conflicts over agency missions and program goals are more likely to surface, leading to agency "repackaging" of information to obscure poor performance. And they questioned whether agencies could provide valid and accurate performance data. Other experts saw potential for use of GPRA in the budget process but expressed caution. These experts noted that the GPRA process can allow agencies and the Congress to renegotiate program goals, thus forcing rigor into federal budgeting and management processes. One expert emphasized that agencies would need to see GPRA information used in decision-making if they were to continue to invest in the initiative. However, if GPRA's exclusive result is to terminate programs, the initiative could suffer a loss of support within the executive branch. Experts also stated that GPRA information would also need to be used outside of the budget process. -------------------- \21 Real discretionary spending refers to outlays that are controlled through annual appropriations and adjusted for inflation. The President's fiscal year 1998 budget projects a decline in real discretionary spending from 1998 through 2002. OBSERVATIONS ------------------------------------------------------------ Letter :6 While GPRA has incorporated critical lessons from the past, the Congress and the executive branch will face certain challenges in their efforts to connect resources to results in the federal government. These challenges cannot be addressed by either the executive or legislative branch alone; all those involved in the resource allocation process must play a part. In particular, efforts to implement GPRA must address the following issues: -- The Congress and the executive branch will need to explore what can be expected of a performance budgeting system. GPRA can inform the budget process and change the nature of its dialogue by more routinely introducing performance information into decision-making. But, GPRA cannot be expected to eliminate conflict inherent in the political process of resource allocation, and final decisions will appropriately take into account many factors, including performance. -- The Congress and the executive branch must acknowledge that it takes time to develop goals, outcomes, and measures that are valid and acceptable to a range of stakeholders. All participants must take full advantage of the iterative planning and reporting processes defined by GPRA. Immediate expectations regarding budgetary impact and the ease of performance measurement must be tempered with long-term involvement and commitment to achieving GPRA's purposes. -- The Congress and the executive branch must recognize the difficulties associated with devising a system that integrates performance and budget information. GPRA provides for such integration through the program activity structure of the federal budget. Both the budget and GPRA processes must be better aligned, requiring adjustments and accommodations. In some cases, agencies may need to develop effective crosswalks between strategic plans and the budget; in other cases, agencies and the Congress may decide to change the program activity structure in the budget. Improved financial reporting and auditing as required by the Chief Financial Officers Act will further strengthen the cost basis and reliability of data underlying the link between performance information and the budget. Over the longer term, GPRA can become a powerful tool for the hard budgetary choices that the Congress and the administration will face in the coming years. In addition to improving attention on the performance of individual program activities, GPRA can be used to address one of the more intractable problems of the federal government--that of duplicative programs that cut across federal missions and agencies. The Congress and the administration could use GPRA as the vehicle to devise a framework that compares and integrates decisions that affect related programs. In this manner, GPRA's focus on governmentwide performance can offer an important alternative to across-the-board reductions and better inform choices among competing budgetary claims. ---------------------------------------------------------- Letter :6.1 We are sending copies of this report to the Chairmen and Ranking Minority Members of the Senate Committee on the Budget; House Committee on the Budget; Senate Committee on Appropriations; House Committee on Appropriations; Subcommittee on Government Management, Information and Technology, House Committee on Government Reform and Oversight; the Director of the Office of Management and Budget; and other interested parties. We will also make copies available to others on request. The major contributors to this letter were Michael J. Curro, Carolyn L. Yocom, and Linda F. Baker. If you have any questions, I can be reached at (202) 512-9573. Paul L. Posner Director, Budget Issues OBJECTIVES, SCOPE, AND METHODOLOGY =========================================================== Appendix I The specific objective of our work was to compare and contrast the key design elements and approaches of GPRA with those of similar past initiatives in order to identify potential challenges for GPRA implementation. To identify past federal performance budgeting initiatives, we used the following criteria: (1) the initiative occurred after World War II, (2) the initiative was implemented governmentwide, and (3) the initiative asserted (either initially or ultimately) a relationship between performance information and the federal budget process. Based upon these criteria, we identified four prior federal initiatives: federal performance budgeting initiatives derived from the first Hoover Commission; the Planning-Programming-Budgeting-System (PPBS); Management by Objectives (MBO); and Zero-Base Budgeting (ZBB). Our work did not address performance budgeting initiatives that were limited to a few programs or agencies, nor did we address initiatives that were planned but never fully implemented. For example, this approach excluded the end-results budgeting efforts in the Forest Service during the 1980s and President Ford's Presidential Management Initiatives. To collect information on GPRA and on the four prior federal initiatives, we used a qualitative research design. In making our review of each prior initiative, we conducted extensive literature searches, including pertinent legislative histories, hearings, and committee prints. For GPRA, we collected information on its legislative history as well as other relevant information including OMB guidances, selected pilot performance plans and reports, and available reviews of GPRA implementation efforts to date. We compiled information on the context, implementation approach, and results of each of the prior initiatives. To compare and contrast these analysis results with GPRA, we summarized our findings for each initiative, then compiled a set of observations relevant to GPRA design and implementation. From this work we identified a set of potential challenges for GPRA implementation as well as relevant observations based on past initiatives. To compare the results of our analysis with GPRA implementation experiences to date, we contacted selected individuals in the executive and legislative branches and other experts from outside government. We selected these individuals based on their knowledge, experience, and interest in GPRA. We asked them to review the identified challenges and observations and participate in one of three panels: (1) an executive panel of individuals with direct responsibility for implementing GPRA and representing agencies covering a range of functions and program types (e.g., regulatory, direct service provision, grant administration, research and development); (2) a legislative panel composed of staff from authorizing, budget, and appropriations committees in the House of Representatives; and (3) a panel of individuals from the National Academy of Public Administration (NAPA), academia, and former government officials with expertise in GPRA or prior performance budgeting initiatives. We asked the panelists to review our observations and indicate the extent to which the challenges we identified held true for the programs and/or budgets under their purview or within their experience. We also asked panelists to discuss what approaches had been used or might be considered to mitigate these concerns. To assure maximum candor, individuals were informed that there would be no attribution of their comments to them or their organizations. We conducted our work in Washington, D.C., between October 1996 and March 1997. We requested comments on a draft of this product from the Director of OMB. On March 3, 1997, we met with designated OMB officials and discussed and incorporated changes based upon their comments. THE FIRST HOOVER COMMISSION ========================================================== Appendix II CONTEXT After World War II, America was left with a wartime organizational bureaucracy and a huge national debt that exceeded the gross domestic product (GDP). Reorganization planning evolved as a systematic means of reducing federal spending while allaying concerns that such reductions would cause a return to the depression of the 1930's. The President and the Congress explored various reorganization efforts, the most effective and well known being the Commission on the Organization of the Executive Branch, more commonly referred to as the first Hoover Commission, established by law in 1947. The Declaration of Policy in the act creating the first Hoover Commission (61 Stat. 246, July 7, 1947) focused on promoting economy, efficiency and improved services in the executive branch of government. The Commission was charged with the structural reorganization of departments and agencies and the President's managerial authorities; it published 19 reports with over 270 recommendations in the Spring of 1949. With estimates of the number of implemented recommendations being as high as 196, the first Hoover Commission is considered to have been highly successful. One recommendation deemed successfully implemented was that for performance budgeting, which the Commission defined as follows: "Under performance budgeting, attention is centered on the function or activity--on the accomplishment of the purpose--instead of on lists of employees or authorizations of purchases . . . . this method of budgeting concentrates congressional action and executive direction on the scope and magnitude of the different Federal activities. It places both accomplishment and cost in a clear light before the Congress and the public." Performance budgets as prescribed by the Hoover Commission were to provide more comprehensive and intelligible information to the President, the Congress, and the public. And, the Commission recommended that attention should shift away from government inputs--items of expense, lists of federal employees--to government outputs--its accomplishments, activities, and their related costs. IMPLEMENTATION APPROACHES Both the executive and legislative branches of government made efforts to implement a performance budget. In the executive branch, initial work on a performance budget began in 1949 when the Bureau of the Budget (BOB) began preparation for the 1951 budget.\1 BOB issued a statement to the Congress about the unique nature of the 1951 budget presentation, and pledging its support for a performance-type budget suggested by the Hoover Commission and others: "While the 1951 budget may be described as the first performance budget, it will be far from perfect, and we hope that we can improve it immeasurably in later years." The 1951 budget submission was a distinct change from prior Presidential budgets. One of the more significant changes made was in the "obligations by activities" section of the budget. This section provided (1) listings of the programs or activities imbedded within a budget account, (2) separated operating and capital expenses, and (3) established breakouts for grants, and other fixed charges as well. Prior to the 1951 budget, less than 45 percent of all budget accounts contained obligation by activity subdivisions; after the 1951 budget, all accounts did. The 1951 budget also included narrative statements on program and performance for each account. Narrative statements varied in their approach, some presenting workload and unit cost information and others simply describing activities within the budget account. Finally, the 1951 budget replaced detailed lists of civilian positions and salaries that accompanied each account with summary information on employment levels. Most executive agencies charged with implementation had high expectations for performance budgets as a means of better defining, presenting, and executing the budget. Performance budgets were expected to align programs and activities in a uniform manner and assist managers in making trade-offs between--and within--particular programs. Agencies also viewed performance budgets as correcting budgeting and accounting weaknesses and improving the administration and oversight of programs. And, some agencies saw the submission of budgets on a program and functional basis as a simplification of the federal budget. However, some agencies did provide more cautionary statements regarding the implementation of performance budgeting. In particular, agencies expressed concern regarding whether--or how--to define different functions and activities consistently. Agencies also noted that the requirements for performance budgeting were adding to rather than substituting for their current budget and reporting requirements. Agency comments regarding the requirements for performance budgeting were mixed, with some expressing concern that requirements were too rigid and others stating that requirements were very generally and broadly defined. Congressional efforts to enact performance budgeting requirements were contained in two laws. The first was the National Security Act Amendments of 1949 (63 Stat. 578, August 10, 1949) which set performance budgeting requirements for the newly created Department of Defense (DOD) specifically patterned after the Hoover Commission's recommendation for a performance budget. That act added Title IV, the "Promotion of Economy and Efficiency Through Establishment of Uniform Budgetary and Fiscal Procedures and Organizations," to the National Security Act of 1947 and statutorily mandated the implementation of a performance budget similar in form to the President's fiscal year 1951 budget. New Section 403, "Performance Budget," stated: "The budget estimates of the Department of Defense shall be prepared, presented, and justified, where practicable . . . so as to account for, and report, the cost of performance of readily identifiable functional programs and activities, with segregation of operating and capital programs. . . ." And, as far as practicable, the Defense budget estimates and authorized programs were to be presented in a comparable form and follow a uniform pattern. The use of a performance budget was expected to correct weaknesses in budget formulation and presentation as well as improving the administration and management of authorized programs. And, BOB expected that a uniform pattern of accounts would allow comparisons across the services that were currently difficult to obtain. A second law, the Budget and Accounting Procedures Act of 1950 (BAPA, 64 Stat. 832, September 12, 1950), ultimately provided a less prescriptive definition of performance budgeting for governmentwide application. Early versions of this bill had contained detailed definitions of performance budgeting very similar to that of Title IV legislation. However, during congressional deliberations, the specific language for performance budgeting was removed from the bill. The conference report notes that the term performance budget was considered surplusage--words in a statute which add nothing to the force and legal effect of the statute--and might result in an interpretation more restrictive than intended by the Congress. BOB also supported BAPA's less prescriptive language, arguing that (1) the executive branch was already implementing performance budgeting and (2) specific performance budgeting language would appear too rigid and make it difficult to proceed with future budgeting improvements. Thus, the final enacted version of BAPA did not contain the term "performance budgeting." Instead, the final language stated in part: "The Budget shall set forth in such form and detail as the President may determine-- (a) functions and activities of the Government;" RESULTS The Congress considered that the Hoover Commission recommendation for performance budgeting was instituted on a governmentwide basis with the passage of BAPA. The second Hoover Commission, established on July 10, 1953 (67 Stat. 184), noted that performance budgeting was first used generally in the budget for fiscal year 1951. Reflecting on the implementation of performance budgeting, the second Commission observed that many programs did not have adequate cost information and suggested that budget activities and organization patterns be made consistent and accounts established to reflect this pattern; and, that budget classifications, organization, and accounting structures should be synchronized. DOD performance budgeting efforts in the 1950's did work towards a consistent presentation of budget accounts that led to the current budget structure of DOD. Comptrollers were established in DOD and the Services with the aim of enhancing the development of adequate budget preparation and review. Each Service was required to develop similar systems which allowed for some general comparisons between the services and standard classifications of cost categories were developed.\2 Although it did not specifically mention performance budgeting, BAPA is generally credited with advancing several important changes to federal budget practices. The statute institutionalized efforts to report sub-account level information to the Congress through the obligations by activity sections, now termed program activities.\3 A greater amount of performance information was placed into the President's budget, primarily output based work-load and unit cost information. BAPA also required additional coordination between agencies, created management devices such as working capital funds, delineated responsibilities for budgeting and accounting between the executive and legislative branches, and emphasized the need for a close relationship between accounting, management, and programming activities. Despite the successes cited, concerns remained that the budget did not adequately link programs with their costs. The report of the second Hoover Commission summarized these concerns as follows: "The installation of performance budgeting in the Federal agencies has met with varying degrees of success. . . . performance budgeting has encountered practical difficulties greater than originally contemplated and in some cases created congressional dissatisfaction with respect to program classification and accounting support." In 1954, Arthur Smithies, noted chronicler and analyst of the budget, clarified this issue by distinguishing between a performance budget and a program budget. "Congressmen themselves are dissatisfied with the present form of the budget. They feel they have lost something by the performance budget and have not gained much . . . . Unless the performance budget can evolve into a true program budget, the Congress may decide to revert to the old system and console itself with the fiction that it has no programmatic responsibilities . . . . While the preparation of a meaningful program budget is a task of immense difficulty, and may never be wholly successful, there can be little doubt that further progress without direction is both feasible and desirable." -------------------- \1 The executive branch acted on Hoover Commission recommendations to change the President's budget prior to legislative enactment of the Budget and Accounting Procedures Act of 1950. In fact, the executive branch argued that congressional action was not necessary, since the budget presentation was already being changed. \2 While DOD budgets continue to reflect the performance budgeting requirements developed, the term performance budget was repealed when the National Security Act Amendments and other statutes were codified in 1962. The legislative history of this codification (P.L. 87-651, September 7, 1962) notes that its passage was not intended to make any substantive change to the law, but to bring up to date Title 10 of the U.S. Code. \3 In fact, GPRA requires that annual performance plans cover each program activity set forth in the budget of an agency. PLANNING-PROGRAMMING-BUDGETING-SYSTEM (PPBS) ========================================================= Appendix III CONTEXT In January of 1965, President Johnson described the nation's economic performance as "a creditable record of achievement." From 1961 to 1964 the economy had been growing in real terms at an average annual rate of over 5 percent. Average annual inflation was just over 1 percent during this period, while unemployment was roughly constant at 5 percent. There was some concern about annual federal deficits, which in 1962 reached $7 billion, or 1.3 percent of GDP. A Planning-Programming-Budgeting-System (PPBS) was seen as a means of building upon the Nation's economic strength by modernizing the management tools used in the federal government. Proponents of PPBS believed that efficiencies and improvements in government operations could be achieved through a common approach for (a) establishing long range planning objectives, (b) analyzing the costs and benefits of alternative programs which would meet these objectives, and (c) translating programs into budget and legislative proposals and long-term projections. President Johnson considered PPBS a technique for controlling federal programs and budgets, rather than "having them control us." Furthermore, an earlier introduction of a PPBS-type system in DOD in 1961 was deemed a significant improvement over previous budget practices. Prior to PPBS, the DOD system was highly decentralized and resource formulation and allocation processes across the services were duplicative, inequitable, and limited to consideration of a single budget year. Initially termed a "program package-program element" system, DOD's PPBS activities provided a means of evaluating and deciding among major alternative methods of accomplishing military missions. Planning horizons were also extended with the development of a 5-year defense plan. On August 25, 1965, President Johnson announced his intention to introduce PPBS on a governmentwide basis, asserting that three major objectives would be achieved: "(1) It will help us find new ways to do jobs faster, to do jobs better, and to do jobs less expensively. (2) It will insure a much sounder judgment through more accurate information, pinpointing those things that we ought to do more, spotlighting those things that we ought to do less. (3) It will make our decision-making process as up-to-date, I think, as our space-exploring program." IMPLEMENTATION APPROACHES There were distinct differences between DOD approaches and the subsequent governmentwide implementation of PPBS. DOD implementation involved several hundred analysts and over 10 years of contractor-assisted development efforts. DOD introduced three key phases of activity for implementing PPBS: (1) reviewing requirements, (2) formulating and reviewing programs, extended several years into the future, and (3) developing annual budget estimates. The first two phases were continual, year-round efforts that resulted in a 5-year program plan for the entire defense establishment. In phase three, the budget year requirements established in the 5-year program plan are separated out into an annual budget request. In contrast to this phased approach used at DOD, governmentwide implementation of PPBS was expected to be accomplished in less than 6 months. On October 12, 1965, less than 2 months after the formal announcement of PPBS, the Bureau of the Budget (BOB) issued Bulletin 66-3 which provided agency guidance and instructions for implementing PPBS. Overall, 22 executive departments and establishments were mandated and 17 smaller agencies were encouraged to implement PPBS. Bulletin 66-3 gave agencies 10 days to designate an official responsible for their PPBS system and to report their choice to BOB. Within the next 20 days, agencies were to make tentative decisions on their broad program categories. Agency instructions, procedures, or regulations regarding PPBS implementation were to be forwarded to BOB within the next 2 months. A final Program Structure, approved by the director of the agency, was expected by February 1, 1966. Program Structures were the basic foundation of the PPBS system, designed to provide a coherent statement of a national need, an agency's authority to fill that need, and the activities planned to meet that need. BOB expected agencies to categorize all operations and activities in output oriented terms reflecting each agency's objectives. Three subdivisions of activities were available within the Program Structure: (1) program categories, defined as activities with similar broad missions, (2) program subcategories, defined as subdivisions of narrower objectives, and (3) program elements, defined as the specific products (e.g., goods and services) contributing to agency objectives. For example, if education is a sample Program Structure, a program category might be secondary education; subcategories might include college preparatory and vocational activities; and program elements might include facilities, books, and teachers. Three documents were expected to provide data on Program Structures: Program and Financial Plans (PFP), Program Memoranda (PM) and Special Studies. The PFP were similar to the DOD 5-year plan, containing multiyear descriptions of program objectives and accomplishments in quantitative nonfinancial terms related to the universe of need. PM were expected to describe agency program categories, summarize PFP data, and delineate recommended programs. Agencies were to illustrate how they would achieve national needs, showing costs and effectiveness of alternative objectives, program types, and levels of operation. Furthermore, PM should include any assumptions and uncertainties on the cost and criteria used to support agency recommendations and estimates. Special Studies were expected to vary greatly in scope and were carried out in response to agency top management or BOB inquiries, or at the initiative of analytic staff. Contrary to expected time frames, PPBS implementation proceeded slowly--even after several years of effort. In November of 1966, President Johnson issued a memorandum to Cabinet members and agency heads stating that too many agencies had been slow in establishing PPBS and that PPBS had not been used to make top management decisions. The President urged personal participation of agency heads and instructed the Director of BOB to review and report on agency progress in implementing PPBS. Nevertheless, fully 2 years into implementation, agency directors and former BOB officials testified that implementation was proceeding more slowly than hoped. Some agencies characterized their efforts as in the beginning stages or as requiring several more years before achieving notable results. Others reported that new information systems had to be developed or devised in order to track data on a program or mission basis. As originally designed, PPBS information systems were not expected to correlate to the Presidents' budget submission to the Congress. Instead, agency operating budgets--used to allocate resources and control day-to-day operations--were expected to conform gradually with PFP. Hence, BOB did not expect changes to the President's budget or to the internal submission of annual budget requests to BOB. Bulletin No. 66-3, the first guidance on implementing PPBS, specifically noted: "The introduction of the Planning, Programming, and Budgeting system will not, by itself, require any changes in the form in which budget appropriation requests are sent to Congress. Further, this Bulletin is not to be interpreted to set forth changes in the format of annual budget submissions to the Budget Bureau." However, to affect resource allocation decisions made within the executive branch, PPBS reports were timed to occur with the BOB budget preparation schedule. PM and Special Studies were expected to be used during the BOB spring review of the budget, when agencies and BOB would develop initial estimates of budgetary need and PFP was expected to be used during the fall as agencies developed annual budget requests for BOB. The result was two tracks of budget information: one which addressed the new PPBS requirements and one which addressed the existing BOB requirements for submitting the President's budget to the Congress. This separation between Program Structures and the President's budget created an implementation burden that later BOB bulletins tried to address, primarily by devising a more concrete link between PPBS and the budget. In July 1967, a second BOB bulletin (No. 68-2) directed agencies to provide a crosswalk--or a reconciliation--between their PPBS and appropriations structures. The crosswalk was to be sufficient to ensure that the budget submission was consistent with the intent of the program decisions. In 1968, the Congress requested and received an accompanying commentary to BOB's third bulletin (No. 68-9); the commentary noted that the then-current "two-track system" of program and appropriation structures was confusing and causing an undue burden. Agencies were asked to consider changing their PPBS program structures so as to avoid crosswalks and integrate PPBS and appropriations structures. Subsequent BOB guidances made procedural changes to the PPBS system, primarily limiting the scope and magnitude of reporting requirements for agencies and increasing staff hiring and training. Although originally allowed to include unlimited program proposals without regard to agency budget levels, the PFP requirements became limited to budgeted activities. Noting that many PM lacked analysis of major alternatives, policy decisions, or strategies directed towards specific outputs, BOB dramatically reduced its requests for major policy issues presented in PM documents. Further, BOB provided agencies extra preparation time for PM, and pledged assistance with the analysis and review of major policy issues. Lastly, during the first two years of implementation, almost 900 PPBS-specific positions were created, of which almost 400 were filled through new hires. Four years into implementation, over 4,500 staff had attended PPBS training sessions. RESULTS While DOD continues to use PPBS procedures today, the governmentwide initiative begun with such great promise in 1966 was formally discontinued in 1971 with remarkably little comment. Some observers and participants faulted the implementation process, contrasting DOD's 10 years of preparation with a significantly shorter governmentwide implementation period. A former agency official charged that PPBS was implemented indiscriminately, with agencies lacking the capability to perform PPBS activities, and BOB lacking the competence to guide them. Others said PPBS failed to garner the necessary support it needed because it affected the balance of power between the executive and legislative branches. PPBS participants and observers cited many problems developing measures and analysis techniques, as well as incorporating results into decision-making practices. Congressional hearings reviewed executive approaches to estimating, measuring, and valuing benefits, ultimately recommending the use of standard interest rates and discount policies. A GAO report cited several obstacles to relating output measures to program benefits; for example the report noted that the increased use of grants meant that program outputs could not be obtained due to a "rather loose and intermittent" federal control over grantees' program performance.\1 Some members of Congress questioned the broader purposes and accomplishments of PPBS as a decision-making tool, particularly in light of the impact of assumptions on analysis results; they further noted that their lack of access to PPBS documents placed them at a disadvantage in considering resource allocation questions. Some agencies cautioned that PPBS analysis could not substitute for inherently political decisions such as the allocation of resources among different priorities (e.g., health v. education); others asserted that decisions for certain federal functions--such as foreign affairs--could not be relegated to systems analysis. Other observers found PPBS unrealistic because it attempted to improve decision-making without recognizing the differing goals and interests of the decisionmakers. Over 3 years into PPBS implementation, the Joint Economic Committee of the Congress published a compendium of papers on the analysis and evaluation of public expenditures in PPBS.\2 In this compendium, an Assistant Director for Program Evaluation at BOB noted that expectations for PPBS needed to be constrained by certain realities of the federal environment, namely -- Governments operate with limited resources, and the demand for these resources always exceeds the available supply. -- Past resource commitments place heavy constraints on current budgets, providing limited control over resource allocation. -- Workable program measurement techniques are difficult to achieve, particularly given the complexity and size of the federal government. -- Implementation of new ideas can be slowed by the size of government, the inherent uncertainties of its tasks, and the high degree of coordination needed. -- Often there are political and moral claims made on the federal government which do not necessarily reflect an interest in cost effectiveness or efficiency. -- The resource allocation process in government is not well linked to planning, as these activities serve different needs and respond to different time frames. -- Once a budget is established, there is minimal accountability for performance.\3 Although it failed as a governmentwide performance budgeting initiative, PPBS is credited with instituting improvements in federal program management. PPBS allowed agencies to reappraise their mission and functions; accumulate better information on inputs, outputs, and their relationship to objectives; and increase top official interest over planning, budgeting, and performance. Furthermore, decisionmakers increased the use of systems analysis, recognizing its value as a means of better understanding outputs, benefits, and costs. Finally, PPBS left a long-standing legacy of increases in the amount and quality of program evaluation in the federal government. Despite the immense implementation difficulties--a truncated start-up, significant increases in paperwork, problems measuring program benefits and costs, and complex crosswalks to link program and budget structures--few individuals argued against the goals of PPBS. Some argued for its continuation, asserting that the goals and purposes of PPBS were critical to improving government operations. At a congressional hearing in 1970, one former HEW official summarized this view in the following manner. ". . . Rekindle the spluttering flame of PPB[S] . . . . In my judgment PPB[S] is absolutely right in concept. It requires more sustained support from the Congress, the White House, and the BOB. It requires patience. Its message and value is care in considering what the Government has done and might do. New initials will be needed but the job must be done."\4 -------------------- \1 Survey of Progress in Implementing the Planning-Programming-Budgeting System in Executive Agencies (B-115398, July 29, 1969). \2 The Analysis and Evaluation of Public Expenditures: the PPB system, a compendium of papers submitted to the Subcommittee on Economy in Government of the Joint Economic Committee, Congress of the United States, U.S. Government Printing Office, Washington: 1969. \3 "The Status and Next Steps for Planning, Programming, and Budgeting," by Jack W. Carlson, Assistant Director for Program Evaluation, BOB. \4 Statement of William Gorham, formerly Assistant Secretary for Program Coordination at HEW, in a hearing before the Subcommittee on Economy in Government of the Joint Economic Committee, Congress of the United States, 91st Congress, Second Session, June 2, 1970. MANAGEMENT BY OBJECTIVES (MBO) ========================================================== Appendix IV CONTEXT During the 1960's a bipartisan consensus developed that federal management needed improvement. A study requested by President Johnson in 1966 and carried out by the Heineman Task Force criticized the federal government's management of the new Great Society programs. The Task Force recommended strengthening the management responsibilities of the then-Bureau of the Budget (BOB). In 1970 President Nixon proposed changing BOB into a new Office of Management and Budget (OMB), with the new agency expected to give greater attention to federal management issues. To gain greater administrative control over major executive branch departments and agencies, President Nixon proposed a new governmentwide initiative: Management by Objectives (MBO). MBO was a popular management technique used in the private sector and had also been implemented at the Department of Health, Education, and Welfare during the President's first term. MBO was intended to centralize goal-setting decisions while at the same time allowing managers to choose how to achieve the goals. It focused on tracking progress toward goals previously agreed upon between a supervisor and subordinate. IMPLEMENTATION APPROACHES President Nixon formally initiated MBO in an April 18, 1973, memorandum to 21 agencies, which included the 11 cabinet departments and constituted about 95 percent of the budget and federal employees. President Nixon stated: "I am now asking each department and agency head to seek a sharper focus on the results which the various activities under his or her direction are aimed at achieving. . . . This conscious emphasis on setting goals and then achieving results will substantially enhance federal program performance." A follow-up memo to the MBO department heads from the Director of OMB further explained that the new initiative aimed at better communication, faster identification of problems, and greater accountability of managers to supervisors. Ultimately, the OMB Director stated, MBO would lay the groundwork for the President to decentralize more responsibility to the agencies. In his April 1973 letter, the President asked each agency to propose the 10 or 15 most important objectives--referred to as "presidential objectives"--to be accomplished in the coming year; the goal was to identify 100 presidential objectives. Different agencies were given different deadlines, varying between 2 and 8 weeks, to submit proposals. Subsequently, agencies were told that their search for objectives need not be limited to their proposals to the President. Agencies were encouraged to identify additional objectives, to track progress towards achieving them, and to use MBO in all aspects of their operations. OMB was to play a key role in implementing MBO. As part of MBO implementation, a new position within OMB was created: the "management associate." Thirty management associates with varying backgrounds, some with government experience and some without, were hired. Their responsibilities would include providing day-to-day assistance to the departments in preparing objectives, tracking progress, working closely with OMB budget examiners, and providing technical assistance to agency staff and OMB top management to help implement the initiative. In addition, staff were specially selected to implement MBO at the agencies and were generally located between the Office of the Secretary and program managers. OMB statements emphasized that the initiative was to be conducted with a minimum of paperwork. Face-to-face meetings were to be held roughly every 2 months between top OMB and agency staff. The meetings were to focus on agency progress in achieving objectives, problems requiring top management attention, and any changes to objectives. Some existing OMB requirements were eliminated as a way of encouraging agency acceptance of the new initiative. OMB gave agencies some guidelines on their proposals for presidential objectives. In proposing presidential objectives, agencies were to consider the importance to the President's agenda, measurability, and the ability to achieve the objective without additional resources and within 1 year. Agencies were to identify objectives on their own--that is, without intervention by OMB--and were asked to develop action plans with specific milestones for accomplishing objectives. All objectives were to be linked to the organizational units that would be held accountable for achieving them. If circumstances warranted, objectives could be changed during the year. OMB would review agencies' proposed presidential objectives as well as track progress toward achieving them. In its first year, no explicit connection of MBO to the budget process was attempted. MBO fell far short of expectations during its first year. Although 20 of the 21 MBO agencies had identified presidential objectives and 18 had progress tracking systems in place by the end of the first year, many other important implementation steps were not achieved. For example, management conferences were held, although not as often as originally planned with 4 to 6 months passing between conferences for some agencies. Despite OMB's intention to address this problem, scheduled meetings continued to be canceled frequently. And, as MBO reviews were increasingly done at the staff level, rather than at the OMB and agency head level, MBO paperwork increased. At OMB, tensions initially developed between the new management associates and OMB's budget examiners; this eased to some extent as the management associates found that monitoring agency objectives was not a full-time task, especially given the associates' lack of control over agency actions. Increasingly the management associates became involved in non-MBO tasks such as doing special studies. Most importantly, presidential involvement in MBO also faltered during 1974, affecting agency implementation and acceptance of MBO. In the second year of MBO, an attempt was made to re-emphasize MBO by linking objectives with agency budget submissions. In a February 1974 meeting, OMB informed agency heads that their 1976 budget requests were to be based on their presidential and agency (secretarial) objectives. OMB hoped that this would increase the permanence of MBO and encourage more explicit statements of the purposes for which money was to be spent. In June 1974, OMB asked the 21 MBO agencies to identify selected objectives in the letters transmitting their budget requests to OMB; these objectives were to be discussed in depth in the budget justifications. Agencies were told to "be prepared to provide" outlay estimates and "preliminary" schedules of milestones upon request, but were not required to include action plans. In August 1974, President Nixon resigned and, shortly after taking office, President Ford endorsed agencies' proposed 1975 Presidential objectives. These were the last presidential objectives requested under MBO. RESULTS Although certainly affected by President Nixon's resignation, the MBO initiative suffered from its initial separation from existing budget formulation processes and from problems in identifying and measuring objectives. Efforts in the second year to tie the MBO initiative to the budget's priority setting processes were quickly overwhelmed by its early demise. The President's request that agencies focus on results and express those results in measurable terms did not make the practice of performance measurement any easier. For various reasons agencies found this difficult to do. Not surprisingly, as initially submitted to OMB, agencies' objectives were often vaguely worded (e.g., "the abolition of crime in society" or "to make the U. S. Merchant Marine the most competitive in the world") and not easily measurable. In addition, agency objectives often dealt with matters not achievable within a single year (such as finding a cure for cancer) or were beyond the control of agency managers (such as improving water quality), making accountability problematic. Despite these issues and its brief life as a formal initiative, proponents believe that MBO had positive results in both the short and long term. For the administration that proposed it, the MBO initiative enhanced its ability to explain the President's agenda to the public--for example, the emphasis on transferring more federal power to cities and states. Some OMB staff and agency officials found MBO valuable as an internal agency management process, helping to clarify goals and associated activities. To some extent, the basic concepts of MBO--negotiating goals and holding subordinates accountable for achieving them--have survived in federal management practices. In addition, the potential of MBO as a tool for articulating presidential agendas and linking them with the budget was later confirmed by a similar initiative under President Bush; this initiative included publishing presidentially approved objectives, the resources needed to achieve them, and relevant accomplishments in the President's Budget. And issues raised during MBO concerning the difficulties inherent in identifying and measuring federal outcomes would remain for later initiatives to address. ZERO-BASE BUDGETING (ZBB) =========================================================== Appendix V CONTEXT In the mid 1970s, the annual deficit was a matter of public debate. By 1977 the annual deficit had been above $50 billion for 2 years, reaching a post-World War II high of $73.8 billion for fiscal 1976. A general sense existed that federal spending was out of control, with much of it no longer subject to annual appropriations but driven by permanent entitlement programs and multiyear budgetary authority. During 1976, the Congress and Candidate Jimmy Carter had responded to the new budget situation. Beginning in the spring the Congress held hearings on proposals for so-called "sunset" legislation that would have required periodic zero-base reviews of all federal programs by their congressional authorizing committees. Sunset proposals, however, did not become law. While campaigning for the presidency, Jimmy Carter promised to balance the budget within his first term and to reform the federal budgeting system, which he characterized as "inefficient, chaotic, and virtually uncontrollable by either the President or the Congress." To these ends he had promised to introduce zero-base budgeting (ZBB), which he had used as Governor of Georgia and which also had been discussed in sunset hearings. In fall 1976, congressional appropriations committees asked selected independent agencies to pilot test the applicability of ZBB concepts to legislative decision-making. Used in private industry as well as in some state and local governments, ZBB in theory required expenditure proposals to compete for funding on an equal--starting from "zero"--basis. ZBB prepares a detailed identification and evaluation of all activities together with alternatives, and spending necessary to achieve desired plans and goals. Where federal budgeting in recent years had made incremental changes to an accepted base of past spending, ZBB in contrast sought to look below the base, evaluating the efficiency and effectiveness of current operations and comparing the needs of one program against the needs of other programs that might be of higher priority. ZBB also looked to a greater involvement of program managers in budgeting as a way to identify new efficiencies and to incorporate better analysis into budget decision-making. IMPLEMENTATION APPROACHES On February 14, 1977, shortly after his inauguration, President Carter issued a memorandum to the heads of executive departments and agencies mandating use of zero-base budgeting for all fiscal year 1979 agency budget requests. The memorandum mandated that a new ZBB budget process would replace--not simply accompany or link to--the existing executive branch budget formulation process for all budget proposals in the immediately upcoming budget cycle. Consistent with an emphasis in ZBB theory on a close link between planning with budgeting, federal planning and budgeting under ZBB were to be done at the same time, in a single process. In contrast to its implementation of PPBS and MBO, OMB did not add or create a special staff for ZBB. Federal managers and budgeteers were expected to implement the new initiative. ZBB would not affect budget materials provided to congressional appropriations or authorizing committees, nor would it change the form of the President's Budget. Formal implementation steps were taken within 2 months of the memorandum. On March 21, 1977, OMB sent agencies draft ZBB guidelines for comment, issuing final guidance on April 19 as Bulletin 77-9. In effect, agencies were given a lead time of about 6 months before final budget submissions were due to OMB. Agencies were to set up their own ZBB systems using the steps outlined in the Bulletin as a framework. Among other new requirements, agencies were asked to identify the "decision units" for which budget requests would be made. A decision unit was to be -- "at an organizational or program level at which the manager makes major decisions on the amount of spending and the scope, direction, or quality of work to be performed." -- "not so low in the structure as to result in excessive paperwork and review . . . [nor] so high as to mask important considerations and prevent meaningful review of the work being performed." -- "normally . . . included within a single account, be classified in only one budget subfunction, and to the extent possible, reflect existing program and organizational structures that have accounting support." In all cases, the guidance stated, the identification of the decision units was to be determined by the information needs of top management. Budget requests for each decision unit were to be prepared by their managers, who would (1) identify alternative approaches to achieving the unit's objectives, (2) identify several alternative funding levels, including a "minimum" level normally below current funding, (3) prepare "decision packages" according to a prescribed format for each unit, including budget and performance information, and (4) rank the decision packages against each other in a series of steps, beginning with program managers and proceeding up the hierarchy. The results of the ZBB process would be agency budget justifications and rankings, with the latter required to be submitted to OMB but not to the Congress. With OMB's approval, agencies could consolidate decision units as a means to minimize paperwork and the review burden on top management. The guidance also required agencies to set objectives and performance indicators at the beginning of their ZBB process. Top and program managers were to set objectives as "explicit statements of intended output, clearly related to the basic need for which the program or organization exists." Concurrently they were to identify the key indicators to be used in measuring performance and results. These should be "measures of effectiveness, efficiency, and workload for each decision unit. These measures can often be obtained from existing evaluation and workload measurement systems." Indirect or proxy indicators could be used if these systems did not exist or were under development. A "lack of precise identification and quantification of such objectives," however, would "not preclude the development and implementation of zero-base budgeting procedures." Despite considerable variation in how agencies implemented ZBB, some patterns can be discerned. Some agencies tended to associate their decision units with their account structure or, within their account structure, with their program activities. Some agencies did not identify minimums below current funding, and many identified minimums as an arbitrary percentage of current funding, generally between 75 and 90 percent. Agencies also made use of the option to consolidate decision units and often set initial decision units at high organizational levels (e.g., the division level or higher). Lastly, one study of several agencies found that fewer than half the decision packages examined had quantifiable accomplishments, workload, or unit cost information. The next year, in May 5, 1978, OMB issued Circular A-115, which revised some aspects of the ZBB process. Addressing problems with objectives and performance information, OMB now urged agencies to use the results of their performance evaluations in analyzing alternative methods of accomplishing objectives and in analyzing anticipated accomplishments identified with each level of performance. The circular also strengthened language dealing with the objective-setting requirement. The guidance on selecting decision units, preparing rankings, and consolidation was clarified. A requirement to train staff before they participated in the ZBB process was also added. In other respects, however, ZBB requirements were unchanged. For example, no provision was made for a separate planning phase, and the requirement to prepare decision packages for all budget requests, including those for mandatory programs, remained. The budget that resulted from agencies' and OMB's first year of ZBB efforts disappointed some observers. Few significant budgetary actions were identified as resulting from ZBB, and some questioned the utility of the many hours spent by program managers, budgeteers, and top managers on ZBB. In the following year, agency budget justifications to OMB continued to be prepared using ZBB, but agency budget justifications to the Congress continued to be prepared as in the past, largely without reference to agencies' ZBB information. As the Carter presidency proceeded, less and less attention was devoted in the Budget Message to the role and claimed achievements of ZBB. On August 7, 1981, in the first year of the succeeding administration, OMB rescinded circular A-115 requiring agencies to have ZBB systems. Some ZBB requirements, however, survived beyond the formal life of the initiative. Requirements for agencies to identify "decision units" and prepare consolidated rankings remained until May 1986. A requirement to identify three funding levels lasted even longer, remaining until 1994, as did an OMB option to request that the agency present a "consolidated" ranking of "program elements and related funding levels." RESULTS In one sense, ZBB was successfully implemented: all agencies submitted the required paperwork on time. By the end of ZBB's first budget year, agencies had prepared about 25,000 internal decision packages and submitted about 10,000 of these to OMB. But in essential ways federal ZBB had not been an exercise in zero-basing a budget. The widespread use of arbitrarily chosen percentages to identify alternative funding levels, rather than analysis based on program knowledge and performance information, precluded genuine zero-basing, as did consolidation and selection of initial decision units at high levels in the organization. From the beginning, paperwork burden for federal managers constituted a significant implementation problem. One study estimated that paperwork increased, on average, 229 percent in ZBB's first year. In addition to the ZBB packages, agencies had to prepare separate budget materials, often using different categories, for OMB, appropriations, and authorizing committees. Preparing crosswalks between these added to agency burden. Agencies believed that inadequate time had been allowed to implement the new initiative. The requirement to compress planning and budgeting functions within the timeframes of the budget cycle had proven especially difficult, affecting program managers' ability to identify alternative approaches to accomplishing agency objectives. Some agency officials also believed that the performance information needed for ZBB analysis was lacking. Available information concerned processes and activities, not how well these processes and activities performed. Agencies also questioned the need to prepare and rank decision packages for programs whose spending levels were outside their control. For example, the Department of Health, Education, and Welfare did not identify minimum levels for social security and other programs where it believed spending was uncontrollable; Treasury stated it saw little use in ranking decision packages for interest on the debt since the interest would have to be paid in any case. Paperwork and other burden and technical difficulties were compounded by agency perceptions that OMB had not used the results of agencies' ZBB efforts in its budget decision-making. In Congress, the results of the congressionally requested ZBB pilots, made public in June 1977 cast doubts on ZBB's suitability as a potential tool for congressional decision-making. One major thrust of the pilots had been to see whether ZBB rankings--comparing priorities of "decision packages" against one another--could be used by appropriators to identify the impact of budget cuts. The results of the pilots were not encouraging. In one pilot, the agency had failed to set minimums below current funding for over one-third of its decision units and refused to rank its decision packages because the process-oriented program activity structure of the agency's budget was too interdependent to permit meaningful ranking. The lack of cost accounting information needed to identify alternative funding levels was also cited as a technical problem. Finally, the level of burden and paperwork was a problem for both for agencies and appropriators. In one typical case, 362 pages were needed for an agency's ZBB-based budget justification versus 72 pages for its non-ZBB justification. The results of the congressional pilots were largely consistent with later agency experiences. No mechanism existed, however, to incorporate lessons learned from the congressional pilots into executive branch ZBB implementation. By the time OMB sent agencies a survey in October 11, 1977, seeking their views on implementation problems and proposed solutions, the gaps between ZBB's initial promise and its first year results were becoming apparent. Despite implementation problems and the relatively short time span in which all its elements were required, federal ZBB has been credited with some positive results. Some participants in the budget process as well as other observers attributed certain program efficiencies, arising from the consideration of alternatives, to ZBB. Interestingly, ZBB established within federal budgeting a requirement to present alternative levels of funding linked to alternative results--a requirement that lasted until 1994. OVERVIEW OF THE GOVERNMENT PERFORMANCE AND RESULTS ACT ========================================================== Appendix VI GPRA seeks to promote greater public confidence in the institutions of government through a better reporting and accounting for the outcomes of federal programs. As stated in the act, the goals of GPRA are to "(1) improve the confidence of the American people in the capability of the Federal Government, by systematically holding Federal agencies accountable for achieving program results; (2) initiate program performance reform with a series of pilot projects in setting program goals, measuring program performance against those goals, and reporting publicly on their progress; (3) improve Federal program effectiveness and public accountability by promoting a new focus on results, service quality, and customer satisfaction; (4) help Federal managers improve service delivery, by requiring that they plan for meeting program objectives and by providing them with information about program results and service quality; (5) improve congressional decisionmaking by providing more objective information on achieving statutory objectives, and on the relative effectiveness and efficiency of Federal programs and spending; and (6) improve internal management of the Federal Government."\1 From these broad purposes, a system of interrelated plans and reports provides the basis for linking federal resources and results, with requirements and new concepts piloted before governmentwide application. STRATEGIC PLANS GPRA requires each agency to develop strategic plans covering a period of at least 5 years. Agencies' strategic plans must include the agency's mission statement; identify long-term general goals, including outcome-related goals and objectives; and describe how the agency intends to achieve these goals through its activities and through its human, capital, information, and other resources. Under GPRA, agency strategic plans are the starting point for agencies to set annual program goals and to measure program performance in achieving those goals. To this end, strategic plans are to include a description of how long-term general goals will be related to annual performance goals as well as a description of the program evaluations used in establishing goals. As part of the strategic planning process, agencies are required to consult with the Congress as well as solicit the views of other stakeholders. Agencies' first strategic plans are to be submitted to the Director of OMB and the Congress by the end of fiscal year 1997. Strategic plans must be updated at least every 3 years. ANNUAL PERFORMANCE PLANS GPRA also requires each agency to prepare an annual performance plan that includes the performance indicators that will be used to measure "the relevant outputs, service levels, and outcomes of each program activity" in an agency's budget. The annual performance plan is to provide the direct link between strategic goals outlined in the agency's strategic plan and what managers and employees do day-to-day. When an agency believes it is not possible to express a measurable goal for a program activity, the agency may ask OMB's authorization to use a nonquantifiable goal. In addition, GPRA allows agencies to aggregate, disaggregate, or consolidate program activities for purposes of performance planning. These plans are also to be used by OMB to develop an overall federal performance plan for the federal budget, which is to be submitted each year to the Congress with the President's budget. The first annual performance plans are to be submitted to OMB in the fall of 1997, with the first overall federal performance plan due for fiscal year 1999. ANNUAL PERFORMANCE REPORTS Ultimately, GPRA will require that each agency prepare an annual report on program performance for the previous fiscal year. In each report, agencies are to review and discuss performance compared with the performance goals established in annual performance plans. When a goal is not met, agencies are to explain the reasons the goal was not met; plans and schedules for meeting the goal; and, if the goal was impractical or not feasible, the reasons for that and the actions recommended. Actions needed to accomplish a goal could include legislative, regulatory, or other actions or, when the agency found a goal to be impractical or infeasible, a discussion of whether the goal ought to be modified. The report is also to include the summary findings of program evaluations completed during the fiscal year covered by the report. Agencies' first performance reports for fiscal year 1999 are due to the President and the Congress no later than March 31, 2000.\2 MANAGERIAL FLEXIBILITY In crafting GPRA, the Congress also recognized that managerial accountability for results is linked to managers having sufficient flexibility, discretion, and authority to accomplish desired results. GPRA authorizes agencies to apply for managerial flexibility waivers in their annual performance plans beginning with fiscal year 1999. The authority of agencies to request waivers of administrative procedural requirements and controls is intended to provide federal managers with more flexibility to structure agency systems to better support program goals. The nonstatutory requirements that OMB can waive under GPRA generally involve the allocation and use of resources, such as restrictions on shifting funds among items within a budget account. Agencies must report in their annual performance reports on the use and effectiveness of any GPRA managerial flexibility waivers that they receive. IMPLEMENTATION APPROACH: PHASING-IN AND PILOTING OF REQUIREMENTS GPRA calls for phased implementation, as described above, beginning with selected pilot projects in performance goals and managerial flexibility in fiscal years 1994 through 1996. These pilots are expected to develop experience with GPRA processes and concepts before implementation begins governmentwide in 1997. As of March 1997, 68 pilot projects for performance planning and performance reporting were under way in 28 agencies. OMB also is required to select at least five agencies from among the initial pilot agencies to pilot managerial accountability and flexibility for fiscal years 1995 and 1996; however, OMB did not do so. GAO is required to report on governmentwide readiness for implementation by June 1, 1997; OMB is required to report on the costs, benefits, and usefulness of the performance planning and measurement pilots by May 1, 1997, identifying any recommended changes in GPRA requirements. GPRA also requires OMB to select at least five agencies, at least three of which have had experience developing performance plans during the initial GPRA pilot phase, to test performance budgeting for fiscal years 1998 and 1999. Performance budgets to be prepared by the pilot agencies are intended to provide the Congress with information on the direct relationship between proposed program spending and expected program results and the anticipated effects of varying spending levels on results. OMB is required to report on these pilots by March 31, 2001. OMB's report is to assess the feasibility of performance budgeting, recommend whether legislation requiring performance budgets should be proposed, and identify any other recommended changes to GPRA requirements. -------------------- \1 P.L. 103-62, sec. 2. \2 For fiscal years 2000 and 2001, agencies' reports are to include performance data beginning with fiscal year 1999. For each subsequent year, agencies are to include performance data for the year covered by the report and 3 prior years. *** End of document. ***