Implementation of the Government
Performance and Results Act of 1993

The Government Performance and Results Act of 1993 (GPRA) is intended to bring about a fundamental transformation in the way government programs and operations are managed and administered. The United States is at the mid-point of a four year period for making any needed changes in government processes, systems, and practices before the law takes full effect in 1997. This paper examines the United States experience to date.

INTRODUCTION

1. On August 3, 1993, President Clinton signed into law the Government Performance and Results Act.

2. The main features of this law are: -- a requirement for Federal departments and agencies to prepare strategic plans, beginning with an initial plan to be submitted to the Office of Management and Budget (which is an agency within the Executive Office of the President) and to Congress by September 30, 1997. -- a requirement that Federal departments and agencies prepare annual performance plans, setting out specific performance goals for a fiscal year, starting with a performance plan for fiscal year 1999. (The Federal Government's fiscal year begins October 1 and ends the following September 30. Fiscal year 1999 begins on October 1, 1998.) -- a requirement that the Office of Management and Budget (OMB) prepare an annual government-wide performance plan, which is based on the agency annual performance plans. The government-wide performance plan is to be a part of the President's budget and is transmitted to Congress. In the agency and government-wide performance plans, the levels of program performance to be achieved will correspond with the program funding level in the budget. The first of these plans will be for the fiscal year 1999 budget, which Congress should receive in February, 1998. -- a requirement that Federal departments and agencies submit an annual program performance report to the President and Congress, and which compares actual performance with the goal levels that were set in the annual performance plan. The annual report is due six months after the end of a fiscal year. The first report, covering fiscal year 1999, is to be submitted by March 31, 2000. -- provisions giving managers greater flexibility in managing by allowing the waiver of various administrative controls and limitations. In return, managers are expected to be more accountable for the performance of their programs and operations.

3. In the years prior to these government-wide requirements coming into effect, the law provides for several sets of pilot projects. The performance measurement pilot projects test and demonstrate whether the structure and specifications for the annual performance plan and program performance report work as intended. The managerial accountability and flexibility pilot projects examine the practical application of the managerial accountability and flexibility concept on Federal managers and their staff.

4. Within the Federal Government, GPRA applies to all 14 Cabinet departments, virtually all independent establishments (agencies), and all government corporations. A government corporation is an entity either owned or controlled by the Federal Government. Departments and establishments have component units, which may be called a bureau, administration, or service, and are covered by being a part of a department or independent establishment. In this paper, the term "agency" is hereafter used to mean a department, independent establishment, or government corporation.

5. The change that GPRA seeks to bring about would place much greater emphasis within the government on what programs are actually accomplishing, and how well the accomplishments match with the programs' purpose and objectives. In the United States, this is called 'Managing for Results'. To this juncture, the primary focus has more usually been on the amount of budget resources obtained, and what these buy: staff on-board, for example, new equipment, supplies, or the number of grants, procurements, and transactions that could be made.

6. GPRA centers on the role and responsibility of Federal managers (of which there are several tens of thousands within the Federal government). Other reforms have usually focused on particular functions (such as accounting and auditing) and which are often support or ancillary activities in contrast to an agency's core programs.

7. Many (perhaps most) Federal officials are mainly evaluated not on what their programs accomplished, but on whether they followed the administrative rules and internal practices that governed their conduct as managers. Under GPRA, this adherence still matters, but one's role and responsibility for a successful or failing program should be of greater importance.

   PART 1: BACKGROUND

   Earlier efforts

8. GPRA is not without its historical ancestors. These can be traced over the past 30 years, and, somewhat ominously, largely failed to take root. Indeed, if any had succeeded, this paper would have a different topic and theme.

9. The first was the Program, Planning, and Budgeting System (PPBS), which was introduced in the 1960's and sought to extend the program management and budgeting scheme being used by the Department of Defense to the government-at-large. What Vietnam- era 'whiz kids' in the Pentagon were able to accomplish in the Department of Defense with the operations research focus of PPBS could not be readily replicated elsewhere. PPBS was succeeded in the 1970's by Management by Objective (MBO) and then by Zero- Base-Budgeting (ZBB). Within the Federal Government, both MBO and ZBB soon became artifacts of that decade, although an evolved form of PPBS continues to be used in the Department of Defense.

10. During the 1980's, there were three significant initiatives: productivity improvement, quality management, and a short-lived resurrection of MBO. The productivity improvement effort quickly became mired in difficulty: problems arose in defining what units would be measured, compounded by the widespread incapacity of government financial systems to cumulate and assign costs for the units. Quality management continues to this day, although in an evolving and more focused form.

    Will history repeat?

11. Is GPRA yet another reform that will flower briefly and soon fade? What makes GPRA different from these earlier initiatives, and why might it succeed when the others have not?

12. An important distinguishing aspect of GPRA is that it is a law; the others were established through Presidential directive. GPRA is thus a creation not only of the President, but of the Congress as well. Seen another way, laws such as GPRA can be enduring monuments to governance; Presidential directives are more fragile edifices. Indeed, most of the earlier initiatives began and ended with a Presidential term.

13. By design, GPRA is not trying to transform immediately the whole Executive branch of government. Many of these other efforts attempted to do so, and quickly. The combination of haste and scope became the path of failed ambition. Major change requires time: staff often need to adapt, and systems and processes must be adjusted. Several initiatives sought to expand a much smaller effort that had worked well for a particular type or level of function into government-wide application. A full- scale instant immersion strategy was followed. But in not allowing for lessons learned, experience gained, and a trial and error approach, the onset of problems and flaws in the course of putting the reform in place first bred agency reluctance and then resistance to the reform itself.

14. In the past, too much was expected from existing systems and processes for measuring, reporting, and analyzing information. Only in recent years with widespread use of desktop computers and linked networks of computers, has the collection and presentation of timely performance information moved from being a paper- and staff-intensive effort to something that can be done in a more practical manner.

15. Performance information that managers use in managing programs and operations forms the underpinning for GPRA performance measurement. Some earlier efforts used this (or similar) information in a more limited way, such as analyses for decision-making. While such uses can be very important, the data was often special purpose, developed and used by only a few staff. These staff had little or no ownership for any goals or commitments associated with the information. Any tie to what managers would do (or accountability for how they managed) was remote, if it existed at all.

16. Perhaps most significantly, these are different times. Throughout previous decades, funding for Federal programs was generally stable, if not growing. Now, funding is mainly decreasing, and programs are increasingly being challenged to explain their purpose and demonstrate their value. Program performance now counts for much more than in the past.

17. The public's view of government has also changed. Several commentators have declared that American government is currently beset with three deficits: a budget deficit, a performance deficit, and a trust deficit. The performance deficit is indicated by a public that doesn't know what its government is doing. The trust deficit marks a public which has low confidence in its government doing the right thing. More and better information on what the government is doing and how well it is doing can address the trust and performance deficits.

    Immediate antecedents

18. GPRA's origin can primarily be traced to two separately conceived proposals put forward about five years ago. One proposal was described in the Management Report (January 1989) of then President Reagan. This report contained a chapter, prepared by OMB staff, titled "Government of the Future" and which outlined the basic structure of what would become four years later, GPRA.

19. The other seed for GPRA sprang from Congress. In 1991, Senator William Roth first introduced the legislation that would eventually become the Government Performance and Results Act. This was the other origin of GPRA. The proposed law was largely framed on the experience of one American city -- Sunnyvale, California -- which had successfully been using many of GPRA's main features for over a decade. The following year, the legislation was extensively rewritten and 19 other Senators (nearly equally representing both political parties) joined Senator Roth in sponsoring the proposed law. In late 1992, the legislation was approved by the Senate, but the House of Representatives took no action. In 1993, newly-elected President Clinton gave early and strong support for the proposed legislation. GPRA was subsequently passed by both the Senate and House of Representatives without objection, and was promoted and supported by members of both political parties.

20. The Chief Financial Officers Act became law in 1990, and several features of this Act also helped lay the groundwork for GPRA. The first was an instruction to agencies that program performance information was to be included in the annual financial statements required by this Act. (This performance information was retrospective, covering a past fiscal year.) The second was a joint OMB/agency effort which developed common performance measures for over 15 functions similar to those being routinely done in the private sector. These commercial-type functions included areas such as debt and credit management, real property management, electric power generation and distribution, supply and inventory management, and insurance.

21. The Executive branch did not rely solely on the Sunnyvale experience when it participated with the Congress in redrafting the proposed law, or in assessing whether GPRA would be feasible or practical at a national government level. Any confidence that GPRA might be carried out successfully primarily arose from two sources: the experience in developing common performance measures for the annual financial statements and the increased inclusion of performance information in these statements; and from similar initiatives that were taking hold in other countries. From a U.S. perspective, much weight was placed on a conclusion that these other national initiatives were still being advanced a half decade or more after being started (most of the previously referenced Federal efforts had already languished within such a time-period) and overall, seemed to be succeeding. The OECD Public Management Committee was especially useful in identifying these other national initiatives.

22. Nothing in GPRA is new or revolutionary to the American experience. GPRA's major elements can be found in most businesses, where these have been used for decades but for a different ultimate objective: profit or loss. What GPRA seeks to create is the government's counterpart to the business 'bottom line'.

    Parallel efforts

23. As GPRA became law, other initiatives were also coming to the forefront of Federal governmental reforms. The most prominent of these is the National Performance Review (NPR), a wide-ranging look at how government could be made to work better, and led by the Vice President.

24. The NPR has embraced GPRA's purpose and concept. Several parts of the NPR relate directly to performance. These include:

    -- Customer service standards. Agencies have established about 2,000 standards. The customer service standards reflect a major Federal emphasis on improving the delivery of government services and products to their users and beneficiaries. The customer service standards continue an emphasis on quality improvement and management that began last decade. Not all the standards are measurable. In time, those that are should be incorporated as performance goals in GPRA plans and reports.

    -- Performance agreements. These are agreements between the President and Cabinet Secretary or the head of an independent agency. The concept is based on performance-based employment contracts that are being used in several other countries, with several important differences. These agreements are at the highest level, and not between a minister and an agency director, which is the usual starting point elsewhere. The Presidential agreements are also not used to reward or penalize individual performance. Based on the Presidential agreement, agencies are then to create other agreements between the agency head and subordinate officials, and which 'cascade' down through the organization. About ten performance agreements have been signed to date. Some of these have cascaded down through an agency. and, at these tiers, can be linked to personnel evaluation systems and pay incentives. The intention is to expand the number of agreements, and include more performance-related goals and objectives that are in an agency's strategic plan or annual performance plan.

    -- Reinvention labs. Over 200 reinvention labs were created in government agencies. These labs are testing ways to streamline administrative processes, and reduce and eliminate unnecessary controls on Federal managers. The main focus of the labs has been on controls and rules established by the agency itself, and not on government-wide requirements.

    -- Performance partnerships. Performance partnerships are negotiated agreements between the Federal Government and individual States or local governments, in which the Federal Government provides the State or local government with greater flexibility in administering a program in exchange for greater accountability for program performance and results. Apart from social security and veterans programs, the Federal Government directly delivers few benefits or services to the public. More often, the States or local government provide these services, with the cost being funded entirely or partly by the Federal Government. Performance partnerships can bring more attention on the value and impact of these Federal funds. At this time, a small number of partnerships are being negotiated.

25. Some assert that there is no more important measure to be included in an annual performance plan than the cost per unit of service, unit of output, or unit of result. However, existing financial systems in many government agencies lack the capacity to assign, cumulate, or report costs. The government is addressing this shortcoming in two ways. First, the Federal Accounting Standards Advisory Board has recommended and OMB has issued an accounting standard requiring cost accounting in the agencies. Second, the government is completing a definition of the financial system capabilities and improvements that will be needed to achieve this cost accounting standard.

    PART 2. THE MAIN ELEMENTS OF GPRA AND THEIR STATUS

26. An overview of the design of GPRA reveals several underlying premises and aspects:

    -- Virtually the entire Executive branch is subject to the Act. Provision is made for classified plans and reports from several agencies where the information within must be protected for national security reasons.

    -- The departments, independent establishments, and government corporations that must fully meet GPRA requirements number about 75. Approximately 60 smaller agencies (those with $20 million or less in annual spending) may be relieved by OMB from having to meet all the requirements.

    -- GPRA is a very large umbrella, covering not only the expenditure of monies for programs and operations, but regulation and tax expenditures when the latter relate to the achievement of performance objectives. (Tax expenditures are allowances in the tax code, such as credits and deductions, and represent revenues foregone.)

    -- Because there is neither a single nor uniform approach to performance measurement across the government, GPRA avoids over-specification or rigidity in how agencies meet the Act's requirements. Flexibility is key if agencies are to adapt and fit their own particular processes and activities into a larger performance framework.

    -- Almost everything that is done by the government can be measured in some way. Certainly, there are a few activities which are not worth measuring from a performance standpoint (interest payments on debt, for example), or where an overly finite level of measurement makes little sense. And everything will never be measured equally well. To avoid unproductive debate on how best to categorize or classify any particular measure, GPRA defines only two types of performance measures: output and outcome measures. Both were defined sufficiently broadly to cover other measures, such as impact, attribute, efficiency, etc.

    -- The Act does not establish or assign responsibility for GPRA implementation to a particular organization or unit within an agency. In the law, the individuals given responsibility for carrying out GPRA are the Director of OMB and the agency heads.

    -- Agencies should not create a new GPRA bureaucracy. Agencies are admonished not to 'reinvent the wheel', but to use and apply, to the extent appropriate, existing processes and products when developing the plans and reports required by GPRA.

    -- Plans and reports are to be brief and concise. Hectares of forest should not be cut to prepare voluminous piles of paper, and which only intimidate the prospective reader.

    -- Federal agencies do not generally require a substantial investment in new systems or procedures to manage programs effectively. To agencies seeking investment funding for this purpose, two questions should be raised: have managers regularly examined how their programs were doing, and acted on this information? and, what measures are they using for this purpose? Few agencies are likely to concede that either their managers have never managed, or used a fundamentally flawed set of measures in managing. While there are legitimate needs to improve and expand the set of measures, or to collect better data more quickly, these should be a small fraction of current agency budgets.

    -- The Federal budget traditionally includes descriptive material showing the relationship between spending levels and program activity or achievement levels. However, this is not performance budgeting in the GPRA context. GPRA requires a third set of pilot projects (during fiscal years 1998 and 1999) which test performance budgeting. In these pilot projects, optimization analyses will be done, presenting choices and tradeoffs between different levels of performance for the same or different budget levels. Cost accounting data will be needed. Because of substantial uncertainty about how well the performance budgeting pilot projects will work, or interest in having budgets crafted in such a fashion, GPRA defers until the next century a decision on whether the government should produce a budget based on these performance budgeting principles.

    Strategic Plans

27. GPRA requires that an agency's strategic plan contain the following elements:

    -- a comprehensive mission statement;

    -- a description of general goals and objectives and how these will be achieved;

    -- a description of the relationship between performance goals in the annual performance plan and the general goals and objectives in the strategic plan;

    -- an identification of key external factors that could affect achievement of the general goals and objectives; and

    -- a description of program evaluations used, and a schedule for future evaluations.

28. Strategic plans cover a minimum of six years, and are to be revised and updated at least every three years.

29. Strategic plans provide the foundation for carrying out all other GPRA requirements. The strategic plans:

    -- state an agency purpose: why an agency and its programs exist, and what will be accomplished and when.

    -- define the long-range course of the agency with sufficient precision to guide the short-term actions of agency managers.

30. GPRA does not specify that a government-wide strategic plan be prepared, nor are multi-agency plans for cross-cutting functions envisioned. The absence of such plans dictates that appropriate coordination be done among those agencies whose activities and operations form part of a larger inter-agency program. In these instances, the objective is to assure that individual agency goals are synchronized and harmonized with related goals in the plans of other agencies. The agencies are working jointly to develop a common, consistent approach in setting goals for the inter-agency programs. Joint groups already started cover areas such as trade, research and development, land and resources management, and credit programs.

31. Development of strategic plans should be used as the time for broad discussion about the agency's future programmatic direction and priorities. Consultation with Congress is required during plan development, and agencies are also to seek the views of other interested or potentially affected parties. By minimizing, if not avoiding, a recurring annual debate about program direction and purpose, program managers are given a stable basis from which to set specific annual goals in the performance plans. Strategic plans should not be a chosen means for raising basic policy questions, such as altering eligibility requirements for beneficiaries.

32. While full agreement or general consensus on the strategic plan content is to be hoped for, this may not always happen. Agencies may have to make difficult choices from among competing or differing views. To underscore both the importance of consultation and to record that all relevant and germane views were considered, agencies are to summarize the consultation that occurred and any contrary views received during plan preparation.

33. Some agencies have years of experience in preparing strategic plans. Others have yet to write their first. But no previously prepared strategic plan has yet been identified as meeting all GPRA requirements. While several existing plans already appear to meet some of the required GPRA elements, such as a mission statement and a well-defined set of general goals and objectives, what generally is lacking is coverage of external factors, the relationship between the strategic plan and the annual performance plan, and satisfying the required consultation with Congress. The experienced agencies are likely to re-examine their role and missions in a fundamental way when developing GPRA strategic plans, and not simply re-package existing plans.

    Annual Performance Plans

34. An annual performance plan contains the following elements:

    -- one or more performance goals for each of the programs and operations covered in the plan. (A performance goal is a target level of performance expressed as a tangible, measurable objective against which actual achievement can be compared, including a goal expressed as a quantitative standard, value, or rate. Also, a plan need cover only the major programs and operations in an agency.)

    -- performance indicators that will be used in measuring outputs and outcomes. (A performance indicator is a particular value or characteristic, and is used when a goal is not self-measuring.)

    -- a description of the means to be used to verify and validate measured values.

    -- a brief description of the operational processes, skills, and technology, and the human, capital, information, or other resources required to meet the performance goals.

    -- any proposed waivers of administrative requirements and controls to give managers greater flexibility.

35. In choosing the goals and indicators to be included in the plan, agencies should be guided by the following principles:

    -- Goals and indicators should primarily be those used by program managers to determine how well a program or activity is doing in achieving its intended objectives.

    -- Include measures that will be useful to agency heads and other stakeholders in framing an assessment of what the program or activity is accomplishing.

36. Performance goals in the annual performance plan are derived from the general goals in the strategic plan. As such, a performance plan defines the incremental progress in a particular year toward achieving or sustaining the strategic plan's goals. Often, the goals in a performance plan will be outputs, while the general goals are more likely to be outcomes.

37. Many of the general goals in a strategic plan will not be fully achieved until years in the future. Quite probably, because of shifting tenure and turnover, many of those setting these goals will not be in positions of direct responsibility when achievement ultimately occurs. The 12 month timespan of the annual performance plan allows the government to fix accountability for one year's progress and achievement on individuals and organizations with the reasonable expectation that managers will hold their position over this period.

38. Performance measurement pilot projects are being used to demonstrate whether agencies can prepare annual performance plans that meet GPRA requirements, and to identify any difficulties agencies are encountering in plan preparation. (There are no pilot projects for GPRA strategic plans.)

39. Pilot projects were officially designated in all 14 Cabinet departments and an equivalent number of independent agencies. Some agencies have more than one pilot project, and over 70 individual pilot projects are covered by these designations. In some instances, agencies have supplemented the designated list with their own informal or internal pilot projects.

40. The pilot projects range from the very large -- the entirety of the Social Security Administration and the Internal Revenue Service -- to the very small. As pilot projects are a learning experience for the agencies, few restrictions or limits were put on who could be designated. Nearly 450,000 civilian and military employees are covered by these pilot projects. While this is a large scale endeavor, a more critical facet of the designations is that the pilot projects cover all but two of the 30 major functions of government. The two functions not covered are direct delivery of medical and nursing care, and electric power generation and distribution.

41. Depending on when they were designated, the pilot projects have submitted up to three separate annual performance plans. With the final set of plans submitted in the Spring of 1995, assessments are now being done on both the agency experience and the overall quality of these plans. Some initial impressions:

    -- Enough pilot projects were able to produce performance plans of sufficient quality and covering a variety of functions to prove that good plans can be produced across the government. Expectations are that a fourth or more of the individual pilot projects will be selected as illustrative examples of promising practices. The examples are not perfect plans, but a selection of parts from the better plans which other agencies might consider adapting or emulating in their own plans.

    -- A significant number of pilot projects encountered substantial problems. About ten percent of the pilot projects were eventually withdrawn. Reasons for withdrawal included inherent difficulty in defining program goals, reorganization of the program or staff, and the failure of other parties (outside the agency) to produce the performance information that was needed to prepare the plans and the program performance reports.

    -- Many of the plans contained weak performance measures. These measures included statements purporting to be goals, but which were no more than inherently unmeasurable declarations; goals where performance could not be calculated or determined because one or more values were omitted (much like a formula with only one integer); goals which described the procedural steps being taken to set future goals and measure performance; and goals that were unrelated to the main purposes of a program.

42. A concern that the poorer plans might be indicative of future plan quality when plans are submitted 'for real' in 1997 has prompted OMB to take certain steps. These actions include:

    -- an orientation for all OMB professional staff on performance measurement, to better prepare these staff in assisting agencies in selecting appropriate and useful performance measures.

    -- an accelerated effort to have agencies develop and provide more and better performance information with the fiscal year 1997 and 1998 budgets. In the Spring of 1995, OMB conducted a Spring Review on performance in which the availability and value of performance information for some key government programs was assessed.

    -- giving consideration to having the fiscal year 1998 budget process be a 'dry run' for fiscal year 1999, when GPRA requirements take effect. Agencies have been alerted to expect early calls for providing some parts of the strategic plans in the Summer of 1996, and full descriptions of the performance measures agencies expect to use in their fiscal year 1999 performance plans may be submitted with the fiscal year 1998 budget. This timetable gives OMB and the agencies at least a year to forge agreement on the measures that will be used in fiscal year 1999.

    -- renewing the emphasis on agencies quickly beginning to collect current performance data so that baseline and trend information is available when the specific performance targets for fiscal year 1999 are set. Without baseline information, calculating the planned performance levels can be little more than guesswork.

    Annual Program Performance Reports

43. The annual program performance report should contain the following elements:

    -- a comparison of the actual performance achieved with the performance level(s) specified for each performance goal and performance indicator in the annual performance plan.

    -- if a performance goal was not met, an explanation of why the goal was not met, along with either: (a) the plans and schedules for achieving the performance goal in the future, or (b) a statement that the performance goal as established is impractical or infeasible, and expressing the agency's intention to modify or discontinue the goal.

    -- the summary findings of any program evaluations completed during the fiscal year.

    -- a description of the use, and an assessment of the effectiveness, of any waiver of administrative requirements and controls in achieving performance goals. (This is the managerial accountability and flexibility portion of the report.)

44. By completing one cycle of planning, doing, measuring, and evaluating, and helping begin a new cycle, program performance reports close the loop. What programs (and managers) achieved or failed to achieve is factored into the next round of plans, both performance plans and strategic plans. The adjustments, revisions, and corrections that flow from information in an annual report renew the plans, making them current and real. The report's retrospective scrutiny of what was actually achieved also serves as a self-correcting device. Goals found to be too ambitious will be lowered, and goals which consistently underestimate what really can be accomplished will be raised.

45. As the reports are sent to the President and Congress, and are available to public, accountability and performance will become quite visible, -- perhaps uncomfortably so.

46. The performance measurement pilot projects are also being used to demonstrate whether agencies can prepare annual program performance reports that meet GPRA requirements, and to identify any difficulties agencies are encountering in preparing these reports.

47. Pilot projects that prepared a performance plan for fiscal year 1994 have submitted program performance report for that year. These reports are currently being reviewed. In most instances, the quality of the report reflects the quality of the plan. In some instances, agencies are able to report actual performance even though performance goals in the performance plan lacked a measurable value.

    Managerial Accountability and Flexibility

48. The conceptual entirety of GPRA is fused from three basic elements: (1) strategic planning, (2) performance measurement (which covers goal-setting, measurement, and reporting), and, (3) greater flexibility for managers in return for greater accountability for results.

49. This third element -- managerial flexibility -- is introduced by waiving selected administrative requirements and controls. These administrative and procedural requirements are prescribed by agency rules and directives. (GPRA gives no new authority to waive administrative requirements established in law.) Requirements and controls that may be waived are those imposed across the government by an agency which has a central management role, and cover areas such as financial management, personnel, supply, buildings, etc. The Department of the Treasury, the General Services Administration, OMB, and the Office of Personnel Management are considered to be central management agencies.

50. A key conclusion reached after examining the experiences of other governments with reforms similar to GPRA is that incentives are essential if managers are to change how they manage and become more accountable. Fewer rules, fewer controls, and more latitude in how one uses and directs resources and organizations to accomplish program and policy goals were seen as key incentives. Recognition of superior performance through individual financial reward is another. (Imposing a financial penalty for poor performance is more difficult under the current system.)

51. While the managerial flexibility element of GPRA attempts to mirror what other countries are doing, a manager's autonomy under GPRA is likely to be less than the discretion given his or her counterparts in some other countries, where they can decide how and on what to spend on running costs, or to set compensation rates.

52. A set of managerial accountability and flexibility pilot projects are to test how important this third element could be in transforming how agencies are managed and programs are administered. No pilot projects have yet been chosen, but review of eight nominated pilot projects and their requested waivers will be completed in October 1995.

53. On the whole, the eight nominations raise uncertainty about whether the pilot projects will fully test the managerial accountability and flexibility concept.

54. The nominations are a mixed lot. Several nominations are very narrow, seeking only to waive a single requirement, some are a mixture of potentially significant changes and requirements which are more an annoyance than a substantive burden. Some prospective waivers have been overtaken by time (the requirements to be waived no longer exist, see the reference to the National Performance Review below). Other requested waivers are tangential, at best, to the notion of managerial flexibility.

55. While it is too early to draw any conclusions, some speculative explanations have been voiced about why this set of pilot projects is turning out this way. These include:

    -- The major downsizing of the Federal workforce. Still underway, this downsizing may have made agencies uncertain about whether relief from limits on how many staff an agency could employ was appropriate or needed.

    -- A preemptive de-regulation by central management agencies. GPRA was enacted prior to completing the first phase of the National Performance Review. Several agencies, most particularly OPM, have cut substantially the requirements imposed on other agencies. One can't waive what is no longer required.

    -- Agencies may have discovered that the most burdensome requirements and controls are self-inflicted, and not imposed by central management agencies.

    -- Perhaps, the Federal Government may be an exception to the prevailing principle that managers need incentives, such as flexibility, if they are to be more accountable for results. Or perhaps, it is still too early for most managers to determine what they would like in terms of added flexibility.

    -- Some agencies may have found that developing good performance measurement pilots was task enough, and the managerial flexibility pilot projects were overloading their capacity at this point in time.

    PART 3: GENERAL OBSERVATIONS

    Some general findings

56. The concept and basic structure of GPRA appears to be sound. No fundamental flaws have yet been discovered. This should not be surprising, because GPRA concepts are neither new nor unproven.

57. Enough agencies are making enough progress to indicate that GPRA can be done. However, GPRA is still in a formative stage, and the reality of what this law actually means for agencies, managers, the Congress, and the public is yet undetermined.

58. A few agencies are beginning to introduce new standards for evaluating the performance of senior managers and officials. These new standards place much greater weight on how well managers are managing their programs, and on what programs achieve. Changes in how managers are evaluated and rewarded is critical to affecting any change in how they manage.

59. The large-scale cutting of red tape and unnecessary rules by the National Performance Review is forming a basic underpinning for managerial accountability and flexibility. Personnel and procurement are two areas where major reductions are occurring.

60. Scaling up to government-wide implementation is not that great a mountain to climb.

    -- About 25 percent of the Executive branch employees are covered by the performance measurement pilot projects when Postal Service personnel and members of the armed forces are removed from the calculation.

    Some general expectations

61. GPRA will be successful by disappearing! The transformation set in course by GPRA will be completed when agencies and managers carry out its principles and practices as every-day routine, and not because a law requires them to do so.

62. Implementation of GPRA should and will be uneven across the agencies. If GPRA is easily or perfectly carried out everywhere, then this Act has brought no great change. As a rising tide lifts all boats, time, experience, and comparison should raise the quality of all agency plans and reports to a satisfactory level.

63. Perfection is the enemy of the good. Agencies should strive to do the best they can now. To delay until an elegant design or process is put in place, or until all unknowns are explained, simply means nothing will ever happen. Public and Congressional expectations about the value and usefulness of plans and reports are likely to be modest at first, and grow as better examples appear. Substantive plans and reports, even if incomplete, will likely fare better in these forums than those which are seen as without substance.

64. There is no risk without failure. Managers and agencies will make mistakes. Not all strategies will work; every program will not succeed. GPRA blends accountability with entrepreneurship. Without some tolerance for errors, managers will be risk-avoiders. A natural tendency to focus on faults and failings must be capped. Some mistakes -- not all -- should be forgiven. Some may lead to changes in managers. A few mistakes will lead to corrective steps within the systems and processes. If the answer to every failure is to place more controls and bounds within the management structure, then the managers remaining in government will be those who view their sole responsibility to be that of following rules.

65. Collectively, managers must have a sense of stewardship for their programs. (In other countries, this proprietary interest may be termed ownership.) GPRA rides on the shoulders of many thousands of Federal managers, on their conclusion that what GPRA seeks to do is a help, not a hindrance. GPRA brings fundamental change, and change is not always welcome. Some managers will be uncomfortable, others may resist, but initial indications are that many managers view GPRA as a challenge and opportunity. GPRA probably asks more from them, but also allows them to demonstrate their managerial skill and capacity, and to highlight the value and worth of the programs and operations they oversee and direct.

66. Agencies and OMB must link GPRA plans and reports to budget formulation and execution processes in a decisive way. OMB's recently issued instructions on the preparation and submission of strategic plans underscore this principle.

    Some general concerns.

67. An attempt to make GPRA a budgetary ax would confirm the worst fears of agencies and managers that this Act is not about changing how programs are managed, but about fewer programs and less money. Certainly, as program performance is reviewed, programs that achieve little or nothing may be eliminated. But not always. With changes, some failing programs can succeed. With more funding, underperforming programs may do much better. Managers should believe that the evaluative judgments arising from GPRA will produce remedies as well as terminations.

68. Managers are frequently reluctant to be accountable for what they do not control. In many Federal programs, the States, local governments, or other parties administer the programs and provide the services. (This year, Congress may lessen further the requirements and standards that Federal agencies can establish for some of these programs.) Actions of these other entities can lie beyond the span of control of a Federal manager. If not within a manager's span of control, such programs are usually within his or her span of influence. The general goals in a strategic plan will cover programs and activities within both span of control and span of influence. The bridge not yet done is how to bring performance for span of influence activities into the annual performance plans. The performance partnerships mentioned earlier can be one means, and several of the performance measurement pilot projects are joint Federal-State efforts. In these pilot projects, the States are setting goals for their own performance. Collaboration and cooperation will be key. A complicating concern is a perception that this bridge might create a future national, performance-based management system of gigantic proportions, involving and affecting Federal agencies, States, local governments, and others. There are no plans for such a system; indeed, it would be resisted as infeasible if not impractical. More critically, the idea that Federal managers would, in some manner, set performance levels, such as service standards, for State and local governments is untenable politically. The prevailing view in that Federal mandates should decrease, and mandates that are not Federally funded should disappear.

69. A lack of Congressional interest is likely to be fatal. What is not used eventually is no longer produced. If and how Congress will use GPRA plans and reports is a test to come.

70. Past, current, or future program performance must become a factor in funding choices and decisions. If the government or public view is that simply spending money on a program or problem is performance enough, than the effectiveness or efficiency of programs (their real performance) is of little consequence.

71. Since GPRA became a law, other performance-related initiatives have gotten underway. Others began earlier or loom on the horizon. While this might become too much of a good thing in this still-formative stage of GPRA implementation, the greater peril is that many separate, parallel efforts could fragment organizational responsibilities and staff, duplicate products, waste effort and time, and confuse. As Benjamin Franklin, America's emissary to France during the War for American Independence, stated, "We must all hang together, or assuredly we shall all hang separately." OMB is developing an overall framework that would consolidate various plans and reports and link these with processes for forming and executing the budget. This framework could possibly encompass a dozen or more of the current performance-related efforts, including GPRA and other initiatives (such as financial management and those of the National Performance Review previously referenced).