71409 The World Bank PREMnotes July 2012 NUMBER 19 Special Series on Performance-Informed Budgeting in the U.S. National Government: An Evolutionary Approach and a Work in Progress Philip G. Joyce The United States, at the national level of government, has been trying to identify stronger links between performance and funding for at least 50 years. The most recent two presidents had fundamentally different approaches to performance-based reforms. The administration of George W. Bush embraced a top-down, comprehensive approach to performance, embodied by the President’s Management Agenda and the Program Assessment Rating Tool (PART). The Obama administration has delegated more of the agenda to the agencies and has abandoned the PART in favor of a more in-depth, targeted approach to evaluation. Continuing challenges in the United States include creating incentives for focusing on the long term rather than the short term, making expanded use of performance information for budget decision making, and simultaneously focusing on performance improvement and reducing unsustainable budget deficits. For at least 50 years, the United States has been preparation of the budget by departments and attempting to establish stronger links between agencies, culminating in the president’s budget; indicators of program success or failure and fund- approval of the budget by the Congress; budget ex- ing. Particularly since 1993, when new legislation ecution as programs and policies are implemented; set the stage for the development of more perfor- and audit and evaluation after implementation. mance measures, and throughout the Clinton, Performance information can be produced and Bush and Obama administrations, the United used at any of these stages of the budget process. States has been emphasizing the production and Thus, performance-informed budgeting occurs when use of performance information in the budget pro- performance information is present and used cess. Program evaluation also has a long history in when budget decisions are made and evaluated. the country, starting in the 1960s and continuing This context is important to understanding to the present. George W. Bush’s administration progress that has been made in the United States, (hereafter the Bush administration) emphasized since most analyses that have been conducted on comprehensiveness in performance measurement efforts to link performance information and the and program evaluation. The Obama administra- budget process have reached a similar conclusion. tion, on the other hand, has preferred that agencies The practice of performance-informed budget- develop a smaller number of key performance ing, these analyses argue, is much more evident measures, and that program evaluations be more in budget execution—that is, in the management in-depth and targeted. of resources—than in the allocation of resources. As is true for most countries, the budget The continuing challenge is in increasing the use process in the United States has many stages: the of performance data to inform governmentwide FROM THE POVERTY REDUCTION AND ECONOMIC MANAGEMENT NETWORK budgetary choices. The odds of success in expand- The national level of government is character- ing the use of performance data will depend on the ized by three equal branches of government—the ability of reformers to develop appropriate incen- executive branch (headed by the president and tives to encourage powerful budget actors to move made up of agencies reporting directly to the presi- away from the status quo and to embrace a more dent), the Congress (made up of two bodies—the results-focused approach for the budget process. House of Representatives and the Senate) and the This research note highlights: judicial branch, which includes the U.S. Supreme • The peculiar political and procedural envi- Court and the lesser federal courts. There are 435 ronment that budget reform needs to navi- members of the House of Representatives, and gate in the United States; their positions (seats) are up for election every • The historical development of performance two years—seats are apportioned to states based budgeting in the 20th century, culminating on population. There are 100 senators, two from in the passage and implementation of 1993 each of the 50 states. Under the U.S. Constitution, Government Performance and Results Act Congress makes the laws (with the approval of the (GPRA); president), the president executes the laws, and • Experience since 2000, particularly with ini- the judiciary interprets the laws (Lowi and Gins- tiatives (the President’s Management Agenda burg 2006). In this fragmented system, divided [PMA] and the PART) in the Bush admin- government has been the norm. There are only istration and developments in the Obama two significant political parties—the Republicans administration, including the passage of the and the Democrats—and in only 12 of the 42 years GPRA Modernization Act in 2010; and between 1970 and 2012 has the same political • Observations concerning the challenges that party controlled the presidency and both houses the United States continues to face in its ef- of Congress. forts to establish a more performance-focused The federal fiscal year runs from October 1 culture for its budget process. to September 30. Federal spending for fiscal year (FY) 2012, which ends September 30, 2012, is The Political and Procedural approximately $3.6 trillion (24 percent of GDP). Context for Performance- This spending is broadly divided into mandatory spending, which consists mainly of legal entitle- Informed Budgeting in ments that do not need annual legislative approval the United States (60 percent of the budget) and discretionary The United States has a federal system of gov- (appropriated) spending, which represents the re- ernment. There is one national government, maining 40 percent (Congressional Budget Office 50 state governments, and almost 90,000 local 2012a). The annual budget process starts with the governments. The governmental system is one of president proposing a budget by the first Monday “shared sovereignty,� where the states, under the in February, which includes recommendations on U.S. Constitution, have independent taxing and taxing and spending (Schick 2008). This budget spending authority. The local governments are results from an executive branch process in which “creatures of the state�: their powers to tax and federal agencies request funding to be included in spend are provided by state constitutions (Lee, the president’s proposed budget. This process of Johnson, and Joyce forthcoming). There are a executive branch review normally takes from six wide variety of performance measurement and to nine months and is managed by the Office of budgeting practices in states and localities. Some Management and Budget (OMB), which reports are quite sophisticated, while others are not. State directly to the president. Congress is not bound performance-budgeting practices in the United in any way by the president’s recommendations. States were quite ably outlined in a previous Once the president’s budget reaches Congress, note in this series (Moynihan 2012). This note Congress has from that point until the end of addresses only the national level, although there September to complete its work on the budget. are many sectors (education, health, environment, Congressional action is largely divided into three transport) where decisions made at the national areas: level have significant implications at the state and • Passage of a budget resolution, which pro- local level. vides the broad framework, including levels 2 PREMNOTE JULY 2012 of revenue and spending for the coming fis- In fact, writing almost 50 years ago, Allen cal year and several future fiscal years. This Schick (1966) characterized budget reform in resolution, because it is establishing rules the United States as divided into three distinct for congressional action, does not have to be phases tied to three functions of budgeting— signed by the president. control, management, and planning. After the • Passage of 12 appropriation bills that finance control-oriented reforms of the early part of the the discretionary portion of the budget. 1900s came the first recorded attempts to bring • Passage of bills affecting revenues and manda- program performance into the budget process tory spending, only if changes to the status with the creation of so-called “performance bud- quo are anticipated. geting� in the 1950s. This management-focused The budget process is fragmented and highly reform attempted to connect spending to the partisan. The end result is that budget reforms outputs purchased, but not to the outcomes that that are embraced by the executive branch fre- resulted. Next, the 1960s brought the advent of quently face tepid or hostile reactions from Con- the planning-programming-budgeting system gress, particularly when the legislature fears that (PPBS), the goal of which was to bring a results its budgetary prerogatives are being threatened. focus into the budget process. The PPBS, which The United States currently has a large and grow- got its start in the Defense Department, could not ing debt. At the end of FY2012, the debt held by be successfully applied to domestic agencies. Fol- the public is projected to exceed $10 trillion, or lowing the implementation (and withdrawal) of almost 68 percent of GDP (Congressional Budget the PPBS, other reform efforts continued through the 1970s, including management by objectives, Office 2012b). All budgetary debates for the fore- which was advocated by the Nixon administra- seeable future will be conducted with that federal tion, and zero-base budgeting, which was the debt—and questions concerning how to deal with preferred reform of the Carter administration it—in the background. The United States also has (Harkin 1982). a chronic problem in meeting budget deadlines; The 1980s saw little in terms of performance only three times since 1977 have all appropriation reforms, but an important milestone of the 1990s bills become law prior to the start of the fiscal year. was the passage of the Government Performance The budget resolution has also failed to be enacted and Results Act (GPRA) in 1993. GPRA differed for 5 of the last 10 fiscal years, including the past from past reforms in one important respect—it was two (FY2012 and FY2013). a law, rather than simply a presidential initiative. It required agencies to: Budget Reform, 1900–2000 • Develop strategic plans outlining long-term Budget reform in the United States has a long goals and describing results expected to be history. Early efforts, which began in state and achieved; local governments in the late 1800s, were largely • Publish performance plans, consistent with focused on attempting to reduce corruption by strategic plans, that would specify both per- introducing budget controls. Practically speak- formance indicators and performance targets, ing, this involved the establishment of line item consistent with planned goals; and budgets and the development of laws aimed at • Report on performance annually, specifically limiting spending, as well as a focus on preaudits by providing data on actual performance rela- and postaudits of spending. At the federal level, tive to the goals and measures established in the highlight of these efforts was the passage of the the strategic and performance plans. Budget and Accounting Act of 1921, which intro- While GPRA’s ultimate stated goal was the duced the requirement of a presidential budget use of performance data in the budget process, its proposal and established the president’s budget main legacy, starting during the Clinton adminis- office (now the OMB). While these efforts were tration through to the present, was an increase in successful in establishing more budget controls, the supply of performance information. The Clin- they did not attempt to have the budget process ton administration also pursued a separate reform focus on what was being purchased with the funds agenda embodied by the National Performance allocated to those line items. Review (later the National Partnership for Rein- JULY 2012 PREMNOTE 3 venting Government), which was spearheaded by agencies were evaluated according to a set of es- Vice President Al Gore (Joyce 2003). tablished criteria on each of these five dimensions using a “traffic light� system, where “green� meant The Bush and Obama that agencies complied with all of the criteria, and Administrations’ Approaches to “yellow� or “red� implied a progressively worse level of performance. Performance-Informed Budgeting In a post mortem, the Bush administration When the George W. Bush administration took of- argued that this system was highly successful judg- fice in early 2001, it argued that GPRA had largely ing by changes in traffic light scores from the begin- failed to change budget decision making, and it ning to the end of the administration, as indicated articulated its own set of management priorities in table 1 (Joyce 2011). The table indicates that, embodied in the President’s Management Agenda while in the baseline evaluation in 2001 only one (PMA), which was published in September of agency received a “green� in any management area 2001 (OMB 2001). This agenda included five (National Science Foundation’s financial perfor- governmentwide management reforms: mance), the number of greens increased to 72 by • Strategic investment in human capital, which 2008. Some agencies apparently made particularly was designed to focus on preparing federal dramatic improvements. The State Department, agencies for the necessary task of replacing for example, which was scored as “red� in each of retiring federal workers; the five categories in 2001, improved to “green� • Competitive sourcing, which was proposed to in every one by 2008. In addition, there were expand the use of contracting in government; three other departments who were all “green� in • Improved financial performance, which 2008—the Environmental Protection Agency, the focused primarily on improving financial Department of Labor, and the Social Security Ad- management, mainly indicated by improved ministration. There are some reasons to question audit results; whether progress was quite this dramatic; agencies • E-government, which attempted to follow may have learned to game the systems somewhat, international trends in improving service and scores were subjective, with no replicable delivery through the use of technological methodology for arriving at the score. Some of resources; and the standards also evolved a bit from beginning to • Budget and performance integration, designed end; it was arguably easier to “get to green� in some to continue to improve performance infor- cases at the end than it had been at the beginning mation while also allocating and managing because the standards had been relaxed. resources to achieve results. The second significant reform initiative of For each of these five management areas, the the Bush administration was the creation of the Bush administration graded 26 agencies using a PART, first unveiled for use in the FY2004 budget scorecard administered by OMB. Each quarter, process. The PART took the “program� as the unit Table 1. Comparison of 2001 and 2008 Scores under the President’s Management Agenda during the Administration of George W. Bush 2001 Distribution 2008 Distribution Increase in Management area Green Yellow Red Green Yellow Red “green� scores Human capital 0 3 23 14 11 1 +14 Competitive sourcinga 0 0 26 12 13 1 +12 Financial performance 1 4 21 16 4 6 +15 E-government 0 9 17 11 9 6 +11 Budget and performance 0 3 23 19 7 0 +19 integrationb Total scores 1 19 110 72 44 14 +71 Source: This table is reprinted from Joyce 2011. a. This was called “commercial services management� in 2008. b. This was called “performance improvement� in 2008. 4 PREMNOTE JULY 2012 of analysis, and evaluated these programs (there attention to cross-cutting approaches to were approximately 1,000 identified) through the government performance, where more than use of a questionnaire that requested information one program, or more than one agency, con- from agencies on characteristics of their programs tributes to the achievement of results (GAO (Mark and Pfeiffer 2011). The PART score for 2004, 2005). It was a departure from GPRA, each program was weighted based on responses to which focused at the agency or department these 25 to 30 questions, which covered: level. • Program purpose and design—are they clear and • While arguably there was some evidence of defensible? (20 percent) use of PART data for decision making in the • Strategic planning—does the agency set valid executive branch (particularly during devel- annual and long-term goals? (10 percent) opment of the president’s budget), it got an • Program management—does the agency apathetic or sometimes hostile response from exercise sound financial management and Congress. Reformers found it difficult to engage in program improvement efforts? produce data that could be used to improve (20 percent) programs and to inform resource allocation • Program results—does the program deliver (GAO 2004, 2005). results based on its goals? (50 percent) • The PART did not adequately account for Agencies filled out the questionnaire, with the differences among different types of their answers reviewed (approved) by OMB. The programs, and in particular, assumed that program was eventually “scored� as falling within programs had measureable outputs and one of five categories: effective (85–100), moder- outcomes. This puts programs conducive ately effective (70–84), adequate (50–69), inef- to quantitative measurement in an advanta- fective (0–49), and results not demonstrated (if a geous position (Gueorguieva et al. 2009). program lacks adequate measures, it falls into this • The PART created a lot of work for both fed- category regardless of its score; Gilmour 2006). eral agencies and OMB. In fact, even though Approximately one-fifth of the 1,000 programs some agencies viewed the PART as a creation were scored each year over the first five years that of OMB, many OMB budget examiners were the PART was in effect. The specific goal of the not really sorry to see it go and questioned PART was to evaluate every program; thus the whether the benefits were really worth all of Bush administration sacrificed evaluation depth the effort (Redburn and Newcomer 2008). to gain a comprehensive view of all programs. President Obama took office in January of As with the PMA, the Bush administration 2009 at one of the most challenging times in had its own evaluation of the PART, which put U.S. history. The Obama performance agenda has the reform in a positive light. Specifically, while focused on at least four separate initiatives over the percentage of ineffective programs stayed the first year and a half: (i) evaluating the effects roughly the same (at 3 to 5 percent) over the seven of economic stimulus, (ii) reducing the deficit, years of the PART, there was a substantial increase (iii) pursuing targeted performance goals, and (iv) in the percentage of programs rated effective or conducting targeted program evaluations. moderately effective between 2002 and 2008. In The American Recovery and Reinvestment 2002, 30 percent of programs were rated effective Act (ARRA, or “stimulus bill�) was passed in early or moderately effective; in 2008, that figure was 2009. The most recent estimate of the total cost of 51 percent. Perhaps as significantly, the results not ARRA puts the figure at approximately $831 bil- demonstrated category decreased by two-thirds, lion in tax reductions and spending cuts designed from 50 percent in 2002 to 17 percent in 2008 to stimulate the economy (CBO 2012a). The main (Bush Administration 2008). goal of the ARRA is to employ individuals who External evaluations of the PART have otherwise might be unemployed because of the been more mixed. Among the more significant recession. The Obama administration is focused, conclusions: therefore, on counting the number of jobs “saved • The PART’s highly decentralized approach or created� and has developed a detailed method- (with the “program� as the unit of analysis) ology for tracking both spending and jobs. encouraged a “stove-piped� view of govern- Second, because the federal budget deficit has ment management, which discouraged exceeded $1 trillion for each of the last three fiscal JULY 2012 PREMNOTE 5 years—2009, 2010, and 2011 (prior to 2009, it of programs and practices, isolating the effect of had never reached even $500 billion)—the Obama Government action from other possible influenc- administration has a stated goal of reducing this ing factors� (OMB 2010, 90–92). budget deficit in the medium term, as part of a The other potentially significant develop- longer-term effort to return deficits to their 40- ment during the Obama administration was the year historical average of 3 percent of GDP. As passage of the GPRA Modernization Act, which part of a demonstration of the administration’s was signed into law by the president in 2010. The desire to cut back on spending, it has identified update of GPRA was heavily informed by evalua- programs in its budgets, including the most recent tions conducted since 2004 by the Government one for FY2013, that it believes should have fund- Accountability Office (GAO), and the update ing reduced or eliminated because of inadequate intended to use agencies’ experiences under the performance. In the context of the overall budget, act to improve its effectiveness. Specifically, the these proposals are relatively modest, adding up new law creates a clearer link between strategic to less than 1 percent of total federal spending. plans, programs, and performance information. It Third, federal agencies were instructed to requires quarterly performance reporting rather establish a small number of “high priority perfor- than the annual reporting that had been the norm mance goals,� reversing the Bush administration’s under the original GPRA. The intent of the law is “top-down� approach to agency performance to to increase program managers’ use of performance an approach where agencies are driving more of information during decision making. It validates, the specific agenda for performance. The original legislatively, what had become a performance emphasis was on goals that could be accomplished management infrastructure that includes chief over 12 to 24 months. The approach started operating officers, program improvement officers, with the identification of a small number of per- a governmentwide performance improvement formance goals for each agency. In the FY2011 council, and a governmentwide performance Web budget, those goals were identified, measures were site, www.performance.gov. The law will be imple- established, and targets were set. Table 2 shows mented over a number of years, and Congressional buy-in will still be necessary to remove significant examples of these goals, measures, and targets for barriers—in legislation and in practice—to program five agencies: the Department of Education, the performance (Kamensky 2011). Department of Homeland Security, the Depart- ment of Veterans Affairs, the National Science Foundation, and the Social Security Administra- Key Issues and Challenges tion (Joyce 2011). After review, it is clear that What to make of all of this activity? It is unques- the vast majority of these goals are specific and tionably true that the past 20 years have seen the measurable. Some of the goals are more ambigu- most sustained attention to performance than any ous (for example, what is a school “system with period in U.S. history. This is manifested in the rigorous processes for determining teacher ef- enactment of two separate laws and three succes- fectiveness�?) than others (it is presumably easy sive presidential administrations that have made for the Social Security Administration to track performance measurement efforts central to their progress toward a goal of achieving “an average management agendas. The U.S. political system, speed of answer of 264 seconds by the national because of its fragmented nature, is resistant to 800-number�). systematic, sustained approaches to performance. Finally, a significant commitment in time and Instead, the tendency is for each new presidential resources has been devoted to program evaluation, administration to put its own stamp on the perfor- in part to assist with the identification of what mance agenda. Thus, the performance approach of works and what does not. The Obama administra- the United States is more appropriately described, tion argues that as important as performance goals as is policy making in general, as “evolutionary� and performance measures can be, performance rather than “revolutionary.� That is, there is no information “can answer only so many questions. single watershed event that created a new “perfor- More sophisticated evaluation methods are re- mance era,� but rather progress occurs as a series of quired to answer fundamental questions about small steps, many of which have different names the social, economic, or environmental impact and requirements. 6 PREMNOTE JULY 2012 Table 2. Examples of High-Priority Performance Goals in President Obama’s FY11 Proposed Budget Department or agency Performance goal Performance target Department of • Evidence-based policy: Imple- • Increase by 2/3 the number of Department Education mentation of a comprehensive discretionary programs that use evaluation, approach to using evidence to performance measures and other program data inform the Department’s policies for continuous improvement and major initiatives • Improve the quality of teaching and • Increasing by 200,000 the number of teachers learning for low income and minority students who are being recruited or retained to teach in hard- to-staff subjects and schools in systems with rigorous processes for determining teacher effectiveness. Department of • Countering terrorism and enhanc- • Wait times for aviation passengers (Target: Less Homeland Security ing security than 20 minutes by 2012). • Improve the efficiency of the pro- • Decrease the number of days spent in custody cess to detain and remove illegal by criminal aliens before they are removed from immigrants from the United States the United States from 43 to 41 days in 2010. Department of • Eliminate veterans homelessness • In conjunction with HUD, reduce the homeless Veterans Affairs veteran population to 59,000 by June 2012 on the way to eliminating veteran homelessness • Improve the quality, access, and • By the end of 2011, 97 percent of eligible pa- value of mental health care pro- tients will be screened at required intervals for vided to veterans by December post-traumatic stress disorder. 2011 National Science Improve the education and train- • By the end of 2011, at least six major NSF Foundation [NSF] ing of a Science, Technology, STEM workforce development programs at the Engineering, and Mathemat- graduate/postdoctoral level have evaluation ics (STEM) workforce through and assessment systems providing findings evidence-based approaches that leading to program re-design or consolidation includes collection and analysis for more strategic impact in developing STEM of performance data, program workforce problem solvers, entrepreneurs, or evaluation and other research. innovators. Social Security • Increase the number of online • By 2012, achieve an online filing rate of 50 Administration applications percent for retirement applications. .Achieve [SSA] 44 percentage of total retirement claims filed online in 2011. • Improve SSA’s customers’ service • Achieve an average speed of answer of 264 experience on the telephone, in seconds by the national 800-number. field offices, and online Source: OMB (2010), reprinted from Joyce (2011). This evolutionary approach has led some to of a general upward trend in attention to per- view each reform as a failure when it is invariably formance concerns throughout the entire 20th replaced by a new reform—with a new name—by a century—each reform developed capacity and subsequent administration. This seems an overly made it more likely that future reforms would be harsh conclusion. Analyses have frequently not implemented. It is fair to say, for example, that been conducted at enough distance (in terms of the reform efforts of the Bush administration, time or perspective) from the “reform� to permit while reflecting its own specific approach and a real evaluation of effects. Viewed through the values, were built on the foundation that preceded lens of history, these reforms can be seen as part it, particularly GPRA. By the same token, even JULY 2012 PREMNOTE 7 though the Obama administration jettisoned the at times (this was particularly true for the Bush performance initiatives, the Obama reform PART). This has occurred despite the leg- built on experience with those reforms. There are islative underpinning of GPRA; there is, in sometimes significant ideologically differences fact, very little knowledge of the law or the between administrations, however. Most recently, successor Modernization Act in Congress. this was reflected in the Bush administration’s It is very difficult for decision makers in embracing of private sector practices, an approach Congress, who have a vested interest in that was deemphasized by both the Clinton and the status quo, to accept such reforms as Obama administrations. legitimate. Further, no broad consensus While many specific observations can be exists on how to measure performance and made about the experience of the U.S. federal use that information for budgeting. Part of government with performance management, in this is certainly influenced by the fact that the interest of space, this note focuses on eight performance data have not been reported conclusions: by the executive branch to Congress in a (i) The fragmented nature of the U.S. political way that members of Congress or congres- system has encouraged a fragmented ap- sional staff find useful. Long, detailed per- proach to performance. The most important formance reports are much more likely to budgetary decisions, on an annual basis, collect dust on congressional shelves than are those made in the appropriations pro- to be used to inform decision making. cess. This process, however, involves the (iii) Despite the lack of congressional attention, consideration of many different pieces substantial progress has been made at the of legislation, with no requirements for agency level, particularly in the production of any systematic consideration of perfor- performance data, but also in the use of these mance. The various executive branch data for management. The most systematic performance reforms do not require any evidence of performance information use specific response by Congress. The two is by executive branch agencies to manage versions of the GPRA were championed resources once they have been received by congressional committees with no sig- from Congress (Joyce 2003). Even a cur- nificant role in the budget process. While sory review of agency reports accessible each of these laws amended the Budget from www.performance.gov (the Web site and Accounting Act of 1921 (relating to created by the Obama administration for the president’s budget proposal), neither performance reporting) uncovers a great had any significant effect on congressio- many high-quality measures and suggests nal budgeting. In fact, the Congressional substantial improvement in the identifica- Budget and Impoundment Control Act tion of appropriate performance measures of 1974, which creates the framework for from what was present when GPRA was congressional budgeting, is largely silent on first enacted in 1993. Further, studies by matters of performance. There have also GAO and others have found a great deal been very limited efforts to bring accrual of evidence that agencies are using per- concepts, as opposed to financial reporting, formance data to make internal resource into budgeting. management decisions, including, but not (ii) Twenty years of sustained attention to per- limited to, decisions concerning the alloca- formance has not encouraged Congress to tion of government grants. systematically use performance data. The (iv) Even within the executive branch, the United separation of powers between the legisla- States has yet to resolve whether it is more use- tive and executive branches in the United ful to take a top-down or bottom-up approach States—and the fragmented political to performance budget reform. The Bush ad- structure—is also reflected in its approach ministration’s PMA and the PART were de- to performance-informed budgeting. cidedly more top down than the approach Congress has paid very little attention to of the Obama administration, which performance when making budget deci- has embraced a more agency-centered sions, instead expressing open hostility approach to goal setting in performance 8 PREMNOTE JULY 2012 measurement. This is a bit more reminis- ward performance monitoring rather than cent of the Clinton administration’s Na- evaluation, which normally has a longer- tional Performance Review (Redburn and term focus. Further, programs without Newcomer 2008). It is an open question, quantitative measures may find themselves however, whether the administration will at a disadvantage in the budget process. be able to sustain its process, and whether (vi) The Obama administration has made trans- it will be clearly superior to a more top- parency a major emphasis, but it remains to down approach. There are a couple of be seen how many advances will in fact be potential pitfalls. First, while the workload made. The transparency of PART, where associated with the PART caused the rank all results could be viewed on the OMB and file at OMB to shed no tears over its Web site, was a surprising development demise, there is some question as to how to long-time observers of OMB. There much OMB will actually let the agencies was substantial analysis of PART results drive the agenda. It is a simple fact that by academics and other outsiders, largely OMB is the only entity within the execu- because they were easily accessible. The tive branch with the capacity to manage a PMA was less criticized, not because it governmentwide process. Further, for all of was inherently better, but because no its advantages in terms of getting potential one outside of the Bush administration buy-in from agencies, the Obama approach had any idea how a given agency got to be may actually put less pressure on the agen- green, yellow, or red. The transparency cies to perform unless OMB imposes its of the Obama effort has to be called into will (at which point, of course, it starts to question, since readers of the FY2012 and look like a top-down process). FY2013 budgets cannot find any actual (v) In U.S. performance-informed budgeting, performance information to date related there continues to be a focus on short-term to the high-priority performance goals. If and quantitative measurement. The Obama progress on a set of 12-to-24 month perfor- administration sent a couple of clear signals mance goals cannot be disclosed to anyone suggesting that delivering results sooner outside the administration one to two years is superior to policies that take longer after they were first unveiled, this is not to deliver. This was clearly the impera- a positive development for transparency tive in the evaluation of stimulus funds, and accountability. While it is true that a where short-term job creation was the lot of information is potentially accessible priority—an understandable focus, as this through www.performance.gov, this is is part and parcel of economic stimulus. more of a “data dump� than a transparency The high-priority performance goals are initiative; it is hard to imagine a citizen be- also focused fairly explicitly on the short ing able to find useful data if he or she did term. This potentially creates a problem not already know where (that is, in which for research and development agencies, or agency publication) to find it. others who are pursuing results (such as (vii) Thus far, the U.S. approach is probably best health outcomes) where the payoff may characterized as “supply focused� rather than come later. It is worth asking whether the “demand driven.� Both the two versions administration has simply been encourag- of GPRA and the various presidential ing agencies to pursue policies that can initiatives, while extolling the benefits demonstrate results in time to catalogue of the use of performance information, them to support the president’s reelection. have been mostly focused on technocratic Further, strong preference continues to be concerns—the production of that informa- expressed for priorities with results that tion. Implicitly, the argument has been that can be measured quantitatively. There are supply will promote demand. There has a great many programs where the quest for been very little attention paid to creating quantitative measurement may miss the the incentives that would promote the use main point of the program. In particular, of information for budgeting. Arguably, such an emphasis may create a tendency to- both the PART and the PMA in the Bush JULY 2012 PREMNOTE 9 administration did create some incentives reduction. This will be particularly true if—as because agency or program “scores� were seems inevitable—some effort is made to rein in influenced by evidence of use. This was spending as part of a broader effort to reduce the largely focused on budget execution; no federal budget deficit. systematic incentives were created for In this context, performance is likely to be budget preparation and approval. used not as part of an effort to improve programs, (viii) The performance agenda must coexist with the but rather as justification for eliminating them. pressures to reduce spending and the deficit. Here, history tells us that the fact that a given Increasingly, performance measures and administration is committed to evidence-based evaluation will be used to identify pos- decision making does not necessarily mean that sible areas of reduction. This could bring more decisions, at least at the macrolevel, will be about the second coming of the debate based on evidence. There may instead be a tempta- that defined early Clinton administration tion to engage in across-the-board (“cheese slicer�) discussions concerning the recommenda- budgeting. Even here, in a context in which reduc- tions of the National Performance Review— ing spending is a priority, weak claims may have were they intended to identify places to a marginally more difficult time being sustained. cut the budget or to create more effective management of programs? The quandary About the Author is as follows—if agencies believe that the Philip Joyce is Professor of Management, Finance, and purpose of the performance agenda is to Leadership at the Maryland School of Public Policy. help them manage more effectively, they He is Editor of Public Budgeting and Finance, and will be more likely to be supportive of the a Fellow of the National Academy of Public Admin- program than if they believe it is just a tool istration. Joyce’s research mainly focuses on two to be used to justify cutting budgets. A issues—links between performance information shift from a management to budget-cutting and the budget, and the U.S. congressional budget emphasis can already be seen between the process. His most recent book is The Congressional FY2011 and later (FY2012 and FY2013) Budget Office: Honest Numbers, Power and Policy- budgets. In the latter, the administration making (2011). He has consulted extensively for acknowledges that it proposed reducing or various organizations, including the International eliminating some programs that it supports Monetary Fund and the World Bank. in the interest of deficit reduction. If this emphasis continues, it will create the real Acknowledgments possibility of a defensive response on the part of federal agencies. The author thanks colleagues Gladys Lopez-Ace- vedo (Senior Economist, PRMPR), Jim Brumby Conclusion (Sector Manager, PRMPS), Keith Mackay (Consul- tant, PRMPR), and Philipp Krause (Budget Unit, For all of the attention paid to performance ODI)for their comments and support. and the resources devoted to its measurement and reporting, the age-old question—will per- References formance information be used for budgeting, and by whom?—has not yet been resolved in the Bush Administration. 2008. “White House, Perfor- mance Improvement: What We’ve Achieved Over United States. Consistently, Congress, or at least the Past Six Years.� Originally published on Bush the committees responsible for funding deci- White House Web site, but no longer available. sions, has tended to be fairly hostile to attempts CBO (U.S. Congressional Budget Office). 2012a. to change the dialogue around the federal bud- The Budget and Economic Outlook: Fiscal Years get. Here, divided government, and the general 2013–2022. Washington, DC: U.S. Government antipathy of conservative Republicans (who are Printing Office. ———. 2012b. Updated Budget Projections: Fiscal Years likely to continue to control at least the House 2012 to 2022. Washington, DC: U.S. Government of Representatives after the 2012 election) to Printing Office. public programs and government spending, may GAO (U.S. Government Accountability Office). 2004. provide some impetus for data-driven budget Performance Budgeting: Observations on the Use 10 PREMNOTE JULY 2012 of OMB’s Program Assessment Rating Tool for the Lee, Robert, Ronald Johnson, and Philip Joyce. Forth- FY2004 Budget. GAO-04-174, Washington, DC. coming. Public Budgeting Systems, 9th Edition. ———. 2005. Performance Budgeting: PART Focuses Atten- Lowi, Theodore, and Benjamin Ginsburg. 2006. tion on Program Performance, But More Can Be Done American Government, 9th Edition. W. W. Norton to Engage Congress. GAO-06-28, Washington, DC. and Company. Gilmour, John. 2006. Implementing OMB’s Program As- Mark, Katherine, and John R. Pfeiffer. 2011. sessment Rating Tool. Washington, DC: IBM Center “Monitoring and Evaluation in the United States for the Business of Government. Government: An Overview.� Evaluation Capac- Gueorguieva, Vassia, Jean Accius, Carmen Apaza, et. ity Development Working Paper Number 26, al. 2009. “The Program Assessment Rating Tool Independent Evaluation Group, World Bank, and the Government Performance and Results Washington, DC. Act: Evaluating Conflicts and Disconnections.� Moynihan, Donald. 2012. “Performance Management American Review of Public Administration 39: 225. in U.S. State Governments.� PREMnote, Special Harkin, J. M. 1982. “Effectiveness Budgeting: The Series on Nuts and Bolts of M&E Systems, World Limits of Budget Reform.� Policy Studies Review 2 Bank, February. (3): 112–26. OMB (U.S. Office of Management and Budget). 2001. Joyce, Philip. 2003. Linking Performance and Budget- The President’s Management Agenda. Washington, DC. ing: Opportunities in the Federal Budget Process. ———. 2010. “Program Evaluation.� In The Budget of the Washington, DC: IBM Center for the Business of United States Government, Fiscal Year 2011: Ana- Government. lytical Perspectives, 90–92. Washington, DC: U.S. ———. 2011. “Transparency and Accountability in the Government Printing Office. Federal Budget: How Is the Obama Administration Redburn, F. Stevens, and Kathryn Newcomer. 2008. Building on the Legacy of Federal Performance- Achieving Real Improvement in Program Performance Informed Budgeting?� Public Administration Review and Policy Outcomes: The Next Frontier. National 71 (3; May/June): 356–67. Academy of Public Administration, October. Kamensky, John. 2011. “GPRA Modernization Act Schick, Allen. 1966. “The Road to PPB: The Stages of Explained, Part 1.� IBM Center for the Business Budget Reform.� Public Administration Review 26 of Government, January 6, http://www.busines- (4): 243–58. sofgovernment.org/blog/business-government/ ———. 2008. The Federal Budget, 3rd Edition. Washing- gpra-modernization-act-2010-explained-part-1. ton, DC: Brookings. This note series is intended to summarize good practices and key policy findings on PREM-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank. PREMnotes are widely distributed to Bank staff and are also available on the PREM Web site (http://www. worldbank.org/prem). If you are interested in writing a PREMnote, email your idea to Madjiguene Seck at mseck@worldbank.org. For additional copies of this PREMnote please contact the PREM Advisory Service at x87736. This series is for both external and internal dissemination JULY 2012 PREMNOTE 11