Such objections, in my view, stem from a basic misunderstanding of the purpose of collecting performance data and of bringing those data into the budget process. The misunderstanding has been encouraged by many advocates of reform, including elected officials. These supporters sometimes claim that having more performance information will make budgeting easier, as if there is some magic algorithm—some automatic connection between performance data, on the one hand, and budget allocations, on the other. The point of more and better performance data is not to establish some automatic connection, but to make the budget process—an inherently political process—more informed by considerations of performance. For this reason, I prefer to use the more accurate and less ambitious term performance-informed budgeting in place of the more grandiose, and less precise, performance-based budgeting.
This discussion relies on a paper written for the IBM Center for Business and Government in 2004 entitled Linking Performance and Budgeting: Opportunities in the Federal Budget Process (Download joyce_report1.pdf). What appears below is a summary of the basic argument of the paper, and a brief update on the practice of performance-informed budgeting at the federal level in the United States, on the occasion of the presentation of the last budget proposal by the current President, George W. Bush. In March, I will write a separate blog presented the results of a nationwide study of performance-informed budgeting in the 50 state governments.
A Framework for Performance-Informed Budgeting
In democratic governments, budget cycles have four stages:
- preparation (usually the development of a budget proposal by a chief executive);
- approval (ratification by the legally responsible body, such as a parliament);
- execution (the implementation of the budget by agencies or ministries); and
- audit/evaluation (the review of actions after the fact).
If the budget process is to become more informed by performance, such a transformation from traditional budgeting involves simultaneously considering two factors at each state of this process
- the availability of appropriate information—on strategic direction, results and costs—in order to make budgeting more results-focused
- the actual use of that information to make budget decisions.
Many discussions of performance-informed budgeting fall short precisely because they focus on a portion of the process. In particular, it is common to see such reforms identified as “failures” because performance information is not seen as having an important influence on the proposed budget by the chief executive, or on decisions to allocate resources by a legislative body. This is an overly narrow view of the budget process, and impedes our ability to successfully study and articulate the many possible situations where budget and performance information can and should be integrated. Looking at the full budget process enables one to recognize that there are important questions to be asked regarding the availability and use of performance information at each stage of the traditional budget process.
The preoccupation with elected officials is consistent with a view that policymaking in the formal sense involves only those officials (Congress and the President, in the U.S.). It fails, however, to acknowledge the informal use of discretion—which is also policymaking—that occurs in agencies or line ministries. A given government or agency might have or make use of performance information at one stage of the process, independent of what might happen at other stages of the process. For example, agencies might make substantial use of performance information in building the budget, while other actors (central budget offices, legislatures) make little or no use of that information at subsequent stages. Conversely, the absence of performance concerns in preparation and approval would not prevent a given agency from using its discretion to execute its budget by considering the effects of different execution strategies on its goals and objectives (i.e. applying outcome measures).
Current Practice in the National Government in the United States
The current efforts to better integrate performance information into the budget process occur in the context of many such reforms that have been attempted in the past. In the United States, this began as early as the 1950s, when “performance budgeting” encouraged more attention to outputs and their relationship to inputs. It continued in the 1960s, with attention to “program budgeting”, which focused for the first time on outcomes, or results. The results orientation has continued, in fits and starts, to the present day, through such reforms as zero-based budgeting (from the 1970s), management by objectives (1980s) and the Government Performance and Results Act and various other legislative and administrative reforms (1990s).
While each of these reforms may have failed to deliver the full range of benefits promised, they can (viewed through the lens of history) be seen as part of a general upward trend in attention to performance concerns throughout the entire 20th century—each reform taught us things, developed capacity and made it more likely that future reforms would be implemented. Most recently, the 1990s reforms laid the groundwork that was present when the Bush administration took office, and are demonstrative of the widespread and deep commitment of the federal government to the concept of performance management. These reforms have, as of yet, been less successful in integrating the use of performance information into government decision processes. It is this shortcoming that is central to the Bush administration’s management and budgeting initiatives.
The connection of resources and results has been a major concern, at least in terms of domestic spending programs, throughout the Bush administration. This has manifested itself in two separate management reforms. The first, called “Budget and Performance Integration” (BPI) is one of five pillars of the Bush management agenda (the others focus on financial management, human resources management, electronic government, and contracting out). The second, the Program Assessment Rating Tool (PART), is an attempt to evaluate 1000 federal programs to determine the extent to which they achieve their desired results.
President's Management Agenda
Under the President’s Management Agenda (PMA), federal agencies are evaluated according to an established set of criteria, and are evaluated four times per year and scored (according to a “traffic light” system, where agencies are awarded “green”, “yellow” and “red”, based on the extent to which they meet the established criteria). For BPI, a “green” score involves taking all of the following actions:
- Senior agency managers meet at least quarterly to examine reports that integrate financial and performance information, and use this information to make decisions.
- Agencies have strategic plans containing a limited number of outcome oriented objectives, and these are consistent with these reports and performance reviews (the PART).
- Full budgetary cost charged to mission accounts/activities; marginal cost of changing performance goals is accurately estimated.
- Agency uses PART evaluations to direct program improvements and PART ratings used to justify funding request
Despite the stringency of the standards, there is substantial evidence of progress. While in the FY03 budget, 23 of 26 agencies were graded as “red”, while only 3 agencies (EPA, DOT, and SBA) even achieved “yellow” status (suggesting compliance with some, but not all standards). In the most recent evaluation there was substantial improvement. By the first quarter of FY08, there were 14 “greens”, 12 “yellows”, and no agency with a “red” score.
PART
The Bush administration’s other government-wide performance-based initiative, the Program Assessment Rating Tool (PART) was first unveiled for use in the fiscal year 2004 budget process. The PART takes the “program” as the unit of analysis, and attempts to determine whether programs are successful in meeting their stated objectives. OMB has defined approximately 1000 programs throughout the federal government. Over five years, OMB reviewed each of these 1000 programs and provided each a score in terms of the extent to which the program was successful.
The PART is a menu-driven device that attempts to evaluate all programs according to a consistent set of criteria. The programs are evaluated according to program purpose and design; strategic planning; program management; and program results and accountability. This fourth area is the most crucial, as it accounts for fully 50 percent of the PART “score”. The PART reviews are to inform budget decisions (at least in the executive branch) and should feed into actions and proposals designed to improve performance.
Perhaps the most significant finding in the first round of PART evaluations was that more than one-half of the programs reviewed did not have the data necessary to demonstrate whether or not they were successful in achieving results. It is important to note that this does not mean that these programs were not effective; rather, it means that they could not prove their effectiveness. By the FY08 budget, however, this number had shrunk to 22 percent. Further, it is clear that, at least at the executive level, these ratings have been used to justify funding requests. For the FY08 budget, there was a clear relationship between PART scores and funding requests. By and large, the higher the score, the higher the percentage increase for the program. On the other hand, the Congress has either ignored the PART or has been openly hostile to it. This is primarily because there are many powerful members of Congress who see the attention to performance as a potential threat to their ability to control the allocation of resources, and particularly to direct those resources based on constituent concerns (that is, largely based on WHERE the inputs flow, rather than WHAT the program has accomplished).
At the same time, the attention to results at the level of the agency or program has carried with it a recognition that, despite the hesitancy of the Congress to embrace a connection between performance and resource allocation, there are many ways that agencies can used the discretion provided to them under law. In the use of this discretion—discretion to make grants, to allocate resources to geographic subunits, to contract out for the production of services, or to manage employees—agencies can consider performance in the management of resources (in the allocation of resources in budget execution). Further, agencies have been expanding their capacity to evaluate programs, at the same time that evaluative organizations—such as the Government Accountability Office—continue to focus on performance as an important component of their reviews of agency practices.
Conclusion
President Bush has just submitted his last budget. What is the current state of performance-informed budgeting, and what does the future hold? I think it is fair to say that, at least in terms of the integration of budget and performance, the Bush administration has left the federal government better than it found it. It took the groundwork that had been laid by prior administrations and expanded upon it. Much of the effort prior to the current administration had involved the production of more and better performance information. The Bush agenda built on this to attempt reforms that focused on the use of performance data. There have been some successes, as evidenced by the apparent progress of agencies on the management scorecard.
The PART, on the other hand, has been somewhat of a mixed bag. On the one hand, it may have encouraged agencies to collect information on performance in order to protect themselves from budget cuts. In addition to the lukewarm reception from the Congress, however, there have been other criticisms of the PART. First, critics argue that the “one size fits all” character of the PART questionnaire reduces program evaluation to a rote, formula-driven, enterprise. Second, some have charged that the PART is designed to establish an apparently “scientific” basis for what the administration wanted to do anyway—which is to cut the budgets of domestic programs with which it does not agree.
What is likely to happen in the next administration? The specifics depend almost entirely on who the new President is. Each of the four viable candidates this writing may have a different approach to management. Only one—Mitt Romney—has specific experience as a government executive. Senator John McCain has implicitly embraced the mantra of performance through his long-opposition to Congressional earmarks. If a Democrat—Senator Hillary Clinton or Senator Barack Obama--is elected, it seems likely that he or she will distance themselves from the specific reforms of the Bush administration, especially the PART. But just as the Bush administration took the next logical step—trying to use performance data—after the Clinton administration had focused on increasing the production of these data, it seems that the next administration is likely to try to put its own brand on the continued effort to better link the budget and performance.
