KeywordsPublic Sector Performance Indicator Performance Goal Performance Information Funding Decision
Performance budgeting refers to a budgeting system that takes into account performance information in budgetary decisions.
Performance budgeting (also called performance-based budgeting, results-oriented budgeting, results-based budgeting) is a rather elastic term that encompasses different ways of linking performance information and budgeting in the public sector (Curristine 2005; Schick 2003, 2007, 2014). Performance information (PI), in turn, refers to “systematic information describing the outputs and outcomes of public programmes and organizations … generated by systems and processes intended to produce such information” (Pollitt 2006, p. 39). Outputs are immediate results of government programs, whereas outcomes refer to broader impacts.
In the broadest sense, the label performance budgeting (PB) is used to describe any arrangement where performance information is added to the budget documents, alongside financial information. More specifically, PB takes three different forms: (1) presentational PB; (2) performance-informed budgeting, and (3) direct performance budgeting (e.g., Curristine 2007; Hawkesworth and Klepsvik 2013). In the case of presentational performance budgeting, performance information is published together with budget documents but is not expected to explicitly be taken into account in funding decisions. Instead, the goal of disseminating PI is to promote accountability and increase transparency in the public sector. In performance-informed budgeting, performance information is systematically taken into account in funding decisions, but the link is not automatic: performance indicators alone do not determine the size of the allocations and their weight depends on the circumstances. Direct performance budgeting refers to a system where the amount of resources allocated to an organization directly depend on the results it has achieved. The most extreme case of it is formula funding where the allocations are calculated on the basis of a formula that takes into account the achieved outcomes or outputs.
The adoption of performance budgeting systems is an important reform trend in the organisation for economic cooperation and development (OECD) countries but also elsewhere (Berry 2007; Curristine 2007; Hawkesworth and Klepsvik 2013; Sterck and Scheers 2006). The pioneer in performance budgeting was New Zealand, which started experimenting with it already in the 1980s. Australia, Canada, Denmark, Finland, the Netherlands, Sweden, the United Kingdom, and the United States then began to adopt the reform to varying degrees during the 1990s (Curristine 2005). Many other countries have followed suit since then. Today, roughly two-thirds of OECD countries have issued guidelines for performance budgeting (Schick 2014). However, we still cannot talk about one uniform model of PB (Schick 2014). Each country has adopted its own version, and the elements of the performance budgeting system vary considerably between governments. Furthermore, even the most “advanced” reformers cannot claim that their PB system is “finally ready.” Instead, they are still in the process of changing their systems (Curristine 2005; Schick 2014).
Its promise to increase the efficiency of resource use makes PB particularly attractive during times of financial strain when reformers are looking for ways to “do more with less” and to cut expenditures in a “rational” way (Hawkesworth and Klepsvik 2013; Pollitt 2010). At the same time, the implementation of PB has often faced a number of difficulties, leading to disappointments and even reform reversals. The goal of this chapter is to give an overview of the advantages of PB, the challenges involved in implementing it, and the potential negative consequences it can lead to.
The Advantages of Performance Budgeting
Performance budgeting was brought onto the public sector reform agenda by the New Public Management movement (Berry 2007; Pollitt and Bouckaert 2004). The core goals of New Public Management have been to increase the economic efficiency and effectiveness of public sector organizations. According to its advocates, performance budgeting has a number of advantages (Curristine 2005; Pollitt 2001; Robinson 2007).
First, it is often argued that PB promotes transparency in the public sector and helps to reveal what is actually achieved with taxpayers’ money (Andrews 2004; Berry 2007; Hawkesworth and Klepsvik 2013). If PB is used, public officials have to communicate their performance targets at the beginning of the budget cycle. This, of course, necessitates the reporting of actual performance and the extent to which performance goals were met once the fiscal year has concluded. Doing so is expected to promote increased awareness of what different government programs actually cost and what benefits they accrue.
Second, PB can help to put pressure on public sector organizations to achieve better results. If the public officials know that the size of their budget depends on the results they deliver, they are motivated to be more results oriented in their activities (Ammons 2002; Robinson 2007).
Third, PB is purported to alleviate problems arising from incrementalism in budgeting and reduce inertia in the funding of programs and policies (Robinson 2007). The incremental approach to budgeting – whereby all organizations covered in the budget are given incremental increases in their budget from year to year – is often criticized for not taking into account whether the programs that are receiving funding are actually effective and deserve continued appropriations.
Fourth, if the PB system entails a change in the budget format from an input-based classification (e.g., personnel, supplies, subsidies, investments, etc.) to an outcome- or output-based one, it can increase flexibility in the use of budgetary resources (Andrews 2004; Hawkesworth and Klepsvik 2013; Robinson 2007). Input-based budgeting (especially if the line items are very detailed) can lead to a waste of resources when public officials spend down the foreseen line items even if priorities or funding needs for different activities have changed. If the appropriations are made for outcomes or outputs, the managers in the organization can make spending decisions with a view to what delivers the best results.
Finally, PB is claimed to increase allocative efficiency in the public sector – meaning that resources are directed to activities that attain the largest benefits (Pollitt 2001; Robinson 2007). Performance information allows public officials to make better informed decisions about funding levels for different activities and adjust resource allocations according to changes in government’s priorities and societal needs.
Challenges of Performance Budgeting
There are at least three sets of challenges that performance budgeting faces: (1) It is difficult to measure performance in the public sector; (2) it is difficult to use performance information for budgetary decisions; (3) budgeting is an inherently political process.
First, an important set of challenges faced by performance budgeting is related to the fact that it is often difficult to measure performance in the public sector in a meaningful way. In collecting performance information, public sector organizations can focus either on outputs, outcomes, or both. Collecting information about outcomes can provide valuable insights about the effects of governmental activities on society. However, a change in an outcome indicator may be influenced by a wide array of external factors that are beyond the control of a governmental organization (De Lancer Julnes 2006). For example, the crime rate is influenced by many things aside from the activities and performance of police departments (e.g., a sudden worsening of the economic environment). Thus, it may be difficult to ascertain what exactly the effects of different government programs have been. Pointing to causal linkages is particularly challenging in the case of government programs that take a long time to have an effect (e.g., in education, health care, economic development) (Hatry 2002).
Given the challenges involved in developing and using outcome measures as informational input for budgeting, public organizations may settle for output measures, which are easier (and also cheaper) to collect (Blöndal 2003). Excessive focus on output measures, however, may distract public managers from the actual mission of their organizations (Bohte and Meier 2000). If public officials know that their performance is evaluated according to narrowly defined output indicators, they may be motivated to dedicate most of their efforts to maximizing these indicators at the expense of other activities that have important public value but cannot be easily measured. Second, it can be very challenging to use performance information for budgetary decisions as it is difficult to establish what implications a change in outcome or output measures should have for funding levels (Caiden 1998; Schick 2007). If an organization has failed to achieve a performance goal, does that mean that its budget should be reduced or increased? There may be good arguments for both. On the one hand, the performance measure may have deteriorated because of underfunding; thus, additional funding could help to achieve the goal in the future. On the other hand, one could argue that funding should be taken away from ineffective organizations and allocated to more effective ones. Indeed, performance indicators alone may not be sufficient for giving adequate guidance in such a situation. As Perrin (1998, p. 372) has put it, “Attempting to reduce a complex program or social intervention, such as initiatives in child welfare, economic development, or health promotion, to a small number of quantitative indicators can disguise and mislead rather than inform what is really happening.” Furthermore, several governmental goals tend to be “eternal” (e.g., alleviating poverty, reducing crime, promoting economic development), and even if agencies fail to achieve their stated goals, reducing their appropriations would be an unrealistic sanction given the need for governments to address these problems (Caiden 1998). In addition, using performance information for decisions on budgetary appropriations is complicated because of the attribution problem. Given the complexity of many societal problems, it would be difficult to establish clear causal linkages between resources and outcomes (Schick 2007; Pollitt 2001). For example, if funding for school lunches is increased, it may be hard to evaluate how much the increased funding influenced the health indicators of schoolchildren, as there are many other factors that can play a role here. That, in turn, could make it arduous for the agency requesting additional funding to demonstrate ex ante how exactly additional funding would improve the performance indicators. Also, ex post, if the agency’s target indicators were not achieved, it could be unfair to reduce their funding.
Third, as budgeting is inherently a political process, the decision-makers are likely to be motivated by many other factors – alongside or even at the expense of performance information – when allocating resources: the desire to be reelected, ideological considerations, electoral promises, symbolic arguments, societal values, and the need to ensure transparency and control (Joyce 2008; Schick 2007; Wildavsky 1978, 1992). Thus, even if the performance information provides a strong signal to elected officials that they should reduce or increase funding for an agency, they may be reluctant to do so because of the aforementioned considerations. Furthermore, in many countries a significant portion of the budget is made up of expenditures mandated by law (e.g., social security payments, family benefits, etc.). In the case of such expenditures, it would not be possible to reduce or increase funding according to performance information without changing existing law (Pitsvada and LoStracco 2002). Finally, the budget preparation is usually a time-constrained process in which the participants have to consider a lot of information (Hatry 2002; Schick 2007, 2014). Collecting, analyzing, and interpreting performance information are time consuming and costly (Andrews 2004; Caiden 1998). Thus, under intense time pressure and in the face of limited analytical capacities, officials may not be able to integrate performance information into their budgetary decision-making.
In light of these difficulties, it is not surprising that many empirical studies have observed that performance information is often not used in making budgetary decisions – even if systems for collecting PI are in place (e.g., Curristine 2005, 2007; Joyce 2008; Schick 2007, 2014). The nonuse of performance information for budgetary decisions is particularly pronounced in the case of politicians in the legislature, whereas it is used to some degree in the negotiations between finance ministries and line ministries. The most extensive use of PI tends to occur in the internal decision-making of agencies (Berry 2007; Curristine 2007; Hawkesworth and Klepsvik 2013; Schick 2014). The existing empirical literature also indicates that performance information is more likely to be used at the local than central government level (e.g., Melkers and Willoughby 2005).
While fiscal stress is often viewed as a driver behind the adoption of performance budgeting reforms, the empirical evidence shows that the use of performance information tends to fall during cutbacks (Raudla and Savi 2015; Schick 2014). In principle, performance budgeting could allow decision-makers to undertake more “rational” cutback decisions, e.g., by preserving well-performing programs and trimming less effective ones (Hawkesworth and Klepsvik 2013). In reality, however, cutback decisions are driven by other considerations than performance, e.g., what is legally possible and what is politically feasible and less likely to meet strong opposition or social unrest (Schick 2014).
Another important finding of the existing studies is that different types of decision-makers in the public sector have varying informational needs when it comes to PI in the budget process (Curristine 2007; Hawkesworth and Klepsvik 2013). For example, politicians find formal performance reports focusing on snapshot indicators less useful than deeper, richer information that addresses the complexities and uncertainties of the public sector (Ter Bogt 2004). Generally speaking, the closer to the actual public service provision one moves, the more useful performance information is likely to become in resource allocation decisions (Pollitt 2001). For example, it may be rather difficult for a cabinet to decide whether education should get more funding than health care on the basis of PI. Similarly, a minister of education would find it challenging to use performance information for decisions on how to allocate resources between higher, secondary, and primary education. However, it would be feasible to use performance information in budgetary decision-making within particular schools or universities (Pollitt 2001). Different policy sectors are also likely to have varying informational needs. Using performance information might be more feasible in agencies that are involved in providing tangible services (e.g., road construction, issuing drivers’ licenses and passports) than in agencies that deal with intangible services (e.g., coordination) or regulation (e.g., financial regulation or consumer protection) (Pollitt 2001; Ter Bogt 2004).
The Problems of Adopting a Performance-Based Budget Classification
The challenges described in the previous section apply to all forms of performance budgeting. There are, however, a number of more specific problems (or even dangers) of switching from a traditional input-based budget classification to a performance-based one.
First, if, as a result of the format change, the budget consists of very general and abstract outcome targets and budgetary appropriations for achieving them, it may be difficult to assess whether the outcomes have been actually achieved and whether the use of resources has been appropriate. For example, after an outcome-based budget format was adopted in Australia, the ministries came up with highly aggregate outcome goals (e.g., the entire defense department had only three outcome measures). Dissatisfied with such a budget, the finance minister voiced the following concerns: “Some outcomes are so broad and general as to be virtually meaningless for budget accounting purposes leading taxpayers to only guess what billions of their dollars are being spent on… The Outcomes and Outputs Framework was intended to shift the focus of financial reporting from inputs… to outputs and outcomes, i.e., actual results. While this is worthy in theory, it has not worked. Basic information on inputs was lost in the changeover, and reporting of outcomes is seriously inadequate” (Mackay 2011, p. 23).
The reduction of the number of line items in the budget law may lead to a reduced role of the legislature in the budget process and hence to its decreased capacity to control whether the use of taxpayers’ money has been purposeful. This, in turn, may undermine the transparency and democratic legitimacy of the budget process as a whole. The existing empirical studies have demonstrated that the adoption of performance-based budget documents has indeed undermined the parliament’s role in the budget process, rendered the budget law less transparent for the members, and created difficulties in undertaking ex post controls (e.g., Sterck 2007).
Third, the use of a performance-based budget format gives rise to a situation where the breakdown of the budgetary expenditures varies significantly from year to year (as performance goals may change), making it difficult to compare expenditures over time (Webber 2004).
Finally, the use of a performance-based format may lead to reduced accountability of managers in the public sector. One of the core ideas of performance budgeting is that the managers would be made accountable for results instead of inputs. Given the challenges of establishing causal relationships between the activities of an organization and the societal outcomes (described above), it may, however, be difficult to hold officials accountable for results they cannot wholly control.
The Potential Dangers of Linking Budget Appropriations to Performance Indicators
The critics of presentational performance budgeting have argued that if the public officials know that performance information does not actually influence budgetary appropriations, they have fewer incentives to focus on improving performance. In order to address this problem they suggest to link funding more closely to performance indicators. Establishing more direct linkages between performance and budgetary appropriations, however, can lead to negative consequences.
First, since achieving performance goals in the public sector usually does not depend entirely on the efforts of the agencies, the reduction of funding after the organization has failed to achieve a performance target can create a sense of unfairness and lead to reduced motivation (Mayne 2007; Pollitt 2001).
Second, close linkages between performance targets and financial incentives can encourage “gaming” by the agencies and their officials and, as a result, may lead to worse results than would have been achieved without the performance budgeting system (Bevan and Hood 2006; Hood 2006). Indeed, when the size of budgetary appropriations is clearly linked to performance indicators, agency officials may be motivated to focus only on reaching these indicators and neglect other aspects of their activities (Bohte and Meier 2000). Worse still, in order to maximize their performance indicators, the officials may undertake activities that are socially undesirable (de Lancer Julnes 2006; Mayne 2007). An infamous example of such behavior is the case of hospitals in the United Kingdom. After their funding was made dependent on very specific performance indicators like shorter waiting times, the hospitals started putting up tents in their yards where the emergency care patients had to wait, in order to achieve their target of reduced waiting time in the hospital (Bevan and Hood 2006).
Third, linking funding closely to results may lead to a deterioration in the quality of performance information collected by the agencies (Hood 2006; Perrin 1998; Smith 1999). If the officials know that performance indicators can be used for reducing their budgets, they might have incentives to present embellished or distorted information or to start defining performance indicators in a way that is advantageous to them (but does not necessarily reflect improved performance) (Perrin 1998).
Finally, close linkages between performance indicators and budgetary appropriations can lead to a situation where performance indicators are primarily used for accountability. This, in turn, may undermine the role of performance information in organizational learning. Indeed, several studies have argued that a performance measurement system that emphasizes accountability might in fact work against achieving better results and improved performance (e.g., Halachmi 2002; Hatry 2002). When performance information is primarily used for accountability, the main questions asked are: was the goal achieved, who is responsible for it, and what are the sanctions for not reaching the set goal? When performance information is predominantly used for improving results, the focus of the system is on discovery, learning from experience, innovation, and change.
Thus, in establishing a performance budgeting system, it is important to realize that there might be a trade-off between increasing accountability and improving performance. If the primarily goal of measuring performance is increased accountability, the officials would have incentives not to deviate from the set goals and plans, even if such deviations might be in the public interest. Accountability-oriented performance budgeting systems may prevent public officials from reacting responsively and adjusting strategically to changing circumstances in the environment (Halachmi 2002).
At first sight, it might be difficult to argue against performance budgeting since it is rhetorically convincing and “sounds really good” (Moynihan 2005). Who wouldn’t want to improve performance in the public sector? Who wouldn’t want the government to make more efficient use of taxpayers’ money? Performance budgeting does indeed have several advantages. It can promote transparency, demonstrate what has been achieved with public resources, encourage officials to become more results oriented in their work, and enhance allocative efficiency. However, implementing performance budgeting often faces a number of challenges and even dangers. The potential problems caused by performance budgeting tend to be relatively small in cases where performance information is just added to the budget documents. They can, however, become more severe when the more extreme forms of the reform idea are implemented, especially when funding decisions become directly dependent on whether the organization has achieved a performance target or not. When budgetary allocations are directly linked to the achieved results, public organizations may start gaming the system, which could lead to perverse results. Given the nature of the public sector and complexity of the problems it has to address, it would be more advisable to use performance budgeting as an analytical instrument rather than a decision rule (Schick 2007). If used as an analytical instrument, performance budgeting can enrich the informational basis for making funding decisions. If used as a decision rule (like it is in the case of performance-based budget formats), performance budgeting can become too constraining and lead to unwanted results. Reformers of the budget process should also keep in mind that budgeting is inherently political. Thus, attempts to completely subject it to a technocratic vision of rationality are likely to be unsuccessful.
- Berry FS (2007) Performance budgeting internationally: assessing its merits. In: de Julnes PL, Berry FS, Aristigueta M, Yang K (eds) Practice-based performance management: an international handbook. Sage, Oak Park, pp 285–296Google Scholar
- Curristine T (2007) Performance budgeting in OECD countries. OECD Publishing, ParisGoogle Scholar
- Hatry HP (2002) Performance measurement: fashions and fallacies. Public Perform Manage Rev 26(4):352–358Google Scholar
- Pitsvada B, LoStracco F (2002) Performance budgeting-the next budgetary answer. But what is the question? J Public Budg Account Financial Manage 14:53–74Google Scholar
- Pollitt C (2010) Cuts and reforms. Public services as we move into a new era. Soc Econ 32(1):17–31Google Scholar
- Pollitt C, Bouckaert G (2004) Public management reform: a comparative analysis. Oxford University Press, OxfordGoogle Scholar
- Schick A (2014) The metamorphoses of performance budgeting. OECD J Budg 13(2):1–31Google Scholar
- Wildavsky A (1992) The new politics of budgetary reform. Scott, Foresman, GlenviewGoogle Scholar