This article first appeared in the August 1996 issue of Government Finance Review, the journal of the Government Finance Officers= Association of Canada and the United States, which also has a copyright.

I. INTRODUCTION

Continuing pressures for improved accountability and greater value for money performance have prompted governments at all levels to recognize the need for program performance measurement. Many jurisdictions, particularly in the United States, have even legislated performance measurement, but performance measurement is still at an early stage of development. While many governments are developing formal performance information, there are not many that have implemented it to the point that it is used as a regular feature of management and decision making.

Performance measurement is still greeted with skepticism by some program managers. For some, it is just a "bean counting" exercise, offering little in terms of program planning and performance improvement. The skeptics might have a point if performance measurement is implemented as a stand alone information system, used primarily for accountability reporting of past performance.

To be successful, performance measurement must be more than score keeping. It must be used and supported by a large number of program managers. Managers are the ones who know their programs best and thus what the best measures will be. Managers will be responsible for timely and accurate data collection and reporting, and they will be the ones with the greatest potential to use the measures to improve program effectiveness and efficiency. Performance measurement must therefore become an integral part of managing government programs, and it ought to be used in as many decision making applications as possible.

This article provides government officials with guidance on how to use performance measurement. First a simple framework for developing meaningful and useful performance information is described. Next, the article discusses general principles of results planning, monitoring and evaluation. Then it shows how performance measurement can be used in a variety of management processes including strategic planning, operational planning and control, program evaluation, manager's performance appraisal, resource allocation, and accountability reporting to the public and elected officials.

II. DEVELOPING USEFUL MEASUREMENTS

Managers need to be clear about what constitutes meaningful performance information before they can use it. If performance measurements are to help to achieve results, they must relate directly to the program's mission and its key results. Developing meaningful and useful performance measurements is a three step process which is used to define the program mission, to identify its key results and finally to select measures for the key results. An example appears in Exhibit 1 on the next page.

This methodology can be applied successfully to any government and any level of an organization, and it is relevant to any kind of program - direct service, regulatory, transfer payment, and internal support programs. It follows a program's logic, and thus it is simple for managers to understand and apply. There are three steps:

1. Define the Program Mission.

Every program or service exists for a purpose. It must be clear about what that purpose, or MISSION, is because performance measurement must contribute to achieving the mission. An effective mission statement is succinct and responds to the following three questions:

- WHAT is the product/service provided?
- WHO is the intended primary client or target group?
- WHY is the program of more general benefit?

2. Identify the Key Results of the Program.

The results are statements describing what the program wishes to achieve in support of the mission. Thus, they can be classified into three groups relating directly to the three questions of the mission:

work process outputs (WHAT?) - the direct outputs and efficiency of the activities undertaken
conducting business;

client oriented results (WHO?) - the benefits of the program from the client's point of view - in regulatory programs the mission is to modify the behavior of a target group so it may be appropriate in such cases to speak of impacts rather than benefits

strategic outcomes (WHY?) - the social, economic or cultural conditions that the program seeks to influence

Each of these kinds of results is important and useful for different decisions and processes. But the critical category is client benefits/impacts. Information on the extent to which a program delivers benefits to its clients or the influence it has on its target group is important in every program decision. Just as the customer is the most important person to a business, the client is central to any government program.

3. Develop Performance Measures for the Stated
Results

Performance measures are the means by which results can be planned, monitored and evaluated. Measurement is easier for work process outputs, more difficult for client benefits/impacts and most difficult for strategic outcomes. Also managers can forecast and control performance of work process outputs more than they can for client benefits/impacts and strategic outcomes. But measures should be developed for all three categories because they provide an objective and verifiable method of planning and assessing performance, managers need measures to relate results to resources, and measures provide a means of comparison and prediction.

Not all results can be feasibly measured, of course. Sometimes proxy indicators must be used as indirect measures of performance. For example, it might be possible to survey client satisfaction, but perhaps the program cannot afford a survey. Instead, measuring client behavior or demand for service might be used as a proxy for satisfaction. The measures of strategic outcome shown on Exhibit 1 are proxy indicators. Care must be taken in setting targets for proxy indicators, and a tolerance for variances must be shown in monitoring and evaluating performance. Often it is more appropriate to analyze proxy indicators in terms of the trend of performance over time rather than planned to actual performance in a single year.

III. PRINCIPLES OF USING PERFORMANCE MEASUREMENT

The development of performance measures is not "rocket science". Most managers, using the three step method described above, can develop meaningful and useful performance measures quickly, Similarly, the use of performance measurement is not especially complicated. Planning, monitoring and evaluating performance measures is, in principle, exactly like using financial information.

1. Results Planning

Planning performance can be difficult if performance measures have not been collected previously, because there are no trends, baselines or standards to guide planning. Many governments discover, however, that they already have most of the data they need for performance measurement, and those data are largely accessible on computerized data bases. All of the data needed for the measures shown on Exhibit 1, for instance, were available from existing data bases - some maintained by the county and some by the state. The existence of such data makes it possible to reconstruct performance data for prior years, and the resulting time series of data can be used for planning.

Past performance is only one way to estimate future performance. Other ways include comparison to other government organizations, standards established through work engineering, private and public sector association data, and research into client needs (eg. surveys or focus groups)

Planning does not mean that a program will always be able to achieve its results to 100% of potential. Performance targets will have to be realistically set. In planning performance one must take into account what direction the strategic plan gives, what the policy priorities are, how much demand for service there is, how much money is allocated, and other environmental variables.

In monitoring and evaluating results, performance measurements are no different from financial data. By themselves, measurements do not tell the whole story about program performance, but variances of actual performance from planned performance, or trends over time, do raise red flags about problems or opportunities. Such variances do not per se constitute value judgments about whether the program or its managers performed well or poorly. The important point when examining variances is to understand why they occurred and what the program can do to about them.

Performance can be monitored not only during the year, but over time as well. Consider Exhibit 2. A program that spends within its allocation, typically attracts little if any scrutiny. It is only when it proposes to spend more, that budget agencies raise a funding issue with the program. But what if the results the program achieves are declining? That Aresults issue@ is every bit as important as a funding issue. Without performance measurement, results issues might not be detected.

Program performance often changes only a little year by year; but a small negative variance compounded over several years can develop into a serious problem. Such situations do occur, and might be avoided if results are monitored as well as resources.

Managing results through performance measurement helps to achieve optimum performance from programs. Management often consists of trade offs among competing priorities, and results management is no different. In managing results, managers are faced with situations such as:

- a set standard of quality may restrict the quantity of service that can be provided;

- a priority placed on one result may mean that another result must be de-emphasized

Performance measurement will not make these decisions for us but it will make the trade offs clearer.

Performance measurement will become stronger as it is used more often it is used and for more different purposes. Some applications of performance measurement are externally initiated, and some are management processes internal to the program or department. A major advantage of the framework presented in this article is that the information can be used in a variety of management processes. This notion is illustrated by Exhibit 3 below. Note that all applications require information about client benefits/impacts. These are the results most critical to program success, and of interest to managers, politicians and the public alike. Features of the external and internal applications are discussed below, beginning with the internal management processes.

1. Strategic Planning

Strategic planning is the process which determines the basic mission of an organization, its long term goals, the contribution of each part of the organization and the action that must be taken. The three step process for developing performance measurement contains the main elements of strategic planning, in particular defining the mission and identifying the strategic outcomes. Strategic planning in government, however, often lacks performance measures and, hence, offers weak and general guidance to operating programs. Performance measurement will assist strategic planning by:

- ensuring that programs fit within the strategic plan

- structuring each program along its own strategic lines

- more rigorous planning, monitoring and evaluating of strategic outcomes, and thus

- allowing better deployment of resources and management attention

2. Operational Planning and Control

Operational planning and control is probably the most frequent and important use of performance measurement. The shorter term operational planning and control horizon (eg. 1 year) means that the focus is on the ongoing efficiency and effectiveness of program delivery rather than on strategic outcomes. Therefore the relevant results are those which managers and staff can influence over the shorter term: work process outputs and client benefits/impacts.

It is worth stressing again the importance of results from the clients' point of view. It is possible to have a very busy year without making any difference to clients. If a government organization has any kind of performance measurements, they are usually about work process outputs. Indicators about client benefits or impact are too rare. But consider the following situation where the operational targets are being achieved or exceeded but client benefits are not.

This program is meeting its work plan targets, but it is not effective in delivering value to the clients. Perhaps the "wrong" work processes are being emphasized. Without client oriented performance measures this program might have no indication that it should change.

The main uses of results information in operational planning and control are:

budgeting - if the use of additional resources is proposed, the request must be justified in terms of additional results, or if resources are reduced the impact on results must be shown;
in-year monitoring - to keep the program on course and to avoid nasty surprises;
year-end review - to evaluate performance and to set future direction.

3. Program Evaluation

Program evaluation is a more formal and rigorous assessment of a program than is operational planning and control, and it takes place at much longer intervals. Program evaluation is broader in scope than operational planning and control, questioning the usefulness of and need for programs. Thus strategic outcome and client benefit information is more relevant that work process outputs.

Performance measures for use by managers in program planning and control are no substitute for program evaluation. But the method described in this article for developing performance measures will supply much of the data needed for a typical program evaluation. Performance measures developed for program management can be of great value in program evaluation because they will:

establish a framework for analyzing the program, getting evaluations started earlier and fostering an evaluation mind set among managers and staff;

provide a time series of performance data, reducing the cost of evaluations and making the findings more conclusive;

help in identifying programs that need to be evaluated, because the measurements will reveal performance problems as they develop.

4. Manager's Performance Appraisal

There is a difference between holding a program accountable and holding a manager accountable. The distinction is that while the program must ultimately be seen in light of its strategic outcomes, these results are influenced by more than one government program and therefore are by definition beyond the manager's personal control.

It is fair to hold the manager accountable for client benefits/impacts and work process outputs because even though control may not be absolute, it is still the manager's job to organize the work processes so that the benefits or impacts are delivered to the clients. But strategic outcomes are subject to too many outside influences. So for the purposes of a program manager's performance appraisal, strategic outcomes are not relevant.

Performance contracting is receiving renewed attention today as a means of strengthening accountability and increasing managers' motivation to improve program performance. Performance contracting is impossible without performance measures. Performance measures, particularly if they include measures of client benefit/impact as well as work process outputs and efficiency, can facilitate performance contracting by allowing targets to be set and objectively assessed, and by focusing individual effort on organizational goals.

5. Accountability to Political Governing Bodies

Many governments adopt performance measurement precisely because the political governing body has demanded it as a means to strengthen accountability. The problems facing government today, and the programs we use to address them, are many and complex. Politicians cannot possibly have personal and detailed knowledge about all of them, yet they are required to make decisions about them. It seems eminently reasonable that a political decision maker should be able to ask from any program, and receive immediately:

- a statement of the mission of the program

- statements describing the results the program is striving to achieve

- performance measures which set forth the plans and actual performance in achieving those results

Politicians are elected to make policy decisions. So in providing performance information to them, it is not enough simply to report work process output and efficiency data. Politicians are rightly and legitimately interested in the benefits programs provide to clients and in the greater good those programs are doing for the community or society.

6. Resource Allocation

Although performance measurement will help to bring more rationality and analysis to the resource allocation process, resource allocation will never be a wholly rational, analytical process. In government there is no common denominator for comparing alternative investments. Also, in most governments, the size of the discretionary budget is quite small; ninety percent of the budget or more is allocated before the budget making process even begins because of legal obligations (eg. debt service), legislated entitlements, and tradition. In government resource allocation is a competition for limited funds among non-comparable demands. Funds are allocated by means of pressure that advocacy groups bring to bear on the politicians.

A government department has an advocacy role to play because it has a mandate to affect a certain segment of society's activities. Thus, a department can apply pressure. Pressure can take many forms, and not all forms are legitimately available to the department. But the source of pressure or influence that the department can deploy is information, because the government department probably has more information about a given issue than any other group.

It is unfortunate most funding requests from government departments are made in operational terms ("More money is needed to do more work"), as this argument rarely impresses politicians today. There is much work that can be done and little money with which to do it.

In making a funding request or disputing a budget cut, the program must provide information that makes sense to politicians. The money will be spent on work process outputs, to be sure. But the funding request must be justified in terms of the delivery of benefits to the client and the outcomes the program will produce for the community or society. This is way to ensure politicians listen to and understand a funding request.

Many factors contribute to a political resource allocation decision, and results information is just one of those factors. So there is no guarantee that a request for funds justified in terms of results will be successful. The chances are better, however, if politicians are provided a results oriented business plan, which in turn can be justified to the voters and which can be used to monitor progress and achieve success.

7. Communication with the Public

It is commonplace to observe that the public is cynical about the performance of government programs. Yet, in the experience of the author of this article, performance measures will reveal that the majority of government programs already are performing well. For those that are not, performance indicators will be instrumental in bringing about improved results. Like politicians, the public not be too interested in the amount and type of work programs do; work process output information is useful to program managers, but it is not meaningful to the public. Citizens want to know how programs are affecting their lives and the quality of life of society. Client benefit/impact and strategic outcome information are more important to them.

The public pays good money to fund government programs. Those programs are achieving results. Without performance indicators, governments do not have an effective means of conveying this to the public. Performance indicators should be used in public communications because citizens deserve to know that they are being served well.

Performance measurement can be very useful in a wide variety of management processes. Different kinds of performance information is useful for different kinds of decisions. Performance measurement cannot make the decisions. It will not replace managers' or politicians' experience, training and judgement, but it will give a new dimension to the decision making process.

Government programs exist in order to improve or maintain some aspect of life. Government employees are entrusted with resources so that they can produce program results. Thus, the fundamental reason for developing performance measures is to help governments to produce results.

Performance measurement must therefore not be seen as an score keeping system isolated from the mainstream of government management. Rather, it must be a tool serving a variety of management processes. Strategic planning, operational planning and control, program evaluation and managers' performance appraisal are all processes that work much more effectively with performance measures. Accountability, resource allocation and communication with the public will also be strengthened through the use of performance measures. Performance measures must be clearly linked to and support these processes.

The development and use of performance measurement is not complicated. Meaningful and useful performance information can be developed by following a simple program logic. Furthermore, performance measures can be used in a manner similar to the use of financial information. The environment of government today requires the use of performance measurement. Governments must give equal and formal attention to results as well as resources.