Linking Performance Measures to Resource Allocation.

What do you do after you develop performance measures and targets? This article describes Dallas County, Texas' strategy for ensuring that measures are not only published but also studied, discussed, and utilized to make corrective actions and distribute resources in a way that supports the

Although the literature on performance measurement is full of references to the importance of connecting the measurement system to the budget, there is very little practical advice on how to accomplish this connection gracefully and effectively. The so-called feedback loop that connects the results of a performance measurement process to a resource allocation process always has been the trickiest part of a complete outcomes-based management system.

To be sure, the setting of objectives, the design of measures, the establishment of targets, and the reporting and data-gathering process have their own complexities, but by-and-large most governments have learned how to accomplish these tasks, and good examples are now readily available to assist any jurisdiction willing to commit to the process. Where most governments stumble is in using these measures to make a real difference in management, personnel, business procedures, or resource allocation. For the most part, feedback becomes an ad-hoc process of publishing measures annually and hoping they will be discussed as part of an annual budget process.

In attempting to deal with this problem, Dallas County, Texas, has developed a progressive pressure system that makes performance measurement and comparison with targets a daily instrument of shared management concern in a non-threatening, non-punitive way.

The term shared management refers to the reality of urban county government in Texas, with its United Nations-like structure. All elected officials report both to their own constituencies, who expect good services, and to the elected Commissioners Court, who allocate resources and guard the tax rate. County government in Texas has no hierarchical management structure like most cities and corporate organizations. The elected County Commissioners serve as full-time managers, as well as legislators.

Add to this diffuse organizational structure the influence of a state legislature that controls about half of the counties' $600 million in revenues, a chief accountant (in Texas, they are known as county auditors) who is accountable to the state district judges in each county (there are 37 in Dallas County), and you have the ingredients for a structure that could easily complicate efforts to manage with a focus on results.

But the system seems to work well in Dallas County by any measure of governmental success: low taxes, superior bond ratings, support from the local press, low attendance at public hearings (it means we are trusted to do the right thing), good services, and elected officials who are routinely re-elected. Dallas County Commissioners have become results-oriented not because government was perceived as ineffective, but rather because they sought to continually improve.

Outcome Measurement

Dallas County began regular quarterly performance reporting in the early 90s (see Exhibit 1). The first set of reports were little more than workload and financial measures, prepared without involvement by the operating departments. This system of reports was implemented sequentially and generally without the typical mission evaluation and goal- setting that is usually recommended prior to the onset of performance measurement. This was due in part to the organizational framework of Texas county government, which derives most of its "missions" from state law rather than from agenda-setting established by local elected officials.

The performance measure system had evolved into four quarterly reports by 1997. Two of these reports were specialized outcome reports, developed by the Office of Budget and Evaluation to focus on particularly troublesome areas of county government.

Juvenile Department Outcomes. The first true outcome report to be developed was a comprehensive and honest recidivism report for the county's $40 million dollar Juvenile Department. Recidivism is the ultimate outcome measure for a therapeutic juvenile justice system that has a mission of rehabilitating youthful offenders. The report includes data on behavior and progress up to 24 months after release from care and couples the outcomes with extensive demographic, social, and educational data to help juvenile authorities make changes to programs that do not work effectively.

Early versions of this report were discussed periodically but not systematically. However, the very first report to be published showed a surprisingly high level of recidivism in the two years following extensive (and expensive) treatment at the county's Youth Village. The planned doubling of the capacity of the facility was immediately shelved and the county turned to privatized programs with cash recidivism guarantees. This was the first important use of performance measures to directly affect resource allocation.

Judicial System Internal Benchmarking. The county began producing a regular judge-by-judge efficiency comparison for the 72 district and county judges and justices of the peace who depend on the Commissioners Court for resources. These reports are made public despite the potential political liability for elected judges and the limitation associated with using efficiency as a useful stand-alone measure of the quality of justice.

This report compares the cost-per-case disposition of each judge with costs of every other judge hearing a similar docket, with the understanding that judges have a large degree of control over their dockets and the costs they incur. This was a rare opportunity to compare similar operations without having to go outside of the organization, a form of internal benchmarking.

After an initial period of concern over the use of the report, the judges became concerned that the comparisons would be fairly presented, and they made numerous contributions to the presentation technique. Some even began finding ways of cutting their costs in order to improve their rankings relative to their colleagues. For example, certain judges began using part-time contract court reporters to replace full-time county employees, a decision entirely at their sole discretion under state law.

To these two specialized quarterly reports and the more general Management Report of output and financial data was added a bimonthly Capital Projects Report which focuses on scheduling milestones for construction projects. These reports were the backbone of the performance reporting system until 1997, when the last step toward full outcome measurement was initiated. [1]

Comprehensive Performance Measures and Targets. In 1997, every county department was trained in outcome measurement and was asked to develop such measures and targets--but only if they were requesting new or expanded programs. In 1998, all departments were required to submit outcome measures as a part of the budget process, even if they were not requesting new resources. These measures and targets have been reported quarterly since their development.

Once the reports became routinely available, the Commissioners Court turned its attention toward the important task of ensuring that the measures were constantly used by management during the year and were available during the formal resource allocation process. Each year since the first publication of the report, the progressive pressure system has been refined until it reached its current configuration.

Progressive Pressure

Exhibit 2 diagrams the progressive pressure that is applied to managers to focus their operations on outcomes. The system assumes that managers generally want their operations to be visible and efficient and that budget-makers want to place limited resources where they will do the most good. While this may appear to be an oversimplification, the system of discussing and modifying the budget to improve outcomes has assumed goodwill of all parties. Dallas County has learned that if goodwill and trust are lacking, any technique will probably fail. Conversely, when the chemistry of trust is good, the technique often does not matter. Therefore, the progressive pressure model supports and encourages the relationship building that is ultimately the most important requirement of governmental improvement.

Visibility. Progressive pressure begins with visibility and frequent review. To that end, all management reports are made available to everyone, even as they are being constructed. Managers have a chance to look at their peers' reports on a county-wide Intranet, and completed quarterly reports are placed on the county's Web page. Sensitive reports such as the Judicial Management Report are clearly marked "draft" until all affected parties have an opportunity to comment. Footnotes and narrative explanations are used to explain nuances in the data.

Grading. Each quarterly performance measure is informally graded against its target by the Office of Budget and Evaluation (OBE), and given a visible indication in the form of a check mark if it appears to be on course (Exhibit 3). Due regard is given to seasonal fluctuations and/or extraordinary events that may make a target temporarily unrealistic. Furthermore, both the department and the budget analyst are provided the opportunity to offer comment on mitigating circumstances.

Narrative Explanation and Summary Statements. A narrative statement is created for each department's performance data by the budget analyst and the department head. The narrative is condensed into simple phrases, which lead the reader to certain significant accomplishments or potential problem areas (Exhibit 4). These summary tables allow a reader to concentrate on the most significant issues without studying the entire report. It is these problem areas that become the target of the next level of progressive pressure.

Troublesome Trends Identification. Each quarter, the Office of Budget and Evaluation (OBE) prepares two discussion lists from the graded performance measure report. The first list contains items that represent troublesome trends in outcomes and/or missed targets. Examples would be a falling rate of successful prosecutions, lower-than-expected recruiting successes, or a reduction in child support payments distributed.

The second list contains performance data that suggest that the external environment is worsening. This list is an early warning that factors unrelated to the county's management may portend future problems. Examples would be increasing referrals to the truancy courts or additional numbers of state prisoners backing up in county jail. The list of troublesome trends (that are under management's control) and the list of negative environmental trends (that are not) are made available for discussion by the Commissioners Court.

The troublesome trends list is further subdivided into an "active" list and a "watch" list. The Office of Budget and Evaluation (OBE) prepares the recommendations and the Commissioners Court approves the assignment of each of the troublesome trends to one of the two categories. A watch list item is held "on the table" until the next quarter, at which time new watch list items may be added, old items may be dropped, or old items may be escalated to the next level of progressive pressure. Departments are notified by OBE that the Commissioners Court has requested that one or more performance measures have reached watch-list status, and their attention is called to this area.

For those items which the commissioners find the most troublesome, a five-signature letter is prepared and sent to the department head, indicating that a trend has been noticed and inviting (not requiring) the department head to join them in evaluating the trend and understanding the ramifications of the problem.

For example: a letter to the county's Chief Medical Examiner contained the following paragraph: In particular, your department received upgraded positions in the Forensic Biology-DNA Lab with the expectation that more analyses would be performed, thus generating more charges. This program actually performed fewer analyses in FY99 than in FY98 and generated substantially lower charges. In addition, your department has not initiated a customer satisfaction survey, as agreed.

At the conclusion of the letter, the department head is requested to provide additional explanatory information directly to the Commissioners Court, and the services of OBE are offered if needed. Each letter is tailored by OBE to point out specific deficiencies in the particular data in question for the department being addressed.

No department head or elected official has yet refused this invitation, and the occasion of the five-signature letter has usually evoked thoughtful responses to the shared problem. Department heads are aware that their subsequent request for resources will be considered by the commissioners. They seem to welcome the opportunity to tell their story early in the year in a non-competitive environment rather than during budget and tax rate deliberations.

Initially, there was a tendency on the part of department heads to interpret the five-signature letter as a punitive letter or a warning of substandard performance. The Commissioners Court has worked hard to dispel this notion. The initial tendency toward negativity has disappeared; a mutual respect and a team attitude toward the solution of a shared problem has surfaced.

Performance Forum. The next step in progressive pressure is a public meeting called a Performance Forum; the Commissioners Court conducts such a forum monthly with an agenda established by the Office of Budget and Evaluation. The Performance Forum which follows the publication of each quarterly performance report (i.e., the January, April, August, and October meetings) has as its sole agenda a discussion of the quarterly results, and selection of items to be placed on the watch and active lists. The Commissioners Court agrees on the lists and also may authorize congratulatory letters to departments that appear to be meeting or exceeding performance goals. Also at this meeting, OBE presents the list of county functions experiencing external workload pressure, as an early warning for subsequent budget discussions.

The Performance Forums for the two following months are devoted to responses from departments who received invitations to discuss their troublesome trends or missed performance targets. This forces the departments to confront their problems early in the year and includes the Commissioners Court in the strategy to solve the problem. Often a shared strategy for correcting the problem will emerge from the informal session.

Since the county's fiscal year begins on October 1, the annual budget workshops replace the August Performance Forum. The third-quarter performance data is made available on the same form that accompanies any request for new resources. It is the job of OBE to ensure that performance measures are connected to requests by reminding the Court of prior-year performance agreements and/or structuring new measures to capture the outcomes of any new resources requested. Often, performance targets are projected with and without new resources, so that the Court can have a measure of the value to the community of any new resources applied.

Results. Dallas County has developed a pattern of sending six to 10 letters each quarter to department heads whose measures fail to meet their targets. Approximately twice that many items receive "watch list" status. Thus far, there has been complete positive compliance with the requests for additional information. The result has been that questionable programs are being discussed informally at least once during the year at a Performance Forum.

Resources Redirected

Several large programs have been modified, and at least two significant programs have been abandoned as a result of the progressive pressure model. One was a program to provide aftercare to youth who are released after spending up to one year at the county's Youth Village, a therapeutic program intended to return youthful offenders to productive patterns of living. Targets were established for successful outcomes, and when they were not reached, the program was cancelled by consensus of the Commissioners Court and juvenile authorities. Similarly, a program to bring specialized protective services to the siblings of children involved in the justice system has been abandoned because its performance targets were not met. Several other programs have been modified in an attempt to improve their outcomes, with final disposition of the programs awaiting additional outcome data.

The Commissioners Court could have unilaterally modified or abandoned these programs, and in so doing might have left a lingering resentment in the mind of the programs' advocates. The outcome measures told the story, with due regard for all attempts to make the programs work, and all sides concurred that resources should be redirected.

At least one program, a substance abuse program for inmates, has been expanded as a direct result of outcomes that exceeded expectations. The success of the program in meeting its targets to prevent a return to incarceration encouraged the sheriff and the Commissioners Court to double the capacity of the program.

Other results are more subtle and harder to quantify. They flow from the public discussion of the performance targets, resulting in early management attention being focused on an identified problem area. It seems likely that many projects are improved simply by being the topic of management effort that would not otherwise be directed in the same manner.

Lessons Learned

Dallas County's use of gentle and progressive pressure to manage with a focus on outcomes has been effective despite the lack of clear authority lines among elected officials. Resource allocation decisions are made with constant reference to performance data and management's track record of meeting their targets. The key to the success of the system has been to use it more often than once a year as an early warning that management attention is required.

Numerous smaller changes in resource allocation can be traced directly to the data accumulated in the performance measurement systems. The largest programmatic changes have been discussed above. Many smaller adjustments have occurred as a result of targets having been established, reported, discussed, and graded. Performance measures and targets are now an integral part of virtually every resource allocation discussion.

County Commissioners have learned to maintain a humble attitude toward performance measures, because despite management's best efforts, the measures are often ambiguous. Seldom is a measure so clear that some mitigating circumstance or nuance is not relevant. This is the reason that give-and-take discussions in a non-threatening environment are so valuable. The county has begun to use the term performance indicator rather than performance measure in recognition of the possible imperfection in some of the data.

A healthy distrust for any particular performance indicator, however, should not be used as ammunition to diminish the importance of a system of useful and objective data to assist with resource allocation. Taken as a whole, a performance system is a key element in producing valuable results for constituents. The benefits of a measurement system are derived primarily from the discussion, not the creation, of the performance data.

PHILIP B. SCHEPS, PH.D. has been the Budget Officer of Dallas County, Texas, for 11 years and serves as faculty for the GFOA "Benchmarking and Performance Measures" training course. He has also served as Director of Finance for the City of Austin, Texas. He can he reached at pscheps@dallascountry.org for comments or questions concerning this article.

NOTE

(1.) Dallas County's quarterly performance reports are available on the Office of Budget and Evaluation's section of www.dallascounty.org

EVOLUTION OF DALLAS COUNTY'S PERFORMANCE MEASUREMENT SYSTEM

1992 * Management Report--Volume I quarterly workload and financial measures

1994 * Juvenile Recidivism Report quarterly outcomes related to Juvenile Department Programs

1995 * Management Report--Volume II quarterly judge-by-judge efficiency comparison for 92 courts

1996 * Major Projects Review bi-monthly review of status of all infrastructure projects

1997 * Countywide training on performance measures

* Outcome measures required for new resources

* Management Report--Volume Ill outcome measures for all departments

1998 * Outcome measures required for all departments

* Performance Forum and progressive pressure

1999 * Outcome measures used in budgeting

Exhibit 2 PROGRESSIVE PRESSURE TO ACHIEVE OUTCOMES

Weakest

* Preparation of objectives and measures

* Establishing/negotiating targets

* Annual compilation of data and public disclosure

* Quarterly gathering of data and public disclosure of interim results

* Preparation of written narratives for each measure

* "Grading" of results by outside agency such as Office of Budget and Evaluation

* Preparation of quarterly "watch" list by outside agency

* Preparation of quarterly "question" list by outside agency

* Written invitation from governing body to explain interim results

* Presentation/discussion of response at public meeting

* Consideration of measures and targets along with resource allocation of compensation decisions

* Specific contract for service with "automatic" consequences in staffing, gain-sharing, and/or compensation

Strongest