Benchmarking Tools: An Application to Juvenile Justice Facility Performance

• Published date: 01 December 2002
• Date: 01 December 2002
• Author: Ronald C. Nyhan
• DOI: 10.1177/0032885502238679
• Subject Matter: Articles

THE PRISON JOURNAL / December 2002Nyhan / BENCHMARKING TOOLS

BENCHMARKING TOOLS:

AN APPLICATION TO JUVENILE

JUSTICE FACILITY PERFORMANCE

RONALD C. NYHAN

Florida Atlantic University

This article explores the use of data envelopmentanalysis (DEA) to compare techni-

cal efficiency among juvenile justice facilities in the state of Florida. Comparative

analysis, sometimes called benchmarking, helps to improve efficiencyand effective-

ness of government service providers by identifying the presentlevel of productivity,

diagnosing deficiencies, identifying opportunities for improvement, and developing

strategies for change. Making service comparisons has implications for budgeting,

contracting, granting, and other resource allocation decisions in the public sector.

Using data taken froma recent Florida Juvenile Justice Accountability Board report,

DEA is used to assess the relativeefficiency of 35 halfway house facilities operated by

the state and under private contract. The article illustrates the use of DEA to deter-

mine efficiently operated facilities, estimate targets for improvement of inefficient

facilities, and identify those facilities that are potential models for improvement.

Finally, the relative performance of state-operated versus privately contracted half-

way house facilities is evaluated.

The public sector is constantly trying to improve the efficiencyand effec-

tiveness of the goods it produces and the services it provides. The Interna-

tional City/County Management Association and the Governmental

Accounting Standards Board (Fountain & Roob, 1994) have established

guidelines for defining and reporting on performance measures. Severalgov-

ernment initiatives, including the Government Performance and Results Act

of 1993, seek to identify and report on inputs, outputs, and most important,

outcomes of government activities. Federal, state, and local governmentsare

recording a plethora of performance measures. In many cases, these data are

used to make temporal comparisons (i.e., differences in the same organiza-

tional unit from one period to another). This is helpful to compare differences

and make relatively simple judgments about improvements.Some data com-

parisons are also being made at the subgovernment level, such as different

THE PRISON JOURNAL, Vol. 82 No. 4, December 2002 423-439

DOI: 10.1177/0032885502238679

© 2002 Sage Publications

423

precincts of police departments or fire service facilities in a city. In some

cases, comparisons are being made across jurisdictions (International City &

County Management Association, 1997). By comparing and contrasting per-

formance, an expectation exists that “best practice” units can be identified

and used as benchmarks for improving the efficiency, quality, and effective-

ness of all government units (e.g., Martin & Kettner,1996; Nyhan & Martin,

1999).

Although much effort was expended in defining and reporting perfor-

mance measures to be used for comparisons, less effort is evident in develop-

ing analytic tools to address the myriad of problems in making multiattribute

comparisons. This article will address the traditional tools used by public

managers, introduce a new tool, data envelopment analysis (DEA), and then

use it in conjunction with regression analysis to evaluate the relativeperfor-

mance of privately and publicly managed juvenile justice facilities in

Florida.

TRADITIONAL COMPARATIVE

PERFORMANCE MEASUREMENT

Two approaches, ratio and regression analyses, historically have been

employed to make comparisons of government units using performance-

related data. Ratio analysis involves computing data pairs consisting of one

performance variable (output or outcome) divided by one input variable

(usually cost related). These ratios are then used to rank order the units. The

higher the ratio of a given government unit, the higher is the relativeperfor-

mance. However,when there are several ratios, the interpretation of the most

efficient across all measures becomes very complex and its utility is highly

problematic. For example, comparing performance of different police pre-

cincts on the basis of multiple measures such as crime incidents per citizen,

arrests per officer, cost of clearance, and conviction rates can be confusing

because some units have high ratios in one area and low ones in others. What

is needed is a single scalar measure that combines the performance of indi-

vidual units across all ratios. One method of accomplishing this is through

the use of weighted averages. In this method, each of the performance ratios

can be assigned a weight reflective of its relative importance based on the

judgments of policy makers and evaluators.The best performer will have the

highest weighted average of all ratios. However, the task of assigning

weights to the ratios is highly subjective and is one of the most difficult and

controversial tasks policy makers and evaluators can undertake.

424 THE PRISON JOURNAL / December 2002