Dea PDF

.
GLOSSARY
DATA ENVELOPMENT ANALYSIS
A TECHNIQUE FOR MEASURING

THE EFFICIENCY OF
GOVERNMENT SERVICE DELIVERY
Steering Committee for the Review of

Commonwealth/State Service Provision
1997
vii
. GLOSSARY
 Commonwealth of Australia 1997
ISBN: 0 646 33533 2
The Steering Committee for the Review of Commonwealth/State Service

Provision welcomes the use of this information paper in furthering the aims of
its terms of reference. The information paper is copyright and may be used as
permitted by the Copyright Act 1968 provided appropriate acknowledgment of
the source is published. Any inquiries should be directed to the Steering
Committee Secretariat.
Secretariat
Steering Committee for the Review of Commonwealth/State Service Provision
Industry Commission
LB2 Collins Street East Post Office
Melbourne VIC 8003
Level 28
35 Collins Street
Melbourne VIC 3000
Telephone: 03 9653 2100 Facsimile: 03 9653 2199

E-mail: IC@mail.indcom.gov.au
An appropriate citation for this report is:

Steering Committee for the Review of Commonwealth/State Service Provision
1997, Data Envelopment Analysis: A technique for measuring the efficiency of
government service delivery, AGPS, Canberra.
viii
. GLOSSARY
PREFACE
Effective government service provision benefits from the support of rigorous

measurement techniques. Data Envelopment Analysis (DEA) is an analytical
tool that can assist in the identification of best practices in the use of resources
among a group of organisations. Such identification can highlight possible
efficiency improvements that may help agencies to achieve their potential.
The aim of this paper is to promote a better understanding of the DEA
technique. It explains DEA’s conceptual underpinnings, how to interpret the
output from DEA models and its strengths and weaknesses. Also, through the
use of case studies on hospitals, dental services, police, motor registries, and
corrective services, this paper provides a practical guide to developing and
refining a DEA model and interpreting of results.
This paper is directed at those responsible for providing government services
and those accountable for their delivery in a cost effective manner. It should
encourage people to think about how more detailed and rigorous analysis of
performance can assist in improving the efficiency with which resources are
used to provide essential services to the community.
DEA can be a very useful analytical technique by providing an important ‘first
step’ tool in comparative analysis. But users also need to recognise its
limitations as an input to the development of public policy. Its theoretical
predictions of potential efficiency gains may not be translatable into actual gains
when factors such as service quality, fundamental differences between
individual services and the costs of implementing changes are fully accounted
for. Non-efficiency objectives such as access and equity are also important
policy considerations for government, against which efficiency benefits will
inevitably be balanced.
The Steering Committee wishes to thank the service agencies that were involved
in the case studies and their staff for their enthusiasm and assistance. I would
also like to thank, on behalf of the Steering Committee, the DEA Working
Group which was responsible for preparing the paper.
Bill Scales, AO
Chairperson
Steering Committee for the Review
of Commonwealth/State Service Provision
ix
. GLOSSARY
CONTENTS
DEA Working Group VI
Abbreviations VII
Glossary IX
1 Improving the performance of government service providers 1
2 What is data envelopment analysis? 9
3 How do you calculate DEA? 25
4 Overview of the case studies 39
5 Case studies 51
Appendixes
A Technical guide to DEA 123
B Programs for the application of DEA 131
References and further reading 135
x
. GLOSSARY
DEA WORKING GROUP
Jurisdiction Member
New South Wales
NSW Treasury Mr John Pierce (Convenor)
NSW Treasury Mr Roger Carrington
NSW Treasury Ms Nara Puthucheary
NSW Treasury (from January 1997) Ms Kathryn Kang
Victoria
Department of the Treasury, Victoria Mr Barry Bond
Department of the Treasury, Victoria Mr Dick Elvins
Department of the Treasury, Victoria Ms Kerry Macdonald
Queensland
Queensland Treasury Ms Trish Santin
Queensland Treasury Ms Jennifer Pallesen (until January 1996)
Queensland Treasury Mr Rick Stankiewicz (until January 1996)
Tasmania
Department of the Treasury and Finance Mr Richard Mackey (until March 1996)
Northern Territory
Department of the Chief Minister Mr Phil Temple
NT Treasury Mr Don Parker
Commonwealth Government
Department of Finance Mr Richard Mackey (from April 1996)
Local Government
Local Government & Shires Associations of NSW Mr Murray Kidnie
Local Government & Shires Associations of NSW Ms Lorraine Slade
Secretariat
Industry Commission Mr Jeff Byrne
Industry Commission Mr Rob Bruce
Industry Commission Mr Tendai Gregan
External reviewers
Australian Bureau of Statistics Mr Steven Kennedy
Tasman Asia Pacific Dr Denis Lawrence
University of Georgia Professor Knox Lovell
xi
. GLOSSARY
ABBREVIATIONS
CRS Constant returns to scale
DEA Data envelopment analysis
RTA Roads and Traffic Authority (NSW)
SCNPMGTE Steering Committee on National Performance

Monitoring of Government Trading Enterprises
SCRCSSP Steering Committee for the Review of

Commonwealth/State Service Provision
TFP Total factor productivity
WIES Weighted inlier equivalent separations
xii
. GLOSSARY
GLOSSARY
Allocative Whether, for any level of production, inputs are used

efficiency in the proportion which minimises the cost of
production, given input prices.
Benchmarking The process of comparing the performance of an
individual organisation against a benchmark, or ideal,
level of performance. Benchmarks can be set on the
basis of performance over time or across a sample of
similar organisations, or against some externally set
standard.
Best practice In this context, the set of management and work
practices which results in the highest potential, or
optimal, quantity and combination of outputs for a
given quantity and combination of inputs
(productivity) for a group of similar organisations.
Best practice can be identified at a number of levels,
including organisational, national and international.
Congestion A situation in which an organisation has unwanted or
surplus inputs and would incur a net cost to reduce
those inputs. For example, redundancy payments
associated with reducing staff levels will result in a
net cost to an organisation if they are higher than the
savings in wages for a given period.
Cost efficiency Where an organisation is technically efficient and
allocatively efficient and, hence, produces a given
quantity, quality and mix of outputs at minimum
possible cost given existing knowledge of
technologies and people’s preferences.
Data Envelopment A linear programming technique which identifies
Analysis (DEA) best practice within a sample and measures efficiency
based on differences between observed and BEST
PRACTICE units. DEA is typically used to measure
technical efficiency.
Decision Making The organisations or units being examined in a DEA
Units study. In public sector studies, these units may not be
commercial or profit-making entities.
Dynamic efficiency Success with which producers alter technology and
products following changes in consumer preferences
and productive opportunities.
13
. GLOSSARY
Effectiveness Degree to which the outputs of a service provider

achieve the stated objectives of that service — for
example, the extent to which hospitals are meeting the
demand for non-elective surgery. In the case of
government service providers, the government
normally sets such objectives.
Efficiency Degree to which the observed use of resources to
produce outputs of a given quality matches the
optimal use of resources to produce outputs of a given
quality. This can be assessed in terms of technical
efficiency and allocative efficiency.
External operating Factors which affect the providers of outputs that are
environment not in the direct control of managers — for example,
weather, client wealth and in some cases input prices.
Human services Services provided by core government agencies, such
as education, health, welfare and justice.
Linear program A set of linear mathematical equations for which a
solution can be obtained subject to an upper bound
(maximisation) or a lower bound (minimisation).
Non-scale technical Proportion of technical efficiency which cannot be
efficiency attributed to divergences from optimal scale (scale
efficiency); sometimes known as managerial
efficiency or pure technical efficiency.
Outputs Goods and services provided to entities or persons
outside the production unit.
Partial productivity Ratio of the quantity of an output (or the combined
indicator quantities of a subset of total outputs) to the quantity
of an input (or the combined quantities of a subset of
total inputs) where some inputs or outputs are not
included. For example, output per employee does not
include the other inputs required to produce the
output, such as raw materials, semi-finished goods
and capital.
Peers In DEA studies, the group of best practice
organisations with which a relatively inefficient
organisation is compared.
Production frontier The curve plotting the minimum amount of an input
(or combination of inputs) required to produce a
given quantity of output (or combination of outputs).
14
. GLOSSARY
Production Relationships incorporated in production processes

technology which determine the manner in which inputs can be
converted to outputs.
Productivity Measure of the physical output produced from the use
of a given quantity of inputs. This may include all
inputs and all outputs (total factor productivity) or a
subset of inputs and outputs (partial productivity).
Productivity varies as a result of differences in
production technology, differences in the technical
efficiency of the organisation, and the external
operating environment in which production occurs.
Returns to scale Relationship between output and inputs. Returns can
be constant, increasing or decreasing depending on
whether output increases in proportion to, more than
or less than inputs, respectively. In the case of
multiple inputs and outputs, this means how outputs
change when there is an equi-proportionate change in
all inputs.
Scale efficiency The extent to which an organisation can take
advantage of returns to scale by altering its size
towards optimal scale (which is defined as the region
in which there are constant returns to scale in the
relationship between outputs and inputs).
Slacks The extra amount by which an input (output) can be
reduced (increased) to attain technical efficiency after
all inputs (outputs) have been reduced (increased) in
equal proportions to reach the production frontier.
This is a feature of the piece-wise linear production
frontier derived when using DEA.
15
. GLOSSARY
Technical efficiency Conversion of physical inputs such as labour services

and raw materials or semi-finished goods into
outputs. Technical efficiency is determined by the
difference between the observed ratio of combined
quantities of an entity’s output to input and the ratio
achieved by best practice. It can be expressed as the
potential to increase quantities of outputs from given
quantities of inputs, or the potential to reduce the
quantities of inputs used in producing given quantities
of outputs.
Technical efficiency is affected by the size of
operations (scale efficiency) and by managerial
practices (non-scale technical efficiency). It is defined
independent of prices and costs.
Total factor Ratio of the quantity of all outputs to the quantity of
productivity (TFP) all inputs. TFP can be measured by an index of the
ratio of all outputs (weighted by revenue shares) to all
inputs (weighted by cost shares).
Yardstick Competition over performance levels where no
competition market exists for the goods or services concerned.
This competition relies on performance indicators.
16
1 IMPROVING THE PERFORMANCE OF
GOVERNMENT SERVICE PROVIDERS
Measuring the efficiency of government service provision can play an

important role in achieving improvements in the performance of
Australia’s human service delivery. Data envelopment analysis is a
technique that can be used to assist in identification of best practice
performance in the use of resources, highlight where the greatest gains
may be made from improvements in efficiency, and help agencies
achieve their potential.
1.1 Improving performance in government funded service

delivery
Government funded services contribute about 10 per cent to Australia’s gross
domestic product and affect the daily lives of all Australians. For example,
governments make available education and training, health and community
services; they maintain law and order; and they provide subsidised care and shelter
for people in need. Improvements in the performance of these government funded
‘human’ services have the potential to deliver major social and economic benefits
and to improve our capacity to address social needs more effectively.
Human service delivery performance is coming under increased scrutiny as part of
the ongoing process of microeconomic reform. In the absence of market
contestability, comparative performance reporting provides a way of introducing
competitive pressures for government service providers. Developing and reporting
performance indicators is crucial to identifying performance improvements, and
thus guiding decision making. Comparative performance reporting is typically
undertaken cooperatively to assist all participants to improve their performance; it
is not focused on determining if a specific level of performance is being achieved.
The Steering Committee for the Review of Commonwealth/State Service
Provision, established in 1993, is developing and reporting jurisdiction based
performance indicators for human services funded by governments. The
performance indicator reports of the Steering Committee, released in 1995 and
1997, have concentrated on the areas of health, public housing, education, justice
and community support services.
1
1.2 Who uses performance measures?

Performance varies in human service provision across jurisdictions and within
jurisdictions. Providing an indication of how much performance differs and which
organisations are the best performers is potentially of value to the providers of
funds and the clients of these services — members of the community — as well as
to those managing the service provision — governments, departments and service
providers. Comparable performance measures are of most value when they:
• are linked to service objectives and aspects of provision for which there is
responsibility and accountability; and
• relate to aspects of service provision for which there is limited competitive
market pressure. The resulting comparisons of performance measures, or
yardstick competition, can provide an alternative form of pressure for
improved performance.
Concerned citizens are able to use publicly available information on the
performance of different service providers to make governments more accountable
for the expenditure of taxpayer funds, and to exercise client choice more
effectively.
Governments can use performance measures to:
• stimulate policy development by highlighting the effect on the performance
of government determined aspects of the operating environment (for example,
client choice, extent of competition);
• facilitate monitoring of public sector managerial performance and improve
accountability within government;
• promote ‘yardstick competition’ by providing a means of comparing the
performances of those responsible for similar aspects of service provision
where there is little direct competition in input and/or output markets;
• analyse the relationships between agencies and between programs, to allow
governments to coordinate policy across agencies (for example, the
interrelationships between policing, courts and correctional services); and
• assist the resource allocation/budgeting process by providing a means of
allocating funding based on agreed plans for improved performance, rather
than on the assumption that performance should equal past levels.
Comparative performance measurement is also a powerful management tool for
both agency managers (such as department heads) and individual service provider
managers (for example, hospital or police station managers).
Managers can use performance measurement to:
• identify differences in performance; and
2
1 IMPROVING THE PERFORMANCE OF GOVERNMENT SERVICE PROVIDERS
• focus attention on other organisations which may be performing better. More

detailed exercises may help identify the practices being used by the other
organisations, along with the extent to which those practices could usefully be
adopted in the operations under review.
The objective of comparative performance measurement is to facilitate a program
to improve performance, not to provide a simple grading of service providers.
Identifying major gaps in performance can provide the impetus for an organisation
to fundamentally rethink how it does things. There has been much focus on the
‘continuous improvement’ approach to managing government organisations, but
this approach may limit managers to looking at only small changes to procedures.
If a quantum leap is needed, rather than an incremental change, then monitoring
can help managers focus on where a major overhaul of the organisation’s
operations may be required. Most organisations have at least some aspect of their
operations from which others can learn, and in the absence of direct competition,
sharing information is the best way of transferring best practice.
The process of performance measurement has the value of identifying performance
variations, and hence providing encouragement and direction for performance
improvement. There are also two wider benefits of the process which can be
equally important in supporting performance improvements.
First, measuring performance requires a clear understanding and articulation of the
resources being used, and the outputs being produced, in the process of providing a
service. Making the inputs and outputs transparent can allow a critical assessment
of why particular resources are being used to provide particular outputs, clarifying
service provision objectives and priorities.
Second, attempting to measure performance provides a heightened awareness of
data shortcomings for managers and policy makers. If data deficiencies are
catalogued and advertised the quality of data may be improved, and the ability to
better measure performance enhanced.
There may be a hesitancy to try new approaches to measuring the performance of
human services given concerns that data are not of sufficient quality. However, as
these points illustrate, a useful start can usually be made on performance
measurement with data that are currently available — waiting for the perfect data
may lead to extensive delay, and the use of available data is often a catalyst for
developing better quality data.
No single performance measure or technique can provide the whole answer;
quantitative analysis involves significant assumptions and limitations.
Consequently, it may be desirable to use the results of several approaches, both
quantitative and qualitative, to judge how a particular agency is performing overall
and what needs to be done. Thus, the Steering Committee is interested in the
3
development and application of new techniques and approaches to performance

measurement.
1.3 Concepts of effectiveness, efficiency and productivity

Assessing the performance of human service providers is a complex task. The
Steering Committee has developed a framework of effectiveness and efficiency
with which to assess government funded service provision.
Effectiveness is the extent to which outputs of service providers meet the objectives
set for them by governments. An example is whether hospitals meet the waiting
time targets set for elective surgery.
Efficiency is the success with which an organisation uses its resources to produce
outputs — that is the degree to which the observed use of resources to produce
outputs of a given quality matches the optimal use of resources to produce outputs
of a given quality. This can be assessed in terms of technical, allocative and
dynamic efficiency. (Definitions of the different types of efficiency are developed
in Chapter 2 and are included for reference in the Glossary.)
Improving the performance of government service relies on both efficiency and
effectiveness. A government service provider might increase its measured
efficiency at the expense of the effectiveness of its service. For example, a hospital
might reduce the inputs used for cleaning but service the same number of patients.
This could increase the apparent efficiency of the hospital but reduce its
effectiveness in providing satisfactory outcomes for patients. Therefore, it is
important to develop effectiveness indicators for government service providers. For
example, are patients being re-admitted to hospitals at unacceptable rates?
Effectiveness is more fully discussed in the Steering Committee report for 1997
(SCRCSSP 1997).
This paper focuses on the assessment of the technical efficiency of government
service providers. Technical efficiency is determined by the difference between the
observed ratio of combined quantities of an organisation’s output to input and the
ratio achieved by best practice. Producing the maximum output or consuming the
minimum inputs, as compared to what is technically feasible, is an essential step for
service providers to be able to best meet their objectives.1
1 See Pestieau and Tulkens (1993) for a fuller discussion of the relationship between
technical efficiency and the ability of public enterprises to achieve their objectives.
4
The technical efficiency of an organisation depends on its level of productivity, or

the ratio of its outputs to inputs.2 The following could lead to productivity
improvements:
• the adoption of technological advances; and/or
• increases in efficiency through for example the removal of unnecessarily
restrictive work practices or better management; and/or
• a change in the operating environment in which production occurs.
All agencies use a range of inputs, including labour, capital, land, fuel and
materials, to produce services. If an agency is not using its inputs in a technically
efficient manner, it is possible to increase the quantities of outputs without
increasing inputs, or to reduce the inputs being used to produce given quantities of
outputs.
The Steering Committee has begun the task of assessing efficiency by compiling
information on simple, partial productivity indicators — ratios of output to input
which do not include all outputs and inputs. Efficiency indicators reported so far
include, for example, measures of recurrent expenditure and staff and/or capital per
unit of a particular output.
Partial productivity measures and recurrent costs per unit of output are used widely
because they are simple to calculate, but they need to be interpreted with caution.
By definition these measures are always only partial in that they do not account for
the relationships and trade-offs between different inputs and outputs. This is a
significant limitation in their application to government service delivery, which
typically involves multiple inputs and outputs. For example, if labour inputs are
replaced by capital inputs, labour productivity is likely to increase while capital
productivity declines. To assess whether the agency has become more efficient
overall, output must be measured against both labour and capital inputs.
Several partial productivity measures and recurrent costs per unit of output may be
used collectively to obtain a broad picture of efficiency. However, the presentation
of a large number of partial measures will typically be difficult to comprehend and
interpret if some indicators move in opposite directions over a given period of time.
This reinforces the value of more comprehensive summary measures of efficiency.
Partial measures may provide important information on specific aspects of
operation, but it is important to see how the agency is performing overall relative to
comparable organisations producing similar outputs.
2 Productivity is an absolute concept, measured by the ratio of outputs to inputs, while

efficiency is a relative concept, measured by comparing the actual ratio of outputs to
inputs with the optimal ratio of outputs to inputs.
5
Given the shortcomings of partial indicators, governments are adopting more

comprehensive indicators of performance for government trading enterprises and
government service providers. In the case of government trading enterprises, the
techniques adopted include total factor productivity (TFP) indexing (see Box 1)
and data envelopment analysis (DEA).
Box 1: Performance measurement in government trading

enterprises
The Steering Committee on National Performance Monitoring of Government Trading
Enterprises was established following the July 1991 Special Premiers’ Conference to develop
a national performance monitoring scheme for government trading enterprises across the
Commonwealth, states and territories. The Committee has developed a consistent suite of
financial and non-financial performance indicators of efficiency and effectiveness, which are
reported by participating government trading enterprises annually (see SCNPMGTE 1996).
To measure the performance of a government trading enterprise, increasing use is being made
of total factor productivity (TFP) indexing — a procedure which combines all outputs and
inputs into a comprehensive measure of overall productivity. Important to this process, the
Steering Committee on National Performance Monitoring of Government Trading
Enterprises published a guide to using TFP and examples of its application in several case
studies (SCNPMGTE 1992).
An international benchmarking program for key Australian infrastructure industries, started at

the Bureau of Industry Economics in 1991 and now being undertaken by the Industry
Commission, has used DEA, TFP and partial productivity measures for benchmarking.
However, the TFP technique is not generally applicable to service provision,

because it requires a price for each output’s and inputs and output prices often
cannot be identified for many government services. Thus, DEA is being used more
for government service providers. As well as being able to handle multiple services
and inputs, DEA does not require information on the price of services or inputs,
making it particularly applicable to government service provision.
DEA is a linear programming technique that identifies the apparent best providers
of services by their ability to produce the highest level of services with a given set
of inputs, or to produce given services with the least amount of inputs. Other
service providers receive an efficiency score that is determined by their
performance relative to that of the best performers. The technique can also
determine whether the main source of inefficiency is the scale of operations or the
managerial capabilities and effort of the service provider. Further, it can
incorporate variables to account for environmental factors which might influence
6
the productivity of a service provider but which are beyond its control — for
example, the education or wealth of clients.
Like any empirical technique, DEA has limitations of which practitioners need to
be mindful (these are discussed in more detail in the following chapters). DEA
results provide the maximum benefit when they are interpreted with care. In
general, they should be considered as a starting point for assessing the efficiency of
the service providers within a sample. Indications of possible sources of relative
inefficiency can guide further investigation to determine why there are apparent
differences in performance. This information can be used to inform the managers of
individual service providers, administrators and policy makers.
Finally, it is important to recognise that performance measures will inevitably
evolve through time. Gaining experience in formulating and using the measures,
the agency will refine the set to better meet its requirements. Agencies might start
off with relatively simple measures and progress to more sophisticated measures as
they gain experience and as they initiate the collection of better quality data.
1.4 Objectives of this paper

The Steering Committee established a Working Group to promote the application
of DEA to government services. The DEA Working Group has produced this
information paper to further understanding of the technique and to outline its
application in a number of case studies focussing on how relative efficiency is
measured and on how results can be used to enhance performance.
The following chapter provides a relatively non-technical explanation of the
principles behind DEA. This is followed by a simple example of how to calculate
DEA in Chapter 3, and an overview of five case studies of the application of DEA
(highlighting practical issues encountered) in Chapter 4. Chapter 5 contains the
case studies, which cover Victorian hospitals, Queensland oral health services for
students, and NSW corrective services, police patrols, and motor registries.
Appendix A supports the second and third chapters, and Appendix B lists software
options for applying DEA.
Performance measurement of human service delivery is still in its infancy. Much
remains to be done in getting the necessary data systems consistently in place, and
in resolving the precise nature and value of many human service outputs and how
to measure them. DEA is useful for improving the performance of government
service delivery by advancing our understanding of key efficiency drivers and
identifying examples of good practice.
7
2 WHAT IS DATA ENVELOPMENT ANALYSIS?
Data envelopment analysis provides a means of calculating apparent

efficiency levels within a group of organisations. The efficiency of an
organisation is calculated relative to the group’s observed best
practice. This chapter explains the basic concepts behind DEA and
provides a simple graphical example. Some extensions to the DEA
model, allowing the sources of inefficiency to be identified, are also
discussed.
2.1 Data envelopment analysis and different efficiency

concepts
Typically using linear programming, DEA calculates the efficiency of an
organisation within a group relative to observed best practice within that group.
The organisations can be whole agencies (for example, Departments of Health),
separate entities within the agency (for example, hospitals) or disaggregated
business units within the separate entities (for example, wards).1
To discuss DEA in more detail it is necessary to look at the different concepts of
efficiency. The most common efficiency concept is technical efficiency: the
conversion of physical inputs (such as the services of employees and machines)
into outputs relative to best practice. In other words, given current technology,
there is no wastage of inputs whatsoever in producing the given quantity of
output. An organisation operating at best practice is said to be 100 per cent
technically efficient. If operating below best practice levels, then the
organisation’s technical efficiency is expressed as a percentage of best practice.
Managerial practices and the scale or size of operations affect technical
efficiency, which is based on engineering relationships but not on prices and
costs.
Allocative efficiency refers to whether inputs, for a given level of output and set
of input prices, are chosen to minimise the cost of production, assuming that the
organisation being examined is already fully technically efficient. Allocative
efficiency is also expressed as a percentage score, with a score of 100 per cent
indicating that the organisation is using its inputs in the proportions which
1 Given that DEA is particularly well suited to government service and other non-profit
organisations, as well as private sector firms, the individual units examined are often
referred to as decision-making units rather than firms.
9
would minimise costs. An organisation that is operating at best practice in

engineering terms could still be allocatively inefficient because it is not using
inputs in the proportions which minimise its costs, given relative input prices.
Finally, cost efficiency refers to the combination of technical and allocative
efficiency.2 An organisation will only be cost efficient if it is both technically
and allocatively efficient. Cost efficiency is calculated as the product of the
technical and allocative efficiency scores (expressed as a percentage), so an
organisation can only achieve a 100 per cent score in cost efficiency if it has
achieved 100 per cent in both technical and allocative efficiency.
Figure 2.1: Illustration of different efficiency concepts
Labour
• B
• A
Locus of points of
• A’ minimum input use
• needed to produce
A ’’ given output
C
•
Budget line
Capital
O
These concepts are best depicted graphically, as in Figure 2.1 which plots
different combinations of two inputs, labour and capital, required to produce a
given output quantity. The curve plotting the minimum amounts of the two
inputs required to produce the output quantity is known as an isoquant or
efficient frontier. It is a smooth curve representing theoretical best engineering
practice. Producers can gradually change input combinations given current
technological possibilities. If an organisation is producing at a point on the
2 Cost efficiency is sometimes extended to include a third measure called dynamic

efficiency: the degree to which producers respond to changes to technology and products
following changes in consumer preferences and productive opportunities.
10
isoquant then it is technically efficient. The straight line denoted as the budget
line plots combinations of the two inputs that have the same cost. The slope of
the budget line is given by the negative of the ratio of the capital price to the
labour price. Budget lines closer to the origin represent a lower total cost. Thus,
the cost of producing a given output quantity is minimised at the point where the
budget line is tangent to the isoquant. At this point both technical and allocative
efficiencies are attained.
The point of operation marked A would be technically inefficient because more
inputs are used than are needed to produce the level of output designated by the
isoquant. Point B is technically efficient but not cost efficient because the same
level of output could be produced at less cost at point C. Thus, if an
organisation moved from point A to point C its cost efficiency would increase
by (OA–OA'')/OA. This would consist of an improvement in technical
efficiency measured by the distance (OA–OA')/OA and an allocative efficiency
improvement measured by the distance (OA'–OA'')/OA'. Technical efficiency is
usually measured by checking whether inputs need to be reduced in equal
proportions to reach the frontier. This is known as a ‘radial contraction’ of
inputs because the point of operation moves along the line from the origin to
where the organisation is now.
2.2 Operationalising the concepts

The smooth curve in Figure 2.1 representing theoretical best practice typically
cannot be calculated from observed data. Rather, data usually are only available
on a group of organisations which gives limited information on theoretical best
practice. First, it is unknown whether any of the organisations in the group, or
sample, are achieving outright best practice. Second, the sample points will not
cover all of the range of possible input combinations.
There are several ways to use the data from the sample to try and approximate
the smooth curve in Figure 2.1. Early attempts used ordinary least squares
regression techniques, that plot an average curve through the sample points.
However, this was not satisfactory because an individual organisation’s
efficiency was compared with an average level of performance in the sample
rather than an estimate of best practice within the sample. This led to attempts to
approximate best practice in the sample by estimating frontiers.
The two techniques used to estimate the frontier are DEA and stochastic frontier
analysis. The focus in this report is on DEA, which is a deterministic means of
constructing a ‘piece-wise linear’ approximation to the smooth curve of Figure
2.1 based on the available sample. In simple terms, the distribution of sample
11
points is observed and a ‘kinked’ line is constructed around the outside of them,
‘enveloping’ them (hence the term data envelopment analysis).
Stochastic frontier analysis is an alternative approach using regression
techniques. It tries to take account of outliers which either are very atypical or
appear to be exceptional performers as a result of data measurement errors. The
relevance of stochastic frontier analysis to budget sector applications is limited
to those situations in which a single overall output measure or relatively
complete price data are available. This is not often the case for service
providers, so stochastic frontier analysis is not covered in this information
paper. (See Fried, Lovell and Schmidt 1993 for a discussion of stochastic
frontiers.)
DEA is often only used to calculate the technical efficiency of government
services. The DEA approach to calculating technical efficiency can be shown
with a simple numerical example: a sample of five hospitals that use two
inputs — nurses and beds — to produce one output — treated cases. Obviously
the inputs and outputs of a real hospital are considerably more complex, but this
simplification may be a good starting point for actual as well as illustrative
examples — for instance, the input ‘beds’ might serve as a proxy for the amount
of capital inputs used by the hospital. The hospitals are likely to be of differing
sizes; to facilitate comparisons, input levels must be converted to those needed
by each hospital to produce one treated case. The hospital input and output data
are presented in Table 2.1.
Table 2.1: Illustrative hospital data

Nurses per Beds per
Hospital Nurses Beds Treated cases treated case treated case
1 200 600 200 1 3
2 600 1200 300 2 4
3 200 200 100 2 2
4 600 300 200 3 1.5
5 500 200 100 5 2
The five hospitals range in size from 200 to 1200 beds, and there is a similarly
large range in the numbers of nurses, beds, treated cases, and nurses per treated
case and beds per treated case. Given the large discrepancies in the five
hospitals’ characteristics it is not obvious how to compare them or, if one is
found to be less efficient, which other hospital it should use as a role model to
improve its operations. The answers to these questions become clearer when the
data for nurses per treated case and beds per treated case are plotted in Figure
2.2, where data are abstracted from differences in size.
12
The hospitals closest to the origin and the two axes are the most efficient, so a
‘kinked’ frontier can be drawn from hospital 1 to hospital 3 to hospital 4. For
the moment, the parts of the frontier above hospital 1 and to the right of
hospital 4 are drawn by extending the frontier beyond these points parallel to the
respective axes. The kinked frontier in Figure 2.2 envelopes all the data points
and approximates the smooth isoquant in Figure 2.1 based on the information
available from the data.
Figure 2.2: Illustrative hospital input–output data
Beds per treated case

6
4 2
3 1
2’
2 5
3 4
5’
1
0
0 1 2 3 4 5 6
Nurses per treated case
Which are the most efficient or best practice hospitals in the sample? Hospitals
1, 3 and 4 are on the efficient frontier, so are assumed to be operating at best
practice. However, hospitals 2 and 5 are north-east of the frontier, so are
considered to be less efficient. This is because they appear to be able to reduce
their input use and still maintain their output level compared with the
performance of the best practice hospitals. For example, hospital 2 could reduce
its use of both inputs by one third before it would reach the efficient frontier at
point 2’. Similarly, its technical efficiency score is given by the ratio 02’/02
which is equal to 67 per cent in this case. This is because the ‘hypothetical’
hospital 2' has a value of 1.33 for nurses per treated case and a value of 2.67 for
13
beds per treated case. In terms of actual input levels, hospital 2 would have to
reduce its number of nurses from 600 to 400 and its number of beds from 1200
to 800. At the same time, it would have to maintain its output of 300 treated
cases before it would match the performance of the hypothetical best practice
hospital 2’.
But how is the hypothetical best practice hospital 2’ derived? It is formed by
reducing the inputs of hospital 2 in equal proportions until reaching the best
practice frontier. The frontier is reached between hospitals 1 and 3 in this case,
so the hypothetical hospital 2’ is a combination, or weighted average, of the
operations of hospitals 1 and 3. If hospital 2 is looking for other hospitals to use
as role models to improve performance, then it should examine the operations of
hospitals 1 and 3 because these are the efficient hospitals most similar to itself.
In DEA studies these role models are known as the organisation’s ‘peers’.3
The other less efficient hospital — hospital 5 — is in a different situation. It is
north-east of the efficient frontier, but contracting its inputs in equal proportions
leads to the hypothetical hospital 5', which still lies to the right of hospital 4 on
the segment of the frontier which was extended parallel to the nurses per treated
case axis. Thus, the peer group for hospital 5 solely consists of hospital 4
because it is the only one which ‘supports’ that section of the frontier on which
the hypothetical 5' lies. But hospital 5' is not fully efficient because the number
of nurses per treated case can be reduced, while the number of beds per treated
case is held constant, thus moving from 5' back to 4. That is, to maximise its
efficiency given the available data, hospital 5 has to reduce one input more than
the other. In this special case, a radial contraction of inputs means that the
frontier is reached, but a further reduction of one of the inputs can be achieved
without a reduction in output. This extra input reduction available is known in
DEA studies as an input ‘slack’. Thus, it is important in DEA studies to check
for the presence of slacks as well as the size of the efficiency score.
It is relatively easy to implement this simple example of data envelopment
analysis in a two-dimensional diagram. However, with a larger number of inputs
and outputs and more organisations, it is necessary to use mathematical
formulae and computer packages. Using the same principles, an example of how
to implement a more complex analysis is given in Chapter 4 and the technical
details behind DEA are briefly presented in Appendix A. Before moving on to
look at some extensions to the basic DEA model outlined above, some of the
questions DEA can help agency managers answer are briefly reviewed.
3 The term ‘peers’ in DEA has a slightly different meaning to the common use of the word
peer. It refers to the group of best practice organisations with which a relatively less
efficient organisation is compared.
14
2.3 What questions can DEA help us answer?

By providing the observed efficiencies of individual agencies, DEA may help
identify possible benchmarks towards which performance can be targeted. The
weighted combinations of peers, and the peers themselves may provide
benchmarks for relatively less efficient organisations. The actual levels of input
use or output production of efficient organisations (or a combination of efficient
organisations) can serve as specific targets for less efficient organisations, while
the processes of benchmark organisations can be promulgated for the
information of managers of organisations aiming to improve performance.
The ability of DEA to identify possible peers or role models as well as simple
efficiency scores gives it an edge over other measures such as total factor
productivity indices.
Fried and Lovell (1994) listed the following as questions that DEA can help to
answer for managers:
• How do I select appropriate role models to serve as possible benchmarks
for a program of performance improvement?
• Which production facilities are the most efficient in my organisation?
• If all my operations were to perform according to best practice, how many
more service outputs could I produce and how much could I reduce my
resource inputs by, and in what areas?
• What are the characteristics of efficient operating facilities and how can
they guide me in choosing locations for expansion?
• What is the optimum scale for my operations and how much would I save
if all my facilities were the optimum size?
• How do I account for differences in external circumstances in evaluating
the performance of individual operating facilities?
The simple model of DEA already outlined can satisfy the first four of these
questions. To answer the last two, some extensions to the model are needed.
2.4 Extensions to the DEA model

By making the DEA model a little more complicated, the range of topics it can
explore is increased. Particularly interesting is the decomposition of the
technical efficiency score into components resulting from: the scale of
operations; surplus inputs which cannot be disposed of; and a residual or ‘pure’
technical efficiency. A further extension which is often important is to allow for
differences in operating environments; this involves trying to adjust for factors
15
which might be beyond managers’ control, and which thus possibly give some
organisations an artificial advantage or disadvantage. Each of these issues is
addressed in turn below. A technical treatment of these topics is presented in
Appendix A.
2.4.1 Scale efficiency

The simple example presented in Section 2.2 was based on the assumption of
constant returns to scale. Given this assumption, the size of the organisation is
not considered to be relevant in assessing its relative efficiency. Small
organisations can produce outputs with the same ratios of input to output as can
larger organisations. This is because there are no economies (or diseconomies)
of scale present, so doubling all inputs will generally lead to a doubling in all
outputs. However, this assumption is inappropriate for services which have
economies of scale (or increasing returns to scale). In these services, doubling
all inputs should lead to more than a doubling of output because producers are
able to spread their overheads more effectively or take advantage of purchasing
items in bulk. For other services, organisations might become too large and
diseconomies of scale (or decreasing returns to scale) could set in. In this case, a
doubling of all inputs will lead to less than a doubling of outputs. It would be to
an agency’s advantage to ensure that its operations are of optimal size — neither
too small if there are increasing returns nor too large if there are decreasing
returns to scale.
If it is likely that the size of service providers will influence their ability to
produce services efficiently, the assumption of constant returns to scale is
inappropriate. The less restrictive variable returns to scale frontier allows the
best practice level of outputs to inputs to vary with the size of the organisations
in the sample. This is demonstrated using the simplified one input (medical
staff), one output (treated cases) example shown in Figure 2.3.
16
The constant returns to scale frontier is the straight line emanating from the
origin (OBX), determined by the highest achievable ratio of outputs to inputs in
the sample, regardless of size. The variable returns to scale frontier (VAABCD)
passes through the points where the hospitals have the highest output to input
ratios, given their relative size, then runs parallel to the respective axes beyond
the extreme points. The scale efficiency of an organisation can be determined by
comparing the technical efficiency scores of each service producer under
constant returns to scale and variable returns to scale.
The distance from the respective frontier determines technical efficiency under
each assumption. The distance between the constant returns and the variable
returns frontiers determines the scale efficiency component. Technical
efficiency resulting from factors other than scale is determined by the distance
from the variable returns frontier. Thus, when efficiency is assessed under the
assumption of variable returns, the efficiency scores for each organisation
indicate only technical inefficiency resulting from non-scale factors. Technical
efficiency scores calculated under variable returns, therefore, will be higher than
or equal to those obtained under constant returns.
Figure 2.3: The production frontier and returns to scale
Treated cases
X
Ev
O
w wD
m C
Bw
E
Ec
T mm wE
Ev
A
Ac
T m w
A
O A A M
E
C V Medical staff
17
This can be demonstrated using the examples in Figure 2.3.

• Hospital B is the only one that has no scale or non-scale inefficiency under
either assumption. It represents the optimal scale within the sample.
• Hospitals A, C and D are scale inefficient but do not have any inefficiency
resulting from non-scale factors under the variable returns assumption. For
example, the scale efficiency score of hospital A is determined by the ratio
of distances TAAc/TAA, which is less than one. Hospital A has increasing
returns to scale because it would approach the optimal scale in the sample
if it increased its size. Hospitals C and D are producing outputs with
decreasing returns to scale and are too large to be considered scale
efficient, with hospital D being the furthest from optimal scale.
• The technical inefficiency of hospital E under constant returns (TEEc/TEE)
is made up of both scale inefficiency (TEEc/TEEv) and non-scale technical
inefficiency (TEEv/TEE).
2.4.2 Input and output orientation

Another issue that can be illustrated in Figure 2.3 is the question of output and
input orientation. The examples so far have been input oriented — that is, by
how much can inputs be reduced while maintaining the same level of output?
However, the corresponding output-oriented question could be equally
important — by how much can output be increased while keeping the level of
inputs constant? The latter question may be more relevant for many government
service providers, particularly those supplying human services, as the
community often wants more of these services while budgetary pressures make
it difficult to increase inputs.
In Figure 2.3 the input-oriented technical efficiency score for hospital E under
variable returns to scale was given by the ratio of distances TEEv/TEE. The
technical efficiency score for hospital E, using an output orientation and again
assuming variable returns to scale, is given by the ratio of distances MEE/MEEvO.
If an organisation is technically inefficient from an input-oriented perspective,
then it will also be technically inefficient from an output-oriented perspective.
However, the values of the two technical efficiency scores typically will differ,
as will the presence and extent of slacks.
Depending on whether an input-saving or output-expanding orientation is
utilised, the peers for hospital E will also differ. Its peers are hospitals A and B
under input orientation but hospitals B and C under output orientation. This
reflects the fact that hospital E can learn different things from the two sets of
peers. Hospital C is better at producing more output from a roughly similar
18
input level to that of hospital E, while hospital A produces less output than does
hospital E but uses considerably fewer inputs.
2.4.3 Input congestion

In many situations agencies will not be able to dispose of ‘unwanted’ inputs. For
example, union restrictions may inhibit reductions in the workforce, or
government controls may prohibit reducing certain inputs. To cover situations
such as these, Färe, Grosskopf and Lovell (1983) introduced the concept of
congestion efficiency (See Figure 2.4). In the previous examples, the constant
returns to scale isoquant is eventually parallel to both axes (as in Figure 2.2).
This reflects the assumption that an input that cannot be profitably used can be
disposed of at no cost. In contrast, under congestion, the inability to dispose of
unwanted inputs increases costs.
Figure 2.4: Measuring input congestion

Capital
A
•
B
•
Variable returns to scale isoquant
with congestion
• D
• Variable returns to scale isoquant
•
C with costless disposal
0 Labour
In Figure 2.4, congestion is assumed to be present in the use of labour. Thus,

instead of the frontier eventually running parallel to the labour axis, congestion
is reflected in a frontier which slopes upwards from the labour axis. In Figure
2.4, an organisation operating at point B would have congestion inefficiency
equal to OC/OD, whereas a firm operating at point A would be congestion
efficient.
19
After decomposing the constant returns technical efficiency score into

components resulting from scale efficiency and congestion efficiency, a residual
or ‘pure’ technical efficiency score remains. This residual score largely indicates
the scope for efficiency improvements resulting from less efficient work
practices and poor management, but may also reflect differences between
operating environments.
2.4.4 Adjusting for operating environments

The environment in which a service provider operates can have an important
influence on its relative performance if other providers are operating in different
environments. Many of these operating environment factors are not under the
control of managers, and ignoring them in assessing performance may lead to
spurious results. Climate, topography, the socioeconomic status of a
neighbourhood, government restrictions and the degree of unionisation, for
example, can affect performance but be beyond management control.
The efficiency score of a police station in a poor area, for example, may not be
comparable with the score of a police station in a well-to-do area. This may be
misleading if the level of crime is lower in well-to-do neighbourhoods and if the
police stations’ output is measured by the number of incidents attended and
arrests. Thus, it may be important to adjust for the impact of the respective
socioeconomic environments on incidents attended and arrests. Only then may it
be possible to determine which police station is being more successful at
transforming inputs (the number of police and cars) into outputs (the number of
incidents attended and arrests).
But how to adjust for these differences in operating environments which are
beyond management control? The main options in DEA studies are to:
• only compare organisations which operate in a similar operating
environment. This may sound attractive but it often dramatically reduces
the potential number of comparison partners and, hence, much information
is lost: for instance, the main lessons may come from organisations that
operate differently in a dissimilar environment;
• only compare the organisation with other organisations in a similar or less
favourable operating environment. This overcomes some of the problems
of the preceding method but still ignores potential lessons from more
favourable operating environments;
• include the operating characteristic as part of the DEA calculation. This
method is useful where the direction of influence of the characteristic is
obvious. However, the characteristic has to be a continuous variable and,
20
by including more variables in the analysis, the efficiency scores tend to be

automatically inflated. This method was used by the Bureau of Industry
Economics (BIE 1994b) to adjust for the effects of climate on the observed
efficiency of gas suppliers in different countries. Suppliers in colder
climates have a higher demand for gas and are able to achieve better
capital use than suppliers in warmer countries. By including a measure of
degree-days — the number of days per year which deviate from an average
temperature by more than a specified range — the Bureau was able to
make more like-with-like comparisons;
• employ a two-stage procedure which uses econometric methods to
estimate the relationship between the characteristic and the efficiency
scores.4 The efficiency scores can be adjusted on the basis of this
relationship. The advantages of this approach are that it can accommodate
several characteristics, makes no prior assumptions about the direction of
influence, and allows for tests of statistical significance — something
which usually is not possible in DEA studies. The technique is similar to
that used in the NSW case studies of police services and motor registries
presented later in this report.
Some of these adjustment methods are examined in more detail in Appendix A.
2.5 Advantages and limitations of DEA

The main advantages of DEA are that:
• it can readily incorporate multiple inputs and outputs and, to calculate
technical efficiency, only requires information on output and input
quantities (not prices). This makes it particularly suitable for analysing the
efficiency of government service providers, especially those providing
human services where it is difficult or impossible to assign prices to many
of the outputs;
• possible sources of inefficiency can be determined as well as efficiency
levels. It provides a means of ‘decomposing’ economic inefficiency into
technical and allocative inefficiency. Furthermore, it also allows technical
inefficiency to be decomposed into scale effects, the effects of unwanted
inputs which the agency cannot dispose of, and a residual component;
4 The efficiency scores have a truncated distribution between zero and one, so it is necessary
to use Tobit rather than ordinary least squares regression techniques. (See the NSW Police
Patrols case study in Chapter 5 for an explanation of the regression techniques.)
21
• by identifying the ‘peers’ for organisations which are not observed to be

efficient, it provides a set of potential role models that an organisation can
look to, in the first instance, for ways of improving its operations. This
makes DEA a potentially useful tool for benchmarking and change
implementation programs. This role is strengthened by DEA’s ability to
incorporate differences in operating environments beyond management
control and, thus, to make more like-with-like comparisons.
However, like any empirical technique, DEA is based on a number of
simplifying assumptions that need to be acknowledged when interpreting the
results of DEA studies. DEA’s main limitations include the following.
• Being a deterministic rather than statistical technique, DEA produces
results that are particularly sensitive to measurement error. If one
organisation’s inputs are understated or its outputs overstated, then that
organisation can become an outlier that significantly distorts the shape of
the frontier and reduces the efficiency scores of nearby organisations. In
regression–based studies, the presence of error terms in the estimation
tends to discount the impact of outliers, but in DEA they are given equal
weight to that of all other organisations. It is important to screen for
potential outliers when assembling the data. One useful check is to
scrutinise those organisations whose output-to-input ratios lie more than
about two-and-a-half standard deviations from the sample mean. This
approach is used in some of the case studies presented later in the report.
• DEA only measures efficiency relative to best practice within the
particular sample. Thus, it is not meaningful to compare the scores
between two different studies because differences in best practice between
the samples are unknown. Similarly, a DEA study that only includes
observations from within the state or nation cannot tell us how those
observations compare with national or international best practice.
• DEA scores are sensitive to input and output specification and the size of
the sample. Increasing the sample size will tend to reduce the average
efficiency score, because including more organisations provides greater
scope for DEA to find similar comparison partners. Conversely, including
too few organisations relative to the number of outputs and inputs can
artificially inflate the efficiency scores. Increasing the number of outputs
and inputs included without increasing the number of organisations will
tend to increase efficiency scores on average. This is because the number
of dimensions in which a particular organisation can be relatively unique
(and, thus, in which it will not have similar comparison partners) is
increased. DEA gives the benefit of the doubt to organisations that do not
have similar comparison organisations, so they are considered efficient by
22
default. There are different rules as to what the minimum number of

organisations in the sample should be; one rule is that the number of
organisations in the sample should be at least three times greater than the
sum of the number of outputs and inputs included in the specification
(Nunamaker 1985).
Despite these limitations, data envelopment analysis is a useful tool for
examining the efficiency of government service providers. Just as these
limitations must be recognised, so must the potential benefits of using DEA (in
conjunction with other measures) be explored to increase our understanding of
public sector performance and potential ways of improving it.
23
3 HOW DO YOU CALCULATE DEA?
The focus in this chapter is on key considerations involved in

specifying the outputs and inputs attributed to each organisation in a
DEA study and its most appropriate coverage. The simple hospitals
example of the previous chapter is extended to illustrate how to apply
the DEA linear programming formulae. Key results provided by the
DEA output and their uses are also discussed.
The simple example presented in the preceding chapter of five hospitals
producing one output (treated cases) using two inputs (nurses and beds) served
to illustrate the basic concepts behind DEA. It was also sufficiently simple to
solve graphically. However, this simple method is inappropriate for more
realistic and, consequently, more complex situations. Once the number of inputs
and outputs and the number of organisations in the sample increases, linear
programming methods are needed to calculate DEA.
This chapter expands the hospital example — twenty hospitals, two outputs and
two inputs — to illustrate how these linear programming techniques are used.
This sample is still much simpler than most actual studies would be, but remains
of manageable size for the purpose.
3.1 Specifying outputs, inputs and coverage

The first step in a DEA study is deciding on its most appropriate scope.
Sometimes the study’s most important contribution is providing managers with
the discipline of having to specify their outputs and inputs and how they can
best be measured.
It is essential to include managers with a sound understanding of the process of
the organisations being examined from the early stages of model development.
The measures of outputs and inputs should be as comprehensive as possible: not
including some output dimensions will disadvantage those organisations which
are relatively efficient at producing those outputs. Important trade-offs
invariably have to be made. Including too many different outputs and inputs —
particularly if there are not many organisations in the sample — will tend to
inflate the efficiency scores because there is more scope for each organisation to
be relatively unique. The organisations might then be considered efficient by
default. As a result, the ideal selection includes the smallest number of output
25
and input measures that adequately captures all essential aspects of the
organisation’s operations.
The process of developing a final model of service production is often iterative,
with different combinations of inputs and outputs, and sometimes measures of
inputs and outputs, trialed before a final model is reached. This ensures the most
appropriate measures, and inputs and outputs, are utilised in the assessment of
relative efficiency and also allows the sensitivity of the model to different
specifications to be tested.
Outputs
Government agencies deliver a wide range of outputs, and it can be difficult to
specify all of them and to account for differences in their quality. However,
outputs of service deliverers can generally be classified into those that are
proactive and those that are reactive.
• Reactive outputs are often those most readily associated with a particular
service — for example, police attending a crime scene, or a hospital
providing treatment for admitted patients.
• Proactive outputs are often equally as important in the delivery of the
service, but less readily identified and measured — for example, time
spent by police gaining the confidence of their community, or a hospital
providing an education and immunisation program. Proactive outputs are
also related to providing a contingent capability for the community — for
example, hospitals providing casualty departments to respond to and cope
with unexpected accidents and natural disasters.
Both the reactive and proactive outputs should be taken into account. The
quality of the outputs provided, relative to that of other providers in the sample,
should also be considered in any efficiency assessment, or managers may be
able to increase apparent efficiency at the expense of output quality. This is in
addition to the need to assess the effectiveness of the overall service being
provided (discussed in Section 1.3).
The quality of reactive outputs and the level and quality of proactive outputs are
often reflected in the outcomes achieved by the service overall — for example:
the degree to which a community feels safe within a particular area will reflect
the quality of police reactions to incidents and crimes, and the degree to which
police have gained the community’s confidence; the quality of treatment in a
hospital can be reflected in the proportion of patients returned to surgery
unexpectedly; and the output and quality of an education and immunisation
26
program is likely to be reflected in a reduced incidence of the targeted disease in

the community.
Outcome indicators are often associated with the effectiveness of a service, but
it is possible to use them in a DEA assessment of efficiency where it is not
possible to measure the proactive outputs of service providers directly. They are
also useful where there is scope for differences in the quality of outputs, to
ensure that quality is not ignored in the efficiency analysis.
In the hospital example being used to illustrate DEA concepts, simply using the
number of treated cases (as in the preceding chapter) will not adequately capture
the full scope of the hospital’s role. It is only a measure of the reactive part of
the hospital’s contribution to the community without accounting for the
proactive side in terms of education, immunisation services and provision of a
contingent capability. Concentrating on the reactive output side is unlikely to be
adequate. Examples of outputs used in DEA studies which aim to capture the
quality of service provision outputs are included in the case studies of hospitals,
police patrols and motor registry offices (See Chapter 5).
The functions of hospitals differ markedly, with some providing basic services
and others providing more resource-intensive specialist care. In efficiency
comparisons, ignoring the fact that some hospitals provide more intensive care
for acute cases would disadvantage a small country hospital, for example, which
only provides care for non-acute cases and transfers its acute cases to larger
metropolitan hospitals. To account for this aspect of hospital operations, the
scenario’s one output (the total number of treated cases) is changed to two
outputs (the number of non-acute cases and the number of acute cases). This
will produce more like-with-like comparisons. To keep the example simple, the
proactive dimension, which could involve a measure of community health
levels, for example, is omitted.
Labour
The desirable measure of labour inputs is that which most accurately reflects the
amount of labour used by each organisation. Total hours worked might be the
most suitable measure in many cases. However, many organisations do not keep
records of hours worked, so the number of full-time equivalent staff is often the
best available measure. Both measures are preferable to the simple number of
persons employed, which may be misleading if the average number of hours
worked per employee varies considerably between the organisations.
However, physical measures of labour input do not capture differences in the
quality of labour. This can be addressed by disaggregating the number of hours
or full-time equivalents into different types of labour, such as administrative and
27
operational. In the example, the labour input is measured by the number of full-
time equivalent nursing staff.
An alternative to using a direct measure of the quantity of labour input is to
deflate each organisation’s total labour costs by an appropriate labour price. To
be accurate, this approach requires a good estimate of the average labour price
each organisation faces: for example, an organisation that must pay overtime to
employees may have relatively higher labour costs than an organisation that
does not.
Capital
Measures of capital input are subject to considerable variation and can be a
potential weakness in efficiency measures. There is little consensus on the best
means of calculating the price and quantity (and thus cost) of capital inputs in
any one period. This is a particularly important issue for those government
business enterprises where capital inputs generally account for a large
proportion of production inputs. Capital inputs may also be relatively important
for many government service providers such as hospitals and schools.
The difficulty in measuring capital inputs is that the entire cost of a durable
input purchased in one accounting period cannot be charged against that
period’s income. Rather, the capital item will provide a flow of services over a
number of years. How much of the purchase price should be charged to each
period then has to be determined, along with how interest and depreciation costs
should be allocated.
There are a variety of methods for calculating the annual cost of capital and the
quantity of capital input. The declining balance method is often used in
government business enterprise studies, and relies on having an accurate market
valuation of the organisation’s assets at one point in time (see Salerian and
Jomini 1994). However, many government service providers often have little
information available on the value of their capital assets. As a result, many
government service efficiency studies rely on simple measures of the overall
capital used by each organisation. If possible, the capital measures used should
provide some insights into the sources of inefficiency that may be associated
with the use of capital inputs. This could include purchasing too large a quantity
of capital, paying too high a price for capital, purchasing the wrong type of
capital, or using an incorrect mix of other inputs with the capital available.
In the hospital example, the number of beds in the hospital was initially a proxy
for the hospital’s total capital inputs — buildings, land, operating theatres, x-ray
equipment and so on. Clearly, this is not a very accurate proxy, but such simple
measures are a useful starting point in many government service studies,
28
provided their limitations are recognised. As noted in Chapter 1, using available

data to start the process is often the best catalyst for ensuring mechanisms are
put in place to systematically collect the data necessary to construct better
measures. For the purposes of illustrating how to calculate DEA, the number of
beds per hospital is still used to approximate each hospital’s overall capital
inputs.
Materials and services inputs

Ideally, the analysis should account for all the inputs used by each organisation,
just as it should measure all aspects of their output. As well as labour and
capital inputs, all organisations use a range of materials and services varying
from electricity to pharmaceuticals in hospitals, food in prisons, and electricity
to run computers in agency offices. These miscellaneous items are usually
aggregated in efficiency studies and deflated by a representative price index.
Ideally, the price index should account for differences in the prices faced by
each organisation — otherwise, those organisations facing relatively high prices
will be disadvantaged because expenditure of a given number of dollars will be
translated into a larger input quantity using an average price.
In the hospital example, materials and services costs are not included separately.
This is equivalent to assuming that they are used in fixed proportions to the
labour and capital inputs.
Coverage
The coverage of a DEA efficiency study depends on the overall aims of the
study, the availability of potential comparison partners, and the availability of
data. Inevitably, trade-offs have to be made and some degree of pragmatism is
always required. If an organisation is sufficiently large it may choose to start
with an in-house study measuring the efficiency of different business units
performing similar functions — for example, different hospitals within a health
department. Alternatively, comparisons could be made at a more aggregate level
but this would normally involve including similar organisations in different
jurisdictions and/or countries.
Ideally, the more organisations included in the sample the better the explanatory
power of the DEA model — there will be fewer organisations found efficient by
default. Typically, there will also be more to learn from including a more
diverse range of organisations. However, the cost of possibly including too
much diversity is that comparisons may no longer be sufficiently like-with-like.
This may require adjustment for differences in operating environments to ensure
that the study is both fair and credible.
29
The appropriate scope of a study is usually a matter of what type of organisation

is involved. A study being undertaken by an agency itself may concentrate on
individual processes in detail, whereas one undertaken by a government
monitoring agency may concentrate on overall performance at the aggregate
level. In all cases, three things should be kept in mind. First, it is often better to
start with the available information, rather than waiting for the perfect data set
(although data needs to be reliable for valid conclusions to be drawn). Second,
the limitations of the study should always be remembered, and the specification
should be refined if necessary. Third, DEA is only one of a number of
techniques that can be used in assessing overall performance.
3.2 DEA formula and a simple example

The remainder of this chapter contains illustrations of how to apply DEA to an
extended data set, presenting the constant and variable returns to scale cases and
calculating scale efficiency scores for each of the twenty hospitals.
There are several different ways to present the linear programming problem for
DEA. The formulae for other DEA extensions, including input congestion and
allocative efficiency, are shown in Appendix A. In most cases, they involve
relatively straightforward modifications to the basic formulae described here.
The simplest general presentation for the version of DEA where assumptions
include constant returns to scale, and an objective of minimising inputs for a
given level of output (an input-orientated version of DEA), proceeds by solving
a sequence of linear programming problems:
(1) Minimise En with respect to w1, ...wN, En
subject to:
∑
N
j =1
w j yij − yin ≥ 0 i = 1,K, I
∑
N
j =1
w j x kj − En xkn ≤ 0 k = 1,K, K
wj ≥ 0 j = 1,K, N
where there are N organisations in the sample producing I different outputs (yin
denotes the observed amount of output i for organisation n) and using K
different inputs (xkn denotes the observed amount of input k for organisation n).
The wj are weights applied across the N organisations. When the nth linear
program is solved, these weights allow the most efficient method of producing
organisation n’s outputs to be determined. The efficiency score for the nth
30
organisation, En*, is the smallest number En which satisfies the three sets of
constraints listed above. For a full set of efficiency scores, this problem has to
be solved N times — once for each organisation in the sample.
This seems a daunting formula: does it really make any intuitive sense? The
less than transparent nature of the DEA formula has contributed to DEA’s
reputation as being a bit of a ‘black box’ which people have trouble
understanding — and the above formula is one of the simpler ways of
presenting it! But it does make intuitive sense once the maths is penetrated.
The above formula is saying that the efficiency score for the nth organisation
should be minimised subject to a number of constraints. The factors that can be
varied to do this are the weights wj and the score En itself. The weights are used
to form the hypothetical organisation lying on the frontier. The constraints are
that the weighted average of the other organisations must produce at least as
much of each output, as does organisation n (the first set of constraints above),
while not using any more of any input than does organisation n (the second set
of constraints above). The third set of constraints simply limits the weights to
being either zero or positive.
Relating this back to the simple diagram in Figure 2.2, the process is simply one
of looking at all the possible combinations of weights on the other organisations
that will produce a point on the frontier such as 2'. The efficiency score is being
minimised because it represents the smallest proportion of existing inputs that
organisation n can use and still produce its existing output if it was using the
best practice observed in the sample. It is desirable to be as close to the origin as
possible to ensure being on the frontier: that is, both the weights and the
efficiency scores are systematically varied to contract each organisation as close
to the origin as possible while the contracted point is still a weighted average of
some of the other organisations. Thus, point 2 can be contracted as far as point
2': closer to the origin than 2', the point cannot be formed as a weighted average
of any of the other points and is not feasible. In the example in Figure 2.2, this
gave hospital 2 an efficiency score of 67 per cent. Points 1, 3 and 4 cannot be
contracted any closer to the origin while remaining weighted averages of other
points, so they achieve efficiency scores of 100 per cent.
Extended hospital data set

How does this apply to the expanded hospitals example (Table 3.1)? The two
outputs are the numbers of minor and acute treated cases, while the two inputs
remain the numbers of (full-time equivalent) nursing staff and beds.
31
Table 3.1: Two output, two input hospital data

Hospital number Minor cases Acute cases Nurses Beds
1 150 50 200 600
2 225 75 600 1200
3 90 10 200 200
4 160 40 600 300
5 50 50 500 200
6 75 75 320 150
7 200 50 375 450
8 350 100 400 320
9 400 90 550 480
10 250 300 900 660
11 350 350 850 720
12 350 400 720 940
13 275 375 900 850
14 220 40 250 370
15 300 10 115 250
16 320 275 600 590
17 375 230 550 710
18 230 50 200 280
19 290 90 450 410
20 360 70 415 575
The DEA formula for the first hospital in the two output, two input, twenty
hospitals example (data listed above) would be:
(2) Minimise E1 with respect to w1, w2, … , w20 and E1
subject to:
150w1 + 225w2 + 90w3 + … + 230w18 + 290w19 + 360w20 – 150 ≥ 0
50w1 + 75w2 + 10w3 + … + 50w18 + 90w19 + 70w20 – 50 ≥ 0
200w1 + 600w2 + 200w3 + … + 200w18 + 450w19 + 415w20 – 200E1 ≤ 0
600w1 + 1200w2 + 200w3 + … + 280w18 + 410w19 + 575w20 – 600E1 ≤ 0
w1 ≥ 0, w2 ≥ 0, w3 ≥ 0, … , w18 ≥ 0, w19 ≥ 0, w20 ≥ 0
The first constraint requires that the weighted average of the output of minor
treated cases, less hospital 1’s output of 150 minor treated cases, be greater than
or equal to zero. This means that the hypothetical frontier hospital for hospital 1
has to produce at least 150 minor treated cases. Similarly, the second constraint
requires that the frontier hospital for hospital 1 produce at least fifty acute
32
treated cases. The third and fourth constraints require the hypothetical hospital
to not use any more than hospital 1’s 200 nurses and 600 beds, respectively.
Solving this system of equations is not trivial and requires a computer program.
A number of specialised and general computer packages can be used to conduct
data envelopment analysis (see Appendix B).
The results obtained from solving this DEA problem are presented in Table 3.2.
The efficiency scores estimate the extent to which both inputs would need to be
reduced in equal proportions to reach the production frontier. In addition, for
some hospitals, after both inputs have been reduced in equal proportions, one
input could be reduced still further without reducing output (these are referred
to as ‘slacks’ in the DEA literature).1 The table also contains the peer group for
each hospital, the peer weights and the peer count — the number of times this
hospital appears in the peer group of other hospitals (excluding itself).
Hospital 1 obtains an efficiency score of 0.63 or 63 per cent (see Table 3.2).
That means that it appears that it could be able to reduce its number of nurses
and beds by 37 per cent and still produce its 150 minor treated cases and fifty
acute treated cases to operate at observed best practice. In practical terms, this
means that hospital 1 would have to reduce its number of nurses by 75 to a new
total of 125 and its number of beds by 224 to a new total of 376. The peer group
and peer weights columns indicate that the best practice for hospital 1 is given
by a weighted average of 80 per cent of hospital 15 and 20 per cent of hospital
12. However, as evident from the input slack columns, as well as reducing both
nurses and beds by 37 per cent, hospital 1 has an additional 176 beds. That
means that to remove all the apparent waste and inefficiency relative to
hospitals 15 and 12, hospital 1 would appear to have to reduce its number of
beds to a new total of 200.
Overall, six hospitals achieve efficiency scores of 100 per cent. It is evident
from the peer count column that all of the apparently efficient hospitals appear
in peer groups for other hospitals (and thus, none are efficient by default).
However, it is far more likely that hospitals 15, 8, and 16 are truly efficient
because they are peers for seven or more other hospitals in the sample. Hospitals
6, 11 and 12 each appear in only two or three peer groups, so there could be
scope for them to improve their efficiency further even though they receive
efficiency scores of 100 per cent.
1 In the example above, the model is run with the assumption that the objective is to
minimise inputs for a given level of output. If the model is run with the assumption that the
objective is to maximise output then slacks would reflect the amount that an output can be
increased, after all outputs have been increased in equal proportions to reach the
production frontier (see Figure 2.2).
33
Table 3.2: Constant returns to scale DEA results for the twenty
hospitals
Hospital Efficiency Labour Beds Peer Peer Peer
number score slacks slacks group weights count
1 0.63 0 176 15, 12 0.4, 0.1 0
2 0.31 0 76 15, 12 0.5, 0.2 0
3 0.39 22 0 15, 8 0.2, 0.1 0
4 0.48 123 0 15, 8 0.1, 0.4 0
5 0.50 37 0 6 0.7 0
6 1.00 0 0 6 1 2
7 0.46 0 0 8, 15, 16 0.2, 0.4, 0.1 0
8 1.00 0 0 8 1 8
9 0.75 26 0 15, 8 0.3, 0.9 0
10 0.93 0 0 11, 6 0.7, 0.8 0
11 1.00 0 0 11 1 2
12 1.00 0 0 12 1 3
13 0.94 0 0 16, 11 1.0, 0.3 0
14 0.59 0 0 15, 16, 8 0.6, 0.1, 0.1 0
15 1.00 0 0 15 1 11
16 1.00 0 0 16 1 7
17 0.90 0 0 16, 15, 12 0.3, 0.5, 0.4 0
18 0.85 0 0 8, 16, 15 0.1, 0.1, 0.6 0
19 0.71 0 0 8, 15, 16 0.6, 0.2, 0.1 0
20 0.62 0 0 15, 16, 8 0.8, 0.2, 0.2 0
At the other end of the spectrum, with the lowest observed efficiency, hospital 2
appears from the data in Table 3.1 to be grossly over-resourced relative to its
output. It has the highest number of beds by far and the fifth equal highest
number of nurses but only produces a modest number of minor and acute treated
cases. However, it is less obvious from the raw data that the hospital with the
second lowest efficiency score — hospital 3 — would be a poor performer
because it is considerably smaller. This highlights the advantage of DEA as a
systematic way of measuring relative efficiency within the whole sample.
3.3 Introducing scale effects

One simple addition to the DEA formulae above enables the change to variable
returns scale. This change relaxes the simplistic assumption that inputs normally
will move in exact proportions to the scale of operations: it allows for the
existence of economies and diseconomies of scale.
34
The additional constraint is that the weights in the DEA formula must sum to
one. From Figure 2.3 the variable returns frontier is the tight fitting frontier
VAABCD compared with the less restrictive constant returns frontier OBX.
Introducing this constraint has the effect of pulling the frontier in to envelop the
observations more closely. The variable returns DEA problem for the first
hospital in the twenty hospital data set is given by:
(3) Minimise E1 with respect to w1, w2, …, w20 and E1
subject to:
150w1 + 225w2 + 90w3 + … + 230w18 + 290w19 + 360w20 – 150 ≥ 0
50w1 + 75w2 + 10w3 + … + 50w18 + 90w19 + 70w20 – 50 ≥ 0
200w1 + 600w2 + 200w3 + … + 200w18 + 450w19 + 415w20 – 200E1 ≤ 0
600w1 + 1200w2 + 200w3 + … + 280w18 + 410w19 + 575w20 – 600E1 ≤ 0
w1 + w2 + w3 + … + w18 + w19 + w20 = 1
w1 ≥ 0, w2 ≥ 0, w3 ≥ 0, … , w18 ≥ 0, w19 ≥ 0, w20 ≥ 0
The measure of scale efficiency (illustrated in Figure 2.3) can be derived by

taking the ratio of the constant returns to the variable returns efficiency scores.
If the value of this ratio is one, then the hospital is apparently operating at
optimal scale. If the ratio is less than one then the hospital appears to be either
too small or too large relative to its optimum size. To determine whether it may
be too small or too large requires running a third variant of DEA subject to
‘non-increasing’ returns. This corresponds to fitting the frontier OBCD in
Figure 2.3. By comparing the variable and non-increasing returns scores for
those hospitals which appear to be not at optimal scale, it is possible to identify
on which part of the frontier they fall. If the variable and non-increasing returns
scores are the same then the hospital would be on the segment of the frontier
BCD, and thus would be too large relative to its optimum size. If the variable
returns score is higher than the non-increasing returns efficiency score, then the
hospital is on the segment of the frontier VAAB, and thus would be too small
relative to its optimum size. To calculate the non-increasing returns version of
DEA, the constraint in (3) that the weights sum to one is replaced with the
constraint that their sum must be less than or equal to one (see Appendix A).
The results for the DEA run with variable returns to scale for the twenty
hospitals are presented in Table 3.3. The average size of the efficiency scores is
higher in the variable returns case — 87 per cent compared with 75 per cent for
constant returns (see Section 4.8.4 for a discussion of the meaning of average
efficiency scores). There are now nine hospitals achieving an efficiency score of
100 per cent, although of the three additional efficient hospitals compared with
35
constant returns, one does not appear in any peer counts. This indicates that this
hospital — hospital 3 — was found apparently efficient by default because there
are no other hospitals of comparable size.
Table 3.3: Variable returns to scale DEA results for the twenty
hospitals
Hospital CRTS VRTS Scale Too small/ Peer Peer
number efficiency efficiency efficiency too big group count
1 0.63 0.89 0.71 too small 15, 12 0
2 0.31 0.36 0.87 too small 15, 12 0
3 0.39 1.00 0.39 too small 3 0
4 0.48 0.63 0.77 too small 6, 15 0
5 0.50 0.75 0.67 too small 6 0
6 1.00 1.00 1.00 – 6 7
7 0.46 0.56 0.82 too small 6, 12, 15 0
8 1.00 1.00 1.00 – 8 1
9 0.75 1.00 0.75 too big 9 1
10 0.93 0.93 1.00 too big 11, 6 0
11 1.00 1.00 1.00 – 11 2
12 1.00 1.00 1.00 – 12 6
13 0.94 0.98 0.96 too big 12, 11 0
14 0.59 0.72 0.83 too small 15, 12, 6 0
15 1.00 1.00 1.00 – 15 8
16 1.00 1.00 1.00 – 16 1
17 0.90 1.00 0.90 too big 17 1
18 0.85 0.99 0.86 too small 15, 12, 6 0
19 0.71 0.74 0.97 too small 8, 16, 6, 15 0
20 0.62 0.93 0.67 too big 17, 15, 9 0
The average scale efficiency score is 86 per cent. The hospitals that are not of
optimal size comprise nine that appear to be too small and five that seem too
big. There are some apparent anomalies in this — for instance, hospital 2, which
was identified as being the worst performer as a result of its inadequate output
for a relatively large amount of inputs, is still the least efficient under variable
returns but the results suggest that it is too small rather than too big. Clearly,
apparent anomalies such as this would have to be followed up with more
detailed analysis in an actual study.
36
3.4 Conclusion
This discussion has covered some of the main issues to consider before
undertaking a DEA efficiency study, and an example of how to calculate DEA
for a group of twenty hospitals. A more technical description of DEA and
various extensions is presented in Appendix A. In Appendix B, the computer
programs to calculate DEA information such as that presented in this chapter
are outlined.
Chapter 4 contains an overview of case studies where DEA has been used to
assess the relative efficiency of a range of human services. The case studies are
presented in detail in Chapter 5.
To summarise the main issues to consider and anticipate before undertaking a
DEA study, the following questions based on Fried and Lovell (1994) are worth
asking:
• What should the unit of observation be — the aggregate organisation or
business units within the organisation?
• What are the organisation’s main outputs and inputs?
• What characteristics of the operating environment are relevant?
• What should the comparison set be — within the city, within the state,
national or international?
• What time period should the study take?
• Are all outputs and inputs under management control?
• What do you tell the managers of an apparently less efficient organisation?
• What would you say if you were the manager of an apparently less
efficient organisation?
• What should you do with an organisation that is apparently less efficient
because it is too small or too large?
37
4 OVERVIEW OF THE CASE STUDIES
In this chapter specific reference is made to case studies in which

data envelopment analysis was applied:
• acute care services in Victorian hospitals;
• Queensland oral health services for school students;
• NSW correctional centres;
• NSW police patrols; and
• NSW Roads and Traffic Authority (RTA) motor registries.
4.1 Introduction
The models used to assess efficiency are outlined below, along with practical
issues that were encountered in applying DEA. The following points should be
kept in mind when examining the case studies:
• the case studies are work in progress, with the ways in which the models
could be improved highlighted where appropriate;
• it is not possible to compare efficiency scores across case studies — each
is specific to the sample of service providers included in the study;
• the issues raised in this section are not comprehensive. The case studies
(presented in full in Chapter 5) contain more detail on preparing a DEA
study and interpreting results; and
• while the case studies presented in this report are based on organisations
for which State governments are responsible, it would be equally
appropriate to use DEA to assess efficiency at other levels of government
and, where data were available and comparable, across jurisdictions.
4.2 Acute care services in hospitals in Victoria
4.2.1 DEA model

The study incorporated 109 hospitals in Victoria for 1994-95. Given differences
in input data availability and expected differences in operating structures, the
sample was split into metropolitan/large country hospitals (including teaching,
research and base hospitals) and small rural hospitals (excluding base hospitals).
39
An output orientation was used to reflect the objective of hospitals to provide

the highest level of care with given resources. The DEA model included the
following inputs and outputs.
Inputs
• Labour, disaggregated into the number of full-time equivalent medical and
non-medical officers.
• Consumables, such as pharmaceuticals and electricity, measured by
recurrent non-labour expenditure.
Outputs
• Number of patients treated by each hospital expressed in terms of
weighted inlier equivalent separations (WIES). This measured the number
of separations (when a patient leaves the hospital) weighted by the
expected resources required to treat each case. These were aggregated into
three categories based on the degree of complexity of each WIES.
• Unplanned re-admissions rate (an imperfect proxy for the quality of care).1
4.2.2 Some results and issues for consideration

The study suggested that the relatively less efficient metropolitan/large rural
hospitals may be able to increase their outputs by an average 11 per cent while
holding inputs constant, with size generally having little apparent influence on
efficiency.
Those small rural hospitals which appeared relatively less efficient could
potentially increase all their outputs by an average 33 per cent, using the same
level of inputs. These hospitals, on average, could possibly increase their
outputs by a further 29 per cent if they were producing services at the optimal
size in the sample.
Overall, the results suggested there was probably more scope for improvement
to best practice in the sample of small rural hospitals than in the sample of
metropolitan/large country hospitals. Closer analysis showed that there was a
greater range of performance in small rural hospitals, and that scale efficiency
was an important determinant of technical efficiency. This is likely to be a result
of rural hospitals facing overall lower demand (because they have fewer clients
within their catchment areas than metropolitan hospitals) yet still having to
maintain a level of ‘readiness’ to meet potential demand as it arises. This type of
1 The inverse of the unplanned re-admission rate was used to reflect fewer unplanned re-
admissions being a preferable output to higher unplanned re-admissions.
40
information could be used to consider the appropriate sizes of rural hospitals in

the context of the broader objective of providing equitable access to hospital
services.
In the metropolitan/large country sample, a high proportion of hospitals were
found to have the same observed efficiency, reducing the overall explanatory
power of the model. This could have been as a result of the omission of capital
from the model which would bias the results towards hospitals with relatively
high capital usage. These hospitals were more likely to be able to produce
higher outputs with lower levels of the measured inputs such as staff and raw
materials. The model could be improved by including some measure of the
capital input of hospitals.
4.3 Queensland oral health services for school students
4.3.1 DEA model

The study covered the provision of oral health services to school services in
thirteen regions in Queensland. The two smallest regions were excluded because
they were deemed not to be as comparable as the other regions. The sample size
was expanded to thirty-three by including data for each of the eleven remaining
regions for three years. An input orientation was considered the most
appropriate by Queensland Health, including the following inputs and outputs.
Inputs
• Labour, measured by the number of days worked, disaggregated into
dental officers, dental therapists and dental assistants.
• Other consumables, measured by non-salary recurrent expenditure.
Outputs
• The number of general completed treatments.

The study found that the apparent efficiencies of the oral health care units were
relatively similar. Most units achieved efficiency scores of greater than 80 per
cent — that is, they may be able to reduce inputs by up to 20 per cent while
maintaining the same number of completed treatments if they operate at what
appears to be best practice.
41
The performance of individual units — whether apparently good, average or

poor performers — tended to be consistent over the three–year period.
However, performance of one region deteriorated from being efficient in
1992-93 to having the lowest efficiency score (70 per cent) in 1994-95. This
reflected a significant decline in the number of treatments provided, combined
with a relatively large increase in non-labour expenditure over the three–year
period.
Further examination of why some regions appeared to perform consistently
better or worse than others would be useful. Factors that could be considered are
whether an important output for some regions had been excluded from the
study, or whether environments differed over the time period of the study.
Consistently good performers could be examined to identify the types of
management practices that were more likely to lead to efficiency in providing
those services.
4.4 Correctional centres in NSW
4.4.1 DEA model

There are significant differences in the resources used to run maximum and
minimum security centres. Therefore the study was limited to minimum security
correction centres. There were only eleven similar centres in NSW, and data for
each organisation for up to the past five years was included in the study,
increasing the sample size to an acceptable level. This approach was valid
because the NSW Department of Corrective Services advised that there was
minimal change in the management of inmates over this period.
The model was input oriented, with efficiency relative to best practice measured
in terms of how inputs could be reduced without a reduction in outputs. The
DEA model included the following inputs and outputs.
Inputs
• Labour — the number of full-time equivalent custodial and other
correctional officers.
• Capital — the number of beds.
• Other inputs such as food, clothing and other consumable goods and
services, measured by recurrent expenditure on these goods and services.
42
Outputs
• The number of inmates, disaggregated into those eligible for conditional
leave of absence and other inmates, because management of the latter was
more resource intensive.
• The number of inmate receptions in each correctional centre (a measure of
the turnover of inmates — a resource intensive activity unevenly
distributed over the centres).
• The number of hours spent by inmates in personal development programs
(to reflect the level of these services provided to inmates).

It was found that the correctional centres in the sample, on average, may be able
to reduce their inputs by about 4 per cent without reducing outputs if they could
all operate at what appeared to be best practice. If the correctional centres could
achieve optimal scale, then they may be able to reduce inputs by a further 4 per
cent.
This study had a relatively high proportion of correction centres that are
defined as efficient by default (about 20 per cent of the managerially efficient
correction centres) because it had a relatively small sample compared with the
number of outputs and inputs used in the analysis. To overcome this problem,
further analysis could include correction centres from other states to increase the
sample size. Alternatively the number of inputs and outputs in the analysis could
be reduced.
The study identified one correctional centre as having a marked reduction in
apparent efficiency over the time period — from appearing to be efficient it
became the apparently least efficient in the sample. Further investigation found
that the centre had been converted from a male to a female facility in the year in
which it was found relatively less efficient, with inmate numbers declining by
around 40 per cent without a commensurate reduction in inputs.
4.5 Police patrols in NSW
4.5.1 DEA model

The study covered 163 police patrols in NSW. A patrol could include one or
several police stations within a specific geographic area. An input oriented
model was used to reflect the objective of police patrols to provide effective
43
policing with minimum inputs. The DEA model included the following inputs
and outputs.
Inputs
• Labour — the number of staff disaggregated into police officers and
civilian employees.
• Capital — the number of police cars in each patrol.
Outputs
• Number of arrests.
• Responses to major car accidents.
• Responses to incidents measured by recorded offences.
• Number of summons served.
• The number of kilometres travelled by police cars.
The first four outputs refer to the reactive aspects of policing. The last output —
kilometres travelled by police cars — covers some of the proactive, or
preventative, aspects of policing. (A visible police car can reassure the
community and prevent crime.)
Environmental factors
Factors identified which may affect the apparent efficiency of a patrol but which
were beyond the control of management were:
• the proportion of people aged 15 – 19 years within a patrol’s jurisdiction;
• the proportion of public housing in a patrol’s jurisdiction; and
• whether a patrol was a country or metropolitan patrol.
Given the above inputs and outputs, patrols with higher proportions of young
people and public housing were expected to appear to be relatively more
efficient, because they were likely to respond to more crime and have less idle
time. Country patrols, with larger, less populated areas, were expected to appear
relatively less efficient compared with metropolitan patrols because they
required more inputs to provide a similar service.

Police patrols, on average, might be able to reduce their inputs by 13.5 per cent
while maintaining their output levels and operating size. Scale inefficiency did
not appear to be a major source of input reduction. However, if it were possible
to restructure patrols to achieve their optimal size there may be further input
44
savings, on average, of 6 per cent. The measured efficiency of police patrols did
not appear to be influenced by the environmental variables using this model.
It is not clear how the quality of police work influences the level of the outputs
included in the model. Crime prevention is a major output of police patrols but
is difficult to measure. It is conceivable that a patrol identified as efficient by
DEA, because it had a high number of crime related activities relative to its
inputs, was ineffective in crime prevention. Further work is required to improve
the measurement of proactive policing to fully capture this aspect of police work
in efficiency measurement.
4.6 Roads and Traffic Authority motor registry offices in NSW
4.6.1 DEA model

The study covered 131 motor registry offices in NSW. An input orientation was
used, given that registry managers could not control the demand for services,
and thus the level of outputs. Their objective was therefore to meet the given
demand with the least resources. The DEA model included the following inputs
and outputs.
Inputs
• Labour, measured by the total number of hours worked by all staff.
• Capital, measured by annual expenditure on computers (a key capital input
for registry offices). The Roads and Traffic Authority considered that this
measure reflected the number of computers in each office because most
computer equipment was acquired at the same time and expenditure for
that year was used.
• Other consumables, such as licences, plates and postage, measured by
expenditure.
Outputs
• Number of transactions performed in each office, weighted by the average
time taken for each type of transaction.
• Average waiting time for customers, which was the relevant measurable
variable reflecting the quality of service received by customers in registry
offices.2
2 The reciprocal of waiting time was used to reflect that a shorter waiting time was a
preferred output to a longer waiting time.
45
Environmental factors
Two factors which were considered to be outside the control of registry office
managers but which could influence the relative efficiency of each registry were
whether:
• it was open for Saturday trading; and
• it processed data for motor registry agents which did not have access to
computer services.
The presence of either factor was expected to increase the relative efficiency of
offices, because they were likely to allow relatively more transactions to take
place with the same level of staff.

The results suggest that motor registries may be able to produce the same level
of measured output with 15 per cent fewer inputs relative to best practice. Size
of registry offices appears to have only a minor impact on efficiency.
Both environmental factors were found to have a positive impact on measured
relative efficiency. However, these effects were not found to be significant, so
the efficiency scores for motor registries were not adjusted.
Future studies of RTA registry offices will use computer terminal numbers,
rather than expenditure on computers, as a proxy for capital input.
4.7 General observations
4.7.1 Coverage
The organisational unit used in all of the case studies was the unit from which
services are actually delivered. At this level of decision making:
• managers are generally responsible for how inputs are used to produce
outputs;
• the organisations being assessed generally have access to similar types of
resources and are expected to complete similar tasks; and
• there are generally enough organisations within a jurisdiction to allow
comparisons to be made (where this was not the case, time series data
were included to increase the sample).
46
4.7.2 Inputs
Labour is most often measured by full-time equivalent staff, and raw materials
are most often measured by recurrent expenditure on goods and services.
However, it was consistently difficult to identify an appropriate and accurate
measure for capital. Most often, a proxy was used as the only available
comparable data. Limitations in assessing the capital input to the service
provision process need to be considered when assessing results. Improving data
bases on the significant levels of capital utilised in the provision of human
services is necessary for improvements in the assessment of overall performance
in these areas.
4.7.3 Outputs
Careful consideration needs to be given to measuring and including the
proactive or preventative outputs of organisations in the analysis. Examples
include the crime prevention activities of police and the public health programs
of hospitals. Where these outputs are not included in the model, service
providers that are highly proactive will be penalised in the efficiency assessment
if these activities are effective in reducing the need to provide reactive services.
Indicators of effectiveness, such as those reported by the Steering Committee in
the Report on Government Service Provision 1997, need to be considered in
conjunction with an assessment of technical efficiency.
The quality of the outputs being measured should be considered. This is often
very difficult, but is necessary to ensure that higher measured efficiency has not
been achieved by providing services at a lower quality than previously provided.
4.7.4 General issues to be considered in interpreting and

presenting results
The simple average of efficiency scores across a sample may not necessarily
indicate the potential for overall efficiency improvement. A less efficient
organisation which is a large user of inputs, for example, has a greater potential
(if it were to become efficient) to reduce the total inputs used across the whole
sample, than does a smaller user of inputs with the same efficiency score —
even though both will have the same effect on the average score. If average
scores are linked to the magnitude of potential reductions in inputs or
47
expansions in outputs overall, efficiency scores need to be weighted by the

inputs or outputs in question.3
In addition, the efficiency scores overstate the relative efficiencies of
organisations in a DEA study. The efficiency scores represent the extent to
which all inputs can be reduced proportionately to reach the production frontier.
But some organisations may be able to reduce some inputs even further, without
reducing output.4 It is important to consider reductions in the use of these inputs
when assessing both the sample and individual organisations within the sample.
Generally, the omission of any particular input favours those organisations
which use above average amounts of that input. Likewise, organisations for
which a high proportion of their output is not measured will appear to be less
efficient. In presenting DEA results, it is important to place any efficiency
assessment in the context of the overall objectives of the organisations being
assessed. There may be a relatively high level of inputs compared with outputs
in some service outlets because, for instance:
• a given level of inputs is required to provide a service which is used by
relatively fewer people; and
• a given level of inputs is required to ensure potential demand can be met,
but this level of ready capacity is used relatively less frequently than in
other areas.
These situations are most likely to occur where the catchment area of clients for
the service provider is not highly populated, such as in rural and remote areas.
Organisations may be technically efficient for their size, but it may not be
possible for them to achieve the economies of scale within their catchment areas
that are available in more populated areas. Thus, it is important to assess the
importance of scale efficiency on the technical efficiency of organisations in the
sample. These issues become less important if organisations are more alike, and
can be accounted for to some degree by using environmental variables such as
location or population density.
3 The average efficiency score with variable returns to scale presented in Section 3.3 for the
hypothetical twenty hospitals is 87 per cent. This implies a potential reduction in beds and
nurses, on average, of 13 percent across all hospitals. However, the efficiency scores are
not evenly distributed across hospitals of different sizes. After taking into account the
distribution of efficiency scores across hospitals of different sizes based on beds, for
example, the sum of the weighted efficiency scores ([beds in hospital X/total beds] *
efficiency score) indicates that the total number of beds across the sample could be
reduced by 15 per cent, rather than 13 per cent.
4 These inputs are described in the DEA literature as slacks (see Figure 2.2 and the
Glossary).
48
Finally, as the case studies illustrate, the available data for service providers’
inputs and activities are often not fully consistent or comprehensive. In order to
improve the data bases for service providers, there is a need to document any
data deficiencies so that these may be addressed for future assessments of
performance.
DEA results provide the maximum benefit when they are interpreted with care.
In general, they should be considered as a starting point for assessing the
efficiency of the service providers within a sample. Indications of possible
sources of relative inefficiency can guide further investigation to determine why
there are apparent differences in performance. This information can be used to
inform the managers of individual service providers, administrators and policy
makers.
49
5 CASE STUDIES
5.1 Technical efficiency in the hospitals of Victoria1
5.1.1 Summary
This report details the results of a study of the technical efficiency of a sample
of acute care public hospitals in Victoria. The study uses DEA to explore
relative efficiency of all hospitals in the sample.
The objectives of this study were to demonstrate the potential for using DEA as
a benchmarking tool for measuring the performance of acute services in
Victorian public hospitals.
Annual data for 1994-95 was provided by the Victorian Department of Human
Services on 109 hospitals, including teaching hospitals. The inputs and outputs
used are set out in Table 5.1.1.
Table 5.1.1: Preferred model specification
Inputs Outputs
Full-time equivalent non-medical staff WIES with intensity rate < 0.2 (Y1)
Full-time equivalent medical staff WIES with intensity rate ≥ 0.2 and < 0.4 (Y2)
Non-salary costs WIES with intensity rate ≥ 0.4 (Y3)
Inverse of the unplanned re-admission rate
A weighted inlier equivalent separation (WIES) measures the number of

separations of a given complexity. A WIES is similar, but not equivalent, to a
diagnostic related group separation (DRGS). It measures different acute care
cases by their degree of expected resource intensity, ranging from minor
treatments (Y1) through to complex cases (Y3). For example, Y1 equals the
total number of WIES figures for episodes of care which required 0 – 0.2
WIES3 per day during each episode. (WIES3 refers to the way in which WIES
1 Researched and written by Tendai Gregan and Rob Bruce of the Industry Commission.
Comments from Dr Graeme Woodbrigade, Paul D’Arcy, Professor Knox Lovell and Dr
Suthathip Yaisawarng are gratefully acknowledged. However any errors or omissions are
the responsibility of the authors.
51
are measured in 1996). Y2 and Y3 are similarly defined, with the intensity rates
given in the above Table.
The unplanned re-admission rate was included to account for the objective of
hospitals to maintain acceptable standards of quality of care while seeking
efficiency improvements. Unplanned re-admission rates are a proxy for the
quality of care in a hospital, but are not an ideal measure. Future studies should
seek to incorporate more accurate measures of the quality of care in hospitals.
The model was run using an output maximisation orientation. Initially, it was
run using the full sample under the assumption of constant returns to scale.
Relaxing this assumption produced a variable returns to scale model which
allowed the issue of scale inefficiency to be examined. Given differences in data
available at hospital level for inputs, and expected differences in operating
structures, the sample was split in two: metropolitan/large country hospitals
(including teaching, research and base); and small rural hospitals (excluding
base hospitals). Constant and variable returns to scale model runs were then
conducted for each sub-sample.
Detailed results for each model are included in Annexes A5.1.1–A5.1.5. These
results include information on: technical efficiency scores; the extent and nature
of scale efficiency scores; as well as actual and target values for inputs and
outputs.
In summary, the difference for metropolitan/large country hospitals between the
most and least efficient seems small. Twenty–four out of thirty–seven hospitals
made up the efficient frontier. The average relative efficiency score for hospitals
not on the frontier was 1.11, with the average hospital potentially able to
increase its outputs by 11 per cent, holding all inputs constant. In addition, after
increasing all outputs by 11 per cent, some large hospitals may still be able to
increase one or more output by up to 25.3 per cent. Scale efficiency of 1.05 for
metropolitan/large country hospitals indicates, on average, that size appears to
have little influence on efficiency.
For small rural hospitals, the results suggest that the dispersion between
efficient and less efficient hospitals may be wide. Fourteen out of sixty–nine
hospitals made up the efficient frontier. Small rural hospitals which were not on
the frontier had an average efficiency score of 1.33, and appear to be able to
increase all their outputs by 33 per cent, using the same level of inputs. In
addition, there appeared scope for some hospitals to increase between one and
three outputs by between 4.4. per cent and 26.8 per cent. Scale efficiency of
1.29 for small rural hospitals indicates, on average, that size may have had some
influence on efficiency.
52
5 CASE STUDIES
The models used were developed in consultation with the Victorian Department
of Human Services. Advice was sought on hospital inputs, outputs and
indicators of quality of service. Initially, the sample included all hospitals, but
Department input on the relevance of some peers and the relative efficiency
scores indicated that there were some problems in the input data across the
whole sample. The input of the Department led to the splitting of the sample,
which was also supported by expected differences in the operating structures of
metropolitan/large country and small rural hospitals. The subsequent models of
metropolitan/large country hospitals and small rural hospitals were validated by
the Department as providing a plausible analysis of the relative efficiency of
Victorian hospitals.
The sensitivity of the two models was tested by changing the measure of labour
inputs from full-time equivalent staff to salary costs. The efficiency scores and
the hospitals appearing on the frontier varied little when this was done,
indicating that staff costs appeared to be reasonably consistent within each of
the sub-samples. These tests support the hypothesis that the model
specifications used are a reasonable representation of the production technology
used by large and small Victorian hospitals.
5.1.2 Background
DEA has been used to analyse the relative efficiency of hospitals in NSW (NSW
Treasury 1994), and the United States (Banker, Das and Datar 1989, Burgess
and Wilson 1993, Valdmanis 1992), among others. For an extended
bibliography of DEA health studies, see Seiford (1994).
This study was conducted by the Industry Commission in consultation with the
Victorian Department of Human Services. The Department is responsible for
the funding, monitoring and evaluation of the State’s hospitals. The Department
was interested in investigating whether DEA could be used as a tool for
benchmarking relative hospital efficiency. This study includes information on
casemix (the WIES data) because it provides rich information on different types
of hospital outputs and facilitates like-with-like comparisons.
A single year’s data was used to test the feasibility of DEA as a management
tool for measuring hospital efficiency. Discussions held between the Industry
Commission and Department officers allowed the Commission to learn about
53
the operations of Victorian hospitals, develop an appropriate model

specification, and interpret the results.2
5.1.3 Data
Table 5.1.2 shows the types of data used to construct the DEA models (actual
data in Annexes 5.1.1–5.1.5). The data were supplied by hospitals in returns to
the Department and a casemix data-base, comprising information for 1994-95
on:
• the resources used to provide inpatient acute care services;
• the percentage of all such cases which result in the unplanned re-
admission of the patient; and
• the number of inpatient acute care services, grouped by case severity and
length of treatment.
Detailed definitions of each item are given below.
Table 5.1.2: Victorian hospitals data, 1994-95

Inputs Units
X1: Full-time equivalent non-medical staff (metropolitan & large country Number
hospitals only)
X2: Full-time equivalent medical staff (metropolitan/large country hospitals only) Number
X3: Total full-time equivalent staff Number
X4: Non-salary costs $’000

X5: Non-medical salaries (metropolitan/large country hospitals only) $’000
X6: Medical salaries (metropolitan/large country hospitals only) $’000
X7: Total salaries $’000
Outputs Units
Y1: WIES with intensity rate < 0.2 (minor) Number

Y2: WIES with intensity rate × 0.2 and < 0.4 (moderate) Number
Y3: WIES with intensity rate × 0.4 (complex) Number
Y4: Unplanned re-admission rate Percent
The study focused on hospital inpatient acute care services, which make up the
majority of total hospital services. (Over the sample, an average of 82 per cent
2 The Commission sincerely appreciates the support given to the project by the Department,
in particular Ms Fatima Lay, Mr Tony Simonetti and Mr John Iliadis of the Acute Health
Care Division.
54
5 CASE STUDIES
of inputs were devoted to acute inpatient care.) The Department considered that
the measures for non-acute and outpatient services, such as bed days, were not
of a level that explained the output of hospitals as well as those used for acute
inpatient services, which account for case severity and length of stay.
For each hospital, estimates of the inputs used to provide acute inpatient
services were derived by multiplying each of the total inputs by the share of
acute inpatient services in total hospital costs.3
DEA is sensitive to outliers, which are observations that are not typical of the
rest of the data.4 Outliers can arise from either measurement or reporting error,
or may reflect significant efficiencies being achieved by particular hospitals.
Alternatively, outliers may identify hospitals which use different production
technologies. In the first case, outliers should be removed from the data, and in
the latter instances, hospitals should be checked to determine whether they have
access to an omitted input or use different technology. All the inputs and outputs
in the full sample of 109 hospitals were screened for potential outliers using the
technique discussed in Section 2.5. The potential outliers were referred to the
Department, who advised that three hospitals had measurement errors. These
three were removed to form the sample of 106 hospitals used in the model runs.
The remaining potential outliers were judged to be free of measurement or
recording errors, and to be comparable to the rest of the set, and were retained in
the sample.
Inputs
Valdmanis (1992) and Burgess and Wilson (1993) used physical inputs, such as
the number of full–time equivalent staff by skill category; the number of beds as
a proxy for capital; the number of admissions; and the number of visits by
3 An initial analysis was carried out excluding information on non-acute hospital outputs,
but including the inputs used to provide these services. This led to biased results. It was
found that hospitals which provided relatively more non-acute services — as indicated by
the share of non-acute services in the total budget — appeared to be relatively inefficient
compared with hospitals which concentrated on providing acute care services. When
inputs used to provide non-acute services were excluded by estimating the quantities of
inputs used for acute services only, it was found that the efficiency scores improved for
hospitals that provided relatively more non-acute services. If these estimates still contain
some inputs used to provide non-acute services, then it can be expected that there will be a
degree of bias against hospitals which provide relatively more non-acute services. The
extent of this bias will depend on the size of estimation error. However, it was judged that
any error — and thus bias — would be small, given the accuracy of the budget share data
used to split acute and non-acute services.
4 See Section 2.5 for a discussion of the impact of outliers on DEA results.
55
physicians. In contrast, Banker, Das and Datar (1989) used cost information,
broken down by labour type and non-labour resources, to measure inputs.
Although physical measures are preferred to cost measures because DEA
measures physical productivity, this study used both types to test whether there
was a significant difference in the results.
Full-time equivalent staff

Full-time equivalent staff were used to measure labour input. Given the shift
work nature of hospitals and the prevalence of part-time employment, data on
the number of full-time equivalent staff gives a more accurate indication of the
amount of labour used to provide services than does a simple count of the
number of staff employed or the cost of labour.
The cost of labour is an alternative measure of the resources used, but the staff
measure used was preferable for two reasons.
First, the Department of Human Services advised that salary expenditure per
full-time equivalent staff member could be expected to vary significantly
between city and rural hospitals. Where this was the case, differences in
measured expenditure would reflect the prevailing regional wage rates, the level
of training of staff, and the physical quantities of labour used to perform any
given service. The staff measure was likely to be more homogeneous across
regions than was expenditure because it was not influenced by wage rates.
Greater homogeneity allows for better comparisons of the actual physical
product than does cost measures.5
Second, for hospitals aiming to minimise costs, they had to employ the least
physical quantity of each input to produce a given level of output.
The labour data was split into two classes:
1. non-medical full-time equivalent staff, directly employed by hospitals (that is,
nurses, nurse assistants, cleaners, management and administration staff); and
5 For example, if a city and country hospital both use one doctor hour to treat a patient for a
broken leg, then the measure of both their physical products would be 1 (equal 1 broken
leg treatment / 1 doctor hour). However, if cost data rather than quantity data is used and
doctor’s wages are lower in the city than in the country, then the ‘productivity’ of the
country doctor would mistakenly appear to be lower. If the hourly wage in the city is $45
and the country wage is $50, then the city hospital’s ‘productivity’, (1/[$45×1]), 0.022, is
greater than that in the country (1/[$50×1]), 0.020. In fact, both hospitals are equally
efficient in their provision of services, but the relatively higher costs in the country may
reflect, among other things, a less competitive market for labour and thus higher wages.
56
5 CASE STUDIES
2. medical full-time equivalent staff (doctors, specialists) directly employed by

hospitals.
This broadly reflected the different skills and functions of labour used in
hospitals.
The choice of non-medical and medical full-time equivalent staff was based on
the traditional division of labour used in hospitals: nurses and doctors. The vast
majority of ‘non-medical’ staff are nurses, who provide general care to patients,
usually under the direction of doctors. The remainder of staff in this category
provide general hospital support and administration services. ‘Medical’ full-
time equivalent staff comprise interns (trainee doctors), doctors, surgeons and
specialists directly employed as officers of the hospital.
The medical category excluded persons providing medical services to the
hospital on a fee-for-service basis, who are referred to as visiting medical
officers. The input of visiting medical officers, and possibly some non-medical
staff, was captured in the contract fees paid to them, which were included in the
non-salary costs (X4) of each hospital. Ideally, these should have been captured
in a full-time equivalent measure, but such information was not available.
Salary costs
Financial information on the costs of labour was also provided. Labour costs
were divided into the same categories as staff: non-medical staff salaries and
medical staff salaries.
Good information on these categories was available for metropolitan/large
country hospitals, but was patchy for small rural hospitals because these do not
typically employ medical staff directly. Given that they use visiting medical
officers, rather than salaried doctors, the data on medical full-time equivalents
and the corresponding medical costs were zero. Accordingly, two separate
models of small rural hospitals were used: one using total full-time equivalent
and the other using total salaries. A pooled sample of all hospitals, large and
small, also used total full-time equivalents as the labour input measure.
However, for the reasons set out in Section 5, this sample was split into
metropolitan/large country hospitals and small rural hospitals.
Non-salary costs
Inputs other than labour are important for providing acute care hospital services.
These were captured in non-salary costs, which accounted for the remaining
inputs — other than capital — used in the production of hospital services.
57
Non-salary costs = maintenance + contract costs (visiting medical officers) +

electricity + gas + water + consumables (bandages, drugs, etc.) +
superannuation
Fixed capital is a significant input in providing care. A measure of capital was
not included in the model because comprehensive and accurate information on
the stock of capital assets was not available. This has some implications, which
are examined below, for interpreting the model.
Outputs
Other studies (Burgess and Wilson 1993; Valdmanis 1992) have measured
outputs using the number of inpatient hospital bed days, the number of
surgeries, and the number of emergency treatments. Like Banker, Das and Datar
(1989), this study used casemix data. However, this study differs in that the data
were adjusted for length of stay. Time adjusted casemix data was preferable to
bed days because first, it is more homogeneous across hospitals, and second, it
captures casemix adjusted for severity of illness and the expected resources
required to treat patients.
Weighted Inlier Equivalent Separations (WIES)

WIES is a measure of case intensity (diagnostic related group) adjusted by the
normalised patient length of stay (inlier equivalent separations, or IES).
Formally:
WIES = IES × DRG weight
where:
• each DRG represents a class of patients with similar clinical conditions
requiring similar hospital services. A more detailed explanation of DRGs
is given in the National Health Data Dictionary (NHDC 1995) and Eagar
and Hindle (1994);
• DRG weights are an index of case complexity based on clinical history —
for example, a leg fracture has a lower DRG weight than a liver transplant;
and
• IES represent the ‘normal’ length of time for which a patient will stay in
hospital, for every type of DRG.6 A case which is in this ‘normal’ interval
6 The ‘normal’ length of stay is given by the DRG average length of stay, which is based on
historical records and current medical practice. The low boundary point (LBP) is set to
one third of the DRG average length of stay and the high boundary point (HBP) is set at
three times the DRG average length of stay. Values of low and high boundary points are
58
5 CASE STUDIES
is called an inlier and given an IES equal to one. Cases which take less
time are weighted lower and those which take more time are weighted
higher.7
The WIES used in this study include all acute care services to inpatients: acute
care; palliative care; and alcohol and drug treatment. They exclude services for:
nursing home care; aged care; psychiatric and certain types of rehabilitation.
(See Appendix 7 of HCS 1994.)
There are over 500 DRGs and thus WIES. To apply DEA, these WIES groups
were aggregated into three categories which reflect the different casemixes
handled by different types of hospitals:
• WIES with an intensity rate less than 0.2 (minor)
• WIES with an intensity rate greater than or equal to 0.2 and less than 0.4
(moderate); and
• WIES with an intensity rate greater than or equal to 0.4 (complex)
Despite the advantages of using WIES figures over traditional variables, the
casemix classification system is not perfect. The casemix formulations have
been upgraded continually since inception to make them as comprehensive and
accurate as possible. To the extent that not all acute care activities may be
captured by the WIES figures, the DEA results presented in this report should
be interpreted with caution.8
Unplanned re-admission rate

The quality of hospital outputs is a defining characteristic of the care provided.
It is difficult to measure the quality of care. The typical surrogate measures
include mortality rates, re-infection rates and unplanned re-admission rates.
rounded to whole numbers. In addition, the maximum value of a high boundary point is
limited to 100 days.
7 Specifically, a case which is less than the low boundary point of the ‘normal’ length of
stay is given an IES equal to the actual length of stay divided by the low boundary point.
Similarly, a case which is greater than the high boundary point of the ‘normal’ length of
stay is given an IES equal to one plus the fraction given by the number of days above the
one high boundary point divided by two times the DRG Average Length of Stay.
8 A model with a single output variable, total WIES, was tested and found to be
unsatisfactory because it yielded inappropriate benchmarking partners. For example, it
gave small rural hospitals which treat simple cases mainly as peers for large teaching
hospitals treating much more complicated cases. The preferred model has WIES separated
into three classes of casemix. It was judged that the increased number of outputs gave a
more plausible mix of peers, and did not unduly inflate either the efficiency scores of
hospitals or the number of hospitals that were efficient by default.
59
None of these data were readily available. However, it was found possible to
construct re-admission rates for the hospitals using the Casemix database. This
was done by assuming that an unexpected return to the same hospital within
twenty-eight days of the patient previous episode of care (which may or may not
be related to the first episode of care) was a re-admission. The lower the number
of these re-admissions, the higher the quality of care arguably.
Two criticisms of using the unplanned re-admission rate as a proxy for quality
are that:
• the method used to calculate the rates tends to overstate the actual rates,
because many re-admissions may be clinically unrelated to the first
episode of care; and
• hospitals with a more complex casemix have a higher probability of
unplanned re-admissions, biasing the results against these hospitals.
However, when using DEA to measure relative efficiency, hospitals are
compared only with those hospitals which produce a similar mix of outputs,
given input levels, ensuring that those with higher levels of complex cases and
unplanned re-admissions are compared with each other only.
This variable was included in the model in recognition of the fact that hospitals,
in seeking improvements in efficiency, wish to maintain or improve standards of
service. The unplanned re-admission rate has been regularly used as a quality
indicator since the introduction of casemix funding in 1993.
Unplanned re-admission rates have been used as an indicator of hospital
effectiveness (SCRCSSP 1995), but this study used the rates in the measurement
of hospital efficiency. The assumption of the study was that an increase in
output using the same quantity of inputs, and at least maintaining the same
quality standards, was a true increase in efficiency, whereas the same increase in
output with a fall in quality might not have meant that there had been an
efficiency increase necessarily. This is because quality is a defining
characteristic of any output — it is easier and less resource intensive to produce
low quality rather than high quality output. Therefore, ignoring the quality
dimension results in a flawed view of any measured efficiency increases.
Nevertheless, care is required in interpreting these results.
This case study measured efficiency in terms of hospitals’ ability to increase
outputs using the same quantity of inputs, that is, the model was output oriented.
Because the unplanned re-admission rate is a ‘negative’ output (that is, an
increase is undesirable), the inverse was used in the analysis. Maximising the
inverse of the unplanned re-admission rate is the same as minimising the
unplanned re-admission rate.
60
5 CASE STUDIES
5.1.4 Model specification and orientation

Five models were run, each with a different sample and input variables, but the
same four outputs (Y1 – Y4).
1. All metropolitan, large country and small rural hospitals, with inputs: total
full-time equivalent staff (X3) and non-salary costs (X4).
2. Metropolitan/large country hospitals only, with inputs: non-medical full-time
equivalent staff (X1), medical full-time equivalent staff (X2) and non-salary
costs (X4).
3. Small rural hospitals, only with inputs: total full-time equivalent staff (X3)
and non-salary costs (X4).
4. Metropolitan/large country hospitals only, with inputs: non-medical salaries
(X5), medical salaries (X6) and non-salary costs (X4).
5. Small rural hospitals only, with inputs: total salaries (X7) and non-salary
costs (X4).
An output orientation was chosen after consultation with the Victorian
Department of Human Services. Thus, the relative efficiency of hospitals was
measured on their potential to increase outputs (given their existing level of
inputs) relative to best practice in the sample. There were three reasons for this
choice of orientation:
• the existence of waiting lists for metropolitan acute care indicates that
productivity improvements would be best directed to increasing outputs,
rather than decreasing inputs;
• in rural areas, medical facilities are provided to a relatively small
population, with often limited demand, on the grounds of equity of access
to essential services. This means that managers of small rural hospitals
may have little scope to reduce their use of inputs; and
• Victorian acute care hospitals are funded on the basis of the outputs they
provide, so the incentive is to maximise outputs rather than reduce inputs.
However, funding is based on expected average resource use for particular
services, so hospital managers must also ensure efficient input use to remain
within budget constraints. Each hospital forms a contract with the Department
of Human Services for an agreed target level of WIES. Hospitals that produce
more services than this level are not funded for those extra services.
On balance, the incentive of maximising the services provided was considered
to be the most appropriate driver of productivity gains.
61
In consultation with the Department, models 2 and 3 were determined to be the

most appropriate. The results of all models are discussed in Section 5.1.5, with
the analysis focused on models 2 and 3.
5.1.5 Results and conclusions
Technical efficiency scores

The principal results reported in this section were derived by imposing the
assumption of variable returns to scale on each of the models outlined above
(See Chapter 2).
The technical efficiency scores indicate which of the hospitals are calculated by
the model to be on the efficient (best practice) frontier (those with a score of
one), and which are calculated to be less efficient relative to hospitals on the
frontier (those with scores greater than one). The higher the score, the higher the
potential increase in output (while maintaining inputs) relative to best practice.
Technical efficiency scores only refer to relative performance within the sample.
Hospitals given an efficiency score of one are efficient relative to all other
hospitals in the sample, but may not be efficient by some absolute or world
standard necessarily.
Scale efficiency scores

The impact of scale on relative efficiency was also assessed. The effect of size
on efficiency was analysed using a three stage process. First, the models were
run assuming constant returns to scale. Second, a comparison of the results for
constant returns to scale and those for variable returns to scale allowed an
assessment of whether the size of a hospital had an influence on its technical
efficiency. Finally, to assess the nature of any scale inefficiency, each model
was run under the assumption of non-increasing returns to scale. Comparing
these final results with results for variable returns to scale enabled hospitals to
be described as having increasing, decreasing or constant returns to scale. For a
detailed explanation of this three stage process, see Section 2.4.1 and
Appendix A.
Model 1 results: all hospitals

Annex 5.1.1 sets out the results of model 1. The variable returns to scale case
had twenty-seven hospitals (25 per cent) making up the efficient frontier. None
of these had scope to increase one output further, so all were truly efficient
relative to all hospitals in the sample.
62
5 CASE STUDIES
The average efficiency score of the seventy-nine hospitals off the frontier was
1.29, indicating that these hospitals on average may be able to increase all their
outputs by 29 per cent using the same amount of inputs.
After analysing the results of this model and consulting with the Department, it
was decided that the results did not accurately reflect the Department’s
expectations of relative efficiency within Victorian acute care hospitals. Nearly
all metropolitan/large country hospitals were relatively less efficient and
therefore had small rural hospitals as their peers, or benchmark partners.
This was because there are two important differences in the way that
metropolitan/large country and small rural hospitals operate:
1. Use of medical staff. Small rural hospitals use visiting medical officers
instead of salaried doctors, so they appear to use relatively fewer full-time
equivalent staff to produce their outputs than do metropolitan/large country
hospitals. This resulted in nearly all metropolitan and large country hospitals
being off the efficient frontier, along with small rural hospitals that did have
salaried doctors. In several instances, small rural hospitals who employed no
doctors were significant peers for major teaching hospitals and specialist
research hospitals.
2. Costs. The Department advised that small rural hospitals face significantly
different costs from metropolitan and large country hospitals, which would
affect the quantities of physical inputs they employ.
Given the data difficulties and the significant differences in operating
procedures and costs faced by metropolitan/large country hospitals compared
with small rural hospitals, the sample was split and models 2 and 3 were run.
Model 2 results: Metropolitan/large country hospitals

had twenty-four hospitals (69 per cent) making up the efficient frontier. With
the exception of one, all were unable to increase a single output or reduce an
input further, so were apparently truly efficient relative to all hospitals in the
sample.9 The one hospital that was able to increase an output and reduce inputs,
and four others were apparently ‘efficient by default’, meaning that they were
9 One hospital on the frontier appeared to have scope to reduce its use of non-medical full-
time equivalent staff and non-salary costs, and increase production in the output of Y2.
After consultation, it was revealed that this hospital had special research functions which
may not have been fully captured in the model specification. This view was supported by
the fact that this hospital did not appear as a best practice peer for any of the inefficient
hospitals. Thus, the hospital was on the frontier by default.
63
on the frontier as a result of some unique characteristics in the use of inputs and
production of outputs which were not explained by the model specification.
A key feature of this model was the high proportion of hospitals on the efficient
frontier.
The average relative efficiency score of the thirteen hospitals off the frontier
was 1.11, indicating that on average these hospitals could potentially increase
all their outputs by 11 per cent using the same amount of inputs.
Average scale efficiency of 1.05 indicated that non-frontier hospitals, on
average, might be able to increase their outputs by 5 per cent beyond their best
practice targets under variable returns to scale, if they were to operate at
constant returns to scale. In addition, it was found that most were apparently
larger than the optimal efficient size derived by the model.
The apparent efficiency of non-frontier hospitals was also influenced by the
extent to which it appears possible to reduce an input, or expand an output, after
all outputs have been expanded uniformly to place the hospital on the
production frontier.10 The extent that it seems possible to reduce an input or
expand an output was determined by multiplying the efficiency score of each
hospital by its actual level of output or input and then determining the difference
between this figure and the target level for the input or output. The total scope
for changing each output or input was then expressed as a percentage of the
total actual outputs (or inputs), thereby giving an indication of the relative size.
An output oriented study such as this typically reports only how much each
output may be increased after all outputs have been increased in the proportion
given by the efficiency score. However, this study also reports apparently excess
inputs because their existence in an output oriented study indicates that there is
potential to not only increase output to best practice levels using the same
quantity of inputs, but to increase it using fewer inputs. This potential may never
be realised, depending on the cause of the excessive input. Apparently excessive
use of an input can reflect a low demand for hospital services in a region and the
inability of managers to reduce inputs because they are bound to labour
agreements or need to provide equitable access to essential services.
In addition to the potential for an average 11 per cent increase in all their
outputs as indicated by the efficiency score, non-frontier hospitals may be able
to increase output further in two of the four output categories. The model
suggests non-frontier hospitals may be able to increase their production of Y2
10 For an explanation see Section 2.2.
64
5 CASE STUDIES
by an average of 12 per cent and the rate of unplanned re-admissions by a

further 25 per cent, on average, relative to best practice.
Non-frontier hospitals may also be able to reduce the use of some inputs. The
model suggests it may be possible to reduce non-medical staff by 3 per cent,
medical staff by 6 per cent and non-salary costs by 9 per cent.
Excluding the five hospitals which were apparently efficient by default, over 50
per cent of hospitals appeared to be efficient. Those hospitals on the frontier
were among the larger hospitals, and were likely to be those with access to
capital equipment which was unaccounted for in this model. The efficiency
scores were therefore biased towards those hospitals which used more capital
relative to those that produce the same output with less capital. That is, hospitals
with relatively high capital intensities were more likely to make up the frontier,
because the partial productivity of their other inputs will be higher relative to
that of other hospitals. A model which included capital and a larger sample of
large teaching and research hospitals from interstate or overseas would better
lend itself to analysis of relative efficiency of these metropolitan/large country
hospitals.
Model 3 results: small rural hospitals

had fourteen hospitals (20 per cent) making up the efficient frontier. None of
these could further reduce inputs, or expand outputs, so all were truly efficient
relative to all hospitals in the sample.
The average efficiency score of the fifty-five hospitals off the frontier was 1.33,
indicating that these hospitals, on average, may be able to increase all their
outputs by 33 per cent using the same amount of inputs.
Average scale efficiency of 1.30 indicates that non-frontier hospitals, on
average, may be able to increase their outputs by 30 per cent beyond their best
practice targets under variable returns to scale, if they were to operate using
constant returns to scale. In addition, the results suggest that most hospitals
were larger than the optimal size implied by the model.
As with metropolitan/large country hospitals, the efficiency of non-frontier
hospitals was influenced by the scope to further reduce individual inputs or
expand outputs, beyond that reflected by the efficiency score. On average, non-
frontier hospitals could expand Y2 by 23 per cent and Y3 by 27 per cent.
Unplanned re-admissions could be reduced by 10 per cent and Y1 increased by
4 per cent.
65
The scope for expanding output of Y2 and Y3 was large, but it should be
recalled that these classes of output represent more complicated cases. Many
small rural hospitals would not have the facilities or qualified staff to treat Y2
and Y3 cases, and such cases would be passed on to metropolitan/large country
hospitals typically. Accordingly, the apparently significant scope for increasing
output that the model generates for these outputs should be interpreted
cautiously. However, the remaining scope for increasing the other outputs
suggests that small rural hospitals would be able to increase their number of
basic treatments (Y1) and lower their unplanned re-admission rates (beyond the
33 per cent given by the average efficiency score).
The model suggests that non–frontier hospitals may be able to reduce total full-
time equivalent staff by 2.8 per cent and non-salary costs by 4.7 per cent. This
suggests that if less efficient hospitals operated according to best practice in the
sample, then they might not only be able to expand their output using the same
amount of inputs, but may be able to produce more output using 2.8 per cent
less of total full-time equivalent staff and 4.7 per cent less of non-salary costs.
However, this does not account for the practical limitations of reducing inputs,
such as contracted labour, or for possible constraints on the demand for outputs
of many small rural hospitals.
Sensitivity analysis: models 4 and 5

To test the robustness of the models to changes in the measurement of inputs,
models 2 and 3 were run with salary expenditure instead of full-time equivalent
staff. Detailed results are given in Annexes 5.1.4 and 5.1.5.
Changing the way in which labour was measured had a minor impact on model
results. The hospitals assessed to be on the frontier were largely the same; along
with the average efficiency scores, scale efficiency scores and the scope to
expand some outputs and decrease inputs. This suggests that:
• wage rates appear to be reasonably consistent in each of the sub-samples;
and
• the model is robust in its labour specification.
Further models were run using traditional measures of output, such as adjusted
length of stay and acute and non-acute bed days. Analysis and consultations
with the Department indicated, as expected, that these did not capture outputs as
accurately as did the outputs used in the preferred models 2 and 3.
66
5 CASE STUDIES
5.1.6 Future studies

This case study provides a sound starting point for using DEA to assess the
efficiency of acute care services in Victorian hospitals. In the development of
further models, areas in which the modelling could be improved include:
• the capital input of hospitals;
• a more accurate indicator of the quality of care provided by hospitals; and
• inclusion of public/private patient mix to determine the effect on the
efficiency of different patient mixes.
67
Annex 5.1.1: Model 1 - all hospitals
Efficiency Scale
Score Score Actual inputs Input targets Actual outputs Output targets
HOSPITAL PHI SE Total FTE Non-salary Total FTE Non-salary Unplanned Unplanned
(VRS) staff costs staff costs re-admission re-admission
($’000) ($’000) WIES1 WIES2 WIES3 ratea WIES1 WIES2 WIES3 ratea
HP2 1 1.56 2013.14 48702 2013.14 48702 12115.02 17554.22 16422.44 6.63 12115.02 17554.22 16422.44 6.63
HP3 1 1.41 1463.73 28483 1463.73 28483 3788.66 8428.51 13520.01 13.35 3788.66 8428.51 13520.01 13.35
HP4 1 1.32 1151.83 20069 1151.83 20069 9239.72 10683.23 4250.59 10.28 9239.72 10683.23 4250.59 10.28
HP5 1 1.27 424.62 8870 424.62 8870 1077.78 2511.97 4565.45 4.36 1077.78 2511.97 4565.45 4.36
HP6 1 1.56 2890.62 47790 2890.62 47790 11984.28 16655.07 21799.12 11.27 11984.28 16655.07 21799.12 11.27
HP7 1 1.35 913.22 18310 913.22 18310 5029.88 8718.33 7372.95 8.83 5029.88 8718.33 7372.95 8.83
HP39 1 1.00 * 30.45 1388 30.45 1388 192.21 530.17 724.29 1.32 192.21 530.17 724.29 1.32
HP8 1 1.45 903.5 19687 903.5 19687 6336.38 8167.83 7478.84 9.13 6336.38 8167.83 7478.84 9.13
HP10 1 1.47 1483.4 29820 1483.4 29820 10686.42 13277.6 10082.34 15.50 10686.42 13277.6 10082.34 15.50
HP11 1 1.33 516.22 10355 516.22 10355 3993.22 5014.07 4125.03 8.76 3993.22 5014.07 4125.03 8.76
HP12 1 1.59 99.93 3360 99.93 3360 1041.53 1049.94 977.04 3.09 1041.53 1049.94 977.04 3.09
HP14 1 1.52 2712.85 62241 2712.85 62241 11965.57 17454.41 26061.18 11.33 11965.57 17454.41 26061.18 11.33
HP16 1 1.15 260.62 6121 260.62 6121 2042.32 3225.78 2015.74 4.36 2042.32 3225.78 2015.74 4.36
HP18 1 1.27 175.17 5119 175.17 5119 1544.5 2153.06 1930.04 5.39 1544.5 2153.06 1930.04 5.39
HP20 1 1.35 540.22 12709 540.22 12709 3370.37 5409.52 4997.13 7.29 3370.37 5409.52 4997.13 7.29
HP21 1 1.64 1203.49 30153 1203.49 30153 9210.84 10761.21 8953.54 9.89 9210.84 10761.21 8953.54 9.89
HP41 1 1.06 61.48 1743 61.48 1743 589.63 907.89 733.33 6.47 589.63 907.89 733.33 6.47
HP43 1 1.00 * 47.11 535 47.11 535 407.77 414.09 459.04 12.64 407.77 414.09 459.04 12.64
HP31 1 1.59 933.97 25399 933.97 25399 6852.34 8689.47 7673.53 7.55 6852.34 8689.47 7673.53 7.55
HP32 1 1.61 562.69 17053 562.69 17053 3014.11 5041.9 5692.09 5.49 3014.11 5041.9 5692.09 5.49
HP35 1 1.16 308.2 8551 308.2 8551 2996.7 4160.34 2592.08 7.65 2996.7 4160.34 2592.08 7.65
HP67 1 1.00 * 39.23 268 39.23 268 381.34 321.74 264.98 8.76 381.34 321.74 264.98 8.76
HP78 1 1.00 * 19.44 654 19.44 654 406.03 341.74 267.95 7.11 406.03 341.74 267.95 7.11
HP89 1 1.00 * 3.94 76 3.94 76 120.55 44.67 39.84 11.96 120.55 44.67 39.84 11.96
HP98 1 1.00 * 43.8 301 43.8 301 439.95 377.51 331.13 12.99 439.95 377.51 331.13 12.99
HP101 1 1.00 * 32.5 97 32.5 97 134.14 65.55 79.56 15.13 134.14 65.55 79.56 15.13
HP102 1 1.27 11.6 224 11.6 224 115.11 55.31 30.24 5.14 115.11 55.31 30.24 5.14
68
Annex 5.1.1: Model 1 - all hospitals (continued)
Efficiency Scale
HOSPITAL PHI SE Total FTE Non-salary Total FTE Non-salary Unplanned Unplanned
(VRS) staff costs staff costs re-admission re-admission
HP29 1.01 1.82 207.81 4312 207.81 4312 1903.56 1498.89 980.27 11.14 1924.65 2233.81 1120.71 7.48
HP19 1.02 1.37 690.92 14636 690.92 14636 4157.2 6423.26 5763.47 9.53 4231.85 6538.61 5866.97 8.18
HP17 1.03 1.82 1678.45 52260 1678.45 39830.2 7489.94 10300.59 15834.22 9.20 7684.67 11367.47 16245.89 7.63
HP46 1.03 1.33 82.91 1823 82.91 1823 837.29 825.09 645.54 6.55 862.91 927.9 665.29 6.35
HP52 1.03 1.54 127.74 2920 127.74 2920 1264.75 1121.69 741.2 9.34 1297.82 1494.88 830.58 7.32
HP1 1.06 1.72 766.64 18413 766.64 14693.74 6024.45 5976.86 2298.6 7.89 6363.4 7653.81 3447.73 7.47
HP27 1.06 1.67 362.99 11793 362.99 9461.96 2406.49 3182.76 3351.36 7.84 2554.34 3602.91 3557.27 6.05
HP63 1.07 1.25 37.91 783 37.91 783 479.42 371.82 306.31 12.53 512.1 464.95 327.43 8.02
HP86 1.08 1.43 49.02 880 49.02 880 528.49 384.14 281.65 8.32 570.5 542.76 372.91 7.70
HP92 1.08 1.52 52.09 1254 52.09 1254 598.46 452.37 370.54 6.33 647.07 649.44 409.49 5.86
HP24 1.09 1.45 404.55 12359 404.55 11600.88 2490.45 4204.52 3400.87 9.92 3093.79 4569.38 3695.99 6.80
HP15 1.1 1.39 202.71 5376 202.71 5376 1672.47 1958.33 1700.72 3.92 1844.51 2530.19 1875.66 3.56
HP26 1.1 1.49 355.84 10277 355.84 9345.04 2258.65 3201.23 3211.35 6.74 2476.68 3510.25 3521.34 5.95
HP44 1.1 1.61 148.29 2654 148.29 2654 1182.27 995.63 886.83 9.03 1295.52 1448.69 971.78 8.24
HP58 1.1 1.33 53.93 1039 53.93 1039 539.58 435.03 431.47 7.35 593.62 599.27 474.68 6.68
HP60 1.12 1.35 19.1 191 7.84 191 104.16 39.18 63.17 7.23 122.28 72.37 70.61 6.47
HP100 1.12 1.01 30.14 1114 30.14 1114 225.94 433.02 499.96 3.26 297.58 486.88 562.15 2.04
HP94 1.13 1.00 12.82 288 12.82 288 135.71 132.13 154.17 10.87 167.68 150.18 174.81 6.24
HP47 1.14 1.28 26.04 815 26.04 815 409.13 296.46 212.45 9.73 463.01 421.2 312.41 7.12
HP33 1.14 1.03 43.6 970 43.6 970 356.29 491.96 444.1 5.54 444.41 561.07 506.49 4.68
HP79 1.14 1.69 168.23 4087 168.23 4087 1462.39 1257.31 910.45 8.56 1667.94 2054.26 1182.98 7.39
HP36 1.15 1.30 222.06 6645 222.06 6480.56 1533.92 2410.66 1895.01 3.69 1899.64 2774.54 2181.06 3.21
HP85 1.15 1.16 11.97 235 11.97 235 170.85 109.83 61.49 9.19 196.45 142.6 130.52 7.99
HP71 1.16 1.47 24 505 24 505 331.53 154.61 165.56 12.24 385.12 320.1 260.36 8.42
HP83 1.17 1.03 37.36 873 37.36 873 397.5 436.37 316.71 8.67 463.85 509.21 416.86 7.43
HP103 1.18 1.28 15.63 290 15.63 290 203.02 123.53 95.56 9.20 238.95 184.66 161.75 7.82
HP40 1.19 1.39 121.42 2071 121.42 2071 875.23 873.46 739.83 8.38 1043.34 1172.31 881.93 7.03
HP30 1.19 1.30 194.97 4166 194.97 4166 817.25 1579.32 1593.13 9.40 1059.48 1873.64 1890.03 2.42
69
Efficiency Scale
HOSPITAL PHI (VRS) SE Total FTE Non-salary Total FTE Non-salary Unplanned Unplanned
staff costs staff costs re-admission re-admission
HP42 1.2 1.45 117.4 2583 117.4 2583 959.87 913.57 769.19 9.83 1149.96 1338.52 921.52 8.21
HP95 1.2 1.35 85.75 2052 85.75 2052 733.66 714.1 682.57 7.59 879.65 992.03 818.39 6.33
HP34 1.21 1.30 248.38 6920 248.38 6920 1663.77 2414.44 2045.02 7.86 2034.21 2931.74 2483.17 5.85
HP62 1.21 1.47 27.01 445 27.01 445 268.73 170.98 127.37 6.87 326.01 299.46 270.32 5.67
HP74 1.21 1.45 22.6 305 13.69 305 120.69 66.38 101.18 5.65 146.51 122.68 122.83 4.66
HP76 1.21 1.16 7.68 115 5.84 115 71.98 45.12 39.54 11.11 123.41 54.39 47.67 9.22
HP99 1.22 1.64 12.43 226 12.43 226 137.96 39.69 25.44 8.04 168.75 130.83 132.55 6.57
HP9 1.23 2.27 453.1 16626 453.1 11343.66 3041.92 2166.17 2751.16 3.76 3742.01 4733.75 3384.33 3.06
HP22 1.25 1.47 286.45 6608 286.45 6608 1996.59 2176.24 1792.74 6.14 2495.76 3230.32 2240.95 4.91
HP25 1.25 1.47 530.76 12806 530.76 12806 3340.42 4055.14 3567.89 8.42 4174.04 5576.59 4458.27 6.74
HP23 1.27 1.00 12.95 167 12.95 167 84.31 90.74 97.7 12.77 186.09 126.83 123.33 10.12
HP75 1.27 1.52 247.08 5632 247.08 5632 1648.62 1646.87 1682.25 9.55 2101.2 2665.26 2144.06 7.10
HP48 1.29 1.33 119.83 2223 119.83 2223 872.49 873.23 569.76 10.13 1124.03 1212.9 734.02 7.86
HP59 1.29 1.03 38.11 933 38.11 933 314.25 402.98 358.04 11.86 434.76 520.47 462.43 4.84
HP64 1.29 1.03 37.33 675 37.33 675 356.46 347.33 274.72 13.68 461.58 449.76 365.76 8.83
HP90 1.29 1.49 18.38 286 18.38 286 216.43 98.85 91.91 18.73 278.51 209.16 175.45 10.33
HP55 1.3 1.30 115.27 2984 115.27 2984 903.6 1027.98 824.33 9.66 1170.77 1432.79 1068.06 7.06
HP73 1.31 1.12 13.24 173 13.24 173 160.33 74.12 82.39 15.34 209.7 137.54 118.22 11.43
HP93 1.31 1.15 14.83 294 14.83 294 179.26 123.56 67.25 10.20 235.3 181.18 159.08 7.78
HP84 1.32 1.10 35.06 733 35.06 733 326.56 294.25 300.1 11.81 432.54 421.63 397.5 7.87
HP45 1.33 1.25 47.4 964 47.4 964 435.45 371.31 302.99 10.86 579.52 556.65 403.23 8.16
HP69 1.33 1.00 15.51 291 15.51 291 178.18 141.65 125.51 15.29 245.27 188.06 166.63 8.50
HP53 1.36 1.39 25.75 438 25.75 438 270.87 155.9 103.27 11.43 368.59 304.37 246.81 8.40
HP80 1.36 1.05 31.94 763 31.94 763 277.96 316.31 248.89 5.90 393.5 430.07 403.04 4.34
HP72 1.37 1.23 37.61 1300 37.61 1244.69 391.38 392.16 281.84 6.71 537.44 538.51 446.09 4.89
HP87 1.41 1.28 8.3 146 6.25 146 56.82 30.87 42.28 11.19 121.79 62.13 59.75 7.92
HP91 1.41 1.20 39.68 839 39.68 839 318.06 306.83 251.07 5.83 448.99 478.48 414.82 4.13
HP96 1.41 1.67 17.95 378 17.95 378 188.17 72.58 88.97 9.21 265.97 225.43 202.71 6.51
70
Efficiency Scale
HOSPITAL PHI (VRS) SE Total FTE Non-salary Total FTE Non-salary Unplanned Unplanned
staff costs staff costs re-admission re-admission
HP51 1.44 1.69 26.84 568 26.84 568 276.28 113.36 127.27 9.37 397.5 351.54 296.14 6.51
HP105 1.44 1.41 32.94 482 32.94 482 261.79 174.83 128.81 8.14 375.83 349.58 310.52 5.67
HP97 1.45 1.03 26.79 529 26.79 529 189.32 229.4 174.22 8.53 360.45 333.55 303.47 5.86
HP28 1.46 1.30 269.47 6124 269.47 6124 1146.75 1902.48 1731.84 8.09 1679.25 2785.91 2536.03 2.02
HP56 1.46 1.22 124.39 2750 124.39 2750 696.43 918.69 837.8 7.10 1013.49 1336.93 1219.22 4.31
HP50 1.48 1.92 31.7 490 31.7 490 275.51 89.17 79.36 10.99 406.93 357.5 305.18 7.44
HP66 1.48 1.67 30.15 1968 30.15 946.9 338.5 159.09 117.97 13.04 502.12 483.37 354.15 7.13
HP49 1.5 1.28 77.56 1296 77.56 1296 508.28 477.76 204.86 10.37 762.65 783.95 486.47 6.92
HP61 1.53 1.00 20.51 439 20.51 439 162.48 169.48 166.73 10.64 272.14 257.36 253.19 5.39
HP65 1.53 1.18 17.89 322 17.89 322 174.61 118.82 86 12.05 267 210.58 179.53 7.88
HP82 1.53 1.03 6.31 89 5.55 89 80.13 36.19 16.78 17.15 129.98 55.34 47.79 11.21
HP88 1.53 1.72 32.92 560 32.92 560 267.18 139.98 119.72 9.36 409.66 382.24 340.8 6.11
HP37 1.6 1.52 849.95 15990 849.95 15990 2578.66 3261.6 4483.67 8.22 4113.43 5866.12 7152.27 5.15
HP57 1.61 2.04 35 743 35 743 297.75 129.24 114.41 11.27 479.2 439.46 329.44 7.00
HP68 1.66 1.82 27.81 487 27.81 487 242.37 101.73 87.18 21.14 402.66 338.41 278.96 8.83
HP106 1.69 1.00 28.69 537 28.69 537 186.51 203.79 191.94 10.33 357.29 344.03 324.03 5.32
HP54 1.81 2.44 58.45 952 58.45 952 311.58 121.62 129.48 8.68 564.24 600.51 463.84 4.79
HP81 1.83 1.00 8.45 102 8.45 102 77.77 33.38 40.06 22.22 156.54 82.35 73.06 12.06
HP77 1.96 2.00 18.15 350 18.15 350 118.46 53.51 46.08 11.11 232.7 210.25 206.16 5.66
HP70 2.08 1.52 9.46 158 9.46 158 79.15 21.22 15.29 19.31 164.96 102.73 94.36 9.26
NON-FRONTIER HOSPITALS
TOTAL 11540 285793 11515 259100 75018 85619 81538 88985 112898 98756
AVERAGE 1.29 1.33 9.67 6.74
No. of less efficient hospitals 79
% less efficient hospitals 75
Notes: * indicates hospital is the optimal size defined by Constant Returns to Scale
a the inverse of this was used as an output in the model
Slacks are derived by the product of the efficiency score and the actual input or output, subtracted from the target for the input or output.
71
Annex 5.1.2: Model 2 - Metropolitan/large country hospitals
Efficiency Scale
score score Actual inputs Input targets Actual outputs Output targets
HOSPITAL PHI (VRS) SE Non- Medical Non- Non- Medical Non- Unplanned Unplanned
medical FTE staff salary medical FTE staff salary re-admission re-admission
FTE staff costs FTE staff costs WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($‘000) ($‘000)
HP1 1 1.18 715.62 51.02 18413 715.62 51.02 18413 6024.45 5976.86 2298.6 7.89 6024.45 5976.86 2298.6 7.89
HP2 1 1.18 1739.3 273.84 48702 1739.3 273.84 48702 12115.02 17554.22 16422.44 6.63 12115.02 17554.22 16422.44 6.63
HP3 1 1.05 1283.44 180.29 28483 1283.44 180.29 28483 3788.66 8428.51 13520.01 13.35 3788.66 8428.51 13520.01 13.35
HP4 1 1.00 * 1063.22 88.61 20069 1063.22 88.61 20069 9239.72 10683.23 4250.59 10.28 9239.72 10683.23 4250.59 10.28
HP5 1 1.00 * 374.17 50.45 8870 374.17 50.45 8870 1077.78 2511.97 4565.45 4.36 1077.78 2511.97 4565.45 4.36
HP6 1 1.00 * 2600.09 290.54 47790 2600.09 290.54 47790 11984.28 16655.07 21799.12 11.27 11984.28 16655.07 21799.12 11.27
HP7 1 1.00 * 812.3 100.92 18310 812.3 100.92 18310 5029.88 8718.33 7372.95 8.83 5029.88 8718.33 7372.95 8.83
HP8 1 1.05 816.95 86.55 19687 816.95 86.55 19687 6336.38 8167.83 7478.84 9.13 6336.38 8167.83 7478.84 9.13
HP9 1 1.47 425.8 27.3 16626 346.38 27.3 9594.13 3041.92 2166.17 2751.16 3.76 3049.74 3477.16 2758.23 3.75
HP10 1 1.09 1334.27 149.13 29820 1334.27 149.13 29820 10686.42 13277.6 10082.34 15.50 10686.42 13277.6 10082.34 15.50
HP11 1 1.00 * 458.25 57.97 10355 458.25 57.97 10355 3993.22 5014.07 4125.03 8.76 3993.22 5014.07 4125.03 8.76
HP12 1 1.00 * 99.93 0 3360 99.93 0 3360 1041.53 1049.94 977.04 3.09 1041.53 1049.94 977.04 3.09
HP14 1 1.08 2390.54 322.31 62241 2390.54 322.31 62241 11965.57 17454.41 26061.18 11.33 11965.57 17454.41 26061.18 11.33
HP16 1 1.00 * 255.15 5.47 6121 255.15 5.47 6121 2042.32 3225.78 2015.74 4.36 2042.32 3225.78 2015.74 4.36
HP18 1 1.00 * 165.65 9.52 5119 165.65 9.52 5119 1544.5 2153.06 1930.04 5.39 1544.5 2153.06 1930.04 5.39
HP20 1 1.01 500.1 40.12 12709 500.1 40.12 12709 3370.37 5409.52 4997.13 7.29 3370.37 5409.52 4997.13 7.29
HP21 1 1.14 1084.98 118.51 30153 1084.98 118.51 30153 9210.84 10761.21 8953.54 9.89 9210.84 10761.21 8953.54 9.89
HP24 1 1.15 391.57 12.98 12359 391.57 12.98 12359 2490.45 4204.52 3400.87 9.92 2490.45 4204.52 3400.87 9.92
HP29 1 1.00 * 204.93 2.87 4312 204.93 2.87 4312 1903.56 1498.89 980.27 11.14 1903.56 1498.89 980.27 11.14
HP30 1 1.00 * 191.31 3.67 4166 191.31 3.67 4166 817.25 1579.32 1593.13 9.40 817.25 1579.32 1593.13 9.40
HP31 1 1.19 851.89 82.08 25399 851.89 82.08 25399 6852.34 8689.47 7673.53 7.55 6852.34 8689.47 7673.53 7.55
HP32 1 1.05 540.56 22.13 17053 540.56 22.13 17053 3014.11 5041.9 5692.09 5.49 3014.11 5041.9 5692.09 5.49
HP35 1 1.00 * 291.91 16.29 8551 291.91 16.29 8551 2996.7 4160.34 2592.08 7.65 2996.7 4160.34 2592.08 7.65
HP36 1 1.08 216.96 5.1 6645 216.96 5.1 6645 1533.92 2410.66 1895.01 3.69 1533.92 2410.66 1895.01 3.69
HP17 1.01 1.09 1456.91 221.54 52260 1456.91 191.45 38216.56 7489.94 10300.59 15834.22 9.20 7543.63 10997.04 15947.71 7.64
HP19 1.02 1.01 623.58 67.34 14636 623.58 67.34 14636 4157.2 6423.26 5763.47 9.53 4223.49 6525.68 5855.37 8.10
72
Annex 5.1.2: Model 2 - Metropolitan/large country hospitals (continued)
Efficiency Scale
medical FTE staff salary medical FTE staff salary re-admission re-admission
FTE staff costs FTE costs WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($‘000) staff ($‘000)
HP13 1.04 1.05 778.54 91.21 17141 778.54 89.99 17141 5598.75 6678.03 6422.84 10.21 5835.91 7455.77 6694.91 9.49
HP15 1.05 1.03 191.99 10.72 5376 191.99 10.72 5376 1672.47 1958.33 1700.72 3.92 1762.53 2149.19 1792.3 3.72
HP23 1.06 1.03 242.95 4.13 5632 211.96 4.13 5514.16 1648.62 1646.87 1682.25 9.55 1739.93 2560.16 1775.42 3.86
HP27 1.06 1.10 345.99 17 11793 345.99 17 10640.68 2406.49 3182.76 3351.36 7.84 2553.65 3859.51 3556.3 5.99
HP26 1.07 1.09 341.17 14.67 10277 341.17 14.67 10277 2258.65 3201.23 3211.35 6.74 2423.34 3751.04 3445.51 5.48
HP33 1.08 1.01 164.09 4.14 4087 164.09 2.22 4087 1462.39 1257.31 910.45 8.56 1584.23 1420.67 1046.77 5.24
HP25 1.09 1.18 504.42 26.33 12806 461.44 26.33 12806 3340.42 4055.14 3567.89 8.42 3656.91 5159.86 3905.93 6.88
HP34 1.13 1.04 240.43 7.95 6920 240.43 7.95 6920 1663.77 2414.44 2045.02 7.86 1872.9 2763.39 2302.07 4.32
HP22 1.16 1.02 277.77 8.68 6608 277.77 8.68 6608 1996.59 2176.24 1792.74 6.14 2311.89 3412.51 2075.84 4.76
HP28 1.26 1.05 262.16 7.3 6124 250.19 7.3 6124 1146.75 1902.48 1731.84 8.09 1450.03 2490.56 2189.86 6.40
HP37 1.4 1.08 797.08 52.87 15990 655.48 52.87 15990 2578.66 3261.6 4483.67 8.22 3609.49 5699.57 6276.03 5.88
TOTAL 6227.08 533.88 169650 5999.54 500.65 154336.4 37420.7 48458.28 52497.82 40567.93 58244.95 56864.02
AVERAGE 1.11 1.05 8.02 5.98
73
Annex 5.1.3: Model 3 – Small rural hospitals
Efficiency Scale
HOSPITAL PHI SE Total FTE Non- Total FTE Non- Unplanned Unplanned
(VRS) staff salary staff salary re-admission re-admission
costs costs WEIS1 WEIS2 WEIS3 rate a WEIS1 WEIS2 WEIS3 ratea
($‘000) ($‘000)
HP39 1 1.06 61.48 1743 61.48 1743 589.63 907.89 733.33 6.47 589.63 907.89 733.33 6.47
HP41 1 1.00 * 30.45 1388 30.45 1388 192.21 530.17 724.29 1.32 192.21 530.17 724.29 1.32
HP43 1 1.79 148.29 2654 148.29 2654 1182.27 995.63 886.83 9.03 1182.27 995.63 886.83 9.03
HP44 1 1.00 * 47.11 535 47.11 535 407.77 414.09 459.04 12.64 407.77 414.09 459.04 12.64
HP46 1 1.00 * 32.5 97 32.5 97 134.14 65.55 79.56 15.13 134.14 65.55 79.56 15.13
HP52 1 1.00 * 19.44 654 19.44 654 406.03 341.74 267.95 7.11 406.03 341.74 267.95 7.11
HP55 1 1.37 82.91 1823 82.91 1823 837.29 825.09 645.54 6.55 837.29 825.09 645.54 6.55
HP67 1 1.25 93.6 1643 93.6 1643 621.72 887.73 801.96 8.11 621.72 887.73 801.96 8.11
HP78 1 1.00 * 3.94 76 3.94 76 120.55 44.67 39.84 11.96 120.55 44.67 39.84 11.96
HP89 1 1.59 127.74 2920 127.74 2920 1264.75 1121.69 741.2 9.34 1264.75 1121.69 741.2 9.34
HP98 1 1.00 * 43.8 301 43.8 301 439.95 377.51 331.13 12.99 439.95 377.51 331.13 12.99
HP101 1 1.00 * 39.23 268 39.23 268 381.34 321.74 264.98 8.76 381.34 321.74 264.98 8.76
HP102 1 1.67 115.27 2984 115.27 2984 903.6 1027.98 824.33 9.66 903.6 1027.98 824.33 9.66
HP104 1 1.27 11.6 224 11.6 224 115.11 55.31 30.24 5.14 115.11 55.31 30.24 5.14
HP56 1.02 1.75 124.39 2750 124.39 2271.85 696.43 918.69 837.8 7.10 951.54 933.24 851.07 6.37
HP40 1.04 1.59 121.42 2071 109.72 2071 875.23 873.46 739.83 8.38 906.65 904.82 766.39 8.08
HP92 1.04 1.56 52.09 1254 52.09 1254 598.46 452.37 370.54 6.33 623.33 565.25 431.95 6.08
HP38 1.05 1.02 26.42 1139 26.42 1139 156.39 398.55 362.44 1.64 175.72 428.59 575.82 1.56
HP63 1.05 1.27 37.91 783 37.91 783 479.42 371.82 306.31 12.53 504.96 433.08 335.75 8.46
HP86 1.05 1.47 49.02 880 49.02 880 528.49 384.14 281.65 8.32 554.58 492.1 390.6 7.93
HP42 1.06 1.64 117.4 2583 117.4 2419.85 959.87 913.57 769.19 9.83 1013.02 991.85 811.78 8.08
HP58 1.08 1.35 53.93 1039 53.93 1039 539.58 435.03 431.47 7.35 581.35 562.69 464.87 6.82
HP95 1.09 1.49 85.75 2052 85.75 2052 733.66 714.1 682.57 7.59 800.38 940.2 744.64 6.96
HP48 1.11 1.54 119.83 2223 107.65 2223 872.49 873.23 569.76 10.13 965.5 966.32 718.38 9.15
HP60 1.12 1.35 19.1 191 7.84 191 104.16 39.18 63.17 7.23 122.28 72.37 70.61 6.47
74
Annex 5.1.3: Model 3 - Small rural hospitals (continued)
Efficiency Scale
HOSPITAL PHI (VRS) SE Total FTE Non- Total FTE Non- Unplanned Unplanned
staff salary staff salary re-admission re-admission
costs costs WEIS1 WEIS2 WEIS3 rate a WEIS1 WEIS2 WEIS3 rate a
($‘000) ($‘000)
HP94 1.12 1.28 26.04 815 26.04 792.09 409.13 296.46 212.45 9.73 458.36 389.27 296.79 7.21
HP100 1.12 1.01 30.14 1114 30.14 1114 225.94 433.02 499.96 3.26 297.58 486.88 562.15 2.04
HP79 1.13 1.01 12.82 288 12.82 288 135.71 132.13 154.17 10.87 167.68 150.18 174.81 6.24
HP47 1.14 1.03 43.6 970 43.6 970 356.29 491.96 444.1 5.54 444.41 561.07 506.49 4.68
HP85 1.15 1.16 11.97 235 11.97 235 170.85 109.83 61.49 9.19 196.45 142.6 130.52 7.99
HP71 1.16 1.47 24 505 24 505 331.53 154.61 165.56 12.24 385.12 320.1 260.36 8.42
HP83 1.17 1.03 37.36 873 37.36 873 397.5 436.37 316.71 8.67 463.63 508.97 416.88 7.43
HP103 1.18 1.28 15.63 290 15.63 290 203.02 123.53 95.56 9.20 238.95 184.66 161.75 7.82
HP62 1.21 1.47 27.01 445 27.01 445 268.73 170.98 127.37 6.87 326.01 299.46 270.32 5.67
HP74 1.21 1.45 22.6 305 13.69 305 120.69 66.38 101.18 5.65 146.51 122.68 122.83 4.66
HP76 1.21 1.16 7.68 115 5.84 115 71.98 45.12 39.54 11.11 123.41 54.39 47.67 9.22
HP99 1.22 1.64 12.41 226 12.41 226 137.96 39.69 25.44 8.04 168.7 130.86 132.69 6.57
HP75 1.26 1.01 12.95 167 12.95 167 84.31 90.74 97.7 12.77 186.09 126.83 123.33 10.12
HP64 1.28 1.04 37.33 675 37.33 675 356.46 347.33 274.72 13.68 456.7 445.01 367.05 8.70
HP59 1.29 1.03 38.11 933 38.11 933 314.25 402.98 358.04 11.86 434.76 520.47 462.43 4.84
HP90 1.29 1.49 18.38 286 18.38 286 216.43 98.85 91.91 18.73 278.51 209.16 175.45 10.33
HP72 1.3 1.30 37.61 1300 37.61 1100.37 391.38 392.16 281.84 6.71 508.03 509.05 412.36 5.17
HP45 1.31 1.28 47.4 964 47.4 964 435.45 371.31 302.99 10.86 569.2 508.05 396.05 8.21
HP73 1.31 1.12 13.24 173 13.24 173 160.33 74.12 82.39 15.34 209.7 137.54 118.22 11.43
HP84 1.31 1.11 35.06 733 35.06 733 326.56 294.25 300.1 11.81 429.27 436.39 394.49 8.44
HP93 1.31 1.15 14.83 294 14.83 294 179.26 123.56 67.25 10.20 235.3 181.18 159.08 7.78
HP69 1.33 1.00 15.51 291 15.51 291 178.18 141.65 125.51 15.29 245.27 188.06 166.63 8.50
HP53 1.36 1.05 31.94 763 31.94 763 277.96 316.31 248.89 5.90 393.5 430.07 403.04 4.34
HP80 1.36 1.39 25.75 438 25.75 438 270.87 155.9 103.27 11.43 368.59 304.37 246.81 8.40
HP49 1.38 1.39 77.56 1296 70.45 1296 508.28 477.76 204.86 10.37 702.55 660.36 522.73 7.51
HP91 1.39 1.22 39.68 839 39.68 839 318.06 306.83 251.07 5.83 442.63 454.49 422.05 4.19
75
Efficiency Scale
HOSPITAL PHI SE Total FTE Non- Total FTE Non- Unplanned Unplanned
(VRS) staff salary staff salary re-admission re-admission
costs costs WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($‘000) ($‘000)
HP87 1.41 1.28 8.3 146 6.25 146 56.82 30.87 42.28 11.19 121.79 62.13 59.75 7.92
HP96 1.41 1.67 17.95 378 17.95 378 188.17 72.58 88.97 9.21 265.97 225.43 202.71 6.51
HP51 1.44 1.41 32.94 482 32.94 482 261.79 174.83 128.81 8.14 375.83 349.58 310.52 5.67
HP105 1.44 1.69 26.84 568 26.84 568 276.28 113.36 127.27 9.37 397.5 351.54 296.14 6.51
HP66 1.45 1.03 26.79 529 26.79 529 189.32 229.4 174.22 8.53 360.45 333.55 303.47 5.86
HP97 1.45 1.72 30.15 1968 30.15 878.09 338.5 159.09 117.97 13.04 490.95 418.87 314.75 7.28
HP50 1.48 1.92 31.7 490 31.7 490 275.51 89.17 79.36 10.99 406.93 357.5 305.18 7.44
HP65 1.52 1.01 20.51 439 20.51 439 162.48 169.48 166.73 10.64 272.14 257.36 253.19 5.39
HP61 1.53 1.18 17.89 322 17.89 322 174.61 118.82 86 12.05 267 210.58 179.53 7.88
HP82 1.53 1.72 32.92 560 32.92 560 267.18 139.98 119.72 9.36 409.12 380.35 341.41 6.12
HP88 1.53 1.03 6.31 89 5.55 89 80.13 36.19 16.78 17.15 129.98 55.34 47.79 11.21
HP57 1.59 2.08 35 743 35 743 297.75 129.24 114.41 11.27 472.25 416.45 337.19 7.11
HP68 1.66 1.82 27.81 487 27.81 487 242.37 101.73 87.18 21.14 402.66 338.41 278.96 8.83
HP106 1.69 1.00 28.69 537 28.69 537 186.51 203.79 191.94 10.33 357.29 344.03 324.03 5.32
HP54 1.74 2.56 58.45 952 58.45 952 311.58 121.62 129.48 8.68 543.15 531.85 485.14 4.98
HP81 1.82 1.01 8.45 102 8.45 102 77.77 33.38 40.06 22.22 156.54 82.35 73.06 12.06
HP77 1.96 2.00 18.15 350 18.15 350 118.46 53.51 46.08 11.11 232.7 210.25 206.16 5.66
HP70 2.08 1.52 9.46 158 9.46 158 79.15 21.22 15.29 19.31 164.96 102.73 94.36 9.26
TOTAL 1962.18 41598 1906.37 39644.25 16975.67 14466.23 12221.35 21903.43 20780.96 18028.18
AVERAGE 1.33 1.30 10.39 7.14
Notes : * indicates hospital is the optimal size defined by Constant Returns to Scale
76
Annex 5.1.4:Model 4 – Metropolitan/large country hospitals
Efficiency Scale
HOSPITAL PHI (VRS) SE Non- Medical Non- Non- Medical Non- salary Unplanned Unplanned
medical salaries salary medical salaries costs re-admission re-admission
salaries ($’000) costs salaries ($’000) ($‘000) WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($’000) ($‘000) ($’000)
HP2 1 1.19 70781 22636 48702 70781 22636 48702 12115.02 17554.22 16422.44 6.63 12115.02 17554.22 16422.44 6.63
HP3 1 1.05 44275 23291 28483 44275 23291 28483 3788.66 8428.51 13520.01 13.35 3788.66 8428.51 13520.01 13.35
HP4 1 1.00 * 42997 9132 20069 42997 9132 20069 9239.72 10683.23 4250.59 10.28 9239.72 10683.23 4250.59 10.28
HP5 1 1.00 * 12689 5528 8870 12689 5528 8870 1077.78 2511.97 4565.45 4.36 1077.78 2511.97 4565.45 4.36
HP6 1 1.00 * 68551 17730 47790 68551 17730 47790 11984.28 16655.07 21799.12 11.27 11984.28 16655.07 21799.12 11.27
HP7 1 1.00 * 30419 8515 18310 30419 8515 18310 5029.88 8718.33 7372.95 8.83 5029.88 8718.33 7372.95 8.83
HP10 1 1.08 52685 13802 29820 52685 13802 29820 10686.42 13277.6 10082.34 15.50 10686.42 13277.6 10082.34 15.50
HP11 1 1.00 * 19931 6390 10355 19931 6390 10355 3993.22 5014.07 4125.03 8.76 3993.22 5014.07 4125.03 8.76
HP12 1 1.00 * 3526 0 3360 3526 0 3360 1041.53 1049.94 977.04 3.09 1041.53 1049.94 977.04 3.09
HP14 1 1.10 86757 27362 62241 86757 27362 62241 11965.57 17454.41 26061.18 11.33 11965.57 17454.41 26061.18 11.33
HP16 1 1.00 * 10210 396 6121 10210 396 6121 2042.32 3225.78 2015.74 4.36 2042.32 3225.78 2015.74 4.36
HP18 1 1.00 * 7998 644 5119 7998 644 5119 1544.5 2153.06 1930.04 5.39 1544.5 2153.06 1930.04 5.39
HP19 1 1.00 * 19943 6008 14636 19943 6008 14636 4157.2 6423.26 5763.47 9.53 4157.2 6423.26 5763.47 9.53
HP21 1 1.00 * 54996 208 30153 54996 208 30153 9210.84 10761.21 8953.54 9.89 9210.84 10761.21 8953.54 9.89
HP26 1 1.04 11650 986 10277 11650 986 10277 2258.65 3201.23 3211.35 6.74 2258.65 3201.23 3211.35 6.74
HP29 1 1.00 * 6815 203 4312 6815 203 4312 1903.56 1498.89 980.27 11.14 1903.56 1498.89 980.27 11.14
HP30 1 1.00 * 5903 284 4166 5903 284 4166 817.25 1579.32 1593.13 9.40 817.25 1579.32 1593.13 9.40
HP31 1 1.16 36800 3282 25399 36800 3282 25399 6852.34 8689.47 7673.53 7.55 6852.34 8689.47 7673.53 7.55
HP32 1 1.01 20018 1445 17053 20018 1445 17053 3014.11 5041.9 5692.09 5.49 3014.11 5041.9 5692.09 5.49
HP35 1 1.00 * 11299 1943 8550 11299 1943 8550 2996.7 4160.34 2592.08 7.65 2996.7 4160.34 2592.08 7.65
HP36 1 1.03 7678 458 6645 7678 458 6645 1533.92 2410.66 1895.01 3.69 1533.92 2410.66 1895.01 3.69
HP8 1.01 1.03 33599 7825 19687 33599 7825 19687 6336.38 8167.83 7478.84 9.13 6395.87 8244.52 7549.06 8.24
HP15 1.01 1.01 6815 649 5376 6815 649 5376 1672.47 1958.33 1700.72 3.92 1696.78 1986.8 1725.44 3.66
HP24 1.01 1.11 14886 1048 12359 14886 1048 10625.44 2490.45 4204.52 3400.87 9.92 2814.74 4238.69 3428.5 5.32
HP17 1.03 1.04 50991 18822 52260 50991 13894.35 35555.65 7489.94 10300.59 15834.22 9.20 8555.86 12209.24 16381.78 7.52
HP13 1.04 1.03 29539 8476 17141 29539 8386.06 17141 5598.75 6678.03 6422.84 10.21 5839.76 7491.07 6699.33 9.58
HP1 1.05 1.22 29352 5091 18413 29352 5091 15311.47 6024.45 5976.86 2298.6 7.89 6316.22 7522.39 3730.99 7.52
77
Annex 5.1.4:Model 4 - Metropolitan/large country hospitals (continued)
Efficiency Scale
medical salaries salary medical salaries salary re-admission re-admission
salaries ($’000) costs salaries ($’000) costs WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($’000) ($‘000) ($’000) ($‘000)
HP9 1.05 1.54 16579 2716 16626 16579 2716 11222.81 3041.92 2166.17 2751.16 3.76 3196.86 3980.37 3499.2 3.58
HP20 1.05 1.00 20608 5920 12709 20608 5920 12709 3370.37 5409.52 4997.13 7.29 3528.23 5662.89 5231.19 6.49
HP23 1.05 1.00 8501 181 5632 8470.39 181 5632 1648.62 1646.87 1682.25 9.55 1736.95 2314.06 1772.38 3.91
HP34 1.05 1.04 8539 545 6920 8539 545 6920 1663.77 2414.44 2045.02 7.86 1753.1 2544.08 2154.82 3.90
HP22 1.07 1.05 8925 644 6608 8925 644 6507.42 1996.59 2176.24 1792.74 6.14 2144.12 2590.11 1925.21 4.16
HP33 1.09 1.00 5623 273 4087 5623 151.4 4011.11 1462.39 1257.31 910.45 8.56 1589.91 1366.95 1003.23 5.55
HP25 1.13 1.08 17287 2327 12806 17287 2327 12806 3340.42 4055.14 3567.89 8.42 3771.94 5185.44 4028.79 7.36
HP27 1.21 1.06 15969 1360 11793 15969 1360 11793 2406.49 3182.76 3351.36 7.84 2900.29 3835.84 4039.04 4.12
HP28 1.28 1.02 8956 571 6124 8956 571 6124 1146.75 1902.48 1731.84 8.09 1523.75 2428.48 2210.66 6.34
HP37 1.58 1.00 29589 5545 15990 23115.69 5545 15990 2578.66 3261.6 4483.67 8.22 4063.03 5674.92 7064.63 5.22
TOTAL 250458 52471 187109 243954.1 47331.81 161723.5 41769.12 50428.01 51869.17 46920.02 62805.84 59741.25
AVERAGE 1.11 1.06 7.93 5.79
% less efficient hospitals 56.76
78
Annex 5.1.5: Model 5 – Small rural hospitals
Efficiency Scale
HOSPITAL PHI SE Total Non-salary Total Non-salary Unplanned Unplanned
(VRS) salaries costs ($’000) salaries costs re-admission re-admission
($’000) ($’000) ($’000) WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
HP39 1 1.00 * 1169.0 1388 1169 1388 192.21 530.17 724.29 1.32 192.21 530.17 724.29 1.32
HP41 1 2.33 2878.0 1743 2878 1743 589.63 907.89 733.33 6.47 589.63 907.89 733.33 6.47
HP43 1 1.12 817.0 535 817 535 407.77 414.09 459.04 12.64 407.77 414.09 459.04 12.64
HP44 1 3.13 5022.0 2654 5022 2654 1182.27 995.63 886.83 9.03 1182.27 995.63 886.83 9.03
HP46 1 2.63 3007.0 1823 3007 1823 837.29 825.09 645.54 6.55 837.29 825.09 645.54 6.55
HP52 1 3.13 4798.0 2920 4798 2920 1264.75 1121.69 741.20 9.34 1264.75 1121.69 741.2 9.34
HP66 1 1.00 * 165.0 1968 165 1968 338.50 159.09 117.97 13.04 338.5 159.09 117.97 13.04
HP67 1 1.00 * 422.0 268 422 268 381.34 321.74 264.98 8.76 381.34 321.74 264.98 8.76
HP89 1 1.00 * 147.0 76 147 76 120.55 44.67 39.84 11.96 120.55 44.67 39.84 11.96
HP98 1 1.00 * 693.0 301 693 301 439.95 377.51 331.13 12.99 439.95 377.51 331.13 12.99
HP101 1 1.00 * 191.0 97 191 97 134.14 65.55 79.56 15.13 134.14 65.55 79.56 15.13
HP102 1 1.20 408.0 224 408 224 115.11 55.31 30.24 5.14 115.11 55.31 30.24 5.14
HP104 1 2.08 2997.0 1643 2997 1643 621.72 887.73 801.96 8.11 621.72 887.73 801.96 8.11
HP55 1.01 3.45 4944.0 2984 4941.7 2749.35 903.60 1027.98 824.33 9.66 1211.84 1040.82 834.63 9.14
HP56 1.02 3.13 4212.0 2750 4212 2278.6 696.43 918.69 837.80 7.10 961.42 941 852.84 6.98
HP95 1.03 2.63 2910.0 2052 2910 1760.34 733.66 714.10 682.57 7.59 754.1 734 701.59 4.48
HP40 1.04 2.78 3795.0 2071 3721.51 2071 875.23 873.46 739.83 8.38 906.65 904.82 766.39 8.08
HP92 1.04 2.56 1737.0 1254 1737 1117.98 598.46 452.37 370.54 6.33 619.78 567.84 438.14 6.12
HP38 1.05 1.05 1053.0 1139 1006.21 1139 156.39 398.55 362.44 1.64 175.72 428.59 575.82 1.56
HP42 1.05 3.23 4189.0 2583 4189 2449.27 959.87 913.57 769.19 9.83 1008.46 959.81 808.12 5.00
HP86 1.1 2.22 1578.0 880 1578 880 528.49 384.14 281.65 8.32 580.47 525.82 456.57 7.58
HP48 1.11 2.78 4157.0 2223 3771.93 2223 872.49 873.23 569.76 10.13 965.5 966.32 718.38 9.15
HP60 1.12 1.28 344.0 191 285.64 191 104.16 39.18 63.17 7.23 122.28 72.37 70.61 6.47
HP58 1.13 2.13 1784.0 1039 1784 1039 539.58 435.03 431.47 7.35 611.18 566.06 488.72 6.49
HP78 1.18 1.89 1248.0 654 1152.07 654 406.03 341.74 267.95 7.11 477.52 453.64 431.54 6.04
HP47 1.19 1.82 1519.0 970 1519 970 356.29 491.96 444.10 5.54 451.21 583.58 533.26 4.67
HP94 1.19 1.85 972.0 815 972 610.8 409.13 296.46 212.45 9.73 487.54 423.3 330.2 8.16
79
Efficiency Scale
HOSPITAL PHI SE Total Non-salary Total Non-salary Unplanned Unplanned
(VRS) salaries costs salaries costs re-admission re-admission
($’000) ($’000) ($’000) ($’000) WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
HP63 1.2 1.96 1363.0 783 1363 783 479.42 371.82 306.31 12.53 573.15 496.29 384.11 10.48
HP74 1.21 1.32 483.0 305 414.4 305 120.69 66.38 101.18 5.65 146.51 122.68 122.83 4.66
HP76 1.21 1.16 267.0 115 202.29 115 71.98 45.12 39.54 11.11 123.41 54.39 47.67 9.22
HP83 1.23 1.96 1385.0 873 1385 805.02 397.50 436.37 316.71 8.67 490.2 538.14 467.97 7.03
HP100 1.27 1.35 1369.0 1114 1369 1114 225.94 433.02 499.96 3.26 351.21 551.84 637.15 2.56
HP72 1.34 2.08 1435.0 1300 1435 907.12 391.38 392.16 281.84 6.71 525.13 526.18 471.4 5.00
HP99 1.34 1.49 487.0 226 394.98 226 137.96 39.69 25.44 8.04 184.75 123.97 91.95 6.00
HP75 1.35 1.05 371.0 167 262.95 167 84.31 90.74 97.70 12.77 215.43 152.35 132.21 9.43
HP59 1.37 1.75 1378.0 933 1378 924.72 314.25 402.98 358.04 11.86 431.58 553.44 491.72 4.90
HP49 1.38 2.38 2459.0 1296 2183.43 1296 508.28 477.76 204.86 10.37 702.55 660.36 522.73 7.51
HP62 1.38 1.69 947.0 445 611.25 445 268.73 170.98 127.37 6.87 369.8 368.87 352.61 5.00
HP85 1.39 1.28 471.0 235 398.33 235 170.85 109.83 61.49 9.19 237.14 176.44 138.13 6.62
HP45 1.4 2.22 1626.0 964 1626 964 435.45 371.31 302.99 10.86 607.54 549.21 436.89 7.78
HP71 1.4 1.61 873.0 505 873 505 331.53 154.61 165.56 12.24 465.52 408.48 338.21 8.72
HP87 1.41 1.28 293.0 146 229.41 146 56.82 30.87 42.28 11.19 121.79 62.13 59.75 7.92
HP64 1.43 1.67 1305.0 675 1269.05 675 356.46 347.33 274.72 13.68 510.01 496.95 432.2 9.56
HP91 1.46 1.96 1529.0 839 1305.96 839 318.06 306.83 251.07 5.83 463.61 485.98 491.93 4.00
HP103 1.46 1.41 550.0 290 441.25 290 203.02 123.53 95.56 9.20 295.96 248.77 209.33 6.31
HP80 1.51 1.61 1166.0 763 1114.86 763 277.96 316.31 248.89 5.90 419.55 477.43 478.26 3.91
HP84 1.52 1.61 1228.0 733 1228 733 326.56 294.25 300.10 11.81 496.67 476.1 456.42 7.76
HP82 1.53 1.03 217.0 89 167.57 89 80.13 36.19 16.78 17.15 129.98 55.34 47.79 11.21
HP105 1.53 1.75 1012.0 482 778.32 482 261.79 174.83 128.81 8.14 401.05 403.46 396.05 5.32
HP73 1.55 1.06 411.0 173 286.95 173 160.33 74.12 82.39 15.34 248.04 180.45 149.8 9.91
HP79 1.56 1.09 500.0 288 402.64 288 135.71 132.13 154.17 10.87 312.71 271.75 240.6 6.96
HP51 1.59 1.79 890.0 568 890 568 276.28 113.36 127.27 9.37 439.41 418.14 374.68 5.89
HP53 1.59 1.54 858.0 438 808.17 438 270.87 155.90 103.27 11.43 430.11 401.56 375.74 7.20
HP50 1.6 1.75 1106.0 490 890.07 490 275.51 89.17 79.36 10.99 441.18 414.02 389.25 6.86
80
Efficiency Scale
HOSPITAL PHI SE Total Non-salary Total Non- Unplanned Unplanned
(VRS) salaries costs ($’000) salaries salary re-admission re-admission
($’000) ($’000) costs WEIS1 WEIS2 WEIS3 ratea WEIS1 WEIS2 WEIS3 ratea
($’000)
HP88 1.64 1.96 1069.0 560 954.01 560 267.18 139.98 119.72 9.36 438.13 426.76 411.59 5.71
HP93 1.64 1.33 606.0 294 443.96 294 179.26 123.56 67.25 10.20 294.57 248.84 210.41 6.21
HP57 1.71 2.08 1174.0 743 1174 743 297.75 129.24 114.41 11.27 509.38 465.8 382.34 6.59
HP54 1.75 2.70 1814.0 952 1628.84 952 311.58 121.62 129.48 8.68 543.83 525.69 508.32 4.97
HP96 1.75 1.67 663.0 378 497.21 378 188.17 72.58 88.97 9.21 329.86 311.16 284.04 5.25
HP97 1.8 1.39 1000.0 529 812.32 529 189.32 229.40 174.22 8.53 393.52 413.18 413.43 4.73
HP81 1.84 1.01 247.0 102 187.46 102 77.77 33.38 40.06 22.22 152.41 80.01 73.82 12.06
HP90 1.88 1.10 709.0 286 541.95 286 216.43 98.85 91.91 18.73 406.13 346.61 295.17 9.98
HP61 1.92 1.37 636.0 322 460.07 322 174.61 118.82 86.00 12.05 335.58 297.21 258.11 6.27
HP106 1.94 1.20 717.0 537 717 537 186.51 203.79 191.94 10.33 372.15 396.01 372.98 5.00
HP68 2 1.59 1001.0 487 1000.86 487 242.37 101.73 87.18 21.14 485.32 423.38 378.13 10.56
HP69 2.04 1.08 534.0 291 438.16 291 178.18 141.65 125.51 15.29 362.76 312.15 262.78 7.51
HP65 2.1 1.16 829.0 439 605.38 439 162.48 169.48 166.73 10.64 370.33 367.38 349.76 5.07
HP77 2.24 1.67 591.0 350 482.45 350 118.46 53.51 46.08 11.11 264.82 240.11 217.3 4.97
HP70 2.28 1.37 298.0 158 276.59 158 79.15 21.22 15.29 19.31 180.79 112.74 90.63 8.45
TOTAL 67365 40284 62767.5 38577.85 17043.20 14648.88 12371.33 23695.40 22768.90 20418.37
AVERAGE 1.47 1.59 10.28 6.76
No of less efficient hospitals 56
% less efficient hospitals 81.16
81
5 CASE STUDIES
5.2 Assessment of the performance of oral health services for

Queensland school students
5.2.1 Introduction 11
The Queensland Treasury is undertaking pilot studies in the Queensland public
sector to apply DEA. DEA is particularly useful to public sector managers
because it does not require inputs or outputs to be priced.12
The first study, in conjunction with Queensland Health, applied DEA to
determine the relative performance of units providing oral health services to
Queensland students from 1992-93 to 1994-95. Oral (or dental) health services
are administered through thirteen geographical regions and undertaken in fixed
and mobile dental clinics which visit each school at least once a year. The aim is
to examine and treat each child to achieve acceptable oral health.
Data requirements
DEA measures the efficiency of service providers relative to those included in
the sample only, so more observations will lead to better results usually. When
there are few observations, the service providers being compared are more
likely to be unique in the combinations of inputs used and outputs produced, and
the model will determine a larger number of the providers as efficient.
Increasing the number of inputs or outputs in the analysis exacerbates this
problem because there is more potential again for providers to be unique within
the sample.
There is no strict minimum number of observations required to undertake a
DEA but a general rule for the minimum sample size is the sum of the number
of inputs and the number of outputs multiplied by three. For example, a study
such as oral health services, which has five inputs and one output, would require
a minimum of eighteen observations: that is five inputs plus one output,
multiplied by three. Relative efficiency scores tend to decrease as the sample
size increases, improving the explanatory power of the model.
11 By Patrizia Santin-Dore and Jennifer Pallesen of the Economics Division of the

Queensland Treasury. The assistance of Steve Shackcloth and Ian Proud from Queensland
Health in providing data and feedback on the model’s results is particularly appreciated.
12 For a brief overview of the background, theory and application of DEA see Santin-Dore, P.
and Pallesen, J. 1995, ‘Data envelopment analysis: an overview’, Queensland Economic
Review, September Quarter, pp. 34–37.
83
DEA can be applied:

• in cross-section, comparing a number of organisations at one point in time;
• as a time-series, measuring the performance of a particular organisation
over time; or
• as panel data, combining cross-section and time–series data, (that is,
comparing a number of organisations over time).
Using panel data is a good way to increase the sample size. However, if there
has been a significant change in technology over the sample period, it is
difficult to assess whether increases in productivity reflected in rising average
efficiency scores each year are a result of improvements in technical efficiency
or technological change. Expenditure data also needs to be deflated by an
appropriate price index.
5.2.2 Model specification for oral health services

For the oral health services study, annual data for the three-year period 1992-93
to 1994-95 was provided by Queensland Health on the thirteen Queensland
regions, giving a sample of 39 units.13
Readily available data on oral health services are listed in Table 5.2.1. There
were no significant changes in technology over the study period.
Table 5.2.1: Oral health services data available

Variable Units
Inputs
Dental officer days Number
Total expenditure on dental officers $
Dental therapist days Number
Total expenditure on dental therapists $
Dental assistant days Number
Total expenditure on dental assistants $
Total expenditure $
Labour related costs $
Non-labour related costs (= total expenditure less labour related costs) $
Student enrolments Number
Output
Treatments completed Number
13 Panel data were used in this study because this offered the largest possible sample size of
thirty-nine observations.
84
5 CASE STUDIES
One of the most important steps in undertaking a DEA study is choosing the
inputs and outputs to be used in the model. The inputs and outputs must relate to
the objectives of the organisation, be consistent across organisations, and be
quantifiable.
The objectives or desired outcomes of oral health services were to provide
dental treatment to as many students as required services and to undertake
preventative care.14 The only data available on oral health services for
measuring output were the number of treatments completed.15 This sufficiently
measures the first objective. The second objective of preventative care was not
assessed because it requires a measure of service quality which was not
available. The treatments completed in each region vary in complexity, time and
resources used. However, because the different types of treatments completed in
each region were not recorded, all treatments were regarded as equal by the
model.
The inputs used to provide oral health services are labour and capital.16 The
labour inputs were divided into the number of dental officer days, dental
therapist days and dental assistant days. Labour inputs were measured in
physical quantities rather than dollars because wage rates vary between regions,
and between dental officers, therapists and assistants. If salary expenditures had
been used, then differences in expenditure would reflect not only the physical
quantity of labour used to perform any given service, but also the prevailing
wage rates.
For example, if dentists in city and country regions spent one hour to treat a
patient for a filling, then the measure of each of their physical products in each
case would be equal to one. However, if cost data were used and the country
dentist’s wages were lower, then the productivity of the city dentist would
appear to be lower mistakenly.
Capital input was measured by non-labour related costs which were calculated
by subtracting labour related costs from total expenditure. Queensland Health
determined that this was the only way of measuring capital with the available
data. (Also, when using dollars over a number of years, the data needs to be
deflated using an appropriate price index.)
14 Over the study period, oral health services were provided to: preschool to year 7 in 1992-
93; preschool to year 8 in 1993-94; and preschool to year 9 in 1994-95.
15 Treatments are any dental procedures performed on patients.
16 Capital is used to refer to all non-labour related costs.
85
The number of student enrolments in each region was also included to account
for differences in potential demand.
The model was run in two formats:
• input minimisation, holding output constant and determining the minimum
level of inputs necessary to achieve that level of output; and
• output maximisation, holding inputs constant and determining the
maximum output that can be produced for that given level of inputs.
For each of these formats, the model was run with the assumption of constant
returns to scale initially (that is, output increases in equal proportion to an
increase in inputs, for example, a 10 per cent increase in inputs results in a 10
per cent increase in output). By holding returns to scale constant, it is assumed
that all regions are operating at a scale appropriate to their situation.
This assumption was then relaxed to allow for variable returns to scale (that is
an increase in inputs can result in a greater or lesser increase in output). Under
this assumption, the model’s efficiency scores are adjusted to remove
differences resulting from operating at a less efficient scale, with any remaining
inefficiency attributable to other factors.
Variations of the model specification

The model was run a number of times using different combinations of inputs
and regions with the same output. This allowed the assumptions of constant
returns to scale and variable returns to scale and their impact on the results to be
tested and compared.
In consultation with Queensland Health, it was determined that the most
appropriate variation was the one which excluded the two smallest regions
(because they were deemed too small to be directly comparable with the other
regions) and total enrolments (because this was not an input over which service
providers had direct control).
Therefore, the preferred variation of the model was specified as follows: inputs
were the number of dental officer days, dental therapist days and dental assistant
days, and non-labour related costs; and output was the number of treatments
completed. Given the exclusion of the two smallest regions, the sample size was
reduced to thirty-three units. The details below focus on this specification.
The model was then run in input and output maximisation mode with constant
returns to scale and variable returns to scale applied in each case, giving four
sets of results.
86
5 CASE STUDIES
5.2.3 Results and conclusions
Technical efficiency scores

The technical efficiency scores indicate which of the regions are deemed to be
efficient (those given a score of 100) and which are deemed to be less efficient
relative to those that are efficient (those with scores of less than 100). The lower
the score, the less efficient the region rates relative to the most efficient.
It is important to note that the scores are relative — that is, those given a score
of 100 are efficient relative to the rest of the regions in the model, but might not
be operating efficiently by some absolute standard or standards elsewhere
necessarily.
In summary, all runs showed that the vast majority of the regions in the sample
appear to be performing reasonably well. The gap between the efficient and less
efficient regions was relatively small, with most regions achieving technical
efficiency scores of higher than 80 per cent.
Overall performance of regions

Results from the input minimisation model suggested that:
• assuming constant returns to scale, nine out of thirty-three units operated
relatively efficiently, that is: region 10–9317, region 11–93, region 3–94,
region 7–94, region 10–94, region 6–95, region 7–95, region 9–95 and
region 10-95;
• assuming variable returns to scale, inefficiency in seven units could be
attributed to operating at an inappropriate scale of operation, that is, they
were either too big or too small to operate efficiently. These units were
region 4–93, region 7–93, region 9–93, region 5–94, region 9–94, region
4–95 and region 5–95; and
• the remaining seventeen units were technically less efficient.
Technical efficiency scores for the input minimising cases under constant
returns to scale and variable returns to scale are depicted graphically in
Figures 5.2.1 and 5.2.2, respectively.
17 Region 10–93 means region 10 in 1992–93. Similarly, region 10–94 means region 10 in
1993–94 and region 10–95 means region 10 in 1994–95.
87
Figure 5.2.1: Technical efficiency scores for input minimisation under

constant returns to scale
REGION11
REGION10
REGION9
REGION8
REGION7
REGION6
REGION5
REGION4
REGION3
REGION2
REGION1
50 60 70 80 90 100
1992-93 1993-94 1994-95
Under variable returns to scale (Figure 5.2.2), regions 1, 2 and 8 seemed to

consistently perform less efficiently over the three years. Regions 3, 4, 5 and 6
improved their relative performance over the period, while regions 7, 9 and 10
seemed to be efficient in all three years.
However, the performance of region 11 deteriorated from a 100 per cent ranking
in 1992-93 to the lowest ranking of 70 per cent in 1994-95. This reflected:
• a significant decline in the number of treatments provided, from over
19 000 in 1992-93 to just over 14 000 in 1994-95 (partly a result of a fall
in the number of enrolments for the region); and
88
5 CASE STUDIES
• a large increase in non-labour related costs, from around $76 000 in 1992-
93 to over $416 000 in 1994-95. There is evidence that some items of non-
labour related costs not included in 1992-93 were included in following
years, which may explain the large increase.
Figure 5.2.2: Technical efficiency scores for input minimisation under

variable returns to scale
REGION11
REGION10
REGION9
REGION8
REGION7
REGION6
REGION5
REGION4
REGION3
REGION2
REGION1
50 60 70 80 90 100
1992-93 1993-94 1994-95
Changing the orientation of the model to one of output maximisation had little
effect on the technical efficiency scores and almost no effect on the rankings of
the regions.
Peers and targets
DEA can also suggest peers and target input and output levels for each region.
Peer regions are those which have been ranked as efficient by the model, and
89
which a less efficient region may seek to use as a guide for improving its
performance. In calculating targets, the actual level of input used is compared
with the target input level calculated by the model, along with the percentage
improvement needed to achieve the target. For example, the model suggested
region 10–93 and region 7–95 as peers for region 1, which performed less
efficiently over the three years. Region 1 can look at the input and output levels
of these peers to gain insights into how it can improve its performance.
5.2.4 Conclusions
In summary, while the analysis does not test Queensland providers against an
external standard, the vast majority of the units in the sample appear to be
providing oral health services at better than 80 per cent efficiency. Alterations to
the combination of inputs used did not significantly affect the results of the
model, although the initial inclusion of the two smallest regions did result in
these being determined as more efficient than the other regions, distorting the
overall results.
For the preferred model specification, the gap between the efficient and less
efficient units was relatively small. Regions generally showed improvement
over the three–year period of the study.
Improving the performance of government service providers such as oral health
services should not be based on efficiency alone. A government service provider
might increase its efficiency by sacrificing the effectiveness of its service, so it
is important to develop effectiveness indicators as well.
90
5 CASE STUDIES
Forward — the NSW case studies

NSW Treasury and major budget sector agencies are beginning to use DEA to
help establish benchmarks to improve the efficiency of government service
provision.
NSW Treasury has completed a number of DEA studies on the technical
efficiency of NSW police patrols (local police districts), corrective services, and
motor registries. It has also completed a pilot study on the technical efficiency
of local courts in NSW. Studies have commenced to provide insights into the
technical efficiency of NSW hospitals and metropolitan fire brigades.
The DEA studies presented in this information paper were prepared by various
members of a Treasury team that is developing performance indicators for the
budget sector.18
The views expressed in these studies are the authors’ and do not reflect those of
NSW Treasury, the participating government agencies or the NSW Government
necessarily.
5.3 Technical efficiency of corrective services in NSW
5.3.1 Introduction
The number of inmates in Australia has grown steadily over recent years,
reflected in higher rates of imprisonment. Governments are ensuring that
inmates serve longer sentences by abolishing prison remission (SCRCSSP
1995). In 1988, the NSW Government increased the sentences for crimes and
abolished prison remission. Consequently, the daily average number of inmates
in prisons (correction centres) rose from 4124 in 1987-88 to 6279 in 1994-95.
The information presented in Table 5.3.1 indicates that the rate of imprisonment
in NSW increased from 101.9 inmates per 100 000 adults in 1988-89 to 135.9
inmates per 100 000 in 1994-95. Only the Northern Territory and Western
Australia have a higher rate of imprisonment. NSW Government expenditure on
18 The team has included Roger Carrington, Nara Puthucheary, Deirdre Rose and Suthathip
Yaisawarng. John Pierce, Secretary of NSW Treasury, supervised the project while he was
the Executive Director of State Economic Revenue Strategy and Policy. NSW Treasury
would like to thank the NSW Police Service, the Department of Corrective Services, and
the Roads and Traffic Authority for their assistance in preparing the studies, and Tim
Coelli (University of New England) for his useful suggestions on earlier drafts of the
studies.
91
prisons and corrective services increased from $239 million to $367 million
over 1988-89 to 1994-95 (ABS 1995a) — a real increase of 30 per cent.19
Table 5.3.1: Estimated total prisoners per 100 000 adults, 1988-89 to
1994-95
NSW 1 Vic Qld SA WA Tas NT ACT 2
1988-89 101.9 68.1 116.0 77.9 135.5 76.9 363.0 10.6
1989-90 115.0 69.8 106.6 81.5 138.9 70.1 351.3 11.1
1990-91 129.3 69.1 101.5 87.2 152.3 70.8 394.5 11.1
1991-92 134.2 66.9 94.9 97.2 155.3 76.1 397.8 9.4
1992-93 135.9 66.8 89.0 101.5 150.0 74.5 373.4 7.5
1993-94 137.9 73.9 94.6 108.7 165.1 71.9 384.6 8.6
1994-95 135.9 71.8 109.2 118.6 164.8 74.2 393.9 8.6
1 NSW figures exclude periodic detainees.
2 ACT figures are only remandees. ACT sentenced prisoners are included in NSW figures.
Source: Steering Committee for the Review of Commonwealth/State Service Provision (1995).
The NSW Government has developed several policies to reduce the cost and
improve the effectiveness of corrective services. Inmates are held in the lowest
possible level of security where appropriate, so over 50 per cent of inmates are
held in minimum security correction centres. Security posts in maximum
security correction centres are being replaced with electronic security
equipment, and alternatives to incarceration, such as community service orders,
are now available to the courts to deal with fine defaulters. A home detention
scheme commenced in 1996-97 for suitable minimum security inmates; it will
impose liberty restrictions similar to incarceration. Several personal
development programs such as vocational education and training, drug and
alcohol rehabilitation, and work release programs help inmates prepare for their
return to the community.
This paper reports the progress in measuring the technical efficiency of
corrective services in NSW using DEA (Lovell 1993). Farrell (1957) suggested
that there are two aspects to overall economic efficiency — technical efficiency
and allocative efficiency. Technical efficiency describes the physical
relationship between inputs and outputs; it reflects the ability of an organisation
to produce maximum outputs given inputs, or to produce certain outputs with
least inputs. Allocative efficiency, on the other hand, measures the optimal mix
of inputs (or outputs) for an organisation given observed prices.
19 The implicit price deflator for Government final consumption expenditure in New South
Wales in 1989-90 (ABS 1995b) is used to derive this figure.
92
5 CASE STUDIES
The study does not consider the effectiveness of correction centres to reduce
recidivism. Recidivism not only depends on the attitudes and skills acquired by
inmates while in the correctional system (which may be of limited influence for
inmates serving short sentences), but is also influenced by factors outside the
control of the corrective system such as family and community support (for
example, access to public housing and other social services, employment, and
vocational education and training).
Section 5.3.2 explains the production of corrective services and Section 5.3.3
presents the initial results of the technical efficiency of corrective services.
Section 5.3.4 contains conclusions about the technical efficiency of corrective
services, and outlines further initiatives to improve the analysis.
5.3.2 The production of corrective services

The Department of Corrective Services carries out the orders of the courts by
incarcerating inmates until they are lawfully discharged. It aims to manage
inmates in a humane, safe and secure environment, and provide personal
development programs for inmates that focus on the causes of their crime and
help them return to the community as law-abiding and productive citizens.
These broad objectives of corrective services provide a focus to specify the
outputs of corrective services.
In NSW, correction centres are classified as maximum, medium or minimum
security. As inmates enter the corrective system, they are classified to the lowest
appropriate security rating which is consistent with the Department’s
responsibility to protect the community. This study focuses on minimum
security correction centres, which account for the majority of corrective
facilities in NSW, and hold most of the state’s inmates.
Other categories of correction centres are not included in the study because they
use different technology and resources to manage inmates. For example,
maximum security correction centres have complex electronic surveillance
systems and some of the older facilities have watch towers staffed by armed
guards. Moreover, a ready-armed response is available to deal with serious
incidents, such as escapes by inmates. Medium security correction centres have
a ready-armed response only during the day. In both maximum security and
medium security correction centres, inmates are held in secure cells and
enclosed in a secure perimeter to prevent escape. By contrast, inmates in
minimum security correction centres face limited physical barriers and may
require little supervision by custodial officers.
93
One measure of confined inmates in a minimum security correction centre is the

average daily inmate population. However, such a centre can hold several
categories of inmates who require different inputs for their incarceration. The
various categories are:
• inmates who are held in secure cells overnight but face limited physical
barriers during the day;
• inmates who are not restrained by physical barriers but are supervised by
staff when entering the community. Their cells are typically not secured,
although accommodation blocks are locked overnight; and
• inmates who are allowed to enter the community without supervision and
whose accommodation is less secured compared with the accommodation
for the other categories of inmates.
The last category is cheaper to manage than the first category of inmates for
several reasons.20 They have conditional leave of absence to participate in work
release, education or weekend release programs, allowing them to enter the
community without custodial supervision. Inmates participating in work release
programs pay board which further reduces the cost of their incarceration.
Moreover, these inmates are held in less secured accommodation overnight,
which costs less to construct than does secured accommodation. To account for
this, we include the average daily number of inmates that are eligible for
conditional leave of absence without custodial supervision and the average daily
number of other inmates in the analysis.
A third measure of confinement — the number of inmates received by a
correction centre — is considered in the analysis because it affects the operating
costs of the correction centre. New receptions and discharges require additional
resources to provide reception and induction programs, administrative
processing and pre-release programs. Moreover, new inmates, even if
transferred from other centres, require additional supervision and support while
adapting to new circumstances.
The Department of Corrective Services provides a range of personal
development programs, such as vocational education and training and drug and
alcohol rehabilitation, to help prepare inmates for their return to the community.
The time that inmates spend in personal development programs measures the
quantity of personal development training that inmates receive; it does not
reflect how well these programs prepare the inmates for their return to the
20 There are some differences in the cost of managing the second category of inmates and the
other categories, but not as large as the difference in the cost of managing secured inmates
and inmates eligible for conditional leave of absence.
94
5 CASE STUDIES
community. Information is available only on the time that inmates spend in

education programs or industrial activities, such as saw milling and farming,
which provide them with additional vocational skills and some income. To
reduce the number of variables in the analysis, the time that inmates spend in
these activities is combined to measure the personal development programs
delivered by the Department of Corrective Services. The motivation for
reducing the number of variables in the analysis is explained in Section 5.3.3.
A minimum security correction centre uses a variety of inputs to manage
inmates, including staff, capital, food, clothing, and other consumer goods and
services, such as power and water. The operation of a correction centre requires
several types of labour such as custodial officers, industrial officers (who teach
vocational skills), medical and other professional officers, and maintenance and
clerical staff. Custodial officers and industrial officers are the largest group of
employees. They oversee security in the correction centre and deliver the main
development program, inmate employment. Therefore, for the purposes of this
study, custodial and industrial officers are classified as custodial officers and are
measured by full-time equivalent staff numbers. Medical staff are excluded
from the study because the health budget for correction centres is provided by
the NSW Department of Health. There is no information on the other labour
used in minimum security correction centres.
Capital is measured by the design capacity for each correction centre, which is
the number of inmates that a correction centre can hold. This is usually
determined by the number of beds in a correction centre. Other recurrent
expenditure (that is, less wages and salaries) measures inputs such as food,
clothing and other consumer goods and services.
In summary, the outputs included in this study on corrective services were the
average daily number of inmates eligible for conditional leave of absence, the
average daily number of other inmates, the number of inmate receptions, and the
number of hours that inmates spend in personal development programs. The
inputs included the number of beds in a correction centre, full-time equivalent
custodial staff numbers, and other recurrent expenditure.
The study has similar features to a DEA study on correctional institutions in
New York State (Yaisawarng and Vogel 1992). That study developed several
models of corrective services that have comparable inputs to those used in this
study. However, it assumed that some inputs, like recurrent expenses, were
adjustable while holding others, such as capital, constant. Further, the study
focused on confinement as the output of correctional institutions. Yaisawarng
and Vogel used either a single measure (the average daily inmate population) or
95
multiple measures to reflect the different inputs required to manage violent

inmates, less violent inmates, and people awaiting sentence or arraignment.
Ganley and Cubbin (1992) considered outputs other than confinement in their
study of correctional institutions in the United Kingdom, but did not consider
the personal development of inmates. In addition to outputs of confinement (that
distinguish between sentenced and unsentenced inmates), they included a
quality indicator for confinement (the degree of overcrowding), and attempted
to measure the quality of the supervision provided by correctional officers by
the number of serious offences committed by inmates. Ganley and Cubbin used
labour costs and non-labour costs to measure inputs: they considered these costs
could be adjusted.
5.3.3 Technical efficiency of corrective services

There are seventeen medium security and minimum security correction centres
in NSW that incarcerate minimum security inmates. Six correction centres were
excluded from the study. They included the correction centres which hold both
medium security and minimum security inmates (because they use substantially
different resources to manage the medium security inmates) and the minimum
security correction facilities attached to medium security or maximum security
centres (because it is not possible to isolate their inputs).
The remaining eleven minimum security correction centres were a relatively
small sample to generate sensible DEA results if data for only one year were
used in the analysis. Therefore, panel data for 1990-91 to 1994-95 were used to
increase the sample. This was possible because the Department of Corrective
Services argued that there was minimal change in the technology (which
included inmate management) used in minimum security correction centres over
this period. Most correction centres were evaluated five times, that is, once for
each year. However, two correction centres were converted from medium
security in 1993-94, so they were evaluated only for each year that they were a
minimum security centre.
There is incomplete information for some correction centres on the time that
inmates participate in education programs in a particular year. To overcome this
problem, the proportion of total inmate hours to the number of months for
which information is available was assumed to prevail over the year. An implicit
price deflator for NSW Government consumption (ABS 1996) was used to
convert the nominal other recurrent expenditure into real other recurrent
expenditure.
96
5 CASE STUDIES
DEA has several limitations. It is sensitive to outliers, and the omission of

important services or inputs produces biased results. Further, DEA includes
random occurrences in its measures of efficiency — for example, errors in the
data. To reduce the risk of data errors influencing the results of the study, the
data were screened for potential outliers using descriptive statistics for the
services (outputs) and the inputs and also by identifying service–input ratios that
were further than two-and-a-half standard deviations from their sample means.
Potential outliers that were identified this way were referred to the Department
of Corrective Services for comment. The Department subsequently confirmed
that these observations were not a result of errors in the data. Table 5.3.2
presents descriptive statistics for each service and input variable in the sample.
Table 5.3.2: Descriptive statistics for minimum security correction

centres[a]
Standard
Mean deviation Minimum Maximum
Outputs
Inmates eligible for conditional leave 47.74 6.57 7 187
of absence (no.)
Other inmates (no.) 134.60 12.09 28 408
Receptions (no.) 704.32 90.09 111 2 386
Personal development programs 120 565 10 009.66 31 657.10 338 014

(hours)
Inputs
Beds (no.) 181.55 17.20 64 453
Custodial staff (full time equivalent 65.67 5.30 30 150.50

no.)
Real other expenditure ($’000) 1 003.77 66.51 290.60 2 261.37
[a] Forty-seven observations.
The Department of Corrective services has little control over the number of
inmates that it manages. It must incarcerate and help rehabilitate inmates with
least inputs. Therefore, an input-oriented DEA model is used to determine the
technical efficiency of minimum security correction centres. A similar approach
is used by Yaisawarng and Vogel (1992) and Ganley and Cubbin (1992). Details
on the method to calculate the technical efficiency of correction centres is
presented in Appendix A and Lovell (1993).
97
Table 5.3.3: Summary statistics: efficiency of minimum security

correction centres
Technical Pure technical Scale
efficiency efficiency efficiency
No. of observations 47 47 47
Mean 0.930 0.964 0.963
Standard deviation 0.018 0.012 0.009
Minimum value 0.593 0.666 0.757
Number of efficient
correction centres 20 29 20
The results presented in Table 5.3.3 suggest that pure technical efficiency
(managerial efficiency) and scale efficiency were equal sources of lower
efficiency for correction centres.
Correction centres may be able to produce, on average, the same outputs with
approximately 4 per cent fewer inputs. However, this result needs to be
interpreted with care. Managerial efficiency is influenced by the sample and the
number of outputs and inputs included in the study. If DEA cannot compare a
correction centre with other correction centres, then it is deemed efficient by
default, which tends to increase the average managerial efficiency score. This
study had a relatively high proportion of correction centres that were apparently
efficient by default (about 20 per cent of the managerial efficient correction
centres) because it had a relatively small sample compared with the number of
outputs and inputs used in the analysis. To overcome this problem, future
analysis could include correction centres from other states to increase the
sample, or alternatively the number of outputs or inputs used in the analysis
could be reduced.
About 13 per cent of correction centres appeared to require larger reductions in
inputs, compared with the average reduction in inputs, to become managerially
efficient. The least efficient correction centre would appear to have to reduce its
inputs by around 33 per cent. This correction centre was converted from a male
facility to a female facility in 1994-95. Inmate numbers declined by about 40 per
cent without a commensurate reduction in inputs. Before its conversion to a
female facility, the centre appeared to be managerially efficient in 1990-91 and
1991-92, was above average managerial efficiency in 1992-93, and was
marginally below average managerial efficiency in 1993-94.
The average apparent scale efficiency of correction centres was 96 per cent,
which suggests they might be able to reduce inputs by a further 4 per cent to
achieve optimal scale. The information presented in Table 5.3.3 suggests that
about 43 per cent of the correction centres were scale efficient. About 60 per
98
5 CASE STUDIES
cent of the scale inefficient correction centres exhibited increasing returns to

scale. The Department of Corrective Services might wish to examine the scope
to combine these centres into larger centres that possess optimal scale. The
remaining correction centres exhibited decreasing returns to scale and would
appear to require a reduction in size to achieve optimal scale.
The managerial efficiency of a minimum security correction centre is overstated
if, after an equi-proportionate decrease in all inputs, some excess inputs (or
slacks) are still present. About 36 per cent of the correction centres had at least
one such input. The information presented in Table 5.3.4 suggests that
correction centres with particular excess inputs, on average, may be able to
reduce the inputs further by 10.63 beds, 2.33 full-time equivalent staff and
$145 620 in real other recurrent expenditure, and produce the same outputs.
However, the excessive use of inputs was relatively minor compared with total
inputs.
Table 5.3.4: Input slacks in correction centres

Number of Total slacks as a
correction centres Mean percentage of total
with slacks inputs
Beds (no.) 7 10.63 1.00
Custodial staff (full-time
equivalent) 5 2.33 less than 1.00
Real other expenditure
($’000) 13 145.62 4.01
5.3.4 Conclusion
The analysis suggests that minimum security correction centres, on average,
might be able to produce the same outputs with 4 per cent less inputs. Moreover,
if all correction centres were of optimal scale, they might be able to reduce
inputs by a further 4 per cent. However, operational imperatives relating to
centre location requirements and meeting the needs of specific offender groups
need to be taken into consideration.
Care is required in interpreting the results because a relatively high number of
correction centres were apparently efficient by default, which contributed to the
high mean managerial efficiency of the correction centres. Further, there were
deficiencies in some data, especially the information for inmate personal
development.
The technical efficiency scores for individual correction centres, the associated
information on peers, and the effectiveness indicators for corrective services
99
developed by the Steering Committee for the Review of Commonwealth/State

Service Provision (1995) provided the Department of Corrective Services with
an objective framework to judge the extent to which, and speed with which, it
might be able to improve the technical efficiency of minimum security
correction centres. The Department discovered inefficiencies in some correction
centres that it was unaware of before the DEA exercise. It will review these
centres to discover the cause of the inefficiencies.
NSW Treasury will seek to include minimum security correction centres in
other states in future analysis of the technical efficiency of corrective services in
NSW. Increasing the sample might reduce the number of correction centres that
are apparently efficient by default. Moreover, a larger sample might improve the
benchmarks and targets for the efficient provision of corrective services. An
alternative to setting better benchmarks and targets for corrective services is to
seek further guidance from the Department of Corrective Services to reduce the
number of outputs and inputs used in the analysis. However, this approach risks
excluding important variables in the provision of corrective services, which
would produce biased results.
Further work is required to improve some of the information used in this
exercise. The Department of Corrective Services is aware of the deficiencies of
its statistics on the time inmates spend in personal development programs and is
examining ways to improve this information.
5.4 Performance measurement of police services in NSW
5.4.1 Introduction 21
The purpose of this paper is to examine the technical efficiency of the Police
Service in 1994-95, using a two-stage procedure. In the first stage, DEA is used
to compute technical efficiency for all police patrols.22 In the second stage,
Tobit regression is used to analyse external factors or operating environments
which might explain the variation in apparent technical efficiencies across
21 This is an edited version of Roger Carrington, Nara Puthucheary, Deirdre Rose and
Suthathip Yaisawarng 1997, ‘Performance Measurement in Government Service Provision
– the Case of Police Services in NSW’, Journal of Productivity Analysis (forthcoming).
22 A police patrol is an organisation unit which is responsible for providing services to a
designated area within the community. The designated area is also referred to as the
“patrol”.
100
5 CASE STUDIES
police patrols. The results of this study can be used to assist the NSW Police
Service in delivering better and efficient services to the community.
The motivation of this study originates from the introduction of comprehensive
financial reforms by the NSW Government to help ensure government service
provision, such as health, law and order and education, is efficient and effective.
The main reform is contractual (or performance) budgeting, an approach
whereby the Government enters into agreements with government service
providers to purchase services that assist in achieving government policy
objectives, rather than funding services according to historical expenditure
patterns. Performance measurement is necessary to complement contractual
budgeting to provide an incentive for government service providers to become
more effective and efficient. According to Pierce (1995), performance
indicators provide information that makes government service providers more
accountable to Parliament. They also promote yardstick competition in the
provision of government services that face little competition, acting as a
powerful internal management tool to examine reasons for poor performance.
Section 5.4.2 discusses the operation of the NSW police patrols, Section 5.4.3
presents the empirical results, and Section 5.4.4 summarises the findings and
demonstrates the use of DEA to improve the performance of a major
government service provider.
5.4.2 Assessing the performance of the NSW police service

Law and order is a high priority of the NSW Government. It has sought to allay
community concern over public safety by employing more police officers. The
Government announced in the 1995-96 Budget that the number of police
officers available for duty would increase by 650 over the following four years
to 13 407.
Community perceptions on public safety are influenced by social, economic and
institutional factors that are beyond the control of the NSW police service.
Nevertheless, public safety is an outcome that the police service seeks to
influence by providing several services to the community. Under the auspices of
the Steering Committee, and in conjunction with other Australian state and
federal police services, it has developed several objectives for its services:
• to protect, help and reassure the community;
• to prevent crime; and
• to enforce the law.
101
Effectiveness indicators for the objective of protecting, helping and reassuring

the public include the number of hospitalisation and fatal road crashes, the
percentage of vehicles stolen which were recovered in the same year, and the
number of complaints about police behaviour. Crime rates reflect the police
service’s effectiveness in preventing crime. However, crime rates need to be
interpreted with care because they are influenced by factors beyond the control
of police. Success in bringing offenders to justice and the number of deaths in
custody are examples of effectiveness indicators of law enforcement developed
by the Steering Committee. The Steering Committee published information on
the number of deaths in police custody only because there is a lack of
comparable national information for the other indicators.
The Steering Committee developed a limited number of efficiency indicators for
police services — for example, the average total police vehicle cost per
kilometre. The unit cost of police vehicles in NSW declined from 35 cents per
kilometre in 1992-93 to 31.5 cents per kilometre in 1994-95. However, the
Steering Committee argued that there are several difficulties in developing a
comprehensive suite of efficiency indicators for police services based on unit
cost. Outputs for some police activities, especially crime prevention, are
difficult to define, and police can deliver several services simultaneously, so it
is difficult to isolate the inputs for each service, (for example, the public
presence of police arresting an offender also reassures the community and
prevents crime).
Effectiveness and efficiency indicators are important tools that can assist in
improving the performance of police services, but this paper focuses on the
efficiency of police services only. Efficiency scores were calculated for each
police patrol using DEA. This technique provides a single measure of efficiency
for each patrol. This efficiency score reflects the success of a patrol in
producing a variety of services using a number of inputs. To a large extent, this
overcomes the problem of allocating inputs to specific services (which is
required for unit cost analysis). DEA identified the apparently best patrols by
their ability to produce the greatest number of services with a given set of
inputs, or to produce certain services with least inputs. Patrols received an
efficiency score that was determined by their performance relative to the best
patrols. The information on the set of best patrols (peers) that are compatible
with a specific less efficient patrol is useful for the less efficient patrol to
identify ways in which it might improve its efficiency.
It is important, although difficult, to specify completely and correctly the
activities of units analysed. The omission of important variables produces
biased results. A knowledge of police duties provides an insight into the
102
5 CASE STUDIES
services that a police patrol provides the community, and focuses attention on
the development of an ideal model for measuring police efficiency. In the
following two sub-sections, the inputs and outputs that should appear in a DEA
model of the NSW police service are listed, and the available data that are the
variables selected for use in the empirical analysis are discussed.
Production of police services

The NSW police service organises its command to deliver its services into four
geographical regions of NSW that include twenty-five districts and 165 patrols,
special agencies (such as the Drug Enforcement Agency) and regional support
units (such as the Forensic Services Group). Most police services are delivered
by police patrols.
A police patrol typically consists of a patrol commander, detectives, general
duty police (which include beat police) and civilian employees. The patrol
commander is responsible for the operations of a patrol. The commander
provides leadership, develops the strategic plan for the patrol, and investigates
complaints about police behaviour. General duty police conduct random breath
testing of motor vehicle drivers, attend major car accidents, arrest people, issue
traffic infringement notices, maintain lost property exhibits, handle missing
persons inquiries, secure prisoners in custody, and respond to calls for
assistance. Beat police gather intelligence on criminal activities, arrest people,
control crowds at sporting events and demonstrations, obtain warrants and serve
summons, maintain a public presence to prevent crime and reassure the public
of their safety, and conduct follow-up interviews of victims of crime. Detectives
investigate more serious criminal matters, arrest people, and attend to coronial
matters. Civilians provide clerical support that includes answering telephone
calls and sending radio messages. Highway patrols use the infrastructure of
patrols — police stations, for example — in their endeavours to reduce the
number and severity of road accidents. However, highway patrols are often a
separate police unit to police patrols, so they are excluded from the analysis.
The NSW police service classifies the services of a police patrol into two broad
categories: reactive policing, which covers law enforcement activities; and
proactive policing, which covers activities that protect, help and reassure the
public and prevent crime. Random breath testing and patrolling crime spots are
examples of proactive policing.
If there were no data limitations, the DEA model of police patrols would cover
all the outputs that a patrol produces to deliver services to the public. Given that
there are several service activities in each category, a case mix index which
reflects the quality of services would be created. Aggregate output indexes for
103
reactive policing and proactive policing, which use appropriate case mix
indexes as weights, would represent ideal output measures in the DEA model.
The specification of inputs is much more clear cut. Police patrols use several
main inputs to provide services to the community: police officers,23 civilian
employees and capital equipment, such as police cars and computers.
The sample
This paper uses a sample of 163 police patrols in 1994-95.24 The NSW police
service has comprehensive data on outputs for reactive policing, but little
information on proactive policing (which accounts for about 40 per cent of
patrol work). It also has no information on the outputs of civilian employees.
Further, the NSW police service does not have reliable information on the time
that police spend on their activities. After discussions with the NSW police
service, the major measured outputs of police patrols were included in the DEA
study.
The reactive policing outputs of a patrol are responses to incidents, arrests,
serving of summons and attendances at major car accidents. These were
measured by the number of cases25 and were included as output variables in the
DEA study. However, there are several caveats associated with data for arrests
credited to a patrol and a patrol’s response to incidents.
The information on the arrests performed by a patrol includes arrests made by
other NSW police agencies such as special operations groups and the Drug
Enforcement Agency, and the Australian Federal Police. Moreover, arrests by
23 The NSW police service cannot isolate the outputs and inputs for detectives and general
duty police (which include beat police). It suggested that they can be combined into one
category: namely, police officers.
24 One Sydney patrol is excluded from the analysis because it is the central jail for offenders
waiting to appear before a court. The NSW police service also decided to include a police
water patrol into a nearby (regular) patrol for the purposes of this study.
25 Previous studies (for example, Darrough and Heineke 1979; Gyimah-Brempong 1987,
Gyapong and Gyimah-Brempong 1988, Levitt and Joyce 1987) used arrests or clearances
as proxies for police outputs. This is only one aspect of the reactive policing of the NSW
police service. Further, the police service rejected clearances as a meaningful measure of
output of police patrols. A crime is cleared when police have sufficient evidence to lay a
charge against those who committed the crime; they are not required to make an arrest or
serve a summons to clear the crime. Moreover, it is possible for police patrols to increase
their clearances merely by recording additional charges against the person or people that
committed the crime. If clearances are used as a measure of output, then there is a risk that
police patrols will focus on crimes that are easy to clear and ignore serious crimes that are
harder to solve.
104
5 CASE STUDIES
these police agencies to solve criminal activities in several patrols are credited
to each patrol. The NSW police service cannot separate these arrests from the
arrests affected by a patrol alone. Consequently, the outputs of some patrols are
artificially inflated by these arrests. Wrong arrests also overstate the outputs of a
patrol.
The NSW police service’s information on incidents is limited to recorded
offences. This information is likely to understate a patrol’s response to incidents
because it excludes those incidents for which police either decide that no further
action is required or issue a warning to the offenders.
Output variables for proactive policing were difficult to obtain. The total
kilometres travelled by police cars captures some aspects of proactive policing.
It reflects a police presence in the community to reassure the public, and a
visible police car prevents crime. However, it ignores the proactive policing that
beat police do on foot in metropolitan patrols. Information is not available for
alternate measures of proactive policing such as the number of random breath
tests conducted by a patrol or the number of criminal intelligence reports filed
by beat police. Darrough and Heineke (1979), Gyimah-Brempong (1987), and
Gyapong and Gyimah-Brempong (1988) assumed that non-crime activities
(such as traffic control and emergency first aid care) are related to the size of a
community, and used the population of the community to measure these
services. The NSW police service argued that the official population in a patrol
does not accurately reflect the proactive policing provided by a patrol because
the population of a patrol can swell considerably as people enter its jurisdiction
for work or entertainment. Moreover, even if accurate figures on a population
were available, it still must be unrealistically assumed that each police patrol
provides a similar proportion of proactive policing relative to the other services
it provides the community.
Similar to most existing studies, two types of labour input were used: number of
police officers and the number of civilian employees as of 30 June 1995.26 The
number of employees assigned to a patrol included people on extended leave
(for example, sick leave, long-service leave or seconded leave to other police
units). Therefore, care is required when interpreting the results. A patrol may
appear relatively less efficient because it had a higher proportion of its
personnel on extended leave. Further, a patrol that consistently overworked its
staff might appear more efficient compared with a similar patrol for which staff
26 Actual number of hours worked is more desirable but this information was not available.
105
worked normal hours. Capital input was measured by the number of police
cars.27
In summary, the DEA model of NSW police patrols included five output
variables: the number of arrests, responses to offences recorded, serving of
summons, attendances at major car accidents, and the kilometres travelled by
police cars. The three inputs used were: the number of police officers, the
number of civilian employees and the number of police cars. The sample
included 163 police patrols for 1994-95. Table 5.4.1 presents descriptive
statistics for each output and input in the sample.28
Although the efficiency scores obtained from solving DEA represent the ability
of management to convert inputs into outputs at the current scale of operation, it
is possible that some other external environmental factors beyond the control of
the management may affect their measured efficiency. The study looked to
determine which external factors had some influence upon variations in pure
technical efficiency across police patrols and in which direction. A second-stage
analysis was used to explain the variation in DEA technical efficiency scores
from the first stage.29 This used the Tobit procedure to estimate the relationship
between pure technical efficiency scores and operating environmental factors
unrelated to the inputs used in the DEA model.30
Specifically, the following model was estimated:
27 The number of computers or other equipment installed in patrols or the floor space of
buildings occupied by a patrol could be included in the measure of capital if the
information was available.
28 DEA is susceptible to outliers, so output–input ratios were computed for each patrol, and
the value that exceeded two-and-a-half standard deviations from the sample mean was
considered a potential outlier. Potential outliers were referred to the NSW police service
who checked the data and confirmed there were no obvious measurement or reporting
errors. Burgess and Wilson (1993) discussed the nature of outliers and their impact on
DEA efficiency scores. Wilson (1995) suggested a way to detect influential observations,
which is a computer–based technique and is appropriate when it is not possible to access
the first-hand data or too costly to check all data points. This was not the case in this study.
29 The method differed from that used by Levitt and Joyce (1987) who directly included
environmental variables in their DEA study of UK police authorities. Their method
required an assumption on how each environmental variable affected efficiency. This
assumption precluded the test of its impact. McCarty and Yaisawarng (1993) discussed
advantages and disadvantages of these two approaches.
30 Given that the pure technical efficiency scores are censored from above at one, the
ordinary least squares regression produces biased and inconsistent results (Judge et al.,
1988).
106
5 CASE STUDIES
TEVRS = α + Zβ + u
where TEVRS is a (J x 1) vector of pure technical efficiency scores for all J
patrols, the scalar α and the (R ∗ 1) vector β are unknown parameters to be
estimated. Z is a (J ∗ R) matrix of environmental factors and u is a (J ∗ 1) vector
of residuals.
The NSW police service suggested several environmental variables, or
noncontrollable inputs, that might affect the measured efficiency of a patrol but
that are beyond the control of management.31 First, police observe that most
offenders are young people aged 15 to 29 years. A patrol with a higher
proportion of young people in its jurisdiction is likely to respond to more
incidents compared with a similar patrol with a lower proportion of young
people in its jurisdiction. Second, a patrol with a high proportion of public
housing in its jurisdiction is likely to respond to more incidents than a similar
patrol in an area with a lower proportion of public housing. Finally, country
patrols usually cover a larger area than metropolitan patrols. They require
additional cars and staff (above the level of resources justified for the services
they provide to the community) to permit the NSW police service to provide an
effective police presence in country areas which is comparable to the service it
provides in metropolitan areas.
The proportion of young people in a patrol area and the proportion of
government housing in a patrol area were derived from 1991 census data. A
dummy variable was used to specify the location of a patrol, where a value of
zero indicated a patrol was located in a metropolitan area and a value of one
indicated a patrol was located in the country.
Patrols with a higher proportion of young people or a higher proportion of
public housing in their area, or both, were expected to appear more efficient
than similar patrols facing lower proportions of these socioeconomic conditions
because they were relatively busy responding to more crime (that is, they had
less idle time). Some of their additional work might not be reflected in measured
outputs because some incidents they investigated warranted a warning to
offenders only. Nevertheless, police patrols with a higher proportion of these
environmental variables were expected to have higher measured outputs.
31 Gyimah-Brempong (1989) used a sample of police departments in Florida, United States in

1982 and 1983 to study the impacts of environmental variables on the production of public
safety. The author found that a higher proportion of non-whites in a police jurisdiction
raised the cost of supplying police services to achieve a certain level of public safety,
which is measured as the inverse of the crime rate of the community. A greater proportion
of comparatively highly educated people in a locality reduced the cost of supplying police
services to achieve a certain level of public safety in a locality.
107
Country patrols were expected to appear relatively less efficient compared with
metropolitan patrols because they required more inputs to provide an effective
service.
5.4.3 Performance of the NSW police patrols

An input orientated DEA model was chosen to calculate the overall technical
efficiency scores for police patrols because the objective of the NSW police
service is to provide effective policing with least inputs to the community.
However, caution is required in interpreting the results. An apparently less
efficient patrol was not necessarily an ineffective patrol. Given the DEA
specification, a patrol with relatively low measured outputs might have reduced
crime through better policing. However, relatively low measured outputs could
have resulted from poor policing (that is, police did not pay proper attention to
crime or did not report their response to crime). Alternatively, poor policing
might have caused higher measured outputs — for instance, a police patrol
could have increased its measured law enforcement outputs by ignoring or
paying little attention to police procedures to solve crime, thus, encouraging
crime which further increased the measured outputs of law enforcement. With
these caveats in mind, summary statistics of the various measures of input-
orientated technical efficiency are presented in Table 5.4.2.32
First, considering the pure technical efficiency, police patrols, on average, may
be able to produce the same level of measured outputs with 13.5 per cent fewer
inputs, holding the current input ratios constant. Using a Z-test, the null
hypothesis that the sample mean was one at the 5 per cent level of significance
was rejected. However, about one-third of the patrols appeared to need to
reduce their inputs by less than the sample average if they were to become
efficient. Moreover, approximately 35 per cent of patrols were apparently
radially efficient.
Disaggregating the results by location reveals that country patrols, on average,
appeared to be more efficient than metropolitan counterparts. This was contrary
to expectations, and may be partly because the kilometres that beat police did on
foot in metropolitan patrols were excluded from the measure for proactive
policing. Using a Kruskal-Wallis test, it was found that the null hypothesis that
there was no significant difference in the pure technical efficiency for the
32 The software DEAP developed by Coelli (1995) is used to calculate the various measures
of technical efficiency.
108
5 CASE STUDIES
metropolitan and country patrols at the 5 per cent level of significance could not
be rejected.33
The mean pure technical efficiency score overstated the efficiency of police
patrols, if there appears excessive use of some inputs, beyond that reflected by
the efficiency scores. Table 5.4.3 presents a summary of the excess in each input
after radial technical inefficiency is removed. This reveals the scope for further
non-radial reductions in inputs once a police patrol operates on the production
frontier. Holding the level of police services constant, on average, it would
appear that police patrols could reduce their use of police cars by 1.96 cars;
sixty-two patrols may be able to reduce the number of civilian employees by
2.24 persons; and eight patrols might be able to reduce number of police
officers by 11.32 officers. Their excessive use of inputs accounted for 1 to 13
per cent of total inputs. The apparent excessive use of civilian employees by
almost 40 per cent of patrols in the sample may be because civilian output was
excluded from the specification of the DEA model. Some of the excess inputs
may have been converted into other outputs provided by police patrols which
were not measured in this study.
It is shown in Table 5.4.2 that average scale efficiency of 0.94 suggested further
potential input savings of 6 per cent if a police patrol could operate at the
constant returns to scale technology. Investigating the distribution of scale in
Table 5.4.4 reveals that 18 per cent of patrols appeared to already operate at the
appropriate scale. Approximately half of the patrols in the sample appear to be
experiencing decreasing returns to scale, and could be reduced in size. On the
other hand, about one-third of the patrols seemed to be experiencing increasing
returns to scale, and these may be able to be consolidated with other small units
to achieve the optimal size. The comparison of the metropolitan and country
police patrols gives a slightly different picture: the apparently scale inefficient
metropolitan police patrols were roughly split between increasing and
decreasing returns, while there were twice as many apparently scale inefficient
country police patrols operating on the decreasing returns region relative to
those on the increasing returns range. However, an across-the-board downsizing
of larger patrols may not be justified, and it may be more appropriate to consider
the patrols on a case-by-case basis before any restructuring policy is
implemented.
To determine whether environmental factors might affect the measured
efficiency of police patrols, the pure technical efficiency scores were regressed
33 The Kruskal-Wallis statistic (adjusted for ties) had a value of 2.83 for overall technical
efficiency, 3.24 for pure technical efficiency and 0.73 for scale efficiency. The associated
p-values for these statistics were 0.093, 0.072 and 0.394.
109
against the proportion of young people in a patrol, the proportion of government

housing in a patrol, and the location of a patrol. The Tobit results in Table 5.4.5
explains about 6 per cent of the variation in pure technical efficiency scores, and
none of the coefficients of the explanatory variables were significant at the 5 per
cent level of significance. The insignificance of location confirmed the prior
finding based on the Kruskal-Wallis test. Consequently, it was concluded that
the measured efficiency of police patrols was not influenced by these
environmental variables.
5.4.4 Conclusions
The NSW Government is implementing contractual budgeting for government
service providers to encourage the delivery of efficient and effective services.
Performance measurement of government service providers is necessary to
allow the community to reap the full rewards of contractual budgeting. It
encourages providers to improve their efficiency and effectiveness because this
information makes them more accountable to Parliament, and it promotes
yardstick competition in the provision of government services that face little
competition.
NSW Treasury is using DEA to measure the efficiency of major government
service providers. Furthermore, NSW Treasury is encouraging these agencies to
acquire knowledge of the DEA so as to be able to maintain the DEA models
developed by Treasury and to use the results of the studies in their corporate
planning and internal resource allocation.
This study suggested that NSW police patrols, on average, might be able to
produce the same measured outputs with 13.5 per cent less inputs at the current
scale and using their inputs efficiently. However, the average reduction masked
the fact that the apparent reduction in inputs for about one-third of the patrols to
become technically efficient would be less than the average reduction, and that
35 per cent of the patrols already appeared to be technically efficient. No
significant difference was found in the technical efficiency of country patrols
and metropolitan patrols. The technical efficiency scores for some patrols may
have overstated their technical efficiency because they had excess inputs beyond
that reflected in their efficiency scores, which accounted for 1 per cent to 13 per
cent of total inputs. Care is required in interpreting these results because it is
unknown how the quality of police work influences the measured outputs of a
police patrol. Nevertheless, the results provided indicative information on the
technical efficiency of NSW police patrols. Scale inefficiency was not a major
source of input reduction. However, if it is possible to restructure the police
patrols, the potential input savings, on average, may be 6 per cent. The
110
5 CASE STUDIES
restructure of the police patrols should be implemented only after individual

police patrols’ responsible boundaries are carefully investigated. Some police
patrols may be able to benefit from downsizing; and others from expansion.
The DEA results and the effectiveness indicators developed by the Steering
Committee provide the NSW police service with an objective input to
development of policies to improve the performance of police patrols. To
ascertain the reasons for the apparent divergence in the technical efficiency of
police patrols, NSW Treasury has provided the NSW police service with details
on the measured technical efficiency of individual patrols and information on
peers for each patrol which appears to be less efficient. The effectiveness
indicators provide further information on the performance of police patrols.
Given this information, the NSW police service could commence a series of
operations audits to determine if the apparent differences in technical efficiency
of police patrols is a result of poor policing, proactive policing or thorough
policing by some patrols. This will allow the NSW police service to judge the
extent to which, and speed with which, it may be able to improve the technical
efficiency of its patrols.
Further work is required to improve the measures of some of the outputs and
inputs used in this study. The measure of proactive policing (the kilometres
travelled by a police car) did not capture all the aspects of this form of policing
and probably disadvantaged metropolitan patrols because it ignores beat duty
done on foot. The study’s measures for labour did not allow for differences in
the quality of labour, but this problem probably cannot be resolved quickly.
Still, information on the full-time equivalent hours of police officers and
civilian employees would provide a better indication of the labour input used by
a police patrol compared with the number of employees in a patrol. The NSW
police service recently introduced a financial system that indicates to which
patrol an employee is attached for every pay of the year. This should permit the
NSW police service to develop information on the full-time equivalent hours
worked by police officers and civilian employees in police patrols. However, the
financial system cannot provide details on the computers and other equipment
installed in police patrols, so an improvement in the measure of capital (the
number of police cars) is not imminent.
111
Table 5.4.1: Descriptive statistics for NSW police patrols [a]

Standard
Mean deviation Minimum Maximum
Outputs
Offences 3670.31 2345.08 360 12 395
Arrests 938.70 625.53 145 3 215
Summons 101.76 104.14 6 596
Major car accidents 450.05 284.05 31 1 663
Kilometres travelled by police cars 422 265 233 598.10 127 146 1268 555
Inputs
Police officers 50.92 26.17 9 127
Civilian employees 6.57 5.99 0 41
Police cars 10.37 5.43 3 34
[a] 163 observations
Table 5.4.2: Summary statistics: efficiency of metropolitan and country

police patrols
Efficiency measures
Technical Pure technical
efficiency[a] efficiency[b] Scale efficiency[c]
Mean
Total 0.8129 0.8650 0.9408
Metropolitan 0.7944 0.8464 0.9399
Country 0.8408 0.8929 0.9422
Standard deviation
Total 0.1460 0.1395 0.0787
Metropolitan 0.1551 0.1543 0.0779
Country 0.1271 0.1120 0.0804
Minimum value
Total 0.4446 0.4477 0.5782
Metropolitan 0.4446 0.4477 0.5782
Country 0.6226 0.6422 0.6811
Number of efficient patrols
Total 29 57 29
Metropolitan 16 31 16
Country 13 26 13
Number of observations
Total 163 163 163
Metropolitan 98 98 98
Country 65 65 65
[a] Constant returns to scale model
[b] Variable returns to scale model
[c] Constant returns to scale technical efficiency/variable returns to scale technical efficiency
112
5 CASE STUDIES
Table 5.4.3: Input slack variables

Number of Total slack as
patrols with percentage of
Inputs slacks Mean total input
Police officers 8 11.32 1.09
Civilian employees 62 2.24 12.98
Police cars 10 1.96 1.15
Table 5.4.4: Summary of returns to scale of NSW police patrols

Increasing Constant returns Decreasing
returns to scale to scale returns to scale Total
Metropolitan 39 16 43 98
Country 16 13 36 65
Total 55 29 79 163
Table 5.4.5: Results of Tobit regression

Normalised Standard Regression
Variable coefficient [a] error t - ratio coefficient
Proportion of young people that live in
or visit a patrol –3.4063 2.4865 –1.3700 –0.6495
Proportion of government housing in a
patrol –2.3768 1.3310 –1.7857 –0.4532
Location of patrol 0.1354 0.1988 0.6809 0.0258
Constant 5.6237 0.7181 7.8303 1.0724
Pure technical efficiency score of a patrol [b] 5.2441 0.3917 13.387
Log-likelihood function = –23.6772

Standard error of estimate (σ) = 0.1906
Squared correlation between observed and expected values = 0.0666
[a] Normalised coefficient = regression coefficient/ = βi/σ.
[b] Dependent normalised coefficient = 1/σ
113
5.5 Technical efficiency of NSW motor registry offices
5.5.1 Introduction
The Roads and Traffic Authority (RTA) is responsible for licensing drivers,
registering vehicles, promoting road safety and traffic management, and
constructing, maintaining and enhancing roads in NSW. The RTA allocated
about $144 million in 1995-96 to motor registries which predominantly oversee
the first two responsibilities.
The RTA differs from other NSW budget sector agencies because it has access
to a defined pool of funds from Commonwealth grants, user charges, and
hypothecated state taxes such as the motor vehicle weight tax and fuel levies.
Despite having a more defined revenue stream, the RTA is still subject to
Government direction and control. Its operations must encompass current
Government policies and initiatives such as improvements in resource allocation
and efficiency reviews.
The RTA has an extensive array of performance indicators to monitor and
improve the effectiveness and efficiency of the delivery of motor registry
services. Customer satisfaction is used as an indicator of effectiveness.
Efficiency indicators include total cost per weighted transaction,34 weighted
transactions per labour hour, and the time spent by customers in registries.
These are useful efficiency measures but they can vary for reasons other than
inefficiency, such as the scale of the motor registry, different input
combinations used by registries, and the environment in which services are
delivered.
More sophisticated techniques, such as DEA, assess the efficiency of a motor
registry by comparing its inputs to produce services, and take into account its
scale and its operating environment in examining that efficiency. Better
information on the efficiency of motor registries provides the RTA with
additional opportunities to free funds for other uses such as road maintenance,
or to provide the same services with less reliance on state motor vehicle weight
taxes or fuel levies. This paper examines the scope for the RTA to improve the
efficiency of its motor registries using DEA.
34 Motor registries may perform up to 150 different types of transactions. Each type of
transaction is weighted by the average time taken to perform it.
114
5 CASE STUDIES
Related studies
During this study, it was unknown whether there were other studies that
measured the efficiency of motor registries using DEA. However, there was a
proposal to measure the performance of Queensland Transport using DEA
(National Road Transport Commission 1994). The proposal aims to develop
performance indicators for four areas of Queensland Transport’s operations —
road maintenance, road construction, system stewardship and driver licensing
and vehicle registration. Customer service centres provide driver licensing and
vehicle registration services. The proposal considers that customer services are
the output of customer service centres, and that labour, capital and materials are
the inputs. Tasman Asia Pacific (1996) prepared a report for the National Road
Transport Commission that included an assessment of the efficiency of
Queensland customer service centres after the completion of this analysis. The
study has one output for customer service centres total, minutes of service (the
number and type of transactions weighted by the average time for each type of
transaction) and two inputs, total labour costs and other operating costs which
exclude rates, rent and capital purchases because there is incomplete
information on these costs.
There is a substantial body of literature on financial institutions (banks and
credit unions), post offices and pharmacies using DEA. Motor registries operate
in an analogous manner to these service providers because they provide counter
services and form part of a branch network. Therefore, these studies provided a
guide to specifying the outputs and inputs used in this study.
Studies of the efficiency of financial institutions have similar measures for
inputs. However, their measures for outputs differ from those used in this study
because the monetary transactions that take place in financial institutions are of
a different nature from those in motor registries.
Ferrier and Lovell (1990), Aly et al (1990), Fried, Lovell and Eeckhaut (1993),
Ferrier et al (1993), and Fried and Lovell (1994) considered the important inputs
of financial institutions were labour, raw materials and capital. These inputs
were either combined and measured in aggregate operating expenditure or
measured individually — for example, labour was measured in staff numbers or
the wage bill; raw materials were measured by expenditure on this input, or as
non-labour operating expenditure; and capital was measured by rental costs or
by its book value. Outputs included a variety of loans and bank deposits which
were measured in physical quantities or in dollars.
Doble (1995) presented a model of post offices in the United Kingdom. Labour,
measured in full-time equivalent hours by counter clerks, was the only input
used in the study. Doble excluded hours of work done by branch managers,
115
arguing that the management of staff did not have a direct effect on the
production of counter transactions.
A post office handles approximately 190 different types of transactions.
Different types of transactions require different amounts of staff time, so
transactions are weighted by the average time to complete each type of
transaction. Doble included nine categories of weighted transactions, such as
issuing stamps and vehicle licences, as outputs in his study. The large sample of
1281 post offices allowed Doble to include a large number of outputs in the
DEA analysis and still obtain sensible results. The study included an output for
the quality of service provided by post offices, which was measured by the
average time that a customer waits for service.
The Färe, Grosskopf and Roos (1995) study of state–owned Swedish
pharmacies included several outputs, such as preparing prescriptions, delivering
drugs, and selling special articles and food for people with disabilities (which
were measured by the number of services provided by pharmacy staff); selling
consumer goods (which was measured by the number of sales, or transactions);
and conveying information on drugs (which was measured by the hours spent
collecting, preparing and conveying the information). The Swedish Government
required the pharmacies to meet certain quality of service standards, so several
attributes of the provision of pharmacy services were included in the model:
business hours, the percentage of prescriptions filled in one day, and the time
that customers wait for service. Inputs included pharmacists and technical staff,
measured by the hours they worked during the year, and non-labour operating
expenditure.
5.5.2 Provision of motor registry services
Outputs
A motor registry may perform up to 150 different transactions, which include
the issue and renewal of driver licences, motor vehicle registration, number
plates, firearm licences and driver licence testing. The RTA records all the
transactions conducted by motor registry counter staff. The different
transactions require similar staff skills but different amounts of time, so the total
number of transactions might not reflect the resources used. Thus, the total
number of transactions weighted by the average time spent to perform each type
of transaction was adopted as a proxy for the services provided by motor
registries.
The total number of weighted transactions did not reflect the quality of the
service provided. One aspect of the quality of service provided by a motor
116
5 CASE STUDIES
registry is the time that a person queues for service: the longer a person waits,
the more likely they are to be dissatisfied with the quality of service. Twice a
year, the RTA measures this waiting time in a motor registry. In this study,
waiting time was calculated as the average of the two surveys.
However, waiting time is an output that registry management should minimise,
so the inverse of waiting time is used in the analysis. Doble (1995) used an
alternative method of inverting waiting time: the average time that a person
waits for service in a post office was subtracted from the highest average time
that a person waits for service. This indicator of quality now measured the time
that a customer did not wait for service compared with the maximum time they
could wait for service.
Inputs
Motor registries use people, capital equipment and raw materials (such as
stationery) to provide their services. Labour was measured by the total hours
that staff work in the year, capital was measured by the value of computer
equipment, and raw materials were measured by the expenditure on these items.
The total hours that staff work in a year included the work of managers and
supervisors, permanent and casual staff, and staff on loan to the registry. It
excluded recreation, sick and special leave, training away from the registry, staff
on loan to other registries, and managers away attending meetings or
participating in quality improvement teams to improve the performance of
motor registries in a particular region.
The capital of a motor registry included computer terminals, photocopiers,
telephones and buildings. However, the RTA had incomplete information on
these assets; it suggested that the number of computer terminals was a good
proxy for the capital used by a motor registry. However, information on the
number of computer terminals actually used to serve customers was not
available. The bulk of the computer terminals were installed in registries in
1991, and only minor investment in computer equipment had occurred since that
date. Therefore, the value of computers installed in a motor registry in 1991
should have reflected the number of computer terminals it used, provided the
number of terminals had not been altered in response to significant changes in
the demand for registry services.
The main raw materials used to produce transactions included licences, plates,
postage and stationery. Total expenditure on these items was used to measure
this input.
117
In summary, the DEA model for motor registries had weighted transactions and
the reciprocal of waiting time as outputs, and the total staff hours, the value of
raw materials and the value of computers in 1991 as inputs.
Influence of environmental variables

Some factors beyond the control of management may influence the efficiency of
motor registries. The RTA identified two environmental variables which were
considered in this study: whether the registry was open on a Saturday, and
whether the RTA motor registry did data entry for councils and police stations
that operate as agencies of the RTA in remote regions of New South Wales.
To improve customer service, the RTA extended business hours of motor
registries by opening selected registries statewide on Saturday mornings. This
allowed these registries to use their capital more effectively. These registries
faced heavy demand on Saturdays. It was estimated that about two-thirds of
Saturday customers came from the Monday to Friday business of the same
registry. The remaining third was estimated to come from business of other
registries that did not open on Saturday. Given that most customers choose to
visit a motor registry rather than use the mail to complete their business, the
additional business to Saturday traders lowered the demand for services in other
registries that did not open that day. Moreover, motor registry managers had
limited scope to adjust labour to reductions in demand for Monday to Friday
trade, because they had to employ full-time staff to cover the peak period of
demand for their business.35 Therefore, a motor registry that traded Monday to
Friday could appear less efficient, compared with a similar registry that traded
Monday to Saturday.
All but one of the agencies did not have access to the central computer system
of the RTA, so other registries (either nearby country registries or registries
requiring extra work) processed their transactions into the RTA database.
Labour and capital were likely to be used more efficiently when processing an
agency’s transactions because staff did not serve customers. Further, agency
work could be processed during non-peak periods of customer demand,
enabling labour and capital to be used more fully. Accordingly, agency work
allowed parent registries to process more transactions than non-parent registries,
for the same amount of inputs.
35 The RTA is in the process of negotiating for greater flexibility in staffing patterns so
registry managers can schedule staff to meet the varying demand for services during the
day.
118
5 CASE STUDIES
5.5.3 Data and estimation of technical efficiency

Registry managers have no control over the demand for the registry’s services.
Therefore, their main objective is to handle transactions with the least inputs.
This implies minimising the resources used to complete each transaction while
maintaining the quality of service. Accordingly, a DEA model with an input-
minimising orientation was used to estimate the technical efficiency of motor
registries. There were 137 motor registries in NSW in 1994-95, but only 131
registries were included in this study. (Registries that were to close during 1994-
95 were excluded, because there was incomplete data for these registries.) Forty
registries opened on Saturdays and eighteen performed data entry for agencies
of the RTA.
Outlier analysis
DEA is susceptible to outliers, which are observations which are not typical of
the rest of the data. The production frontier estimated by DEA is determined by
extreme input and output points, so a single outlier can heavily influence the
measures of technical efficiency. Outliers may arise for two reasons. Outliers
may arise from errors in the data caused by measurement error or reporting
error. Alternatively, if the data is viewed to come from a probability distribution
then it is possible for the data to have observations with a low probability of
occurring. Outliers may reflect important phenomena which would go unnoticed
if they were excluded from the analysis (Burgess and Wilson 1993).
The data for motor registries were screened for potential outliers by examining
the summary statistics for each output and input, which are presented
Table 5.5.1.
Table 5.5.1: Descriptive statistics for NSW motor registry offices[a]

Standard
Mean Deviation Minimum Maximum
Outputs
Weighted transactions 179 932 118 850 24 080 607 784
Reciprocal of waiting time (1/minutes) 0.43 0.88 0.10 10.00
Inputs
Labour (hours) 14 029 9 136 2 809 41 906
Raw materials ($) 66 789 47 452 7 379 286 675
Computers ($) 91 060 36 884 41 163 234 048
[a] 131 observations
Furthermore, output–input ratios were calculated for each motor registry and the
values were checked using a two-and-a-half standard deviation rule. That is, any
119
observation which was observed to be greater than two-and-a-half standard

deviations from the sample mean was considered a potential outlier. Potential
outliers were referred to the RTA who checked the data and confirmed there
were no obvious measurement or reporting errors. Consequently, no
observations were excluded from the analysis.
5.5.4 Results
Summary statistics of the various measures of technical efficiency are presented
in Table 5.5.2. The method for calculating the technical efficiency and its
components is presented in Appendix A. The results presented suggest that pure
technical efficiency was the main source of technical inefficiency rather than
scale inefficiency. Pure technical efficiency indicated the possible improvement
in the use of inputs to produce the same outputs that the RTA could achieve
without altering the scale of its motor registries. This could be called managerial
efficiency. On average, it appeared that motor registries may be able to produce
the same level of measured outputs with 15 per cent fewer inputs.
Table 5.5.2: Summary statistics: technical efficiency of NSW motor

registry offices
Technical efficiency
(scale and pure Pure technical Scale
technical efficiency)[a] efficiency[b] efficiency [c]
Mean 0.81 0.85 0.95

Standard deviation 0.10 0.09 0.07
Minimum 0.55 0.63 0.64
Number of efficient motor 6 14 6
registries
[a] Constant returns to scale model
[b] Variable returns to scale model
[c] Constant returns to scale model/variable returns to scale model
Motor registries, on average, appeared to be 95 per cent scale efficient. If motor

registries could adjust to their optimal scale, then they may be able to further
reduce inputs, on average, by 5 per cent. The results in Table 5.5.3 indicate that
the majority of registries which appeared to be less efficient were experiencing
increasing returns to scale, suggesting that they were too small rather than too
big.
There may be social, demographic or geographic reasons for a motor registry
being a particular size. If there were no barriers to amalgamation or separation
of registries, then information on scale efficiency could assist the RTA in
120
5 CASE STUDIES
determining the optimal size of their registries. If barriers do exist, information

on scale efficiency indicates the costs incurred in maintaining the existing level
of service provision in a particular region.
Table 5.5.3: Summary of returns to scale on NSW motor registries

Increasing Constant Decreasing
returns to scale returns to scale returns to scale
Number of motor registries 75 6 50
The mean pure technical efficiency score could overstate the efficiency of motor
registries if some inputs are used excessively, beyond that reflected in the
efficiency scores. About 70 per cent of motor registries have excessive input
use. Table 5.5.4 reveals the scope for further non-radial reductions in inputs
(termed as ‘slacks’) once a motor registry operates on the production frontier.
Motor registries with such excessive inputs may be able to reduce their labour,
on average, by 616 hours, their raw materials by $11 152 and their computer
terminals by $11 274. Excessive inputs as a proportion of total inputs accounted
for about 9 per cent of raw materials, 5 per cent of computers and less than 1 per
cent for labour.
Table 5.5.4: Slack input variables

Number of registries Total slacks as a percentage of
Input with slacks Mean total inputs
Labour (hours) 6 616 less than 1
Raw materials ($) 67 11 152 9
Computers ($) 51 11 274 6
There are several methods for including environmental variables in a DEA

analysis. These are discussed in Chapter 2. This study used a two-step procedure
to analyse external factors of operating environments which might have
explained the variation in technical efficiencies across motor registries. The
two-step procedure required that inputs and environmental variables were not
correlated (Lovell 1993), and so Saturday trading and agency work were not
strongly correlated with the inputs used in this study.
The pure technical efficiency scores were regressed against the environmental
variables that indicated whether the registry conducted agency work or traded
on Saturdays. Given that the pure technical efficiency scores were truncated at
one, and that ordinary least squares estimation of censored data produced biased
and inconsistent results, a Tobit procedure was used (Judge et al. 1988).
121
The algebraic version of the Tobit model is presented in equation (1) and the
estimated model is presented in Table 5.5.5.
(1) TEi = β0 + β1 AGENCYi + β2 SATi + u i i = 1, ... ,131
TEi is the pure technical efficiency score of the i-th registry and ui is a normally
distributed disturbance term. AGENCYi and SATi are binary variables for the
i-th registry. A value of one in the AGENCY variable indicates that the registry
conducts agency work while a zero indicates otherwise. Similarly, a one in the
SAT variable indicates that it trades on Saturdays while a zero indicates that it is
not open on Saturdays.
The sign of the coefficients indicates the direction of influence of the
environmental variables, and the ratio of the estimated coefficients to their
standard errors (t-ratios) indicates the strength of the relationship between
efficiency and each variable. The squared correlation between the observed and
expected values indicates how much of the variation in efficiency scores can be
explained by the environmental variables (agency work and Saturday trading).
Table 5.5.5: Results of the Tobit regression

Normalised Standard error t-ratio Regressionn
coefficient coefficient
SAT 0.1678 0.1921 0.8736 0.0161
AGENCY 0.1941 0.2577 0.7531 0.0187
CONSTANT 8.7864 0.6001 14.64 0.8445
TE 10.4040 0.7026 14.81
Log-likelihood function = 90.4220

Mean-square error = 0.0076
Mean error = 0.0019
Squared correlation between observed and expected values = 0.0091
The signs on the estimated coefficients were as expected. Both variables had a
positive influence upon the level of pure technical efficiency. However, neither
variable was significant at the 5 per cent level of significance. The equation
explained about 1 percent of the variation in pure technical efficiency scores.
Based on these results, the pure technical efficiency scores were not adjusted for
the influence of agency work and Saturday trading.
122
5 CASE STUDIES
5.5.5 Conclusions
The analysis of the technical efficiency of motor registries, as measured,
suggested, on average, that they may be able to produce the same level of
outputs with 15 per cent fewer inputs. Pure technical efficiency appeared to be
the main source of inefficiency rather than scale inefficiency. However, if motor
registries could achieve optimal scale they could further reduce inputs, on
average, by 5 per cent. Care is required in interpreting the results because there
were weaknesses with the measure for capital. Nevertheless, the results
provided indicative information on the technical efficiency of motor registries.
To improve the measure of capital, the RTA has surveyed each motor registry to
obtain the number of computer terminals it uses to process transactions. Further,
it is developing weightings for agency work. This will reduce the potential for
these transactions to improve the technical efficiency of registries that process
agency work. These improvements will be included in future studies that
determine the technical efficiency of motor registries.
NSW Treasury has provided the RTA with the technical efficiency scores for
individual motor registry offices and associated peer information from this study
as a systematic framework for raising and addressing questions about the
performance of their motor registries.
123
APPENDIX A TECHNICAL GUIDE TO DEA
DEA is the term used by Charnes and Cooper (1985) to describe a non-
parametric approach to measuring efficiency. Diewert (1993) and Zeitsch and
Lawrence (1996) recently reviewed this technique, and the following discussion
borrows from their work.
A1 Technical efficiency
There are several different ways to present the technical efficiency linear
programming problem for DEA. The simplest presentation for the input-
oriented, constant returns to scale version of DEA is:
(A1) Minimise En
w1,...,wN,En
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − E n x kn ≤ 0 k = 1 ,...., K
w j ≥ 0 j = 1 ,...., N .
where there are N organisations producing I different outputs yin for i = 1, … , I

using K different inputs xkn for k = 1, … , K. The wj are weights applied across
the N organisations and the solution, En*, is the efficiency score of the nth
organisation. The linear program solves for the convex combination of the N
data points that can produce at least the observation n output and use at most
En* times the observation n combination of inputs. To get a full set of efficiency
scores, this problem has to be solved for each of the N organisations.
The linear programming problem for the output-oriented, constant returns to
scale version of DEA is similar to the above problem, except that it takes the
123
DATE ENVELOPMENT ANALYSIS
convex combination of observations that uses no more inputs than organisation

n and produces the maximum amount of outputs. The output-oriented DEA
linear programming problem is given by the following equation:
(A2) Maximise Fn
w1,...,wN,Fn
Subject to:
∑
N
j =1
w j y ij − F n y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − x kn ≤ 0 k = 1 ,...., K
w j ≥ 0 j = 1 ,...., N .
The first constraint indicates that the output of the hypothetical weighted
average has to be at least as great as n’s output scaled up by the factor Fn. The
second set of constraints state that the weighted average of the inputs cannot be
any larger than n’s input.
Returning to the input-oriented case, the constant returns to scale technical
efficiency score can be decomposed into three components — one due to a sub-
optimal scale of operations (scale efficiency); a second due to an inability to
dispose of ‘surplus’ inputs (congestion efficiency); and a residual or ‘pure’
technical efficiency. To form these measures, the DEA linear programs in (A1)
need to be re-run under the assumptions of variable returns to scale and variable
returns to scale with congestion.
124
APPENDIX A: TECHNICAL GUIDE TO DEA
The variable returns to scale DEA linear program is given by:

(A3) Minimise Sn
w1,...,wN,Sn
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − S n x kn ≤ 0 k = 1 ,...., K
∑
N
j =1
w j = 1
w j ≥ 0 j = 1 ,...., N .
As noted in Chapter 3, the extra constraint that the weights must sum to one has
the effect of pulling in the frontier to form a tighter envelope around the data.
The DEA linear programming problem under variable returns to scale with
congestion is given by:
(A4) Minimise Cn
w1,...,wN,Cn
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − C n x kn = 0 k = 1 ,...., K
∑
N
j =1
w j = 1
w j ≥ 0 j = 1 ,...., N .
125
The effect of placing an equality on the input constraint is to allow the frontier
to ‘bend backwards’ as in Figure 3.4. In technical terms, the assumption of ‘free
disposability’ of inputs is removed. This means that an organisation cannot
costlessly get rid of inputs to move down to the segment of the frontier that runs
parallel to the axes in Figure 3.2.
The three components of technical efficiency can now be defined as follows:
(A5) Scale efficiency = En / Sn
(A6) Congestion efficiency = Sn / Cn
(A7) Residual efficiency = Cn
The product of (A5), (A6) and (A7) is simply the constant returns to scale
efficiency score, En, in the original DEA model (A1).
As noted in Chapter 2, a scale efficiency score of less than one does not indicate
whether the organisation is bigger or smaller than its optimal size. To establish
this, an additional variant of DEA — one subject to non-increasing returns to
scale — must be run. The DEA linear programming problem for the non-
increasing returns to scale case is given by:
(A8) Minimise Rn
w1,...,wN,Rn
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − R n x kn ≤ 0 k = 1 ,...., K
∑
N
j =1
w j ≤ 1
w j ≥ 0 j = 1 ,...., N .
If the scale efficiency score is less than one and En and Rn are equal, then n is
subject to increasing returns to scale and would need to increase its size to reach
126
its optimum scale. If En is less than Rn , then n is subject to decreasing returns to

scale and would need to reduce its size to reach its optimum scale.
The technical efficiency DEA problems outlined in this section only require
information on output and input quantities. They do not use any information on
output or input prices. As noted in the chapters, the difficulty of allocating
prices to human services outputs makes DEA a relatively attractive technique
compared with total factor productivity. However, even for human services,
information on input prices and costs is often available, allowing an
organisation’s cost and allocative efficiency to be calculated.
A2 Cost and allocative efficiency

If the input prices for each organisation are known, then the cost efficiency
score for each observation can be calculated by solving N linear programs of the
form:
∑
K
(A9) Minimise k =1
pkn x kn
w1,...wN,x1n,...,xKn
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − x kn ≤ 0 k = 1 ,...., K
w j ≥ 0 j = 1 ,...., N ,
where p1n , … , pKn are the input prices for the K inputs that unit n faces.
This linear program chooses the input quantities that minimise n’s total costs
subject to a feasibility constraint and assuming that the input prices it faces are
fixed. The feasibility constraint requires that the weighted average which forms
the hypothetical efficient organisation has outputs at least as great as n’s and
* *
inputs no greater than n’s. The solution vector to (A9) x1n ,...,
is x Kn is n’s cost-
minimising level of inputs given its input prices and output level.
127
The technical efficiency scores derived from the linear programming problem
(A1) can be combined with the solutions to the cost-minimising linear
programming problems (A9) to form measures of the cost and allocative
efficiency of each organisation. Specifically, cost efficiency is found by
dividing the costs that would be faced by an organisation if it used the cost
minimising level of inputs by its actual costs. Thus:
∑k =1 pkn xkn* / ∑k =1 pkn xkn .

K K
(A10) Cost efficiency for the n-th observation=
A score of one for this index would indicate that an organisation is cost
efficient.
Allocative inefficiency is calculated by dividing costs faced by an organisation
assuming it used the cost-minimising level of inputs by costs assuming the
organisation used the technically efficient level of inputs. Thus:
∑ / En ∑k =1 pkn xkn
K * K
(A11) Allocative efficiency = k =1
pkn xkn
where En is the technical efficiency score derived from the linear programming
problem (A1).
From (A11) it can be seen that an organisation’s cost efficiency is the product of
its allocative efficiency and its technical efficiency.
A3 Accounting for operating environment differences

There is always a trade-off in DEA studies between ensuring that a like-with-
like comparison and maximum use is made of the information available to learn
how to improve performance. Limiting the study to like-with-like comparisons
leads to the comparisons being ‘fairer’ and perhaps more readily acceptable to
managers. However, a diverse range of operating environments may be useful in
the study to provide a wider range of ideas and operating styles from which
managers could learn.
In most cases, at least some allowance will need to be made for differences in
the organisations’ operating environments. As noted in Chapter 2, there are
several different ways to do this. The simplest way is to restrict the comparison
set to other organisations that have similar or less favourable operating
environments. However, this selection process can be arbitrary and excludes a
lot of the information that might be available. More sophisticated ways of
128
allowing for operating environment differences involve either single or multiple

stage adjustment processes.
In the single stage adjustment methods, the DEA linear programming problem
directly incorporates the operating environment characteristic that is being
adjusted for. Again, there are several ways this can be done. One option is to
simply include the characteristic in a manner analogous to the other inputs. This
assumes that the characteristic can be radially contracted as can the other inputs,
which is unlikely to be realistic in most cases. For instance, in the example of
the impact of an area’s socioeconomic status on schools, a school cannot change
(at least in the short run) the socioeconomic status of its neighbourhood. Rather,
it is more appropriate to take this as being fixed. In that case, the characteristic
needs to be included as a constraint in the linear program, but not one into
which the efficiency score enters. This reflects the fact that the organisation has
no control over the characteristic. This is known as including the characteristic
as a non-discretionary input. The linear program for this problem is as follows:
(A12) Minimise Vn
w1,...,wN,Vn
Subject to:
∑
N
j =1
w j y ij − y in ≥ 0 i = 1 ,...., I
∑
N
j =1
w j x kj − V n x kn ≤ 0 k = 1 ,...., K
∑
N
j =1
w j z j − zn≤ 0
∑
N
j =1
w j = 1
w j ≥ 0 j = 1 ,...., N .
where zn is the value of the operating environment characteristic in question for

unit n, n=1,…, N. This specification can be extended to include multiple
129
operating environment characteristics but, as with any DEA specification,

including more constraints will tend to inflate efficiency scores. This technique
also requires the operating environment characteristic to be a continuous
variable.
Two-stage adjustment procedures allow more flexibility than the above
procedure. These techniques typically carry out the initial DEA calculation
without referring to operating environment characteristics. Then, they regress
the efficiency scores from the DEA problem against a set of operating
environment characteristics using the Tobit regression technique to allow for the
truncated distribution of the efficiency scores. The DEA scores can be adjusted
to ‘standardise’ for the particular characteristic. This produces a set of
efficiency scores assuming that all organisations were operating with the same
degree of this characteristic. This approach has the advantage of not requiring
the direction of the characteristic’s impact on the efficiency scores to be
specified in advance. It also means that statistical tests can be carried out on the
strength of the relationship between the characteristic and efficiency levels.
Some practitioners have extended the two-stage adjustment procedure to three
and four stages to allow for slacks as well as radial inefficiency. Fried, Schmidt
and Yaisawarng (1995) described one such approach in detail, as well as
reviewing different approaches to adjusting for operating environment
differences.
130
APPENDIX B PROGRAMS FOR THE
APPLICATION OF DEA
There are a number of software options for running DEA. These can
be categorised as specialist DEA software and other software which
has the capacity to conduct linear programming and which can be
customised to perform DEA. Some examples and contact points are
listed below.
B 1 Specialist DEA software

This software can be purchased from the following contact points.
• DEAP — Tim Coelli, University of New England, Armidale, Australia
Fax 067 73 3607, Email: tcoelli@metz.une.edu.au
• Frontier Analyst — Bernard Petter, Strategic Retail Directions
Fax 03 9574 8882, Web site: http://www.scotnet.co.uk/banxia/famain.html
• Frontier Analyst — Marjory Sweeney, Banxia Software, Glasgow
Fax: +44 141 552 5765, Web site: http://www.scotnet.co.uk/banxia
• IDEAS — Shirley Shmering, 1 Consulting (US), Fax: + 413 256 1211
• PASS — Dr C.T. Clark, Productivity Assessment Consultants, Educational
Productivity Council, University of Texas, U.S.A.
Fax +1 512 301 1931
• Warwick DEA — Antreas Athanassopoulos, Warwick University (UK),
Fax +44 203 52 3719
Web site: http://www.csv.warwick.ac.uk/~bsrlu/dea/deas/deas1.htm
B2 Linear programming software

This software has been developed to run linear programming problems
specifically.
• General Algebraic Modelling System (GAMS)
Web site: http://www.gams.com/docs/intro.htm
• LINDO — Web Site: http://www.lindo.com/products.html
This software has a linear programming option.
• Microsoft Excel (using the solver tool)
131
• SAS — Web site: http://www.sas.com/www-bin/jump.pl

• SHAZAM — Web Site: http://shazam.econ.ubc.ca/
Example of programming for DEA using Shazam

This section is a brief review of how to program the variable returns to scale
DEA problem in the Shazam econometrics package. The programs were run
using version 7 of Shazam and the data for the twenty hospitals listed in
Table 3.1.
Using matrix notation, the linear programming module of Shazam is set up to
solve problems of the form:
Maximise c’x
Subject to the constraints:
Ax ≤ b
x≥0
The x vector contains the coefficients for which Shazam solves (the wi’s and the
En in the terminology of Chapter 3) and the term c’x is the objective function of
the linear program. Because the objective function only contains the En term, all
the coefficients in the c vector will be zero except for the last one which is equal
to one. An option within the linear program command allows the problem to be
changed from maximising the objective function to minimising it.
The problem can be thought of in terms analogous to equation system (1) in
Chapter 3. The A matrix contains the output values, input values and the
summation coefficients corresponding to all the terms in the constraints of (1),
up to but not including the last term before the inequality sign. The last terms
before the inequality sign in the output and input constraints change for each
DEA problem being run, because they contain the information specific to the
organisation being examined in the individual run.
To automate the process so all twenty DEA problems are solved in the one
Shazam run, a number of do loops are introduced. The input file for the Shazam
run is:
FI 22 A20.TXT
FI 23 B20.TXT
FI 24 C20.TXT
FI 25 D20.TXT
READ (22) A / ROWS=6 COLS=20
READ (23) B / ROWS=6 COLS=1
READ (24) C / ROWS=21 COLS=1
132
APPENDIX B: PROGRAMS FOR THE APPLICATION OF DEA
READ (25) D / ROWS=6 COLS=1

DO #=1,20
COPY A:# B / FROW=1;2 TROW=1;2
COPY A:# D / FROW=3;4 TROW=3;4
MATRIX D=-1@D
MATRIX AD=A|D
?LP C AD B / MIN ITER=100 PRIMAL=S#
ENDO
SMPL 1 21
GENR T=TIME(0)
GENR SCORE=0.0
DO #=1,20
GENR SCORE(#)=S#(21)
ENDO
FORMAT(11F7.4)
PRINT T S1-S10 / FORMAT
PRINT T S11-S20 / FORMAT
PRINT T SCORE / FORMAT
STOP
The first commands nominate the files which contain the data. The data are then
read in matrix form. The A matrix contains the output and input data for the first
twenty terms in the constraints in Chapter 3’s equation system (1). The first do
loop is then formed, which enters the information specific to each organisation
found in the terms immediately before system (1)’s inequality signs. The first
copy command moves the organisation’s output data into the first element of the
b vector (this is slightly different to the way in which (1) is set out, but it has an
equivalent effect). Shazam only allows less than or equal to constraints, so all
the output quantities are multiplied by –1 before being entered into the A
matrix. This makes the less than or equal to constraint on the negatives
equivalent to a greater than or equal to constraint on the positive output
quantities. Similarly, because Shazam does not explicitly have an equal to
constraint, constraints for the sum of the weights — one less than or equal to
and the other greater than or equal to — are included. The only way that both
constraints can be satisfied is by equality.
The second copy command and the following two matrix commands move the
input information for each organisation into the last terms before the inequality
signs in (1). This gives the 21 x 6 matrix AD.
The DEA linear program runs are done and the solution to each run in an s
vector is saved, the first twenty elements of which contain the weights for the
133
run and the twenty-first element of which is the efficiency score for that
organisation. A new score variable, which just contains all the efficiency scores
is formed, and finally the s and score vectors are printed. The s vectors identify
the peer group (those organisations with non-zero weights) and their relative
contribution to forming the hypothetical best practice target for the organisation
in question.
The following tables reproduce the A, b, and d matrices read into the program.
The c matrix was described above.
Table B1: Example of twenty hospitals — A matrix

Column Column Column Column Column Column Column Column Column Column
1 2 3 4 5 6 7 8 9 10
Row 1 –150 –225 –90 –160 –50 –75 –200 –350 –400 –250
Row 2 –50 –75 –10 –40 –50 –75 –50 –100 –90 –300
Row 3 200 600 200 600 500 320 375 400 550 900
Row 4 600 1200 200 300 200 150 450 320 480 660
Row 5 1 1 1 1 1 1 1 1 1 1
Row 6 –1 –1 –1 –1 –1 –1 –1 –1 –1 –1
Column Column Column Column Column Column Column Column Column Column
11 12 13 14 15 16 17 18 19 20
Row 1 –350 –350 –275 –220 –300 –320 –375 –230 –290 –360
Row 2 –350 –400 –375 –40 –10 –275 –230 –50 –90 –70
Row 3 850 720 900 250 115 600 550 200 450 415
Row 4 720 940 850 370 250 590 710 280 410 575
Row 5 1 1 1 1 1 1 1 1 1 1
Row 6 –1 –1 –1 –1 –1 –1 –1 –1 –1 –1
Table B2: Example of twenty hospitals — b and d vectors

b vector d vector
Row 1 0 0
Row 2 0 0
Row 3 0 0
Row 4 0 0
Row 5 1 0
Row 6 –1 0
134
REFERENCES AND FURTHER READING
General references
Ali, A.I, and Seiford, L.M. 1993, ‘The mathematical programming approach to
efficiency analysis’, in Fried, H.O., Lovell, C.A.K, and Schmidt, S.S.
(eds) 1993, The Measurement of Productive Efficiency, Oxford
University Press, New York.
Aly, H.Y., Grabowski, R., Pasurka, C., and Rangan, N. 1990, ‘Technical, scale
and allocative efficiency in US banking: an empirical investigation’,
Review of Economics and Statistics, LXXII, pp. 211–18.
Anderson, T. 1996, ‘A data envelopment analysis (DEA) home page’,
Internet: http://www.emp.pdx.edu/dea/homedea.html
BIE (Bureau of Industry Economics) 1993, Data Envelopment Analysis: An
Explanation, AGPS, Canberra.
—— 1994a, International Performance Indicators: Electricity Update 1994,
Research Report 54, AGPS, Canberra.
—— 1994b, International Performance Indicators: Gas Supply, Research
Report 62, AGPS, Canberra.
—— 1995, International Benchmarking: Overview 1995, Research Report
95/20, AGPS, Canberra.
Bogetoft, P. 1995, ‘Incentives and productivity measurements’, International
Journal of Production Economics, vol. 39, pp.67–81.
Charnes, A. and Cooper, W.W. 1985, ‘Preface to topics in data envelopment
analysis’, Annals of Operations Research, vol 2, pp. 59–94.
Charnes, A., Cooper, W.W., and Rhodes, E. 1978, ‘Measuring the efficiency of
decision making units’, European Journal of Operational Research,
vol. 2, pp. 429–44.
Coelli, T. 1996, ‘DEAP - a data envelopment analysis (computer) program’,
University of New England Centre for Efficiency and Productivity
Analysis, Internet: http://www.une.edu.au/econometrics/deap.htm
Diewert, W.E. 1993, ‘Data envelopment analysis: a practical alternative?’, in
Swan Consultants (eds), Measuring the Economic Performance of
Government Enterprises, Proceedings of a Conference held at the
Sheraton International Airport Hotel, 12 February.
135
Dyson, R.G., Thanassoulis, E. and Boussofiane, A. 1995, ‘A DEA

(data envelopment analysis) tutorial,
Internet: http://www.csv.warwick.ac.uk/~bsrlu/dea/deat/deat1.htm
Emrouznejad, A. 1995, ‘Ali Emrouznejad’s DEA home page’
Internet: http://www.csv.warwick.ac.uk/~bsrlu/
Färe, R., Grosskopf, S. and Lovell, C.A.K. 1983, ‘The structure of technical
efficiency’, Scandinavian Journal of Economics, vol.85, pp. 181–90.
——, —––, Norris, M., and Zhang, Z. 1994, ‘Productivity growth, technical
progress, and efficiency changes in industrialised countries’,
American Economic Review vol. 84, pp. 66–83.
Fried, H.O. and Lovell, C.A.K. 1994, ‘Measuring efficiency and evaluating
performance’, Workshop at Regent Hotel, Sydney.
——, and Schmidt, S.S. 1993, The Measurement of Productive Efficiency,
Oxford University Press, New York.
——, Schmidt, S.S. and Yaisawarng, S. 1995, Incorporating the operating
environment into a measure of technical efficiency’, Paper presented
to the Bureau of Industry Economics seminar, Canberra.
Ganley, J.A. and Cubbin, J.S. 1992, Public Sector Efficiency Measurement:
Applications of Data Envelopment Analysis, North-Holland,
Amsterdam.
Grosskopf, S. 1993, ‘Efficiency and productivity’, in Fried, H.O., Lovell,
C.A.K, and Schmidt, S.S. (eds), The Measurement of Productive
Efficiency, Oxford University Press.
Innovation Network 1993, ‘Applying benchmarking to higher education: part I
definitions’, Internet: http://www.innovnet.com/ben1.htm
Judge, C.G., Hill, R.C., Griffiths, W.E., Luthepohl, H. and Lee, T. 1988,
Introduction to the Theory and Practice of Econometrics, John Wiley
and Sons, New York.
Klein, R. and Carter, N. 1989, ‘Performance measurement: a review of concepts
and issues’, in Public Finance Foundation (ed), Performance
Measurement: Getting the Concepts Right, London.
Koopmans, T.C. 1951, ‘An analysis of production as an efficient combination
of activities,’ in Koopmans, T.C. (ed), Activity Analysis of Production
and Allocation, Cowles Commission for Research in Economics,
monograph no. 13, John Wiley and Sons, New York.
136
Lawrence, D. 1993, ‘What performance measures should we use?’, in Swan

Consultants (eds), Measuring the Economic Performance of
Government Enterprises, Proceedings of a Conference held at the
Sheraton International Airport Hotel, 12 February.
Lovell, C.A.K., 1993a, ‘Production frontiers and productive efficiency’, in
Fried, H.O., Lovell, C.A.K. and Schmidt, S.S. (eds) 1993, The
Measurement of Productive Efficiency, Oxford University Press, New
York.
——, 1993b, ‘Linear programming approaches to the measurement and analysis
of productive efficiency’, Working Paper 94–393E, College of
Business Administration, University of Georgia, Athens, Georgia.
Nakamura, A.O. and Warburton, W.P. 1996, ‘Developing high performance
public sector performance measures’, [mimeo].
Nunamaker, T.R. 1985, ‘Using data envelopment analysis to measure the
efficiency of non-profit organisations: a critical evaluation’,
Managerial and Decision Economics, vol.6, no. 1, pp. 50–8.
Pestieau, P. and Tulkens, H. 1993, ‘Assessing and explaining the performance
of public enterprises’, Finanz Archiv, vol.50, no. 3, pp. 293–323.
Pierce, J. 1995, Public sector performance measurement – A policy makers’
perspective. Presented to the Symposium on Public Sector
Performance Measurement, University of New England Conference
on Efficiency and Productivity, Armidale, November.
——, and Puthucheary, N. 1997, Using performance measures to drive change
within the public sector: the NSW experience’, Presented at Leading
for Outcomes 4, Kooralbyn Hotel Resort, Brisbane.
Rouse, P. 1996, Alternative Approaches to the Treatment of Environmental
Factors in DEA: an Evaluation’, Department of Accounting and
Finance, University of Auckland.
Salarian, J. and Jomini, P. 1994, ‘Total Factor Productivity, Annual User
Charges, Capital Stocks and Cost Allocation: A Synthesis’, The
Australian Economic Review, 3rd Quarter.
SCNPMGTE (Steering Committee on National Performance Monitoring of
Government Trading Enterprises) 1992, Measuring the Total Factor
Productivity of Government Trading Enterprises, Canberra.
—— 1996, Government Trading Enterprises Performance Indicators: 1990-91
to 1994-95, AGPS, Canberra.
SCRCSSP (Steering Committee for Review of Commonwealth/State Service
Provision) 1995, Report on Government Service Provision 1995,
AGPS, Canberra.
137
—— 1997, Report on Government Service Provision 1997, Vols. 1 & 2, AGPS,

Canberra.
Seiford, L.M. and Thrall, R.M. 1990, ‘Recent developments in DEA: the
mathematical programming approach to frontier analysis’, Journal of
Econometrics, vol. 46, pp. 7–38.
Senge, P. 1990, The Fifth Discipline: The Art and Practice of the Learning
Organization, Currency Doubleday, New York.
Smith, P. 1990, ‘The use of performance indicators in the public sector’,
Journal of the Royal Statistical Society, vol. 153, no.1, pp. 53–72.
Tasman Asia Pacific 1996, Measuring the Productivity Performance of State
Road Authorities: A Case Study of Queensland Transport, Report
Prepared for the National Road Transport Commission.
White, K. 1993, Shazam Econometrics Computer Program version 7.0,
McGraw Hill, Internet: http://shazam.econ.ubc.ca/
Zeitsch, J. and Lawrence, D. 1996, ‘Decomposing economic inefficiency in
base-load power plants’, Journal of Productivity Analysis, vol.7, pp.
359–78.
——, ——, Salerian, J. 1994, ‘Comparing like with like in productivity studies
- apples, oranges and electricity’, Economic Record, vol.70, no. 209,
pp.162–70.
Victorian hospitals case study

Banker, R., Das S. and Datar S. 1989, ‘Analysis of cost variances for
management control in hospitals’, Research in Government Non-
Profit Accounting, vol. 5, pp. 269–91.
Burgess, J. and Wilson, P. 1993, ‘Technical efficiency in veterans
administration hospitals’, in Fried, H., Lovell, K. and Schmidt, S.
1993, The Measurement of Productive Efficiency — Techniques and
Applications, University Press, New York, pp. 335–351.
Eagar, K. and Hindle, D. 1994, A Short Introduction to Casemix, National
Casemix Education Series no. 1, February, Department of Human
Services and Health, Canberra.
HCS (Health and Community Services) 1994, Casemix Funding for Public
Hospitals 1994-95, Victorian Department of Health and Community
Services, Melbourne.
138
NHDC (National Health Data Committee) 1995, National Health Data

Dictionary, Version 4.0, Canberra: Australian Institute of Health and
Welfare.
NSW Treasury 1994, ‘Applicability of data envelopment analysis for measuring
service provision performance in Australia’, Report to Steering
Committee on Commonwealth/State Service Provision, 16 November
[unpublished mimeo] Industry Commission, Melbourne.
SCRCSSP (Steering Committee for the Review of Commonwealth/State
Service Provision) 1995, Report on Government Service Provision
1995, Industry Commission, Melbourne.
Seiford, L. 1994 ‘A DEA bibliography (1978–92)’ in Charnes, A., Cooper, W.,
Lewin, A.Y., and Seiford, L.M. (eds) 1994, Data Envelopment
Analysis: Theory, Methodology and Applications, Kluwer Academic
Publishers, London.
Valdmanis, V. 1992, ‘Sensitivity analysis for DEA models’ Journal of Public
Economics, vol. 48, pp. 185–205.
Queensland oral health case study

Banker, R. D., Das, S. and Datar, S. M. 1989, ‘Analysis of cost variances for
management control in hospitals’, Research in Governmental and
Nonprofit Accounting, vol 5, pp 268–91.
Charnes, A., Cooper, W., Lewin, A. Y. and Seiford, L. M. 1994, Data
Envelopment Analysis: Theory, Methodology and Applications,
Kluwer Academic Publishers, United States.
Chilingerian, J. A. and Sherman, H. D. 1990, ‘Managing physician efficiency
and effectiveness in providing hospital services’, Health Services
Management Resources, vol 3, no. 1, pp 3–15.
NSW correctional services case study

ABS 1995a, Government financial statistics, Unpublished data AGPS,
Canberra.
— 1995b, National Accounts, cat. no. 5242.0, AGPS, Canberra.
— 1996, National Accounts, cat. no. 5242.0, AGPS, Canberra.
Farrell, M.J. 1957, ‘The measurement of productive efficiency’, Journal of the
Royal Statistical Society, A CXX, Part 3, pp. 253–90.
139
Ganley, J.A. and Cubbin, J.S. 1992, Public Sector Efficiency Measurement:
Applications in Data Envelopment Analysis, North Holland,
Amsterdam.
Lovell, C.A.K. 1993, ‘Production frontiers and productive efficiency’, in Fried,
H.O., Lovell, C.A.K. and Schimdt, S.S. (eds), The Measurement of
Productive Efficiency, Oxford University Press, New York.
Yaisawarng, S. and Vogel, S. 1992, ‘Input technical efficiency of local
correctional agencies in New York State’, Unpublished manuscript,
Union College, New York.
NSW police patrols case study

Afriat, S. 1972, ‘Efficiency estimation of production functions’, International
Economic Review, vol.13, pp. 563–98.
Burgess, Jr, J.F. and Wilson, P.W. 1993, ‘Technical efficiency in veterans
administrative hospitals’, Fried, H.O., Lovell, C.A.K. and Schmidt,
S.S. (eds), The Measurement of Productive Efficiency, Oxford
University Press, New York, pp. 335–51.
Carrington, R., Connolly, P. and Puthucheary, N. 1996, ‘Improving the
performance of NSW government agencies’, Paper presented to the
25th Annual Conference of Economists, Australian National
University, Canberra, 22–26 September.
Coelli, T. 1995, ‘A guide to DEAP version 1.0: a data envelopment analysis
program’, [Mimeo], Department of Econometrics, University of New
England, Armidale, New South Wales.
Darrough, M.N. and Heineke, J.M. 1979, ‘Law enforcement agencies as
multiproduct firms: an econometric investigation of production
costs’, Public Finance 34, pp. 176–95.
Före, R., Grosskopf, S. and Lovell, C.A.K. 1985, The Measurement of
Efficiency of Production, Kluwer-Nijhoff Publishing, Boston.
Grosskopf, S. 1986, ‘The role of reference technology in measuring productive
efficiency’, The Economic Journal, vol. 96, pp. 499–513.
Gyapong, A.O. and Gyimah-Brempong, K., 1988, ‘Factor substitution, price
elasticity of factor demand and returns to scale in police production:
evidence from Michigan’, The Economic Journal vol. 54, pp. 863–78.
140
Gyimah-Brempong, K., 1987, ‘Economies of scale in municipal police

departments: the case of Florida’, The Review of Economics and
Statistics, vol. 69, pp. 352–56.
—— 1989, “Production of public safety: are socioeconomic characteristics of
local communities important factors?”, Journal of Applied
Econometrics, vol. 4, pp. 57–71.
Judge, G.G., Hill, R.C., Griffiths, W.E., Luthepohl, H. and Lee, T., 1988,
Introduction to the Theory and Practice of Econometrics. John Wiley
and Sons, New York.
Levitt, M.S. and Joyce, M.A.S. 1987, The Growth and Efficiency of Public
Spending, Cambridge: University Press, Cambridge.
McCarty, T.A. and Yaisawarng, S. 1993, ‘Technical efficiency in New Jersey
school districts’, In Fried H.O., Lovell, C.A.K. and Schmidt, S.S.
(eds), The Measurement of Productive Efficiency, Oxford University
Press, New York , pp. 271–287.
NSW Government 1995, Financial Statement, Sydney.
NSW Treasury 1996, Performance of NSW Government Businesses, Sydney.
Pierce, J. 1995, ‘Public sector performance management - a policy maker’s
perspective’, Paper presented to the New England Conference on
Efficiency and Productivity, University of New England, Armidale,
New South Wales, 23–24 November.
Wilson, P.W. 1995, ‘Detecting influential observations in data envelopment
analysis’, The Journal of Productivity Analysis, vol. 6, pp. 27-45.
NSW motor registries case study

Aly, H.Y., Grabowski, R., Pasurka, C. and Rangan, N. 1990 ‘Technical, scale
and allocative efficiencies in US banking: an empirical investigation’,
Review of Economics and Statistics, vol. 72, pp. 221–18.
Burgess Jr, J.F. and Wilson, P.W. 1993 ‘Technical efficiency in veterans
administrative hospitals’ in Fried, H.O., Lovell, C.A.K. and. Schmidt,
S.S. (eds), The Measurement of Productive Efficiency, Oxford
University Press, New York.
141
Doble, M. 1995 ‘Measuring and improving technical efficiency in UK post

office counters using data envelopment analysis’, Annals of Public
and Cooperative Economics, vol. 66, pp. 31–64.
Färe, R., Grosskopf, S. and Roos, P. 1995 ‘Productivity and quality changes in
Swedish pharmacies’, International Journal of Production
Economics, vol. 39, pp. 137–47.
Ferrier, G.D., Grosskopf, S., Hayes, K.J., and Yaisawarng, S. 1993, ‘Economies
of diversification in the banking industry’, Journal of Monetary
Economics, vol. 31,. pp. 229–249.
—— and Lovell, C.A.K. 1990 ‘Measuring cost efficiency in banking:
econometric and linear programming evidence’, Journal of
Econometrics. vol. 46, pp. 229–245.
Fried, H.O. and Lovell, C.A.K. 1994, Credit Union Service, Filene Research
Institute, Madison.
——, ——, and Eeckhaut, P.V. 1993, ‘Evaluating the performance of US credit
unions’, Journal of Banking and Finance, vol. 17, pp. 251-65.
Judge, C.G., Hill, R.C., Griffiths, W.E., Luthepohl, H. and Lee, T. 1988,
Introduction to the Theory and Practice of Econometrics, John Wiley
and Sons, New York.
Lovell, C.A.K., 1993a, ‘Production frontiers and productive efficiency’, in
Fried, H.O., Lovell, C.A.K. and Schmidt, S.S. (eds) 1993, The
Measurement of Productive Efficiency, Oxford University Press, New
York.
National Road Transport Commission 1994, Aggregate Performance Measures:
A Case Study of Queensland Transport. Data Assessment and
Proposed Methodology. Interim Report, Working Paper 17.
Tasman Asia Pacific 1996, Measuring the Productivity Performance of State
Road Authorities: A Case Study of Queensland Transport, Report
Prepared for the National Road Transport Commission, Tasman Asia
Pacific, Melbourne.
142

Dea PDF

Uploaded by

Copyright:

Available Formats

Dea PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dea PDF

Uploaded by

Copyright:

Available Formats

.

DATA ENVELOPMENT ANALYSIS

A TECHNIQUE FOR MEASURING

Steering Committee for the Review of

 Commonwealth of Australia 1997

ISBN: 0 646 33533 2

The Steering Committee for the Review of Commonwealth/State Service

Telephone: 03 9653 2100 Facsimile: 03 9653 2199

An appropriate citation for this report is:

Effective government service provision benefits from the support of rigorous

DEA Working Group VI

1 Improving the performance of government service providers 1

2 What is data envelopment analysis? 9

3 How do you calculate DEA? 25

4 Overview of the case studies 39

A Technical guide to DEA 123

B Programs for the application of DEA 131

References and further reading 135

DEA WORKING GROUP

CRS Constant returns to scale

DEA Data envelopment analysis

RTA Roads and Traffic Authority (NSW)

SCNPMGTE Steering Committee on National Performance

SCRCSSP Steering Committee for the Review of

TFP Total factor productivity

WIES Weighted inlier equivalent separations

Allocative Whether, for any level of production, inputs are used

Effectiveness Degree to which the outputs of a service provider

Production Relationships incorporated in production processes

Technical efficiency Conversion of physical inputs such as labour services

Measuring the efficiency of government service provision can play an

1.1 Improving performance in government funded service

1.2 Who uses performance measures?

• focus attention on other organisations which may be performing better. More

development and application of new techniques and approaches to performance

1.3 Concepts of effectiveness, efficiency and productivity

The technical efficiency of an organisation depends on its level of productivity, or

2 Productivity is an absolute concept, measured by the ratio of outputs to inputs, while

Given the shortcomings of partial indicators, governments are adopting more

Box 1: Performance measurement in government trading

An international benchmarking program for key Australian infrastructure industries, started at

However, the TFP technique is not generally applicable to service provision,

1.4 Objectives of this paper

Data envelopment analysis provides a means of calculating apparent

2.1 Data envelopment analysis and different efficiency

would minimise costs. An organisation that is operating at best practice in

Figure 2.1: Illustration of different efficiency concepts

2 Cost efficiency is sometimes extended to include a third measure called dynamic

2.2 Operationalising the concepts

Table 2.1: Illustrative hospital data

Figure 2.2: Illustrative hospital input–output data

Beds per treated case

2.3 What questions can DEA help us answer?

2.4 Extensions to the DEA model

2.4.1 Scale efficiency

Figure 2.3: The production frontier and returns to scale

This can be demonstrated using the examples in Figure 2.3.

2.4.2 Input and output orientation

2.4.3 Input congestion