Data envelopment Analysis is a Mathematical Programme for measuring
performance efficiency of organizational units. The organizational
units, termed as decision-making units (DMU) can be of any kind:
manufacturing units, a set of schools, banks, hospitals, power plants,
police stations, prisons, a set of firms etc. DEA has
been unsuccessfully applied to measure the performance efficiency of
these different kinds of DMUs which share a common characteristic - that
they are non-profit organization where measurement of performance
efficiency is difficult.
In this book, author not only lays the foundation of the technique of
DEA effectively, but also dedicates a significant portion of the book to
illustrate it through full-fledged applications implemented by the
author himself. It covers the latest developments in this field and
lists out resources. A "not-too-rigorous" mathematical treatment makes
this book an easy rapid reading text. A chapter dedicated to do's and
don'ts from the implementation point of view and solutions to the
selected problems at the end of the book makes it very useful. The book
emphasizes on practice rather than abstract theory.
The book though
disappoints readers having a keen eye for details, due to a number of
typographical errors, use of variables without defining them and
inappropriate sequencing of problems within exercises and across
chapters. The inadvertent insertion of technical details unnecessarily
at numerous instances might distract or annoy readers. The matrix
representation of the formulations need not have been left in the main
body of the text. Similarly discussion on the non-Archimedean constant
could have been extracted out to an appendix.
The author begins
with very lucid and clear exposition of terms, concepts aided with
graphical illustrations. He elucidates the applicability of DEA to
Not-For-Profit organizations with reasoning and elaborates the
underlying philosophy of DEA. The generalized mathematical formulation
by first principles leading to formulation of fractional DEA program
through a simple example elicits interest among readers. The CCR model
laying the first stepping stone in the formal theory of DEA is
immediately followed by dual DEA formulation which capitalize the
interest aroused till then.
The relationship between output
(input) maximizing (minimizing) multiplier DEA and the input (output)
oriented envelopment DEA in chapter 2 could have been better illustrated
through a solved example. The two-stage optimization procedure is yet
another distraction which could have been avoided by taking it appendix
as it is not used further in text. In the section on interpretation of
dual, the example involving firms PQR & T is inserted unnecessarily
which creates discontinuity and concludes in an adhoc fashion.
In
line with the chronology of the theory of DEA, the concept of "economies
of scale" is introduced subsequently in chapter 3, which includes basic
concepts of CRS and VRS, technical and scale efficiencies, estimation
of MPSS and their detailed investigation. This helps the reader
understand the evolution of DEA to real life adaptation and its
increasing acceptance. This chapter ends with ample examples for the
reader to have a good grasp of the theory. This chapter introduces the
BCC model quite succinctly backed by good references, but is plagued
with a number of typographical errors. With this chapter, the basic
theory and evolution of DEA during its foundation years is covered
exhaustively and the text is fairly easy to understand as the shift to
application of theory beyond this chapter. Besides a good understanding
of the dual LP theory is a prerequisite to develop a sound foundation.
The
variants of basic models and the recent developments have been covered
in chapter 4 of the book which covers Multiplicative DEA models,
Additive models, Window Analysis and Malmquist Productivity Index
Approach. Similarly, variants in these models, such as use of
non-discretionary input / output, treatment of categorical variables,
incorporation of judgmental and subjective weights, are discussed. These
models and variants are briefly introduced but examples for select few
are illustrated. Examples for each type of variant could however have
made the chapter more effective. The chapter ends with three exercise
problems and a note for advanced readers on other DEA models which
establishes its linkage with multi criteria decision making models. This
evokes appreciation for utility of DEA extensions.
The best of
the book lies in the chapter on computer support for DEA, wherein the
computational features of DEA are taken into account along with the
resources available in print and web media. DEA software is presented in
the fifth chapter which is aptly detailed with finer implementation
level details like compatibility with the OS, the embedded solver,
limits on the size of the problem, interfacing with other common
software etc. The format of data input, specification of variables,
syntax of command, and snapshots of the various intermediate and final
outputs along with their interpretation make very interesting and
engaging reading. This is very helpful for readers who want to try out
DEA. The software of varying degrees of robustness and complexity are
demonstrated namely DEAP, EMS, GAMS formulation and Excel serving the
vast range of readers on the spectrum of elementary to advanced level.
The
chapter on DEA bibliography and application takes the reader to the
world of "DEA in practice" which helps one understand the issues
involved in problem formulation, implementation, sensitivity analysis of
model, post DEA analysis, validation and reaffirmation of the results
through techniques such as regression etc. The problems attempt to
demonstrate full-fledged implementation which range from comparative
performance of schools to productivity assessment of State transport
undertakings in India, from comparative risk assessment of energy
systems and energy efficiencies of transport modes in India to Carbon
dioxide emissions of countries. This wide range of problems adds to the
richness of the text especially when the studies have been carried out
by the author himself.
The book concludes with a chapter on do's
and don'ts for undertaking DEA studies. Some fundamental considerations
enlisted for characterizing DMUs, selection of input/output and DEA
model, post DEA procedures, strengths and weaknesses of DEA would
enhance confidence of amateurs. The chapter climaxes with a graph
clearly illustrating the severe limitation of DEA in case of DMUs having
same peer members, leaving the reader wanting for more.
A book
covering so much is expected to be shallow, but it makes easy reading
and opens the doors to the field of DEA. The most notable feature of the
book is its use of many small case studies, comprehensive bibliography
and illustration of software usage. "DEA made simple" or "DEA in 24
hours" could have been more appropriate titles.
- Paperback: 202 pages
- Publisher: SAGE Publications Pvt. Ltd (August 18, 2003)
- Language: English
- ISBN-10: 076199761X
- ISBN-13: 978-0761997610
Tidak ada komentar:
Posting Komentar