Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
"Are we really in charge?": An analysis of the principal -agent relationship
(USC Thesis Other)
"Are we really in charge?": An analysis of the principal -agent relationship
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
“ARE WE REALLY IN CHARGE?”
AN ANALYSIS OF THE PRINCIPAL AGENT RELATIONSHIP
by
John Richard Gray
A Dissertation Presented to the
FACULTY OF THE SCHOOL OF POLICY, PLANNING,
AND DEVELOPMENT
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PUBLIC ADMINISTRATION
August 2002"'
Copyright 2002 John Richard Gray
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3094329
Copyright 2002 by
Gray, John Richard
All rights reserved.
®
UMI
UMI Microform 3094329
Copyright 2003 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
SCHOOL OF POLICY, PLANNING, AND DEVELOPMENT
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089
This dissertation, written by
John Richard Gray
under the direction o f his Dissertation
Committee, and approved by all its
members, has been presented to and
accepted by the Faculty o f the School o f
Policy, Planning, and Development, in
partial fulfillment o f requirements for the
degree o f
DOCTOR OF PUBLIC ADMINISTRATION
^ Chairperson
Dean
Date
DISSERTATION COMMITTEE
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DEDICATION
To my wife, without whose support this would never have been possible; to
my children, Allison and Patrick, who challenged me to live up to the standards I
set for them, thank you.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
I would like to express my appreciation to my committee members, Dr.
Tom Pieplow, Dr. Ralph Bledsoe, and to the chair of ray committee, Dr. William
Petak, for their support, encouragement and guidance. Their willingness to reach
out and guide me through the dissertation process was invaluable.
I would also like to express my gratitude to Dr. John McLaughlin, who
offered his support for my proposal and encouragement that the entire project could
be completed. Finally, my thanks to my friend Dr. Carl Smolka, who worked
alongside me through the course work and led the way through the dissertation
process.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
DEDICATION ............................................................................................. ii
ACKNOWLEDGEMENTS .............. ill
LIST OF TABLES ........................................................................................... viii
LIST OF FIGURES.............................................................. ix
ABSTRACT .............................................................................................. xi
Chapter
I INTRODUCTION............................................................................... 1
The Growth in Government and Outsourcing............................. 2
Principal Agent Paradigm............................................................ 7
Relationship to Larger Trends in Government ....................... 10
Study Objective ........................................................................... 11
H. LITERATURE REVIEW ................................................................... 13
Principal Agency Theory Writings................................................ 14
Principal Agent Theory ........................................................ 14
S. A. Ross—Economic Theory of Agency......................... 14
K. Eisenhardt—Control as Organizational Determinant . . . 15
I. Pratt and R. Zeckhauser—An Overview ....................... 16
K. Eisenhardt—Review of the Principal Agent Theory . . . 19
Principal Agent Applied to Government............................... 20
D. Good and G. Miller—Principal Agent Applied to the
Public Sector................................................................... 20
S. Globerman and A. Vining—Application to Info
Technology........................................... 22
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Evolving Principal Agent in Government..................................... 23
Contracting Government Functions ..................................... 25
V. Heyman—Concerns from the 1960s............................. 25
E. S. Savas—Expansion of Contracting ......................... 26
I. Sharkansky—Pros and Cons .......................... 27
D. Kettl—Pros and Cons .................................................. 27
H. Stretton and L. Orchard—Overview........................... 29
R. Moe—Legalistic Concerns........................................... 30
I Kirkpatrick and M. Lucio—Public Choice Theory 31
H. B. Millward—The Hollow State................................... 32
Privatization .............................................................. 33
Competition Theory.................................................................... 34
Competition Writings............................................................ 35
I. Rehfiiss—The Case for Principal A gent......................... 35
J. Pack—Summarizing the Theories and Practices............ 36
C. Kent and D. Kettl—The Case Against ......................... 37
Competition in Commercial Activities ................................. 38
Directly Applicable Studies.......................................................... 40
HI. IMPORTANCE TO PUBLIC ADMINISTRATION......................... 46
New Paradigm of Management.................................................... 49
Changes in Federal Employment.................... 52
Implications to the Work Force....................................... 54
IV. SPECIFIC RESEARCH QUESTIONS .............................................. 56
Does Principal Agent Apply? ...................................................... 57
Do Participants Understand Principal Agent? ............................. 58
Is There a T rend?......................................................................... 59
Answering the Questions ............................................................ 60
V. RESEARCH METHODOLOGY........................................................ 63
Research Methodology Options .................................................. 63
Methodologies Eliminated.................................................... 65
Methodology Selected.......................................................... 66
Study Protocol............................................................................. 68
Definition of Variables .................................................. 71
Questionnaire Design............................................................. 74
Procedure ........................................................ 76
V
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Data Collection Approach ............................................. 78
Analytical Techniques ............................................. 79
Limitations of the S tudy............................................... 81
Threats to External Validity.................................................. 81
Threats to Internal Validity .......... 83
VI. DATA ............................................................................................. 87
The Study Locations.................................................................... 87
Rock Island Arsenal.............................................................. 89
Red River Army Depot ........................................................ 90
Anniston Army Depot .......................................................... 91
Letterkeimy Army D epot...................................................... 91
Study Respondents...................................................................... 93
Organizational Structures ...................................................... 96
Rock Island Arsenal ............................................................ 98
Red River Army Depot ........................................................ 106
Anniston Army Depot .......................................................... 108
Letterkenny Army D epot...................................................... I ll
Questionnaire D ata.................... 114
General Environment...................... 115
Satisfaction with Current Structure .............. 116
Variables............................................................................... 119
Asset Specificity................................................... 120
Goal Alignment ............................................................ 122
Moral H azard..................................................................... 126
Information Asymmetry .................................................... 129
Transaction C osts.................... 131
Incentives and Coercives.................................................... 135
VII. ANALYSIS ............................................................................... 138
Background Information ...................................................... 138
Pattern Matching ................................................................... 140
Organizational Pattern Matching ......................................... 141
Patterns in Variables ............................................................ 144
Explanation-Building ......................................... 148
Organizational Patterns .................................................... 149
Patterns in Variables ............................................................ 152
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VIII FINDINGS AND CONCLUSIONS............................ 159
Findings ......................................... 159
Does the Principal Agent Relationship Apply to
Organizations where Government lias Implemented
Contractual Arrangements for Commercial Activities? . . . 160
Do the Participants Understand the Inherent Forces that
Detract from an Optimal Efficient Organization?.............. 162
Can Common Trends of Coping with these Forces Be
Found that Show the Principals are Managing the
Organization for Optimum Efficiency? ............................. 164
Conclusions ................................................................................. 167
Applicability to Theory and Implications to Broader
Areas ...................... 167
Applicability to Other Sites ............................... 169
Limitations of the Study—Limitations of the D a ta ............... 170
External Threats to Validity ....................................... 170
Internal Threats to Validity ........................ 172
Other Areas for Future Research ................................... 174
SELECTED BIBLIOGRAPHY........................... 176
APPENDICES
1 . Acronyms................... 183
2. Questionnaire............................................................................... 185
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
1. Idealized Comparison of Traditional and Economically Competitive
Agencies......................................................................................... 36
2. Internal and Externa! Threats to Study Validity ...................................... 86
3. Study Location Characteristics................................................................. 93
4. Data on the Variables................................................................................ 120
5. Pattern Matching of Organizational Structures........................................ 141
6. Pattern Matching of Variables...................................................... 145
7. Explanation-Building of Organizational Structures.................................. 152
8. Explanation-Building of Patterns in Variables . ...................................... 158
9. Threats to Study Validity.......................................................................... 170
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF FIGURES
1. Organizational Structures........................................................................ 8
2. Principal Agent M odel.................................................. 20
3. A-76 Study Process Model .............................. 48
4. Research Methodology Approaches....................................................... 64
5. Research Methodology Approach ......................................................... 67
6. Explanation of Theory Approach...................................................... 71
7. Traditional Command Organizational Structure.................................... 96
8. Organizational Hierarchy within the Traditional Command
Structure.......................................................................... 98
9. Rock Island Arsenal First Generation Principal Agent
Organization ........................................................................... 99
10. Rock Island Arsenal First Generation Lines of .Responsibility and
Accountability................................................................................. 100
11. Rock Island Arsenal Second Generation Principal Agent
Organization ................................................................................... 103
12. Rock island Arsenal Third Generation Principal Agent
Organization ................................................................................... 105
13. Rock Island Arsenal Fourth Generation Principal Agent
Organization .................... 105
14. Red River Depot Information Technology Organization....................... 107
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15. Anniston Depot Information Technology Organization.......................... 109
16. Anniston Public Works Organization ............... 110
17. Letterkenny Army Depot .................................................. 112
18. Letterkenny Depot Organization ..................................................... 114
19. Time Continuum of Principal Agent Organization at Study
Locations ................ 140
20. Conclusion ........................................ 166
X
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
This research focuses on the increasing shift to organizations structured
under the principal agent model within the U.S. Department of Defense. The use of
the Office of Management and Budget’s Circular A-76 competitions to determine
who will provide government services is causing management and organizations to
shift from a traditional bureaucratic structure to the principal agent structure. The
principal agent model can operate regardless of whether the agent is a governmental
organization or a private sector contractor. The research examined if those engaged
in the A-76 process understood the principal agent model, it’s variables and if it
was consistently applied regardless of the agent organization.
Three research questions guided this study. They are: (1) Does the principal
agent relationship apply to organizations where government has implemented
contractual arrangements for commercial activities? (2) Do the participants
understand the inherent forces that detract from an optimal efficient organization?
and (3) Can common trends of coping with these forces be found that show the
principals are managing the organization, for optimum efficiency?
The results of the research showed that, when the principal organizations
had a private sector firm as the agent, the model, was applied consistently across all
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
variables except the use of incentives and coercives. However, when the agent was
an in-house government organization, the organizations deviated in implementation
of the expected organizational structures and also in the way they handled the
variables associated with the principal agent model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER I
INTRODUCTION
It is the objective of administrative study to discover, first,
what government can properly and saccessfuUy do, and,
secondly, how it can do these proper things with the
utmost possible efficiency and at the least possible cost
either of money or of energy. (Woodrow Wilson, 1887)
Woodrow Wilson set in place the direction of public administration 115 years
ago that government should do the “proper things”; and, just as importantly, it should
do these with the “utmost efficiency and at the least cost.” Ever since then, there
have been questions about how to attain “utmost efficiency.” Despite the perception
of a vast sea of federal bureaucrats, the reality is that government is a mix of civil
servants and contractor employees working together. This is an evolution from when
government employees provided virtually everything the organization required and
was vertically integrated. At one time, government employees worked in semi-
isolation from private-sector firms providing internal services and external outputs
for public consumption. Janitors, plumbers, sewer plant operators, locksmiths,
telephone operators and a host of others employed within civil service worked side
by side with analysts, computer specialists, scientists, procurement professionals and
program managers.
I
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At military installations, which operated as small cities, tills self-sufficiency
and desire to provide all manner of sendees manifested itself to a greater degree than
in other federal agencies. Military installations were often in isolated locations and
did not have ready access to commercial services and private industry suppliers.
However, as the U.S. entered the last quarter of the 20s* Century, the economy
developed a robust infrastructure; installations were not as isolated; and industry' was
eager to provide many of the services being provided by government employees.
When the Vietnam conflict ended, the federal government began to depend more on
the private sector for basic services. In the name of efficiency, the first round of
outsourcing occurred in the mid-1970s. Many of the functions that went to the
private sector were the most basic jobs—stock clerks at the commissary, janitorial
services, and phone operators. More complicated functions, such as facilities
management, data processing, health services and education, were retained in-house
with government employees providing these services.
The Growth in Government and Outsourcing
The use of private-sector firms is longstanding government policy. In 1955,
the Eisenhower administration issued Bureau of the Budget’s Bulletin 55-4, stating:
“The Federal government will not start or carry on any commercial activity or
provide a service or product for its own use if such a product or service can be
procured from private enterprise through ordinary business channels.”
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The Office of Management and Budget (OMB) formalized this policy in OMB
circular A-76. It said:
In a democratic free enterprise system, the Government should
not compete with its citizens. The private enterprise system,
characterized by individual freedom and initiative, is the primary
source of national economic strength, In recognition of this
principle, it has been and continues to be the general policy of the
Government to rely on competitive private enterprise to supply
the products and services it needs.
Presidential administrations for the past forty-five years have sought to provide a
government that is cost-efficient and supports the public policy of using the private
sector to provide commercially available services.
The exact point in time when the salaries of federal government employees
dropped below what was spent for contractor labor is debatable. By 1980, the cost of
the federal government’s contracted work force exceeded that of the federal
government civilian work force by nearly $20 billion (Goodsell, 1983). In 1994,
OMB reported the cost of contractor services exceeded the federal payroll by $30
billion. Simultaneously, the overall cost of government grew from under $100
billion per year in 1939 to over Si trillion in 1993, measured in 1986 constant dollars
(Rosen, 1995). This growth has been in all areas; defense, education and the
multitude of social services Americans demand. As the number, scope and size of
federal programs has grown, the staff of federal employees has remained relatively
constant (Guttman & Barry, 1976). Government funds these programs but does not
usually provide direct service delivery with civil servants.
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Along with the increase in funding has been a dramatic increase in the use of
contractors. Numerous studies by the General Accounting Office on the cost
effectiveness of using contractors or government employees have concluded that data
is often unreliable and unavailable (General Accounting Office, 1990, 2000). There
is little doubt that, in certain areas, especially low skill functions, the benefits and
pay scales of civil servants exceed those of the private sector; and there are cost
savings in moving this work to contractor performance. Many of these savings were
reaped is the early rounds of outsourcing. However, with the large-scale downsizing
of the 1990s and the impacts of technology on work processes, the government was
no longer able to clearly relate savings in labor to outsourcing efforts alone (General
Accounting Office, 1997).
When government changes to a competitive process to determine who
provides its services, it is no longer a monolithic organizational structure operating as
a single integrated organization. The organization splits into the oversight function
that is the principal and the performing function that is the agent. The principal must
remain government employees; this is the decision making function. The agent can
be a private-sector firm, or it can be an in-house government organization. In trying
to become more efficient and reduce the total cost of government, a different
organizational approach is being implemented—a contract based organizational
structure. When the contract is the basic unit of an organizational relationship, the
principal agent model is in place. Moving towards a principal agent organization has
been ongoing for over 30 years.
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The popular press and trade journals are filled with editorials and opinions on
the potential savings and the need to outsource more work to the private sector.
Under the Clinton Administration, several senior-level groups recommended
increased outsourcing. The Outsourcing and Privatization Integrated Product Team,
headed by then Deputy Secretary of Defense John White, proposed turning work in
the seven areas of base commercial activities, depot maintenance, materiel
management, education and training, finance and accounting, data centers and
military housing over to the private sector (Donnelly, 1995). The Defense Science
Board Task Force on Privatization, drawn from the military and from private
industry, recommended moving similar work to the private sector (Erlich, 1996).
The presumption is that the private sector is more efficient, even though government
organizations win approximately 50% of the competitions (General Accounting
Office, 2000). It is the competitive process, not who provides the services, that
creates the opportunity for gaining efficiencies (Center for Naval Analysis, 1996).
The competition process sets in place the principal agent organizational model,
regardless of whether a private-sector firm or an in-house government organization
provides the services.
There is little doubt that the outputs of government will continue to be
provided by both private-sector contractors and civil servants. Many of the jobs
required to run military installation functions are referred to as “Commercial
Activities,” because they are commonly found in the commercial sector. These were
identified in the A-76 circular and include equipment maintenance and repair, utility
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
operations, supply operations, information technology, graphic arts and mailroom
operations, property maintenance, education and training center operations, and
engineering services. The specific job list is exhaustive and cuts across all
departments within the Executive Branch. It is these Commercial Activities that are
being extensively competed. The 1998 Federal Activities Inventory Reform (FAIR)
Act required federal agencies to develop an inventory of all jobs that could be
performed by the private sector. Within the Department of Defense (BOD), 504,000
out of a potential 662,900 jobs are included in that inventory. The DOD has the
largest number of employees and is the most actively involved of all the Executive
Branch Departments in Commercial Activity studies or the general efforts of
“Outsourcing.” Almost 112,000 jobs are currently under review, and another
118,000 jobs will be studied under the A-76 process within the next five years (Friel,
2001). No one knows for sure just how many contracted employees work for the
DOD today, although it is definitely larger than the number of federal employees and
will continue to grow (Nakashima, 2001).
The DOD conducted 1,661 cost comparison studies under A-76 between
October 1978 and December 1986. When the study results were implemented, the
estimated savings of $612,557,000 were reported. Forty-eight percent of the studies
determined government employees should continue performing the work, and 52% of
the time the work went to the private sector (General Accounting Office, 1988b).
With the end of the Cold War and the shift in Presidential priorities, there was
constant pressure to reduce Defense spending. A series of commissions from the
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1983 President’s Private Sector Survey on Cost Control, known as the Grace
Commission, to the 1995 Defense Commission on Roles and Missions supported the
idea that government was inefficient and that the market’s competitive forces could
reduce the cost of government.
President Clinton, under his 1992 National Performance Review- (NPR) drive
to improve the efficiency of government, proposed the reduction of 272,000 federal
employees. In 1995, DOD directed the Military Services to conduct additional
competitions under A-76 (General Accounting Office, 1999). The 1997 Quadrennial
Defense Review study assumed a potential 20% cost savings in all contractible
functions. The assumption was that, whether the government or the private sector is
the producer, the cost would be 20% less. Efficiency and cost savings were the
drivers, and the effectiveness of the ensuing organizational changes is never
mentioned.
Principal Agent Paradigm
The procedure used to determine who provides government services divides
the macro level organization into two distinct camps—those under a form of contract
and those overseeing the contract. This is the classical principal agent relationship.
In some cases, the agent may be a private-sector firm providing the services; and, in
other cases, government employees continue delivering the services and outputs.
The principal establishes the goals regardless of who is the agent. This is a very
different structure, when compared to the traditional monolithic, vertically integrated
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
organization of government. Figure 1 depicts this difference. When the principal
agent structure is adopted, the principal clearly identifies what outputs are expected
and then measures the quantity and quality of outputs. The agent is expected to
perform to an acceptable level of quality and quantity, regardless of whether it is an
in-house governmental organization or a private-sector contractor.
The military and the Department of Defense are built upon a concept of
integrated organizations under a unified command and control structure. For the
most part, this concept forms the basis for the civilian organizational structures
supporting the military forces. In comparison, other federal, state and local
governments have been built upon a concept of utilizing contractors for carrying out
many of their primary and secondary functions. Moving functions from government
Traditional /-Principal Agent
■ P rin cip a l
> 3
Agent
Agent
Figure 1. Organizational Structures
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
employees to contractor employees has occurred over the past decades. Today, we
take for granted that janitors, computer repairers, telephone service people, and
cafeteria workers are commonly not government employees. Two and three decades
ago, government employees performed these functions for their organizations.
Many functions are classified as “inherently governmental.” These are work
processes, policymaking and decision-making that cannot or should not be contracted
out. Where this line should be drawn as to what is or is not inherently governmental
is not always clear. For example, functions, such as highway construction or delivery
of mental health care, are considered a government responsibility but have
traditionally been secured from the private sector. Writing a speech may not be
inherently governmental, unless it addresses or sets policy. Contracting decisions are
inherently governmental, but writing the scope of work and evaluating contractor
performance are not. Maintaining government finances may or may not be
governmental; that seems to be up for debate.
As this trend in using competition to decide whether contractor or
government employees will perform its functions continues evolving, government
must shift management styles to accommodate this different organizational approach.
The principal agent model is the evolving underlying management approach. In his
article “Contracting at the Core,” Steve Kelman (2001) talks about the need for
government contracting offices to focus on this new way of doing business where
they contract for support, rather than physical things. To highlight the size of what is
occurring, he states that Defense, Energy and NASA spend a majority of their funds
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on contracted products and services. The Department of Energy actually contracts
94% of its budget.
The evidence points to a shift in government to where the principal agent
model is beginning to predominate. This change to a principal agent organizational
model requires a change in management approaches, regardless of whether the
government organization or a private-sector firm is the agent. This requires an
understanding of information asymmetry, moral hazard, incentive structures and
asset specificity. The purpose of this research is to determine if organizations are
implementing and operating under the principal agent model, as theory would
predict.
Relationship to Larger Trends in Government
Victor Heyman, in his 1961 article “Government by Contract: Boon or
Boner?” toid us: “Much is known about problems of administering the regular
government agencies, but very little is known about administering the large numbers
of people and activities brought into the federal service under contract.” This
continues to hold true today, and issues of control and monitoring by the principal
need to be understood. Principal agent structures create a need to manage transaction
costs and develop an incentive system that encourages increased efficiencies and
avoids the pitfalls of shirking and moral hazard.
This increasing use of principal agent organizations relates to greater trends
in government and the shift towards the “Hollow State” (Miiiward, 1996). The
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
“Hollow State” refers to a government that provides funding, but relies on contracts
with private firms and non-profit organizations to actually provide services. The
debate continues as to how much government should rely on the private sector as it
works for the betterment of society. The question is: How far should government
head towards a “Hollow State”? It is a question of whether government agencies
have adopted a management approach that best enables them to efficiently operate
under the principal agent organizational model.
Study Objective
This study examines the principal agent organizational relationships where
commercial activities competitions have been held. It will look at two populations:
(1) where a private-sector firm under contract is the agent and (2) where government
employees are the agent. It will delineate the principal agent model (depicted in
Figure 2 in Chapter 2) and compare it to what is found in the organizations. It will
determine if there are variations of the principal agent model and if these have
positive or negative implications. Principal agent theory states that issues, such as
shirking, moral hazard and transaction costs, may impede the economic efficiencies
of operation. The study will examine these and determine if they exist, to what
extent they exist and the impact they have.
There have been many studies performed by the General Accounting Office
and others on the outsourcing initiatives of government. There are large numbers of
articles and opinions on both sides of the value of the A-76 Commercial Activities
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
program. The savings are often disputed numbers, but the trend Is to expand the use
of the principal agent approach within government. The value of this study is to
analyze how principal agent theory has been put into practice at several locations. By
examining the variables of principal agent theory' and the results of implementation
under different conditions, this organization level analysis will give insight
unavailable in today’s literature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER II
LITERATURE REVIEW
There is nothing so useful as a good theory. (Kurt Lewin,
1935)
The literature review draws together a diverse collection of authors. The
review begins with the writings on the Principal Agent Theory, which developed the
theory and where the theory applies. Principal agent theory has gained acceptance in
only the last 25 years. The literature is not as extensive as for other theories, and it
has primarily been applied to the corporate world. This study extends the theory into
government organizational structures. The second section of the literature review
examines the macro level trend towards a reliance on the private-sector support for
delivery of government outputs. There has been extensive writing on this topic for
several decades. The third section focuses on competition theory literature and how
this impacts government organizations. The fourth section reviews the studies and
analyses of competing government functions. There have been many studies of the
overall A-76 program that have analyzed the macro level results of the program and
at specific locations where competitions were conducted. The outcome of the A-76
competition is an organization built on the principal agent model. The literature
review will pull together the macro level trends of competing government services
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with the writings on A-76 and how government is evolving to principal agent
organizational structures.
Principal Agent Theory Writings
The concept of principal agent theory began with the organizational structure
of the firm. How does the owner or stockholder organize so that employees work in
the interest of the owners? In the case of the federal government, the stockholders
are citizens, Congress and the current administration. Early corporate organizations
were based on bureaucratic models or a clan approach where the family ran it. As
corporate organizations grew and became more complex, absent ownership through
stock became the norm. The challenge was to structure the firm organizationally so
that it was economically efficient for the stakeholders—owners, managers and
workers. The writings of Chester Bernard and Ronald Coase in the 1930s set the
basic foundation for principal agent theory, although they did not develop the theory.
They looked at the structure of the firm and roles of the executive and tried to
establish how to best align their actions with the goals of the stockholder owners.
Principal Agent Theory
S. A. Ross—Economic Theory of Agency
In his 1973 article, “The Economic Theory of Agency: The Principal
Problem,” S. A. Ross originated the modern principal agent theory. He approached
the idea of agency from an economist’s point of view in developing equations to
reach Pareto efficiency. His intent was to find the Pareto efficient situation where an
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
alternative is equally preferred by all over alternatives, and at least one stakeholder
prefers it above the other alternatives. Ross viewed the firm as a monolithic entity.
The basic problem was in opening up the Mack box that was the firm; he finds
conflicting goals between shareholders, managers and workers, all wanting a larger
share of the surplus generated by the firm. Ross attempts to develop a series of
equations that lead to the Pareto efficient solution. When only limited information is
available, it is almost impossible to develop the equations concerning utility
functions, states of the world and competitive markets that lead to a Pareto efficient
structure. The key question of how to devise operating arrangements and incentive
systems that yield maximum output is very complex. Ross developed the basic
concepts and wrote the first of the economic-based articles developing complex
equations to explain the principal agent theory. The difficulty in quantifying the
variables makes application of the equations to individual situations almost
impossible.
K. Eisenhardt—Control as Organizational Determinant
One of the leading authors on principal agency theory is Kathleen Eisenhardt.
Her 1985 article, “Control: Organizational and Economic Approaches,” provides a
clear comparison between the traditional integrated organization and the principal
agent model. Her determinant on the most appropriate type of organization is the
ability to measure outcomes or behaviors. She compares agency-styled organizations
to more traditional integrated organizations. Her research was focused on incentives
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and control mechanisms for smaller retail chain stores. Eisenhardt found that the
organizational structure should he driven by ability to design control mechanisms and
to define the desired results. Behavior- versus outcome-based results and the ability
to have adequate measurement systems determine the appropriate organization. “The
agency problem is to determine the optimal contract for the agent’s service.” She
found four primary differences between a principal agent and the traditional
hierarchical organizational approach. First is that, under principal agent, the reward
structure is explicitly tied to performance; and, under an organizational approach, it
is more implicit and less directly tied to performance. Secondly, the principal agent
focuses on costs and the cost of measurement; while, under an organizational
approach, the emphasis is on ease of measurement. Thirdly, a principal agent
assumes a divergence of preferences, while the organizational approach assumes a
convergence. Lastly, under the organizational approach, information is viewed as
available at only one level of detail; under principal agent, it is viewed as having
variations in the detail of information, but all at different levels of cost.
J. Pratt and R. Zeckhauser—An Overview
In 1985, John Pratt and Richard Zeckhauser compiled a series of monographs
in their book, “Principals and Agents: An Overview.” This was the one of the first
and largest collections of writings on the subject of agency theory (agency theory and
principal agent theory are interchangeable) and resulted from a conference held at the
Harvard Business School. Their view is: “The agency relationship is a pervasive fact
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of economic life . . . at least one significant component of almost all transactions.”
They attempt to simplify the theory and identify the underlying concepts of
information, incentives and agency structure. Their discussion on information
asymmetry and the need for monitoring lays a framework for why there are different
variations of principal agent. When the principal has access to performance data and
market data, the need for monitoring decreases, and understanding the results of the
agent’s actions increases. Similarly, the incentives developed for a principal agent
structure are directly related to the information available. For a simple example,
when the school is the principal and the teacher is the agent, should the reward be
based on seniority, the number of students taught or some measure of the quality of
teaching? The level of information between the principal and the agent drives the
structure of the relationship. External influences, such as the resources the teacher
has in her classroom and the children’s home environment, need to be considered.
This is a simple example of the complexities of the principal agent approach. The
book consists of eight chapters that cover the spectrum of principal agent issues;
however, many are isolated to specific industrial sectors.
Within Pratt and Zeckhauser (1985), two chapters have particular interest.
The first is Kenneth Arrow’s “The Economics of Agency.” Similar to Ross, Arrow
looks to economics to describe the principal agent relationship. However, when he
factors in situations where there is information asymmetry—what he called hidden
information—and hidden action—otherwise called moral hazard—the economic
complexities expand considerably. Developing a single set of useful equations
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
becomes extremely complex. “Contractual relations are frequently a good deal more
complicated than the simple models of exchange of commodities and sendees at
fixed prices would suggest.. . . Most importantly, the theory tends to lead to very
complex fee functions.” Arrow concludes that the economic complexities of the
principal agent relationships drive the contractual mechanism to follow much more
simple structures. Contracts between the principal and the agent follow more custom
than true economics. The contract is not related to the desired outcome on a Pareto
efficient basis, but often follows a common practice. For example, the way the
principal (the patient) pays the agent (the physician) is in no way related to the
outcome.
The second chapter of interest is Harrison White’s “Agency as Control.”
White, a sociologist, explains agency theory as one group’s extension of control over
another. Where Ross and Arrow focus on the economic and financial control
structures, White looks to interpersonal and interorganizational relationships to
explain agency. His examples of the Roman Empire and the early Christian Church
to explain the control of a principal over the actions of an agent brings a unique
perspective not offered by other authors. White suggests that the development of
personal ties and dependencies can be a strong factor in agency and control
mechanisms.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
K. Eisenhardt—Review of Principal Agent Theory-
Four years after her first article on agency, Kathleen Eisenhardt (1989)
summarized the literature of Principal Agent Theory. Various authors had related the
theory to the fields of accounting, economics, finance, marketing, political science,
organizational behavior and sociology. The heart of principal agent theory is the use
of the contract as the means of control Principal agent assumes both participants
have an independent agenda, motive and desired outcome. The ability to align the
actions of the agent with the goals of the principal will determine the efficiency and
effectiveness of the structure. Eisenhardt answers the question of what comprises
principal agent theory, establishes it is a valid theory and defines where can it be
applied. Figure 2 summarizes graphically how principal agent theory works.
The theory applies when the organizational relationship has an agent with
defined outputs and the principal in a control position. It applies in most business
arrangements and under A-76. It is the formal performance requirements—most
often the contract—which is the basis for principal agent theory. The variables of
goal alignment, asset specificity, etc., identified under principal agent, influence the
success of the organizational structure.
Eisenhardt ties together the theoretical writings on principal agent and
compares these back to organizational theory authors. She highlights the value of
looking at agency theory from both an economic and an organizational perspective to
understand the dynamics involved. She concludes that one must look at the principal
1 9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
agent from varying perspectives to understand what is occurring and suggests
additional research from an organizational view.
lafiwmaag Varkbies
.Infonnation
Asymmetry
Ineenlives &
Coerdves
Ifaasacioa
Costs
Figure 2. Principal Agent Model
Principal Agent Applied to Government
D. Good and G. Miller—Principal Agent Applied to the Public Sector
Most literature on principal agent focuses on private-sector firms and covers a
multitude of industries from advertising to manufacturing. David Good (1992)
extended the research into the public sector looking at public transit agencies. He
found two distinct approaches: (1) an organizational model, where management was
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
integrated within the organization, and (2) a principal agent model where the transit
agency management was contracted. His results showed a consistent 2.5% efficiency
gain, when the principal agent model was employed.
Gary Miller (2000) also extended the theory to the public sector. He focuses
on the issue of whether a principal agent relationship is appropriate in public
agencies. The true principal in government is the citizenry, but it is totally separated
from the agent. The traditional government bureaucracy stands in as the principal’s
representative. The principal agent relationship sub-optimizes the organization,
because it creates so many players with differing agendas and goals. For example,
the agent desires to maximize profit, while bureaucracy may want to expand the
service levels and its base of power. The citizen may want a low cost provision of
services, yet that goal may never be made clear. This is different from business,
where the stockholder principal has one primary goal—financial gain. Miller sees a
lack of credible commitment by the bureaucrat to the goals of the principal. His
recommended solution reaches back to the Weber theory of a bureaucracy where the
“professional elite,” such as the Forest Service, Public Health Service or Department
of Transportation engineers, direct the work of the organization and reflect what is in
the best interests of the citizenry. These professionals’ goals are most closely aligned
with those of the principals they serve, while being insulated from many of the
influences of a political environment.
Good and Miller both examined devising incentive structures for
organizations producing public goods. The literature is at the theoretical level,
2 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
applying the macro level principal agent theory to government organizational
structures. Good concludes that the incentives within the organizational structures of
both private- and public-sector managers encourages them to engage is non-cost-
minimizing behaviors.
S. Globerman and A. Vining—Application in Info Technology
Globerman and Vining (1996) provide taxonomy for evaluating the
contracting of government information management organizations. They set the
variables as asset specificity, contestability, and task complexity. The complexity of
these three variables should drive whether the function is appropriate to outsource
and operate under the principal agent model Ideally, the goals of the agent and
principal are similar, limiting any goal alignment problems. If the priorities are
aligned, there are little bargaining or transaction costs for prioritizing and
accomplishing work. The downside is that the agent can become a monopoly
provider. Inefficiencies are not penalized by the marketplace, and the principal agent
problems may be increased. While Globerman and Vining apply this specifically to
the information technology environment, it is applicable to most other government
functions.
These authors provide the origins and an overview of principal agent theory.
Ross defined the theory in mathematical concepts; Eisenhardt applied it in the retail
industry; and Good applied it to the public sector. The authors showed how principal
agent sets in place a unique organizational approach. Several authors have applied
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the theory to government organizations in transportation and information technology
areas. When the private sector provides government services, it is under a
contractual arrangement. When a government organization acts as the agent, the
contract is replaced with regulations, a performance work statement, and
memorandum of agreement or verbal instructions. The arrangement between
principal and agent should be dependent upon the asset specificity, information
asymmetry .and the level of outcome uncertainty. As the complexity of these
variables increases, the difficulty of managing increases, as do the transaction costs.
This study compares the two structures of either a government organization
or a private-sector firm as the agent. The principal agent model applies, because
there is a defined performance level and an oversight role provided by a principal
organization. This contrasts to when government provides the services as a vertically
integrated producer. This vertical integration is designed to limit transaction costs
and increase organizational efficiencies. The advantages include a faster response
time, less formal contract definition, less bargaining and the absence of a profit
motive.
Evolving Principal Agent in Government
To understand principal agent within government requires an understanding
of the much larger effort to outsource the production and support functions of
government. It is on a worldwide basis and encompasses many different approaches.
There is a great deal written on government as an inefficient producer but very little
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on the principal agent relationship within government, its impact on efficiency, and
how to manage under this organizational model. The focus of this research is on an
organization that is a consumer of products from the private sector but also as a semi-
autonomous producer. Throughout history, the U.S. Army used government
employees in its federal arsenals and plants to produce many of the goods it uses. A
second approach was owning the infrastructure, such as the tank and aircraft plants,
while using private firms to produce the goods. Within its military installations are
commissaries, laundries, vehicle shops, personnel departments, training ranges and
all forms of organizations required to support the soldiers, sailors, airmen and their
families. However, as the United States entered the last quarter of the 20f l l Century, it
was believed that private firms could produce those same outputs at a lower cost.
The literature is replete with writings showing that market-driven functions are
efficient and that government monopolies are inefficient (Niskanen, 1971;
Brocherding, 1977; Savas, 1987). The macro level trend is to use contractual
arrangements under the principal agent structure. At any one installation, that
structure may be in effect dozens of different times as the principal (government)
contracts with a variety of agents for many different functions. Across the entire
U.S. Army and the U.S. government, this principal agent structure is in place
thousands of different times.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Contracting Government Functions
V. Heyman—Concerns from the 1960s
In 1961, Victor Heyman began to question if contracting government
functions was the right thing to do. His Public Administration Review article,
“Government by Contract: Boon or Boner?” looked primarily at research and
development efforts. He felt that, many times, government had to contract, because
they could not offer salaries to reach the best and brightest. He saw advantages in
contracting to universities for basic research and to industry for applied research,
particularly for weapons projects leading to production contracts. When there is a
need for short-term projects, contracting makes infinitely more sense than hiring
permanent employees. Heyman felt, if the government did not retain the expertise
within, it would not have the knowledge base to oversee the contracts; this creates a
condition of information asymmetry. Secondly, he was concerned about cost control.
Private-sector firms’ incentive is profit, so the incentive for cost plus contracts is to
increase the cost base. For other types of contracts, there are incentives focused on
profit maximization more than successful completion; this creates adverse selection.
In adverse selection, the agent chooses a path of action the principal would not
support, except the principal has insufficient knowledge to direct the agent in the
most efficient manner. Heyman had several other concerns, but his focus was the
loss of knowledge and control and uncontrolled cost growth. Heyman concluded
there is no simple formula to determine when contracting is better than providing the
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
service with government employees. Rather, farther stud}/ is needed on how best to
manage this new' paradigm.
E. S. Savas—Expansion of Contracting
E. S. Savas (1982) quotes General Accounting Office studies, where savings
have been achieved and quality has improved through competition and moving the
production to the private-sector firm. He cites as examples laundry services at
military bases, water supply and waste refuse handling at the municipal level, and
more traditional government functions, such as weather forecasting, education and
social services. Savas proposes going beyond the concept of outsourcing and stresses
allowing the marketplace to directly provide many services. His examples of waste
hauling, weather forecasting, forest management, water supply, transportation and
health care all show improvement in quality and decreases in cost, when they are
moved from government to the private sector. Savas looks at the function in terms of
exclusion and consumption and marketplace availability for providing the service.
His belief is that, through competition, an efficient structure can be established. He
is a strong proponent of the principal agent model across all levels of government.
He supports expansion of contracting to non-traditional areas, such as education and
fire fighting. Savas acknowledges the problems of corruption, excessive profits,
higher transaction costs, limited flexibility and other shortcomings, but he believes
the benefits outweigh the potential problems.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I Sharkansky—Pros and Cons
Ira Sharkansky (1980) states that the use of private sector providers as
contractors to do government work is a distinctly American approach. He refers to
contracted functions as being on the margins of government as opposed to being a
core function. He details the advantages of rising contractors, lower cost structure,
weakened power of government employee unions, reduced facility ownership
expenses and reduced long term commitment. A disadvantage is, in markets with
limited competition, the potential exists for greater costs than when government
organizations perform the work. Secondly, there is the lack of accountability and
control, when work is contracted to a private-sector firm. These concerns are similar
to the information asymmetry and adverse selection problems that Victor Heyman
identified 19 years earlier. Sharkansky quotes studies that conclude that there is very
little data or evaluation of the costs, quality, flexibility or effectiveness of using
contractors. The point of Ms writing is that the delivery of government services is in
a transition mode to delivery by contractors. TMs is the transition to the principal
agent model. What Sharkansky foresaw in 1980 continued to expand over the next
two decades.
D. Ketti—Pros and Cons
Where Savas has evidence that shrinking government results in a better and
more efficient society, Donald Ketti (1993) acknowledges both, the advantages and
the disadvantages of government supported by large groups of contractors. He brings
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to light the management weakness of using contractor support and the potential for
greater waste and abuse of taxpayer funds. He highlights the fact that, when GAO or
OMB looks to programs with significant problems, 80% involve contracted
operations. The problems often result from poorly specified contracts and weak
oversight by management—factors that lead to problems in any principal agent
arrangement. Ketti identifies asymmetrical information, shirking, lack of asset
specificity and the ex post costs of monitoring to explain the problems government
encounters, when it relies on contractors.
Ketti brings to light the problems of a public market versus a private market.
If a firm does 98% of its business with the government, is it really a private-sector
firm? If there is only a single producer for an item (defense weapons, as an
example), is there really a competitive market? Ketti identifies both supply-side and
demand-side imperfections in the public market. On the supply side, lack of the
existence of a market, lack of competition in the marketplace and externalities can
cause problems. Often there is no market except the government, as in nuclear
weapons production. The government is the marketplace, because there is no other
demand for the product; you have a single buyer and a very limited number of sellers.
Finally, externalities, such as environmental pollution caused by private-sector firms
doing government business, are passed off; because responsibility goes back to the
government. On the demand side, asset specificity and information asymmetry are
problems. In many cases, government agencies cannot clearly define the product, as
in research and development programs, to enable competitive market development.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
No one knows exactly what it will take to clean up a toxic waste dump, and this
creates more of a collaborative partnership than a contractual arrangement. Without
an arms-Iength relationship, the competitive market breaks down. If there is limited
or asym m etrical information and the agent is more knowledgeable than the principal,
the competitive forces fall apart, and the agent becomes dominant. Ketti identifies
the primary pitfalls of principal agent and the need for the government managers to
have a thorough knowledge of how to operate under this organizational structure.
H. Stretton and L. Orchard—Overview
Hugh Stretton and Lionel Orchard (1994) provide a broad theory-based text
on the provision of public services. They Integrate several different theories Into a
coherent understanding of the forces affecting the provision of government services.
Beginning with theories on monopolistic producers, they Incorporate the work of
Mancur Olsen’s “The Rise and Decline of Nations” on group behavior and the role of
Incentives for efficient production. Their book focuses on the national economy
level, but many of the thoughts are applicable at the organizational level. They
integrate economics and social science to support their theory and propose that the
economy will work best with a mix of public and private provisions of sendees.
Those in the “for profit” sector may be better at providing many of traditional
government functions using private-sector processes and value structures. Other
functions require a different set of values, commitment and environment that comes
with government providing services. Stretton and Orchard conclude that it is critical,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
when designing the organization to provide government services, that economic
theory be integrated with organizational theory.
R. Moe—Legalistic Concerns
Ronald Moe (1987) was with the Congressional Research Service, when he
wrote that privatization is the single most influential concept in public administration
of the 1980s. His article, “Exploring the Limits of Privatization,” looks at the
movement of work from the governmental sector to the private sector. The
“Privatization Movement,” as Moe refers to it, has its origins in the anti-government
sentiment, and there are no clear bounds where it ends. The authors Moe cites
(Savas, 1982; Butler, 1985; Hanke, 1985), who support this movement, view the
private and public sectors as being very much alike and do not identify what
differentiates the two sectors. Moe looks for a legal basis to clearly define what
should or should not be performed by private-sector employees. He stresses the need
for an underlying theory that provides direction on what can be contracted and what
should be retained for public-sector performance. His secondary concern is the
ability of government to manage contractor performance.
The assignment of functions must take into consideration the need
for government to attract and retain officers and employees with
sufficient knowledge and training to properly contract for,
monitor, evaluate, and use the highly technical products from the
private sector. Government, in order to avoid being “captured” by
contractors must have a skilled cadre of technicians and
managers.
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Moe cites problems, such as those between NASA and Thiokol during the
Challenger disaster, as an example of over-contracting and in trying to define who is
liable and accountable, when things go wrong. Moe is the only author who goes to
the legalistic roots of public administration for direction on what is properly
performed by government employees versus the private sector. He does not look to
economics or public sentiment for direction, but he looks to a legal basis. He sees
this paradigm shift as a challenge for the entire field of public administration and
concludes that the leaders have not provided direction to the practitioners.
I. Kirkpatrick and M. Lucio—Public Choice Theory
This shift to a principal agent model is international in scope. Ian Kirkpatrick
and Miguel Lucio (1996) write of the structural factors and the public choice
theorists driving this change. As the demands for government services grew after
World War II, particularly for social services, the tax revenues failed to keep up. As
taxes increased, the pressures to decrease the cost side of government continued to
rise. Public choice theorists argued that bureaucracies were inherently inefficient,
and the only way to lower costs was to allow competition and the market to select the
provider.
Whilst there can be no doubt that various preconditions make
market relations more efficient than bureaucratic approaches, the
overall thrust of the New Right public choice theory was (and is)
that, overall, the former are likely to lead to more efficient, better
quality public services.
3 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It was the pressure to deliver a higher level of services while holding taxes in check
that initiated the drive to contract government services. This, in tom, leads
government to a principal agent operation as more services are delivered under
contract.
H. B. Millward—The Hollow' State
In 1996, H. Brinton Millward edited “Symposium on the Hollow State:
Capacity, Control, and Performance in Interorganizational Settings ” This is a
collection of writings about government evolving to funding private-sector firms and
no longer delivering the services it requires. The phrase “hollow state” evolved from
the “hollow corporation” introduced by Business Week magazine in 1986. The
hollow corporation integrated the end product and replaced internal production with
a network of subcontractors. He highlights the flexibility, low fixed cost and smaller
bureaucracy inherent in a hollow organization. He also highlights the potential
problems of negotiating, coordinating, monitoring, writing and enforcing contracts,
information asymmetry, and monopoly practices. Millward was identifying the
advantages and disadvantages of moving government towards a principal agent
relationship. He highlighted the ex ante and ex post costs without calling them by
this taxonomy. Millward envisioned a time when most of government operated
under the principal agent model, and only the principal remained as government
employees.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Privatization
The term “privatization” is often associated with the principal agent model.
The definitions of privatization vary by author; but, of the three basic approaches,
only one incorporates the principal agent model. Calvin Kent (1987) defines
privatization as eliminating government as the provider of goods and services,
replacing this with private-sector firms and creating a market. One example is when
local government stops collecting refuse and turns it over to private firms. Stephen
Moore and Stuart Butler (1988) add asset sales as a second definition of
privatization. This has been popular in Britain and New Zealand with the sale of
coal and steel industries, airlines, telecommunications and eves national forests. The
government gets a monetary return, and the private sector now provides the industry.
In the United States, there have been proposals to privatize National Airport, the
Naval Petroleum Reserves and even the air traffic control system. At the local level,
water supply, tax collection and even elementary education have had trials.
Government involvement through regulation may continue, but government as the
provider ends. The rationale is that private industry can provide these services more
efficiently and with a higher quality.
Privatization has a third definition referred to as “Outsourcing.” Outsourcing
is contracting to the private sector for what was formerly provided by government
employees. This is the most common form of transitioning from a government
provider to a private-sector provider. It is what many Americans think of when they
discuss privatizing government, what Donald Ketti refers to as “government by
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
proxy.” The government finances and controls the level of service, but delivery is by
private firms. Examples include building maintenance, military base management,
research and development analyses and even prison security. Government is
responsible for delivery of the output and does this through contracts. Government
retains responsibility for accomplishment but is not the service provider. This is the
classic principal agent model concept.
Moore and Butler (1988) and others advocate expansion of outsourcing,
focusing on the 11,000 commercial activities performed by federal employees.
Proponents cite many successes, particularly at the municipal level, to support their
position. In many instances, there has been both an improvement in service and a
reduction in costs. Opponents (Gable & Moyer, 1996) cite the lack of cost controls,
the unreliability of firms to deliver due to labor strife, and the lack of success in past
contracting efforts as reasons to retain a governmental work force. Opponents also
cite many examples where service delivery has failed or costs have increased.
Competition Theory
The movement towards outsourcing is championed by those that believe
competition forces organizations to operate at a most efficient level. Competition is
the market mechanism that produces the least cost, highest quality, and innovative
designs and products. The writings extolling the virtues of a market based,
competitive environment are voluminous (Ketti, 1993; Savas, 1982; Osborne &
Gabler, 1992; Office of Management and Budget, 1988).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Competition Writings
J. Rehfoss—The Case for Principal Agent
John Rehfoss (1991), in “A Leaner, Tougher Public Management? Public
Agency Competition with Private Contractors,” summarizes the case for the principal
agent model as the primary way to organize. His rationale is from the public choice
economists who “consider public agencies monopolistic, budget aggrandizers, and
bureaucratic imperialists.” He supports Ms view with the earlier cited literature to
document the expanded use of competition trends at all levels of government.
Rehfoss believes this need for competition is driven by the differing focus of
organizations. Table 1 is from Rehfoss and shows the focus on improving service
and lowering costs of the competitive agency versus the focus on accountability and
stability of the traditional agency. This leads Mm to conclude that government
should continue in competitive efforts and further implementation of the principal
agent model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 1
Idealized Comparison of Traditional and Economically Competitive Agencies
Traditional Agencies mp ■
• Emphasis on Accountability ® Emphasis on Efficiency
• Single Supplier of Services • Many Suppliers of Service
• Hierarchical Focus • Client Focus
• Emphasis on Stability and
Predictability
• Emphasis on Adaptation and
Innovation
* Limited Competition among
Agencies
• Increasing Competition among
Agencies and Private Actors
j ® Emphasis on Policy • Emphasis on Service Provision j j
J. Pack—Summarizing the Theories and Practices
Janet Rothenberg Pack (1987) reviewed the theories and practices covered at
the 1986 University of Pennsylvania conference on “The Privatization of Public
Sector Services.” Much of what she discusses are the premises Savas and others use
for expanding the role of the private sector in providing government services. She
highlights the problems in developing good contractual arrangements and the
possible problems of moral hazard and adverse selection. Pack believes that simple
cost comparison decisions, such as whether a public hospital should do its own
laundry or contract for the service, are uninteresting. Privatization becomes
interesting when issues of equity or merit goods are considered, such as whether
education or health care should be delivered by government or contracted. One of
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the primary differences between using a government or contract provider is what
happens when problems occur—the accountability problem Ronald Moe raised. A
second issue is how will situations be handled when the service must be altered and
who bears the costs. The concepts of a total cost, including contingency and
transaction costs, must be included when deciding to outsource. Janet Pack
questions whether the potential ex post costs have been properly factored into the
decision process.
C. Kent and P. Ketti—'The Case Against
Calvin Kent (1987) develops the opposite side of the argument—the case
against competition and private-sector provision of services. The need to generate
profits, pay taxes, the cost of capital, and other non-income-producing efforts, such
as marketing, should make private-sector firms less competitive. Donald Ketti
(1993) also writes extensively about the potential pitfalls of going to a contract-based
structure. Ketti frames the discussion using the principal agent model. He covers the
issues discussed by other authors and highlights the “Smart Buyer Problem.” Ketti
contends that government agencies (the principal) are often not trained or
knowledgeable enough about the function of the agent to develop the best contractual
document. This puts the principal at such a disadvantage for moral hazard issues that
the transaction never reaches an efficient level. Kent counterbalances this with the
idea that government bureaucrats, not facing competition, are motivated by self-
interest to improve their personal position with expanded power or additional salary.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Success in the bureaucracy can be measured in budget size and employees, not in
customer satisfaction or low cost. “The very nature of bureaucratic provision
militates against efficiency.”
There is a wide body of literature that provides multiple reasons on both sides
of the argument for and against contracting of government services. The principal
agent structure results from the competitive process and, hopefully, provides the
service delivery system that offers the greatest benefit to the citizenry.
Competition in Commercial Activities
The commercial activity outputs produced for government are not the
classical “public goods.” They are private goods consumed by a public organization.
They meet the tests of rivalry in consumption and excludability in ownership. These
are goods and services that are available from the private sector and for which a
market exists. Competition and implementing the principal agent model are
expected to produce services of an equal quality and lower cost. The ability to reach
the efficient cost level is dependent upon such factors as asset specificity, market
position and dominance of the current provider, the size and barriers of entering the
market and the frequency of purchases.
The literature supporting the private sector providing government services
more efficiently is very broad. Competition brings the principal agent structure into
place and brings the benefits of efficiency. Savas (1982) is one of the more
outspoken authors on the benefits of outsourcing. He looks to several functions
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where the benefits of competition have been shown and cites examples in refuse
collection, power generation, transportation, social services, and fire protection.
Studies of municipal refuse collection show cost reductions in Phoenix and
Minneapolis of 30% or more. The key to a more efficient operation is not who
provides the service, because quite often city employees were more efficient, but that
the competition determines the provider. This competition brings the principal agent
model into place.
Stretton and Orchard (1994) provide a balanced view that shows examples of
the private sector as the more efficient producer, examples of private-sector
shortcomings and also where the public entity is the more efficient producer. The
Office of Management and Budget (1988) summarized the concept as follows:
Competition is the driving force behind quality and economy of
operations in the private sector. Private sector managers are
continually challenged by competitors who may force them out of
business if they do not operate in the most efficient manner. This
constant competitive pressure forces managers to be innovative
and flexible as they promote performance based management to
serve their customers. Government managers, in normal
operations, do not encounter the same pressures for efficiency that
private sector managers do. They have few baselines for
comparisons and do not face legislative and fiscal constraints that
force them to operate inefficiently. Competition with the private
sector can highlight these inefficiencies and consequently help
identify the changes necessary either to streamline the government
operation or to determine whether the private sector can more
efficiently perform the service.
To have a competitive market requires a number of buyers and sellers and a
contractual relationship. The product must be clearly defined so buyer and seller
understand exactly what is being purchased. In buying commercial products, such as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
office supplies, computers, or vehicles, this is relatively simple. When buying
facility and grounds maintenance for a 20,000-acre military installation, it becomes
difficult because of a lack of clear output, unknown requirements and a less than
folly competitive market. It also requires sustained commitment, oversight and
contract management. The General Accounting Office (1992) testified: “Contractors
are receiving bonuses for mediocre performance and are charging the federal
government for millions of dollars in costs, such as employee parties, tickets for
sporting events, and alcoholic beverages, that are unallowable.. . . ”
Directly Applicable Studies
Over time, there have been many studies and analyses of contracting of
government functions. Some have looked at individual sites, while others have
looked at the issue on a broader scale. Each time a government function is brought
under a contractual arrangement, the principal agent model Is put in place.
Regardless as to whether the provider is a private firm or a government organization,
the contractual arrangement puts the principal agent relationship in place.
A literature search for studies dealing with contracting government functions
was conducted of library databases at the Army War College, the Defense Technical
Information Center, and university databases with professional journal abstracts.
Academic journals, such as Public Administration Review, have published general
articles on the contracting of government functions, primarily from an administrative
theory perspective (Ferris & Graddy, 1986; Bozeman, 1988; Prager, 1994). The most
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
prolific author has been the General Accounting Office, with numerous reports on all
aspects of contracting government functions. Between August 1980 and August
1987, GAO completed 65 different reports, which specifically addressed the A-76
process. As befits the General Accounting Office, the focus of each study was on
answering a particular request from Congress. Using an auditing approach, the GAO
studies would try to verily savings, document problems, and review program
effectiveness. The early focus was on commercial activity functions but more
recently focused on other cost competition and outsourcing efforts. Other agencies,
such as the Army Audit Agency or the DOD Inspector General, have also completed
studies; but the scope of the analysis is typically limited to a particular competitive
action. The Office of Management and Budget, the Center for Naval Analysis and
the Congressional Budget Office have done some analyses examining the
competition process and summary level results (Office of Management and
Budget, 1988; Center for Nava! Analyses, 1993, 1996, 1997; Congressional Budget
Office, 1995). During the early 1990s, there was a series of public-private
competitions for depot maintenance work, and this generated several articles (Gabler
& Moyer, 1996, Congressional Budget Office, 1995). Across the board, the articles
and studies looked at the quantitative cost savings and ignored the organizational
changes. In none of this literature was the concept of a principal agent relationship
discussed.
There were a significant number of commercial activities studies completed
in the late 1970s. The early-published work, completed right after these studies were
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
finalized, concluded that large savings were possible. In 1982, the Congressional
Budget Office, in 1983, the National Academy of Public Administration and, in
1984, the OMB completed studies supporting potential savings ranging from $l-$3
billion from competing government commercial services. In a 1986 review, GAO
(1986) supported this analysis. GAO (1988b) looked at the studies completed within
DOD, because 80% of the federal government savings came from DOD to identify
which functions had achieved the greatest savings and where additional savings
could be found. These analyses reviewed the results but did not focus on how the
organizations were operating under the principal agent model.
A review of the Navy’s results with the Commercial Activities (CA) program
examined the projected savings but not the actual results (Carrick, 1988). Garrick
believes that putting government managers in a competitive environment gained the
desired results of increased efficiencies. The Navy kept the in-house government
organization as the service provider in 58% of the functions studied. Carrick
concluded: “CA competitions induce negative spillover effects and require the
diversion of scarce management time. Thus, the key to economically achieving
government efficiency consists of structuring appropriate internal incentives for
public managers to identify and implement ‘best-practice’ production methods.” His
goal was to gain the benefits of the competitive process without the detrimental
impacts of competition. Carrick did not look at long-term management of the
organizations under principal agent but focused on the actual competition.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GAO was also asked to look at actual results to see if the planned savings
from the competitions were achieved. Some of these studies showed that cost
savings were not materializing as planned. At Ft. Eustis, $13.5 million in projected
savings were never found, and contractor costs were expected to exceed in-house
costs by $600,000 (General Accounting Office, 1988a). A review of the Ft. Sill
logistics commercial activities study showed a cost growth of at least $14.5 million,
and as high as $20 million, despite the initial decision to contract out based upon a
projected $2.7 million in savings over the 5-year contract period (General
Accounting Office, 1991a). These are two examples of moral hazard that occurred
once the organization went to a principal agent model and the total costs were
quantified.
The 1988 study by the OMB, titled “Enhancing Government Productivity
Through Competition: A New Way of Doing Business Witten the Government to
Provide Quality Government at Least Cost,” cited an average of 30% in savings from
the original government cost. The savings average 20% when the government
organization won and 35% when the private sector won. A 1993 Center for Naval
Analysis study, “Analysis of the Navy’s Commercial Activities Program,” cited
similar savings. At the agency level, GAO (1990b) studied the Government Printing
Office and concluded that a lack of competition resulted in increased costs and
inefficiencies. The Defense Commission on Roles and Missions was established in
the 1994 Defense Authorization Act. It produced a 1995 report, “Directions for
Defense,” and, using the OMB and CNA analysis, claimed there was a huge potential
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for cost savings to modernize the defense forces by contracting basic services and
functions, such as depot maintenance (CORM, 1995). In summary, by the 1990s, the
general opinion was that competition and outsourcing could yield large dollar savings
in government operations. These are the larger forces Kirkpatrick identified as
driving governments around the world to using contracts and the principal agent
model. However, in all of these studies, the focus is on costs savings, and does not
look at the need to change the way organizations are managed and if management is
effective in making this change to the principal agent structure.
The data on cost savings through contracting is, at times, contradictory. In
the highly technological defense areas, competition has brought costs down
(Congressional Budget Office, 1995). However, other instances have shown that the
movement of work to the private sector, even under competition, has actually
increased costs (General Accounting Office, 1994). These studies look at the broad
scope of the issue and attempt to draw generalized conclusions. One problem is the
lack of analysis of the ex ante and ex post costs. The inclusion of transaction costs
under a principal agent relationship would bring further clarity to the cost efficiency
of outsourcing and competition. Because of the limited data, the accuracy of claimed
savings is questionable, and the general perception that competition alone drives
costs lower in all instances is suspect. GAO (1990a) performed an analysis and
concluded that the data available did not support the savings claims. GAO (1996b)
began to question the perception that outsourcing would yield big savings. They
found that the underlying assumption of potentially large savings by shifting depot
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
maintenance work to private industry was unsupported. In 1997, DOD projected
savings from outsourcing to be at least $2.5 billion annually. GAO (1997a) produced
a report questioning these projections. “Although DOD’s commercial activities
database provides an initial projection of savings from outsourcing competitions, few
studies have been done to determine actual savings realized over time.”
Although much has been written on the principal agent model, the theory of
competition to lower costs and the A-76 program for contracting government
functions, very little has been done to integrate these three areas. Large structural
changes in government over the last five decades and such developments as public
choice theory are pushing government organizations to competing functions.
Whether the winner of the competition is a government organization or a private
sector-firm, the principal agent model comes into being. Although studies have been
done of locations where competition has occurred, and even across multiple sites,
none could be found that studied the organizational structures and the principal agent
relationship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER HI
IMPORTANCE TO PUBLIC ADMINISTRATION
When administrative historians some years hence study the
1980s, they are likely to conclude that “privatization” was the
single most influential concept of the decade. Their studies
will undoubtedly portray public administration as being
profoundly altered by several ideas that have collectively
become known as privatization. These same historians are
also likely to conclude that public administration, as an
intellectual field and profession, was in disarray and decline
with relatively little capacity to direct its own destiny in the
1980s. (Ronald C. Moe, 1987)
The writings and research In Chapter 2 discussed a shift in government
operations from an integrated organization to one structured under the principal agent
model. The competition process is used to define the work needing to be
accomplished. The agent is a sub-component of the original integrated organization,
having won the role through the competitive process. Very little is known about the
transition and implementation to the principal agent organization. No research on the
residual organization and its ability to undertake the role of principal and manage the
agent was found in the literature. Many studies have been done on the cost savings
of the A-76 process, but there is only limited research on the organizational
effectiveness of its implementation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The federal policy to rely on the private sector for the provision of goods and
services available began is 1955 in a Bureau of the Budget Bulletin, which was re
issued in 1957 and 1960. OMB issued the first Circular A-76 in 1966 that
established the policy on contracting for commercial activities. Circular A-76 was
revised in 1967, 1979, and 1983. In 1983,.,OMB issued a supplemental procedures
handbook for conducting the A-76 cost comparison studies. As the program gained
in prominence and more functions were contracted to private-sector firms, Congress
enacted several sets of restrictions to limit the program and stop the loss of federal
employment. These restrictions had the effect of limiting the program until 1996,
when most expired. In 1996, the A-76 handbook 'was again revised, and many of the
procedures were designed to streamline the competition process. From October 1995
to January 1997, DOD announced plans for A-76 studies covering 34,000 jobs and
that another 100,000 would be started within the following six years.
The A-76 process is simple in concept but has proven difficult in execution.
There have been numerous studies to reduce the process’ length and complexity
(Center for Naval Analyses, 1996; General Accounting Office, 1988b, 1995). The
size, scope and impact of the program attracts continued attention from industry.
Figure 3 depicts the basic study process. The first step is developing the performance
work statement and quality assurance plan. This is followed with a management
study of the existing government operation to improve processes, insure proper
staffing levels and pay grades, which yields the Most Efficient Organization (MEO).
The in-house government cost proposal is then built based upon the MEO—the
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
output of the management study. This entire process takes from one to three years,
depending upon the complexities. The PWS then forms the basis for a contract
solicitation, where private-sector firms are offered the opportunity to develop their
proposals. The private-sector proposals are then cost-compared to the in-house
government proposal to decide who will become the service provider.
Cost
Study
Decision
Perfbtstsace Work
Statement and Quality
Assurance. Plan.
I
Contractor
Proposals
Figure 3. A-76 Study Process Model
There are clear differences in the cultures of government and private-sector
organizations. Government organizations focus on service and are constrained by a
myriad of regulations and laws from both the Executive and Legislative Branches.
Private-sector firms focus on profit, growth of the business base and have a greater
flexibility in personnel management. The A-76 process tries to compete these two
different organizations on an even playing field to find the most efficient producer.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The process tries to bring the government culture focus in line with that of the private
sector, meeting the contract requirements at the least cost levels.
Government organizations are continuing down this evolutionary path begun
almost 50 years ago. The current Bush Administration is following in the footsteps
of the Reagan, Bush and Clinton administrations before it and has a goal of placing
additional civil service jobs up for competition. This is taking the federal
government away from an integrated command and control management mode! to
one dominated by contractual agreements. The processes and management skills
needed for each organizational structure are different, and to transition from one to
the other requires a different management approach. The public administration
community has recognized the shift but has provided very limited analysis and
writings on how to manage the organizational changes through this transition.
New Paradigm of Management
The ultimate result changes the idea of the “Hollow State” described by
Millward (1996). Under this concept, government is not the direct provider but, as a
matter of public policy, contracts for virtually all outputs and services. When the
Hollow State concept is applied to local, state or federal government, it means they
provide governance, while the private sector produces the required services. The
Hollow State relies on reports, evaluations and feedback to monitor outputs, but the
providers are the private-sector firms. This eliminates nrnch of the government
bureaucracy and hierarchy. It is patterned after the hollow corporation concept that
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
integrates the output of many subcontractors but produces very little as the prime
production firm. This Idea takes government out of the direct producing role. The
government manager establishes contract networks to provide sendees. These
networks are inherently weaker forms of control than the traditional bureaucratic
hierarchy. Managers are continually faced with the problem of coordinating and
focusing the efforts of the various contracts and contractors. Savas (1982) Identified
potential problems with this concept as a lack of responsiveness, the potential for
fraud, and an inability to control the size of government. He questioned whether the
benefits of privatization for cost efficiency will actually occur or If they will
disappear in contractual modifications and other changes.'
The concept of the Hollow State has been around for many years. Since
World War II, every major public Initiative the federal government has launched,
including Medicare, Medicaid, environmental cleanup, anti-poverty programs,
sewage treatment, highway construction, and job training, has been government
managed and performed by the private sector (Ketti, 1993). The Hollow State
concept takes this to an extreme where the private sector is the first line provider of
virtually all government services. These would include local services, such as
garbage and parking meter collections, recreation services and emergency services.
At the state level, this would include motor vehicle registrations, social services
provision, managing state parks and highway maintenance. At the federal level, it is
grants management, regulation writing, tax collection, data processing and facilities
management. At the limit, everything that is not “inherently governmental” is
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
contracted. Inherently governmental could be limited to policymaking, contracting
decisions, legislative actions, military actions, and affairs of state; however, even
these are becoming less restricted.
The theory implications for public administration go back to Woodrow
Wilson and the quest to serve the citizenry with “the utmost efficiency and at the
least cost.” The options range from the Hollow State concept, with services provided
by contractors, to where government employees directly provide most of the needed
sendees. The outcome will probably fall between these extremes, but the underlying
shift to the principal agent model is based upon the assumptions of cost savings
through competition. Politically, the movement for the past three decades at the
federal level is to reduce the most visible component of government size—-its
employees. Although total expenditures for contractor support continue to climb and
now exceed $108 billion annually, the focus remains on federal employment levels
(General Accounting Office, 1995). The effort is to reduce the cost of government
sendees, not reduce the scope of government. Competition processes may change
who provides the service, but the financing continues through government
revenues-— taxes.
The potential significance of this study Is its ability to relate the principal
agent theory to the actual implementation. The competition process changed the way
services were delivered to the principal agent under a contractual arrangement;
however, the reality may be that some organizations never migrate to the principal
agent model Under principal agent, research is needed to understand how Incentives
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
operate to motivate the organization to Increase economic efficiency. The question
of how implementing the principal agent model impacts the performance of the
larger organization Is an issue. Using this Information, policymakers can better
understand the direction in which government organizations are evolving and
develop organizations designed around the principal agent model. It can also help
provide an understanding of transaction costs and organizational structures as the
environment changes from a traditional bureaucratic model to a principal agent
model.
Changes in Federal Employment
The National Performance Review had a goal of reducing federal
employment by 272,900. Between 1993 and 1996, 240,000 of these reductions were
achieved, creating the smallest federal work force since John F. Kennedy was
President. This downward trend In federal employment has continued, though not as
dramatically. There are many “inherently governmental” functions the private sector
cannot perform. OMB defines these as “so intimately related to the public interest as
to mandate performance only by Federal employees.” As the federal employment
levels drop, those functions not inherently governmental continue to migrate to
contractor performance. Kelman (2001) states that NASA spends 78% and the
Department of Energy almost 94% of their budgets on contractor-provided goods and
sendees. In its M y 2001 issue, Government Executive Magazine reports that
contracted man-years now approach 231,000 annually for the Army alone.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The decline in federal employment and commensurate growth in the
contractor support is a long-term trend that will continue. The result is a greater
reliance on the principal agent model, where the government employee fulfills the
role of principal. Where contracting has historically been seen as a support function,
contract negotiation and administration becomes the primary role of the principal,
making this a core function of government (Kelman, 2001). This requires different
recruitment, training and career paths for government employees. Entry-level
employment in trades and non-professional areas will disappear almost entirely from
federal employment. This will eliminate much upward potential for those not in the
professional ranks. No longer will the large number of diverse employment
opportunities exist within government, but these will shift to the private-sector
provider. Production functions will continue to shift away from government.
Historically, A-76 studies have resulted in about a 50% win rate by the
private sector (General Accounting Office, 1988, 1997). These functions are then
competed again every five years. If the function goes to the private sector, the
government does not put forth an in-house proposal in follow-on solicitations, so the
function remains forever under private-sector performance. If the function is initially
awarded to an in-house government provider, five years later, it is competed again,
with some percentage moving to the private sector. Ultimately, there is a continually
declining base of in-house government organizations acting as the agent.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Implications to the Work Force
This trend has two very important implications. The first relates to how
government organizations operate and manage. The second relates to the makeup of
the work force. At the macro level, these trends are forcing government into an ever™
growing number of principal agent structures. The model is being employed in areas
where it was not utilized in the past, probably by managers that are unaware of how
to best operate under this paradigm (General Accounting Office, 1992). Evidence of
such problems was found after the first large round of changes to contractor
performance (Genera! Accounting Office, 1988, 1991), and such problems can be
expected as more functions move to the principal agent model. This shift requires a
change in the day-to-day government processes, changes in the training of the staff
working under this structure and, ultimately, a change in management style. This is a
shift from an organization directly responsible for the mission to one responsible
through, the oversight of the agent and the contractual arrangements. It requires not
only a different approach but also different types and skills of employees.
As these functions transition to the private sector, the number .and types of
employees coming into government service will change. There will be a greatly
reduced opportunity for those looking for entry-level employment without the more
sophisticated skills of contract management. There will always be the need for the
blue-collar and pink-collar work force within the agent, but government may no
longer have the entry-level positions that offer the stepping stones to higher-level
white-collar positions. The problems of information asymmetry, where the agent is
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
more knowledgeable about the function than the principal, will grow. This problem
exists in many technical areas, and it will expand. As the work shifts to a private-
sector provider, the upward opportunities within the principal organization will
diminish, simply because there are so few employees. The principal is typically
under 10% of the size of the agent's organization, so the ability to broaden in job
experience or progress in upward mobility will both be restricted.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER IV
SPECIFIC RESEARCH QUESTIONS
The major insight of agency theory is that control loss is
inevitable in agent-principal relationships. (Barry M.
Mitnick, 1992)
The first three chapters established the thrust of the research, the existing
literature and the reason why this topic is important. This chapter presents a detailed
discussion of the research questions.
1. Does the principal agent relationship apply to organizations where
government has implemented contractual arrangements for commercial activities?
2. Do the participants understand the inherent forces that detract from an
optimal efficient organization?
3. Can common trends of coping with these forces be found that show the
principals are managing the organization for optimum efficiency?
The outcome of an A-76 competition is a contractual arrangement with a
principal agent organizational relationship, regardless of the winner. A “contractual
arrangement” is in place when there is a formal performance work statement
identifying outputs and quality indicators to be produced by the agent. This study
explores the theory of principal agent relationships and organizational arrangements
where the government has implemented A-76 competition of commercial activities.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The research question is in three parts. It begins by asking If the principal agent
model applies, then if It Is understood by those affected, and, finally, are there
commonalities in how the model is implemented.
Does Principal Agent Apply?
The first research question is to determine if the principal agent relationship
exists in the manner that theory would predict. Within the question, “principal agent
relationship” is the organizational model depicted in Figure 2 in Chapter 2, as
described by Kathleen Eisenhardt. The principal and agent have clearly delineated
outputs and objectives. The “contractual arrangements” is the formal document that
establishes outputs and quality Indicators. If the agent is a private-sector firm, the
document Is a legal contract; but, if the agent is an In-house government
organization, the document is the performance work statement. “Commercial
activities” include those functions identified in OMB circular A-76. The two
functions chosen include: (1) base operations public works such as facilities and
equipment maintenance—base supply—and (2) information technology.
Data-gathering and analysis include looking at how several installations
manage the principal agent relationship. The organizational structure would be
expected to have separate principal and agent sub-organizations and a distinct control
structure. Control and measurement data collection systems should be in place to
monitor performance. The research will look along a time continuum to see how
different installations have implemented the principal agent model. This effort will
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
answer the first part of the research question: Does the principal agent relationship
apply to organizations where government has implemented contractual arrangements
for commercial activities?
Do Participants Understand Principal Agent?
The second part of the research question investigates if those involved in the
principal agent relationship understand what impacts the organizational efficiency.
“Inherent forces” are the six variables that impact the principal agent structure: goal
alignment, asset specificity, moral hazard, information asymmetry, incentives and
coercives, and transaction costs. “Optimal efficiency” is when the organization
operates at a level when all changes to any variable will not increase the total output
of the organization—a Pareto efficient state. The question asks if the participants
understand the forces that can detract or improve the operation of principal agent
organizations.
The research examines each of the six variables at each installation. The
participants are asked for their understanding of asset specificity, moral hazard, and
information asymmetry and how these variables impact the organizations’ efficiency.
Additional information on how the organization copes with these variables and what
has been done to manage them to reach a high level of efficiency is investigated.
The incentive and coercive systems are researched to see if they are designed
to meet the goals of the principals and to optimize the efficiency of the organization.
If incentive and coercive structures were in place, it indicates management
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
understands the underlying theory and is striving to work the principal agent
relationship at an efficient level. Without a system of incentives and coercives, the
principal has limited control or influence over the agent’s actions. Although the
principal always has the alternative of canceling the contract and competing it again,
this is a very drastic step. A well-designed set of incentives and coercives can
motivate the agent to more efficient performance.
Each variable can impact on the transaction costs. The research asks if the
organization understands the transaction costs, examines the opportunities for these
costs to occur and what is done to minimize these costs. The ex ante costs axe sunk
costs and cannot be recovered; however, the ex post costs axe directly related to the
quality of the contract document and the performance monitoring system. The
research will examine whether the organizations have acknowledged these ex post
costs and are managing the total transaction costs to reach an optimum efficiency.
There will not be an effort made to quantify the costs but, rather, to see if the
principal organizations have an understanding and actively to attempt to minimize
these costs. This will answer the second part of the research question: Do the
participants understand the inherent forces that detract from an optimal efficient
organization?
Is There a Trend?
Finally, the study will look for preferred management approaches under a
principal agent relationship. The phrase “common trends of coping” refers to
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
common organizational structures, similar performance monitoring systems, similar
incentive systems or other approaches that are common across several installations.
Each installation has subjected the same functions to the A-76 competitive process
with the goal of reaching a more efficient structure. A common approach to
managing the influencing variables could be expected. If there is a variety in
approaches, the question of why did they take a different approach will be addressed.
This will attempt to answer the final part of the research question: Can common
trends of coping with these forces be found that show the principals are managing the
organization for optimum efficiency?
Answering the Questions
The specific research Questions were developed from the literature in Chapter
2. The larger macro level changes affecting government are occurring far above the
installation and organizational levels. These installation level organizations have
been operating as an integrated structure for almost 60 years. The changes are being
driven from outside the organization and are causing fundamental transformations in
how the organization is managed. Changing to a principal agent model of operation
can cause upheaval within the organization, and the benefits are often hard to
quantify. Therefore, resistance to this change could be expected, unless there is no
alternative as when a private-sector firm wins the contract. The organization that
becomes the principal may also have difficulty in transitioning to an oversight role.
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The data-gathering and analysis will answer whether the principal agent model
operates in the way theory predicts or if there are variations occurring.
Thed ata-gathering looks at the organizations to define the status quo
operating mode and the site-specific history of principal agent implementation. The
A-76 process directs the implementation of the principal agent model after the
competition is completed. For those locations where the government is both
principal and agent, there are questions of clear boundary delineations of both
parties. Theory would suggest a quasi-contract that lays out specific levels of output
and quality, and there would be distinct organizational boundaries for the principal
and the agent. Where the private-sector firm is the agent, a contract will be in place
with these defined quality and output measurement systems. The data-gathering
defines the organizational relationships and how the organizations work with the
principal agent model variables.
The analysis focuses on common trends across the organizations looking for
validation of theory. The A-76 competitive process, the variables of the principal
agent model, and the operating environments are common to all sites. The analysis
effort is to determine if practice follows theory; or if there is a deviation and why this
occurs. The data is developed from questionnaires and interviews, is primarily
qualitative, and the analysis lias very limited numeric quantification. However, the
trends and results developed can provide a validation of principal agent theory in
practice.
6!
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The principal agent theory says that there are transaction costs associated with
operating under this structure. The ex ante costs are for putting the structure in place.
Once the contract is in place, there may be cost increases, renegotiating and
monitoring costs. The research examines the ex post costs of renegotiation and
monitoring to determine if the goals of the principal and agent are aligned to
minimize conflict and costs. These do not appear in current thinking for savings
projections and future cost competitions (General Accounting Office, 1995).
Issues of interna! and external validity discussed in Chapter 5 bound the
research questions and the study. The organizations studied are a sample of U.S.
Army industrial installations and a subset of common functions performed at each
location. The research does not look at how the A-76 competitions were conducted
nor how the functions perform their mission. The focus is limited to how the
organizations have structured the principal agent relationsMp and their understanding
of how it should operate. The intent is to validate the theory against what is
occurring at the locations, not to develop new theory.
The conclusions will look to validate the organizational model under
principal agent theory based upon the data and analysis. The U.S. Army and OMB
have written guidance not only on how to conduct A-76 studies but also on how to
implement the results. Although not specifically calling out the principal agent
model, this is the desired end organizational state. This is a dramatic organizational
change, and the question being asked is how well have organizations adapted to this
new structure.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER V
RESEARCH METHODOLOGY
It is not the fruits of scientific research that elevate man
and enrich his nature, but the urge to understand, the
intellectual work, creative or receptive. (Albert Einstein,
n.d.)
Research Methodology Options
The research methodology uses a comparative approach examining principal
agent relationships of similar operations and organizations. Many different
methodologies were considered from a program evaluation methodology (Wholey,
Hatty, & Newcomer, 1994) and from a research methodology (O’Sullivan & Rasse!,
1995). Certain methodologies are inappropriate because of the lack of controls and
the variance in location. Selection of the study design was through a process of
eliminating those approaches that were not applicable and then selection of the most
appropriate technique. The research methodology combines elements of the design
for description and comparative analysis.
Elizabeth O’Sullivan and Gary Rassel (1995) identify two categories of
research design: (1) design for description and (2) design for explanation. The
design for description is used by researchers to explain what they are observing and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
may suggest causality. Within the design for description category, they identify three
techniques: (1) cross-sectional studies, (2) time-series analysis and (3) case studies.
These techniques may be used independently of one another or in combination.
Within design for explanation are the experimental and quasi-experimental
designs—methods where the researcher has some control over the subjects and the
situation.
S. Isaac and W. B. Michael (1995) identify nine categories of research design:
(1) historical, (2) descriptive, (3) developmental, (4) case or field study,
(5) correlational, (6) causal-comparative, (7) true experimental, (8) quasi-
experimental and (9) action. Although the taxonomy differs from that of O’Sullivan
and Rassel, the basic categories are very similar. Figure 4 shows how the
researchers’ categorizations relate.
O F Sullivan a n d Rassel Isaac and Michael
D esim .forD esariB tioB H isto rical
G ross-Sectional Studies'
Tim e S eries A nalysis
Case Studies
Descriptive.
'Developmental
F ie ld Study
Experim ental D esigns'
Quasi'-Experim ental. D esigns
Causal^Cdxnparaiive
True Experim ental
QijaribExperimentat
A ction
eo sreiational
Figure 4. Research Methodology Approaches
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Methodologies Eliminated
Selection of a research methodology involves a process of eliminating some
approaches and selection of the most appropriate approach. O’Sullivan and Rassel’s
design for an explanation methodology and the quantitative-based methodologies of
Isaac and Michael are not appropriate. The experimental designs and quasi-
experimental designs require a controlled environment and control of the variables to
establish a cause/effect relationship. The lack of a control group, the inability to
structure an experiment or control the variables makes these two methods
inappropriate. The correlational method investigates the extent to which variations
in one variable affect other variables. The mathematical analysis is rigorous and
requires numerically-oriented data-gathering. This study will use surveys and some
scalar data, but it is insufficient to perform valid statistical analysis or establish
correlation. The causal-comparative method relates cause and effect relationships
and the influence of variables upon outcomes. The study objective is not to develop
a causal relationship but, rather, to focus on determining if the principal agent
relationships exist as theory predicts. The mathematical analysis is descriptive, not
prescriptive.
The action method involves applying an interaction to the study group during
the analysis period. As an example, this could be some sort of a training program to
improve performance. There is no intent to influence any of the ongoing operations
under study. The historical method reconstructs past history to support a hypothesis.
It was not used, because it builds off data from an extended time period to explain a
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
hypothesis. This analysis is more current and examines a shorter span of time. For
these reasons, the historical, correlational, causal-comparative, true experimental,
quasi-experimental and action research methods were eliminated.
Methodology Selected
The analysis methodology selected utilizes Isaac and Michael’s descriptive,
developmental and field study approaches. The study involves several locations and
gathers organizational and operational data on their principal agent model
implementation. The data is used to develop insights into operations, identify
commonalities and trends, and identify differences relative to the theoretical model.
Figure 5 shows graphically that the selected study approach integrates
elements of all three of the descriptive, developmental and field study methods of
Isaac and Michael. The descriptive, developmental research and field study research
methods form a more articulated approach to the design for the description research
methodology of O’Sullivan and Rassel. Although not quite reaching the depth of
detail of a multi-site case study, the data is gathered directly from field sites and is
used to develop a descriptive analysis across time. The data was collected using
persona! and telephone interview's, survey questionnaires and personal observations.
Qualitative analytical techniques, such as pattern matching and basic statistical
analysis of scalar data, were used to complete the descriptive research. The end
result is a multi-location study to determine if the principal agent relationships exist
as theory predicts and how these function in real world conditions.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
O ’S u llivaii and Design, fo r D escrip tion
R asse!
Isaac and H iclia e l
Figure 5. Research Methodology Approach
Figure 5 relates the three methodologies of Isaac and Michael to the design
for description of O’Sullivan and Rassel. It is a qualitative analysis, not a
mathematically quantifiable analysis. The descriptive method develops extensive
data about a population or study situation to describe the characteristics, conditions
and problems. It is literally used to describe what is taking place rather than
determine causes or explanations of what is occurring; however, information
gathered from a descriptive research design may be used to test hypotheses or explain
relationships. Survey research is one descriptive research method that develops
information without drawing conclusions. The study uses the descriptive approach to
develop the background information in which each installation operates.
The developmental research method looks across time to investigate patterns
and change. This approach does not go into the same level of detail as the
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
descriptive method and focuses more on variables and their change over time.
Longitudinal, trend and time study are typical developmental methods. The research
question for this study focuses on the principal agent relationship but explores
changes within the organizations to see if this relationship matures or changes.
Elements of tWs method are incorporated into the study looking for trends and
external influences.
The field study or case study method develops extensive data focused on a
given organizational unit. It is often difficult to extrapolate the conclusions of a field
study to other sites. GAO used a modified case study approach when they performed
studies at Ft. Sill (General Accounting Office, 1991a), Jacksonville Naval Air Station
(General Accounting Office, 1989) and other sites. Almost all of the GAO studies
looked primarily at cost data without looking at the underlying organization
structures. The General Accounting Office’s individual studies look at problems and
issues but provide no overall trends that might apply across locations or through
time. This study integrates aspects of the field study method but does not go into the
depth of a traditional, single-site case study. This study incorporates elements of
these three methodologies in a qualitative analysis approach to answer the research
questions.
Study Protocol
The design for description combines elements of three methodologies:
(1) descriptive, (2) developmental, and (3) field study. A descriptive methodology is
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
used to establish the existing conditions using primary data gathered directly from
original sources, not secondary data collected by others. Using the principal agent
theory as a baseline model, the analysis compares the predicted results to what is
occurring within the organizations under study. Looking along a time continuum, a
developmental method examines changes over time and if these changes follow the
principal agent model. The data is gathered in sufficient detail to allow a modified
field study approach to determine trends or commonalities across sites. These three
methodologies are used in conjunction as a design for description to validate the
principal agent model with the organizations’ data.
Mary Ann Schrier (1994) discusses a series of data collection techniques
when verifying and analyzing theory'. Several of these, such as forms analysis and
expert judgment, are inappropriate. Gathering extensive qualitative data in an
iterative approach offers the best opportunity to focus data collection efforts. The
initial effort is a questionnaire presented to multiple individuals to validate the
information and arrive at a consensus of what is occurring at each site. The design of
the questionnaire is to elicit both qualitative and quantitative information. The
questionnaire was followed with interviews of key individuals to develop a more in-
depth understanding. Finally, limited observation enabled the researcher to have a
more complete understanding and fill in any information voids. This three-tiered
approach attempted to maximize the data gathering and continually focus it on the
research question analysis.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Different texts provide direction on constructing and conducting surveys to
improve the data reliability and usefulness (Fink & Kosecoff, 1985; Rhea & Parker,
1992). The initial contacts gained approval and identified the key personnel to
participate in this study. Prior to distribution, individuals who had been involved in
the A-76 process reviewed the questionnaires. The questionnaires were delivered
electronically via E-mail. The questions followed the study objectives to determine
if a principal agent relationship exists, the influence of the variables, and the
identification of common trends. Both scalar and open-ended responses were
requested. As a follow-up to the questionnaires, interviews using open-ended
questions requiring detailed responses obtained additional data. The intent was to
clarify specific areas and develop adequate data for analysis and develop conclusions.
Survey questionnaire responses tend to be shortened and information often
left out when using only written input. The follow-on interviews, either in person or
via telephone ferther clarified the information. The interview process developed
additional details and insight. The interview's were more open-ended and were not
conducted until the written questionnaires were received. This enabled the research
to focus on information gaps.
Robert Yin (1994) developed a classical approach to qualitative analysis. He
proposed pattern matching as the primary approach -when validating an underlying
theory for what is occurring. In this study, the identification and explanation of the
principal agent theory is the focus and drives the study protocol and analysis. Yin
writes that it is the research questioning of the theory that focuses the data-gathering
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and analytical techniques. Figure 6 depicts graphically the concept of using theory
explanation as the study approach to direct the research and analytical techniques.
Principal Ageutj
Heoij
Research.
Question
t
Data Gathering ■Analytical Analysis
Techniques Techniques Results
Figure 6. Explanation of Theory Approach
Definition of Variables
The variables under analysis are posed in classical principal agent theory.
Data was gathered on asset specificity, goal alignment, moral hazard, incentives and
coercives, monitoring and transaction costs as the dependent variables. The
independent variable was the organizational design implemented under the principal
agent model. There are two independent variables—either the private-sector firm or
the in-house government organization as the agent. The questionnaire defined each
variable and then explored the understanding by the respondents, the magnitude of
the impact and complexity of that variable.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Asset specificity is the degree to which the agent's performance or output can
be specified. When purchasing products, such as a computer, asset specificity is
extremely high. However, when buying services, such as research into new
technologies, it becomes difficult. The operation of a military installation requires
defining levels of maintenance for equipment and facilities, defining the quality of
outputs and the responsiveness of staff covering a vast array of services. The
prevailing performance contracting concept establishes the desired outcome without
defining how to accomplish the work. High asset specificity defines the quantity and
quality of the performance as finitely as possible.
Goal alignment is a measure of how well the goals of the agent and the
principal match up. A simple example of goal incongruence is when the agent
strives to maximize profit and the principal strives for cost minimization. Certain
arrangements instill this conflict, and others minimize the conflict. When the agent
is an internal organization, the goals may not be aligned, and sub-optimization may
occur. When the agent is an external firm, goal conflict can occur, thus, decreasing
the economic efficiency of the process. If the goals of both principal and agent are
aligned, this reduces monitoring and other transaction costs.
Moral hazard occurs when the agent takes an inefficient action or provides
distorted information, because his interests are not aligned with the principal’s. One
type of moral hazard is ‘ '"profit maximization,” where the agent attempts to charge
more than necessary to extract the maximum from the principal. A second type of
moral hazard is “shirking,” where the employee does less than is required in the
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
performance work statement. The agent expends less effort or performs less work
than was agreed to with the principal A third type of moral hazard originated in the
insurance industry. This occurs when the agent feels unconstrained to take risks that
would not normally be taken, because the consequences are mitigated. For example,
a driver is less careful because he has insurance. All of these types of moral hazards
influence the principal agent relationship.
information asymmetry is when information is not shared perfectly between
the principal and the agent. The common situation is when the agent has a more
intricate knowledge of operations than the principal. The agent is performing daily
operations and understands in detail the problems and opportunities. The principal is
often removed and not knowledgeable of what is required. However, the principal
can have more information than the agent; and this, too, can cause information
asymmetry.
incentives and coercives are actions taken to encourage superior performance
or discipline poor performance. These would normally be financial incentives and
penalties in the traditional contract. Incentive clauses that reward performance above
the minimum are used to motivate the agent to exceed and provide higher than
expected value. Incentives for improvements in responsiveness, output quantities or
quality? measures are integral to the principal agent model. Without these, the agent
has no incentive to do more than meet the minimum performance criteria. Coercives
are the penalty options the principal may invoke, if the agent does not meet the
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
minimum performance criteria. Termination of the contract is the most drastic and
final option, although financial withholding penalties are a more common option.
Monitoring is the oversight and information-gathering mechanism the
principal puts in place to insure the agent is achieving the goals. Monitoring consists
of quality surveillance and assurance plans, performance and production reports, and
other inspection processes. In a principal agent relationship, monitoring processes
are in place to validate contract performance and the alignment between the goals of
the principal and the agent’s actions. The intensity of the monitoring reflects the
level of goal alignment. A vertically integrated organization may require less
monitoring because of less external goal conflict.
Transaction costs are those costs associated with instituting the principal
agent relationship. These costs axe ex ante, occurring before the relationship is
implemented, and ex post, occurring after the relationship has been put in place. Ex
ante costs are the costs of drafting the agreement and negotiations of the agreement.
The ex post costs include categories such as monitoring and shirking, and costs for
haggling or re-negotiation in the event of non-performance. This study will look to
identify if the principals understand these transaction costs enough to actively
manage them.
Questionnaire Design
Development of questionnaires is more of an art than a science. The
researcher strives to develop information that is useful without leading or limiting
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
those being Interviewed. lean Converse and Stanley Presser (1986) suggest that
simplicity Is the key to a successful survey questionnaire. Their recommendation is
to focus on close-ended questions and use the open-ended questions to explore
certain aspects In much more detail. The advantage of the close-ended approach Is
the uniformity of the answers, while the open-ended questions allow the interviewee
to inject more detail and explanation. They recommend techniques, such as using
examples in the body of the question to provide clarity and improve the
understanding.
The use of examples, the grouping of questions around a particular aspect and
the asking of redundant questions to verify earlier answers improve survey accuracy.
A technique to determine the strength of commitment Is the use of the scalar
approach. Converse and Presser recommend not extending the scale into too many
levels and to provide a “no opinion” option. Without the “no opinion” option, the
interviewee is forced to express a preference when one may not exist.
Louis Rhea and Richard Parker (1992) write that the three key elements to a
good questionnaire are clarity, comprehensiveness and acceptability. The survey
instrument must ask questions that cannot be interpreted differently by respondents.
The survey must sufficiently cover all aspects of the topic, leave no gaps in data and
try to insure additional questions are not needed. If statistical analysis is performed,
the questionnaire and sample size must meet certain criteria to be acceptable and
valid.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Procedure
A preliminary discussion with senior management personnel at each site
preceded development of the survey questionnaire. From this, the key issues and the
areas for exploration in the questionnaire were developed. This preliminary effort
helped structure the questionnaire and focus the study methodology. The guidelines
developed by Rhea and Parker, Converse and Presser, Fink and Kosecoff, and others
say there is no one right approach to developing a questionnaire and conducting a
survey. It is more ait than science, more logic than rule-based. Several concepts,
such as leading explanations, clarifying the use of the information, grouping
questions and using closed-ended rather than open-ended questions, were common in
ail texts. These concepts were incorporated into the survey instrument and the
follow-on interviews.
The initial questionnaire was followed by personal interviews of key
respondents. The concept was to reduce the immediate effort upon the respondents
and improve the quality of the responses. The longer the questionnaire, and the more
time required to respond, can detract from respondents’ willingness and ability to
provide detailed information. The questionnaires used at each site were identical.
Each participant was asked to respond independently without collaborating or
consulting other staff members. Keeping the responses independent is important so
their perspectives do not influence the input of others.
The questionnaire was used to establish the organizational climate, the
structure under which the organization works and the general principal agent
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
relationship. The focus was to determine how the relationship Is structured and how
the participants view the success of the relationship. This questionnaire established
the independent variables of organizational structure and explored the respondents’
level of knowledge on the dependent variables. As much detail as possible is
gathered, and there is a greater use of closed-ended, scalar questions. The scalar data
can be used for basic quantitative, descriptive analysis. A seven-point scale was used
for preferences on most questions. By using scalar responses, the respondents were
able to answer a number of detailed questions and establish the strength of their
beliefs. On a limited number of questions, historical data on costs or frequency of
action were requested. This descriptive methodology gathers the base data, setting
the foundation of the existing conditions.
The follow-on interviews were more open-ended. They focused on
operational interactions between the principal and agent organizations and built the
history of how the organizational structure evolved over time. This attempted to fill
in gaps in information from the first questionnaire and was more field-study-oriented.
It looked to establish any uniqueness at each location, but also to establish
commonality. Principal agent theory would imply significant commonality,
regardless of the independent variable, and this phase looked for data to support or
discredit this idea.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Data Collection Approach
This research surveyed a limited number of military industrial installations,
and the sites selected were reasonably comparable. The size of staff, basic mission,
and geographic locations all made comparison possible. The independent variable
was whether the agent was a contractor or in-house organization. Sites were selected
that provided both alternatives.
Data collection was done remotely through surveys and personal interviews.
There is a lack of historical records or existing data sources for much of the
information required for the analysis. The geographic dispersion of sites limited the
observation. The questionnaires provided electronically via E-mail and interviews by
telephone were the primary data sources.
Respondents participating and providing information cannot be selected at
random, but are selected from a pool of individuals with specific knowledge of the
organization. Senior management recommended the respondents for their intimate
knowledge of the organization’s operations and the A-76 study. They were
representative of both the principal and agent organizations. Each individual was
provided a short summary of the study purpose and what their participation would
entail prior to beginning the data collection.
Information gathered through surveys and personal opinions is less subject to
rigorous analysis than directly observable, countable data elements. In developing an
accurate representation of each site, data was gathered from multiple individuals to
arrive at a consensus view of the conditions that exist. The principals, agents, and
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
monitors (those who oversee or report on the agent’s performance but are not the
principal) were asked to respond. TWs trianguiation approach of data-gathering from
multipie levels reinforces the points commonly agreed to and also serves to identify
any disconnects in perceptions.
Survey responses were supplemented with personal interviews. Most of these
interviews were by telephone, although some were in person. The follow-on
interviews were focused on the final section of the research question—whether
principals were managing to optimize the efficiencies of the entire organization.
TWs did not lend itself towards survey data but required a more detailed and
comprehensive input best captured in an interview.
Analytical Techniques
Figure 6 depicts the progression of the research question influencing the data-
gathering technique that then feeds the analytical technique. Pattern matching, is
one of the four dominant techniques for explanation of theory (Yin, 1994). Principal
agent relationships should have a high level of commonality across locations. If the
patterns found in the research match those predicted by the theory, this adds internal
validity to the study. By using multiple sites, common findings reinforce the
conclusions. This is referred to as cross-site analysis. Using matrices is a technique
to order data and to focus on identifying patterns across multiple sites in a
comparative way that trends and patterns can emerge. The goal is to establish
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
common patterns that would support the underlying concepts of principal agent
theory (Miles & Huberman, 1984).
First, a baseline is established for each site. Because multiple individuals are
being asked to participate from each site, a single, homogenous view of what is
occurring may not immediately emerge. Using a double series of questionnaires and
interviews enables clarification. The use of pattern matching and cross-site analysis
enables the identification of commonalities supporting the principal agent theory or
conversely identifies what does not support the theory.
Michael Patton (1990) refers to this analytical approach as “Logical
Analysis.” It is the distillation and ordering of data into a logical presentation to
establish patterns. He refers to the works of such early researchers as Donald
Campbell in developing both two- and three-dimensional matrix analysis approaches.
This study uses only the two-dimensional approach.
Following the pattern-matching analysis is an explanation-building analysis.
This Yin (994) technique builds upon the pattern matching. Explanation-building is
in a narrative form and endeavors to explain the data. It attempts to explain the
causal links that are often complex and difficult to measure. Explanation-building
bridges from the theoretical proposition to the data-gathering and attempts to explain
why the results occurred from a theoretical perspective. If the research data does not
match the theoretical expectations, it is the explanation-building that aids in
understanding the variance and external factors that may be influencing the situation.
The explanation-building is where these differences become understood.
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Limitations of the Study
To add to the body of knowledge in public administration, the study results
must be valid. There are fourteen different threats to the validity of a study. The
conclusions for the four sites under study must be correct; this is the internal validity.
The ability to apply these conclusions to other sites is the externa! validity. The
following sections discuss the study validity threats in detail.
Threats to External Validity
The study population is comprised of commercial activity functions at Army
industrial installations. These are functions common to running a major installation
and include facility maintenance, supply support, engineering, etc., and consist of
both blue-collar trades work and white-collar office work. The threats to study
validity are both internal and external (O’ Sullivan & Rassel, 1995). External validity
is the appropriateness of applying the results to other organizations. The six threats
of external validity include unique program features, effects of selection, effects of
setting, effects of history, effects of testing, and the reactive effect of experimental
arrangements. The study design made three of the threats insignificant, while the
remaining three—effects of selection, effects of history and effects of setting—pose
potential issues.
Unique program features occur when there is an appearance of commonality
of the areas under study; but, in fact, significant differences exist. TWs makes the
extrapolation of the study results to other populations erroneous. All of the functions
8 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
selected at each Installation are a part of the A-76 study process. The rales and
guidelines for establishing the contracts and the MEG are standard across all sites.
The basic functions at each site are common to the operation of any military
industrial installation. The focus on organizational principal agent relationships
serves to identify and minimize the influence of unique program features.
Effect of testing is when there is a possibility that a pretest has influenced the
outcome. There is no pretest in this study. The use of descriptive methodologies
eliminates this possible impact on validity. The data gathered is observational on the
implementation of the principal agent structure, and the study will not Impact those
structures or attempt to change them.
Reactive effects of experimental arrangements refer to impacts of using an
experimental or quasi-experimental method and the subjects’ reactions when they
know they are under study. This is not applicable, because this is not experimental,
and the use of data gathered from actual results rather than a constructed experiment
void these concerns.
The effects of selection can Impact the external validity, when the study sites
are not representative of the larger population. These functions are common to ai
installations and have clearly defined outputs. There is a reasonably competitive
market available, and they are ail commercial-type functions. Application of these
results to areas that do not have these attributes, such as depot maintenance where
there is a poorly developed competitive market, may be inappropriate.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The effects of history may also have an impact. Impacts of history could
include the changes within the A-76 program, organizational development initiatives
at government activities or the effects of programs, such as the National Performance
Review and Quadrennial Defease Review. The changes in government over the last
decade may have influenced the current operating environment. TMs may have
altered the organizational dynamics over time and made any findings inappropriate, if
applied at other sites.
Effects of setting refer to situations that may impact the extrapolation to other
locations. For example, what is found at an industrial setting and rural location may
not apply to an administrative setting in an urban location. The sites and functions
for this study were selected with a basic commonality of mission and organizational
structure for cross-comparison. The functions are common to all military
installations. However, the sites are somewhat unique, because they are industrial-
based rather than troop stationing or training installations. The analysis results
should be representative of how principals and agents interact, and the installations’
industrial mission may limit the external validity.
Threats to Internal Validity
Threats to internal validity concern the reliability of the independent variable
causing the impact on the dependent variables. Internal validity is confirmation that
the conclusions are correctly based on the findings. There are eight different threats
to internal validity, of which history, maturation and selection are a concern. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
threats that are not a concern are statistical regression, testing, instrumentation,
design contamination and experimental mortality. These threats focus on the
measurement instrument and analysis.
Statistical regression is a concern, when the study subject is an outlier to a
normal population. The selected sites were chosen because they are representative.
The use of multiple sites balances the individual case study approach that might fail
to identify aberrations. By using four sites with common characteristics, statistical
regression problems are avoided. The threat of testing influencing the data is
negated, because the subjects are already operating under their principal agent
relationship, and this study documents that relationship. There is no testing; only
interviews and survey data is collected. Instrumentation is a threat, if the survey
instrument changes during the course of the study. The short duration of the study
and the control exercised by the researcher negates this threat.
Design contamination is when the subjects under study act differently,
because they know they are under study. TMs analysis of organizational
relationships is a documentation of what is occurring, and changes should not occur
because of any study influences. The use of multiple individuals at each site
mitigates problems with design contamination. Experimental mortality is when
functions dropout of the study. Each site was contacted before the study began to get
senior management support, and the study methodology was designed to minimize
impact on the sites. All of the locations remained a part of the study population, and
experimental mortality was not a problem.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The remaining three threats are potential limitations on the internal validity.
History can be both an external and internal validity issue. The threat to internal
validity is when something other than the independent variable cannot be ruled out as
the source of variation. Each of these sites lias approached the establishment of the
principal agent relationship under different circumstances. Some were significantly
reducing staff due to Base Realignment and Closure actions, while simultaneously
undertaking A-76 competitions. Others were growing and adding staff. There is no
time when these sites were static and not influenced by larger external forces.
However, during the interview process, questions are asked to establish if the
interviewees see any effects of history.
Maturation is related to the history as the installations become more
knowledgeable about the entire outsourcing process. These are natural evolutionary
changes that occur over time. One of the study sites has had a longer history of
experience with the principal agent relationship and over time may have evolved to a
different level than sites with less experience. Maturation occurs when the
researcher is looking for change due to the independent variable when, in fact, it is
due to evolutionary changes over time. Again, by tailoring some of the interview
questions, the influence should be identified.
Selection is a minor concern, because the sites are selected for their
commonality of function, size and geographic locations. They may be systematically
different than other locations because of the type of work or the types of contractual
arrangements instituted. However, these sites were chosen to eliminate the obvious
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
problems that might occur from selection problems to minimize this threat to internal
validity.
Table 2 summarizes the threats to validity and categorizes them as potential
or insignificant. Those that have the potential to impact the study will be discussed
further in Chapter 7.
Table 2
Internal and External Threats to Study Validity
External
Validity
Threats
Potential
Threats
Effects of Selection
Effects of History
1
Insignificant
Threats
Effects of Setting
Unique Features
Effects of Testing
Reactive Effects
Internal
Potential
Threats
History' 1
Maturation 1
Selection 1
Validity
Threats
Insignificant
Threats
Statistical Regression Testing I
Instrumentation
Design Contamination
Experimental Mortality
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER VI
DATA
It is a capital mistake to theorize before one has data. (Sir
Arthur Conan Boyle, n.d.)
The Study Locations
The Department of Defense has 189 major installations. Outsourcing under
the A-76 process is occurring at virtually every site. The study gathered data only
from Army industrial installations. These sites form a subset within the larger
population of military installations. They are dominated by civilian employment
with a very limited military population. They are typically removed from major
metropolitan areas.
The Army’s depots and arsenals have completed or are conducting A-76
competitions of their public works functions. Some have been contracted to private
firms for years; others have competed and retained the functions in-house. The total
sample size is five maintenance depots, three arsenals and six storage depots. Four
of these locations were selected as study sites: (1) Rock Island Arsenal, (2) Red River
Army Depot, (3) Anniston Army Depot, and (4) Letterkenny Army Depot. The six
smaller storage depots were eliminated from consideration; personnel reductions
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
over the past decade render them almost non-existent. The sites vary from a
population of slightly under 2,000 to 4,000 employees. The locations vary
geographically but perform the same basic functions.
The four study sites are reasonably comparable in physical characteristics and
mission. These sites are in varying stages of implementing the principal agent model
as an outcome of their A-76 competitions. Rock Island Arsenal had converted to
contractor support over a decade earlier and has extensive experience managing
under this structure. Both Red River Army Depot and Anniston Army Depot
completed their A-76 solicitations within the past two years and are relatively new at
operating under principal agent. Both of these sites continue to use government
employees to perform the functions. Finally, Letterkenny Army Depot is still
completing the A-76 process and does not have a decision whether the provider will
be a contractor or in-house personnel. These four study sites stretch across a time
continuum for implementing the principal agent model. They extend from a
Letterkenny organization still in the planning stage, through those that are early on in
their experience, to the Rock Island organization that has had many years’ experience
operating under principal agent. The following sections provide a physical
description, demographics and a short background on the functions included in each
site’s A-76 studies.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Rock Island Arsenal
Rock Island Arsenal is located on a 964-acre Island on the Mississippi River
between Rock Island, Illinois, and Davenport, Iowa. It Is the largest government-
owned weapons manufacturer and arsenal In the western world, employing almost
1,200 people. The primary mission is the manufacture of machined weapon system
components, such as gun mounts for tanks. The Arsenal is also the home to 4,000
other DOD employees. A major mission is supplying the base operations support,
such as information technology, facility maintenance, etc., to the tenant activities.
During the 1980s, many functions underwent A-76 competitions and were
awarded to contractors for performance. They have been re-competed every five
years and, over time, have been awarded to different winning companies. Once the
functions were awarded under contract, the government never again proposed an In-
house organization as a potential competitor. The functions that are performed by
these contractors include most public works functions, motor pool operations,
household goods movement, audiovisual, mailroom and similar functions. They do
not have a single contractor or a single contract but, rather, have multiple contracts
with different firms. Four different firms were identified as performing the agent
functions for Rock Island Arsenal; while, at the other installations, they would have
consolidated to one or two firms performing this work.
In 1997, another round of A-76 competition was initiated to compete the
entire work force of 1,200 employees under a single contract. The unique aspect of
this proposal was that it not only included traditional base operations functions, but
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
also the core mission of the arsenal—manufacturing of components for military
weapons. As the study progressed, decisions were made to eliminate the weapons
manufacturing mission from the scope. This reduced the ongoing A-76 efforts to
information technology and some remaining logistics supply operations. During the
data-gathering phase of this research, Rock Island Arsenal announced that the
government had the most cost-effective proposal and that government employees
would continue to perform both of these functions.
Med River Army Depot
Red River Army Depot is located in Texarkana in northeast Texas. It covers
19,000 acres and is adjacent to the Lone Star Ammunition Plant. The primary
mission of Red River Army Depot is the rebuilding of Bradley Fighting Vehicles and
Multiple Launch Rocket Systems to like-new condition. They have an ammunition
storage mission and certification of PATRIOT and HAWK missiles. The depot
employs 1,500 civilians and has another 1,000 people working for various tenant
organizations on the installation.
Red River began their A-76 studies in 1997 and went to solicitation in 1999.
They had two separate solicitations—one for information technology and the other
for public works and facility maintenance. Red River received no bidders from the
private sector on either solicitation. Although not unique, the total lack of bidders is
somewhat unusual. With the lack of bidders, the installation canceled the study and
took personnel actions to move to their MEO. Canceling is somewhat different from
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
winning an A-76 competition, because the installation is under no obligation to
implement the oversight QAE required under A-76.
Anniston Army Depot
Anniston Army Depot is located near Anniston in northeast Alabama. It was
constructed in 1942 and consists of 25 square miles of land, 46 miles of railroad, 87
miles of fencing, 266 miles of roadway, over 2,100 buildings, 2,650 federal
employees and a contractor work force of several hundred. The primary mission of
Anniston Army Depot is the rebuilding of tracked vehicles, such as the M-l tank,
ammunition storage and chemical weapons storage.
Anniston began its A-76 studies in 1997 and finished in 2000. They studied
facility maintenance, motor pool, boiler plant, property accountability and other
public works functions in one study and information technology, mailroom, and
audiovisual operations in a second study. There were multiple bidders for both
solicitations. The studies both resulted with in-house wins and federal employees
continuing to perform the functions. Staff reductions of approximately 20% were
part of the proposals of the government organization.
Letterkenny Army Depot
Letterkenny Army Depot is located in Chambersburg, Pennsylvania.
Constructed in 1942, it consists of approximately 18,000 acres, 51 miles of railroad,
150 miles of paved roads and over 4 million square feet of buildings under roof. The
work force has 1,100 federal employees working for Letterkenny Army Depot, 500
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
federal employees who are installation tenants working for other DOD organizations,
and 200 contractor employees working on various defense contracts. The primary
mission of Letterkenny is the rebuilding of air defense and tactical missile systems,
such as PATRIOT, and ammunition storage.
Letterkenny is one of the last depot installations to complete this round of A-
76 competitions. They began the process in 1998 but had not completed the
solicitation during the research data-gathering. The initial plan had been for a single
large study that included a great many functions, including resource management,
information technology, acquisition and public works. However, over time, all but
public works, audiovisual support and mailroom operations were deleted from the
competition scope. Letterkenny then decided to have two solicitations, the larger for
public works functions and a separate, much smaller solicitation for audiovisual and
mailroom support. The public works function encompasses about 175 people doing
a variety of tasks. These include: facility maintenance; trades, such as carpenters,
plumbers and electricians; supply management; environmental management; snow
plowing; motor pool operations and a host of other similar functions. It also includes
the administrative planning and oversight of this wide variety of work along with
managing many different subcontractors.
Letterkenny represents the opposite end of the time continuum from Rock
Island Arsenal. Where Rock Island has a very mature principal agent relationship,
Letterkenny is still in the beginning stages of planning the transition to a principal
agent model. Data collection focused on how the organization plans to operate once
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the A-76 solicitation is complete and the principal agent model is adopted. Table 3
summarizes the study locations’ characteristics.
Table 3
Study Location Characteristics
Rock Island
Arsenal
Red River
Army Depot
Anniston
Army Depot
Letterkenny 1
Army Depot |
Home State Illinois Texas Alabama Pennsylvania 1
Prime
Mission
Manufacture
Weapon
Components
Rebuild
Bradley
Fighting
Vehicles
Rebuild
M-l Tanks
Rebuild 1
PATRIOT Air 8
Defense 1
Systems
Agent is
Government
| or
1 Contractor
Both Government Government Unknown
Study Respondents
The initial contact at the four study locations was through the installation
Commander. He then requested that senior installation personnel support the study
and respond to the questionnaire. The questionnaire was E-mailed to these
individuals, and responses came back within a few weeks. Responses from Rock
Island Arsenal came from multiple individuals representing every function that had
undergone an A-76 competition. Initially, both Anniston and Red .River personnel
saw the questions as pertaining only to where the agent was a contractor. They
viewed the organizational relationships at their location in a traditional structure
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rather than a principal agent model Anniston and Red River provided fewer
questionnaire responses, because they felt these responses adequately represented
what was occurring. Additional telephone interviews were conducted with these
individuals to clarify and expand their input. At Letterkenny Army Depot, where the
A-76 process had not yet been completed, interviews were the exclusive method of
data collection.
At Rock Island, the Commercial Activities Office supervisor offered to
coordinate responses and provide information for the study. He circulated the
questionnaire to both his staff members and the staff of the functions already
operating under principal agent structure—facilities maintenance, audiovisual and
mailroom, and equipment management. Five individuals answered the
questionnaire. Follow-up interviews were conducted with the Commercial Activities
Office supervisor, the management consultant supporting the current studies and
several of the individuals involved in managing the functions under contract. The
interviews occurred over several weeks and several different sessions.
At Red River Army Depot, two individuals who had been deeply involved in
the A-76 study coordinated responses. One individual was a supervisor in the
information technology function and was part of the agent structure; the second
individual was in the public works function and was part of the principal
organization. The respondents had been involved in developing the performance
work statements, the most efficient organization and the implementation of the
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
principal agent structure. The initial responses to the questionnaire left information
gaps, but follow-on interviews adequately filled in missing information.
At Anniston Army Depot, the primary respondent was part of the principal
organization charged with oversight. Three individuals from Anniston responded—
one from each function that had been under A-76 and one from the oversight group.
Initially, they did not see the relationship of the principal agent variables to their
situation, because Anniston had won their competitions, and government employees
continued to perform the functions. Follow-on interviews brought to light details on
how they were organized and operated.
At Letterkenny Armv Depot, interviews were conducted with the A-76
Program Manager, the senior staff of the functions under study, and the staff who
conducted the A-76 study. The interview questions were those used for the other
installations. The purpose was to understand how the organization intended to
implement the A-76 study results and if the participants were prepared to enter into a
principal agent organizational structure. Because Letterkenny had not yet
implemented the results of the A-76 competition nor knew the outcome, it offered
the opportunity to observe an organization at the beginning of the process. Seven
interviews were conducted. Since there was no agent in place, there was no
quantitative data on the satisfaction with agent performance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Organizational Structures
Figure 7 shows the traditional command organizational structure that was
standard at all sites prior to the A-76 competitions. It was an integrated organization
that did not operate under a principal agent model but operated under an internal
command and control structure that is traditional in bureaucracies. All first-line
organizations were equal, usually designated as directorates reporting to the
installation commander.
Figure 7. Traditional Command Organizational Structure
The subordinate organizations can be categorized into three basic types: (1) a
mission organization, (2) a contractible base operations organization, or (3) a GIN
base operations organization. The mission organization is the reason the installation
exists; it is the primary production organization. At Rock Island Arsenal, it is the
manufacturing organization; and, at depots, it is maintenance or the ammunition
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
handling organizations. These organizations are considered non-contractible. TMs
definition of being non-contractible is often by statue rather than by the inherently
governmental definition applied to other organizations. By statute, no less than 50%
of the funding for depot maintenance can be allocated to military depots. Almost
50% of the total dollars appropriated for Army depot maintenance are contracted
with private-sector firms, such as Boeing or Raytheon, making the government depot
maintenance organizations non-contractible. The second organizational category is
for contractible base operations functions. These are the organizations that may be
competed under A-76 competition. They are the focus of this study, because they are
moving towards the principal agent model. The third category is for those
governmental in nature (GIN) base operations functions, such as personnel,
contracting and accounting. These organizations are involved in the government
decision-making and management functions that make them non-contractible by
definition.
The traditional organizational structure is a hierarchy under a military
Commander. At the top organizational levels are the directorates, followed by
division level organizations, branch level organizations, and, when the organization
is large enough, there are section level organizations with the first-line supervisor.
This subordinate level structure varies based on the size of the organization.
Organizations are along functional fines; so, for small functions, such as contracting;
first-line supervision may be at the directorate level. Figure 8 shows the
organizational hierarchy within the traditional command structure.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In s ta lla tio n
C o m m a n d e r
O r e c t o n a t e
# 1
1 1
Directorate
#2 1 #3
D ivision: D iv isio n
L e v e l ##1. L e v e l # 2
B r a n c h B r a n c h
L e v e l # 1 L e v e l. # 2
JL
■ k .
.S e c tio n S e c ti o n
L e v e l :##1 L e v e l #2
Figure 8. Organizational Hierarchy within the
Traditional Command Structure
Rock Island Arsenal
Rock Island Arsenal has worked under a principal agent mode for well over a
decade, using a contractor as the agent. Over time, the Arsenal has gone through an
evolution in its organizational structure. The first A-76 competition in the 1980s was
a single, multi-function study won by a private-sector firm. The mission
organizations remained unchanged. The Arsenal had a single principal organization
to oversee all aspects of the agent’s performance. The agent had a multi-level
organization structure performing multiple functions. Figure 9 shows how this
organization operated.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
.Agent
Option
Figure 9. Rock Island Arsenal First Generation
Principal Agent Organization
The A-76 competition consolidated multiple functions under a single large
contract with the General Dynamics Corporation. A single principal organization
was to oversee all aspects of the contract and direct the agent’s performance. The
principal organization was staffed with individuals skilled in contract administration,
but not overly knowledgeable of the technical aspects of the work. This created
problems in directing the agent and accountability issues. The principal’s knowledge
base remained in functional organizations, while the responsibility for oversight of
the agent was in the administrative principal organization. An example of this
convoluted line of responsibility was the mailroom operation. The Director of
Information Management was responsible for all technical aspects of the operation.
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In Figure 10, the Director of Information Management is represented as one of the
principal management organizations. The principal management organizations do
not have direct authority over the agent responsible for accomplishing the work.
Direction needed to be relayed through the administrative principal that served as a
consolidated oversight organization for the contract. The administrative principal
worked through the agent’s multi-level structure to reach the agent organization
performing the work. This led to the problem issues described in the literature, such
as information asymmetry and increased monitoring costs. This structure was not
designed by the Rock Island Arsenal but was forced upon them by headquarters in
Washington.
Agent 0r§.
F u n c t i o n 1 Function 2
Principal ignst.
Function 1
Figure 10. Rock Island Arsenal First Generation
Lines of Responsibility and Accountability
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The installation undertakes a single large contract, because it is often easier to
administer, limits the number of acquisitions that are required, and may also limit the
number of firms interested in competing. Aggregating multiple, disparate functions
necessitates the competing contractors be large firms, to form partnerships or to use
many subcontractors. When there is a limited period to respond to the solicitation,
this strategy may provide an advantage to the in-house organization, which has had
years to develop their proposal to accomplish the work.
Outsourcing and A-76 competitions are seldom initiated by an installation.
When the installation initiates these A-76 competitions, it is at the insistence of the
higher level Army Headquarters whose desire is to reduce costs. The historical cost
reductions described in the literature review drive the external pressures to conduct
these competitions. The in-house organization has years to prepare their Most
Efficient Organization proposal and has years of experience in performing the
function. Thus, the strategy of a single large acquisition is perceived to offer some
advantage to the in-house organization.
If the A-76 competition awards the work to a private-sector firm, it is
recompeted only among contractor firms, usually on five-year cycles. The
government does not offer an in-house organizational proposal for the follow-on
solicitations. Once the decision is made to outsource, government organizations are
not in a position to bring the work back in-house. Rock Island Arsenal decided that,
for the second generation of contractor-provided services, it would break the single
large contract into several smaller contracts and compete each separately. There
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were several reasons why this course of action was chosen. First, using smaller
contracts, the bidding could be opened to small and disadvantaged businesses to
undertake the less complex and capital-intensive functions. Secondly, using multiple
firms meant that, if one particular function was being performed at an unacceptable
level, that contract alone could be terminated and opened back up for solicitation. In
the single large contract approach, this was not possible. Finally, the Arsenal
changed the entire approach to managing the agents. By creating multiple contracts
with multiple agents, Rock Island decentralized the management and placed the
responsibility for administrative oversight and direction with the principal
organizations having the technical knowledge and the responsibility for its
accomplishment.
Figure 11 reflects how Rock Island Arsenal evolved a second-generation
principal agent organizational structure. The smaller contracts were aligned and
managed directly by the functions they supported. The knowledge of what is
required, the oversight of production and the responsibility and accountability lay
within a clear line without an intermediate administrative layer. This has multiple
firms acting as agents to multiple principal organizations. The principal staff
believes this has resulted in a higher level of quality and overall responsiveness by
the agents. They are now singularly focused on their individual functional
requirements and are not subsumed into the larger omnibus organization, as when
there was a single provider.
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 11. R ock Island A rsenal Second Generation
Principal Agent O rganization
In 1997, Rock Island Arsenal was directed to do another A-76 competition
covering the entire organization, including its core-manufacturing mission along with
its base operations missions. This would have placed the entire Arsenal under a
principal agent model; and, if the work were awarded to a contractor, it would have
been an extension of the “Hollow Government” concept to its ultimate limit. This
model already exists in many of the ammunition manufacturing plants that are
government owned and contractor operated. However, in 1999, political pressures
caused a change in direction, and the core-manufacturing mission was removed from
consideration for outsourcing under A-76. This left two functions: (1) a small
logistics function of about 15 man-years and (2) a much larger information
technology function of about 150 man-years for study. In September 2001, after
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
going through the A-76 competitive process, government employees won the
competitions and retained both functions for in-house performance. In both cases, a
principal agent structure will be installed.
For the information technology function, the adopted structure has the agent
as a separate organization under the Director of Information Technology’s direction.
There was a designated program manager specified in the work statement, regardless
of whether the performance is by government employees or contractor. This program
manager is responsible for all the production functions of the agent organization.
Alongside the agent organization is a retained GIN organization whose responsibility
is to oversee the quality assurance evaluation (QAE) and Contracting Officer’s
Representative (COR) functions regardless if the agent was an in-house government
or contractor organization. There are a limited number of other GIN functions, such
as budgeting and long-range strategic planning, retained within the Office of the
Director of Information Technology at the organizational pyramid peak. Figure 12
shows this organization with all government employees, those functioning as the
principal and those that are the agent.
The smaller logistics function is structured similarly to the information
technology organization—government employees function as the agent, and the
principal is a government employee in a separate organization. What makes this
structure unique is that, within the larger organization, there are multiple agents,
including both contractors in one organization and government employees in a
second, both under the direction of the principal. This is shown in Figure 13.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dir. of
I n t o T e c h n o l o g y
'.GIN Function'^.
A g e n t
Prog. Mgr.
Into Technology
I
Figure 12. Rock Island Arsenal Third Generation
Principal Agent Organization
D f j f
S K F iR f io n s
Figure 13. Rock Island A rsenal Fourth Generation
Principal Agent Organization
T
A g e n t
Function 1
J L
3 [
A gen t
Functio-n 2
OAE/CQR
Furictiorv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Rock Island Arsenal is somewhat unique in its evolution under the A-76
process and the principal agent model. Because It has been Involved in the process
for many years, the organizational structure has evolved to accommodate about every
variation imaginable. The intermixing of contractors and government employees as
agents across a multitude of functions has created an organizational structure that fits
the principal agent model but is much more complex because of the number of
different agents Involved. Rock Island has implemented the principal agent model
over two decades in multiple situations. The respondents stated they have been able
to successfully implement the principal agent model throughout the organization, but
It has been a long and evolutionary process.
Red River Army Depot
Red River initiated their most recent A-76 studies in 1997. The installation
undertook two separate studies—one for Information technology and a second for
public works functions. After two years of study and going through the sohcitation
process, no firms offered proposals for either function. At that point, Red River
canceled the A-76 studies and moved directly to their MEO. This reduced the staff
size by about 30% and changed the organizational structure. Figure 14 shows the
structure of the Director of Information Management organization after
implementation o f the MEO.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Oiuf
I r t l a i p w t
SINFiikS b
A i i S f s i w i .
f f l l B i i S
ir ts A ii System s H
A |ilfyB £lis
Istsjratei TscMogy D »;
ApirttaStts
1
_ L _
Ajaiil
Function!
Figure 14. Red River Depot Information Technology Organization
The Information Systems Division and the Integrated Technology Division
perform the contractible functions identified in the PWS. This organizational
structure is similar to what existed before MEO implementation, simply reduced in
staff size. If Red River had contracted for the PWS outputs, then the Director of
Information Management and the Admin Systems Division would have become a
single organization. The Director of Information Management is also the supervisor
of the Admin Systems Division; so, in many respects, it operates as a single
organization. The two divisions that operate as the agent—Information Systems
and Integrated Technology—are viewed as detachable organizational blocks that can
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
be replaced by a contractor. The QAE function is in the Admin Systems Division,
which performs the monitoring of the agent organizations.
The Red River Directorate of Public Works has a structure similar to the
Directorate of Information Management. Certain subordinate organizations operate
as the agent alongside the organizations that perform GIN functions. The program
management approach at Red River is different from Rock Island, when government
employees perform the agent function. At Rock Island, there is a single program
manager designated to oversee the agent organizations. Red River chose not to
consolidate the agent functions under a single program manager but clung to the
more traditional structure, where each subordinate organization reports directly to the
Director—the senior principal staff person. Although Red River went through the
process of A-76 and was moving towards a principal agent model, it chose a hybrid
organizational structure that was more traditional. The agent functions were not
consolidated under one manager, but remain spread through several subordinate
division-level organizations.
Anniston Army Depot
Like Red River, Anniston began it’s a-76 studies in 1997. Anniston also
initiated two separate studies-— one for information technology and one for public
works. Anniston had a large number of bidders, but the in-house proposals were
selected for both functions. Anniston reduced the staff levels by slightly more than
20%, but the organizational structures remained intact. After the decision was made
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
selected for both functions. Anniston reduced the staff levels by slightly more than
20%, but the organizational structures remained intact. After the decision was made
to retain the government work force as the agent, some minor changes to the
organization MEO occurred. For the most part, this involved retaining certain GIN
functions intermixed in the lower level organizations with many of the MEO agent
functions. This intermixing of GIN and MEO functions has eliminated the clear
lines of a principal agent relationship. Figure 15 shows the information technology
organization at Anniston.
O in o f
Figure 15. Anniston Depot Information Technology Organization
The Director of Public Works organizational structure is similar to the
information technology structure with no clear delineation between principal and
agent and an intermixing of GIN and MEO functions across the organization. The
DPW also included several subordinate organizations that were excluded from the A-
1 0 9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76 study. They were considered supporting the maintenance mission yet integrated
within the base operations DPW organization. This intermixing of GIN and MEO
organizations and functions within the organization makes development of a strong
principal agent relationship difficult, if not impossible. Figure 16 shows the
Anniston Directorate for Public Works organization unchanged after the A-76,
simply with fewer employees.
Dir, o f Public. W Vortcs
Planning and
R esource 'Office
Equip fVfaoasjem ent
Account ability Div.
W op d , Rairtt and.
W eta 1 D ivf s ion'
Mobile Equipm ent
'MtainftprtancP Division
R oads and Grounds
Divisions
Electric, Refrigeration
and Pitsmbitig D*v.
Fabrication Division
GIN Mission Support
Prod met ion fE .ciu .iP :
fWlairttdnance' Division
■GlN: M ission Support;
Figure 16. Anniston Depot Public Works Organization
Anniston did not set up a QAE/COR function within the organizations
performing the work. Instead, they assigned the QAE/COR to a staff of two
individuals within their Resource Management organization. The individuals
assigned this task were intimately involved with developing the PWS and conducting
the A-76 study but are not technical experts. The DPW organization employs over
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
there is no external work count reporting system. This is in contrast to Rock Island
and Red River, where the QAE function was an integral part of the principal’s effort
to manage the agent. This emasculation of the QAE function and the intermixing of
principal and agent functions has inhibited the development of the principal agent
structure. Anniston operates in the traditional bureaucratic structure and has made
minimal progress migrating to the principal agent organization.
Letterkenny Army Depot
Letterkenny was in the midst of their A-76 study solicitation during data
collection. The study progress had been delayed by almost two years for a variety of
reasons. The interviews with management showed a strong reluctance to the A-76
process, a distrust of contractors (and, by implication, the entire principal agent
model), and a belief that the process was not in the organization’s or the employees’
best interests. There was a concern that, should the in-house MEO win the A-76
competition, the work force would be unable to adapt, causing production problems
that could not be overcome. The alternative of having a contractor work force was
viewed as even less desirable, because an in-house MEO would allow the existing
employees to maintain their employment, seniority and benefits.
There was resentment by both employees and management that not all of the
installation organizations were subject to the A-76 process. This appeared to cause
morale and productivity problems. The MEO was developed with the minimum
staffing set at the lowest pay grades to accomplish the work. Although the in-house
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
morale and productivity problems. The MEO was developed with the minimum
staffing set at the lowest pay grades to accomplish the work. Although the in-house
proposal had not yet been made public because the competition process continued, it
was causing resentment among the work force. These staffing and grade reductions
were not done in the organizations not being studied under A-76.
Figure 17 shows the existing organizational structure of Letterkenny Army
Depot. It is a vertically integrated organization aligned by function reporting to the
installation commander. In the 1980s, data processing, public works and a few
smaller functions had undergone A-76 competitions and were retained in-house.
Only the telephone system operation under the A-76 process was awarded to a
private-sector firm—Coredev Corporation. This five-person function operated in the
principal agent mode, under a single government employee’s direction.
Traditional Cotstmaiid .Orgasiizmioaai Structure
1 ^ ^ 'ia ^ ‘ S rirn a S ^ -
M is s io n
O rg an e^o tr# i
Principal Agent Organization w a d tr A - 7 6 Process
Conta'actoal. Jaiaftgranaai.
1
Mission Mission Principal Agent
Organisation ®t ■Oigahfeatiof} W2 OrganizMleri
Organization
Figure 17. Letterkenny A rm y Depot
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For this A-76 competition, Letterkenny had planned on placing the majority
of the base operations functions within the MEO. These functions included: public
works, resource management, information technology, and many miscellaneous
functions. This concept followed what Rock Island Arsenal had done in its first
generation approach to A-76. There would be one, very large agent organization
performing multiple functions, reporting back to a principal organization with
responsibility for QAE/COR oversight. These organizations would work alongside
the remaining GIN and mission organizations.
As the study progressed, many of the functions were removed from the A-76
competition until basically only public works remained. These changes were
directed by headquarters organizations that believed specific functions should not be
contracted. The MEO was viewed as a stand-alone organization, whether it was
staffed with government or contractor employees. Even though the study scope was
reduced, Letterkenny did not follow the second-generation Rock Island approach of
integrating the MEO into a functional organization. Letterkenny chose to stay with a
principal that split the responsibility for technical decision-making in one
organization and the administrative oversight responsibility in a second organization.
In Figure 18, the decision-making responsibility for such things as repairs to specific
buildings is in the Director of Public Works. The organization performing the work
is the MEO Organization, while the organization responsible for oversight and
monitoring the agent is the Technical Oversight Group, QAE\COR function. The
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
structure is very similar to what Rock Island put in place in their first generation o f
A -76 competitions.
I rsstasffattoo C o m m a n c te r
I 'T e m p o r a r y ' o r g 1 —
M E O . O r g a n iz a t io n T e c h n i c s o v e r s i g h t ' © r o u p
© I N f c e a e o p e r a t i o n s
. Q A E t e O R -fu n c tio n
D ir . o f fvf S n t e r e i n c e
Cat f« 5 c o r e " m issio n
D ir..o f. P r a d u c t ' A s s u r a n c e
D ir . or
1 r ifa -m atforT M am sa g em e n t
D ir. o f R e s o u r c d s
■Dir. .cf L a w D ir. o f C o n t r a c t in g .
€3! N'. fea sts o p e r a t io n s '
O iN ■ ■ P e s © o p e r a tio n s -
Dir.. o f .F * u b iic . ■ W o r k s
© i M ' b a s e o p e r a t io n s :
Figure 18. Letterkenny Depot Organization
Questionnaire Data
The questionnaire was divided into several subsections. The first, General
Environment, established the respondents’ positions in the organization, their
backgrounds and how their organization was structured. The second section was
designed to gain an understanding o f how satisfied the respondents were with the
principal agent structure in their organization. The third and final section asked
about each o f the principal agent model variables and specific questions on how w ell
these were understood and being managed.
1 1 4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
General Environment
The general environment questions established who the respondents were and
their understanding of the principal agent model The majority of respondents were
staff personnel who worked within the principal organization. The size of their
respective agent organizations ranged from a dozen employees to approximately 250.
At Rock Island, where the agent organization Is contracted, the principal staff has
read the performance work statement, performance plans, and has received formal
training in their responsibilities. At the Red River and Anniston locations, the
principal staff had read the PWS but had no training for their role in oversight In the
agent organizations. At Letterkenny, personnel who would be in the principal
organization had not been Identified; there was no plan for formal training; and
virtually no one had read the PWS.
At the sites where government employees functioned as the agent, there were
three indicators that the principals took their responsibilities less seriously than if a
contractor had been the agent. First, no formal training in the oversight role of the
principal organization was provided at any location. TMs was the opposite of when
the agent was a contractor and most of the principals had some formal training.
Secondly, there were no written procedures followed on how the principal was to
manage the agent. The understanding was that the agent was expected to perform the
outputs In the PWS. This was in contrast to the formal Quality Assurance Plans and
inspection plans used when the agent was a contractor firm. Finally, the cost of
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
monitoring and the plan to monitor the agent was more formalized and more
stringent when the agent was a contractor.
Satisfaction with Current Structure
Satisfaction with the existing structure was consistent across all locations.
Regardless of whether the agent was a contractor or government employees, there
was a consistent high satisfaction level response. Once the A-76 study process itself
was completed and the transition to the end state organization in place, the
satisfaction level was high at all locations for ail functions.
Letterkenny has not yet completed the A-76 competition; there was no agent
selected. Letterkenny expressed extreme dissatisfaction with the entire process and
believes it will not improve performance. The dissatisfaction had little to do with
actual service delivery but was rooted in the concern for the impact on employees.
The perception was that the wage and benefit package offered by the federal system
would exceed that of any potential contractor. The employees and management were
concerned about the impact on individuals’ retirement, wages and benefits should a
contractor win the competition. There was concern about performance, regardless of
whether it was by a contractor or government employees. If the agent was a
contractor, the concern was there would not be adequate trained and skilled
employees available to do the work. Staff expressed a concern that the PWS would
restrict the ability to assign work that had been overlooked or unforeseen. This
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
would result in cost increases to the government, and the agent might attempt to hold
them hostage—a concern of excessive post ante transaction costs.
There was also a concern that the government employees would not perform
up to the PWS level. Letterkenny had undergone a massive Reduction In Force in
1999. Many of the public works employees had been placed there under seniority
rules and were only now gaining the foil range of experience and training to work at
a foil performance level. The average age of employees was 48 years old with 23
years’ seniority. They had become accustomed to benefits, such as compressed work
schedules, the ability to take vacation at their discretion, and having back-up and
support personnel. If the organization won the A-76 study, there would be a 20% to
30% staff reduction, and the federal personnel RIF rules would apply. The most
qualified individuals could be displaced by less qualified individuals. Older workers
with only 1-2 years left until retirement could displace the younger workers. The
luxury of having multiple trained personnel could go away. The need to focus on
performance metrics was of major concern. As one senior manager said: “I am as
concerned if we win A-76 as if we lose.” If a contractor would win, there is a
contract in place and vehicle to demand performance. If the government would win,
the result is a decrease in staff, a lower wage scale for employees, less flexibility in
the work environment, and a higher performance level expected.
At Red River and Anniston, the federal employees comprise the agent
organization. The satisfaction level at these locations continues to be high (average 6
on a 7-point scale), despite the reduction in staff by almost 30%. The primary issue
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
brought up was multiple individuals that provided in-depth support were no longer
available in many areas. In the past, there were several individuals capable of
performing a task; the reductions left the organization one deep in many areas. TMs
caused concern that, when a key person departs or becomes unable to work, the
organization has no backup support. There is also no ability to bring in additional
staff to train as replacements. The establishment of fixed personnel ceilings and
costs under an MEO structure limits the organization’s ability to meet changes in its
environment. Finally, management is forced to manage the work force much more
closely. In the past, employees could take vacation time as they pleased. With the
MEO in place, there is close monitoring of who can be released, closer monitoring of
work effort and outputs, more need to prioritize and schedule work load, and much
less ability to respond to surges in work or unforeseen events. These are the same
issues expressed by the Letterkenny management. The Red River and Anniston staff
stated they were working through these issues; their concern was for the future, when
the current staff began retiring and their ability to attract a replacement work force.
An Anniston manager stated: “We won this time, but we will never win again.”
At Rock Island, the contractors have been the agent for many years in three of
the five functions. Current satisfaction level is high in each of these three functions
(6 out of a 7-point scale). However, the respondents discussed earlier experiences
with several contractors that were not as positive. They mentioned instances where
the contractor hired less than fully capable staff, had not been fully knowledgeable of
the PWS requirements and had not been responsive to the requirements of the
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
principal. The remediation was to hold extensive sessions with the contractor to try
and improve performance. A second step was to implement the contract penalty
clauses; and, at least once, there was a change in contractors. This did not happen
often but is reflective of past experiences. The current contractors, acting as the
agent, are providing acceptable performance levels and equal the rating provided at
Red River and Anniston, (6 out of 7-point scale). There are no evaluations of federal
employees as the agent at Rock Island; they have not yet moved to the MEO and do
not have a performance history under the principal agent model.
Variables
Each study respondent was asked specific questions on the variables that
influence the principal agent model. The respondents understood the general
concepts of each variable; however, they had not put them together looking at how
they influenced the total organization’s operation. On many of the variables, there is
no data source, and the respondents made generalizations and observations. Each
variable is discussed in detail, and a summary of the results is shown in Table 4.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4
Data on the Variables
Rock Island Red River Anniston Letterkenny
Asset
Specificity High High High High
Goal
Alignment Not Always Always Always Not Always
Moral
Hazard Some None None Expected
Information
Asymmetry Some No Issue No Issue Unknown
Transaction
Costs Unknown Unknown Unknown Unknown
Incentives &
Coercives Coercives None None Coercives
Asset Specificity
Asset specificity is the degree to which the output can be specified. For all
locations, the agent provides support services with multiple outputs. Development of
the PWS is the most time-consuming aspect of the entire A-76 process, often taking a
year or more. To do it right requires installing a work count system that develops
valid historical records of each output. Time constraints prohibited this in all
locations, and interviews of the work force to gain approximate quantities work were
used.
The agent’s required outputs are documented for each location in the PWS
for each function. The PWS defines both what is required and a level of acceptable
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
quality. In the case of Letterkenny, there were almost 1,100 different outputs
identified. The A-76 solicitation approach defined the output desired, not the method
for accomplishing the task. For example, the output defined may be “Process
Reports of Discrepancy, Form 364 in the estimated quantity” or “Perform preventive
maintenance on all boiler and plumbing IAW applicable technical specifications.”
The basic PWS was 268 pages in length. In addition, there were eighteen
attachments listing such things as approved chemicals for use, vehicles requiring
maintenance, and a list of applicable technical manuals and mandatory Army
regulations. There were an additional thirteen sections of Contract Data
Requirements Lists (CDRL) requiring the winning organization to submit
performance reports. During the solicitation period, the bidding contractors
submitted a series of questions requesting clarification of the PWS that resulted in
six amendments to the solicitation.
The experience of Letterkenny was typical at all four installations. Without
stating how to perform the work, the principals defined the required outputs and
quantities in sufficient detail to enable them to manage the agent organization. The
principals prepared to manage a contractor as the agent because of the historical
precedent that contractors won 50% to 60% of the solicitations. With that in mind,
the PWS was developed in significant detail with required timeliness and quality
indicators for each output. This established high asset specificity. Rock Island and
Red River responded with a score of 7, and Anniston responded with a 6 (out of 7),
when asked if they felt the outputs had been adequately specified. Similar responses
1 2 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of 6 and 7 were reported, when asked if the MEO/agent clearly understood the
requirements. There were no instances identified where there were
misunderstandings. On the question on the adequacy of the work count system
capturing the output data, Anniston and Red River were lower in satisfaction—5 out
of 7-point scale—while Rock Island evaluated theirs at a 6 or 7.
Goal Alignment
Goal alignment is the measure of how well the goals of the agent and the
principal match up. At Rock Island, the questionnaire data revealed that performance
of the PWS was the primary goal of both principal and agent. When asked about
commitment to performance of the PWS, the values for both principal and agent
were in the upper 6-7 on a 7-point scale. When asked about a goal to exceed
performance levels in terms of timeliness or quality, the principal organizations were
neutral, indicating basic levels of satisfaction. The agent organizations had little
interest in improving performance levels. The obvious area where there was direct
conflict between the goals of the principal and the agent was in profit maximization
and cost reduction. The principal had a mid-level desire to reduce the cost of
performance, which would typically reduce the agent’s staff levels. The agent, on the
other hand, had a mid to high level of desire to maximize profit and grow the
business base. This was not viewed as detrimental to the working relationships but,
rather, as a natural state for any profit-seeking organization.
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At Anniston and Red River, the consensus was that the goals of the principal
and the agent were highly aligned. In both cases, the agent organizations were folly
Integrated Into the principal organization from which they originated. The agent
organizations viewed themselves as Integra! to their function of information
technology or public works but also the larger depot mission of Army maintenance.
The employees remained a part of the personnel system that set wages and the
performance rating system. Other than having fewer co-workers in their function,
very little changed for these employees. In addition, there was no alternative
organization for them to transfer their loyalties and goals. They had no corporate
relationship, no profit maximization opportunity and no ability to grow the business
base, so there was no viable alternative to which the individuals or the agent
organization could align goals that might be in conflict with the principal. The
Anniston approach of intermixing the principal and agent functions at al levels
within the organization further served to align the goals of the agent with higher
order goals of the principal. Both organizations reported a score of 7 on questions of
goal alignment.
Letterkenny is approximately one year from implementation of the principal
agent model. The respondents have little concern of goal alignment, If the existing
federal employees win the A-76 competition. Their responses indicated a possible
better alignment than exists today. The current state of uncertainty, who would have
employment, what their pay rate would be, and the resentment towards those not
under the threat of losing their jobs has created an environment where the desire to
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
meet the larger overall goals of the organization has deteriorated. At times, there is a
defeatist attitude where the employees are resolved that they will lose their
employment, so there is little desire to work for the good of the organization. The
organizations currently performing agent functions continue to operate, hut the desire
to meet the PWS performance levels is limited, because some employees believe they
will not have a job within the coming year.
The Letterkenny staff is significantly concerned about goal alignment, if the
agent is a contractor. The A-76 solicitation is a fixed price solicitation; the
contractor will be paid one price for full performance of the PWS. Under many of
the base operation contracts at other sites, the approach was a cost plus fixed fee,
which guarantees the contractor a profit level and, because costs are covered, reduces
the risk to the agent. The fixed price concept moves the risk from the principal to the
agent. The concerns expressed by the Letterkenny staff are two-fold. First, the
contractor would immediately begin looking for ways to amend the contract claiming
work was outside the scope and, therefore, entitled to additional payments. This was
something that, if the in-house organization won, it would be unable to easily
accomplish. There is very little opportunity for the in-house organization to increase
costs. Secondly, there was an incentive for the contractor to “shirk” and perform
below the desired performance levels, having less personnel than necessary to stay
within the fixed price bid. The Letterkenny staff was concerned that there was
insufficient ability to direct the agent to perform, short of canceling the contract.
This would then lead to a second solicitation, and they were sure the price would
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
increase significantly, because there would no longer be a viable in-house
organization as a competitor.
From the Letterkenny employee perspective, one goal is to have as positive a
work environment as possible. Along with the typical benefits of health care,
retirement and vacation, there were amenities designed to improve family life and
morale. The employees have a flexible work schedule where they can arrive anytime
between 6:00 AM and 9:00 AM to begin their workday. They have a compressed
work schedule alternative where they can choose to work eight- hour shifts, ten-days
per pay period, nine- hour shifts, nine days per pay period or ten hour days, eight
days per pay period. They have the ability to take vacation time in one-hour
increments at their discretion. All of these quality-of-life benefits would be subject
to change, if the contract were awarded a contractor. There was a concern that, if the
contractor would win, the employees would be offered their current jobs working for
the contractor but at a substantially reduced pay rate. With a work force average age
of 48, an economy in recession and limited alternative opportunities within the
community, the employees were concerned that they would take a substantia!
reduction in wages and about the impact this would have on their life styles and
families. In summary, both the employees and the management at Letterkenny had
significant concerns about the future alignment of goals, if a contractor won the A-76
solicitation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Moral Hazard
Moral hazard occurs when the agent takes an inefficient action, because his
interests are not aligned with those of the principal. Moral hazard can occur as a
result of efforts for profit maximization or employee shirking. Moral hazard for
profit maximization cannot occur, when the agent is an in-house government
organization. There is no profit opportunity for the government employees;
therefore, this type of moral hazard cannot occur at Red River, Anniston or at Rock
Island, where the agent is a government organization.
At Rock Island, when the contractor was the agent, the questionnaire data
revealed that, for the most part, the agent tried to perform the work in an acceptable
maimer with a reasonable profit margin. There were some instances when the agent
tried to maximize profit levels. One example was the use of an inflated overhead
above allowable levels that would increase the profit margin. A second example was
the recommendation for a minor construction project that was unnecessary and
would have also increased profit levels. These examples were seen more as
aberrations in the day-to-day operations, rather than a standard way of doing
business. The principal organizations believed they knew of the opportunities where
such things could occur, and the close monitoring of the agent reduced the likelihood
of this occurring. The monitoring that found such efforts of profit maximization also
deterred future attempts.
Letterkenny is concerned about the potential for profit maximization.
Because there is no decision on who will be the agent organization, there is no
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
working relationship and only concern for possible pitfalls and problems. One of the
pillars of Letterkenny’s strategic plan is to reduce operating costs. This A-76 effort
is seen as one way that will reduce costs, regardless of the winner. However, the
fixed price approach means, if a contractor wins, the only way to increase profit
levels is to add costs to the contract or to reduce costs below what was anticipated
when the bid was made. Letterkenny management believes their in-house proposal is
low enough that any competitor will have very little profit margin. For that
competitor to meet the expected market level profit margins, they will need to
somehow expand the scope of the contract and increase the costs of doing business.
The interview process confirmed a fear of the unknown at Letterkenny, compared to
an existing operation at the other sites.
The second type of moral hazard is shirking, when the agent avoids some
level of work. Respondents from all sites stated this was not a significant problem.
On a scale of 1 to 7, with 7 being insignificant, the response was generally a 6 or 7.
Similar to the response to the questions on profit maximization, there were isolated
instances identified, when the Rock Island contractor agent did shirk. They did not
move some equipment and perform certain tasks when they should have. Whether
this was a deliberate failure to perform or an error in performance is unknown. The
check on shirking is the use of spot inspections, client feedback and the principal
having a good quality control effort. In the case of Rock Island, the principal agent
model has been in place for many years, and the mature relationship and experienced
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
operation under this model has enabled them to build in checks and balances to limit
the opportunities for moral hazard to occur.
At Red River and Anniston, the concept of shirking by the MEO employees
acting as the agent was considered very remote. The employees were integrated into
the organizational structure, and there was no advantage to be gained, if shirking did
occur. It was viewed as management’s responsibility to see that this did not occur,
just as it had been prior to A-76 bringing the principal agent approach into place.
This did not mean there was not limited shirking before the A-76 study or that it
would never occur again. The respondents indicated that there was no benefit to the
MEO as an organization, and they saw no evidence of it occurring. However, the
work count systems and quality monitoring did not exist in similar detail and
oversight to what Rock Island has put in place. In the follow-up interviews, the
respondents said there could be some shirking occurring without their knowledge
because of limited monitoring, although they believed this was unlikely.
Letterkenny was not concerned with shirking as a problem. Its intention was
to utilize the same work count system and performance monitoring regardless as to
whether the agent was a contractor or government employees. For each output, they
have expected performance levels and feedback mechanisms as part of their control
measures. Their belief was that the Quality Assurance Plan and planned monitoring
were sufficient checks to insure shirking would not occur. The fixed price contract
meant that, as long as the work was performed, it was the agent’s responsibility to
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
insure that the employees worked diligently. The principal’s responsibility was to
insure that the agent organization met the performance criteria.
Information Asymmetry
Information asymmetry occurs when information is not shared perfectly
among the principal and agent. At Red River and Anniston, the agent was integrated
into the larger functional organization with direct oversight by the principal. Work
direction, quantity of outputs, quality and technical information flows horizontally
and vertically within the organization. There was no benefit for either the principal
or the agent work groups to withhold information. None of the respondents at either
site indicated there was any problem with accessing and sharing information. It
appears the integration of the organizations forces a sharing of information of all
types.
Rock Island has encountered some information asymmetry in their principal
agent operation. When the agent is a contractor, there are times when the technology
of the function has made significant changes, but the principal has not kept abreast.
The principal is focusing on performance in accordance with the contract PWS,
which is to be in place five years. During that time, the technical expertise within the
principal may leave the organization and not be replaced, or the agent may become
more technically advanced and the principal may not keep up. This is a problem
during the execution years of the contract but becomes more significant, when the
function needs to be re-solicited after 5 years. At this point, the principal no longer
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
has years of experience in performing the work and is no ionger the expert in what
needs to be done. Rock Island found this lack of hands-on experience causes
problems in finding ways to improve the PWS and eliminate less efficient processes
and unnecessary outputs. The scope of the problem is related to the competence of
the QAE/CGR oversight organization. If this QAE/COR organization has devolved
to an administrative function, collecting work counts and monitoring customer
complaints, the technical expertise becomes absent. However, if the QAE\COR took
a very active role in maintaining technical expertise with a thorough knowledge of
the performance of the function and an active involvement in what is required of the
agent, then this problem could be overcome.
Letterkenny did not foresee problems with information asymmetry. It
expected significant reporting requirements and the retained GIN level of expertise to
offset any potential problems. Although Letterkenny removed information
technology from the A-76, and the other three sites retained it for in-house
performance, this was one area discussed as having the potential for an information
asymmetry problem. The information technology field is changing very rapidly,
requires a high level of expertise and has multiple niches of specialization; this could
create an information asymmetry problem. The principal organization was
envisioned to be very small, held by the senior person through RIF procedures. This
did not necessarily put in place the most qualified QAE\COR staff. If this were a
small staff, any departures would create knowledge gaps that could not be easily
filed. At a more macro level, the Army Materiel Command had placed a hold on
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
further A-76 studies of information technology organization, because of these kinds
of issues. This was the prime reason Letterkenny dropped information technology
from the A-76 study.
Transaction Costs
Transaction costs are those costs associated with instituting the principal
agent relationship. The questionnaire asked about the ex ante, ex post and the
monitoring costs. All respondents had a limited knowledge of these costs. For the
ex ante costs, each location used a consultant for developing the PWS and contract
documents. The consultant’s cost typically approached $1 million per site.
However, the consultant worked side by side with a full-time C A Office of federal
employees. These CA Offices were usually set up on a temporary basis but could
exist for several years before being eliminated. The staff could range from 6 to 10
full-time personnel supplemented by up to a dozen part-time personnel. In addition,
the people working the A-76 study conducted countless interviews of the work force
to define the outputs and performance factors for the PWS. There was some attempt
to capture the federal employees’ labor hours expended on the A-76 study. When
asked if this captured most or all of the expended labor, the answer was “no, it did
not.” The respondents did not know what value this data would have; although, at
the end of the A-76 process, a final report would include a dollar figure for the
study’s cost. The consensus was that the cost of the stud)*’ far exceeded any number
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reported and the value of the effort. The preference was a directed 20% staff
reduction and to be finished with the entire A-76 process.
One key point made during the interviews was that the entire A-76 process
was forced upon the installation and was not optional. The ex ante costs were
somewhat irrelevant, because there was no funding to support the study effort but,
rather, a reallocation of personnel resources within the organization. The original
plan from each installation was to maximize the use of the consulting contractor, but
DA limited the funding to below what the installations felt was needed to properly
conduct the study. This forced the installations to assign additional staff to support
the study process. Each of the installations saw as their primary goal to win the A-76
competitions and retain the jobs for their co-workers. Therefore, assigning additional
staff to complete the process and build the MEO was given a high priority. However,
tracking the labor hours and the associated costs were not a priority.
In a separate discussion with senior DA staff, they had a need for good cost
data on all aspects of the A-76 process. They felt there was a great value in knowing
what the ex ante, ex post and monitoring costs were. The A-76 process was seen as a
way to reduce total costs to the Army; and, without good cost data, they could not
substantiate the value of the program. There was a general feeling that the
installations were not closely controlling the study costs, and expeditiously bringing
the studies to closure, thereby increasing the ex ante costs. Many costs were never
recorded, and there was a lack of consistency in the cost data.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
All installations had a general idea of their monitoring costs. These were
typically quoted in dedicated staff rather than dollars. If the contractor was the agent,
then a COR function was staffed to insure compliance with the contract. At
Letterkenny, Figure 18 shows a stand-alone Technical Oversight Group to serve as
the QAE\€OR, regardless of the agent. The plan is to develop unique job
descriptions and staff the organization for monitoring the agent. This was different
from the Rock Island approach that decentralized the monitoring function. Where
the agent is a contractor, Rock Island had staff within the larger parent functional
organization, whose function was monitoring the agent. When the agent consists of
federal employees, Rock Island has the monitoring function staffed in a closely
aligned, but separate, organization.
Red River did not set up a separate monitoring organization but placed the
personnel within the larger parent organization. Red River’s organizational approach
of having an agent organization that was separable from the larger parent
organization made it logical to place the monitoring function in the parent
organization. Anniston chose a different approach. With the in-house win of the A-
76, they chose to assign a very small monitoring staff. They placed it in a separate
organization with the premise that the work count report was the primary oversight
requirement. At each installation, the monitoring costs can be calculated from
existing records in man-hours and then converted to dollars. However, there was no
budget identified or upper-level management attempt to reduce these costs.
Interviewees reported that no one even asked what the monitoring costs were.
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The third type of transaction costs investigated were the ex post costs of re
negotiation. One subcategory is any renegotiation for non-performance or for
changes in the scope of work. A second subcategory is the cost to conduct the
competition process at the end of the five-year performance period. In the first
category, only Rock Island had experienced the need to renegotiate for failure to
perform. There were no records and no idea what the cost of this renegotiation effort
was. The labor accounting system did not capture costs at this level. In the second
category of changes in scope, there were instances at Anniston and Rock Island
where this had occurred. In both cases, there were no records and no estimates of the
costs of the renegotiation. They could tell the cost increase for the performance
changes but not the cost to negotiate.
The final ex post transaction costs are for renegotiation at the end of the
performance period. Rock Island has gone through this process several times but has
no estimate of the costs of renegotiation. The process consists of revalidating the
PWS, making changes as necessary and then competing the work. Rock Island does
not track the costs to enable identification of the Individual solicitation costs. It is
much less than starting from a zero base, where there is no PWS, and does not entail
developing the In-house proposal MEO; so the costs are much less, If the function is
already under contract.
For every installation that has an in-house MEO with government employees,
there is a mandatory review by the Army Audit Agency each year. Because Army
Audit is not part of the Installation, the installation has no view of the costs of these
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
audits. The installations do not track the supporting manpower and costs dedicated
to these follow-up audits. This is another category of transaction monitoring costs,
which the Army incurs but which is not captured as a cost.
Letterkenny, Anniston and Red River had conducted A-76 studies on the
majority of their public works functions two decades earlier. For the most part, the
work conditions had changed so significantly in terms of technology and
requirements that the efforts to develop the PWS and MEO from that era were
useless. Red River also stated that the work count system it went to after
cancellation of the A-76 actually will make it harder to develop a PWS in five years.
The work count system they had historically used has been reduced, and it will have
less detailed data in five years to conduct the follow-on A-76 competitions.
Letterkenny, on the other hand, said it was going to a much more detailed work count
system that would make it much easier to develop the PWS in five years.
In summary, none of the installations know or track the transaction costs.
They do not attempt to manage or minimize these costs in a concerted effort and
view the requirement to compete the function in five years as too distant to worry
about today. There is no plan to minimize these costs in the future and no plan to
deal with the follow-on A-76 competitions to reduce the transaction costs.
Incentives and Coercives
Incentives and coercives are those actions taken to encourage performance or
discipline poor performance. For those installations where the MEO is the in-house
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
government organization, there are no incentives to performance. As one manager
said: “The incentive is they have a job.” Typically, under the principal agent
structure, there would be an incentive structure to encourage and reward the agent for
performance above a minimum level. In many contracts, this would be a
performance award; but, where the agent is the in-house organization, this does not
exist. At Letterkenny, Red River and Anniston, cash performance awards are made
to ail employees—both the principal and agent groups—based on the depot’s
financial results in their core mission. The agent organization employees have
virtually no impact on this financial metric, but this installation-wide award system is
in place at all three locations.
For installations where the agent is a contractor, there can be incentives to
performance, but none were found in use. There are contracts that include an
incentive award fee for performance that exceeds the minimum levels. In all four
installations, the approach was a fixed price or cost plus contracts and did not include
incentive awards. The objective of the principal organization is to limit the contract
cost, and none of the installations developed an incentive scheme to reward and
motivate performance. The overriding idea is that a price had been agreed to; the
monthly payment is incentive enough for performance. The principals believe the
contract requirements will be met, or payment withheld, and that acceptable
performance limits have been set, so there is no need for incentives.
There are coercives in these contracts. There is a payment deduction for
minor performance shortfalls; and, ultimately, there can be a contract cancellation for
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
major problems. Rock Maud has used both types of coercives over the years. They
have imposed financial penalties and have also cancelled a contract for non
performance. Where the agent is the in-house government organization, there are
few coercives for poor performance. There is the ultimate coercive of termination
and moving the function to a contractor, although this has never been done. The
Army Audit has this authority, if the MEO failed to perform or if the installation has
dedicated more resources than the original proposal. The penalty of termination
moves the burden back to the principal organization to go through the entire
solicitation process again. This further increases the transaction costs and creates the
personnel and political turmoil such a decision brings. The preference is to attempt
to cure the performance deficiencies and continue with the current performing
organization.
In discussions with all four sites, the concept of coercives for the in-house
MEO seemed foreign. The expectation was that the organization would meet
performance levels, and there was no need for coercives, except the possible ultimate
termination. The senior management mentioned the federal personnel rules, existing
labor management contracts and other limitations that limited their ability to develop
coercives. For the most part, this was a topic that apparently had not been thought
through the way it would have been if the agent were a contractor.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER VII
ANALYSIS
Knowledge must be gained by ourselves. Mankind may
supply us with the facts; but the results, even if they agree
with previous ones, must be the work of our mind.
(Benjamin Disraeli, n.d.)
Background Information
During the 1980s, all four installations conducted A-76 competitions and
outsourced the operation of their telephone system. Rock Island Arsenal competed
several additional functions, such as vehicle maintenance, facility maintenance, and
mail operations. These all went to contractor performance under the principal agent
model. The three depots held A-76 competitions on similar functions but kept them
in-house and maintained a traditional organizational structure. Although the Army’s
regulatory guidance was to treat the in-house wins similar to a principal agent
structure with oversight monitoring and performance measurement, it did not
happen.
During the 1980s, A-76 competitions were planned as a continuing way of
operation, and the installations established Commercial Activities Offices. However,
with the Congressional restrictions and the reduced emphasis on A-76 in the late 80s
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and early 90s, these staffs were disbanded. In the latter part of the Clinton
Administration, there was a renewed emphasis; and, for the past four years, each
installation has been actively engaged in A-76 studies. Each installation engaged a
management consultant to develop the Performance Work Statements and ancillary
documents. Government support to the studies was on a temporary assignment basis,
with the intent of disbanding the Offices when the studies concluded.
Each installation is at a different stage in the A-76 process along a time
continuum. Figure 19 shows this difference. Letterkenny is still within the A-76
competition process. In approximately one year, Letterkenny will alter its
organizational structure and incorporate a principal agent model, regardless of
whether government or contractor employees is the agent organization. Red River
Army Depot and Anniston Army Depot completed their A-76 process and have
moved to their succeeding organizational structures. Rock Island has two distinct
groups of organizations. The first group has been under the principal agent model
using contractors as the agent for over a decade (referred to as first generation in
Figure 19). The second group, (referred to as second generation, in Figure 19) has
completed the A-76 process recently and is in the process of moving to the principal
agent model. The second generation group is composed of government employees.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
'Rock Island
Second Generation
Lfette:
A rm j
Red River &'
Anniston Army Depot
Rock Island
First Geaewtloa
A -76.
'tetiafef!
A-76
Coiupetwl
Pnnqpal Agent
IffiM ated
A-76 iH /M
Reeossfsetttioa
Mature Pmcipal h g p t
Figure 19. Time Continuum of Principal Agent Organization
at Study Locations
Pattern Matching
The research question in Chapter 4 was: Does the principal agent relationship
apply to organizations where government has implemented contractual arrangements
for commercial activities? To answer this: Specific attributes and patterns are
expected in the organizational structures and operations. The A-76 process dictates
that a principal agent model be adopted. It would be expected that, in implementing
A-76 in the same functions, there would be a high degree of commonality of
organization and approach. The analysis uses a matrix as a pattern-matching
technique, looking to identify commonalities in organizational structure and in the
variables. Because the data is not all numeric, the valuations are qualitative,
indicating levels of relationship.
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Organizational Pattern Matching
The principal agent model has a principal identified for oversight and
management of the agent. The agent would be a distinct organization with a clearly
defined set of outputs. The agent organization would have identifiable leadership
responsible and accountable for the organization’s outputs and performance. A
formalized work count system with performance reports would ensure the agent was
accomplishing the appropriate workload. There would be formalized quality
assurance and evaluation to insure the agent performed in a satisfactory manner.
These are the organizational patterns that theory would expect, when the principal
agent mode! is implemented under the A-76 study process. Table 5 shows how each
location conforms to the expected pattern when compared to the theoretical model.
Table 5
Pattern Matching of Organizational Structures
r — Theoretical Rock Island Red River Anniston Letterkenny
Organizational
Separation Yes Yes Yes No Yes
Responsibility
Chain thru
Agent Manager Yes Yes Partial No Yes
Work Count
System Yes Yes No No Yes
9 Operational
1QAE Function Yes Yes No No Yes
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Iii Table 5, the theoretical column is what would be expected from
implementation of the principal agent structure. The theoretical model does not
differentiate between using an in-house organization or a contracted external
organization as the agent, nor does it differentiate along a time continuum. Rock
Island matches the expected pattern in all aspects. They have a clear organizational
separation for both those functions contracted and those retained in-house. In each
case, they have established a program manager to be responsible and accountable for
the agent organization’s outputs and performance. To oversee the agent
organization’s performance, Rock Island has implemented a detailed work count
reporting system and a quality assurance function.
Letterkenny has not yet implemented their structure, but the planning process
is on a course for instituting the classical principal agent model. The organizational
concept is the same, regardless of whether the agent is an in-house or contractor
performance. Letterkenny intends to have a separate agent organization with a
separate principal staff. They plan for a work count system of almost 1,100 outputs
and a detailed quality assurance plan to oversee the agent’s work.
Red River and Anniston have not folly matched the predicted principal agent
model structure. In both cases, the agent is an in-house organization. Anniston, in
particular, chose not to separate the principal and agent and retained the integrated
organizational structure that existed before the A-76 process. Anniston intermixed
the principal and agent functions within the organization. Responsibility for
performance has not been clearly delegated and assigned to agent organizations but,
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rather, remains interspersed across the organization. The work count system is in
place; however, there is no consistent external reporting to a quality assurance and
evaluation function. The QAE function is basically non-existent.
Red River has made partial accommodation to the principal agent model. For
the most part, the agent functions were placed within a separate organization and
then integrated witWn the larger parent organization. When Red River chose to keep
the agent organizations within the larger organization, responsibility for performance
was not unified to a single manager of the agent organization. Red River reduced the
work count system to a much smaller level, which makes it difficult to monitor
outputs. The QAE functions were retained in a separate organization from the agent,
but the authority to direct improvements and monitor performance are reduced.
When the agent is or could become an external contractor organization, the
principal agent structure has been put in place. There is no single organizational
structure for the principal agent model. When the agent is a contractor, there have
been organizational options of centralizing and decentralizing the principal function.
Rock Island’s migration from a centralized to a more decentralized approach
indicates certain benefits, although not significant enough to outweigh other options.
What seems to drive the choice of structure is the consolidation of multiple functions
under one contract. When Rock Island and Letterkenny had a large, multi-function
study, they chose a centralized principal. When Rock Island, Anniston and Red
River competed smaller functions, they chose a more decentralized organizational
structure for the principal.
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
When the agent is an in-house organization, there is no standard
organizational structure. The principal agent roles were virtually unique to each
installation. Anniston chose to integrate into a horizontally and vertically integrated
organization. Letterkenny plans to retain separate and distinct organizations
mirroring the structure, if the agent was a contractor. Rock Island and Red River
chose to keep the principal function within the parent organization but to break it out
from the agent into a companion organizational element. Within the agent
organizations at each location, there was no common organizational structure.
Although reasonably common in mission and physical characteristics,
organizationally, each set in place an internal organization that was unique to its
location. Even though each was performing a similar function, i.e., information
technology, and each developed a Most Efficient Organization, each organization
was unique.
Patterns in Variables
Where the principal agent model has been implemented, the variables should
be approached consistently across all sites. Each site would be expected to have a
high degree of asset specificity. The principal should strive to align the goals of the
agent with their own. The principal tries to reduce instances of moral hazard and
information asymmetry. The goal of implementing the principal agent model is to
minimize costs to the overall organization. Consequently, the principal must
understand the transaction costs and attempt to manage them. Similarly, the
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
principal would be expected to put in place incentives to Improve performance and
encourage the agent to exceed minimal performance levels. The use of coercives
should be similar and tied to the incentives. Table 6 compares the variables at each
location to the theoretical principal agent mode!.
Table 6
Pattern Matching of Variables
Theoretical Sock Island te d Hlver Anniston Letterkenny
Asset
Specificity
1 1
High High High High High
Goal
Alignment M ot Always Not Always Always Always Not Always 1
Moral
Hazard Expected Some None None Expected 1
Information
Asymmetry Issue Some No Issue No Issue Unknown
Transaction
Costs Known Unknown Unknown Unknown Unknown
Incentives &
Coercives Present Coercives None None Coercives
The model predicts a very high level of asset specificity, which was found at
all sites. A-76 is a contracting process requiring a detailed PWS and contractual
requirements. Because of this, each site spent up to several years developing In
detail what was required of the agent organization. The level of specificity was
extremely high at all sites, although identifying how specific outputs were to be
accomplished was limited. The role of the principal was to oversee the
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
accomplishment of the outputs, not to dictate the method of performance to the
agent.
Goal alignment is how well the goals of the principal and agent agree.
Principal agent theory would envision some incongruence in the goals of the two
organizations. The categorization for this variable- was either “always” having the
goals aligned or “not always” having the goals in alignment. Rock Island lias
experienced some lack of goal alignment in the past, and this was a concern of
Letterkenny. Both viewed the contractor goats as potentially being in some level of
conflict with the goals of the principal. Red River and Anniston both viewed the
agent organizations as having a close alignment with the principal’s goals. The
retention of the in-house work force that had been socialized over time and had the
same norms as the larger organization overcame possible goal conflict. In the case of
Anniston, the total integration of both the principal and the agent organizations
insured a commonality of goals and a focus on a common good. The issue of goals
not being aligned between the principal and the agent arises when one can gain at the
expense of the other. At Red River and Anniston, any gain by either the principal or
the agent was viewed as mutually beneficial to the larger organization. At
Letterkenny and Rock Island, the potential existed for any gain by the agent being
offset by a greater burden placed on the principal.
Moral hazard and information asymmetry have a similar pattern to goal
alignment. The principal agent model would predict that there is the potential for
both to occur. The pattern found is consistent across both variables. Where Rock
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Island has experienced some levels of moral hazard and information asymmetry,
Letterkenny has concerns that this will be an issue in the future. In dealing with a
contractor as agent. Rock Island has found instances of both variables but has also
overcome these as a problem. Letterkenny is concerned that these could become
problem issues, if they go to a contractor agent. Red River and Anniston, on the
other hand, have no problems or concerns with moral hazard or with information
asymmetry. Their organizational approach of integrating the agent and the in-house
use of employees has overcome these potential issues.
The installation should identify and manage transaction costs to minimize the
total costs. The recurring competitions every five years and the detailed labor and
accounting systems already in place would indicate that this is a simple process.
However, not one of the study locations does a good job of capturing any portion of
the transaction costs. Across ail sites, limited or no detailed transaction cost data
were kept. The ex ante costs for consultants to develop the PWS was the only known
cost element at each site. Summation of labor costs for the ex ante, monitoring or ex
post costs was not available at any site. This was the only variable where all sites
responded the same but opposite to what principal agent theory would predict.
The use of incentives and coercives is to encourage the desired performance.
Not one of the sites had an incentive structure based on the agent’s performance or
plans to implement one. TWs was true regardless if the agent was a contractor firm
or in-house employees. Government employees did have incentive systems available
to them, but they did not tie back to the PWS outputs for the agent. The common
14?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
logic for a lack of incentives was that there was a contract in place, and the agent is
obligated to perform to that level Superior performance is accepted but not
encouraged through incentives.
Coercives of withholding of payment for failure to adequately perform exist
when the agent is a contractor firm. These coercives were neither specific nor set in
place prior to performance failure but were set in motion after some level of failure
occurred. Principal agent would have the coercives be established and known by
both parties prior to a performance problem occurring.
At Red River and Anniston, there basically is no set of coercives for the agent
organization. Withholding payment is not an option. There is the ultimate penalty of
moving the entire function to contractor performance, if the organization does not
perform, but this has never occurred. There axe individual employee performance
ratings that can reflect on poor performance, but there are no coercives at the
organizational level. Red River and Anniston have retained the bureaucratic
approach of encouraging good performance and correcting poor performance.
Explanation-Building
Explanation-building follows the pattern matching. Explanation-building
bridges from the theoretical proposition and looks to explain what has been found in
data-gathering and why this has occurred. If the research data does not match the
theoretical expectations, it is the explanation-building that aids in understanding the
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
variance and external factors that may be influencing the situation. The explanation-
building is where these differences become understood.
Organizational Patterns
The variation in organi.zati.onal patterns and the willingness to embrace the
principal agent model is directly related to whether the agent organization is a
contractor. Rock Island has long experience with the principal agent model, since
moving functions to contractor performance in the early 1980s. That set in place a
contractor organization working side by side with government employees but always
a separate work group. Their reporting chain was clearly distinct and there was
always a subordinate overseer relationship between the contractor agent and the
government principal. Rock Island started out with a clear segregation between the
principal and the agent organizations that carried forward. With the succeeding
rounds of A-76, the organization was comfortable with the model and continued
adopting the principal agent organizations. The agent structure was set in place as a
stand-alone organization with either government or contractor employees.
Letterkenny follows the Rock Island pattern, establishing its principal and
agent organizations as unique, stand-alone organizations. Letterkenny is preparing
for a contractor agent organization and is building a principal organization to serve in
the oversight mode. Simultaneously, Letterkenny is building an in-house MEO
proposal that is the counterpart to a contractor agent organization. The preparation
149
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for the principal role overseeing the contractor is driving the Letterkenny
organizational structure.
Both Red River and Anniston developed a more integrated organization
structure. Both know there will be no contractor and that the existing government
employees will serve as the agent. It will be no less than five years until either
installation will address A-76 in these functions again. Two trends appear at work at
these two sites that influence the organizational structures. First, they had
successfully operated under an integrated approach for many years. The principal
agent model is a result of the A-76 contracting process and not an organizational
structure willingly entered into. The comfort level was very high with the existing
structure; there were clear lines of authority and an integrated operating environment.
By retaining the traditional structure, the senior management retained direct lines of
control throughout the organization. Secondly, the role of the principal becomes
somewhat unnecessary under the traditional structure. By retaining the traditional
structure, the oversight by the principal is very limited and, in some cases, non
existent. It appears that organizations cling to the more traditional integrated
organizational structure unless driven to the principal agent model. Having a
contractor as the agent forces the organization to a principal agent approach. Even
though the A-76 process pushes the organization towards a principal agent model
with a defined set of PWS outputs, quality assurance monitoring and external
oversight, the organizations prefer to hold to the traditional integrated approach.
150
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Under the principal agent model, there are two basic organizational options
for the principal. The first is a centralized principal—the organizational alignment in
the first generation at Rock Island, and proposed at Letterkenny. The advantages are
that it centralizes contract management, quality assurance evaluation and technical
direction into a single organization. The primary driver for this approach is the size
and complexity of the functions the agent performs. Both Rock Island and
Letterkenny started with large, complex, multi-function performance work
statements, and the role of the principal was best fulfilled by a single organization.
The second organizational option for the principal is a decentralized
approach. Rock island evolved into this over time as it created several single
function contractor agents. Red River and Anniston both planned a decentralized
approach for their principal organization. The initial plan for both installations was a
decentralized principal, and both had single function A-76 studies. However, after
the in-house win, Anniston centralized their principal. The choice of using a
centralized principal organization is based on whether the agent performs single or
multi functions. When the agent performs a single function, the principal is usually
in a decentralized organization structure. When the agent performs multi functions,
the principal is usually in a centralized organization structure (Table 7).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 7
Explanation-Building of Organizational Structures
Contractor Agent Government Agen
Accepts Principal Agent Rock Island
Letterkenny
--------------------------------
Rock Island
Letterkenny
Resists Principal Agent Red River
Anniston 1
Centralized Principal Letterkenny
1
Letterkenny
Anniston
Decentralized Principal Rock Island Rock Island
Red River |
Patterns in Variables
Asset specificity is high in ail locations. This is attributable to the contractual
A-76 process that required extensive work to develop the PWS. This high level of
asset specificity means that the outputs of the agent organizations are highly
quantifiable and measurable. This pattern was consistent at all sites regardless as to
whether the in-house government or contractor organization was the agent.
The three variables of goal alignment, moral hazard and information
asymmetry have a common pattern. Where the agent is an in-house government
organization at Red River and Anniston, none of these three variables is considered a
problem. One key was the retention of employees who have worked in the
organization for many years and have developed a high level of trust. There is no
gain to be made by the agent staff, if they engaged any of these three variables. This
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
integration of employees into the norms and goals of the organization builds a trust
level, which, under a principal agent model, counters the negative attributes of these
variables. The lack of competing interests made the Idea of goal alignment and profit
maximization irrelevant at these installations. Under A-76 and the principal agent
model, the governmental organization cannot undertake some of the classical
negative actions, such as budget maximization or developing new mission functions.
All of this is controlled under the PWS; the principal agent model limits any efforts
to undertake these actions.
At Red River and Anniston, there is the potential for individuals to have
conflicting personal goals and to shirk. Collectively, as an organization, the agent
would not have the opportunity to engage in such activities. Information asymmetry
could occur, but again the respondents felt this was highly unlikely. The flow of
information both vertically and horizontally and the presence of principal individuals
Involved in the technical aspects of the function mitigates the Information asymmetry
risks.
Rock Island and Letterkenny took different views of the goal alignment,
moral hazard and information asymmetry variables. Rock Island is dealing with
contractor agents, and Letterkenny is preparing for that end state. When the agent is
a contractor, the negative aspects of the principal agent variables can become an
issue. Rock Island saw some instances of each and had to Increase monitoring and
oversight efforts to overcome these problems. The contractor agent exists as a profit-
making entity, very different from the reason government employees are organized to
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
perform a set of tasks. The contractor typically has a goal to maximize profits, which
may create goal conflicts with the principal. An inherent lack of trust built into the
structure of two distinct organizations with overseer and subordinate roles creates
both the potential and the concern over these issues.
In the beginning of the principal agent relationship, the potential for conflict
seems greater, as In the Rock Island experience. At this point, the contractor agent
has just eliminated employment for some number of government employees.
Although usually offered jobs with the contractor, they typically prefer to stay with
government employment and look for other government positions. The new
employees of the contractor agent have no working relationship and no trust built up
with the principal. Not having performed the jobs in the past, there is a level of
Inexperience and learning curve to overcome. Usually, there is a higher start-up cost
for the agent and pressure to keep costs down. These can lead to goal alignment and
shirking Issues. The principal begins with limited faith in the ability of the agent to
perform. The agent is at the disadvantage, new to the work and usually having
proposed doing the work for 20% to 30% less than it had been done before. This all
leads toward a perceived or real problem with goal alignment and moral hazard.
These trust issues and start-up problems occur less, when the agent is the government
organization.
The Rock Island experience showed that the concern for moral hazards
declined over time. The animosity of displacing co-workers and the trauma of the A-
76 process declines. The contractor agent desires to do a good job to continue
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
performance of the contract. A working relationship and a trust level builds up
between the agent and the principal. The monitoring efforts identified areas of
concern, and these were resolved. If problems are too great to overcome, the contract
is canceled, competed again, and the process begins anew. However, at Rock Island,
the issues were overcome and a working relationship established. As opposed to a
Wgh level of trust within the family, as found with Red River and Anniston, the Rock
Island principal agent relationship is more like one with your favorite car mechanic;
you trust Mm but are never entirely sure that he has not billed just a little too much or
has not made a repair that was not entirely needed.
When the agent is a contractor, the issue of information asymmetry is as great
or greater than the moral hazard and goal alignment issues. The principal has
potentially reduced the oversight staff to where technical knowledge begins to
deteriorate. The principal is no longer folly aware of all that is happening witMn the
function. The information technology area is an example, where knowledge of the
broader industry trends is needed, and yet the diminished principal staff my not be
able to stay abreast. At the more basic day-to-day operations, the principal staff may
not be aware of agent performance in geograpMcaliy separate areas and may not have
the ability to validate quality of performance or work counts. This is why there are
potential information asymmetry problems in the agent’s execution of the work.
Similarly, the information asymmetry issue becomes a problem at the end of
the five-year execution period, when the contract needs to be competed again. No
longer does the principal have a foil staff of individuals knowledgeable of the details
155
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of performance. As long as the performance requirements do not change, this may
require only some adjustments to the PWS. However, if there are major changes in
performance requirements, such as in the information technology field, this may
require a major effort. The agent now has the knowledge base, and the principal
must develop the output requirements without ever having performed the function.
As time progresses, and ten or fifteen years have passed, there may be no one
working for the principal that has ever performed the work. The information
asymmetry problem may increase over time. If the function is relatively simple with
limited work counts, such as a mailroom operation, this problem may never be great.
However, if the function is large and complex, such as information technology, this
can become a major problem for the principal to manage the agent.
A consistent pattern across all four sites also appears for dealing with
transaction costs and incentives and coercives. None of the sites is tracking the
transaction costs or working to minimize the total costs of implementing A-76 and
the principal agent model The lack of good cost data, the lack of effort to track
ongoing costs and the lack of a focus to reduce future costs were apparent. An
explanation is that the installation organizations have not accepted this as a
permanent change to operations, and this type of competition will continue
indefinitely. Although aware of the political pressures to continue competitions,
none of the interviewees seemed aware of larger trends in government and were
concerned more with the immediate issue of their particular competition.
156
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Management looked at each component of the transaction costs separately and did
not know or attempt to minimize the total.
Similarly, none of the four sites attempted to design an incentive or coercive
package to influence the performance of the agent. This was a concept that seemed
foreign to many of the interviewees. There was consensus that coercives may be
necessary, but a structure was not in place. Incentives to encourage higher
performance levels were ignored. When the agent was a contractor, the approach
was that acceptable levels of performance were established; anything over that was
unnecessary and not worth additional funding. When the agent was the in-house
government organization, the approach was the same.
In summary, the variables can be broken into two subgroups. Table 8 shows
how these, relate. The first group consists of asset specificity, transaction costs and
incentives and coercives. These variables were approached similarly, whether the
agent is a contractor or in-house government employees. All three are directly tied to
the contracting process. The asset specificity is part of the PWS; the incentives and
coercives are part of the solicitation; and the transaction costs are the result of the
contracting process. Going forward into the A-76 process, the installations did not
know who would emerge as the agent, and each was treated equally, if the variable
was addressed at all.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 8
Explanation-Building of Patterns in Variables
Contractor Agent Government Agent
Asset Specificity High High
Transaction Costs Untracked Untracked
1 Incentives & Coercives None None
Goal Alignment Issue Non-issue i
Moral Hazard Issue Non-issue
Information Asymmetry Issue Non-issue
The second subgroup consists of goal alignment, moral hazard and
information asymmetry. The prime factor that explains the approach to managing
these variables is whether the agent is a contractor or in-house government
organization. When the agent is a contractor, the principals recognize and attempt to
manage these variables. If the agent is a contractor, there is an immediate we/they
situation that divides the organization. When the agent is an in-house government
organization, the approach is almost family-like, assuming common goals and high
levels of trust and cooperation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER VHI
FINDINGS AN© CONCLUSIONS
When a subject Is highly controversial. . . one cannot hope to
tell the truth. One can only show how one came to hold
whatever opinion one does hold. One can only give one’s
audience the chance of drawing their own conclusions as they
observe the limitations, the prejudices, the idiosyncrasies of
the speaker. (Virginia Woolf, n.d.)
The research presented the literature on competition as a mechanism to make
government operations more efficient. The purpose was to validate whether
implementation follows what theory would predict. Principal agent theory underlies
the organizational structures that come out of these competitions. Previous studies
have shown that competition generally reduces costs but did not look at the
implementation of the principal agent structure. Four research sites were chosen to
study implementation of the principal agent model following the A-76 competition
process. The study of four sites and common functions in different stages of
implementation provided the opportunity to understand how theory is implemented
in operational settings.
Findings
Chapter 4 stated the specific research question in three parts. Each part
addressed a different aspect of principal agent implementation and is addressed
159
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
separately in the following subsections. The principal agent theory and the overall
Army guidance for the A-76 program would indicate that each location should
respond affirmatively to each question. The management of the A-76 competition
process at the installation, level and the lack of specific direction allowed the
installations to create their own unique implementation of the principal agent model.
Does the Principal Agent Relationship Apply to Organizations
where Government has Implemented Contractual Arrangements
for Commercial Activities?
The research looked at two aspects OF this question—the organizational
structure and the variables that are common to a principal agent structure. Theory
predicts a clear separation of principal and agent organizations and a clear line of
responsibility for oversight by the principal. This should occur regardless of whether
the agent organization is an in-house government organization or a private-sector
firm under contract.
Implementation was found to be dependent on whether the agent is an in-
house government organization or a private-sector firm. When the agent is a private-
sector firm, the principal agent model is applied in its classical approach. There is a
defined principal organization that implements its oversight and quality assurance
roles. However, when the agent is an in-house government organization,
implementation strays from the principal agent and reaches back to the traditional
bureaucratic structure. The organizations seem less willing to give control to a
principal group and have a strong desire to integrate the agent into the larger parent
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
organizations. This desire to integrate blurs the lines of responsibility but also
eliminates certain areas of contention. Implementation should not be dependent
upon which organization is the agent and contradicts the principal agent theory that
underlies the A-76 process.
When looking at the principal agent model variables, there is a mixture of
results. Three of the variables were consistent at all sites. Asset specificity is high, at
all locations because of the great amount of work placed on developing the
performance work statement. This is folly consistent with principal agent theory. At
ail sites, there was very little or no effort devoted to instituting incentives and
coercives or to manage transaction costs. TWs is not consistent with principal agent
theory. The installations seemed unable or unwilling to develop an incentive system
that could be applied equally to both the in-house government organization and the
private-sector competitors. The limitations of the civil service system hinders the
development of coercives that can be applied to the in-house organization. Because
of these difficulties, all sites deviated from the theory that incentives and coercives
should be in place for the principal agent model. Secondly, none of the locations
accounted for or worked to reduce the transaction costs.
The remaining three variables of goal alignment moral hazard and
information asymmetry are tied to the type of organization that is the agent. At Red
River and Anniston, where the in-house organization was the agent, there was no
recognition of these variables or recognition that they could be a potential problem.
At Rock Island, where the agent was a private-sector firm, and at Letterkenny, where
161
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the agent was not yet selected, the principals recognized these variables as possible
areas of conflict and had a plan to address them.
In summary', the answer to the first question— “ Does the principal agent
relationship apply to organizations where government has implemented contractual
arrangements fo r commercial activities? ”—is dependent upon what organization
serves as the agent. When the agent is a private-sector firm, the principal agent
model applies. Organizationally, and from the manner in which the principals
address the variables, compliance Is what theory would predict. The exception Is a
lack of an Incentive structure and managing transaction costs. Conversely, when the
agent organization is in-house government employees, there is a limited effort to
implement the principal agent structure organizationally. The installations create
hybrid organizations that contain aspects of the traditional bureaucratic structures.
Except for asset specificity, when the agent is an in-house government organization,
the variables associated with the principal agent theory are not acknowledged.
Do the Participants Understand the Inherent Forces that Detract
from an Optimal Efficient Organization?
This question explores the participants’ depth of knowledge and
understanding of the principal agent model variables. Asset specificity is the single
variable clearly understood at all sites. The contractual nature of the A-76 process
and the high potential of the agent being a private-sector firm drove each installation
to detail every output and the associated quality performance levels. Whether the
162
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
agent was an in-house government organization or a private-sector firm had no
relationship to the asset specificity.
None of the sites expressed an understanding or desire to develop a system of
incentives and coercives other than withholding payment for performance. In
chapter 2, the importance of incentives to folly implement the principal agent model
(Eisenhardt, 1985) was discussed. Without the incentive and coercive structure, the
agent will strive to meet the minimal acceptable level. Data from the sites revealed
that there was no incentive or coercive system in place but, more significantly, that
none of the installations had given much thought to developing one. There was little
understanding of how a system of incentives could be structured to encourage higher
levels of performance. This indicates a lack of understanding of the incentive
system’s value and a lack of understanding how incentives and coercives fit into the
principal agent model.
Similarly, none of the sites was attempting to manage transaction costs.
There was no history of the ex ante costs, monitoring costs or ex post costs. The
sites did not recognize the competition process as an ongoing effort that will extend
over time nor the need to manage these costs. The goal is to minimize the total
transaction costs; but, without knowing what these costs are, and without planning
for future efforts, this variable is not even acknowledged.
For the three variables of goal alignment moral hazard and information
asymmetry, the knowledge level of the principal organization was related to the agent
organization. All three of these variables are addressed if the agent is a contractor
163
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and basically ignored if the agent is an in-house government organization. When the
agent was a private-sector firm, the principals acknowledged the existence of the
variables and their potential impact on the principal agent relationship. The
principals developed quality assurance mechanisms and work count reporting
systems to mitigate the possible impact of the variables. In contrast, the two sites
that had in-house organizations as agents did virtually nothing to recognize these
variables. The idea that the agent might have conflicting goals or information
asymmetry was countered by the structure that integrated the agent into the larger
organization. The result of the organizational integration may be the mitigation of
the potential conflict between the principal and agent.
In summary, the principals had little recognition of these variables. They
expressed little understanding of their impact on their organizations or how to
manage the variables to improve performance. The active management of the
variables occurred only when the agent was not an integral part of the larger
organization.
Can Common Trends of Coping with these Forces Be Found
that Show the Principals are Managing the Organization for
Optimum Efficiency?
The question presumes there is an approach to the principal agent model that
is better than others and would be found at multiple sites. The approach should be in
place regardless of the agent. The only commonality across all sites was the high
164
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
degree of asset specificity. This is a contractual requirement, but also provides the
very basic building block of the principal agent relationship.
There were no other common trends identified. When the agent was a
private-sector firm, the principal generally behaved as theory predicts. There was
little understanding of principal agent theory in its larger context, but the principal
organizations generally understood, the concepts and variables. Where government
entities are experienced in using contractors, it was evident they understood the
principal agent relationship. When the agent was an in-house government
organization, there was substantial deviation from the principal agent model. They
clung to the traditional organizational structures and did not significantly attempt to
implement a principal agent approach. This lack of drive to implement principal
agent could have two explanations. The first is management’s desire to hold to the
traditional bureaucratic structure, because they are more familiar and comfortable
with this mode of operation. Management was never provided training on the
advantages of principal agent and the need to change organizational models.
Management prefers an integrated organization that offers more flexibility and does
not separate employees into two camps of those under and those not under the agent
approach. The second reason is that there are certain inefficiencies in the principal
agent approach that management may be trying to avoid. The reporting
requirements, the need for a quality assurance function, and the monitoring costs are
all by-products of the principal agent model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In summary, there are no common trends identified across the sites that might
demonstrate that the principals are managing the principal agent approach for
optimum efficiency. Organizationally, there Is no commonality or in how the
variables are managed. There is little demonstration that those in the principal
organizations understand the theory behind the organizational approach or have
knowledge of the potential pitfalls. Figure 20 summarizes what was found for the
specific research question. It was expected that a “YES” would have been applicable
to every block in the matrix, if the sites had understood and applied the principal
agent approach. Instead, only when the agent is a private-sector firm is the principal
agent model put into place.
Does i»ri»c.% S ;al agent,
reiatio'nsitip ■ apply"?
YES' NO
X > o p:r.m c% >ats YES T S T O
»n«fer3fc an « t forces'?
Art 'ttiere'.coiwmozi
isro ■ * s r o
tx-esK ts?
Figure 20. Conclusion
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Conclusions
The findings of the study answered the research question. The following
sections discuss how these findings can apply to other situations, locations and the
limits of the study.
Applicability to Theory and Implications to Broader Areas
The research examined implementation of the principal agent mode! under
the A-76 program. As a result of the research, the researcher concludes that
organizations tend to avoid implementation where they have an alternative.
Organizations do not folly implement and do not seem to folly comprehend the
complexity of operating under this organizational model. The use of the contract as
the unit of analysis for the in-house organization is resisted. Where the organization
did implement principal agent, it quickly learned what was required to manage under
this approach but did not embrace every aspect, especially the concept of incentives
and coercives.
When dealing with an in-house organization as agent, the variables dealing
with goal alignment, moral hazard and information asymmetry lose much of their
significance. The principal organizations strive to integrate the agent into the
organization, and the lack of competing goals encourages the in-house agent to align
with the principal. The theory may need to be expanded to recognize subsets—one
dealing with when the agent is external to the organization, such as a contractor, and
a second when the agent is a subordinate organization within the larger parent
167
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
organization. In this case, the unit of analysis remains the contract, but the oversight
and monitoring functions are altered. The overall approach of developing specific
performance outputs and then managing these is the heart of the principal agent
model, but deviations seem inevitable when the agent is an internal part of the
organization. In this case, the model, may need to focus away from the variables of
goal alignment and information asymmetry and focus almost solely on the need for a
performance monitoring system.
The broader implications relate to the general movement of government to a
principal agent and performance-based organization. The theory is sound, but the
implementation has tremendous resistance at the lowest organization levels. Those
who design programs such as A-76 are not charged with implementation. Those who
must implement have not been convinced that this is a better way to organize and
operate. This resistance places delays and problems in the process, and the
organizations reach back to retain their traditional hierarchical bureaucratic
structures. Ultimately, the resistance that occurs at the organizational level may alter
the larger trend to movement towards the Hollow State. The movement to the
Hollow State is driven from the top, not from the bottom. It is often a political and
economic decision, not one willingly pursued by the organization. Time will tell if
the principal agent model is widely adopted in government organizations or if it is a
passing fad of using competition to determine who should provide government
services.
168
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Applicability to Other sites
The results of this study seem applicable to other military installations. When
expanded to other sites, the findings point to larger issues. Across all sites, there is
limited understanding of the principal agent model, the variables and how to manage
under this organizational structure. This highlights the need for additional training
for installation personnel implementing this type of organization. The lack of
understanding on how to structure a system of incentives and coercives inhibits
effective management of the agent’s performance. Additionally, the lack of any real
knowledge of the transaction costs throughout the life of the A-76 process is
indicative that the organizations are implementing the process without managing for
maximum efficiency. Understanding the principal agent model variables is required
for the model to work effectively.
During the data collection process, it was apparent that there was no learning
from the experiences of one site to another. It would be appropriate for a “lessons
learned” session of the individual sites to assist other installations in their
implementation. This is not occurring. There is very little training offered in the
process beyond the development of the PWS and the contracting process. How to
organize and operate under principal agent is left to the devices of the individual
installations. This lack of common learning is reflective of a lack of understanding
that the process is neither simple nor unique to any one location.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Limitations of the Study—Limitations of the Data
Chapter 5 identified the threats to the study validity, and these are
summarized in Table 9. The primary threats are the effects of history,, selection and
maturation. There was very limited mathematical analysis, and the use of the design
for description approach made the threats of selection and history the most
significant limitations.
Table 9
Threats to Study Validity
External
Validity
Threats
Potential
Threats
Selection
History
Insignificant
Threats
Setting
Unique Features
Testing
Reactive Effects
Potential History
Threats Maturation
Internal Selection
Validity
Threats
Statistical Regression Testing
Insignificant Instrumentation
Threats Design Contamination
Experimental Mortality
External Threats to Validity
The external threats to validity relate to applying the study results to other
locations. External validity was increased through the use of multiple sites; a larger
sample size increases the probability that the results will be common. Chapter 5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
identified why the threats of unique program features, effects of setting, effects of
testing and reactive effects are considered insignificant. Unique program features did
not apply, because the A-76 program is neither unique nor limited. The A-76
program is being used at hundreds ofDOD locations; and, though each location is
unique, the approach is common.
Using multiple geographic locations and common functions minimized the
effects of the setting. The information technology and public works functions are
common in all military installations. There is a significant private-sector industry
that makes these functions readily contractible to contractor performance. Using
multiple locations and functions that are neither unique to the military nor specific
locations mitigates the effects of the setting.
The other two threats to external validity deemed insignificant are the effects
of testing and reactive effects. The first occurs when a pretest has influenced the
outcome. There is no pretest in this study. The use of descriptive methodologies
eliminates this possible impact on validity. The reactive effect refers to the artificial
environment created when testing occurs or when the subjects know they are being
studied. In the study design, there was no testing. The descriptive approach gathered
data on the historical and current processes and organizational design. The
participants were not reacting to the study; they were reporting the situation as it
existed.
The two threats initially identified as having a potential impact on the
external validity were the effects of selection and the effects of history. The effect of
171
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
selection impacts the study, if the selected sites are unrepresentative. The four
selected sites are Army industrial sites and form a subset within the larger population
of military.7 installations. The use of common functions of information technology
and public works operations found at every military installation was designed to
offset the effects of selection.
The threat of the effect of history7 relates to looking at a particular time period
that may not be representative across time. To offset this threat, the study design
used multiple locations across a time continuum of implementing A-76 competition
results. In particular, the Rock Island experience across two decades and different
series of A-76 competitions helped mitigate the effects of history. Examining the
organizational approaches across time and not restricting the study to the current
process provides a better understanding of how history has impacted the process.
Internal Threats to Validity
Internal threats to validity concern the data reliability to support the study
conclusions. There are eight different threats to internal validity of which history,
maturation and selection are a concern. The factors that are not a concern are
statistical regression, testing, instrumentation, design contamination and
experimental mortality. These threats focus on the measurement instrument and
analysis. The study design for description and the data-collection approach limited
the numeric data. There was no testing or statistical regression. None of the sites
was eliminated, and the data-collection instrument did not change, so the threats of
172
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
experimental mortality and instrumentation did not exist. Design contamination
occurs when the study participants know they are under study and act differently than
they would normally act. The use of historical information in the study approach
eliminated any potential design contamination.'
Three potential threats to internal validity include history, maturation and
selection. The threat of history occurs, when events other than those identified could
affect the variables. Maturation effects are the natural changes that may take place
over time. Selection effects relate to the way subjects were chosen, which could
affect the way they respond. All three of these could have some level of influence on
the study. The use of multiple locations and the extensive literature review of macro
level trends minimized the effects of history. The use of interview's and open-ended
questions to explain why events occurred provided the opportunity to identify other
forces that were not considered. The effects of selection are similar in both internal
and external validity. Recognition of the potential threat to internal validity
influenced site selection to specifically minimize any impacts.
The threat of maturation may influence the study results. The evidence of
Rock Island Arsenal and its evolving implementation of the principal agent model
indicate there is no static approach. The time period for implementing the principal
agent approach is in decades. The findings at Anniston show that the evolution to
principal agent may also be limited to the functions contracted to private-sector
firms. The single study location at Rock Island, where there appears to be an
evolution in implementing principal agent, is insufficient to draw any conclusions as
173
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to whether this is an effect of maturation. This change could be a trend to be found
at all sites over time, but there is insufficient data to demonstrate this as a larger
trend.
Other Areas for Future Research
Two areas are proposed for future research: (1) why do organizations have
such hesitancy to adopt the principal agent model and (2) what types of training
would be useful for organizations preparing to go under the principal agent model?
The larger organizational shift to a principal agent model and the need for more
efficient structures is happening throughout government. The A-76 competition
process was designed to put into place a principal agent approach, regardless of the
producer. It is evident there is resistance to adopting the principal agent approach, if
there is an alternative. From the literature review and the analysis, it is apparent
there has been no research into why this resistance is occurring.
A second proposed area for future research follows the first. The question
addresses what types of training would be useful for organizations preparing to go
under the principal agent model. It appears that part of the resistance to moving to a
principal agent model is the lack of knowledge of the underlying organizational
theory. With the shift to principal agent implementation across government
organizations, those being asked to manage under this model should be trained in the
benefits and potential problems with this approach. This additional research would
focus on what training should be provided and to whom so that the organizations can
174
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
understand the principal agent paradigm, how to manage the organizations and the
associated variables.
1 7 5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
SELECTED BIBLIOGRAPHY
Arrow, Kenneth J. (1985). The economies of agency. In John W. Pratt & Richard I.
Zeckhauser, Principals and agents: The structure o f business (pp. 37-55).
Boston, MA: Harvard Business School Press.
Bozeman, Barry (1988, March-April). Exploring the limits of public and private
sectors: Sector boundaries as Maginot Line. Public Administration Review,
48 (2), 672-673.
Brocherding, Thomas (1977). Budgets and bureaucrats: The sources o f government
growth, Durham, NC: Duke University Press.
Butler, S. (1985). Privatizing federal spending: A strategy to eliminate the budget
deficit. New York: Universe Books.
Garrick, Paul M. (1988). New evidence on government efficiency. Journal o f Policy
Analysis and Management, 7(3), 518-528.
Center for Naval Analyses (Alan I. Marcus) (1993). Analysis o f the Navy’ s
commercial activities program. Alexandria, VA: Center for Naval Analyses.
________ (Carla E. Tighe, Samuel D. Kleinman, lames M. Jondrow, & R. Derek
Turnkey) (1996a, October). Outsourcing and competition: Lessons learned
from DOD commercial activities program. Alexandria, YA: Center for
Naval Analyses.
_ _ _ _ _ (R. Derek Turnkey, Robert P. Trost, & Christopher M. Snyder) (1996b,
December). Analysis o f DOD's commercial activities program. Alexandria,
YA: Center for Naval Analyses.
________ (Carla E. Tighe, James M. Jondrow, John D. Keenan, Carl S. Moore,
LCdr. Christopher M. Reeger, Matthew T. Robinson, & R. Derek Turnkey)
(1997, January). Case studies in DOD outsourcing. Alexandria, VA: Center
for Naval Analyses.
176
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Commission on the Roles and Missions of the Armed Services (1995). Directions
fo r defense. Washington, DC: Government Printing Office.
Congressional Budget Office (1995, July). Public and private roles in maintaining
military equipment at the depot level Washington, DC: Government Printing-
Office.
Converse, Jean M., & Presser, Stanley (1986). Survey questions: Handcrafting the
standardized questionnaire. Newbury Park, CA: SAGE Publications, Inc.
Donnelly, John (1995, November 6). Senior panel weighs privatization possibilities.
Defense Week, 9.
Eisenhardt, Kathleen M. (1985, February). Control: Organizational and economic
approaches. Management Science, 31 (2), 134-149.
________ (1989, January). An assessment and review. Academy o f Management
Review, 14 (1), 57-74.
Erlich, Jeff (1996, January 28). Task force pushes for lean DOD. Defense News, 29.
Ferris, James, & Graddy, Elizabeth (1986, July-August). Contracting out: For what?
With whom? Public Administration Review. 46(4), 332-344.
Fink, Arlene, & Kosecoff, Jacqueline (1985). How to conduct surveys. Newbury
Park, CA: SAGE Publications, Inc.
Friel, B. (2001). DOD puts 500,000 jobs on outsourcing list. Government Executive.
Gable, Mark & Moyer, Bruce (1996). Defense depots: To privatize or not to
privatize. Federal Managers Quarterly (2), 10-13.
General Accounting Office (1986, December 31). Potential savings from private
sector cost comparisons. Washington, DC: Government Printing Office.
___ (1988a, October 31). No savings from conti'acting fo r support services at
Fort Eustis, Virginia. Washington, DC: Government Printing Office.
_ _ _ _ _ _ (1988b, November 28). DOD’ s experience in contracting out
commercially available activities. Washington, DC: Government Printing
Office.
177
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(1989, February 10). Contracting out public works functions at
Jacksonville Naval Air Station. Washington, DC: Government Printing
Office.
(1990a, March 15). DOD's reported savings figures are incomplete and
inaccurate. Washington, DC: Government Printing Office.
(1990b, September 26). Monopoly-like status contributes to inefficiency
and ineffectiveness. Washington, DC: Government Printing Office.
(1991a, February 11). Expected savings are not being realized in Ft. Sill's
logistics contract. Washington, DC: Government Printing Office.
(1991b, November). Government contractors: Are service contractors
performing inherently governmental functions? Washington, DC:
Government Printing Office.
(1992, December 3). Cost-effective contract management requires
sustained commitment. Washington, DC: Government Printing Office.
_ _ (1994, December 9). Aerospace guidance and metrology center: Cost
growth and other factors effect closure and privatization. Washington, DC:
Government Printing Office.
__ (1995, March 29). An overview o f the federal contracting-out program.
Washington, DC: Government Printing Office.
_ (1996a, April 17). Privatization and the debate over the public-private
mix. Washington, DC: Government Printing Office.
_ (1996b, July 15). Commission on roles and mission’ sprivatization
assumptions are questionable. Washington, DC: Government Printing
Office.
(1997a, March 11). Challenges confronting DOD as it renews emphasis
on outsourcing. Washington, DC: Government Printing Office.
(1997b, March 12). Challenges facing DOD as it attempts to save billions
in infrastructure costs. Washington, DC: Government Printing Office.
(1999, February). Results o f recent competitions. Washington, DC:
Government Printing Office.
178
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
__________ (2000, December). Results o f A-76 studies over the past 5 years.
Washington, DC: Government Printing Office.
Globerman, Steven, & Vinlng, Aidan (1996, November-December). A framework
for evaluating the government contracting-out decision with an application to
information technology. Public Administration Review, 56(6), 577-586.
Good, David H. (1992, April). Productive efficiency and contract management:
Some evidence from public transit agencies. Public Finance Quarterly, 20
(2), 195-216.
Goodsell, Charles T. (1982). The case fo r bureaucracy. Chatham, NJ: Chatham
House.
Girttman, Daniel, & Wilkter, Barry (1976). The shadow government. Hew York:
Pantheon Books.
Hanke, Steven (1985). Privatization: Theory, evidence and implementation.
Proceedings o f the Academy o f Political Science, 55(4), 101-113.
Reyman, Victor K. (1961, March-April). Government by contract: Boon or boner?
Public Administration Review, 21 (2), 59-64.
Isaac, S., & Michael, W. B. (1995). Handbook in Research and evaluation. San
Diego, CA: Educational and Industrial Testing Services.
Kalman, Steven (2001, July 3 0). Government Executive Magazine.
Kent, Calvin A. (1987). Entrepreneurship and the privatizing o f government New
York: Quorum Books.
KettI, Donald F. (1993). Sharing power, public governance and private markets.
Washington, DC: The Brookings Institution.
Kirkpatrick, Ian, & Lucio, Miguel Martinez (1996, Spring). The contract state and
the future of public management. Public Administration Review, 74, 1-8.
Lewin, K. (1935). A dynamic theory ofpersonality: Selected papers. New York:
McGraw-Hill Book Company.
Miles, Mathew B., & Huberman, A. Michael (1984). Qualitative data analysis: A
sourcebook o f new methods. Beverly Hills, CA: SAGE Publications, Inc.
179
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Miller, Gary (2000, April). Credible commitment and efficiency in the design of
public agencies. Journal o f Public Administration Research and Theory, 10
(2), 289-328.
Millward, H. Brinton (1996). Symposium on the Hollow State: Capacity, control,
and performance in interorgamzational settings. Journal o f Public
Administration Research and Theory, 6 (2), 193-195.
Mitnick, B. M. (1992). The theory of agency and organizational analysis. In T S f . E.
Bowie & R. E. Freeman, Ethics and agency theory (pp. 75-96). Hew York:
Oxford University Press.
Moe, Ronald C. (1987, November-December 31). The limits of privatization.
Public Administration Review, 47, 453-460.
Moore, Stephen, & Butler Stuart (1988). Privatization: A strategy fo r taming the
federal budget. Washington, DC: Heritage Foundation.
Nakasfaima, Ellen (2001, April 3). Pentagon hires out more than in. Washington
Post.
Niskanen, William (1971). Bureaucracy and representative government. Chicago:
Aldine-Atherton, Inc.
O’Sullivan, Elizabeth, & Rassel, Gary R. (1995). Research methods fo r public
administrators. White plains, NY: Longman Publishers.
Office of Management and Budget (1988, August). Enhancing governmental
productivity through competition: A new way o f doing business within the
government at least cost. Washington, DC: Government Printing Office.
Osbome, David, & Gabler, Ted (1992). Reinventing government: How the
entrepreneurial spirit is transforming the public sector, from schoolhouse to
statehouse, city hall to the Pentagon. Reading, MA: Addison-Wesley.
Pack, Janet Rothenberg (1987). Privatization of public-sector services in theory and
practice. Journal o f Policy Analysis and Management, 6 (4), 523-540.
Patton, Michael Quinn (1990). Qualitative evaluation and research methods.
Newbury Park, CA: SAGE Publications, Inc.
Prager, J. (1994), Contracting out government services: Lessons from the private
sector. Public Administration Review, 54(2), 176-184.
180
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Pratt, John W., & Zeckhauser, Richard J. (1985). Principals and agents: An
overview. In John W. Pratt & Richard J. Zeckhauser, Principals and agents:
The structure o f business (pp. 1-37). Boston, MA: Harvard Business School
Press.
Rehfiiss, John (1991, Summer). Tougher public management? Public agency
competition with private contractors. Public Administration Quarterly, 239-
252.
Rhea, Louis M., & Parker, Richard A. (1992). Designing and conducting survey
research. San Francisco, CA: Jossey-Bass Publishers.
Rosen, Harvey S. (1995). Public finance. Boston: Richard D. Irwin, Inc.
Ross, S. A. (1973, May). The economic theory of agency: The principal problem.
American Economic Review, 62(2), 134-139.
Savas, E. S. (1982). Privatizing the public sector: How to shrink government.
Chatham, NJ: Chatham House.
_ _____ (1987). Privatization: The key to better government Chatham, NJ:
Chatham House Publishers.
Schrier, Mary Ann (1994). Designing and using process evaluation. In Joseph 3.
Wholey, Harry P. Hatry, & Kathryn E. Newcomer, Handbook o f practical
program evaluation (pp. 40-68). San Francisco, CA: Jossey-Bass, Inc.
Sharkansky, Ira (1980, March-April). Policy making and service delivery on the
margins of government: The case of contractors. Public Administration
Review, 40 (2), 116-123.
StrettOfi, Hugh, & Orchard, Lionel (1994). Public goods, public enterprise, public
choice. New York: St. Martin’s Press.
White, Harrison C. (1985). Agency as control. In John W. Pratt & Richard J.
Zeckhauser, Principals and agents: The structure o f business (pp. 187-213).
Boston, MA: Harvard Business School Press.
Wholey, Joseph S., Hatry, Harry P., & Newcomer, Kathryn E. (1994). Handbook o f
practical program evaluation. San Francisco, CA: Jossey-Bass, Inc.
Yin, Robert K. (1994). Case study research. Thousand Oaks, CA: SAGE
Publications, Inc.
181
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDICES
182
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX 1
ACRONYMS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACRONYMS
CA—Commercial Activity
CAI—Commercial Activities Inventory
CAMIS—Commercial Activities Management Information System
CNA— -Center for Nava! Analysis
COM—Contracting Officer’s Representative
CPAS—Congressional Plan of Action Summary
DA—Department of the Army
DOD—Department of Defense
DQJM— -Directorate of Information Management
DPW—Directorate of Public Works
GAO—General Accounting Office
GIN—Governmental In Nature
MEO—Most Efficient Organization
NPR—National Performance Review
OMB—Office of Management and Budget
PWS—Performance Work Statement
QAE—Quality Assurance Evaluation
RIF—Reduction In Force
TDA—Table of Distribution and Allowances
184
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX 2
QUESTIONNAIRE
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
QUESTIONNAIRE
General Environment
This questionnaire is focused os establishing the way organizations at your
installation operate. There are two main organizations involved, those whose
performance is established under a performance work statement, referred to as the
Most Efficient Organization (MEO) and those who are responsible for overseeing the
performance of the MEO. The MEO could be either a contractor or an in-house
federal organization. I refer to those overseeing the MEO as the “principal.” The
questions are asked of your installation. If you feel there is more information you
can offer on any question, please do so. First, tell me a few things about yourself and
your installation.
Would you describe your position as senior management, middle management, or
staff?
Are you a member of the principal organization overseeing the MEO’s performance?
YorN
Are you a member of the MEO organization performing the work? Y or N
If you are neither part of the principal nor MEO organizations, please describe your
relationship to these organizations.
For your installation, is the MEO organization contracted to a private firm or is it part
of the federal civil service? Contracted or civil service If contracted,
what is the name of the firm?
Do you have multiple organizations performing as an MEO, i.e., different contractors
for different function? N or Y If Yes, please explain what firms perform what
functions.
What is the approximate staff size of the MEO organization?
50-100 100-150 150-200 200+
How long have you operated in this operational structure, under an MEO or
contracted?
0-2 years 2-5 years 5+ years
186
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Circle the functions performed by the MEO.
Facility and grounds maintenance
Facility engineering
Boiler plant operations
Environmental waste handling
Motor pool operations
Installation supply activities
Property accountability functions
Safety inspections
Public affairs
Audiovisual operations
Mailroom operations
Information technology functions
Other, please describe
Have you personally read the Performance Work Statement?
Y or N or only those sections that apply to my area
Have you personally read the proposal performance plan submitted by the MEO
organization? Y or N
For the most part, do the principal and the MEO organizations work side by side in
the same setting? Even though the blue-collar staff may work in many locations, do
the white-collar staffs work in the same area?
Have you personally ever received any classroom training on how to function when
organized with a principal directing work and an MEO organization performing the
work? If so, what training was received?
Does your organization have written procedures defining how the principal and MEO
organizations should interact and work together? Y or N
Please identify what these procedures are (written SOPs, MO As, etc.)
I will ask a series of questions about your satisfaction with the current structure.
Some will be comparing to a prior organization when federal employees working for
your installation Commander performed all of these functions. On a scale of 1-7
with 1 being very unsatisfactory to 7 being extremely satisfied, please answer the
following;
How satisfied are you with the quantity of MEO outputs?
How satisfied are you with the quality of MEO outputs?
187
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
H orn- satisfied are you with the responsiveness of the MEO?
How satisfied are you with the knowledge level of the MEO staff?
Thinking back to before the MEO structure was put in place are you, using the 7
point scale with a 7 being much more satisfied and 1 being very dissatisfied, how do
you compare the present performance of the MEO to the earlier operation where the
function was performed by in house staff? In some cases, these are the same
individuals, just fewer of them, in other cases tWs may be comparing to the current
contract provider.
What is your comparative satisfaction with the quantity7 of work outputs?
What is your comparative satisfaction with the quality of work outputs?
What Is your comparative satisfaction with the responsiveness of the MEO?
What is your comparative satisfaction with the MEO staff knowledge levels?
Variables
I am going to ask some questions of the operation at your installation. The first set
deals with “asset specificity.” TWs is the clarity with which the MEO’s work is
directed and the outputs clearly known. For example, a task might be “provide clean
restrooms.” Yet there may be large variations in how often these are cleaned and the
level of cleanliness. Again using the 7-point scale with 1 being very unsatisfactory or
poorly done to 7 being extremely satisfied or very well done, please answer the
following:
Thinking of the many outputs the MEO performs, do you feel these are adequately
described and required in either the PWS or day to day communication?
In your opinion, using the 1-7 scale, is the work count system adequate to capture the
outputs of the MEO? Can you describe the work count or work
reporting system?
On the 1-7 scale, how well do you feel the MEO knows exactly what is expected?
There are inevitable disagreements with the MEO over what Is a required output.
Using the 1-7 scale please estimate the magnitude of the problem, with 1 being a
significant problem and 7 being Insignificant?
188
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thinking back to when the work was performed by federal employees not under an
MEO, how would you rate your satisfaction?
Approximately how many different outputs does the MEO perform?
Under 100 100-500 500-1000 over 1000 no idea
Are there reasonably clear definitions of what is required, how often and quality
indicators for each output? Y or N
“Goal alignment” is the level of agreement between the principal and the MEO to
have the same priorities and goals. There is some natural conflict in that the MEO
contractor may want to maximize profits while the principal may desire to minimize
costs. This may lead to a lack of common goals. Using the 1-7 scale with 7 being
strongly agree, 4 being no opinion and 1 being strongly disagree, what are your
thoughts on the following statements.
The goal of the principal is for the MEO to perform the PWS outputs.
The goal of the principal is for the MEO to exceed the PWS outputs and improve
operations at the installation.
The principal is willing to pay more for higher quality or responsiveness.
The principal is working or desires to reduce the costs of the MEO and is willing to
accept a lower performance level.
The goal of the MEO is to meet the PWS requirements.
The goal of the MEO is to exceed performance requirements.
The goal of the MEO is to maximize profit.
The goal of the MEO is to grow in business base or staff.
If a better and less costly way of performing work is found, the MEO is willing to
reduce those costs and pass the savings on to the principal.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
“Moral hazard” is what can occur after the PWS is in place and the MEO is not
freely observed so taking actions that may not be in the best interest of the principal
may occur. This occurs when the MEO is not freely observed and may take actions
more in their own interests or because the MEO has a different knowledge level than
the principal. Moral hazard falls into three categories: Profit Maximization, Shirking
and Information Asymmetry. “Profit maximization” is when the MEO engages in
actions that increase their profit or staff more than is required. An example might be
a cost plus effort where they convince the principal more effort is required than really
is necessary.
Is there a profit motive for the MEO?
Have you been aware of any actions by the MEO that were for the benefit of the
MEO at the expense of the larger organization? Can you give any
examples?
How does your organization insure the M E O does not work for self-fulfillm ent at the
expense of the principal organization?
“Shirking” is when the MEO avoids some level of work. For example, he may not
shovel snow, waiting for it to melt. He might be required to clean the restrooms 5
days a week, but only cleans them 4. If the principal requires the work, the MEO
may have a tendency to avoid the work because it may help increase his profit or
lower his work effort.
Using the 1-7 scale with 1 being a significant problem and 7 being insignificant, do
you see attempts on the part of the MEO to “shirk”? Can you give any
examples?
Do you have any instances of when responsiveness or quality was less than required
by the PWS?
Can you give examples? Was this due to shirking?
Were than any corrective actions initiated?
Can you describe the quality control structure to insure the MEO performs all that is
required?
Thinking back to when the work was performed b y federal employees not under an
MEO, would you rate shirking a greater, lesser or same problem?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
“Information asymmetry” occurs when the MEO has greater knowledge than the
principal and thus is able to direct or influence decision-making. It is similar to
when you take your car to the mechanic, you are the principal, he is the MEO, but
because he has greater knowledge, he really influences decisions you make.
Using the 1-7 scale with 1 being the MEO and 7 the principal, who has greater
knowledge of what is done, how well it is done and what needs to be accomplished?
Do you see any problems or conflicts with the current information level of each
party? If so, could you describe your concerns?
Does the principal keep the MEO and MEQ’s staff informed of the long-term goals
and perspectives? Are there times the principal’s knowledge of
the long-term goals might alter the MEO’s actions? Can you give any
examples?
How do you balance the principal’s need for detailed information on the MEO’s
work processes with the separation of the organizations?
Thinking back to when federal employees not under an MEO performed the work,
did you see any similar information imbalances?
“Incentives and coercives” are those actions the principal is able to take to reward
or penalize the MEO based upon performance. Typically, if the MEO is a contractor,
these are monetary in nature with the ultimate coercive being the cancel of the
contract .and bringing on a new firm. If the MEO is federal employees, the incentives
may be very different.
Are there incentives provided to the MEO? Can you describe these?
On a 1-7 scale, do you believe the incentives influence day-to-day performance?
Typically, to influence performance, rewards incentives should be a frequent
reflection of performance, monthly or quarterly. Are you incentives designed to
reward short-term performance or more long term, annual evaluation?
Are you incentives for the MEO geared to the individual, the cost center, or the entire
agent organization?
Are there coercives available to the principal? Can you describe these?
Has the principal ever invoked coercives? Can you explain why and did the
desired result come about?
1 9 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
On a 1-7 scale, do you believe the coercives Influence day-to-day performance?
Are your incentives for the MEO geared to the individual, the cost center, or the
entire MEO organization?
“Monitoring” is the action taken to provide oversight to the MEO. It is a cost to the
principal that exists solely because there is only a limited trust the MEO will perform
all that is required. Ideally, both parties work to minimize this cost.
Monitoring focuses on three aspects, quantity, quality and cost. Does your
organization have a formal output work count system of the MEO performance?
Can you describe this system?
Does your organization have a formal quality reporting system of the MEO
performance?
Can you describe tWs system?
Does your organization have a cost accounting system of the MEO that enables you
to have a cost per output?
This would be similar to an Activity based costing system. Can you describe this
system?
Can you provide an idea of the cost to monitor the output of the MEO?
Thinking back over the last several years, has this cost increased, decreased, or
remained the same? Thinking to the future, do you foresee any changes in
the cost of monitoring?
Has your organization taken or plan to take any actions to reduce the cost of
monitoring?
Using the 1-7 scale with 7 being very useful and 1 being worthless, how useful is the
monitoring of the MEO function?
“Transaction costs” are the costs of having an A-76 process and structure. When
there was a single, integrated organization, transaction costs were low, command and
direction came from the senior management. With an A-76, principal and MEO
structure there were the ex ante costs, the initial costs to get a contract or MEO in
place. After the structure is in place there are ex post costs. These include
monitoring costs, possible negotiation and renegotiation costs, processing costs if
192
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
there is an increase of some kind. Under an activity based costing concept, we would
know what these costs amount to.
Do you have a cost to initiate the first MEO? This would include PWS development,
contracting and transition costs?
How often do you plan to renegotiate or recompete this MEO structure?
Does your installation principal have a measure of the costs to monitor and work
with the MEO? These would be above normal direction of the work and monitoring
of work. This may be for things such as dispute resolution.
Has there been a need to negotiate or renegotiate the PWS or contract? Could you
explain the circumstances? Do you have a measure of this cost? How did you
measure the costs??
Have there been any unforeseen cost increases after the MEO structure was put in
place??
Is there a process or mechanism in place to limit future cost increases?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Information overload: Exploring management of electronic mail
PDF
Adapting and applying a mission-focused strategic framework for emergency management
PDF
An analysis of health risk selection and quality of care under Medicare fee -for -service and Medicare managed care health care systems
PDF
An organizational history of the National Aeronautics and Space Administration: A critical comparison of administrative decision making in two pivotal eras
PDF
Determining acceptable seismic risk: A community participation-based approach
PDF
A configuration study of multiagency partnerships as practiced in Taipei City government
PDF
Banking on the commons: An institutional analysis of groundwater banking programs in California's Central Valley
PDF
An analysis of third -party regulatory audit systems for medical device safety
PDF
A preliminary assessment of the influence of certain variables on the implementation of recommendations by scientific -technical advisory committees
PDF
Executive spending power: Flexibility in obligation and outlay timing as a measure of federal budgetary and policy control
PDF
Executive development and effectiveness: A study of mobility assignment experiences of career executives in federal research and development agencies
PDF
Evaluation of professional services consultants in rural government
PDF
Federal mentor -protege programs: A pilot study on protege survival and growth
PDF
Historical perspectives and future horizons of local government managers and the International City /County Management Association
PDF
Do uses of human resource information technology (HRInT) tools in federal organizations improve their human resource management productivity?
PDF
Determining risk propensity of government program managers for high risk /high payoff projects
PDF
Interoperability among complex defense computer -based systems: The Navy case
PDF
Educational reform for private state -approved schools in California
PDF
Bridled boldness: Reengineering the Federal Bureau of Prisons. Lessons learned
PDF
Charter schools' influence on public school administrators' innovative behaviors
Asset Metadata
Creator
Gray, John Richard
(author)
Core Title
"Are we really in charge?": An analysis of the principal -agent relationship
School
School of Policy, Planning and Development
Degree
Doctor of Public Administration
Degree Program
Public Administration
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,political science, public administration
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Petak, William (
committee chair
), Bledsoe, Ralph (
committee member
), Pieplow, Tom (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-537858
Unique identifier
UC11339787
Identifier
3094329.pdf (filename),usctheses-c16-537858 (legacy record id)
Legacy Identifier
3094329.pdf
Dmrecord
537858
Document Type
Dissertation
Rights
Gray, John Richard
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
political science, public administration