. ‘ . . M A . ‘ . . . ‘ A ‘ . Jfim‘
.51.} , V , ‘ , a?!

. . . . a
3.": . . . . , . , . . . . . ‘ g .,uu.o.mun.m

. hlm‘g
«3%..3
e . .,

o.‘ r;
7%..)

, ‘v
W,
1"

3.
5

o

v :34
\fi:

.. n
v i . . 7:.
. .735”: n A .tvr
v . at

.nqlnrl. x

 

 

 

 

 

 

 

 

' I" . PMS

:2an

 

'"UBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

BUSINESS CONTINUITY IN THE SUPPLY CHAIN:
PLANNING FOR DISRUPTIVE EVENTS

presented by

Andrea McGee Prud’homme

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Marketing and Supply Chain
Marggement

 

 

W k W xrviqLCn

Major Professor’ 5 Signature
5/ Cl [100%

Date

 

MSU is an afflnnative-action, equal-opportunity employer

 

— ‘.~.—.---—--¢-a-.—t—-—r«——

.‘-‘—.-.-.n‘-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE
Am 1) a 235

 

 

BUSINESS CONTINUITY IN THE SUPPLY CHAIN: PLANNING FOR
DISRUPTIVE EVENTS

By

Andrea McGee Prud’homme

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
For a degree of
DOCTOR OF PHILOSOPHY
Department of Marketing and Supply Chain Management

2008

ABSTRACT

BUSINESS CONTINUITY IN THE SUPPLY CHAIN: PLANNING FOR
DISRUPTIVE EVENTS

By

Andrea McGee Prud’homme

This study examines the planning firms engage in to manage supply chain
risk to maintain or improve the stable flow of goods, services, information, and
funds through the supply chain for enhanced firm performance. Two different
theories are utilized to examine opposing premises related to the occurrence and
management of disruptive events. Specifically, Normal Accident theory is used
to examine the role of supply chain attributes of complexity, dependency and
uncertainty on the use of supply chain risk management planning, the occurrence
of disruptive events, and firm performance. High Reliability Theory is used to
examine the role of both effective business practices and concern for disruptions
on supply chain risk management planning, the occurrence of disruptive events,
and firm performance.

The theories are tested using data obtained from supply chain
management professionals associated with the American Purchasing Society. A
survey instrument was used to obtain information on the perceptions of supply
management professionals regarding their firms supply chain attributes, the use
of supply chain continuity planning, the occurrence of supply chain disruptions
and firm performance. The survey was administered via electronic and

traditional paper mail, with an overall survey response rate of 9.5%.

The findings indicate that firms that engage in continuity planning activities
are positively associated with higher levels of firm performance. Firms that
perceive they have high levels of dependencies within the supply chain are more
likely to engage in risk management activities, although firms with higher levels of
complexity are more likely to experience supply chain disruptions. In additional,
formal supply chain continuity planning, especially when shared with other supply

chain partners, is more effective than informal planning.

Copyright by
Andrea McGee Prud’homme
2008

DEDICATION

While there were many people that provided training, encouragement and
support through my process of earning a Ph.D., there are a few people without
whose support I would never have had the courage to even begin or to at times
simply preserver. I dedicate my dissertation to my family: without their love,
encouragement, advice and shoulders to sometimes cry on, I would not be in the
position to write these words of dedication.

My mother is an amazing woman. Her life long and unfailing example of
hard work, perseverance, and ability to rise to what ever the occasion demanded,
coupled with her belief in my ability and her support and encouragement to
pursue my goals, serves as my personal inspiration and benchmark. Without her
as my rock, my guidepost, and my personal champion, I would not have had the
strength and confidence to spread my wings. My mother has also given me the
gift of wonderful sisters. Their support, encouragement and friendship have also
provided me with the resources to continue to grow and achieve. I can not
imagine how I could have come this far without the constant knowledge that my
mother and sisters are there to help me up when I stumble, to hold my hand
when I need someone to walk beside me, and to help me soar when I want to fly.

My grandparents, who passed on while I was in my Ph.D. program, had
always been a source of support and encouragement. I was blessed to grow up
with a second set of parents to help me to be the best person I could be. Their
unconditional love, joy in my successes and influence is greatly missed, and I

deeply regret that they are not able to share in this achievement.

ACKNOWLEDGMENTS

Obtaining a Ph.D. is a long journey that begins well before the first day of
the first seminar course. I equate the journey to a series of doors opening and
through which I was encouraged and allowed to pass. For me the journey began
subtly and I did not recognize where the first open door would eventually lead.
Dr. Billy M. Thornton supported and mentored me during my time as an adjunct
faculty Colorado State University. He gently planted the idea that I should
continue my education to make the transition from industrial professional to
become a university professor. While at CSU, Dr. Madeline (Mellie) Pullman
opened the next important door for me by advising me on Ph.D. programs and
introducing me to the person that would be the gate keeper of the many
subsequent doors through which I would travel. Dr. Kenneth K. Boyer served as
my advisor, teacher, mentor, coach, champion and friend through my years of
study at Michigan State University. He opened doors that would otherwise have
remained closed, he helped me get through doors that I didn’t think I could fit
through, and pressed me through doors that I didn’t want to pass through. I
could not have done this work without his unfailing guidance and support. When
I found myself at what I thought was a dead end, Dr. David J. Closs opened a
door that I didn’t know existed, allowing me the opportunity to continue on my
journey, for which I am eternally grateful. Dr. Roger J. Calantone provided
invaluable knowledge and guidance to help me with the technical aspects of my
training and research. Dr. George A. Zsidisan’s sharing of his knowledge in the
area of risk management, his support as a friend and his introduction into CAPS
are also most greatly appreciated. And finally, there were many times that l

vi

didn’t think I could pass through another door or I was ready to shut a door
myself. At these points, my family, my closest friends, my fellow Ph.D. students,
and Dr. Michael A. Gross helped to remind me that the journey is long, it is hard,
there are points at which we stumble, and there are hills that seem
insurmountable; but that it is achievable and that I can keep traveling by taking it
one step at a time.

To all of these people I owe a great debt, for without them opening doors
that need opened and by encouraging me to keep passing through them, I would
not have been able to complete this first long stage in the ongoing journey to
become a creator and disseminator of knowledge.

My heartfelt appreciation and deepest gratitude goes to all of these people
for opening doors and for walking along parts of my journey with me. I could not

have successfully made the trip without their support.

vii

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES Y
LIST OF FIGURES Ki
1. INTRODUCTION 1
1.1 INTRODUCTION 1
1.2 PROBLEM MOTIVATION 3
1,3 DEFINITIONQ 7
1.4 RESEARCH QUESTIONS 8
1.5 RESEARCH METHODOLOGY OVERVIEW 10
1.6 RESEARCH CONTRIBUTInNs 11

2. LITERATURE REVIEW 12
2.1 RISK 13
2.1.1 Business Risk 13
2.1.2 Risk in the Supply Chain 17
2.1.3 Risk Planning 25

2.2 NORMAL ACCIDENT THEORY 31
2.3 HIGH RELIABILITY THEORY 38
2.4 SYNTHESIZING THE LITERATURE 44

3. MODEL DEVELOPMENT 55
3.1 CONSTRUCT DEVELOPMENT 56
3.1.1 Supply Chain Complexity 57
3.1.2 Supply Chain Dependency 59
3.1.3 Effective Practices 61
3.1.4 Concern for Disruptions 63
3.1.5 Continuity Planning 64
3.1.5.1 Profile the Supply Chain 65

3.1.5.2 Risk Identification 66

3.1.5.3 Risk .‘ ‘ 67

3.1.5.4 Risk Management Strategies 68

3.1.5.5 Risk Management Practices 69

3.1.5.6 Audit of Risk Management Plans 71

3.1.6 Disruptive Events 72
3.1.7 Firm Performance 74
3.1.8 Demographic Information 76

3.2 HYPOTHESES 76
3.2.1 Grounded in Normal Accident Theory 77
3.2.2 Grounded in High Reliability Theory 79
3.2.3 Impact of Risk Management Planning 80

3.3 SCALE DEVELOPMENT AND VALIDATION 81

viii

4. RESEARCH METHODOLOGY ................................................................................. 87

4.1 GROUNDED THEORY METHOD ................................................................................................ 87
4.2 RESEARCH DESIGN ................................................................................................................ 88
4.3 ORGANIZATIONAL SPONSORSHIP ............................................................................................ 92
4.4 DATA COLLECTION ................................................................................................................. 97
4.4.1 Target Sample ............................................................................................................. 97
4.4.2 Questionnaire Design ................................................................................................ 100
4.4.2.1 Construction of the Survey ................................................................................. 102

4.4.2.2 Survey Pre-testing .............................................................................................. 103

4.4.2.3 Survey Implementation ....................................................................................... 104

4.4.3 Response Rate .......................................................................................................... 110

5. RELIABILITY AND VALIDITY ................................................................................ 114
5.1 DEMOGRAPHICS .................................................................................................................. 114
5.2 MISSING DATA ..................................................................................................................... 118
5.3 DATA NORMALITY ................................................................................................................ 123
5.4 RESPONSE BIAS .................................................................................................................. 124
5.4.1 Non-response Bias .................................................................................................... 124
5.4.2 Common Method Bias ............................................................................................... 125

5.5 CONFIRMATORY FACTOR ANALYSIS ...................................................................................... 127
5.5.1 Measurement of Constructs ....................................................................................... 128
5.5.2 CFA Model Fit ............................................................................................................ 129
5.5.3 Construct Validity ....................................................................................................... 141
5.5.3.1 Convergent Validity ............................................................................................ 141

5.5.3.2 Discriminant Validity ........................................................................................... 143

5.6 RELIABILITY AND VALIDITY SUMMARY .................................................................................... 143

6. DATA ANALYSIS ................................................................................................... 147
6.1 RESULTS OF PATH MODEL TESTING ..................................................................................... 147
6.2 RESEARCH IMPLICATIONS ..................................................................................................... 156
6.3 MANAGERIAL IMPLICATIONS .................................................................................................. 160
6.3.1 Role of Supply Chain Complexity, Dependency and Uncertainty ............................. 160
6.3.2 Formal versus Informal Risk Management Plans ...................................................... 161
6.3.3 Profiling the Supply Chain ......................................................................................... 162
6.3.4 Risk Reduction Techniques ....................................................................................... 163
6.3.5 Importance of Effective Business Practices .............................................................. 163

6.4 DISCUSSION ........................................................................................................................ 166
6.5 RESEARCH QUESTIONS ........................................................................................................ 166
6.6 FUTURE RESEARCH ............................................................................................................. 169
APPENDIX B: COPY OF SURVEY CONTACT MATERIALS .................................... 180
APPENDIX C: SCALE ITEM RESULTS FROM SURVEY ......................................... 184
APPENDIX D: SUMMARY OF FINAL REVISED HYPOTHESES .............................. 190
APPENDIX E: SUMMARY OF DATA ANALYSIS STATISTICS ................................ 191
BIBLIOGRAPHY ......................................................................................................... 192

LIST OF TABLES

TABLE 2.1 SUMMARY OF DEFINITIONS IN INTERACTION/COUPLING MATRIX ........................................ 33
TABLE 2.2 SUMMARY OF SUPPLY CHAIN RISK LITERATURE .............................................................. 47
TABLE 3.1 SUMMARY OF SUPPLY CHAIN COMPLEXITY SCALE ITEMS ................................................. 58
TABLE 3.2 SUMMARY OF PERCEIVED SUPPLY CHAIN DEPENDENCY SCALE ITEMS .............................. 61
TABLE 3.3 SUMMARY OF EFFECTIVE PRACTICES SCALE ITEMS ......................................................... 63
TABLE 3.4 SUMMARY OF PROFILE THE SUPPLY CHAIN SCALE ITEMS ................................................. 66
TABLE 3.5 SUMMARY OF RISK IDENTIFICATION SCALE ITEMS ............................................................ 67
TABLE 3.6 SUMMARY OF RISK ASSESSMENT SCALE ITEMS ............................................................... 67
TABLE 3.7 SUMMARY OF RISK MANAGEMENT STRATEGIES SCALE ITEMS .......................................... 69
TABLE 3.8 SUMMARY OF RISK MANAGEMENT PRACTICES SCALE ITEMS ............................................ 70
TABLE 3.9 SUMMARY OF AUDIT RISK MANAGEMENT PLANS SCALE ITEMS ......................................... 72
TABLE 3.10 SUMMARY OF DISRUPTIVE EVENTS SCALE ITEMS ............................................................ 74
TABLE 3.11 SUMMARY OF FIRM PERFORMANCE SCALE ITEMS ............................................................ 75
TABLE 4.1 SUMMARY OF SURVEY RESPONSE RATE ISSUES ............................................................ 101
TABLE 4.2 RESPONSES RATES ...................................................................................................... 111
TABLE 5.1 INDUSTRY SECTOR OF SURVEY RESPONDENTS .............................................................. 115
TABLE 5.2 INDICATORS OF FIRM SIZE ............................................................................................. 116
TABLE 5.3 RESPONDENrs’ FUNCTIONAL RESPONSIBILITY ................................................................ 117
TABLE 5.4 RESPONDENTs’ LEVEL OF RESPONSIBILITY .................................................................... 117
TABLE 5.5 RESPONDENTs’ ORGANIZATIONAL LEVEL ....................................................................... 117
TABLE 5.6 SUMMARY OF MODEL FIT INDICES .................................................................................. 131
TABLE 5.7 ITEMS REMOVED FROM ASSOCIATED CONSTRUCTS ......................................................... 136
TABLE 5.8 SUMMARY OF FIT METRICS OF FINAL CFA MEASUREMENT MODEL .................................. 140
TABLE 6.1 RESULTS OF H1 TO H8 AND MODEL FIT STATISTICS ....................................................... 150
TABLE 6.2 SUMMARY OF RESEARCH IMPLICATIONS ......................................................................... 159
TABLE 6.3 SUMMARY OF MANAGERIAL IMPLICATIONS ...................................................................... 165

LIST OF FIGURES

 

 

 

 

 

 

 

 

 

 

 

FIGURE 1.1 THE DISRUPTION PROFILE 9
FIGURE 2.1 ENTERPRISE VULNERABILITY MAP 15
FIGURE 2.2 FINANCIAL PLANNING FOR RISK 17
FIGURE 2.3 POSSBILE IMPACT ON THE FRIM 18
FIGURE 2.4 THE ITERATIVE CYCLE 0F LEARNING IN DISRUPTION MANAGEMENT ................................. 26
FIGURE 2.5 INTERACTION/COUPLING MATRIX 33
FIGURE 3.1 PROPOSED MODEL 55
FIGURE 3.2 MODEL WITH HYPOTHFSFS 81
FIGURE 3.3 SCALE DEVELOPMENT PROCESS 84
FIGURE 4.1 DATA COLLECTION TIMFI INF 91
FIGURE 6.1 FULL MODEL SIGNIFICANT PATHS 153
FIGURE 6.2 NORMAL ACCIDENT THEORY MODEL SIGNIFICANT PATHS 154
FIGURE 6.3 HIGH RELIABILITY MODEL SIGNIFICANT PATHS 155

 

xi

1. INTRODUCTION

1.1 Introduction

This study examines the planning firms engage in to manage supply chain
risk to maintain or improve the stable flow of goods, services, information, and
funds through the supply chain for enhanced firm performance. There has been
limited empirical research on the topic of continuity planning and risk
management occurring specifically within the area of supply chain management.
Previous research has been primarily conceptual with limited application of
theory or contribution to theory building. The use of theory here, specifically
Normal Accident Theory and High Reliability Theory, provides a more rigorous
scientific assessment regarding the drivers of supply chain risk management
planning and its impact on the reduction of disruptive events and levels of firm
performance. The theories are tested by means of empirical examination of the
role of perceptions of supply chain complexity, uncertainty and dependency, and
the use of general effective business practices on the planning activities
undertaken to manage disruptive events and improve firm performance. The
findings provide managers insight into the creation and content of business
continuity plans that are used to better manage supply chain disruptions for
enhanced firm performance.

There is growing academic interest in supply chain risk management, as
evidenced by the special issue of Production and Operations Management

entitled “Risk Management in Operations" (Seshadri & Subrahmanyam, 2005)

 

 

and a recent call for papers by Journal of Operations Management for the issue
“Risk Issues in Supply Chain Management." In addition, there are several recent
books targeted at practitioners to better understand the causes of, and guidance
for, planning for disruptive events (Barnes, 2001; Bazerman & Watkins, 2004;
M'rttelstaedt, 2005; Regester & Larkin, 2005; Sheffi, 2005). Finally, there have
been several highly publicized events that clearly demonstrate the need for firms
to engage in better continuity planning. A sampling of recent events include
incidents that have impacted the logistical flow of goods, such as the 1997 strike
by workers at United Parcel Service (UPS) who at the time controlled 80 percent
of package deliveries in the US. (Krause, 1997); and the 2002 closure of 29 US.
Pacific coast ports for 10 days that left over 200 ships anchored off the coast
unable to dock (Michels, 2002). Events can impact component materials used
for production of other goods, such as the 1999 earthquake in Taiwan that
damaged finished inventory and work in processes, as well as severely impacting
production of silicon chips (TSMC, 1999). Materials already in consumer
possession may be impacted, such as the 2006 recall of over six million Sony
batteries used in the laptop computers of Dell, Hewlett-Packard, IBM, Apple, and
Toshiba. An early estimate of the cost for Sony was in excess of $250 million
just for the Dell and Apple negotiated recalls (Sony Corporation, 2006). In short,
there are multiple events that can have a considerable negative impact on the
flow of goods and services through a supply chain.

Clearly there is a need for both academics and practitioners to better

understand the role risk management plays in business performance, which may

be driven by the prevalence of a multitude of events that can disrupt the supply
chain. Rigorous research is needed to better understand the role of supply chain
continuity planning on firm performance and for theory building to provide more

insight into this critical business area.

1.2 Problem Motivation

The focus of specific topics within the general area of supply chain
management (SCM) research is constantly changing. The quality interest from a
decade ago developed into extensive total quality management (TQM) research.
The original focus on material price reduction evolved into a concern for the total
cost of ownership (TCO). More recently, interest regarding the role of supply
chain risk and business continuity management has increased (Hutchins, 2003).

Firms must plan to manage business interruptions, defined by the
Business Continuity Institute as “any event, whether anticipated (i.e., public
service strike) or unanticipated (i.e., blackout) which disrupts the normal course
of business operations at an organization's location." Management of the
unexpected is one of the greatest challenges that business must deal with
(Weick & Sutcliffe, 2001 ), but the difficulty of anticipating unexpected events
makes this an especially challenging business responsibility. It has also been
argued that better understanding of disruptive events to reduce their frequency
and impact is the ethical responsibility of managers (Chapman, Christopher,
Juttner, Peck, & Wilding, 2002). Increased visibility and awareness of the

vulnerability of supply chains has caught the attention of senior management,

who are now more likely to recognize the strategic importance of managing
disruptions (Kleindorfer & Van Wassenhove, 2004). Continuity plans must be
developed not only for events that are likely to happen but in order to account for
the ambiguity and complexity found in today’s supply chains, for events that are
speculative and in the realm of ‘could be’ (Hamel & Prahalad, 1994; Pich, Loch,
& Meyer, 2002), as well as for those with potentially greater visibility. However,
much of the supply chain improvement efforts undertaken by firms have been
focused on reducing costs and increasing efficiency, rather than on specifically
managing supply chain risk (Hendricks & Singhal, 2005b).

To secure supply chain performance, firms must plan to both mitigate and
manage disruptive events that could result in potential losses. Losses include
those that directly impact the firm’s reputation, functionality or its financial
performance. To better manage such disruptions, there is a need for enhanced
understanding of how firms create plans to both protect the supply chain from the
occurrence of events and to diminish the impact should events occur. This
understanding should include the process firms follow in risk management
planning and the content of the plans created for upstream, internal and
downstream supply chain continuity management. The scope of continuity
planning has widened from a more narrow focus on ‘disaster recovery’ and
preserving tangible assets to include a more comprehensive and proactive
‘business continuity’ approach. This more holistic view includes processes,
networks, flows, relationships and affiliations for ongoing business functionality

(Rodetis, 1999). This broadening of scope also includes moving beyond supply

disruption and price fluctuation to included less tangible issues of brand,
reputation and ethical issues (Khemani, 2007).

The need for increased planning to reduce and manage disruptive events
is driven by a few critical factors. As firms have moved to a supply chain model
that includes sourcing and market diffusion on a more global scale, the result is a
flow of goods that is often physically elongated entailing longer and more variable
lead times, with business entities that are more geographically dispersed (Sheffi,
2001; Hillman, 2006); in short, supply chains have become more complex. This
complexity may lead to a higher likelihood of a disruptive event that could
seriously impact a single supply chain member or ripple through the supply chain
to negatively impact additional nodes (Choi, Dooley, & Rungtusanatham, 2001).
On the other hand, if a firm is able to recognize the hazards within its own
organization and the supply chain within which it operates it may be possible to
create effective plans to mitigate the probability and the resulting negative impact
of disruptive incidents. It is the complexity of supply chain structure which may
influence the occurrence of business interruptions and the potential for well laid
plans to positively influence firm performance that serves as the basis for this
research.

Given the large number of events that can occur in a supply chain either
inside a firm or in conjunction with its upstream and downstream partners, it is
nearly impossible to predict all possible types of interruptions that could be
experienced (Lewis, 2003) or even to predict all possible sources of disruptions

(Christopher & Lee, 2004). Varying supply chain configurations result in variable

‘shapes’ of supply chain risk (sources, nature, triggers, scale, speed of onset,
and severity of impact) that increase the challenges of managing risk (Ritchie &
Brindley, 2007a). It is therefore prudent for managers to take proactive
measures in order to effectively 1) extend the time until an event of serious
magnitude occurs, 2) to take steps to mitigate the effects of an event and finally,
3) to recover from the effects of such an event to increase the likelihood of
ongoing profitability and survival of the organization.

The relatively nascent level of scholarly study of the topic of risk
management as it specifically relates to the supply chain presents both
opportunities and challenges. Much of the research to date in this area has been
conceptual and practitioner focused with some theory testing (Smeltzer & Siferd,
1998; Svensson, 2000; Zsidisin & Ellram, 2003; HellstrOm, 2005; Zsidisin,
Melnyk, & Ragatz, 2005) rather than on theory building. There is also a
corresponding lack of research to define relevant constructs; some preliminary
constructs have been created (Svensson, 2002b), but there is no consensus for
generally accepted constructs that have undergone rigorous testing in multiple
studies. Therefore further research should entail the creation, testing and
possible modification of constructs and corresponding scale items to examine the
area of risk and business continuity within the context of the supply chain.

There has been work examining the types of disruptive events that firms
attempt to provide protection from (Zsidisin, 2003), and there are also practitioner
focused publications that provide advice for what should be included in risk

management or business continuity plans (Elkins, Handfield, Blackhurst, &

Craighead, 2005; Sheffi & Rice Jr., 2005). However there has been little work to
methodically and rigorously examine the risk management planning process and
its impact on the occurrence of disruptive events and firm performance. This
study examines the application of Normal Accident Theory and High Reliability
Theory to: 1) understand if and how firms are engaged in risk management
planning to provide protection from disruptive events; 2) determine if the planning
reduces the frequency or impact of disruptive events; 3) determine if the planning
improves firm performance; and 4) identify theory and tools that can enhance

supply chain continuity knowledge.

1.3 Definitions

Risk and disruption management requires the review of several streams of
literature including operations management, supply chain management, risk
management, organizational behavior, and others. These literature streams
often use differing terminology or use the same terms with a slightly different
nuance of meaning. Therefore, it is important to define the terms to be used here
in discussing these concepts, and the specific constructs that are developed for
model testing. The following is a brief discussion of the terms used for this study
and their associated meanings, within this bounded context:

- Incident, event, disruption, disturbance, and occurrence are used
interchangeably to refer to an experience that interrupts or completely

stops the flow of goods, services, information, and funds through a

supply chain or which results in other negative quantitative or
qualitative consequences for a firm or supply Chain.

. Entity, associate, member, and partner are used interchangeably to
refer to parties doing business with each other and that comprise
separate business nodes within a supply chain. Although the term

‘partner’ is often used to denote a specific type of long-term contractual
and cooperative business relationship, within this study it does not
imply any special type or configuration of interaction or link between
business entities.

- Chance, probability, and likelihood are used interchangeably when
discussing risk and the possibility of an unfavorable event occurring.

- Risk is used in general terms to indicate that there are various calculable
probabilities surrounding a variety of possible events and their outcomes
that could have a negative effect on the flow of goods, services, funds and

information resulting in some level of quantitative or qualitative loss for the
firm.

1.4 Research Questions

The disruption profile in Figure 1.1 illustrates the stages that occur when a
firm experiences an event that inhibits normal business operations (Sheffi & Rice
Jr., 2005). Planning has been identified as occurring in the preparation or steady
state of the disruption profile. When an event is detected, a first response occurs
to manage the initial impact. Preparation for recovery is meant to mitigate the
event's full effect by limiting the depth and length of the performance decline.
Finally, recovery efforts are implemented to return the firm to previous (or

improved) performance levels.

Figure 1.1 The Disruption Profile

      
 
   

PLANNING Disruption Length
A

 

 

  
  
 
 
 
     
     

 

 

 

 

 

 

 

Long
0 ................................................................................... 8‘: ........... Term
I - - 0
g . . Ilnmal Impact 0 Impact
5 Preparatron I . .
E Steady State I
g ‘ ': Preparation
3 " for r overy
.E 3
a I
s I
0 2
'3 Event E :\ Full
g Occurs : : Impact
.. x = ,
\ k j
V V Time
Impact Lags Length of Recovery

Modified from Sheffi and Rice (2005)

The purpose of this research is to better understand the processes and
techniques firms utilize in the planning stage of supply chain continuity
management, before an event occurs. In order to focus and bound this research
project, ex post activities used to manage potentially negative consequences
after an event has occurred to return to a steady state are excluded from explicit
examination. Within this study there is no focus on any particular type of
disruption. Although there has been work specifically focused on the prevention
of events that are the result of malevolent human actions and are studied within
the area of supply chain security (Closs & McGarrell, 2004), the intention here is
to examine the general planning activities used by firms to improve the potential

for business continuity regardless of the source of the disruptive event.

The focus on the supply chain risk management planning firms undertake

within the study is built upon the alternative theories of Normal Accident (Perrow,

1984; Perrow, 2004) and High Reliability (Roberts, 1989). The literature review

and interactions with industry experts suggest the following research questions:

1.
2.

What factors drive firms to engage in supply chain continuity planning?

How are firms undertaking the creation of supply chain continuity plans
(process)? What are firms including in supply chain continuity plans
(content)? What are the ‘best practices’?

To what extent are firm’s engaged in supply chain continuity planning
upstream, internally, downstream or some combination of directions?

Does increased supply chain continuity planning result in a reduced
frequency and impact of disruptive events and in improved firm
performance?

Does supply chain complexity, uncertainty or dependency influence the
processes and effectiveness of supply chain continuity planning
processes or results?

These questions have not been definitively addressed in the extant supply

chain management literature, yet represent important aspects of business

planning within the complex and uncertain environment of contemporary supply

chains.

1.5 Research Methodology Overview

This empirical study utilizes a survey to gather data from supply chain

management practitioners. Members of a professional organization for supply

management are utilized as the sample in this study. A survey instrument was

created based upon constructs and scale items generated from current

knowledge regarding risk management, supply chain management, complexity,

1O

and Normal Accident and High Reliability Theories. Structural equation modeling

(SEM) is utilized to examine the proposed model.

1.6 Research Contributions

This study seeks to advance supply chain continuity planning and risk
management by examining the process of planning for disruptive events which
can mitigate the frequency and impact of occurrences negatively impacting a firm
and its supply chain. A comprehensive literature review synthesizes findings and
theory from several streams of literature to provide insights into this complex
area. The literature review and the study findings can be used as the foundation
for a stream of research examining the use of logistics, procurement, inventory
management, etc., in supply chain risk management and continuity planning. The
findings will be useful to firms with exposure to unplanned abrupt natural events
(e.g.: floods, tornadoes, earthquakes), unplanned for but intentional human
events (e.g.: terrorism or sabotage); or unplanned unintentional events (e.g.: fires
or power outages). In addition, the application of alternative theories that have
been used in management but have had limited application or testing in supply

chain planning furthers theoretical development in the field.

11

2. LITERATURE REVIEW

Several streams of literature and research form the basis of this study. A
general examination of risk in business is followed by a discussion of the specific
role it plays within the context of a supply chain. Risk is a multidimensional
concept, specific aspects of which have been extensively explored in areas such
as insurance, finance, psychology, and organizational behavior. The intention
here is not to fully explore risk as it relates to all business and behavioral areas,
but to provide a general working definition of risk and the specific role of risk in
supply chain continuity planning. After a discussion of the relevant risk literature,
an examination of Normal Accident Theory and High Reliability Theory is
conducted. Normal Accident Theory and High Reliability Theory represent a
balance that businesses should consider when undertaking risk management in
order to create appropriate continuity plans. Briefly, Normal Accident Theory
states that all complex systems are subject to failure (Perrow, 1984; Perrow,
1999) while High Reliability Theory states that proper system structure and
management can prevent failures (Roberts, 1989; Weick, Sutcliffe, & Obstfeld,
1999; Weick 8. Sutcliffe, 2001). The application of these two theories within the
context of the supply chain as a system will be examined, and serves as the
theoretical basis of the proposed model to be tested, along with the associated

constructs.

12

2.1 Risk

The Merriam-Webster Dictionary defines risk as “the possibility of injury or
loss”. A common measurement of risk is based upon the probability or frequency
of an event and upon the magnitude of the impact should the event occur.
Quantitatively this can be represented as Risk = Probability * Impact (Weber &
Milliman, 1997; Jia, Dyer, & Butler, 1999; Pich et al., 2002) where the assumption
is that the impact is negative and involves some form of loss for the firm. The
risk of loss is a result of the interactions of processes, people, systems, or
external events that lead to failure or inadequate performance (Manning &
Gurney, 2005). While there may be uncertainty about the actual outcome, the
various probabilities of the potential outcomes are non-zero and may be known.
An important element that is missing from this definition of risk is the ability to
exert some control over the possibility and magnitude of loss exposure (Weber &
Milliman, 1997). Therefore, it is worthwhile to explore ways in which firms can
create plans to minimize or control risk for improved performance and ongoing

business functionality.

2.1.1 Business Risk

Risk management and business continuity have been examined in a
number of business disciplines including facilities management (Hardy &
Roberts, 2003; Castillo, 2004; Gill, 2006), information systems (Cerullo 8. Cerullo,
2004), banking (Bielski, 2003; Manning 8 Gurney, 2005), real estate (Temba &

Jeff, 2002; Foster & Dye, 2005), accounting (Alonso & Boucher, 2001), project

13

management (Pich et al., 2002), and insurance (Urrutia, Vu, Gronewoller, &
Hoque, 2002). However, the specific discipline in which a research stream of risk
management is grounded will influence the approach used and the particular
aspects of risk being studied. Risk in these different streams has been examined
as the consequence of an actual choice, as affecting the human or managerial
decision making process, as dependent upon the context of the situation, and as
the preferences or risk propensities of managers and other decision making
individuals (Weber & Milliman, 1997; Conchar, Zinkhan, Peters, & Olavarrieta,
2004). It is beyond the scope of this study to provide an exhaustive review of the
study of ‘risk’ in the business, psychology and organizational behavior literature.

For the purposes of the current study and as indicated earlier, ‘risk’ is defined as:

the various calculable probabilities surrounding a variety of
possible events, and their outcomes, that could have a negative
effect on the flow of goods, services, funds and/or information at
one of more supply chain nodes and that results in some level of
loss for the firm

It is important to note that for the purpose of this study, continuity planning is not
to be limited the need to control or lessen the frequency or impact of very large
scale, catastrophic events that seriously damage a firm. While these types of
high-profile events can without doubt individually have a considerable impact on
an organization and often receive a great deal of publicity, they are not the only
type of event that firms should seek to provide protection from. Smaller events
can also adversely impact a business, especially if they occur repeatedly

resulting in a large cumulative effect. Included in this study are disruptions of any

14

size or source which negatively impact the firm or the functionality of the supply
chain.

The management of risk for business continuity should take into account
events ranging from the potentially catastrophic (low or high frequency and
substantial impact) events to the less disastrous (low frequency and minimal
impact) events, as shown in the enterprise vulnerability map in Figure 2.1.

Figure 2.1 Enterprise Vulnerability Map

 

High

     

3 4

Minimal Substantial
Consequences l Severity

Frequency I Probability

 

 

Low

 

Research to date has tended to focus greater attention on risky events that have
a higher likelihood and a larger negative effect (Zsidisin et al., 2005). While it is
imperative that firms plan to protect against incidents that occur in area 2 (high
frequency/substantial impact), if there are enough events in area 4 (low
frequency/substantial impact) or area 1 (high frequency/minimal impact) the
cumulative results can still be very detrimental (Hallikas, Karvonen, Pulkkinen,
Virolainen, & Tuominen, 2004; Sheffi, 2005) and come to result in a major

damaging effect on the firm. The primary goal of risk management is to try to

move from areas 1, 2 and 4 into area 3 for a reduction in both the frequency and
impact of events, while also considering the cost/benefit of the possible
strategies and outcomes for making such shifts.

Because it is not possible to eliminate all possible sources of loss, firms
will often include the financial impact of some negative events into their budget or
even create and fund specific hazard accounts for identified risks based upon a
cost/benefit analysis. ‘Significant risk’ represents a possible hazard that would
require the expenditure of capital (rather than budget) to cushion against the
resulting impact of an event or which would produce considerable damage to the
firm (e.g. reputation) and must also be part of the financial planning process.
Therefore firms must create financial plans to account for the variety of events
that could occur (Figure 2.2), from the relatively small and mundane to much
larger incidents that could be substantially harmful to the firm (Manning and
Gurney, 2005). This would include explicitly budgeting for potential expected or
higher frequency events that cumulatively can have a large negative impact,
while also attempting to create plans to reduce their occurrence. In addition, it
may be necessary to set aside funds for lower frequency events that individually
can have a large, detrimental impact. Finally, there are events that are so
unlikely that even though their occurrence would be catastrophic for the firm, little
if any financial planning would occur, since current assets could be better used in

other ways.

16

Figure 2.2 Financial Planning for Risk

Expected losses,

 

“ economic reserves as
g part of the budget
“6
3. Losses not part of the
:5 budget, requiring Losses so unlikely,
0 economic capital no financial planning
3 is undertaken
2
D.
—>

 

Cost of loss
Adapted from Manning and Gurney (2005)

2.1.2 Risk in the Supply Chain

An established measurement of risk is based upon the probability or
frequency of an event and upon the magnitude of the impact should the event
occur (Weber & Milliman, 1997; Jia et al., 1999; Pich et al., 2002). An important
element in the evaluation of risk is the ability to precisely determine the
probability of an outcome (Conchar et al., 2004). When it is not possible to
determine the probability distributions within given scenarios, the lack of
information creates an ambiguous situation (Ghosh & Ray, 1997) for a decision
maker. Given the extremely large number and variety of disruptive supply chain
events that could occur (considering disruptions from the environment, upstream,
internal, or downstream sources) combined with the number of entities that may
experience events, it is evident that a supply chain operates in a highly

ambiguous environment.

Risk taking requires confidence in decision choices. However, ambiguity
can erode that confidence because a lack of information can be disturbing and
uncomfortable (Heath & Tversky, 1991 ) for a decision maker. It has been
demonstrated that there are elevated perceptions of risk in ambiguous
environments, and that ambiguity itself may be perceived as risk (Ghosh 8. Ray,
1997) where an increase in the range, variance or expected loss of possible
events leads to an increase in perceived risk (Jia et al., 1999). Within a supply
chain, disruptions can span the continuum of impact on the firm from no or
negligible impact through varying degrees of effort to restore normality to serious
long-term damage or even the possible demise of the firm, as shown in

Figure 2.3. It has been empirically demonstrated that poor previous
experiences with risk reduces self-efficacy and increases a desire for risk
avoidance; conversely high self-efficacy has been associated with an increase in
risk taking (Krueger Jr. 8. Dickson, 1994). It has also been argued that risk is
closely related to the potential for losses (Conchar et al., 2004). It seems
reasonable that when organizations perceived that they are vulnerable to supply
chain disturbances, especially if there is high variation in the possible type,
source and impact, that a greater focus would be placed on planning to reduce
the probability and impact of future events in an effort to reduce future negative

impacts and losses.

Figure 2.3 Possbile Impact on the Frim

 

‘ No Negligible Moderate effort Major effort Serious Demise
impact impact to regain to regain long term of firm
normality normality impact

18

 

Supply chains are a complex collection of business entities involving
multiple players performing a series of activities to supply, transform and deliver
goods and services, including exchanges of information and funds, to meet
market demand (Chen & Paulraj, 2004). While a comprehensive definition would
include the reverse or backward flow of goods, only the fonlvard flow is
considered in the scope of this study. This supply chain definition can be further
expanded to include their adaptive and complex nature as a collection of firms,
entities or agents that seek to maximize profitability and increase firm longevity
through the exchange of goods and services (Choi et al., 2001). Although it can
be argued that most supply chains are really supply networks, the term supply
chain will be utilized within this study, and the examination of a true supply
network is not attempted.

As a large-scale system (due to the number of members), supply chains
exist within a paradox of autonomy and dependence. Each individual supply
chain entity is an autonomous enterprise, yet is also interdependent upon the
other autonomous entities (suppliers and customers) with which it must do
business (Grabowski & Roberts, 1997). This creates a substantial challenge in
the management of a supply chain which stems from the need to control the
boundary Spanning activities that occur between a firm and the upstream and the
downstream tiers that form the supply chain. The autonomy of firms and their
decisions, combined with the fact that their actions (or lack thereof) also impact

the other firms upon which they are mutually dependent through linked business

19

processes, contributes to supply chain vulnerability to disruption when an event
(or series of events) occurs at one point in the system that inhibits the normal
business operations and flow of goods and services at that point, or possibly
beyond. The impact of such events may ultimately propagate through the supply
chain creating difficulty or an inability to continue to meet customers' needs as
scheduled (Spekman & Davis, 2004; Wu, Blackhurst, & O’Grady, 2007).
Although the likelihood that a serious incident would occur at any particular
location is rather small, the number of firms that are part of a supply chain result
in an increased collective probability of a disruptive occurring at some point in a
given supply chain that would result in a negative impact on other members of
the supply chain (Sheffi, 2005).

Recent changes in supply chains as a result of globalization and the
outsourcing of multiple functions (manufacturing, logistics, IT services,
warehousing, etc) have created longer, more complex and further dispersed
networks (Cavinato, 2004; Christopher & Lee, 2004; Hillman, 2006).
Globalization of supply chains means that firms have a greater flow of goods and
services moving across national borders and need to respond to differing
regulatory requirements (Closs & McGarrell, 2004). Growth in the outsourcing of
business activities has resulted in firms having a greater dispersion of processes
being performed outside of the firm’s immediate control (Bowersox, Closs, &
Stank, 2000). Deregulation in a number of industries and the increasingly rapid
rate of technological change have also increased the environmental complexity

Of supply chains and the turbulence that firms face (Siggelkow & Rivkin, 2005).

20

As firms become more dependent upon the actions and activities of other
members of the supply chain, the level of the organization's exposure to
disruptive events becomes increasingly dependent upon those same external
links (Pfeffer & Salancik, 1978; Zsidisin, Panelli, & Upton, 2000). These
changing and increasing risks require new management responses and
processes (Brindley & Ritchie, 2004; Christopher & Lee, 2004). Simultaneously,
supply chains are also experiencing the prevalence of practices such as lean
systems and JIT, which have led to a consolidation of firms’ operations and
supplier bases, leaner operational practices, and tighter inventory policies that
remove the slack that serves as buffers between processes and firms. Both of
these trends contribute to an increase in supply chain fragility and vulnerability to
disruptions (Kleindorfer & Van Wassenhove, 2004; Zsidisin et al., 2005). The
best efforts of business to make supply chains more efficient and to run with
lower costs may have had the unintended consequence of increasing the risk of
a disruption occurring or in the possible expansion of the impact magnitude when
an event occurs (Sheffi, 2005).

As firms continue to try to ‘do more with less’ in efforts to reduce costs and
increase efficiency, fewer resources are available for proactive environmental
scanning. In general, employees need to have time available to identify and
consider possible events to make proactive plans, with the implication that they
are therefore freed from spending all of their working hours on direct, tactical
daily tasks. Another problem is that investments for the improvement of a

process (in this case improving supply chain continuity) often do not yield

21

immediate, directly linked or easily visible results. It may not be possible to know
and analyze the quantitative or qualitative impact of an event that could have
happened, but didn’t. This lack of a direct, easily traceable link between efforts
and results, combined with an inability to capture the savings achieved by
avoiding an event, may reduce the desire and likelihood of firms to make efforts
or investments in supply chain risk management (Repenning & Sterman, 2001).
A reduced level of available attention for, and investment in, managing non-
recurring events and activities may reduce the costs of daily or routine
transactions, but incurs a trade-off cost by leaving fewer resources available to
plan for unexpected events. This may increase the potential for damage from
surprise events that disrupt the supply chain (Lampel & Shapira, 2001; Weick &
Sutcliffe, 2001).

The management of operational risk within the supply chain has growing
practical concern yet is an under-researched area within operations and supply
chain management literature (Killian, 2002; Peck, 2005), in spite of empirical
evidence demonstrating the negative impact of supply chain disruptions on the
firm. Even a relatively small, single disruption of short duration has been shown
to have a detrimental effect on a firm’s financial performance and stock price. An
announcement of a product or shipment delay can have a swift impact, reducing
stock prices by over 8% immediately after an event, and with a long term
reduction of up to 20% over the following six months (Hicks, 2002). Hendricks
and Singhal conducted a study resulting in empirical evidence of the substantial

negative impact of supply chain ‘glitches‘ or disruptions on firms. Their first

22

findings relate to the profitability and cash flow of the firm (Hendricks & Singhal,
2005a) where they found that firms experience a mean 6.92% reduction in sales
growth, a 10.66% increase in costs and a 13.88% increase in inventory levels
following a supply chain glitch. Interestingly, the findings are consistent across
multiple industries and are apparently unrelated to the cause or source of the
disruption. In addition, there is a damaging impact on market perceptions of the
equity risk of the firm (Hendricks & Singhal, 2005b). An increased perception of
firm riskiness may cause investors to demand higher rates of return and increase
the finn's cost of capital. Further, customers and suppliers may demand
assurances and impose contractual guarantees for supply chain performance.
Disruptions of all kinds and across industries appear to result in sizeable financial
and potential relationship penalties for firms who do not manage supply chain
continuity well.

Catastrophic events on a larger scale resulting in major losses (and often
considerable negative publicity coverage), such as the Tylenol tampering in 1982
or the 1989 Exxon Valdez oil spill, have also been shown to have an immediate
negative impact on stock prices of 5-11%, and long term impact of up to 15% for
firms that are less effective in managing the recovery process (Knight & Pretty,
1998). Conversely, it has been shown that announcements of the
implementation of supply chain management tools, technologies or processes
that are viewed as effective or ‘best’ practices can have a positive impact on
financial performance and stock values (Filbeck, Gorman, Greenlee, & Speh,

2005).

23

There is clearly strong evidence that effective supply chain management,
specifically including the improvement of supply chain continuity, can help an
organization to avoid unfavorable financial implications and market perceptions
that may result from publicly known disturbances in the flow or timing of goods or
service availability. This is especially important in light of the apparent
perception of the market that it is irrelevant as to the cause, reason or industry in
which the event occurs; all firms are expected to maintain a smooth and
uninterrupted flow of goods and services through their supply chains (Hendricks
8 Singhal, 2005a).

In addition to adversely impacting shareholders, an operational disruption
can have a detrimental impact on employees, customers, suppliers, and other
stakeholders (Lewis, 2003; Hendricks 8. Singhal, 2005b; Ritchie & Brindley,
2007b). Poor management of a disruption can also adversely impact a firm’s
reputation, hurt brand image and equity, firm credibility, and damage customer
relationships. In contrast, handling such events well can eventually lead to
stronger branding in the long term (Rodetis, 1999; Chopra & Sodhi, 2004).
Finally, there are regulatory requirements impacting fiduciary activities (e.g.
Sarbanes-Oxley) or the movement of goods (e.g. Home Land Security) that
require firms to more explicitly address risk planning and management (Hutchins,
2003; Manning & Gurney, 2005). Overall, firms have ample incentive to engage
in effective planning to reduce the occurrence of disruptive events.
Unfortunately, while there is strong evidence of a negative link between supply

chain disruptions and firm performance, there has been little scientific

24

examination of the planning that firms engage in to reduce the frequency and
impact of their occurrence.

It is important to note that not all disruptions lead to long term or
permanent damage to the firm. Take for example Nokia and Ericsson; both
experienced a massive supply side disruption when the facility of a key supplier
to both firms had a sizable fire in a manufacturing facility. Nokia, working closely
with the supplier, was able it implement plans that allowed for a more rapid
recovery than Ericsson. Within a year of the incident, Ericsson’s market share
had dropped to the point that it withdrew from the market and Nokia’s market
share grew substantially when Ericsson, its major competitor, left (Norrman 8
Jansson, 2004; Sheffi, 2005). Given the demonstrated impact of even a small
disruption, the potential of a large unscheduled event such as a disruptive
disaster that completely shuts down a facility, or which severs a link in the supply
chain in which the firm is embedded, creates a strong motivation for firms to
create plans to better deal with such events. Proper continuity planning may be
able to help firms to avoid undesirable events that in turn may lead to financial
repercussions. Planning well may even allow the firm to emerge stronger

following an incident than it was before (Norrman 8 Jansson, 2004).

2.1.3 Risk Planning

The effective management of disruptions is a multi-step process that
should begin long before an event actually occurs (Helferich 8 Cook, 2002). The
common first step is mapping the supply chain to better understand the network

of linkages in which the firm is imbedded. This is followed by the identification of

25

the types of events that could occur and their possible ramifications (Chopra 8
Sodhi, 2004), which may vary with the business environment and structure of a
specific supply Chain. It is also important to identify the signals that may indicate
a disastrous event is developing and that action is need to either prevent it or to
mitigate its impact (Mitroff, 1988).

Once possible sources and impacts of events have been analyzed, the
firm should create plans that allow for possible avoidance of events. When
avoidance is not possible, plans must be put in place that reduces the impact
when events do occur (Lewis, 2003; Kleindorfer 8 Saad, 2005). The project
management literature (Pich et al., 2002) adds still further to this with a final
element of learning for continuous improvement in all of the steps to mitigate and
manage an event (Figure 2.4). The purpose of continuity planning for ongoing
supply chain functionality is first to try to prevent events from happening with
early detection and action, and secondly to reduce the severity of the event,
allowing the firm to recover to a level of supply chain performance that is, at a

minimum, no less than before the event occurred.

Figure 2.4 The Iterative Cycle of Learning in Disruption Management

 

 

 

 

Learning
I I I I Event I I
.. . . Preventionl
Planning P MItIgatIon 9 DetectIon 5 Preparation b Response 5 Recovery

 

 

 

 

 

 

 

 

 

 

 

 

 

 

26

The impact of disruptive events on a supply chain can manifest in multiple
forms. The first and perhaps the most obvious form is a disturbance in the
availability of physical goods or materials. This can be either a deficiency, as in a
stock-out or shortage; or it can be an over abundance of inventory that is not
immediately needed, is unwanted or is obsolete (Cavinato, 2004; Christopher 8
Lee, 2004; Spekman 8 Davis, 2004). The issues associated with a lack of
material may lead to missing scheduled customer shipments, expedited
processing and shipping, or the need to purchase materials from non-usual
sources — often at a higher price. Having more inventory than desired leads to
an increase in carrying costs, driven primarily by the opportunity cost of capital
and the possibility of obsolescence which impacts the cash flow and investment
capacity of the firm.

A second manifestation of supply chain risk deals with the flow of
information (Spekman 8 Davis, 2004) including a lack of material or product
visibility within the pipeline, especially when the supply chain is physically
elongated due to off-shore sourcing and/or servicing a global customer market
(Christopher 8 Lee, 2004). There could also be the distortion of demand
information as it flows through the supply chain (Forrester, 1961; Lee,
Padmanabhan, 8 Whang, 1997). Information on market intelligence of events
that are occurring in the business environment (e.g. competitor actions, changes
in regulatory requirements, etc.) is also important for effective management of
the supply chain (Cavinato, 2004), and can be used to detect possible precursor

signals that an event may be impending or beginning to unfold.

27

A third impact is financial, where a flow of funds is parallel to the flow of
materials. Many times supply chain location decisions are based upon
advantageous tax allowances, currency valuation or other financial incentives;
when these change there may be an impact on the cost structure of the supply
Chain. Supply chain members can be susceptible to changing material pricing,
currency fluctuations, effects of hedging, credit worthiness of supply chain
associates, timeliness of payments, and the financial stability of suppliers and
customers (Cavinato, 2004; Spekman 8 Davis, 2004; Sheffi, 2005).

A fourth manifestation is relational and includes the appropriateness of
supplier and customer selection, the types of relationships firms engage in, the
degree of interdependence, and the tendency of supply chain entities to engage
in opportunism. An important aspect of the selection of supply chain business
relationships is the ability to discover and bring to market advances in
technology, products, services and processes, as well as the ability of supply
chain entities to protect the image, brand equity, and reputation of the firm
through appropriate behavior, business practices and corporate social
responsibility. Appropriate supply chain partner firms should be selected based
upon their ability to advance the goals of the firm in addition to their ability to
provide protection from undesirable disruptions or negative business practices
(Cavinato, 2004; Spekman 8 Davis, 2004; Sheffi, 2005).

The fifth dimension encompasses hazards to physical assets such as
substantial damage or destruction of a facility from causes such as a flood, fire,

or severe weather. It may also take the form of product or process failure that

28

may cause harm to individuals or to the environment which creates significant
liability and undesirable press for the firm (Sheffi, 2005).

Although disruptions can manifest in multiple forms, it is the dependencies
between firms that may allow disruptions to propagate across entities, making
them susceptible to disruptions in the flow of goods and services when an event
occurs at some point in the supply chain. There are three types of dependencies
within a supply chain. The first relates to time, where there are time lags
between processes to move goods or funds, to provide services, or to share and
utilize information. These time lags are further exacerbated by global supply
chains which have longer physical paths for goods to travel, and conversely by
the shortened ‘clockspeed’ of many product life cycles (Kleindorfer 8 Van
Wassenhove, 2004). The second dependency relates to functionality. Firms
perform unique activities which serve to complement other firms in the network.
The third dependency is relationships, where business activities rely on the
interaction of processes within the supply chain (Wilding, 1998; Svensson, 2004).
These three dependencies put supply chain members at risk if an event occurs
which impacts the ability of one member to continue to engage in normal
business activities to maintain the flow of goods and services through the supply
chain.

No organization is entirely immune from experiencing a disruptive event.
There is a compounded probability that an event will occur at some individual
node in a supply chain due to the increasing number of entities that each have

their own distinctive vulnerabilities and corresponding chance of an incident.

29

Therefore, firm’s must take steps to build a supply chain that is both secure
(reduced likelihood or lower impact of events) and resilient (able to adapt,
respond and recover when events do occur) to minimize the total impact of
disruptive events on the supply chain (Rice Jr. 8 Caniato, 2003; Sheffi, 2005).
However, recent research suggests that, surprisingly, only 5%-25% of firms listed
on the Fortune 500 are formally prepared to handle a crisis event or major
disruption (Mitroff 8 Alpaslan, 2003) and less than 20% of firms examine their
supply network at least annually (Hillman, 2006). This lack of planning occurs
despite industry statistics that reveal the number of natural and human induced
disruptive incidents has been increasing (Rodetis, 1999). Recently, we have
experienced events that have had substantial impact over wide geographic and
economic areas. Examples include several highly visible natural disasters (e.g.
the devastating 2004 Tsunami in the Indian Ocean and 2005 hurricanes Katrina
and Rita in the US. Gulf region), economic and managerial disasters (e.g. 2001
Enron and 2002 WorldCom bankruptcies in the US, and 2007/8 sub-prime
mortgage crisis), and infrastructure failures (e.g. 2003 power outages that lasted
for several days in the northeastern US. and southeastern Canada). In addition,
there are the large scale threats of future events, such as the potential pandemic
of Avian Flu.

Not all disruptions have such widespread impact as those mentioned
above. It is possible that the impact could be quite localized, such as the recent
experience of General Motors when an exclusive supplier of stamped metal parts

decided to terminate the business relationship (Chappell, 2006); or when a group

30

of 33 employees, representing three quarters of the family owned printing
company’s workforce won a Powerball Lottery on a group ticket, resulting in a
payout of over $2.6 million each (Delaware Lottery Games, 2004). It is therefore
likely that a growing number of firms, from large to small, are now more aware of
the need to create plans to prevent and mitigate the impact of disruptive events
that would impede the continuity of flow of goods and services through the supply
chain. Given the need for firms to engage in incident planning and continuity
management, and the primary focus of research to date on supply chain
disruptions from a very general or high-level view (Blackhurst, Craighead, Elkins,
8 Handfield, 2005), there is a need to more closely examine the ways in which
firms manage exposure to disruptions and how plans are created to prevent and

recover from possible events for ongoing performance.

2.2 Normal Accident Theory

This section will provide a brief overview of Normal Accident Theory
(Perrow, 1984; Perrow, 1999). Although this theory has not had widespread
direct application in operations or supply chain management literature, several of
its underlying concepts, such as complexity, interdependencies, and coupling
have been utilized. A review of the use of the key elements of Normal Accident
Theory and an application within the context of a supply chain follows.

Normal Accident Theory as put forth by Perrow (1984; 1999) offers a
pessimistic view of systems which states that any complex and interdependent

system (e.g., a supply chain) will eventually fail, especially when combined with

31

dangerous technologies. An important element of Normal Accident Theory
(NAT) is that the focus is not on the choices of an individual decision maker or
the activities (or lack of activity) of a single operator, but is instead based upon
the structure and complexity of inteflinking processes and organizations. The
interactiveness of processes or entities and the degree to which they are tightly
coupled to each other are the two factors that make a system more prone to
experiencing negative events. Table 2.1 provides an overview of linear and
complex interactiveness and of tight and loose coupling within the framework of
NAT as shown in Figure 2.5. In general, the more complex the interactions,
defined as sequences of actions and their results that may be unfamiliar,
unplanned, unexpected, and are not easily visible or understood, the more
vulnerable the system. The tighter the coupling, defined as little slack or buffer

between entities or processes, the more vulnerable the system.

32

Table 2.1 Summary of Definitions in Interaction/Coupling Matrix

Linear System

Complex System

 

- Spatial segregation

- Dedicated connections

- Segregated subsystems

- Easy substitutions

- Few feedback loops

- Singly purpose, segregated controls
- Direct information

- Extensive understanding

Loose Coupling

- Proximity

- Common-mode connections

- Interconnected subsystems

- Limited substitutions

- Feedback loops

. Multiple and interacting controls
. Indirect information

. Limited understanding

Tight Coupling

 

- Processing delays possible

- Order of sequences can be clanged
- Alternative methods available

- Slack in resources possible

- Buffers and redundancies fortuitously

applied
- Substitutions fortuitously available

- Delays in processing not possible
~lnvafiantsequences
- Only one method to achieve goal

- Little slack possible in supplies,

equipment or personnel

- Buffers and redundancies are

designed in and deliberate

- Substations of supplies, equipment,

personnel are limited and designed in

 

Taken from Perrow, 1999, pages 88 and 96

Figure 2.5 Interaction/Coupling Matrix

 

 

 

 

 

 

S
k 1 2
O)
.E
'o.
3
O
o 3
o 4
tn
o
o
.1
Linear Complex
Interactions

(Perrow, 1999)

It is the dependencies within a complex system that increase vulnerability
and the virtual inevitability of a ‘normal accident.’ A normal accident may occur
quite rarely, the normality comes not from the frequency of occurrences, but from
the view that a failure is a ‘normal’ or expected part of the system’s existence.
Within the framework of NAT, an accident is defined as reasonably substantial
“damage to a defined system that disrupts the ongoing or future output of that
system” (Perrow, 1999, pg. 64). Perrow further goes on to define ‘incidents’ as
disruptions that impact a portion of the system, such as a part (the smallest
component of the system, eg. a valve) or a unit (a related collection of parts).
Thus the Specific term ‘accident’ is reserved for very serious events that impact
major subsystems or the entire system, although incidents or disruptions on a
smaller scale that may occur more frequently are also of concern; especially if
there is a cumulative effect.

A major element of NAT is the requirement of complexity. Choi and Hong
(2002) provide a concise definition where supply chain complexity is a function of
the variety or structural differentiation of the system. Complexity contributes to
vulnerability within a system in two ways. First, as the complexity of the system
increases, so does the possibility that there will be a substantial weakness built
into at least one part of the system. Second, as the system becomes more
complex with a greater number of linkages or larger geographic dispersion
between linked entities, the ambiguity within the system increases. Heightened
ambiguity means that individuals operating in the system are less able to

understand and be conscious and aware of events that are occurring in other

34

parts of the system, including the possible ramifications of those events. In the
most basic sense, complexity is a function of the structural variety and the
degree of dispersion between entities, the linkages of system elements, and is
impacted by the ‘tightness’ of those linkages (Chapman, 2005). The structural
variety refers to the number of subsystems, departments, or divisions within the
system (Daft, 1989). In a supply chain, this would be the number of suppliers,
customers and channels of distribution (horizontal complexity), the number of the
supply and demand tiers (vertical complexity), and the number of operating
locations and their geographic dispersion (spatial complexity) (Choi 8 Hong,
2002; Daft, 1989). Complexity related to the coupling of the linkages within a
system depends on how loose (high levels of uncertainty) or tight the
relationships are. This also relates to the ‘load’ or amount of effort required to
monitor or control the relationships (Dooley 8 Van de Ven, 1999). In a supply
chain, this would be demonstrated by the number of relationships along the
continuum of arms-length to full partnerships that require differing levels of
attention and resource commitment.

A supply chain is more than a part of a complex system, it is a
complex adaptive system (CAS) (Choi et al., 2001) and it is therefore undergoing
frequent changes to physical structure and infrastructural linkages, further adding
to its complexity. A CAS can be loosely described as a system that has a
coherent form, changes over time, and is not dependent upon the deliberate
actions of a single entity for management or control (Holland, 1995). Key

elements of a CAS include the following:

35

. The system must consist of entities or agents that have purposeful and
meaningful interactions based upon actions and events.

0 The emergent actions of individual entities are a reaction to the
simultaneous and parallel actions of other system entities.

o The actions of agents impact other agents due to the interconnectivities
of entities in the system.

a The entities often do not have total freedom of action or autonomy due
to negative feedback processes that regulate and control their actions
(e.g. contractual obligations).

o Dynamism exists as agents enter or leave the system, and as they
adapt and change, which then induces other entities to respond with
adaptation and change.

. There are also dynamic environmental influences that impact specific
entities and the system, but which exist external to the CAS (currency
fluctuations, regulatory requirements, etc.).

These interacting changes, reactions and adjustments do not follow a linear
pattern; large changes may lead to small impacts or small changes may lead to
large impacts. Finally, although it would be logical to assume that the outcomes
of a CAS would be chaotic and random, there are in fact patterns of behavior that
do allow for some form of control and influence (Choi et al., 2001; Sterman,
2001). These characteristics tie into the view that a supply chain is an organized
subset of firms within a much larger network of firms supplying and buying goods
and services that is also influenced by the environment in which the individual
firms and the supply chain operate. See Choi et al. (2001), pg 358, for a table
that clearly links the general elements of a CAS specifically to a supply network
(since a supply chain is a subset of a supply network, it is assumed that similar
associations apply).

Within the framework of NAT, the tighter the coupling, defined as time

dependency, path inflexibility and unavailability of slack resources (Perrow,

36

1999), the higher the interactive complexity of the system and therefore the more
susceptible the entire system is to a negative event that occurs in any part of the
system (Rice Jr. 8 Caniato, 2003). Serious failures can occur as a consequence
of the interaction of diverse system elements, such as the environment,
equipment, operators, processes and procedures, and supplies and materials
(Jermier, 2004), all of which are active components of a supply chain. Due to the
coupling of processes and entities, failure of the system is often the result of the
interaction of many small failures rather than a single, large scale failure (Perrow,
1999)

The coupling elements in NAT of time dependency and path inflexibility
are specifically applied to a supply chain by Svensson (2004) as time legs and
co-dependent process functionality and relationships. The lack of availability of
slack resources and of changing (reduced) levels of inventories (Sterman, 2001)
is manifestations of the current trends of lean and JIT practices. This creates a
susceptibility to disruptions that in essence allows a domino effect where an
event, even a small one, occurs in one part of the chain and potentially grows as
it travels through the chain, leading to a larger failure at some other point
(Roberts, Bea, 8 Bartles, 2001; Elsinger, Lehar, 8 Summer, 2006). The complex
and cascading linkages that exist may make the system more sensitive to events
and result in more disruptions or ‘catastrophes’ than might otherwise be intuited
(Choi et al.,_2001). This contagion of impact is especially true when there are
large scale events that are exogenous to the system (Elsinger et al., 2006). In

light of the trend over the past two decades for supply chains to become leaner

37

and more tightly coupled, complex and dispersed, supply chains may now exhibit
the unintended consequence of having become more fragile and more
susceptible to disruptive events that may ultimately have larger negative impacts

(Rice Jr. 8 Caniato, 2003; Sheffi, 2005).

2.3 High Reliability Theory

High Reliability Theory (HRT) is an optimistic counter to NAT in which
Roberts (1989) posits that in high reliability organizations (HRO) the occurrence
of ‘accidents’ is very rare and it is not implicitly clear that all systems are subject
to failure. This view was further expanded by Weick, Sutcliffe and Obstfeld
(1999) who put forth the basic tenet that with correct management organizations
can stave off ‘normal accidents’ indefinitely. In general, HROs operate under
difficult conditions but experience fewer accidents than might be expected, even
when faced with the possibility of a propensity of unexpected events due to
technologies, complexity and variable demand (Weick 8 Sutcliffe, 2001). The key
point is that these organizations place a greater emphasis on the importance of
system reliability rather than on efficiency (Weick 8 Roberts, 1993). The view Is
that if systems are better understood for more appropriate and effective
management, potential hazards can be detected earlier and reigned in before a
situation spirals out of control into a disaster (Chapman, 2005). It is important to
note that both NAT and HRT do not focus on individual behaviors and choices,
but on systems, organizational structure, and the linkages that bind them

together (Rosa, 2005).

38

Organizations themselves have become leaner, with fewer employees
doing more work and with their attention spread more thinly. The resulting
reduced attention that can be given to non-tactical and non-recurring events may
reduce the costs of routine transactions, but at the cost of increased damage that
may result from an unexpected event (Lampel 8 Shapira, 2001) which is not
detected early enough to prevent damage. Thus, there is a need to effectively
plan to stave off disruptive events in almost all organizations. The specific plans
used by business can be implemented to varying degrees at the discretion of
management based upon a firm’s business environment, strategy and goals.
Firms in some industries are critical to the general infrastructure of a community
or the country (e.g. hospitals, airports, utilities, financial services and public
works) and may operate under strict regulatory mandates that require them to
engage in effective business continuity planning (Rodetis, 1999; Bielski, 2003).
However, most firms are not in industries that are under the same externally
imposed requirements and are free to pursue varying levels of risk management
and planning.

The majority of research on HROs has focused on organizations that are
required to have exceptionally high levels of reliability, where even a small lapse
in performance can have catastrophic results, such as a US Navy aircraft carrier
(Rochlin, Porte, 8 Roberts, 1998), nuclear power plants and the United States
Federal Aviation Agency Regional Air Traffic Control Center (Klein, Bigley, 8
Roberts, 1995) or Pacific Gas and Electric Company (Roberts, 1990). There is

little research examining firms experiencing a multitude of low impact events over

39

time which may still result in a major cumulative effect (Lewis, 2003) or that are
not in such ‘mission critical’ environments as those operating in the industries
previously mentioned. It stands to reason that organizations not engaged in such
high risk activities or with such high need for near perfect continuity would still
have a need for high reliability in a competitive business environment. There is
evidence that firms with high risk products (e.g. chemicals) or with rapidly
changing products (e.g. high-tech) examine their supply networks most often
(Hillman, 2006), but this does not preclude firms from other industries or with
other products from the need to find ways to provide some protection from
disruptive events (Waller 8 Roberts, 2003).

While the study of HROs has been focused on single organizations or
other types of single entities, HRT is also applicable to systems that span across
organizational boundaries. Within the context of HRT, a supply Chain represents
a “large-scale system” composed of a network of humans utilizing technical
systems to complete tasks that support the goals and missions of multiple
entities (Grabowski 8 Roberts, 1997), as goods and services flow through the
system. This also means that many of the processes that are used in HROs to
mitigate risk for improved system reliability and business continuity are applicable
to a larger, multi-entity supply chain. Grabowski and Roberts (1997; 1999) argue
that most of the challenges within a large-scale system, such a supply chain, are
driven by a paradox of autonomy and interdependence. Each member within a
supply chain exists autonomously as a self-managed entity with unique

objectives, assets, processes, and organizational culture; however, each

40

individual supply chain member is simultaneously interdependent upon the
actions, behaviors and goods and services of other entities operating within the
supply chain. This paradox drives many of the issues that the supply chain
faces, and the solutions that are implemented.

One issue that must be addressed is the need for effective
communications. The challenge specifically in a supply chain is that the entities
are most often distributed both spatially and organizationally, making face-to-face
communications difficult. This puts an increased burden on understanding
cultural needs and mental models to clarify goals, responsibilities, relationships,
and to develop trust across organizations. It can also make it difficult to gather
relevant data from all pertinent sources (Grabowski 8 Roberts, 1997; Roberts et
al., 2001). Therefore, the entities within a supply chain must create methods for
effectively communicating operating status and be able to recognize when an
event or situation requires communication with other supply chain members.

A second issue is the impact of decisions that may generate unintended
consequences. A single firm or a dyad in a supply chain may make a change
that is perceived as beneficial, but that ends up impacting another member of the
network in an unanticipated and possibly negative way. A third issue to
overcome is that separate supply chain entities each have their own norms,
processes, policies, goals, and reward structures that are used to reinforce their
unique business strategy and culture; however, as members of a larger scale
system, it is necessary to work across organizational boundaries to form an

overall focus on system reliability beyond the specific needs of individual entities.

41

This would include the development of reward and incentive practices that span
across firm boundaries to encourage multiple supply chain members to take into
account the desirable benefits of reliability and the avoidance of the costs of
failure not just for their own firm, but for other firms in the supply chain (Roberts
et al., 2001). This requires oversight and the use of checks and balances for
feedback to ensure the system is running appropriately, especially if there are
long incubation periods for events to permeate into serious system disruptions.
Finally, the structure of the system must be designed to allow for flexibility to
adapt to a changing environment as well as having built in redundancy for
system slack that serves to create buffers between processes and entities.

Five traits have been identified by Weick and Sutcliffe (2001) as the
distinguishing characteristics of HROs. Collectively they call these five attributes
‘mindfulness’. This definition includes the purposeful decision to create an
organizational structure that allows for better visibility of unexpected events at an
earlier stage so that the expansion of incidents can be halted or managed for
reduced overall impact. It also Includes a concerted effort to continuously update
the understanding of the organization and its relationship with its environment for
ongoing interpretations of its contexts, issues and potential problems, and
possible solutions. The key advantage of mindfulness for HROs is the ability to
recognize events while they are still in the earliest stages, even if they are
providing only faint signals or signs of their existence, for improved prevention,

mitigation and management of disruptions.

42

The five traits of mindfulness identified by Wieck and Sutcliffe (2001) are
summarized as:

- Preoccupation with failure: HROS experience a lower than expected
number of disasters due to an obsession with failures of all sizes,
rather than a hubristic focus on success. This is accomplished
through the timely reporting of errors and near misses, followed by
an examination of root causes for continuous improvement.

- Reluctance to simplify interpretations: There is not an attempt to
simplify all information because HROS acknowledge the complex,
unstable, unknowable and unpredictable environment in which they
operate.

- Sensitivity to operations: There is a focus on the operational,
reporting and structural processes along with constant examination
to uncover points of vulnerability before an event occurs.

- Commitment to resilience: HROS learn from previous events,
where an examination of holistic information and sensitivity to
operations are combined with a commitment to resilience, so that
they can quickly recover from incidents. The key point is not that
HROS are completely event free and that errors never occur, but
that they are able to rapidly respond and recover, so that events
are not disabling or recurring.

- Deference to expertise: HROS cultivate diversity in order to acquire
greater breadth of knowledge; decision making is placed in the
hands of those that have the greatest expertise to make decisions
and manage processes.

HROS take a proactive approach to managing risk. They undertake
continuous network analysis and monitoring in order to identify changes in the
network and potential resulting changes in sources of risk. This allows for the
reduction of exposure to possible disruptive events while also taking advantage
of Changes that may create potential opportunities (Hillman, 2006). A key
element of network scanning is the identification of the early warning signals that
many events have, long before the situation reaches the stage of eminent

negative impact (Mitroff, 1988).

43

In general, HROS are acutely aware that they are open to the occurrence
of unfavorable events, and they also actively seek ways to control, reduce or
eliminate their sources of vulnerability. A focus on managing processes, or
‘mindfulness’, creates an environment in which individuals are aware of the
possible events that should be monitored, they know their role in the process on
a larger scale, and they understand the actions that should be taken when event
signals are detected. In short, they plan for events so that they are better
prepared to act when an incident occurs to ensure the highest possible level of

ongoing business functionality.

2.4 Synthesizing the Literature

Supply chain risk literature has focused primarily on supplier management
(Zsidisin, 2003; Zsidisin, Ellram, Carter, 8 Cavinato, 2004; Giunipero 8
Eltantawy, 2004; Hallikas 8 Virolainen, 2004; Hillman, 2006;) and on the demand
or the downstream side of risk (Fisher, Hammond, Oberrneyer, 8 Raman, 1994;
Lee et al., 1997; Sodhi, 2005). There is a much more limited stream that
examines internal disruptions (Castillo, 2004; Gill, 2006). An October 2003
examination of supply chain risk management literature found that since 1990,
74% of the articles focus on the upstream or supply side, 69% considered the
downstream or demand side, and only 10% examined internal elements (some
articles examined more than one direction of risk management) (Paulsson,

2004). There is however limited research that holistically examines risk and

44

continuity planning from the perspective of a single firm and that explicitly and
simultaneously considers upstream, internal, downstream, and environment
sources of possible disruptions. In addition, while It is accepted that a supply
chain is embedded within a larger and more complex network, there has been
little research examining the impact of complexity on performance within a supply
chain context (Milgate, 2001). Nor has there has been work that explicitly
examines the relationship between supply chain complexity and continuity
planning for disruptive events for the reduction of disruptive events and their
impact on firm performance. A summary of the supply chain risk management

literature is provided in Table 2.2

45

Table 2.2.

This study seeks to fill the current research gaps by specifically examining

the following areas:

1. Explicit examination of the supply chain continuity planning
processes that firms are engaged in.

2. Examination of three possible sources of motivation for supply
chain continuity planning: perceptions of vulnerability, perceptions
of supply chain complexity and the use of continuity planning as an
effective strategic practice.

3. Examination of the impact of supply Chain continuity planning on
firm performance and the reduction of supply chain disruptions

46

Table 2.2 Summary of Supply Chain Risk Literature

 

 

 

 

 

 

 

using single
case studies
from furniture,
auto, retail, 8
pre-mfg house
industries

 

 

of atomistic source of
disturbances that is more
frequently categorized as
quantitative than qualitative.
Reviews methods to reduce
vulnerability and variation

 

 

 

Supply
Author(s) Study Purpose Key Findings Chain Theory Focus
Direction
(Fisher, Case study of With high complexity Demand Risk
Hammond, Sport (product variety) and high response/
Obermeyer, Obermeyer demand uncertainty (long reduction
and Raman, and the use of lead time, fashion products) activities
1997) supply and it is key to find use ‘accurate
demand response’, faster lead time
matching and higher flexibility to
better meet demand
(Lee, Conceptual ‘Bullwhip' effect consists of Demand Risk
Padmanabhan, discussion of four sources: demand signal response!
Iand Whang, demand processing, rationing game, reduction
1997) uncertainty order batching, and price activities
variations. Gives possible
mitigation strategies
(Smeltzer and Examines Linkage of purchasing Supply Transaction Risk
Siferd, 1998) proactive administrative actions with cost response/
purchasing as their possible risk and economics reduction
risk mgt within consequences. activities
the context of Identification of the
TCE and conditions necessary for the
Resource purchasing function to
Dependency manage risk, which may
model. Case involve activities beyond
studies of 24 their normal scope
firms
(Zsidisin, Examines risk Risk assessment is Supply Risk
Panelli and assessment, relatively common, either assessment
Upton, 2000) contingency formally or informally. The
plans and creation of contingency
buffering as plans is not often used.
utilized by Many firms reported little
purchasing corporate investment of
professionals, time or resources. Buffering
Case studies is a common mitigation
of nine firms strategy
(Svensson, Conceptual Disturbances to supply Supply Channel Sources
2000) examination of chain have source, category Theory and
the construct and time dimensions. First assessment
of supply chain of Svensson’s vulnerability of supply
vulnerability, frameworks. Very high level chain risk

 

47

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

Data from the
Swedish auto
industry.

 

 

disturbances),
consequence, and trend.

 

 

 

Supply
Author(s) Study Purpose Key Findings Chain Theory Focus
Direction

(Warren and Conceptual Provides examples of Supply Risk

Hutchinson, examination of potential attacks on _ responsel

2000) security for electronic systems and chain reduction
e-commence applicable counter measures activities

(Johnson, Post hoc Lessons from managing risk Supply Risk

2001) ‘case’ of toy in toy industry; presents chain responsel
industry techniques to manage reduction

demand and supply: activities

product extensions; rolling

mix; licensing; outsourcing;

producing in multiple

countries

Sheffi, 2001) Conceptual Advocates the use of Supply Risk

examination of operational redundancies chain responsel
security and and modified organizational reduction
terrorism structure, specifically the activities
issues in the use of public-private
supply chain partnerships

(Hallikas, Examines risk Illustrates the use of internal Supply Risk

Virolainen, and assessment audit and computer aided assessmen

Tuominen, process. Case cause and effect analysis to t

2002) study of two analysis risk in a buyer
firms supplier dyad

(Chapman, Conceptual Supply chain risk mgt Supply Stages of

Christopher, examination of consists of the fours steps chain continuity

Juttner, Peck, the process of planning

and Wilding, steps of risk 'risk identification

2002) mgt ' risk assessment

'continuity mgt, coordination
'learning from experiences.

(Hicks, 2002) Examines Stock prices fell an average Supply Financial
stock price of 8% immediately after impact of
following an announcement of a product disruptions
announcement or shipment delay, and had
of supply chain a long term (6 month)
delay impact of up to 20%.

(Svensson, Conceptual Operationalization of a Supply Sources

20023) examination of construct of vulnerability and
vulnerability of based upon service level assessmen
in/outbound (absence of disturbances), t of supply
logistics flows. deviation (presence of chain risk

 

48

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

Supply
Author(s) Study Purpose Key Findings Chain Theory Focus
Direction
(Svensson, Conceptual Operationalization of the Supply Sources
2002b) examination of construct of supply chain and and
supply chain vulnerability based on time, Demand assessmen
vulnerability in technology, knowledge, t of supply
terms of time 8 social, economic/judicial, chain risk
relationship market, 8 IT dependencies.
dependencies. Created typology scenario
Based on data matrix and identified key up
from the and down stream
Swedish dependency relationships
automobile
industry.
(Lewis, 2003) Examines risk Creation of a model of risk Supply Sources
from an input based upon the classic and and
—» output operations of Input —> Demand assessmen
model transformation —» output tof supply
perspective, model, the dimensions of chain risk
based on case causal event and negative
studies of four consequence. Lack of
firms. visibility hampers risk mgt;
layers of loss impact
operations, customer, and
direct and generic
stakeholders.
(Rice and Conceptual Provides prescriptive basic Supply Risk
Caniato, 2003) discussion of and advanced responses to And responsel
actions firms increase physical, info and Logistics reduction
should take to freight security through the activities
increase use of flexibility and
resiliency. redundancy. IDs possible
Based upon actions based upon specific
anti-terrorism failure modes of production
and security supply, HR, transportation
research. and communications
(Russell and Conceptual Lists five tenets of: Supply Risk
Saldanha, discussion of partnering with freight and responsel
2003) logistics moving firms, overseas Logistics reduction
practices of partner knowledge, access activities
security aware to mode-shifting capabilities,
logistics using multiple
communication channels,
adoption of military concepts
of agility, reservists and pre-
l positioning

 

49

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

SUPP'Y
Author(s) Study Purpose Key Findings Chain Theory Focus
DIrection

(Zsidisin, 2003) Identified the Literature summary of Supply Sources
characteristics factors affecting supply risk and
of inbound perceptions and supply risk assessmen
supply, characteristics. t of supply
implications of Identification of significant chain risk
perceptions of market and supplier
risk; classified characteristics for supply
risks. Six case management managers.
studies.

(Zsidisin and Use of Agency Examines the role of the Supply Agency Risk

Ellram, 2003) Theory purchasing function in Theory responsel
framework to controlling risk from the reduction
manage supply side through buffering activities
supply risk and behavioral-based

management techniques.
Summary of literature on
supply risk research.

(Castillo, 2004) Case study of Uses Boeing’s model, details Internal Sources,
business some issues to be examined assessmen
continuity in continuity planning, t and
planning at including infrastructure and reduction of
Boeing personnel issues risk

(Cavinato, Conceptual Supply chain consists of five Internal Risk

2004) examination of internal chain/network assessmen
logistics risk, constructs: physical, t
and call for financial, informational,
broadening relational, and innovational.
definition of Four categorizations of
risk relevant product/supply

costs are presented as are
four types of supply risks

(Chopra and Conceptual Identification of the drivers Supply Sources

Sodhi, 2004) examination of of supply chain risk, the chain and
importance of implications for various assessmen
proactively mitigations strategies and t of supply
managing SC the risk/reward trade-off chain risk
risk curve for managing risk.

(Sinha, Conceptual Presents a generic Supply Risk

Whitman, and examination of prescriptive methodology for chain responsel

Malzahn, 2004) risk with mitigating risks in an reduction
consideration aerospace supply chain and activities
of multiple minimizes conflicting
supply chains objectives using integrated
simultaneously definition (IDEFO) modeling

 

50

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

Supply
Author(s) Study Purpose Key Findings Chain Theory Focus
Direction
(Finch, 2004) Examines Increased levels of risk are Supplier Sources
relationship associated with increased and and
between firm inter-organizational Customer assessmen
size and level networking, especially with t of supply
of risk. Case small to medium sized chain risk
studies and enterprises.
review of ‘grey’
literature.
(Giunipero and Conceptual Investments in risk mgt are Internal Investment
Eltantawy, examination of determined by level of levels to
2004) the level of product technology, need reduce risk
investment in for security, criticality of
risk mgt that supplier, and level of
firms should experience with the material
undertake. and/or supplier
(uncertainty).
(Hallikas, Conceptual When the dependency Supply Risk
Karvonen, examination of between companies responsel
Pulkkinen, risk In increases, they become reduction
Virolainen, supplier more exposed to the risks of activities
and Tuominen, networks other companies.
2004)
Hallikas and Conceptual Examines use of supplier Supply Transaction Risk
irolainen, examination of selection and relationship cost responsel
2004) network and management as methods to economics reduction
risk manage/reduce risk. activities
management
(Spekman and Conceptual Identifies six dimensions of Supply Sources
Davis, 2004) discussion of supply chain risk as the flow chain and
supply chain of goods, information, funds, assessmen
management firm reputation, relationships, tof supply
risk. and security of information chain risk
systems, Strong tie to
partner selection.
(Svensson, Examination of Dependencies categorized Supplier Sources
2004) horizontal and into three areas of time, and and
vertical functional 8 relational. Each Customer assessmen
corporate has internal or external tof supply
vulnerability. focus, and is either atomistic chain risk
Based up data or holistic. ID causes of up 8
from the down stream vulnerabilities.
Swedish Degree of tier level visibility
automobile firms have is key component
industry. of vulnerability perceptions.

 

51

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Supply
Author(s) Study Purpose Key Findings Chain Theory Focus
Direction
(Zsidisin, Examination of Firms use a multitude of Supply Agency Sources
Ellram, Carter the process of techniques to asses supply Theory and
and Cavinato, assessing risk. Firms focus on supplier assessment
2004) supply risk. quality issues, supplier of supply
Case studies performance and preventing chain risk
of seven firms. supply interruptions.
(Hellstrbm, Centralized 8 Creation of a normative Internal Contingenc Risk
2005) decentralized model for employee y Theory ma. "ageing”
risk mgt based involvement in risk reduction t decision
upon a consisting of five styles making
contingency (autonomous, modified
model of autonomous, segmented,
leadership unitary 8 group decision)
(Hendricks 8 Event study After supply chain ‘glitch' Internal Financial
Singhal, 2005a) methodology there is 6.92% reduction in impact of
to examination sales growth, 10.66% cost disruptions
of firm stock increase, and 13.88%
prices after a inventory level increase.
supply chain Consistent across industries,
problem disruption sources.
(Hendricks 8 Event study Generally worse Internal Financial
Singhal, 2005b) methodology performance for firms before impact of
to examination the announcement of supply disruptions
of firm stock chain problems compared to
prices after a benchmark portfolios. After
supply chain announcement, increase in
problem firm's equity risk of 13.5%.
(Kleindorfer Development Creation of ‘SAM-SAC’ Supply Sources
and Saad, of a framework: firms should chain and
2005) conceptual engage in establishing assessmen
framework security stds, classifying t of supply
based upon assets 8 processes, rank 8 chain risk
joint risk prioritize assets, and iterate
assessment for CI to reduce frequency of
and mitigation. disruptions and increase
Utilizes data absorptive capacity.
from US. Necessary implementation
Chemical conditions: environmental ‘fit'
Industry. and effective relationships
(Peck, 2005) Sources of risk Identified four levels of SC Supply Systems Sources
affecting the risk (product 8 process, chain Theory and
supply chain, infrastructure, environment, assessmen
using a interorganizational, to t of supply
systems theory explain risk scope and chain risk
approach dynamism

 

52

 

 

Table 2.2 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

Author(s) Study Purpose Key Findings Scuhpapl‘y Theory Focus
Direction

(Sheffi and Long term, Graphic disruption profile Supply Sources

Rice, 2005) comprehensive and the stages of prep, . and
research into event, first response, initial chain assessmen
disruption impact, full impact, recovery I of supply
management. preparation, recovery, and chain risk
Focus on Iong-tem impact. Illustrates
conceptual vulnerability framework
identification of along axis of dismption
common traits probability 8 disruption
of high/under consequences to identify a
performing vulnerability profile. Input
firms —» transformation —» output.

(Sodhi, 2005) Use of Creation of a model to allow Demand Risk
modeling to company to reallocate responsel
manage capacity among different reduction
demand risk products and thus manage activities

demand/inventory risk

(Zsidisin and Conceptual Early supplier involvement is Supply Agency Risk

Smith, 2005) examination of proposed to reduce risk in Theory responsel
the role of new product development by reduction
early supplier 'reducing uncertainty activities
development 'improving info flow
in managing 'creating goal congruency
supply risk. 'limiting adverse selection
Uses a case 'allowing buying firm to
study of a better monitor supplier firm
single firm. activities

(Zsidisin, Examination of High degree of convergence Supply Institution Business

Melnyk and how 8 why of practices around al Theory continuity

Ragatz, 2005a) firms create principles of awareness, plans
business prevention, remediation, and and Open
continuity knowledge management. Systems
plans, using Propose that business Theory
case studies of continuity planning will grow
three firms in prevalence, firms will
selected as benchmark high performers
having high and it will evolve into a
continuity standard purchasing
planning. practice

(Zsidisin, Examination of Increased risk driven by Supply Sources

Ragatz, and the impact of shorter product life-cycles, and

Melnyk, 2005b) supply chain ‘leaning’ of the supply chain, assessment
management high customer expectations, of supply
trends on risk. more info linkages facilitated chain risk

by IT. SC continuity
lplanning framework.

 

53

 

 

Table 2.2 (cont’d)

 

Supply

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Author(s) Study Purpose Key Findings Chain Theory Focus
Direction

(Tomlin, 2006) Single product, Supplier's percentage Supply Risk
two suppliers uptime and the nature of the responsel
model to disruptions (frequent but reduction
examine the short versus rare but long) activities
impact of are key determinants of the
disruptions optimal sourcing strategy

(Gill, 2006) Conceptual Decentralized workplaces Firm Risk
discussion of can be advantageous for responsel
decentralized business continuity reduction
workplaces activities

(Hillman, 2006) Conceptual Results from larger Supply Sources
discussion of research study that and
areas 8 identifies the area of supply assessment
sources of risk chain risk that are foremost of supply
in supply base in the minds of practitioners chain risk

(Kiser and Conceptual Identifies external 8 internal Supply Risk

Cantrell, 2006) discussion of risks. Provides six steps for responsel
process of managing risk: profile supply reduction
understanding base, assess vulnerability. activities
and controlling evaluate implications, id
supply risk mitigation actions, perform

cost/benefit analysis 8
implement plans

(Swink and Survey of firms Increasing focused Supply Transaction Sources

Zsidisin, 2006) to examine commitment provides cost and
risks/benefits benefits when existing economics assessment
of focused commitment is low. Over Agency of supply
commitment commitment can lead to risk 8 chain risk
strategy with that off-sets the transactional Knowledge
suppliers. and scale-related benefits. based view

(Craighead. Conceptual Calls into question practices Supply Sources

BIaCkflUlSl. .. examination of of supply base reduction, and

.. " ' g of SC global sourcing and use of assessment

3337+)Iandfield, density, node supply clusters of supply
criticality and chain risk 8
complexity with Risk
mitigation responsel
capabilities of reduction
recovery and activities
warning

 

 

3. MODEL DEVELOPMENT

Research is an iterative process grounded in theory and existing
knowledge that extends our understanding though the use of data to test and
modify existing theories and to create new theories (Handfield 8 Melnyk, 1998;
Meredith, 1993). The proposed conceptual model, as shown in Figure 3.1,
served as the starting point of the iterative process undertaken in this study. It
represents a recursive, cross-sectional model using reflective constructs. The
original model was tentative and It was expected to undergo revision based upon
findings from the data. A discussion of the literature reviewed and applied to
create the specific constructs in the proposed model follows.

Figure 3.1 Proposed Model

Supply Chain
Complexity

(CPLX)

Supply Chain
Dependency

 

 

 

   
 
    
       
 

  

Firm
Performance
(FMPF)

 
 

Disruptive
Events
(DE)

 

 

 

 

  
 
     
     

Effective
Practices
(EFPR)

   

Concern for
Disruptions
CNRN)

55

3.1 Construct Development

Good empirical research requires good measurement as its foundation,
where multi-item constructs are developed with a sound conceptual and strong
theoretical base. A two-step normative process for construct development, as
outlined in Stratman and Roth (2002) serves as the basis for the construction of
construct scales in this research. The exploratory nature of this study and the
absence of scales which have been tested in other studies makes it especially
important that a rigorous and scientific process of construct development and
scale creation occurs as a necessary predecessor to empirical research of this
topic (Sethi 8 King, 1991).

The constructs (or latent variables) consist of a set of reflective scale
items (or manifest variables) meant to capture a specific concept represented by
the construct. The manifest variables (MV) are captured through an extensive
literature search and from practitioner input as the representative scale items
meant to capture the latent variable (LV) construct to be tested. Since structural
equation modeling (SEM) will be used to test the model, at least three MVs are
needed for each LV to ensure the model has statistical identification (Bentler 8
Chou, 1987; Marsh, Balla, 8 McDonald, 1988); there are no single indicator
constructs which may hamper measurement reliability (Bentler 8 Chou, 1987)
and model identification. A larger ratio of MVs to LVs will also aid in
compensating for a potentially small sample (Marsh et al., 1988). However, an
overly large number of MVs inhibits parsimony (Gerbing 8 Anderson, 1984), so a

balance has to be achieved with a sufficient number of MVs to capture the LV

56

without the ratio becoming unnecessarily large. The constructs developed for
use in this study are grounded in previous theoretical and empirical studies. The
literature used for the creation of each LV and the associated measurement MVs

are detailed in the following sections.

3.1.1 Supply Chain Complexity

Complexity can be loosely defined as the variety and uncertainty
contained within a system (Flynn 8 Flynn, 1999; Frizelle, 1998). The complexity
of the system is a key element of its susceptibility to disruptive events (Choi et
al., 2001). Operationally, the complexity of a supply chain Is a function of the
individual attributes of uncertainty, technological intricacy, and system structure
and their influence on each other.

Specifically, research has been done on the uncertainly of demand due to
information distortion as data flows upstream through supply chain entities, often
referred to as the ‘bullwhip effect’ (Davis, 1993; Fisher et al., 1994; Forrester,
1961; Lee 8 Billington, 1995; Lee et al., 1997). There is also general uncertainty
of expected demand that is inherent in the process of forecasting (Chase,
Jacobs, 8 Aquilano, 2004; Sheffi, 2005). There can be uncertainty about the
supply of material into the firm due to long physical paths along which goods
must travel that increases lead times with greater variation in delivery times
(Hallikas 8 Virolainen, 2004).

Technological intricacy of the products and processes used can drive
greater complexity within a supply chain. The complexity of the technology and

processes used to make a product contribute to difficulty in gathering relevant

57

information and in controlling the processes (Flynn 8 Flynn, 1999; Khurana,
1999; Tatikonda 8 Montoya-Weiss, 2001). The complexity of the products
relates not only to the processes used in manufacturing, but to the depth of the
bill of material used to make the product. A large bill of material can serve as a
proxy measure of both the complexity of the products the firm provides and the
number of possible suppliers required to supply many different types of materials
(Choi 8 Hong, 2002).

Finally, the number (fonrvard, backward and within-tier) and dispersion of
players in the system (Daft, 1995; Beamon, 1999; Choi 8 Hong, 2002;
Craighead, Blackhurst, Rungtusanatham, 8 Handfield, 2007) influences the
structural complexity of the supply chain. The greater the number and the more
geographically dispersed supply chain entities are, the more difficult it is to know
the status of each supply chain node and to manage the flow of information,
goods and services to ensure supply chain continuity. The scale items and

literature source for the construct are summarized in Table 3.1.

Table 3.1 Summary of Supply Chain Complexity Scale Items

 

Scale Items Source

Products and processes (Khurana, 1999; Randall and Ulrich, 2001; Tatikonda
and Montoya-Weiss, 2001; Choi and Hong, 2002)

Market (Zsidisin, 2003; Chen and Paulraj, 2004; Hallikas and
Virolainen, 2004)

Supply chain structure (Daft, 1995; Beamon, 1999; Flynn and Flynn, 1999;
Choi and Hong, 2002; Craighead et al., 2007)

Uncertainty (Wilding, 1998; Flynn and Flynn, 1999)

 

58

3.1.2 Supply Chain Dependency

Vulnerability is defined by the Merriam-Webster Dictionary as “capable of
being. . .wounded or open to attack or damage”. A supply chain is vulnerable
when it is possible that events could occur which would be detrimental to the
firm, including potential disruption in normal business activities, an interruption in
the flow of goods and services, a negative financial impact of increased costs or
reduced revenue, or harm to the firm’s reputation. Vulnerability exists when
disturbance in the flow of goods and services occurs and the impact can not be
completely absorbed (Svensson, 2002a).

Dependencies within the supply chain can also result in ‘parallel
interactions’ that occur within the system between different supply chain entities
within the same supply chain tier. There is a ripple effect where the poor
performance of one supply chain member impacts another seemingly unrelated
member. For example the late delivery of material from one supplier may result
in a different material originally requested from another supplier to not be needed
or consumed as originally scheduled (Wilding, 1998). Or the late material may
need to be expedited from an alternative source.

Sources of disturbances in the flow of goods can be generally grouped as
atomistic or holistic. An atomistic disturbance is one that has a direct impact on
the firm driven by the performance of a first-tier partner or the logistics link
between the firm and a fist-tier partner. A holistic disturbance has an impact
beyond the first-tier relationship as a disturbance ripples through the supply

chain. Additionally, disturbances can be categorized as quantitative or

59

qualitative. A quantitative source includes a lack of needed volume or availability
of goods, possibly resulting in a stock-out. A qualitative source includes incorrect
or poor quality materials (Svensson, 2000). Due to the obscurity of events that
occur upstream and downstream within the supply chain beyond the first tier,
most firms are only explicitly aware of events that transpire atomistically
(Svensson, 2004). Therefore, firms may have difficulty in gauging their
vulnerability beyond their internal operations and their dependency upon first-tier
upstream and downstream partners.

In an effort to Operationalize the construct of supply chain vulnerability due
to supply chain linkages, Svensson (2002b) built upon earlier work to identify the
following dependency dimensions:

- Technical: adaptation to align business practices and the use of
compatible equipment and technology

- Time: use of time-based need or the synchronization of business
activities

- Knowledge: mutual sharing and learning from each other to create
knowledge of problem solving

- Social: interaction between firms is based upon personal relationships
and the strength of the social atmosphere and personal chemistry of
execufives

. Economicfludicial: formal, often written, agreements between firms

- Market: image and status of a firm which may impact the image and
status of other firms, including general goodwill

- Information Technology. the investment in and use of compatible
hardware and software for communication, including common IT
standard protocol

Ambiguity can also contribute to perceptions of vulnerability. Ambiguity is
exacerbated by uncertainty about possible changes in demand patterns and

supply availability, and the non-linear and parallel interactions that occur as the

60

impact of events ripple through the supply chain (Wilding, 1998; Peck, 2005). It
is therefore appropriate to include both ambiguity and the multiple dimensions of
dependence in a construct to measure perceived supply chain vulnerability. The

scale item topic and literature source are summarized in Table 3.2.

Table 3.2 Summary of Perceived Supply Chain Dependency Scale Items

 

 

Scale Items Source

Alignment of business activities (Svensson, 2002b)

Time-based synchronization (Svensson, 2002b)

Knowledge exchange (Svensson, 2002b)

Close personal relationship (Svensson, 2002b)

Economically dependent (Svensson, 2002b)

Reputation influence (Svensson, 2002b)

Adaptation of IT investments (Svensson, 2002b)

Knowledge of first tier partner activities gigfisgggg) Ray, 1997; Svensson, 2004;
Knowledge of partner activities beyond (Svensson, 2004)

the first tier

Awareness of partner strengths and (Ghosh and Ray, 1997; Peck, 2005)
weaknesses

 

3.1.3 Effective Practices

Firm performance can not be attributable to planning for business continuity
alone; therefore a construct to capture the use of business practices that have
been linked to improved business performance must also be included. There
have been numerous studies attempting to identify the effective or ‘best’
operation management and supply chain practices that lead to higher business
performance (Vickery, Drbge, 8 Markland, 1993; Flynn, Schroeder, 8

Sakakibara, 1995; Sakakibara, Flynn, Schroeder, 8 Morris, 1997; D'Avanzio, von

61

Lewinski, 8 Van Wassenhove, 2003). It has further been found that many ‘best
practices’ are dependent upon the environment, firm goals, and interaction
effects between those practices and supply chain members (Ketokivi 8
Schroeder, 2004a). Firms must engage in effective practices that improve
internal processes as well as crossing organizational boundaries to engage other
supply chain firms in the improvement of external and shared processes (Stank,
Keller, 8 Closs, 2001; Spekman 8 Davis, 2004; Filbeck et al., 2005). There is a
plethora of research examining specific practices in operations and the supply
chain that are related to firm performance (Davies 8 Kochhar, 2002). A
complete review of effective practices is beyond the scope of this research, so
those practices which have been identified in studies as representing effective or
‘best’ practices, with consideration of the timeliness of the study, are utilized
here (e.g. TQM has become so ubiquitous, it is a standard rather than ‘best’
practice) The effective practices used in this study consist of both internally and
externally focused methods that have shown to have a positive relationship with
performance (Stank et al., 2001; Davies 8 Kochhar, 2000; Wisner, 2003; Griffis,
Cooper, Goldsby, 8 Closs, 2004; Ketokivi 8 Schroeder, 2004a; Spekman 8
Davis, 2004; Laugen, Acur, Boer, 8 Frick, 2005; Li, Rao, Ragu-Nathan, 8 Ragu-
Nathan, 2005). The scale item topic and literature source are summarized in

Table 3.3.

62

Table 3.3 Summary of Effective Practices Scale Items

 

Scale Items
New Product Development
Productivity and waste elimination

Outsourcing non-core competencies
Cross-functional orientation
Acquisition of proprietary assets
Strategic partnerships

Information sharing

Supplier certification

Customer relationship management

Source

(Salvador, Forza, Rungtusanatham, and
Choi, 2001; Ketokivi and Schroeder, 2004a;
Laugen et al., 2005)

(Ketokivi and Schroeder, 2004a;
Laugen et al., 2005)

(Laugen et al., 2005)
(Ketokivi and Schroeder, 2004b)
(Ketokivi and Schroeder, 2004a)

(Salvador et al., 2001; Ketokivi and
Schroeder, 2004a; Laugen et al., 2005;
Li et al., 2005)

(Davies and Kochhar, 2000; Wisner, 2003;
Li et al., 2005)

(Davies and Kochhar, 2000;
Salvador et al., 2001)

(Wisner, 2003; Li et al., 2005)

 

3.1.4 Concern for Disruptions

There is ample evidence of the direct costs and of other direct negative

impacts on firm financial performance due to a supply chain disruption

(Hendricks 8 Singhal, 2005a, 2005b). There is also evidence that handling

disruptions well can lead to improved firm performance over the long run

(Norrman 8 Jansson, 2004; Sheffi, 2005). However, investing in supply chain

resiliency to reduce the likelihood and impact of disruptive events requires a

concern for supply chain disruptions. Because supply chain disruptions can

come for a variety of directions, firms should be concerned about disruptions

from upstream and internal sources (suppIY). customer sources (demand) and

business environment sources.

63

3.1.5 Continuity Planning

The stages of the disruption profile, as defined by Sheffi and Rice (2005)
include preparation, first response, management of the initial impact, preparation
for recovery, and the recovery process. The creation of a continuity plan should
include securing the people in the organization, core assets (systems, facilities,
processes, equipment, and infrastructure), and business relationships within the
supply chain (Foster 8 Dye, 2005). This research focuses on the planning that
firms undertake when evaluating sources of risk and creating plans that will
reduce the likelihood and impact of events that may interrupt normal business
functionality. The Disaster Recovery Institute defines business continuity and
business continuity management as follows:

business continuity. the ability of an organization to provide service

and support for its customers and to maintain its viability before,
during, and after a business continuity event

business continuity management: a holistic management process
that identifies potential impacts that threaten an organization
and provides a framework for building resilience with the
capability for an effective response that safeguards the interests
of its key stakeholders, reputation, brand and value creating
activities. The management of the overall program through
training, rehearsals, and reviews, to ensure the plan stays
current and up to date

Continuity planning is a multifaceted process that in its entirety consists of
six distinct elements. Therefore, the construct used here must also include the
specific elements of a comprehensive supply chain continuity plan. They are
summarized as (Chopra 8 Sodhi, 2004; Clark 8 Harman, 2004; Hallikas 8

Virolainen, 2004; Kiser 8 Cantrell, 2006; Sheffi 8 Rice Jr., 2005):

64

Profile the supply chain: examine the existing supply chain to
determine the number and location of suppliers, production and
warehouse facilities, transit lanes, customers and their
associated linkages

Risk identification: investigate possible sources of risk in the supply
chain that could impede the flow of goods and services, or
normal business operations

Risk assessment: determine the probability and impact of identified
potential disruptions

Risk management strategies: general strategies of acceptance,
management, reduction, pooling, sharing, transferring and
elimination of supply chain risk should be examined for fit with
the firm and the supply chain, along with cost/benefit analysis of
various strategies

Implementation of risk management plans: specific techniques and
practices selected as most appropriate are implemented based
upon the selected risk management strategies

Audit of risk management plans: plans should be examined for
effectiveness, including sharing plans with appropriate supply
chain partners.

The construct of planning for events is formed by the six specific elements
of risk management planning outlined above. Each has a unique scale meant to
assess the degree to which the firm is engaged in a particular element of overall
supply chain continuity planning, and will be discussed at greater length in the

following sections.

3.1.5.1 Profile the Supply Chain

The first stage is to profile the supply chain, or create a ‘map’ of the
existing system. This includes the identification of suppliers, production and
service points, warehouses, logistics lanes, and customers that are crucial for
ongoing business functionality in terms of their locations, linkages and volumes

(Gardner 8 Cooper, 2003). Understanding the locations and criticality of

65

suppliers, logistics providers, internal operations, and customers is the first step
of the creation of viable continuity plan (Barnes, 2001; Bazerman 8 Watkins,
2004; Elkins et al., 2005; Hillman, 2006; Sheffi, 2005). The scale item topic and

literature source are summarized in Table 3.4.

Table 3.4 Summary of Profile the Supply Chain Scale Items

 

Scale Items Source

Identify supply chain (Gardner and Cooper, 2003; Elkins et al., 2005;
partners and transit lanes Sheffi and Rice, 2005a; Hillman, 2006)

Create a supply chain ‘map’ (Gardner and Cooper, 2003; Hillman, 2006)

 

3.1.5.2 Risk Identification

Supply chain susceptibilities exist due to time, functional and relationship
dependencies (Kleindorfer 8 Van Wassenhove, 2004). It is therefore necessary
for firm’s to explicitly consider possible susceptibilities when selecting new
suppliers and customers as well as in the evaluation of existing business
relationships (Hillman, 2006). Although the uninterrupted flow of goods and
services is often the most obvious concern, there are qualitative issues that
should also be considered. The reputation or brand equity of the firm can be
impacted by the poor selection of a supply chain associate who may be engaged
in questionable or undesirable business practices (Finch, 2004; Spekman 8
Davis, 2004). The selection and evaluation of potential supply chain partners
should consider not only the physical sources of risk, but the behavioral as well.
An important part of evaluating the risk of a supplier or customer involves

understanding the sources of potential fragility and defining the types of events

66

that could transpire. Zsidisin (2003) provides a summary of event types from the
supply side, many of which the firm may also experience in its role as a supplier
to its own customers. The scale item topic and literature source are summarized

in Table 3.5.

Table 3.5 Summary of Risk Identification Scale Items

 

Scale Items Source

Risk as part of selection and (Finch, 2004; Spekman and Davis, 2004;
evaluation of supply chain partners Hillman, 2006)

Identification of partner vulnerabilities (Zsidisin, 2003; Finch, 2004)
Information from multiple sources (Zsidisin, 2003; Finch, 2004)

 

3.1.5.3 Risk Assessment

A common measurement of risk is based upon the probability or frequency
of an event and upon the magnitude of the impact should the event occur (Jia et
al., 1999; Pich et al., 2002; Weber 8 Milliman, 1997; Zsidisin et al., 2004). The
process of a comprehensive appraisal of risk exposure should identify potential
losses, including the likelihood and the quantitative (financial) and qualitative
(non-financial) impact that might be the result from a disruptive event, for a
comprehensive appraisal of risk exposure (Yates 8 Stone, 1992). The scale item

topic and literature source are summarized in Table 3.6.

Table 3.6 Summary of Risk Assessment Scale Items

 

 

Scale Items Source

Probability (Weber and Milliman, 1997; Pich et al., 2002;
Zsidisin et al., 2004)

Impact (Yates and Stone, 1992; Weber and Milliman, 1997;

Zsidisin et al., 2004)

 

67

3.1.5.4 Risk Management Strategies

There are a number of general strategies firm’s can use to try to cope with
supply chain risk, including accept, manage, share, transfer, reduce, pool or
eliminate it altogether (Allaire 8 Firsirotu, 1989; Chopra 8 Sodhi, 2004; Hillman,
2006; Johnson, 2001; Rudi, 2001; Sheffi, 2001; Sheffi, 2005). The strategies can
be briefly summarized as:

- Accept: Determine that there is little to be done to change or
manage the frequency or impact of events, and simply carry on
with business.

- Manage: Firms can manage demand variability with active price
adjustments, and collaborative planning, forecasting and
replenishment (CPFR); or supply variability with continuous
replenishment programs (CRP). Both demand and supply
management can be done via activities such as information
sharing and collaboration.

- Share: Risk sharing can take the form of agreements, often
explicitly contained in contracts that include buy-back
agreements, revenue sharing, automatic price adjustments, and
other such activities.

- Transfer: Require other supply chain partners to take on the risk,
such as supplier owned inventory, volume or price guarantees,
and penalties for lack of performance. It may also include
activities such as sub-contracting for additional capacity.

- Reduce: Working with supply chain partners to determine sources
of vulnerabilities and eliminating those that can be. Those that
can’t be require the identification of early detection mechanisms
so that action can be taken early to mitigate the impact of the
impending event.

- Pool: On the demand side this may include reduction of product
mix, the use of postponement, or use of aggregated and
centralized inventory management. On the supply side this may
include buyer consortiums or keiretsu type structures.

- Eliminate: The source of risk has been completely removed to have
a zero probability of occurrence or zero impact.

68

Because it is not feasible to make efforts to reduce or eliminate all possible
sources of risk, it is important for firms to undertake cost/benefit analysis to
determine which types of risk management efforts are most appropriate (Elkins et
al., 2005). The scale item topic and literature source are summarized in Table

3.7.

Table 3.7 Summary of Risk Management Strategies Scale Items

 

 

Scale Items Source

Accept (Sheffi, 2005)

Manage (Allaire and Firsirotu, 1989; Sheffi, 2005)

Share (Sheffi, 2005; Hillman, 2006)

Transfer (Allaire and Firsirotu, 1989; Sheffi, 2005; Hillman, 2006)

Reduce (Johnson, 2001; Chopra and Sodhi, 2004; Sheffi, 2005;
Hillman, 2006)

Pool (Johnson, 2001; Rudi, 2001; Chopra and Sodhi, 2004;
Hillman, 2006)

Eliminate (Sheffi, 2005; Hillman, 2006)

 

3.1.5.5 Risk Management Practices

The two primary techniques that firm’s can implement to better control risk
are redundancy and flexibility (Pich et al., 2002). Redundancy is a form of
buffering (Sheffi, 2005), and can take the form of utilizing multiple suppliers for
key materials or by holding safety stock of materials (Christopher 8 Lee, 2004;
Zsidisin 8 Ellram, 2003) and/or of finished goods in order to de—couple supply
chain tiers (Hillman, 2006). Redundancy is often used to maintain the capability

of normal business functionality, even if a portion of the process fails (Elkins et

69

al., 2005; Rochlin et al., 1998). Flexibility can take the form of having the ability
to adjust to changing conditions or the form of standardization (Chopra 8 Sodhi,
2004; Li et al., 2005; Rice Jr. 8 Caniato, 2003; Rochlin et al., 1998; Zsidisin 8
Ellram, 2003). This could consist of specific contractual relationships that include
the ability to adjust product mix, timing and volumes of material. It may also
include standardization of processes and facility layout so that multiple facilities
can meet firm needs, or through cross training of employees (Sheffi, 2005). The
implementation of continuity management plans must address both directions
from which disruptions can originate: disruptions in the availability of supply from
upstream and internal sources, and disruption in the desired demand for goods
from downstream sources (Johnson, 2001). The scale item topic and literature

source are summarized in Table 3.8.

Table 3.8 Summary of Risk Management Practices Scale Items

 

Scale Items Source

Holding safety stock (Zsidisin and Ellram, 2003; Chopra and Sodhi, 2004;
Christopher and Lee, 2004; Hillman, 2006)

Have multiple suppliers (Zsidisin and Ellram, 2003; Chopra and Sodhi, 2004;
Christopher and Lee, 2004; Elkins et al., 2005)

Have multiple facilities (Elkins et al., 2005; Sheffi, 2005)
Employee cross-training (Ketokivi and Schroeder, 2004a; Sheffi, 2005)

Standardization (Chopra and Sodhi, 2004; Sheffi, 2005)
Postponement (Chopra and Sodhi, 2004; Li et al., 2005)
Flexibility (Chopra and Sodhi, 2004; Sheffi, 2005)

 

7O

3.1.5.6 Audit of Risk Management Plans

Once plans are in place, the final step is to test the plans to ensure that
they function as expected and that everyone involved in the process has a clear
understanding of the activities that must be completed when an events begins to
manifest, and what steps must be taken to prevent further damage (Mitroff,
1988). Because a supply chain involves links between autonomous business
entities, it is important that the plan is shared with, and reviewed by, key partners
both upstream and downstream, as well as within the firm (Grabowski 8 Roberts,
1997). This includes understanding the signs that would trigger activation of the
risk response plan, “fire drills’ to test plans, and modification of plans as the
supply chain and environment changes (Elkins et al., 2005; Elsinger et al., 2006;
Hillman, 2006). Testing of the plan should take place before an event occurs to
ensure that critical information is included and appropriate responses and
activities are well understood (Rodetis, 1999). Because the network in which a
supply chain is imbedded is subject to nearly constant change, it is important to
reexamine the entire process, beginning with periodically updating the supply
chain profile and map. This is necessary as suppliers, customers, and
competitors enter or leave the market and as the market environment is altered
due to changing currency exchange rates, government regulation, consumer
preferences, technological changes, and material availability, firms must regularly
re-examine their supply chain to identify changes that may require adjustments to
previous plans (Bielski, 2003; Sheffi, 2005). The scale item topic and literature

source are summarized in Table 3.9.

71

Table 3.9 Summary of Audit Risk Management Plans Scale Items

 

 

Scale Items Source
Review plans with key players (Mitroff, 1988; Grabowski and Roberts, 1997)
Test plans (‘fire drills’) (Elkins et al., 2005; Elsinger et al., 2006; Hillman, 2006)
Periodic plan review (Bielski, 2003; Sheffi, 2005)

 

3.1.6 Disruptive Events

Events that could disrupt the supply chain can come from two generic
directions; from either outside or inside the supply chain. Events from inside the
supply chain can be further broken down into three directions, for a total of 4
directions from which a firm can experience a disruption. Events that are
generated from outside the supply Chain, can be called 1) ‘environment events’
(economic or technological changes, competitor actions, etc.); or events can be
generated by actions that occur within the supply chain from 2) upstream events
(late deliveries, lack of material or service availability, etc.); 3) internal events
(lack of adequate capacity, poor processes performance or output, etc.); and 4)
from downstream events (changes in demand mix, timing or volume requested)
(Chen 8 Paulraj, 2004; Davis, 1993; Sheffi, 2005; Sodhi, 2005; Zsidisin, 2003).

Environment events represent those incidents that originate from actions
or incidents outside of the supply chain in the larger business environment and
over which supply chain members may have little or no control (Christopher 8
Lee, 2004). These can include governmental regulation, competitor actions,

changes in the economic environment (Zsidisin, Ragatz, 8 Melnyk, 2005; Zsidisin

72

et al., 2005b), the number of qualified suppliers, technological changes (Chen 8
Paulraj, 2004), copyright or patent infringement, and counterfeiting

Upstream events are disruptions that impact the flow of goods and
services into the organization. They can be either an inability of suppliers to
meet the firm’s needs or due to a logistical problem where in-bound materials
may be lost, damaged or delayed (Chopra 8 Sodhi, 2004). These incidents may
be due to the actions of a firm’s direct supplier, or may be the result of an event
that occurs further upstream and that impacts the chain of supply that feeds the
firm (Sheffi, 2005). The inability of the supplier to meet the needs of the firm
could be caused by capacity issues related to volume, mix or timing of demand
or to technological changes (Zsidisin, 2003). Firms may be impacted by global
shortages of some material as demand increases from developing countries,
such as India or China (Hillman, 2006) or due to the time lag needed to build
plants to increase capacity (Sheffi, 2005). Although there is a tendency to
consider the inbound supply chain as consisting of materials, it does include
important services such as the continuous supply of utilities such as power and
water; an interruption in the flow of these services can also have a serious impact
on supply chain performance.

Disruptive events that are internal to the firm may occur when there is a
problem with output volume or quality levels that inhibits the ability to meet
demand (Chopra 8 Sodhi, 2004), due to situations such as a capacity shortage,
equipment or process failure, or poor quality performance. It is also possible for

a firm to experience a disruption in internal processes due to personnel events

73

such as a labor strike, widespread illness, or a number of employees leaving the
company simultaneously (Sheffi, 2005). It may be possible for the firm to resolve
the situation before the customer becomes aware of a problem. However, a
more serious case is when the customer also is impacted by the issue the firm is
experiencing.

Downstream, or demand side, events are generated by customers, such as
dramatic changes in requested product mix, volume or timing (Chopra 8 Sodhi,
2004) that the firm may have difficulty responding to. Customer needs are often
dynamic, requiring supply chains to adjust to meet changes in demand (Chen 8
Paulraj, 2004). Both large increases and decreases in volume, as well as
changes in timing or the mix of demand, can have a detrimental impact on the

firm. The scale item topic and literature source are summarized in Table 3.10.

Table 3.10 Summary of Disruptive Events Scale Items

 

Scale Items Source
Environment Events (Zsidisin, 2003; Chen and Paulraj, 2004; Zsidisin et al., 2005b)
Upstream Events (Zsidisin, 2003; Chopra and Sodhi, 2004; Sheffi, 2005)
Internal Events (Chopra and Sodhi, 2004; Sheffi, 2005)
Downstream Events (Chen and Paulraj, 2004; Chopra and Sodhi, 2004; Sheffi, 2005)

 

3.1.7 Firm Performance

Asking direct questions about firm performance may lead to high non-
response bias (Boyer 8 Verrna, 2000), especially when the queried individual
either does not know the answer or is not comfortable sharing potentially

sensitive information. Therefore, researchers often ask performance related

74

questions in an indirect way. These are termed quasi-perceptual measures,
where the content is exactly defined (e.g. market share or sales in units) but a
relative unit of measurement (e.g. a 1-7 ‘compared to industry performance’
scale) allows the informant to use their discretion in comparing their firm to their
competitors (DrOge, Jayaram, 8 Vickery, 2004; Ketokivi 8 Schroeder, 2004b;
Vickery et al., 1993) and does not involve revealing possibly proprietary
information. The performance measures used here are built upon quasi-
perceptual measures as used in other relevant studies (Anand 8 Ward, 2004;
Beamon, 1999; Boyer, 1999; Randall 8 Ulrich, 2001; Ward 8 Duray, 2000;
Wisner, 2003). While detailed performance measures specific to supply chain
performance, such as on-time delivery or fill rates, have been used quite often
(Griffis et al., 2004), the measures used here are meant to capture the general
performance of the firm. The assumption Is that the ability to effectively plan for
supply chain disruptions will have a positive impact on the firm. The scale item

topic and literature source are summarized in Table 3.11.

Table 3.11 Summary of Firm Performance Scale Items

 

 

Scale Items Source
Sales (Beamon, 1999; Anand and Ward, 2004)
Profitability (Boyer et al., 1997; Beamon, 1999)

Overall competitive position (Wisner, 2003)
Return on assets (ROA) (Randall and Ulrich, 2001)
Return on sales (ROS) (Randall and Ulrich, 2001)

 

75

3.1.8 Demographic Information

Respondents are asked general demographic questions about their
functional job responsibility and level with in the firm, the industry the company
operates in, and the size of the company in terms of revenue and number of
employees. These questions will be used to examine the response constituency
for possible group comparison. Demographic specifics are not part of the model
or the specific hypotheses to be tested, but may be useful for examination of

response by group, such as large and small firms.

3.2 Hypotheses

The following section discusses the hypotheses to be tested, based upon
the initial research questions, a review of the literature and the proposed model.
The influence of supply chain complexity and dependency on risk management
planning and the occurrence of disruptions are grounded in Normal Accident
Theory. The influence of the use of effective business practices and a concern
for supply chain disruptions on risk management planning, the occurrence of
disruptions and firm performance are grounded in High Reliability Theory. The
relationships between risk management planning, the occurrence of disruption,
and firm performance are based on the review of operations and supply chain

management literature.

76

3.2.1 Grounded in Normal Accident Theory

The complexity of the system and environment in which a supply Chain
operates may influence planning, as firms try to control their supply chain risk for
improved business continuity and firm performance. Practices such as
outsourcing and globalization have created supply chains that are more complex
and posses higher levels of interaction uncertainty, creating more complex and
therefore more risky systems (Sheffi 8 Rice Jr., 2005). This is especially true as
supply chains become longer and more dispersed with increased reliance on
other supply chain members (Bowersox et al., 2000; Cavinato, 2004; Zsidisin et
aL,2000)

Using Normal Accident Theory as a foundation, the greater the complexity
of the system, the greater the likelihood of a system disruption. As the number of
individual entities that make up a specific supply chain increases, the greater the
cumulative probability that an event will occur in any single part of the chain that
will impact other supply chain members. Yet firms have difficulty knowing the
activities that occur beyond their first-tier partners (Svensson, 2004). It is
therefore expected that when firms are participants in supply chains that are
more complex, there will be a greater need to engage in continuity planning to
better mitigate the occurrence of disruptive events on both a large scale (‘normal
accidents’) and on a smaller scale (Craighead et al., 2007).

Supply chain vulnerability is related to the multiple forms of dependency
that firms have with other supply chain members (Svensson, 2002a) and the

ambiguity of the system due to a lack of visibility (Peck, 2005; Svensson, 2004)

77

that may allow disruptions to business operations and the flow of goods and
services. Organizations that are more tightly linked to other members of the
supply chain network are more susceptible when a problem occurs with another
supply chain node (Hallikas et al., 2004). There is strong empirical evidence that
such disturbances can have substantial negative short and long term quantitative
and qualitative effects that firms want to avoid (Hendricks 8 Singhal, 2005a;
Lewis, 2003). In addition, there is evidence that handling disruptions well can
contribute to a reduction in negative results and that there are also positive
impacts that allow the firm to become more competitive after a disruption
(Chopra 8 Sodhi, 2004; Norrman 8 Jansson, 2004). The effective management
of supply chain disruptions represent a significant business challenge, and one
that senior management is increasing recognizing as being of strategic
importance (Kleindorfer 8 Van Wassenhove, 2004; Weick 8 Sutcliffe, 2001).
Therefore, it is expected that perceptions of dependency will be positively related

to higher levels of comprehensive planning for business continuity.

H1 a: Higher levels of perceived supply chain complexity are positively
associated with higher levels of continuity planning

H1 b: Higher levels of perceived supply chain dependency are positively
associated with higher levels of continuity planning

H2a: Higher levels of perceived supply chain complexity are positively
associated with greater frequency and impact of disruptive events

H2b: Higher levels of perceived supply chain dependency are positively
associated with greater frequency and impact of disruptive events

78

3.2.2 Grounded in High Reliability Theory

In a competitive business environment, firms must successfully implement
effective practices for improved firm performance, and there is ample empirical
evidence of this linkage. Using a cumulative capabilities view, as firms engage in
practices that improve firm performance, there are likely to be multiple positive
outcomes from the effective practices (Dostaler, 2001). It has also been argued
that business continuity planning is a subset of effective practices (Hendricks 8
Singhal, 2005b). Using High Reliability Theory as a foundation, firms that are
more ‘mindful’ of the need to improve reliability place a greater emphasis on the
proactive management of supply chain risk (Weick 8 Sutcliffe, 2001). As part of
mindful practices, there is ongoing environmental scanning and engagement in
continuous improvement efforts. In addition, there is some evidence that the
market responds favorably to implementation of supply chain management
practices that are considered to be effective or ‘best’ (Filbeck et al., 2005).
Therefore, the following hypotheses will be tested:

H3a: Higher levels of the use of effective practices is positively
associated with higher levels of continuity planning

H3b: Higher levels of concern for disruptions is positively associated
with higher levels of continuity planning

H4a: Higher levels of the use of effective practices is positively
associated with firm performance

H4b: Higher levels of concern for disruptions is positively associated
with firm performance

H5a: Higher levels of the use of effective practices is negatively
associated with disruptive events

H5b: Higher levels of concern for disruptions is negatively associated
with disruptive events

79

3.2.3 Impact of Risk Management Planning

There is evidence of the negative financial impact of supply chain
disruptions, as well as possible negative qualitative impacts (Chopra 8 Sodhi,
2004; Hendricks 8 Singhal, 2005a; Hendricks 8 Singhal, 2005b; Hicks, 2002;
Knight 8 Pretty, 1998). There can be increased costs from expediting, excess
material, idle capacity, and rework (Zsidisin et al., 2004). If it is accepted that
disruptive events contribute to poor firm performance, it is possible that the
avoidance of such disturbances would result in better firm performance. It can
therefore be expected that firms that are engaged in supply chain continuity
planning may experience a reduced level of frequency and impact of disruptive

events and exhibit higher firm performance (Craighead et al., 2007).
H6: Higher levels of continuity planning are positively associated with
higher levels of firm performance

H7: Higher levels of continuity planning are negatively associated with
disruptive events

H8: More frequent and larger magnitude dismptions are negatively
associated with firm performance

A summary of the hypotheses in the proposed model are shown in Figure 3.2.

80

Figure 3.2 Model with Hypotheses

Supply Chain
Complexity
Supply Chain
Dependency

.. ‘ eeeeeeeeee 08‘.

 

 

 

....... 4%
H1a,b + \ '6
a -- x Firm

Strategies

H8a,b -

Disruptive
Events

 

 

 

 

 

Concern for
Disruptions

 

3.3 Scale Development and Validation

The determination of constructs, the creation of their respective scale
items and their validation should be a rigorous, multi-step process. The selection
and development of constructs should be based upon the theory utilized and the
hypotheses to be tested (Flynn, Sakakibara, Schroeder, Bates, 8 Flynn, 1990) as

well as on existing literature (Menor 8 Roth, 2007). Statistical validation occurs

81

during the data analysis stage; however it is helpful to pre-examine the
constructs, the scales, and the survey instrument before data collection to
eliminate as many barriers to validation as possible. A three step process for the
examination of validity is outlined by O’Leary-Kelly and Vokurka (1998). First,
before data collection occurs, assessment of content validity is undertaken to
demonstrate that the scale items are linked to their parent constructs both
logically and theoretically. The second validity test involves examining the data
to determine construct validity or the empirical assessment of the indicators for
unidimensionality, reliability, and validity. Finally, the nomological validity is used
to determine the structural relationships between constructs. The process to be
used before data collection is discussed in the following sections, and the
process utilized after data collection will be discussed in the Data Analysis
chapter.

Although assessment of content validity is a single step in the overall
validity testing process, scale development for content validity is itself a three-
stage process. The three-stage scale development process is based upon
Churchill (1979) as utilized by Stratman and Roth (2002) who provide a
compressed overview (see Figure 3.3) and must occur before data collection.
Both of these works provide guidelines for rigorous and systematic construction
of scales for scientific investigation. During the first stage, multi—item scales
consisting of at least three items are used to asses the constructs. Multi-item
scales are necessary to adequately capture each of the constructs which

measure multifaceted business processes. Because supply chain research

82

occurs in organizations of different sizes and in different industries, it is possible
that there may be differences in the applied definitions of constructs or in
terminology need in scale development (Hensley, 1999). In addition, because
the scale items are based upon the judgment of the respondent, multiple
measures are needed to improve convergence (Perreault 8 Leigh, 1989).

However, it has also been shown that there are costs associated with
asking respondents to answer items that appear to be redundant or when the
construct appears to have too many questions; respondents tend to become
aggravated and may not read and respond to each question resulting in
responses that may not be accurate reflections of the survey respondent.
Therefore, researchers must carefully balance the need for multiple items to
appropriately measure complex concepts while also considering the small
incremental information obtained with each additional question when weighed
against the possibility of discouraging respondents and compromising
measurement reliability (Drolet 8 Morrison, 2001).

The specific constructs utilized in this study are based upon theory, a
review of the literature and input of industry expert practitioners (Menor 8 Roth,
2007). With this input, scale items were created to support the defined

constructs (Hensley, 1999).

83

Figure 3.3 Scale Development Process

 

Specify Domain of Constructs
- Literature review <-
- Interviews with experienced practitioners

J

Generate Item Pool

- Literature review

- Existing scales

- Suggestions from practitioners
V

Purify Measures

- Independent panels of expert judges

. Manual sorting instrument

. Measures of inter-judge agreement

 

 

 

 

 

 

 

 

 

 

 

 

 

Adapted from Churchill (1979) and Stratman and Roth (2002)

Initial pre-testing of the scale items requires that they be purified in order
to eliminate any obvious sources of measurement error and to achieve
substantive (‘face’) validity. An approach that is especially appropriate for new
scale development (Hensley, 1999) uses a manual sorting technique to pre-
establish tentative unidimensionality of the constructs and scale reliability and
validity (Hinkin, 1995; Li et al., 2005; Moore 8 Benbasat, 1991; Stratman 8 Roth,
2002). This in an iterative process where independent judges are given short
descriptions of the constructs and a randomized list of all of the scale items as
generated from the literature and modified from the previous sorting round. In
each round, the judges are asked to assign each of the scale items to what they
believe to be the appropriate construct. Modification of scale items occurs based
upon the success or failure of the assignment of individual items to the intended

parent construct. As part of this process, the judges will also be asked to

84

suggest additional items that may better capture the construct than those already
created, or to eliminate items that appear to be redundant or inappropriate.

It is expected that when appropriate scales are developed that proper
sorting of the items into the intended constructs would occur at a greater rate
than would occur due to chance. Pre-test scale reliability based upon the
judgments from the manual sorting of the instrument can be done utilizing
different methods. One method, as outlined by Moore and Benbasat (1991)
simply examines the number of correct scale item to construct placements. The
higher the percentage of correctly placed items, the higher the degree of inter-
judge agreement, although there is no established guideline of an ‘acceptable’
ratio of placements (Li et al., 2005). Perreault and Leigh (1989) advocate a
method that measures the observed proportion of agreement between judges.
An advantage of this method is that the number of defined categories is included
in the calculations, eliminating the conservative results found in earIier measures
(Rust 8 Cooil, 1994).

Another issue that can impact the validity of survey responses is the
problem of ‘yea-saying’, where respondents like to answer ‘yes’ or “agree” with all
responses. This issue can be reduced by rewording questions so that the
direction of some questions varies. Questions that are positive or negative in
nature should be interspersed in a random nature throughout the questionnaire.
Reversing some questions requires respondents to answer some questions with
a ‘yes’ or to ‘agree’, and to answer others with a ‘no’ or to ‘disagree’, helping to

reduce yea-saying bias (Rosenthal 8 Rosnow, 1991).

85

Yea-saying is different from social desirability. Social desirability is a
phenomenon where self-reporting respondents prefer to answer questions in a
way that allows them to present themselves in a favorable light. One way that
researchers attempt to circumvent this type of response is to offer the
respondents anonymity and to keep responses confidential. However, in
situations where respondents were not initially concerned with confidentiality,
explicitly mentioning it may sensitize them to this issue and raise their suspicions
regarding the survey and how it could be used (Rosenthal 8 Rosnow, 1991).
Therefore, in this research, there will be an explicit statement that explains how
the findings will be used and that individual responses will not be shared, with the
results reported only in aggregate.

Most of the scale items will require the respondents to answer based upon
a 7-point attitude scale. These items are primarily anchored by "strongly agree”
and “strongly disagree” with a neutral mid-point. There are also some questions
that ask respondents to estimate the frequency and impact of disruptive events
that have occurred in the previous 12 months. A list of the scale items and the
associated construct is contained in Appendix A. Although the concepts
embedded within the constructs and the resulting scale items are based upon
existing literature and theory, the format and wording of the questions as placed
within the questionnaire are based upon the guidelines for question writing and
survey construction as outlined by Dillman (2007) and elucidated in the following

Chapter.

86

4. RESEARCH METHODOLOGY

This chapter outlines the scientific research design and methodology used
in this study. Research that is conducted in a rigorous and scientific manner
provides a strong foundation for study replication and generalizability of findings.
The remainder of this chapter outlines the use of grounded theory methodology
and the research design employed. Description is provided of the organizational
sponsorship, the target sample, the process used to gather data from survey

respondents, and the data analysis techniques utilized.

4.1 Grounded Theory Method

The systematic approach for empirical research in operations
management as detailed by of Flynn, Sakakibara, Schroeder, Bates and Flynn
(1990) and Handfield and Melnyk (1998) serves as the basis for the process
followed here. These studies represent key works that provide thorough and
methodical guidelines for conducting scientific research in supply chain
management. The prescribed steps include:

1) Establishment of a theoretical foundation

2) Selection of an appropriate research design in keeping with the
research questions

3) Use of an appropriate data collection method with careful consideration
of the advantages and drawbacks of possible Choices

4) Implementation including sample selection, instrument design and
testing, and data gathering

5) Data analysis using predetermined methodologies

6) Summarization of the results for eventual publication for knowledge
sharing and building.

87

This study utilizes a grounded theory approach (Glaser 8 Strauss, 1967),
which is the general view that research begins with a strong theoretical
foundation to examine data gathered from the field for the refinement of
theoretical views based on rigorous analysis of the data leading to the support,
enhancement or modification of existing theory. This approach is well suited to
situations that rely on a social system where human behavior is a major
component of the business environment under study, and thus of the research
questions. It is also appropriate when qualitative research techniques are
employed.

The grounded theory approach is utilized here for three reasons. First, the
constructs in the proposed model have limited, or no, specific usage in previously
published research, have not previously been well defined, and do not have
generally accepted measurement constructs. Second, given the lack of prior
examination of the constructs, there is a corresponding lack of knowledge of the
relationships among the constructs, as well as a lack of specific scale items to
capture the constructs. Third, there is a need for empirical data gathering to
begin to understand this area in order to formulate observation into higher-level
generalizations and theory development (Meredith, 1998; Wacker, 1998; Forza,

2002;; Rungtusanatham, Choi, Hollingworth, Wu, 8 Forza, 2003).

4.2 Research Design

Following the grounded theory method, the first step is to establish a
theoretical foundation described in Chapter 2 Literature Review and Chapter 3

Model Development. Next, an appropriate research methodology should be

88

selected based on theory and the research questions. Flynn et al. (1990) and
Handfield and Melnyk (1998) provide guidelines for a systematic process to
conduct empirical research in operations management. The theoretical
foundation of this study is grounded in Normal Accident Theory and High
Reliability Theory. Use of these theories from the management literature
supports the need, and call, to utilize and incorporate theory from other
disciplines to more accurately reflect the cross-functional nature of operations
and supply chain management (Meredith, Rauri, Amoako-Gyampah, 8 Kaplan,
1989), as well as the more recent call to incorporate more behavioral theory in
operations management research (Bendoly, 2006). These theories have not
been well researched within a supply chain, and therefore provide an opportunity
to expand our understanding of these theories and their application in more
diverse contexts.

Given the exploratory nature of the research, a survey is an appropriate
method for gathering data from a broad range of individuals for increased
generalizability. The use of data from actual practitioners in a supply chain
reflects the applied nature of operations and supply chain management and can
help to close the gap between theory and practice (Flynn et al., 1990). The
survey used here gathered data from a broad range of individuals, since there
was no a priori intention to specifically examine particular types of disruptions,
risk management planning practices, or firm responses to disruptions. Therefore,
study participants did not consist of a pre-selected group of firms or individuals,

but from a heterogeneous sample that has likely experienced a wide range of

89

disruptive events and that engage in varying levels of supply chain risk
management planning. The intention is to formalize research in the area of
supply chain risk management with a strong theoretical base, and to better
understand the current state of planning that firms are engaged in to reduce the
likelihood and impact of disruptive events. Individuals will be asked relevant
questions about the activities and performance of their own firm, and about
activities between their firm and their customers and suppliers; therefore, the unit
of analysis in this study is the corporation as it manages risk both within its
supply chain consisting of its own operations and between its suppliers and
customers, and from sources external to the supply chain.

The next steps, identification of appropriate data sources and the actual
data collection, proved to be the most time consuming portion of this study. The
timeline for all of the activities from initial identification of potential organizations
for sampling through the Closure of the survey are detailed in Figure 4.1. The
process of selecting and obtaining a data source is detailed in section 4.3
Organizational Sponsorship; sample selection and the survey implementation
process are detailed in section 4.4 Data Collection. The final steps in the Flynn
et al. (1990) process, the analysis of the data for reliability and validity, are

discussed in Chapter 5 Reliability and Validity.

90

Figure 4.1 Data Collection timeline

 

£:ou:oamo.-:o: anmm Loam: anmm 2552.... E3 >023 mcozmucmmco
EEG—:0: __m 2 3 2% =2 2 53 Co 820:3: mczmomoa Em 360.3
Eom wEoofioa =mEm Siam :22 823 2:989”... o_:o=oo_m cos—8E5: Emow 5:5 89:00
2. __.E< nu :93: a no“. Ian :2. 3. :3. con ~00
pomo_o macaw ma Emm o_:o:8o_o-:o: “me“ $23 5:296: when. :_ EoEoocmm :o_~oo_om
:o_fio__oo o_:o:om_o 2 2 E8 Emu 2:28pm. @953 3:255:35 .mcozmucmmco
Son. 28 $23 :32 “mo: OBS 3-2. :3 623 Co 8:02 Em _m=_:_
no. 2. >2: m =E< 2. :22: no. :2. Eoan_o>ou 2.8m. mo. «new

.62 - .oo

91

4.3 Organizational sponsorship

Data collection from appropriate data sources is necessary to obtain
information to adequately test hypotheses and increase generalizability of the
findings. The data source identified for this study was individuals involved in the
management of a portion of the supply chain of their organization. This is an
appropriate choice because professionals involved in the management of a firm’s
supply chain are quickly impacted by disruptions to supply and demand when
they must adjust operational plans to accommodate Changes to supply and
demand plans. Because of their immediate and important role in responding to
disruptions, they may frequently be involved in supply chain risk management
and continuity planning.

One approach for contacting such individuals is through the use of one of
the professional organizations that serve a variety of individuals working in
different functional fields. Under the umbrella of supply chain management, such
organizations include those serving individuals in supply management
(purchasing or procurement), logistics (transportation or warehousing), or
operations management (production planning, manufacturing engineering, etc.)
There are multiple organizations of varying sizes and specializations that target
each of these general supply chain areas. A list of ten organizations was initially
selected based on two criteria: 1) each needed to have a large enough
membership base to provide an adequate sample size and 2) having members

that would likely have the appropriate knowledge and experience in the

92

management and performance of a supply chain to be able to suitably responded
to the survey questionnaire. The ten identified organizations were:

0 American Management Association (AMA)

0 American Purchasing Society (APS)

0 Business Continuity Planners Association (BCPA)

0 Business Continuity Institute (BCI)

- Council of Supply Chain Management Professionals (CSCMP)
. International Customer Service Association (ICSA)

. Project Management Institute (PMI)

. Society of Manufacturing Engineers (SME)

. SOLE - The International Society of Logistics

. Warehouse Education and Research Council (WERC)

This list was reviewed by academicians with knowledge and expertise regarding
the specific and differing Characteristics of these groups. Several were removed
because of particular limitations - either it was unlikely that the organization
would wish to participate in this study or that the membership was not as
appropriate a target sample as originally anticipated. Contacting the short list of
organizations required a great deal of time and effort. It was sometimes difficult
to get into direct touch with the appropriate individual responsible for coordinating
research projects within the organization. This often required multiple electronic
and telephone communiqués to reach the correct person and discuss the study.
After much explanation of the study and several rejections for participation, two
organizations agreed to participate in the study. However, as the time for the
survey approached, one of the organizations (BCI) pulled out of the study,
leaving the American Purchasing Society as the single organizational sponsor

willing to participate that possessed an appropriate membership base.

93

The American Purchasing Society (APS) serves a large member base of
over 1,600 individuals; as well other professionals who are not members but still
utilize its services. The description of the APS from their website is:

“The American Purchasing Society is an organization of buyers,

purchasing managers, executives, and others interested in the

purchasing profession. We were founded in 1969 and have
members from every state and 28 countries worldwide. Our
objective is to improve the business purchasing function through
education and our certification program. We were the first
organization to establish a certification program for professional
buyers and purchasing managers and our Certified Purchasing

Professional (CPP) program is unique because we not only

measure the competence of the applicants through a written

examination, but we conduct reference checks to evaluate the
applicant's business reputation. The American Purchasing Society's
educational objectives are achieved through training programs and

our own educational publications of interest to business and the

purchasing community.” (American Purchasing Society)

In addition to its certification program, it offers numerous courses covering topics
such as negotiation skills, purchasing law, cost and price analysis, supplier
selection, and other professional skill and knowledge enhancement classes.
They also conduct surveys of their membership to obtain current information on
professional trends and best practices. The APS is an organization committed to
the continuing education of its members and to industry, and therefore an
appropriate partner for the data needed for this study.

The American Purchasing Society agreed in late 2006 to participate in and
champion the study. Their support was important beyond simply providing
contact information was important because by endorsing and supporting the
study, survey recipients may perceive greater relevance, which lends additional

encouragement to participate and therefore increase the response rate (Aaker,

94

Day, 8 Kumar, 1998; Frohlich, 2002). APS management expressed enthusiasm
for the study, providing evidence that the topic is seen as relevant and of intrinsic
interest to their membership. In addition, APS allowed use of their logo and the
name and signature of the APS president on all correspondence with the target
sample. In exchange for providing their support, the sponsoring organization
was promised a summary report of findings.

Although the American Purchasing Society initially agreed to participate in
the study, a great deal of trust building had to occur before full access was
granted to all of the data needed for this study. Unfortunately, the APS has
participated in studies in the past where promises made by the researchers were
not kept. Thus, there were several weeks of negotiation needed to obtain their
agreement. As part of this process, coupled with a desire to protect their
members’ privacy, the initial contact information for APS individuals did not
include company name or job title information. However, after a large number of
people responded quickly to the survey, combined with the encouraging
comments written by respondents on the survey, the APS became more
enthusiastic and agreed to provide all the desired contact information.

The APS originally provided contact information, including both e-mail and
traditional addresses, when available for their 1,674 members and several weeks
later for an additional 3,764 individuals who had recently contacted APS but who
were not members (for a total of 5,438). Contact information for some individuals
was limited to only name and e-mail information. After the survey process had

begun with responses that indicated the respondents were very interested in the

95

study APS agreed to provide additional information, including job title,
organization name and addresses. A willingness to provide more information is a
function of both the evidence of interest by the membership in the study, and the
continued trust building occurring between the APS and the researcher. The
additional contact information would have been useful to have a priori so that
individuals with explicit characteristics that made them unsuitable for this study
(e.g. inappropriate industry or job title) could have been eliminated from the
contact data base. The use of this additional post hoc data will be discussed in
greater detail later.

The large data base of individuals provided by APS provides a sufficient
and appropriate target sample to adequately gather information on the constructs
to be examined and to provide adequate power for data analysis. Comparable
empirical studies have used similar target groups (Krause 8 Ellram, 1997;
Melnyk, Sroufe, Calantone, 8 Montabon, 2003; Zsidisin 8 Ellram, 2003). A
contact sample of supply management professionals is considered appropriate
because these managers are often involved in supply chain risk management
and continuity planning. In addition, supply chain disruptions impact the flow of
goods through the firm, and procurement professionals are in a position to
quickly see evidence of either a shortage or a build up of material that would be

the result of disrupted flows.

96

4.4 Data Collection

4.4.1 Target Sample

The sample selected for data collection represents supply chain
professionals who match the a priori objective of querying individuals working in
heterogeneous supply chains. As such, there was a concerted effort to eliminate
members from the full database that identified themselves as consultants or
academics, individuals from organizations categorized as service, consultancy,
educational, and other potential respondents that are undesirable for inclusion in
this study. The careful qualification of participants provides a more appropriate
sample for the study, and ensures that the general topic of the research is of
interest to the survey participants (Aaker et al., 1998). Of the original 5,438
individuals for which the APS provided contact information, 2,239 where
identified as inappropriate for this study, after the additional contact information
was provided. Because this data was not available in the initial database
supplied by the APS, elimination of some individuals deemed as inappropriate for
the study occurred post hoc (Fredendall, Hopkins, 8 Bhonsle, 2005). This post
hoc approach is less than ideal because of the effort wasted in obtaining
responses that were unusable for this specific study (although still of use for the
APS), but rigorous in the sense that while reducing the sample size, it provides
much greater reliability of the final response data, findings and conclusions.

In examining the information contained in the full-data contact list, several
criteria were developed to identify individuals that were not appropriate for

inclusion in this study, such as job title or organization name. Individuals were

97

removed based upon a job title that indicates a potential for limited knowledge of
the supply chain, their firm’s supply chain strategy or of firm performance (e.g.
purchasing coordinator, administrator, expediter, consultant, etc.) (Huber 8
Power, 1985). Individuals were also removed based upon the name of the
organization they were affiliated with (e.g. City of XXX, other government
organizations, educational institutions, etc.). The remaining individuals made up
the qualified sample consisting of 3,199 (58.8% of original list) individuals.
Because the initial survey contact was made before it was possible to screen out
unqualified individuals, any response from this category of respondent was also
removed, resulting in the elimination of 22 responses from the early stages of
data collection. The use of post hoc elimination of responses trades some
statistical power for greater reliability and validity.

Required sample size is a function of the power needed to detect and
reject a poor model (MacCallum, Browne, 8 Sugawara, 1996; McQuitty, 2004). A
sample size that is too small may result in estimates of parameters that are too
low, bias in some model goodness of fit statistics, and increased uncertainty in
replication and statistical power (Shah 8 Goldstein, 2006). A larger number of
sample respondents provide more flexibility in the types of analysis that can be
done, including the ability to potentially compare groups of respondents.
Therefore, an a prion“ design consideration was obtaining a suitable sample
which provides an adequate number of responses. This is a function of the size
of the contact sample and the anticipated response rate. It is important to also

recognize that some responses received may be unusable for a number of

98

reasons, including having the respondent fail to complete the entire
questionnaire, missing data within the response, or outliers. The possibility of the
occurrence of any or all of these issues requires some buffer to be designed into
the target sample so that the elimination of unusable cases would still leave an
adequate response sample size.

However, there can be difficulty in obtaining a large percentage of
responses from those requested to participate. There is a perception that many
potential respondents have reached a saturation point, where individuals are
asked so often to participate in a study or take a survey that they are simply no
longer interested (Bickart 8 Schmittlein, 1999). The reduction in response rates
is especially applicable to samples that are drawn from frequently utilized
sources, such as the major professional organizations like the Council of Supply
Chain Management Professional (formerly Council of Logistics Management)
(Griffis, Goldsby, 8 Cooper, 2003) or the Institute for Supply Management (ISM).
In this light the relative lack of use of the APS is one of its major positive
attributes within the context of this study. These managers have been contacted
less frequently through their organization contact information to participate in
studies, thus are likely to be more responsive to this study. In addition, there is
evidence that response rates in published supply chain research has been
decreasing to levels as low as 4.23%, especially for surveys which use contact
information from frequently targeted organizations (Griffis et al., 2003; Faught,
W., 8 Whitten, 2004). Concerns with low response rates can be mitigated with

careful selection of survey respondents, rigorous methodology and diligent

99

follow-up (Boyer 8 Verma, 2000; Frohlich, 2002; Dillman, 2007). The qualified
sample used in this study is large enough to provide the variety of potential
respondents needed and even with a relatively lower response rate will still allow

for adequate power to carry out the desired statistical analysis.

4.4.2 Questionnaire Design

The guidelines of Dillman (2007) serve as the template for the 1)
construction of the questionnaire, 2) pre-testing and 3) data collection
implementation. Careful design of the physical survey instrument has a
significant impact on the ability to adequately capture the constructs being
studied, as well as on response rates. Pre-testing is an important part of survey
research, and when the constructs have had little direct previous usage, it
becomes even more important. Pretesting is done to eliminate problems
associated with potential confusion in question wording and format, errors in
spelling or other typos, and in the general understandability of the questionnaire.
Pilot testing will assist in exposing possible flaws in question wording,
terminology or meaning before full survey implementation (Flynn et al.,
1990;Dillman, 2007).

Having a sufficient portion of the targeted sample respond to the survey
instrument is critical for increased validity and generalizability of the findings
(Flynn et al., 1990). The process of designing the survey is multifaceted and
must be carefully managed to adequately capture the concepts under study and
to encourage a high response rate. There are several elements that have been

found to have the greatest influence on actual survey participation (Aaker et al.,

100

1998; Frohlich, 2002; Dillman, 2007). The specific items and a brief comment on

how they were addressed a priori to survey implementation are shown in Table

4.1 with a detailed discussion following in the next sections.

Table 4.1 Summary of Survey Response Rate Issues

 

Issue

 

Perceived amount of work required
(related to survey length) and ease of
response

Intrinsic interest in the topic
Credibility of the sponsoring
organization

Level of induced motivation (the

incentives used)

Multiple contacts using mixed media

Personalization

How addressed

Survey questions limited to 92 items, six
demographic questions, and three optional
open-ended questions. Survey available on
line via a custom web site link, or paper
survey sent with pre-addressed and postage
paid return envelope.

The sponsoring organization expressed
interest and support for the study, including
‘championing’ the research.

APS is highly respected in supply chain
management and possesses members with
appropriate skills and knowledge.

Incentives were 1) a summary copy of the

study results; 2) a copy of a relevant supply
chain risk management article; and 3) entry
into a drawing for a small monetary reward.

At least four and up to six contacts via
electronic and paper mediums. Posting in
APS newsletter and direct contact sent
electronically and paper via traditional mail,
including post cards with the survey web-link
address and a paper copy of the survey.

All messages addressed with personal
salutation consisting of both first and last
name and electronic copies of both APS
president and researcher signatures.

 

101

4.4.2.1 Construction of the Survey

There are several key elements for increasing response rate outlined by
Aaker (1998) and Dillman (2007) that should be undertaken in the construction of
the survey. First, the questionnaire should be respondent-friendly to decrease
the time and effort required for respondents to complete the survey. This would
include an attractive and easy to follow layout of the survey with questions that
are clear and easily comprehensible, as well as easy to respond to (a copy of the
survey, including the format as sent to the target sample is included in Appendix
A. The number of questions asked will also influence respondent’s perceptions
of the ease of completing the survey. The questionnaire used here consists of
92 questions (plus six for demographics and three optional open ended) meant to
gather information about perceptions of supply chain complexity, uncertainty and
dependency; effective practices and concern for disruptions; aspects of supply
chain continuity planning; firm performance; and disruptive events. Gathering
demographic information will help to ensure the most accurate information
regarding the individual’s status with in the firm and about the firm. There are
also three optional, open-ended questions which allow respondents to provide
information on ‘best’ practices and to share their opinions on supply chain risk
management issues.

The length of the survey seeks to strike a balance between parsimony and
comprehensively gathering information by containing a sufficient number of
questions to adequately capture the constructs being examined while not

becoming so long that participants are unwilling to complete (or even begin) the

102

full questionnaire. Longer questionnaires often result in a reduced response rate,
with each additional question potentially reducing the number of responses by
0.05% (Heberlein 8 Baumgartner, 1978). However, if the survey is too short,
response rates may also be reduced because respondents do not perceive that
the survey is a serious attempt to gather information (Adams 8 Gale, 1982).
Insufficient information may also impede the ability to adequately examine the

concepts under study, reducing the validity and reliability of the findings.

4.4.2.2 Survey Pre-testing

As part of the initial pre-testing process, several individuals were asked to
complete the survey and time themselves to determine the average length of
time need to respond to the questionnaire. Individuals were selected primarily
from other doctoral students and several supply chain professionals who were
not included in the APS contact database. Because the survey was to be
distributed both electronically and in paper form, the survey was pre-tested using
both forms. Ten individuals completed the survey, and reported a mean time of
14.5 minutes to complete the survey, although none responded to the optional
open-ended questions. No significant difference between the times to complete
the electronic and paper survey was found. In addition, the pre-testers made
several recommendations regarding minor Changes in wording to improve

readability and flow.

103

4.4.2.3 Survey Implementation

The letter asking the survey contacts to complete the survey contained
instructions for completing the survey, including information that completing the
survey should take no longer than 20 minutes. It is important to note that in the
section of the survey that allowed respondents to provide open-ended feedback
on the survey, a very small number of individuals (six, representing only 1.6%)
commented that the survey was long or had too many questions. The survey
length may have deterred some individuals from undertaking or completing the
survey, especially for the paper survey when it was possible to see the length of
the entire questionnaire (the electronic survey only displayed a single page or
grouping of questions as a time). However, it is very important to note that nine
(2.9%) commented that they liked the survey and thought the topic important and
10 (3.3%) individuals specifically asked to receive a summary copy of the study
even though such a report was promised to the respondents as an incentive for
completing the study; this enthusiasm would seem to indicate that the topic is of
interest to supply management professionals. This supports other empirical
research which found that two of the most important factors positively influencing
survey response was the saliency of the topic to the target sample members and
the sponsoring organization (Larson, 2005), both of which are achieved in this
survey.

The frequency and type of contact (e-mail, phone, traditional mail, etc.)
with the target sample is a second area of interest in designing a sampling plan.

Dillman (2007) recommends multiple points of contact using mixed methods.

104

There is not a specific recommendation of the types of contact to be used,
therefore, the contact methods utilized in this study were slightly modified as the
actual data collection occurred and in response to the specific requests of the
participating organization. The first contact with the target sample occurred
electronically. This method has several advantages. A primary advantage of
using electronic surveys is a reduction in the expected effort required for the
respondents to complete the questionnaire, a key element in improving response
rates (Aaker et al., 1998; Frohlich, 2002). Another important advantage is the
ability to avoid data transcription errors that can occur when the responses from
a completed paper survey are later entered into an electronic form for data
analysis. In addition to ease of implementation and speed of receiving
responses, there is evidence that data utilizing paper and electronic collection
processes does not generate problematic differences in results and can be used
to support the same conclusions (Griffis et al., 2003), making it an attractive data
collection method (Klassen 8 Jacobs, 2001).

Multiple electronic and paper mailings were sent to further aid in
increasing the response rate, thereby reducing non-response bias. Multiple
mailings allow for non-response bias testing by comparing the results from early
and late respondents where late respondents serve as a proxy for non-
responders (Armstrong 8 Overton, 1977; Dillman, 2007). It is important to know
if non-respondents are significantly or systematically different from respondents,
as this will compromise the generalizability of the findings (Flynn et al., 1990). In

general, an increased number and variety of contacts, and a greater ratio of

105

contacts that are made using paper methods (letter, postcards, etc.) will result in
higher data collection costs. It is therefore important to carefully consider the
expense per collected response when determining the number and mode of
communications to be sent out (Kaplowitz, Hadlock, 8 Levine, 2004).

The first contact with the targeted sample was indirect when the APS
posted a notice in their monthly newsletter that a research project was being
undertaken to study risk management. This notice also encouraged readers to
respond when they received the survey invitation, in effect ‘priming’ them for the
survey to come. The first direct communication with the target sample for which
e-mail contact information was available was an electronic pre-notice message
sent to the identified sample respondents a short time before the survey was
conducted, to provide a brief explanation of the importance of the study, the
value of the individual’s response, and indicating the support of the APS. The
use of pre-notification has been found to be even more important when e-mail is
being used, due the to amount of ‘junk mail’ and ‘spam’ that many people receive
(Porter 8 Whitcomb, 2003; Kaplowitz et al., 2004). The concern regarding spam
was addressed by having the pre-notice sent using the APS and university logo,
and the name and signature of the organization president and lead researcher on
the electronic communication to alert individuals to expect the survey. APS
sponsorship helped impress upon recipients the support of the parent
organization and the importance of their response.

The pre-notice was followed electronically a few days later by an

electronic form of cover letter. This letter more fully outlined the purpose of the

106

study, provided a web-link to access an electronic version of the questionnaire,
described important information on the voluntary nature of participation (in
accordance with Michigan State University’s UCRIHS human subjects research
protocol), and expressed appreciation for completing the survey and information
about the response incentive.

In addition, all electronic messages were sent using ‘mass mailing’
software that provides a couple of key features meant to address the issues of e-
mail ‘spam filtering’ and the possibility of invalid electronic addresses. The
software utilized sends a pre-signal to the address to confirm that it is valid. In
addition, each message is sent individually so that they are not received as part
of a mass distribution on the host system and they are sent with an individual
return e-mail account. In this case, a new secured e-mail account
(msurisk@bus.msu.edu) was established specifically to be used for sending
these messages and to allow for the receipt of any queries that individuals might
have about the survey. This proved to be quite useful, as some of the target
sample had e-mail systems that upon receipt of a message from an unknown
source sent an automatic reply to the sender address to request that the sender
undertake some form of identification or process to establish the message as
actually coming from an individual. The use of an e-mail address that is survey
specific also allowed for the receipt of automatic messages that indicated if an
individual was no longer part of the organization, allowing for maintenance of the
qualified contact list. Unfortunately, it was not possible to determine if any

messages were stopped and excluded by the recipients’ e-mail filtering software.

107

Therefore, the number of successful contacts made is less than or equal to that
reported. As a result, the response rate reported is a conservative estimate.

The timing of electronic communications is also important. The initial
invitation to participate in the study was sent mid-week and the messages were
sent in the early morning hours (US. Eastern Time) so that the messages would
be available when the recipients opened their e-mail in the morning. Both the
daily and weekly timing of survey invitations have been shown to increase
response rates (Faught et al., 2004). In addition, the use of a web-link for survey
response has been shown to increase response rates (Klassen 8 Jacobs, 2001).

Several days later, a second electronic message with a similar cover letter
and web-link to the survey was sent, followed by a third message after several
more days. Those individuals with electronic contact information were therefore
contacted up to four times: first with the pre-notice and subsequently with the
invitation to participate in the survey and the link to the actual survey. If an
individual provided their contact information when responding to the survey, their
name was removed from subsequent mailings [by the end of the data collection,
255 individuals (84%) provided their complete contact information]. This reduced
the intrusion of multiple contacts for participants, with the intent of curtailing the
perceived imposition on their time.

Personalization of all correspondence sent is another important aspect to
improving response rates. This would include using the individual’s name in the
salutation and hand written or electronic signatures (rather than simply typed

names) in the Closing. It is important to note that personalization now has a

108

reduced impact, especially in electronic communications (Porter 8 Whitcomb,
2003) due to the ease of using ‘mail-merge’ and other programs to easily insert
predetermined personal data into what is otherwise a form letter. An alternative
method for making the individual feel important could be by telling them that they
are part of a select group asked to participate in this survey and utilizing a closing
time of the survey to create a sense of urgency. This is another point where the
support of the sponsoring organizations can be used to improve response rates.
All contact with the target sample, electronic and paper, included a personal
salutation to the recipient.

Finally, some sort of incentive can be used to further encourage
individuals to respond to a questionnaire. Three incentives were used to
encourage respondents to complete the survey. One of the incentives utilized
here was the promise of a summary copy of the findings. Copies of the summary
report were sent to the individuals that provided contact information, and to the
APS. The APS can then decide if the report should be sent directly to all
members or simply posted on their web—site for access by all those who are
interested. Due to the low initial response rate, a second incentive was included.
Access to a copy of the paper “18 Ways to Guard Against Disruption” (Elkins et
al., 2005) was promised and provided. Permission to utilize this paper as an
incentive was granted by the authors and by the publisher. The final incentive
utilized was financial. Respondents were given the opportunity to enter into a
drawing for an individual reward of a $50 Visa gift card. If an individual chose to

respond anonymously to the survey, there was no way to enter them into the

109

drawing, and no automatic or predetermined identifying information was attached
to any survey sent or to any response.

After it was decided that no further responses would be generated by
electronic contact (the pre-notice and three invitations to complete the survey), a
paper invitation and questionnaire was sent to the non-responders. The paper
invitation was printed on Michigan State University letter head, with a copy of the
APS logo and with the names and signatures of both the APS president and lead
researcher. This letter contained the same information as the electronic letter
(instructions for completing the survey, web-link information, UCRIHS disclosure
and voluntary nature of participation, etc.). Also included was a pre-addressed
business reply envelope. Given the large number of survey packages sent, and
the number of electronic contact already undertaken, only a single set of paper

surveys were sent out.

4.4.3 Response Rate

A variety of techniques were used to increase the response rate, in
keeping with the guidelines of Dillman (2007). As previously mentioned, all
contact with the individuals included the logos of both the APS and Broad
Graduate School of Management at Michigan State University as well as the
signatures of both the APS president and MSU lead researcher (copies of survey
contact material is in Appendix A.). A pre-survey notice was sent to individuals,
and both electronic and paper surveys were used. Incentives for participation

included the ability to request a copy of the survey results, access to a

110

practitioner article on the subject of supply chain risk management and the ability
to register for a Chance to win a gift card worth $50.

Overall, the final response rate was 9.5% (303 responses from a qualified
sample of 3,199). Response rates of approximately 10% are considered
acceptable when using similar professional organizations as the target sample
(Melnyk et al., 2003), although lower response rates are often used in supply
chain research (Griffis et al., 2003). Of these, 246 (81%) responded
electronically and 57 (19%) responded by returning the paper survey. The paper
survey sent out also included information for accessing the electronic survey,
which some respondents chose to utilize, resulting in a lower paper survey
response rate. The order and wording of the questions was identical, so this is

not a concern. A summary of the response rates is shown in Table 4.2.

Table 4.2 Responses Rates

APS Non-APS
Members Members

 

Complete Database 1,674 3,764

Qualified Sample 1,199 2,000

Qualified Responses 188 115

Qualified Response Rate 15.7% 5.8%
Total Qualified 3,199
Total Qualified Responses 303
Total Qualified Response Rate 9.5%

 

As can be seen, the response rate for the APS members is much higher
than for non-members. This result is not surprising, since there is evidence that
response rates are higher when there is high saliency of the research topic and

identification with the sponsoring organization (Larson, 2005). As previously

111

mentioned, APS regularly conducts surveys of their membership, and they have
an organizational mission of education and service which includes providing
members with current research in the field of supply management.

Another issue that is important to discuss is the use of actual APS members
and non-members. During initial contact with APS regarding participation in the
study, the membership number originally provided by APS was in fact a
combination of both members and non-members. This inflation of membership is
partially a result of the desire to appear as a more substantial alternative to more
well-known organizations, such as the Institute for Supply Management (ISM,
formerly NAPM). The quote of the combined contact sources was an unwelcome
surprise for the researcher. The use of members and non-members in this study
has both drawbacks and advantages. The most obvious is the higher response
rate from APS members. The contact sample of all qualified individuals if they
were members, and using the same response rate a members, would have
provided a much larger response sample (extrapolated to be ~500) which in turn
would allow for a greater range of data analysis techniques and increased
possibility of sub-group analysis. Conversely, the inclusion of both organizational
members and non-members provides a more heterogeneous supply professional
sample which increases support for the generalizability of the findings.

In summary, the data collection was carried out utilizing contact information
of supply management professional provided by the American Purchasing
Society. Initial contact with the APS began in September 2006. Because of the

approaching holiday season and a perception that many business professionals

112

would be exceptionally busy at the end of the calendar year, the start of the
actual data collection was held until the beginning of 2007. APS provided a
notification of the impending survey in their January 2007 newsletter. An e-mail
pre-notice of the survey was sent to all the individuals for whom the APS had
provided contact information. The mass e-mail software program used is able to
identify undeliverable e-mail addresses prior to sending out the messages
inviting the individuals to participate in the study and providing a link to the
survey website. The high initial response rate encouraged the APS to provide
more detailed information on the contact list already supplied. This increased
visibility raised questions regarding the appropriateness of some of the sample,
resulting in 2,239 people being removed from the original sample of 5,438 for a
remaining qualified sample size of 3,199. Through the month of March and April
several e-mail invitations and post card reminders where sent to illicit a higher
response rate. In addition, a full paper copy of the survey was sent to the APS
non-responders in late March. In total 246 people responded electronically via
the survey website and 57 people responded to the traditional paper survey, for a
total of 303 responses (9.5% response rate). This response rate is in keeping
with similar studies (Griffis et al., 2003; Melnyk et al., 2005). , and as detailed in
the analysis of the collected data that follows in the next Chapters is large enough

to provide adequate statistical power (McQuitty, 2004).

113

5. RELIABILITY AND VALIDITY

Before advanced methods of data analysis can begin, it is necessary to
examine the data utilizing fundamental statistical techniques to 1) determine the
degree to which the data adheres to the basic assumptions underlying
multivariate methods; 2) evaluate and adjust for missing data; and 3) identify
outliers (Hair Jr., Black, Babin, Anderson, 8 Tatham, 2006). In addition, it is
important to assess the demographics of the respondents to determine if the
response sample adequately represents the target population. The following
sections discuss steps undertaken to ascertain the appropriateness of the
sample data for use. All statistical information reported in the following sections

was developed using SPSS 15.0, unless othenrvise noted.

5.1 Demographics

The target sample for this study is a broad cross-section of supply chain
professionals. Therefore, individuals with differing levels of responsibility within
their firm and representing varying sizes and a range of industries are included.
Possession of current demographic information about the respondents allows for
comparison of clusters of responses along dimensions such as level of
responsibility/influence within the firm, firm size, or industry (Flynn et al., 1990).
The following summarizes the demographics provided by the survey
respondents.

The contact information provided by the APS did not include data on each

individual’s industry. There were also no a pn'on' assumptions regarding linkages

114

between supply chain disruptions and specific industries. Therefore, there was
no need, nor useful way, to eliminate individuals from participating in the survey
based on industrial sector, other than obvious elimination based upon the
individual’s firm affiliation as reported to APS. As can be seen from Table 5.1,
the respondents represent a wide range of industries, although the majority
identified themselves as working in manufacturing. Surveys from individuals in
industries such as communications or finance indicated that their area of
responsibility included supply chain activities and are therefore appropriate for

inclusion in this study.

Table 5.1 Industry Sector of Survey Respondents

 

 

lnduitw Sector 3 Percent
Agriculture 1 0.3
Communications 4 1 .3
Construction 10 3.3
Finance, Insurance, Real Estate 9 2.9
Government 1 1 3.6
Health Services 12 4.0
Hospitality 8 Food Service 8 2.6
Manufacturing 171 56.4
Retail or Wholesale 29 9.6
Transportation 8 Warehousing 6 2.0
Utilities 3 1.0
Other 38 12.5
Missing __1 0.3
Total 303

 

Firm representation spans across a wide spectrum of sizes, as measured
by both firm revenue and number of employees, as seen in Table 5.2. There are
several respondents that indicated they were not authorized to provide revenue

information, resulting in this specific item having the largest number of missing

115

responses. There are no respondents that omitted both revenue and

employment information.

Table 5.2 Indicators of Firm Size

 

Annual Reventfi g Percent

Less than $10 million 33 10.8
$10 to 249 million 84 27.7

$250 to 499 million 57 18.8
$500 to 749 million 28 9.2
$750 million to 1 billion 83 27.4
Over $1 billion 4 1.3
Missing _1;4_ 4.6

Total 303

 

Employees n Percent

Less than 100 65 21.5
100 to 249 48 15.8
250 to 999 73 24.1

1,000 to 4,999 54 17.8

5,000 to 9,999 17 5.6
Over 10,000 43 14.2
Missing __3_ 1.0

Total 303

 

Most of the individuals report that they have functional responsibility for
supply management (purchasing, procurement, buying, etc.) as Table 5.3
indicates. There are however 36 individuals that report different functional
responsibilities other than supply management. However, the functional area of
responsibility reported is closely enough related to still remain relevant to this
study. In addition, as can be seen in Table 5.4, survey respondents are primarily
in a managerial and supervisory level of responsibility, with several individuals at

the Director or Vice president level. Finally, Table 5.5 illustrates that respondent

116

organizational level is almost equally split between corporate level and strategic
business unit along with plant levels. The diversity of respondents provides

support for the generalizability of the findings.

Table 5.3 Respondents’ Functional Responsibility

 

 

Function Responsibility g Percent
Engineering 3 1.0
lT/IS 4 1.3
Logistics 1 0 3.3
Production Control/Planning 13 4.3
Supply Management 265 87.5
Warehousing 5 1.7
Missing _2 0.6
Total 303

 

Table 5.4 Respondents’ Level of Responsibility

 

Level of Responsibility n Percent

Vice President 10 3.3
Director 46 15.2
Manager 142 45.6

Supervisor 26 8.6
Individual Contributor 72 23.8
Other 7 2.3
Missing _3_ 1 .0
Total 303

 

Table 5.5 Respondents’ Organizational Level

 

Organizational Level II Percent

Corporate 144 47.5
Strategic Business Unit 88 29.0
Plant 59 19.5

Other 8 2.6
Missing __4 1.3
Total 303

 

117

5.2 Missing Data

Missing data can present a problem for data analysis. This is particularly
true with surveys where data gathering issues such as response fatigue may
result in more blank responses toward the end of the questionnaire or if
individuals do not feel comfortable or lack the knowledge to respond to specific
items or a set of items. The issue of missing data is similar to non-response
bias. In non-response bias, the researcher is concerned with patterns and
causes for not participating in a study and with missing data the concern is also
for patterns and causes of non-response within the collected data. From a
practical standpoint when cases with incomplete responses are removed from
the data, a reduction in usable sample size may limit the ability to robustly
conduct data analysis, or require the gathering of additional data to reach an
appropriate sample size for adequate power. From a substantive standpoint,
missing data may occur non-randomly, indicating bias and affecting the validity
and generalizability of the findings (Hair Jr. et al., 2006; Tsikriktsis, 2005).

A four-step process to identify and handle missing data is outlined in Hair,
Black, Babin, Anderson, and Tatham (2006) and serves as the guide for the
missing data analysis process utilized in this study. The first step involves
determining the type of missing data. Second is examining the extent of missing
data by case and by item, and searching for patterns. Third is determining the
level of randomness of the missing data. The final step is the selection of an

appropriate imputation method to estimate the missing data, if needed. The

118

findings and outcomes of applying these steps are detailed in the following
sections.

Some missing data is to be expected in almost all empirical data collection
undertakings. Data that can be considered ‘ignorable missing data’ occurs at
random and does not require specific remedies. This type of missing data is a
result of imperfect representation In the sample of the complete target population,
‘skip patterns’ where respondents skip over sections or questions they are
unable to answer or are not applicable in a survey instrument and data that is
censored because it is inconsistent across the sample or not applicable to the
current study. Of greater concern is missing data that is not ignorable and can
be caused by either known or unknown processes. Identifiable processes that
can explicitly cause missing data are often related to the survey instrument or
data collection procedures. They may include a failure of respondents to
complete the entire questionnaire or an identified data entry error. Unidentifiable
processes, by definition, are more difficult to detect and to remedy and frequently
are related to the respondents’ actions. These may include respondents’
unwillingness to answer sensitive questions or inability to provide the requested
information. Care taken in the development of the survey instrument and the
data collection methods may reduce the likelihood of substantial levels of non-
ignorable missing data.

When non-ignorable missing data is found, it is necessary to determine
the extent of the missing data. If the level of missing data is low enough, even if

it occurs non-randomly, there may be no affect on the data analysis results. The

119

assessment of missing data should be conducted to find both individual items
and cases with larger than desirable levels of missing data. It is assumed that
some data will be missing; however, when data is missing at low levels or in a
random pattern, it is often possible to retain the items and the cases in the study;
however high levels or non-randomly missing data presents a problem and must
be addressed before further data analysis can proceed.

The level of missing data can be determined by assessing the percentage
of items (representing survey questions) with missing data and the number of
cases (representing respondents) with missing data above a pre—determined
threshold. The goal is to determine if there are any cases or items with levels of
missing data that are high enough to cause concern, or that indicate an obvious
omission pattern. The survey sample was examined using the ‘Missing Value
Analysis’ module in SPSS 15.0 (excluding demographic and open ended
questions) to determine the extent and pattern of missing data for both items and
cases. Of the 92 items examined, nine (10%) had no missing data and the
remaining 83 items had missing data in less than 5 percent of the cases. No
specific patterns of missing data at the item level were observed. Given the low
levels of data missing per item (survey question), the fitl within general guidelines
and the lack of a discemable pattern (Hair Jr. et al., 2006), no items were
removed due to missing data.

Of the 303 cases included, 182 (60%) had no missing data, 65 (21%)
cases were identified as missing data on a single item, 51 (17%) cases were

missing data for two to eight items, and five (2%) cases had data missing for nine

120

or more items. Individual cases which have less than 10 percent of the items
missing data are considered usable, as long as the data is missing randomly
(Hair Jr. et al., 2006). In order to protect against possible bias the five cases
missing more than 10 percent of items were eliminated, bringing the sample size
to 298 cases.

The number of cases with missing data here is somewhat higher than the
average found in a recent study of missing data from articles published in the
Joumal of Operations Management between 1993 and 2001 (Tsikriktsis, 2005).
Interestingly, 67 percent of the articles examined did not report any information
regarding missing data in their samples. It is therefore difficult to determine if
missing data occurs frequently but is simply rarely explicitly addressed, if the
issue is simply ignored as surmised in (Tsikriktsis, 2005), or if authors are
reluctant to report high levels of missing data. Based upon the high number of
items on the survey, it was determined that further examination of the missing
data was warranted to determine the usability of the remaining sample data.

The third step in diagnosing missing data is to determine the randomness
of the missing data. The test for missing completely at random (MCAR) requires
comparing the observed pattern to determine that it does not differ from a
random pattern. Little’s MCAR test has a significance value of 0.67, indicating
that the missing data here is MCAR (Hair Jr. et al., 2006).

The identification of missing data as MCAR allows for the broadest choice
from possible imputation methods to determine the value of missing data points

based upon the values of other items and cases in the sample. There are

121

several methods of imputation that are dependent upon the amount of data that
is missing, the sample size and the level of conservativeness of the researcher (a
summary of multiple methods can be found in Hair, Jr. et al, page 63). For
example, the most conservative method would be to eliminate all items and
cases with missing data. Other methods would include substitute missing data
based upon similar cases or calculating missing data using regression analysis.
The simplest calculation method is to utilize mean substitution, but the
disadvantages of this method are a reduction in the variance of the distributions
and in the observed correlations, as well as some distortion of the data
distributions. lmputation by regression has the advantage of utilizing the actual
relationships among variables, but may generate replacement values that fall
outside of the bounded Likert scale range utilized here.

After careful consideration of the benefits and disadvantages of data
imputation methods, a model-based method was selected to impute the missing
data in this study. The advantage of this method in this situation is that it allows
for the least bias by providing the best representation of the original distribution
of the item values. The lack of bias is important for the generalizability of the
findings. In addition, this method can accommodate both random and non-
randomly missing data (Hair Jr. et al., 2006). The imputed data values were
created using the EM (expectation-maximization) algorithm to estimate the
means, the covariances, and the Pearson correlations of the missing data. This
is a two step iterative process, where the first step (“E”) computes expected

values using maximum likelihood based upon on the observed data and the

122

current estimates of the parameters. The second step (“M”) step uses the new
parameter estimates from step ‘E’ to re-estimate the missing data, and continues
until convergence of parameter estimates is achieved (Tsikriktsis, 2005; SPSS
15.0, 2007). The remainder of the data analysis is conducted using the data set

with the missing data imputed using the EM process.

5.3 Data Normality

Both univariate and multivariate statistical analysis is based on
fundamental assumptions of the data characteristics. When these assumptions
are found not to hold true for the data, it is often necessary to utilize various data
transformation methods or to alter the types of data analysis that can be
conducted. Normality refers to the distribution shape of the data. Meeting this
assumption is important, since a large degree of non-normality inhibits the use of
many basic statistical analysis tools (Hair Jr. et al., 2006). Two measures used
to examine the shape of the data distribution are kurtosis (the tendency to a
‘peaked’ or ‘flat’ distribution) and skewness (the tendency to have a
disproportionate amount of data in the left or right ‘tail‘ of the data). High levels
of kurtosis, usually bounded by cut-off value of +/- 3, indicate that the variance
observed in the data is due to extreme, but infrequent observations; both tails of
the distribution are elongated with a ‘flatter’ probability distribution. High levels of
skewness, usually bounded by +/-1, indicate a tendency of the data to have
either an elongated right or left tail. Observed together, kurtosis and skewness
test the centered symmetry of the data distribution. In examining the kurtosis of

the data, only one item falls beyond the cut-off value of +/- 3; no items fall beyond

123

the skewness cut-off value of +/- 1 (Bollen, 1989; Hair Jr. et al., 2006). Use of
the Kolmogorov-Smimov and Shapiro-Wilk tests, both of which test the goodness
of fit of the sample distribution with a normal and uniform distribution, indicate
normality at a significance level of p <.01 for all of the items (a subsequent
multivariate test of kurtosis with the confirmatory factor analysis achieved a
Mardia-Based Kappa of 0.10). In summary, there is no strong evidence that the

data collected is non-normal.

5.4 Response Bias

5.4.1 Non-response Bias

In order to assess potential non-response bias, an ANOVA comparison of
early responders and later responders was conducted. This method assumes
that late responders are similar to non-responders (Armstrong 8 Overton, 1977).
In this case, late responders are individuals who received multiple invitations to
participate, thus they likely would have been non-responders if not contacted
multiple times, providing support for the assertion that they are likely to be similar
to non-responders. The response sample was segmented into early, mid and
late responder groups, each consisting of approximately 1/3 of the response
sample. This was done to clearly separate early and late responders.

The exact time and order is known for the electronic responses due to the
date and time stamp assigned as the data is recorded into the electronic file that
received the data. The exact timing and order of response for paper surveys is

less precise, since there are varying delivery lead times (determined by outbound

124

and inbound travel distance, and by internal mail processing routines on both the
respondent and research team mail handling systems); however, paper
responses were entered into the data base in the order received for a close
approximation of order and timing of actual response. In examining the response
timing stratification, 104 (34%) of responses were categorized as ‘early’ and 112
(37%) as ‘Iate’. An ANOVA of the item responses revealed that only 2 (2%) of
the items had statistically different means and standard deviations at a p < .05
significance level. Given the large number of items (92) in the questionnaire and
the low number of items that exhibit statistical difference, there is no substantial

indication that non-response bias is present.

5.4.2 Common Method Bias

Common methods bias is the difference between the observed and the
actual construct relationships, resulting in possible inflation or deflation of the true
relationship between constructs. The resulting discrepancy between observed
and actual constructs occurs as an artifact of the instrument used to gather the
data, which inadvertently impacts or influences the respondents’ reported
measures or scores. The method used to gather data allows systematic variance
that confounds the variance associated with the traits being measured (Doty 8
Click, 1998). Although the respondents were given the option to respond
electronically or by paper, the questions asked were identical and presented to
the respondent in the same order. Therefore, although the data collection
technologies are different, the respondents are utilizing a single instrument. Data

used in this sample was collected from single respondents within a firm,

125

examination of the contact information revealed that no individuals with the same
organizational name/identification are included in the sample. Given the focus of
the survey and the use of single respondents, there is the possibility of the
results gathered to potentially exhibit common method bias.

Podsakoff, MacKenzie, Lee, and Podsakoff (2003) offer several
techniques for controlling common method bias a priori as part of the survey
instrument creation and data collection design. Suggestions used in this
research include the protection of respondent anonymity, which reduces
evaluation apprehension and aids in increasing answer honesty and reducing the
occurrence of social desirability, leniency and acquiescence in responses.
Another precaution taken was the careful construction and multiple rounds of
testing undertaken on the scale items for improved comprehension, avoidance of
‘double-barreled’ questions, decomposition of complicated questions into simpler
and more focused questions, and the avoidance of specific syntax that may lead
to question ambiguity. Careful construction of scale items and survey questions
was undertaken as an a priori precaution to mitigate common method bias, but it
is still important to test for the presence of bias in the survey results.

Examination of the data for common method bias was conducted using a
confirmatory factor analytic approach of Hannon’s one factor test. The test
operates on the assumption that if common method bias is present, a model
containing a single latent factor would account for all manifest variables with fit
equal to a multifactor model. Thus, a single factor model with worse model fit

than the hypothesized measurement model would indicate an absence of

126

common method bias. The model fit values for a one factor model yielded a 12

= 9065.67 with df = 2410 compared to 12 = 4294.78 with df = 2319 for the

measurement model. Given the considerably worse fit for the unidimensional
model, there is no indication that common method bias is a concern. A second
point is that the model contains full mediation, which also serves to reduce the
impact of common method bias (Podsakoff, MacKenzie, Lee, and Podsakoff,

2003)

5.5 Confirmatory Factor Analysis

The data analysis undertaken here follows the two-step approach
advocated by Anderson and Gerbing (1988). The first step is to perform a
confirmatory factor analysis (CFA) on the measurement model to examine the
data for convergent and discriminate validity. CFA is used when there is a priori
assignment of variables to their associated constructs. Confirmatory
measurement models specify the relationships between the observed variables
(measurements) and the underlying constructs (factors) with the constructs
allowed to freely correlate. A properly specified, converging measurement model
is a necessary precursor to examination of the relationships between constructs
(Anderson 8 Gerbing, 1982; Bagozzi 8 Heatherton, 1981 ). Once a proper
measurement model is achieved, the second step is to examine the structural
relationships of the factors in the path model (this will be further discussed in the
Chapter 6: Data Analysis). A summary of the CFA model fit is found in table 5.6

and discussed in detail in the following sections.

127

5.5.1 Measurement of Constructs

Before a CFA can be conducted, it is necessary to determine if the sample
size is adequate to achieve a desired level of statistical power. Adequate power
is needed to detect close model fit. Using McQuitty’s (2004) guidelines, the
degrees of freedom (df = 1363) and sample size (n=303) the data set are
adequate to achieve power of at least 0.80. Another consideration is the ratio of
manifest variables (MV) to associated latent variable (LV) (Shah 8 Goldstein,
2006). A MV is an observed variable (the items from the survey) which are
intended to measure specific unobserved LVs. Latent variables are conceptual
constructs that cannot be directly measured, and therefore require the use of
multiple MVs to adequately capture the underlying LV. It is generally accepted
that single indicator LVs are undesirable, and that a minimum of three MVs per
LV should be used. A larger ratio of MVs to LV can compensate for a smaller
sample size (Marsh et al., 1988), but requires a larger sample size when there
are more parameters to estimate to achieve adequate power and also reduces
model parsimony (Anderson 8 Gerbing, 1984). The model used here has 92
MVs for 13 LV constructs, for a MV to LV ratio of 7.08 All initial LVs have three
or more MVs, with no single indicator LVs.

The ratio of sample size to the number of estimated parameters is also an
important consideration in CFA. When the sample size to estimated parameters
ratio is too low, there is a higher level of susceptibility to overestimated model fit
and lower reliability. The use of model fit estimates that are less biased by

smaller sample size can alleviate this issue (Hu 8 Bentler, 1998). While the

128

general guideline is have a sample size to parameter estimate of 5:1, a recent
study of operations management research found that 35.7% of the models had a
lower ratio (Shah 8 Goldstein, 2006). The ratio of sample size to estimated
parameters here is slightly below the 5:1 guideline at 3.3:1, but in keeping with

other operations management research.

5.5.2 CFA Model Fit

After determination of adequate conditions as described above including
1) a minimum of three MVs per LV with no single indicator LVs, and 2) evidence
of an acceptable sample size for the number of estimated parameters, the next
step is to conduct a CFA analysis. The first step of Anderson and Gerbing’s
(1988) two step process is to assess the psychometric properties of the
measurement models via CFA. Using CFA, the overall fit of the measurement
model can be assessed using a variety of fit indices. In general, fit indices can
be categorized as absolute (the degree to which the sample data is reproduced
by the measurement model) and incremental (the improvement when the
hypothesized model is compared to a restricted baseline model) (Bollen, 1989;
Hu 8 Bentler, 1998). A summary of the most commonly used fit indices is shown
in Table 5.6. The specific types of fit indices within these two categories are
identified in the order of most common usage in operations management
research according to the findings of Shah and Goldstein (2006).

There are several absolute fit measures, the most basic and frequently
used of which is the 12 statistic, which quantifies the difference between the

observed and estimated covariance matrices. Within the context of SEM, lower

129

12 values indicate support of the model as representative of the data. The

goodness of fit index (GFI) was developed to address issues of large sample size
influencing model fit results and adjusted GFI (AGFI) increase as fit improves, up
to a limit of one; generally accepted minimum levels range from 0.90-0.95 (N. W.
Browne 8 Cudeck, 1989). AGFI takes into account model complexity by
adjusting the GF I by a ratio of the total degrees of freedom to those used in the
model. Used less frequently is root mean squares of approximation (RMSEA),
which attempts to correct the tendency of rejecting models with either large
sample sizes or a large number of observed variables. Both the Root mean
square residual (RMSR) and Standardized Root Mean Residual (SRMR) rely
upon standardized results and may make model comparisons more difficult.
RMSR and SRMR decrease as the measurement model fit improves, and are
bounded by zero with acceptable values in the area of 0.08. These metrics are
sometimes called ‘badness of fit’ measurements where high values indicate poor
model fit. RMSR tests the accuracy of the prediction of the variance for
individual items; SRMR uses standardized RMSR values and is useful when

comparing fit across models (Hair Jr. et al., 2006).

130

 

w 3.0 E .08:
0.8.0:. 8:.0> 02:08
.0:9: 8: 00:8: 82

00.0 08:00 8 88.... 80.
0.000.: 89:08.98 08. 89.080

3:

Eu 0: .98..» 2.0080080 .08. 0.5.0.. .0083

Table 5.6 Summary of Model Fit lndices

.E .08: .

0890:. 828 8:9: 8.8050888. 2.8.8.00 .002: .0. 8.800000 :00. 80:.

:0. o 8 00:0. 8:82 .009: 0. 02880 0.08:. 89:05.38 80. 80.08.. .5 .....z :8: 88.05. E 0>:0.8:.00
2.8.9.80 .000:.

m 00.0 E .08: .000:. 0: 02:88 .8888 :0: 0:0 8:: .0: 88> NR :02.

0:00.05 8:.0> 8:9: .88 .0. E 88:58.08: 802:0: 00:08.8 8 0:01 80:. 2... 00::02
28.88 030.00.: as. 2.0

08... 2 8.88.. _88. .o 2.0. .3 _8 08.2 :0... E .o

:. . .00 :05 .030. 2.0003 0.0008 80.:0E.0.8 08. 89.088 0:88:80 800:. 880.03. 880000 8832

 

.002: :0: .0 0:88: 0>:0:.0:.0 o: 800:. 8:88 0 8 E 8.8:.00 ”80.05 g... 38:88:.

 

2002:.

.0208. 00:00:80 0:..E080 0: 2._.:< :0:0:..x0.8<

.9_0:..0:-:0: 0:0 88050809.: E0000: 8 00.000 8 .0..m.

mod w .08: m. 830.. .0008 .00.». 0.980 .3 880:0 80.. .8 E 8 0.00. 000.08 0808.". 0.080 805.801

2...... 55.; .o a 880 9.2%.

”E .08 8:00.05 80.0., :00:8:.00 .0008". :005.

090. 5.3 ..E 8 80:08.. .08.: .0. 0.50:0 :0:0~_0.00:0:m mwsm 800.0805 800. 890.0805
00.00.0808 0:8. 00:0..0200 8:0E:8

0.0 3.08. 08.:0 0:0 82080 80.20: 80.2% .8080.

0.000.: 0.8800 2 20055 0.808. 0:: 8 0m0.0>< 0.003 800.2 80m.

8.89.8.0 05.0800 :0

9.0 w .08: 0. .033 .08 .0008. 8:.E0xm.

w omd E .000:

0808:. 8:.0> .0:9: 2 8 80:0 0: 000 280:8. 08 08:00 .090. 0.8 0.0800 .090. :00.
.F 0: o 8 00:0”. 08 08:00 0. 02:800. :5 .6 82.0 9.08. 08.80.88. 80:8m. 8 80:0 05 80.000 0: 880.2 E 8 08:80.0
9.08098:
0:0..0>:_::. 0.08.8.8 80:85 008880 88...:8
.8000 0:88 0: .E .08: 88...:8 0.0:. . 0N... 0:0 82080 0:: 80.20:
0:00.05 80.8 .08.. 0.8.00 .090. 0: 3:880 0580:. E 2mm 88: 2.00.888 2:0 00:08.8 0:. 2.8000 0: 0003 NR

131

 

80.50.: 00:0..0>00 002080 0:0 00.88. 05.00.80 .080 0.980 0:. 80:00.00. .009: 00.88. 0:: :03 30: 8088... ”8.5000: 0... 05.023

 

0000.0:

80.? £0.50 8008020005 800E023

 

Incremental fit indices are used to compare the research model to a
reference model of either the worst case (null) model or to an ideal model that
perfectly represents the modeled phenomena under study. The most commonly
used incremental fit indices are comparative fit index (CPI) and normed fit index
(NFI). CFI is an improvement over the NFI and is less sensitive to model
complexity, with minimum acceptable fit value of approximately 0.90 (Hair et al.,
2006). Less commonly used in operations management research are non-
normed fit index (NNFl or TLl), and incremental fit index (lFI or BL86) (Shah &
Goldstein, 2006).

The acceptable cut-off values for each specific fit measurement are an
issue of continuing discussion in research methodology literature within a variety
of social science disciplines (Bentler 8: Chou, 1987; Bollen & Lennox, 1991; Hu &
Bentler, 1998). Compounding this debate are numerous other factors, including
sample size, the number of latent factors and the number of their corresponding
manifest variables, the degree to which a study is more or less exploratory as
opposed to confirmatory, and researcher sensitivity to Type I or Type II errors. In
addition, it is also often necessary to examine two or more fit measurement
values simultaneously since each has its own unique strengths and weaknesses,
and it often occurs that a marginal fit value with one measurement may be off-set
by an acceptable value on another measurement (Hu & Bentler, 1998; Yuan,
2005)

The choice of appropriate fit measures is also a function of sample size,

as many indices are sensitive to issues inherent to the examination of data from

132

smaller samples. Hu and Bentler (1998) suggest that researchers use one of two
sets of indices to assess model fit depending on the sample size. The sample
size of 298 cases in this study is above the 250 cases cut-off that is often used to
determine which fit indices would be most appropriate. Given the lack of
definitive guidelines for the use of specific fit indices, it is best to report the
results of multiple fit measures to allow for the most comprehensive
understanding of model fit. At a minimum, it is suggested that researchers report

12 with the degrees of freedom, and RMR and RMSEA which reflect the residual

differences between the input and implied (model) matrices. These measures
denote the ability of the tested model to predict the matrix covariance (Shah &
Goldstein, 2006).

Initial CFA examination of the proposed model resulted in poor model fit,
so item purification was undertaken. The initial CFA yielded model fit statistics of
12 = 8111.02, df= 3990, CFI = 0.714, SRMR = .072, and RMSEA = .059; none
of these results are within acceptable ranges for model goodness of fit values. In
addition, a number of items failed to achieve acceptable factor loadings (A
values). Items that exhibited the worst fit within the model were eliminated one
by one, based upon the Lagrange Multiplier (LM) evaluation and the

standardized factor loadings. Of the original 92 items, 17 were eliminated. The
resulting measurement model achieved fit statistics of 12 = 2602.20, df = 1363,
CFI = 0.87, SRMR = .052, and RMSEA = .055 (CI: 0.52, 0.58). The difference

between original CFA model and modified CFA model are A 12 = 5508.82 and A

133

df = 2627, demonstrating a substantial improvement in overall fit of the modified
model over the original proposed model.

Because the scales used here have had limited or no testing in previous
empirical research, it was expected a priori that there would be attrition of scale
items within constructs. This is part of the iterative nature of ongoing
improvement of construct validity and scale refinement (Menor & Roth, 2007).
After scale reduction, all but two of the constructs still maintain the needed MV to
LV ratio of not less than 3:1; which is discussed in detail below. The items
removed from their associated constructs are indicated in Table 5.7 and
explained in detail below.

Another important outcome of the examination of the proposed constructs is
the finding that two constructs appear to capture more than a single concept and
should each be split to better represent their intended concept. The construct
related to the occurrence of disruptions, Disruptive Events, required modification.
It was determined that measuring all directions of disruptions within a single
construct was not appropriate. It appears that disruptions in supply (incoming
material and firm output) should be separated from demand disruptions (from
customers). Thus, the more discriminate constructs of Disruptive Events —
Supply and Disruptive Events — Demand are utilized. The use of two separate
supply chain constructs that discriminate between up-stream (supply) and down-
stream (demand) differences is supported in other supply chain research
(Frohlich & Westbrook, 2001). The distinction between the directionality of the

disruption could be because the sample consists primarily of individuals focused

134

on the supply side of their firm’s supply chain; or it may be that there is indeed a
difference in supply and demand disruptions that can not be precisely determined
from the current data and warrants further study. This results in a demand
disruption scale which contains two items.

The second construct that requires amendment is Supply Chain
Complexity. Upon analysis, it appears that the concept of uncertainty which was
thought to be imbedded within complexity is better represented by a separate
construct of Supply Chain Uncertainty. Each modified construct still meets the

minimum criteria for no less than three scales items.

135

Table 5.7 Items Removed from Associated Constructs

 

Supply Chain Complexity
0 Complexity of the products sold to customers (e.g. uses new or frequently changing
technology)
. Complexity in the business environment (e.g. number of competitors

Supply Chain Dependency
0 Our employees have close personal relationships with their supply chain partner
counterparts
. We are economically dependent on our supply chain partners
. Our reputation in the market place is affected by our supply chain relationships

Concern For Disruptions
. Level of concern for market/business environment caused disruptions (e.g.
regulatory issues, currency fluctuations, material cost instability, economic change,
competitor’s actions, etc.)

o In your opinion, how would you rate the vulnerability of your supply chain
0 You hear about your competitors experiencing supply chain disruptions

Effective Practices

. We concentrate on core competencies and outsource some activities (e.g. lS/IT,
facility maintenance, warehousing, etc.)

Profile the Supply Chain
. We have a profile (or map) of our supply chain

Risk Identification
c We know the particular vulnerabilities of our customers
0 We use multiple sources of information to identify supply chain risk

Audit of Risk Management Plans
. We have informal supply chain continuity plans for our supply chain
c We have identified signals that we should watch for that would indicate a disruption
may be starting or about to occur

Risk Management Strategies
Risk Management Practices

Change in Disruption

. FREQUENCY: Change in disruptions from upstream/supplier sources (material
shonages

o IMPACT: Change in disruptions from upstream/supplier sources (material
shonages

o FREQUENCY: Change in disruptions from business environment sources (changes
in technology

. IMPACT: Change in disruptions from business environment sources (changes in
technology

 

136

Modifications to the original hypotheses are appropriate, as the basic
underlying relationships between supply chain complexity, uncertainty, planning,
the occurrence of supply chain disruptive events, and firm performance are
maintained, although they are now examined with a greater degree of specificity
through the use of the modified and added constructs. All further analysis of the
constructs, model fit and path strength reflect the alteration of the original
constructs and hypotheses. Specifically, the hypotheses incorporate the

following changes:

H2a: Higher levels of perceived supply chain complexity are positively
associated with greater frequency and impact of disruptive events

H2b: Higher levels of perceived supply chain dependency are positively
associated with greater frequency and impact of disruptive events

Both now reflect the truncated construct of supply chain complexity and the

relationship to both disruptive events from supply and demand sources so that,

H2as: Higher levels of perceived supply chain complexity are positively
associated with greater frequency and impact of disruptive events -

sumo/y

H2aD: Higher levels of perceived supply chain complexity are positively
associated with greater frequency and impact of disruptive events —
demand

H2bs: Higher levels of perceived supply chain dependency are positively
associated with greater frequency and impact of disruptive events -

supply
HZbD: Higher levels of perceived supply chain dependency are positively
associated with greater frequency and impact of disruptive events -
demand
The separation of disruptive events into supply and demand constructs also

requires the following alterations of their relationships with other constructs of the

137

use of effective practices, concern for disruptions, the use of planning, and firm

performance.

H5as: Higher levels of the use of effective practices are negatively
associated more frequent and larger impact disruptive events -

supply

H5ap: Higher levels of the use of effective practices are negatively
associated more frequent and larger impact disruptive events -
demand

H5bs: Higher levels of concern for disruptions are negatively associated
more frequent and larger impact disruptive events — supply

H5bp: Higher levels of concern for disruptions are negatively associated
more frequent and larger impact disruptive events - demand

H75: Higher levels of continuity planning are negatively associated with
more frequent and larger impact disruptive events — supply

H70: Higher levels of continuity planning are negatively associated with
more frequent and larger impact disruptive events - demand

H83: More frequent and larger impact disruptions — supply are negatively
associated with firm performance

H80: More frequent and larger impact disruptions - demand are
negatively associated with firm performance

In addition, the construct of Supply Chain Uncertainty which was extracted for the

Supply Chain Complexity construct also requires the addition of the hypotheses

H1 c: Higher levels of perceived supply chain uncertainty are positively
associated with higher levels of continuity planning

H2cs: Higher levels of perceived supply chain uncertainty are positively
associated with greater frequency and impact of disruptive events -

swmw

H2cD: Higher levels of perceived supply chain uncertainty are positively
associated with greater frequency and impact of disruptive events —
demand

The full list of hypotheses as modified and tested is contained in Appendix D:

Summary of Final Revised Hypotheses.

138

Before statistical analysis of construct validity is undertaken, a basic
examination of nomological and face validity should be conducted, especially
given the number of manifest items that have been removed from the data set.
Nomological validity is assessed by investigating the correlations between the
measurement constructs. Constructs that are thought to be related to each other
should have higher correlations than constructs thought to have less of a
relationship. This is important specifically in this model that is both highly
complex and that has several constructs that measure different concepts that are
theorized to be highly related to each other. An examination of the matrix of the
correlations of the constructs reveals that those constructs that are thought to be
related also exhibit higher correlations than those that are thought to be less
closely related (Table 5.8).

The establishment of face validity is a fundamental part of the survey
instrument design, so should occur well before data collection and analysis, as it
did in this study. However, given the number of items removed, it is advisable to
revisit the face validity of the constructs and scales. As items were removed,
careful attention was given to the possibility of changing the underlying
conceptual meaning of the constructs or in the potential for omission of a key
element of the construct. In reviewing the remaining scale items and constructs,
with several risk management researchers, it was determined that face validity

was maintained within the more parsimonious measurement model.

139

Table 5.8 Summary of Fit Metrics of Final CFA Measurement Model

666 v 6 a 28566.6 .

_mcommfi 9: 95.6 Eon E 3583 m_ 6:228 zoom .8 w>< 9: .6 62 96:3 9:.
862me mo:m:m> mm8m>< u m>< 5:596”; 627.600 .6595 u mo. ”29m 9:02:80 u a E26300 2355 u am ”.6802

Ammod .Nmod H_UV mmod n (mmfim H«mod n 1.2mm 58.0" EU ”mom? n gt Nomdoow u N Hzu— Eco—2

 

 

E6 .26-.§6-.mm6 66.6r86 .36 .86 .666 86- 666 B6- 86 86 ~66 6: l6? 85555“. Em .fi
36 .666 .2663- 26-36- 36:26:66.6 .666 666 :6 86 66.6 63 666 newswoacofiamafi
26 :6666- 366.6. 666 36:66.6 .36 :66 86 R6 666 N? 63 2668-26.32.65?
$6 $6 $6 .666 .86 .666 666- 26 36- 666 86 $6 6: t... 668868 .6
$6 $6 36 .36 .666 26- 666 666- 666 666 666 63 66+ 68865 96.85 6
666 .R6 .56 .666 666 .666 .26 :6 $6 86 63 9.6 Ea 662 .5
26 .666 .86 666 .26 S6 86 36 666 .3 8+ “828823.63 6
$6 .86 666 .26 36 $6 666 666 SP 666 868528.me 6
86 666 666 36 86 #6 R6 666 $6 2626 520 2&8 .6
666 .666 .26 $6 :6 S6 66.. 63. 222.85 £20 26666 6
«66 .666 36 R6 66.6 a: :6. 268.658 520 £666 .6
36 66.6 666 E6 63 66... 266662925868 .F
a. 2 2 6 6 N 6 6 e 6 N F m>< mo. 6 am :8: 262.;

 

«3292 .33... Boa—65:50 «5 Eat E .2605. __Eo>o new «occur—goo .mpumzmfi «_mom

140

5.5.3 Construct Validity

Construct validity is the degree to which the proposed theoretical latent
constructs are adequately measured by a set of measurement items. To
evaluate convergent and discriminant validity, the approach outlined by Fornell
and Larcker (1981) is adopted. Convergent validity is determined by examination
of factor loadings, construct reliability and variance extracted. Discriminant
validity is determined through examination of correlations and comparison of the
variance extracted with the squared correlation between constructs (Hair Jr. et

aL,2006)

5.5.3.1 Convergent Validity

Factor loadings (i. values) that are high indicate that the items being

measured converge onto a singly construct. In addition, they should also
achieve statistical significance. Both conditions are necessary, because a
statistically significant factor loading may only achieve relatively low strength. It
is therefore recommended that factor loading be statistically significant and have
a standardized factor loading of 0.50 or higher, with an ideal bottom cut-off of
0.70 (Hair Jr. et al., 2006). Standard factor loadings were used to identify scale
items that did not demonstrate acceptable levels of convergent validity, with all
but two exhibiting loadings above 0.50. The two items that were retained with

values below 0.50 have i. values of 0.46 (change in frequency of upstream

disruptions) and 0.47 (change in impact of upstream disruptions) quite close to

the cut-off value and both statistically significant. They remain in the Change in

141

Supply Disruptions construct for the purposes of this study, but warrant further
refinement in subsequent research.

Following the framework of Fornell and Larcker (1981), the construct
reliability (CR), another indicator of convergent validity, is calculated based on
the following equation. Here i. is equal to the standardized loading between an
observed variable and the latent construct of interest and 2 represents the

standardized error of the estimate.

(202

 

(2).)2 + z Var 8

Construct reliability is considered a better measure than Cronbach’s coefficient
alpha, which is highly sensitive to sample size and number of items in the
construct. Reliability estimates greater than .70 are considered to be acceptable
(Nunnally, 1978), although lower estimates are acceptable, provided other
indicators of construct validity are good (Hair Jr. et al., 2006). Cronbach’s alpha
is shown for all factors, but we rely on CR. All of the constructs in this study
achieve a CR 2 0.70, with the exception of Level of Concern for Disruptions
which has a CR = 0.65 and Change in Disruptions - Supply with has a CR = .55.
While these do not strictly meet the lower bound cut-off value, given the nature of
the research, these values are generally considered acceptable.

A third test of the construct validity for each latent construct is assessed by
calculating the average variance extracted (AVE) for each construct. AVE is

calculated with the following equation:

142

20.2)

AVE: 2
X(X)+2Vars

 

AVE values greater than 0.50 indicate that the variance captured by the construct
is greater than the variance due to measurement error, providing evidence of

convergent validity (Fornell & Larcker, 1981).

5.5.3.2 Discriminant Validity

In order to provide evidence of discriminant validity the shared variance
between any given pair of constructs must fall below the average variance
extracted for the constructs (Fornell & Larcker, 1981). This would indicate that
the item measures better explain the latent construct with which it is associated
than it does another latent construct (Hair Jr. et al., 2006). The test for this is to
compare the squared correlation estimate between two constructs with the
variance extracted. In order to more easily fit the information into a single
summary table of the CFA test statistics, the square root of the variance
extracted is included along the diagonal of Error! Reference source not

found.Table 5.8.

5.6 Reliability and Validity Summary

Examination of the sample data is necessary to confirm that the basic
underlying assumptions required for multivariate statistical testing are met before
more advanced data analysis can begin. These include assessing the
demographics of the respondents is an appropriate representation of the target

sample and allows for generalizability of the results. Some missing data is to be

143

expected, but the levels of missing data by item and case should be low and with
a random pattern. In addition, the collection process should not introduce non-
response bias and should control for common method bias. The data should
exhibit a normal distribution, or data transformation may be necessary to carry
out required analysis.

The data collected was obtained from a qualified target sample of 3,199
supply chain professionals. Of these, 303 responded to the survey, for a
response rate of 9.5%. Examination of the demographics of the respondents
indicates that they are all involved in supply chain management from a wide
variety of industries and from a range of firm sizes, as shown in Table 5.1 and
Table 5.2. This supports the generalizability of the results, since no particular
industry or size of firm is disproportionally represented in the data.

Although the sample size obtained is adequate to undertake the data
analysis techniques to be used, it is still important to ensure completeness of the
data examination investigation of missing data. There are low levels of data
missing by item and by case, although five cases were found to have more than
10 percent of the responses to items missing and were therefore eliminated. The
remaining sample was tested and determined to have data missing completely at
random (MCAR) with no underlying bias in the data. This allowed for imputation
of the missing data using an expectation-maximization (EM) algorithm which
allows for the best representation of the original distribution with the least bias
(Hair Jr. et al., 2006). With a complete data set, it was determined that there

were no outliers, and that the data should be inspected for normality. Levels of

144

both kurtosis and skewness were found to be below the acceptable cut-off levels,
indicating the data has a normal probability distribution. It was determined via an
ANOVA test of early and late responders that there was no evidence of non-
response bias. Finally, the study and the data collection instrument were
carefully designed to take into account the risk of common method bias. An
inspection of the data using Harmon’s one factor test indicates that common
method bias is not of concern within the data.

The testing undertaken on the sample data indicates that it is possible to
proceed with further data analysis. This first analysis undertaken was a two-step
confirmatory factor analysis as prescribed by Anderson and Gerbing (1988). The
results of the CFA indicated that modification of several constructs. Specifically,
the construct of Supply Chain Complexity, which contained elements of
uncertainty, should have these items removed to form a new construct of Supply
Chain Uncertainty. In addition, the construct of Disruptive Events which
originally combined both supply and demand disruptions should be separated
into two distinct constructs of Disruptive Events - Supply and Disruptive Events —
Demand. The separation of these constructs into more discriminant and refined
concepts reduces the risk of common method bias (Podsakoff, MacKenzie, Lee,
and Podsakoff, 2003).

After the required changes to the constructs and their individual items, the
data was examined for convergent and discriminate validity as prescribed by

Fornell and Larcker (1981). Although Cronbach’s alpha (or) is often used, more

robust tests are available and were used in this study. Specifically, the i. of

145

individual items, calculation of construct reliability and AVE indicate that the data
does posses convergent validity. Further testing demonstrated the presence of

discriminant validity. The results of these tests can be found in Table 5.8

146

6. DATA ANALYSIS

Structural equation modeling (SEM) is the method selected for model
testing. SEM represents a special case of covariance structure modeling that is
used to evaluate an a prion' set of linear relationships (the ‘path model’) between
variables that may be observed (manifest variables) or unobserved (latent
variables). It is used here to examine both the construct measurement and
structural relationship components of the research. There have been several
recent guidelines in focused literature on the application of this powerful
technique, such as strategic management (Shook, Jr., Hult, & Kacmar, 2004),
logistics (Garver & Mentzer, 1999) and international business (Hult et al., 2006).
The recently published guidelines for operations management by Shah and
Goldstein (2006) serve as the skeleton for this study of the steps that should be

completed and the data to be reported.

6.1 Results of Path Model Testing

The results of hypothesis testing for the Full model are provided in Table
6.1, which includes the parameter estimates, level of significance and model fit
statistics (details for the NAT and HRT model are available from the author). In
addition, the graphical presentation of the models and the significant paths are
shown in Figure 6.1, in Figure 6.2 for the NAT Model and in Figure 6.3 for the
HRO model, all of which will be discussed in detail. It is worth reiterating that the
results of the CFA analysis indicated that it some of the originally designed

constructs required modification. Most notably, the separation of two compound

147

constructs into four focused constructs, and the consolidation of a second order
construct. The original Supply Chain Complexity construct was divided into
Supply Chain Complexity (of products, processes, market and supply chain
structure) and Supply Chain Uncertainty (in supply and demand); the occurrence
of Disruptive Events was divided based upon the source of the disruption into
Disruptive Events — Supply and Disruptive Events - Demand. The second order
construct of Continuity Planning was reduced from six elements to four.

The resulting full model consists of three constructs based upon NAT:
Supply Chain Complexity (CM PLX), Supply Chain Uncertainty (UNCT) and
Supply Chain Dependency (DPND); two constructs based upon HRT: Effective
Practices (EFPR) and Concern for Disruptions (CNRN); a reduced second order
construct of Continuity Planning (PLAN); separate constructs to represent
Disruptive Events — Supply (DES) and Disruptive Events - Demand (DED); and
lastly, Firm Performance (FMPR).

The primary purpose of hypothesis testing is to determine the efficacy of
the use of risk management planning in a supply chain to reduce the frequency
and severity of supply chain disruptions, a reduction that is hypothesized to
positively relate to firm performance. Antecedents to continuity planning may be
influenced by a firm’s perceptions that it is vulnerable to disruptions based upon
the supply chain structure in which it operates, or conversely that diligent
management may be able to reduce vulnerability. These concepts are grounded
in NAT and HRT theory, and tested within a full model and individually as

separate models. Based upon the theoretical foundation of NAT, it may be

148

inferred that when a supply chain is complex with high uncertainty and
dependency between supply chain entities, there is little that a firm can do to
prevent the occurrence of disruptions within the supply chain (the occurrence of
‘normal accidents’). Conversely, HRT suggests that with proper concern for
disruptions and the use of effective business practices and continuity planning,
firms may be able to reduce their susceptibility to supply chain disruptions to
move toward higher supply chain reliability.

Overall fits for all three models are marginal, but do provide consistent
insights worth noting. First all three models indicate strong support for the
mediating role of continuity planning, and the positive relationship of Continuity
Planning with Firm Performance (H6). The second notable consistent finding is
the lack of support for the expected relationship between Concern for Disruptions
and the occurrence of Disruptive Events — Supply and — Demand; H5bsp
hypothesized that the relationship would be negative (higher levels of concern
are associated with lower levels of disruptive events) however, in both the Full
and HRT models, there is a significant positive relationship between these
constructs. An important third result was also unexpected, specifically the lack of
a significant relationship between Disruptive Events —Supply and —Demand, and
Firm Performance (H839). All results for the three models are further discussed
in the following sections, including a summary of hypothesis support, path A

values, significance levels and overall model fit in Table 6.1.

149

Table 6.1 Results of H1 to H8 and Model Fit Statistics

 

 

liypothesis Full Model NAT Model HRT Model
H1a: (CPLX —>PLAN) + .12* support .13“ support
H1b:(DPND—+PLAN) + .50** support .64** support
H1c: (UNCT—>PLAN) + .07 .07
H2as:(CPLX—>DE5) + .19** support .28*** support
H2aD: (CPLX—>DED) + .16* support .20*** support
H2bs:(DPND—>DES) + -.04 -.12
H2bD:(DPND—>DED) + .03 -.02
H2cs: (UNCT-6 DES) + .04 .26*** support
H2cD: (UNCT—> DED) + .13* support .20*** support
H3a: (EFPR—+PLAN) + .29" support .58*** support
H3b: (CNRN—+PLAN) + .06 .16*** support
H4a: (EFPR—>FMPF) + .23*** support .25*** support
H4b: (CNRN—+FMPF) + -.03 -.03
H5as: (EFPR—>DES) - -.11 .03
H5aD: (EFPR—>DED) - -.04 -.03
H5bs: (CNRN—’DES) - .33*** reverse .46*** reverse
H5bD: (CNRN—’DED) - .13* reverse .30*** reverse
H6: (PLAN—>FMPF) + .16*** support .29*** support .16“ support
H75: (PLAN—>DES)- -.14 -.03 -.07
H70: (PLAN—+DED)- -.13* support -.14* support -.11
H85: (DES—+FMPF)- -.06 -.08 -.06
H80: (DED—>FMPF)- -.03 -.06 -.04

 

2 2
1(364) = 1274.9 1(217) '—' 801.0

OH = .783, CFI = .807; OH = .839;
RMSEA =.080 RMSEA =.092 RMSEA =.095
Notes: + indicates hypothesis has been modified to reflect changes in constructs

2
1(715) = 2068.6

2
*** p<.01, ** p<.05, * p<.10 Z(df)

 

CPLX = Supply Chain Complexity, PLAN = Supply Chain Risk Planning,

UNCT = Supply Chain Uncertainty, DES = Disruptive Event — Supply,

DED = Disruptive Event — Demand, DPND = Supply Chain Dependency,

EFPR = Effective Practices, CNRN = Concern for Disruptions, FMPF = Firm Performance

150

Within the Full Model shown in Figure 6.1, which includes both NAT and
HRT, statistically significant associations with the use of Continuity Planning are
Supply Chain Complexity (H1 a) and Supply Chain Dependency (H1 b), and the
use of Effective Practices (H3a). This would indicate that when there are higher
perceptions of complexity and entity dependency within the supply chain, firms
are more likely to engage in planning activities meant to manage supply chain
risk. In addition, firms that are engaged in higher levels of a variety of generally
accepted effective practices are also more likely to engage in planning activities.
Interestingly, there is no significant relationship between Supply Chain
Uncertainty (H1c) or Concern for Disruptions (H3b) and Continuity Planning.
These findings provide some support for the efficacy of both NAT and HRT
theory within this context. It would seem that firms can, through effective
continuity management, positively impact disruptive events (support for HRT)
although they will not be able to completely eliminate all disruptive events
(support for NAT).

It is important to note that H5bsp predicted a negative relationship between
the Concern for Disruptions and Disruptive Events - Supply or - Demand. Using
HRT as a basis, the expectation was that firms with greater concern for
disruptions would experience disruptions at a lower rate or with lower magnitude.
However, the significant positive relationship between the two constructs
indicates that the potential cause and effect relationship may be that the
occurrence of disruptive events leads to higher levels of concern for disruptions.

This would seem to indicate that managers should in fact be concerned about

151

disruptive events, since they seem to be experiencing them. In addition, given
the market intolerance for disruptions to the supply chain, it is imperative that
managers learn to better manage their supply chain risk for increased supply
chain continuity. Although correlation can be determined, causality can’t with the
cross-sectional data used here. Future research utilizing longitudinal data may
allow for more definitive conclusions regarding the relationship between the
occurrence of, and concern for, disruptive events.

Continuity Planning is significantly related to lower levels of Disruptive
Events -Demand (H70) but not Disruptive Events -Supply (H73). There is also
strong support for the relationship between engagement in Continuity Planning
and higher levels of Firm Performance (H6). Surprisingly, there is no support
that Disruptive Events - Supply or - Demand are associated with lower levels of

Firm Performance (H859).

152

Figure 6.1 Full Model Significant Paths

Supply Chain
Complexity

Uncertainty
Dependency
Firm

 

Continuity
............. . Planning
Profile Risk ID.._': 6. Disruptive
'- ............. . 5:. ......... __ ,,,,,,,,,,,, -' v Events -Demand
{Assess Z_ Plan /

/ /
/ / Events -Supply
Effective / / / /

Practices / /
/

Concern for — significant

Disruptions — -> significant opposite of hypo
non-significant (not shown)

Within the NAT Model shown in Figure 6.2, there is support for several of
the hypotheses grounded in NAT. Consistent with the Full model, both Supply
Chain Complexity (H1a) and Supply Chain Dependency (H1 b) are significantly
positively associated with Continuity Planning while Supply Chain Uncertainty
(H1c) does not have a significant relationship. Also consistent with the Full
model, Supply Chain Dependency does not have a significant relationship with
Disruptive Events - Supply or - Demand (H2bs and H2bD). In a slight difference
from the Full model are the findings that higher perceptions of Supply Chain

Complexity (H2as_D) and Supply Chain Uncertainty (H2csp) which exhibit a

153

stronger positive associated with higher levels of both Disruptive Events - Supply
and — Demand disruptions. This is an interesting finding because a key element
of NAT is the importance of dependency, which does not exhibit a significant
association in the hypothesized models.

The relationships between other constructs in the Full model and NAT
model are also similar. Continuity Planning (H6) has a significant positive
relationship with Firm Performance, a weak significant negative relationship with
Disruptive Events — Demand (H70) and no significance with Disruptive Events -
Supp/y (H75). In addition, there is no significant relationship between Disruptive
Events - Supply or — Demand (H839) and Firm Performance.

Figure 6.2 Normal Accident Theory Model Significant Paths

Supply Chain
Complexity

Supply Chain
Uncertainty
Supply Chain
Dependency Firm

L ------- Performance

Continiuity 'E'm Disruptive
------------ .Plam mg .-- : Events - Demand
Profile 2- .......... Risle.‘ E
if???“ .--..‘.’.‘.a.?.. ........................ ‘ Disruptive
Events - Supply

—> significant
....... > non-significant

    

 

 

154

A comparison of the Full model with the High Reliability Model in Figure 6.3
highlights consistent findings for most of the hypothesized paths, with a few
notable exceptions. Effective Practices (H3a) has a stronger positive association
with Continuity Planning than in the Full model, and Concem for Dismptions
(H3b) has a strong positive association not seen in the Full model. Confirmation
of these hypotheses provides support for the efficacy of HRT in this context.

Also similar to the Full Model, the HRT Model achieved significant positive
associations between Concern for Disruptions and Disruptive Events — Supply
and -Demand, a finding contradicting the hypothesized negative relationship
(H5bs,o)-

Figure 6.3 High Reliability Model Significant Paths

  
    
 

  
   

Firm
Performance

  
 
 

 

Continuity
............... Planning
Profile Risk ID

"""""""""""""
.................
OOOOOOOOOOOOOOOOO

Assess! Plan

  
  

 
    

 
   

   
 

 

 
   

  
 

 
     

 

; Disruptive
........... Events-Demand
............. /§
..... .. / s
......... /
............ /.,..
................. ....... .1 ......
PractIces ' Events - Supply
/ / ’

0°... /
0.. / /
Concern for / ’ . .
Disruptions _’ SlgnIficant
------- > non-significant

— -> significant opposite of hype

155

In discussing the results of model testing, there are a few salient points
worth mentioning. The most important finding is that across models, there is a
significant positive relationship between Continuity Planning and Firm
Performance. This relationship is strongest in the NAT model that does not
include any other constructs that could positively impact firm performance.
However, in the HRT model, the relationship between Effective Practices and
Continuity Planning is stronger. This may indicate that activities undertaken to
reduce and manage supply chain risk are part of a set of effective business
practices.

Another important finding, that clearly supports NAT, is the strength of the
relationship of Supply Chain Dependency and Continuity Planning. In fact, this is
the strongest relationship in the models. Although Supply Chain Complexity is
also important, it appears to be clear that firms understand the basic premise of
NAT, which is that it is not the action (or inaction) of a single operator in the
chain, but the interactions between entities when an event occurs that cause

‘accidents’ to happen and move through a system.

6.2 Research Implications

The proposed models used in this study achieve marginal fit statistics, but
are still useful given the nascent level of research on risk management planning
within a supply chain setting. In addition, the use of NAT and HRT has rarely
been undertaken in supply chain research, so constructs that have been
rigorously examined in multiple studies and contexts were not available.

Therefore, this research makes several contributions to the field: 1) starting the

156

process of construct creation and validation, 2) providing insights that could be
used to examine and refine the role of complexity, uncertainty and dependency,
and 3) examining the role of continuity planning and other effective practices in
reducing events that disrupt supply chain functionality and that likely impact
overall firm performance.

The findings here provide support for the utility of both NAT and HRT in
explaining factors that influence firms to undertake planning efforts to reduce
supply chain risk. In addition, the data provide insight on the impact of planning
efforts on the occurrence of disruptions and firm performance. Consistent with
NAT, firms with higher perceptions of supply chain complexity and dependency
engage in higher levels of continuity planning. Within this study, supply chain
complexity includes structural elements (number of internal divisions and
facilities, production facilities and warehouse, and external links of suppliers and
customers), geographic dispersion and multiple tiers of nodes. Supply chain
dependency includes elements of time-based synchronization, knowledge of
partner activities, awareness of partner strengths and weaknesses, alignment of
activities, the exchange of knowledge and investment in information exchange.
Interestingly, uncertainty of supply availability and demand variation does not
have a significant impact on planning. It is possible that uncertainty is accepted
as a ‘normal’ state with a perception that supply chain managers have little ability
to take action to influence a reduction in uncertainty.

There is also support for HRT through the strong positive relationship of

both concern for disruptions and the use of effective business practices on

157

engagement in supply chain risk management planning activities. Within this
study, effective business practices were limited to those that have been used in
multiple recent studies and that are less ubiquitous than TQM, Lean or JIT.
Those practices identified for inclusion here incorporate activities related to new
product development, sharing high quality information, utilization of customer
relationship management, interdepartmental cooperation and strategic supply
chain management. While it can be argued that risk management may be an
effective business practice, the analysis here indicates that there is reason to
measure these constructs separately.

As expected, the use of a wide variety of effective business practices is
positively and strongly associated with higher levels of firm performance,
although not with a reduction in disruptive events from either supply or demand
sources. An important finding is that the anticipated negative relationship
between concern for disruptions and the occurrence of disruptions was shown
instead to be significantly positive. This raises questions about the link between
these constructs. Using HRT, it is expected that firms that are more concerned
about reliability would undertake activities to reduce the occurrence of events
that would negatively impact continuous supply chain operation. However it
appears that more frequent and larger magnitude of supply chain disruptions are
associated with elevated levels of concern. It is important to note that SEM
analysis does not allow for conclusive cause-and-effect relationships, because it
is based upon covariance relationships which do not incorporate a temporal

element.

158

Table 6.2 Summary of Research Implications

 

Issue

Impact

Recommendation

 

Few previously defined
or tested constructs.

Little previous
examination of the role
of complexity,
uncertainty and
dependency on supply
chain risk.

Little previous
examination of the use
of specific risk
management practices
utilized by firms to
improve supply chain

performance continuity.

Many of the constructs used here Creation and initial testing of

were created from a review of
salient literature. Although great
care was taken in designing and

pre-testing the constructs, several

required modification.

NAT theory states that complex,

uncertain and dependent systems

are prone to ‘normal accidents’.
Supply chains exhibit these

attributes and are therefore prone

to failure. However, it is unclear
the degree to which these
attributes are related and the
impact on disnIptive events.

There is wide variety in the

practices firms can choose to use

specifically to manage risk, but
little research to determine which

are most effective in reducing risk

based upon the business
environment and types of risk
faced.

There may be a possible It is possible that there is an

linkage with risk
management practices
and other general
‘effective’ business
practices.

Neither NAT nor HRO
have been frequently

interaction effect between
practices that are ‘general’ and
that are specific to risk
management.

Evidence of efficacy for the
competing theories of Normal

utilized in operations or Accidents and High Reliability

supply chain research.

There is a high
correlation between
concern for disruptions
and the occurrence of
disruptive events.

The study reveled results the
opposite of expected regarding

the relationship between concern

for and the occurrence of
disruptive events.

constructs in this research can
serve as the basis for further
definition and refinement of
relevant constructs.

Refined and tested definitions of
complexity, uncertainty and
dependency within a supply
chain context are needed. In
addition, a deeper
understanding of the role of
complexity, uncertainty and
dependency within a supply
chain context is needed.

Further examination of risk
management practices in a
variety of business setting is
needed, perhaps utilizing
Contingency Theory.

Examination of the role of
specific risk management
practices and general ‘effective’
business practices is needed to
capture their specific
contribution and their combined
contribution to firm performance.

Further testing of these theories
is needed to better understand
their usefulness in explaining
risk management in supply
chains; or to develop a unifying
theory.

More research is needed to
better understand the potential
causal relationship between
these two constructs.

 

159

6.3 Managerial Implications

Disruptions from multiple sources and across industries can lead to sizable
market penalties for firms that don’t manage this important business area well.
The reason or cause of a supply chain disruption is not a critical factor; the
market expects an uninterrupted flow of goods and services. In addition to the
market and financial consequences that a firm may incur, there can also be a
detrimental impact on employees, suppliers, customers and other supply chain
stakeholders. Unfortunately, it is also possible that the efforts undertaken by
firms to create leaner supply chains, increase efficiency, and reduce costs may
have unintended consequences of creating supply chains that are more
vulnerable to disruptions because of the elimination of safety buffers.

Although the management of the unexpected can be a difficult challenge for
business to contend with, supply management professionals do not perceive that
they must accept supply chain disruptions as part of doing business. There is a
clear perception that they can positively affect both the frequency and impact of
interruptions to their supply chains through proactive risk management, even if
they are in supply chains that are complex and have uncertainty of supply or

demand.

6.3.1 Role of Supply Chain Complexity, Dependency and Uncertainty

In general firms that reported supply chains that were more complex, had
higher levels of supply and demand uncertainty, and greater dependency on

supply chain partners experienced more supply chain disruptions. Supply chain

160

complexity includes greater number and larger geographic dispersion of supply
chain nodes, both of which influence the length and width of the supply chain.
Interestingly, unlike the complexity of the supply chain structure, the complexity
of the products does not appear to have a significant impact. This complexity is
also associated with higher levels of uncertainty in both the supply of incoming
material, output for internal sources, and demand from customers. Both supply
chain complexity and uncertainty are correlated with higher levels of concern for
disruptions, higher perceptions of supply chain vulnerability to disruptions and to
a higher actual occurrence of disruptions.

The correlation of supply chain complexity and the occurrence of supply
chain disruption may be influenced by the industry in which a firm operates.
Firms that are more complex also report more frequent knowledge of their
competitors experiencing disruptions. The data in this study supports the
premise that firms that seek to manage supply chain risk, may be able to create a
source of differentiation and competitive advantage, as well as mitigate or avoid
negative financial consequences (Norrman & Jansson, 2004; F ilbeck et al., 2005;

Hendricks 8. Singhal, 2005a; Hendricks & Singhal, 2005b).

6.3.2 Formal versus Informal Risk Management Plans

There was much higher incidence of the use of informal risk management
plans, but individuals that reported a higher usage of formal plans also had
significantly higher levels of firm performance. Formal plans were also

substantially more likely to undergo regular internal review and, more importantly,

161

more likely to be shared with both up and down-stream supply chain partners.
The usage of formal plans is more effective than informal plans in improving firm
performance, especially if formal plans are reviewed internally, undergo testing
via “fire drills’ and are regularly shared with both internal and external supply
chain participants. It would appear that the very process of creating, sharing and

maintaining formal plans places greater focus on this critical business activity.

6.3.3 Profiling the Supply Chain

A key first step in the process of supply chain risk management is having an
up-to-date and comprehensive profile (or map) of the supply chain. Having an
accurate profile of the supply chain is correlated with knowledge of the unique
vulnerabilities of suppliers, important transit lanes, a firm’s own facilities, and
customers. It could be that the very process of explicitly identifying a firm’s
supply chain and the relationships between the nodes and the transit lane
linkages illuminates the vulnerabilities and potential sources of risk. The explicit
examination and mapping of the supply chain for sources of risk has a high
correlation to the ability of the firm to also assess the probability and potential
magnitude of a disruption, from both a quantitative and qualitative impact. An
additional potential advantage of examining the supply chain and the node
linkages is the ability to identify and utilize multiple sources of information that
could provide ‘signals’ of impending or unfolding interruption to normal supply

chain functionality.

162

(A

El

6.3.4 Risk Reduction Techniques

The simplest and easiest to implement forms of buffering, safety stock and
supply flexibility (using at least two suppliers and having a cross-trained
workforce), are the most commonly used supply chain risk reduction techniques.
In keeping with JIT and lean practices, pushing safety stock back on to suppliers
is used more frequently than a firm holding their own buffer stock. It is important
to note that while the use of multiple suppliers may provide some protection from
an interruption in supply should a single supplier fail to provide needed material
or service, multiple suppliers also equates to variation in inputs, which in turn
may lead to problems with internal conversion processes.

Other practices that may be used to manage risk, such as material
standardization or the postponement of final product configuration are less
commonly used. These practices require greater effort to implement, due to the
need for cooperation across functional areas within the firm (product engineering
for re-design, procurement for changing material requirements, manufacturing
engineering for process re-design, etc.) and well and cooperation outside the firm

to suppliers and customers.

6.3.5 Importance of Effective Business Practices

Firms that engage in a wide variety of ‘good’ business practices, even if not
explicitly undertaken to manage supply chain risk, are less likely to experience
supply chain interruptions. There appears to be some interaction affect that
makes enhances firm performance when firms have higher levels of both general

effective business practices and specific risk reduction activities. Both externally

163

focused and internally focused practices appear to be important. Externally,
firms that report a higher commitment to providing accurate, timely and credible
information to their supply chain partners and to the development of mutually
beneficial partnerships are more successful in managing risk. Sharing
information, even if not specifically undertaken for risk management, is positively

correlated with fewer supply chain disruptions.

164

Table 6.3 Summary of Managerial Implications

 

Issue

Recommendation

 

Publically known supply chain disruptions
result in:

. 6.9% I in sales growth

0 10.7% T in cost

0 13.9% T in inventory

0 13.5% T in equity risk

Firms should utilize a general framework for

developing firm specific risk management
plans.

More frequent and impactful disruptions are

associated with supply chains that have:
0 T complexity

o T uncertainty

. T dependency

Focus on efficiency, lean and JIT activities

often results in:

0 Removal of buffers

0 Increased susceptibility to supply chain
disruptions

Informal risk management plans are less
effective because they:

0 Are less comprehensive in scope

. Undergo less periodic review & testing

. Are less likely to be explicitly shared with

relevant parties

Potential positive and cumulative interaction
effect between general ‘effective' business

and risk management practices

0 Supply chain disruptions should not be
accepted as a ‘normal’ state of
business

0 Firms must undertake risk
management to avoid costly disruption
penalties

Effective risk management planning
includes:

Profiling the current supply chain
Identifying potential sources of n'sk
- Assessing potential risk impact
Preparing formal plans

Sharing plans with partners

Comprehensive assessment of supply
chain complexity, uncertainty and
dependencyto:

0 Identify sources of vulnerability

. Create appropriate solutions

Firms need to fully understand their
processes and sources of risk, including:
. Material flows

. Process and entity dependencies

o Judicious use of buffers

Firms should invest in the process of
creating, sharing and maintaining formal
plans, rather than relying on informal
plans

Firms should not discount the impact of
risk management as a subset of effective
business performance management

 

165

6.4 Discussion

This study examines the risk management planning activities that firms may
undertake in an effort to reduce vulnerability to supply chain disruptions and to
create supply chains that have greater reliability in the flow of goods, services,
information, and funds through the supply chain for improved firm performance.
Building upon limited previously published theory driven empirical research within
this topic area creates both challenges and opportunities. Most notably the need
to examine multiple literature streams related to complexity, uncertainty,
dependency, operational practices, and risk management. The advantage of
searching outside traditional literature streams is the ability to incorporate ideas
from a broader span of thought. In addition, it allows for the creation of
constructs that incorporate a more holistic view of concepts. Notably, this study
has synthesized several streams of literature to create a framework for supply

chain risk planning.

6.5 Research Questions

In undertaking this study, several research questions were identified to
drive the research. The questions are restated below, along with the findings.

1. What factors drive firms to engage in supply chain continuity planning?

Firms that perceive they operate in supply chains with higher levels of
complexity and dependency are more likely to engage in planning efforts to
reduce the occurrence and frequency of supply chain disruptions. Uncertainty

in supply and demand has no significant relationship with planning. In addition,

166

 

firms that engage in a wide variety of effective business practices are more

likely to also engage in continuity planning.

2. How are firms undertaking the creation of supply chain continuity plans
(process)? What are firms including in supply chain continuity plans
(content)? What are the ‘best practices’?

Firms appear to have a clear picture of the locations of their suppliers,
customers and their own facilities, but less so of their transit lanes. They also
believe they have a good understanding of the vulnerabilities of suppliers and
their own facilities, and consider vulnerabilities in selecting new or evaluating
current suppliers; again there is a lack of understanding of transit lane
vulnerabilities. Although the vulnerabilities are fairly well identified, firms do not
perceive there is adequate knowledge of the probability or impact of different
types of potential supply chain disruptions. Firms are more likely to engage in
informal planning, although formal plans are more highly correlated to a
reduction in disruptions. In general firms feel they are able to manage their
supply chain risk well, and use the practices of cross-training and supplier

agreements most often.

3. To what extent are firm ’3 engaged in supply chain continuity planning
upstream, intemally, downstream or some combination of directions?

The results of this study indicate greater engagement in planning activities
that are focused internally and upstream to suppliers, rather than externally
towards the business environment or downstream to customers. This result is

not unexpected, since the survey respondents consist primarily of professionals

167

in the procurement area of their firm’s which would likely influence the results.
However, since supply chain disruptions can occur from a wide variety of
sources, further research utilizing professionals with a differing or wider supply

chain focus would be beneficial.

4. Does increased supply chain continuity planning result in a reduced
frequency and impact of disruptive events and in improved firm
performance?

There is little conclusive evidence that planning directly affects the
frequency or impact of disruptive events, a disappointing result. There is
however significant support for the significant positive relationship between
higher levels of planning and firm performance. Given the strong positive
relationship between planning and firm performance and the use of effective
practices and firm performance, there maybe an interaction effect between
planning and effective practices that impacts firm performance. This inability to
provide support for causality is a limitation of cross-sectional research, and an
area for future study. It is also possible that, similar to the Sand Cone Model of
capabilities (Ferdows & Demeyer, 1990), these create a stronger influence

when combined.

5. Does supply chain complexity, uncertainty or dependency influence the
processes and effectiveness of supply chain continuity planning processes
or results ?

Higher levels of supply chain complexity and uncertainty are associated
with higher levels of supply and demand disruptions, and there is no significant

association from dependency. However, the strong positive relationship between

168

planning and firm performance indicates there is potential for planning to mitigate

the impact of the supply chain environment.

6.6 Future Research

Future research in the area of supply chain risk management has several
opportunities for further development. These areas include the continued
refinement of constructs that examine the supply chain structure and
environment in areas that may impact the vulnerability of the supply chain to
disruptive events. Specifically, constructs to capture the structural complexity of
the supply, the dependency that exists between nodes, and the uncertainty in
supply and demand need to be advanced and validated through multiple
studies. Conceptual work has begun in these areas, but theory building and
subsequent testing will require robust constructs.

In addition, there is very limited work that tests the planning activities that
firms do and should engage in for improved supply chain resiliency and firm
performance. There is a great deal of prescriptive advice provided to
professionals through a wide variety of trade publications and main stream
books. While there are many anecdotal stories that serve as examples of
superior or inferior responses to supply chain disruptions, there is little empirical
testing of the use of specific processes and their outcomes.

This study utilized procurement professionals; future research should entail
gathering data from a broader set of supply chain professionals, including
operations and supply chain managers and individuals with greater knowledge

of downstream supply chain functionality. By gathering information from supply

169

chain management professionals in a wider variety of functional areas, it will be
possible to compare and contrast the perceptions of risk and risk management
from the perspective of a broader range. In addition, the results will be more
generalizabile with less reliance on the perspective of individuals that are
primarily supply or up-stream focused.

Another important extension of the current study would be undertaking a
longitudinal study to determine the relationship of risk management planning at
a point in time on the occurrence of disruptive events and firm performance
from that point forward. The current cross-sectional study does not allow for
evidence of a cause-effect relationship between risk management planning
activities and the influence the activities may have on reducing disruptive
events or improving firm performance. The ability to more definitively establish
such relationships through a longitudinal study would provide further insight into
the specific role of risk management planning activities, along with other
effective business practices, on the frequency and magnitude of disruptive
events and on firm performance.

Unfortunately, such a study was not possible within the limitations of this
research. A longitudinal study that measures such complex business practices
would require a time difference between measurement points of at least one
year in order to allow the results of planning activities to take effect. In
addition, it is possible that disruptive events that have a large enough impact to
be memorable to supply management professionals do not occur frequently, so

allowing for a longer time differential between initial and subsequent data

170

gathering points will help to increase the likelihood that a difference in the

occurrence of disruptive events is substantial enough to be captured.

171

Appendix A: Copy of Survey

(items reduced from original size to meet dissertation page formatting requirements)

ER’04
9:3 H 43’:

f .

     
       

  

\ll( .'I ll( i.-\\l Sl-\lt
I*\le RsIII

 

Instructions for the survey participant

Thank you in advance for your participation in this study examining supply chain risk management
and continuity planning.

The enclosed survey should take approximately 20 minutes or less to complete. Your participation
is completely voluntary and we greatly appreciate your support. The knowledge of professionals
such as your self is vital to find ways to improve business performance.

As you take the survey, please keep the following in mind:

0 All individual response will be kept strictly confidential. Only the survey team will have access
to individual responses; all reports will summarize the data and report findings in aggregate.

. If you are unsure of the exact answer to a question, please provide your best estimate.

0 If possible, please try to answer all the questions. The more of the questionnaire you answer,
the better able we are to gather meaningful information and find trends. However, if you are
uncomfortable with a specific question, you may leave it blank.

0 As you respond to the questions, please base your answers on your knowledge of a supply
chain for a specific product line/family you understand well.

0 On the last @page of the survey, you will have the opportunity to register for the drawing for one
of five Visa Gift Cards of $50 each.

Again, thank you for helping us to better understand how firms manage supply chain risk and in
finding ways to better manage this important business issue.

If you have any questions or comments about this study, please feel free to contact us at:

Andrea Prud’homme Rich Hough, CPP
Michigan State University American Purchasing Society
(517) 432-5535 x272 support@american-purchasing.com

SC_Risk@bus.msu.edu

You indicate your voluntary agreement to participate by completing and returning the survey. All
individual responses will be kept strictly confidential and will not be shared with any individual or
organization beyond the research team. All analyses, reports and presentations of the data
generated from this study will be reported at an aggregate level. If you have questions or
concerns regarding your rights as a study participant, you may contact - anonymously if you wish -
Dr. Peter Vasilenko, Chair of Behavioral & Social Science Institutional Review Board by phone
(517) 355-2180, fax (517) 432-4503, or e-mail irb@msu.edu

172

You will be asked questions about supply chain disruptions and the management of supply chain risk
that is meant to reduce disruptions. For the purposes of this survey, a supply chain disruption can be
small or large and Is movent that reduces, delays or stops the flow of goods through the
supply chain or that requires extra resources to maintain or resume the flow of goods when an
event occurs.

Supply chain focus: When answering the questions in this survey, it is important to focus on a
specific supply chain that you have good knowledge about. Please indicate in the space below the
product line or product family that you will be thinking about as you answer the survey questions:

 

(this question is meant to focus your responses and will not be used to specifically identify you or your firm)

NONE moderate HIGH

 

1. Level of concern for market / business environment caused disruptions

 

(e.g. regulatory issues, currency fluctuations, economic change, 1 2 3 4 5 6 7
competitor’s actions, material cost instability, etc.)
2. Level of concern for upstream / supplier caused disruptions (e.g. late 1 2 3 4 5 6 7

deliveries, inability to meet demand, quality issues, material allocation, etc.)

3. Level of concern for internally caused disruptions (e.g. interruption or delay 1 2 3 4 5 6 7
in production output due to quality, equipment or personnel issues; inventory
issues; capacity shortage; etc.)

4. Level of concern for downstream / customer caused disruptions (e.g. 1 2 3 4 5 6 7
changes in product requirements, changes in timing, volume or mix of
demand, etc.)

 

 

 

Extremely Extremely
LOW moderate HIGH
1. Complexity of the products sold to customers (e.g. uses new or frequently
changing technology, large bill of material, quality specifications that are 1 2 3 4 5 6 7
difficult to meet, several different versions or options of products, etc.)

 

2. Complexity in the business environment (e.g. number of competitors,
different customer niches, dynamic and frequently changing conditions, etc.)

 

3. Overall complexity of the supply chain structure in terms of the number of
production or warehouse facilities where your product is built or stored

 

4. Overall complexity of the supply chain structure in terms of the number of
different divisions, business units, departments, or managers of your firm 1 2 3 4 5 6 7
that are involved in your supply chain

 

5. Overall complexity of a wide supply chain with numerous suppliers,
customers, distribution points, or multiple channels

 

6. Overall complexity of a long supply chain with multiple tiers or echelons of
suppliers and customers.

 

7. Overall complexity in terms of the geographic dispersion (regionally,
nationally or globally) of suppliers, production or warehouse facilities, 1 2 3 4 5 6 7
customers, or distribution channels.

 

8. Overall uncertainty of supply of material (e.g. uncertainty in availability,
quality or lead time of purchased items)

 

9. Overall uncertainty of output (e.g. variable quality, volume or timing of output
from production)

 

10. Overall uncertainty of demand (e.g. variation in timing, mix or volume of
customer demand; customers changing product requirements)

 

173

Please indicate the level to which you agree or disagree with the following statements as they
apply to your supply chain.

 

 

Strongly Strongly
DISAGREE neutral AGREE
1. We know our suppliers, including their locations and inter-relationships 1 2 3 4 5 5 7
within our supply chain
2. We know our customers, including their locations and inter-relationships 1 2 3 4 5 5 7

within our supply chain

 

3. We know our production and warehouse facilities, including their locations 1 2 3 4 5 5 7
and inter-relationships within our supply chain

 

4. We have identified the transit lanes used to ship our material and products 1 2 3 4 5 5 7

 

 

 

 

We have a profile (or map) of our supply chain 4 7
Strongly Strongly
DISAGREE neutral AGREE
1. We know the particular vulnerabilities of our suppliers 1 2 3 4 5 5 7
2. We know the particular vulnerabilities of the primary transit lanes for 1 2 3 4 5 5 7

incoming material and inventory

 

3. We know the particular vulnerabilities of our primary production points and 1 2 3 4 5 5 7
our warehouses

 

4. We know the particular vulnerabilities of our customers 1 2 3 4 5 5 7

 

We know the vulnerabilities of the primary transit lanes used for outbound 1 2 3 4 5 5 7
inventory and products

 

6. We explicitly consider vulnerabilities when selecting new supply chain 1 2 3 4 5 5 7
partners

 

7. We explicitly consider vulnerabilities when evaluating current supply chain 1 2 3 4 5 5 7
partners

 

We use multiple sources of information to identify supply chain risk

 

We have identified signals that we should watch for that would indicate a 1 2 3 4 5 5 7
disruption may be starting or about to occur

 

Strongly Strongly
DISAGREE neutral AGREE

1. We have a process for assessing the probability or frequency of different 1 2 3 4 5 6 7

types of supply chain disruptions

 

2. We have a process for assessing the magnitude or impact of different 1 2 3 4 5 5 7
types of supply chain disruptions

 

3. We have a process to capture the financial impact of a supply chain 1 2 3 4 5 5 7
disruption (e.g. cost of expediting, late penalties, overtime, lost sales, etc.)

 

4. We have a process to capture the qualitative impact of a supply chain 1 2 3 4 5 6 7
disruption (e.g. loss of good will, damage to firm reputation, impact on
supply chain relationships, etc.)

 

174

A supply chain continuity plan is used to identify, manage and reduce the frequency and impact of
disruptions that may occur to the supply chain

 

 

 

 

 

 

Strongly Strongly

DISAGREE neutral AGREE
1. We have formal supply chain continuity plans for our supply chain 1 2 3 4 5 5 7
2. We have informal supply chain continuity plans for our supply chain 1 2 3 4 5 5 7
3. We review our supply chain continuity plans internally 1 2 3 4 5 5 7
4. We have ‘fire drills' to test our supply chain continuity plans 1 2 3 4 5 5 7
5. We periodically review supply chain continuity plans with our suppliers 1 2 3 4 5 5 7
6 1 2 3 4 5 6 7

. We periodically review supply chain continuity plans with our customers

 

Please indicate if there has been any change in the FREQEUNCEY and IMPACT of supply chain
disruptions in the following areas over the past 12 months:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FRE ENCY IMPACT
LESS MORE much much
often same oftenlSMALLER same LARGER
1. Change in disruptions from business environment
sources (changes in technology, government 1 2 3 4 5 6 7 1 2 3 4 5 6 7
regulation, currency fluctuations, changing
commodity prices, competitor actions, etc.)
2. Change in disruptions from upstream/supplier
sources (material shortages, poorsupplier 1 2 3 4 5 6 7 1 2 3 4 5 6 7
delivery or quality performance, in-bound
logistics, etc)
3. Change in disruptions from internal sources
(insufficient capacity, quality problems, leailure,1 2 3 4 5 5 7 1 2 3 4 5 5 7
personnel issues, etc.)
4. Change in dismptions downstream/customer
sources (out-bound logistics issues, changes in 1 2 3 4 5 5 7 1 2 3 4 5 5 7
mix, timing or volume of customer demand, etc.)
Significantly Significantly
LOWER same HIGHER
1. Compared to ourcompetitors, our sales are 1 2 3 4 5 5 7
2. Compared to our competitors, our profitability is 1 2 3 4 5 5 7
3. Compared to our competitors, our overall competitive position is 1 2 3 4 5 5 7
4. Compared to ourcompetitors,our return on assets (ROA) is 1 2 3 4 5 5 7
5. Compared to our competitors, our return on sales (ROS) is 1 2 3 4 5 5 7

 

175

In the next two sections you will be asked questions regarding the general types of STRATEGIES
that your firm may use to manage supply chain risk. You will also be asked questions about some
specific PRACTICES that could be used to implement general strategies. This list is po_t meant to
imply that all firms should use any or all of these strategies or practices.

Strongly 35009117
STRATEGIES DISAGREE neutral AGREE

1. There is little we can do to reduce supply chain disruptions, so we accept 1 2 3 4 5 6 7
them as part of doing business

 

 

2. In general, our supply chain risks are well managed 1 2 3 4 5 6 7

 

3. We utilize a strategy of sharing supply chain risk with our supply chain 1 2 3

partners (e.g. buy back agreements, cost or revenue sharing, etc.) 4 5 6 7

 

4. We have strategically transferred most of our supply chain risk to supply
chain partners (e.g. vendor managed inventory, contractual performance 1 2 3 4 5 6 7
agreements, etc.)

 

5. In general, we try to reduce supply chain risk as much as possible 1 2 3 4 5 6 7

 

6. We have been able to pool supply chain risk (e.g. use of buyer
consortiums, centralized warehousing, trans-shipments, postponement or 1 2 3 4 5 6 7
delayed finalization of finished products, etc.)

7. We have been able to eliminate most of our SC risk 1 2 3 4 5 6 7

 

 

Please indicate to what extent you agree or disagree that your firm uses the following
practices specifically to reduce the risk or impact of supply chain disruptions:

Strongly Strongly
PRACTICES DISAGREE neutral AGREE

1. Whenever possible, we use two or more suppliers for our material needs,
. 1 2 3 4 5 6 7
to reduce rIsk

 

 

 

 

 

 

 

 

 

2. We store material and goods in multiple warehouses to reduce risk 1 2 3 4 5 6 7

3. We hold our own safety stock of raw materials/components 1 2 3 4 5 6 7

4. We have suppliers hold safety stock of materials/components for us 1 2 3 4 5 6 7

5. We hold our own safety stock of finished or semi-finished goods 1 2 3 4 5 6 7

6. We have multiple production lines/facilities for critical processes or 1 2 3 4 5 6 7
products to decrease the risk of having only a single location

7. We have specific agreements with our suppliers to respond to changes in 1 2 3 4 5 6 7
product mix, volume or delivery schedules

8. Our workforce is cross-trained to perform a variety of jobs as needed 1 2 3 4 5 6 7

9. We can make our products in more than one location due to processes 1 2 3 4 5 6 7
standardization across facilities, production lines or work cells

10.We standardize our materials/components across product types/lines 1 2 3 4 5 6 7

 

11.We utilize postponement, where operations or activities (e.g.
customization, final configuration, etc.) are moved to a later point in the 1 2 3 4 5 6 7
supply chain

 

12.We offer a wide portfolio of different versions or options of our product 1 2 3 4 5 6 7
to meet the needs of different customer segments

 

176

Please indicate to what extent you agree or disagree that your firm uses the following practices
as part of general business practices unrelated to risk management:
Strongly Strongly
DISAGREE neutral AGREE

1. We implement processes to improve new product development (NPD) (e.g. 1 2 3 4 5 6 7

modularization, design for manufacturability, early supplier involvement, etc.)

 

2. We concentrate on our core competencies and outsource some processes 1 2 3 4 5 5 7

and activities (e.g. lS/IT, facility maintenance, warehousing, etc.)

 

3. We have cross-functional cooperation between departments/functions

 

. We develop or acquire proprietary equipment, processes, and/or products

 

5. We develop strategic supply chain partnerships, where appropriate, that are 1 2 3 4 5 5 7
designed for on-going mutual benefit

 

 

 

 

6. We are committed to providing quality information to our supply chain 1 2 3 4 5 5 7
partners that is accurate, timely, adequate, and credible.

7. We utilize a supplier certification process for key materials or suppliers 1 2 3 4 5 5 7

8. We share critical/proprietary information with our supply chain partners 1 2 3 4 5 5 7

9. We utilize customer relationship management (CRM) for the purpose of 1 2 3 4 5 5 7

managing customer complaints, building long-term relationships with
customers, and improving customer satisfaction

 

 

 

Strongly Strongly
DISAGREE neutral AGREE
1. We have a strong time-based synchronization with our supply chain 1 2 3 4 5 5 7
partners
2. We have knowledge of activities of our first tier supply chain partners 1 2 3 4 5 5 7

 

3. We have knowledge of activities of our supply chain beyond the first tier 1 2 3 4 5 5 7

 

 

 

 

 

 

 

 

 

4. Our employees have close personal relationships with their supply chain 3 4 5 7
partner counterparts

5. We are aware of the strengths of our supply chain partners 1 2 3 4 5 5 7

6. We are aware of the weaknesses of our supply chain partners 1 2 3 4 5 5 7

7. We are economically dependent on our supply chain partners 1 2 3 4 5 5 7

8. We have a high degree of alignment of business activities with our supply 1 2 3 4 5 5 7
chain partners

9. Our business activities are developed through knowledge that is 1 2 3 4 5 5 7
exchanged with our supply chain partners

10.We adapt our IT investments to better fit with our supply chain partners 1 2 3 4 5 5 7

11.0ur reputation in the market is affected by our supply chain relationships 1 2 3 4 5 5 7

Very Very

LOW moderate HIGH

12. In your opinion, how would you rate the vulnerability of your supply chain
to disruptions (any event that impacts the flow of goods/services or that
has a negative financial impact to maintain or resume the flow of
good/services)

1234567

 

OFTEN sometimes NEVER
13.You hear about your competitors experiencing supply chain disruptions 1 2 3 4 5 6 7

 

 

177

DEMOGRAPHIC DATA

 

This section seeks to gather general demographic information about you and your firm. This
information will not be used to specifically identfl Lou or your firm, but allows for general
grouping of responses based on criteria such as industry or company size to try to find trends.

1. Please indicate the primary industry your firm is involved in:

Communications I; Manufacturing
Construction 'L; Retail or Wholesale Sales
Finance, Insurance or Real Estate I ‘ TransportationNVarehousing
- Government I ,‘ Utilities
I . Health Services I I Other:

.2

F? Agriculture IT Hospitality/Food Service
I“

l_

l

l

 

N

. What is the annual gross revenue of your company?
I; Less than $10 million
I : $10 to 249 million
L." $250 to 499 million
II $500 to 749 million
I": $750 million to 1 billion
I,‘ Over $1 billion

3. How many employees does your company have?
I: Less than 100
I; 100 to 249
I : 250 to 999
I"; 1,000 to 4,999
I I 5,000 to 9,999
L Over 10,000

.5

What is your primary functional area or department?
I E Engineering I Production control/planning
I . Facilities management I; Project management
I] IT/IS T7 Risk/Business continuity management
I“? Logistics fl Warehousing
I J Manufacturing/production If _- Other:
I 3 Procurement/purchasing

 

5. What position level are you in your company?
[I VP
I”. Director
I Manager
I . Supervisor
I ‘ Individual Contributor
I Other:

 

6. What level of your organization do you work in?
E Corporate Level
p Strategic Business Unit Level
I . Plant Level
I Other:

 

178

Please include any comments about supply chain risk management that you think are important:

 

 

 

 

 

Are there any best practices that you would like to share?

 

 

 

 

 

What issues or topics do you think need more exploration or research?

 

 

 

 

 

THANK YOU FOR COMPLETING THE SURVEY, YOUR INPUT IS GREATLY APPRECIATED

If you would like to be entered for the chance to win one of the ten $50 Visa gift cards, please
provide the information below. This information will only be used to send the gift card to you, if
your name is drawn. You may still send the survey anonymously if you wish.

Name:

Address:

 

 

 

 

State: Zip Code:
e-mail:

 

 

Phone:

 

179

Appendix B: Copy of survey contact materials

«MARSTAT» «FIRST» «LAST»
«COMPANY»

«CADDRESS»

«CITY», «STATE» «ZIP»

Dear «FIRST»:

The American Purchasing Society and researchers from Michigan State University have
teamed together to study supply chain risk management. Your professional knowledge and
experience is invaluable in helping us to better understand what drives firms to engage in
supply chain risk management planning, the types of planning firms under take, and the effect
of such plans on the occurrence of supply chain disruptions.

Your input into this study is vitally important, regardless of the level of planning your
firm engages in or the frequency or size of supply chain disruptions you may experience.
We need data from as many professionals as possible to determine how firms try to manage
this process and to identify ‘best practices' so firms can improve to reduce the likelihood and
impact of a wide variety of disruptive incidents - insights that you can use within your own firm!

In exchange for your valuable input to the survey, we are providing you with the following
opportunities:

For your firm: A summary report of the findings will be made available to APS members to
allow you to begin improving your supply chain.

For you: As a special thank you for your participation, you will have the opportunity to enter a
drawing for one of ten Visa Gift Cards of $50 each.

You may return the survey in the enclosed postage paid envelope, or you may complete the
survey electronically at: www.msurisk.org

Thank you for your input on this important business topic.

.» " “I. /
‘77s“ 7 ””377 W4
l 1‘.” .

Richard H. Hough,.CPP . Andrea M. Prud’homme
American Purchasmg SOCIety Michigan State University
support@american-purchasing.com SC_risk@bus.msu.edu

\II( ‘I III 7.-\\II \II\II
. \ I \l \l I

   

You indicate your voluntary agreement to participate by completing the survey. All individual responses will be
kept strictly confidential and will not be shared with any individual or organization beyond the research team. All
analyses, reports and presentations of the data generated from this study will be reported at an aggregate level.

If you have questions or concerns regarding your rights as a study participant, you may contact anonymously if

you wish Dr Peter Vasilenko, Chair of Behavioral Social Science Institutional Review Board via.
phone (517) 355- 2180 fax (517) 432—4503 or e-mail irb@msu. edu

180

 

  

“o MNI'sIKNIIH

'4 l-
a f.

   

 

”O H p967
Dear XXX:

The American Purchasing Society and researchers from Michigan State University are teaming
together to study supply chain risk management. Your knowledge and experience is invaluable in
helping us to better understand what drives firms to engage in risk management, the types of
planning firms under take, and the effect of such plans on the occurrence of supply chain
disruptions.

Your input into this study is vitally important, regardless of the level of planning your firm
engages in or the frequency or size of supply chain disruptions you may experience. We
need data from as many professionals as possible in order to determine ‘best practices’- insights
that you can use within your own firm. .

In exchange for your valuable input to the survey available at the link below, we are providing you
with the following opportunities:

For your firm:

1) A summary report of the findings will be made available to APS members.

2) A short article, 18 Ways to Guard Against Disruption, is available to you at the end of the
questionnaire if you complete the survey electronically.

Foryou:
1) As a special thank you for your participation in the survey, you will have the opportunity to
enter a drawing for one of ten Visa Gift Cards of $50 each.

You may return the survey in the enclosed postage paid envelope. Or you may log on to the
website to complete the survey electronically: www.msurisk.org

Thank you for your support!

Richard H. Hough,.CPP . Andrea M. Prud'homme
American Purchasrng Socrety Michigan State University
support@american-purchasing.com SC_risk@bus.msu.edu

You indicate your voluntary agreement to participate by completing the online survey. All individual responses will be
kept strictly confidential and will not be shared with any individual or organization beyond the research team. All
analyses, reports and presentations of the data generated from this study will be reported at an aggregate level.

If you have questions or concerns regarding your rights as a study participant, you may contact - anonymously if you

wish - Dr. Peter Vasilenko, Chair of Behavioral, Social Science Institutional Review Board by phone (517)355-2180, fax
(517)432-4503, or e-mail irb@msu.edu

181

Copy of e-mail message sent:
We need your help!

The American Purchasing Society and researchers from Michigan State University are teaming
together to study supply chain risk management. Your knowledge and experience in supply chain
and materials management is critical to helping us to better understand what drives firms to
engage in risk management, the types of planning that firms under take, the effect of such plans
on the occurrence of supply chain disruptions.

A short article, 18 Ways to Guard Against Disruption, is available to you at the end of the
questionnaire

As a special thank you for your participation in the survey, you will have the opportunity
to enter a drawing for one of ten Visa Gift Cards of $50 each.

Please go to the following website to complete the survey:

 

http:/Inebula.bus.msu.edulSC Risklsc risk.htm

 

 

 

Please feel free to share this opportunity with other supply chain management
professionalsl

Your input into this study is vitally important, regardless of the level of planning your firm engages
in or the frequency or size of supply chain disruptions you may experience. We need data from as
many professionals as possible in order to determine best practices so that firms can put plans in
place to reduce the likelihood and impact of a wide variety of disruptive incidents - insights that
you can use within your own firm with the findings from this study that we will share with you!

Please take a bit of time out of your busy day to help us better understand how we can more
effectively reduce supply chain disruptions.

Andrea Prudhomme and Ken Boyer
Michigan State University
SC_Risk@bus.msu.edu

Rich Hough
American Purchasing Society

All information that you provide is strictly confidential; name/address information will be utilized only to mail the gift cards.
Your privacy will be protected to the maximum extent allowable by law. All reports and data from this survey will be an
aggregate summary of the results, but no individual will be identified by name or address.

You indicate your voluntary agreement to participate by completing the online survey. Please email or call if you have any
questions regarding this study. You may also contact Peter Vasilenko, Ph.D., Chair University Committee on Research
Involving Human Subjects (UCRIHS) at (517) 355-2180 or ucrihs@mus.edu. Thank you for your cooperation. We look
forward to hearing from you.

182

Copies of postcards sent to contact sample
(shown larger than actually sent to meet dissertation formatting guidelines)

 

£381
0" . 0%

Q 1» Supply Chain Study

  
  

... .....

    

.. ‘;:.:.,‘;.i;/'° \Iltjlll(.\\' \HII
«I-..’¢ t\l‘~ti\\,l":
‘90 H A 9‘77
The American Purchasing Society and Michigan State University are teaming together to study supply chain risk
management. We need your valuable experience and professional opinions to help us better understand this
important business issue.

 

Your input into this study is vitally important, regardless of the level of planning your firm engages in or the frequency
and size of supply chain disruptions you may experience.

The final results of this study will be made available to APS members to provide insights that you can use within
your own firm to reduce supply chain disruptions.

In a few days you will receive a survey by mail, along with a postage paid envelope for you to return your completed
survey. Please take a bit of time out of your busy day to help us enhance our understanding supply chain risk
management, and how APS members can better plan for and recover from supply chain disruptions.

To make your participation as convenient as possible, you can also electronically access the survey immediately at:
www.msurisk.org

Thank you for your participation in this important study!

Rich Hough, CPP Andrea Prudhomme
American Purchasing Society Michigan State University
support@american-purchasing.com SC_risk@bus.msu.edu

 

 

The American Purchasing Society and Michigan State University will soon begin the analysis of the
responses they have received so far from their study of supply chain risk and continuity management
survey.

If you haven’t already done so, there is still time for you to participate in this important study!! By
completing the survey, your valuable experience and professional opinions will provide us with important
insight into this critical business issue.

Your contribution to this study is vitally important, regardless of the level of planning your firm engages
in or the frequency and size of supply chain disruptions you may experience.

The final results of this study will be made available to APS members to provide you with insights
that you can use within your own firm to reduce supply chain disruptions.

Please take a bit of time out of your busy day to help us enhance our understanding of supply chain risk
management, and how APS members can better plan for and recover from supply chain disruptions.

To make your participation as convenient as possible, you may electronically access the survey:
www.msurisk.org
If you would like a hard copy of the survey, please contact Andrea at SC_risk@bus.msu.edu or

(517) 432-5535 x272 and a paper survey with a preaddressed I postage paid return envelope will be
mailed to you right away.

Thank you for your participation in this important study!

Rich Hough, CPP Andrea Prud'homme
American Purchasing Society Michigan State University
«‘5'54
W.,; BROAD

   

‘1 W. O (_QItAIHAII \(;tlt)0l

01‘ xi] Ci (LI 1777.7

 

MICI "CAN S'ls‘tTl-

IITJI\"[I‘-,5li)

 

 

 

 

183

Appendix C: Scale Item Results from Survey

All response options are a Likert type scale of 1 to 7. Anchor and midpoint terminology
detailed below. Items in all italics were dropped from scales.

(.34

SUPPLY CHAIN COMPLEXITY
(scale: 1 = Extremely Low 4 = Moderate 7 = Extremely High) Std )1
(:1 Complexity of the products sold to customers (e.g. uses new or n/a

frequently changing technology, large bill of material, quality
specifications that are difficult to meet, several different versions or
options of products, etc.)

We offer a wide portfolio of different versions or options of our n/a
product to meet the needs of different customer segments

Complexity in the business environment (e.g. number of competitors, n/a
different customer niches, dynamic and frequently changing
conditions, etc.)

Overall complexity of the supply chain structure complexity in terms 0.69
of the number of production or warehouse facilities where your
product is built or stored

Overall complexity of the supply chain structure in terms of the 0.66
number of different divisions, business units, departments, or
managers of your firm that are involved in your supply chain

Overall complexity of a wide supply chain with numerous suppliers, 0.77
customers, distribution points, or multiple channels

Overall complexity of a long supply chain with multiple tiers or 0.76
echelons of suppliers and customers.

Overall complexity in terms of the geographic dispersion (regionally, 0.59
nationally or globally) of suppliers, production or warehouse
facilities, customers, or distribution channels.

SUPPLY CHAIN UNCERTIANTY

I" Q

\r I.)

L :2)

(scale: 1 = Extremely Low 4 = Moderate 7 = Extremely High) Std 1

Overall uncertainty of supply of material (e.g. uncertainty in 0.74
availability, quality or lead time of purchased items)

Overall uncertainty of output (e.g. variable quality, volume or timing 0.82
of output from production)

(:10 Overall uncertainty of demand (e.g. variation in timing, mix or 0.66

volume of customer demand; customers changing product
requirements)

184

 

Mean Std Dev
4.40 1.67
4.79 1.91
4.57 1.37
4.01 1.51
3.88 1.64
4.26 1.54
3.93 1.53
4.39 1.57

Mean Std Dev

4.61

4.28

4.50

1.58

1.58

1.53

SUPPLY CHAIN DEPENDENCY

 

 

 

 

 

(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std 1 Mean Std Dev
D1 We have strong time-based synchronization with our supply chain 0.74 4.61 1.54
partners
D2 We have knowledge of activities of our first tier supply chain partners 0.79 5.07 1.42
I33 We have knowledge of activities of our supply chain beyond the first tier 0.75 4.35 1.55
W Our employees have close personal relationships with their supply chain n/a 4.23 1. 73
oartner counterparts
2’35 We are aware of the strengths of our supply chain partners 0.65 5.24 1.30
D6 We are aware of the weaknesses of our supply chain partners 0.53 5.02 1.38
.07 We are economically dependent on our supply chain partners n/a 4-50 7-53
{)8 We have a high degree of alignment of business activities with our 0.70 4.55 1.47
supply chain partners
[)9 Our business activities are developed through knowledge that is 0.68 4.58 1.53
exchanged with our supply chain partners
DIOWe adapt our IT investments to better fit with our supply chain partners 0.54 3.78 1.72
E?- z 1 Our reputation in the market place is affected by our supply chain n/a 5.17 1. 62
relationships
LEVEL OF CONCERN FOR DISRUPTIONS
(scale: 1 = None 4: Moderate 7 = High) Std )1 Mean Std Dev
A C 1 Level of concern for market/business environment caused disruptions n/a 4.64 1.5 7
(e.g. regulatory issues, currency fluctuations, material cost instability,
economic change, competitor’s actions, etc.)
AC2Level of concern for upstream/supplier caused disruptions (e.g. inability 0.67 5.53 1.34
to meet demand, late deliveries, quality issues, material allocation, etc)
KIC'3LeveI of concern for internally caused disruptions (e.g. interruption or 0.71 4.73 1.68
delay in production output due to quality, equipment or personnel issues;
inventory issues; capacity shortage; etc.)
+1C4Level of concern for downstream / customer caused disruptions (e.g. 0.64 4.45 1.63
changes in product requirements, changes in timing, volume or mix of
demand, etc.)
(scale: 1 = Very Low 4 = Moderate 7 = Very High)
[_J- :5: In your opinion, how would you rate the vulnerability of your supply chain n/a 4.45 1.37
(scale: 1 = Often 4: Sometimes 7 = Never)
2'3 You hear about your competitors experiencing supply chain disruptions n/a 4.29 1.39

185

w
(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std 1

I P1 We implement processes to improve new product development (NPD) 0.70
(e.g. modularization, design for manufacturability, early supplier
involvement, etc.)

51“.? We concentrate on core competencies and outsource some activities n/a
(e.g. IS/IT, facility maintenance, warehousing, etc.)

3:133 We have cross-functional cooperation between different 0.65
departments/functions

ii .54 We develop/acquire proprietary equipment, processes, or products 0.61

:4 ; ‘5 We develop strategic supply chain partnerships, where appropriate, 0.77

that are designed for on-going mutual benefit

8176 We are committed to providing quality information to our supply chain 0.78
partners that is accurate, timely, adequate, and credible.

P7 We utilize a supplier certification process for key materials/suppliers 0.63
1:3 We share critical/proprietary information with supply chain partners 0.58

1.139 We utilize customer relationship management (CRM) for the purpose 0.65
of managing customer complaints, building long-term relationships
with customers, and improving customer satisfaction

PROFIL_E THE SUPPLY CHAIN

 

(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std 1
net We know our suppliers, including their locations and inter- 0.75
relationships within our supply chain
$31132 We know our customers, including their locations and inter- 0.58
relationships within our supply chain
s»:‘~:;1 We know our production and warehouse facilities, including their 0.76
locations and inter-relationships within our supply chain
WW! We have identified the transit lanes used to ship materials/products 0.65
:1ch We have a profile (or map) of our supply chain n/a
MNTIFICMION
(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std 1
12311 We know the particular vulnerabilities of our suppliers 0.78

Ii=2 We know the particular vulnerabilities of the primary transit lanes for 0.72
incoming material and inventory

We know the particular vulnerabilities of our primary production points 0.79
and our warehouses

34.4 We know the particular vulnerabilities of our customers n/a

Fifi We know the vulnerabilities of the primary transit lanes used for 0.60
outbound inventory and products

We explicitly consider vulnerabilities when selecting new supply chain 0.76
partners

We explicitly consider vulnerabilities when evaluating current supply 0.75
chain partners

:8 We use multiple sources of information to identify supply chain n'sk n/a

LJ

LID

186

Mean Std Dev

4.76

4. 74
5.04

4.90
5.26

5.45

4.87
4.63
4.71

Mean
5.80

5.50
5.87

5.34
4. 78

m
5.01
4.80
5.11

4. 63
4.95

5.37
5.31

5.07

1.74

1.63
1.57

1.71
1.39

1.44

1.93
1.77
1.79

m
1.14

1.23
1.17

1.42
1.64

$3.52
1.35
1.42
1.39

1.47
1.36

1.48
1.50

1.61

 

RISK ASSESSMENT

 

 

 

 

 

(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std )1 Mean Std Dev
531 We have a process for assessing the probability or frequency of 0.89 3.90 1.70
different types of supply chain disruptions
$331.2 We have a process for assessing the magnitude or impact of different 0.89 4.02 1.67
types of supply chain disruptions
i-iA3 We have a process to capture the financial impact of a supply chain 0.74 4.22 1.75
disruption (e.g. cost of expediting, late penalties, overtime, lost sales,
etc)
“TM We have a process to capture the qualitative impact of a supply chain 0.82 3.99 1.76
disruption (e.g. loss of good will, damage to firm reputation, impact on
supply chain relationships, etc.)
AUDIT OF CONTINUITY PLANS
(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std )1 Mean Std Dev
Pl We have formal supply chain continuity plans for our supply chain 0.80 3.78 1.83
.11: r52 We have informal supply chain continuity plans for our supply chain n/a 4.52 1.52
I533 We review our supply chain continuity plans internally 0.83 4.20 1.77
I‘~?"4 We have ‘fire drills' to test our supply chain continuity plans 0.66 2.66 1.61
25 We periodically review our supply chain continuity plans with our 0.77 3.69 1.81
suppliers
W6 We periodically review our supply chain continuity plans with our 0.64 3.07 1.62
customers
919 We have identified signals that we should watch for that would indicate n/a 4. 75 1.56
a disruption may be starting or about to occur
RISK MANAGEMENT STRATEGIES
(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std )1 Mean Std Dev
3:"? 1' 1 There is little we can do to reduce supply chain disruptions, so we n/a 2.83 1.65
accept them as part of doing business
5T7. In general, our supply chain risks are well managed .70 4-55 1-42
913 We utilize a strategy of sharing supply chain risk with our supply .54 3.85 1.76
chain partners (e.g. buy back agreements, cost or revenue sharing,
etc.)
i‘é'i4 We have strategically transferred most of our supply chain risk to .52 3.72 1.76
supply chain partners (e.g. vendor managed inventory, contractual
performance agreements, etc.)
811.15 In general, we try to reduce supply chain risk as much as possible .68 5.25 1.51
3310 We have been able to pool supply chain risk (e.g. use of buyer .62 3.79 1.80
consortiums, centralized warehousing, trans-shipments,
postponement or delayed finalization of finished products, etc.)
.317 We have been able to eliminate most of our supply chain risk .80 3.58 1.68

187

 

 

RISK MANAGEMENT PRACTICES

(scale: 1 = Strongly Disagree 4 = Neutral 7 = Strongly Agree) Std A Mean Std Dev
P1 Whenever possible, we use two or more suppliers for our material n/a 4.98 1. 77
needs, to reduce risk
152 We store material and goods in multiple warehouses to reduce risk n/a 3. 66 2.04
P3 We hold our own safety stock of raw materials/components n/a 4. 45 1.91
P4 We have suppliers hold safety stock of materials/components for us n/a 4.80 1. 72
P5 We hold our own safety stock of finished or semi-finished goods n/a 4.56 1.84
P6 We have multiple production lines/facilities for critical processes or .60 3.72 2.01
products to decrease the risk of having only a single location
P7 We have specific agreements with our suppliers to respond to .60 4.51 1.84
changes in product mix, volume or delivery schedules
58 Our workforce is cross-trained to perform a variety of jobs as needed .60 4.85 1.66
P9 We can make our products in multiple locations due to standardization I .63 3.93 2.00
processes across facilities, production lines or work cells
P10 We standardize our materials/components across product types/lines .61 4.53 1.79
I311 We utilize postponement, where operations or activities (e.g. .66 3.84 1.63

customization, final configuration, etc.) are moved to a later point in
the supply chain

CHANGE IN DISRUPTION -- SUPPLY
(scale: 1 = Less Often 4 = Same 7 = More Often) Std x

DFZ FRE QUENCE Y: Change in disruptions from upstream/supplier n/a
sources (material shortages, poor supplier delivery or quality
performance, in-bound logistics, etc.)

DF3 FREQUENCEY: Change in disruptions from internal sources (quality 0.88
problems, insufficient capacity, IT failure, personnel issues, etc.)

Di 2 IMPACT: Change in disruptions from upstream/supplier sources n/a
(material shortages, poor supplier delivery or quality performance, in-
bound logistics, etc.)

DI3 IMPACT: Change in disruptions from internal sources (insufficient 0.88
capacity, quality problems, IT failure, personnel issues, etc.)

 

CHANGE IN DISRUPTION — DEMAND

Mean Std Dev

4. 33

3.94

4. 33

3.94

1.56

1.64

1.56

1.64

(scale: 1 = Much Smaller 4 = Same 7 = Much Larger) Std )1 Mean Std Dev

DF 1 FREQUENCY: Change in disruptions from business environment n/a
sources (changes in technology, government regulation, currency
fluctuations, changing commodity prices, competitor actions, etc.)

DF4 FREQUENCEY: Change in disruptions downstream/customer sources 0.98
(out-bound logistics issues, changes in mix, timing or volume of
customer demand, etc.)

D I1 IMPACT: Change in disruptions from business environment sources n/a
(changes in technology, govemment regulation, currency fluctuations,
changing commodity prices, competitor actions, etc.)

D14 IMPACT: Change in disruptions downstream/customer sources 0.87
(outbound logistics issues, changes in mix, timing or volume of
customer demand, etc.)

188

4.27

3.88

4.34

3.97

1.58

1.58

1.49

1.60

FIRM PERFOMRANCE

(scale: 1 = Significantly Lower 4 = Same 7 = Significantly Higher) Std )1 Mean

 

Std Dev
iJMICompared to our competitors, our sales are 0.62 4-35 1-35
‘3’?»il-BCompared to our competitors, our profitability is 0.78 4-86 1-38
*V’fiTSCompared to our competitors, our overall competitive position is 0.77 5.05 1.31
iT-I..1‘?4Compared to our competitors, our return on assets (ROA) is 0.95 4.76 1-28
Plat-‘I‘Compared to our competitors, our return on sales (R08) is 0.96 4.77 1-30

189

 

 

Appendix D: Summary of Final Revised Hypotheses
Normal Accident Theory Hypotheses

H1 a: Higher levels of perceived supply chain complexity are positively associated with higher
levels of continuity planning (CPLX —*—+ PLAN)

H1 b: Higher levels of perceived supply chain dependency are positively associated with higher
levels of continuity planning (DPND —+—> PLAN)

H1 c: Higher levels of perceived supply chain uncertainty are positively associated with higher
levels of continuity planning (UNCT —i—-) PLAN)

H2as: Higher levels of perceived supply chain complexity are positively associated with greater
frequency and impact of disruptive events — supply (CPLX ——-+—> DES)

HZaD: Higher levels of perceived supply chain complexity are positively associated with greater
frequency and impact of disruptive events - demand (CPLX ———+—-) DED)

H2bs: Higher levels of perceived supply chain dependency are positively associated with
greater frequency and impact of disruptive events — supply (DPND ——+——> DES)

H2bD: Higher levels of perceived supply chain dependency are positively associated with
greater frequency and impact of disruptive events - demand (DPND ——+-—) DED)

H2cs: Higher levels of perceived supply chain uncertainty are positively associated with greater
frequency and impact of disruptive events - supply (UNCT—+—> DES)

H209: Higher levels of perceived supply chain uncertainty are positively associated with greater
frequency and impact of disruptive events — demand (UNCT —-+—) DED)
High Reliability Theory Hypotheses

H3a: Higher levels of the use of effective practices are positively associated with higher
levels of continuity planning (EFPR —+—> PLAN)

H3b: Higher levels of concern for disruptions are positively associated with higher levels
of continuity planning (CN RN —+> PLAN)

H4a: Higher levels of the use of effective practices are positively associated with firm
performance (EFPR —+—> FMPF)

H4b: Higher levels of concern for disruptions are positively associated with firm
performance (CNRN ——+—> FMPF)

H5a3: Higher levels of the use of effective practices are negatively associated more
frequent and larger impact disruptive events - supply (EFPR —;> DES)

H5ao: Higher levels of the use of effective practices are negatively associated more
frequent and larger impact disruptive events — demand (EFPR —'—) DED)

H5bs: Higher levels of concern for disruptions are negatively associated more frequent and
larger impact disruptive events - supply (CN RN —") DES)

HSbD: Higher levels of concern for disruptions are negatively associated more frequent
and larger impact disruptive events — demand (CNRN —‘—> DED)
Planning Hypotheses

H6: Higher levels of continuity planning are positively associated with higher levels
of firm performance (PLAN —+—+ F MPF)

H75: Higher levels of continuity planning are negatively associated with more frequent and
larger impact disruptive events — supply (PLAN —‘> DES)

H79: Higher levels of continuity planning are negatively associated with more frequent
and larger impact disruptive events - demand (PLAN —"> DED)

H83: More frequent and larger impact disruptions — supply are negatively associated with
firm performance (DES —;—> FMPF)

H80: More frequent and larger impact disruptions — demand are negatively associated
with firm performance (D ED —'—) FMPF)

190

 

 

Appendix E: Summary of Data Analysis Statistics

 

Issue

Result

Reference

 

Final sample size

Final response rate
Statistical analysis software
SEM Software and input data
Missing data

Normality

Number of manifest variables

Number of latent variables

Manifest to latent variable ratio

Correlated measurement
errors

Recursiveness

Evidence of model
identification
Power determination

CFA model fit

Full path model fit

298 usable responses

9.5%
SPSS 15.0
EQS 6.1, raw data table

lmputation using EM method

Acceptable kurtosis (< +l-3),
skewness (<+/- 2) and
Mardia—Based Kappa of 0.084

Hypothesized = 6, and 1 second

order construct of 6 variables

After CFA = 8, and 1 second order

construct of 4 variables
92 to start, 75 after CFA

At least 3:1, other than one
construct with 2:1

None found

Recursive model

df = 2,319 for CFA model
df = 715 for path model

Power adequate at >0.80

1,223.9, = 4294.784

CFI =0.83; SRMR = 0.55;
RMSEA = 0.054

112,,” = 2068.6

CFI = 0.78; SRMR = 0.14;
RMSEA = 0.080

191

(Hu & Bentler, 1998;
MacCallum et al., 1996;
Marsh et al., 1988;
Rigdon, 1995)

(Bentler & Wu, 2003)
(Hair Jr. et al., 2006)

(Bollen, 1989; Hair Jr. et
al., 2006)

(Rigdon, 1995)

(MacCallum et al., 1996;
McQuitty, 2004)

(M. W. Browne &
Cudeck, 1993; Hu &
Bentler, 1999)

(M. W. Browne &
Cudeck, 1993; Hu &
Bentler, 1999)

F11

 

 

BIBLIOGRAPHY

Aaker, D. A., Day, G. 8., & Kumar, V. (1998). Marketing research (6th ed.). New
York: John Wiley & Sons, Inc.

Adams, L. L., & Gale, D. (1982). Solving the quandary between questionnaire length
and response rate in higher educational research. Research in Higher Education,
17(3), 231-240.

Allaire, Y., & Firsirotu, M. E. (1989). Coping with strategic uncertainty. Sloan
Management Review, 30(3), 7.

Alonso, F., & Boucher, J. (2001). Business continuity plans for disaster response.
The CPA Journal, 71 (1 1), 60.

American Purchasing Society. , 2007, from http://www.american-purchasing.com

Anand, G., & Ward, P. T. (2004). Fit, flexibility and performance in manufacturing:
Coping with dynamic environments. Production and Operations Management,
13(4), 369.

Anderson, J. C., & Gerbing, D. W. (1982). Some methods for respecifying
measurement models to obtain unidimensional construct measurement. Journal
of Marketing Research, 19(4), 453.

Anderson, J. C., & Gerbing, D. W. (1984). The effects of sampling error on
convergence, improper solutions, and goodness-of-fit indices for maximum
likelihood confirmatory factor analysis. Pyschometrika, 49, 155-173.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A
review and recommended two-step approach. Psychological Bulletin, 103(3), 411.

Armstrong, J. S., & Overton, T. S. (1977). Estimating nonresponse bias in mail
surveys. Journal of Marketing Research, 14(3), 396-402.

Bagozzi, R. P., & Heatherton, T. F. (1981). Evaluating structural equation models
with unobservable variables and measurement error: A comment. Journal of
Marketing Research, 18(August), 375-381.

Barnes, J. C. (2001). A guide to business continuity planning. West Sussex,
England: John Wiley & Sons, Ltd.

Bazerman, M. H., & Watkins, M. D. (2004). Predictable surprises: The disasters you
should have seen coming. Boston, MA: Harvard Business School.

192

 

 

Beamon, B. M. (1999). Measuring supply chain performance. lntemational Journal of
Operations & Production Management, 19(3), 275.

Bendoly, E. (2006). Incorporating behavioral theory in OM empirical models. Journal
of Operations Management, 24(6), 735-736.

Bentler, P. M., & Chou, C. P. (1987). Practical issues in structural equation
modeling. Sociological Methods and Research, 16(1), 78-117.

Bentler, P. M., & Wu, E. (2003). E08 6.1 for windows. Encino, CA: Multivariate
Software, Inc.

Bickart, B., & Schmittlein, D. (1999). The distribution of survey contact and
participation in the United States: Constructing a survey-based estimate. Journal
of Marketing Research, 36(2), 286-294.

Bielski, L. (2003). Business continuity planning must step up. American Bankers
Association. ABA Banking Journal, 95(11), 52.

Blackhurst, J., Craighead, C. W., Elkins, 0., & Handfield, R. B. (2005). An empirically
derived agenda of critical research issues for managing supply-chain disruptions.
International Journal of Production Research, 43(19), 4067.

Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.

Bollen, K. A., & Lennox, R. (1991). Conventional wisdom on measurement: A
structural equation perspective. Psychological Bulletin, 110(2), 305.

Bowersox, D. J., Closs, D. J., & Stank, T. P. (2000). Ten mega-trends that will
revolutionize supply chain logistics. Journal of Business Logistics, 21(2), 1.

Boyer, K. K. (1999). Evolutionary patterns of flexible automation and performance: A
longitudinal study. Management Science, 45(6), 824.

Boyer, K. K., & Verma, R. (2000). Multiple raters in survey-based operations
management research: A review and tutorial. Production and Operations
Management, 9(2), 128.

Brindley, C., & Ritchie, B. (2004). Introduction to supply chain risk. In C. Brindley
(Ed.), Supply chain risk 0. Hampshire, UK: Ashgate Publishing Limited.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K.
A. Bollen, & J. S. Long (Eds), Testing structural equation models (pp. 136-162).
Newbury Park, CA: Sage.

Browne, N. W., & Cudeck, R. (1989). Single sample cross-validation indices for
covariance structures. Multivariate Behavioral Research, 24(4), 445-455.

Castillo, C. (2004). Disaster preparedness and business continuity planning at
193

Boeing: An integrated model. Joumal of Facilities Management, 3(1), 8.

Cavinato, J. L. (2004). Supply chain logistics risks: From the back room to the board
room. lntemational Journal of Physical Distribution & Logistics Management,
34(5), 383.

Cerullo, V., & Cerullo, M. J. (2004). Business continuity planning: A comprehensive
approach. lnforrnation Systems Management, 21(3), 70.

Chapman, J. (2005). Predicting technological disasters: Mission impossible?
Disaster Prevention and Management, 14(3), 343.

Chapman, P., Christopher, M., Juttner, U., Peck, H., & Wilding, R. (2002). Identifying L
and managing supply-chain vulnerability. Logistics and Transport Focus, 4(4), 59. -

Chappell, L. (2006, April 24, 2006). Fed-up supplier tells GM: We quit. Automotive
News, pp. 8. I:

 

Chase, R. B., Jacobs, F. R., & Aquilano, N. J. (2004). Operations management for
competitive advantage (10th ed.). New York: McGraw-Hill Irwin.

Chen, I. J., & Paulraj, A. (2004). Towards a theory of supply chain management: The
constructs and measurements. Journal of Operations Management, 22(2), 119-
1 50.

Choi, T. Y., Dooley, K. J., & Rungtusanatham, M. (2001). Supply networks and
complex adaptive systems: Control versus emergence. Journal of Operations
Management, 19(3), 351.

Choi, T. Y., & Hong, Y. (2002). Unveiling the stmcture of supply networks: Case
studies in Honda, Acura, and DaimlerChrysler. Journal of Operations
Management, 20(5), 469.

Chopra, 8., & Sodhi, M. S. (2004). Managing risk to avoid supply-chain breakdown.
Sloan Management Review, 46(1), 53.

Christopher, M., 8. Lee, H. (2004). Mitigating supply chain risk through improved
confidence. lntemational Journal of Physical Distribution & Logistics
Management, 34(5), 388.

Churchill, G. A. (1979). A paradigm for developing better measures of marketing
constructs. Journal of Marketing Research, 16(3), 64-73.

Clark, J., 8: Harman, M. (2004). On crisis management and rehearsing a plan. Risk
Management, 51(5), 40.

Closs, D. J., & McGarrell, E. F. (2004). Enhancing security throughout the supply
chain, IBM Center for the Business of Government.

194

Conchar, M. P., Zinkhan, G. M., Peters, 0., & Olavarrieta, S. (2004). An integrated
framework for the conceptualization of consumers' perceived-risk processing.
Academy of Marketing Science Journal, 32(4), 418.

Craighead, C. W., Blackhurst, J., Rungtusanatham, M. J., & Handfield, R. B. (2007).
The severity of supply chain disruptions: Design characteristics and mitigation
capabilities. Decision Sciences, 38(1), 131.

Daft, R. L. (1989). Organization theory and design. New York: West Publishing.

Daft, R. L. (1995). Organization theory and design (5th ed.). St. Paul, MN: West
Publishing Company.

D'Avanzio, R. H., von Lewinski, H., & Van Wassenhove, L. N. (2003). The link
between supply chain management and financial performance. Supply Chain
Management Review, 7(6), 4047.

Davies, A. J., & Kochhar, A. K. (2000). A framework for the selection of best
practices. lntemational Journal of Operations & Production Management, 20(10),
1203.

Davies, A. J., & Kochhar, A. K. (2002). Manufacturing best practice and performance
studies: A critique. lntemational Journal of Operations & Production Management,
22(3), 289.

Davis, T. (1993). Effective supply chain management. Sloan Management Review,
(Summer), 35-46.

Delaware Lottery Games. (2004). Printing company employees scoop sixth largest
POWERBALL® jackpot. Dover, Delaware:

Dillman, D. A. (2007). Mail and intemet surveys: The tailored design method. New
York: John Wiley & Sons, Inc.

Dooley, K. J., & Van de Ven, A. (1999). Explaining complex organizational dynamics.
Organization Science, 10(3), 358-372.

Dostaler, l. (2001). Beyond practices: A qualitative inquiry into high performance
electronics assembly. Production and Operations Management, 10(4), 478.

Doty, D. H., & Glick, W. H. (1998). Common methods bias: Does common methods
variance really bias results? Organizational Research Methods, 1(4), 374-406.

Drbge, C., Jayaram, J., & Vickery, S. K. (2004). The effects of internal versus
external integration practices on time-based performance and overall firm
performance. Journal of Operations Management, 22(6), 557.

Drolet, A. L., 8. Morrison, D. G. (2001). Do we really need multiple-item measures
in service research? Journal of Service Research, 3(3), 196-204.

195

Elkins, 0., Handfield, R. B., Blackhurst, J., & Craighead, C. W. (2005). 18 ways to
guard against disruption. Supply Chain Management Review, 9(1), 46.

Elsinger, H., Lehar, A., & Summer, M. (2006). Risk assessment for banking systems.
Management Science, 52(9), 1301-1314.

Faught, K. S., W., G. K.,Jr, & Whitten, D. (2004). Doing survey research on the
internet: Yes, timing does matter. The Journal of Computer lnforrnation Systems,
44(3), 26.

Ferdows, K., Demeyer, A., (1990). Lasting improvements. Journal of Operations
Management, 9(2), 168.

Filbeck, G., Gorman, R., Greenlee, T., & Speh, T. (2005). The stock price reaction to
supply chain management advertisements and company value. Journal of
Business Logistics, 26(1), 199-216.

Finch, P. (2004). Supply chain risk management. Supply Chain Management, 9(2),
183.

Fisher, M. L., Hammond, J. H., Obermeyer, W. R., 8. Raman, A. (1994). Making
supply meet demand in an uncertain world. Harvard Business Review, 72(3), 83.

Flynn, B. B., & Flynn, E. J. (1999). Inforrnation-processing alternatives for coping
with manufacturing environment complexity. Decision Sciences, 30(4), 1021.

Flynn, B. B., Sakakibara, S., Schroeder, R. G., Bates, K. A., & Flynn, E. J. (1990).
Empirical research methods in operations management. Journal of Operations
Management, 9(2), 250.

Flynn, B. B., Schroeder, R. G., & Sakakibara, S. (1995). The impact of quality
management practices on performance and competitive advantage. Decision
Sciences, 26(5), 659.

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with
unobservable variables and measurement error. JMR, Journal of Marketing
Research (pre- 1 986), 18(000001 ), 39.

Forrester, J. (1961). Industrial dynamics. New York: MIT Press John Wiley & Sons,
Inc.

Forza, C. (2002). Survey research in operations management: A process-based
perspective. International Journal of Operations & Production Management, 22(2),
152.

Foster, S. P., & Dye, K. (2005). Building continuity into strategy. Journal of Corporate

196

Real Estate, 7(2), 105.

Fredendall, L. 0., Hopkins, C. D., & Bhonsle, A. (2005). Purchasing's intemal service
performance: Critical external and internal determinants. Journal of Supply Chain
Management, 41(2), 26.

Frizelle, G. (1998). The management of complexity in manufacturing. London:
Business Intelligence Limited.

Frohlich, M. T. (2002). Techniques for improving response rates in OM survey
research. Journal of Operations Management, 20(1), 53.

Frohlich, M. T., & Westbrock, R. (2001). Arcs of integrations: An lntemational study
of strategies for linking supply chains. Journal Of Operations Management, 19(2),
185-200.

Gardner, J. T., & Cooper, M. C. (2003). Strategic supply chain mapping approaches.
Journal of Business Logistics, 24(2), 37-64.

Garver, M. S., & Mentzer, J. T. (1999). Logistics research methods: Employing
structural equation modeling to test for construct validity. Journal of Business
Logistics, 20(1), 33.

Gerbing, D. W., & Anderson, J. C. (1984). On the meaning of within-factor correlated
measurement errors. Journal of Consumer Research, 11, 572-580.

Ghosh, D., & Ray, M. R. (1997). Risk, ambiguity, and decision choice: Some
additional evidence. Decision Sciences, 28(1), 81.

Gill, T. J. (2006). Workplace continuity: How risk and technology will affect facilities
strategy. Journal of Facilities Management, 4(2), 110.

Giunipero, L. C., & Eltantawy, R. A. (2004). Securing the upstream supply chain: A
risk management approach. lntemational Journal of Physical Distribution &
Logistics Management, 34(9), 698.

Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: Strategies
for qualitative research. New York: Aldine.

Grabowski, M., 8. Roberts, K. (1997). Risk mitigation in large-scale systems:
Lessons from high reliability organizations. California Management Review, 39(4),
152.

Grabowski, M., & Roberts, K. H. (1999). Risk mitigation in virtual organizations.
Organization Science, 10(6), 704.

Griffis, S. E., Cooper, M., Goldsby, T. J., & Closs, D. J. (2004). Performance
measurement: Measure selection based upon firm goals and information reporting

197

needs. Journal of Business Logistics, 25(2), 95.

Griffis, S. E., Goldsby, T. J., & Cooper, M. (2003). Web-based and mail surveys: A
comparison of response, data and cost. Journal of Business Logistics, 24(2), 237.

Hair Jr., J.,F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006).
Multivariate data analysis (Sixth ed.). Upper Saddle River, N.J.: Pearson Prentice
Hall.

Hallikas, J., Karvonen, l., Pulkkinen, U., Virolainen, V., 8: Tuominen, M. (2004). Risk
management processes in supplier networks. lntemational Journal of Production
Economics, 90(1), 47.

Hallikas, J., & Virolainen, V. (2004). Risk management in supplier relationships and
networks. In C. Brindley (Ed.), Supply chain risk (pp. 43-65). Hampshire, UK:
Ashgate Publishing Company.

Hamel, G., 8. Prahalad, C. K. (1994). Competing for the future. Boston, MA: Harvard
Business School Press.

Handfield, R. B., & Melnyk, S. A. (1998). The scientific theory-building process: A
primer using the case of TQM. Journal of Operations Management, 16(4), 321.

Hardy, V., 81 Roberts, P. (2003). lntemational emergency planning for facilities
management. Journal of Facilities Management, 2(1), 7.

Heath, 0., 8: Tversky, A. (1991). Preference and belief: Ambiguity and competence
in choice under uncertainty. Journal of Risk and Uncertainty, 4, 5-28.

Heberlein, T. A., & Baumgartner, R. (1978). Factors affecting response rates to
mailed questionnaires: A quantitative analysis of the published literature.
American Sociological Review, 43(4), 447-462.

Helferich, O. K., & Cook, R. L. (2002). Securing the supply chain. Oak Brook, IL:
Council of Logistics Management.

Hellstrbm, T. (2005). A decision model for involvement in vulnerability reduction.
Disaster Prevention and Management, 14(2), 196.

Hendricks, K. B., & Singhal, V. R. (2005a). Association between supply chain
glitches and operating performance. Management Science, 51 (5), 695.

Hendricks, K. B., & Singhal, V. R. (2005b). An empirical analysis of the effect of
supply chain disruptions on long-run stock price performance and equity risk of
the firm. Production and Operations Management, 14(1), 35.

Hensley, R. L. (1999). A review of operations management studies using scale
development techniques. Journal of Operations Management, 17(3), 343.

198

Hicks, M. (2002). When the chain snaps. eWeek, 19(7), 35.

Hillman, M. (2006). Strategies for managing supply chain risk. Supply Chain
Management Review, 10(5), 11.

Hinkin, T. R. (1995). A review of scale development practices in the study of
organizations. Journal of Management, 21(5), 967-988.

Holland, J. H. (1995). Hidden order. Reading, MA: Addison-Wesley.

Hu, L., & Bentler, P. M. (1998). Fit indices in covariance structure modeling:
Sensitivity to under-parameterized model misspecification. Psychological
Methods, 3(4), 424-453.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indicies in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling, 6, 1-55.

Huber, G. P., & Power, D. J. (1985). Retrospective reports of strategic-level
managers: Guidelines for increasing their accuracy. Strategic Management
Journal, 6(2), 171.

Hult, G. T. M., Ketchen Jr., D. J., Cui, A. S., Prud’homme, A. M., Seggie, S. H.,
Stanko, M. A., et al. (2006). An assessment of the use of structural equation
modeling in international business research. In D. J. Ketchen, 8. D. D. Berg (Eds),
Research methodology in strategy and management, vol. 3. Oxford, United
Kingdom: Elsevier JAl.

Hutchins, G. (2003). Risk management in the supply chain. Annual Quality Congress
Proceedings, 57, 49-58.

Jennier, J. M. (2004). Complex systems threaten to bring us down. Organization 8.
Environment, 17(1), 5.

Jia, J., Dyer, J. S., 81 Butler, J. C. (1999). Measure of perceived risk. Management
Science, 45(4), 51 9.

Johnson, M. E. (2001). Learning from toys: Lessons in managing supply chain risk
from the toy industry. California Management Review, 43(3), 106.

Kaplowitz, M. D., Hadlock, T. D., & Levine, R. (2004). A comparison of web and mail
survey response rates. Public Opinion Quarterly, 68(1), 94.

Ketokivi, M. A., 81 Schroeder, R. G. (20043). Manufacturing practices, strategic fit
and performance: A routine-based view. lntemational Journal of Operations &
Production Management, 24(1/2), 171.

Ketokivi, M. A., & Schroeder, R. G. (2004b). Perceptual measures of performance:

199

Fact or fiction? Journal of Operations Management, 22(3), 247.

Khemani, K. (2007). Bringing rigor to risk management. Supply Chain Management
Review, 11(2), 67.

Khurana, A. (1999). Managing complex production processes. Sloan Management
Review, 40(2), 85.

Killian, L. M. (2002). An introduction to methodological problems of field studies in
disasters. In R. A. Stallings (Ed.), Methods of disaster research, Xlibris
Corporation.

Kiser, J., & Cantrell, G. (2006). 6 steps to managing risk. Supply Chain Management
Review, 10(3), 12.

Klassen, R. D., & Jacobs, J. (2001). Experimental comparison of web, electronic and
mail survey technologies in operations management. Journal of Operations
Management, 19(6), 713.

Klein, R. L., Bigley, G. A., & Roberts, K. H. (1995). Organizational culture in high
reliability organizations: An extension. Human Relations, 48(7), 771.

Kleindorfer, P. R., & Saad, G. H. (2005). Managing disruption risks in supply chains.
Production and Operations Management, 14(1), 53.

Kleindorfer, P. R., & Van Wassenhove, L. N. (2004). Managing risk in global supply
chains. In H. Gaigon, & J. Kimberly (Eds.), The alliance on globalization
()Cambridge University Press.

Knight, R. F., & Pretty, D. J. (1998). Value at risk: The effects of catastrophes on
share price. Risk Management, 45(5), 39-41.

Krause, C. (1997). Package deal, Public Broadcasting Service.

Krause, D. R., & Ellram, L. M. (1997). Success factors in supplier development.
International Journal of Physical Distribution & Logistics Management, 27(1), 39.

Krueger Jr., N., 8. Dickson, P. R. (1994). How believing in ourselves increases risk
taking: Perceived self-efficacy and opportunity recognition. Decision Sciences,
25(3), 385.

Lampel, J., 81 Shapira, Z. (2001). Judgmental errors, interactive norms, and the
difficulty of detecting strategic surprises. Organization Science, 12(5)

Larson, P. D. (2005). A note on mail surveys and response rates in logistics
research. Journal of Business Logistics, 26(2), 211.

Laugen, B. T., Acur, N., Boer, H., & Frick, J. (2005). Best manufacturing practices:
What do the best-perfonning companies do? lntemational Journal of Operations &
200

Production Management, 25(2), 131.

Lee, H. L., & Billington, C. (1995). The evolution of supply-chain—management
models and practice at Hewlett-Packard. Interfaces, 25(5), 42.

Lee, H. L., Padmanabhan, V., 8. Whang, S. (1997). lnforrnation distortion in a supply
chain: The bullwhip effect. Management Science, 43(4), 546.

Lewis, M. A. (2003). Cause, consequence and control: Towards a theoretical and
practical model of operational risk. Journal of Operations Management, 21(2),
205.

Li, S., Rao, S. S., Ragu-Nathan, T. S., & Ragu-Nathan, B. (2005). Development and
validation of a measurement instrument for studying supply chain management
practices. Journal of Operations Management, 23(6), 618.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and
determination of sample size for covariance structure modeling. Psychological
Methods, 1(2), 130-149.

Manning, 8., & Gurney, A. (2005). Operational risk within an insurance market.
Journal of Financial Regulation and Compliance, 13(4), 293.

Marsh, H. W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit indexes in
confirmatory factor analysis: The effect of sample size. Psychological Bulletin,
103(3), 391.

McQuitty, S. (2004). Statistical power and structural equation models in business
research. Journal of Business Research, 57(2), 175-183.

Melnyk, S. A., Sroufe, R. P., Calantone, R. J., & Montabon, F. (2003). A model of
site-specific antecedents of ISO14001 certification. Production and Operations
Management, 12(3), 369.

Menor, L. J., 8. Roth, A. V. (2007). New service development competence in retail
banking: Construct development and measurement validation. Journal of
Operations Management, 25(4), 825-846.

Meredith, J. (1993). Theory building through conceptual methods. lntemational
Journal of Operations & Production Management, 13(5), 3.

Meredith, J. (1998). Building operations management theory through case and field
research. Journal of Operations Management, 16(4), 441.

Meredith, J. R., Rauri, A., Amoako-Gyampah, K., & Kaplan, B. (1989). Alternative

research paradigms in operations. Journal of Operations Management, 8(4), 297-
326.

201

 

Michels, S. (2002). On the waterfront, Public Broadcasting Service.

Milgate, M. (2001). Supply chain complexity and delivery performance: An
international exploratory study. Supply Chain Management, 6(3/4), 106.

Mitroff, l. I. (1988). Crisis management: Cutting through the confusion. Sloan
Management Review, 29(2), 15.

Mitroff, l. l., & Alpaslan, M. C. (2003). Preparing for evil. Harvard Business Review,
81(4), 109.

Mittelstaedt, R. E.,Jr. (2005). Will your next mistake be fatal? avoiding the chain of
mistakes that can destroy your organization. Upper Saddle River, NJ: Wharton
School Publishing.

Moore, G. C., & Benbasat, l. (1991). Development of an instrument to measure the
perceptions of adopting an information technology innovation. Information
Systems Research, 2(2), 192-222.

Norrman, A., & Jansson, U. (2004). Ericsson's proactive supply chain risk
management approach after a serious sub-supplier accident. lntemational Journal
of Physical Distribution & Logistics Management, 34(5), 434.

Nunnally, J. C. (1978). Psychometric methods (2nd ed.). New York: McGraw-Hill.

O'Leary-Kelly, S. W., & Vokurka, R. J. (1998). The empirical assessment of construct
validity. Journal of Operations Management, 16(4), 387.

Paulsson, U. (2004). Supply chain risk management. In C. Brindley (Ed.), Supply
chain risk (pp. 79-96). Hampshire, UK: Ashgate Publishing Limited.

Peck, H. (2005). Drivers of supply chain vulnerability: An integrated framework.
International Journal of Physical Distribution & Logistics Management, 35(3/4),
210.

Perreault, W. D.,Jr, & Leigh, L. E. (1989). Reliability of nominal data based on
qualitative judgments. Journal of Marketing Research, 26(2), 135.

Perrow, C. (1984). Normal accidents: Living with high risk technologies (First ed.).
New York: Basic Books.

Perrow, C. (1999). Normal accidents: Living with high risk technologies (Second
ed.). Princeton, NJ: Princeton University Press.

Perrow, C. (2004). A personal note on normal accidents. Organization &
Environment, 17(1), 9.

Pfeffer, J., 8. Salancik, G. R. (1978). The external control of organizations: A

202

 

 

Pich, M. T., Loch, C. H., & Meyer, A. D. (2002). On uncertainty, ambiguity, and
complexity in project management. Management Science, 48(8), 1008.

Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common
method biases in behavioral research: A critical review of the literature and
recommended remedies. Journal of Applied Psychology, 88(5), 879-903.

Porter, S. R., & Whitcomb, M. E. (2003). The impact of contact type on web survey
response rates. Public Opinion Quarterly, 67(4), 579.

8

Randall, T., & Ulrich, K. (2001). Product variety, supply chain structure, and firm
performance: Analysis of the US. bicycle industry. Management Science, 47(12),
1588.

Regester, M., & Larkin, J. (Eds.). (2005). Risk issues and crisis management: A
casebook of best practice (3rd ed.). London: Chartered Institute of Public
Relations.

Repenning, N. P., & Sterman, J. D. (2001). Nobody ever gets the credit for fixing
problems that never happened: Creating and sustaining process improvement.
California Management Review, 43(4), 64.

Rice Jr., J. B., & Caniato, F. (2003). Building a secure and resilient supply network.
Supply Chain Management Review, 7(5), 22.

Rigdon, E. E. (1995). A necessary and sufficient identification rule for structural
models estimated in practice. Multivariate Behavioral Research, 30(3), 359-383.

Ritchie, B., & Brindley, C. (2007a). An emergent framework for supply chain risk
management and performance measurement. The Journal of the Operational
Research Society, 58(11), 1398.

Ritchie, B., 8. Brindley, C. (2007b). Supply chain risk management and performance:
A guiding framework for future development. lntemational Journal of Operations &
Production Management, 27(3), 303-322.

Roberts, K. H. (1989). The significance of Perrow's normal accidents: Living with
high-risk technologies. Academy of Management Review, 14(2), 285-289.

Roberts, K. H. (1990). Managing high reliability organizations. California
Management Review, 32(4), 101.

Roberts, K. H., Bea, R., & Bartles, D. L. (2001). Must accidents happen? lessons
from high-reliability organizations. The Academy of Management Executive,
15(3), 70.

Rochlin, G. l., Porte, T. R. L., & Roberts, K. H. (1998). The self-designing high-
reliability organization: Aircraft carrier flight operations at sea. Naval War College
Review, 51(3), 97.

203

reliability organization: Aircraft carrier flight operations at sea. Naval War College
Review, 51(3), 97.

Rodetis, S. (1999). Can your business survive the unexpected? Journal of
Accountancy, 187(2), 27.

Rosa, E. A. (2005). Celebrating a citation classic - and more. Organization &
Environment, 18(2), 229.

Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods
and data analysis (2nd ed.)McGraw-Hill, Inc.

Rudi, N. (2001). Some models of risk pooling. (Ph.D., University of Pennsylvania). ,
123.

Rungtusanatham, M. J., Choi, T. Y., Hollingworth, D. G., Wu, Z., & Forza, C. (2003).
Survey research in operations management: Historical analyses. Journal of
Operations Management, 21(4), 475.

Rust, R. T., & Cooil, B. (1994). Reliability measures for qualitative data: Theory and
implications. Journal of Marketing Research, 31 (1 ), 1.

Sakakibara, S., Flynn, B. B., Schroeder, R. G., & Morris, W. T. (1997). The impact of
just-in-time manufacturing and its infrastructure on manufacturing performance.
Management Science, 43(9), 1246.

Seshadri, S., & Subrahmanyam, M. (2005). Introduction to the special issue on "risk
management in operations". Production and Operations Management, 14(1), 1.

Sethi, V., & King, W. R. (1991). Construct measurement in information systems
research: An illustration in strategic systems. Decision Sciences, 22(3), 455.

Shah, R., & Goldstein, S. M. (2006). Use of structural equation modeling in
operations management research: Looking back and forward. Journal of
Operations Management, 24(2), 148.

Sheffi, Y. (2001). Supply chain management under the threat of international
terrorism. International Journal of Logistics Management, 12(2), 1.

Sheffi, Y. (2005). The resilient enterprise: Overcoming vulnerability for competitive
advantage. Cambridge, MA: MIT Press.

Sheffi, Y., & Rice Jr., J. B. (2005). A supply chain view of the resilient enterprise. MIT
Sloan Management Review, 47(1), 41.

Shook, C. L., Jr., K., Hult, G. T. M., & Kacmar, K. M. (2004). An assessment of the
use of structural equation modeling in strategic management research. Strategic
Management Journal, 25(4), 397.

204

Smeltzer, L. R., & Siferd, S. P. (1998). Proactive supply management: The
management of risk. International Journal of Purchasing and Materials
Management, 34(1), 38-45.

Sodhi, M. S. (2005). Managing demand risk in tactical supply chain planning for a
global consumer electronics company. Production and Operations Management,
14(1), 69.

Sony Corporation. (2006). Statement regarding Sony's support of apple 's recall of
lithium ion battery packs used in apple notebook computers

Spekman, R. E., & Davis, E. W. (2004). Risky business: Expanding the discussion
on risk and the extended enterprise. lntemational Journal of Physical Distribution
& Logistics Management, 34(5), 414.

SPSS 15.0. (2007). SPSS user's guide book

Stank, T. P., Keller, S. B., & Closs, D. J. (2001). Performance benefits of supply
chain logistical integration. Transportation Journal, 41(2/3), 32-47.

Sterman, J. D. (2001). System dynamics modeling: Tools for learning in a complex
world. California Management Review, 43(4), 8.

Stratman, J. K., & Roth, A. V. (2002). Enterprise resource planning (ERP)
competence constructs: Two-stage multi-item scale development and validation.
Decision Sciences, 33(4), 601.

Svensson, G. (2000). A conceptual framework for the analysis of vulnerability in
supply chains. International Journal of Physical Distribution & Logistics
Management, 30(9), 731.

Svensson, G. (2002a). A conceptual framework of vulnerability in firms' inbound and
outbound logistics flows. International Journal of Physical Distribution & Logistics
Management, 32(1/2), 110.

Svensson, G. (2002b). A typology of vulnerability scenarios towards suppliers and
customers in supply chains based upon perceived time and relationship
dependencies. International Journal of Physical Distribution & Logistics
Management, 32(3/4), 168.

Svensson, G. (2004). Key areas, causes and contingency planning of corporate
vulnerability in supply chains: A qualitative approach. lntemational Journal of
Physical Distribution 8. Logistics Management, 34(9), 728.

Tatikonda, M. V., & Montoya-Weiss, M. M. (2001). Integrating operations and
marketing perspectives of product innovation: The influence of organizational
process factors and capabilities on development performance. Management
Science, 47(1), 151.

205

Temba, M., & Jeff, M. (2002). Corporate real estate and business continuity: An
integrated enterprise conceptualization. Joumal of Corporate Real Estate, 4(4),
348.

Tsikriktsis, N. (2005). A review of techniques for treating missing data in OM survey
research. Journal of Operations Management, 24(1), 53.

TSMC. (1999). TSMC to start partial operation with increased power supply,
from http:l/www.tsmc.com/tsmcdotcom/PRListingNewsArchivestAction

Urrutia, J. L., Vu, J., Gronewoller, P., & Hoque, M. (2002). Nonlinearity and low
deterministic chaotic behavior in insurance portfolio stock returns. Journal of Risk
and Insurance, 69(4), 537.

Vickery, S. K., Dr6ge, C., & Markland, R. E. (1993). Production competence and
business strategy: Do they affect business performance? Decision Sciences,
24(2), 435.

Wacker, J. G. (1998). A definition of theory: Research guidelines for different theory-
building research methods in operations management. Journal of Operations
Management, 16(4), 361.

Waller, M. J., 8. Roberts, K. H. (2003). High reliability and organizational behavior:
Finally the twain must meet. Journal of Organizational Behavior, 24(7), 813.

Ward, P. T., & Duray, R. (2000). Manufacturing strategy in context: Environment,
competitive strategy and manufacturing strategy. Journal of Operations
Management, 18(2), 123.

Weber, E. U., & Milliman, R. A. (1997). Perceived risk attitudes: Relating risk
perception to risky choice. Management Science, 43(2), 123.

Weick, K. E., & Roberts, K. H. (1993). Collective mind in organizations: Heedful
interrelation on flight decks. Administrative Science Quarterly, 38(3), 357-381.

Weick, K. E., & Sutcliffe, K. M. (2001). Managing the unexpected: Assuring high
performance in an age of complexity. San Francisco: John Wiley & Sons, Inc.

Weick, K. E., Sutcliffe, K. M., & Obstfeld, D. (1999). Organizing for high reliability.
Research in Organizational Behavior, 1(21), 81 -1 23.

Wilding, R. D. (1998). The supply chain complexity triangle: Uncertainty generation
in the supply chain. International Journal of Physical Distribution & Logistics
Management, 28(8), 599.

Wisner, J. D. (2003). A structural equation model of supply chain management
strategies and firm performance. Journal of Business Logistics, 24(1), 1-25.

206

management strategies and firm performance. Journal of Business Logistics,
24(1), 1-25.

Wu, T., Blackhurst, J., & O'Grady, P. (2007). Methodology for supply chain disruption
analysis. lntemational Journal of Production Research, 45(7), 1665-1682.

Yates, J. F., & Stone, E. R. (Eds.). (1992). The risk construct. New York: John Wiley
& Sons.

Yuan, K. (2005). Fit indices versus test statistics. Multivariate Behavioral Research,
40(1), 115.

Zsidisin, G. A. (2003). Managerial perceptions of supply risk. Journal of Supply
Chain Management, 39(1), 14.

Zsidisin, G. A., & Ellram, L. M. (2003). An agency theory investigation of supply risk
management. Journal of Supply Chain Management, 39(3), 15.

Zsidisin, G. A., Ellram, L. M., Carter, J. R., & Cavinato, J. L. (2004). An analysis of
supply risk assessment techniques. International Journal of Physical Distribution
8 Logistics Management, 34(5), 397.

Zsidisin, G. A., Melnyk, S. A., & Ragatz, G. L. (2005). An institutional theory
perspective of business continuity planning for purchasing and supply
management. lntemational Journal of Production Research, 43(16), 3401.

Zsidisin, G. A., Panelli, A., & Upton, R. (2000). Purchasing organization involvement
in risk assessments, contingency plans, and risk management: An exploratory
study. Supply Chain Management, 5(4), 187.

Zsidisin, G. A., Ragatz, G. L., & Melnyk, S. A. (2005). The dark side of supply chain
management. Supply Chain Management Review, 9(2), 46.

Zsidisin, G. A., Smith, M. E. (2005). Managing supply risk with early supplier

involvement: A case study and research propositions. Journal of Supply Chain
Management, 41(4), 44.

207

        

EEEEEEEEE

IIIIIIIIIIIIIIIiI’IIIIIIIlIllliIIlIIiljllllllI

3 1293 0 956 5 95