mitigating supply chain disruptions- essays on lean management, interactive complexity, and tight coupling

In this dissertation, I consider disruptions as accidents and use organizational accident theory to address how supply chain disruptions can be prevented by understanding the role of lea

MITIGATING SUPPLY CHAIN DISRUPTIONS:

ESSAYS ON LEAN MANAGEMENT, INTERACTIVE COMPLEXITY

AND TIGHT COUPLING

* * * * * The Ohio State University

2006

Dissertation Committee:

Professor Peter T Ward, D.B.A Adviser Approved by

Professor James Hill, Ph.D

Professor Paul Nutt, Ph.D

Professor David A Schilling, Ph.D Professor Steven J Spear, D.B.A Adviser

Graduate Program in Business Administration

Copyright by Kathryn A Marley

2006

ABSTRACT

The prevalence and cost implications of supply chain disruptions is the motivation for a considerable amount of academic and practitioner literature (e.g., Rice and Caniato, 2003; Hendricks and Singhal, 2003; Blackhurst et al., 2005; Hendricks and Singhal, 2005a, 2005b; Kleindorfer and Saad, 2005; Sheffi, 2005; Tang, 2006; Tomlin, 2006) In this dissertation, I consider disruptions as accidents and use organizational accident theory to address how supply chain disruptions can be prevented by understanding the role of lean management, interactive complexity and tight coupling within a system (Perrow, 1984, 1999a) I accomplish this through three related essays I address the theoretical basis for lean management conceptually in the first essay In the second and third essays, I address empirically the effects of interactive complexity and tight coupling

on the likelihood of supply chain disruptions, and the impact that different levels of these conditions have on reducing supply chain disruptions

Although lean management has attracted a great deal of attention within academic and practitioner literature, there is little research that addresses why lean management appears to work in practice In the first essay, I address the theoretical basis for lean management by drawing insights from research that considers how complex systems achieve reliability Specifically, I consider two organizational accident theories - Normal Accident Theory (NAT) and High Reliability Theory (HRT) NAT suggests that, in the

absence of countermeasures, a high degree of interactive complexity and tight coupling lead to accidents (Perrow, 1984, 1999a), while HRT argues that organizations facing these conditions may be vulnerable to accidents but can manage these conditions through application of countermeasures (Roberts, 1990a) Firms practicing lean management achieve improved operational performance by removing complexity from processes (Womack et al., 1990; Womack and Jones, 1996) Therefore, we attempt to make a theoretical contribution by connecting the observable attributes apparent in lean

management with the measurable performance being achieved to suggest why and under what conditions these attributes contribute to high levels of performance

In our second essay, we consider the impact of interactive complexity and tight coupling on supply chain disruptions Although some disruptions are the result of

abnormal events, such as hurricanes, fires, or intentional acts, the focus of this research is

on “normal” supply chain disruptions We suggest that disruptions are likely to occur under conditions of a high degree of interactive complexity and tight coupling (Perrow,

1984, 1999a) To accomplish this, we estimate the levels of interactive complexity and tight coupling of various processes in a steel processing plant and relate these to the likelihood of supply chain disruptions The results indicate that there is a significant process complexity effect, thus suggesting that process simplification can be an effective countermeasure to preventing supply chain disruptions

In our third essay, we aim to understand the structural changes that firms can make to mitigate supply chain disruptions According to NAT and HRT, the likelihood

of disruptions can be reduced by making structural changes to reduce interactive

complexity, reduce tight coupling, or attack both simultaneously (Perrow, 1999b;

Roberts, 1990a) To understand which approach works best, we compare the proportion

of supply chain disruptions from groups of processes from a steel processing plant with varying levels of interactive complexity and tight coupling We find significantly fewer disruptions under conditions of low process complexity and tight coupling and no fewer disruptions when processes are simplified and buffered with additional inventory Because lean management involves simplifying processes with reduced slack, our results support the benefits of adopting lean management practices to improve supply chain performance

Dedicated to Gregg and Colton, the loves of my life and my best friends

You fill my life with more joy and happiness than I could ever hope for

Thanks for being a part of the journey with me!

ACKNOWLEDGMENTS

I would like to thank my advisor Peter Ward for his support, advice, friendship, and mentoring throughout my Ph.D program

Thanks also to my committee members James Hill, Paul Nutt,

David Schilling, and Steven Spear

Your helpful comments and guidance are very much appreciated

Thanks to my parents and family for their endless love and support

Thanks to Gopesh and Sowmya Anand for being such helpful, supportive, loving, and loyal friends

Friends like you are irreplaceable!

The Marley’s will miss you!

Thanks to the Fisher College of Business, Center for Operational Excellence and

Lean Enterprise Institute for their financial support

VITA

July 23, 1974……… Born - Sewickley, Pennsylvania

1996………B.A Christian Ministries/Social Service, Grove City College 2000………… M.B.A., Management, University of Akron

2003………M.A., Operations Management, The Ohio State University

PUBLICATIONS Marley, K.A., Collier, D.A., Meyer-Goldstein, S., 2004 The Role of Clinical and Process Quality In Achieving Patient Satisfaction in Hospitals, Decision Sciences 35 (3), 349-

369

FIELDS OF STUDY Major Field: Business Administration

Minor Field: Quantitative Psychology and Logistics

TABLE OF CONTENTS

Page

Abstract……… ii

Dedication……… v

Acknowledgments……… vi

Vita……… vii

List of Tables……… xi

List of Figures……… xii

Chapters: 1 Introduction……… 1

2 Normal Disruptions and Ordinary Processes……… 7

2.1 Lean Management……… 11

2.1.1 The how of lean management……… 13

2.1.2 The why of lean management ……… 15

2.1.3 Theoretical consistencies with lean management………… 21

2.2 Matrix of Choices……… 25

2.3 Propositions……… 30

2.4 Illustration……… 34

2.5 Conclusion……… 36

3 Interactive Complexity, Tight Coupling, and Disruption-free

Performance………46

3.1 Literature Review……… 48

3.1.1 Supply Chain Disruptions……… 48

3.1.2 Organizational Accident Theories……… 53

3.1.3 Hypotheses……… 57

3.2 Methods……… 61

3.2.1 Sample………61

3.2.2 Dependent Variable - Supply Chain Disruptions………… 62

3.2.3 Independent Variable - Interactive Complexity……… 63

3.2.4 Independent Variable – Tight Coupling……… 66

3.3 Analysis and Results……… 67

3.4 Discussion……… 69

3.5 Conclusion……… 76

3.6 Limitations and Future Research……… 77

4 The Impact of Lean Management on Disruptions……… 84

4.1 Literature Review……… 87

4.1.1 Supply Chain Disruptions……… 87

4.1.2 Normal Accident Theory and High Reliability Theory…… 91

4.1.3 Hypotheses……… 95

4.2 Methods……… 101

4.2.1 Sample………101

4.2.2 Dependent Variable - Supply Chain Disruptions………… 102

4.2.3 Independent Variable - Interactive Complexity……… 102

4.2.4 Independent Variable –Tight Coupling……… 104

4.3 Analysis and Results………106

4.4 Discussion………109

4.5 Conclusion………115

4.6 Limitations and Future Research……… 116

List of References……… 123

LIST OF TABLES

2.1 Rules of the Toyota Production System……… 42

2.2 Strategies of High Reliability Organizations……… 43

2.3 Consistencies between NAT dimensions and lean management………… 44

2.4 Consistencies between HRT dimensions and lean management………… 45

3.1 Definition of steps in steel processing……… 80

3.2 Descriptive Statistics and Correlations……… 81

3.3 Logistic Regression Analysis……… 82

3.4 Rules of the Toyota Production System……… 83

4.1 Definition of steps in steel processing……… 119

4.2 Descriptive statistics – Process Complexity and Tight Coupling………… 120

4.3 Descriptive statistics – Product Complexity and Tight Coupling………… 121

4.4 Difference Between Proportions……… 122

LIST OF FIGURES

2.1 The tools and practices of lean management……… 38 2.2 The interaction of the tools and rules of lean management……… 39 2.3 Normal Accident Theory framework……… 40 2.4 Normal Accident Theory framework – Mitigating Disruptions………… 41

3.2 Interaction between Tight Coupling (Inventory) and Probability of a

Disruption at different levels of complexity……… 79

CHAPTER 1

INTRODUCTION

The significant cost and operational implications of supply chain disruptions are

the subject of a considerable amount of academic and practitioner research (e.g., Rice and

Caniato, 2003; Hendricks and Singhal, 2003, 2005a, 2005b; Blackhurst et al., 2005;

2005a, 2005b; Kleindorfer and Saad, 2005; Sheffi, 2005; Tang, 2006, Tomlin, 2006) For

our purposes, we define a supply chain disruption as a stoppage in production at a

customers’ plant caused by activities that occur at a supplier plant upstream Supply chain

disruptions may be the result of activities such as poor scheduling and planning,

communication, machine or transportation breakdowns, quality or computer problems, or

employee errors (Blackhurst et al., 2005; Sheffi, 2005) Because supply chain

performance can be considered a good metric, we consider the structural changes that

firms can make to reduce the likelihood of supply chain disruptions

In his article in the Academy of Management Review, Whetten (1989) discusses

the necessary elements of a theoretical contribution Building off the work of Dubin

(1978) and others from theory development disciplines, he describes a complete theory as containing four elements – (1) what, (2) how, (3) why, and (4) who, where, when The

“what” involves identifying which factors should be included to provide an explanation

of the phenomena Here, it is important to balance comprehensiveness with parsimony; including all pertinent concepts, while eliminating those that add little understanding The

“how” draws relationships by adding order and pattern to the factors In addition,

causality may be introduced According to Whetten, the “why” step is the “theoretical glue that welds the model together.” It is here where the “underlying psychological, economic, or social dynamics that justify the selection of the factors and the proposed causal relationships” is determined Lastly, the “who, what, when” step invoke the necessary limitations of the proposed theory This, most often, comes through testing of the first three steps

Using Whetten’s (1989) framework, this dissertation attempts to add a theoretical basis to lean management by considering why adopting lean management can reduce the likelihood of supply chain disruptions The essence of lean management is the creation of

a culture that encourages learning and, thus continuous incremental process improvement (Spear, 2002) The “what” and “how” of lean management are well-established in

literature They include Total Quality Management, Just-in-Time Production, Total Preventive Maintenance, and Employee Involvement (McKone et al., 1999; Cua et al., 2001; McKone et al., 2001; Shah and Ward, 2003) In addition, Spear and Bowen (1999) describe “how” the tools lead to operational advantage in their discussion of the “Rules of the Toyota Production System.” Adding theory to lean management requires us to go beyond “what” tools and practices comprise lean and “how” these tools lead to

operational advantage for companies who practice them To make a theoretical

contribution we must address “why” it is that lean works by understanding the logic

Within our study, we consider the impact of adopting lean management on the likelihood of supply chain disruptions Within the supply chain disruption literature, many researchers suggest countermeasures to mitigate supply chain disruptions These countermeasures include strategies such as adding redundancy and flexibility (Rice and Caniato, 2003; Sheffi, 2005), monitoring supplier relationships (Toby, 2005), and

increasing inventory levels (Chopra and Sodhi, 2004; Tomlin, 2006) However, largely neglected in extant literature is a discussion of how simplifying processes through lean management can reduce the likelihood of supply chain disruptions

To understand the relationship between lean management and supply chain disruptions, we draw insights from organizational theory that addresses how catastrophic accidents can be avoided This literature is useful to our study because both safety accidents and supply chain disruptions involve a “disruption to the ongoing or future output of a system” (Perrow, 1984, 1999a) Within the organizational accident literature,

we consider two theories, Normal Accident Theory (NAT) and High Reliability Theory (HRT) The basic notion of NAT is that accidents are inevitable under conditions of tight coupling and a high degree of interactive complexity (Perrow, 1984, 1999a) Tight coupling refers to the level of slack or buffer within the system and interactive

complexity refers to the way that “parts within a system are connected and interact” (Sagan, 1993) To reduce catastrophic potential, NAT researchers suggest that firms either add slack to their system or reduce complexity (Perrow, 1999b) HRT researchers address the various countermeasures that some organizations have adopted to remain remarkably error-free despite operating under conditions of tight coupling and a high

such as increasing training and redundancy and decentralizing decision-making are suggested Both theories provide a lens to explain the lack of disruptions achieved by successful practitioners of lean management

In Chapter 2, we conceptually develop the argument of why lean management works in practice Within this chapter, we draw comparisons between the characteristics

of lean management firms and research in NAT and HRT to suggest how the dimensions

of interactive complexity and tight coupling impact the success of firms practicing lean management Specifically, we argue that reducing complexity while maintaining slack gives firms operational advantage over firms with more complex processes Reducing complexity enables firms to streamline their processes, eliminate waste and increase the overall reliability of the system (Womack and Jones, 1996; Spear, 2004) In addition, operating with small buffers of inventory allows firms to be responsive to changes within their systems because process steps are not decoupled by slack resources (Levy, 1997)

We conclude this chapter with propositions suggesting that firms practicing lean

management will be characterized by low levels of interactive complexity and tight coupling and that these dimensions will lead to a reduced likelihood of supply chain disruptions In addition, we suggest that some of the strategies suggested by HRO

researchers enhance the reliability of lean management firms, providing additional protection against disruptions

The last step in Whetten’s (1989) discussion on what constitutes a theoretical contribution involves testing of the proposed theory Therefore, the next two chapters of the dissertation are essays empirically testing the relationship between interactive

we present literature on supply chain disruption mitigation countermeasures and suggest that the NAT and HRT dimensions of a high level of interactive complexity and tight coupling lead to increased supply chain disruptions Within our analysis, we consider interactive complexity to be comprised of process and product complexity In addition,

we characterize the level of tight coupling in the system as the amount of inventory before, within, and after processes Our research hypotheses for Chapter 3 are that a high level of interactive complexity and low inventory individually lead to an increased

likelihood of supply chain disruptions, as well as the interaction between the conditions

To test these hypotheses, we use logistic regression on archival data from 451 orders from a steel processing plant We measure the level of product complexity, process complexity, and inventory for each order and determine whether or not these orders caused disruptions at customer plants The results from this analysis provide insights into how reducing complexity can be considered an internal countermeasure to mitigate supply chain disruptions

In Chapter 4, we consider how altering the dimensions of interactive complexity and tight coupling can be considered a mitigation strategy to reduce the likelihood of supply chain disruptions Within the organizational accident literature, researchers suggest that to reduce the likelihood of disruptions, firms can reduce complexity, increase slack, or perform both simultaneously (Perrow, 1984, 1999a; Roberts, 1990a) We

suggest that firms practicing lean management are successful because they have reduced the complexity within their systems concomitant with a reduction in slack (Womack et al., 1990; Womack and Jones, 1996) Therefore, in this study, we argue that there will be

complexity and low inventory than for orders characterized by (1) high process

complexity and low inventory, (2) high process complexity and high inventory, and (3) low process complexity and high inventory Using a dataset of orders from the steel processing plant, we test our hypotheses using a standardized z test The results from this analysis will provide managers with insights on how lean management can be used as a strategy to mitigate disruptions

Within this dissertation, we attempt to make a theoretical contribution by

understanding the interplay of lean management principles, interactive complexity and tight coupling and their effect on supply chain disruptions Drawing insights from

organizational accident literature, we present a conceptual and empirical argument for why lean management works in practice We address empirically the relationship

between interactive complexity and inventory levels and supply chain disruptions, as well

as how altering the levels of interactive complexity and inventory can be considered a countermeasure for reducing supply chain disruptions

CHAPTER 2

NORMAL DISRUPTIONS AND ORDINARY PROCESSES

Lean management has attracted a great deal of attention within academic and practitioner literature because it suggests an approach that leads to both high performance and continuous improvement (e.g., Womack et al., 1990; Womack and Jones, 1996; Spear and Bowen, 1999; Shah and Ward, 2003; Hines et al., 2004; Liker, 2004; Bhasin and Burcher, 2006; Liker and Morgan, 2006) The essence of lean management is the creation of a culture that encourages learning and, thus continuous incremental process improvement through simplifying the way that work is performed (Womack and Jones,

1996, Spear, 2002) This emphasis on continuous improvement is achieved in part by applying a set of well delineated principles to identify and eliminate waste from all aspects of the organization (Womack and Jones, 1996) Within literature there is no shortage of examples of firms that have adopted lean management practices Although the roots of lean management can be found in the Toyota Production System (Sugimori, 1977; Monden, 1981, Spear and Bowen, 1999; Liker, 2004), firms in industries such as healthcare (Bushell et al., 2002), steel (Proctor, 1997; Dhandapani et al., 2004;

shipbuilding (Storch and Lim, 1999), textile and clothing (Bruce et al., 2004) and

publishing (Huls, 2005) are finding ways to apply lean management principles and achieve operational success

Within these firms, a well-defined set of lean management tools such as pull systems, quality management programs, preventative maintenance programs and

employee involvement programs are consistently applied (Shah and Ward, 2003) However, a question that remains unresolved is why adopting these lean management tools leads to improved performance In other words, are there certain conditions that exist within the firm that drive the successful application of lean practices? Determining these conditions not only provides a theoretical basis for lean management, but also enables managers to better understand the structural and infrastructural characteristics that support a lean management initiative

To address the theoretical basis for lean management, we argue that many of the firms that adopt lean can be characterized as complex work systems Complexity

manifests itself in a variety of ways within these organizations – appearing within

individual processes, products, and relationships with suppliers and customers Firms managing these complex work systems pursue lean management to achieve improved reliability, responsiveness, and productivity (Levy, 1997; Pavnaskar et al., 2003; Spear, 2004) However as complexity increases, firms become more vulnerable to experiencing disruptions (Roberts and Libuser, 1993) We argue that the theoretical basis for lean management lies in understanding how to overcome the complexity inherent in these systems to achieve disruption-free performance

Within the organizational theory literature, there is a body of knowledge that addresses how to manage complex systems when the consequences from accidents or disruptions are very great We refer to this literature as organizational accident theory and consider contributions from two theories - Normal Accidents Theory (NAT) (e.g., Sagan, 1993; Perrow, 1984, 1999a) and High Reliability Theory (HRT) (e.g Roberts, 1990a, 1990; Weick and Sutcliffe, 2001) Thus, to help explain how lean management appears

to lead to success in avoiding disruptions in more prosaic operational settings, we draw insights from models developed to describe the dynamics of managing risk in situations that are inherently perilous

A common thread throughout organizational accident literature is the notion of managing interactive complexity Interactive complexity “refers to the presence of unfamiliar or unplanned and unexpected events in a system that are either not visible or not immediately comprehensible”(Marais et al., 2004) As firms and their supply chains expand and the number of communication and technological links increase, they become more interactively complex Dealing effectively with interactive complexity is at the heart of the literature on reliability in organizations facing perilous risk

Charles Perrow has emphasized the vulnerability of complex systems to

unintended and unanticipated consequences Perrow’s book, Normal Accidents, (1999), describes how these undesirable outcomes are directly the result of systems having so many elements, connected in so many ways, that no one is capable of understanding fully the system’s structure or dynamics To avoid the inevitability of a (normal) accident under conditions of very tight coupling or little slack, firms should attempt to reduce the number of elements and number of interconnections within the system to make itself more tractable (Perrow, 1999b)

High reliability organization researchers have approached a question similar to that of NAT but have reframed it slightly from “how do complex organizations fail?” to

“how do complex organizations succeed?” They characterize “high reliability

organizations (HROs)” as those that have very few accidents despite operating under complex and demanding conditions (Roberts, 1990a, 1990b; Weick and Sutcliffe, 2001) Proponents of high reliability theory (HRT) recognize that characteristics such as high interactive complexity and tight coupling can be potentially damaging to an organization; however, they believe that strategies such as redundancy, accountability, responsibility, and a “culture of reliability” can mitigate the effects (Roberts, 1990a)

Addressed in NAT and HRT are two constructs that are fundamental to the

concept of lean management – interactive complexity and organizational slack Because both theories also emphasize several principles that are fundamental to the ideas of lean management (standardization, reliability, learning, training), they provide a conceptual basis for understanding why the tools and practices inherent to lean management are so

effective in achieving disruption-free performance in operations where the consequences

of failure take a financial rather than human toll We discuss the important concepts in NAT and HRT and how they can be applied to explain why and under what conditions lean management serves to avoid disruptions that were previously considered “normal.”

The essence of lean management is the notion of simplifying processes Womack and Jones (1996) suggest that simplification can be accomplished by “identifying all activities that are muda (waste) and eliminating them.” Wasteful activities are those that

do not add value from the customer perspective By singling out the value added steps in the process, the flow between activities becomes less jumbled and more linear and

problems and defects are easily identified and eliminated (Womack and Jones, 1996; Rother and Shook, 2003) Lean management firms focus on continuous improvement that extends within the plant and throughout the enterprise (Womack and Jones, 1994)

The notion of simplification extends to the programs that comprise lean

management Previous research has addressed the specific tools and techniques that are practiced in lean systems that, when adopted by firms, can contribute to high

performance It is well established that there are four elements within lean management– Just-in-Time (JIT), Total Quality Management (TQM), Total Productive Maintenance (TPM), and Employee Involvement (EI) (McKone et al., 1999; Cua et al., 2001; McKone

et al., 2001; Shah and Ward, 2003) JIT accounts for the flow of production and includes such elements as reduced setups and lot sizes, pull systems, and kanbans (Davy et al., 1992; Flynn et al., 1995; White, 1999; Cua et al., 2001; McKone et al., 2001; Shah and Ward, 2003; Kannan and Tan, 2005) TQM describes an organization-wide focus on quality, not only within the creation of the product or service, but also in the development

of procedures to address how employees can improve the quality of their work Practices such as Statistical Process Control, Benchmarking, Process Capability Analysis are included here (Flynn et al., 1995; Hackman and Wageman, 1995; Powell, 1995; Ahire et al., 1996; Dow, 1999; Cua et al., 2001; McKone et al., 2001; Shah and Ward, 2003)

TPM involves autonomous and planned maintenance activities that support a company’s efforts in keeping equipment in proper shape throughout its entire life, by involving all levels of employees in regular procedures (McKone et al., 1999) Activities such as housekeeping (5S), cross-training maintenance, and schedule compliance are included (McKone and Weiss, 1998; Cua et al., 2001; McKone et al, 2001; Shah and Ward, 2003) EI accounts for the role of employees in achieving the goals and objectives

of the above programs Developing cross-functional teams, empowering employees and decentralizing decision-making are ways that firms accomplish the goals of lean

management (Pil and MacDuffie, 1995; MacDuffie, 1996; Cua et al., 2001; Shah and Ward, 2003)

In addition to these four elements, firms practicing lean management place critical importance on their relationships with customers and suppliers (Womack et al., 1990; Womack and Jones, 1996) This involvement is often referred to as the “lean enterprise.” Value is defined by the customer, created by the firm, and delivered with the help of the supplier network (Womack and Jones, 1994; Womack and Jones, 1996) Accomplishing this may involve supplier development programs or associations, customer involvement

in product design, or supplier rationalization (Womack et al., 1990; Levy, 1997;

MacDuffie and Helper, 1997) The tools and practices of lean management are shown in Figure 2.1

2.1.1 The how of lean management

The practices that comprise lean manufacturing only tell a part of the story A pressing issue for operations management researchers and practitioners is explaining

“how” implementing lean practices leads to improved performance Lean management advocates believe that implementing the aforementioned lean tools is not sufficient Rather, the key to competitive advantage is through rates of improvements that are both greater than those of their competitors and experienced across a broader range of

processes and functions (Spear and Bowen, 1999; Spear, 2002)

This was made apparent by Spear and Bowen (1999) in their extensive study of the Toyota Production System, the archetypal lean management model Over the years, Toyota had achieved success through its implementation of lean management and they continue to be open in sharing their practices with others However, imitators have found

it difficult to duplicate Toyota’s success Spear and Bowen argue the key to Toyota’s

success was not in its cultural roots nor the tools and practices they adopted but rather in

the tacit knowledge developed organically through application of four rules (Table 2.1)

Lean management as embodied by the Toyota Production System combines doing work with learning to do work better This is accomplished because all work, be it in design, logistics, manufacturing, administration, training, and so forth, is conducted as a series of nested, ongoing experiments (Spear, 2004) Expectations are made clear before work is done While work occurs, tests embedded in the work make it apparent

immediately when results contrary to expectations are occurring When a problem occurs,

it is immediately addressed both to dampen its effects and prevent the problem from propagating in unexpected ways in downstream operations Employees are expected to discover the underlying reason for a problem as a prelude to redesigning the work to prevent the problem from recurring This immediate resolution of problems means that the organization is continuously investigating the assumptions underlying its work design that has been challenged by results Furthermore, redesign of work is carried out in a highly structured, scientific, experimental manner so that the assumptions in the redesign, just like the assumptions in the initial design are quickly tested in practice (Spear, 2002, Spear, 2004)

The lean tools and practices and Spear and Bowen’s work rules interact to form the mechanism that drives lean management The four lean management programs can

be divided into two concepts with corresponding rules The practices of JIT, TQM, and TPM, and Supply Chain Practices are best explained by understanding the way that work pathways, and connections are specified and employee involvement entails making improvements under the instruction of teachers within the firm (Figure 2.2)

2.1.2 The why of lean management

To better understand “why” lean management leads to improved performance, we draw on concepts from NAT and HRO theories We accomplish this by examining the dynamics that cause catastrophic safety accidents and relate these dimensions to more prosaic disruptions that occur within plant or service operations Perrow’s definition of

an accident is useful because he defines an accident as a “failure in a subsystem or system

as a whole that damages more than one unit and in doing so disrupts the ongoing or future output of the system.” Because a missed shipment by a plant within a supply chain can cause a stoppage in the production or services of a downstream customer, this

definition could be applied to both manufacturing and service facilities Because of the strong emphasis on operational efficiency in firms pursuing lean management, the

frequency and severity of supply chain disruptions can be a good performance metric for firms

Normal Accident Theory (NAT) resulted from an organizational analysis by Charles Perrow on the Three Mile Island disaster (Perrow, 1984, 1999a) In his study of high-risk complex technological systems, he introduced the notion that accidents are

inevitable or “normal” when two conditions exist - a high degree of interactive

complexity and tight coupling NAT theory suggests that when these conditions are present, humans are unable to anticipate accidents When there is a high degree of complexity, even small seemingly independent failures can interact in unexpected ways that can not be anticipated by process designers nor understood by operators If a highly complex system has very little slack or buffer between stages, the inevitable failures will intensify uncontrollably causing a serious accident (Perrow, 1984, 1999a; Rijpma, 2003)

Interactive complexity consists of two dimensions High interactive complexity refers to processes of “unfamiliar, unplanned and unexpected sequences, and either not visible or immediately comprehensible.” Low interactive complexity refers to “expected and familiar production or maintenance sequences, and those that are quite visible even if unplanned.” Tight coupling refers to the amount of slack or buffer that exists between system components and ranges from high to low High tight coupling implies that there

is little slack or buffer within the system or it is not possible to delay processing, while low tight coupling refers to excess slack, buffers or time (Perrow, 1984, 1999a)

Figure 2.3 captures the NAT framework Systems such as nuclear power plants are found in the upper right quadrant of the grid, or where there is both a high level of interactive complexity and high tight coupling According to NAT, organizations that exist in this quadrant, where the two conditions are present, will eventually experience accidents Therefore, it is important for firms to make the necessary changes in the form

of either adding slack or reducing complexity to mitigate against disasters, however Perrow (1999b) acknowledges that doing both in unison will most likely make the system inefficient

Other researchers have used the NAT as a framework for their research Wolf (2001) examines the relationship between complexity, coupling and safety performance

in a sample of 36 petrochemical plants In a subsequent study, Wolf (2005) controls for the effects of complexity and coupling while investigating the relationship between resource availability and resource commitment on safety performance Gephart, (2004) traces how insights from NAT are used to study risk sense making and ecosystems

accidents, while Sagan (1993) examines the vulnerability of the nuclear weapons

management in the context of NAT and Weick (2004) uses NAT to frame the problem of the misdiagnosis of West Nile Virus by the Center for Disease Control and Prevention

HRT generally acknowledges that organizations may be prone to accidents under conditions of tight coupling and high interactive complexity (Roberts, 1990a) Therefore, HRT researchers have devoted a great deal of effort to understanding what

countermeasures firms can adopt to prevent accidents when these two conditions exist Much of the research has focused on specific organizations that have achieved nearly

perfect safety records, in spite of these conditions Roberts (1990a, 1990b) defines high reliability by stating that “within the set of hazardous organizations there is a subset which has enjoyed a record of high safety over long periods of time One can identify this subset by answering the question, ‘how many times could this organization have failed resulting in catastrophic consequences that it did not?’ If the answer is on the order

of tens of thousands of times the organization is ‘high reliability.”

Because of their similar themes, there has been considerable debate over the merits of HRT and NAT (Sagan, 1993; LaPorte and Rochlin, 1994; LaPorte, 1994;

Sagan, 1994; Bain, 1999) Sagan (1993) refers to the HRT advocates as “optimistic” because they believe that accidents can be prevented by adopting certain strategies, and refers to NAT advocates as “pessimistic” because they believe that despite adopting strategies, accidents are inevitable However, the two theories generally are consistent with respect to causal variables NAT ascribes tight coupling and interactive complexity

as precursors to accidents HRO theorists acknowledge that interactive complexity and tight coupling may leave organizations vulnerable to catastrophic risk, however they

address the ways high reliability organizations manage these conditions to avoid or delay

disruptions (e.g., Roberts, 1990a, 1990b, Weick et al., 1999; Weick and Sutcliffe, 2001)

As discussed within the context of the Normal Accidents Theory, complexity within an organization involves such characteristics as confusing interactions, complex technology and indirect information channels (Roberts, 1990a) HROs respond to these often unexpected situations by adopting specific operational goals promoting stability and safety through all levels of the organization To deal with unexpected interactions,

employees are involved in extensive training using simulations of a wide range of

possible scenarios that they may encounter In addition, redundancy is added to systems

to provide backup support in the event that one part or unit fails (Roberts, 1990a, 1990b)

To handle complex technologies, managers give lower level employees a high degree of responsibility and accountability to make decisions and fix problems that occur within their domain (Roberts et al., 1994) Operators are trained to keep potentially interacting systems separate to avoid unexpected interactions The complexity of indirect

information is alleviated through the addition of direct information sources connecting key areas within the organization (Roberts, 1990a, 1990b)

Tight coupling may refer to time dependent processes or little slack (Perrow,

1984, 1999a) Redundancy enables organizations to maintain reliability under strict time pressures by increasing the number of employees assigned to certain tasks at busy times when delays are not possible Also, having back-up systems in place protects against unexpected maintenance failures (Roberts, 1990a) When there is little slack, HROs use bargaining and negotiation and system flexibility Bargaining and negotiation refers to upper level employees gathering information from employees negotiating with them on how “to reach a minimax solution” to meet simultaneous demands Creating a flexible system provides employees with a variety of options to pursue when the unexpected occurs (Roberts, 1990b) The HRO strategies used to managed complexity and tight coupling are summarized in Table 2.2

An extension to the High Reliability Theory is the idea that there are specific processes that organizations adopt to manage unexpected events Weick and Sutcliffe (2001) attribute high reliability to “mindfulness.” By this, they mean that the

organizations are very capable of identifying changes in circumstance, recognizing when these require changes in behavior, and actually changing behavior so that the change in circumstances is handled without causing disruption High reliability organizations (HROs) achieve “mindfulness” through five distinct mindsets (1) They are exceptionally sensitive to failures in that they are continuously looking for what might be going wrong, (2) both when things are going right and when something goes wrong, they are careful to avoid simplifying their mental models and explanations, (3) related to the first two points, they are focused on operations, the processes by which work is done, (4) they are

concerned with resilience, responding to problems quickly to dampen their effects, and (5) they are deferential to expertise, allowing those best able to solve problems to make decisions rather than deferring to authority that is based solely on hierarchical status (Weick et al., 1999; Weick and Sutcliffe, 2001)

A nuclear aircraft carrier is considered an example of a mindful organization Operating in what is what is often called “the most dangerous four and one-half acres in the world,” employees aboard carriers are exposed to large machinery, dangerous

weapons, high speeds, and constant activity 24 hours a day (Weick and Sutcliffe, 2001)

In addition, personnel face many complex interactions within their work and significant time pressure between activities (Roberts, 1990a) Despite these conditions, nuclear aircraft carriers can be considered mindful organizations because of the way they achieve

reliability in their dangerous environment Landings are graded and televised within the ship to improve performance, and near misses are discussed and analyzed promptly after occurring Aircraft are checked thoroughly and by multiple personnel to ensure their safety Officers communicate continuously with pilots in flight, gathering information on the status of the aircraft Flight crews value “the importance of routines and predictable behavior, as well as doing what they are told” to adhere to a commitment to resilience Lastly, pilots with more relevant knowledge about their aircraft may override the

commands of higher-ranking officers if safety is in question (Weick and Sutcliffe, 2001)

2.1.3 Theoretical consistencies with lean management

To draw insights from these theories on lean management, we must be able to relate safety accidents to more prosaic supply chain disruptions As stated previously, the definition of an accident presented in the NAT and HRT is useful because of its emphasis

on the disruption to the future output of a system Because of the strong emphasis on operational efficiency in firms practicing lean management (e.g Womack et al., 1990; Shah and Ward, 2003; Liker, 2004) the frequency and severity of supply chain

disruptions can be a good performance metric for firms

Tight coupling, as defined by Perrow, is a central concept in lean management As firms attempt to eliminate the amount of waste and non-value added time within their organization, they often focus on reducing the buffers and slack within their systems (Levy, 1997) These buffers may come in the form of finished goods, work-in-process and raw materials inventory (Liker, 2004) Value stream mapping, a key lean

management tool, helps firms to balance their production steps to reduce the temporal

delays between them and remove excess inventory, characterized as “overproduction” (Rother and Shook, 2003) After eliminating inventory buffers, managers’ attention can

be focused on “improving the quality of inputs, keeping tight control over the production process, reducing lead and cycle times at every stage, reducing lot sizes and set-up times, and shortening product development cycles” (Levy, 1997)

Because the market rewards firms that use less capital in the form of inventory or other slack resources, most firms strive to reduce slack resources Literature suggests that firms practicing lean management have been particularly successful in achieving such reductions (Womack and Jones, 1996; Liker, 2004) We argue that firms practicing lean management are able to reduce slack and thus tighten coupling without exposing themselves to the risk of additional disruptions They achieve these reductions in slack concomitantly with a systematic reduction in interactive complexity that offsets increased probability of system disruption A high degree of interactive complexity exists when there are a number of control parameters that exist within an operation that have the potential to interact in unexpected ways (Wolf, 2001, 2005) In other words, as the number of potential interactions between people, machines, and process increases, it becomes more difficult for operators to perceive and handle unexpected failures Under lean management principles, system design favors simple linear systems as opposed to systems with numerous points of interaction and unnecessary feedback loops (Womack and Jones, 1996) Work is setup in u-shaped cells to eliminate extensive (and wasteful) travel between production steps This also enables workers to focus on their specific tasks without the potential for many unfamiliar interactions and needless feedback loops

(Liker, 2004) Information between production steps is transferred directly via a kanban system, for example, that precisely signals the quantity of products required by the next production stage (Rother and Shook, 2003) Standardizing work by creating specific instructions on how specific tasks are to be completed improves the quality at each step

of the process by reducing the level of variability (Spear, 2004) Reducing complexity through implementing lean management also aids firms in reducing the time it takes to complete work and increase their productivity and lower costs through improved quality (Krafcik, 1988) Table 2.3 summarizes our discussion of the consistencies between NAT and lean management

A preoccupation with failure is essentially a method of learning for High

Reliability Organizations Any chance they can gain information from a failure or a close call is translated into an opportunity for improvement A critical component of this

concept is encouraging employees to report errors and to suggest improvements (Weick

et al., 1999) Similarly, Toyota and other lean management practitioners encourage employees to seize every defect as a learning opportunity They also explicitly teach workers how to improve by following the rules of the scientific method (Spear and

Bowen, 1999) A central organizational function that is empowered in this way is the maintenance department Because it is in such close contact with failures at all levels of development, workers can create a large “database for learning” that is useful throughout the firm (Weick et al., 1999) Lean firms use Total Productive Maintenance programs in

a similar manner By carving out time within the workday to clean and maintain

equipment, workers become more familiar with the equipment’s mechanics and more educated on how to make adjustments (McKone and Weiss, 1998)

In arguing for mindfulness, Weick and Sutcliffe (2001) recommend resisting the urge to make simplifications to come to quicker conclusions They advise that there be involvement from people with diverse backgrounds and experience and from multiple functions to solve problems This enhances the learning that can come from the failure

by increasing the variety of opinions contributing to it (Weick et al., 1999) Similarly, lean management firms’ pursuit of continuous improvement requires the involvement of all functions within the organization to succeed (Liker, 2004) As Weick and Sutcliffe prescribe, lean management firms choose to devote the time and resources necessary to solve problems thoroughly by conducting ongoing experiments Such controlled

experimentation allows the organizations to develop and evolve standardized procedures that make the work less complex with fewer steps and fewer handoffs (Spear, 2004)

Weick and Sutcliffe (2001) suggest cross-training to achieve flexibility and

balance Lean management practice emphasizes extensive cross-training of employees to aid in problem solving and reduce variability of work (Shah and Ward, 2003; Liker, 2004) All levels of the organization contribute in a lean environment in pursuing the goal of continuous improvement (Fairris and Tohyama, 2002) The key programs in lean, JIT, TQM, TPM, are process-focused making this characteristic of mindfulness an

essential component and goal in the adoption of lean This also relates to the fourth characteristic of mindfulness - a commitment to resilience Resilience involves

responding to errors, diagnosing problems, and learning from them (Weick and Sutcliffe,

2001) Once again, lean firms operate under the pursuit of zero defects and continuous improvement (Womack et al., 1990) This is accomplished through ongoing

experiments, training, and employee empowerment When a problem occurs,

workarounds are not encouraged, but rather a careful analysis of the cause of the problem and its possible solution is pursued (Spear, 2004)

Lastly, mindfulness involves a “deference to expertise” meaning that decisions can be made by the person with the most expertise pertaining to the situation, regardless

of their rank within the organization (Weick and Sutcliffe, 2001) This allows high reliability organizations to quickly react to problems without pursuing a chain of

authority One of the “Rules of the Toyota Production System” emphasizes that

improvements should be made at the operational level of the organization where

problems occur under the guidance of a teacher The use of a teacher helps assure that organizational goals are being addressed while still employing the expertise of process level workers (Spear and Bowen, 1999) We summarize the consistencies between

HRO/mindfulness theory and lean management in Table 2.4

uninformed on the specifics of the menu Prior to the guests arriving, the staff scrambles

to prepare the food due to limited manpower and time Throughout the evening, large orders come in simultaneously, leaving the staff incapable of handling the complexity of the orders under such tight time pressures, and the guests leave hungry and unhappy

One option that quadrant IV firms have is to move toward quadrant I Such a move implies simplifying by removing interactions, while adding slack or buffer to their systems The plausibility of this choice, however, is in question, because it seems to involve overcompensating By simplifying, failures are made more visible and

comprehensible so adding buffers in the form of time, space, inventory or resources may add high costs to systems where it is not necessary Here, a chef simplifies the menu significantly which allows the current staff to be able to prepare the food in a timely and uncomplicated manner However, he hires additional staff and requires that they arrive several hours early, therefore experiencing significant labor costs that can be deemed unnecessary and wasteful

Another choice that firms have is to move to quadrant II by remaining complex but adding slack or buffer into their system This may be an option for some

organizations operating in environments where tight coupling is not considered a

competitive advantage and simplifying is structurally impossible However, once again, there will be costs associated For firms, choosing this option, decreased disruption potential will come at a price and in environments that emphasize cost cutting, this option may not be desirable In this scenario, the chef retains the complicated menu, however,

he prepares in advance by hiring additional staff to compensate for the complexity of the

meal The staff arrives early, sufficiently informed on the work they should be

completing, however, because of the number of workers involved, there are a

significantly higher number of interactions between them and the chance for disruption remains What distinguishes this scenario from the tightly coupled, interactively complex scenario, is the opportunity to recover Extra time is built into the schedule to account for mishaps that may occur The additional labor hours, however, come at a price for the restaurant management

The last option is moving to quadrant III, which involves remaining tightly coupled, but decreasing complexity by simplifying operations While loosening up coupling may seem like a first line of defense to protect against disruptions, in many situations it may increase the complexity of the process because the number of

interactions between people and resources increases Without inventory buffers, firms are forced to discover more timely and efficient ways of doing work Because

complexity hinders these efforts, simplification becomes a likely course of action Here, the chef chooses a menu of items that are tasteful, yet simple to prepare He

systematically organizes the steps to prepare each item and utilizes a well-informed and skilled staff Guests receive hot meals on time by a manageable number of workers By reducing the complexity of the meals, the chef was able to reduce labor costs by not hiring additional workers and in turn met the expectations of his clientele

There is a logical relationship between the option just described and the essence

of lean management Firms that adopt lean management principles are committed to meeting the needs of their customers in the most efficient way possible This often