Human Failure, Organizational Change & Culture
The relationship of organizational change and human failure on Hofstede’s cultural dimensions
Because human performance is also dependent on cultural factors, this thesis investigates whether the established relationship between organizational changes and human failure features a cultural pattern of occurrence as well. In order to render an acceptable degree of comparison, the loss events are aligned on Hofstede’s cultural dimensions, power distance, uncertainty avoidance, individualism, masculinity and long-term orientation.
This study concludes, that a society’s uncertainty avoidance and its individualism are related to the occurrence of large human failure loss events. While a society’s high uncertainty avoidance is negatively correlated, a society’s high individualism is positively correlated with human failures.
It is further proposed, that a large power distance often prevents a workforce from committing human failures when their organization is changing. Trust in the vertical hierarchy gives them security. On the other side, high individualism aggravates human failures during organizational changes. The employees know that they are on their own, and that they have nobody to rely upon in insecure times.
1.1. The Framework of the First Study
1.2. The Framework of the Second Study
1.3. The Shape of the Thesis
2. The Three Elements of the Study
2.1. Human Failure
2.2. Organizational Change
Culture and Organizations
3.1. Global Loss Event Database
3.2. Human Factor Database
Cause of Event
3.3. Analyzing Culture – Existing Measurements and their
Advantages and Disadvantages
Kluckhohn & Strodtbeck’s Cultural Orientations
Laurent’s Cultural Diversity of Western Conceptions of Management
Hall’s Compass Model
3.4. Hofstede’s Culture Dimensions
Individualism – Collectivism
Masculinity – Femininity
Long-term orientation – Short-term orientation
Replication and Validation
3.5. Statistical Methodology
4. Analysis and Results
4.1. Human Failure
4.2. Organizational Change - Human Failure
4.3. National Culture – Organizational Change – Human Failure
Organizational Change – Human Failure
National Culture – Organizational Change – Human Failure
International Standard Industrial Classification (ISIC)
Figure 1 Influences on Human Performance
Figure 2 Human Failure
Figure 3 Culture
Figure 4 Form for establishing Cause of Event
Figure 5 Form for establishing Organizational Changes
Figure 6 Hall’s Compass Model
Figure 7 Human Failure and Non-Human Failure Loss Events over Time and Severity
Figure 8 Number of Loss Events over Time Categories
Figure 9 Percentage of Human Failure in Top 100 and Bottom 100 Loss Events
Figure 10 Severity of Human Failure Loss Events in different Industries
Figure 11 Severity of Human Failure and Non-Human Failure Loss Events in different Industries
Figure 12 Severity of Human Failure and Non-Human Failure Loss Events in Petrochemical Industry
Figure 13 Severity of Human Failure and Non-Human Failure Loss Events in Energy Supply Industry
Figure 14 Human Failure Loss Events and detected Organizational Changes
Figure 15 Non-Human Failure Loss Events and detected Organizational Changes
Figure 16 Human Failure and Non-Human Failure Loss Events – with and without Organizational Changes
Figure 17 Percentage of detected Organizational Changes in connection with the 50 largest and 50 lowest Human Failure Loss Events
Figure 18 Human Failure Loss Events over Time Categories
Figure 19 Number of detected Organizational Changes with Human Failures
Figure 20 Severity of Human Failure Loss Events wit Organizational Changes
Figure 21 Occurrence of different Types of Organizational Changes
Figure 22 Human Failure and Non-Human Failure Loss Events – with different Types of Organizational Changes
Figure 23 Severity of Human Failure Loss Events with different Types of Organizational Changes
Figure 24 Fallacy of the Study
Figure 25 All analyzed countries with a high score on the cultural dimension describing Uncertainty Avoidance
Figure 26 Number of analyzed Loss Events per Country
Figure 27 All analyzed countries with a high score on the cultural dimension describing Power Distance
Figure 28 All analyzed countries with a high score on the cultural dimension describing Individualism
Table 1 Macro Factors
Table 2 Types of Organizational Change
Table 3 Kluckhohn & Strodtbeck’s Value Orientations and their Variations
Table 4 SWOT-Analysis for Kluckhohn & Strodtbeck’s Value orientation
Table 5 SWOT-Analysis for Laurent’s National Cultural Differences
Table 6 SWOT-Analysis for Hall’s Compass Model
Table 7 Power Distance Index for all analyzed countries
Table 8 Uncertainty Avoidance Index for all analyzed countries
Table 9 Individualism Index for all analyzed countries
Table 10 Masculinity Index for all analyzed countries
Table 11 Long-Term Orientation Index for all analyzed countries
Table 12 SWOT-Analysis for Hofstede’s Cultural Dimensions
Table 13 Human Failure and Severity – Correlation
Table 14 Human Failure and Organizational Change – Observed Frequencies in absolute Numbers and Percentages
Table 15 Human Failure and Organizational Change – Expected Frequencies in absolute Numbers
Table 16 Human Failure and Organizational Change - Correlation
Table 17 Human Failure and different Types of Organizational Change – Association
Table 18 Human Failure and different Types of Organizational Change – Correlation
Table 19 Severity of Human Failure and different Types of Organizational Change – Correlation
Table 20 Human Failure and Cultural Dimensions – Correlation
Table 21 Human Failure, Organizational Change and Cultural Dimensions – Correlation
Table 22 Human Failure, Types of Organizational Change
and Cultural Dimensions – Correlation
Table 23 International Standard Industrial Classification
Changes of organizational structures in a company influence the behavior of its personnel (Ferrie et al., 1998). Layoffs, mergers, restructurings and reorganizations can cause stressful situations and anxiety within the workforce (Vaherta et al., 2004; Kivimäki et al., 2000), which in turn may lead to unintended actions with possibly undesirable consequences. Should these actions result in an accident, the cause of the possible ensuing losses is referred to as human failure. Every system dependent on human performance is prone to such human failures, whereas this condition has to be regarded as normal and inevitable (Williamson & Webb, 1993). But the proof of what exactly provoked the unsafe act, can seldom be conclusively established and classified; too complex are psychological behavior patterns, too coarse the granularity of available data. Additionally, the retrospective ascertainment of contextual factors at the moment of the unsafe act and the theoretically predetermined causal entanglement between action and consequence prove to be difficult to accomplish (Reason, 1990).
As part of an internship at Swiss Reinsurance Company Ltd. (Swiss Re) – one of the world’s largest reinsurer, providing diversified reinsurance products and financial services – in the department of Risk Engineering Services Casualty, I was able to conduct a study on the topic of “Organizational Changes and Human Failure” (see Mellert & Wilke, 2008). This work mainly consisted of building up a database comprising detailed information about 500 large loss events, and a following analysis of the accumulated data. The combination of the fact, that due to a lack of time a relevant part of available data remained unstudied, and a rising interest for this field of research, led to a continuation of the project. Not only should the remaining cases be analyzed as the ones before, but the whole study was to be extended and complemented with one more question to be answered: What role does the national culture at the organizations location play, when it comes to organizational changes and the prevention of possible human failures? In the following section, the underlying research questions and hypotheses shall be explained and a rough outline of the whole project will serve as a basis for a better understanding of the subsequent chapters.
1.1. The Framework of the First Study
Amongst others, the Health and Safety Executive (HSE) – a non-governmental organization in the United Kingdom conducting research in the field of workplace health and occupational risks – describes that human performance in a system where man interacts with technology is influenced on different levels (HSE, 2005; Badke-Schaub et al., 2008).
illustration not visible in this excerpt
Figure 1 Influences on Human Performance (after HSE, 2005)
On the lowest level stands the individual with its own attitude, its emotions and risk perception, etc. (see also Rundmo, 2000; Dörner, 2008). On the second level, the job itself with its rules, the task, the workload and supervision, etc. influence and steer human performance (see Figure 1). On the top level the organization with its culture, communication and leadership, completes the list of essential influences on human behavior and human performance in man-machine systems (also see Nishitani, 1996; Dalijono, 2006). Based on this model, the focus of the first study laid in the connection of the top, organizational level, with the lowest, the individual level.
The purpose of the first analysis was to establish, whether there exists a relationship between organizational changes and human failure loss events. Loss events are to be understood as incidents, which lead to a decrease or a total loss of material and immaterial commodities. With this question in mind, the following emerging hypothesis was tested (see Mellert & Wilke, 2008):
Organizational Changes influence Human Performance in a way that they lead to Loss Events.
This assumption was analyzed on the basis of available loss information of Swiss Re. The data stemmed from Swiss Re’s Global Loss Events database (GLE), which at the time of data collection comprised over 32’000 large global loss events. In the subsequent data preparation, natural perils, marine, road and aviation losses were excluded from the dataset. This first selection left 2’652 loss events in the main database. After another filtering process, which expelled the loss events with the lowest severity, the sample consisted of man-made losses that where higher than $54 Mio. In order to answer the research question, the study was established in two major steps:
i. In the first run, the dataset of 500 GLE was scanned for the actual cause of event, segregating loss events, which were due to human failure, from losses that could not be traced back to a human failure.
ii. In the second step, all loss events were again examined in detail to identify temporally relevant and predefined organizational changes within the concerned company, evident before the actual loss event.
In both runs, text analysis of contemporary and current articles in the press, of business and, if available, accident or investigation reports was applied to accomplish the task.
At that stage it was possible to detect several relations between organizational changes and loss events that were due to human failure. Although they all will be further discussed in more detail in their appropriate chapters in the study at hand, a small overview shall now be given. Like this the forthcoming arguments can be better understood. Among the prominent research results of the first study were:
- In more than two thirds of all losses that could be determined as a human failure, some forms of organizational changes could be detected in the affected organization prior to the loss event.
- Human failures with preceding organizational changes led to greater average loss amounts.
- The relevant organizational changes that could be detected before the actual human failures, varied from industry to industry.
1.2. The Framework of the Second Study
Given these findings, new questions about the relationship between organizational changes and human failure arise, asking for a deeper understanding of these relationships and moreover, trying to fathom where exactly lies the reason for human beings to fail in an organization. A more intensive study of the interlinkage of the top level of the man-machine model with the bottom level would render deeper insights into the mechanism of human failure. It could be clarified what processes of organizational change cause which psychological behavior patterns within a workforce. In particular the spatial and cultural aspect can not be neglected, as different industries are more prevalent in certain areas of the world, whereas other industries simply do not exist at any given place. Based on the presumption, that the main part of the workforce of an organization springs from the local society where the organization has its branch, the culture of the organization’s social environment plays a major role when analyzing human behavior (Laurent, 1983).
With this preset in mind, a lot of new questions emerged after having completed the first part of the study:
- Do the consequences of human failure loss events not only depend on the industry but also on the region where the organizational change is happening?
- Is there a regional pattern of occurrence of the interdependency between organizational change and human failure? What about cultural influences?
- Is the “openness” of a society’s culture towards the unknown a mitigating factor for organizational changes to cause stressful situations within a workforce?
- Or is a society where “everybody looks for his own” more vulnerable to negative aftermaths of organizational changes and possible human failures?
Out of these manifold questions a very distinct hypothesis began to take form that will be tested in the present study:
The relationship between organizational change and human failure loss events shows a geographic and cultural pattern of occurrence.
To get a clear classification of different characteristics of cultures, and due to the importance of his theories in management sciences and cross-cultural studies, after a thorough review of other studies, Geert Hofstede’s work on different cultural dimensions was used to describe the differences in organizational thinking. He argues in his work, that human beings develop “mental programs” in their early childhood, which then later are strengthened in schooldays and workplace-organizations. He then assumes, that these mental programs contain components of national culture that are very well distinguishable from one another (Hofstede, 2001b). By analyzing over 116’000 questionnaires, obtained within subsidiaries of one multinational business organization (IBM) in 72 countries, he was able to identify five dimensions, along which the countries could be positioned according to their differences in values. He describes that these dimensions were empirically verifiable, statistically independent, and that they were also validated with data stemming from 140 other studies (see also Shane, 1995; Merrit, 2000). The fact, that the returned questionnaires within one multinational organization produced such explicit differences in values, strengthens the argument, that the measured phenomena were the result of national differences. The reviewed employees from one country to another were similar in almost every aspect but nationality. As these national values are an integral part of a company’s value system, the described dimensions can also be used to analyze the organizational culture, which then is specific to organizations across countries (Hofstede, 1985).
Hofstede’s culture dimensions will be applied in this study in order to have distinctive features to differentiate between aspects of one culture that can be measured relative to another culture. Like this, the occurrence of organizational change and human failure can be arranged spatially and culturally. Besides general insights into cultural influences on human behavior, the results may also be used for specific tasks. When it comes to assessing risks, Swiss Re is dependent on a fundamental understanding of organizational structures inherent to the observed company. Particularly the questioning of the workforce, but also the study of the organizational architecture can give information about the corporate culture. If an organization is about to execute changes to its organizational structure, the knowledge that processes of organizational change may affect human performance and that the outcome of these changes – amongst other things – is dependent on the organizational culture, can be an important decision criterion, whether an insurance is underwritten or not. In this sense, a study on the above described relationship and its cultural dependency, will allow better evaluation of risks to be made. Moreover, the detection of possible shortcomings in the management of change can further be used in the prevention of possible human failure loss events.
On the other hand, human factor research with a special focus on culture is able to offer deeper insights into cultural factors influencing and steering human performance. Additionally, it has the potential to broaden and strengthen the general knowledge about the effects, which organizational changes exert, on a workforce and how these changes are unequally perceived by employees with a different cultural background. Hence, such a comparative study on cultural effects is rendering a contribution to the scientific field of cross-cultural psychology.
1.3. The Shape of the Thesis
The following study is divided into three main sections, each based on the preceding one. They set the scene for the work as a whole. First, theoretical thoughts will be introduced and essential definitions be given (Chapter 2.), to facilitate comprehension of subsequent chapters. A methodical part (Chapter 3.), where the set-up of the study will be explained, is followed by a section dedicated to the analysis and actual results of the project (Chapter 4.). This will be administered by visualizing a large part of the findings and explaining the major relationships with the help of charts and graphs. Through this, comprehension of the sometimes rather complex relations and interdependencies will hopefully be enhanced.
Chapter 2. considers theoretical questions and will give essential definitions for the three elements that form the frame of the project, whereas Chapter 2.1. defines what actually is meant by the term “human failure”, and will dig a little deeper into the field of human behavior and human error studies. The structure of this chapter is mainly based on the studies of James Reason, the long-time researcher and theorist on the nature of human error, stress, ergonomics and human reliability (see Reason, 1990). After introducing the lowest level of the man-machine model, Chapter 2.2. will aim its focus at the top level and will serve the understanding of organizational change. Chapter 2.3. will then once again heighten and expand the approach and will closely look at the issue of national culture and will distinguish it from the term organizational culture. At this stage, also a connection to economic geography will be drawn, explaining the spatial characteristic of economic behavior.
Chapter 3. will serve the purpose of giving an understanding of the applied methods. Chapters 3.1. and 3.2. illustrate the layout of the main database GLE where data was derived from, and respectively the composition of the secondary, human factor database especially designed for the study. Existing measurements of culture, as well as their advantages and disadvantage are presented in Chapter 3.3. Based on criteria outlined in this chapter, it will be decided to apply Hofstede’s cultural dimensions, because this scheme suits best the purpose of investigating the hypothesis. Chapter 3.4. will then comment the introduction of Hofstede’s culture dimensions, and declare why and how they were assigned to each country where a loss event occurred. Also the main critics and limitations of Hofstede’s culture dimensions shall be discussed. In the last Chapter 3.5. of this methodical part, the underlying statistical methodology is revealed, while complementary and alternative statistical methods are discussed as well.
Finally, the last main section, Chapter 4., deals with the analysis and actual results that could be obtained, and where possible, will visualize them. Here again, not only to keep consistency but also to aid comprehension, the results will be discussed in stages: Chapter 4.1. outlines the distribution of human failure loss events over the sample, and discusses its frequency and severity over different industries, before Chapter 4.2. inserts the level of the organization and tries to clarify the relationship between organizational change and human failure. It will be argued to what extent the established correlations are able to explain a definitive link of human failure with organizational changes. In Chapter 4.3., the results of the incorporation of the different dimensions of national culture are shown, and explained, which of these dimensions might serve as an explanation for a possible negative influence that organizational changes exert on human performance.
Chapter 5. summarizes the main findings, highlights the limitations of the study and raises open questions, before in the last Chapter 6., an outlook in the direction of further, necessary studies on the influence of national culture on organizational behavior is ventured.
2. The Three Elements of the Study
Based on the above described model where the human being comes into contact with technology in a man-machine system, human performance is influenced on different levels (HSE, 2005; Dalijono, 2006; Badke-Schaub et al., 2008). As the theoretical background of this work is fundamentally based on this assumption, the following section is, meant to derive and explain the most essential definitions for understanding and justifying the design of the study. The thoughts are depicted in a bottom-up approach, starting on the level of the individual, where the term of human failure forms the core issue. The illustration of what is understood by organizational change and its characteristics – the top level of the model – is followed by a definition of culture and its sub-entity of organizational culture.
2.1. Human Failure
One only needs to screen the latest newspapers, to come across articles dealing with “human error”, “human failure”, “human factor” or “human mistake”. There seems to be a never ending list of terms for classifying accidents caused by humans. In contrast to this group, there are also wordings for accidents that can not be directly linked to a human being – in such a case, the actual cause is labeled as “technical error”, “machinery breakdown”, “system failure” etc. Where exactly has the line between human and machine to be drawn, especially when man is working in a tight relationship with machinery? It becomes clear that failure, whatever forms it may take, has different meanings for different people, so that an encompassing and universally accepted definition of failure – be it human or technical – has not yet been accomplished, nor does it seem, that this mission ever will be (Hofinger, 2008). Still, attempts to classify different forms of erroneous actions have been made, where James Reason’s terminology of human behavior and different error types is one among the most prominently featured in contemporary literature (see Reason, 1990). One major advantage of Reason’s methodology is, that he not only classifies errors, but also describes general human actions which can lead to unsafe actions (Hofinger, 2008). This approach allows important psychological insights and may be helpful in circumventing human failure.
This chapter is far-off from reinventing the wheel, but it hopefully will be in James Reason’s sense, to slightly adapt his profound definitions of human error, to get a much simpler but still satisfying taxonomy for the use of this work. After all, it was Reason (1990) himself, who mentioned, that human failure terminology usually is made for specific purposes. It therefore will always show small discrepancies, depending on what reason and intention it initially was made.
A short glance at history reveals some timeless wisdoms concerning human performance. The Romans already knew that to be mistaken is nothing but human, and even Goethe’s Faust understood that man errs as long as he doth strive, and probably nobody would contradict these two phrases nowadays. Especially the latter hits the mark, as it is nothing but natural, that actions in everyday life do not always reach the intended goal, although this would have been possible. If one thinks about the task of driving one’s own familiar car up a sinuous mountain road, it seems, that there is a multitude of ways where this undertaking actually can go wrong: wrong gear, inappropriate speed, improper lane-changing etc. Making a proper separation between the different forms of failure may seem not feasible and even unnecessary. But if looked at closely, it gets obvious that when speaking of human failure there needs to be an intention (e.g. reaching the top of the mountain), an action (e.g. driving the car) and a consequence (e.g. a car accident) before a useful classification of human failure is possible (Reason, 1990). This approach also goes along with models within the field of social geography, which deal with theories of action and the intentionality of human behavior (see Werlen, 1997; 2007).
With the help of these three terms, different types of human failure can be segregated, and so they form important pillars in most of the prevailing definitions of human failure. Hofinger (2008) describes human failures as deviation from a behavior which is considered as correct, or as a negative divergence from a targeted goal. Also Reason (1990, p.9) speaks of “…occasions in which a planned sequence of mental or physical activities fails to achieve its intended outcome…”. According to Hagen & Mays (1981) the definition of human failure includes additionally the dimension of a negative outcome, where the term is extended to such a degree that the actual unsafe act in a predefined task can lead to partial damage of material and property, or to a breakdown of process flow. Based on these definitional thoughts, a slightly different, definition adapted to the study was generated. Emphasis was placed upon the fact, that all pillars of the most common definitions also occur in this one:
Human failure describes all intentional and unintentional human actions, which deviate from any given, predefined task, and which lead to an unintended outcome and undesirable consequences.
It has to be understood, that these undesirable consequences (e.g. accidents) seldom are the result of one single failure of a human being, but more likely occur through a combination of events and inadequate decisions (Hughes & Kornowa-Weichel, 2004; Besco, 2004). Human failures can happen at any level of a process, whether it is at the individual level or at the organizational level (Hofinger, 2008).
In order to deepen the insights and to establish more profound understanding, human failure was further sub-divided into different, so-called unsafe acts that could be the cause of an unwanted event. The ascertainment of these more specific types of human failure in the subsequent study was to be accomplished by text analysis of articles in newspapers and business reports. Because with this procedure an exact determination of the actual unsafe act is highly delicate, the distinction had to be kept as simple as possible. For this reason, only three types of human failure were distinguished, whereas the prerequisite “intention” was used as the distinctive feature. Errors were segregated from mistakes and violations (also see HSE, 2008a).
Classifications which describe the causes of failures are mostly based on psychological causes within a persons mind, that steers its behavior (Hofinger, 2008). Additionally they are numerous and argumentative (see Hollnagel, 1983). Norman (1981) discusses in his work the categorization of errors, and based on the assumption that schemata control action sequences, he distinguishes errors that are due to errors in the formation of intention, faulty activation of schemata and loss of intention. Norman termed these so-called “attention failures” as slips, and they were complemented by Reason (1990) with the term lapses, for describing “memory failures”. In the study at hand, such a fine differentiation was neither strived for, nor would it have been helpful for the following analysis. The two expressions were united under the label “error”. The fact, that both slips and lapses are unintended actions, justifies their connection. The characteristics of errors, which is that certain actions do not go according to a plan, can be summarized in following definition:
Errors are totally unintended human actions that do not go as planned and occur during a planned sequence of a task, regardless of whether or not the plan was adequate to achieve its goal.
Again it must be underlined, that this form of an unsafe act is absolutely unintended, as the executor of the action knows that the action is not appropriate. Hence they can occur because of many reasons: stress, bad design and organization of the workplace or simply due to time-of-day effects are only few to name out of a lot (see Dalijono et al., 2006; Blatter & Cajochen, 2007). To point this out, a look at the above mentioned example of driving a familiar car up a steep, winding mountain road, may help in understanding. Although you know perfectly well, that you need to shift into lower gear the steeper the climb gets, you nevertheless change into fourth gear instead of the correctly intended second gear. You were so well acquainted with your own car, that this unintended outcome of your action – the stalling of the engine – astonishes you. This example makes it clear, that errors always happen and are unlikely to be eliminated though training or better technology (see also HSE, 2005; Foord & Gulland, 2006).
In contrast to the above discussed errors, mistakes are failures of judgment. This means, that the plan for achieving a specific goal possesses deficiencies, or that even the processes involved in the selection of the objective were flawed in the first place (Reason, 1990). In other words, mistakes occur when we only think that we are doing the right thing. This fact makes the big difference between an error and a mistake: with the first one the action and the outcome are unintended, whereas with the latter, the outcome is in fact also undesired, but the plan for reaching the goal was wanted – it was just not the appropriate one.
Mistakes are defined as failures of judgment or decision-making. These (intended) actions occur where the wrong thing is done, believing it to be right, and they appear in situations where human behavior is based on remembered rules or familiar procedures. Since these decisions in unfamiliar situations can lead to misdiagnoses and miscalculations, training is the key to avoiding mistakes.
The more detailed definition of this form of an unsafe act, shows that mistakes are hard to detect and prevent (Hofinger, 2008). To amplify these thoughts and to make them more sizeable, another example with the car and the mountain road shall be given. While cruising up the described mountain road, you remember the rule that your instructor once gave you: “the steeper the climb gets, the lower the gear you want to use”. After completing a narrow curve, you find yourself confronted with a sharp incline of the road ahead, which presses you on for a quick decision. You recall the rule of thumb, given by your driving instructor many years ago and shift down into second gear while approaching the steep climb. Once you are in the climb, you notice that you overreacted, and that third gear would have been more appropriate to negotiate this part of the road. Your initial judgment was simply mistaken. A little more practice in driving mountainous roads could have prevented you, from making your engine scream and yourself from being all shaken up.
Errors and mistakes are also seen as such by the persons who committed them. With the benefit of hind-sight, they know that their initial intention of reaching a goal was not successful due to their own failure; either because of the erroneous action itself or because the plan for achieving the outcome had deficiencies from the start. Taking one more rung up the ladder of intention, one reaches the level, where actions are carried out, that are not considered as being erroneous or mistaken by the agent because they occur absolutely intentional. Violations are intentional deviations from certain rules of behavior, that are rarely willfully malicious but just forms of bending the rules (Shappell & Wiegmann, 2000). Because the person who violates certain rules seldom looks at himself as a violator, his actions therefore are only considered as erroneous by his social milieu, making the latter an important precondition for describing violations (Reason, 1990). Errors and mistakes can occur in isolation, violation can not because they need forms of rules and values that can be broken. The respective origins for these willful disregards of rules and regulations are manifold and can lie in time or cost pressure, but also in the consideration of effectiveness against security measures (Hofinger, 2008).
Violations are intentional, but not necessarily reprehensible actions, which in order to achieve a goal, deviate from rules or procedures despite the possible consequences. They are rarely willfully malicious.
To go back to the demonstrative model with the car and the mountain road: Whilst driving up the winding and at some places very dangerous road with overhanging and numerous blind bends, you check your watch and notice, that you are late for your appointment on top of the mountain. Even though you see the ostentatious and eye-catching signs along the road, forbidding any overtaking maneuvers, you risk to pass the slow truck driving in front of you. Despite your awareness of the possible consequences for yourself – the danger of an oncoming vehicle, or a fine if you get caught by a traffic control – you take the risk of overtaking in order to achieve your intended goal. Maybe there is light traffic on this road, so that there is only a residual risk of both an oncoming car and the police conducting traffic control on such a solitary road, and you can reach the mountaintop punctual to be on time for your meeting. With this example, another important fact with violations emerges: as long as there are no negative consequences following your violations, you even will be rewarded for not following the rules, and they can become routine (HSE, 2005).
illustration not visible in this excerpt
Figure 2 Human Failure Different kinds of unsafe acts can be distinguished according to the degree of intention. With errors, the outcome is totally unintended, while with violations the possible consequences are known to the violator. Mistakes are a composition of both other forms in a sense, that the plan for achieving a goal was intended, but the outcome was not the desired one. (after HSE, 2005; Shappell & Wiegmann, 2000)
2.2. Organizational Change
After having described the individual aspect of the influences on human performance, the focus will now be shifted towards the uppermost level of the man-machine system, concerning the organizational factors. Although popular lore has a clear meaning and idea of what is understood by the term organizational change, it is notwithstanding sometimes very difficult to come down with a conclusive definition. Even more, since the wording consists of two separate, equally tricky to define concepts – organization and change. This chapter therefore serves the purpose of first delivering a definition of what is meant by “organization”, and then combining it together with the idea of change to the expression of “organizational change”. Several characteristics of this phenomena will be specified so that in the end, some thoughts on how such organizational changes are to be managed so that a risk of failing is minimized, can be developed.
Defining organization seems as hard a term to set limits to as it is with failure. Depending on what purpose the terminology is used for and on which specification it is focused, the definition may describe an organization either as a socio-technical system or as a process (Buerschaper, 2008). Because in this thesis the main focus will be laid upon business companies, the term organization will be used to describe mainly such forms of socio-technical systems. Its meaning shall now be deduced and a definition be given.
Statt (1991) understands organizations as groups of people, which are joined together for the purpose of reaching certain, common goals. Reflections in the work of Buchanan & Huczynski (2004) steer in the same direction. They also talk of pursuing collective goals as the actual sense of the social arrangement. Achieving controlled performance is an important requirement, which decides whether the task is accomplished successfully or not. Another definition of organization given by Mullins (2005), widens the term and includes the existence of a structure which frames the interactions between the groups and individuals. This structure is created and sustained by management – another expanding of the definition – to provide order and to direct the efforts to reach the corporate goals.
Especially the last definition of Mullins (2005), underlines the fact, that an organization as a socio-technical system is directly and very tightly linked to the process of organizing, which then becomes essential for the system to exist. According to Weick (1995), organizing as a process means, to reach a consensus on how to deal with events and problems within and without the borders of a system, including therefore a range of abilities to think and communicate. It becomes clear that these kinds of structuring activities prevent complex systems from being hurled into utter chaos and disorder. Thus, organizing is a continuous process, where acting in groups is crucial and where different, individual patterns of thoughts develop and specific forms of organizational communication arise (Badke-Schaub, 2008; Buerschaper, 2008).
Picking up the important ideas of these theories and assembling them together, leads to the following definition of what is understood by organization in this work:
Organization designates on the one hand a socio-technical system, whose activities are purposed to achieve certain collective goals, and on the other hand describes processes of organizing as necessary activities of the human entity within the system.
This view of an organization, is mainly based on the assumption, that an organization is a closed system. It is set within a wider environment of further systems, where the latter provides inputs (e.g. material, resources etc.) to the organization and at the same time forms the catch basin for the organizations outputs (e.g. goal achievement, employee satisfaction etc.) (Senior & Fleming, 1997).
In 1964, Bob Dylan put it into words and sang it out to the audience, that the times they are a-changing. In doing so, he showed, that the demanded and dearly longed-for change of the social order but also the way of thinking at that time, had many faces. Thus, change is a widely spread term for describing a discrepancy between a given past-situation and an actual present-situation in time, and is used in a multitude of very different contexts: it shows up in every day’s speech as well as in debates during presidential elections. If organizations are regarded as open systems, it is possible, that changes of conditions in its environment can have effects on the organizations set-up, making a sudden emergence of changes within the organization likely (Senior & Fleming, 1997). In this study, though, organizations are considered as closed systems, and as a consequence of this, changes within an organization are thought to be planned rather than just happening. Another possibility, not followed in this thesis, would be to analyze organizations embedded in a significantly changing environment. This implies, that organizational changes have to be seen as systemic, rather than isolated (Wischnevsky, 2004). In contrast, this study assumes, that in organizations change is caused with intent, creating a condition that did not already exist (Ford & Ford, 1995). Hence, the theoretical construct of organizational change is based on Lewin’s (1951) concept of change process, where he defines change as proceeding in three phases of unfreezing, moving and refreezing.
According to this model, an organization carrying through organizational changes, first has to disturb the status quo by unfreezing habitual ways of thinking and behavior of its employees. This phase might include laying off that part of the workforce that holds on to maintain the usual situation, and strengthening the other part that welcomes the announced change. In the second stage of this model, the organization gets to the point to actually conduct the intended changes, so that it can be moved to the new situation. This makes it at the same time possible to replace the disestablished modes of thinking with new types of behavior by individuals and new structures and strategies. After completing this step of a change process, the concerned organization needs to institutionalize and stabilize the realized changes, by monitoring and protecting the organization from falling back into the old habits.
Now it is of course interesting to ask for the underlying reason, why and how changes are applied in the first place. It gets clear, that such a transformation of an organization’s structure, with the replacement of old and long-established habits through new and still unevaluated modes of organizational behavior, always makes the undertaking vulnerable to unforeseen events. Especially in process operations of hazardous industries, organizational changes such as downsizing the number of employees are always a risk of putting the remaining workforce under additional stress and workload, which then again impacts the human performance and safety on the job (Brabazon & Conlin, 2001; HSE, 2008d). Hence, in scientific literature on organizational change it is often noted, that answering this question about the why and the how, acquires a lot of effort, but is worth trying in order to unify the field (Armenakis & Bedeian, 1999).
Corresponding to the above mentioned characteristic of change, it has got to do with the wish of altering the present-situation and creating a new, maybe more promising condition. Dunphy (1996) discusses in his paper on organizational change in corporate settings, that the most common trigger for a planned change is usually some form of a failure to create flexible and adaptable organizations. Ironically, Weick & Quinn (1999) argue in their work that even though a failure to adapt can be seen as a trigger for change, organizational change never starts because it never stops. This thought indicates, that organizations, in order to stay adaptive, need to change constantly and all the time. They even proclaim, for researchers to shift their vocabulary from “change” to “changing”, making it more obvious, that change is happening always and all around, and never can be a terminated process. Even though these ideas prove to be inspiring, this is an approach that is not followed in the present study, due to reasons of practicability: the analyzed organizational changes had to be assumed as temporary processes with a definite start and end in order to be able to make a statement about the discrepancy between a past- and present-situation. Supposing that change is a never ending process, would not allow a predication to be made about a “before-and- after” or “with-and-without” condition. The paper at hand is more in the sense of Amburgey et al. (1993), who after having completed their study on 1’011 Finnish newspaper organizations over 193 years, suggest, that organizational changes lead to an increase of organizations failing in their purpose. They accept that an organization at one time shows no organizational change, and at another time it does. Their findings strengthen the above mentioned vulnerability of an organization during a process of organizational change.
With Dunphy’s (1996) statement in mind, that there is no all-embracing and globally accepted theory of organizational change, the following definition of organizational change was still attempted and will be used throughout the following paper:
Organizational change describes all processes in an organization and its direct environment, which lead to changes of its organizational structures and of its habitual modes of thinking and thus influence the behavior of its workforce.
Again it is underlined, that organizational changes are understood as rather static in their appearance, meaning, that at one moment they are absent, and in the subsequent moment they are detectable. Within these two limiting points, organizational change is seen as rather dynamic in its development and progression. This goes along with the general perception, that organizational changes often come in waves with different characters, directions and intensities. Once more, it has to be stressed, that organizational change is not assumed to be something negative per se and may sometimes also be regarded as being complementary with skilled workers (Caroli & Van Reenen, 2001). Common sense would forbid human beings inflicting something bad on themselves, if there was not the hope and aspiration of improving the situation. In this study it is simply argued, that because of the disruptive nature of changes, the concerned organizations are put at a higher risk of failure, due to their higher vulnerability (Amburgery et al., 1993). Especially in high-risk organizations, the corporation’s leadership is challenged to conduct a management of change which ensures a high degree of process safety (Grote, 2008). The reader must bear in mind, that a variety of factors determine whether organizations fail and that major organizational changes only represent one parameter (also see Wischnevsky, 2004). It also needs to be stressed, that there are various processes that can be counted as organizational change, common sense or a quick glance at the latest newspaper tells you, that lay offs are just as organizational changes as reorganizations or privatization. What exactly now is understood as organizational change and what not, is further outlined and justified in Chapter 3.2. which deals with the methodical part of the study.
Quite often, people returning from holidays in another country, report their relatives about differences they were able to detect in the other society’s behavior. They do not mean to judge whether the uncovered diversity is good or bad – they just noticed once more, that other people’s way of thinking varies from their own. Thinking differently is not something to avoid nor fear, but something exciting and fascinating, yet sometimes also exhausting and controversial, in short something worth looking at more deeply. These patterns of different ways of thinking are very commonly referred to as culture, a phenomenon charged with many meanings and connotations, and one among many factors responsible for the complexity of human behavior (Betancourt & López, 1993). There is a large variety of theoretical thoughts on the topic of culture to be found in scientific literature. It is neither the intention of this chapter to dig down all the way to the roots of cultural research and question their correctness or validity, nor would it be possible and appropriate to master its complexity in such a short form. Depending on what areas of research one focuses on, be it anthropology, history, psychology or cultural- and social-geography, several, slightly divergent definitions can be discerned. Hence, this chapter only serves the purpose of rendering a useful culture definition for the purpose of this study. It first concentrates on the meaning of culture on a national level, and then discusses its characteristics in organizations. The final definition will be derived and summarized from diverse definitions originating from different fields of research.
In cultural-geography, a discipline examining the interdependency between culture and space, place and its surrounding natural landscape, culture is often described as a collective inventory of meanings which is lived through material and symbolic customs in everyday live (Knox & Marston, 2001). This “inventory of meanings” that is practiced everyday, can be translated into norms and values that are shared by the members of a society. Hayton et al. (2002) define culture as a “set of shared values, beliefs, and expected behaviors” (p. 33), whereas Helmreich & Merritt (1998) understand culture as “a complex framework of national, organizational and professional attitudes and values within which groups and individuals function” (p.1). It stands out, that in the last definition of Helmreich & Merritt (1998), there is already some form of localization and isolation of the term culture to the rather artificial unit “nation”. It underlines the problem of finding a universally valid definition for culture. The term “nation” is to be understood as a group of people who share the same religion, language, history or political identity, and who avow themselves to these similarities. It is not necessary for a nation to be concentrated in one region (Knox & Marston, 2001). Of course it will do no justice to the nature of the phenomenon culture to confine it strictly to national borders, as there always will be transboundary forms of culture, as well as several different cultures within one nation. Most notably in times of globalization, a multitude of different cultures may congregate at the same place (Werlen & Wälty, 1995). Hence, culture does not stop at national borders nor do dissimilar cultures unify just for the sake of artificially drawn boundaries. In order to study cultural differences, though, very often the only units available to compare cultures are nations (Hofstede, 1998a), especially if the comparison is thought to be worldwide and therefore including numerous countries, relevant data are mostly available only at that level. This challenge of localizing the faults where one culture ends and another starts, makes the whole undertaking a very delicate issue and may lead to ambivalent and conflicting discussions about cultural regions where certain cultural systems prevail (see Huntington, 1993). Arising ethnocentrism – an exaggerating tendency to think one’s own values to be superior to others – becomes then an important issue, because very often, the spatial distribution of religions is taken as a criteria to distinguish. So, strictly speaking, nations should not be taken as societies. Thus in this work, the term culture is only applied in reference to societies and not nations, although they are named after their nationality.
The extensive examination of numerous definitions of culture by anthropologists Kroeber & Kluckhohn (1952) is concluded in a rather detailed, but nevertheless still widely accepted definition. They conceive culture as “patterned ways of thinking, feeling and reacting, acquired and transmitted mainly by symbols, constituting the distinctive achievements of human groups, including their embodiment in artifacts; the essential core of culture consists of traditional (i.e. historically derived and selected) ideas and especially their attached values” (p. 181). Again, this definition includes values and symbols, which makes them core elements in the attempt to make culture more palpable. This view can be also reconciled with the thoughts of contemporary and acclaimed sociologist Anthony Giddens, who sees culture as consisting of values, norms to be followed and commodities and artefacts created by a certain group (Giddens, 1994).
The following analysis is strongly based on the work of Geert Hofstede (see Chapter 3.4.). In order not to intermingle his cultural understanding with other definitions of culture, and thus to gain consistency, it is Hofstede’s definition that will be used for describing the characteristics of culture in this thesis. Hofstede & Hofstede (2005) summarize the above mentioned patterned ways of thinking, feeling and reacting, and rename them software of the mind. Through this, they indicate that these elements are acquired in an individual’s early life and further reinforced at school and life in organizations when growing up. This so-called mental programming then steers peoples behavior throughout its life in groups and organizations. This point of view does not differ distinctively from Eitzen & Zinn (1995) who claim that culture is actually learned behavior, acquired during the process of socialization, where individuals internalize the characteristics of their influencing cultural environment.
Culture is the collective programming of the mind that distinguishes the members of one group or category of people from another (Hofstede, 2001b, p. 9).
In this sense, the “mind” in Hofstede’s definition is to be understood as the “thinking, feeling and reacting” in Kroeber’s & Kluckhohn’s definition. Their commonness lies then in the fact, that both definitions see culture as a system of collectively held values. Hofstede (1998b) describes values as “broad tendencies to prefer certain states of affairs over others” (p. 478), whereupon Eitzen & Zinn (1995) narrow the expression by saying that “these are the criteria the members of society use to evaluate objects, ideas, acts, feelings, or events as to their relative desirability, merit, or correctness” (p. 127). In both cases we conclude, that human beings are valuing beings and that they consciously or unconsciously evaluate others and themselves. Values, together with symbols, heroes and rituals may be used to describe the visible manifestations of cultural differences.
illustration not visible in this excerpt
Figure 3 Culture Collective activities (rituals), objects with a particular meaning (symbols) and persons with highly valued characteristics (heroes) are influencing the actual core of a culture: values. All together, during the entire life of a person, are responsible for the mental programming of the members of a culture, which can be termed as national culture (after Hofstede, 2001b)
Following the concept of Hofstede (2001b), symbols are seen as transmitters of culture, and may have various forms to communicate its meanings. Words, gestures, pictures or objects can carry sometimes very complex meanings which are only understood by members of a society who share the same culture. Heroes represent in a sense all the qualities which are highly prized in a culture. They can be dead, alive or even imaginatory, and because they combine all the valued characteristics, they serve as role-models. Collective activities that are seen as socially essential are called rituals, and they do not serve the purpose of reaching a certain outcome, but are carried out for their own sake. These three practices lead to a tighter bond to the collectively shared norms and values of a culture (see Figure 3).
Culture and Organizations
Up to now, the theoretical thoughts on culture restricted themselves to a national level, intended to render criteria for delineating one culture from another. If the level of analysis is lowered it is also possible to detect forms of cultures uncoupled from nationality, such as cultures within organizations (e.g. safety culture as described by Davies & Spencer (2001) or Fleming (2001)). It is for instance clear that, an organization may feature certain ways of acting and interacting that distinguishes itself from other organizations even within the same region or nation. When searching for articles in scientific papers but also daily newspapers that describe cultural aspects of organizations, one will find a lot of synonymous expressions that deal with the phenomenon of a common way of thinking in an organization: Corporate culture, organizational climate etc. try to put in words “how things are being done around here” (HSE, 2008c). This study confines itself on using the term organizational culture. Coming from this direction, it gets obvious, that organizational culture does not differ very much from national culture, except of course, in the different level of analysis. It is therefore necessary to draw a dividing line between national and organizational culture. To obtain a useful definition of what is understood by organizational culture is again dependent on the scientific direction one is coming from. Once more it would present a major achievement to deduce a universally valid definition, which is not at all targeted by this study. Merely a small overview of what is generally meant with organizational culture and how it differs from culture associated to nations, shall be given.
Helmreich & Merritt (1998) define organizational culture as “the values, beliefs, assumptions, rituals, symbols and behaviors that define a group, especially in relation to other groups or organizations” (p. 109). They further comment that some scientific disciplines consider organizational culture to be easily managed, while others look at it as something very hard to manipulate because it was historically grown. Also Jones et al.(2005) follow a similar understanding when they understand organizational culture as mainly consisting of commonly shared assumptions, values and artefacts. Another way to look at it, the more so since the meaning of culture is retained and only the level of analysis changed, is to define organizational culture in the sense of Hofstede (1998c):
Organizational culture is the collective programming of the mind which distinguishes the members of one organization from another (Hofstede, 1998c, p. 2)
This definition is in line with the definition of national culture given in the previous chapter. The only difference is, that the given terms “group” and “category” are replaced by “organization”. Although this similarity might imply that organizational and national culture are the same – amongst others they are in a way historically evolving and socially constructed – they are actually not, because organizations will probably never reach the depth of socially shared understanding as common in national cultures. This may be attributed to the fact, that organizational culture is typically learned in adulthood (Wilkins & Ouchi, 1983). The study of organizational cultures enables a management of organizations, and the knowledge that organizational culture channels the influence of its surrounding national culture (Helmreich & Merritt, 1998) is helpful when conducting organizational changes. Despite the importance of comprehending organizational cultures, this topic is not further pursued here, because this study will not penetrate down to this level; the scarcity of available data would simply not allow it.
At this point it is also possible to draw a connection to the field of economic geography, which is describing the spatial order and the spatial organization of economy. Dicken & Lloyd (1999) mention in their theoretical book on economic geography, the importance of looking at the strategic behavior of companies with a geographical perspective. Organizational changes very often include the relocation of an organizations subsidiary, or even the closure and reopening of a branch at a new place because of the new location’s advantages. Whether an organization changes its structure or not, therefore also very often contains a spatial and cultural aspect, which may not be neglected. The decision about an organization’s location is tightly linked to the geographical arrangement of the economy. Dicken & Lloyd (1999) emphasize though, that organizational changes, like e.g. a new location of a subsidiary, must not be analyzed totally isolated from their surrounding, for the reason that they are not conducted in an isolated manner in the first place.
Economic geography also assumes that economic behavior is embedded in social structures and therefore inseparably related to its surrounding context (Bathelt & Glückler, 2003). Hence it influences the organization’s performance (Hansen & Wernerfelt, 1989). It is a logical consequence, that organizations are not only dependent on their formal and informal networks and contacts with other corporations at the same location, but that they are also strongly influenced by the cultural value systems prevailing at that place. This so called concept of embeddedness of organizations describes this phenomenon as the result of development of social interactions (Bathelt & Glückler, 2003). Granovetter (1985) discusses in his paper about social structures and embeddedness, the fact, that economic behavior is never only between isolated actors but that it is always a part of social interaction-systems. These social systems of interaction of course vary from place to place and are therefore to a certain degree based on values and norms of the local population.
In order to get an explanation for the development of economic and social structures, regulation theories try to describe the interactions between economic-technical and social-institutional structures and processes (Sedlacek, 2007). It is assumed that among other institutional connections, norms, rules, conventions and the distribution of power are forming the environment for economic behavior (Bathelt & Glückler, 2003). With this approach it is again possible to interpret the importance of culture in economic behavior. It highlights that companies are always standing in a tight relationship with the social institutions at a location, which are ultimately linked to the prevailing values – the core elements of culture. Through this perspective it gets clear, that social regulation is interacting with economic accumulation, and that cultural issues are influencing organizations in their purpose to reach a certain goal. One theoretical approach that needs to be referenced here, is the one used by GREMI (Groupe de Recherche Européen sur les Milieux Innovateurs) (see Bathelt & Glückler, 2003). It was examining mainly organizations in regions which had a big innovation potential due to a clustering of high-tech companies, and brought them in connection with the social structures present there (see Camagni, 1995). It therefore also highlights, that social institutions are decisive for generating and encourage know-how and economic success.
This chapter as a whole introduced the main elements which form the theoretical pillars of the study, which are:
1) Human Failure
2) Organizational Change
After the presentation of the man-machine model, the definitions of human failure, organizational change and culture were given in a bottom-up approach. Additionally, it was tried to accentuate their linkage to each other and highlight their interdependency. The analysis in the subsequent chapters will follow this approach and will take reference to this chapter when needed.
This section outlines the applied methods and describes the available and finally used data so that the setting and chosen procedures of the study are clear. As already presented in the introductory chapter, this study intended to investigate the effects that organizational changes exert on human performance and whether this relation showed any cultural pattern of occurrence. This research question required a two-stepped study, first defining whether a loss event was due to human failure or not, and secondly if organizational changes occurred prior to the loss event. The applied methods for operationalizing the variables and also the decision process will be discussed accordingly. Already existing methods to categorize national culture which were considered for utilizing in this study will be presented and reviewed. At last, finally, the indices of Hofstede’s culture dimensions (see Hofstede, 2001b) were adjoined to all the countries where loss events happened and which also were included in Hofstede’s work. This allowed an analysis of the cultural influences on the reviewed relationship.
Hence, the next chapters serve the purpose of describing the structure of the main database GLE where the sample was extracted. This is followed by an explanation of the build-up of the human factor database. Three alternative attempts of making culture measurable are presented in the form of SWOT-analyses, before the rather critical discussion of Hofstede’s dimensions and their applicability in this study is opened. The last chapter will explain why correlations where used as a statistical tool and highlight the limitations of such a statistical methodology when it comes to analyze multicausal, social phenomena.
3.1. Global Loss Event Database
Swiss Re’s Global Loss Event database captures large loss events, either man-made or caused by natural catastrophes. To be included, the loss events have to exceed certain defined thresholds, regardless of any Swiss Re involvement. It is used in all existing claims systems within Swiss Re. A Global Loss Event (GLE) is comparable to a newspaper article about an event, as it captures a set of crucial information like date, name, location, loss amount and if possible a short description of the event. Though, reliable information about the actual cause is sometimes rather scarce, because investigations trying to reveal the true reason for the accident take a long time to close – sometimes several years. Very often the outcomings can not be supplemented in the database due to a large number of new loss events, and a shortage of staff to accomplish the task of periodic updating every single case. Therefore, the database serves as an information pool as well as an enabler for various purposes, such as event based reporting or retro collection of loss events, similar to the study at hand.
GLE database comprises loss events with large economic impact or direct effect on the income statement of Swiss Re. Thus, there are two segregated groups of loss event contained: loss events independent of Swiss Re share, and others with Swiss Re involvement. Following criteria have to be fulfilled for a loss event to be included into the database:
with no Swiss Re share:
- Loss events with no Swiss Re share have to exceed CHF 10 Mio. insured loss
with Swiss Re share:
- Loss events with Swiss Re share have to exceed CHF 1 Mio. individual loss (except loss event in aviation and space travel have to exceed $1 Mio.)
- Losses belonging to ALCs (Associative Loss Complexes are claims issues which receive widespread coverage in the media and which may have an impact on Swiss Re. The potential does not arise from one single loss but from the sum of losses with a similar cause or a similar effect. The name ALC is a Swiss Re internal term and not a common insurance language term.)
- Known accumulation of claims within Swiss Re Group (e.g. hurricanes, winter storms or floods)
- Co-insurance claims
Events with thresholds lower than those mentioned above can be opened at any time, if necessary, however, the event has to be compliant with the guidelines. As stop loss treaties and other aggregate covers are not “one event”-bound and rarely triggered by only one event, they do (in 90% of all cases) not qualify for a GLE. If there are uncertainties as to whether the claim needs a GLE or not, the GLE Clearinghouse can be approached. The Clearinghouse does daily research and captures larger events due to natural catastrophes. Furthermore, it checks the data quality and validates all new GLEs. If, for example, a double entry is detected, all contributing units are asked via mail to relink their claims to the valid GLE.
For this study on organizational changes and its consequences on human failure loss events, natural perils and traffic loss events (marine, road, aviation) were excluded. The dataset comprised at the time of data taking (April 2008) over 32’000 cases. The reason for excluding traffic loss events, lay in the fact, that in this field already a large number of studies has been published on human performance and its organizational and cultural influences (also see Wiedemann & Badke-Schaub, 2008; Dekker, 2006; Helmreich & Merritt, 1998; Stokes & Kite, 1994). This first selection left 2’652 loss events in the main database.
Since Swiss Re was interested to conduct a study about the largest loss events only, another partition was made, based on the insured loss amount: only man-made losses with an insured loss amount equal or higher than $ 54 Mio. were to be included in the analysis. This second filtering step left 626 large loss events remaining. From these remaining events, only the 500 greatest were used for further analyzing.
To get a certain degree of comparability of the economic indicator over time, all insured loss amounts were adjusted with a, for insurance lines of businesses common, annual inflation rate of 4%. This adjustment to the year 2008 led to a new distribution of insured loss amounts, ranging from $55.6 Mio. up to $70’954 Mio. The largest loss event, which by the way was caused be a major class action considering a large period of time, clearly had to be treated as an outlier and needed to be removed from further analyses. Besides this exclusion, other loss events that could not be parameterized with industrial classification codes – due to a lack of information on the loss event –had to be removed. The actual sample in the end comprised 414 loss events, which were distributed over 54 countries worldwide. When extracting data from its original database GLE, satellite loss events unfortunately slipped through the filter’s meshes. The reason for excluding them, lay in the fact, that firstly, loss events in connection with aviation were not to be analyzed and secondly, because the search for the actual unsafe act and the untangling of the indeed very complex organizational interrelations, proved to be too hard to establish conclusively within the field of space aviation. Besides they represent high-risk ventures and feature hard-to-decipher man-machine interactions.
3.2. Human Factor Database
To accomplish the task of identifying a relation between human failure and organizational change, methods and a tool had to be developed that would guarantee an unobstructed import and a subsequent easy handling of data. For this reason a secondary database was set up in Microsoft Access, which with the help of tables and forms enabled analyses to be carried out. In a first step of this procedure, the cause of event was determined, whereas in a second step, the perspective was directed towards the organizational background of every loss event. A proper distinction of the following issues was achieved through complying the depicted theoretical thoughts in chapter 2.
Cause of Event
The retrieved sample included the most relevant information for the study and then was imported into a Microsoft Access database. The specific information for every single loss event was arranged in the usual manner for tables, so that an easier subsequent evaluation was permitted. To complement the database, additional columns were established, that would define the actual cause of event and if possible distinguish, whether the loss was due to an error, mistake or violation. This step was tackled through archival research and a following text analysis of literally any newspaper articles, business, accident and investigation reports dealing with the loss that could be gathered. If it was mentioned that the actual cause was revealed as being the consequence of a human error, mistake or violation, the case was labeled as “Human Failure” and accordingly “Error”, “Mistake” or “Violation” (see Figure 4). Very often though, news coverage from one source differs from another – the first identifies the reason as an error, the second as a mistake, the third as a slip etc. An accurate determination of the unsafe act sometimes can get very demanding indeed. To elude such inconsistencies in terminology, another identifier was established. To indicate, that according to archives a loss event in fact was a human failure, but the ultimate unsafe act was not establishable with the methods applied, the case was labeled as “Human Failure” and “Unknown” (see Figure 4).
If loss events could not be attributed as a human failure, since the cause was for example due to a technical defect, or most commonly, because the reason simply could not be conclusively established, the case was classified with the label “Human Failure Unknown”. Another possibility for naming the two different groups might have been, “Human Failure” if it was determinable and “Others” if it was not. It came to the decision, that “Human Failure Unknown” was to be used, in order to accentuate, that this group to date also contained loss events with indeterminable causes. As already mentioned before, from now on the rather bumpy expression “Human Failure Unknown” will be referred to as “Non-Human Failure”. Like this, some of its clumsiness may be removed, and in return, comprehension for the matter can be gained. This way of classifying the two different causes leads to a rather restrictive procedural method. Because loss events were only counted as a “Human Failure Loss Event” when a human failure could be conclusively established, the ratio between “Human Failure” and “Non-Human Failure” would get very conservative. It is very likely, that in the group of “Non-Human Failure” there are certain loss events which are due to a human failure, that could just not be detected (also see Chapter 4.1.).
illustration not visible in this excerpt
Figure 4 Form for establishing Cause of Event
To accomplish an analysis on the impact of human failure on different industries, the respective industrial sector had to be ascertained where a loss event occurred. Therefore, a two-digit industrial classification code for each loss event was determined and attached. This assignment was based on and followed the rules of the classifications registry provided by the United Nations, which keeps information on statistical classifications regularly updated and revised. For the purpose of this work, the latest revision (Rev. 4) of the International Standard Industrial Classification (ISIC) was used. In their preface of revision 4, United Nations describe its intention as follows: “The International Standard Industrial Classification of All Economic Activities (ISIC) is the international reference classification of productive activities. Its main purpose is to provide a set of activity categories that can be utilized for the collection and reporting of statistics according to such activities.” (United Nations, 2008, p.iii). The classification is subdivided in a hierarchical way, classifying into sections at the lowest level, two-digit divisions, three-digit groups and four-digit classes at the most detailed level. Since the determination of the exact economic activity of the organization may get rather difficult because they sometimes operate in various, overlapping segments, the industrial classification in the study at hand confined itself to the application of the two-digit division only (see Table 23). This was thought appropriate, moreover, because a higher specificity would not necessarily bring further insights for the research question.
After having completed this step, in a second run, the focus was put on the corporate background of each loss event. It needed to be determined, whether organizational changes did occur prior to the accident or not. However, the simple stating that organizational changes, as described in Chapter 2.2., could be detected or not, would certainly not be sufficient, to draw meaningful conclusions. As already mentioned in the definitional part, there are a multitude of forms in which organizational changes can manifest themselves. One can imagine how delicate a task it may be, to draw a preassigned and immutable dividing line, and count the closure of a small department in a big corporation as an organizational change, while classifying the downsizing of an anyway small enterprise as “no organizational changes”. As this examples shows, it is first of all, always a questions of relativity, which processes may be seen as relevant, and which not. Like every organization has its own distinct sizes in terms of number of workforce, turnover or architecture, in the same manner vary the consequences of the executed organizational changes. The dismissal of a small number of employees in a small organization with a big turnover, does not lack in relevance compared to an overall restructuring of a multinational company, with numerous subsidiaries. Also organizational changes must be looked upon not as something isolated, but as dynamic phenomena embedded in sequences of events that are interrelated (Barkema & Schijven, 2008). Due to the complexity of this impediment, the idea of defining certain relevance-criteria which had to be fulfilled, only to justify afterwards all the exceptions that undoubtedly would have to be made, was surrendered right at the start of the study. Instead, a lot of common sense was deployed in the decision process and the underlying arguments were justified and documented in the database, where the detected organizational changes are described. The only dichotomous variable that had to be fulfilled before organizational change could be treated as substantial, was time. Through discussions with experts on organizational change it was decided, that organizational changes could be regarded as being relevant, if they where taking place within a 2-year period prior to the date of the loss event. Organizational changes were assumed to take effect within this 2-year frame; if they were conducted outside of this period, they were considered as being deprived of their impetus and of no other relevance for the analysis. Furthermore, the choice of such a 2-year period goes along with other studies on the impact of organizational changes (see Wischnevsky, 2004; Romanelli & Tushman, 1994; Virany et al., 1992).
Besides the temporal relevance of organizational changes, it also had to be defined, which forms and processes of organizational changes were to be looked at more closely. For this reason, a list of predefined organizational changes was compiled, which then was used to assess the occurrence of various organizational changes. Because these processes happen on the uppermost level of the earlier mentioned man-machine system, the assembled processes were renamed under the term “Macro Factors”. Through literature research (also see HSE, 2003) and again many discussions with experts in the field of organizational change, the decision fell on 13 macro factors that would determine whether a loss event was labeled as “with organizational changes” or not. Following, the predefined list will be outlined and each macro factor shortly introduced: