SlideShare a Scribd company logo

The use of pro-active versus re-active risk management practices for managing supply chains

P
Pieterjan Tilleman

This document is a master's thesis submitted by Pieterjan Tilleman to the University of Ghent in 2013-2014. The thesis examines the use of proactive versus reactive risk management practices for managing supply chains. It includes a literature review on risk management, supply chain management, and supply chain risk management. The thesis presents a conceptual framework analyzing the influence of proactive and reactive risk strategies, moderated by supply chain management practices, design complexity, and other factors. Primary data was collected through a survey of Belgian companies and analyzed using statistical techniques including factor analysis and hypothesis testing. The results provide insight into how different approaches to risk management impact supply chain performance.

1 of 115
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
UNIVERSITEIT GENT FACULTEIT ECONOMIE EN BEDRIJFSKUNDE ACADEMIEJAAR 2013 – 2014 The use of pro-active versus re-active risk management practices for managing supply chains Masterproef voorgedragen tot het bekomen van de graad van Master of Science in de Toegepaste Economische Wetenschappen: Handelsingenieur Pieterjan Tilleman onder leiding van Prof. Ann Vereecke Begeleider: Evelyne Vanpoucke
The use of pro-active versus re-active risk management practices for managing supply chains
I UNIVERSITEIT GENT FACULTEIT ECONOMIE EN BEDRIJFSKUNDE ACADEMIEJAAR 2013 – 2014 The use of pro-active versus re-active risk management practices for managing supply chains Masterproef voorgedragen tot het bekomen van de graad van Master of Science in de Toegepaste Economische Wetenschappen: Handelsingenieur Pieterjan Tilleman onder leiding van Prof. Ann Vereecke Begeleider: Evelyne Vanpoucke
II CLAUSE OF CONFIDENTIALITY PERMISSION Ondergetekende verklaart dat de inhoud van deze masterproef mag geraadpleegd en/of gereproduceerd worden, mits bronvermelding. Pieterjan Tilleman
III Acknowledgements Vooreerst wil ik Evelyne Vanpoucke van harte bedanken voor de volle steun, de vele raad, de nodige tijd en de snelle beantwoording van vele emails die mij op het goede pad hebben geleid. Vervolgens mag ik een dank aan mijn vriendin Louise uiten voor de hulp met de bibliografie en de steun en toeverlaat wanneer het eens wat minder vlotte. Ook mag ik de medewerking van de vele operations en supply chain managers van de deelgenomen Belgische bedrijven aan dit onderzoek niet vergeten. Al was het soms een niet voor de hand liggende opdracht om een vragenlijst van een dergelijke omvang ingevuld te krijgen, doch zonder deze data zou dit werk niet tot stand gekomen zijn. Tenslotte, dank ik mijn familie voor de volharding en de enorme steun en in het bijzonder mijn vader voor het nalezen van dit werk.
IV Table of Contents Table of Contents ................................................................................................................................. IV List of Tables....................................................................................................................................... VIII List of Formulas.................................................................................................................................. VIII List of Figures .........................................................................................................................................1 1 Introduction ...................................................................................................................................1 2 Literature research.........................................................................................................................3 2.1 Risk Management...................................................................................................................3 2.1.1 The Nature of Supply Chain Risk.....................................................................................3 2.1.2 Risk management .........................................................................................................11 2.2 Supply Chain Management...................................................................................................22 2.3 Supply Chain Risk Management ...........................................................................................27 2.4 Supply Chain Security Management.....................................................................................30 2.5 Conclusion............................................................................................................................32 3 Conceptual framework.................................................................................................................33 3.1 Environment.........................................................................................................................33 3.2 Risk perception or representation of risk .............................................................................35 3.3 Proactive versus Reactive Risk Strategies.............................................................................36 3.4 Moderators and Mediators ..................................................................................................39 3.4.1 Supply chain management practices. ...........................................................................39 3.4.2 Design and complexity of the supply chain...................................................................41 3.4.3 Continental differences ................................................................................................46 3.4.4 Comparison with 3 years ago........................................................................................46 3.5 Model diagram .....................................................................................................................47 4 Data Collection .............................................................................................................................48 4.1 Introduction..........................................................................................................................48 4.2 Results..................................................................................................................................49 4.3 Variables and constructs ......................................................................................................50 4.4 Exploratory factor analysis (EFA)..........................................................................................52 5 Methodology and Analysis ...........................................................................................................53 5.1 General model......................................................................................................................53 5.1.1 Confirmatory Factor Analysis (CFA) of the general model ............................................53 5.1.2 Comparison with alternative frameworks ....................................................................56 5.1.3 Descriptive statistics.....................................................................................................57
V 5.1.4 Hypothesis testing ........................................................................................................58 5.1.5 Multicollinearity ...........................................................................................................59 5.2 Input from mediator variables..............................................................................................60 5.2.1 Framework ...................................................................................................................60 5.2.2 Confirmatory factor analysis.........................................................................................61 5.2.3 Descriptive statistics with mediator variables ..............................................................61 5.2.4 Structural path significance ..........................................................................................62 5.2.5 Mediating effect and hypothesis testing ......................................................................62 5.3 Input from moderator variables ...........................................................................................64 5.3.1 Descriptive statistics and differences for the moderator variables ..............................64 5.3.2 Moderating effect and moderated mediation..............................................................67 5.4 Some comparison with 3 years ago......................................................................................71 5.5 Hypothesis summary............................................................................................................72 6 Overall Conclusion........................................................................................................................73 7 Limitations and possibilities for future research ..........................................................................75 8 References....................................................................................................................................76 9 Appendices...................................................................................................................................80 Appendix 1: PWC results ..................................................................................................................80 Appendix 2: Participated Belgian companies ...................................................................................81 Appendix 3: Internal consistency scale measurement .....................................................................82 Appendix 4: Descriptive statistics.....................................................................................................85 Appendix 5: Different frameworks ...................................................................................................88 Appendix 6: Confirmatory Factor analysis for alternative framework formulations ........................89 Appendix 6: T-statistics for mediators..............................................................................................93 Appendix 7: Sobel test......................................................................................................................94 Appendix 8: Paired Sampled T-statistics...........................................................................................96 Appendix 9: Multi-group moderation and moderated mediation..................................................100 Appendix 10: Statistics for differences with 3 years ago ................................................................102 Variable paired samples t-test....................................................................................................102 Paired sample t-tests for relationships.......................................................................................105
VI Abbreviations list DV: Dependent Variable ERM: Enterprise Risk Management FMEA: Failure Mode and Effect Analysis IV: Independent Variable PEST- variable: variable that signifies Political, Economical, Social & Technological trends in the environment PORTER-variable: variable that takes a look at the company’s competitive forces RM: Risk Management SC: Supply Chain SCD&C: Supply Chain Design and Complexity SCM: Supply Chain Management SCRM: Supply Chain Risk Management SCSM: Supply Chain Security Management SCVM: Supply Chain Vulnerability Map SME: Small and Medium Enterprise TQM: Total Quality Management
VII List of Figures Figure 1: risk aspects .............................................................................................................................5 Figure 2: Dimensions of risk ...................................................................................................................6 Figure 3: risk sources..............................................................................................................................7 Figure 4: vulnerability - efficiency relation (left), vulnerability - simplicity (right)..................................9 Figure 5: risk management framework ................................................................................................11 Figure 6: fault tree analysis for the AVIA example ...............................................................................13 Figure 7: risk matrix for the additional probability of detecting risks...................................................14 Figure 8: risk mitigations strategies......................................................................................................19 Figure 9: the concept of Supply Chain Risk Management (Blos et al., 2009)........................................27 Figure 10: a SCRM framework..............................................................................................................29 Figure 11: PEST anlysis .........................................................................................................................34 Figure 12: Porter's Five Forces Model ..................................................................................................35 Figure 13: Supply Chain........................................................................................................................41 Figure 14: Supply Network ...................................................................................................................42 Figure 15: a more complex and adaptive network...............................................................................42 Figure 16: supply information network................................................................................................44 Figure 17: Model diagram ....................................................................................................................47 Figure 18: proportion of participated industries ..................................................................................49 Figure 19: proportion of participated countries...................................................................................49 Figure 20: General Model.....................................................................................................................53 Figure 21: framework with general SCM mediator between environment and risk perception (first framework) ..........................................................................................................................................60 Figure 22: Framework with general SCM mediator between risk perception and risk management (second framework) .............................................................................................................................61 Figure 23: Mediating effect ..................................................................................................................63 Figure 24: Environmental factors between Europe and Asia ...............................................................64 Figure 25: Risk Probability and Impact for Europe and Asia.................................................................65 Figure 26: calculation for the multi-group moderation t-statistic and p-value.....................................68
VIII List of Tables Table 1: risk criticality matrix ...............................................................................................................14 Table 2: risk management actions........................................................................................................17 Table 3: the business from a biological view........................................................................................43 Table 4: Environmental constructs.......................................................................................................50 Table 5: Risk perception variables........................................................................................................50 Table 6: Proactive and Reactive Management .....................................................................................50 Table 7: Supply Chain Management Practices......................................................................................51 Table 8: Supply Chain design and complexity.......................................................................................52 Table 9: CFA summary table.................................................................................................................54 Table 10: Discriminant validiy ..............................................................................................................55 Table 11: CFA summary table for comparison with other frameworks................................................57 Table 12: outer model T-statistics........................................................................................................58 Table 13: inner model loadings and T-statistics ...................................................................................58 Table 14: Multicollinearity ...................................................................................................................59 Table 15: Risk management for Europe and Asia.................................................................................65 Table 16: Risk management for lower and higher complex networks..................................................66 Table 17: determination coefficients ...................................................................................................67 Table 18: Mean, standard errors and T-statistics for the two groups for moderation .........................70 Table 19: hypothesis summary table....................................................................................................72 List of Formulas Formula 1: Sobel test statistic ..............................................................................................................62 Formula 2: t-test for multi-group moderation and moderated mediation...........................................68
1 1 Introduction In many business environments, networking in supply chains is almost an inevitable solution to help companies respond fast to market changes. A lot of opportunities are accompanied with networking. Examples are lower transaction costs, ability to concentrate on core skills, lower capital investments, sharing sunk costs, greater flexibility and access to key technologies. So the use, the meaning and the practices of the concept supply chain management became important. However, increased network cooperation does increase the dependency between organizations and as a consequence of the advantages above companies become more exposed to the risks of other companies. Hence networking causes transfer of risks between several companies from a supplier- customer viewpoint. It may decrease some risks but unfortunately increase others. Inevitably partners must share their risk among them as a solution to mitigate their risks and to succeed in their operations. Therefore today’s industries must operate under extreme caution and the concept of supply chain risk management was born. The need of the concept became useful after series of crises and catastrophes had attracted public attention like natural disasters, political and economic instabilities, terrorist attacks and many more. Secondly modern supply chains seem to be more vulnerable than ever: increased competitive pressure in the business environment and globalization of markets. Counterfeiting products has increasingly entered the supply chain and harms a company’s product and reputation. The financial crisis has brought companies to be very suspicious and seek to ensure their business and operations continuity. Nowadays they struggle more than ever from facts like supplier insolvency and less access to credit that especially impact the less financially stable companies. There are plenty number of relevant examples. Automobile manufacturer Land Rover found itself in serious trouble after its only supplier of chassis frames, UPF-Thomson, suddenly and unexpectedly folded further supply delivery (Sheffi & Rice Jr, 2005; Wagner & Bode, 2009). Electronic company Ericsson faced dramatic problems with a huge impact, after a fire at a sub-supplier and has implemented an entire new organization with new supplier risk management tools (Norrman & Jansson, 2004). Ford, Toyota and DaimlerChrysler experienced massive disruptions to the flow of materials into their North-American assembly plants within a few days after the terrorist attack of 9/11 due to border shut-downs (Sheffi & Rice Jr, 2005). Globalization compelled firms to make their supply chains more efficient, resilient or more responsive by outsourcing or off shoring activities, sourcing in low-cost countries, collaborations with
2 other partnerships, decreasing inventory and so on. But all these activities can be associated with a higher level of risk and supply chain sensitivity. A report stated that companies suffering from supply chain disruptions experienced 33-40% lower stock returns relative to their industry benchmarks. Consequently it can also negatively impact a firm’s brand image and reputation. In addition severe disruptions like the Fukushima nuclear disaster have healthy and safety risk consequences. So concluded we feel the need for risk management in a relatively unstable world on the one hand and an increasingly flexible supply chain on the other hand. This work gives an introduction into the risk management world. We will discuss the main concepts of risk, the perceptions, the difference between proactive and reactive mitigation strategies, the supply chain practices that are needed to stay resilient and many more on the basis of existing literature. Secondly we will employ our gained knowledge to build and test a framework that make use of several aspects of the environment, risk concepts and supply chain practices and complexity. Data was collected in many industries to achieve this goal. Additionally we will try to do some investigation for progressions in the past three years and compare continents, in particular European and Asian countries. In the end we will come up with some meaningful conclusions that fit with our model and outcomes, give some limitations and suggestions for future research.
3 2 Literature research First we will dig in to the world of risk management with all his facets. Thereafter we will consider supply chain management concept and finally we will end up with the meaning of supply chain risk management and some thoughts of supply chain security management. 2.1 Risk Management 2.1.1 The Nature of Supply Chain Risk Introductory case Consider a random company; named AVIA. AVIA is a manufacturing company that produces metallic aircraft components for the aviation industry. It is operating its activities since a long time. Through these years it has maintained its supplier base, the metal industry and reached its few customers from the aircraft assembly industry. Suppose now that because of tensions in the commodity market, the company that is responsible for the supply on metal parts has defaulted to deliver the needed products. Company AVIA can appeal on a few minor companies but this amount is not sufficient. You as a company decide to produce further your semi-finished goods with the little supply of metal parts. Suppose then a major customer refuse to do further business with you because you have augmented your prices due to your increased variable costs per part produced. Or you are unable anymore to deliver the requested components on time because time goes by until you receive your parts. On top of that a fire caught your plant and a third of your machinery capacity has been demolished. Reinvesting in new assembly equipment is accompanied with a lot of costs. As a consequence you see your benefits declining and you end the year with a very negative profit and loss account for the AVIA Company. Risk Functions which generate the possibility of beneficial effects or profit often include risks. This is certainly the case with business activities. So risk can here be for example: 1) Your supplier fails in delivering the needed metal parts such that you operate under capacity. 2) Problems in fulfilling customer deliveries arise because you cannot deliver your aircraft components on time. 3) Because of cost considerations you increase your prices. Customers quit business with AVIA which results in a too low or inappropriate demand. 4) Due to a fire at the production plant, it is difficult to get back on track which resides in the difficult re-management of its costs, resources, development and flexibility.
4 The above situations are all examples that contain risk. We can already make a first distinction between demand side risks and supply side risks. The third risk is an example of the first category and the first risk is an example of the latter. These are risk categories that are internal to the supply chain whereas the fourth risk is an example of an external to the supply chain risk. But what is risk actually? Risk is a characteristic of decisions that is defined as the extent to which there is uncertainty about whether potentially significant and/or disappointing outcomes of decisions will be realized (Sitkin & Pablo, 1992). So risks suggest variation in the distribution of possible outcomes, their likelihoods and their subjective values (Wagner & Bode, 2008). According to Kahneman & Tversky’s “prospect theory” individual risk behavior is determined how the situation is framed. For example if individuals are protecting prior gains will be more risk averse. In financial risk management, risk is considered as having an upside and downside potential of possible outcomes according to a normal distribution with a two-sided variance. In contrast the aim of this work is to approach risks in supply chains which can be better stated, considering the severe impact of disruptions, as being purely negative. According to several authors, risk is considered in this manner and that corresponds best from a supply chain consideration (Wagner & Bode, 2009). Zsidisin (2003) contains a broad definition of risk applied to the supply chain: the probability of an incident associated with inbound supply from individual supplier failures (quality, delivery, relationships and price) or the supply market occurring, in which its outcomes result in the inability of the purchasing firm to meet customer demand or cause threats to customer life and safety. What is good in this definition is that it mentions the distinction between supply and demand side risks. It mentions some supply risks but these are not exhaustive. Production capacity constraints on the supply market, technological changes with the supplier, product design features, to mention a few can also play a role. It also assesses the risk aspects which are important to understand risk. Risk aspects In the former definition we find 2 important aspects that constitute risk namely the extent or the impact of outcomes and the possibility or potentially significance that may or may not be disappointing of these outcomes. These are convenient aspects of risk because according to the Bayesian theory when you multiply these two figures for each outcome you get the distribution and thus the severity of each possible outcome. Therefore we can split risk in 4 categories shown in figure 1 below. We consider the enterprise’s vulnerability the highest when both the likelihood and the impact of disruption are high whereas rare, low impact events require less action to mitigate. Furthermore disruptions that combine high probability and low impact are part of the daily operations in the normal flow of business. On the other hand risks with low likelihoods but high
5 consequences need a concrete planning and interference that is outside the daily business operations. The fire at our plant is an example of a low probability, high impact risk. A reduced demand of your aircraft assemblies could be an example of high probability low impact risk. It is important that you recognize this risk in function of your company because the probability and impact differs among different corporations. A strike in one plant of the Airbus corporations has a lower impact on a big multinational with several aviation plants than the same strike in our little AVIA plant. Figure 1: risk aspects Kleindorfer and Saad (2005) suggest besides impact and probability, the speed of the possible risk can also play a role. Speed can be understood as the rate at which the event leading to loss happens, the rate at which these losses happen and how quickly the risk event is discovered by the company. Furthermore the frequency or how often a similar kind of risk event happens. In our situation how often will our major supplier fail to deliver our metal parts? If this occurs too frequently our company may lose his reputation, several customers will abandon AVIA and in the long run our company may be even going out of business. At last, Griffis and Whipple (2012)consider the probability of risk detection as an additional aspect of risk and adds a third dimension besides impact and probability. For instance, a high-likelihood/high- impact risk that is also extremely difficult to detect, warrants a substantially different risk management strategy than a high-likelihood/high-impact risk that can be more readily detected (see further). Low impact High probability High impact High probability Low impact Low probability High impact Low probability
6 Figure 2: Dimensions of risk Risk Drivers and Risk Natures At some point you as a manager believe something exists in your business operations environment and will lead to a particular risk event and a serious impact could occur (P. G. Smith & Merritt, 2002). This is what we call risk drivers. Risk drivers can further increase the risk experienced by the supply chain participants (Jüttner, Peck, & Christopher, 2003). The tensions on the commodity market in metal parts supply or serious changes in a foreign currency exchanges rate are examples of risk drivers. They can lead to a serious risk event and it’s important to watch out and keep in mind that such events, although at first sight these seems to be far from your business, can cause problems to your firm. Competition and globalization increase risk indirectly whereas outsourcing which can results in increasing complexity can have a direct effect on risks (Jüttner et al., 2003). In essence, risk drivers are the start of causal pathways that ends up in risks. A further distinction can be made according to Kleindorfer and Saad (2005) on the nature of the risks or risk sources. These are variables (networking, environmental and operational) which cannot be predicted with certainty and which impact on the supply chain outcome variables (Jüttner et al., 2003). We can make a distinction on risks that come from coordinating supply and demand e.g. supplier fails in delivering the needed part, on the one hand. On the other hand risk arises from normal activities. These can be further subdivided in operational risk and risk arising from natural hazard, terrorism or political instability or called disruptive risks. These latter are risks with low probability, high-consequence of outcomes whereas the former has a higher probability of outcomes. However, most of the quantitative models are designed for managing operational risk. So there is a need towards more disruption risk models. Examples of the former are equipment malfunctions or human centered issues from strikes to fraud. An example of the latter is the fire at AVIA Company. A summary of risk sources can be found in the figure taken from Jüttner et al. (2003). Environmental
7 and organizational risk sources have an impact towards the supply chain whereas network risk sources are the risk sources of the supply chain (Jüttner et al., 2003). Figure 3: risk sources Manuj and Mentzer (2008) on the other hand divides the sources of risk in 4 categories (Supply, Demand, Operational and Security Risk) each with their risk event examples. Supply risks can contain besides the above examples supplier opportunism and inbound product quality. Or the supplier can default in flexibility to deliver the metal parts for AVIA just-in-time. Demand risk can be that our aircraft partner’s demand is very variable or a competitor from the assembled aircraft parts industry negotiates a more interesting demand with our metal parts suppliers. Through this AVIA is losing market share. Another example, but occurs more in food products supply chains, is that these products result in a weather-related demand uncertainty. For example, the demand of ice-cream is the highest when it’s a warm weather. The authors, Chen and Yano (2010), suggest a very flexible contracting scheme to optimize the distribution of risks between the manufacturer and retailer. This can be achieved for example through weather related rebate contracts to mitigate demand uncertainty. Next risks can be seen from an operational point of view like the risk on product quality failures. At last security risks or currency risks can also play a role. Melnyk, Rodrigues, and Ragatz (2009) added also information/technology risks (C. S. Tang, 2006), financial risks and legal/regulatory risks (Wagner & Bode, 2008).
8 Disruptions “Supply chain disruption is the unintended, anomalous event that materializes somewhere in the supply chain and threatens the normal course of business operations (Wagner & Bode, 2009)”. In other words disruption is anything that unexpectedly affects your supply chain. According to Sheffi (2005) these events follow a disruption profile in a predictable way in terms of its effect on company performance. From this perspective general problems can be roughly divided between deviation, disruption and disaster (Gaonkar & Viswanadham, 2004). Whereas the former can be more seen from an operational view, in essence a variation in lead times or demand within the supply chain from their expected or mean value, the latter two problems deal more on the environmental problems. With a disruption is the structure of a supply chain radically changed and with a disaster is the supply chain shut down temporary or irrecoverable. The authors formulated 2 mathematical optimization models that deal with only deviation and disruption problems because modeling disasters is simply impossible. Random events, land natural disasters like tropical storms and earthquakes can be best estimated from historical data for their possible occurrence. The likelihood of accidents on the other hand can also be estimated from industry data, prior events and the enterprise particular safety programs and implementations. Lastly, the probability of intentional disruptions such as job actions, strikes or sabotage) is the most difficult to estimate because the likelihood is a function of the specific company’s decisions and actions. Then, there is the difference between several kinds of storms according to Altay and Ramirez (2010). The impact of damage from windstorms and floods seem to be dramatically lower from that of an earthquake in terms of operational Cash Flow. The authors give the reason for the better predictability of these former 2 climate events and firm’s ability to prepare their firms in advance for them. Earthquakes damages a lot and makes recovery very slow and do not allow preparation time. In addition they show that the impact of natural disruptions is dependent on the firm’s position in the supply chain. The disasters that can be prepared can be planned for the upstream partners. There stock is accrued in advance and can be sold to downstream partners where these have opposite total asset turnovers. A solution to overcome this problem with the downstream partners is supply-chain wide risk practices because a firm that is not prepared will disrupt the operations of the rest of the supply chain.
9 Melnyk et al. (2009) proposes a discrete event computer simulation model that is based on the decomposition of a supply chain disruption in several facets like for example the quantity loss, time period, periodicity, profile breath & location of a disruption and the output level of its recovery towards the supply chain performance. They concluded that the use of classical statistical analysis is rather limited since they do not deal with the time dimension of disruptions. Because of the transient behavior of the process intervention analysis using time series is more appropriate in their study (Melnyk et al., 2009). A general rule should be to include a combination of methodologies in order to make a comparison. Perry (2007) builds a disaster response model after the 2004 Tsunami in Thailand that is for a part transferable to a business context. They highlight the logistic aspects (expertise and efficiency) and the need for quick information by extensive communication and local knowledge to deal with disasters quickly. This can be the case when some manufacturing activities are outsourced in a distant country. Vulnerability It seems according to Wagner and Bode (2009) that probability of risks are determined by supply chain characteristics (density, complexity, criticality, …) and consequently their vulnerability both as part of as well as across the entire supply chain. Vulnerability is defined according to (Blaikie, 1994;(Wagner & Bode, 2009) ) as a company’s capacity to anticipate, cope with, resist, and recover from the impact of a natural hazard. Several characteristics of the supply chain increase or decrease the vulnerability of the supply chain. For example extreme leanness and efficiency is very effective for a company’s operations and reliability towards their customers but may result in an increasing level of vulnerability. While lean management can provide several advantages in cost reductions and efficiency, it makes companies more hazardous to risk vulnerability and velocity. Consequently establishing back-up systems and maintaining reasonable slack can increase the level of readiness in managing risk. One can make a vulnerability efficiency operations vulnerability simplicity supply network Figure 4: vulnerability - efficiency relation (left), vulnerability - simplicity (right)
10 trade-off between robustness and overall efficiency to cope the level of risk. And because of the supply chain is only as secure as its weakest link minor movements can entail serious disruptions which makes the supply chain very vulnerable. Second, to reach more leanness or customer made products provided to worldwide demand, often this is coupled with an increasing complex network. But this must be paid off towards increasing vulnerability. We shall further see that one can overcome the vulnerability in complex network by being more resilient. Sheffi proposed a supply chain vulnerability map (SCVM) with four quadrants namely financial, strategic, hazard and operational vulnerability. Strategic vulnerability means the vulnerability when a new product is introduced. Hazard vulnerabilities are the internal as well as external risk drivers previously described. Operations vulnerability focus on the supply chain as for example distribution network failures. The framework is constructed in a manner that items of a category placed in the centre are very important and those on the edge less important. The goal of this framework is that each of the categories has a property to find, quantify and minimize risk (Blos, Quaddus, Wee, & Watanabe, 2009). Wagner and Bode (2006) found evidence of the effect of supply chain vulnerability drivers are positive towards more supply chain risk. This is the case for supplier dependence, single and multiple sourcing. So firms must according to Wagner and Bode (2006) avoid dependences and improve the robustness of a company’s chain. Meanwhile the choice of single or global sourcing must be done through a risk-benefit analysis (see further). Vulnerability is not in every industry the same. The aircraft manufacturing industry operates in an extreme risk environment, characterized by high levels of commercial, technological and political risk as well as the inherent product safety issues (Haywood & Peck, 2003). Interviewed companies from the author’s research acknowledged that their supply chain is most vulnerable during times of change as the risk profiles affecting their supply chains were also changing, but also that change is a constant state in their supply chain activities. Aircraft companies never experienced a steady-state resulting in increased supply chain change management (Haywood & Peck, 2003). To reduce a company’s weakness Asbjørnslett (2009) suggest to take a vulnerability analysis. It’s a top down analysis and its main focus is towards the system mission and the survivability of the system (Asbjørnslett, 2009). The essential steps to take this analysis is first to search for possible threats and their consequences, next the company must bring back their system to new stability by aligning adequate resource and last determine the disruption or the time the stability is again established. It gives a complete proactive vulnerability analysis framework that works in two rounds.
11 First the manager tries to understand the threats and risks, analyses and rank the consequent possible scenarios and is left over with a set of critical vulnerable elements in a first round. These require additional specific analysis that needs reduced fragility by adding appropriate resource to mitigate their criticality to them in a second round of investigation. 2.1.2 Risk management Risk management is defined as identifying and assessing the probabilities and consequences of risks, and selecting appropriate risk strategies to reduce the probability of, or losses associated with, adverse events (Manuj & Mentzer, 2008). The execution of an overall risk management process is useful for companies because managers tend to focus solely on critical performance targets, which affect the way they manage risk (C. S. Tang, 2006). The need for more supply chain and risk management has certainly become clear after the PWC investigation (Levi, Vassiladis, & Kyratzoglou, 2013). In their research they categorize enterprises in 4 levels of achievement of supply chain and risk management. They grouped the two lower and two higher levels together to reach some conclusions. Appendix ? gives some results from their study and show the percentage of companies with more than 3 incidents that suffered an impact of 3% or higher on their performance as a result of supply chain disruptions. First companies that invested in an advanced risk and supply chain management level are better equipped towards risks than lower risk management levels. Framework for a general structure of the risk management process Figure 5: risk management framework Supply risk management contains several steps and can be seen on the figure 5 (Hallikas, Karvonen, Pulkkinen, Virolainen, & Tuominen, 2004), (Griffis & Whipple, 2012; Zsidisin, 2003),). The different steps will be discussed successively. Risk Identification Risk Assesment, Evaluation and prioritization Risk Management actions and Mitigation strategies Risk Monitoring and Strategy Sharing
12 Risk Identification Equipment interruptions, quality failures and supply fluctuations. These are common strong signals of risks in manufacturing systems. The main focus of risk identification is to recognize future uncertainties to be able to manage these scenarios proactively in a later stage. Chopra and Sodhi 2004 identified nine broad categories of supply chain risks: disruptions, delays, systems, forecasts, intellectual property, procurement, receivables, inventory and capacity. Furthermore according to Manuj and Mentzer (2008) it is recommend that, once identified, risks should be segmented by specific characteristics in order to create a risk profile. You can categorize them in domestic or global risks. Once you start to investigate and identify risk, a common approach is to start with a brainstorming session with the management team with a diversity of people from sales, marketing, quality and finance if possible of your business. It can be helpful sometimes to get your session accompanied with your supply chain partners or major customers ((P. G. Smith & Merritt, 2002), (Preston G Smith, 2002)). By brainstorming you can base your business on the past as well as you can ask if everyone can think of success factors and wonder themselves what can go wrong? Actions that can be performed to discover risks are for example (Mullai, 2009): - Identify risk generating activities - Identify and formulate problems - Determine the background to determine the context - Define (technical, analytical) boundaries for the study - Collect relevant risk-related data and information Daimler Chrysler had to quit production for several days because of a defective fuel injector that came from their supplier Bosch, so the former company claimed his supplier for delivering the wrong part. Bosch claimed that it didn’t make mistakes but instead pointed at his supplier Federal Mogul for their faulty sockets which in turn found his supplier Dupont guilty for delivering defective granulates (Henke, 2009). A practical approach in finding the origin of disruptions is the use of the Tree model. This allows you to find the underlying root causes for today’s disruptions but also by using “what if” scenarios to get the root cause for future uncertainties (Griffis & Whipple, 2012). Ask yourself “what could go wrong at this point that would prevent us from achieving success”, especially for projects. Ultimately in a later stage this can form the basis for a comprehensive scenario planning approach (Sheffi & Rice Jr, 2005). Also a risk simulation can be done or a sensitivity analysis can be performed to check if some crucial parameters or outputs change in different scenarios. It is important to find
13 the causes because they require different modes of prevention and have also different potential impacts. An example of a fault tree analyses with root causes for our case of AVIA is given below: Figure 6: fault tree analysis for the AVIA example Secondly, we can address the reliability tools from Total Quality Management (TQM) to discover risks. One tool that can be used for Risk Identification is Failure Mode Effects and Criticality Analysis (FMECA). In essence this tool aims at performing bottom-up analyses of processes to determine where systems might fail, and then to either design out or improve detection of these potential failure points. The advantage is that this procedure moves from reactive to a more proactive means of equipment maintenance in an effort to reduce equipment breakdown and failure. But this tool is relatively absent from the supply chain literature because it often lacks the assessment/evaluation factors such as probability and likelihood. A better approach would then be Failure Modes and Effects Analysis (FMEA) in a supply chain context. With this method you have to identify and rank potential failure modes of a design or manufacturing process but its disadvantage is that it does not take criticality into account and thus does not completely address the potential impact of a risk (Griffis & Whipple, 2012). Especially for the identification of possible catastrophic events, Knemeyer, Zinn, and Eroglu (2009) applied this risk management framework for low probability high impact events. Companies have to determine the key supply locations with highest probability of threats and a list of them. Approaches that can help them are “internal assassin” whereby a manager who thinks as a terrorist and thinks Lost sales opportunity with aircraft industry metal part stock out lead time delay tensions in the commodity market lack of alternative sources of supply only minor suppliers available reduced production fire at the plant
14 about how to carry out threats against a firm and the “wheel of crises” whereby certain possible consequences of crises are discussed where the wheel stops. Risk Assessment, evaluation and prioritization Risk analysis or the assessment of a risk event is nothing else then weighting or measuring the subjective probability of a risk event and the potential consequences of it from the viewpoint of the enterprise. In a later stage companies should tailor the responses and strategies will be taken to reduce either their probability or their consequences. These two aspects can be used to develop a risk map or a risk criticality matrix: the probability or sometimes called the criticality index (how critical is a possible risk for your company) and the impact or several severity classifications. Negligible impact Marginal impact Critical impact Catastrophic impact Low Least Emphasis Probability High Most Emphasis Table 1: risk criticality matrix The aim of this matrix is to evaluate each risk on their emphasis and ultimately prioritize this risks to mitigate and map them in the matrix. Griffis and Whipple (2012) notices an incomplete picture and suggest that the probability of detecting these risk factors should be admitted in the traditional two- by-two matrix used by many other authors. (see also risk aspects). How can this additional factor be integrated? This is done again with a two-by-two matrix. Figure 7: risk matrix for the additional probability of detecting risks The manager can, for a specific risk factor, assess the ease in which the occurrence of that risk factor can be monitored (from easy to difficult) on the x-axis. On the y-axis, the lead time, from short to I Least Emphasis II III IV Most Emphasis
15 long, between detection and realization of the risk is depicted. For example if our metal supplies come by ship you have to take care and map several sources of risks. An example of the first quadrant can be a mechanical failure because it may have little to no advance warning of problems but once occurred in most cases the technical staff is capable in solving these kinds of failures. Weather fluctuations like a dangerous storm are immediately detected due to the accurate weather forecasts nowadays such that a vessel is able to take an alternative route and avoid the storm. The risk of piracy at last is difficult to monitor and characterizes with an immediate recognition of the detection, resulting in greater emphasis. Sometimes the firm can draw a tolerance threshold line that divides the risks you will manage actively form those that will not be managed, after which the risks identified are sorted by expected loss (P. G. Smith & Merritt, 2002). In other words, the company selects the maximum risk criteria it can afford. That’s another way of prioritizing risks when the firm has to cope with a lot of risks, especially minor risks. Prioritization is important as firms often focus only on recurring but low- impact risks at the expense of paying attention to high-impact but less-probable risks (Griffis & Whipple, 2012). In a further stage risk can also be compared against the selected risk evaluation criteria (Mullai, 2009) and further be ranked by criticality or severity. For the assessment and estimation of a catastrophe the use of simulation and optimization can be recommended (Knemeyer et al., 2009). Other estimation methods are for example the opinion of experts combined with historical data, suitable for aircraft incidents or with the opinion of decision makers, eligible for other types of catastrophic events like nuclear reactor meltdowns. The game theory whereby an optimal strategy has to be determined between the objective function of the attacker and the constraints of the firm can be used to simulate terrorism. The output of these approaches should be a list of key locations with estimated potential loss values. Risk Management Actions and Mitigation strategies Much research is done about management actions, strategies and action plans against risks. An attempt to give a reasonable overview follows. (Kleindorfer & Saad, 2005) (Manuj & Mentzer, 2008), (C. S. Tang, 2006), (Griffis & Whipple, 2012)) Under risk management actions we understand the general used strategies towards the risks perceived. They are risk taking, risk transfer, risk reduction and risk elimination respectively. Within each risk management action several mitigation strategies can be used in succeeding this action.
16 Firstly managers can choose to take the risk. Reasons therefore can be that the risk may be perceived to be low and the company is willing to accept the risk because of very little consequences for the firm. According to two German researchers, during the financial crises times, a lot of companies accepted their risks in this country. This was more the case with manufacturing companies. But after the crisis these companies shifted towards a more comprehensive approach of risk mitigation. This is in contrast with service companies who despite the crisis stayed to perform more risk acceptance strategies (Blome & Schoenherr, 2011). Alternatively, managers can opt to transfer their risks from one company to another or subdivide their risks over several companies. This may reduce the total risk in the network if the company takes the risk can cope with it better than the company transferring it resulting in having a large supplier network. But the downsides of this risk are the high switching and administrative costs and their availability when changing from supplier. These are therefore part of transaction-specific investments. Furthermore it may decrease opportunities to achieve economies of scale. Furthermore risk can be shared in contracts with the intention of better coordination with channel partners, collaborative forecasting and collective replenishment planning which increase supply chain visibility and encourages further analysis of individual risks. They can be managed generally by developing a common network strategy, sharing best practice modes of action and contract policies. Moreover, several situations exist that there might be some risk but the company takes the needed effort to reduce it as much as possible. A common used mitigation approach could be the use of non- performance penalties built into contracts. If our metal part supplier doesn’t succeed to deliver the demanded parts, price reductions will be used as stated in the contract. Examples of other security mechanisms used to reduce risk include monitoring techniques, such as audits of supplier’s quality checks, inspections of random materials, and tracking of key performance indicators (KPI’s). At last, risk elimination may be appropriate when the firm cannot live further with this risk and must be completely discarded. AVIA decided to quit assembling their aircraft parts with that old machine that produces much defects. Sometimes choosing an appropriate risk strategy means changing current operating models or practices. This means that you systematically review your ‘inventory’ of risk procedures and controls with the aim to improve risk management practices. An example is the centralized versus localized approach of manufacturing to mitigate risks and increase benefits.
17 Table 2: risk management actions To deal with catastrophic events the company can draw a catastrophic risk management matrix which maps the key locations in the same manner as normal risks to detect appropriate risk strategies for each threat on the list like for example move a location, buy insurance, assume risk and so on (Knemeyer et al., 2009). Chaos theory may additionally provide some help to formulate appropriate catastrophic risk strategies. Wagner and Bode (2009) makes a difference between cause and effect oriented supply chain risk management practices. The first are preventive in nature. We think of information security, physical security and freight security. For example AVIA can switch in advance to a more financial stable supplier to reduce the risk of a sudden supplier default. Or the company can relocate their manufacturing plants to safer regions to avoid natural hazards like tropical storms or tsunamis near the coast. The second practice contains measures aiming at minimizing the level of damage in case of a risk event occurrence, e.g. insurance companies. The disadvantage of these companies is that they do not always understand supply chain risks and it’s difficult for them to insure a company’s own facility against disruptions from their suppliers at multiple locations. But there seems to be progress in this field: They are now providing business interruption insurance for disruptions occurring at a supplier’s facilities (Alvarenga & Lehman, 2012) for named suppliers but unfortunately don’t cover the whole network of suppliers and subcontractors. Buffering strategies, financial risk reserves and product redesign are other examples of this practice. Most of the risk handling activities proposed in the literature are rather effect-oriented than cause- oriented. As risk mitigation strategies require costly investments in equipment as well as human resources, it is important to know which mitigation strategies offer the greatest protection from risks in a certain situation. risk taking • ignoring the risk when developing a mitigation strategy risk transfer •large supplier base risk sharing •channel coordination •collaborative forecasting •collective replenishment planning risk reduction •non performance penalties •audits of supplier's quality checks risk elimination •remove machines •stop operate unhealthy production processes
18 A possibility for determining the favorable assessment costs is mapping them against the benefits from risk mitigation strategies, stated in the framework of Shavell (Kleindorfer & Saad, 2005). This results in a tradeoff between the cost of acquiring reliable information on risks and the benefits of mitigation activities. At optimum, a balance must be struck between the marginal costs and benefits of better risk assessment. So when do we have to use these mitigation strategies? Risk identification and assessment give a more specific indication on where to focus the actions. According to Griffis and Whipple (2012), strategies such as monitoring and risk taking can be used when the likelihood and potential impact are low and the ability to detect the risk is easy. When selecting a monitoring strategy, either risk reduction or elimination, you can choose to perform random inspections of products to detect errors or the risk maybe perceived low so that managers could choose to take the risk. When the opposite is true, more aggressive risk mitigation strategies like complete risk elimination need to be considered. Examples here are avoiding dangerous shipping routes or to quit outsourcing and to manufacture the product in-house to have more product and process control. If the likelihood of a risk occurring is low and detection may be easy, but the impact of the risk could be significant, then a postponement strategy may be appropriate if the event causing the risk can be postponed until more control by the focal firm is established. More control can be established through vertical integration or imposing contractual obligations on suppliers (Jüttner et al., 2003). The clearest example of a postponement strategy is producing in modular form. The advantage is that you can push your semi-finished product from surplus to deficit areas. A company that uses a postponement strategy is the computer manufacturer Dell. They produce computer hardware in modular form and let their customers and firms decide which functionalities and properties they must contain. One can also perform a demand postponement strategy and shift the demand across products towards their customers (C. S. Tang, 2006) such as a price strategy. The opposite of postponement is called speculation or also called selective risk taking and is also an option here. When you perform a speculation strategy, you build up inventory to buffer against the specific risk. In cases where impact of a risk may be low, but likelihood of occurrence is high, and the ability to detect the risk in advance is difficult, a firm may select an imitation strategy and source with the same supplier because if one firm is exposed to this risk, all firms are. A flexibility strategy at last could be used when the likelihood of risk occurrence is high and detection is easy. This could be achieved through multiple sourcing. This strategy requires some adaptation for
19 the company because in a general culture where the focus lies more on core competencies and value creation more single sourcing relationships have emerged (Blome & Henke, 2009). Strategic partnerships and alliances are an example of this sourcing strategy. Secondly, companies don’t have always the choice to choose between single and multiple sourcing. For example, when you have the choice of only one supplier, because of intellectual protection, then you have a sole sourcing relationship. This mitigation strategy is sometimes called hedging in a supply chain context because the company has a globally dispersed portfolio of suppliers. Another way to provide flexibility is the adoption of standard processes and the use of interchangeable and generic or modular parts. Finally using simultaneous instead of sequential processes in key areas as production/distribution speeds up the recovery phase after a disruption. Figure 8: risk mitigations strategies Tomlin (2009) determines also the optimal adaptation strategies when probability of supplier and customer failure is high or low. The used strategies are supplier diversification, contingent sourcing, which is adding a supplier which is only used in case the main supplier fails to deliver and demand switching. Diversification should be executed when demand uncertainty increases. Furthermore when the firm faces an increasing supplier failure probability or faces a high level of risk aversion it should opt for a contingency strategy. Demand switching is appropriate in case of a low supply risk. Mitigation strategies can change over time. Suppose company AVIA is plagued with several recalls for their assemblies because of metal parts affected with corrosion in an earlier stage. Because it isn’t always easily detectable, current employed postproduction testing is no longer efficient anymore. Rater than a strategy of control, through frequent testing, a strategy that uses severe penalties with
20 their suppliers for recalled products is more appropriate. Another possibility can be to use an elimination strategy and look for another supplier. After the appropriate mitigation strategies are determined, it is recommend developing prevention and contingency plans to reduce the risk in likelihood of occurrence and impact severity. Wagner and Bode (2009) suggest that continuity or recovery plans are important tools to ex-ante optimize the ‘firefighting’ after a disruption. These contain for example the radical design of products and the layout of the manufacturing processes. As a review to the selected mitigation strategies, some principal criteria can be addressed: efficacy or the degree to which risks are eliminated, feasibility or the aligning of the right mitigation strategy to the appropriate risk and efficiency which relates to the cost-effectiveness which was explained above. Risk Monitoring and Strategy Sharing Monitoring your risks means identifying the potential increasing trends in their probability or consequences in the future. After implementing mitigation strategies you will find that some risks are closed where the risk event has been prevented or other risks remain where the risk event had happened despite the prevention plans implemented (P. G. Smith & Merritt, 2002). Nowadays companies can employ real-time risk monitoring capabilities along with techniques to track key supply chains flows. These tools can speed response in case of numerous unplanned events. A lot of electronic and high-tech companies, who have very dependent supply chains, have integrated these tools into their standard supply chain management practices. Improving the traceability of the supply chain leads to organizations that follow key performance indicators through the entire supply chain and consequently identify risk not only with their first-tier suppliers but also with their sub- contractors ((Alvarenga & Lehman, 2012), see also supply chain security). A company first follows the risk management steps described above and analyzes its network-related risks internally. In the second phase the partners should identify the areas of risk management that require joint effort and where risks strategies should be shared. As enterprises are connected in a network, they are dependent on each other so it can be useful to share entirely or partially risk management processes and to develop collaborative means to manage the risk and communicate their views on risks. It is important that the individual risk management processes are supplemented by a collaborative process. Sheffi (2005) even argues that competitors should collaborate to control common risks.
21 Moreover in complex network environments mutual risk identification and assessment can be seen as tools for creating the risk profile of the entire network on the basis of the partners' risk profiles (Hallikas et al., 2004). The primary tool employed by the Japanese to implement closer supplier co- ordination and individual supplier development is cross-exchange of staff between buyers and suppliers. This requires the benevolence of the enterprise of exchanging inter-organizational information towards risks & rewards sharing and knowledge transfer. But it gives the firm the possibility to perform a benchmarking exercise and it generates supply-chain wide visibility of vulnerabilities and it should give the firm incentives to identify and implement disruption management systems. Conclusion Now that we have briefly described the risk management process framework, one can argue of its need. As risk mitigation strategies require costly investments in equipment as well as human resources, it is important to know if these strategies pay off towards risks of all kind. As Jüttner et al. (2003) noted, there is a supply chain trade-off decision between delivering high customer value and managing possible risks. A trade-off between extra risk mitigation costs and less costs of delivering high quality and on-time products as a main principle of supply chain management. Kleindorfer and Saad (2005) investigate if investments in risk management activities yield towards frequency or severity of accidents. With the use of variables like regulatory programs, facility characteristics and community demographics, they determined whether observed accidents in the chemical sector decreases with the use of these risk programs as mediator variables as a consequence of more severe regulatory programs, more hazardous facility characteristics or the financial structure of the company. The investigation indeed found evidence of this relationship. Dani (2009) suggests this risk management framework must be an iterative process and should not stop with one investigation of risk but instead repeat the exercise to study new issues and risks identified after the analysis of the event. Furthermore this exercise must be aligned at the strategic level of the company and according to the strategic objectives to have a clear understanding. Concluded, it is important to update the possible risk sources and strategic objectives in line with the risk event or mitigation strategies that may be adapted according to possible new discovered risk issues (Dani, 2009). We will further see that the adoption of a risk management strategy will foster the use of a proactive supply chain approach. Mullai (2009) takes it a step further and claims that the process can start at any point. The major steps of the framework (risk analysis, evaluation and mitigation) are interactive, change-responding, can be accomplished simultaneously and are aligned through risk communication (Mullai, 2009).
22 2.2 Supply Chain Management The first definition of the supply chain management is dated from the early 1980’s and compromises the following: “a ‘standard’ supply chain is a system compromising of materials, goods and information (including money), which pass within and between organizations, linked by a range of tangible and intangible facilitators, including relationships processes, activities and integrated (information) systems” (Peck, 2006) While our approach is not to give a full overview of the supply chain with all his aspects, we will nevertheless give you some information of some practical aspects in this domain that can be linked or have relationships with the risk management domain. Bullwhip effect An important issue for the supply chain is that you have to take into account the major consequences of the Bullwhip effect. Essentially, the bullwhip effect depicts the phenomenon in which the orders exhibit an increase in variability up the supply chain, even when the actual customer demands were fairly stable over time (Sterman 1989; (C. S. Tang, 2006). Cisco systems Inc. wrote off 2.5 billion in inventory due to a lack of communication among its downstream supply chain partners (Spekman and Davis, 2004(C. S. Tang, 2006). The increase in variability of the orders up the supply chain can cause many problems for the upstream partners including higher inventory, lower customer service level, inefficient use of production and transportation capacities, etc. The more distance between suppliers and the final consumer in the supply chain, the more these demand changes are compounded (Fine 1998; Lee, Padmanabhan, and Whang 1997(C. S. Tang, 2006). In order to mitigate the bullwhip effect, one needs to identify the root causes (C. S. Tang, 2006) which can be done in the first step, risk identification, of the risk management process. Secondly, many companies have switched from “local” suppliers to “low cost” and often distant suppliers on the basis of overhead cost optimization, without considering the cost of risks caused by this strategic change. Larger companies now buy from smaller suppliers in very remote areas of the globe. The extended supply chain now has many additional points of potential failure, enlarging the bullwhip effect and requiring new approaches to risk management. Companies face longer logistics lead times as well as new and unfamiliar risk profiles encompassing natural disasters, epidemics, and social, political or monetary instability (Alvarenga & Lehman, 2012).
23 An agency theory perspective The agency theory perspective justifies the differences in the objectives and risk preferences of the two parties: the principal (purchasing organization) and agent (suppliers), as well as information asymmetries. Both parties undergo an agreement with risk sharing (Zsidisin & Ellram, 2003). The aim to consider the relationship in this perspective is to reduce the purchasing firm’s risk of moral hazard and adverse selection. The first means the risk of the lack of the supplier, aware of not by the purchaser, to lever the agreed upon effort to meet customer demand. Adverse selection means the inaccurate assessment or misrepresentation form the purchaser of the sometimes unknown supplier abilities to meet customer requirements (Zsidisin & Ellram, 2003). An example of moral hazard is the unwillingness to further invest in appropriate infrastructure needed to produce the metal parts for AVIA. An example of the latter is that AVIA, unaware of the major investments in new metal production equipment, keeps further purchasing the parts from old machinery that are produced with minor quality. In essence, suppliers and buyers have to be aware of opportunistic behavior risk. This comes down to the breaking of their mutual informal agreements & contracts between the partners within the supply network in the pursuit of competitive advantage and profit (Seiter, 2009). The author proposes action programs like more communication quality, better partner selection and mutual sharing of cost accounting information to reduce opportunistic behavior directly or indirectly through reduced information asymmetry. Make-or buy decision Don’t try to force the manufacturing of complementary assets in-house when you can outsource particular needed competences in a more cost-advantageous way. For many companies, “make or buy” decisions have been chosen in favor of buying, not making. While this reduces manufacturing overhead costs, companies lose oversight of key governance and management competences and strategies. As a consequence this might introduce unknown (new) risks into the supply chain. Periodic risk rebalancing is therefore essential (Alvarenga & Lehman, 2012). Outcome-based versus Behavior-based management techniques In order to align the objectives of both agents and principals several management techniques or practices are available. These can be split in two categories (Celly & Frazier, 1996; Zsidisin & Ellram, 2003). Outcome-based management techniques address the importance of coordination of outcomes and results such as sales growth or sales in relation to targets. According to Zsidisin and Ellram (2003), the
24 use of buffer oriented techniques is an example of that group. Inventories can be held either by the purchasing firm resulting in internal safety stock or by suppliers which is supplier-managed inventory, or both. Rather than reducing the likelihood of a harmful event, firms employ buffers to reduce the disruptive effect of supply risk events. Therefore this approach is short-term oriented. Behavior based management techniques addresses behaviors such as customer education activities or selling techniques with distributers from the supplier personnel thereby signaling important objectives and suggesting specific distributor actions (Celly & Frazier, 1996). It focuses on processes, emphasizing ‘task and activities” that lead to a reduction in supply risk and is therefore long-term oriented. The findings of Celly and Frazier (1996) were that supplier personnel rely too much on outcome- based efforts when coordinating relationships with distributors. Zsidisin and Ellram (2003) on the other hand found that there exists a partially positive relationship for behavior techniques with perceived supply chain risk and this was not supported for buffer oriented techniques. The implementation of buffers is done regardless of the extent of perceived supply risk. Although each of the efforts has its downsides, outcome-based may be sometimes inappropriate in some situations and behavior-bases may be sometimes costly to the firm. An emphasis solely on maintaining buffers to manage supply risks can harm business profitability due to capital devoted to inventories and Celly and Frazier (1996) find the risk of demand fluctuation is being positively related to behavior-based management efforts. On the other hand buffers give you a form of assurance and behavior based contracts facilitates communication and mutual goal alignment. An inclusion of the two efforts will be optimal. Examples of the management techniques are given below: Outcome Buffer Behavior - communication about sales growth, market share, … - managing inventory - supplier certification - quality management programs implementation - target costing - supplier development The advantage of supplier certification is that they reduce the need for the purchasing organization to conduct time-consuming inspections. Implementation of quality management programs is also an
25 option but certification improves already the abilities of the supplier to satisfy the quality expectations of the purchasing firm. Also target costing is interesting where supplier and purchaser determine ways to drive cost out of its products. At last supplier development can meet the purchasing organization’s short or long-term supply needs. Behavior-based efforts may be predominant in franchise channel systems because inter-firm interaction should be relatively high and relationship termination is more difficult. In conventional channel systems, outcome-based efforts may be predominant because of the need to keep coordination costs down. In addition we will find more buffer-oriented approaches within the use of transactional supplier relationships whereas the use of behavior outcome-base approaches is more adopted with cooperative supplier relationships (Blome & Henke, 2009). Making your supply chain redundant versus flexible According to Sheffi and Rice Jr (2005), the following difference can be made. Redundancy activities include safety stock, the deliberately use of multiple suppliers or back-up sites. Adding redundancy is in a way needed for every day’s operability but flexibility strategies on the other hand gives more competitive advantage in the marketplace. Flexibility can be seen from 3 facets: supply, in-house conversion and distribution. It aims first of all on correct alignment of the supplier relationship with the procurement strategy whether you opt for a single supplier or for multiple suppliers or for a single supplier for each critical part. Secondly using standard processes and having multiple locations with built-in inter-operability. This allows a firm to operate in another plant once one is disrupted or the replacement of sick operators. On the customer or distribution side at last, managers face a choice about which customers to serve first after a disruption. Managers can thereby decide on several criteria like the vulnerability, the profitability and costs of all customers. Strategies like postponement or producing semi-finished products, described above, can be used. Managers in addition are often reluctant to invest in flexibility measures because they don’t see directly or hardly can estimate the impact on risk mitigation (C. Tang & Tomlin, 2009). So unlike redundancy, flexibility can also improve the competitive advantage of companies in times without a disruption. Flexibility at last is indeed important when your company executes a lot of projects because the project manager must be able to change or redefine several aspects during project execution. Furthermore projects ensure the coordination between client, contractor and supply chain. Hence, project execution translates into flexible management. But gaining more on flexibility and consequently efficiency implies the risk of losing sight and therefore as we have seen before can increase vulnerability. Companies producing components for the automotive sectors are an example of industries that focus on project management (Gaudenzi, 2009). It provides a risk management framework that is also applicable for projects.
26 C. Tang and Tomlin (2009) investigates how much flexibility is needed to mitigate supply chain risk. On the basis of mathematical models they conclude that only a small amount of flexibility strategies (e.g. multiple suppliers, flexible supplier contracts, postponement, responsive pricing and flexible manufacturing processes) is required to mitigate risk. From their PWC research (Levi et al., 2013), it seems that companies focus on flexibility and customer service levels on the one hand and other companies on cost reduction and efficiency on the other hand. It seems according to their study that those former are better coped against risks (appendix). It’s also worth noting that 80% of the cost-efficient companies face high variable supply chain lead times given that low variability is mostly one of the key drivers of an efficient operating strategy. The resilient supply chain Once gone through all the major risk management steps and approaches the ultimate goal is to build a resilient supply chain. A supply chain that is both able to absorb disruptions and risks, can proactively manage its supply chain and sees a possibility to turn the threats of a disruption in a major advantage can be seen as a resilient enterprise. Resilience is not the same as robustness. The robustness of a company is the ability to resist from an accidental event, retain to its same stable situation as it had before and stick to his initial mission (Asbjørnslett, 2009). In contrast resilience aims at a new stable situation and has the adaptive ability instead of being resistant. Building a resilient enterprise should be seen from a strategic level and changes the way a company operates and increases its competitiveness. Two important aspects determine a company’s resilience (Sheffi & Rice Jr, 2005). First is the market position. Is the industry competitive or has the company a lot of market power? Else it depends on the responsiveness of the supply chain. When companies are not so responsive they might risk losing market share. Responsive companies otherwise can increase market share or once they have a reasonable amount of market power they can lock in their leadership. Corporate culture In the search for the resilient enterprise, it is important not to underestimate the contribution of culture to an organization (Sheffi & Rice Jr, 2005). Empowering front-line employees to take initiative and guide actions is a possible step in building a suited corporate culture. Secondly Japanese lean principles from the Toyota Manufacturing System can be used like Poka Yoke and Hijunka. These concepts are in essence that one must learn from errors and fixe the root causes. Companies can minimize the risk of possible disruptions by paying attention to small problems as indications of major disruptions. At last continuous communication between all layers of a company can foster the good working of the company.
27 To encourage the need and use of risk management a disruption can be simulated and employee reactions are monitored and used in training (Kleindorfer & Saad, 2005; Sheffi & Rice Jr, 2005). This is the exercise of role playing with red and blue teams. The red team represents in most cases competitors, rivals or supply chain experts, equipped with whatever information is available. They attempt to attack the supply chain to cause major disruption. On the other hand you have the blue team who tries to mitigate or countermeasures those actions which are cost effective against the Red Team scenarios. This way of training can enhance the risk culture in the firm. This lacks yet in a lot of firms according to the authors of the paper from Cagliano, De Marco, Grimaldi, and Rafele (2012). Sheffi (2008) proposes that governments can introduce cultural aspects into disruption mitigation. Cultural changes on a societal level happen several times every century so the government should encourage and drive these trends to improve resilience along the chain. 2.3 Supply Chain Risk Management Research in supply chain risk management is still in an early stage and has been around for about a decade now. As a consequence here is a huge diversity in topics, opinions, and research methodologies in the field of supply chain risk management. Furthermore there is still a lot of variation towards the meaning of supply chain risk management according to several established focus groups. According to them, supply chain risk management is seen as a subset of Supply Chain Management (SCM) and also as a subset of Enterprise Risk Management (ERM) which can be seen in the figure below. But what compromises the field of supply chain risk management (Sodhi, Son, & Tang, 2012)? Supply chain operations and risk management processes go hand-in hand and complement one another. Figure 9: the concept of Supply Chain Risk Management (Blos et al., 2009) According to Wagner and Bode (2009), Supply chain risk management (SCRM) contains the field of activity seeking to eliminate, reduce and generally control pure risks in supply chains. But what do
28 these risks contain? According to the focus group from the research from Sodhi et al. (2012), risks in this field should concern majorly with dealing with unknown events or dealing with disruptions or disasters with low probability and high-impact. This was the opinion of almost the half of the researchers. Secondly it should concern dealing with risk within supply chain operations, according to 20% of the researchers (Sodhi et al., 2012). Kajüter, 2003 states: “Supply chain risk management is a collaborative and structured approach to risk management, embedded in the planning and control processes of the supply chain, to handle risks that might adversely affect the achievement of supply chain goals.” In this definition the focus is rather laid on the supply chain. Norrman and Lindroth, 2002 assess the focus of supply chain risk management more on a practical approach. Supply chain risk management can be defined as, “to collaboratively apply risk management process tools with partners in a supply chain to deal with risks and uncertainties caused by, or impacting on, logistics related activities or resources” Nevertheless this concept is still missing many pieces that have to be found and linked together to get a comprehensive approach of supply chain risk. SCRM is often not always established as a distinct function or department in companies. Businesses do not agree on how to integrate these risks into their decision-making processes, the risk function is typically “headquarters-centered” and there seems not be a risk regulation that covers the supply chain. However it’s necessary to integrate risk management into operations, strategic and sales planning. The work from Jüttner et al. (2003) summarizes well the concept:” “the identification and management of risks for the supply chain, through a coordinated approach amongst supply chain members, to reduce supply chain vulnerability as a whole.” And can be summarized using the figure from them.
29 Figure 10: a SCRM framework As a consequence it shouldn’t come as a surprise that research in this field leads to often unambiguous conclusions and a large variety of results. But SCRM implementation has brought up: Ericson has implemented a SCRM matrix organization that spans the entire company, from the corporate and strategic level to the functional and finally process-oriented operational level. It’s in addition worth to address the importance of smaller companies. Most SME’s are more exposed to supply chain risks whilst simultaneously disadvantaged by lack of management resources, structures, processes and expertise as opposed to bigger firms (Henke, 2009; Jüttner & Ziegenbein, 2009). In essence, for SME’s the adoption of a comprehensive risk management program can be the same except for some aspects due to their limited possibilities. Whereas larger firms can have the time and the access to control all their supply chain partners not strictly limited to 1st tiers, SME’s are encouraged to select their most important one for further investigation. In most cases information is gathered from their personal network and there is no planned supplier data collection but instead these companies rely more on social interactions with their suppliers (Ellegaard, 2008). It’s not practical for them to map their entire supply chain (Jüttner & Ziegenbein, 2009). Consequently before the risk identification process, Jüttner and Ziegenbein (2009) require the mapping of supply chain vulnerability against their strategic importance to investigate. Furthermore the adoption of expensive risk management tools is not always value adding for them. That is because tools often are designated for specific tasks and are highly sophisticated and therefore not always suitable for SME’s. In summary a 3-phase risk management approach for SME’s is given. The tactics towards the risk aspects are according to a case study from Ellegaard (2008) as follow: probability reduction has the highest priority and effect reduction through multiple sourcing was not
30 practiced at all. Also small companies don’t gather much information from their partners that can help reduce risk. A company must regularly check their activities in a manner that they fulfill their strategic objectives dictated by the firm’s vision. A framework that can aid to measure performance is the Balanced Scorecard of Kaplan and Norton. Their model doesn’t contain only financial objectives but incorporates also internal, customer and innovative & learning perspectives. When dealing with supply chain performance, no primary use of financial performance measures is advised (Sheffi & Rice Jr, 2005).This framework which focuses on performance measurement as a consequence should be integrated with supply chain risk management. Risk and performance are directly related because the higher the risk taken the higher the expectation of suitable returns (Sheffi & Rice Jr, 2005). As more resources and attention is given to supply chain risk management an aligning with performance management is indispensable. Besides the electronic multinational Ericsson (Norrman & Jansson, 2004) another company that has successfully implemented a risk management strategy and accompanying supply chain culture is the Nisan Motor Company (Schmidt & Simchi- Levi, 2013, (Levi et al., 2013)). After serious harmful problems due to an earthquake and Tsunami in Japan, they incur a $ 200 billion loss. “Nissan’s production capacity was perceived to have suffered most from the disaster compared to its competitors.” But the company managed to increase their production with 9.3 % in comparison with an overall decrease with its competitors. How was this company able to manage this? First they implemented an integrative risk management framework with risk identification as early as possible, assessment and performed countermeasures against them. Second the plant has formulated a continuous ready plan with their suppliers. Moreover the team was empowered with local decision management. In addition the company structured its supply chain as flexible as possible. At last, there was extended enterprise visibility and warnings to between internal and external business functions (Levi et al., 2013). 2.4 Supply Chain Security Management Supply Chain Security Management (SCSM) is defined as “the application of policies, procedures and technology to protect supply chain assets from theft, damage, or terrorism and to prevent the introduction of unauthorized people or weapons of mass destruction into the supply chain. Security practices can be control measures implemented in several processes and certainly at several gates throughout the supply chain where products arrive (Voss & Whipple, 2009).
31 The difference with Supply chain risk management is the fact that Supply chain security management focuses more on prevention of contamination, damage or destruction of the supply chain assets and products by policies, procedures and technologies whereas the former investigates more the likelihood of outcomes being susceptible to disruptions that can damage the supply chain (Autry & Sanders, 2009). This paper provides a dynamic capability framework that can be used to implement security practices. These capabilities include processes, technology and human resources. For example the use of radio frequency identification is a technology capability. With the use of the Radio Frequency Identification (RFID) technology and their tracking and tracing capabilities, abilities rise to identify disruptions quickly. Furthermore sensitive control systems often can identify a disruption before its cause is apparent. The management of inventory is an example of a process capability and governmental security initiatives are human resource capabilities. The problem with risk mitigation activities, in particular those that don’t start at the place of original production, is that technical control allows hidden action of participants in the process chain (Mau & Mau, 2009). Therefore a supply chain wide security control system or management is needed. A comprehensive control platform is required to manage supply chain security effectively. The goal for supply chains, those that doesn’t consist of a complex network or single supply chains, is to secure worldwide and realize complete traceability of all involved products and inputs at all the supply chain levels resulting in continuous transaction data in both directions, upstream and downstream. To trace all the necessary information, it is useful to make use of an independent database. In this way effects on the products of your firm can not be overseen thanks to a centralized data management system. The basic thought for implementing security initiatives are basically the same as for the use of risk mitigation strategies. Advanced or high proactive security initiatives pay off and improve security and firm performance but most firms have not progressed beyond basic physical security measures such as infrastructure management and therefore have not derived the service benefits from higher level security measures (Voss & Whipple, 2009). Consequently this creates a dilemma in an optimal tradeoff between cost of implementation and efficiency of the initiatives but also for firms that face greater customer or government requirements. More security improvements sometimes can lead to a decrease of flexibility and firm performance but aim to create a secure supply chain that maintain advanced security processes and procedures.
32 2.5 Conclusion Concluded, risk mitigation measures can be seen and implemented from 3 different levels. The strategic level which focus on for example alternative suppliers, at the tactical level (e.g. improved demand forecast) as well as the operational level (e.g. business continuity plans). Although risk management practices and business continuity planning is left too much to security or insurance professionals in companies, it should be noted as a strategic initiative and must be implemented in an integrated risk framework approach. As a consequence this risk practice helps to build a resilient enterprise and supply chain.
33 3 Conceptual framework The goal of this work is to determine how several environmental aspects (e.g. market size, bargaining power) may have an impact on the managers or the firm’s risk perception by which we mean the probability of risk and their impact. Furthermore we investigate if these perceptions might influence risk mitigation strategies performed by the firm. In particular we will make the distinction between more proactive versus more reactive performance towards risks. At last we want to examine if certain mediators and moderators do play a role in selecting that particular strategy. On the one hand we dive into how varied a firm does perform supply chain management practices and their choice towards risk management. On the other hand we look at the design and the complexity of the supply chain and see if any conclusions can be made with this aspect towards risk. While previous research (Ellis, Henry, & Shockley, 2010) has determined the impact of environment on risk perception, investigation towards the use of proactive versus reactive management is still a gap in the literature that hasn’t been explored yet. Instead Ellis et al. (2010) investigated overall supply disruption risk towards the search of alternative sources of supply. 3.1 Environment Several papers deal with the environment as an important factor for the determination of risk (Ellis et al., 2010) (Ritchie & Brindley, 2009). Ellis et al. (2010) chooses also for the integration of some environmental factors like technological uncertainty and market thinness but also for some company specific characteristics like item customization and item importance. The environmental construct here proposed is as follow: PEST-analysis An exercise for the company in determining the strength of the environment and the external factors can be the adoption of a so called PEST analysis. This analysis contains the macro-environmental aspects of Political forces, Economic forces, Social forces and Technological forces. It’s an environmental scanning component that can help you to determine trends outside the venture. The purpose is to detect driving factors and uncertainties.
34 Figure 11: PEST analysis Porters competitive model and SWOT-analysis A second approach to identify your environment is the so famous Porter’s five forces model. This framework provides you and let you think about you’re major outside forces that come into play with your venture and determines the dynamic tensions between players. The aim of this tool is to perform a competitive analysis because each player may have a sufficient amount of power. It highlights problem areas as well as possible opportunities or whether you have a competitive advantage or you have to accept a reasonable amount of risk. An exercise that can complement this investigation is the SWOT-analysis (Strengths, Weaknesses, Opportunities and Threats). The threat of potential new entrants can limit your own company’s activities but there may be entry barriers that could prevent them to start doing business. Second, suppliers may have bargaining power that impacts your business such as a unique product they deliver to you that requires a high degree of specification and specialization. Customers likewise may have power too over your business for instance if they are able to integrate backwards into the supply chain (e.g. manufacturing your product themselves). In addition there may be a threat from substitute products. Note that this doesn’t contain physically the same products but may also encompass a different transportation mode like for example plane versus car. Substitute products can be more or less attractive depending on for example the switching costs a customer must pay. For example the adoption of a new product would require the customer to buy new equipment or additional software and so on. At last, existing rivalry among industry players can make your environment very turbulent. If there are many •technological pace •adoption cycles •R&D •education •demography •work aspects •financial markets •economic cycle •change of industry and markets • laws & regulations • political (in)stbility • government spending Political Economic TechnologicalSocial
35 competitors and there is little product differentiation, the resulting strategy in your industry will be price competition. Figure 12: Porter's Five Forces Model Market variables At last we introduce some market variables that also can be perceived from the environment and are important for the firm. First we account for the perception of the market size: “is it growing or declining rapidly?“ Moreover has the industry market many or few segments resulting in a company’s market span. In addition, “has the industry many or few competitors?” This describes the market concentration. At last, we can focus on the ease to enter the market. “Is it open to new players or closed to new players?” So concluded, our first hypothesis will be: Hypothesis 1: Environment constructs – Risk perception relationship H1: there is a significant relationship between Market-variables, some PEST-analysis variables & PORTER-analysis variables and the probability of risk and their impact. 3.2 Risk perception or representation of risk Risk perception is defined as the decision maker’s assessment of the risk inherent in a situation (Sitkin & Pablo, 1992). This refers to the assessment stage in the risk management process described earlier.
36 Prior research has already proved the relationship between risk perceptions and risk behavior although as a mediator between risk propensity and risk behavior (Sitkin & Pablo, 1992). We will reinvestigate this relationship but consider the perception in both risk probability and risk impact. Hypothesis 2: Risk perception – Risk management relationship: H2a: there is a relationship between the risk perception of risk probability on the choice of pro versus reactive management. H b2: there is a positive association between a higher perception of risk impact on proactive as well as reactive management and the relation on proactive is stronger. We opt to choose the aspects of risk, impact and probability, from a perception point of view rather than objective assessments of risk. That is because a manager will take decisions on the basis of his perceptions of degree of risk and which actions he will do. Risk perception is dependent from several aspects: as earlier stated risk is dependent on how the problem is framed and the problem domain familiarity. Furthermore an individual’s propensity towards risk also influences the perception. At last the homogeneity of the management team and organizational control systems come also into play (Sitkin & Pablo, 1992). We represent the risk from 3 perspectives. First from a supplier point of view: “What is the probability that a key supplier will fail to supply your operations and what will be the impact?” Secondly, risk perception will be gathered from the own manufacturing operations: ”what is the probability and the impact that the company’s operations are interrupted and affects your shipments?” At last probability and effect can be seen from the distribution side: “What is the probability and impact that a company’s shipment operations are interrupted affecting their deliveries?” 3.3 Proactive versus Reactive Risk Strategies The distinction between proactive and reactive risk management can be approached in relation to their timeframe. Proactive can be seen from an ex-ante and reactive from an ex-post perspective. But it emphasizes more than solely this. What is proactive risk strategy actually? Proactive risk management means that you identify risks and do something about them before they affect your project or your business operations. This is in fact the use or the implementation of the risk management framework that was introduced and
37 explained in all its facets in the previous chapter. However, this is easier said than done. Many managers are reluctant to spend time or money on potential problems, since they have plenty of other problems already. Proactive management of risks is a style that is often foreign to many Western managers. Companies fail at risk management in general because they fail at one of the two fundamentals of managing risk well: cross-functionality on the one hand and being proactive on the other hand. Companies make the mistake to believe that innovation is only in Research & Development, and that is where most of the risk lies. A lot of companies lack the cross-functionality capability between departments nowadays. Companies fail at pro-activeness because they wait until late in the project when risks start occurring and they let risk management lapse and as a consequence they perform in a reactive manner. So in concrete by proactive we mean for example a proactive vulnerability analysis or a proactive mitigation strategy or Proactive risk management of supply risk must be seen as a competitive advantage for the firm. And what is reactive risk strategy then? Reactive management is the reverse of proactive management and is the same as firefighting after a disruption. It means that a company waits until situations are almost hopeless and then saving them miraculously or takes the needed actions to counter the risks after the damage is observed. Thereafter the company develops prevention and contingency plans to make the disruptions less severe in the future if those same disruptions or failures occur (Wagner & Bode, 2009). “Being reactive is the default position when a risk materializes” (Dani, 2009). So instead of dealing with risks on a day-to-day basis, responding reactive is in fact being ready when disruptions occur. Activities such as contingency plans that states corrective actions to resolve undesirable consequences or the accumulation of knowledge and experience that can be used when the same problems occur in the future are two example of risk management from a more reactive approach (Ritchie & Brindley, 2009). An example of a more reactive approach is the concept of Supply Chain Event Management (SCEM) (Gaonkar & Viswanadham, 2004). Although this offering doesn’t use a suitable risk framework, it includes supply chain visibility, track and trace and alert messaging which merely address human operator problems and leave him as the only person to resolve the issue. One can consider the distinction between Outcome or Buffer oriented strategies and Behavior based strategies stated earlier already as an indicator in choosing for a more proactive versus reactive risk strategy. Because when you do for example supplier development or supplier certification you proactively manage your risks with your suppliers instead of putting buffers in a reactive manner. The latter is according to Zsidisin and Ellram (2003) not a manner of risk reduction but is a common approach for every company. Honda focuses on continuous process improvement of suppliers
38 through evaluation in a lot of areas, certification and development (Nelson et al. 1998). Their emphasis on process improvement has had enormous results on reducing the probability of supply risk occurrence. The Honda example supports the finding that improving supplier process is a proactive way to manage supply risk (Zsidisin & Ellram, 2003). Also the distinction according to Sheffi and Rice Jr (2005), also early stated the difference between redundancy and flexibility activities that can be projected to the difference between proactive versus reactive strategies. While some redundancy is part of every resiliency strategy, it represents purely cost with limited benefit unless it is needed after a disruption. In that regard adding redundancy can be more seen of a reactive risk approach. Flexibility on the other hand can be justified for a normal business approach without even taking into account the benefits of risk mitigation strategies and cost reduction but results from increasing flexibility are difficult to measure from a cost accounting perspective. Wieland and Wallenburg (2012) suggest the distinction of Supply chain management practices in the following manner. Risk management can be done proactively through supporting the robustness of the supply chain and reactively on the other hand by supporting the agility of the supply chain. Robustness was explained above. Agility means “the ability of a supply chain to rapidly respond to change by adapting its initial stable configuration” (Wieland & Wallenburg, 2012). Instead of being robust, the initial state is adapted which refers to more reactive strategies whereas robustness refers more to more proactive because of the status-quo policy measures taken in advance. Resilience previously explained means then the ability of adaption but with the aim to gain a competitive advantage after the disruption occurs. This can be seen as the cherry on the cake for a firm that is able to whatever disruption occurs, it can make a positive consequence for the firm with it. It was found that being agile has a strong positive effect on the supply chain’s customer value but not directly on business performance while robustness has a strong positive effect on both the business performance and Supply Chain’s customer value (Wieland & Wallenburg, 2012). Case studies proved that robustness is important to handle supplier-side risks while agility is necessary to deal with customer-side risk and therefore the amount of both performances should fit to the firm’s competitive strategy. Jüttner et al. (2003) conducted interviews and noticed that “Still, whilst the drivers are recognized as competitive pressures with risk implications, it appears that the implications are often sorted out on an ad hoc basis as organizations go along. What seems to be missing is a more proactive approach where risk implications are anticipated at an earlier stage.’” The aim in this work is to wonder
39 whether companies are more using proactive risk strategies or industries tend to more respond and recover from disruptions. We will investigate the relationship between risk perceptions and risk management. Ellis et al. (2010) draws a similar relationship in their Theoretical Model of Risky Decision Making but instead focus on the search of alternative sources as management action. Our focus is on the proactive and reactive risk management strategies. 3.4 Moderators and Mediators As moderators for our model we selected two major broader categories that deal with supply chain management namely practices in supply chain management and supply chain complexity and design. But in order to describe these categories we first draw some general considerations in advance. First, not every company is equally enthusiastic in sharing their supply chain management practices or aligning their practices with their major supply chain partners because of confidentially in the company or risk that the supplier will gain confidential information and the risk of abuse of knowledge. Second, in several countries, regulatory or legal authorities can put constraints on both of these moderators. They can contain supply chain-relevant (trade and transportation) laws and policies. This can be an important factor of uncertainty to the setup and operation of supply chains. Several laws and policies include the ability to obtain approvals necessary for supply chain design activities and supply chain operation practices (Wagner & Bode, 2009). 3.4.1 Supply chain management practices. Supply chain management practices are for example sharing information with your suppliers or forming collaborative approaches with the suppliers. Also the practice of system coupling can be seen as a management practice for the supply chain. The degree of a firm’s system coupling takes into account both the intensity and extent to which information about demand, capacity, inventory and scheduling is shared and used by the firm in both directions of the supply chain (Barut, Faisst & Kanet, 2002). These are not necessarily the same as risk management practices but can be part of a proactive or reactive risk management strategy. As explained above, when managers plan to proactive mitigate risks, choosing different suppliers or sharing risk information with suppliers, can be part of a risk management strategy. So, we can say that a condition to perform business will be an effective supply
40 chain and their management. But also performing risk management can be added as an extra value for the firm. To achieve the company’s business objectives, firms require effective use of integrative supply chain management practices. These can include information, knowledge and design integration. On the one hand sharing information between supply chain partners is needed to jointly control risks (as seen above). But on the other hand it can decrease competitiveness. There can be a leak of knowledge and a diminishing control about their inflows and outflows (Narasimhan & Talluri, 2009). Integration with suppliers but also the internal integration and integration with customers prove to be important drivers for a company’s performance in case of customer satisfaction & competitive performance (Zhao, Huo, Sun, & Zhao, 2013). Furthermore Zhao et al. (2013) found out that supply chain risks are negatively related to supply chain integration. We will investigate a similar relationship namely the effect the firm’s perception of risk probability and impact has on the choice of conducting several supply chain management practices. In fact we examine whether firms who conduct a lot of supply chain management practices will choose more for risk management strategies and more proactive or more reactive kind of strategies. Hypothesis 3a: General SCM practices mediator between environment and risk perception H3a: the relationship of the strength of the environment and the strength of risk perceptions is increased by the mediator variable general supply chain management practices. Hypothesis 3b: General SCM practices mediator between risk perception and the risk management strategies H3b: the implementation of general supply chain management practices increases the relationship between the risk perception indicators and the risk management strategies. Ellegaard (2008) conduct a study in the German automotive industry and found that companies with high supply chain performance show a higher degree of supply chain risk management but the relationship can also be interpreted the other way around. Moreover it found a difference between companies using preventive risk mitigation instruments contrary to those using reactive instruments. Using the supply chain vulnerability map based on (Sheffi, 2005) as previously explained, Blos et al. (2009) identified vulnerable impact rates through the four vulnerability quadrants between
41 automotive and electronic companies in Brazil. The automotive industry’s highest problem is the complex production forecast but they facilitate it through modular production. Electronic companies on the other hand are more vulnerable because of the high dependence on the Asian market and the nature of their products. In the light of the growing trend to act more environmentally friendly and more focus on the durability of company performances, we will also look at sustainable practices employed through the supply chain. For example, managers can perform sustainability performance assessments with their suppliers or they can train their supplier’s personnel in sustainability issues. Thus we will add the following hypotheses. Hypothesis 3c: Sustainable SCM practices mediator between environment and risk perception H3c: Sustainable SCM practices do change the relationship between environmental factors and the perceptions of risk. Hypothesis 3d: Sustainable SCM practices mediator between risk perception and risk management strategies H3d: there is an effect on the relationship between risk perceptions and management strategies when taking the sustainable supply chain management practices into account. 3.4.2 Design and complexity of the supply chain The complexity of a supply chain is a function of their different levels and tiers in the supply chain. Tightly coupled supply chains are furthermore of uttermost importance because the chain is characterized by their components wherefore there are few possible substitutions available. According to Christopher and Lee, 2001 such tighter coupled supply chains are likely more prone to disruptions. The traditional approach, first introduced by Porter, is the supply chain in its literal meaning, namely one supplier delivers parts for your own business and you produce your products that will be distributed towards or directly delivered to your customer(s) that supports a value- adding creation towards the whole chain. Figure 13: Supply Chain Supplier Manufacturer Distributor Customer
42 Today, this picture seems somewhat too simplified and today’s business activities are more integrated resulting in a supplier network. Manufacturers appeal on many suppliers and deliver to more than one customer (see figure below) resulting in the term of supply network management (SNM). Figure 14: Supply Network Furthermore companies perform outsourcing activities when realize that it’s cheaper to produce some parts or perform administrative activities elsewhere than in-house. If after some period the knowledge of performing these activities is known and the part production has come down to a fine art it can be advantageous from a cost perspective to do an in sourcing strategy. Furthermore companies don’t live alone anymore. Be aware that off shoring activities can increase lead times (Khan, Christopher, & Burnes, 2009). At last, the creation of joint ventures, alliance partners and more make the design of supply chain not less easy. Consequently supply chain networks are often Figure 15: a more complex and adaptive network
43 complex and adaptive but also very dynamic. They are composed of intricate, sometimes counterintuitive and nonlinear linking elements that include a lot of information and constraints on each other under which transactions, activities and processes take place (Manuj & Mentzer, 2008). Although the concept of value network has grown in past years, from a risk perspective managers should be thinking in terms of networks that create value. Value is transferred from several actors in a value network. These values which are often not in monetary terms translated can be traded for other values with other partners. Impact flows constitute the intangible components (information, knowledge …) shared through partners. Hallikas and Varis (2009) make the comparison with a business ecosystem theory as a conceptualization, meaning “The environment beyond the core business & the value network and can be seen as interdependencies between several business actors and the business environment in which they operate.” You should thus determine the dependencies that are most critical to your business and select your business in your environment that are important enough to make collaborations with in the future. Several business aspects can be interpreted from a biological view: Survival from environmental pollution Being head-to-head with technological change Health of the flora Being able to produce value and share it between members of the system. Receive enough air and sunlight A company should position itself in the network where it can most generate value now and into the future Ecosystem health regulator Powerful business leaders who create platforms for other industry players and link other companies Photosynthesis Business Lifecycle Dependency from other organisms in nature Partnership and complex interplay between competitive and cooperative business strategy Table 3: the business from a biological view
44 In figure 16 you find an example of a value network, here for Google. Hallikas and Varis (2009) have determined the values and impacts with their resulting possible risk through major players in the Google network. The different players in the network represent nodes where each node represents a value-adding activity for another node in exchange for some value for the other party. Risk can occur because partners often transfer unwanted knowledge in close cooperation with each other. Furthermore the risks in the network are dependent on the behavior of the interconnected systems of companies. Global supply chains drive up the level of complexity, which drive up the level of risks. Because of an increasing dynamic complexity and feedback mechanisms these networks have to be aware of risk and consequently can think of the use of risk management. Giving the importance of complexity, as you can read a bit further, we also included this facet as a moderator in our network. It is also important to recognize the indirect relationship because of their increased invisibility these may pose a higher risk for the head company (Hallikas & Varis, 2009). The intangible nature of products in the exchanges of the value network increases the complexity of the supply chain network (Hallikas & Varis, 2009). Less complex supply chains are better able to manage supply risk and can improve performance. Some complexity reduction measures are reducing the number of suppliers and serving varied customers through one integrated supply chain (Hallikas et al., 2004). Kleindorfer and Saad (2005) addresses the effect of alternative supply chain design options on the efficiency of the supply chain to various sources of disruption but this research was not yet related to the use of risk management strategies. Figure 16: supply information network
45 Relying on a single supply source may be risky nowadays. Think of Ericson who has lost 400 million of damages from a fire at Philips, its single supplier (Norrman & Jansson, 2004). So Autry and Sanders (2009) suggests 3 flexible multiple supply source options that are used by the firms to mitigate their risk. One possibility is to develop multiple suppliers each with a certain advantage for the same component. For example a company can rely on a flexible supplier that delivers products at any time for one half and rely for the other part on a cost-efficient supplier. When there is more demand for any given feature (cost versus delivery) the company can rely on the spot market to make the difference. Secondly, once a company uses a sourcing strategy then it’s optimal to use partly local supply sources to supplement their supply base. This can serve as a back-up for the company. At last it can be optimal to rely on suppliers with multiple manufacturing sites at multiple locations, rather than just one site so that supply can be easily shifted from one site to another. Although the majority of authors suggest multiple sourcing against risk, Blome and Henke (2009) tempt to direct the discussion in favor of single sourcing. The major motivations therefore are the cost reductions through e.g. standardization and transaction. Second, dependency on one single supplier doesn’t calculate in higher risk because highly cooperation can provide strategic and competitive advantage for buyer and supplier. The condition therefore is that they are mutually dependent on each other because when there is a one-sided dependency from only one party the relationship becomes more risky. Therefore proactive risk management must also play a role in single sourcing situations. Finally the decision of the number of suppliers should be made separately from the choice of supplier relationship, cooperative versus transactional, according to Blome and Henke (2009). The case of Marks and Spencer has shown that relying on few amounts of close suppliers and implementing direct sourcing instead of cooperating with a large number of independent suppliers, they could proactively mitigate and manage their risks to its supply chain in a better way (Khan et al., 2009). In a later stage they continued to extend their supply base through suppliers that were engaged throughout the product development process instead of addressing to suppliers who take a high share of their margins that M&S had in the past. We will examine the following hypothesizes. Hypothesis 4: Supply chain complexity and design moderator H4a: there is a difference in relationship between the environment indicators and risk perceptions strategies between firms whose amount of complexity and design of the supply chain is higher and those firms who have a lower amount of supply chain complexity and have a lower level supply chain design.
46 H4b: there is a significant difference between high-level designed and more complex supply chain or networks and low-level designed and less complex supply chains for the relationship between risk perceptions and risk management strategies. 3.4.3 Continental differences One can consider if there might be differences between continents. There are certainly differences in the way Asian countries operate their business as compared to European countries. But reflect these differences also in terms of risk management and risk perception. In a later section we will investigate for any differences. Hypothesis 5: Europe versus Asia moderator H5a: there is a difference in the relationships for environment and risk perception between the continents Europe and Asia H5b: there is a significant difference between Europe and Asia for the relationship between risk perceptions and risk management. 3.4.4 Comparison with 3 years ago A lot can change in 3 years, certainly in these sometimes turbulent times. Therefore we will analyze if there are significant differences in the last 3 year as compared to the current implementation. Hypothesis 6: Current implementation versus effort in the last 3 years moderator H6a: there is no significant difference between the policy of 3 years ago and the current implementation for the risk management programs and supply chain management practices that were related with risk perceptions for both Europe and Asia. H6b: there is a difference for the supply chain management practices, general and sustainable, and risk management strategies, proactive and reactive, that were taken 3 years ago and their current implementation related to the risk perception of the firm for both complex as simple network designs.
47 3.5 Model diagram Figure 17: Model diagram An overview of the variables to investigate and their relationships is given above. The general framework is composed of only the general “blue” variables. In a later stage we will examine the intervention of the “green” mediator variables, the SCM practices. In the end we explore the “red” moderator variables.
48 4 Data Collection Now that the conceptual model is extensively explained, it is time to gather some data. Data is needed to test and investigate the relationships between the several constructs in the framework. In the introduction, I will explain the steps and effort needed that has been done to collect data. Thereafter some exploratory factor analysis will be done to validate the several items that form the constructs. 4.1 Introduction To investigate a conceptual framework, there needs to be data and, more preferable, a lot of data. So to get this data, I have participated with the International Manufacturing Strategy Survey or IMSS. It is a research project carried out every 4 years by a global network. This project studies production and supply chain strategies within the manufacturing and assembly industry through a detailed questionnaire administered simultaneously in many countries by local research groups. In order to get worldwide data from this research project I’ve collected industry data for Belgium. This was made possible in cooperation with the Vlerick Business School. Through telephone contact with several companies I’ve tried to get the goodwill from manufacturing managers or plant managers to participate with this research project and their commitment to fill in the detailed questionnaire. The questionnaire used can be found in the appendix. The purpose was to address manufacturing companies from the metallic, electronic and automobile sector. More concrete, the following manufacturing industries have been addressed along with their associated industry codes: - 25: fabricated metal products, except machinery and equipment - 26: computer, electronic and optical products - 27: electrical equipment - 28: machinery and equipment not elsewhere classified - 29: motor vehicles, trailers and semi-trailers - 30: other transport equipment To find the Belgian companies from the appropriate industries I made use of the Amadeus Export Data Belgium Database. I also approached the companies who have participated the previous time for this research project but some companies didn’t exist anymore and other companies had changed their manufacturing managers compared to the past. To get the company info and the managers contact information I consult the Trends Top database. Furthermore I provide an online possibility to fill in the questionnaire through SurveyMonkey so that the managers don’t have to print
49 out the questionnaire format before filling them manually but instead could fill them online whenever they will and further continue when they have some time. 4.2 Results For Belgium I could approach 30 companies in total from the 6 sectors stated above. There are 136 operations mangers from these companies contacted. 24 of them weren’t eligible for data incorporation because the company was going to shut down in less than half a year or the company didn’t fulfill a manufacturing activity or only performs a distribution function. So 30 out of (136 -24 =) 112 managers participated which results in a response rate of about 26, 79%. There ain’t no such thing as a free lunch. After sending the Belgium questionnaires I get the worldwide data from the IMSS. In total 19 different countries have participated through this large- scale research. Table 4.2 gives you an idea of the Industry Code distribution for Belgium and for the whole world. Table 4.3 gives the participated country distribution. Figure 18: proportion of participated industries Figure 19: proportion of participated countries
50 4.3 Variables and constructs For the construct of environment we include the following variables: PORTER PEST Market Competitive rivalry (A2e) Threat of substitution (A2g) Bargaining power of suppliers (A2h) Bargaining power of customers (A2i) Social pressure (A2k) Rate of technological change (A2b) Environmental pressure (A2j) Demand fluctuation (SC3a) Market size (A2a) Market span (A2c) Market concentration (A2d) Market entry (A2f) Table 4: Environmental constructs The first five variables can be derived directly from the Porters five forces model that we described before. As regard the PEST analysis we have items about social force, technological force and environmental pressure. In addition we included some market variables for the respective industries. Finally we can pick demand fluctuation as additional variables to complete the construct. As described before, for probability and impact we have surveyed the following variables: Probability Impact Probability of a key supplier failure (R1a1) Probability of operations failure (R1b1) Probability of shipment failure (R1c1) Impact of a key supplier failure (R1a2) Impact of operations failure (R1b2) Impact of shipment failure (R1c2) Table 5: Risk perception variables For the risk action programs, we have earlier made a distinction between more proactive as well as more reactive strategies. So we made a distinction between these strategies towards the following mitigation strategies: Proactive Risk Management Reactive Risk Management Preventing operations risk (select a more reliable supplier, use clear safety procedures, preventive maintenance) (R2a2) Detecting operations risk (internal or supplier monitoring, inspection, tracking) (R2b2) Responding to operations risk (backup suppliers, extra capacity, alternative transportation modes) (R2c2) Recovering from operations risk (task forces, contingency plans, clear responsibility) (R2d2) Table 6: Proactive and Reactive Management When investigating the questionnaire we could pick several supply chain management practices surveyed. These can be divided in 2 groups namely practices towards general external action
51 programs as well as practices in relation with sustainability management. The former group analyzes both these practices for a company’s key suppliers and customers. We will include these variables as a mediator in our model and see if these variables as a whole strengthen the relationship between Risk Perceptions and the choice of Proactive versus Reactive management. General external action programs Information sharing: (about sales forecast, production plans, order tracking and tracing, delivery status, stock level) (SC6a2 & SC6f2). Development of collaborative approaches: (supplier development, risk/revenue sharing, long-term agreements) (SC6b2 & SC6g2). Joint decision making: (product design/modifications, process design/modifications, quality improvement and cost control) (SC6c2 & SC6h2). System coupling: (vendor managed inventory, just-in-time, Kanban, continuous replenishment) (SC6d2 & SC6i2). Developing an international sourcing strategy: (supplier scouting at the international level, open foreign sales office, develop an international purchasing office or distribution network) (SC6e2 & SC6j2). In relation with sustainability management Suppliers’ sustainability performance assessment: through formal evaluation, monitoring and auditing using established guidelines and procedures (SMh2). Training or education for suppliers’ personnel (SMi2). Joint effort with suppliers to improve their sustainability performance (SMj2). Table 7: Supply Chain Management Practices For supply chain complexity and design we will include the following variables in table 12. We will test if there is a difference between higher complex supply chains or networks and less complex networks towards the adaptation of more proactive risk management strategies. This construct will be applied as a moderator in our model. We can roughly divide our variables in 2 broader categories namely role of the plant and variables in relation to the manufacturing network but we will not use this distinction in our model. Role of the plant Producing in one plant versus producing at multiple plants in the network (G2a). The role of your plant in the network is stable versus the role of your plant in the network is revised and changed flexibly if needed (G2d). To what extent is the plant responsible for the supply chain (procurement, logistics, supplier
52 development) (G3b). The degree of integrated information system in the company wide network (G4c). Manufacturing network Improve the use of technology to support communication with other plants of the network (e.g. ERP integration, shared databases, social networks) (G7d2). Developing a comprehensive network performance management system (based on cost, quality, speed, flexibility, innovation, service level) (G7e2) Change the balance between outsourcing and in-sourcing (PC1b). Table 8: Supply Chain design and complexity As a last we will investigate if there is a moderating effect between European countries and more Asian countries in our model. Next we have grouped items into several groups in the constructs. 4.4 Exploratory factor analysis (EFA) The measurements of the variables were done by using a 5 point Likert-type scale where 1 means low and 5 means high. Before moving to our methodology we first have conducted an exploratory analysis with the use of the Cronbach’s alpha measure of the item factors that completes the constructs as we can see in Appendix 3. For environment we have dropped the last variable “demand fluctuations” since when leaving this one our consistency scale measure increased so in fact the construct will receive a higher alpha when dropping this variable. The same has happen for the construct of Supply chain complexity and design where we deleted the variables “the extent of the responsibility of the plant” and “the changing balance between out-sourcing and in-sourcing”.
53 5 Methodology and Analysis 5.1 General model Figure 20: General Model First we will investigate the general model in too deep. The general model is the model without mediators or moderators and can be seen in figure 20, here above. In a later stage the intervention of mediators and moderators will come into play. 5.1.1 Confirmatory Factor Analysis (CFA) of the general model Here we have a reflective measurement scale because the some indicators that form a construct are correlated and interchangeable. To do a CFA we applied the PLS algorithm with the use of smartPLS. Explanation of target endogenous variable variance  The indicators Risk probability and Risk impact explains 53.6 % for Proactive risk management (R² = 0.536) and 52, 4 % for Reactive risk management (R² = 0.524).  The risk perception variables probability and impact are explained for as little as 8.5% and 7.5% by the environment constructs PORTER, MARKET and PEST.
54 Inner model path coefficient sizes and significance  When looking at the inner model, we can conclude that the effects of Risk impact (0.614 & 0.604) on the both proactive and reactive strategies will be stronger than the effect of risk probability (0.138 & 0.140) on them.  All of the indicators of risk perception will be statistically significant on the risk management strategies.  The market variable will as only environment indicator have a statistically significant relationship towards both perception and impact. The other environment variables PEST and PORTER have a factor loading of less than 0.1 and we predict that those variables will not be significant. Latent Variable Indicators Outher model Loadings Indicator Reliability Cronbach's alpha AVE Market MarketConcentration 0,7323 0,53626329 0,6552 0,4903 MarketEntry 0,7395 0,54686025 MarketSize 0,6516 0,42458256 MarketSpan 0,6733 0,45333289 PEST EnvirPressure 0,7782 0,60559524 0,6863 0,6006 SocPressure 0,733 0,537289 TechnologicalChange 0,8118 0,65901924 PORTER BargainingPowerCust 0,6644 0,44142736 0,6433 0,4713 BargainingPowerSuppl 0,6421 0,41229241 CompetitiveRivalry 0,6793 0,46144849 ThreatSubstitution 0,755 0,570025 RiskImpact Imp_Operations 0,9783 0,95707089 0,9632 0,9316 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9659 0,9362 Prob_Shipment 0,964 0,929296 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9796 0,95961616 0,9593 0,9609 Detecting 0,9809 0,96216481 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9489 0,9513 Responding 0,9764 0,95335696 Table 9: CFA summary table Indicator reliability Indicator reliability shows how reliable the items are that forms the constructs per item. To come to the reliability indicators, we have to quadrate all the outer model loadings. All the indicators show a reliability that is higher than 0.4 which is the minimum level required for exploratory research.
55 However, a value that is more than 0.7 is preferred. This is the case for the Risk perception constructs and risk management constructs which show levels that are higher than 0.9. Internal consistency reliability The measurement of the alpha in a Cronbach’s alpha analysis calculates the reliability of the construct as a whole. This analysis is based on the foundation that every item in a scale must sufficiently be correlated with any other item from the same scale. For this measurement we use the Cronbach’s alpha measurement indicator. It is recommended for your scale internal consistency that Cronbach’s alpha should have a value of 0.6 à 0.7 as a lower limit. We see that this is the case. However we see that this measure is higher with the risk perception and risk management constructs than those for the environment. Convergent validity Convergent validity means when two or more different methods that measure the same concept correlate high. To check for convergent validation we have to look for the AVE measure and we have to assure that this above 0.5. As we look to our latent variables we see that it’s more or less the case. When we round up the variables MARKET and PORTER, we see that these have a value of 0.5. Discriminant validity MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability MARKET 0,700214 PEST 0,5927 0,7749839 PORTER 0,561 0,5419 0,68651 ProactiveRiskMgmt 0,2867 0,222 0,1657 0,9802551 ReactiveRiskMgmt 0,2926 0,2286 0,1639 0,9588 0,975346 RiskImpact 0,2707 0,1828 0,1814 0,7277 0,7195 0,9652 RiskProbability 0,2908 0,1713 0,1829 0,6439 0,6376 0,8235 0,9675743 Table 10: Discriminant validiy Discriminant validation means the opposite of convergent validity and exists when two or more constructs, measured through one or more methods, correlates low and consequently differ sufficiently from each other. To check for discriminant validity we must assure that the square root of the AVE value of each latent variable, in the table showed in bold, must be higher of the correlation of the latent variable with the other variables. In practice the figure in bold must be higher than the values in the according row and column. As we check this for all our variables we see that this is the case for all our construct
56 variables. For example the AVE value of Proactive Risk Management is found to be 0.9609 so its squared root is 0.9803 which is higher than all the values of its row and its column. Conclusion To conclude our confirmatory factor analysis we can say the following. The reliability measures (indicator reliability and Cronbach’s alpha) are both good. For validity measures we say that the environment variables Market and PORTER are somewhat weak for the AVE measure but the discriminant validity of the constructs is ok. 5.1.2 Comparison with alternative frameworks In general, we could have opted to include market entry also in our PORTER construct and include a possible economic indicator in the PEST construct (for example MarketSize) but this threatens the discriminant validity. Indeed, Appendix 5 shows that the discriminant validity is threatened. This is because the same item is used for more than one construct. Furthermore the AVE variable is bad for the PORTER construct. At last several indicator reliabilities are also less than 0.4 which is too low. What reliability and validity should we have if we put all the environmental factors together in one construct? Appendix 5 provides some figures. A lot of environmental reliability indicators are less than 0.4 and even less than 0.3 but the Cronbach’s alpha for environment is actually good in concerning the reliability indicators. For the validity indicator the same conclusion can be made. Although the AVE is very weak (0.3648) the discriminant validity scores are better, especially for environment. So according this concept, we’re stuck in between: 1 of 2 indicators of reliability and validity performs well but the other bad. The follow figure demonstrates the confirmatory analysis for several constructs. At the bottom we added 3 other frameworks as well. We can conclude that our general will be the most fitted.
57 indicator reliability internal consistency reliability convergent validity discriminant validity general model Good Good Neutral Good market entry in PORTER and market size in PEST and included in market Bad Good Bad Bad all environmental factors Bad Good Bad Good market entry in PORTER and market size in PEST but excluded from MARKET Bad Bad Bad Good market entry in PORTER and market size in MARKET but the latter excluded from PEST Bad Good Bad Bad market size in PEST and market entry in MARKET but the latter excluded from PORTER Good Neutral Neutral Good Table 11: CFA summary table for comparison with other frameworks 5.1.3 Descriptive statistics Rivalry among competitors is seen for most firms as relatively high whereas the threat of substitution is seen as much lower. Furthermore customer bargaining power is much higher than the bargaining power of suppliers which in fact stress the distributor or customer as an important player. From the 3 environmental construct, it appears that Market variables are perceived as the most influenced ones to the firm. The other environmental construct variables are jointly a little less important but still above a neutral perception of 3. For the entire world, it seems that the threat of a key supplier failure is the highest possible (2.74) and the risk of shipment operations are the lowest (2.32). As for probability, the same conclusions can be made for impact. But what is more important is that the impacts for the three categories are much higher. It means that companies deal with medium probability risks (+/-2.5) and higher impact (+/-3.5). It seems that most firms see their risks with a lower probability but with higher impact. It appears that risk management has increased since 3 years for the four indicators. Preventing has the highest score and recovering the lowest both for current implementation and 3 years ago.
58 5.1.4 Hypothesis testing Now we will start to do so hypothesis testing and thus check the structural path significance. Outer model T Statistics BargainingPowerCust <- PORTER 2,9481 BargainingPowerSuppl <- PORTER 3,6311 CompetitiveRivalry <- PORTER 3,5914 ThreatSubstitution <- PORTER 4,4393 MarketConcentration <- Market 10,3393 MarketEntry <- Market 10,1612 MarketSize <- Market 4,9895 MarketSpan <- Market 6,8934 EnvirPressure <- PEST 4,4864 SocPressure <- PEST 4,2431 TechnologicalChange <- PEST 5,1977 Imp_Operations <- RiskImpact 261,8998 Imp_Shipment <- RiskImpact 100,6175 Imp_Supplier <- RiskImpact 91,4657 Prob_Operations <- RiskProbability 254,047 Prob_Shipment <- RiskProbability 101,3114 Prob_Supplier <- RiskProbability 104,3062 Recovering <- ReactiveRiskMgmt 130,6781 Responding <- ReactiveRiskMgmt 157,9065 Detecting <- ProactiveRiskMgmt 149,0042 Preventing <- ProactiveRiskMgmt 180,756 Table 12: outer model T-statistics All of the T-statistics are larger than 1.96 so we can conclude that the outer model loadings are highly significant. Inner model Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,2371 0,0629 3,7657 Market -> RiskProbability 0,2763 0,066 4,2393 PEST -> RiskImpact 0,0227 0,0444 0,5158 PEST -> RiskProbability -0,0102 0,0441 0,2755 PORTER -> RiskImpact 0,0417 0,0489 0,7381 PORTER -> RiskProbability 0,0401 0,0482 0,6756 RiskImpact -> ProactiveRiskMgmt 0,6023 0,0843 7,2829 RiskImpact -> ReactiveRiskMgmt 0,593 0,0846 7,146 RiskProbability -> ProactiveRiskMgmt 0,1464 0,0699 1,9812 RiskProbability -> ReactiveRiskMgmt 0,1477 0,0709 1,9755 Table 13: inner model loadings and T-statistics
59 Market variables are significant with a significance level of 5%. The other environmental variables do not have an effect on the perception of impact as well as probability. This supports Hypothesis 1 for market variables but not for PORTER and PEST variables. Risk impact perception is strongly significant on the 5% significance level towards risk management strategies. On the other hand, risk probability is just not significant on the 5% significance level but certainly on the 10 % significance level. This supports the hypothesis 2a for risk impact on both risk management strategies but not really supports the 2b hypothesis. 5.1.5 Multicollinearity Multicollinearity in fact occurs when two predictor variables are highly correlated. The threat of this phenomenon is that coefficient estimates can change drastically in response to small changes in the data. To test the multicollinearity, statisticians often apply the tolerance or Variance Inflation Factor (VIF) values. The tolerance factor must be higher than 0.2 and the corresponding VIF factor higher lower than 5. We see that this is the case for all of our constructs. Coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) -7,45E-06 0,023 0 1 Market 0,072 0,041 0,072 1,755 0,08 0,322 3,104 PEST 0,071 0,033 0,071 2,129 0,034 0,487 2,055 PORTER -0,048 0,034 -0,048 -1,389 0,165 0,459 2,181 RiskImpact 0,603 0,041 0,603 14,659 0 0,32 3,122 RiskProbability 0,123 0,041 0,123 2,966 0,003 0,317 3,159 Table 14: Multicollinearity
60 5.2 Input from mediator variables 5.2.1 Framework Figure 21: framework with general SCM mediator between environment and risk perception (first framework) In the following section we will analyze the effect of supply chain management practices, both the general ones and also the efforts towards more sustainability. Because the degree of supply chain management practices might enforce the relationship between environmental factors and risk perception but might also strengthen the relationship on this perceptions towards risk management strategies, we will investigate both mediating effects. The framework where general SCM practices are mediated between environment and risk perception is given in figure 21. The framework for the general SCM practices mediated between risk probability/impact and proactive/reactive risk management is given in figure 22.
61 Figure 22: Framework with general SCM mediator between risk perception and risk management (second framework) We see that explanation of target endogenous variable variance is increased both for proactive and risk management when we introduced this mediator to additionally explain the endogenous variable. Secondly, the inner factor loadings for the relationship between risk probability and both management strategies have declined below 0.1. We might expect that this relationship will be gone. In appendix 4 we found both frameworks but instead with the interplay of sustainable SCM practices. The R² of this framework has slightly augmented to 55.1 % for proactive risk management and 54.2 % for reactive risk management when considering sustainable practices between risk perceptions and management. 5.2.2 Confirmatory factor analysis If we take a look at appendix 6, we found suitable indicators for all of the factors of the mediators. The determination coefficient R² has increased from 53.6 % and 52.4% to 61.7 % and 60.9% for proactive and respectively reactive management. We see that for the general SCM there are very high scores for Cronbach’s alpha and for the AVE measure. Moreover, the indicator discriminates good with the other constructs. The same can be said for the sustainable SCM. 5.2.3 Descriptive statistics with mediator variables General supply chain management practices Also for supply chain management practices we see that all the indicators have increased from 3 years ago. Yet we see a different ranking from all the practices. What is striking is that Supplier
62 collaboration and sharing information with the supplier scores 2 times as the highest. It seems that firms highlight the importance of sharing information above performing joint actions. Sustainable supply chain management practices When we analyzed the descriptive statistics we see that managers performs on average more on performance assessment with the suppliers and the least on joint efforts with the suppliers with the aim for more sustainability. What we see is that on average these numbers are somewhat lower than the general SCM practices. To compare with 3 years ago all these indicators have increased. 5.2.4 Structural path significance Because we in advance consider SCM practices might mediate both the environment – risk perception relationship as well as risk perception – risk management relationship, we investigated both frameworks. Appendix 6 shows the T-statistics for the inner loadings for the first framework. We consider both the general SCM as well as the sustainable SCM. The outer loadings were all significant and are not retaken here. No differences occur between general and sustainable SCM. PORTER variables stay not significant and all market variables stay significant. In addition PEST variables seem to have now a significant relationship with SCM practices for both mediators and a significant relationship with probability for the sustainable SCM mediator. What is striking is whereas risk probability towards risk management was here above yet significant on the 5 % level for the general model we see that this relationship has been softened. Also for the second framework we can draw some pretty similar conclusions. The outer loadings are once again not retaken but they are all significant. We see that the relationship between risk perceptions and SCM practices as well as their relationship with risk management strategies are all significant. What is more remarkable here is that the significance of the relationship between risk impact and SCM practices differ from general and sustainable SCM practices. It seems like the mediator has some effect on the relationship between these two variables. But can we talk of a mediating effect between our constructs? 5.2.5 Mediating effect and hypothesis testing To test for the mediating effect we can make use of the Sobel test. It tests whether a variable mediates the significant relationship between 2 other variables. To test for this phenomenon we apply his following formula: Formula 1: Sobel test statistic
63 It’s a test statistic that decides whether the variables significant mediates i.e. strengthen or soften the relationship between 2 other variables. Therefore we need some indicators. First we need for the 2 indicators the β’s from the 2 independent relationships, namely a and b, and their standard errors, SE_a and SE_b. For example, if we want to calculate the Sobel test statistic for the general SCM mediator between Market and risk probability for the first framework we must compute their β’s and standard errors. We find this numbers in appendix 7. For a = 0.1848, b =0.4971, SE_a = 0.0599 and SE_b =0.057 we found a Test statistic z = 2.9085 which is significant on the 5% significance level and therefore there is a mediating effect for this relationship and the mediator decreases the dependency from 0.28 to 0.187. So we will now examine if the 3th hypothesis is fulfilled. If we look at appendix 7, we can draw conclusions for both mediators for the first framework. Except for the relationships with the PORTER variables, the mediating effect is significant for the environmental variables on risk perceptions. The market variable is partial mediated because this relationship remains significant. With regard to the PEST variable, the choice of SCM practice is not really clear: The relationship is partial mediated for the general SCM mediator but not for the sustainable SCM mediator. Risk probability might be fully mediated by sustainable SCM practices in relation with the PEST variables. This can be seen as follow: managers perceives the environmental factors (e.g. technological change, social pressure, environmental pressure) and act with its suppliers in a sustainable manner with in turn can adjust their probability of risk but apparently not for the impacts of risk. What about the second framework, we can conclude the following: There is a difference between the effects of the mediating factors on the variables. In fact, when we only consider the 3 sustainable SCM variables, the relationship between risk impact and risk management is not mediated. Second, the relationships are all weakened with the interplay of the mediators except for the impact of risk on pro- and reactive risk management. Apparently managers take their degree of general supply chain management practices into account when performing a particular level of strategy. Figure 23: Mediating effect
64 5.3 Input from moderator variables 5.3.1 Descriptive statistics and differences for the moderator variables Europe and Asia Figure 24: Environmental factors between Europe and Asia Figure 24 shows the difference between European and Asian countries regarding the environmental factors. What we clearly see, is that a lot of environmental factors are perceived as stronger by Asian countries except for customer and perception power. If we look at appendix 8 for paired sample t- statistics, we find more or less the same conclusions. All the environmental variables are significantly different between the continents, except for customer power, which is perceived as lower for Asia, and market entry, which is not perceived as lower for Asia. 3,36 3,33 3,46 3,49 3,86 3,00 2,94 3,11 3,71 3,33 3,24 2,68 3,69 3,62 3,65 3,81 3,90 3,16 3,00 3,23 3,60 3,50 3,36 2,58 3,19 3,20 3,32 3,30 3,83 2,89 2,89 3,02 3,79 3,24 3,16 2,67 market size techn. Change market span market concentration competitive rivalry market entry threat of substitution supplier power customer power env. Pressure social pressure demand fluctuation 0,00 0,50 1,00 1,50 2,00 2,50 3,00 3,50 4,00 4,50 Europe Asia Whole World
65 Figure 25: Risk Probability and Impact for Europe and Asia Here we see that Asia perceives more risk probability than the European companies but instead weight the impact as much stronger than certainly for supplier disruptions. In a later stage we will investigate if this risk perception difference will also have a different relationship towards risk management strategies. For risk probability the indicators are significantly different but for risk impact this counts only for the perception of impact from a supplier default (Appendix 8). Increase / Decrease Risk action programs in the last 3 years (2010 – 2013) Preventing Risk Detecting Risk Responding Risk Recovering Risk Whole World + 3,02% + 5,84% + 6,09% + 6,66% Asia + 11,36% + 15,06% + 12,75% + 14,63% Europe - 2,67% + 0,13% + 1,48% + 1,41% Table 15: Risk management for Europe and Asia Table 15 gives you an idea of the risk management development programs in the last 3 years (from 2010 – 2013). As we can see Asian countries have developed or done more effort the past 3 years in risk action programs despite their already high level in 2010. Europe on the other hand shows a minor increase for detecting and reactive risk management and even a higher decrease for preventing risk. Seems like those countries believe that risk management programs are less fruitful for their operations or they feel themselves in less crisis situations than 3 years ago. Appendix 8 proves that for all the risk management indicators there are significant differences. The highest difference is for responding to risks. At last, for the mediator variables general and sustainable SCM practices, we found in appendix 8 that all the indicators are clearly different. These practices seems to be more implemented by Asian countries. In particular joint decision making with their customers is more performed for Asian
66 countries. What is more appealing is that Asian countries do more effort for sustainable practices, in particular for again joint efforts and also supplier education on durable issues. The same accounts for Supply chain design and complexity. Only integrated information systems seems to have no significant difference between countries. Low versus high complex supply chain designs When we look at appendix 8, we find that nearly all environmental variables are perceived as higher for more complex networks except for customers bargaining power but seems not to be significantly different. Threat substitution is neither significant on the 5% level but appears to be a little higher for more complex networks. Different means occur for all the probability of risk indicators for the complexity moderator. However, it seems that their risk perception of impact differ not so meaningful. Only the risk of shipment disruptions is perceived stronger by higher complex networks. Increase / Decrease Risk action programs in the last 3 years (2010-2013) Preventing Risk Detecting Risk Responding Risk Recovering Risk Low SC complexity -0,51% 1,87% 1,67% 5,58% High SC complexity 2,18% 5,74% 5,90% 7,52% Table 16: Risk management for lower and higher complex networks Furthermore, what about the risk action programs, we see that more complex networks implement more risk action programs than 3 years before. But is this relationship statistically significant different between lower and higher supply chain complexity? If we look at Appendix 8, we can conclude that it certainly is. At last, also general and sustainable SCM practices does in fact differ from their degree of complexity of their supply chains. System coupling appears to have the greatest difference between the two categories. Indeed, lower designed and complex networks don’t always have the availability of loose coupling. Also training for the sustainable SCM is perceived as the greatest difference between the two groups. The lowest difference is for collaborative approaches. It can be seen for firms with a less developed supply chain as another, often cheaper way, to integrate their supply chain.
67 A note for the validity of the Supply chain design and complexity construct and the determination coefficients for the frameworks according to the moderator variables To give an idea of the validity of the construct, appendix 6 provides some figures for the factor analysis. As we can see the indicators satisfied their threshold values and besides discriminates well from the other indicators If we take a look at table 17, we discover that the variability in the model explained by the R² is increased with 5 to 6 % when we only consider European countries. When only modeling Asian countries, the explained variance decreases with7% for proactive risk management but only 3 % for reactive risk management. R² Proactive Reactive Asia 54,50% 58,10% Europe 67,30% 66,10% Low SD&C 34,30% 35,90% High SD&C 6,30% 5,60% Table 17: determination coefficients When considering for the degree of SCD&C, the determination coefficient has drastically decreased. A possible reason for this trend might be the decreased amount for data for this figures. Several data points lack some values for indicators that describes the SC complexity construct. The aim is to investigate this construct rather as a moderator and for the purpose to investigate differences then nearly explain the whole model with these data only. 5.3.2 Moderating effect and moderated mediation Explication for the difference calculation We will test for some moderating variables, which mean if relationships are in fact moderated or different between 2 groups. This is called multi-group moderation and we can see the T-statistics in appendix 8. First our dataset were split into Asian countries and European countries. It seems that for the former group we have 317 observations and for the latter 503.
68 Asia Europe (m-1)^2 99856 Sample Size 317 503 (m+n-2) 818 Sample Mean -0,0215 0,1802 (n-1)^2 252004 Standard Error (S.E.) 0,0821 0,0787 sqrt(1/m+1/n) 0,071712 t-statistic 1,7020 1st half denominator 0,822824 p-value (2-tailed) 0,0891 2nd half denominator 1,908111 sqrt(1st half + 2nd half) 1,652554 Full denominator 0,118508 numerator 0,2017 Figure 26: calculation for the multi-group moderation t-statistic and p-value Formula 2: t-test for multi-group moderation and moderated mediation If we want to calculate for example the significant T-statistic for differences between Europe and Asia for the relationship between the probability of risk and proactive risk management we must use the sample means and standard errors. These can be found in SmartPLS after bootstrapping. To calculate the t-statistic we must fill in the formula given above. Therefore it is easy to split the calculations a bit like as done above. For the moderator supply chain design and complexity the dataset is again split in two. The first group consists of the observations which have an average of the 5 indicators that forms this construct of less than 3. The second group those that are more or equal than 3. The choice of a threshold of 3 was decided because the earlier threshold of 2.5 generates less than 100 observations for the first group. In order to conduct these paired t-tests with some validity, we opt to choose for groups for at least 100 observations. So these t-statistics are calculated in the same way. We can also determine the moderating effect through relationships with mediators or otherwise called the moderated mediation. These can be seen as the total effects between variables. The
69 separate effects which are included under multi-group moderation are not influenced by the SCM practices mediator. The total effects which were the same as their separate effects were not retaken in the moderated mediation table. In the next two sections we will try to see if we can support hypothesis 5 and hypothesis 6. General SCM mediator Risk probability – risk management difference between Europe and Asia There is a significant difference between Europe and Asia in the relationship between Risk Probability and Risk management strategies on the 10% significance level. Europe has a significant relationship between them and Asia totally not. The difference between Asia and Europe is more significant for reactive (1.869) than for proactive risk management strategies (1.702). Risk Probability –risk management & SCM for lower and higher complex supply chain networks There is a significant difference between high-level complex supply chains and low-level supply chains in the relationship between risk probability and risk management strategies as well as supply chain management practices on the 10% significance level. Lower level complex supply chains have a stronger relationship for the probability of risk towards both proactive as well as reactive risk management than higher-level complex networks. The difference between Low and High complex networks is more significant for proactive (1.812) than for reactive (1.702) risk management. The relationship between risk probability and supply chain management practices is significant for high-level supply chains but not for low level supply chains and the difference is also significant. Sustainable SCM mediator SCM – risk impact between lower and higher complex networks in the first framework What is surprising when we analyze the T-statistics is that for the first framework only the sustainable supply chain management practices are different moderated towards risk impact. It seems that lower complex networks have a significant relationship for their sustainable SCM practices towards their perception of their impact of risk.
70 Risk probability and SCM practices – risk management for lower and higher complex networks in the second framework Secondly when analyzing the t-statistics for moderators we see that distinction between risk probability and risk management, both proactive and reactive, goes also for sustainable variables. Furthermore, and this is new, for the sustainable moderator it appears that also the sustainable practices are different between these groups for the relation towards risk management. But in this case it are the high complex networks that shows a significant relationship between these practices and both proactive and reactive management. Risk impact – SCM practices for lower and more complex networks in the second framework Last but not least, we found one moderated relationship on the 5 % significance level, namely the relationship from the risk impact towards sustainable practices. Lower complex systems take in mind on their perception of risk impact to choose the degree of sustainable practices. Mean Values, standard errors and T-statistics for the two groups for multi-group moderation first framework < sustainable SCM Low complexity High complexity SCMpractices -> RiskImpact -0,1185 0,0538 2,0577 0,0138 0,0466 0,451 second framework < general SCM Asia Europe RiskProbability -> ProactiveRiskMgmt -0,0215 0,0821 0,3400 0,1802 0,0787 2,136 RiskProbability -> ReactiveRiskMgmt -0,0402 0,0854 0,5207 0,1877 0,0806 2,1759 second framework < general SCM Low complexity High complexity RiskProbability -> ProactiveRiskMgmt 0,3734 0,1654 2,5179 0,0449 0,098 0,5057 RiskProbability -> ReactiveRiskMgmt 0,3654 0,1608 2,5375 0,057 0,0997 0,6483 RiskProbability -> SCMpractices -0,0927 0,144 0,8437 0,265 0,1278 1,9934 second framework < sustainable SCM Low complexity High complexity RiskImpact -> SCMpractices -0,3231 0,1505 2,2591 -0,0271 0,0725 0,2424 RiskProbability -> ProactiveRiskMgmt 0,3621 0,1671 2,3725 0,0626 0,0824 0,8433 RiskProbability -> ReactiveRiskMgmt 0,3397 0,1633 2,2728 0,0614 0,0855 0,7836 SCMpractices -> ProactiveRiskMgmt -0,0156 0,0708 0,3214 0,2079 0,0724 2,6722 SCMpractices -> ReactiveRiskMgmt 0,0345 0,064 0,3724 0,1723 0,0467 3,4437 Table 18: Mean, standard errors and T-statistics for the two groups for moderation Conclusion Despite that there is not so much found on the 5% significance level, we can make some conclusions. What about sustainable practices, it seems like there are no differences between Europe and Asia. Furthermore it is surprising but the differences between complex networks for sustainable SCM practices and risk impacts are associated with each other in both directions. But both moderated relationships are not equally significant. Apparently low complex networks focus significantly more on sustainable SCM practices to perceive the impact of risk than general SCM practices.
71 Furthermore it appears that lower complex networks more account for their perception of risk possibility to perform their amount of risk management whereas high complex networks not. Also European countries do this strategy and Asia not. What is surprising is that the higher complex supply chains act on their general supply chain management practices in relation to their probability of risk and not at all in the relation towards risk management. It shows that more complex networks do in fact care about their supply chain. Once they perceive a higher risk probability they align their supply chain management practices in the same way. This is not the conclusion for a less complex supply chains. A last remark that we can make is that for sustainable practices, more high complex supply networks focus on their SCM practices whereas low complex supply networks focus more on their risk probability towards both proactive and reactive risk management. 5.4 Some comparison with 3 years ago A limitation of this study and this gathered data is the absence of data for “effort in last 3 years” for some variables i.e. for environment, risk probability, risk impact and some supply chain design & complexity variables. We will now consider to make some conclusions for the last hypothesis (H6). In earlier section we have showed that there were progressions in risk management compared to 3 years ago. We will now execute some paired sample t-test in order to proof these progressions but also take a look at evolutions of general and sustainable SCM practices. As we look through appendix 9 section paired sample t-test for variables, we found evidence for evolution of most of the variables. What about the dataset for Asia, we see that all the indicators have increased in value between 3 years ago and now. For European countries the story is somewhat different. Although these companies have increased on average their reactive risk management programs, proactive risk management has not significantly changed in the past 3 years. Preventing has decreased and detecting increased some but not much. Furthermore some SCM practices proved to be not significant more practice in the past 3 year. This is the case for performance assessment for the company’s sustainability practices with their suppliers (SusPerfAssS is not significant on the 5 % significance level). In addition, only sharing information with the suppliers as well as with customers has increased between the past 3 years for European companies. Other general SCM practices proved not to be significant on the 5 % significance level. At least, when we conclude for the whole world, we can say that all indicators have significantly changed except for developing an internal sourcing strategy with suppliers.
72 There seems to be no significant relationship between current implementation and effort in the last 3 years for all the investigated relationships. To come to this table we applied the same formula as for calculating moderator t-statistics. The highest t-statistic is for the relationship between SCM sustainable practices on reactive risk management for the first framework with low complex networks. Apparently, no significantly different relationships exist between now and 3 years ago. We will now summarize or findings of the proposed hypothesis. 5.5 Hypothesis summary Hypothesis summary H1 Accepted for market, Rejected for PORTER and PEST for both risk perceptions H2a Accepted for both risk management strategies H2b Accepted for both risk management strategies H3a Accepted for market and PEST Rejected for PORTER for both risk perceptions H3b Accepted for both risk perceptions on both risk management strategies H3c Accepted for market and PEST Rejected for PORTER for both risk perceptions H3d Accepted for risk probability on both risk management strategies Rejected for risk impact on both risk management strategies H4a Accepted for the multi-group moderation relationship between sustainable SCM practices and risk impact * Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H4b Accepted for the multi-group moderation relationships between risk probability on both risk management strategies where both SCM practices are mediator * Accepted for the multi-group moderation relationship between sustainable SCM practices and both risk management strategies * Accepted for the multi-group moderation relationship between risk impact and sustainable SCM practices Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H5a Rejected for all multi-group relationships Rejected for all moderated mediation relationships H5b Accepted for the relationship between risk probability on both risk management strategies where general SCM practices is mediator * Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H6a Rejected for both frameworks and both SCM practices H6b Rejected for both frameworks and both SCM practices Table 19: hypothesis summary table Table 19 summarizes the proposed hypotheses from our conceptual framework and their results. ‘Accepting’ denotes here that the test was significant and consequently supports our conceptual framework. The Hypothesis’s noted with a * are the those ones who are significant on the 10% significance level.
73 6 Overall Conclusion “Many companies perceive if they have to choose their effort between costly risk mitigation strategies or more flexible supply chain strategies as if they to choose one of both.” We began this work with declaring several concepts like risks, their aspects and drivers, disruptions, vulnerability and many more. We proposed the framework to implement a comprehensive risk management on the basis of many previous literature researches. Thereby we addressed the fact of risk monitoring and sharing risk information. The concept of SCRM has been linked with SCM, SCSM and ERM. The aim of this work was to investigate the difference between the use of more pro-active and more re-active risk management strategies to manage supply chains against disruptions. A conceptual framework has been presented which includes environmental factors, SCM factors, risk factors and SC Design and Complexity variables. Data was gathered with the aim to test our framework and propose several hypotheses. We have aimed to compose the most reliable model and with the highest validity. Two frameworks have been investigated where both mediator variables were between the constructs of environment and risk perception on the one hand and between risk perception and management on the other hand. Some environmental variables (i.e. market variables) have impact on the perceptions of risks (hypothesis 1). Secondly, we saw a stronger influence of risk impact than risk probability on risk management strategies and most firms perceives a medium risk probability with a higher impact on average. Although risk management has increased by several countries in 3 years, we do not find so much difference for more proactive or reactive risk management implementation. Yet, proactive management is slightly more in favor (hypothesis 2) but not always related with other variables. General SCM practices ameliorate the model a lot, in terms of declared variance. Sustainable SCM slightly improves the model. Apparently managers take their degree of general supply chain management practices into account when performing a particular level of strategy (hypothesis 3). When we analyzed for Europe & Asia and Low versus High SCD&C (hypothesis 4) we found a lot of differences in their environmental and risk perceptions but also in their general and sustainable supply chain and risk management. In fact, Asian countries score a lot higher for nearly all the indicators. Even worse, the practice of preventing risks more proactively has declined for European countries. Asian countries perceive their probabilities of risk as much higher and yet their impact as a lot lower than European countries. Nevertheless these Asian countries implement more risk
74 programs. They are able to compare a highly leveled supply chain with an advanced level of risk management in comparison with European countries. This is the challenge in today’s business operations. When we dig a little deeper, i.e. performed some further analyses in differences between moderators in the several relationships from our conceptual framework, the outcomes are not always explicable. Risk impact might be associated with sustainable SCM practices in both directions and the perception of risk probability differs to their level of risk management implementation. Lower complex SC focus more on cheaper forms of SCM practices like information sharing and collaborative approaches whereas more advanced supply chains focus on joint decision making and system coupling. In addition, lower complex supply chains focus more on risk perception in their relation to risk management implementations. Complex networks do in fact care about their supply chain because their risk management practices are more dependent of these. Once they perceive a higher risk probability they align their supply chain management practices in the same way. This is not the conclusion for a less complex supply chains (hypothesis 5). At least, a lot of variables have increase in 3 years but not significant different relationships were found. Why can it be possible that a lot of risk management efforts don’t really break through or not much changed in their level of implementation in a lot of companies? Perceive these companies as if the economic crisis is gone for some time? And what about preventing natural hazards? First, we must say that a lot of companies see that risk management strategies are not always time and cost-effective. Certainly European countries think that risk management doesn’t always prove to have added value. Secondly, some strategies might reduce flexibility of operations or supply chain management strategies because they could impose business restrictions. Finally, the feeling of many managers that the crises is yet behind our backs may arouse that we must not stick to constantly prevent risks or performing business operations in a cautious way. Besides, a lot of papers in the risk management were written during or after the crisis (i.e. 2008- 2010). One of them is the German research from paper 37. They found that the financial crisis has not influenced the relationship between enterprise risk management and supply chain (risk) management. We found overall relationship between supply chain management practices and proactive risk management for both current implementations as well as for the effort in the last 3 years for the whole world but with no significant differences. This indicates that after the crisis, companies stick to align supply chain practices and risk strategies but not everyone augment them.
75 7 Limitations and possibilities for future research In this work we investigated risk management strategies and divided them in more reactive and more proactive ones. But more risk management practices might upgrade the framework in value. Because, what we have noticed, is that the constructs for proactive and those for reactive action programs don’t discriminate too much. This makes generality of findings more difficult. In future research it might be interesting to further investigate related risk program strategies such the practice of an integrated risk management or the usage of a supply chain or catastrophic vulnerability framework. Second point we like to address is that besides a lot of data that is used for the current implementation, more data for 3 years ago would be fruitful to make more conclusions for the evolutions of more indicators (risk probability, risk impact, environment factors and some supply chain design and complexity variables). In addition future research can be executed for the relationship of more risk indicators. Besides the risk perceptions of risk probability and risk impact, other risk indicators like for example speed of risk, risk frequency, probability of detecting and the perception of overall risk can be further investigated towards risk management. Conclusions have been made for the whole world an some differences for the European and Asia countries. The dataset used contains 14 European countries, good for 503 companies and 4 Asian countries that represent 337 companies. But this dataset does lack some world figures namely only Canada from North America has participated for this research. So other world continents were not represented through this data. Further research can research data for the missing countries here. At last, no statements have been made for the Belgian firms. Although a modest range of 30 companies have participated through this research, several indicators didn’t contain 30 values. It was therefore not recommended to execute statistical analysis. Future research could enlarge the Belgian dataset.
76 8 References Altay, N., & Ramirez, A. (2010). Impact of disasters on firms in different sectors: implications for supply chains. Journal of Supply Chain Management, 46(4), 59-80. Alvarenga, C., & Lehman, E. (2012). Accenture: Managing Risks in the Supply Chain: Reaching New Standards (pp. 1-12): Accenture. Asbjørnslett, B. E. (2009). Assessing the vulnerability of supply chains. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 15-33): Springer. Autry, C., & Sanders, N. (2009). Supply chain security: A dynamic capabilities approach. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 307-329): Springer. Blome, C., & Henke, M. (2009). Single Versus Multiple Sourcing: A Supply Risk Management Perspective. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 125-135): Springer. Blome, C., & Schoenherr, T. (2011). Supply chain risk management in financial crises—A multiple case-study approach. International Journal of Production Economics, 134(1), 43-57. Blos, M. F., Quaddus, M., Wee, H., & Watanabe, K. (2009). Supply chain risk management (SCRM): a case study on the automotive and electronic industries in Brazil. Supply Chain Management: An International Journal, 14(4), 247-252. Cagliano, A. C., De Marco, A., Grimaldi, S., & Rafele, C. (2012). An integrated approach to supply chain risk analysis. Journal of Risk Research, 15(7), 817-840. Celly, K. S., & Frazier, G. L. (1996). Outcome-based and behavior-based coordination efforts in channel relationships. Journal of marketing research, 33(2), 200-210. Chen, F. Y., & Yano, C. A. (2010). Improving supply chain performance and managing risk under weather-related demand uncertainty. Management Science, 56(8), 1380-1397. Dani, S. (2009). Predicting and managing supply chain risks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 53-66): Springer. Ellegaard, C. (2008). Supply risk management in a small company perspective. Supply Chain Management: An International Journal, 13(6), 425-434. Ellis, S. C., Henry, R. M., & Shockley, J. (2010). Buyer perceptions of supply disruption risk: a behavioral view and empirical assessment. Journal of Operations Management, 28(1), 34-46. Gaonkar, R., & Viswanadham, N. (2004). A conceptual and analytical framework for the management of risk in supply chains. Paper presented at the Robotics and Automation, 2004. Proceedings. ICRA'04. 2004 IEEE International Conference on. Gaudenzi, B. (2009). Assessing risks in projects and processes. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 67-82): Springer.
77 Griffis, S. E., & Whipple, J. M. (2012). A Comprehensive Risk Assessment and Evaluation Model. Transportation Journal, 51(4), 428-451. Hallikas, J., Karvonen, I., Pulkkinen, U., Virolainen, V.-M., & Tuominen, M. (2004). Risk management processes in supplier networks. International Journal of Production Economics, 90(1), 47-58. Hallikas, J., & Varis, J. (2009). Risk management in value networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 35-52): Springer. Haywood, M., & Peck, H. (2003). Improving the management of supply chain vulnerability in UK aerospace manufacturing. Paper presented at the Proceedings of the 1st EUROMA/POMs Conference. Henke, M. (2009). Enterprise and Supply Risk Management. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 177-185): Springer. Jüttner, U., Peck, H., & Christopher, M. (2003). Supply chain risk management: outlining an agenda for future research. International Journal of Logistics: Research and Applications, 6(4), 197- 210. Jüttner, U., & Ziegenbein, A. (2009). Supply chain risk management for small and medium-sized businesses. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 199-217): Springer. Khan, O., Christopher, M., & Burnes, B. (2009). The role of product design in global supply chain risk management. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 137-153): Springer. Kleindorfer, P. R., & Saad, G. H. (2005). Managing disruption risks in supply chains. Production and Operations Management, 14(1), 53-68. Knemeyer, A. M., Zinn, W., & Eroglu, C. (2009). Proactive planning for catastrophic events in supply chains. Journal of Operations Management, 27(2), 141-153. Levi, D. S., Vassiladis, C., & Kyratzoglou, I. (2013). Supply chain and risk management; making the right risk decisions to strengthen operations performance: Pwc. Manuj, I., & Mentzer, J. T. (2008). Global supply chain risk management strategies. International Journal of Physical Distribution & Logistics Management, 38(3), 192-223. Mau, N., & Mau, M. (2009). Securing Global Food Distribution Networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 331-344): Springer. Melnyk, S. A., Rodrigues, A., & Ragatz, G. L. (2009). Using simulation to investigate supply chain disruptions. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 103-122): Springer. Mullai, A. (2009). Risk management system–a conceptual model. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 83-101): Springer. Narasimhan, R., & Talluri, S. (2009). Perspectives on risk management in supply chains. Journal of Operations Management, 27(2), 114-118.
78 Norrman, A., & Jansson, U. (2004). Ericsson's proactive supply chain risk management approach after a serious sub-supplier accident. International Journal of Physical Distribution & Logistics Management, 34(5), 434-456. Peck, H. (2006). Reconciling supply chain vulnerability, risk and supply chain management. International Journal of Logistics: Research and Applications, 9(2), 127-142. Perry, M. (2007). Natural disaster management planning: a study of logistics managers responding to the tsunami. International Journal of Physical Distribution & Logistics Management, 37(5), 409-433. Ritchie, B., & Brindley, C. (2009). SCRM and performance–issues and challenges. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 249-270): Springer. Seiter, M. (2009). Behavioural Risks in Supply Networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 235-247): Springer. Sheffi, Y. (2005). Building a Resilient Supply Chain. Harvard business Review, 1(8), 1-4. Sheffi, Y. (2008). Resilience: What it is and how to achieve it: The House Committee on Homeland Security “The Resilient Homeland–Broadening the Homeland Security Strategy. Sheffi, Y., & Rice Jr, J. B. (2005). A supply chain view of the resilient enterprise. MIT Sloan Management Review, 47(1). Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing the determinants of risk behavior. Academy of management review, 17(1), 9-38. Smith, P. G. (2002). Managing risk proactively in product development projects. Paper presented at the IPLNET workshop, Switzerland. Smith, P. G., & Merritt, G. M. (2002). Proactive Risk Management. Executive Book Summaries, 24(11), 1-8. Sodhi, M. S., Son, B. G., & Tang, C. S. (2012). Researchers' perspectives on supply chain risk management. Production and Operations Management, 21(1), 1-13. Tang, C., & Tomlin, B. (2009). How Much Flexibility Does It Take to Mitigate Supply Chain Risks? In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 155-174): Springer. Tang, C. S. (2006). Perspectives in supply chain risk management. International Journal of Production Economics, 103(2), 451-488. Tomlin, B. (2009). Disruption‐management strategies for short life‐cycle products. Naval Research Logistics (NRL), 56(4), 318-347. Voss, D., & Whipple, J. (2009). Food Supply Chain Security: Issues and Implications. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 293-305): Springer. Wagner, S. M., & Bode, C. (2006). An empirical investigation into supply chain vulnerability. Journal of purchasing and supply management, 12(6), 301-312.
79 Wagner, S. M., & Bode, C. (2008). An empirical examination of supply chain performance along several dimensions of risk. Journal of business logistics, 29(1), 307-325. Wagner, S. M., & Bode, C. (2009). Dominant risks and risk management practices in supply chains. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 271-290): Springer. Wieland, A., & Wallenburg, C. M. (2012). Dealing with supply chain risks: Linking risk management practices and strategies to performance. International Journal of Physical Distribution & Logistics Management, 42(10), 887-905. Zhao, L., Huo, B., Sun, L., & Zhao, X. (2013). The impact of supply chain risk on supply chain integration and company performance: a global investigation. Supply Chain Management: An International Journal, 18(2), 115-131. Zsidisin, G. A. (2003). A grounded definition of supply risk. Journal of purchasing and supply management, 9(5), 217-224. Zsidisin, G. A., & Ellram, L. M. (2003). An Agency Theory Investigation of Supply Risk M anagement. Journal of Supply Chain Management, 39(3), 15-27.
80 9 Appendices Appendix 1: PWC results Percentage of companies with more than 3 incidents, that suffered an impact of 3% or higher on their performance as a result of supply chain disruptions. Source: Levi et al. (2013).
81 Appendix 2: Participated Belgian companies company name function IC 1 VCST industrial products Koen Verhaeren Operations Manager 29 2 Dewulf NV Thomas Decan Production Manager 28 3 GEA process engineering NV Rudy Van de Perre Operations Director 28 4 Henrad Luc Geysen Supply Chain Director 25 5 Alliance International Pascal Demarets Vice President Operations 28 6 Esco Couplings Stef Huybrechts Operations Director 28 7 Anglo Belgian Corporation Johan Van de Velde Production Manager 28 8 SKT Bart Bauters Director 28 9 Asco industries Belgium Guido van den Berghe Plant Manager 30 10 Roxell Filip Cauwels Plant Manager 28 11 MCsyncro Donald De Buck Production Manager 29 12 Recticel NV - BL Flexible Foams Dirk Van De Velde Manufacturing Manager 29 13 International automotive components group Koen Hendrickx Purchase Manager & Logistics 29 14 Röchling Automotive Gijzegem NV Patrick Temmerman Plant Manager 29 15 Airkan Steven Coeman Production Manager 25 16 Tenneco Jelle Leenknegt Operations Manager 29 17 Team industries Walter Dumarey CEO 25 18 Alex Profiles BVBA Dave Debosschere Managing Director 25 19 PolyVision NV Theo Vanheusden Director of Manufacturing & Site Manager 25 20 Benteler automotive Luigi Neirynck Production Manager 29 21 Hansen Industrial Transmission Bruno Everaert Operations Manager 28 22 Johnson Controls Filip Bral Quality Manager 29 23 Joris Ide Jan Wauters Operations Manager 25 24 Delta Light Tom Compernolle Head of procurement (CPO) 27 25 Rogers Corporation Geert Verreecke Manufacturing Team Leader 26 26 Dana - Spicer Off-Highway Wim Thuy Manufacturing Manager 28 27 TE Connectivity Cammaert Denis Operations Manager 26 28 Niko NV Goedele Heylen Director Operations 27 29 Engineering Pieter Verhaert Engineering Manager 25 30 DAF Trucks Vlaanderen NV Marc Beets Managing Director 29
82 Appendix 3: Internal consistency scale measurement Environment Risk probability
83 Risk Impact Risk Management
84 Supply Chain Management Practices Supply Chain Design and Complexity
85 Appendix 4: Descriptive statistics Environment Descriptive Statistics N Mean Std. Deviation Variance SubstitutionThreat 828 2,94 1,103 1,216 MarketEntry 838 3,00 1,092 1,192 BargainingPowerS 830 3,11 ,903 ,816 SocPressure 832 3,24 1,062 1,129 EnvirPressure 835 3,33 1,070 1,144 TechnologicalChange 838 3,33 1,003 1,007 MarketSize 840 3,36 ,881 ,776 MarketSpan 833 3,46 1,021 1,042 MarketConcentration 836 3,49 1,077 1,160 BargainingPowerC 832 3,71 ,944 ,891 CompetitiveRivalry 835 3,86 ,918 ,843 Valid N (listwise) 794 Environment constructs Descriptive Statistics N Mean Std. Deviation PEST 828 3,3023 ,76588 MARKET 828 3,4352 ,68962 PORTER 811 3,3258 ,55169 Valid N (listwise) 794 Risk probability Descriptive Statistics N Mean Std. Deviation Variance Prob_Shipment 801 2,32 1,212 1,469 Prob_Operations 804 2,52 1,181 1,396 Prob_Supplier 806 2,74 1,168 1,364 Valid N (listwise) 801
86 Risk Impact Descriptive Statistics N Mean Std. Deviation Variance Imp_Shipment 801 3,37 1,320 1,742 Imp_Operations 804 3,52 1,241 1,540 Imp_Supplier 807 3,66 1,183 1,399 Valid N (listwise) 801 Proactive and Reactive Risk Management (last 3 year) Descriptive Statistics N Mean Std. Deviation Variance RecoveringL3Y 794 3,15 1,077 1,159 DetectingL3Y 796 3,27 ,993 ,986 RespondingL3Y 795 3,28 ,971 ,943 PreventingL3Y 798 3,42 ,958 ,919 Valid N (listwise) 791 Proactive and Reactive Risk Management (now) Descriptive Statistics N Mean Std. Deviation Variance Recovering 794 3,35 1,080 1,166 Detecting 795 3,46 ,957 ,916 Responding 797 3,47 ,966 ,933 Preventing 798 3,52 ,935 ,875 Valid N (listwise) 790 Supply chain Management practices (last 3 year) Descriptive Statistics N Mean Std. Deviation Variance SystemCouplingC 768 2,77 1,238 1,533 SystemCouplingS 777 2,80 1,157 1,338 InterDistrC 762 2,83 1,289 1,663 InterSourcingS 776 2,89 1,226 1,503 CollabApprC 772 3,02 1,124 1,263 SharingInfoC 773 3,05 1,111 1,235 DecisionMakingS 778 3,05 1,022 1,045 DecisionMakingC 772 3,07 1,118 1,249 SharingInfoS 782 3,20 ,991 ,981 CollabApprS 780 3,21 1,030 1,062 Valid N (listwise) 747
87 Supply chain Management practices (now) Descriptive Statistics N Mean Std. Deviation Variance SystemCoupling_C 763 2,93 1,278 1,634 InterSourcing_S 772 2,95 1,201 1,443 SystemCoupling_S 774 2,95 1,190 1,416 InterDistr_C 759 2,96 1,314 1,727 CollabAppr_C 770 3,15 1,146 1,313 DecisionMaking_S 777 3,18 1,040 1,081 DecisionMaking_C 767 3,22 1,127 1,269 SharingInfo_C 772 3,23 1,119 1,253 CollabAppr_S 778 3,33 ,978 ,956 SharingInfo_S 782 3,37 ,974 ,950 Valid N (listwise) 736 Sustainability SCM practices (3 years ago) Descriptive Statistics N Minimum Maximum Mean Std. Deviation SusPerformanceAssesmentS 794 1 5 3,00 1,151 SusTrainingS 795 1 5 2,46 1,247 SusJointEffortS 775 1 5 2,66 1,189 Valid N (listwise) 772 Sustainability SCM practices (now) Descriptive Statistics N Minimum Maximum Mean Std. Deviation SusPerformanceAssesment_S 792 1 5 3,15 1,130 SusTraining_S 787 1 5 2,64 1,281 SusJointEffort_S 768 1 5 2,86 1,235 Valid N (listwise) 765
88 Supply Chain Design and Complexity (last 3 years) Descriptive Statistics N Mean Std. Deviation Variance OneVSMultiplePlants 545 2,82 1,441 2,076 StableVSchangingFlexibilityPlant 547 2,88 1,300 1,689 NetwPerfMgmtSystemL3Y 523 3,06 1,185 1,404 UseOfTechnologyL3Y 524 3,23 1,163 1,352 IntegrInfoSystems 546 3,64 1,291 1,665 Valid N (listwise) 503 Supply Chain Design and Complexity (now) Descriptive Statistics N Mean Std. Deviation Variance OneVSMultiplePlants 545 2,82 1,441 2,076 StableVSchangingFlexibilityPlant 547 2,88 1,300 1,689 NetwPerfMgmtSystem 525 3,22 1,220 1,488 UseOfTechnology 528 3,39 1,112 1,237 IntegrInfoSystems 546 3,64 1,291 1,665 Valid N (listwise) 502 Appendix 5: Different frameworks Framework 1:
89 Framework 2: Appendix 6: Confirmatory Factor analysis for alternative framework formulations market entry in PORTER and market size in PEST and included in market Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE MARKET MarketConcentration 0,7323 0,53626329 0,6552 0,4903 MarketEntry 0,7395 0,54686025 MarketSize 0,6516 0,42458256 MarketSpan 0,6733 0,45333289 PEST EnvirPressure 0,6906 0,47692836 0.7033 0,5191 SocPressure 0,6551 0,42915601 TechnologicalChange 0,785 0,616225 Market Size 0,7445 0,55428025 PORTER BargainingPowerCust 0,6169 0,38056561 0,6961 0,4382 BargainingPowerSuppl 0,6045 0,36542025 CompetitiveRivalry 0,6038 0,36457444 Market Entry 0,7821 0,61168041 ThreatSubstitution 0,6847 0,46881409 RiskImpact Imp_Operations 0,9783 0,95707089 0,9593 0,9609 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9489 0,9513 Prob_Shipment 0,964 0,929296
90 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9809 0,96216481 0,9632 0,9316 Detecting 0,9796 0,95961616 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9659 0,9362 Responding 0,9764 0,95335696 MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability MARKET 0,7002 PEST 0,7135 0,7205 PORTER 0,7331 0,5641 0,6620 ProactiveRiskMgmt 0,2867 0,2439 0,2072 0,9803 ReactiveRiskMgmt 0,2926 0,2515 0,2033 0,9588 0,9753 RiskImpact 0,2707 0,2044 0,2215 0,7277 0,7195 0,9652 RiskProbability 0,2908 0,2038 0,232 0,6439 0,6376 0,8235 0,9676 All environmental factors together Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE Environment MarketConcentration 0,671 0,450241 0,8293 0,3648 MarketEntry 0,6829 0,46635241 MarketSize 0,5758 0,33154564 MarketSpan 0,6062 0,36747844 EnvirPressure 0,6607 0,43652449 SocPressure 0,5619 0,31573161 TechnologicalChange 0,6464 0,41783296 BargainingPowerCust 0,5475 0,29975625 BargainingPowerSuppl 0,5101 0,26020201 CompetitiveRivalry 0,5998 0,35976004 ThreatSubstitution 0,5542 0,30713764 RiskImpact Imp_Operations 0,9783 0,95707089 0,9593 0,9609 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9489 0,9513 Prob_Shipment 0,964 0,929296 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9809 0,96216481 0,9632 0,9316 Detecting 0,9796 0,95961616 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9659 0,9362 Responding 0,9764 0,95335696
91 Environment Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability Environment 0,603987 ProactiveRiskMgmt 0,2778 0,980255 ReactiveRiskMgmt 0,2825 0,9588 0,975346 RiskImpact 0,2623 0,7277 0,7195 0,965194 RiskProbability 0,2703 0,6439 0,6376 0,8235 0,967574 General Supply chain management practices Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE SCMpractices CollabAppr_C 0,9339 0,87217 0,9835 0,8711 CollabAppr_S 0,9513 0,90497 DecisionMaking_C 0,9208 0,84787 DecisionMaking_S 0,9456 0,89416 InfoSharing_C 0,9385 0,88078 InfoSharing_S 0,9435 0,89019 InternDistr_C 0,884 0,78146 InternSourcing_S 0,9324 0,86937 SystemCoupling_C 0,9356 0,87535 SystemCoupling_S 0,9461 0,89511 MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCM practices MARKET 0,7002 PEST 0,5927 0,7750 PORTER 0,561 0,5419 0,6866 ProactiveRiskMgmt 0,2867 0,222 0,1656 0,9803 ReactiveRiskMgmt 0,2927 0,2287 0,1638 0,9586 0,9753 RiskImpact 0,2707 0,1829 0,1814 0,7276 0,7194 0,9652 RiskProbability 0,2909 0,1714 0,1829 0,6438 0,6375 0,8235 0,9676 SCMpractices 0,2466 0,2218 0,1508 0,6633 0,6635 0,5856 0,5332 0,9333
92 Sustainable Supply chain management practices Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE sus. SCM practices SusJointEffort_S 0,9123 0,8322913 0,9247 0,8691 SusPerfAss_S 0,9325 0,8695563 SusTraining_S 0,9514 0,905162 Market PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCM sust. practices Market 0,70021 PEST 0,5927 0,77505 PORTER 0,561 0,5419 0,68651 ProactiveRiskMgmt 0,2867 0,222 0,1656 0,9802551 ReactiveRiskMgmt 0,2927 0,2287 0,1638 0,9586 0,975346 RiskImpact 0,2707 0,1829 0,1814 0,7276 0,7194 0,9652 RiskProbability 0,2909 0,1714 0,1829 0,6438 0,6375 0,8235 0,96757429 SCM sust. practices 0,293 0,2553 0,175 0,5716 0,5836 0,5319 0,4656 0,9322553 Supply chain design and complexity Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE Supply chain Design and complexity NetwPerfMgmtSystem 0,9683 0,9376049 0,9804 0,926 OneVSMultiplePlants 0,9523 0,9068753 IntegrInfoSystems 0,9598 0,921216 UseOfTechnology 0,9729 0,9465344 StableVSchangingFlexibilityPlant 0,9582 0,9181472 Market PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCD&C SCM (sus.) practices Market 0,7002 0 0 0 0 0 0 0 0 PEST 0,5927 0,77505 0 0 0 0 0 0 0 PORTER 0,561 0,5419 0,68651 0 0 0 0 0 0 Proactive 0,2867 0,222 0,1657 0,98025507 0 0 0 0 0 Reactive 0,2926 0,2286 0,1639 0,9588 0,9753461 0 0 0 0 Risk Impact 0,2707 0,1829 0,1814 0,7277 0,7195 0,9652 0 0 0 Risk Prob. 0,2909 0,1714 0,1829 0,6438 0,6376 0,8235 0,96757429 0 0 SCD&C 0,0914 0,1373 0,0461 0,1862 0,2135 0,225 0,207 0,962289 0 Sus. SCM 0,2842 0,2309 0,1626 0,483 0,4881 0,4936 0,5876 0,1696 0,9350936 Gen. SCM 0,2465 0,2218 0,1508 0,6632 0,6634 0,5855 0,5332 0,2669 0,933381
93 Appendix 6: T-statistics for mediators General SCM Sustainable SCM First framework Sample Mean (M) Standard Error (STERR) T Statistics Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,1243 0,0475 2,6506 0,1328 0,0513 2,5876 Market -> RiskProbability 0,1616 0,0485 3,3971 0,1521 0,0521 2,9194 Market -> SCMpractices 0,1448 0,0599 2,4769 0,238 0,0604 3,9435 PEST -> RiskImpact -0,0198 0,0376 0,5132 -0,0269 0,0381 0,7074 PEST -> RiskProbability -0,0242 0,0376 0,6612 -0,0731 0,0351 2,0825 PEST -> SCMpractices 0,161 0,0458 3,5349 0,105 0,0469 2,2379 PORTER -> RiskImpact 0,0357 0,0391 0,9027 0,0429 0,0516 0,8307 PORTER -> RiskProbability 0,0311 0,0373 0,7135 0,0422 0,0517 0,8154 PORTER -> SCMpractices -0,0049 0,0521 0,3483 -0,0263 0,0443 0,5935 RiskImpact -> ProactiveRiskMgmt 0,5982 0,0846 7,2521 0,6136 0,0825 7,4369 RiskImpact -> ReactiveRiskMgmt 0,5892 0,0848 7,1289 0,6042 0,0823 7,3406 RiskProbability -> ProactiveRiskMgmt 0,1499 0,0721 1,9216 0,1385 0,0725 1,9111 RiskProbability -> ReactiveRiskMgmt 0,1508 0,0736 1,9038 0,14 0,0725 1,9329 SCMpractices -> RiskImpact 0,5501 0,0539 10,306 0,4546 0,0629 7,2219 SCMpractices -> RiskProbability 0,4927 0,0539 9,2288 0,5538 0,0583 9,4955 General SCM Sustainable SCM Second framework Sample Mean (M) Standard Error (STERR) T Statistics Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,2357 0,0669 3,541 0,236 0,0645 3,6747 Market -> RiskProbability 0,2768 0,0697 4,0137 0,276 0,0685 4,0861 PEST -> RiskImpact 0,0211 0,0443 0,5186 0,0214 0,0458 0,502 PEST -> RiskProbability -0,0118 0,0438 0,2752 -0,0123 0,0446 0,2702 PORTER -> RiskImpact 0,0429 0,0487 0,7415 0,043 0,0506 0,7131 PORTER -> RiskProbability 0,0409 0,0478 0,6798 0,0406 0,0517 0,6292 RiskImpact -> ProactiveRiskMgmt 0,4468 0,0756 5,9763 0,5893 0,0894 6,809 RiskImpact -> ReactiveRiskMgmt 0,4328 0,0767 5,7265 0,5798 0,0909 6,593 RiskImpact -> SCMpractices 0,4375 0,084 5,4219 0,0345 0,0609 0,4941 RiskProbability -> ProactiveRiskMgmt 0,088 0,0568 1,4498 0,0659 0,0841 0,6369 RiskProbability -> ReactiveRiskMgmt 0,0891 0,0581 1,4234 0,0592 0,0862 0,5507 RiskProbability -> SCM practices 0,1704 0,0734 2,1555 0,5558 0,0668 8,432 SCM practices -> ProactiveRiskMgmt 0,3509 0,0599 5,9235 0,1541 0,0449 3,3612 SCM practices -> ReactiveRiskMgmt 0,3589 0,0587 6,1708 0,1683 0,0446 3,6884
94 Appendix 7: Sobel test Mediator between environment and risk perception Mediation Test generality direct no med direct w med IV-> Med beta Med -> DV beta IV ->med SE Med -> DV SE Sobel test Market -> RiskImpact 0,237 0,133 0,1848 0,5553 0,0599 0,0559 2,946320656 Market -> RiskProbability 0,28 0,187 0,1848 0,4971 0,0599 0,057 2,908512633 PEST -> RiskImpact 0,023 -0,074 0,1228 0,5553 0,0515 0,0559 2,318605859 PEST -> RiskProbability -0,012 -0,045 0,1228 0,4971 0,0515 0,057 2,300044582 PORTER -> RiskImpact 0,036 0,048 -0,0189 0,5553 0,0527 0,0559 -0,35840029 PORTER -> RiskProbability 0,033 0,044 -0,0189 0,4971 0,0527 0,057 -0,35833092 Mediation Test sustainability direct no med direct w med IV-> Med beta Med -> DV beta IV -> med SE Med -> DV SE Sobel test Market -> RiskImpact 0,237 0,133 0,238 0,4546 0,0588 0,0614 3,55154551 Market -> RiskProbability 0,28 0,152 0,238 0,5538 0,0588 0,0574 3,732464065 PEST -> RiskImpact 0,023 -0,027 0,105 0,4546 0,0463 0,0614 2,168380126 PEST -> RiskProbability -0,012 -0,073 0,105 0,5538 0,0463 0,0574 2,207651769 PORTER -> RiskImpact 0,036 0,043 -0,0263 0,4546 0,0434 0,0614 -0,60397116 PORTER -> RiskProbability 0,033 0,042 -0,0263 0,5538 0,0434 0,0574 -0,60479898
95 Mediator between risk perception and risk managements General SCM Mediation Test direct no med direct w med IV-> Med beta Med -> DV beta IV -> med SE Med -> DV SE Sobel test RiskImpact -> ProactiveRiskMgmt 0,138 0,452 0,4552 0,3547 0,084 0,0625 3,919273337 RiskImpact -> ReactiveRiskMgmt 0,14 0,439 0,4552 0,3623 0,084 0,0617 3,982348619 RiskProbability -> ProactiveRiskMgmt 0,614 0,082 0,1583 0,3547 0,0731 0,0625 2,023237295 RiskProbability -> ReactiveRiskMgmt 0,604 0,083 0,1583 0,3623 0,0731 0,0617 2,031763086 Sustainable SCM Mediation Test direct no med direct w med IV-> Med beta Med -> DV beta IV - >med SE Med -> DV SE Sobel test RiskProbability -> ProactiveRiskMgmt 0,138 0,054 0,5629 0,151 0,0611 0,0476 2,999433246 RiskProbability -> ReactiveRiskMgmt 0,14 0,047 0,5629 0,1645 0,0611 0,046 3,333742626 RiskImpact -> ProactiveRiskMgmt 0,614 0,609 0,0301 0,151 0,0572 0,0476 0,519129801 RiskImpact -> ReactiveRiskMgmt 0,604 0,599 0,0301 0,1645 0,0572 0,046 0,520617411
96 Appendix 8: Paired Sampled T-statistics Europe Versus Asia Environment Risk Probability Risk Impact
97 Risk Management General SCM practices Sustainable SCM practices Supply chain Complexity and Design
98 Low Versus High complex networks Environment Risk probability Risk impact Risk management
99 General SCM practices Sustainable SCM practices
100 Appendix 9: Multi-group moderation and moderated mediation First framework: SCM practices as mediator variables between environment and risk perception t-statistics first framework General SCM practices Sustainable SCM practices MULTIGROUP MODERATION Asia VS EU L VS H compl Asia VS EU L VS H compl Market -> RiskImpact 0,6359 0,7554 0,1017 0,6449 Market -> RiskProbability 1,1892 0,8877 0,5667 0,8953 Market -> SCMpractices 1,1295 0,2488 1,1138 1,0071 PEST -> RiskImpact 0,8564 0,3586 1,0135 0,0682 PEST -> RiskProbability 0,8752 1,3912 1,0766 1,6242 PEST -> SCMpractices 0,0793 0,9016 0,0823 0,3683 PORTER -> RiskImpact 0,9234 0,6434 0,5830 0,2117 PORTER -> RiskProbability 0,7361 0,7447 0,5702 0,3206 PORTER -> SCMpractices 0,6403 0,2952 1,2589 0,0545 RiskImpact -> ProactiveRiskMgmt 0,6523 0,2320 0,6822 0,1851 RiskImpact -> ReactiveRiskMgmt 0,8333 0,1418 0,8873 0,1318 RiskProbability -> ProactiveRiskMgmt 1,4035 1,5103 1,4342 1,4014 RiskProbability -> ReactiveRiskMgmt 1,4983 1,5621 1,5379 1,4461 SCMpractices -> RiskImpact 1,5009 0,2412 0,1811 1,7324 SCMpractices -> RiskProbability 0,9471 1,2217 0,3756 0,5853 MODERATED MEDIATION Market -> ProactiveRiskMgmt 0,0921 0,5414 0,1588 0,6779 Market -> ReactiveRiskMgmt 0,1324 0,5600 0,1973 0,6879 Market -> RiskImpact 0,4830 0,7625 0,5295 0,7815 Market -> RiskProbability 0,1686 1,2184 0,1003 0,9775 PEST -> ProactiveRiskMgmt 0,8700 0,6280 0,9759 0,6196 PEST -> ReactiveRiskMgmt 0,8901 0,5499 0,9987 0,5612 PEST -> RiskImpact 0,8096 0,2623 0,8997 0,3300 PEST -> RiskProbability 0,8173 1,3142 0,9025 1,5949 PORTER -> ProactiveRiskMgmt 1,0701 0,4724 1,2789 0,2779 PORTER -> ReactiveRiskMgmt 1,0701 0,4956 1,2794 0,2736 PORTER -> RiskImpact 1,0368 0,7161 1,2493 0,2701 PORTER -> RiskProbability 0,8804 0,8246 1,0696 0,3544 SCMpractices -> ProactiveRiskMgmt 0,6333 0,2016 0,1201 0,9118 SCMpractices -> ReactiveRiskMgmt 0,7378 0,2174 0,0295 0,9469
101 Second framework: SCM practices as mediator variables between risk perception and risk management t-statistics second framework General SCM practices Sustainable SCM practices MULTIGROUP MODERATION Asia VS EU L VS H compl Asia VS EU L VS H compl Market -> RiskImpact 0,4976 1,1442 0,5062 1,1739 Market -> RiskProbability 0,1458 1,3351 0,1494 1,2834 PEST -> RiskImpact 0,7539 0,1577 0,7588 0,1431 PEST -> RiskProbability 0,7227 1,2396 0,7377 1,2143 PORTER -> RiskImpact 0,9955 0,6033 1,0715 0,5718 PORTER -> RiskProbability 0,8138 0,7038 0,8736 0,6983 RiskImpact -> ProactiveRiskMgmt 0,0613 0,1446 0,0035 0,1807 RiskImpact -> ReactiveRiskMgmt 0,1062 0,0357 0,1017 0,1461 RiskImpact -> SCMpractices 1,2544 0,6952 0,9063 2,0087 RiskProbability -> ProactiveRiskMgmt 1,7020 1,8121 1,3502 1,8093 RiskProbability -> ReactiveRiskMgmt 1,8689 1,7024 1,4679 1,6676 RiskProbability -> SCMpractices 0,3864 1,7177 0,9838 1,3541 SCMpractices -> ProactiveRiskMgmt 0,2912 0,0693 1,2534 1,9513 SCMpractices -> ReactiveRiskMgmt 0,6251 0,6917 1,5046 1,7178 MODERATED MEDIATION Market -> ProactiveRiskMgmt 0,1111 0,8848 0,1324 0,8779 Market -> ReactiveRiskMgmt 0,1500 0,9020 0,1742 0,8940 Market -> SCMpractices 0,7655 1,1293 0,4025 0,0584 PEST -> ProactiveRiskMgmt 0,8080 0,5949 0,8049 0,6170 PEST -> ReactiveRiskMgmt 0,8298 0,5160 0,8266 0,5377 PEST -> SCMpractices 0,8885 0,5226 0,8417 0,4072 PORTER -> ProactiveRiskMgmt 1,0257 0,4569 1,0915 0,4078 PORTER -> ReactiveRiskMgmt 1,0282 0,4765 1,0931 0,4329 PORTER -> SCMpractices 0,9701 0,5471 1,1142 0,5982 RiskImpact -> ProactiveRiskMgmt 0,6679 0,3094 0,6310 0,1923 RiskImpact -> ReactiveRiskMgmt 0,8715 0,2051 0,8234 0,1000 RiskProbability -> ProactiveRiskMgmt 1,4946 1,5499 1,3899 1,6031 RiskProbability -> ReactiveRiskMgmt 1,6070 1,5853 1,4914 1,5595
102 Appendix 10: Statistics for differences with 3 years ago Variable paired samples t-test Proactive and Reactive management Whole world Asia Europe Proactive versus Reactive
103 General SCM practices Whole world Asia
104 Europe Sustainable SCM practices Whole world Asia
105 Europe Paired sample t-tests for relationships Relationship paired sample t-test Asia versus Europe Differences now and 3 years ago for the relationship between SCM and RM T-statistic first framework T-statistic second framework General Asia SCMpractices -> ProactiveRiskMgmt 0,329748177 0,848347035 SCMpractices -> ReactiveRiskMgmt 0,349440394 1,17079123 Europe SCMpractices -> ProactiveRiskMgmt 0,046033491 0,092671711 SCMpractices -> ReactiveRiskMgmt 0,024877732 0,081973729 Sustainable Asia SCM sus. practices -> ProactiveRiskMgmt 0,439419177 0,245626532 SCM sus. practices -> ReactiveRiskMgmt 0,430489809 0,479315314 Europe SCM sus. practices -> ProactiveRiskMgmt 0,132049205 0,162158132 SCM sus. practices -> ReactiveRiskMgmt 0,204659152 0,144855217 Relationship paired sample t-test Low versus High complex networks Differences now and 3 years ago for the relationship between SCM and RM first framework second framework General High SCMpractices -> ProactiveRiskMgmt 0,206600542 1,077970981 SCMpractices -> ReactiveRiskMgmt 0,301711699 0,954383032 Low SCMpractices -> ProactiveRiskMgmt 0,847346522 0,191774818 SCMpractices -> ReactiveRiskMgmt 0,841522227 0,054850686 Sustainable High SCM sus. practices -> ProactiveRiskMgmt 0,609047769 0,75769831 SCM sus. practices -> ReactiveRiskMgmt 0,566825692 0,980248788 Low SCM sus. practices -> ProactiveRiskMgmt 1,269361435 0,68087869 SCM sus. practices -> ReactiveRiskMgmt 1,315267878 0,782414819

More Related Content

PDF
Karin-DArcy05-MorganCanon
Rosalyn Karin-DArcy
 
PDF
Entrepreneurship flattens_the_world
Jason Shikomba
 
PDF
leinster senior club of the year
Dennis Bowes
 
PDF
Germany
Jarita01
 
DOC
Richard Filion english
Richard Filion
 
PDF
30 Mind-blowing Stats on Internet of Things: Market Estimates and Forecasts
Amit Rao
 
PDF
multimedia-y_educacion
alexitangel
 
PDF
d000163_Institutional_research_Cronje_Jun2008
Marthie Cronje
 
Karin-DArcy05-MorganCanon
Rosalyn Karin-DArcy
 
Entrepreneurship flattens_the_world
Jason Shikomba
 
leinster senior club of the year
Dennis Bowes
 
Germany
Jarita01
 
Richard Filion english
Richard Filion
 
30 Mind-blowing Stats on Internet of Things: Market Estimates and Forecasts
Amit Rao
 
multimedia-y_educacion
alexitangel
 
d000163_Institutional_research_Cronje_Jun2008
Marthie Cronje
 

Viewers also liked (7)

PDF
honors_paper
Jason Kholodnov
 
PPTX
Norton tech support number
rahulkundu08
 
PDF
Analysis of Cryptography Techniques
editor1knowledgecuddle
 
PDF
Z's Podcasts Tips
Zat Granados
 
PDF
Social Media
charlesshamar
 
PPTX
History of cricket
Khush Ramani
 
PDF
Comparative Analysis of Cryptographic Algorithms and Advanced Cryptographic A...
editor1knowledgecuddle
 
honors_paper
Jason Kholodnov
 
Norton tech support number
rahulkundu08
 
Analysis of Cryptography Techniques
editor1knowledgecuddle
 
Z's Podcasts Tips
Zat Granados
 
Social Media
charlesshamar
 
History of cricket
Khush Ramani
 
Comparative Analysis of Cryptographic Algorithms and Advanced Cryptographic A...
editor1knowledgecuddle
 
Ad

Similar to The use of pro-active versus re-active risk management practices for managing supply chains (20)

DOCX
Tti basic electricity level 1
liniprasad
 
PDF
Vanderstraeten 2013 ph d - studies on the strategy and performance of busi...
FGVcenn
 
PDF
Global Energy Savings from High Efficiency Distribution Transformers
Leonardo ENERGY
 
DOCX
TPM and the effect of health and safety
Martin Munsie
 
PDF
O7-L3 Supply Chain Operations - ICLT Program
labyankof
 
PDF
Productivity in the Container Port Business - Focus on the Mediterranean Range
Feliciana Monteiro
 
PDF
The application of VSM to NPD
Edoardo Bruno
 
PDF
Original
diazdavidh
 
PDF
Fingerprint Biometrics vulnerabilities
Farhan Liaqat
 
PDF
Blockchain in Education. Alexander Grech & Anthony F. Camilleri. Editor Andre...
eraser Juan José Calderón
 
PDF
FRM - U.P.S. Sirisena
Saddha thissa
 
PDF
Rad Tech Handbook 2012
Moultrie Technical College
 
DOCX
Cict6640 sandis wanjala wamalwa
SandisWanjala
 
DOCX
Thesis
kevin roche
 
PDF
Mémoire M1
Olivier Pasquier
 
PDF
IMP³rove – A European Project with Impact – 50 Success Stories on Innovation ...
IMP³rove Academy
 
PDF
DT003a Final Report - The Design and Build of a Non-contact Extensometer for ...
Morven Gannon
 
PDF
Faculty of Engineering, Built Environment and Information Technology:/Ebit pa...
University of Pretoria
 
PDF
Getting started-in-endoscopic-ear-surgery
Prasanna Datta
 
PDF
Project on a Business Plan for the Solar system installation on Residential b...
Rifat Abdul Rahiman
 
Tti basic electricity level 1
liniprasad
 
Vanderstraeten 2013 ph d - studies on the strategy and performance of busi...
FGVcenn
 
Global Energy Savings from High Efficiency Distribution Transformers
Leonardo ENERGY
 
TPM and the effect of health and safety
Martin Munsie
 
O7-L3 Supply Chain Operations - ICLT Program
labyankof
 
Productivity in the Container Port Business - Focus on the Mediterranean Range
Feliciana Monteiro
 
The application of VSM to NPD
Edoardo Bruno
 
Original
diazdavidh
 
Fingerprint Biometrics vulnerabilities
Farhan Liaqat
 
Blockchain in Education. Alexander Grech & Anthony F. Camilleri. Editor Andre...
eraser Juan José Calderón
 
FRM - U.P.S. Sirisena
Saddha thissa
 
Rad Tech Handbook 2012
Moultrie Technical College
 
Cict6640 sandis wanjala wamalwa
SandisWanjala
 
Thesis
kevin roche
 
Mémoire M1
Olivier Pasquier
 
IMP³rove – A European Project with Impact – 50 Success Stories on Innovation ...
IMP³rove Academy
 
DT003a Final Report - The Design and Build of a Non-contact Extensometer for ...
Morven Gannon
 
Faculty of Engineering, Built Environment and Information Technology:/Ebit pa...
University of Pretoria
 
Getting started-in-endoscopic-ear-surgery
Prasanna Datta
 
Project on a Business Plan for the Solar system installation on Residential b...
Rifat Abdul Rahiman
 
Ad

The use of pro-active versus re-active risk management practices for managing supply chains

  • 1. UNIVERSITEIT GENT FACULTEIT ECONOMIE EN BEDRIJFSKUNDE ACADEMIEJAAR 2013 – 2014 The use of pro-active versus re-active risk management practices for managing supply chains Masterproef voorgedragen tot het bekomen van de graad van Master of Science in de Toegepaste Economische Wetenschappen: Handelsingenieur Pieterjan Tilleman onder leiding van Prof. Ann Vereecke Begeleider: Evelyne Vanpoucke
  • 3. I UNIVERSITEIT GENT FACULTEIT ECONOMIE EN BEDRIJFSKUNDE ACADEMIEJAAR 2013 – 2014 The use of pro-active versus re-active risk management practices for managing supply chains Masterproef voorgedragen tot het bekomen van de graad van Master of Science in de Toegepaste Economische Wetenschappen: Handelsingenieur Pieterjan Tilleman onder leiding van Prof. Ann Vereecke Begeleider: Evelyne Vanpoucke
  • 4. II CLAUSE OF CONFIDENTIALITY PERMISSION Ondergetekende verklaart dat de inhoud van deze masterproef mag geraadpleegd en/of gereproduceerd worden, mits bronvermelding. Pieterjan Tilleman
  • 5. III Acknowledgements Vooreerst wil ik Evelyne Vanpoucke van harte bedanken voor de volle steun, de vele raad, de nodige tijd en de snelle beantwoording van vele emails die mij op het goede pad hebben geleid. Vervolgens mag ik een dank aan mijn vriendin Louise uiten voor de hulp met de bibliografie en de steun en toeverlaat wanneer het eens wat minder vlotte. Ook mag ik de medewerking van de vele operations en supply chain managers van de deelgenomen Belgische bedrijven aan dit onderzoek niet vergeten. Al was het soms een niet voor de hand liggende opdracht om een vragenlijst van een dergelijke omvang ingevuld te krijgen, doch zonder deze data zou dit werk niet tot stand gekomen zijn. Tenslotte, dank ik mijn familie voor de volharding en de enorme steun en in het bijzonder mijn vader voor het nalezen van dit werk.
  • 6. IV Table of Contents Table of Contents ................................................................................................................................. IV List of Tables....................................................................................................................................... VIII List of Formulas.................................................................................................................................. VIII List of Figures .........................................................................................................................................1 1 Introduction ...................................................................................................................................1 2 Literature research.........................................................................................................................3 2.1 Risk Management...................................................................................................................3 2.1.1 The Nature of Supply Chain Risk.....................................................................................3 2.1.2 Risk management .........................................................................................................11 2.2 Supply Chain Management...................................................................................................22 2.3 Supply Chain Risk Management ...........................................................................................27 2.4 Supply Chain Security Management.....................................................................................30 2.5 Conclusion............................................................................................................................32 3 Conceptual framework.................................................................................................................33 3.1 Environment.........................................................................................................................33 3.2 Risk perception or representation of risk .............................................................................35 3.3 Proactive versus Reactive Risk Strategies.............................................................................36 3.4 Moderators and Mediators ..................................................................................................39 3.4.1 Supply chain management practices. ...........................................................................39 3.4.2 Design and complexity of the supply chain...................................................................41 3.4.3 Continental differences ................................................................................................46 3.4.4 Comparison with 3 years ago........................................................................................46 3.5 Model diagram .....................................................................................................................47 4 Data Collection .............................................................................................................................48 4.1 Introduction..........................................................................................................................48 4.2 Results..................................................................................................................................49 4.3 Variables and constructs ......................................................................................................50 4.4 Exploratory factor analysis (EFA)..........................................................................................52 5 Methodology and Analysis ...........................................................................................................53 5.1 General model......................................................................................................................53 5.1.1 Confirmatory Factor Analysis (CFA) of the general model ............................................53 5.1.2 Comparison with alternative frameworks ....................................................................56 5.1.3 Descriptive statistics.....................................................................................................57
  • 7. V 5.1.4 Hypothesis testing ........................................................................................................58 5.1.5 Multicollinearity ...........................................................................................................59 5.2 Input from mediator variables..............................................................................................60 5.2.1 Framework ...................................................................................................................60 5.2.2 Confirmatory factor analysis.........................................................................................61 5.2.3 Descriptive statistics with mediator variables ..............................................................61 5.2.4 Structural path significance ..........................................................................................62 5.2.5 Mediating effect and hypothesis testing ......................................................................62 5.3 Input from moderator variables ...........................................................................................64 5.3.1 Descriptive statistics and differences for the moderator variables ..............................64 5.3.2 Moderating effect and moderated mediation..............................................................67 5.4 Some comparison with 3 years ago......................................................................................71 5.5 Hypothesis summary............................................................................................................72 6 Overall Conclusion........................................................................................................................73 7 Limitations and possibilities for future research ..........................................................................75 8 References....................................................................................................................................76 9 Appendices...................................................................................................................................80 Appendix 1: PWC results ..................................................................................................................80 Appendix 2: Participated Belgian companies ...................................................................................81 Appendix 3: Internal consistency scale measurement .....................................................................82 Appendix 4: Descriptive statistics.....................................................................................................85 Appendix 5: Different frameworks ...................................................................................................88 Appendix 6: Confirmatory Factor analysis for alternative framework formulations ........................89 Appendix 6: T-statistics for mediators..............................................................................................93 Appendix 7: Sobel test......................................................................................................................94 Appendix 8: Paired Sampled T-statistics...........................................................................................96 Appendix 9: Multi-group moderation and moderated mediation..................................................100 Appendix 10: Statistics for differences with 3 years ago ................................................................102 Variable paired samples t-test....................................................................................................102 Paired sample t-tests for relationships.......................................................................................105
  • 8. VI Abbreviations list DV: Dependent Variable ERM: Enterprise Risk Management FMEA: Failure Mode and Effect Analysis IV: Independent Variable PEST- variable: variable that signifies Political, Economical, Social & Technological trends in the environment PORTER-variable: variable that takes a look at the company’s competitive forces RM: Risk Management SC: Supply Chain SCD&C: Supply Chain Design and Complexity SCM: Supply Chain Management SCRM: Supply Chain Risk Management SCSM: Supply Chain Security Management SCVM: Supply Chain Vulnerability Map SME: Small and Medium Enterprise TQM: Total Quality Management
  • 9. VII List of Figures Figure 1: risk aspects .............................................................................................................................5 Figure 2: Dimensions of risk ...................................................................................................................6 Figure 3: risk sources..............................................................................................................................7 Figure 4: vulnerability - efficiency relation (left), vulnerability - simplicity (right)..................................9 Figure 5: risk management framework ................................................................................................11 Figure 6: fault tree analysis for the AVIA example ...............................................................................13 Figure 7: risk matrix for the additional probability of detecting risks...................................................14 Figure 8: risk mitigations strategies......................................................................................................19 Figure 9: the concept of Supply Chain Risk Management (Blos et al., 2009)........................................27 Figure 10: a SCRM framework..............................................................................................................29 Figure 11: PEST anlysis .........................................................................................................................34 Figure 12: Porter's Five Forces Model ..................................................................................................35 Figure 13: Supply Chain........................................................................................................................41 Figure 14: Supply Network ...................................................................................................................42 Figure 15: a more complex and adaptive network...............................................................................42 Figure 16: supply information network................................................................................................44 Figure 17: Model diagram ....................................................................................................................47 Figure 18: proportion of participated industries ..................................................................................49 Figure 19: proportion of participated countries...................................................................................49 Figure 20: General Model.....................................................................................................................53 Figure 21: framework with general SCM mediator between environment and risk perception (first framework) ..........................................................................................................................................60 Figure 22: Framework with general SCM mediator between risk perception and risk management (second framework) .............................................................................................................................61 Figure 23: Mediating effect ..................................................................................................................63 Figure 24: Environmental factors between Europe and Asia ...............................................................64 Figure 25: Risk Probability and Impact for Europe and Asia.................................................................65 Figure 26: calculation for the multi-group moderation t-statistic and p-value.....................................68
  • 10. VIII List of Tables Table 1: risk criticality matrix ...............................................................................................................14 Table 2: risk management actions........................................................................................................17 Table 3: the business from a biological view........................................................................................43 Table 4: Environmental constructs.......................................................................................................50 Table 5: Risk perception variables........................................................................................................50 Table 6: Proactive and Reactive Management .....................................................................................50 Table 7: Supply Chain Management Practices......................................................................................51 Table 8: Supply Chain design and complexity.......................................................................................52 Table 9: CFA summary table.................................................................................................................54 Table 10: Discriminant validiy ..............................................................................................................55 Table 11: CFA summary table for comparison with other frameworks................................................57 Table 12: outer model T-statistics........................................................................................................58 Table 13: inner model loadings and T-statistics ...................................................................................58 Table 14: Multicollinearity ...................................................................................................................59 Table 15: Risk management for Europe and Asia.................................................................................65 Table 16: Risk management for lower and higher complex networks..................................................66 Table 17: determination coefficients ...................................................................................................67 Table 18: Mean, standard errors and T-statistics for the two groups for moderation .........................70 Table 19: hypothesis summary table....................................................................................................72 List of Formulas Formula 1: Sobel test statistic ..............................................................................................................62 Formula 2: t-test for multi-group moderation and moderated mediation...........................................68
  • 11. 1 1 Introduction In many business environments, networking in supply chains is almost an inevitable solution to help companies respond fast to market changes. A lot of opportunities are accompanied with networking. Examples are lower transaction costs, ability to concentrate on core skills, lower capital investments, sharing sunk costs, greater flexibility and access to key technologies. So the use, the meaning and the practices of the concept supply chain management became important. However, increased network cooperation does increase the dependency between organizations and as a consequence of the advantages above companies become more exposed to the risks of other companies. Hence networking causes transfer of risks between several companies from a supplier- customer viewpoint. It may decrease some risks but unfortunately increase others. Inevitably partners must share their risk among them as a solution to mitigate their risks and to succeed in their operations. Therefore today’s industries must operate under extreme caution and the concept of supply chain risk management was born. The need of the concept became useful after series of crises and catastrophes had attracted public attention like natural disasters, political and economic instabilities, terrorist attacks and many more. Secondly modern supply chains seem to be more vulnerable than ever: increased competitive pressure in the business environment and globalization of markets. Counterfeiting products has increasingly entered the supply chain and harms a company’s product and reputation. The financial crisis has brought companies to be very suspicious and seek to ensure their business and operations continuity. Nowadays they struggle more than ever from facts like supplier insolvency and less access to credit that especially impact the less financially stable companies. There are plenty number of relevant examples. Automobile manufacturer Land Rover found itself in serious trouble after its only supplier of chassis frames, UPF-Thomson, suddenly and unexpectedly folded further supply delivery (Sheffi & Rice Jr, 2005; Wagner & Bode, 2009). Electronic company Ericsson faced dramatic problems with a huge impact, after a fire at a sub-supplier and has implemented an entire new organization with new supplier risk management tools (Norrman & Jansson, 2004). Ford, Toyota and DaimlerChrysler experienced massive disruptions to the flow of materials into their North-American assembly plants within a few days after the terrorist attack of 9/11 due to border shut-downs (Sheffi & Rice Jr, 2005). Globalization compelled firms to make their supply chains more efficient, resilient or more responsive by outsourcing or off shoring activities, sourcing in low-cost countries, collaborations with
  • 12. 2 other partnerships, decreasing inventory and so on. But all these activities can be associated with a higher level of risk and supply chain sensitivity. A report stated that companies suffering from supply chain disruptions experienced 33-40% lower stock returns relative to their industry benchmarks. Consequently it can also negatively impact a firm’s brand image and reputation. In addition severe disruptions like the Fukushima nuclear disaster have healthy and safety risk consequences. So concluded we feel the need for risk management in a relatively unstable world on the one hand and an increasingly flexible supply chain on the other hand. This work gives an introduction into the risk management world. We will discuss the main concepts of risk, the perceptions, the difference between proactive and reactive mitigation strategies, the supply chain practices that are needed to stay resilient and many more on the basis of existing literature. Secondly we will employ our gained knowledge to build and test a framework that make use of several aspects of the environment, risk concepts and supply chain practices and complexity. Data was collected in many industries to achieve this goal. Additionally we will try to do some investigation for progressions in the past three years and compare continents, in particular European and Asian countries. In the end we will come up with some meaningful conclusions that fit with our model and outcomes, give some limitations and suggestions for future research.
  • 13. 3 2 Literature research First we will dig in to the world of risk management with all his facets. Thereafter we will consider supply chain management concept and finally we will end up with the meaning of supply chain risk management and some thoughts of supply chain security management. 2.1 Risk Management 2.1.1 The Nature of Supply Chain Risk Introductory case Consider a random company; named AVIA. AVIA is a manufacturing company that produces metallic aircraft components for the aviation industry. It is operating its activities since a long time. Through these years it has maintained its supplier base, the metal industry and reached its few customers from the aircraft assembly industry. Suppose now that because of tensions in the commodity market, the company that is responsible for the supply on metal parts has defaulted to deliver the needed products. Company AVIA can appeal on a few minor companies but this amount is not sufficient. You as a company decide to produce further your semi-finished goods with the little supply of metal parts. Suppose then a major customer refuse to do further business with you because you have augmented your prices due to your increased variable costs per part produced. Or you are unable anymore to deliver the requested components on time because time goes by until you receive your parts. On top of that a fire caught your plant and a third of your machinery capacity has been demolished. Reinvesting in new assembly equipment is accompanied with a lot of costs. As a consequence you see your benefits declining and you end the year with a very negative profit and loss account for the AVIA Company. Risk Functions which generate the possibility of beneficial effects or profit often include risks. This is certainly the case with business activities. So risk can here be for example: 1) Your supplier fails in delivering the needed metal parts such that you operate under capacity. 2) Problems in fulfilling customer deliveries arise because you cannot deliver your aircraft components on time. 3) Because of cost considerations you increase your prices. Customers quit business with AVIA which results in a too low or inappropriate demand. 4) Due to a fire at the production plant, it is difficult to get back on track which resides in the difficult re-management of its costs, resources, development and flexibility.
  • 14. 4 The above situations are all examples that contain risk. We can already make a first distinction between demand side risks and supply side risks. The third risk is an example of the first category and the first risk is an example of the latter. These are risk categories that are internal to the supply chain whereas the fourth risk is an example of an external to the supply chain risk. But what is risk actually? Risk is a characteristic of decisions that is defined as the extent to which there is uncertainty about whether potentially significant and/or disappointing outcomes of decisions will be realized (Sitkin & Pablo, 1992). So risks suggest variation in the distribution of possible outcomes, their likelihoods and their subjective values (Wagner & Bode, 2008). According to Kahneman & Tversky’s “prospect theory” individual risk behavior is determined how the situation is framed. For example if individuals are protecting prior gains will be more risk averse. In financial risk management, risk is considered as having an upside and downside potential of possible outcomes according to a normal distribution with a two-sided variance. In contrast the aim of this work is to approach risks in supply chains which can be better stated, considering the severe impact of disruptions, as being purely negative. According to several authors, risk is considered in this manner and that corresponds best from a supply chain consideration (Wagner & Bode, 2009). Zsidisin (2003) contains a broad definition of risk applied to the supply chain: the probability of an incident associated with inbound supply from individual supplier failures (quality, delivery, relationships and price) or the supply market occurring, in which its outcomes result in the inability of the purchasing firm to meet customer demand or cause threats to customer life and safety. What is good in this definition is that it mentions the distinction between supply and demand side risks. It mentions some supply risks but these are not exhaustive. Production capacity constraints on the supply market, technological changes with the supplier, product design features, to mention a few can also play a role. It also assesses the risk aspects which are important to understand risk. Risk aspects In the former definition we find 2 important aspects that constitute risk namely the extent or the impact of outcomes and the possibility or potentially significance that may or may not be disappointing of these outcomes. These are convenient aspects of risk because according to the Bayesian theory when you multiply these two figures for each outcome you get the distribution and thus the severity of each possible outcome. Therefore we can split risk in 4 categories shown in figure 1 below. We consider the enterprise’s vulnerability the highest when both the likelihood and the impact of disruption are high whereas rare, low impact events require less action to mitigate. Furthermore disruptions that combine high probability and low impact are part of the daily operations in the normal flow of business. On the other hand risks with low likelihoods but high
  • 15. 5 consequences need a concrete planning and interference that is outside the daily business operations. The fire at our plant is an example of a low probability, high impact risk. A reduced demand of your aircraft assemblies could be an example of high probability low impact risk. It is important that you recognize this risk in function of your company because the probability and impact differs among different corporations. A strike in one plant of the Airbus corporations has a lower impact on a big multinational with several aviation plants than the same strike in our little AVIA plant. Figure 1: risk aspects Kleindorfer and Saad (2005) suggest besides impact and probability, the speed of the possible risk can also play a role. Speed can be understood as the rate at which the event leading to loss happens, the rate at which these losses happen and how quickly the risk event is discovered by the company. Furthermore the frequency or how often a similar kind of risk event happens. In our situation how often will our major supplier fail to deliver our metal parts? If this occurs too frequently our company may lose his reputation, several customers will abandon AVIA and in the long run our company may be even going out of business. At last, Griffis and Whipple (2012)consider the probability of risk detection as an additional aspect of risk and adds a third dimension besides impact and probability. For instance, a high-likelihood/high- impact risk that is also extremely difficult to detect, warrants a substantially different risk management strategy than a high-likelihood/high-impact risk that can be more readily detected (see further). Low impact High probability High impact High probability Low impact Low probability High impact Low probability
  • 16. 6 Figure 2: Dimensions of risk Risk Drivers and Risk Natures At some point you as a manager believe something exists in your business operations environment and will lead to a particular risk event and a serious impact could occur (P. G. Smith & Merritt, 2002). This is what we call risk drivers. Risk drivers can further increase the risk experienced by the supply chain participants (Jüttner, Peck, & Christopher, 2003). The tensions on the commodity market in metal parts supply or serious changes in a foreign currency exchanges rate are examples of risk drivers. They can lead to a serious risk event and it’s important to watch out and keep in mind that such events, although at first sight these seems to be far from your business, can cause problems to your firm. Competition and globalization increase risk indirectly whereas outsourcing which can results in increasing complexity can have a direct effect on risks (Jüttner et al., 2003). In essence, risk drivers are the start of causal pathways that ends up in risks. A further distinction can be made according to Kleindorfer and Saad (2005) on the nature of the risks or risk sources. These are variables (networking, environmental and operational) which cannot be predicted with certainty and which impact on the supply chain outcome variables (Jüttner et al., 2003). We can make a distinction on risks that come from coordinating supply and demand e.g. supplier fails in delivering the needed part, on the one hand. On the other hand risk arises from normal activities. These can be further subdivided in operational risk and risk arising from natural hazard, terrorism or political instability or called disruptive risks. These latter are risks with low probability, high-consequence of outcomes whereas the former has a higher probability of outcomes. However, most of the quantitative models are designed for managing operational risk. So there is a need towards more disruption risk models. Examples of the former are equipment malfunctions or human centered issues from strikes to fraud. An example of the latter is the fire at AVIA Company. A summary of risk sources can be found in the figure taken from Jüttner et al. (2003). Environmental
  • 17. 7 and organizational risk sources have an impact towards the supply chain whereas network risk sources are the risk sources of the supply chain (Jüttner et al., 2003). Figure 3: risk sources Manuj and Mentzer (2008) on the other hand divides the sources of risk in 4 categories (Supply, Demand, Operational and Security Risk) each with their risk event examples. Supply risks can contain besides the above examples supplier opportunism and inbound product quality. Or the supplier can default in flexibility to deliver the metal parts for AVIA just-in-time. Demand risk can be that our aircraft partner’s demand is very variable or a competitor from the assembled aircraft parts industry negotiates a more interesting demand with our metal parts suppliers. Through this AVIA is losing market share. Another example, but occurs more in food products supply chains, is that these products result in a weather-related demand uncertainty. For example, the demand of ice-cream is the highest when it’s a warm weather. The authors, Chen and Yano (2010), suggest a very flexible contracting scheme to optimize the distribution of risks between the manufacturer and retailer. This can be achieved for example through weather related rebate contracts to mitigate demand uncertainty. Next risks can be seen from an operational point of view like the risk on product quality failures. At last security risks or currency risks can also play a role. Melnyk, Rodrigues, and Ragatz (2009) added also information/technology risks (C. S. Tang, 2006), financial risks and legal/regulatory risks (Wagner & Bode, 2008).
  • 18. 8 Disruptions “Supply chain disruption is the unintended, anomalous event that materializes somewhere in the supply chain and threatens the normal course of business operations (Wagner & Bode, 2009)”. In other words disruption is anything that unexpectedly affects your supply chain. According to Sheffi (2005) these events follow a disruption profile in a predictable way in terms of its effect on company performance. From this perspective general problems can be roughly divided between deviation, disruption and disaster (Gaonkar & Viswanadham, 2004). Whereas the former can be more seen from an operational view, in essence a variation in lead times or demand within the supply chain from their expected or mean value, the latter two problems deal more on the environmental problems. With a disruption is the structure of a supply chain radically changed and with a disaster is the supply chain shut down temporary or irrecoverable. The authors formulated 2 mathematical optimization models that deal with only deviation and disruption problems because modeling disasters is simply impossible. Random events, land natural disasters like tropical storms and earthquakes can be best estimated from historical data for their possible occurrence. The likelihood of accidents on the other hand can also be estimated from industry data, prior events and the enterprise particular safety programs and implementations. Lastly, the probability of intentional disruptions such as job actions, strikes or sabotage) is the most difficult to estimate because the likelihood is a function of the specific company’s decisions and actions. Then, there is the difference between several kinds of storms according to Altay and Ramirez (2010). The impact of damage from windstorms and floods seem to be dramatically lower from that of an earthquake in terms of operational Cash Flow. The authors give the reason for the better predictability of these former 2 climate events and firm’s ability to prepare their firms in advance for them. Earthquakes damages a lot and makes recovery very slow and do not allow preparation time. In addition they show that the impact of natural disruptions is dependent on the firm’s position in the supply chain. The disasters that can be prepared can be planned for the upstream partners. There stock is accrued in advance and can be sold to downstream partners where these have opposite total asset turnovers. A solution to overcome this problem with the downstream partners is supply-chain wide risk practices because a firm that is not prepared will disrupt the operations of the rest of the supply chain.
  • 19. 9 Melnyk et al. (2009) proposes a discrete event computer simulation model that is based on the decomposition of a supply chain disruption in several facets like for example the quantity loss, time period, periodicity, profile breath & location of a disruption and the output level of its recovery towards the supply chain performance. They concluded that the use of classical statistical analysis is rather limited since they do not deal with the time dimension of disruptions. Because of the transient behavior of the process intervention analysis using time series is more appropriate in their study (Melnyk et al., 2009). A general rule should be to include a combination of methodologies in order to make a comparison. Perry (2007) builds a disaster response model after the 2004 Tsunami in Thailand that is for a part transferable to a business context. They highlight the logistic aspects (expertise and efficiency) and the need for quick information by extensive communication and local knowledge to deal with disasters quickly. This can be the case when some manufacturing activities are outsourced in a distant country. Vulnerability It seems according to Wagner and Bode (2009) that probability of risks are determined by supply chain characteristics (density, complexity, criticality, …) and consequently their vulnerability both as part of as well as across the entire supply chain. Vulnerability is defined according to (Blaikie, 1994;(Wagner & Bode, 2009) ) as a company’s capacity to anticipate, cope with, resist, and recover from the impact of a natural hazard. Several characteristics of the supply chain increase or decrease the vulnerability of the supply chain. For example extreme leanness and efficiency is very effective for a company’s operations and reliability towards their customers but may result in an increasing level of vulnerability. While lean management can provide several advantages in cost reductions and efficiency, it makes companies more hazardous to risk vulnerability and velocity. Consequently establishing back-up systems and maintaining reasonable slack can increase the level of readiness in managing risk. One can make a vulnerability efficiency operations vulnerability simplicity supply network Figure 4: vulnerability - efficiency relation (left), vulnerability - simplicity (right)
  • 20. 10 trade-off between robustness and overall efficiency to cope the level of risk. And because of the supply chain is only as secure as its weakest link minor movements can entail serious disruptions which makes the supply chain very vulnerable. Second, to reach more leanness or customer made products provided to worldwide demand, often this is coupled with an increasing complex network. But this must be paid off towards increasing vulnerability. We shall further see that one can overcome the vulnerability in complex network by being more resilient. Sheffi proposed a supply chain vulnerability map (SCVM) with four quadrants namely financial, strategic, hazard and operational vulnerability. Strategic vulnerability means the vulnerability when a new product is introduced. Hazard vulnerabilities are the internal as well as external risk drivers previously described. Operations vulnerability focus on the supply chain as for example distribution network failures. The framework is constructed in a manner that items of a category placed in the centre are very important and those on the edge less important. The goal of this framework is that each of the categories has a property to find, quantify and minimize risk (Blos, Quaddus, Wee, & Watanabe, 2009). Wagner and Bode (2006) found evidence of the effect of supply chain vulnerability drivers are positive towards more supply chain risk. This is the case for supplier dependence, single and multiple sourcing. So firms must according to Wagner and Bode (2006) avoid dependences and improve the robustness of a company’s chain. Meanwhile the choice of single or global sourcing must be done through a risk-benefit analysis (see further). Vulnerability is not in every industry the same. The aircraft manufacturing industry operates in an extreme risk environment, characterized by high levels of commercial, technological and political risk as well as the inherent product safety issues (Haywood & Peck, 2003). Interviewed companies from the author’s research acknowledged that their supply chain is most vulnerable during times of change as the risk profiles affecting their supply chains were also changing, but also that change is a constant state in their supply chain activities. Aircraft companies never experienced a steady-state resulting in increased supply chain change management (Haywood & Peck, 2003). To reduce a company’s weakness Asbjørnslett (2009) suggest to take a vulnerability analysis. It’s a top down analysis and its main focus is towards the system mission and the survivability of the system (Asbjørnslett, 2009). The essential steps to take this analysis is first to search for possible threats and their consequences, next the company must bring back their system to new stability by aligning adequate resource and last determine the disruption or the time the stability is again established. It gives a complete proactive vulnerability analysis framework that works in two rounds.
  • 21. 11 First the manager tries to understand the threats and risks, analyses and rank the consequent possible scenarios and is left over with a set of critical vulnerable elements in a first round. These require additional specific analysis that needs reduced fragility by adding appropriate resource to mitigate their criticality to them in a second round of investigation. 2.1.2 Risk management Risk management is defined as identifying and assessing the probabilities and consequences of risks, and selecting appropriate risk strategies to reduce the probability of, or losses associated with, adverse events (Manuj & Mentzer, 2008). The execution of an overall risk management process is useful for companies because managers tend to focus solely on critical performance targets, which affect the way they manage risk (C. S. Tang, 2006). The need for more supply chain and risk management has certainly become clear after the PWC investigation (Levi, Vassiladis, & Kyratzoglou, 2013). In their research they categorize enterprises in 4 levels of achievement of supply chain and risk management. They grouped the two lower and two higher levels together to reach some conclusions. Appendix ? gives some results from their study and show the percentage of companies with more than 3 incidents that suffered an impact of 3% or higher on their performance as a result of supply chain disruptions. First companies that invested in an advanced risk and supply chain management level are better equipped towards risks than lower risk management levels. Framework for a general structure of the risk management process Figure 5: risk management framework Supply risk management contains several steps and can be seen on the figure 5 (Hallikas, Karvonen, Pulkkinen, Virolainen, & Tuominen, 2004), (Griffis & Whipple, 2012; Zsidisin, 2003),). The different steps will be discussed successively. Risk Identification Risk Assesment, Evaluation and prioritization Risk Management actions and Mitigation strategies Risk Monitoring and Strategy Sharing
  • 22. 12 Risk Identification Equipment interruptions, quality failures and supply fluctuations. These are common strong signals of risks in manufacturing systems. The main focus of risk identification is to recognize future uncertainties to be able to manage these scenarios proactively in a later stage. Chopra and Sodhi 2004 identified nine broad categories of supply chain risks: disruptions, delays, systems, forecasts, intellectual property, procurement, receivables, inventory and capacity. Furthermore according to Manuj and Mentzer (2008) it is recommend that, once identified, risks should be segmented by specific characteristics in order to create a risk profile. You can categorize them in domestic or global risks. Once you start to investigate and identify risk, a common approach is to start with a brainstorming session with the management team with a diversity of people from sales, marketing, quality and finance if possible of your business. It can be helpful sometimes to get your session accompanied with your supply chain partners or major customers ((P. G. Smith & Merritt, 2002), (Preston G Smith, 2002)). By brainstorming you can base your business on the past as well as you can ask if everyone can think of success factors and wonder themselves what can go wrong? Actions that can be performed to discover risks are for example (Mullai, 2009): - Identify risk generating activities - Identify and formulate problems - Determine the background to determine the context - Define (technical, analytical) boundaries for the study - Collect relevant risk-related data and information Daimler Chrysler had to quit production for several days because of a defective fuel injector that came from their supplier Bosch, so the former company claimed his supplier for delivering the wrong part. Bosch claimed that it didn’t make mistakes but instead pointed at his supplier Federal Mogul for their faulty sockets which in turn found his supplier Dupont guilty for delivering defective granulates (Henke, 2009). A practical approach in finding the origin of disruptions is the use of the Tree model. This allows you to find the underlying root causes for today’s disruptions but also by using “what if” scenarios to get the root cause for future uncertainties (Griffis & Whipple, 2012). Ask yourself “what could go wrong at this point that would prevent us from achieving success”, especially for projects. Ultimately in a later stage this can form the basis for a comprehensive scenario planning approach (Sheffi & Rice Jr, 2005). Also a risk simulation can be done or a sensitivity analysis can be performed to check if some crucial parameters or outputs change in different scenarios. It is important to find
  • 23. 13 the causes because they require different modes of prevention and have also different potential impacts. An example of a fault tree analyses with root causes for our case of AVIA is given below: Figure 6: fault tree analysis for the AVIA example Secondly, we can address the reliability tools from Total Quality Management (TQM) to discover risks. One tool that can be used for Risk Identification is Failure Mode Effects and Criticality Analysis (FMECA). In essence this tool aims at performing bottom-up analyses of processes to determine where systems might fail, and then to either design out or improve detection of these potential failure points. The advantage is that this procedure moves from reactive to a more proactive means of equipment maintenance in an effort to reduce equipment breakdown and failure. But this tool is relatively absent from the supply chain literature because it often lacks the assessment/evaluation factors such as probability and likelihood. A better approach would then be Failure Modes and Effects Analysis (FMEA) in a supply chain context. With this method you have to identify and rank potential failure modes of a design or manufacturing process but its disadvantage is that it does not take criticality into account and thus does not completely address the potential impact of a risk (Griffis & Whipple, 2012). Especially for the identification of possible catastrophic events, Knemeyer, Zinn, and Eroglu (2009) applied this risk management framework for low probability high impact events. Companies have to determine the key supply locations with highest probability of threats and a list of them. Approaches that can help them are “internal assassin” whereby a manager who thinks as a terrorist and thinks Lost sales opportunity with aircraft industry metal part stock out lead time delay tensions in the commodity market lack of alternative sources of supply only minor suppliers available reduced production fire at the plant
  • 24. 14 about how to carry out threats against a firm and the “wheel of crises” whereby certain possible consequences of crises are discussed where the wheel stops. Risk Assessment, evaluation and prioritization Risk analysis or the assessment of a risk event is nothing else then weighting or measuring the subjective probability of a risk event and the potential consequences of it from the viewpoint of the enterprise. In a later stage companies should tailor the responses and strategies will be taken to reduce either their probability or their consequences. These two aspects can be used to develop a risk map or a risk criticality matrix: the probability or sometimes called the criticality index (how critical is a possible risk for your company) and the impact or several severity classifications. Negligible impact Marginal impact Critical impact Catastrophic impact Low Least Emphasis Probability High Most Emphasis Table 1: risk criticality matrix The aim of this matrix is to evaluate each risk on their emphasis and ultimately prioritize this risks to mitigate and map them in the matrix. Griffis and Whipple (2012) notices an incomplete picture and suggest that the probability of detecting these risk factors should be admitted in the traditional two- by-two matrix used by many other authors. (see also risk aspects). How can this additional factor be integrated? This is done again with a two-by-two matrix. Figure 7: risk matrix for the additional probability of detecting risks The manager can, for a specific risk factor, assess the ease in which the occurrence of that risk factor can be monitored (from easy to difficult) on the x-axis. On the y-axis, the lead time, from short to I Least Emphasis II III IV Most Emphasis
  • 25. 15 long, between detection and realization of the risk is depicted. For example if our metal supplies come by ship you have to take care and map several sources of risks. An example of the first quadrant can be a mechanical failure because it may have little to no advance warning of problems but once occurred in most cases the technical staff is capable in solving these kinds of failures. Weather fluctuations like a dangerous storm are immediately detected due to the accurate weather forecasts nowadays such that a vessel is able to take an alternative route and avoid the storm. The risk of piracy at last is difficult to monitor and characterizes with an immediate recognition of the detection, resulting in greater emphasis. Sometimes the firm can draw a tolerance threshold line that divides the risks you will manage actively form those that will not be managed, after which the risks identified are sorted by expected loss (P. G. Smith & Merritt, 2002). In other words, the company selects the maximum risk criteria it can afford. That’s another way of prioritizing risks when the firm has to cope with a lot of risks, especially minor risks. Prioritization is important as firms often focus only on recurring but low- impact risks at the expense of paying attention to high-impact but less-probable risks (Griffis & Whipple, 2012). In a further stage risk can also be compared against the selected risk evaluation criteria (Mullai, 2009) and further be ranked by criticality or severity. For the assessment and estimation of a catastrophe the use of simulation and optimization can be recommended (Knemeyer et al., 2009). Other estimation methods are for example the opinion of experts combined with historical data, suitable for aircraft incidents or with the opinion of decision makers, eligible for other types of catastrophic events like nuclear reactor meltdowns. The game theory whereby an optimal strategy has to be determined between the objective function of the attacker and the constraints of the firm can be used to simulate terrorism. The output of these approaches should be a list of key locations with estimated potential loss values. Risk Management Actions and Mitigation strategies Much research is done about management actions, strategies and action plans against risks. An attempt to give a reasonable overview follows. (Kleindorfer & Saad, 2005) (Manuj & Mentzer, 2008), (C. S. Tang, 2006), (Griffis & Whipple, 2012)) Under risk management actions we understand the general used strategies towards the risks perceived. They are risk taking, risk transfer, risk reduction and risk elimination respectively. Within each risk management action several mitigation strategies can be used in succeeding this action.
  • 26. 16 Firstly managers can choose to take the risk. Reasons therefore can be that the risk may be perceived to be low and the company is willing to accept the risk because of very little consequences for the firm. According to two German researchers, during the financial crises times, a lot of companies accepted their risks in this country. This was more the case with manufacturing companies. But after the crisis these companies shifted towards a more comprehensive approach of risk mitigation. This is in contrast with service companies who despite the crisis stayed to perform more risk acceptance strategies (Blome & Schoenherr, 2011). Alternatively, managers can opt to transfer their risks from one company to another or subdivide their risks over several companies. This may reduce the total risk in the network if the company takes the risk can cope with it better than the company transferring it resulting in having a large supplier network. But the downsides of this risk are the high switching and administrative costs and their availability when changing from supplier. These are therefore part of transaction-specific investments. Furthermore it may decrease opportunities to achieve economies of scale. Furthermore risk can be shared in contracts with the intention of better coordination with channel partners, collaborative forecasting and collective replenishment planning which increase supply chain visibility and encourages further analysis of individual risks. They can be managed generally by developing a common network strategy, sharing best practice modes of action and contract policies. Moreover, several situations exist that there might be some risk but the company takes the needed effort to reduce it as much as possible. A common used mitigation approach could be the use of non- performance penalties built into contracts. If our metal part supplier doesn’t succeed to deliver the demanded parts, price reductions will be used as stated in the contract. Examples of other security mechanisms used to reduce risk include monitoring techniques, such as audits of supplier’s quality checks, inspections of random materials, and tracking of key performance indicators (KPI’s). At last, risk elimination may be appropriate when the firm cannot live further with this risk and must be completely discarded. AVIA decided to quit assembling their aircraft parts with that old machine that produces much defects. Sometimes choosing an appropriate risk strategy means changing current operating models or practices. This means that you systematically review your ‘inventory’ of risk procedures and controls with the aim to improve risk management practices. An example is the centralized versus localized approach of manufacturing to mitigate risks and increase benefits.
  • 27. 17 Table 2: risk management actions To deal with catastrophic events the company can draw a catastrophic risk management matrix which maps the key locations in the same manner as normal risks to detect appropriate risk strategies for each threat on the list like for example move a location, buy insurance, assume risk and so on (Knemeyer et al., 2009). Chaos theory may additionally provide some help to formulate appropriate catastrophic risk strategies. Wagner and Bode (2009) makes a difference between cause and effect oriented supply chain risk management practices. The first are preventive in nature. We think of information security, physical security and freight security. For example AVIA can switch in advance to a more financial stable supplier to reduce the risk of a sudden supplier default. Or the company can relocate their manufacturing plants to safer regions to avoid natural hazards like tropical storms or tsunamis near the coast. The second practice contains measures aiming at minimizing the level of damage in case of a risk event occurrence, e.g. insurance companies. The disadvantage of these companies is that they do not always understand supply chain risks and it’s difficult for them to insure a company’s own facility against disruptions from their suppliers at multiple locations. But there seems to be progress in this field: They are now providing business interruption insurance for disruptions occurring at a supplier’s facilities (Alvarenga & Lehman, 2012) for named suppliers but unfortunately don’t cover the whole network of suppliers and subcontractors. Buffering strategies, financial risk reserves and product redesign are other examples of this practice. Most of the risk handling activities proposed in the literature are rather effect-oriented than cause- oriented. As risk mitigation strategies require costly investments in equipment as well as human resources, it is important to know which mitigation strategies offer the greatest protection from risks in a certain situation. risk taking • ignoring the risk when developing a mitigation strategy risk transfer •large supplier base risk sharing •channel coordination •collaborative forecasting •collective replenishment planning risk reduction •non performance penalties •audits of supplier's quality checks risk elimination •remove machines •stop operate unhealthy production processes
  • 28. 18 A possibility for determining the favorable assessment costs is mapping them against the benefits from risk mitigation strategies, stated in the framework of Shavell (Kleindorfer & Saad, 2005). This results in a tradeoff between the cost of acquiring reliable information on risks and the benefits of mitigation activities. At optimum, a balance must be struck between the marginal costs and benefits of better risk assessment. So when do we have to use these mitigation strategies? Risk identification and assessment give a more specific indication on where to focus the actions. According to Griffis and Whipple (2012), strategies such as monitoring and risk taking can be used when the likelihood and potential impact are low and the ability to detect the risk is easy. When selecting a monitoring strategy, either risk reduction or elimination, you can choose to perform random inspections of products to detect errors or the risk maybe perceived low so that managers could choose to take the risk. When the opposite is true, more aggressive risk mitigation strategies like complete risk elimination need to be considered. Examples here are avoiding dangerous shipping routes or to quit outsourcing and to manufacture the product in-house to have more product and process control. If the likelihood of a risk occurring is low and detection may be easy, but the impact of the risk could be significant, then a postponement strategy may be appropriate if the event causing the risk can be postponed until more control by the focal firm is established. More control can be established through vertical integration or imposing contractual obligations on suppliers (Jüttner et al., 2003). The clearest example of a postponement strategy is producing in modular form. The advantage is that you can push your semi-finished product from surplus to deficit areas. A company that uses a postponement strategy is the computer manufacturer Dell. They produce computer hardware in modular form and let their customers and firms decide which functionalities and properties they must contain. One can also perform a demand postponement strategy and shift the demand across products towards their customers (C. S. Tang, 2006) such as a price strategy. The opposite of postponement is called speculation or also called selective risk taking and is also an option here. When you perform a speculation strategy, you build up inventory to buffer against the specific risk. In cases where impact of a risk may be low, but likelihood of occurrence is high, and the ability to detect the risk in advance is difficult, a firm may select an imitation strategy and source with the same supplier because if one firm is exposed to this risk, all firms are. A flexibility strategy at last could be used when the likelihood of risk occurrence is high and detection is easy. This could be achieved through multiple sourcing. This strategy requires some adaptation for
  • 29. 19 the company because in a general culture where the focus lies more on core competencies and value creation more single sourcing relationships have emerged (Blome & Henke, 2009). Strategic partnerships and alliances are an example of this sourcing strategy. Secondly, companies don’t have always the choice to choose between single and multiple sourcing. For example, when you have the choice of only one supplier, because of intellectual protection, then you have a sole sourcing relationship. This mitigation strategy is sometimes called hedging in a supply chain context because the company has a globally dispersed portfolio of suppliers. Another way to provide flexibility is the adoption of standard processes and the use of interchangeable and generic or modular parts. Finally using simultaneous instead of sequential processes in key areas as production/distribution speeds up the recovery phase after a disruption. Figure 8: risk mitigations strategies Tomlin (2009) determines also the optimal adaptation strategies when probability of supplier and customer failure is high or low. The used strategies are supplier diversification, contingent sourcing, which is adding a supplier which is only used in case the main supplier fails to deliver and demand switching. Diversification should be executed when demand uncertainty increases. Furthermore when the firm faces an increasing supplier failure probability or faces a high level of risk aversion it should opt for a contingency strategy. Demand switching is appropriate in case of a low supply risk. Mitigation strategies can change over time. Suppose company AVIA is plagued with several recalls for their assemblies because of metal parts affected with corrosion in an earlier stage. Because it isn’t always easily detectable, current employed postproduction testing is no longer efficient anymore. Rater than a strategy of control, through frequent testing, a strategy that uses severe penalties with
  • 30. 20 their suppliers for recalled products is more appropriate. Another possibility can be to use an elimination strategy and look for another supplier. After the appropriate mitigation strategies are determined, it is recommend developing prevention and contingency plans to reduce the risk in likelihood of occurrence and impact severity. Wagner and Bode (2009) suggest that continuity or recovery plans are important tools to ex-ante optimize the ‘firefighting’ after a disruption. These contain for example the radical design of products and the layout of the manufacturing processes. As a review to the selected mitigation strategies, some principal criteria can be addressed: efficacy or the degree to which risks are eliminated, feasibility or the aligning of the right mitigation strategy to the appropriate risk and efficiency which relates to the cost-effectiveness which was explained above. Risk Monitoring and Strategy Sharing Monitoring your risks means identifying the potential increasing trends in their probability or consequences in the future. After implementing mitigation strategies you will find that some risks are closed where the risk event has been prevented or other risks remain where the risk event had happened despite the prevention plans implemented (P. G. Smith & Merritt, 2002). Nowadays companies can employ real-time risk monitoring capabilities along with techniques to track key supply chains flows. These tools can speed response in case of numerous unplanned events. A lot of electronic and high-tech companies, who have very dependent supply chains, have integrated these tools into their standard supply chain management practices. Improving the traceability of the supply chain leads to organizations that follow key performance indicators through the entire supply chain and consequently identify risk not only with their first-tier suppliers but also with their sub- contractors ((Alvarenga & Lehman, 2012), see also supply chain security). A company first follows the risk management steps described above and analyzes its network-related risks internally. In the second phase the partners should identify the areas of risk management that require joint effort and where risks strategies should be shared. As enterprises are connected in a network, they are dependent on each other so it can be useful to share entirely or partially risk management processes and to develop collaborative means to manage the risk and communicate their views on risks. It is important that the individual risk management processes are supplemented by a collaborative process. Sheffi (2005) even argues that competitors should collaborate to control common risks.
  • 31. 21 Moreover in complex network environments mutual risk identification and assessment can be seen as tools for creating the risk profile of the entire network on the basis of the partners' risk profiles (Hallikas et al., 2004). The primary tool employed by the Japanese to implement closer supplier co- ordination and individual supplier development is cross-exchange of staff between buyers and suppliers. This requires the benevolence of the enterprise of exchanging inter-organizational information towards risks & rewards sharing and knowledge transfer. But it gives the firm the possibility to perform a benchmarking exercise and it generates supply-chain wide visibility of vulnerabilities and it should give the firm incentives to identify and implement disruption management systems. Conclusion Now that we have briefly described the risk management process framework, one can argue of its need. As risk mitigation strategies require costly investments in equipment as well as human resources, it is important to know if these strategies pay off towards risks of all kind. As Jüttner et al. (2003) noted, there is a supply chain trade-off decision between delivering high customer value and managing possible risks. A trade-off between extra risk mitigation costs and less costs of delivering high quality and on-time products as a main principle of supply chain management. Kleindorfer and Saad (2005) investigate if investments in risk management activities yield towards frequency or severity of accidents. With the use of variables like regulatory programs, facility characteristics and community demographics, they determined whether observed accidents in the chemical sector decreases with the use of these risk programs as mediator variables as a consequence of more severe regulatory programs, more hazardous facility characteristics or the financial structure of the company. The investigation indeed found evidence of this relationship. Dani (2009) suggests this risk management framework must be an iterative process and should not stop with one investigation of risk but instead repeat the exercise to study new issues and risks identified after the analysis of the event. Furthermore this exercise must be aligned at the strategic level of the company and according to the strategic objectives to have a clear understanding. Concluded, it is important to update the possible risk sources and strategic objectives in line with the risk event or mitigation strategies that may be adapted according to possible new discovered risk issues (Dani, 2009). We will further see that the adoption of a risk management strategy will foster the use of a proactive supply chain approach. Mullai (2009) takes it a step further and claims that the process can start at any point. The major steps of the framework (risk analysis, evaluation and mitigation) are interactive, change-responding, can be accomplished simultaneously and are aligned through risk communication (Mullai, 2009).
  • 32. 22 2.2 Supply Chain Management The first definition of the supply chain management is dated from the early 1980’s and compromises the following: “a ‘standard’ supply chain is a system compromising of materials, goods and information (including money), which pass within and between organizations, linked by a range of tangible and intangible facilitators, including relationships processes, activities and integrated (information) systems” (Peck, 2006) While our approach is not to give a full overview of the supply chain with all his aspects, we will nevertheless give you some information of some practical aspects in this domain that can be linked or have relationships with the risk management domain. Bullwhip effect An important issue for the supply chain is that you have to take into account the major consequences of the Bullwhip effect. Essentially, the bullwhip effect depicts the phenomenon in which the orders exhibit an increase in variability up the supply chain, even when the actual customer demands were fairly stable over time (Sterman 1989; (C. S. Tang, 2006). Cisco systems Inc. wrote off 2.5 billion in inventory due to a lack of communication among its downstream supply chain partners (Spekman and Davis, 2004(C. S. Tang, 2006). The increase in variability of the orders up the supply chain can cause many problems for the upstream partners including higher inventory, lower customer service level, inefficient use of production and transportation capacities, etc. The more distance between suppliers and the final consumer in the supply chain, the more these demand changes are compounded (Fine 1998; Lee, Padmanabhan, and Whang 1997(C. S. Tang, 2006). In order to mitigate the bullwhip effect, one needs to identify the root causes (C. S. Tang, 2006) which can be done in the first step, risk identification, of the risk management process. Secondly, many companies have switched from “local” suppliers to “low cost” and often distant suppliers on the basis of overhead cost optimization, without considering the cost of risks caused by this strategic change. Larger companies now buy from smaller suppliers in very remote areas of the globe. The extended supply chain now has many additional points of potential failure, enlarging the bullwhip effect and requiring new approaches to risk management. Companies face longer logistics lead times as well as new and unfamiliar risk profiles encompassing natural disasters, epidemics, and social, political or monetary instability (Alvarenga & Lehman, 2012).
  • 33. 23 An agency theory perspective The agency theory perspective justifies the differences in the objectives and risk preferences of the two parties: the principal (purchasing organization) and agent (suppliers), as well as information asymmetries. Both parties undergo an agreement with risk sharing (Zsidisin & Ellram, 2003). The aim to consider the relationship in this perspective is to reduce the purchasing firm’s risk of moral hazard and adverse selection. The first means the risk of the lack of the supplier, aware of not by the purchaser, to lever the agreed upon effort to meet customer demand. Adverse selection means the inaccurate assessment or misrepresentation form the purchaser of the sometimes unknown supplier abilities to meet customer requirements (Zsidisin & Ellram, 2003). An example of moral hazard is the unwillingness to further invest in appropriate infrastructure needed to produce the metal parts for AVIA. An example of the latter is that AVIA, unaware of the major investments in new metal production equipment, keeps further purchasing the parts from old machinery that are produced with minor quality. In essence, suppliers and buyers have to be aware of opportunistic behavior risk. This comes down to the breaking of their mutual informal agreements & contracts between the partners within the supply network in the pursuit of competitive advantage and profit (Seiter, 2009). The author proposes action programs like more communication quality, better partner selection and mutual sharing of cost accounting information to reduce opportunistic behavior directly or indirectly through reduced information asymmetry. Make-or buy decision Don’t try to force the manufacturing of complementary assets in-house when you can outsource particular needed competences in a more cost-advantageous way. For many companies, “make or buy” decisions have been chosen in favor of buying, not making. While this reduces manufacturing overhead costs, companies lose oversight of key governance and management competences and strategies. As a consequence this might introduce unknown (new) risks into the supply chain. Periodic risk rebalancing is therefore essential (Alvarenga & Lehman, 2012). Outcome-based versus Behavior-based management techniques In order to align the objectives of both agents and principals several management techniques or practices are available. These can be split in two categories (Celly & Frazier, 1996; Zsidisin & Ellram, 2003). Outcome-based management techniques address the importance of coordination of outcomes and results such as sales growth or sales in relation to targets. According to Zsidisin and Ellram (2003), the
  • 34. 24 use of buffer oriented techniques is an example of that group. Inventories can be held either by the purchasing firm resulting in internal safety stock or by suppliers which is supplier-managed inventory, or both. Rather than reducing the likelihood of a harmful event, firms employ buffers to reduce the disruptive effect of supply risk events. Therefore this approach is short-term oriented. Behavior based management techniques addresses behaviors such as customer education activities or selling techniques with distributers from the supplier personnel thereby signaling important objectives and suggesting specific distributor actions (Celly & Frazier, 1996). It focuses on processes, emphasizing ‘task and activities” that lead to a reduction in supply risk and is therefore long-term oriented. The findings of Celly and Frazier (1996) were that supplier personnel rely too much on outcome- based efforts when coordinating relationships with distributors. Zsidisin and Ellram (2003) on the other hand found that there exists a partially positive relationship for behavior techniques with perceived supply chain risk and this was not supported for buffer oriented techniques. The implementation of buffers is done regardless of the extent of perceived supply risk. Although each of the efforts has its downsides, outcome-based may be sometimes inappropriate in some situations and behavior-bases may be sometimes costly to the firm. An emphasis solely on maintaining buffers to manage supply risks can harm business profitability due to capital devoted to inventories and Celly and Frazier (1996) find the risk of demand fluctuation is being positively related to behavior-based management efforts. On the other hand buffers give you a form of assurance and behavior based contracts facilitates communication and mutual goal alignment. An inclusion of the two efforts will be optimal. Examples of the management techniques are given below: Outcome Buffer Behavior - communication about sales growth, market share, … - managing inventory - supplier certification - quality management programs implementation - target costing - supplier development The advantage of supplier certification is that they reduce the need for the purchasing organization to conduct time-consuming inspections. Implementation of quality management programs is also an
  • 35. 25 option but certification improves already the abilities of the supplier to satisfy the quality expectations of the purchasing firm. Also target costing is interesting where supplier and purchaser determine ways to drive cost out of its products. At last supplier development can meet the purchasing organization’s short or long-term supply needs. Behavior-based efforts may be predominant in franchise channel systems because inter-firm interaction should be relatively high and relationship termination is more difficult. In conventional channel systems, outcome-based efforts may be predominant because of the need to keep coordination costs down. In addition we will find more buffer-oriented approaches within the use of transactional supplier relationships whereas the use of behavior outcome-base approaches is more adopted with cooperative supplier relationships (Blome & Henke, 2009). Making your supply chain redundant versus flexible According to Sheffi and Rice Jr (2005), the following difference can be made. Redundancy activities include safety stock, the deliberately use of multiple suppliers or back-up sites. Adding redundancy is in a way needed for every day’s operability but flexibility strategies on the other hand gives more competitive advantage in the marketplace. Flexibility can be seen from 3 facets: supply, in-house conversion and distribution. It aims first of all on correct alignment of the supplier relationship with the procurement strategy whether you opt for a single supplier or for multiple suppliers or for a single supplier for each critical part. Secondly using standard processes and having multiple locations with built-in inter-operability. This allows a firm to operate in another plant once one is disrupted or the replacement of sick operators. On the customer or distribution side at last, managers face a choice about which customers to serve first after a disruption. Managers can thereby decide on several criteria like the vulnerability, the profitability and costs of all customers. Strategies like postponement or producing semi-finished products, described above, can be used. Managers in addition are often reluctant to invest in flexibility measures because they don’t see directly or hardly can estimate the impact on risk mitigation (C. Tang & Tomlin, 2009). So unlike redundancy, flexibility can also improve the competitive advantage of companies in times without a disruption. Flexibility at last is indeed important when your company executes a lot of projects because the project manager must be able to change or redefine several aspects during project execution. Furthermore projects ensure the coordination between client, contractor and supply chain. Hence, project execution translates into flexible management. But gaining more on flexibility and consequently efficiency implies the risk of losing sight and therefore as we have seen before can increase vulnerability. Companies producing components for the automotive sectors are an example of industries that focus on project management (Gaudenzi, 2009). It provides a risk management framework that is also applicable for projects.
  • 36. 26 C. Tang and Tomlin (2009) investigates how much flexibility is needed to mitigate supply chain risk. On the basis of mathematical models they conclude that only a small amount of flexibility strategies (e.g. multiple suppliers, flexible supplier contracts, postponement, responsive pricing and flexible manufacturing processes) is required to mitigate risk. From their PWC research (Levi et al., 2013), it seems that companies focus on flexibility and customer service levels on the one hand and other companies on cost reduction and efficiency on the other hand. It seems according to their study that those former are better coped against risks (appendix). It’s also worth noting that 80% of the cost-efficient companies face high variable supply chain lead times given that low variability is mostly one of the key drivers of an efficient operating strategy. The resilient supply chain Once gone through all the major risk management steps and approaches the ultimate goal is to build a resilient supply chain. A supply chain that is both able to absorb disruptions and risks, can proactively manage its supply chain and sees a possibility to turn the threats of a disruption in a major advantage can be seen as a resilient enterprise. Resilience is not the same as robustness. The robustness of a company is the ability to resist from an accidental event, retain to its same stable situation as it had before and stick to his initial mission (Asbjørnslett, 2009). In contrast resilience aims at a new stable situation and has the adaptive ability instead of being resistant. Building a resilient enterprise should be seen from a strategic level and changes the way a company operates and increases its competitiveness. Two important aspects determine a company’s resilience (Sheffi & Rice Jr, 2005). First is the market position. Is the industry competitive or has the company a lot of market power? Else it depends on the responsiveness of the supply chain. When companies are not so responsive they might risk losing market share. Responsive companies otherwise can increase market share or once they have a reasonable amount of market power they can lock in their leadership. Corporate culture In the search for the resilient enterprise, it is important not to underestimate the contribution of culture to an organization (Sheffi & Rice Jr, 2005). Empowering front-line employees to take initiative and guide actions is a possible step in building a suited corporate culture. Secondly Japanese lean principles from the Toyota Manufacturing System can be used like Poka Yoke and Hijunka. These concepts are in essence that one must learn from errors and fixe the root causes. Companies can minimize the risk of possible disruptions by paying attention to small problems as indications of major disruptions. At last continuous communication between all layers of a company can foster the good working of the company.
  • 37. 27 To encourage the need and use of risk management a disruption can be simulated and employee reactions are monitored and used in training (Kleindorfer & Saad, 2005; Sheffi & Rice Jr, 2005). This is the exercise of role playing with red and blue teams. The red team represents in most cases competitors, rivals or supply chain experts, equipped with whatever information is available. They attempt to attack the supply chain to cause major disruption. On the other hand you have the blue team who tries to mitigate or countermeasures those actions which are cost effective against the Red Team scenarios. This way of training can enhance the risk culture in the firm. This lacks yet in a lot of firms according to the authors of the paper from Cagliano, De Marco, Grimaldi, and Rafele (2012). Sheffi (2008) proposes that governments can introduce cultural aspects into disruption mitigation. Cultural changes on a societal level happen several times every century so the government should encourage and drive these trends to improve resilience along the chain. 2.3 Supply Chain Risk Management Research in supply chain risk management is still in an early stage and has been around for about a decade now. As a consequence here is a huge diversity in topics, opinions, and research methodologies in the field of supply chain risk management. Furthermore there is still a lot of variation towards the meaning of supply chain risk management according to several established focus groups. According to them, supply chain risk management is seen as a subset of Supply Chain Management (SCM) and also as a subset of Enterprise Risk Management (ERM) which can be seen in the figure below. But what compromises the field of supply chain risk management (Sodhi, Son, & Tang, 2012)? Supply chain operations and risk management processes go hand-in hand and complement one another. Figure 9: the concept of Supply Chain Risk Management (Blos et al., 2009) According to Wagner and Bode (2009), Supply chain risk management (SCRM) contains the field of activity seeking to eliminate, reduce and generally control pure risks in supply chains. But what do
  • 38. 28 these risks contain? According to the focus group from the research from Sodhi et al. (2012), risks in this field should concern majorly with dealing with unknown events or dealing with disruptions or disasters with low probability and high-impact. This was the opinion of almost the half of the researchers. Secondly it should concern dealing with risk within supply chain operations, according to 20% of the researchers (Sodhi et al., 2012). Kajüter, 2003 states: “Supply chain risk management is a collaborative and structured approach to risk management, embedded in the planning and control processes of the supply chain, to handle risks that might adversely affect the achievement of supply chain goals.” In this definition the focus is rather laid on the supply chain. Norrman and Lindroth, 2002 assess the focus of supply chain risk management more on a practical approach. Supply chain risk management can be defined as, “to collaboratively apply risk management process tools with partners in a supply chain to deal with risks and uncertainties caused by, or impacting on, logistics related activities or resources” Nevertheless this concept is still missing many pieces that have to be found and linked together to get a comprehensive approach of supply chain risk. SCRM is often not always established as a distinct function or department in companies. Businesses do not agree on how to integrate these risks into their decision-making processes, the risk function is typically “headquarters-centered” and there seems not be a risk regulation that covers the supply chain. However it’s necessary to integrate risk management into operations, strategic and sales planning. The work from Jüttner et al. (2003) summarizes well the concept:” “the identification and management of risks for the supply chain, through a coordinated approach amongst supply chain members, to reduce supply chain vulnerability as a whole.” And can be summarized using the figure from them.
  • 39. 29 Figure 10: a SCRM framework As a consequence it shouldn’t come as a surprise that research in this field leads to often unambiguous conclusions and a large variety of results. But SCRM implementation has brought up: Ericson has implemented a SCRM matrix organization that spans the entire company, from the corporate and strategic level to the functional and finally process-oriented operational level. It’s in addition worth to address the importance of smaller companies. Most SME’s are more exposed to supply chain risks whilst simultaneously disadvantaged by lack of management resources, structures, processes and expertise as opposed to bigger firms (Henke, 2009; Jüttner & Ziegenbein, 2009). In essence, for SME’s the adoption of a comprehensive risk management program can be the same except for some aspects due to their limited possibilities. Whereas larger firms can have the time and the access to control all their supply chain partners not strictly limited to 1st tiers, SME’s are encouraged to select their most important one for further investigation. In most cases information is gathered from their personal network and there is no planned supplier data collection but instead these companies rely more on social interactions with their suppliers (Ellegaard, 2008). It’s not practical for them to map their entire supply chain (Jüttner & Ziegenbein, 2009). Consequently before the risk identification process, Jüttner and Ziegenbein (2009) require the mapping of supply chain vulnerability against their strategic importance to investigate. Furthermore the adoption of expensive risk management tools is not always value adding for them. That is because tools often are designated for specific tasks and are highly sophisticated and therefore not always suitable for SME’s. In summary a 3-phase risk management approach for SME’s is given. The tactics towards the risk aspects are according to a case study from Ellegaard (2008) as follow: probability reduction has the highest priority and effect reduction through multiple sourcing was not
  • 40. 30 practiced at all. Also small companies don’t gather much information from their partners that can help reduce risk. A company must regularly check their activities in a manner that they fulfill their strategic objectives dictated by the firm’s vision. A framework that can aid to measure performance is the Balanced Scorecard of Kaplan and Norton. Their model doesn’t contain only financial objectives but incorporates also internal, customer and innovative & learning perspectives. When dealing with supply chain performance, no primary use of financial performance measures is advised (Sheffi & Rice Jr, 2005).This framework which focuses on performance measurement as a consequence should be integrated with supply chain risk management. Risk and performance are directly related because the higher the risk taken the higher the expectation of suitable returns (Sheffi & Rice Jr, 2005). As more resources and attention is given to supply chain risk management an aligning with performance management is indispensable. Besides the electronic multinational Ericsson (Norrman & Jansson, 2004) another company that has successfully implemented a risk management strategy and accompanying supply chain culture is the Nisan Motor Company (Schmidt & Simchi- Levi, 2013, (Levi et al., 2013)). After serious harmful problems due to an earthquake and Tsunami in Japan, they incur a $ 200 billion loss. “Nissan’s production capacity was perceived to have suffered most from the disaster compared to its competitors.” But the company managed to increase their production with 9.3 % in comparison with an overall decrease with its competitors. How was this company able to manage this? First they implemented an integrative risk management framework with risk identification as early as possible, assessment and performed countermeasures against them. Second the plant has formulated a continuous ready plan with their suppliers. Moreover the team was empowered with local decision management. In addition the company structured its supply chain as flexible as possible. At last, there was extended enterprise visibility and warnings to between internal and external business functions (Levi et al., 2013). 2.4 Supply Chain Security Management Supply Chain Security Management (SCSM) is defined as “the application of policies, procedures and technology to protect supply chain assets from theft, damage, or terrorism and to prevent the introduction of unauthorized people or weapons of mass destruction into the supply chain. Security practices can be control measures implemented in several processes and certainly at several gates throughout the supply chain where products arrive (Voss & Whipple, 2009).
  • 41. 31 The difference with Supply chain risk management is the fact that Supply chain security management focuses more on prevention of contamination, damage or destruction of the supply chain assets and products by policies, procedures and technologies whereas the former investigates more the likelihood of outcomes being susceptible to disruptions that can damage the supply chain (Autry & Sanders, 2009). This paper provides a dynamic capability framework that can be used to implement security practices. These capabilities include processes, technology and human resources. For example the use of radio frequency identification is a technology capability. With the use of the Radio Frequency Identification (RFID) technology and their tracking and tracing capabilities, abilities rise to identify disruptions quickly. Furthermore sensitive control systems often can identify a disruption before its cause is apparent. The management of inventory is an example of a process capability and governmental security initiatives are human resource capabilities. The problem with risk mitigation activities, in particular those that don’t start at the place of original production, is that technical control allows hidden action of participants in the process chain (Mau & Mau, 2009). Therefore a supply chain wide security control system or management is needed. A comprehensive control platform is required to manage supply chain security effectively. The goal for supply chains, those that doesn’t consist of a complex network or single supply chains, is to secure worldwide and realize complete traceability of all involved products and inputs at all the supply chain levels resulting in continuous transaction data in both directions, upstream and downstream. To trace all the necessary information, it is useful to make use of an independent database. In this way effects on the products of your firm can not be overseen thanks to a centralized data management system. The basic thought for implementing security initiatives are basically the same as for the use of risk mitigation strategies. Advanced or high proactive security initiatives pay off and improve security and firm performance but most firms have not progressed beyond basic physical security measures such as infrastructure management and therefore have not derived the service benefits from higher level security measures (Voss & Whipple, 2009). Consequently this creates a dilemma in an optimal tradeoff between cost of implementation and efficiency of the initiatives but also for firms that face greater customer or government requirements. More security improvements sometimes can lead to a decrease of flexibility and firm performance but aim to create a secure supply chain that maintain advanced security processes and procedures.
  • 42. 32 2.5 Conclusion Concluded, risk mitigation measures can be seen and implemented from 3 different levels. The strategic level which focus on for example alternative suppliers, at the tactical level (e.g. improved demand forecast) as well as the operational level (e.g. business continuity plans). Although risk management practices and business continuity planning is left too much to security or insurance professionals in companies, it should be noted as a strategic initiative and must be implemented in an integrated risk framework approach. As a consequence this risk practice helps to build a resilient enterprise and supply chain.
  • 43. 33 3 Conceptual framework The goal of this work is to determine how several environmental aspects (e.g. market size, bargaining power) may have an impact on the managers or the firm’s risk perception by which we mean the probability of risk and their impact. Furthermore we investigate if these perceptions might influence risk mitigation strategies performed by the firm. In particular we will make the distinction between more proactive versus more reactive performance towards risks. At last we want to examine if certain mediators and moderators do play a role in selecting that particular strategy. On the one hand we dive into how varied a firm does perform supply chain management practices and their choice towards risk management. On the other hand we look at the design and the complexity of the supply chain and see if any conclusions can be made with this aspect towards risk. While previous research (Ellis, Henry, & Shockley, 2010) has determined the impact of environment on risk perception, investigation towards the use of proactive versus reactive management is still a gap in the literature that hasn’t been explored yet. Instead Ellis et al. (2010) investigated overall supply disruption risk towards the search of alternative sources of supply. 3.1 Environment Several papers deal with the environment as an important factor for the determination of risk (Ellis et al., 2010) (Ritchie & Brindley, 2009). Ellis et al. (2010) chooses also for the integration of some environmental factors like technological uncertainty and market thinness but also for some company specific characteristics like item customization and item importance. The environmental construct here proposed is as follow: PEST-analysis An exercise for the company in determining the strength of the environment and the external factors can be the adoption of a so called PEST analysis. This analysis contains the macro-environmental aspects of Political forces, Economic forces, Social forces and Technological forces. It’s an environmental scanning component that can help you to determine trends outside the venture. The purpose is to detect driving factors and uncertainties.
  • 44. 34 Figure 11: PEST analysis Porters competitive model and SWOT-analysis A second approach to identify your environment is the so famous Porter’s five forces model. This framework provides you and let you think about you’re major outside forces that come into play with your venture and determines the dynamic tensions between players. The aim of this tool is to perform a competitive analysis because each player may have a sufficient amount of power. It highlights problem areas as well as possible opportunities or whether you have a competitive advantage or you have to accept a reasonable amount of risk. An exercise that can complement this investigation is the SWOT-analysis (Strengths, Weaknesses, Opportunities and Threats). The threat of potential new entrants can limit your own company’s activities but there may be entry barriers that could prevent them to start doing business. Second, suppliers may have bargaining power that impacts your business such as a unique product they deliver to you that requires a high degree of specification and specialization. Customers likewise may have power too over your business for instance if they are able to integrate backwards into the supply chain (e.g. manufacturing your product themselves). In addition there may be a threat from substitute products. Note that this doesn’t contain physically the same products but may also encompass a different transportation mode like for example plane versus car. Substitute products can be more or less attractive depending on for example the switching costs a customer must pay. For example the adoption of a new product would require the customer to buy new equipment or additional software and so on. At last, existing rivalry among industry players can make your environment very turbulent. If there are many •technological pace •adoption cycles •R&D •education •demography •work aspects •financial markets •economic cycle •change of industry and markets • laws & regulations • political (in)stbility • government spending Political Economic TechnologicalSocial
  • 45. 35 competitors and there is little product differentiation, the resulting strategy in your industry will be price competition. Figure 12: Porter's Five Forces Model Market variables At last we introduce some market variables that also can be perceived from the environment and are important for the firm. First we account for the perception of the market size: “is it growing or declining rapidly?“ Moreover has the industry market many or few segments resulting in a company’s market span. In addition, “has the industry many or few competitors?” This describes the market concentration. At last, we can focus on the ease to enter the market. “Is it open to new players or closed to new players?” So concluded, our first hypothesis will be: Hypothesis 1: Environment constructs – Risk perception relationship H1: there is a significant relationship between Market-variables, some PEST-analysis variables & PORTER-analysis variables and the probability of risk and their impact. 3.2 Risk perception or representation of risk Risk perception is defined as the decision maker’s assessment of the risk inherent in a situation (Sitkin & Pablo, 1992). This refers to the assessment stage in the risk management process described earlier.
  • 46. 36 Prior research has already proved the relationship between risk perceptions and risk behavior although as a mediator between risk propensity and risk behavior (Sitkin & Pablo, 1992). We will reinvestigate this relationship but consider the perception in both risk probability and risk impact. Hypothesis 2: Risk perception – Risk management relationship: H2a: there is a relationship between the risk perception of risk probability on the choice of pro versus reactive management. H b2: there is a positive association between a higher perception of risk impact on proactive as well as reactive management and the relation on proactive is stronger. We opt to choose the aspects of risk, impact and probability, from a perception point of view rather than objective assessments of risk. That is because a manager will take decisions on the basis of his perceptions of degree of risk and which actions he will do. Risk perception is dependent from several aspects: as earlier stated risk is dependent on how the problem is framed and the problem domain familiarity. Furthermore an individual’s propensity towards risk also influences the perception. At last the homogeneity of the management team and organizational control systems come also into play (Sitkin & Pablo, 1992). We represent the risk from 3 perspectives. First from a supplier point of view: “What is the probability that a key supplier will fail to supply your operations and what will be the impact?” Secondly, risk perception will be gathered from the own manufacturing operations: ”what is the probability and the impact that the company’s operations are interrupted and affects your shipments?” At last probability and effect can be seen from the distribution side: “What is the probability and impact that a company’s shipment operations are interrupted affecting their deliveries?” 3.3 Proactive versus Reactive Risk Strategies The distinction between proactive and reactive risk management can be approached in relation to their timeframe. Proactive can be seen from an ex-ante and reactive from an ex-post perspective. But it emphasizes more than solely this. What is proactive risk strategy actually? Proactive risk management means that you identify risks and do something about them before they affect your project or your business operations. This is in fact the use or the implementation of the risk management framework that was introduced and
  • 47. 37 explained in all its facets in the previous chapter. However, this is easier said than done. Many managers are reluctant to spend time or money on potential problems, since they have plenty of other problems already. Proactive management of risks is a style that is often foreign to many Western managers. Companies fail at risk management in general because they fail at one of the two fundamentals of managing risk well: cross-functionality on the one hand and being proactive on the other hand. Companies make the mistake to believe that innovation is only in Research & Development, and that is where most of the risk lies. A lot of companies lack the cross-functionality capability between departments nowadays. Companies fail at pro-activeness because they wait until late in the project when risks start occurring and they let risk management lapse and as a consequence they perform in a reactive manner. So in concrete by proactive we mean for example a proactive vulnerability analysis or a proactive mitigation strategy or Proactive risk management of supply risk must be seen as a competitive advantage for the firm. And what is reactive risk strategy then? Reactive management is the reverse of proactive management and is the same as firefighting after a disruption. It means that a company waits until situations are almost hopeless and then saving them miraculously or takes the needed actions to counter the risks after the damage is observed. Thereafter the company develops prevention and contingency plans to make the disruptions less severe in the future if those same disruptions or failures occur (Wagner & Bode, 2009). “Being reactive is the default position when a risk materializes” (Dani, 2009). So instead of dealing with risks on a day-to-day basis, responding reactive is in fact being ready when disruptions occur. Activities such as contingency plans that states corrective actions to resolve undesirable consequences or the accumulation of knowledge and experience that can be used when the same problems occur in the future are two example of risk management from a more reactive approach (Ritchie & Brindley, 2009). An example of a more reactive approach is the concept of Supply Chain Event Management (SCEM) (Gaonkar & Viswanadham, 2004). Although this offering doesn’t use a suitable risk framework, it includes supply chain visibility, track and trace and alert messaging which merely address human operator problems and leave him as the only person to resolve the issue. One can consider the distinction between Outcome or Buffer oriented strategies and Behavior based strategies stated earlier already as an indicator in choosing for a more proactive versus reactive risk strategy. Because when you do for example supplier development or supplier certification you proactively manage your risks with your suppliers instead of putting buffers in a reactive manner. The latter is according to Zsidisin and Ellram (2003) not a manner of risk reduction but is a common approach for every company. Honda focuses on continuous process improvement of suppliers
  • 48. 38 through evaluation in a lot of areas, certification and development (Nelson et al. 1998). Their emphasis on process improvement has had enormous results on reducing the probability of supply risk occurrence. The Honda example supports the finding that improving supplier process is a proactive way to manage supply risk (Zsidisin & Ellram, 2003). Also the distinction according to Sheffi and Rice Jr (2005), also early stated the difference between redundancy and flexibility activities that can be projected to the difference between proactive versus reactive strategies. While some redundancy is part of every resiliency strategy, it represents purely cost with limited benefit unless it is needed after a disruption. In that regard adding redundancy can be more seen of a reactive risk approach. Flexibility on the other hand can be justified for a normal business approach without even taking into account the benefits of risk mitigation strategies and cost reduction but results from increasing flexibility are difficult to measure from a cost accounting perspective. Wieland and Wallenburg (2012) suggest the distinction of Supply chain management practices in the following manner. Risk management can be done proactively through supporting the robustness of the supply chain and reactively on the other hand by supporting the agility of the supply chain. Robustness was explained above. Agility means “the ability of a supply chain to rapidly respond to change by adapting its initial stable configuration” (Wieland & Wallenburg, 2012). Instead of being robust, the initial state is adapted which refers to more reactive strategies whereas robustness refers more to more proactive because of the status-quo policy measures taken in advance. Resilience previously explained means then the ability of adaption but with the aim to gain a competitive advantage after the disruption occurs. This can be seen as the cherry on the cake for a firm that is able to whatever disruption occurs, it can make a positive consequence for the firm with it. It was found that being agile has a strong positive effect on the supply chain’s customer value but not directly on business performance while robustness has a strong positive effect on both the business performance and Supply Chain’s customer value (Wieland & Wallenburg, 2012). Case studies proved that robustness is important to handle supplier-side risks while agility is necessary to deal with customer-side risk and therefore the amount of both performances should fit to the firm’s competitive strategy. Jüttner et al. (2003) conducted interviews and noticed that “Still, whilst the drivers are recognized as competitive pressures with risk implications, it appears that the implications are often sorted out on an ad hoc basis as organizations go along. What seems to be missing is a more proactive approach where risk implications are anticipated at an earlier stage.’” The aim in this work is to wonder
  • 49. 39 whether companies are more using proactive risk strategies or industries tend to more respond and recover from disruptions. We will investigate the relationship between risk perceptions and risk management. Ellis et al. (2010) draws a similar relationship in their Theoretical Model of Risky Decision Making but instead focus on the search of alternative sources as management action. Our focus is on the proactive and reactive risk management strategies. 3.4 Moderators and Mediators As moderators for our model we selected two major broader categories that deal with supply chain management namely practices in supply chain management and supply chain complexity and design. But in order to describe these categories we first draw some general considerations in advance. First, not every company is equally enthusiastic in sharing their supply chain management practices or aligning their practices with their major supply chain partners because of confidentially in the company or risk that the supplier will gain confidential information and the risk of abuse of knowledge. Second, in several countries, regulatory or legal authorities can put constraints on both of these moderators. They can contain supply chain-relevant (trade and transportation) laws and policies. This can be an important factor of uncertainty to the setup and operation of supply chains. Several laws and policies include the ability to obtain approvals necessary for supply chain design activities and supply chain operation practices (Wagner & Bode, 2009). 3.4.1 Supply chain management practices. Supply chain management practices are for example sharing information with your suppliers or forming collaborative approaches with the suppliers. Also the practice of system coupling can be seen as a management practice for the supply chain. The degree of a firm’s system coupling takes into account both the intensity and extent to which information about demand, capacity, inventory and scheduling is shared and used by the firm in both directions of the supply chain (Barut, Faisst & Kanet, 2002). These are not necessarily the same as risk management practices but can be part of a proactive or reactive risk management strategy. As explained above, when managers plan to proactive mitigate risks, choosing different suppliers or sharing risk information with suppliers, can be part of a risk management strategy. So, we can say that a condition to perform business will be an effective supply
  • 50. 40 chain and their management. But also performing risk management can be added as an extra value for the firm. To achieve the company’s business objectives, firms require effective use of integrative supply chain management practices. These can include information, knowledge and design integration. On the one hand sharing information between supply chain partners is needed to jointly control risks (as seen above). But on the other hand it can decrease competitiveness. There can be a leak of knowledge and a diminishing control about their inflows and outflows (Narasimhan & Talluri, 2009). Integration with suppliers but also the internal integration and integration with customers prove to be important drivers for a company’s performance in case of customer satisfaction & competitive performance (Zhao, Huo, Sun, & Zhao, 2013). Furthermore Zhao et al. (2013) found out that supply chain risks are negatively related to supply chain integration. We will investigate a similar relationship namely the effect the firm’s perception of risk probability and impact has on the choice of conducting several supply chain management practices. In fact we examine whether firms who conduct a lot of supply chain management practices will choose more for risk management strategies and more proactive or more reactive kind of strategies. Hypothesis 3a: General SCM practices mediator between environment and risk perception H3a: the relationship of the strength of the environment and the strength of risk perceptions is increased by the mediator variable general supply chain management practices. Hypothesis 3b: General SCM practices mediator between risk perception and the risk management strategies H3b: the implementation of general supply chain management practices increases the relationship between the risk perception indicators and the risk management strategies. Ellegaard (2008) conduct a study in the German automotive industry and found that companies with high supply chain performance show a higher degree of supply chain risk management but the relationship can also be interpreted the other way around. Moreover it found a difference between companies using preventive risk mitigation instruments contrary to those using reactive instruments. Using the supply chain vulnerability map based on (Sheffi, 2005) as previously explained, Blos et al. (2009) identified vulnerable impact rates through the four vulnerability quadrants between
  • 51. 41 automotive and electronic companies in Brazil. The automotive industry’s highest problem is the complex production forecast but they facilitate it through modular production. Electronic companies on the other hand are more vulnerable because of the high dependence on the Asian market and the nature of their products. In the light of the growing trend to act more environmentally friendly and more focus on the durability of company performances, we will also look at sustainable practices employed through the supply chain. For example, managers can perform sustainability performance assessments with their suppliers or they can train their supplier’s personnel in sustainability issues. Thus we will add the following hypotheses. Hypothesis 3c: Sustainable SCM practices mediator between environment and risk perception H3c: Sustainable SCM practices do change the relationship between environmental factors and the perceptions of risk. Hypothesis 3d: Sustainable SCM practices mediator between risk perception and risk management strategies H3d: there is an effect on the relationship between risk perceptions and management strategies when taking the sustainable supply chain management practices into account. 3.4.2 Design and complexity of the supply chain The complexity of a supply chain is a function of their different levels and tiers in the supply chain. Tightly coupled supply chains are furthermore of uttermost importance because the chain is characterized by their components wherefore there are few possible substitutions available. According to Christopher and Lee, 2001 such tighter coupled supply chains are likely more prone to disruptions. The traditional approach, first introduced by Porter, is the supply chain in its literal meaning, namely one supplier delivers parts for your own business and you produce your products that will be distributed towards or directly delivered to your customer(s) that supports a value- adding creation towards the whole chain. Figure 13: Supply Chain Supplier Manufacturer Distributor Customer
  • 52. 42 Today, this picture seems somewhat too simplified and today’s business activities are more integrated resulting in a supplier network. Manufacturers appeal on many suppliers and deliver to more than one customer (see figure below) resulting in the term of supply network management (SNM). Figure 14: Supply Network Furthermore companies perform outsourcing activities when realize that it’s cheaper to produce some parts or perform administrative activities elsewhere than in-house. If after some period the knowledge of performing these activities is known and the part production has come down to a fine art it can be advantageous from a cost perspective to do an in sourcing strategy. Furthermore companies don’t live alone anymore. Be aware that off shoring activities can increase lead times (Khan, Christopher, & Burnes, 2009). At last, the creation of joint ventures, alliance partners and more make the design of supply chain not less easy. Consequently supply chain networks are often Figure 15: a more complex and adaptive network
  • 53. 43 complex and adaptive but also very dynamic. They are composed of intricate, sometimes counterintuitive and nonlinear linking elements that include a lot of information and constraints on each other under which transactions, activities and processes take place (Manuj & Mentzer, 2008). Although the concept of value network has grown in past years, from a risk perspective managers should be thinking in terms of networks that create value. Value is transferred from several actors in a value network. These values which are often not in monetary terms translated can be traded for other values with other partners. Impact flows constitute the intangible components (information, knowledge …) shared through partners. Hallikas and Varis (2009) make the comparison with a business ecosystem theory as a conceptualization, meaning “The environment beyond the core business & the value network and can be seen as interdependencies between several business actors and the business environment in which they operate.” You should thus determine the dependencies that are most critical to your business and select your business in your environment that are important enough to make collaborations with in the future. Several business aspects can be interpreted from a biological view: Survival from environmental pollution Being head-to-head with technological change Health of the flora Being able to produce value and share it between members of the system. Receive enough air and sunlight A company should position itself in the network where it can most generate value now and into the future Ecosystem health regulator Powerful business leaders who create platforms for other industry players and link other companies Photosynthesis Business Lifecycle Dependency from other organisms in nature Partnership and complex interplay between competitive and cooperative business strategy Table 3: the business from a biological view
  • 54. 44 In figure 16 you find an example of a value network, here for Google. Hallikas and Varis (2009) have determined the values and impacts with their resulting possible risk through major players in the Google network. The different players in the network represent nodes where each node represents a value-adding activity for another node in exchange for some value for the other party. Risk can occur because partners often transfer unwanted knowledge in close cooperation with each other. Furthermore the risks in the network are dependent on the behavior of the interconnected systems of companies. Global supply chains drive up the level of complexity, which drive up the level of risks. Because of an increasing dynamic complexity and feedback mechanisms these networks have to be aware of risk and consequently can think of the use of risk management. Giving the importance of complexity, as you can read a bit further, we also included this facet as a moderator in our network. It is also important to recognize the indirect relationship because of their increased invisibility these may pose a higher risk for the head company (Hallikas & Varis, 2009). The intangible nature of products in the exchanges of the value network increases the complexity of the supply chain network (Hallikas & Varis, 2009). Less complex supply chains are better able to manage supply risk and can improve performance. Some complexity reduction measures are reducing the number of suppliers and serving varied customers through one integrated supply chain (Hallikas et al., 2004). Kleindorfer and Saad (2005) addresses the effect of alternative supply chain design options on the efficiency of the supply chain to various sources of disruption but this research was not yet related to the use of risk management strategies. Figure 16: supply information network
  • 55. 45 Relying on a single supply source may be risky nowadays. Think of Ericson who has lost 400 million of damages from a fire at Philips, its single supplier (Norrman & Jansson, 2004). So Autry and Sanders (2009) suggests 3 flexible multiple supply source options that are used by the firms to mitigate their risk. One possibility is to develop multiple suppliers each with a certain advantage for the same component. For example a company can rely on a flexible supplier that delivers products at any time for one half and rely for the other part on a cost-efficient supplier. When there is more demand for any given feature (cost versus delivery) the company can rely on the spot market to make the difference. Secondly, once a company uses a sourcing strategy then it’s optimal to use partly local supply sources to supplement their supply base. This can serve as a back-up for the company. At last it can be optimal to rely on suppliers with multiple manufacturing sites at multiple locations, rather than just one site so that supply can be easily shifted from one site to another. Although the majority of authors suggest multiple sourcing against risk, Blome and Henke (2009) tempt to direct the discussion in favor of single sourcing. The major motivations therefore are the cost reductions through e.g. standardization and transaction. Second, dependency on one single supplier doesn’t calculate in higher risk because highly cooperation can provide strategic and competitive advantage for buyer and supplier. The condition therefore is that they are mutually dependent on each other because when there is a one-sided dependency from only one party the relationship becomes more risky. Therefore proactive risk management must also play a role in single sourcing situations. Finally the decision of the number of suppliers should be made separately from the choice of supplier relationship, cooperative versus transactional, according to Blome and Henke (2009). The case of Marks and Spencer has shown that relying on few amounts of close suppliers and implementing direct sourcing instead of cooperating with a large number of independent suppliers, they could proactively mitigate and manage their risks to its supply chain in a better way (Khan et al., 2009). In a later stage they continued to extend their supply base through suppliers that were engaged throughout the product development process instead of addressing to suppliers who take a high share of their margins that M&S had in the past. We will examine the following hypothesizes. Hypothesis 4: Supply chain complexity and design moderator H4a: there is a difference in relationship between the environment indicators and risk perceptions strategies between firms whose amount of complexity and design of the supply chain is higher and those firms who have a lower amount of supply chain complexity and have a lower level supply chain design.
  • 56. 46 H4b: there is a significant difference between high-level designed and more complex supply chain or networks and low-level designed and less complex supply chains for the relationship between risk perceptions and risk management strategies. 3.4.3 Continental differences One can consider if there might be differences between continents. There are certainly differences in the way Asian countries operate their business as compared to European countries. But reflect these differences also in terms of risk management and risk perception. In a later section we will investigate for any differences. Hypothesis 5: Europe versus Asia moderator H5a: there is a difference in the relationships for environment and risk perception between the continents Europe and Asia H5b: there is a significant difference between Europe and Asia for the relationship between risk perceptions and risk management. 3.4.4 Comparison with 3 years ago A lot can change in 3 years, certainly in these sometimes turbulent times. Therefore we will analyze if there are significant differences in the last 3 year as compared to the current implementation. Hypothesis 6: Current implementation versus effort in the last 3 years moderator H6a: there is no significant difference between the policy of 3 years ago and the current implementation for the risk management programs and supply chain management practices that were related with risk perceptions for both Europe and Asia. H6b: there is a difference for the supply chain management practices, general and sustainable, and risk management strategies, proactive and reactive, that were taken 3 years ago and their current implementation related to the risk perception of the firm for both complex as simple network designs.
  • 57. 47 3.5 Model diagram Figure 17: Model diagram An overview of the variables to investigate and their relationships is given above. The general framework is composed of only the general “blue” variables. In a later stage we will examine the intervention of the “green” mediator variables, the SCM practices. In the end we explore the “red” moderator variables.
  • 58. 48 4 Data Collection Now that the conceptual model is extensively explained, it is time to gather some data. Data is needed to test and investigate the relationships between the several constructs in the framework. In the introduction, I will explain the steps and effort needed that has been done to collect data. Thereafter some exploratory factor analysis will be done to validate the several items that form the constructs. 4.1 Introduction To investigate a conceptual framework, there needs to be data and, more preferable, a lot of data. So to get this data, I have participated with the International Manufacturing Strategy Survey or IMSS. It is a research project carried out every 4 years by a global network. This project studies production and supply chain strategies within the manufacturing and assembly industry through a detailed questionnaire administered simultaneously in many countries by local research groups. In order to get worldwide data from this research project I’ve collected industry data for Belgium. This was made possible in cooperation with the Vlerick Business School. Through telephone contact with several companies I’ve tried to get the goodwill from manufacturing managers or plant managers to participate with this research project and their commitment to fill in the detailed questionnaire. The questionnaire used can be found in the appendix. The purpose was to address manufacturing companies from the metallic, electronic and automobile sector. More concrete, the following manufacturing industries have been addressed along with their associated industry codes: - 25: fabricated metal products, except machinery and equipment - 26: computer, electronic and optical products - 27: electrical equipment - 28: machinery and equipment not elsewhere classified - 29: motor vehicles, trailers and semi-trailers - 30: other transport equipment To find the Belgian companies from the appropriate industries I made use of the Amadeus Export Data Belgium Database. I also approached the companies who have participated the previous time for this research project but some companies didn’t exist anymore and other companies had changed their manufacturing managers compared to the past. To get the company info and the managers contact information I consult the Trends Top database. Furthermore I provide an online possibility to fill in the questionnaire through SurveyMonkey so that the managers don’t have to print
  • 59. 49 out the questionnaire format before filling them manually but instead could fill them online whenever they will and further continue when they have some time. 4.2 Results For Belgium I could approach 30 companies in total from the 6 sectors stated above. There are 136 operations mangers from these companies contacted. 24 of them weren’t eligible for data incorporation because the company was going to shut down in less than half a year or the company didn’t fulfill a manufacturing activity or only performs a distribution function. So 30 out of (136 -24 =) 112 managers participated which results in a response rate of about 26, 79%. There ain’t no such thing as a free lunch. After sending the Belgium questionnaires I get the worldwide data from the IMSS. In total 19 different countries have participated through this large- scale research. Table 4.2 gives you an idea of the Industry Code distribution for Belgium and for the whole world. Table 4.3 gives the participated country distribution. Figure 18: proportion of participated industries Figure 19: proportion of participated countries
  • 60. 50 4.3 Variables and constructs For the construct of environment we include the following variables: PORTER PEST Market Competitive rivalry (A2e) Threat of substitution (A2g) Bargaining power of suppliers (A2h) Bargaining power of customers (A2i) Social pressure (A2k) Rate of technological change (A2b) Environmental pressure (A2j) Demand fluctuation (SC3a) Market size (A2a) Market span (A2c) Market concentration (A2d) Market entry (A2f) Table 4: Environmental constructs The first five variables can be derived directly from the Porters five forces model that we described before. As regard the PEST analysis we have items about social force, technological force and environmental pressure. In addition we included some market variables for the respective industries. Finally we can pick demand fluctuation as additional variables to complete the construct. As described before, for probability and impact we have surveyed the following variables: Probability Impact Probability of a key supplier failure (R1a1) Probability of operations failure (R1b1) Probability of shipment failure (R1c1) Impact of a key supplier failure (R1a2) Impact of operations failure (R1b2) Impact of shipment failure (R1c2) Table 5: Risk perception variables For the risk action programs, we have earlier made a distinction between more proactive as well as more reactive strategies. So we made a distinction between these strategies towards the following mitigation strategies: Proactive Risk Management Reactive Risk Management Preventing operations risk (select a more reliable supplier, use clear safety procedures, preventive maintenance) (R2a2) Detecting operations risk (internal or supplier monitoring, inspection, tracking) (R2b2) Responding to operations risk (backup suppliers, extra capacity, alternative transportation modes) (R2c2) Recovering from operations risk (task forces, contingency plans, clear responsibility) (R2d2) Table 6: Proactive and Reactive Management When investigating the questionnaire we could pick several supply chain management practices surveyed. These can be divided in 2 groups namely practices towards general external action
  • 61. 51 programs as well as practices in relation with sustainability management. The former group analyzes both these practices for a company’s key suppliers and customers. We will include these variables as a mediator in our model and see if these variables as a whole strengthen the relationship between Risk Perceptions and the choice of Proactive versus Reactive management. General external action programs Information sharing: (about sales forecast, production plans, order tracking and tracing, delivery status, stock level) (SC6a2 & SC6f2). Development of collaborative approaches: (supplier development, risk/revenue sharing, long-term agreements) (SC6b2 & SC6g2). Joint decision making: (product design/modifications, process design/modifications, quality improvement and cost control) (SC6c2 & SC6h2). System coupling: (vendor managed inventory, just-in-time, Kanban, continuous replenishment) (SC6d2 & SC6i2). Developing an international sourcing strategy: (supplier scouting at the international level, open foreign sales office, develop an international purchasing office or distribution network) (SC6e2 & SC6j2). In relation with sustainability management Suppliers’ sustainability performance assessment: through formal evaluation, monitoring and auditing using established guidelines and procedures (SMh2). Training or education for suppliers’ personnel (SMi2). Joint effort with suppliers to improve their sustainability performance (SMj2). Table 7: Supply Chain Management Practices For supply chain complexity and design we will include the following variables in table 12. We will test if there is a difference between higher complex supply chains or networks and less complex networks towards the adaptation of more proactive risk management strategies. This construct will be applied as a moderator in our model. We can roughly divide our variables in 2 broader categories namely role of the plant and variables in relation to the manufacturing network but we will not use this distinction in our model. Role of the plant Producing in one plant versus producing at multiple plants in the network (G2a). The role of your plant in the network is stable versus the role of your plant in the network is revised and changed flexibly if needed (G2d). To what extent is the plant responsible for the supply chain (procurement, logistics, supplier
  • 62. 52 development) (G3b). The degree of integrated information system in the company wide network (G4c). Manufacturing network Improve the use of technology to support communication with other plants of the network (e.g. ERP integration, shared databases, social networks) (G7d2). Developing a comprehensive network performance management system (based on cost, quality, speed, flexibility, innovation, service level) (G7e2) Change the balance between outsourcing and in-sourcing (PC1b). Table 8: Supply Chain design and complexity As a last we will investigate if there is a moderating effect between European countries and more Asian countries in our model. Next we have grouped items into several groups in the constructs. 4.4 Exploratory factor analysis (EFA) The measurements of the variables were done by using a 5 point Likert-type scale where 1 means low and 5 means high. Before moving to our methodology we first have conducted an exploratory analysis with the use of the Cronbach’s alpha measure of the item factors that completes the constructs as we can see in Appendix 3. For environment we have dropped the last variable “demand fluctuations” since when leaving this one our consistency scale measure increased so in fact the construct will receive a higher alpha when dropping this variable. The same has happen for the construct of Supply chain complexity and design where we deleted the variables “the extent of the responsibility of the plant” and “the changing balance between out-sourcing and in-sourcing”.
  • 63. 53 5 Methodology and Analysis 5.1 General model Figure 20: General Model First we will investigate the general model in too deep. The general model is the model without mediators or moderators and can be seen in figure 20, here above. In a later stage the intervention of mediators and moderators will come into play. 5.1.1 Confirmatory Factor Analysis (CFA) of the general model Here we have a reflective measurement scale because the some indicators that form a construct are correlated and interchangeable. To do a CFA we applied the PLS algorithm with the use of smartPLS. Explanation of target endogenous variable variance  The indicators Risk probability and Risk impact explains 53.6 % for Proactive risk management (R² = 0.536) and 52, 4 % for Reactive risk management (R² = 0.524).  The risk perception variables probability and impact are explained for as little as 8.5% and 7.5% by the environment constructs PORTER, MARKET and PEST.
  • 64. 54 Inner model path coefficient sizes and significance  When looking at the inner model, we can conclude that the effects of Risk impact (0.614 & 0.604) on the both proactive and reactive strategies will be stronger than the effect of risk probability (0.138 & 0.140) on them.  All of the indicators of risk perception will be statistically significant on the risk management strategies.  The market variable will as only environment indicator have a statistically significant relationship towards both perception and impact. The other environment variables PEST and PORTER have a factor loading of less than 0.1 and we predict that those variables will not be significant. Latent Variable Indicators Outher model Loadings Indicator Reliability Cronbach's alpha AVE Market MarketConcentration 0,7323 0,53626329 0,6552 0,4903 MarketEntry 0,7395 0,54686025 MarketSize 0,6516 0,42458256 MarketSpan 0,6733 0,45333289 PEST EnvirPressure 0,7782 0,60559524 0,6863 0,6006 SocPressure 0,733 0,537289 TechnologicalChange 0,8118 0,65901924 PORTER BargainingPowerCust 0,6644 0,44142736 0,6433 0,4713 BargainingPowerSuppl 0,6421 0,41229241 CompetitiveRivalry 0,6793 0,46144849 ThreatSubstitution 0,755 0,570025 RiskImpact Imp_Operations 0,9783 0,95707089 0,9632 0,9316 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9659 0,9362 Prob_Shipment 0,964 0,929296 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9796 0,95961616 0,9593 0,9609 Detecting 0,9809 0,96216481 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9489 0,9513 Responding 0,9764 0,95335696 Table 9: CFA summary table Indicator reliability Indicator reliability shows how reliable the items are that forms the constructs per item. To come to the reliability indicators, we have to quadrate all the outer model loadings. All the indicators show a reliability that is higher than 0.4 which is the minimum level required for exploratory research.
  • 65. 55 However, a value that is more than 0.7 is preferred. This is the case for the Risk perception constructs and risk management constructs which show levels that are higher than 0.9. Internal consistency reliability The measurement of the alpha in a Cronbach’s alpha analysis calculates the reliability of the construct as a whole. This analysis is based on the foundation that every item in a scale must sufficiently be correlated with any other item from the same scale. For this measurement we use the Cronbach’s alpha measurement indicator. It is recommended for your scale internal consistency that Cronbach’s alpha should have a value of 0.6 à 0.7 as a lower limit. We see that this is the case. However we see that this measure is higher with the risk perception and risk management constructs than those for the environment. Convergent validity Convergent validity means when two or more different methods that measure the same concept correlate high. To check for convergent validation we have to look for the AVE measure and we have to assure that this above 0.5. As we look to our latent variables we see that it’s more or less the case. When we round up the variables MARKET and PORTER, we see that these have a value of 0.5. Discriminant validity MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability MARKET 0,700214 PEST 0,5927 0,7749839 PORTER 0,561 0,5419 0,68651 ProactiveRiskMgmt 0,2867 0,222 0,1657 0,9802551 ReactiveRiskMgmt 0,2926 0,2286 0,1639 0,9588 0,975346 RiskImpact 0,2707 0,1828 0,1814 0,7277 0,7195 0,9652 RiskProbability 0,2908 0,1713 0,1829 0,6439 0,6376 0,8235 0,9675743 Table 10: Discriminant validiy Discriminant validation means the opposite of convergent validity and exists when two or more constructs, measured through one or more methods, correlates low and consequently differ sufficiently from each other. To check for discriminant validity we must assure that the square root of the AVE value of each latent variable, in the table showed in bold, must be higher of the correlation of the latent variable with the other variables. In practice the figure in bold must be higher than the values in the according row and column. As we check this for all our variables we see that this is the case for all our construct
  • 66. 56 variables. For example the AVE value of Proactive Risk Management is found to be 0.9609 so its squared root is 0.9803 which is higher than all the values of its row and its column. Conclusion To conclude our confirmatory factor analysis we can say the following. The reliability measures (indicator reliability and Cronbach’s alpha) are both good. For validity measures we say that the environment variables Market and PORTER are somewhat weak for the AVE measure but the discriminant validity of the constructs is ok. 5.1.2 Comparison with alternative frameworks In general, we could have opted to include market entry also in our PORTER construct and include a possible economic indicator in the PEST construct (for example MarketSize) but this threatens the discriminant validity. Indeed, Appendix 5 shows that the discriminant validity is threatened. This is because the same item is used for more than one construct. Furthermore the AVE variable is bad for the PORTER construct. At last several indicator reliabilities are also less than 0.4 which is too low. What reliability and validity should we have if we put all the environmental factors together in one construct? Appendix 5 provides some figures. A lot of environmental reliability indicators are less than 0.4 and even less than 0.3 but the Cronbach’s alpha for environment is actually good in concerning the reliability indicators. For the validity indicator the same conclusion can be made. Although the AVE is very weak (0.3648) the discriminant validity scores are better, especially for environment. So according this concept, we’re stuck in between: 1 of 2 indicators of reliability and validity performs well but the other bad. The follow figure demonstrates the confirmatory analysis for several constructs. At the bottom we added 3 other frameworks as well. We can conclude that our general will be the most fitted.
  • 67. 57 indicator reliability internal consistency reliability convergent validity discriminant validity general model Good Good Neutral Good market entry in PORTER and market size in PEST and included in market Bad Good Bad Bad all environmental factors Bad Good Bad Good market entry in PORTER and market size in PEST but excluded from MARKET Bad Bad Bad Good market entry in PORTER and market size in MARKET but the latter excluded from PEST Bad Good Bad Bad market size in PEST and market entry in MARKET but the latter excluded from PORTER Good Neutral Neutral Good Table 11: CFA summary table for comparison with other frameworks 5.1.3 Descriptive statistics Rivalry among competitors is seen for most firms as relatively high whereas the threat of substitution is seen as much lower. Furthermore customer bargaining power is much higher than the bargaining power of suppliers which in fact stress the distributor or customer as an important player. From the 3 environmental construct, it appears that Market variables are perceived as the most influenced ones to the firm. The other environmental construct variables are jointly a little less important but still above a neutral perception of 3. For the entire world, it seems that the threat of a key supplier failure is the highest possible (2.74) and the risk of shipment operations are the lowest (2.32). As for probability, the same conclusions can be made for impact. But what is more important is that the impacts for the three categories are much higher. It means that companies deal with medium probability risks (+/-2.5) and higher impact (+/-3.5). It seems that most firms see their risks with a lower probability but with higher impact. It appears that risk management has increased since 3 years for the four indicators. Preventing has the highest score and recovering the lowest both for current implementation and 3 years ago.
  • 68. 58 5.1.4 Hypothesis testing Now we will start to do so hypothesis testing and thus check the structural path significance. Outer model T Statistics BargainingPowerCust <- PORTER 2,9481 BargainingPowerSuppl <- PORTER 3,6311 CompetitiveRivalry <- PORTER 3,5914 ThreatSubstitution <- PORTER 4,4393 MarketConcentration <- Market 10,3393 MarketEntry <- Market 10,1612 MarketSize <- Market 4,9895 MarketSpan <- Market 6,8934 EnvirPressure <- PEST 4,4864 SocPressure <- PEST 4,2431 TechnologicalChange <- PEST 5,1977 Imp_Operations <- RiskImpact 261,8998 Imp_Shipment <- RiskImpact 100,6175 Imp_Supplier <- RiskImpact 91,4657 Prob_Operations <- RiskProbability 254,047 Prob_Shipment <- RiskProbability 101,3114 Prob_Supplier <- RiskProbability 104,3062 Recovering <- ReactiveRiskMgmt 130,6781 Responding <- ReactiveRiskMgmt 157,9065 Detecting <- ProactiveRiskMgmt 149,0042 Preventing <- ProactiveRiskMgmt 180,756 Table 12: outer model T-statistics All of the T-statistics are larger than 1.96 so we can conclude that the outer model loadings are highly significant. Inner model Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,2371 0,0629 3,7657 Market -> RiskProbability 0,2763 0,066 4,2393 PEST -> RiskImpact 0,0227 0,0444 0,5158 PEST -> RiskProbability -0,0102 0,0441 0,2755 PORTER -> RiskImpact 0,0417 0,0489 0,7381 PORTER -> RiskProbability 0,0401 0,0482 0,6756 RiskImpact -> ProactiveRiskMgmt 0,6023 0,0843 7,2829 RiskImpact -> ReactiveRiskMgmt 0,593 0,0846 7,146 RiskProbability -> ProactiveRiskMgmt 0,1464 0,0699 1,9812 RiskProbability -> ReactiveRiskMgmt 0,1477 0,0709 1,9755 Table 13: inner model loadings and T-statistics
  • 69. 59 Market variables are significant with a significance level of 5%. The other environmental variables do not have an effect on the perception of impact as well as probability. This supports Hypothesis 1 for market variables but not for PORTER and PEST variables. Risk impact perception is strongly significant on the 5% significance level towards risk management strategies. On the other hand, risk probability is just not significant on the 5% significance level but certainly on the 10 % significance level. This supports the hypothesis 2a for risk impact on both risk management strategies but not really supports the 2b hypothesis. 5.1.5 Multicollinearity Multicollinearity in fact occurs when two predictor variables are highly correlated. The threat of this phenomenon is that coefficient estimates can change drastically in response to small changes in the data. To test the multicollinearity, statisticians often apply the tolerance or Variance Inflation Factor (VIF) values. The tolerance factor must be higher than 0.2 and the corresponding VIF factor higher lower than 5. We see that this is the case for all of our constructs. Coefficients a Model Unstandardized Coefficients Standardized Coefficients t Sig. Collinearity Statistics B Std. Error Beta Tolerance VIF 1 (Constant) -7,45E-06 0,023 0 1 Market 0,072 0,041 0,072 1,755 0,08 0,322 3,104 PEST 0,071 0,033 0,071 2,129 0,034 0,487 2,055 PORTER -0,048 0,034 -0,048 -1,389 0,165 0,459 2,181 RiskImpact 0,603 0,041 0,603 14,659 0 0,32 3,122 RiskProbability 0,123 0,041 0,123 2,966 0,003 0,317 3,159 Table 14: Multicollinearity
  • 70. 60 5.2 Input from mediator variables 5.2.1 Framework Figure 21: framework with general SCM mediator between environment and risk perception (first framework) In the following section we will analyze the effect of supply chain management practices, both the general ones and also the efforts towards more sustainability. Because the degree of supply chain management practices might enforce the relationship between environmental factors and risk perception but might also strengthen the relationship on this perceptions towards risk management strategies, we will investigate both mediating effects. The framework where general SCM practices are mediated between environment and risk perception is given in figure 21. The framework for the general SCM practices mediated between risk probability/impact and proactive/reactive risk management is given in figure 22.
  • 71. 61 Figure 22: Framework with general SCM mediator between risk perception and risk management (second framework) We see that explanation of target endogenous variable variance is increased both for proactive and risk management when we introduced this mediator to additionally explain the endogenous variable. Secondly, the inner factor loadings for the relationship between risk probability and both management strategies have declined below 0.1. We might expect that this relationship will be gone. In appendix 4 we found both frameworks but instead with the interplay of sustainable SCM practices. The R² of this framework has slightly augmented to 55.1 % for proactive risk management and 54.2 % for reactive risk management when considering sustainable practices between risk perceptions and management. 5.2.2 Confirmatory factor analysis If we take a look at appendix 6, we found suitable indicators for all of the factors of the mediators. The determination coefficient R² has increased from 53.6 % and 52.4% to 61.7 % and 60.9% for proactive and respectively reactive management. We see that for the general SCM there are very high scores for Cronbach’s alpha and for the AVE measure. Moreover, the indicator discriminates good with the other constructs. The same can be said for the sustainable SCM. 5.2.3 Descriptive statistics with mediator variables General supply chain management practices Also for supply chain management practices we see that all the indicators have increased from 3 years ago. Yet we see a different ranking from all the practices. What is striking is that Supplier
  • 72. 62 collaboration and sharing information with the supplier scores 2 times as the highest. It seems that firms highlight the importance of sharing information above performing joint actions. Sustainable supply chain management practices When we analyzed the descriptive statistics we see that managers performs on average more on performance assessment with the suppliers and the least on joint efforts with the suppliers with the aim for more sustainability. What we see is that on average these numbers are somewhat lower than the general SCM practices. To compare with 3 years ago all these indicators have increased. 5.2.4 Structural path significance Because we in advance consider SCM practices might mediate both the environment – risk perception relationship as well as risk perception – risk management relationship, we investigated both frameworks. Appendix 6 shows the T-statistics for the inner loadings for the first framework. We consider both the general SCM as well as the sustainable SCM. The outer loadings were all significant and are not retaken here. No differences occur between general and sustainable SCM. PORTER variables stay not significant and all market variables stay significant. In addition PEST variables seem to have now a significant relationship with SCM practices for both mediators and a significant relationship with probability for the sustainable SCM mediator. What is striking is whereas risk probability towards risk management was here above yet significant on the 5 % level for the general model we see that this relationship has been softened. Also for the second framework we can draw some pretty similar conclusions. The outer loadings are once again not retaken but they are all significant. We see that the relationship between risk perceptions and SCM practices as well as their relationship with risk management strategies are all significant. What is more remarkable here is that the significance of the relationship between risk impact and SCM practices differ from general and sustainable SCM practices. It seems like the mediator has some effect on the relationship between these two variables. But can we talk of a mediating effect between our constructs? 5.2.5 Mediating effect and hypothesis testing To test for the mediating effect we can make use of the Sobel test. It tests whether a variable mediates the significant relationship between 2 other variables. To test for this phenomenon we apply his following formula: Formula 1: Sobel test statistic
  • 73. 63 It’s a test statistic that decides whether the variables significant mediates i.e. strengthen or soften the relationship between 2 other variables. Therefore we need some indicators. First we need for the 2 indicators the β’s from the 2 independent relationships, namely a and b, and their standard errors, SE_a and SE_b. For example, if we want to calculate the Sobel test statistic for the general SCM mediator between Market and risk probability for the first framework we must compute their β’s and standard errors. We find this numbers in appendix 7. For a = 0.1848, b =0.4971, SE_a = 0.0599 and SE_b =0.057 we found a Test statistic z = 2.9085 which is significant on the 5% significance level and therefore there is a mediating effect for this relationship and the mediator decreases the dependency from 0.28 to 0.187. So we will now examine if the 3th hypothesis is fulfilled. If we look at appendix 7, we can draw conclusions for both mediators for the first framework. Except for the relationships with the PORTER variables, the mediating effect is significant for the environmental variables on risk perceptions. The market variable is partial mediated because this relationship remains significant. With regard to the PEST variable, the choice of SCM practice is not really clear: The relationship is partial mediated for the general SCM mediator but not for the sustainable SCM mediator. Risk probability might be fully mediated by sustainable SCM practices in relation with the PEST variables. This can be seen as follow: managers perceives the environmental factors (e.g. technological change, social pressure, environmental pressure) and act with its suppliers in a sustainable manner with in turn can adjust their probability of risk but apparently not for the impacts of risk. What about the second framework, we can conclude the following: There is a difference between the effects of the mediating factors on the variables. In fact, when we only consider the 3 sustainable SCM variables, the relationship between risk impact and risk management is not mediated. Second, the relationships are all weakened with the interplay of the mediators except for the impact of risk on pro- and reactive risk management. Apparently managers take their degree of general supply chain management practices into account when performing a particular level of strategy. Figure 23: Mediating effect
  • 74. 64 5.3 Input from moderator variables 5.3.1 Descriptive statistics and differences for the moderator variables Europe and Asia Figure 24: Environmental factors between Europe and Asia Figure 24 shows the difference between European and Asian countries regarding the environmental factors. What we clearly see, is that a lot of environmental factors are perceived as stronger by Asian countries except for customer and perception power. If we look at appendix 8 for paired sample t- statistics, we find more or less the same conclusions. All the environmental variables are significantly different between the continents, except for customer power, which is perceived as lower for Asia, and market entry, which is not perceived as lower for Asia. 3,36 3,33 3,46 3,49 3,86 3,00 2,94 3,11 3,71 3,33 3,24 2,68 3,69 3,62 3,65 3,81 3,90 3,16 3,00 3,23 3,60 3,50 3,36 2,58 3,19 3,20 3,32 3,30 3,83 2,89 2,89 3,02 3,79 3,24 3,16 2,67 market size techn. Change market span market concentration competitive rivalry market entry threat of substitution supplier power customer power env. Pressure social pressure demand fluctuation 0,00 0,50 1,00 1,50 2,00 2,50 3,00 3,50 4,00 4,50 Europe Asia Whole World
  • 75. 65 Figure 25: Risk Probability and Impact for Europe and Asia Here we see that Asia perceives more risk probability than the European companies but instead weight the impact as much stronger than certainly for supplier disruptions. In a later stage we will investigate if this risk perception difference will also have a different relationship towards risk management strategies. For risk probability the indicators are significantly different but for risk impact this counts only for the perception of impact from a supplier default (Appendix 8). Increase / Decrease Risk action programs in the last 3 years (2010 – 2013) Preventing Risk Detecting Risk Responding Risk Recovering Risk Whole World + 3,02% + 5,84% + 6,09% + 6,66% Asia + 11,36% + 15,06% + 12,75% + 14,63% Europe - 2,67% + 0,13% + 1,48% + 1,41% Table 15: Risk management for Europe and Asia Table 15 gives you an idea of the risk management development programs in the last 3 years (from 2010 – 2013). As we can see Asian countries have developed or done more effort the past 3 years in risk action programs despite their already high level in 2010. Europe on the other hand shows a minor increase for detecting and reactive risk management and even a higher decrease for preventing risk. Seems like those countries believe that risk management programs are less fruitful for their operations or they feel themselves in less crisis situations than 3 years ago. Appendix 8 proves that for all the risk management indicators there are significant differences. The highest difference is for responding to risks. At last, for the mediator variables general and sustainable SCM practices, we found in appendix 8 that all the indicators are clearly different. These practices seems to be more implemented by Asian countries. In particular joint decision making with their customers is more performed for Asian
  • 76. 66 countries. What is more appealing is that Asian countries do more effort for sustainable practices, in particular for again joint efforts and also supplier education on durable issues. The same accounts for Supply chain design and complexity. Only integrated information systems seems to have no significant difference between countries. Low versus high complex supply chain designs When we look at appendix 8, we find that nearly all environmental variables are perceived as higher for more complex networks except for customers bargaining power but seems not to be significantly different. Threat substitution is neither significant on the 5% level but appears to be a little higher for more complex networks. Different means occur for all the probability of risk indicators for the complexity moderator. However, it seems that their risk perception of impact differ not so meaningful. Only the risk of shipment disruptions is perceived stronger by higher complex networks. Increase / Decrease Risk action programs in the last 3 years (2010-2013) Preventing Risk Detecting Risk Responding Risk Recovering Risk Low SC complexity -0,51% 1,87% 1,67% 5,58% High SC complexity 2,18% 5,74% 5,90% 7,52% Table 16: Risk management for lower and higher complex networks Furthermore, what about the risk action programs, we see that more complex networks implement more risk action programs than 3 years before. But is this relationship statistically significant different between lower and higher supply chain complexity? If we look at Appendix 8, we can conclude that it certainly is. At last, also general and sustainable SCM practices does in fact differ from their degree of complexity of their supply chains. System coupling appears to have the greatest difference between the two categories. Indeed, lower designed and complex networks don’t always have the availability of loose coupling. Also training for the sustainable SCM is perceived as the greatest difference between the two groups. The lowest difference is for collaborative approaches. It can be seen for firms with a less developed supply chain as another, often cheaper way, to integrate their supply chain.
  • 77. 67 A note for the validity of the Supply chain design and complexity construct and the determination coefficients for the frameworks according to the moderator variables To give an idea of the validity of the construct, appendix 6 provides some figures for the factor analysis. As we can see the indicators satisfied their threshold values and besides discriminates well from the other indicators If we take a look at table 17, we discover that the variability in the model explained by the R² is increased with 5 to 6 % when we only consider European countries. When only modeling Asian countries, the explained variance decreases with7% for proactive risk management but only 3 % for reactive risk management. R² Proactive Reactive Asia 54,50% 58,10% Europe 67,30% 66,10% Low SD&C 34,30% 35,90% High SD&C 6,30% 5,60% Table 17: determination coefficients When considering for the degree of SCD&C, the determination coefficient has drastically decreased. A possible reason for this trend might be the decreased amount for data for this figures. Several data points lack some values for indicators that describes the SC complexity construct. The aim is to investigate this construct rather as a moderator and for the purpose to investigate differences then nearly explain the whole model with these data only. 5.3.2 Moderating effect and moderated mediation Explication for the difference calculation We will test for some moderating variables, which mean if relationships are in fact moderated or different between 2 groups. This is called multi-group moderation and we can see the T-statistics in appendix 8. First our dataset were split into Asian countries and European countries. It seems that for the former group we have 317 observations and for the latter 503.
  • 78. 68 Asia Europe (m-1)^2 99856 Sample Size 317 503 (m+n-2) 818 Sample Mean -0,0215 0,1802 (n-1)^2 252004 Standard Error (S.E.) 0,0821 0,0787 sqrt(1/m+1/n) 0,071712 t-statistic 1,7020 1st half denominator 0,822824 p-value (2-tailed) 0,0891 2nd half denominator 1,908111 sqrt(1st half + 2nd half) 1,652554 Full denominator 0,118508 numerator 0,2017 Figure 26: calculation for the multi-group moderation t-statistic and p-value Formula 2: t-test for multi-group moderation and moderated mediation If we want to calculate for example the significant T-statistic for differences between Europe and Asia for the relationship between the probability of risk and proactive risk management we must use the sample means and standard errors. These can be found in SmartPLS after bootstrapping. To calculate the t-statistic we must fill in the formula given above. Therefore it is easy to split the calculations a bit like as done above. For the moderator supply chain design and complexity the dataset is again split in two. The first group consists of the observations which have an average of the 5 indicators that forms this construct of less than 3. The second group those that are more or equal than 3. The choice of a threshold of 3 was decided because the earlier threshold of 2.5 generates less than 100 observations for the first group. In order to conduct these paired t-tests with some validity, we opt to choose for groups for at least 100 observations. So these t-statistics are calculated in the same way. We can also determine the moderating effect through relationships with mediators or otherwise called the moderated mediation. These can be seen as the total effects between variables. The
  • 79. 69 separate effects which are included under multi-group moderation are not influenced by the SCM practices mediator. The total effects which were the same as their separate effects were not retaken in the moderated mediation table. In the next two sections we will try to see if we can support hypothesis 5 and hypothesis 6. General SCM mediator Risk probability – risk management difference between Europe and Asia There is a significant difference between Europe and Asia in the relationship between Risk Probability and Risk management strategies on the 10% significance level. Europe has a significant relationship between them and Asia totally not. The difference between Asia and Europe is more significant for reactive (1.869) than for proactive risk management strategies (1.702). Risk Probability –risk management & SCM for lower and higher complex supply chain networks There is a significant difference between high-level complex supply chains and low-level supply chains in the relationship between risk probability and risk management strategies as well as supply chain management practices on the 10% significance level. Lower level complex supply chains have a stronger relationship for the probability of risk towards both proactive as well as reactive risk management than higher-level complex networks. The difference between Low and High complex networks is more significant for proactive (1.812) than for reactive (1.702) risk management. The relationship between risk probability and supply chain management practices is significant for high-level supply chains but not for low level supply chains and the difference is also significant. Sustainable SCM mediator SCM – risk impact between lower and higher complex networks in the first framework What is surprising when we analyze the T-statistics is that for the first framework only the sustainable supply chain management practices are different moderated towards risk impact. It seems that lower complex networks have a significant relationship for their sustainable SCM practices towards their perception of their impact of risk.
  • 80. 70 Risk probability and SCM practices – risk management for lower and higher complex networks in the second framework Secondly when analyzing the t-statistics for moderators we see that distinction between risk probability and risk management, both proactive and reactive, goes also for sustainable variables. Furthermore, and this is new, for the sustainable moderator it appears that also the sustainable practices are different between these groups for the relation towards risk management. But in this case it are the high complex networks that shows a significant relationship between these practices and both proactive and reactive management. Risk impact – SCM practices for lower and more complex networks in the second framework Last but not least, we found one moderated relationship on the 5 % significance level, namely the relationship from the risk impact towards sustainable practices. Lower complex systems take in mind on their perception of risk impact to choose the degree of sustainable practices. Mean Values, standard errors and T-statistics for the two groups for multi-group moderation first framework < sustainable SCM Low complexity High complexity SCMpractices -> RiskImpact -0,1185 0,0538 2,0577 0,0138 0,0466 0,451 second framework < general SCM Asia Europe RiskProbability -> ProactiveRiskMgmt -0,0215 0,0821 0,3400 0,1802 0,0787 2,136 RiskProbability -> ReactiveRiskMgmt -0,0402 0,0854 0,5207 0,1877 0,0806 2,1759 second framework < general SCM Low complexity High complexity RiskProbability -> ProactiveRiskMgmt 0,3734 0,1654 2,5179 0,0449 0,098 0,5057 RiskProbability -> ReactiveRiskMgmt 0,3654 0,1608 2,5375 0,057 0,0997 0,6483 RiskProbability -> SCMpractices -0,0927 0,144 0,8437 0,265 0,1278 1,9934 second framework < sustainable SCM Low complexity High complexity RiskImpact -> SCMpractices -0,3231 0,1505 2,2591 -0,0271 0,0725 0,2424 RiskProbability -> ProactiveRiskMgmt 0,3621 0,1671 2,3725 0,0626 0,0824 0,8433 RiskProbability -> ReactiveRiskMgmt 0,3397 0,1633 2,2728 0,0614 0,0855 0,7836 SCMpractices -> ProactiveRiskMgmt -0,0156 0,0708 0,3214 0,2079 0,0724 2,6722 SCMpractices -> ReactiveRiskMgmt 0,0345 0,064 0,3724 0,1723 0,0467 3,4437 Table 18: Mean, standard errors and T-statistics for the two groups for moderation Conclusion Despite that there is not so much found on the 5% significance level, we can make some conclusions. What about sustainable practices, it seems like there are no differences between Europe and Asia. Furthermore it is surprising but the differences between complex networks for sustainable SCM practices and risk impacts are associated with each other in both directions. But both moderated relationships are not equally significant. Apparently low complex networks focus significantly more on sustainable SCM practices to perceive the impact of risk than general SCM practices.
  • 81. 71 Furthermore it appears that lower complex networks more account for their perception of risk possibility to perform their amount of risk management whereas high complex networks not. Also European countries do this strategy and Asia not. What is surprising is that the higher complex supply chains act on their general supply chain management practices in relation to their probability of risk and not at all in the relation towards risk management. It shows that more complex networks do in fact care about their supply chain. Once they perceive a higher risk probability they align their supply chain management practices in the same way. This is not the conclusion for a less complex supply chains. A last remark that we can make is that for sustainable practices, more high complex supply networks focus on their SCM practices whereas low complex supply networks focus more on their risk probability towards both proactive and reactive risk management. 5.4 Some comparison with 3 years ago A limitation of this study and this gathered data is the absence of data for “effort in last 3 years” for some variables i.e. for environment, risk probability, risk impact and some supply chain design & complexity variables. We will now consider to make some conclusions for the last hypothesis (H6). In earlier section we have showed that there were progressions in risk management compared to 3 years ago. We will now execute some paired sample t-test in order to proof these progressions but also take a look at evolutions of general and sustainable SCM practices. As we look through appendix 9 section paired sample t-test for variables, we found evidence for evolution of most of the variables. What about the dataset for Asia, we see that all the indicators have increased in value between 3 years ago and now. For European countries the story is somewhat different. Although these companies have increased on average their reactive risk management programs, proactive risk management has not significantly changed in the past 3 years. Preventing has decreased and detecting increased some but not much. Furthermore some SCM practices proved to be not significant more practice in the past 3 year. This is the case for performance assessment for the company’s sustainability practices with their suppliers (SusPerfAssS is not significant on the 5 % significance level). In addition, only sharing information with the suppliers as well as with customers has increased between the past 3 years for European companies. Other general SCM practices proved not to be significant on the 5 % significance level. At least, when we conclude for the whole world, we can say that all indicators have significantly changed except for developing an internal sourcing strategy with suppliers.
  • 82. 72 There seems to be no significant relationship between current implementation and effort in the last 3 years for all the investigated relationships. To come to this table we applied the same formula as for calculating moderator t-statistics. The highest t-statistic is for the relationship between SCM sustainable practices on reactive risk management for the first framework with low complex networks. Apparently, no significantly different relationships exist between now and 3 years ago. We will now summarize or findings of the proposed hypothesis. 5.5 Hypothesis summary Hypothesis summary H1 Accepted for market, Rejected for PORTER and PEST for both risk perceptions H2a Accepted for both risk management strategies H2b Accepted for both risk management strategies H3a Accepted for market and PEST Rejected for PORTER for both risk perceptions H3b Accepted for both risk perceptions on both risk management strategies H3c Accepted for market and PEST Rejected for PORTER for both risk perceptions H3d Accepted for risk probability on both risk management strategies Rejected for risk impact on both risk management strategies H4a Accepted for the multi-group moderation relationship between sustainable SCM practices and risk impact * Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H4b Accepted for the multi-group moderation relationships between risk probability on both risk management strategies where both SCM practices are mediator * Accepted for the multi-group moderation relationship between sustainable SCM practices and both risk management strategies * Accepted for the multi-group moderation relationship between risk impact and sustainable SCM practices Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H5a Rejected for all multi-group relationships Rejected for all moderated mediation relationships H5b Accepted for the relationship between risk probability on both risk management strategies where general SCM practices is mediator * Rejected for all other multi-group relationships Rejected for all moderated mediation relationships H6a Rejected for both frameworks and both SCM practices H6b Rejected for both frameworks and both SCM practices Table 19: hypothesis summary table Table 19 summarizes the proposed hypotheses from our conceptual framework and their results. ‘Accepting’ denotes here that the test was significant and consequently supports our conceptual framework. The Hypothesis’s noted with a * are the those ones who are significant on the 10% significance level.
  • 83. 73 6 Overall Conclusion “Many companies perceive if they have to choose their effort between costly risk mitigation strategies or more flexible supply chain strategies as if they to choose one of both.” We began this work with declaring several concepts like risks, their aspects and drivers, disruptions, vulnerability and many more. We proposed the framework to implement a comprehensive risk management on the basis of many previous literature researches. Thereby we addressed the fact of risk monitoring and sharing risk information. The concept of SCRM has been linked with SCM, SCSM and ERM. The aim of this work was to investigate the difference between the use of more pro-active and more re-active risk management strategies to manage supply chains against disruptions. A conceptual framework has been presented which includes environmental factors, SCM factors, risk factors and SC Design and Complexity variables. Data was gathered with the aim to test our framework and propose several hypotheses. We have aimed to compose the most reliable model and with the highest validity. Two frameworks have been investigated where both mediator variables were between the constructs of environment and risk perception on the one hand and between risk perception and management on the other hand. Some environmental variables (i.e. market variables) have impact on the perceptions of risks (hypothesis 1). Secondly, we saw a stronger influence of risk impact than risk probability on risk management strategies and most firms perceives a medium risk probability with a higher impact on average. Although risk management has increased by several countries in 3 years, we do not find so much difference for more proactive or reactive risk management implementation. Yet, proactive management is slightly more in favor (hypothesis 2) but not always related with other variables. General SCM practices ameliorate the model a lot, in terms of declared variance. Sustainable SCM slightly improves the model. Apparently managers take their degree of general supply chain management practices into account when performing a particular level of strategy (hypothesis 3). When we analyzed for Europe & Asia and Low versus High SCD&C (hypothesis 4) we found a lot of differences in their environmental and risk perceptions but also in their general and sustainable supply chain and risk management. In fact, Asian countries score a lot higher for nearly all the indicators. Even worse, the practice of preventing risks more proactively has declined for European countries. Asian countries perceive their probabilities of risk as much higher and yet their impact as a lot lower than European countries. Nevertheless these Asian countries implement more risk
  • 84. 74 programs. They are able to compare a highly leveled supply chain with an advanced level of risk management in comparison with European countries. This is the challenge in today’s business operations. When we dig a little deeper, i.e. performed some further analyses in differences between moderators in the several relationships from our conceptual framework, the outcomes are not always explicable. Risk impact might be associated with sustainable SCM practices in both directions and the perception of risk probability differs to their level of risk management implementation. Lower complex SC focus more on cheaper forms of SCM practices like information sharing and collaborative approaches whereas more advanced supply chains focus on joint decision making and system coupling. In addition, lower complex supply chains focus more on risk perception in their relation to risk management implementations. Complex networks do in fact care about their supply chain because their risk management practices are more dependent of these. Once they perceive a higher risk probability they align their supply chain management practices in the same way. This is not the conclusion for a less complex supply chains (hypothesis 5). At least, a lot of variables have increase in 3 years but not significant different relationships were found. Why can it be possible that a lot of risk management efforts don’t really break through or not much changed in their level of implementation in a lot of companies? Perceive these companies as if the economic crisis is gone for some time? And what about preventing natural hazards? First, we must say that a lot of companies see that risk management strategies are not always time and cost-effective. Certainly European countries think that risk management doesn’t always prove to have added value. Secondly, some strategies might reduce flexibility of operations or supply chain management strategies because they could impose business restrictions. Finally, the feeling of many managers that the crises is yet behind our backs may arouse that we must not stick to constantly prevent risks or performing business operations in a cautious way. Besides, a lot of papers in the risk management were written during or after the crisis (i.e. 2008- 2010). One of them is the German research from paper 37. They found that the financial crisis has not influenced the relationship between enterprise risk management and supply chain (risk) management. We found overall relationship between supply chain management practices and proactive risk management for both current implementations as well as for the effort in the last 3 years for the whole world but with no significant differences. This indicates that after the crisis, companies stick to align supply chain practices and risk strategies but not everyone augment them.
  • 85. 75 7 Limitations and possibilities for future research In this work we investigated risk management strategies and divided them in more reactive and more proactive ones. But more risk management practices might upgrade the framework in value. Because, what we have noticed, is that the constructs for proactive and those for reactive action programs don’t discriminate too much. This makes generality of findings more difficult. In future research it might be interesting to further investigate related risk program strategies such the practice of an integrated risk management or the usage of a supply chain or catastrophic vulnerability framework. Second point we like to address is that besides a lot of data that is used for the current implementation, more data for 3 years ago would be fruitful to make more conclusions for the evolutions of more indicators (risk probability, risk impact, environment factors and some supply chain design and complexity variables). In addition future research can be executed for the relationship of more risk indicators. Besides the risk perceptions of risk probability and risk impact, other risk indicators like for example speed of risk, risk frequency, probability of detecting and the perception of overall risk can be further investigated towards risk management. Conclusions have been made for the whole world an some differences for the European and Asia countries. The dataset used contains 14 European countries, good for 503 companies and 4 Asian countries that represent 337 companies. But this dataset does lack some world figures namely only Canada from North America has participated for this research. So other world continents were not represented through this data. Further research can research data for the missing countries here. At last, no statements have been made for the Belgian firms. Although a modest range of 30 companies have participated through this research, several indicators didn’t contain 30 values. It was therefore not recommended to execute statistical analysis. Future research could enlarge the Belgian dataset.
  • 86. 76 8 References Altay, N., & Ramirez, A. (2010). Impact of disasters on firms in different sectors: implications for supply chains. Journal of Supply Chain Management, 46(4), 59-80. Alvarenga, C., & Lehman, E. (2012). Accenture: Managing Risks in the Supply Chain: Reaching New Standards (pp. 1-12): Accenture. Asbjørnslett, B. E. (2009). Assessing the vulnerability of supply chains. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 15-33): Springer. Autry, C., & Sanders, N. (2009). Supply chain security: A dynamic capabilities approach. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 307-329): Springer. Blome, C., & Henke, M. (2009). Single Versus Multiple Sourcing: A Supply Risk Management Perspective. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 125-135): Springer. Blome, C., & Schoenherr, T. (2011). Supply chain risk management in financial crises—A multiple case-study approach. International Journal of Production Economics, 134(1), 43-57. Blos, M. F., Quaddus, M., Wee, H., & Watanabe, K. (2009). Supply chain risk management (SCRM): a case study on the automotive and electronic industries in Brazil. Supply Chain Management: An International Journal, 14(4), 247-252. Cagliano, A. C., De Marco, A., Grimaldi, S., & Rafele, C. (2012). An integrated approach to supply chain risk analysis. Journal of Risk Research, 15(7), 817-840. Celly, K. S., & Frazier, G. L. (1996). Outcome-based and behavior-based coordination efforts in channel relationships. Journal of marketing research, 33(2), 200-210. Chen, F. Y., & Yano, C. A. (2010). Improving supply chain performance and managing risk under weather-related demand uncertainty. Management Science, 56(8), 1380-1397. Dani, S. (2009). Predicting and managing supply chain risks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 53-66): Springer. Ellegaard, C. (2008). Supply risk management in a small company perspective. Supply Chain Management: An International Journal, 13(6), 425-434. Ellis, S. C., Henry, R. M., & Shockley, J. (2010). Buyer perceptions of supply disruption risk: a behavioral view and empirical assessment. Journal of Operations Management, 28(1), 34-46. Gaonkar, R., & Viswanadham, N. (2004). A conceptual and analytical framework for the management of risk in supply chains. Paper presented at the Robotics and Automation, 2004. Proceedings. ICRA'04. 2004 IEEE International Conference on. Gaudenzi, B. (2009). Assessing risks in projects and processes. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 67-82): Springer.
  • 87. 77 Griffis, S. E., & Whipple, J. M. (2012). A Comprehensive Risk Assessment and Evaluation Model. Transportation Journal, 51(4), 428-451. Hallikas, J., Karvonen, I., Pulkkinen, U., Virolainen, V.-M., & Tuominen, M. (2004). Risk management processes in supplier networks. International Journal of Production Economics, 90(1), 47-58. Hallikas, J., & Varis, J. (2009). Risk management in value networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 35-52): Springer. Haywood, M., & Peck, H. (2003). Improving the management of supply chain vulnerability in UK aerospace manufacturing. Paper presented at the Proceedings of the 1st EUROMA/POMs Conference. Henke, M. (2009). Enterprise and Supply Risk Management. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 177-185): Springer. Jüttner, U., Peck, H., & Christopher, M. (2003). Supply chain risk management: outlining an agenda for future research. International Journal of Logistics: Research and Applications, 6(4), 197- 210. Jüttner, U., & Ziegenbein, A. (2009). Supply chain risk management for small and medium-sized businesses. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 199-217): Springer. Khan, O., Christopher, M., & Burnes, B. (2009). The role of product design in global supply chain risk management. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 137-153): Springer. Kleindorfer, P. R., & Saad, G. H. (2005). Managing disruption risks in supply chains. Production and Operations Management, 14(1), 53-68. Knemeyer, A. M., Zinn, W., & Eroglu, C. (2009). Proactive planning for catastrophic events in supply chains. Journal of Operations Management, 27(2), 141-153. Levi, D. S., Vassiladis, C., & Kyratzoglou, I. (2013). Supply chain and risk management; making the right risk decisions to strengthen operations performance: Pwc. Manuj, I., & Mentzer, J. T. (2008). Global supply chain risk management strategies. International Journal of Physical Distribution & Logistics Management, 38(3), 192-223. Mau, N., & Mau, M. (2009). Securing Global Food Distribution Networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 331-344): Springer. Melnyk, S. A., Rodrigues, A., & Ragatz, G. L. (2009). Using simulation to investigate supply chain disruptions. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 103-122): Springer. Mullai, A. (2009). Risk management system–a conceptual model. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 83-101): Springer. Narasimhan, R., & Talluri, S. (2009). Perspectives on risk management in supply chains. Journal of Operations Management, 27(2), 114-118.
  • 88. 78 Norrman, A., & Jansson, U. (2004). Ericsson's proactive supply chain risk management approach after a serious sub-supplier accident. International Journal of Physical Distribution & Logistics Management, 34(5), 434-456. Peck, H. (2006). Reconciling supply chain vulnerability, risk and supply chain management. International Journal of Logistics: Research and Applications, 9(2), 127-142. Perry, M. (2007). Natural disaster management planning: a study of logistics managers responding to the tsunami. International Journal of Physical Distribution & Logistics Management, 37(5), 409-433. Ritchie, B., & Brindley, C. (2009). SCRM and performance–issues and challenges. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 249-270): Springer. Seiter, M. (2009). Behavioural Risks in Supply Networks. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 235-247): Springer. Sheffi, Y. (2005). Building a Resilient Supply Chain. Harvard business Review, 1(8), 1-4. Sheffi, Y. (2008). Resilience: What it is and how to achieve it: The House Committee on Homeland Security “The Resilient Homeland–Broadening the Homeland Security Strategy. Sheffi, Y., & Rice Jr, J. B. (2005). A supply chain view of the resilient enterprise. MIT Sloan Management Review, 47(1). Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing the determinants of risk behavior. Academy of management review, 17(1), 9-38. Smith, P. G. (2002). Managing risk proactively in product development projects. Paper presented at the IPLNET workshop, Switzerland. Smith, P. G., & Merritt, G. M. (2002). Proactive Risk Management. Executive Book Summaries, 24(11), 1-8. Sodhi, M. S., Son, B. G., & Tang, C. S. (2012). Researchers' perspectives on supply chain risk management. Production and Operations Management, 21(1), 1-13. Tang, C., & Tomlin, B. (2009). How Much Flexibility Does It Take to Mitigate Supply Chain Risks? In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 155-174): Springer. Tang, C. S. (2006). Perspectives in supply chain risk management. International Journal of Production Economics, 103(2), 451-488. Tomlin, B. (2009). Disruption‐management strategies for short life‐cycle products. Naval Research Logistics (NRL), 56(4), 318-347. Voss, D., & Whipple, J. (2009). Food Supply Chain Security: Issues and Implications. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 293-305): Springer. Wagner, S. M., & Bode, C. (2006). An empirical investigation into supply chain vulnerability. Journal of purchasing and supply management, 12(6), 301-312.
  • 89. 79 Wagner, S. M., & Bode, C. (2008). An empirical examination of supply chain performance along several dimensions of risk. Journal of business logistics, 29(1), 307-325. Wagner, S. M., & Bode, C. (2009). Dominant risks and risk management practices in supply chains. In R. Ritchie & G. Zsidisin (Eds.), Supply Chain Risk (pp. 271-290): Springer. Wieland, A., & Wallenburg, C. M. (2012). Dealing with supply chain risks: Linking risk management practices and strategies to performance. International Journal of Physical Distribution & Logistics Management, 42(10), 887-905. Zhao, L., Huo, B., Sun, L., & Zhao, X. (2013). The impact of supply chain risk on supply chain integration and company performance: a global investigation. Supply Chain Management: An International Journal, 18(2), 115-131. Zsidisin, G. A. (2003). A grounded definition of supply risk. Journal of purchasing and supply management, 9(5), 217-224. Zsidisin, G. A., & Ellram, L. M. (2003). An Agency Theory Investigation of Supply Risk M anagement. Journal of Supply Chain Management, 39(3), 15-27.
  • 90. 80 9 Appendices Appendix 1: PWC results Percentage of companies with more than 3 incidents, that suffered an impact of 3% or higher on their performance as a result of supply chain disruptions. Source: Levi et al. (2013).
  • 91. 81 Appendix 2: Participated Belgian companies company name function IC 1 VCST industrial products Koen Verhaeren Operations Manager 29 2 Dewulf NV Thomas Decan Production Manager 28 3 GEA process engineering NV Rudy Van de Perre Operations Director 28 4 Henrad Luc Geysen Supply Chain Director 25 5 Alliance International Pascal Demarets Vice President Operations 28 6 Esco Couplings Stef Huybrechts Operations Director 28 7 Anglo Belgian Corporation Johan Van de Velde Production Manager 28 8 SKT Bart Bauters Director 28 9 Asco industries Belgium Guido van den Berghe Plant Manager 30 10 Roxell Filip Cauwels Plant Manager 28 11 MCsyncro Donald De Buck Production Manager 29 12 Recticel NV - BL Flexible Foams Dirk Van De Velde Manufacturing Manager 29 13 International automotive components group Koen Hendrickx Purchase Manager & Logistics 29 14 Röchling Automotive Gijzegem NV Patrick Temmerman Plant Manager 29 15 Airkan Steven Coeman Production Manager 25 16 Tenneco Jelle Leenknegt Operations Manager 29 17 Team industries Walter Dumarey CEO 25 18 Alex Profiles BVBA Dave Debosschere Managing Director 25 19 PolyVision NV Theo Vanheusden Director of Manufacturing & Site Manager 25 20 Benteler automotive Luigi Neirynck Production Manager 29 21 Hansen Industrial Transmission Bruno Everaert Operations Manager 28 22 Johnson Controls Filip Bral Quality Manager 29 23 Joris Ide Jan Wauters Operations Manager 25 24 Delta Light Tom Compernolle Head of procurement (CPO) 27 25 Rogers Corporation Geert Verreecke Manufacturing Team Leader 26 26 Dana - Spicer Off-Highway Wim Thuy Manufacturing Manager 28 27 TE Connectivity Cammaert Denis Operations Manager 26 28 Niko NV Goedele Heylen Director Operations 27 29 Engineering Pieter Verhaert Engineering Manager 25 30 DAF Trucks Vlaanderen NV Marc Beets Managing Director 29
  • 92. 82 Appendix 3: Internal consistency scale measurement Environment Risk probability
  • 94. 84 Supply Chain Management Practices Supply Chain Design and Complexity
  • 95. 85 Appendix 4: Descriptive statistics Environment Descriptive Statistics N Mean Std. Deviation Variance SubstitutionThreat 828 2,94 1,103 1,216 MarketEntry 838 3,00 1,092 1,192 BargainingPowerS 830 3,11 ,903 ,816 SocPressure 832 3,24 1,062 1,129 EnvirPressure 835 3,33 1,070 1,144 TechnologicalChange 838 3,33 1,003 1,007 MarketSize 840 3,36 ,881 ,776 MarketSpan 833 3,46 1,021 1,042 MarketConcentration 836 3,49 1,077 1,160 BargainingPowerC 832 3,71 ,944 ,891 CompetitiveRivalry 835 3,86 ,918 ,843 Valid N (listwise) 794 Environment constructs Descriptive Statistics N Mean Std. Deviation PEST 828 3,3023 ,76588 MARKET 828 3,4352 ,68962 PORTER 811 3,3258 ,55169 Valid N (listwise) 794 Risk probability Descriptive Statistics N Mean Std. Deviation Variance Prob_Shipment 801 2,32 1,212 1,469 Prob_Operations 804 2,52 1,181 1,396 Prob_Supplier 806 2,74 1,168 1,364 Valid N (listwise) 801
  • 96. 86 Risk Impact Descriptive Statistics N Mean Std. Deviation Variance Imp_Shipment 801 3,37 1,320 1,742 Imp_Operations 804 3,52 1,241 1,540 Imp_Supplier 807 3,66 1,183 1,399 Valid N (listwise) 801 Proactive and Reactive Risk Management (last 3 year) Descriptive Statistics N Mean Std. Deviation Variance RecoveringL3Y 794 3,15 1,077 1,159 DetectingL3Y 796 3,27 ,993 ,986 RespondingL3Y 795 3,28 ,971 ,943 PreventingL3Y 798 3,42 ,958 ,919 Valid N (listwise) 791 Proactive and Reactive Risk Management (now) Descriptive Statistics N Mean Std. Deviation Variance Recovering 794 3,35 1,080 1,166 Detecting 795 3,46 ,957 ,916 Responding 797 3,47 ,966 ,933 Preventing 798 3,52 ,935 ,875 Valid N (listwise) 790 Supply chain Management practices (last 3 year) Descriptive Statistics N Mean Std. Deviation Variance SystemCouplingC 768 2,77 1,238 1,533 SystemCouplingS 777 2,80 1,157 1,338 InterDistrC 762 2,83 1,289 1,663 InterSourcingS 776 2,89 1,226 1,503 CollabApprC 772 3,02 1,124 1,263 SharingInfoC 773 3,05 1,111 1,235 DecisionMakingS 778 3,05 1,022 1,045 DecisionMakingC 772 3,07 1,118 1,249 SharingInfoS 782 3,20 ,991 ,981 CollabApprS 780 3,21 1,030 1,062 Valid N (listwise) 747
  • 97. 87 Supply chain Management practices (now) Descriptive Statistics N Mean Std. Deviation Variance SystemCoupling_C 763 2,93 1,278 1,634 InterSourcing_S 772 2,95 1,201 1,443 SystemCoupling_S 774 2,95 1,190 1,416 InterDistr_C 759 2,96 1,314 1,727 CollabAppr_C 770 3,15 1,146 1,313 DecisionMaking_S 777 3,18 1,040 1,081 DecisionMaking_C 767 3,22 1,127 1,269 SharingInfo_C 772 3,23 1,119 1,253 CollabAppr_S 778 3,33 ,978 ,956 SharingInfo_S 782 3,37 ,974 ,950 Valid N (listwise) 736 Sustainability SCM practices (3 years ago) Descriptive Statistics N Minimum Maximum Mean Std. Deviation SusPerformanceAssesmentS 794 1 5 3,00 1,151 SusTrainingS 795 1 5 2,46 1,247 SusJointEffortS 775 1 5 2,66 1,189 Valid N (listwise) 772 Sustainability SCM practices (now) Descriptive Statistics N Minimum Maximum Mean Std. Deviation SusPerformanceAssesment_S 792 1 5 3,15 1,130 SusTraining_S 787 1 5 2,64 1,281 SusJointEffort_S 768 1 5 2,86 1,235 Valid N (listwise) 765
  • 98. 88 Supply Chain Design and Complexity (last 3 years) Descriptive Statistics N Mean Std. Deviation Variance OneVSMultiplePlants 545 2,82 1,441 2,076 StableVSchangingFlexibilityPlant 547 2,88 1,300 1,689 NetwPerfMgmtSystemL3Y 523 3,06 1,185 1,404 UseOfTechnologyL3Y 524 3,23 1,163 1,352 IntegrInfoSystems 546 3,64 1,291 1,665 Valid N (listwise) 503 Supply Chain Design and Complexity (now) Descriptive Statistics N Mean Std. Deviation Variance OneVSMultiplePlants 545 2,82 1,441 2,076 StableVSchangingFlexibilityPlant 547 2,88 1,300 1,689 NetwPerfMgmtSystem 525 3,22 1,220 1,488 UseOfTechnology 528 3,39 1,112 1,237 IntegrInfoSystems 546 3,64 1,291 1,665 Valid N (listwise) 502 Appendix 5: Different frameworks Framework 1:
  • 99. 89 Framework 2: Appendix 6: Confirmatory Factor analysis for alternative framework formulations market entry in PORTER and market size in PEST and included in market Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE MARKET MarketConcentration 0,7323 0,53626329 0,6552 0,4903 MarketEntry 0,7395 0,54686025 MarketSize 0,6516 0,42458256 MarketSpan 0,6733 0,45333289 PEST EnvirPressure 0,6906 0,47692836 0.7033 0,5191 SocPressure 0,6551 0,42915601 TechnologicalChange 0,785 0,616225 Market Size 0,7445 0,55428025 PORTER BargainingPowerCust 0,6169 0,38056561 0,6961 0,4382 BargainingPowerSuppl 0,6045 0,36542025 CompetitiveRivalry 0,6038 0,36457444 Market Entry 0,7821 0,61168041 ThreatSubstitution 0,6847 0,46881409 RiskImpact Imp_Operations 0,9783 0,95707089 0,9593 0,9609 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9489 0,9513 Prob_Shipment 0,964 0,929296
  • 100. 90 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9809 0,96216481 0,9632 0,9316 Detecting 0,9796 0,95961616 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9659 0,9362 Responding 0,9764 0,95335696 MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability MARKET 0,7002 PEST 0,7135 0,7205 PORTER 0,7331 0,5641 0,6620 ProactiveRiskMgmt 0,2867 0,2439 0,2072 0,9803 ReactiveRiskMgmt 0,2926 0,2515 0,2033 0,9588 0,9753 RiskImpact 0,2707 0,2044 0,2215 0,7277 0,7195 0,9652 RiskProbability 0,2908 0,2038 0,232 0,6439 0,6376 0,8235 0,9676 All environmental factors together Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE Environment MarketConcentration 0,671 0,450241 0,8293 0,3648 MarketEntry 0,6829 0,46635241 MarketSize 0,5758 0,33154564 MarketSpan 0,6062 0,36747844 EnvirPressure 0,6607 0,43652449 SocPressure 0,5619 0,31573161 TechnologicalChange 0,6464 0,41783296 BargainingPowerCust 0,5475 0,29975625 BargainingPowerSuppl 0,5101 0,26020201 CompetitiveRivalry 0,5998 0,35976004 ThreatSubstitution 0,5542 0,30713764 RiskImpact Imp_Operations 0,9783 0,95707089 0,9593 0,9609 Imp_Shipment 0,9604 0,92236816 Imp_Supplier 0,9567 0,91527489 RiskProbability Prob_Operations 0,9776 0,95570176 0,9489 0,9513 Prob_Shipment 0,964 0,929296 Prob_Supplier 0,961 0,923521 ProactiveRiskMgmt Preventing 0,9809 0,96216481 0,9632 0,9316 Detecting 0,9796 0,95961616 ReactiveRiskMgmt Recovering 0,9743 0,94926049 0,9659 0,9362 Responding 0,9764 0,95335696
  • 101. 91 Environment Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability Environment 0,603987 ProactiveRiskMgmt 0,2778 0,980255 ReactiveRiskMgmt 0,2825 0,9588 0,975346 RiskImpact 0,2623 0,7277 0,7195 0,965194 RiskProbability 0,2703 0,6439 0,6376 0,8235 0,967574 General Supply chain management practices Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE SCMpractices CollabAppr_C 0,9339 0,87217 0,9835 0,8711 CollabAppr_S 0,9513 0,90497 DecisionMaking_C 0,9208 0,84787 DecisionMaking_S 0,9456 0,89416 InfoSharing_C 0,9385 0,88078 InfoSharing_S 0,9435 0,89019 InternDistr_C 0,884 0,78146 InternSourcing_S 0,9324 0,86937 SystemCoupling_C 0,9356 0,87535 SystemCoupling_S 0,9461 0,89511 MARKET PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCM practices MARKET 0,7002 PEST 0,5927 0,7750 PORTER 0,561 0,5419 0,6866 ProactiveRiskMgmt 0,2867 0,222 0,1656 0,9803 ReactiveRiskMgmt 0,2927 0,2287 0,1638 0,9586 0,9753 RiskImpact 0,2707 0,1829 0,1814 0,7276 0,7194 0,9652 RiskProbability 0,2909 0,1714 0,1829 0,6438 0,6375 0,8235 0,9676 SCMpractices 0,2466 0,2218 0,1508 0,6633 0,6635 0,5856 0,5332 0,9333
  • 102. 92 Sustainable Supply chain management practices Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE sus. SCM practices SusJointEffort_S 0,9123 0,8322913 0,9247 0,8691 SusPerfAss_S 0,9325 0,8695563 SusTraining_S 0,9514 0,905162 Market PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCM sust. practices Market 0,70021 PEST 0,5927 0,77505 PORTER 0,561 0,5419 0,68651 ProactiveRiskMgmt 0,2867 0,222 0,1656 0,9802551 ReactiveRiskMgmt 0,2927 0,2287 0,1638 0,9586 0,975346 RiskImpact 0,2707 0,1829 0,1814 0,7276 0,7194 0,9652 RiskProbability 0,2909 0,1714 0,1829 0,6438 0,6375 0,8235 0,96757429 SCM sust. practices 0,293 0,2553 0,175 0,5716 0,5836 0,5319 0,4656 0,9322553 Supply chain design and complexity Latent Variable Indicators Loadings Indicator Reliability Cronbach's alpha AVE Supply chain Design and complexity NetwPerfMgmtSystem 0,9683 0,9376049 0,9804 0,926 OneVSMultiplePlants 0,9523 0,9068753 IntegrInfoSystems 0,9598 0,921216 UseOfTechnology 0,9729 0,9465344 StableVSchangingFlexibilityPlant 0,9582 0,9181472 Market PEST PORTER Proactive RiskMgmt Reactive RiskMgmt Risk Impact Risk Probability SCD&C SCM (sus.) practices Market 0,7002 0 0 0 0 0 0 0 0 PEST 0,5927 0,77505 0 0 0 0 0 0 0 PORTER 0,561 0,5419 0,68651 0 0 0 0 0 0 Proactive 0,2867 0,222 0,1657 0,98025507 0 0 0 0 0 Reactive 0,2926 0,2286 0,1639 0,9588 0,9753461 0 0 0 0 Risk Impact 0,2707 0,1829 0,1814 0,7277 0,7195 0,9652 0 0 0 Risk Prob. 0,2909 0,1714 0,1829 0,6438 0,6376 0,8235 0,96757429 0 0 SCD&C 0,0914 0,1373 0,0461 0,1862 0,2135 0,225 0,207 0,962289 0 Sus. SCM 0,2842 0,2309 0,1626 0,483 0,4881 0,4936 0,5876 0,1696 0,9350936 Gen. SCM 0,2465 0,2218 0,1508 0,6632 0,6634 0,5855 0,5332 0,2669 0,933381
  • 103. 93 Appendix 6: T-statistics for mediators General SCM Sustainable SCM First framework Sample Mean (M) Standard Error (STERR) T Statistics Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,1243 0,0475 2,6506 0,1328 0,0513 2,5876 Market -> RiskProbability 0,1616 0,0485 3,3971 0,1521 0,0521 2,9194 Market -> SCMpractices 0,1448 0,0599 2,4769 0,238 0,0604 3,9435 PEST -> RiskImpact -0,0198 0,0376 0,5132 -0,0269 0,0381 0,7074 PEST -> RiskProbability -0,0242 0,0376 0,6612 -0,0731 0,0351 2,0825 PEST -> SCMpractices 0,161 0,0458 3,5349 0,105 0,0469 2,2379 PORTER -> RiskImpact 0,0357 0,0391 0,9027 0,0429 0,0516 0,8307 PORTER -> RiskProbability 0,0311 0,0373 0,7135 0,0422 0,0517 0,8154 PORTER -> SCMpractices -0,0049 0,0521 0,3483 -0,0263 0,0443 0,5935 RiskImpact -> ProactiveRiskMgmt 0,5982 0,0846 7,2521 0,6136 0,0825 7,4369 RiskImpact -> ReactiveRiskMgmt 0,5892 0,0848 7,1289 0,6042 0,0823 7,3406 RiskProbability -> ProactiveRiskMgmt 0,1499 0,0721 1,9216 0,1385 0,0725 1,9111 RiskProbability -> ReactiveRiskMgmt 0,1508 0,0736 1,9038 0,14 0,0725 1,9329 SCMpractices -> RiskImpact 0,5501 0,0539 10,306 0,4546 0,0629 7,2219 SCMpractices -> RiskProbability 0,4927 0,0539 9,2288 0,5538 0,0583 9,4955 General SCM Sustainable SCM Second framework Sample Mean (M) Standard Error (STERR) T Statistics Sample Mean (M) Standard Error (STERR) T Statistics Market -> RiskImpact 0,2357 0,0669 3,541 0,236 0,0645 3,6747 Market -> RiskProbability 0,2768 0,0697 4,0137 0,276 0,0685 4,0861 PEST -> RiskImpact 0,0211 0,0443 0,5186 0,0214 0,0458 0,502 PEST -> RiskProbability -0,0118 0,0438 0,2752 -0,0123 0,0446 0,2702 PORTER -> RiskImpact 0,0429 0,0487 0,7415 0,043 0,0506 0,7131 PORTER -> RiskProbability 0,0409 0,0478 0,6798 0,0406 0,0517 0,6292 RiskImpact -> ProactiveRiskMgmt 0,4468 0,0756 5,9763 0,5893 0,0894 6,809 RiskImpact -> ReactiveRiskMgmt 0,4328 0,0767 5,7265 0,5798 0,0909 6,593 RiskImpact -> SCMpractices 0,4375 0,084 5,4219 0,0345 0,0609 0,4941 RiskProbability -> ProactiveRiskMgmt 0,088 0,0568 1,4498 0,0659 0,0841 0,6369 RiskProbability -> ReactiveRiskMgmt 0,0891 0,0581 1,4234 0,0592 0,0862 0,5507 RiskProbability -> SCM practices 0,1704 0,0734 2,1555 0,5558 0,0668 8,432 SCM practices -> ProactiveRiskMgmt 0,3509 0,0599 5,9235 0,1541 0,0449 3,3612 SCM practices -> ReactiveRiskMgmt 0,3589 0,0587 6,1708 0,1683 0,0446 3,6884
  • 104. 94 Appendix 7: Sobel test Mediator between environment and risk perception Mediation Test generality direct no med direct w med IV-> Med beta Med -> DV beta IV ->med SE Med -> DV SE Sobel test Market -> RiskImpact 0,237 0,133 0,1848 0,5553 0,0599 0,0559 2,946320656 Market -> RiskProbability 0,28 0,187 0,1848 0,4971 0,0599 0,057 2,908512633 PEST -> RiskImpact 0,023 -0,074 0,1228 0,5553 0,0515 0,0559 2,318605859 PEST -> RiskProbability -0,012 -0,045 0,1228 0,4971 0,0515 0,057 2,300044582 PORTER -> RiskImpact 0,036 0,048 -0,0189 0,5553 0,0527 0,0559 -0,35840029 PORTER -> RiskProbability 0,033 0,044 -0,0189 0,4971 0,0527 0,057 -0,35833092 Mediation Test sustainability direct no med direct w med IV-> Med beta Med -> DV beta IV -> med SE Med -> DV SE Sobel test Market -> RiskImpact 0,237 0,133 0,238 0,4546 0,0588 0,0614 3,55154551 Market -> RiskProbability 0,28 0,152 0,238 0,5538 0,0588 0,0574 3,732464065 PEST -> RiskImpact 0,023 -0,027 0,105 0,4546 0,0463 0,0614 2,168380126 PEST -> RiskProbability -0,012 -0,073 0,105 0,5538 0,0463 0,0574 2,207651769 PORTER -> RiskImpact 0,036 0,043 -0,0263 0,4546 0,0434 0,0614 -0,60397116 PORTER -> RiskProbability 0,033 0,042 -0,0263 0,5538 0,0434 0,0574 -0,60479898
  • 105. 95 Mediator between risk perception and risk managements General SCM Mediation Test direct no med direct w med IV-> Med beta Med -> DV beta IV -> med SE Med -> DV SE Sobel test RiskImpact -> ProactiveRiskMgmt 0,138 0,452 0,4552 0,3547 0,084 0,0625 3,919273337 RiskImpact -> ReactiveRiskMgmt 0,14 0,439 0,4552 0,3623 0,084 0,0617 3,982348619 RiskProbability -> ProactiveRiskMgmt 0,614 0,082 0,1583 0,3547 0,0731 0,0625 2,023237295 RiskProbability -> ReactiveRiskMgmt 0,604 0,083 0,1583 0,3623 0,0731 0,0617 2,031763086 Sustainable SCM Mediation Test direct no med direct w med IV-> Med beta Med -> DV beta IV - >med SE Med -> DV SE Sobel test RiskProbability -> ProactiveRiskMgmt 0,138 0,054 0,5629 0,151 0,0611 0,0476 2,999433246 RiskProbability -> ReactiveRiskMgmt 0,14 0,047 0,5629 0,1645 0,0611 0,046 3,333742626 RiskImpact -> ProactiveRiskMgmt 0,614 0,609 0,0301 0,151 0,0572 0,0476 0,519129801 RiskImpact -> ReactiveRiskMgmt 0,604 0,599 0,0301 0,1645 0,0572 0,046 0,520617411
  • 106. 96 Appendix 8: Paired Sampled T-statistics Europe Versus Asia Environment Risk Probability Risk Impact
  • 107. 97 Risk Management General SCM practices Sustainable SCM practices Supply chain Complexity and Design
  • 108. 98 Low Versus High complex networks Environment Risk probability Risk impact Risk management
  • 110. 100 Appendix 9: Multi-group moderation and moderated mediation First framework: SCM practices as mediator variables between environment and risk perception t-statistics first framework General SCM practices Sustainable SCM practices MULTIGROUP MODERATION Asia VS EU L VS H compl Asia VS EU L VS H compl Market -> RiskImpact 0,6359 0,7554 0,1017 0,6449 Market -> RiskProbability 1,1892 0,8877 0,5667 0,8953 Market -> SCMpractices 1,1295 0,2488 1,1138 1,0071 PEST -> RiskImpact 0,8564 0,3586 1,0135 0,0682 PEST -> RiskProbability 0,8752 1,3912 1,0766 1,6242 PEST -> SCMpractices 0,0793 0,9016 0,0823 0,3683 PORTER -> RiskImpact 0,9234 0,6434 0,5830 0,2117 PORTER -> RiskProbability 0,7361 0,7447 0,5702 0,3206 PORTER -> SCMpractices 0,6403 0,2952 1,2589 0,0545 RiskImpact -> ProactiveRiskMgmt 0,6523 0,2320 0,6822 0,1851 RiskImpact -> ReactiveRiskMgmt 0,8333 0,1418 0,8873 0,1318 RiskProbability -> ProactiveRiskMgmt 1,4035 1,5103 1,4342 1,4014 RiskProbability -> ReactiveRiskMgmt 1,4983 1,5621 1,5379 1,4461 SCMpractices -> RiskImpact 1,5009 0,2412 0,1811 1,7324 SCMpractices -> RiskProbability 0,9471 1,2217 0,3756 0,5853 MODERATED MEDIATION Market -> ProactiveRiskMgmt 0,0921 0,5414 0,1588 0,6779 Market -> ReactiveRiskMgmt 0,1324 0,5600 0,1973 0,6879 Market -> RiskImpact 0,4830 0,7625 0,5295 0,7815 Market -> RiskProbability 0,1686 1,2184 0,1003 0,9775 PEST -> ProactiveRiskMgmt 0,8700 0,6280 0,9759 0,6196 PEST -> ReactiveRiskMgmt 0,8901 0,5499 0,9987 0,5612 PEST -> RiskImpact 0,8096 0,2623 0,8997 0,3300 PEST -> RiskProbability 0,8173 1,3142 0,9025 1,5949 PORTER -> ProactiveRiskMgmt 1,0701 0,4724 1,2789 0,2779 PORTER -> ReactiveRiskMgmt 1,0701 0,4956 1,2794 0,2736 PORTER -> RiskImpact 1,0368 0,7161 1,2493 0,2701 PORTER -> RiskProbability 0,8804 0,8246 1,0696 0,3544 SCMpractices -> ProactiveRiskMgmt 0,6333 0,2016 0,1201 0,9118 SCMpractices -> ReactiveRiskMgmt 0,7378 0,2174 0,0295 0,9469
  • 111. 101 Second framework: SCM practices as mediator variables between risk perception and risk management t-statistics second framework General SCM practices Sustainable SCM practices MULTIGROUP MODERATION Asia VS EU L VS H compl Asia VS EU L VS H compl Market -> RiskImpact 0,4976 1,1442 0,5062 1,1739 Market -> RiskProbability 0,1458 1,3351 0,1494 1,2834 PEST -> RiskImpact 0,7539 0,1577 0,7588 0,1431 PEST -> RiskProbability 0,7227 1,2396 0,7377 1,2143 PORTER -> RiskImpact 0,9955 0,6033 1,0715 0,5718 PORTER -> RiskProbability 0,8138 0,7038 0,8736 0,6983 RiskImpact -> ProactiveRiskMgmt 0,0613 0,1446 0,0035 0,1807 RiskImpact -> ReactiveRiskMgmt 0,1062 0,0357 0,1017 0,1461 RiskImpact -> SCMpractices 1,2544 0,6952 0,9063 2,0087 RiskProbability -> ProactiveRiskMgmt 1,7020 1,8121 1,3502 1,8093 RiskProbability -> ReactiveRiskMgmt 1,8689 1,7024 1,4679 1,6676 RiskProbability -> SCMpractices 0,3864 1,7177 0,9838 1,3541 SCMpractices -> ProactiveRiskMgmt 0,2912 0,0693 1,2534 1,9513 SCMpractices -> ReactiveRiskMgmt 0,6251 0,6917 1,5046 1,7178 MODERATED MEDIATION Market -> ProactiveRiskMgmt 0,1111 0,8848 0,1324 0,8779 Market -> ReactiveRiskMgmt 0,1500 0,9020 0,1742 0,8940 Market -> SCMpractices 0,7655 1,1293 0,4025 0,0584 PEST -> ProactiveRiskMgmt 0,8080 0,5949 0,8049 0,6170 PEST -> ReactiveRiskMgmt 0,8298 0,5160 0,8266 0,5377 PEST -> SCMpractices 0,8885 0,5226 0,8417 0,4072 PORTER -> ProactiveRiskMgmt 1,0257 0,4569 1,0915 0,4078 PORTER -> ReactiveRiskMgmt 1,0282 0,4765 1,0931 0,4329 PORTER -> SCMpractices 0,9701 0,5471 1,1142 0,5982 RiskImpact -> ProactiveRiskMgmt 0,6679 0,3094 0,6310 0,1923 RiskImpact -> ReactiveRiskMgmt 0,8715 0,2051 0,8234 0,1000 RiskProbability -> ProactiveRiskMgmt 1,4946 1,5499 1,3899 1,6031 RiskProbability -> ReactiveRiskMgmt 1,6070 1,5853 1,4914 1,5595
  • 112. 102 Appendix 10: Statistics for differences with 3 years ago Variable paired samples t-test Proactive and Reactive management Whole world Asia Europe Proactive versus Reactive
  • 115. 105 Europe Paired sample t-tests for relationships Relationship paired sample t-test Asia versus Europe Differences now and 3 years ago for the relationship between SCM and RM T-statistic first framework T-statistic second framework General Asia SCMpractices -> ProactiveRiskMgmt 0,329748177 0,848347035 SCMpractices -> ReactiveRiskMgmt 0,349440394 1,17079123 Europe SCMpractices -> ProactiveRiskMgmt 0,046033491 0,092671711 SCMpractices -> ReactiveRiskMgmt 0,024877732 0,081973729 Sustainable Asia SCM sus. practices -> ProactiveRiskMgmt 0,439419177 0,245626532 SCM sus. practices -> ReactiveRiskMgmt 0,430489809 0,479315314 Europe SCM sus. practices -> ProactiveRiskMgmt 0,132049205 0,162158132 SCM sus. practices -> ReactiveRiskMgmt 0,204659152 0,144855217 Relationship paired sample t-test Low versus High complex networks Differences now and 3 years ago for the relationship between SCM and RM first framework second framework General High SCMpractices -> ProactiveRiskMgmt 0,206600542 1,077970981 SCMpractices -> ReactiveRiskMgmt 0,301711699 0,954383032 Low SCMpractices -> ProactiveRiskMgmt 0,847346522 0,191774818 SCMpractices -> ReactiveRiskMgmt 0,841522227 0,054850686 Sustainable High SCM sus. practices -> ProactiveRiskMgmt 0,609047769 0,75769831 SCM sus. practices -> ReactiveRiskMgmt 0,566825692 0,980248788 Low SCM sus. practices -> ProactiveRiskMgmt 1,269361435 0,68087869 SCM sus. practices -> ReactiveRiskMgmt 1,315267878 0,782414819