5. Risk Assessment and Quantitative Analytics

Introduction and Context

Risk management in a global supply chain depends on more than identifying possible disruptions. Managers must determine which risks deserve immediate attention, estimate their potential consequences, compare alternative responses, and communicate the basis for investment decisions. Risk assessment provides this discipline. Quantitative analytics strengthens it by translating uncertain events and impacts into structured estimates that can support prioritisation, planning, and governance.

Global supply chains are exposed to interconnected risks. A port closure may delay components, create production shortages, increase transport costs, breach customer commitments, and damage reputation. A supplier’s financial distress may be an operational, financial, and strategic risk simultaneously. Because consequences propagate through networks, a simple list of risks is inadequate. Organisations need methods that reveal likelihood, severity, causal relationships, dependencies, uncertainty, and the effectiveness of controls.

This chapter examines the principal tools used for that purpose: probability-impact matrices, failure mode and effects analysis (FMEA), fault-tree and event-tree analysis, bow-tie analysis, risk registers, scenario analysis, Monte Carlo simulation, Bayesian methods, and stress testing. These methods are complementary rather than interchangeable. Some are useful for rapid screening, others for causal analysis, and others for modelling uncertainty across time and supply-chain nodes. Sound practice requires matching the method to the decision, using credible data, documenting assumptions, and recognising the limits of precision.

Learning Objectives

By the end of this chapter, students should be able to:

1. Explain the purpose, terminology, and stages of supply-chain risk assessment.

2. Distinguish qualitative, semi-quantitative, and quantitative approaches.

3. Construct and interpret probability-impact matrices and risk registers.

4. Apply FMEA, fault trees, event trees, and bow-tie analysis to supply-chain problems.

5. Design scenario analyses and interpret Monte Carlo simulation outputs.

6. Explain how Bayesian updating improves risk estimates when new evidence appears.

7. Develop stress tests for severe but plausible supply-chain conditions.

8. Evaluate data quality, model risk, dependencies, ethical issues, and governance requirements.

9. Translate analytical results into managerial priorities, controls, and resilience investments.

From Risk Identification to Risk Measurement

Risk is commonly understood as the effect of uncertainty on objectives. In supply-chain management, objectives may include service continuity, cost, quality, safety, compliance, sustainability, cash flow, and reputation. A risk event is an occurrence that may produce a deviation from those objectives. A hazard is a condition that can generate an event, while a vulnerability is a weakness that increases exposure or reduces the ability to respond. A control is a measure intended to prevent an event, reduce its likelihood, limit its consequences, or accelerate recovery.

Assessment normally follows a cycle:

• establish the decision context and assessment boundaries;
• identify hazards, events, causes, consequences, and existing controls;
• estimate likelihood and impact;
• analyse dependencies, uncertainty, and control effectiveness;
• prioritise risks and select treatments;
• monitor indicators and update the assessment.

The unit of analysis must be defined carefully. “Supplier disruption” is too broad for reliable analysis. A more useful statement might be: “Tier-two semiconductor supplier unable to deliver specified microcontrollers to the focal plant for more than ten working days.” This formulation identifies the source, affected material, location, duration, and threshold of concern.

Three analytical levels are useful. Qualitative assessment uses descriptive categories such as low, medium, and high. Semi-quantitative assessment assigns scores to categories, often multiplying likelihood and impact ratings. Quantitative assessment uses numerical probability distributions, frequencies, monetary values, time estimates, or service-level measures. Quantification does not automatically make an assessment more accurate. A precise-looking estimate based on weak assumptions may be less useful than a transparent qualitative judgement.

Important terminology includes inherent risk, the exposure before considering controls; residual risk, the exposure after controls; risk appetite, the amount and type of risk an organisation is prepared to pursue or retain; risk tolerance, the acceptable variation around an objective; and risk capacity, the maximum risk the organisation can absorb without threatening viability. These concepts should not be confused. A company may have a low appetite for customer safety risk but a greater tolerance for temporary transport-cost increases.

Probability-Impact Matrices: A Screening Instrument

A probability-impact matrix positions risks according to estimated likelihood and consequence. A five-by-five matrix may classify likelihood as rare, unlikely, possible, likely, and almost certain, while impact may range from insignificant to severe. The intersection produces a rating used to screen and prioritise risks.

For example, suppose a distribution centre assesses four risks:

| Risk | Likelihood score | Impact score | Product | Indicative priority |

|---|---:|---:|---:|---|

| Cyberattack halts warehouse systems | 3 | 5 | 15 | High |

| Temporary labour shortage | 4 | 3 | 12 | High |

| Minor packaging defect | 3 | 2 | 6 | Moderate |

| Single-day local traffic delay | 4 | 1 | 4 | Low |

The multiplication is a ranking convention, not a probability calculation. A score of 15 does not mean a 15 per cent chance or a monetary loss of 15 units. Organisations should define rating scales, evidence requirements, thresholds, and escalation rules. Impact may be assessed across several dimensions—financial, operational, customer, legal, safety, environmental, and reputational—and the highest dimension may determine the overall category.

Matrices are valuable during workshops, early-stage screening, and executive communication. Their weaknesses include scale compression, subjective ratings, false precision, and the possibility that a low-frequency catastrophic event receives less attention than a frequent moderate event. Correlations are also hidden. Two risks may appear separately as moderate but become critical if they occur together. Therefore, matrices should initiate analysis rather than conclude it.

FMEA: Prioritising Failure Modes

Failure Mode and Effects Analysis examines how a process, product, or system can fail. For each failure mode, the team identifies effects, causes, existing controls, and ratings for severity, occurrence, and detectability. A traditional Risk Priority Number is calculated as:

RPN = Severity × Occurrence × Detectability

Each factor is often scored from 1 to 10. High severity means serious consequences; high occurrence means the cause is relatively frequent; high detectability means the failure is difficult to detect before it causes harm. Because a high detectability score indicates poor detection, the scale must be defined explicitly.

Consider an export process. A failure mode is an incorrect customs classification. Its effects may include clearance delay, fines, rework, and customer disruption. Causes may include outdated product codes or inadequate review. Existing controls could include system validation and specialist approval. The team then rates the three dimensions and identifies actions such as master-data governance, automated code checks, and periodic audit.

FMEA is especially useful for repetitive processes, warehouse operations, transportation handoffs, quality assurance, and new-product introduction. It encourages cross-functional discussion and links failure modes to controls. However, RPN multiplication has limitations. Different combinations can produce the same score even though their managerial significance differs. A highly severe failure with a low occurrence score may be treated like a frequent minor failure. In practice, severity thresholds, action rules, and professional judgement should supplement the RPN.

Causal and Consequence Models

Fault-Tree Analysis

Fault-tree analysis begins with a top event, such as “production line stopped for more than 24 hours,” and works backward to identify combinations of causes. Logical gates represent relationships. An AND gate means all linked conditions are required; an OR gate means any one may produce the higher-level event. For example, a line stoppage may result from “no material available” OR “equipment unavailable.” No material may result from “supplier failure” AND “no approved substitute,” or from “transport disruption” AND “insufficient safety stock.”

The tree clarifies causal pathways and can support probability calculations when component probabilities and independence assumptions are credible. Independence should never be assumed casually. Weather, infrastructure damage, and transport delay may share a common cause, making simultaneous occurrence more likely than separate estimates suggest.

Event-Tree Analysis

Event-tree analysis begins with an initiating event and moves forward through possible control successes or failures. Suppose the initiating event is a temperature excursion during pharmaceutical transport. Subsequent branches may ask whether the monitoring system detects it, whether staff respond promptly, whether product can be quarantined, and whether replacement stock is available. Each branch produces a distinct outcome, such as no customer impact, partial loss, or full shipment rejection.

Fault trees explain why an event occurs; event trees explain what may happen after it occurs. Used together, they provide a more complete view of prevention, detection, response, and recovery.

Bow-Tie Analysis

Bow-tie analysis combines a fault tree on the left with an event tree on the right. The central top event is connected to threats and preventive barriers on one side, and to consequences and mitigating barriers on the other. For a hazardous-material spill, threats may include poor packaging, vehicle collision, or loading error. Preventive barriers may include packaging standards, training, inspection, and route controls. Consequences may include injury, environmental damage, regulatory action, and service interruption. Mitigating barriers include emergency response, containment equipment, insurance, and alternate distribution arrangements.

The bow-tie is particularly effective for communicating barrier ownership. It also exposes “paper controls”—controls documented in procedures but not verified in operation. Critical controls should have owners, performance standards, testing frequency, failure indicators, and escalation procedures.

The Risk Register as a Management Instrument

A risk register converts analysis into accountable action. A mature register normally includes a unique risk identifier, description, category, causes, event, consequences, affected objectives, inherent rating, controls, residual rating, treatment plan, owner, due date, indicators, review frequency, and status. It may also include dependencies, affected suppliers or facilities, assumptions, and links to business-continuity plans.

A strong risk statement follows the structure: “Because of [cause], [event] may occur, resulting in [consequence for objective].” For example: “Because the principal port depends on a single channel, a prolonged closure may prevent component imports, resulting in production interruption and contractual penalties.” This is more actionable than “port risk.”

The register should not become a static compliance spreadsheet. It should support decisions about inventory, supplier diversification, contract clauses, capacity reservations, cyber controls, insurance, and crisis planning. Key risk indicators might include supplier liquidity measures, overdue quality corrective actions, port congestion, forecast volatility, cyber vulnerabilities, or days of inventory coverage. Indicators require thresholds and response ownership; data without action rules is merely reporting.

Scenario Analysis and Decision Use

Scenario analysis explores coherent combinations of conditions rather than isolated events. A scenario should describe a plausible environment, its assumptions, the mechanisms by which it affects the supply chain, and the decisions to be tested. Useful scenarios include a major supplier bankruptcy, a regional conflict affecting transport corridors, simultaneous cyber and logistics disruption, a demand surge combined with constrained capacity, or a prolonged energy shortage.

Scenario analysis differs from forecasting. Forecasting attempts to estimate what is most likely; scenarios test how the organisation would perform under alternative futures. A scenario set may include a central case, an adverse case, and a severe-but-plausible case. The objective is not to predict the exact future but to reveal vulnerabilities, decision triggers, and options.

A good scenario workshop brings together procurement, operations, finance, sales, information security, legal, and senior decision-makers. Participants should state assumptions explicitly and avoid treating a single narrative as a probability estimate. Outputs may include time to impact, service loss, cash requirement, recovery time, substitution options, and decisions that must be made within the first 24, 72, or 168 hours.

Monte Carlo Simulation: Representing Uncertainty

Monte Carlo simulation models uncertain inputs through repeated random sampling. Instead of using one transport lead time, for example, the model may use a distribution reflecting ordinary, congested, and severely disrupted conditions. Thousands of iterations generate a distribution of outcomes such as total landed cost, delivery delay, stockout probability, or cash requirement.

A simple inventory example illustrates the logic. Suppose weekly demand is uncertain, lead time varies, and supplier delivery reliability is imperfect. For each simulation iteration, the model samples demand, lead time, and delivery performance, calculates inventory movement, and records whether a stockout occurs. After many iterations, the organisation can estimate the probability of a stockout, the percentile of required safety stock, and the expected cost of shortages and excess inventory.

Results should be interpreted through distributions rather than only averages. The 50th percentile represents a median outcome; the 95th percentile is an outcome exceeded in approximately 5 per cent of modelled cases, assuming the model is appropriate. A manager might design transport capacity for the 90th percentile and maintain contingency arrangements for more extreme outcomes. The choice depends on risk appetite and the cost of protection.

Simulation quality depends on input distributions, dependency structures, time horizons, and model logic. Correlated inputs must be modelled where appropriate. For instance, a storm may simultaneously increase lead time, damage inventory, and reduce available transport. Sensitivity analysis identifies which variables contribute most to outcome variability. If the model is highly sensitive to an uncertain input, investment in better data or a control targeting that input may have greater value than increasing model complexity.

Bayesian Updating and Dynamic Beliefs

Bayesian methods provide a formal way to update risk estimates when evidence changes. Begin with a prior belief about a hypothesis, such as the probability that a supplier will fail within twelve months. New evidence—late deliveries, declining credit indicators, audit findings, or improved financial results—changes the belief to a posterior probability.

In simplified form:

Posterior probability is proportional to Prior probability × Likelihood of the evidence.

The value of Bayesian thinking is not mathematical ceremony; it is disciplined learning. A single late shipment should not necessarily transform a stable supplier into a critical risk. Repeated failures, combined with credible financial evidence, should produce a stronger update. Conversely, strong recent performance may reduce concern but does not eliminate structural exposure such as sole sourcing.

Bayesian networks extend this logic to related variables. Nodes may represent supplier financial health, capacity, quality, transport reliability, and customer service. Conditional relationships allow the organisation to estimate how evidence about one node changes beliefs about others. These models are useful where data are incomplete but expert judgement can be elicited transparently. Their limitations include dependence on prior assumptions, data quality, model structure, and the danger of presenting subjective probabilities as objective facts.

Stress Testing and Resilience Thresholds

Stress testing examines whether a supply chain can withstand specified adverse conditions. Unlike routine risk estimation, stress testing deliberately challenges assumptions. It may ask what happens if the main supplier is unavailable for 30, 60, or 90 days; if transport costs double; if demand rises by 40 per cent; or if a critical information system is unavailable for a week.

A stress test should define the shock, propagation mechanism, response assumptions, performance measures, and failure thresholds. Measures may include fill rate, production hours lost, cash burn, recovery time, safety incidents, contractual exposure, or customer prioritisation. Management should distinguish between a system’s capacity to absorb a shock, adapt during disruption, and recover afterward.

Reverse stress testing begins with an unacceptable outcome—such as insolvency, loss of a regulated licence, or failure to meet critical-customer obligations—and asks what combination of events could produce it. This method can reveal hidden concentrations and weak assumptions. For example, a company may discover that it reaches a critical cash position not through one dramatic event but through moderate demand decline, extended receivables, and expedited freight occurring together.

Integrating the Methods into a Decision Process

No single tool provides a complete assessment. A practical sequence is to use a matrix for initial screening, FMEA for process-level failures, fault and event trees for high-consequence causal pathways, bow-tie analysis for barrier management, scenario analysis for strategic uncertainty, simulation for numerical distributions, Bayesian updating for changing evidence, and stress testing for resilience thresholds.

The choice of method should reflect five questions:

1. What decision must be made?

2. What type of uncertainty is present—frequency, magnitude, dependency, or incomplete knowledge?

3. What data and expertise are available?

4. What level of explanation is required for governance or regulation?

5. How will the result change action?

Analytical outputs should be converted into treatment options. These may include avoidance, reduction, transfer, acceptance, sharing, substitution, postponement, redundancy, flexibility, or improved detection. A response should be evaluated not only by risk reduction but also by cost, implementation time, secondary effects, sustainability, and strategic fit. Diversifying suppliers, for example, may reduce concentration risk but increase qualification cost, complexity, emissions, and quality variation.

Illustrative Case: A Disruption in the Electronics Supply Network

Background

The following is a fictional illustrative case. Orion Mobility manufactures electronic control units for commercial vehicles. Its final assembly plant is located in Country A, while a specialised microcontroller is produced by one qualified supplier in Country B. The product has a long qualification cycle, and the component cannot be substituted immediately.

Situation and Problem

Over three months, delivery performance declines from its normal level. The supplier reports labour turnover and difficulty obtaining a key wafer input. At the same time, a major port serving Country B experiences congestion. Orion’s initial probability-impact matrix classifies the risk as “high,” but the executive team needs to decide whether to hold more inventory, qualify a second source, reserve air freight, or redesign the product.

Analysis

The assessment team writes the risk statement: “Because the supplier is single-sourced and dependent on a constrained upstream wafer input, a prolonged supply interruption may stop control-unit production, causing customer delivery failures and contractual penalties.”

An FMEA identifies failure modes including wafer shortage, supplier production stoppage, port delay, incorrect allocation of scarce stock, and failed emergency air shipment. The most severe mode is not the most frequent one: a complete supplier stoppage has low estimated occurrence but extreme production impact. This leads the team to impose a rule that any failure mode with catastrophic severity receives action regardless of its RPN.

A fault tree for “assembly plant stops” shows two principal pathways: insufficient component supply or failure of a separate testing system. Within the component pathway, supplier stoppage combined with insufficient inventory is a major branch. An event tree examines the consequences of a 20-day supply interruption. If the early-warning system detects the problem and emergency stock is released, the plant can maintain priority production. If qualification of an alternate component has not begun, recovery remains slow.

A Monte Carlo model samples demand, supplier lead time, port delay, inventory consumption, and emergency freight availability. It indicates that current inventory provides adequate protection in ordinary conditions but leaves a significant probability of a stockout under correlated supplier and port disruption. Sensitivity analysis shows that the largest improvement comes from reducing qualification time and securing a second supply route, rather than simply adding a small amount of inventory.

A stress test assumes a 45-day supplier outage, a 30 per cent demand increase for priority customers, and restricted air-freight capacity. Under this condition, the organisation cannot serve all customers. Management therefore establishes a customer-priority protocol, negotiates allocation rules in advance, and assesses the legal implications of preferential supply.

Bayesian updating is used as new evidence arrives. Repeated late deliveries and adverse supplier information increase the estimated probability of disruption. The posterior estimate is not treated as a precise truth; it is used to trigger escalation and a review of assumptions.

Strategy and Outcome

Orion adopts a layered response. It begins technical qualification of a second supplier, increases targeted buffer stock for the most critical product variants, reserves limited emergency freight capacity, and requires the incumbent supplier to provide weekly capacity and upstream-risk information. The company also develops a redesign programme to reduce dependence on the specialised microcontroller.

The outcome is not the elimination of risk. Rather, the company reduces exposure, improves detection, clarifies decision rights, and increases recovery options. The investment case is approved because the analysis demonstrates that flexibility and qualification capability address the dominant uncertainty more effectively than inventory alone.

Learning Points

The case demonstrates that a high-level rating should lead to deeper analysis. It also shows why correlated disruptions matter, why severity overrides may be necessary, why simulation should examine tails as well as averages, and why resilience depends on options that can be activated under pressure.

Limitations, Common Errors, and Model Risk

Risk assessment is vulnerable to several recurring errors. Teams may confuse likelihood with frequency, use inconsistent rating definitions, anchor on recent events, ignore near misses, or underestimate low-frequency high-impact risks. They may double-count consequences across risk entries or fail to recognise that several entries share one root cause.

Data limitations are equally important. Historical data may not represent a changed network, new supplier, emerging technology, or unprecedented disruption. Small samples produce unstable frequencies. Supplier-provided data may be incomplete or strategically presented. Expert judgement is valuable but should be documented, challenged, and periodically reviewed.

Model risk arises when the structure, assumptions, coding, or interpretation of a model is wrong. Controls include version management, independent review, sensitivity analysis, validation against known cases, transparent documentation, and clear communication of uncertainty. Decision-makers should never treat a model output as a guarantee.

Other common mistakes include treating a risk register as an audit document, assigning ownership without authority, measuring controls by whether a procedure exists rather than whether it works, and recommending mitigation without calculating implementation constraints. Excessive complexity is also a risk. A transparent model that supports action is generally preferable to a sophisticated model that users cannot understand or maintain.

Governance, Ethics, and Professional Responsibility

Risk analytics influences supplier selection, customer allocation, employee safety, insurance, investment, and continuity of essential goods. It therefore has ethical and governance implications. Data should be obtained lawfully and used for legitimate purposes. Personal information in supplier or workforce datasets must be protected. Automated scores should not create unjustified discrimination against smaller suppliers or firms operating in particular regions.

Boards and senior managers remain accountable for decisions even when analysis is delegated or automated. Significant assumptions, confidence levels, data gaps, and conflicts of interest should be disclosed. A supplier should not be labelled “high risk” solely because of opaque algorithmic output. Where possible, the organisation should provide explanation, review, and correction mechanisms.

Regulated industries may require traceable evidence, validation, business-continuity testing, and retention of records. Contractual risk allocation should be realistic and should not encourage suppliers to conceal information. Ethical resilience also requires considering workers, communities, environmental effects, and access to critical products rather than focusing only on shareholder loss.

Digital Analytics and Emerging Practice

Modern supply-chain risk platforms integrate enterprise-resource-planning data, transport events, supplier portals, weather information, financial indicators, cyber intelligence, and geospatial feeds. Dashboards can display risk exposures, control status, lead-time variability, and early-warning indicators. Process mining can reveal bottlenecks and deviations from standard workflows.

Machine-learning models may detect anomalies in delivery patterns or estimate disruption probabilities. However, predictive accuracy can decline when conditions change. Explainability, drift monitoring, human review, and documented override procedures are essential. Generative AI can assist with extracting obligations from contracts, summarising incidents, and proposing risk-register updates, but it should not independently determine critical risk ratings without validated data and accountable oversight.

Emerging practice is moving toward dynamic risk sensing, digital twins, network graph analysis, and integrated stress testing. These approaches can model multi-tier relationships and propagation across nodes. Their value depends on data interoperability, supplier participation, cybersecurity, and organisational willingness to act on warnings. Better visibility is not resilience unless it is connected to authority, resources, and rehearsed response plans.

Best-Practice Principles

Effective quantitative risk management in global supply chains rests on several principles:

• define risks in relation to objectives and decision thresholds;
• combine qualitative judgement with quantitative evidence;
• use multiple methods when risks are complex or high consequence;
• distinguish inherent exposure from residual exposure;
• model dependencies and common causes explicitly;
• test the sensitivity of conclusions to uncertain assumptions;
• give critical controls named owners and measurable standards;
• update assessments when evidence, networks, or strategies change;
• communicate ranges, confidence, and limitations rather than false precision;
• connect analysis to funded actions, escalation rules, and rehearsals.

Chapter Summary

Risk assessment converts uncertainty in the supply chain into a basis for prioritisation and action. Probability-impact matrices are useful screening tools, while FMEA examines process failure modes and controls. Fault trees analyse combinations of causes, event trees trace consequences, and bow-tie models connect threats, barriers, top events, and outcomes. Risk registers preserve accountability and link analysis to indicators and treatment plans.

Scenario analysis explores coherent alternative conditions, Monte Carlo simulation represents distributions and tail outcomes, Bayesian methods update beliefs as evidence changes, and stress testing evaluates resilience under severe but plausible shocks. These methods must be selected according to the decision, data, uncertainty, and governance requirement. Their value depends on explicit assumptions, recognition of dependencies, model validation, ethical use of data, and conversion of results into practical choices such as redundancy, flexibility, inventory, substitution, or improved controls. Quantitative analytics does not remove uncertainty; it makes uncertainty more visible and more manageable.

Key Terms

Risk: Effect of uncertainty on objectives.

Risk event: An occurrence that may cause deviation from an objective.

Hazard: A condition or source with the potential to generate harm or disruption.

Vulnerability: A weakness that increases exposure or reduces response capability.

Control: A measure intended to prevent, detect, reduce, or recover from a risk.

Inherent risk: Exposure before existing controls are considered.

Residual risk: Exposure remaining after controls are applied.

Risk appetite: Amount and type of risk an organisation is willing to pursue or retain.

Risk tolerance: Acceptable variation around a defined objective.

Probability-impact matrix: Grid used to classify risks by likelihood and consequence.

FMEA: Structured analysis of failure modes, effects, causes, and controls.

Risk Priority Number: Product of severity, occurrence, and detectability scores in traditional FMEA.

Fault-tree analysis: Backward logical analysis of causes leading to a top event.

Event-tree analysis: Forward analysis of possible consequences after an initiating event.

Bow-tie analysis: Model combining causal threats, preventive barriers, consequences, and mitigating barriers.

Risk register: Controlled record of risks, ratings, owners, controls, actions, and indicators.

Monte Carlo simulation: Repeated sampling from input distributions to estimate outcome distributions.

Bayesian updating: Revision of a prior belief using new evidence.

Stress testing: Examination of performance under specified adverse conditions.

Model risk: Possibility that model structure, assumptions, data, or use produces misleading results.

Review Questions

Short Answer Questions

1. Distinguish inherent risk from residual risk.

2. Why is a probability-impact matrix not a probability calculation?

3. What is the purpose of a bow-tie model?

4. Define Monte Carlo simulation in the context of supply-chain risk.

5. What is Bayesian updating?

Long Answer Questions

1. Explain the complementary roles of FMEA, fault-tree analysis, event-tree analysis, and bow-tie analysis.

2. Discuss the design and governance requirements of an effective supply-chain risk register.

3. Explain how scenario analysis and stress testing differ from forecasting.

4. Evaluate the benefits and limitations of Monte Carlo simulation for inventory and service-risk decisions.

5. Discuss the major sources of model risk and the controls used to manage them.

Analytical / Application Questions

1. A firm has a single overseas supplier, volatile demand, and limited safety stock. Design an assessment using at least four methods from this chapter and explain how each would inform a decision.

2. Construct a qualitative bow-tie analysis for a critical shipment arriving late at a manufacturing plant. Identify threats, preventive barriers, consequences, and mitigating barriers.

3. A simulation shows an average delivery delay of three days but a 95th-percentile delay of fourteen days. Explain what this means and recommend how management should use the result.

22. Multiple Choice Questions with Answers

Question 1

What is the principal purpose of a probability-impact matrix?

A. To calculate exact probabilities of disruption

B. To screen and prioritise risks using likelihood and consequence categories

C. To identify only the root causes of a failure

D. To replace all quantitative analysis

Correct Answer: B. To screen and prioritise risks using likelihood and consequence categories

Question 2

In traditional FMEA, a high detectability score generally indicates that a failure is:

A. Easy to identify before causing harm

B. Highly severe but rare

C. Difficult to detect before causing consequences

D. Already controlled effectively

Correct Answer: C. Difficult to detect before causing consequences

Question 3

Which analytical tool begins with a top event and works backward toward combinations of causes?

A. Event-tree analysis

B. Fault-tree analysis

C. Scenario analysis

D. Monte Carlo simulation

Correct Answer: B. Fault-tree analysis

Question 4

What does the right-hand side of a bow-tie analysis primarily represent?

A. Threats and preventive controls

B. Historical frequency data

C. Consequences and mitigating barriers

D. Supplier segmentation categories

Correct Answer: C. Consequences and mitigating barriers

Question 5

A risk register is most useful when it:

A. Records risks without assigning responsibility

B. Links risks to owners, controls, indicators, and treatment actions

C. Contains only financial risks

D. Remains unchanged throughout the year

Correct Answer: B. Links risks to owners, controls, indicators, and treatment actions

Question 6

What is the main purpose of Monte Carlo simulation in supply-chain analysis?

A. To eliminate uncertainty from a decision

B. To generate outcome distributions from uncertain inputs

C. To identify legal obligations in contracts

D. To rank suppliers using only qualitative opinions

Correct Answer: B. To generate outcome distributions from uncertain inputs

Question 7

Bayesian updating is most appropriate when an organisation needs to:

A. Revise a risk belief in light of new evidence

B. Multiply severity and occurrence scores

C. Map preventive and mitigating barriers

D. Set a fixed inventory level without assumptions

Correct Answer: A. Revise a risk belief in light of new evidence

Question 8

Which statement best describes stress testing?

A. It predicts the single most likely future

B. It examines performance under specified adverse conditions

C. It measures only historical supplier quality

D. It removes the need for business-continuity planning

Correct Answer: B. It examines performance under specified adverse conditions

Question 9

A simulation has a low average loss but a very high 95th-percentile loss. This suggests that:

A. Extreme outcomes are impossible

B. The average fully describes the risk

C. Tail outcomes may require contingency planning

D. The model contains no uncertainty

Correct Answer: C. Tail outcomes may require contingency planning

Question 10

Why should correlated disruptions be represented in risk analysis?

A. Because separate risks can share a common cause and occur together

B. Because correlation always reduces total risk

C. Because correlation makes qualitative analysis unnecessary

D. Because correlated risks cannot be controlled

Correct Answer: A. Because separate risks can share a common cause and occur together