We opened with four collapses: Enron, BP's Deepwater Horizon, Wirecard and WeWork. They share many aspects, but arguably differ along two dimensions: was the failure organizational or individual, and was it intended or not. The test for the first dimension — would removing one person have prevented it? — gave us the first tool of the course: ethical failure is more often a property of systems and incentives than of only bad individuals.
Your own answers to the question “You join a company and find that your team hits its targets by doing something legal but harmful to customers. Everyone does it. What do you do?” ran ahead of the material. Every vote went to the same option: do it, but raise concerns internally — nobody chose pure compliance, refusal, or external reporting. The Danske case later in the evening showed exactly what happens when concerns are raised internally and nothing follows.
On what business ethics is: Carr's claim that business is a game with its own rules was the provocation, and the answer is that the law sets a floor, not a ceiling — most of the interesting questions sit in the space the law leaves open. Three normative lenses — consequences, duties and character — are not a menu but three questions to ask of the same decision. Your ratings made the point for us: Pinduoduo landed between 0 and 4, Apple at 5 to 6 — yet articulating why the two differ forces you to separate outcomes from intentions from character, and the lens you pick changes the answer.
Danske Bank's Estonian branch was the international case. Asked what best explains the failure, most of you chose “a system built to produce that result” over “a few people who knew what they were doing” — the same shift from individuals to systems the opening cases taught. On the mechanism question, the votes split evenly between “why the branch protected the portfolio” and “why group management never looked”; both are right, and together they are the mechanism. Donaldson gave us the vocabulary for whose standards apply across borders: some norms travel, some are local, and a small core is non-negotiable.
We closed with Friedman against the stakeholder view. The room leaned away from Friedman — between “profit matters most, but not exclusively” and outright disagreement — which makes Session 2 the test: if firms do have responsibilities beyond profit, how are they managed, and can they be measured? Danske returns on Thursday.
This is the background for Lecture 2, Managing and measuring business ethics. Definitions, the EU reporting frame, and the numbers on why ESG ratings disagree — set out here so you can look them up whenever you need them, rather than us reading them out in class. None of it is examined directly, but you will need it to use ESG data sensibly in the exam.
1. What ESG actually covers
ESG is a set of non-financial criteria used by investors, regulators and other stakeholders to assess a company's risks, opportunities and impact. It is conventionally split into three pillars — but note that the split is a convention, not a theory, and plenty of issues sit across two of them.
Environmental
- Greenhouse gas emissions
- Energy use and efficiency
- Water and waste management
- Climate change adaptation
Social
- Labour rights and working conditions
- Diversity, equity and inclusion
- Human rights in the supply chain
- Consumer privacy and product safety
Governance
- Board composition and independence
- Executive compensation
- Anti-corruption and bribery
- Shareholder rights
One thing worth carrying into the exam: governance is not simply the third pillar. It is the mechanism through which the other two get decided, funded and monitored. A firm with weak governance does not have two good pillars and one bad one — it has no reliable pillars at all.
2. Materiality
Not every ESG issue matters equally for every firm. Materiality is the judgment about which ones do. Carbon emissions are highly material for an energy company and marginal for a software firm; data privacy is critical in tech and largely irrelevant in agriculture. Materiality is therefore sector-specific, and it is a prioritisation under constraints rather than a measurement.
Double materiality means assessing two questions separately rather than assuming they coincide:
- Financial materiality — which issues affect the firm's performance, risk profile and long-term value?
- Impact materiality — on which issues does the firm significantly affect people or the planet?
The two often diverge, and the interesting cases are exactly the ones where they do. Assess them independently; do not assume win-wins.
3. The European regulatory frame
In the EU, ESG reporting has moved from voluntary to mandatory and audited. The three instruments you are most likely to meet:
| Instrument | Who it binds | What it requires |
|---|---|---|
| CSRD / ESRS | Large companies and listed SMEs in scope | Sustainability reporting to a common standard, on a double-materiality basis, subject to assurance |
| SFDR | Financial market participants | Disclosure of how sustainability risks and adverse impacts are handled in investment products |
| EU Taxonomy | Companies and investors in scope | A classification of which economic activities count as environmentally sustainable |
The practical consequence for a company: ethics is no longer only a matter of intent or reputation. It produces disclosures that can be checked, compared and litigated.
4. Why ESG ratings disagree
Different agencies looking at the same company routinely reach different conclusions. This is the single most important thing to know before you use rating data in an analysis — and it is structural, not random.
Credit ratings measure one thing — the probability of default — using a clear definition, standardised metrics and long historical data. ESG ratings measure many things at once, with no shared definition of what is being measured. The disagreement is a property of the construct, not a flaw in one agency's work.
Agreement is also uneven across the three pillars. Environmental scores correlate at about 0.53 and social at 0.42, but governance is the worst at roughly 0.30 — the pillar most people assume is the most objective is the one raters agree on least. Part of the reason is that they do not agree on what belongs inside “governance” in the first place.
Where the disagreement comes from
The largest source is measurement: agencies agree that, say, employee practices matter, but one scores them from turnover data and another from lawsuits filed. The second largest is scope: they include different attributes in the first place. Weighting — the thing most people assume drives the disagreement — explains almost none of it.
The same research also finds a rater effect: an agency's overall view of a company bleeds into how it scores individual categories. If an agency likes a firm, it tends to score that firm's specific attributes generously too. That is a subjective judgment travelling through what looks like an objective number.
What follows for your analysis
- Never treat an ESG score as a measurement. Treat it as one agency's opinion, formed by a method you should name.
- If you use rating data, say which provider it came from and what that provider actually measures.
- Where two providers disagree about the same firm, that disagreement is itself evidence worth discussing.
- Prefer the underlying indicators to the aggregate score wherever you can get them.
5. Check yourself
Five questions, three minutes. If you can answer them, you are ready for the lecture; if one stumps you, the section you need is one scroll up.
1. Two agencies rate the same firm three notches apart. What is the biggest cause — scope, measurement, or weighting?
Measurement — about 56% of the gap: the same attribute scored from different indicators. Scope explains 38%; weighting, the cause most people guess, explains almost nothing (Berg, Kölbel & Rigobon 2022).
2. Which pillar do raters agree on least — E, S, or G?
Governance, at roughly 0.30 — the pillar most people assume is the most objective. Raters do not even agree on what belongs inside “governance”.
3. Double materiality asks two questions. What are they — and why must you assess them separately?
Financial materiality (which issues affect the firm’s value and risk) and impact materiality (where the firm significantly affects people or planet). Separately, because they often diverge — and the divergent cases are exactly the interesting ones.
4. You want to use an ESG score in your assignment. What must you say about it for the use to be defensible?
Which provider it comes from, what that provider actually measures, and how it compares with another rater — where two providers disagree, that disagreement is itself evidence worth discussing. Prefer the underlying indicators to the aggregate score where you can.
5. After CSRD, what can happen to a sustainability claim that could not happen before?
It can be checked, compared and litigated: reporting to a common standard, on a double-materiality basis, subject to assurance. Ethics stops being only a matter of intent or reputation.
Sources. Berg, F., Kölbel, J. F., & Rigobon, R. (2022). Aggregate confusion: The divergence of ESG ratings. Review of Finance, 26(6), 1315–1344. doi.org/10.1093/rof/rfac033 · Larcker, D. F., Tayan, B., & Watts, E. M. (2022). Seven myths of ESG. European Financial Management, 28(4), 869–882. · Edmans, A. (2023). The end of ESG. Financial Management, 52(1), 3–17 — required reading for this session. · CFA Institute (2021), ESG ratings: navigating through the haze.
Page maintained by Steffen Brenner. Cite as: Brenner, S. (2026), Lecture 2 reference: measuring ethics with ESG, Ethics in International Business, vesterbrief.com/business-ethics/.
OxyContin has been on the US market for five years. It is FDA-approved and legally prescribed for the treatment of pain. Annual sales have risen from about $48 million in 1996 to more than $1.1 billion in 2000. Purdue promotes the drug extensively to physicians and has communicated that the risk of addiction among patients treated for pain is very low.
The company has built a large sales operation around OxyContin. Sales representatives receive substantial performance-based compensation and use prescription data to identify physicians with high prescribing volumes. In 2001, Purdue pays about $40 million in sales incentive bonuses.
At the same time, reports of OxyContin abuse, diversion, and addiction are appearing in parts of the United States and attracting increasing media and regulatory attention. Inside Purdue, executives are discussing how to interpret these developments, including whether the emerging problems primarily reflect the properties and promotion of OxyContin, inappropriate prescribing, or misuse of the drug.
Figures from Van Zee (2009), American Journal of Public Health.
Explore before you answer. A defensible answer to the five tasks needs an overview first: what happened when, who did what, and how far the harm reached. The Companion below holds the full case dossier — the studies, the court record, the aftermath. Build your overview with it: start from a button, or ask your own questions. When you feel ready, work the tasks.
Assume Purdue must produce a double-materiality assessment for 2001. Answer these five. Each one is a live poll — vote as you go.
- On the firm's own numbers that year, was addiction risk financially material? What does your answer tell you about relying on financial materiality alone?
- Was it impact-material? Name the evidence that existed in 2001 — not what we know now.
- Purdue’s sales representatives average a $71,500 bonus on a $55,000 base salary — variable pay exceeds base pay — with individual bonuses up to nearly $240,000. Is that bonus structure a Social issue or a Governance issue? Does the three-pillar split help you here, or get in the way?
- Name one indicator that would have caught this. Did it already exist in 2001, and who held it?
- Would any ESG rating have flagged Purdue that year? Use measurement, scope and weights from the briefing to say why or why not.
Then take your answers to the Companion — ask it to find the weakest step in your reasoning, or to argue the position you did not take.
We open Lecture 4 with your answers to this. Work it before class and come with a position — you can vote on each question here, and we show the results live. There is no right answer to the first question, and finding out why is the point of the session.
In June 2024, Novo Nordisk was the most valuable listed company in Europe. Its market capitalisation had passed $600bn, roughly four times its 2020 level, and Danish GDP figures had started to be reported both with and without one company's contribution.
By March 2026 the share price was down about 75% from that peak. Roughly $475bn of market value had gone.
Along the way, three things happened to the people in charge.
May 2025 — the CEO goes
Novo Nordisk announced that Lars Fruergaard Jørgensen, chief executive since 2017, would step down. The stated reason was “recent market challenges and the share price decline since mid-2024.” The announcement also disclosed something companies rarely disclose: that the Novo Nordisk Foundation had initiated the conversation, asking both for an accelerated succession and for greater Foundation representation on the board. Chairman Helge Lund said the strategy remained unchanged and the board was confident in the business plans. Press reports later said the decision leaked before Jørgensen was told.
July 2025 — the new CEO
Mike Doustdar was named CEO, a 34-year Novo veteran who joined the company in 1992 as an office clerk in Vienna. Most analysts had expected an external appointment with US experience. On the same day, the company cut its guidance again. Six weeks later it announced 9,000 job cuts, 5,000 of them in Denmark.
October–November 2025 — the board goes
After what Lund described as an inability to “reach a common understanding” with the Foundation about board composition, seven of the twelve directors — including Lund himself and five independents — announced they would not stand for re-election. The Foundation said the fast-moving environment made immediate renewal necessary. At an extraordinary general meeting on 14 November, its candidates were elected. Lars Rebien Sørensen — Novo's CEO from 2000 to 2016, and until that month chairman of the Foundation — became chairman of the company.
Two numbers from that meeting are worth holding onto. 86.98% of the voting rights were represented — but only 58.65% of the share capital. And on every resolution the votes against numbered zero, with 5–7% abstaining.
ISS had recommended abstention. Norway's sovereign wealth fund abstained. CalSTRS voted against. Denmark's largest pension fund, ATP, said publicly that “from the outside, it has not been a pretty process, tarnishing the company's image.”
How Novo Nordisk is controlled
The Novo Nordisk Foundation wholly owns Novo Holdings A/S, which holds all of Novo Nordisk's A shares plus some B shares: 28.1% of the share capital and 77.3% of the votes. A shares carry ten times the votes per krone that B shares do, and are never traded. The Foundation's own articles oblige it to retain a controlling interest and to hold at least 25.5% of the capital, and the A shares cannot be sold.
No outside investor can acquire Novo Nordisk. No coalition of them can outvote the Foundation.
What else was happening
In December 2024, Novo's next-generation obesity drug CagriSema produced 22.7% weight loss in trial. A strong result — but the company had guided the market to expect around 25%. The shares fell about 20% in a day. In February 2026, a head-to-head trial found CagriSema (23.0%) failed to beat Eli Lilly's already-marketed tirzepatide (25.5%).
Meanwhile, US compounding pharmacies had spent two years legally copying semaglutide during a shortage that Novo's own capacity shortfall helped create — Sørensen later called the delayed response to this “a huge failure.” And in November 2025 a pricing deal with the US administration cut Novo's US list prices by roughly 70%.
Through all of it, sales kept growing. 2025 revenue was DKK 309bn, up 10%.
Nobody disputes any of these facts. What people disagree about is what they add up to.
Answer the six poll questions. They take about two minutes. Answer honestly rather than strategically — we use the spread of answers in the session, not the “right” one.
Sources. Novo Nordisk company announcements of 16 May 2025 (CEO succession), 29 July 2025 (CEO appointment and guidance), 10 September 2025 (restructuring), 21 October 2025 (extraordinary general meeting) and 14 November 2025 (EGM resolutions); Novo Nordisk EGM minutes, 14 November 2025; Novo Nordisk Annual Report 2025; Novo Nordisk Foundation ownership disclosures; company trial announcements on CagriSema of 20 December 2024 and 23 February 2026; US Food and Drug Administration declaratory order on semaglutide, 21 February 2025. Share price and market capitalisation figures are approximate, compiled August 2026.
Page maintained by Steffen Brenner. Cite as: Brenner, S. (2026), Whose fault was Novo Nordisk? Micro-case, Lecture 4, Ethics in International Business, Copenhagen Business School. vesterbrief.com/business-ethics/.
November 2006 – February 2007 — Danske acquires Sampo Bank
Danske Bank agrees to buy Sampo Bank in November 2006 and completes the acquisition in February 2007. The transaction includes Sampo’s Estonian banking operation, which becomes a Danske Bank branch in 2008. It already holds a portfolio of non-resident customers: foreign nationals and foreign-registered companies banking in Estonia while doing business elsewhere.
2007–2015 — the non-resident portfolio
Approximately €200 billion passes through the portfolio over the nine years from 2007 through 2015: funds received from outside parties, held briefly in branch accounts, and transferred on to other recipients. Branch deposits over the same period never exceed about €1 billion.
The three largest customer countries are Russia, the United Kingdom and the British Virgin Islands. Many of the customers are companies rather than private individuals: limited partnerships and limited liability partnerships registered in the UK, and companies registered in the British Virgin Islands and other jurisdictions. Some have little or no apparent business activity where they are registered. The branch has often not established who ultimately owns or controls them. That matters because the bank is supposed to know its customers and identify the people behind the companies, particularly when large amounts of money are moving through their accounts.
In 2013 the portfolio produces 99% of the branch’s profit before credit losses, and at the end of that year the branch holds 44% of all non-resident deposits in the Estonian banking system. Across 2009–2015 it contributes about 2.3% of Danske Bank Group’s pre-tax profit.
June 2007 — warning from the Russian Central Bank
The Russian Central Bank writes to the Danish financial regulator that clients of the Estonian branch are moving billions of roubles a month of doubtful origin, possibly including money laundering. The Danish regulator forwards the letter to Danske’s Executive Board within ten days and notifies its Estonian counterpart. The bank replies that it is compliant.
2012–2014 — concerns raised, inside and outside
Estonia’s financial regulator raises concerns about the branch on more than one occasion. In December 2013 and during 2014 a branch employee sends four reports to Copenhagen describing what he sees. He resigns in 2014.
Danske’s internal audit confirms major problems. The Estonian regulator inspects the branch, finds serious weaknesses and requires changes to local management. The concerns do not travel effectively to the people in Copenhagen who could act on them.
2015 – early 2016 — the portfolio is closed
Danske winds down the non-resident portfolio. The division is closed at the end of 2015, with the last accounts closed in early 2016. None of this is public.
March 2017 — the story becomes public
OCCRP and an international network of journalists, including Berlingske in Denmark, publish reporting on the Russian Laundromat that connects the Estonian branch to the scheme. The reporting brings the branch’s role into public view. Danske’s own investigation dates the public interest in the portfolio to this month: “Since March 2017, the terminated Non-Resident Portfolio at Danske Bank’s Estonian branch has attracted significant public interest.” The scale of the flows is not yet known; that comes from the later investigation.
March – December 2017 — Danske investigates
Following the reporting, Danske commissions an external root-cause analysis of the closed portfolio. On 21 September 2017 the bank acknowledges major deficiencies in controls and governance at the branch, and the investigation is expanded to cover customers and transactions from 2007 to 2015 and to establish who knew what and when. The board formally mandates the full investigation on 8 December 2017. Danish authorities open their own investigations.
September 2018 — the investigation is published
The investigation, conducted by the law firm Bruun & Hjejle, covers roughly 15,000 customers and about 9.5 million payments. It examines in detail about 6,200 of those customers — those hitting the most risk indicators — and finds the vast majority of those customers suspicious. The €200 billion is transaction flow, not a finding of laundering: the payments were not examined individually, no accurate estimate of the suspicious share is given, and the report states that a large part of the payments is expected to have been suspicious. Chief executive Thomas Borgen resigns the same day.
February 2019 — the branch is closed
Estonia’s financial regulator orders Danske Bank out of Estonia. The bank announces that it will also leave the other Baltic states and Russia.
December 2022 — settlement in the United States
On 13 December 2022 Danske Bank pleads guilty in the United States to one count of conspiracy to commit bank fraud, for misleading US correspondent banks about its anti-money-laundering controls. It forfeits $2.059 billion, within a total resolution of about $2.9 billion including payments to the SEC and Danish authorities. It is not a guilty plea to money laundering.
Mechanisms identified in the investigations
The portfolio was exceptionally profitable for the Estonian branch, so the incentive to keep it was local. The branch ran its own IT platform and its own anti-money-laundering procedures, with limited group oversight. The bank often did not establish who ultimately owned or controlled its customers, or what their businesses actually were. Warnings raised in Estonia and inside the bank did not move to the people in Copenhagen who could act on them. And the portfolio was closed before the story became public: the full investigation and root-cause analysis followed the media reporting in 2017.
Two questions — vote before we discuss
- Which level best explains this failure? You answered this in the abstract earlier — answer it again now with a real firm in front of you.
- The portfolio was 99% of the branch's profit but only about 2% of the group's. What does that gap explain?
Return on equity, 2016–2025
Not required for the two questions above.
2022 is the only loss-making year in the series, driven by a DKK 13.8 billion provision for the Estonia matter. Excluding that provision and a goodwill impairment, Danske reported an adjusted return of 6.5%. Return on equity was 13.3% in 2025.
Sources. Bruun & Hjejle, Report on the Non-Resident Portfolio at Danske Bank's Estonian Branch (19 September 2018) — flow figure, customer numbers, branch profit shares. US Department of Justice and SEC, 13 December 2022 — guilty plea, forfeiture, penalties. Finantsinspektsioon (Estonia), 19 February 2019 — closure precept; 2014 inspections. Danish FSA, Report on the Danish FSA's supervision of Danske Bank as regards the Estonia case — regulator timeline and the May 2018 decision. Danske Bank Annual Report 2017 — group return on equity.
Introduction
This 7-day take-home case exam tests your ability to apply theoretical concepts and analytical approaches from Ethics in International Business to a real company. You will not be evaluated on rote reproduction of theory, but on how intelligently and creatively you use it in a practical analysis. The benefit of this format is the chance to integrate course content into a concrete case, deepening your learning.
The case company
Your analysis will focus on The Coca-Cola Company (NYSE: KO), a global leader in nonalcoholic beverages. Its portfolio spans sparkling soft drinks, water, sports and energy drinks, juices, coffee, and tea, and it operates worldwide through a franchised bottling system.
The assignment
An employee-owned pension fund is considering a significant investment in The Coca-Cola Company. The fund seeks competitive financial returns and gives equal weight to the nonfinancial performance of its portfolio companies. As an ethics expert, prepare an evaluation that assesses Coca-Cola’s current ethical performance, identifies future ethical risks across its global operations, and recommends concrete actions — covering leadership, governance, incentives, controls, subsidiary management, and KPIs — to address any weaknesses.
Please write a report that includes the following elements and stays within this scope:
1) Company background & performance. Briefly introduce the company — its strategy and major revenue streams — and provide a performance snapshot with recent financials and the ESG profile (material KPIs, ratings, and disclosures, where available).
2) Ethical risk analysis (home & global). Focus on one or two categories (depth over breadth), e.g., corruption; environment; labor & human rights; product/data harms; market conduct. Using evidence, show where the company meets, exceeds or breaches (i) legal requirements, (ii) ethical standards, and/or (iii) social norms. Choose topics with credible sources. Describe the institutional environment and legitimacy factors (home/host-country rules, enforcement capacity, informal norms, stakeholder expectations, extraterritorial laws) that enable or constrain such behavior. Provide evidence-based analysis of compliance versus violations, use strong sources (regulators and courts, audited reports, reputable journalism, NGO and assurance reports), explain context (institutional environments, enforcement strength, informal practices, extraterritorial exposure such as the FCPA, GDPR and supply-chain due diligence), and show how these factors raise or lower risk.
3) Business ethics management. Which management tools can leadership use to address the weaknesses identified in Section 2? Propose concrete, actionable steps using standard business-ethics practices. If no material weaknesses are found, explain which tools are likely already in force and how they are monitored.
4) Leadership & governance readiness. Assess the organization’s capacity to make ethical, stakeholder-inclusive decisions. Evaluate specifically: top leadership (CEO and board); incentives (pay design and the balance of financial versus nonfinancial and ESG KPIs); culture; subsidiary management; and the media and stakeholder environment (firm visibility and salient controversies).
5) Recommendations & limitations. Recommend changes in leadership and governance (e.g., board composition, incentive redesign, cultural change). Discuss limitations, such as challenges in ESG measurement.
6) Notes on theoretical frameworks. Apply course frameworks (e.g., stakeholder theory; upper echelons theory; institutional theory; Donaldson’s Values in Tension; legitimacy theory; rights-, duty-, and utility-based ethics) to justify specific actions. For each framework, state stakeholder priorities, acceptable trade-offs, and decision rules — then map these to concrete recommendations.
Data sources
Company materials: Investor Relations, Form 10-K, Proxy Statement (DEF 14A), quarterly results, risk factors. Regulatory filings and databases: SEC EDGAR. Bottler disclosures for supply-chain insight: Coca-Cola Europacific Partners (CCEP) and Coca-Cola FEMSA (KOF) annual and sustainability reports. Financial news and data: Google Finance, Reuters, Financial Times, Wall Street Journal, Bloomberg. Policy and oversight references as relevant: US CBP/UFLPA updates, EU Deforestation Regulation guidance, WHO health-policy documents on sugar and marketing to children.
The written report
Maximum 10 pages plus title page. Follow the CBS formal requirements. You may include figures and tables as appendices, with references to them in the main text.
Submission
The deadline is absolute; late submission results in exam failure. As a general rule, extensions are not granted.
Project ethics & AI
This is an individual project. You may speak with others and consult external sources, but the analysis must be your own work. Cite all sources properly. All reports are checked with plagiarism software. Please review the CBS guidelines for the use of generative AI in exams.
Everything in this course feeds one task: analysing a firm that faces an ethical issue. This page compresses the course into that pipeline.
The pipeline
| Step | What it covers | Course basis |
|---|---|---|
| 1. Company and performance snapshot | Strategy, financial baseline, ESG ratings and materiality | Lecture 2 |
| 2. The issue and the evidence | Define the issue; violations versus compliance; allegations kept separate from findings | Lectures 1 and 3 |
| 3. The multinational structure | Parent, subsidiaries and ownership; where the issue sits in the group; the institutional context in home and host country | Lecture 3 |
| 4. Media and visibility | How much scrutiny the firm faces and what that predicts | Lecture 7 |
| 5. Leadership | CEO profile and personality signals | Lecture 4 |
| 6. Governance and incentives | Board map: independence, skills, committees; pay design and its performance measures | Lectures 4 and 5 |
| 7. Culture, ethics management and measurement | Culture signals; the management toolkit, at least three concrete tools; how to measure whether it works | Lectures 2 and 6 |
Prompt templates for data collection
You may use AI for fact-finding the way you would use a search engine. The rule for every prompt below: the answer is a lead, not a source. Verify every fact in the primary document and cite that document, never the AI. Replace the parts in brackets.
Find the latest proxy statement (DEF 14A) of [company]. From it, list every director with age, tenure, independence and committee seats; the number of meetings of each committee; and the CEO pay mix with the performance measures behind the bonus. Give me the link to the filing itself.
Find the current ESG ratings of [company] at Sustainalytics and MSCI and the controversies they list. Then tell me which issues the company itself calls material in its latest sustainability report. Link every source.
Map the group structure of [company]: the parent, the major subsidiaries and their countries, ownership stakes, and where [issue] sits in the group. Note any business-group or state-ownership features. Link the annual report pages you used.
Collect the documented record on [issue] at [company]: regulatory actions, court filings, NGO reports and investigative journalism, each with a date and a source link. Keep allegations separate from findings.
Describe the institutional environment [company] faces on [issue] in [home country] and in [host country]: the relevant laws, the regulators, and the enforcement record. Link primary sources.
Collect observable signals about [CEO name]: interviews, letters to shareholders, award coverage, photo prominence in the annual report, and the pay gap to the next executive. Give me raw observations with sources, not conclusions.
Describe the press coverage of [company] on [issue] over the past three years: which outlets, how often, and whether the tone is negative or neutral. Name the five most substantial articles with links.
Collect employee-review data for [company] from kununu, Glassdoor, Indeed and Blind: the overall score and the category ratings on each platform, with the date you accessed them.
Describe the ethics infrastructure of [company]: code of conduct, ethics training, whistleblowing channel, compliance function and certifications. Note what the company discloses about each and link the sources.
The structure of the report
- Executive summary
- Introduction and motivation
- Methods and AI declaration
- Company and performance snapshot
- Issues and risks
- Leadership and governance
- Recommendations, with owners, KPIs and timelines
- References and appendices
What your grade is based on
Four things: logical reasoning and coherence; application of theory; evidence; structure and clarity. Language quality is not graded unless it impairs clarity.
A strong answer has a clear argument, applies theory rather than mentions it, and uses evidence selectively. Clarity counts for more than the amount of information.
Rules for AI use
Structuring the report with AI is allowed. Declare it in the methods section.
Data collection and fact-finding with AI needs no declaration. Verify and cite the primary sources; do not cite the AI as the source of a fact.
If AI generated text, tables or figures that you use, cite the AI output following the CBS Library APA guidance, and cite the underlying primary sources for the facts.
The project is strictly individual. Quote all of your sources. All reports go through plagiarism software.
Robert Nardelli ran Home Depot from December 2000 to January 2007. The company performed well in those years. The stock did not, and his pay became one of the most criticised packages in America.
The setup
Nardelli was hired within days of losing the race to succeed Jack Welch at General Electric. He had no retail background. Over six years his compensation came to about 240 million dollars. Operating performance was strong: revenue grew from about 46 to about 82 billion dollars and net earnings from 2.6 to 5.8 billion. The share price, however, stayed essentially flat, while competitor Lowe’s roughly doubled. High pay next to a flat stock is exactly the combination that draws negative coverage.
| Year | Reported total compensation, $ million |
|---|---|
| 2001 | 38.9 |
| 2002 | 35.4 |
| 2003 | 28.4 |
| 2004 | 37.2 |
| 2005 | 35.8 |
| 2006 | 134.5 |
Figures as compiled in press reports of the proxy statements. The 2006 figure includes elements of the exit arrangement.
Negative articles on the pay package, per year on average: 4 before the escalation → 27 at its height.
The pressure
Criticism of the pay package built for years and intensified through 2006: shareholder proposals, pension funds demanding change, and sustained negative press. The board did not change the package. One analyst later commented that by firing Nardelli the board had “caved” to critics who were able to enlist the news media.
The meeting
At the annual meeting in May 2006, Nardelli was the only director present. Shareholder questions were limited to one minute each, enforced with timers, and he answered none of them. The meeting lasted about half an hour. The meeting itself produced worse coverage than the pay package had.
The exit
On 3 January 2007 Nardelli resigned. His severance package came to about 210 million dollars. Seven months later he was chief executive of Chrysler.
Questions for the session
- The board saw years of negative coverage and did not change the pay package. What best explains that?
- One director present, one minute per question. Which response type is this?
- Nardelli left with about 210 million dollars. Did media pressure work here?
UnitedHealth Group is the largest health insurer in the United States, with about 400 billion dollars in revenue in 2024. UnitedHealthcare is its insurance unit; Optum is its health-services unit. In the two years before December 2024, the company faced investigative reporting, lawsuits and government investigations. Why did this scrutiny not change the company’s behaviour?
The news coverage
February 2023. ProPublica published an investigation of how UnitedHealthcare denied treatment to a chronically ill student, based on the company’s internal recordings and documents. One recorded line — “we’re still gonna say no” — was widely quoted.
2023. The news site STAT published an investigative series on the use of an algorithm, nH Predict, to cut off payments for post-acute care of elderly patients.
2024. The Wall Street Journal published a data-driven series showing that Medicare Advantage insurers, with UnitedHealth the largest, had collected billions of dollars for diagnoses that doctors never treated. The series was later named a finalist for the 2025 Pulitzer Prize in investigative reporting.
What followed the coverage
November 2023. The estates of two deceased patients filed a class action over the nH Predict denials, citing the STAT reporting. These are allegations from a lawsuit, not findings.
February 2024. The Wall Street Journal reported that the Department of Justice had opened an antitrust investigation into the company. The same month, a ransomware attack on the Change Healthcare subsidiary disrupted claims and payments across the American health system. The group CEO testified before Congress in May.
May 2024. A pension fund filed a class action claiming the company had not disclosed the antitrust investigation while executives, including UnitedHealthcare CEO Brian Thompson, sold more than 15 million dollars in stock. This too is an allegation, not a finding.
October 2024. A US Senate subcommittee reported that UnitedHealthcare’s denial rate for post-acute care had more than doubled between 2020 and 2022.
So the chain was working: journalists found the story, and lawyers and Congress picked it up.
4 December 2024
Brian Thompson was fatally shot outside the New York Hilton Midtown, shortly before the group’s investor day. Police characterised the attack as targeted and premeditated. Police-sourced reporting said three words were written on the ammunition: “delay”, “deny”, “depose” — an echo of the title of a well-known book about insurance claims handling. What followed online was widespread criticism of the insurance industry, not sympathy for the firm.
After December 2024
In May 2025 the group CEO stepped down and the company’s former long-time CEO returned. In July 2025 the company confirmed that it was responding to criminal and civil investigations by the Department of Justice into its Medicare billing practices. The share price fell sharply during 2025.
Questions for the session
- Years of scrutiny preceded December 2024. Which deterrence mechanism failed most fundamentally?
- Choi and Phung argue that newspapers fail to deter corporate illegality where economic institutions blunt the mechanism. Which condition fits this case best?
- You are on the board in early 2024. What should you have done?
- Looking at everything that happened by 2026, what actually changed the company’s behaviour?
Founded in 1850, Lehman Brothers was by 2008 the fourth-largest investment bank in the United States. Through the 2000s it expanded aggressively into risky lending — subprime and Alt-A mortgages, mortgage-backed securities, commercial real estate, funding the expansion with short-term debt — by 2007 the firm was levered at roughly 30 to 1.
On 15 September 2008 Lehman filed the largest bankruptcy in U.S. history, with about $639 billion in assets. The filing became the defining moment of the global financial crisis.
On 10 September 2008, announcing a third-quarter loss of $3.9 billion, Fuld told analysts that “the board’s been wonderfully supportive”. The next day the board approved $100 million in payouts to five executives. Four days later the 159-year-old firm filed for bankruptcy, and the shareholders that board existed to represent lost over $45 billion — a sequence recounted in the Harvard Business School Alumni Bulletin.
Richard Fuld ran Lehman from 1994 until the end — the longest-serving CEO on Wall Street at the time — and he was also chairman of the board.
The table shows the two boards as they stood for election in spring 2008, from the firms’ own proxy statements.
| Lehman Brothers | JPMorgan Chase | |
|---|---|---|
| Director nominees, 2008 | 11 | 12 |
| Independent directors | 10 of 11 | 10 of 12 |
| Chairman and CEO | Combined (Fuld) | Combined (Dimon) |
| Average age | 67.5 | 59.3 |
| Directors aged 76 or older | 4 | 0 |
| Women | 1 | 1 |
| Sitting CEOs of major companies | 0 | 3 |
| Risk committee meetings, fiscal 2007 | 2 | 7 |
| Board meetings, fiscal 2007 | 8 | 9 |
Lehman’s outside directors included a former Sotheby’s CEO, a former IBM chairman, a theatrical producer, the former head of the American Red Cross and the chairman of GlaxoSmithKline. The one nominee with recent banking experience — a retired U.S. Bancorp CEO — stood for election for the first time in April 2008.
JPMorgan Chase’s outside directors that spring included the sitting chief executives of Honeywell, Yum! Brands and Johnson & Johnson, the retired chairman of ExxonMobil, the president of Comcast Cable, and Robert Lipp, a career financial-services executive who had chaired Travelers. The board also included the president of the American Museum of Natural History and a former member of Congress.
Source: the Lehman Brothers and JPMorgan Chase proxy statements of March 2008 (public SEC filings). Larcker & Tayan (2010) (free PDF) analyses Lehman’s board in depth.
Who was in a position to question the bet?
There are five questions on Lehman’s 2008 board. Vote on each. We go through the results in the session.
- Which row of the table best explains why the two firms fared so differently in 2008?
- What was the key weakness?
- Which board mechanism failed most fundamentally?
- You can make one change in 2006. Which do you make first?
- Would a CSR or ESG committee have helped?
Voting opens during the session. Connect the questions to the two board roles from the lecture — monitoring and resources — and to the Endrikat et al. (2021) evidence.
After the collapses of Bear Stearns and Lehman Brothers in 2008, a standard account took hold: the executives of the failed banks were ruined together with their shareholders. Both chief executives held large amounts of stock to the end, and those holdings became worthless. If the people in charge lost fortunes, their pay cannot explain why the firms took the risks.
Bebchuk, Cohen and Spamann collected what the top five executives of each firm actually received between 2000 and 2008: cash bonuses, and the proceeds from selling their own shares. The data come from the firms’ proxy statements and from the filings executives must make when they trade. The paper is The Wages of Failure (2010) — the required reading for this session (free PDF).
| 2000–2008, top five executives | Bear Stearns | Lehman |
|---|---|---|
| Cash bonuses | $327m | $164m |
| Proceeds from selling their own shares | $1,106m | $851m |
| Combined | about $1.46bn | about $1.01bn |
| Equity holdings at the start of 2000 | $798m | $601m |
| The CEO alone (Cayne / Fuld) | $388m | $523m |
The shares the executives still held in September 2008 became worthless. The figures above are cash that had already been received.
Did the executives of the failed banks pay for the failure?
There are five questions. Vote on each. We go through the results in the session.
- The two top teams took out about $2.4 billion in bonuses and share sales while their firms moved toward collapse. What does that fact alone establish?
- Which pay element did the most damage?
- Fuld held his remaining shares to the end and lost heavily in 2008. The paper still counts him far ahead. Why?
- Which reform follows most directly from this evidence?
- Was the problem the amount of pay or the design of pay?
Voting opens during the session. Numbers in the table are from the paper and rounded.
Executive pay only works as well as the measure it is tied to. This briefing covers why measuring a CEO’s contribution is hard, the questions every pay metric must survive, and when ESG metrics hold up.
THE PROBLEM
The perfect performance measure would capture the CEO’s own contribution to firm value. That contribution cannot be observed directly. Every real measure both misses part of it and includes things the CEO did not do — market movements, luck, the work of others. Kerr (1975) called the general failure “rewarding A, while hoping for B”: companies get what they pay for, and it is not always what they wanted.
1. Why not simply pay on the share price?
The share price includes the market and the sector. A CEO can be paid for a boom they did not create, or punished for a crash they did not cause — pay for luck. Measuring performance relative to peers removes part of the problem.
FIVE QUESTIONS FOR ANY PAY METRIC
Market-based or accounting-based? Does it pay for luck? Can it be manipulated? Does it create threshold effects? How noisy is it? Common choices in practice: relative total shareholder return, accounting returns such as ROA, and earnings per share.
2. A bonus pays in full at target and nothing below 95% of target. What behavior does this create?
Threshold effects: strong pressure to just reach the threshold, little reason to go beyond target, and an incentive to move results between periods — including a “big bath” after a year that is already lost.
3. Which is harder to manipulate — earnings per share or total shareholder return?
Total shareholder return. Earnings per share depends on reported earnings and on the number of shares, and management influences both. The market price is harder for management to steer directly.
WHEN ESG METRICS HOLD UP
Tying pay to ESG only works if the metric survives the same tests. Material and verifiable: a clear baseline, a target, and an audit trail. Within the CEO’s control: no reward for outcomes the firm does not drive. Limited discretion: the more subjective the assessment, the closer the bonus moves to a gift. Balance leading and lagging indicators, and spread ESG terms across the bonus and the long-term plans.
4. A bank ties 10% of the bonus to an employee-satisfaction survey. Which tests does that metric fail?
Verifiability — the survey is self-reported and unaudited — and discretion, since management runs it. It may still be a useful signal, but as a bonus metric it is close to a gift.
In the session we apply these tests to a real bonus.
Put yourself in the position described below and answer the question at the end.
You were appointed as the CEO of a Standard & Poor’s 500 firm around three years ago. The company’s current financial performance according to key financial indicators does not significantly differ from the indicators of comparable firms in your industry.
Recently, one of your colleagues informed you that a major news outlet has run an article about your compensation package. The article contains information about the package’s main components such as the fixed pay, bonus, stock options, and restricted stock. Its tone is negative. It argues that the size of the package appears excessively high compared to what is current practice in the industry, and that against the backdrop of the mediocre financial and share price performance, you are clearly overcompensated.
The current compensation contract runs for another two years. Changes that take effect in the following financial year require your consent.
- As a response to the criticism in the article, would you consider reducing the contractual compensation for next year?
Two questions to think about before you read on: How many children do you have? How important are they to you?
You were appointed as the CEO of a Standard & Poor’s 500 firm around three years ago. The company’s current financial performance according to key financial indicators does not significantly differ from the indicators of comparable firms in your industry.
Recently, one of your colleagues informed you that a major news outlet has run an article about your compensation package. The article contains information about the package’s main components such as the fixed pay, bonus, stock options, and restricted stock. Its tone is negative. It argues that the size of the package appears excessively high compared to what is current practice in the industry, and that against the backdrop of the mediocre financial and share price performance, you are clearly overcompensated.
The current compensation contract runs for another two years. Changes that take effect in the following financial year require your consent.
One more question to think about: How likely is it that you would discuss such an article with your children, once they are old enough?
- As a response to the criticism in the article, would you consider reducing the contractual compensation for next year?
We look at the results together in the session.
Carlos Ghosn, born in Brazil to a Lebanese family and educated in France, rose through Michelin and joined Renault in 1996. In 1999 Renault bought 36.8% of Nissan for $5.4 billion, and Ghosn was sent to Tokyo to run the near-bankrupt company. The rescue worked. He became known as Le Cost Killer, appeared as a hero in a Japanese manga, and built the Renault-Nissan-Mitsubishi Alliance into one of the largest car groups in the world.
At the peak, one person held the top of every box: chairman and CEO of Renault, chairman of Nissan, chairman of Mitsubishi Motors, and head of the Alliance that connected them. He chaired all three boards himself.
On 19 November 2018 Ghosn was arrested at a Tokyo airport. Japanese prosecutors charged him with underreporting about $80 million of deferred compensation over eight years, and with breach of trust: shifting personal trading losses to Nissan and directing company payments to business partners abroad. Nissan’s internal investigation reported homes in Beirut and Rio de Janeiro bought for his use through a Netherlands subsidiary, without board approval. Ghosn denies all charges and says he was the target of a plot inside Nissan. None of the charges has been tried in court.
After more than 100 days in detention he was released on bail of about $14 million, under surveillance, his passports surrendered. On 29 December 2019 he left his Tokyo residence, crossed Japan by train, was carried onto a private jet inside an audio-equipment case, and flew via Istanbul to Beirut. Lebanon does not extradite its citizens. Japan issued an arrest warrant, the men who organized the escape were later convicted, and proceedings continue in Japan and France. Ghosn remains in Lebanon.
What kind of culture does a leader like this create?
There are five questions. Vote on each. We go through the results in the session.
- One person led Renault, Nissan, Mitsubishi and the Alliance at once. What is the biggest cultural effect?
- The traits that rescued Nissan and the conduct in the charges look similar. What is the best reading?
- Which of the eight ethical virtues failed most clearly at Nissan?
- Housing, travel and family expenses ran through subsidiaries without board approval. What does that reveal most?
- Ghosn was never tried. Where does accountability stand?
Voting opens during the session. Question 3 uses the eight ethical virtues from Kaptein (2008), covered in the lecture. The charges against Ghosn are allegations; he denies them and none has been tried.
For each question, write down a range — a lowest and a highest value — so that you are 90 percent sure the true answer lies inside it. Keep the range as narrow as you can while staying 90 percent sure. Fill in every range, then press the score button.
We discuss the results in the session.
Whose fault was Novo Nordisk? We opened with your answers to that question. Between June 2024 and March 2026, Novo Nordisk’s market value fell by roughly $475bn — about 75 percent from its peak. Along the way three different people or bodies took the consequence: the chief executive, who stepped down in May 2025; then seven of twelve directors, who did not stand for re-election months later after the Foundation and the board could not agree on renewal; then, at the extraordinary meeting that followed, a former chief executive returned as chairman. It is difficult to tie specific responsibility to the fall in the share price, and the reshuffling of the board and the CEO is not clearly substantiated by the evidence.
Does personality explain it? Research shows that CEOs account for something like 5 to 20 percent of firm performance. Does personality explain part of that share, and does it extend to ethical conduct? We asked whether CEO traits drive firm outcomes, and how humility and hubris get measured in practice rather than just described. That question extends to the board as well: structure matters too, including size, independence, and whether the board has an ESG committee.
The Lehman board illustrates the problem. On 10 September 2008, announcing a $3.9bn quarterly loss, Fuld told analysts “the board’s been wonderfully supportive.” The board approved $100m in payouts to five executives the next day. Four days later Lehman filed the largest bankruptcy in U.S. history. Its board looked conventional on paper — declassified, majority-independent, a full slate of committees — and a side-by-side with JPMorgan Chase’s board that spring showed why paper independence was not the discriminating fact: both boards had a combined chairman and CEO, and Endrikat’s meta-analysis across 82 studies finds that split does not predict outcomes either way. What did differ: Lehman had almost no current financial expertise on its board, and its risk committee met only twice in fiscal 2007, against seven times at JPMorgan.
Calibration and hubris. We ran a short calibration exercise to show that overconfidence can appear even without the conditions that usually produce hubris. A well-calibrated person catches close to 4.5 of 5 inside their own ranges; most rooms catch one, two, or three. It is a demonstration, on yourself, of the same overconfidence a board has to correct for in the executives it oversees.
The questions to take forward. Was Novo Nordisk’s leadership and board turnover a justified response or a search for someone to blame? Does CEO personality explain firm outcomes, ethical conduct, or neither? And what should a board look like if its job is to catch the overconfidence that individuals cannot see in themselves?
What counts as misbehaviour. We started with Uber: the same app, on the same day, was ruled legal in some countries and illegal in others. Law alone cannot define misbehaviour, because law differs across borders and often lags behind the conduct. Misbehaviour is behaviour that violates the law or the norms of the firm’s stakeholders. Multinationals face this problem in a sharper form than domestic firms: distance, organisational complexity and subsidiary autonomy create more places for misbehaviour to occur and fewer eyes watching. A full explanation usually needs three levels — the individual, the organisation, and the institutions around them.
Corruption. The in-class briefing covered the scale of corruption and its drivers: it concentrates where institutions are weak and the stakes are high, and conformity adds pressure — when everyone in a market pays, refusing has a price. That is the multinational’s dilemma: refuse and lose the market, or pay and abandon the norms it carries from home. The evidence says home norms persist: firms from low-corruption countries bribe less abroad, but localisation pulls the other way the longer a subsidiary operates locally.
Danske Bank. We returned to the Estonian branch with the timeline. The warnings existed from 2007 — the Russian central bank’s letter, internal reports, the auditors. They did not travel to the people in Copenhagen who could act. The gap between what the branch did and what the group professed was constant; what changed over the years was who knew, and that is what moves the case from unintentional to negligent to deliberate, depending on the level of the firm and the year you look at.
Siemens and decoupling. Siemens ran systematic bribery behind a model compliance code. The gap between stated values and operating practice is called decoupling, and it is why a policy document proves nothing by itself. Legal and legitimate are two different tests: a firm can pass one and fail the other, and the vote on BP showed how capital allocation can contradict stated commitments while remaining entirely legal.
Does CSR pay? Across roughly two thousand studies the link between corporate social performance and financial performance is rarely negative. Causation remains open — good firms may do CSR rather than CSR making firms good. And measurement is contaminated by decoupling: Danske scored respectably on ESG until March 2017. The score recorded the revelation, not the conduct.
The four questions to take forward. What kind of misconduct is it? Where in the multinational’s structure does it sit? Was decoupling involved? And what does a good CSR score really record?
We opened by returning to Lecture 1 and asking what it actually means to label Danske Bank’s conduct “unethical.” We worked through four tests: consequences (harm that outweighs the benefits under any defensible weighting), duties (a violation that stands independently of outcomes, and a rule that would undermine banking if universalised), character (nine years of routine turned red flags into paperwork — a practice, not a lapse), and the relativist check (local norms may explain acceptance; they do not by themselves justify it). This four-test structure is one way you could organise the normative analysis in an assignment.
Then we dealt with ethics management. There are five ways to manage ethics: governance; standards and risk; stakeholder engagement; culture and training; and monitoring and reporting. Asked how many Danske had, the answer was arguably four, with culture the clearest weakness. The 2007 warning reached the Executive Board; a whistleblower channel existed and was used; internal audit confirmed serious problems. What was missing was therefore not simply architecture, but connection: the mechanisms did not translate information from the part of the organisation generating almost all of the Estonian branch’s profits into effective intervention.
On the question of which stakeholders ethical decision-making should take into account, salience theory predicts that claims receive managerial attention when stakeholders possess power, legitimacy and urgency. Managers tend to respond to how many of these attributes a stakeholder possesses, not necessarily to the moral merit of the claim. At Danske, the framework helps explain why highly profitable customers commanded attention while people harmed by crimes potentially connected to the financial flows had little organisational presence. Whether firms should attend to stakeholders because doing so benefits the firm or because stakeholders are owed consideration is the instrumental–normative divide.
We then moved to measurement. ESG translates parts of the broader CSR debate into measurable indicators, but measurement does not eliminate judgement. ESG ratings from different agencies correlate at about 0.54, compared with roughly 0.96 for credit ratings. Most ESG-rating divergence comes from measurement (56%) and scope (38%), rather than weighting (6%), with governance showing particularly substantial disagreement. For the assignment, we added an AI workflow built around three different objects you can ask AI to produce: existing raters’ scores, your own constructed ESG score, or an investigation of a specific issue or violation. We also added the dated-materiality prompt and its audit check: ask the model to identify every source dated after your chosen as-of year.
Finally, we applied the Purdue Pharma case to materiality. Using only the information available in 2001, the CFO’s risk register could reasonably exclude the issue as financially material at that point. Impact materiality was different: the strongest answer was yes — but genuinely contestable at the time. Evidence of harm was already visible, and some of the strongest confirming evidence existed inside the firm.
We treated the bonus structure primarily as an incentive-design and governance issue rather than simply a social indicator. And when we asked which indicator could have detected the emerging problem, the regulator’s abuse reports were the strongest single answer — but the fuller lesson was that three warning indicators already existed, held by different organisational actors. The measurement problem was therefore not only whether information existed, but whether it was connected, interpreted and acted upon.
Siemens is Germany’s largest engineering company: roughly 428,000 employees, operations in about 190 countries, selling telecommunications networks, power plants and transport systems — often to public customers. You have just joined the corporate compliance function.
Two things have changed recently. Since February 1999, paying bribes to foreign officials is a crime under German law; before that, foreign bribery was not prohibited, and bribes paid abroad were tax-deductible as a business expense. And since March 2001, Siemens is listed on the New York Stock Exchange, which makes it subject to the US Foreign Corrupt Practices Act.
The business, however, runs the way it has always run. In markets where public customers decide contracts, deals are supported by local “business consultants” — agents on success fees whose contracts promise advisory services. Cash desks in Germany allow employees to withdraw large sums for use abroad. Nothing on your desk is marked illegal. Your job is to decide what to check first.
Figures from SEC v. Siemens AG (2008), complaint ¶¶ 7, 12, 28–31.
Answer these five. Each one is a live poll — vote as you go.
- In 1998 a bribe paid abroad was tax-deductible; in 2004 it is a crime. What changed in 1999 — and what didn’t?
- Which driver does the consultant system serve first — access, competitive advantage, incentives, or speed? What evidence from inside the firm would tell you?
- Local units keep paying. Is the pressure coercive (they must), normative (it is how business is done), or mimetic (competitors do it)?
- Design the decoupling test: you may check exactly one thing to distinguish a real compliance program from a façade. What do you check?
- When it later fails: who bears primary responsibility — the manager who pays, the division that budgets, the board that doesn’t ask?
The full story — what investigators found from November 2006 on — is the reveal in class. Resist looking it up until you have voted.
Your task. This briefing gives you the basic concepts and legal background you will need for the Siemens case that follows. As you read, decide how you would answer the four questions at the bottom. Do not use AI or search yet.
1. What counts as corruption?
The standard definition is Transparency International’s: the abuse of entrusted power for private gain. Entrusted power is a role — an official’s authority, a manager’s control over contracts, a bank’s gatekeeper position. Private gain includes indirect gain: a bonus, a profit target, a promotion.
The forms are distinct:
- Bribery — a payment to induce a decision-maker to breach a duty.
- Kickback — a bribe paid out of the proceeds, after the decision.
- Embezzlement — taking entrusted assets directly; nobody else is corrupted.
- Influence-peddling — selling access rather than a decision.
- Facilitation payments — small sums to speed up what the payer is legally entitled to. The gray zone: legal systems disagree about them.
Two distinctions to carry: petty vs. grand corruption (payments at a counter vs. contracts and policy bought at the top), and public vs. private — no official is required. A purchasing manager taking a supplier’s kickback is corruption. Cuervo-Cazurra (2016), required reading, organizes this terrain.
2. Why do borders matter?
An MNE may operate where local practices, enforcement and expectations differ sharply from those at home. But home-country law can follow the firm abroad. At the same time, headquarters must rely on subsidiaries, managers, agents, joint ventures and local partners whom it cannot perfectly observe. The international-business problem is therefore double: different pressures across countries, but responsibility that does not necessarily stop at the border.
- FCPA (US, 1977) — bribing foreign officials became a home-country crime. Inaccurate books and weak internal controls became offenses in themselves.
- OECD Anti-Bribery Convention (signed 1997, in force 1999) — extended the principle to the major exporting economies. Germany criminalized foreign bribery in February 1999; until then, bribes paid abroad were tax-deductible.
- UK Bribery Act (2010) — covers private-to-private bribery, allows no facilitation-payments exception, and makes failure to prevent bribery an offense. Adequate compliance procedures are the only defense.
What this changes for an MNE: local legality does not necessarily protect a firm from home-country anti-bribery law. US listings, US financial-system connections and other jurisdictional links can expose foreign firms to US enforcement. Enforcement can also reach failures and misrepresentations in the systems surrounding misconduct, not only the underlying conduct. Danske Bank, for example, pleaded guilty in the United States to bank fraud arising from misleading US banks about its customers and AML controls, rather than pleading guilty to money laundering itself.
3. Can we measure corruption?
≈ US$1 trillion in bribes per year — the World Bank estimate (Kaufmann), early 2000s. Dated, but traceable.
“Corruption costs 5% of global GDP” — cited everywhere, no traceable methodology; the UN has acknowledged it cannot be verified.
The Corruption Perceptions Index ranks countries by perceived public-sector corruption. It is not incidence data: perceptions lag events, only the public sector is covered, and it measures where bribes are taken — not where they are paid from, laundered, or where the shell companies sit. Denmark tops the CPI; Danske Bank happened anyway. If you use the CPI, say what it measures.
4. Is all corporate misconduct corruption?
No. Corruption is one family of multinational misconduct, not a synonym for unethical business:
- Corruption & undue influence — bribery, kickbacks, procurement fraud, problematic lobbying.
- Labor & human rights — forced and child labor, suppression of organizing, migrant-worker exploitation.
- Environmental harm — illegal discharges, deforestation, unlawful waste exports.
- Product & data harms — unsafe products, deceptive claims, delayed recalls, privacy violations.
Real cases span families — Danske is a governance failure, a facilitation of others’ crime, and a social harm at once. Use the families to sort evidence, not to file companies (Cuervo-Cazurra et al., 2021).
5. The questions we vote on
Decide your answers as you read.
- A supplier secretly pays your purchasing manager 2% of every contract. No public official is involved. Is it corruption?A. No — corruption requires a public official · B. Yes — private-to-private corruption is possible · C. Only if the supplier receives an illegal contract · D. It depends on the CPI score of the country
- Danske Bank’s US guilty plea was for bank fraud rather than money laundering. Why?A. No suspicious money passed through Danske · B. Money laundering is not illegal under US law · C. The established US offense concerned misleading US banks about customers and AML controls · D. Denmark had exclusive jurisdiction over money laundering
- Denmark ranks at or near the top of the CPI. What can you conclude?A. Danish companies rarely participate in corruption abroad · B. Little corrupt money passes through Danish financial institutions · C. Perceived Danish public-sector corruption is comparatively low · D. Danish firms have effective compliance systems
- You find the claim “corruption costs 5% of global GDP” in several reputable reports, but cannot trace the underlying calculation. What should you do in your assignment?A. Cite the most reputable report repeating it · B. Cite several reports to triangulate it · C. Present it as an approximate estimate · D. Do not use it as an established statistic; find a traceable estimate instead
Sources. Cuervo-Cazurra, A. (2016). Corruption in international business. Journal of World Business, 51(1), 35–49 — required reading. · Spencer, J., & Gomez, C. (2011). MNEs and corruption. Strategic Management Journal, 32(3), 280–300 — required reading. · Cuervo-Cazurra, A., Dieleman, M., Hirsch, P., Rodrigues, S. B., & Zyglidopoulos, S. (2021). Multinationals’ misbehavior. Journal of World Business, 56(5). · SEC v. Siemens AG (2008) — German tax-deductibility before 1999. · World Bank/Kaufmann bribery estimate; Transparency International, CPI methodology.
Page maintained by Steffen Brenner. Cite as: Brenner, S. (2026), Lecture 3 reference: corruption and compliance across borders, Ethics in International Business, vesterbrief.com/business-ethics/.
The Business Ethics Companion is a study aid for this course — built to prompt you with follow-up questions rather than hand you answers, and to run live in-class polls tied to each lecture.
It's built and maintained by Steffen Brenner, Associate Professor at Copenhagen Business School, whose research focuses on corporate governance, executive compensation, and the role of AI in strategic decision-making.
Votes are anonymous — no login, no personal data collected, just a random ID stored in your browser.
Questions or feedback: sbr.egb@cbs.dk.