LL2-25 digest — Ch25 Why did business not react with precaution to early warnings?#

Late lessons from early warnings vol. 2 (EEA 2013), pp. 607–620 (text pp. 607–617). Authors: Marc Le Menestrel (business ethics and decision theory, UPF and INSEAD) and Julian Rode (ethics and behavioural economics, UFZ Leipzig). No panels.

Core argument#

This is a cross‑cutting analytical essay, not a case study. It reuses vignettes from both Late lessons volumes to explain why firms facing early warnings did not act with precaution. The headline is: “In virtually all reviewed cases it was perceived to be profitable for industries to continue using potentially harmful products or operations” (p. 607).

The explanation is layered: - Economic rationality. Ethical values count “if and only if” they add to expected profit. Harms to others enter decisions only through three channels: liability, regulation and reputation (p. 608). - Each channel is often weak. Publics are unaware, underrate risks or feel unaffected, and firms can shape opinion when that “appears cheaper” than cutting the risk (p. 609). Regulators depend on firm‑generated information, have promote‑and‑protect mandates, and seat conflicted experts (pp. 611–613). Liability is absent, puts the burden of proof on victims, or is escaped through insolvency (p. 612). - Uncertainty as a resource. Uncertainty is intrinsic to science, evidential thresholds and “acceptable risk” are value judgements, and firms have exploited this to “manufacture uncertainty” (pp. 609–610). - Cognition and self‑deception. Human cognition underweights distant, unexperienced harms. Self‑serving bias turns ambiguity into a “welcome ‘excuse’” (pp. 613–614).

The authors distil three lessons (pp. 615–616): 1. Warnings create value conflicts in which firms expect precaution to cost them. 2. The research and regulatory context lets firms take “political actions”, which aim to change the rules, the evidence base and public opinion in pursuit of profit, as distinct from “business actions” taken within the rules. 3. Psychological and cultural factors, including an ideology that profit‑seeking itself serves society, sustain denial.

Key evidence (all second‑hand, from other chapters and literature)#

• Manufactured doubt (p. 610): Bayer on Gaucho, benzene consultants, Monsanto on PCBs (1960s), Shell (1967), Brush Wellman calling for more research after the beryllium limit was “beyond doubt” ineffective, tobacco, and Exxon “publicly denying the existence of the problem that they had privately identified”. The result: “paralysis by analysis”.
• Research control without misconduct (pp. 610–611): design and publication bias; manipulable vote‑counting; non‑peer‑reviewed symposia; Ethyl and GM as sole funders of lead research “for over 40 years” (with the hedged claim that critics were defunded and threatened); beryllium papers financed “under the names of well‑known academics”.
• Dual mandates (p. 612): USDA, DFO (misnamed “US”; it is Canadian), MAFF, the French Ministry of Agriculture, the US DOE. Biased behaviour is described only for USDA and (hedged) DFO.
• Conflicted experts (pp. 612–613): benzene TLV committee, paid lead consultants advising the Bureau of Mines, tobacco’s subversion of WHO; a pharma claim that rests on a COI‑policy review.
• Externalised costs (p. 612): fisheries, MTBE, UK asbestos. Manville’s 1982 bankruptcy was “a means of dealing with” asbestos claims.
• Culture (pp. 614–615): Manville’s asbestos “blunder was denial” (Sells, 1994), yet in 1986 the firm voluntarily labelled fiberglass as possibly carcinogenic and “benefited” thanks to “a successful indemnification and marketing strategy”. The warning was later judged unwarranted.

Authors’ recommendations (advocacy, untested)#

• Legal, fiscal and financial internalisation, including assurance bonding (pp. 611–612, 616).
• Accept that “it seems unrealistic to believe that complete alignment of business interests with interests of society at large will always be feasible”. Understand rather than blame “with hindsight”, while strengthening accountability; failure to act is “not necessarily” wilful (p. 616).
• Public institutions “could” (a tentative proposal) publicly analyse sector‑specific “dilemmas and temptations” and promote participatory decision frameworks rather than prescribe actions (p. 616).
• Make corporate political actions transparent to “sustain a sound balance of power” (p. 617). Efforts to keep them secret “can be seen as a signal” of “bad faith”.

Main mechanisms#

• Weak feedback of social costs into firm decisions.
• An incentive not to know, plus information asymmetry.
• Competitive and first‑mover pressure (asserted only).
• Exploitable uncertainty.
• Capture of the research agenda.
• Dual mandates and conflicted experts.
• Burden of proof and insolvency that shift costs to society.
• Under‑reaction to unexperienced risks “in most situations”; “in exceptional cases”, temporary over‑weighting after an event occurs (the German post‑Fukushima phase‑out “may have been” one, p. 613).
• Cultures of self‑deception.
• Harmed third parties as early warners (p. 609).

Transferable insights (strength)#

1. Precaution is unlikely where social harms do not feed back into the decider’s accounts, and each channel leaks (pp. 608–609, 611–613). Moderate. Liability and regulatory leaks are illustrated by cases; the reputation leak is mostly argued.
2. Uncertainty can be structurally exploited; “more research” becomes a delay tactic (p. 610). Strong for the pattern, though frequency is not established.
3. Controlling which studies are done biases evidence without fraud, and erodes trust in honest sponsored science (pp. 610–611). Moderate–strong; the collective‑credibility point is suggestive.
4. Dual promote‑and‑protect mandates bias regulators toward demanding proof of causation (p. 612). Moderate. Five dual mandates, but biased behaviour is shown for only two.
5. Burden and standard of proof allocate the cost of uncertainty; latency plus insolvency turns future harms into social costs (p. 612). Moderate. Assurance bonds as the remedy are asserted.
6. Early warnings are cognitively weak signals; self‑interest turns ambiguity into excuse; good people build denial cultures (pp. 613–615). Moderate (lab psychology plus one insider testimony).
7. Distinguish business actions from political actions; the latter undermine “within the rules” defences and society’s warning capacity itself (pp. 611, 615, 617). Moderate.
8. Harmed third parties are natural early warners (p. 609). Moderate, though called a “minority”.
9. Voluntary precaution can pay after a crisis (p. 615). Suggestive: a single case, and one where the warning proved unfounded.
10. Institutional disclosure of dilemmas and transparency of political activity (pp. 616–617). Asserted.

Main caveats#

• Selection on the dependent variable. The cases were chosen as failures, so “virtually all” is close to built in. There is no comparison with firms that acted, and no engagement with the report’s false‑alarm chapter (vol. 2 Ch 2).
• Second‑hand evidence from mixed sources: peer‑reviewed tobacco‑documents studies, a WHO report and an insider HBR account, alongside journalism and advocacy. The authors’ hedges (“it is even said”, “is accused of”, “may have been”) should be kept.
• Self‑citation, including an unpublished mimeo cited for consumer response to disclosure (p. 609).
• Errors. Canada’s DFO is called “US”; cross‑references to Chs 18/20/21 should read 19/21/22; Friedman is quoted without “without deception or fraud”; the Sells quote is ambiguously attributed to “one of the presidents of Manville” but is probably Sells’s own (he headed Manville’s Fiber Glass Group); Lexchin and O’Donovan (2010), a COI‑policy review, is cited for “widespread” conflicts among experts; several references are wrong or missing (Scherer and Palazzo).
• Two causal stories. The calculative account (deliberate doubt‑making) and the unconscious account (bias, culture) are not reconciled. The implicit key is charity for business actions and suspicion for political actions.
• The one success story is double‑edged. The Manville fiberglass case was a false positive, arguably driven by liability. Its “nearly five years” to show the warning unfounded reflects the company’s own reassessment; formal reclassification came much later (IARC 2001, NTP 2011, per a secondary source).
• Omissions. Costs of precaution, differences among firms, and REACH and the Environmental Liability Directive are ignored.
• Later evidence. Supran and Oreskes (2017) found ExxonMobil’s internal and peer‑reviewed documents mostly acknowledged human‑caused warming while its advertorials mostly expressed doubt. That corroborates p. 610.