LL1-15 — Ch15 ‘Mad cow disease’ 1980s-2000: how reassurances undermined precaution#

Report: Late lessons from early warnings: the precautionary principle 1896–2000 (EEA Environmental Issue Report No 22, 2001) Pages: Report pp. 157–167 = PDF pp. 157–167. The printed and PDF numbering coincide for this chapter. What was read: The full text extract, in order, from the first page marker (157) to the last (167). All eleven PDF pages were also checked visually for: - the boxes and Table 15.1 - italics marked “emphasis added” - footnotes

Conventions: - Quotation. The 2001 report carries an all-rights-reserved notice (PDF p. 2). Verbatim quotation is therefore limited to short phrases where the exact wording matters: framing, language, and what officials said at the time. Everything else is paraphrase; the full wording is on the page cited. Most quoted phrases come from UK government documents that the BSE Inquiry released and the authors quote. - [Background – verify]. Marks context from general knowledge, not from the chapter. It is included for fairness or context and must be checked in the hindsight step.

Authors and standpoint#

Authors. Patrick van Zwanenberg and Erik Millstone (p. 157). The chapter does not give their affiliations. [Background – verify: - Both were science-policy researchers at SPRU, University of Sussex. - Millstone was a long-standing critic of UK food-safety governance. - They later developed this analysis into BSE: Risk, Science and Governance (OUP, 2005). - Footnote 13 (p. 163) points to national reports from the EC-funded BASES project. SPRU may have written the UK report, which would make that citation partly self-referential.]

Disciplinary lens. This is a study of policy process and institutional behaviour. It is not a veterinary or epidemiological account. The biology is used only to establish what could and could not be known (pp. 157–158).

Evidence base. Unusually, most of the evidence is primary documents. These are mainly: - BSE Inquiry oral-evidence transcripts (1998) - the Inquiry’s “revised factual accounts” (1999a–f) - a ministerial private-office minute from 1988 - the 16-volume Phillips Report (2000)

Alongside these are Commons committee reports, a House of Commons Library research paper (Barclay 1996), Hansard, and European Parliament inquiry papers. Other sources include: - one laboratory paper (Kimberlin et al. 1987) - a Lancet review (Collinge 1999) - a book chapter (Dealler 1996, in Ratzan (ed.), The mad cow crisis). [Background – verify: Dealler was a microbiologist who publicly criticised government BSE policy; the chapter does not characterise him.] - a consultancy report for the agriculture departments and HM Treasury (DTZ Pieda 1998) - two written statements to the BSE Inquiry (Cooke 1998; Martin 1998), the 1979 RCEP report, the EU-funded BASES project reports, and the Commission’s 2000 White Paper - press and broadcast items: Nature (1990), New Scientist, The Economist, Radio Times, Farmers Weekly online, BBC Panorama (pp. 166–167)

The chapter’s claims about motive rest on internal minutes and testimony. They are not inferred from outcomes. Two kinds of evidence are mixed, though, and they differ in weight: - Contemporaneous documents: the February 1988 minute, the Under Secretary’s 1988 remark, the May 1988 NIBSC minutes, the September 1989 MRM minutes, SEAC’s 1990 advice and draft, and the Meat Hygiene Division advice. - Retrospective testimony given after the 1996 crisis: the CVL scientist’s recollection of confidentiality, the MAFF official on “setting up the Committee”, the Welsh CMO on the PHLS, Southwood’s 1996 remark, and the enforcement representative on Panorama.

The retrospective testimony may carry hindsight, and the chapter does not distinguish the two kinds. The citations also blur the line: some items the chapter presents as contemporaneous are cited only to the 1998 oral-evidence transcripts or to the Phillips volumes. Examples are Meldrum’s June 1988 private remark (transcript, 16 June 1998) and SEAC’s May 1990 advice (transcript, 24 March 1998). From the chapter alone, a reader cannot tell whether the quoted words come from a document of the time or from later testimony. Much of the source base is also the Inquiry’s own “factual accounts” (1999a–f, one of them, 1999f, a draft). [Background – verify: these were compiled by Inquiry staff from government files, so they are one step removed from the underlying papers.]

Stance. The chapter is openly adversarial to the official account. - It opens by rejecting the later claims of policy-makers “directly responsible” before March 1996 that their approach had exemplified an ultra precautionary approach and rigorous science-based policy-making. The sources are the December 1998 Inquiry testimony of Gillian Shephard (misspelled “Sheppard”) and John Gummer (p. 157, fn 11). The chapter does not give their offices. [Background – verify: both were Ministers of Agriculture, Gummer 1989–93 and Shephard 1993–94.] The phrase “ultra precautionary” is not in quotation marks in the source. It is the authors’ paraphrase of that testimony, not a verified quotation from either minister. - The organising diagnosis is structural. MAFF had to promote farming and the food industry and also protect public health (p. 157). - The authors dispute the Phillips Inquiry’s more exculpatory conclusions (Box 15.4, p. 165). - The language is evaluative and at times accusatory: “misrepresentation” (p. 161), “covertly subordinated” (p. 164).

The authors are not maximalists. They accept that: - eradication would have cost around £12–15bn (p. 158) - science alone could not locate the right response (p. 159) - zero risk was unattainable without slaughtering and restocking the entire herd and cleaning out the feed chain (p. 161) - many infected cattle had been eaten before the disease was recognised (p. 164)

Their complaint concerns five things: dishonesty about uncertainty, refusal of cheap partial measures, control of knowledge, weak enforcement, and institutional design.

Components. The 2001 volume has no panels. The chapter has four boxes and one table:

No response from industry, farmers or government is included.

Who wrote the boxes. Boxes 15.1–15.4 are unsigned. Box 15.1 speaks in the authors’ “we” (“of which we are aware”; “We might eventually learn”), and Box 15.4 refers to “this case study”, so these notes treat the boxes as the chapter authors’ text. Only Table 15.1 is explicitly marked as the EEA’s. The attribution of Box 15.4’s rebuttal of Phillips to the authors is therefore probable, not certain.

Section-by-section notes#

15.1 Introduction (p. 157)#

• The authors frame the chapter as a rebuttal of policy-makers’ later self-description, which they paraphrase as ultra precautionary and rigorously science-based (see Stance: not a direct quotation).
• They acknowledge that the account is “necessarily selective”.
• Thesis: MAFF’s two contradictory objectives meant it “failed to meet either”. The objectives were promoting “the economic interests of farmers and the food industry” and protecting public health from food-borne hazards. The claim is framed as what “the evidence cited here suggests”. The long-run harm to industry and the Exchequer comes back in the conclusion (p. 164).

15.2 A new cattle disease (pp. 157–158)#

• Recognition. BSE was officially acknowledged in November 1986. It resembled scrapie, a transmissible spongiform encephalopathy (TSE) endemic in UK sheep. TSEs are poorly understood, untreatable and fatal; CJD is the best-known human form.
• Feed as the vector. MAFF suspected scrapie had spread to cattle through feed containing rendered animal remains. Feed was “quickly confirmed” as the main vector, but the origin of BSE (scrapie, a spontaneous cattle TSE, or something else) remained unclear.
• An inventory of uncertainty (p. 157). This passage matters because it shows the authors do not claim the harm was knowable:
• There was no evidence that eating meat from scrapie-infected sheep could cause CJD, but nobody could be sure BSE came from scrapie.
• Even if it did, BSE need not share scrapie’s transmission characteristics: experiments showed that the host range of a scrapie strain could not be predicted once it had jumped species (Kimberlin et al. 1987).
• Even if BSE was assumed to be pathogenic to humans, several things were unknown: which tissues were infective, at what levels, how this changed over incubation, and whether there was a safe threshold.
• No test could detect infection in live animals before symptoms appeared.
• Early internal recognition. Senior officials saw a possible human risk immediately (p. 157).
• In early 1988 the Under Secretary in MAFF’s Animal Health Group told colleagues there was no evidence people could be infected, but “we cannot say there is no risk”. The disease might cross “another species gap” (pp. 157–158). The authors italicise these two clauses (“emphasis added”). Font extraction from the PDF confirms the italics fall on exactly these clauses.
• Officials themselves distinguished no evidence of risk from evidence of no risk.
• Framing. Urgent decisions about a novel disease with unknown implications were unavoidable (p. 158).

Box 15.1 Early warnings (p. 158)#

1. Feed recycling. Rendered slaughterhouse waste had gone into animal feed since at least the early 20th century, with known risks of recycling and amplifying pathogens. - The 1979 Royal Commission on Environmental Pollution (RCEP) recommended minimum processing standards for rendering. - Labour lost the 1979 election before acting. The incoming Conservative government withdrew the draft rules as unnecessary and excessively restrictive, leaving plant operation to industry. - Standards arrived only after 1996. - The authors are cautious here: the effect of 1979 standards is “not yet clear”. - [Background – verify: the RCEP’s seventh report is usually dated September 1979, after Labour lost the May 1979 election. If so, the box’s sequence (Labour lost the election “before” it could follow the advice) is compressed; the draft rendering rules may have predated the report. Check against Barclay (1996) and Phillips.]
2. A US precedent. In the mid-1970s the USDA excluded carcasses of sheep and goats with, or exposed to, scrapie from human and animal food. This was partly to prevent spread of scrapie to other flocks and partly over a possible scrapie–CJD link. The UK took no equivalent action. - The source is secondhand: a 1998 written statement to the BSE Inquiry by the former Scientific Director of the Moredun Research Institute (Martin 1998), not US documents. - The counterfactual is explicitly conditional: if BSE came from scrapie, similar cheap restrictions “might have prevented” the epidemic.
3. Human transmissibility. MAFF vets recognised the possibility in 1986 but thought the probability “acceptably slight” (the authors’ wording of the vets’ view, not a quotation). - The earliest documented official acknowledgement “of which we are aware” (the authors’ hedge) of a more-than-remote risk came in May 1988. Minutes of a meeting at the National Institute for Biological Standards and Control (NIBSC), attended by senior advisers, record the view that BSE may be transmissible to humans, by analogy with scrapie and CJD.
4. Accumulating evidence: - 1990–95: BSE’s transmission differed from scrapie’s. Most tellingly, from 1990 it was shown to pass to domestic cats via food, and cats are not susceptible to scrapie. - 1995: unusual CJD in very young people. The temporal and geographical association was circumstantial evidence. - 1996–97: “direct evidence indicative of a causal relationship”. This included studies indicating that the pathological and clinical features of BSE and vCJD were identical, and that both differed from scrapie and sporadic CJD.

Analytic note (the note-taker’s framing, not the authors’): the box describes a ladder of evidence: 1. a precedent abroad (mid-1970s) 2. generic hazard (1979) 3. analogy (1988) 4. cross-species transmission in animals (1990) 5. a human epidemiological signal (1995) 6. direct evidence indicative of causation (1996–97)

The collapse of the policy came only near the top (March 1996). Partial measures were, however, taken at the analogy stage (1988–89).

15.3 Initial decisions (pp. 158–159)#

• The range of options (p. 158). Options ran from most to least precautionary, and cost ran in the opposite direction.
• Eradication: slaughtering almost the whole herd, because contaminated feed batches were untraceable, almost all dairy herds had eaten meat and bone meal (MBM), and beef cattle were mostly bred from dairy herds.
• The text puts this “in the order of” £12–15bn. Footnote 12 derives a maximum of £15bn from £865 per cow (the 1996 compensation rate), about 12 million cattle, slaughter and incineration costs, and knock-on effects on employment (Economist, 1996).
• Cheaper intermediate options, which “would have substantially reduced the risks without spending a great deal of public money”: banning animals from affected herds as human food; banning all suspect bovine tissues from the food chain; or, at minimum, banning clinically affected animals as human food.
• The cheapest option was forgone (p. 158). About 1,200 clinical cases were recorded in 1987 and the first half of 1988. True incidence “must almost certainly have been higher”, since the disease was not yet notifiable. Most were sold as human food. Compensating their removal would have cost no more than £1,000 each, about £1.5m in total. (1,200 × £1,000 is £1.2m, so the £1.5m total presumably allows for under-recording. The chapter does not explain it.)
• The authors add that “it is not yet possible to estimate the harm” from eating those animals.
• Science cannot choose the policy (p. 159). “Even if the science had been massively less uncertain”, placing the response on the range, and deciding how costs are split between public and private sources, would have been a political judgement.
• The government’s view (p. 159). As seen from inside government, any regulation, or any admission that British cattle products “might be harmful”, threatened confidence in British beef at home and abroad. Even the “virtually cost-free” option of sharing information outside MAFF might alert consumers and importers. Fear of these consequences and reluctance to spend “dominated” MAFF policy-making for the first 20 months.
• Chief Veterinary Officer (CVO) advice. The CVO’s first briefing to the minister warned of potentially serious implications “not only domestically but for UK exports”. It advised that regulatory restrictions were not appropriate. It warned that “irresponsible or ill-informed publicity” might lead to “hysterical demands” for “draconian” measures and to other countries rejecting UK exports.
• The chapter does not name this CVO. [Background – verify: probably William Rees, Keith Meldrum’s predecessor; Meldrum became CVO in 1988.] Do not conflate him with Meldrum (p. 160).
• Notifiability. Making the disease notifiable, which the authors call “an essential tool for disease surveillance”, was rejected “in part” because, as one official put it, it might imply to the public that officials knew meat or milk was dangerous (Phillips Vol. 3, para. 2.130). The chapter does not say when BSE did become notifiable; it refers only to “Once BSE became notifiable” (p. 164).
• Secrecy (p. 159).
• Central Veterinary Laboratory (CVL) scientists were placed under confidentiality in December 1986.
• In the first half of 1987, information was restricted even within the State Veterinary Service (Phillips).
• Most of the research community, the medical profession, and senior officials and ministers in other departments learned of BSE only in early 1988.
• The minister overrules his officials, February 1988 (p. 159). In February 1988, with the media starting to pay attention and cases rising, senior MAFF officials “changed their views” and recommended slaughter with compensation for clinically diseased cattle. The chapter sets these side by side and does not say the media caused the change. Privately, officials argued that without a slaughter policy the government would be held responsible if BSE later proved transmissible to humans. Agriculture Minister John MacGregor rejected the advice. His private secretary’s minute of 29 February 1988 gives three reasons: 1. he did not see how to proceed without being clear where the “offsetting savings” would come from 2. “More importantly” (the minute’s own ranking, after an ellipsis): the argument that compensated slaughter would help stem the disease was “precisely the one sugar beet growers have been making”, which the ministry had “strongly and publicly” rejected 3. action “would make the export position much worse, not better”

The authors italicise the first and third (“emphases added”); font extraction confirms this. - The authors’ verdict. Policy “was not precautionary”. Its “primary objective” was to limit the short-term harm to the food industry’s profitability and to public spending. - Analytic note: the sugar beet reason shows a public-health decision being judged by its consistency with the ministry’s stance in a separate agricultural compensation dispute. The chapter gives no detail of the sugar beet case. The minute itself ranked this consistency concern above cost.

15.4 Expert advice and regulatory controls (pp. 159–161)#

• The CMO brought in late (pp. 159–160). The first expert committee was set up in mid-1988, at the insistence of the Chief Medical Officer (CMO).
• The CMO had been told of BSE only in March 1988, 17 months after MAFF.
• He was told only because ministers were advised that they needed his support for their decision not to remove clinically affected cattle from the human food chain. Note-taker’s gloss: an effort to borrow health authority for inaction ended by producing expert review.
• Southwood, 20 June 1988 (p. 160).
• The committee, chaired by Sir Richard Southwood, demanded at its first meeting that clinical cases be removed from human and animal food, with compensation.
• The same day MAFF announced a ruminant-to-ruminant feed ban. Pigs and poultry could still eat the protein, though their susceptibility was unknown.
• An all-species ban was considered and rejected because pigs and poultry were the renderers’ main market.
• The authors say senior veterinary officials “were nevertheless aware that their decision was a gamble”. Their evidence is that in June 1988 the CVO, Keith Meldrum, privately told a colleague that ruminant protein fed to pigs “might contain the agent”, and that whether infection would establish and replicate in pigs “is unknown”.
• Result: about six years of cross-contamination between cattle feed and other feed, which “greatly” prolonged the epidemic. This particular claim carries no specific citation.
• Southwood’s self-limitation (p. 160).
• The committee did not recommend controls on the use in food of sub-clinically infected animals, although their tissues would also carry the agent. In the absence of a live test, such controls would have had to cover the whole herd.
• In March 1996, after the crisis broke, Southwood acknowledged that a ban on all cattle brains “might not have been a politically feasible option” in 1988 (the authors’ paraphrase). His own words: “We felt it was a no-goer. They (MAFF) already thought our proposals were pretty revolutionary” (New Scientist).
• Note-taker’s reading: the committee pre-filtered its advice by its own sense of what was politically feasible. This rests on one retrospective remark.
• Specified bovine offal (SBO) ban, November 1989 (p. 160). The ban on brains and other offal from all cattle in human food came nine months after Southwood reported. The authors say it was introduced “only” after two developments: 1. one of the government’s own expert advisers told officials, in confidence, that he had done private consultancy work for the pet-food industry and had then recommended a bovine offal ban in pet food 2. ministers did not want to be “upstaged” by the meat-products and pet-food industries, both of which had told MAFF they would unilaterally remove bovine offal from their products
• Design of the controls (p. 160). By the end of 1989 the feed ban, slaughter of affected animals and the SBO ban were in place. They were designed to reduce risk, not to eradicate the agent.
• Tissue choice. SBO tissues were chosen because they were easy to remove and of low commercial value, not because only they were infective. There were no bovine data; analogy suggested wider infectivity.
  • Lymph nodes and peripheral nerves: “almost certainly” highly infectious, but could not practicably be removed.
  • Liver: lower infectivity by analogy, but commercially valuable.
• Calves exempt (pp. 160–161). Cattle under six months were excluded from the SBO ban. Calf carcasses were not normally split in abattoirs, so removing their spinal cords would have raised abattoir costs. The chapter implies this was the reason but does not state it outright. The exemption assumed no cow-to-calf transmission. The authors call that implausible, since scrapie passes from ewe to lamb and MAFF did not fund maternal-transmission research until 1989.
• Exposure during the delay (p. 161). From mid-1988 to the end of 1989, an estimated 30,000 infected cattle at least halfway through the average incubation period were eaten. This is a single estimate from Dealler (1996), which the chapter does not examine. [Background – verify: Dealler was a microbiologist who publicly criticised the government’s BSE policy, and the figure is a model projection.]

15.5 Constructing a house of cards (pp. 161–164): the thesis section#

• A convenient hypothesis (p. 161). In 1987 policy-makers adopted the hypothesis that BSE was an “innocuous version of scrapie”. They “struggled to remain wedded to it” against accumulating evidence because it allowed a message that BSE posed no threat to human health. MAFF claimed its reassurances were fully backed by science; the authors call that “a misrepresentation”.
• Private advice versus public statements (p. 161).
• Policy-makers were “repeatedly told” that certainty of no risk was impossible. The warnings came both from their own expert advisers and from “the wider scientific community”.
• In May 1990 the Spongiform Encephalopathy Advisory Committee (SEAC) told policy-makers two things: that it “would not be justified” to state categorically there was no risk to humans, and that it was “not appropriate to insist on a zero risk”. The advice cuts both ways: it rules out a claim of no risk, and it also rules out demanding zero risk.
• On 7 June 1990 the Agriculture Minister told the Commons there was “clear scientific evidence that British beef is perfectly safe” (Hansard col. 906). The reference list dates the Hansard item 8 June. The minister is not named [Background – verify: John Gummer].
• The authors’ charge is that policy-makers claimed a certainty “which they knew to be unavailable”. In this narrow sense the charge is well documented: they had been told certainty was unavailable. What the documents do not settle is what ministers believed about the size of the risk. That inference about motive is the contested part (see Limitations).
• Claims of total control (p. 161). “Occasionally” policy-makers conceded that BSE might not be an innocuous form of scrapie. When they did, according to the authors, “they always claimed” that the November 1989 controls kept all potentially contaminated material out of the food chain. The one citation is the Radio Times (1992); the reference list identifies the speaker as Keith Meldrum (issue of 31 May). “Always” is the authors’ generalisation from that example.
• MAFF scientists and expert advisers had made sure that senior policy-makers knew the controls only reduced exposure (BSE Inquiry 1999e, para. 275, “for example”).
• A 1990 draft SEAC document on beef safety, intended for the CMO, included statements “such as” these two: some edible offals that had “on rare occasions” shown low titres of infectivity were outside the ban; and no scientist could give “an absolute guarantee of safety” for British (or Irish) beef.
• The draft was circulated in the Department of Health and MAFF for amendments. The MAFF official who forwarded it told ministers that “the most potentially inflammatory pieces of drafting in earlier versions” had been edited out. The words “(including the citations above)” sit in parentheses inside that quotation. They appear to be the authors’ insertion, like “(pigs)” and “(MAFF)” elsewhere in the chapter, so the link to those specific passages is the authors’ own. Who did the editing is not stated.
• Phillips described the approach to information provision as one “whose object was sedation” (Vol. 1, para. 1179).
• The trap (p. 161). Claiming total safety made every further measure dangerous. Any new measure, “no matter how useful or cheap”, risked raising questions about the government’s reassurances, and might provoke doubts about the logic of not adopting further, more expensive controls. Zero risk was unattainable without slaughtering and restocking the whole herd and clearing the feed chain. Partial risk reduction therefore became “difficult from the point of view of presentation”. “Given the rhetoric”, the only credible options were two: draw a line and maintain it gave total protection, or try to eradicate the agent.
• Key causal claim (pp. 161–162). Many measures were rejected “not because of their immediate costs” but because they threatened the reassuring message.
• Mechanically recovered meat (MRM), September 1989 (p. 162). MRM was “widely acknowledged” to contain residual pieces of potentially highly infectious nervous tissue. Banning it would have made explicit the risk from peripheral nervous tissue, most of which could not practicably be removed. Minutes of a September 1989 MAFF meeting record that the possible danger raised by “several of those consulted” was recognised. They also record “an expression of the illogicality of what was being done”, and of how easy it would be to have to concede dangers in tissues beyond the proposed ban. They conclude: “It was agreed not to raise it”.
• Head meat, February 1990 (p. 162). The Institute of Environmental Health Officers (IEHO), whose members enforced slaughterhouse controls, told MAFF that none of the existing methods of removing brains could avoid contaminating head meat. It recommended removing head meat before splitting the skull, which the authors call “sensibl[e]”.
  • Agriculture ministers had “qualms” about the practice and felt it ought to be banned.
  • Civil servants argued “successfully” that no new controls should be introduced. A senior Meat Hygiene Division official explained to the Minister for Food that amendment regulations would fuel debate on BSE and lead to demands for similar action on spinal cords. A ban on splitting spinal columns would have “grave consequences for the industry and for the export trade”. And “Nor would it end with spinal cords”: attention would turn to nerve trunks and lymph nodes, which cannot be removed.
  • Note: here the roles are reversed from February 1988. Then officials pressed for precaution and the minister refused; now ministers wanted precaution and officials blocked it. Resistance was not confined to one level of the ministry.
  • The authors’ summary is that “a cheap, simple and risk-reducing measure was avoided” to maintain the reassuring message and “to counteract political pressure for ever more precautionary controls”.
• The authors’ counterfactual (p. 162). If MAFF had publicly acknowledged some of the uncertainties and risks, admitted that some controls might be too impractical or too expensive, and so accepted that “a lower but acceptable level of risk might have to be tolerated”, policy-makers “might have found it easier” to introduce precautionary measures without threatening the ministry’s credibility. The authors hedge this; it is untested.
• Misstated reasons undermine enforcement (p. 162). The authors say policy-makers “sometimes misrepresented their reasons for introducing regulations” to keep the narrative intact. For example, MAFF insisted the SBO ban was not necessary on scientific grounds (the citation is “see, for example” the Commons Agriculture Committee report of 1990). That made it hard to convince industry and other stakeholders that the ban mattered for public health.
• In 1995, unannounced visits by enforcement officers found “some 48 %” of the abattoirs visited failing to comply with the SBO rules.
• An enforcement representative told Panorama (1996) that the message had been that “there probably wasn’t a problem anyway”, so the rules were “maybe a bit of window dressing”.
• The causal link from the official message to non-compliance is the authors’ inference, supported by that one testimony.
• Curating advice and excluding institutions (pp. 162–163).
• Expert advisers were “carefully selected”. Those who did not share the ministry’s policy framework, or who might refuse to accept restrictions on disclosure, were excluded. The evidence is one official’s retrospective Inquiry testimony: one has to turn to external bodies “to try to give some credibility to public pronouncements”, and “Really the key to it is setting up the Committee, who is on it, and the nature of their investigations”. The chapter names no individual expert who was excluded.
• The Public Health Laboratory Service (PHLS), the UK’s established surveillance body for new and emerging diseases, was “always excluded” from BSE policy. The Welsh CMO recalled that opposition to involving it rested on anxiety that this would be “tantamount to admitting the possibility of a human health risk”.
• Controlling research (pp. 163–164). Senior policy-makers “attempted to ensure” that the commissioning, conduct and reporting of research were tightly controlled:
• Many key experiments were never started or were “seriously delayed”; information and evidence were “sometimes withheld”; and data and materials were not always shared.
• A cheap random abattoir survey of sub-clinical infection was never run for food-chain cattle. The only UK survey was in 1999, on cattle excluded from food.
• A scrapie-to-cattle feeding experiment began only in 1996.
• MAFF owned all infected brains after notification and showed “extraordinary reluctance” to supply senior US scientists.
• Principle: precaution means producing and sharing more information (p. 164).

Box 15.2 The European dimension (p. 163)#

• Spread and divergence. BSE spread through trade in animals and feed.
• Ireland, Portugal and France controlled domestic production in the 1990s.
• Belgium, the Netherlands and Italy focused on traded goods.
• Responses differed irrespective of development level or case numbers.
• The institutional claim. The authors say a full account of how and why national responses differed is “beyond the scope of this case study”. Nevertheless, where responsibility sat was “fundamental” to how the problem was defined, evaluated and responded to. “In general”, jurisdictions combining industrial sponsorship with consumer protection (e.g. Ireland) were less precautionary than those that split or shared them (e.g. Austria). The source is the BASES project. The footnote marker is printed “(19)”; the rendered page shows no other in-text marker for fn 13, whose text (BASES national reports) sits at the foot of p. 163, so “(19)” is evidently a misprint for fn 13. No data are shown.
• The Commission. Before 2000 BSE sat with DG III (internal market) and DG VI (agriculture).
• The UK notified its June 1988 feed ban. EU-wide legislation followed six years later.
• The UK kept exporting contaminated feed. MBM exports to the EU rose from 12,553 t (1988) to 25,005 t (1989), and some was fed to cattle.
• The UK refused a Commission request for an export ban in summer 1989.
• The Commission later claimed it had lacked the legal basis “prior to the Single European Act” to ban UK MBM exports itself, a claim the European Parliament disputed. Instead it invited Member States to ban imports nationally. The Netherlands already had a ban; Portugal did not ban UK MBM until “much later”.
  • [Background – verify: the Single European Act entered into force in 1987, before the 1989 request, so the box’s phrasing is puzzling. It may refer to completion of the single market in 1993. Check against European Parliament (1996).]
• An EU-wide ban on feeding MBM to ruminants came in 1994. The Commission banned exports of UK MBM “(and all other cattle products)” in 1996.
• The European Parliament concluded that the Commission “consistently subordinated” animal and public health to maintaining the internal market.
• Reforms after 1996.
• Scientific advice moved to DG XXIV, renamed DG SANCO. In 2000, regulatory functions were split from sponsorship and moved to DG SANCO.
• The 2000 White Paper proposed a European Food Authority. Member States made similar reforms.
• Status in 2000. Cases were rising in France and Ireland, and first cases had appeared in Germany and Spain. Some jurisdictions would face challenges “for many years”.

15.6 The failures and eventual collapse of the policy edifice (p. 164)#

• Why the narrative unravelled after Southwood:
• the evidence grew less reassuring
• MAFF’s control of information was imperfect
• outside actors made their own decisions

The authors add that this happened “despite MAFF’s vigorous attempts to sustain the policy in the face of its inherent weaknesses and accumulating counter-evidence”. - Reactive tightening. “By late 1995 a lengthy series of events and evidence” had obliged MAFF to tighten regulations progressively, “in a reactive, rather than in an anticipatory, fashion”. This describes cumulative tightening over the years before late 1995, not a single episode in late 1995. By then a growing share of the national and global scientific community was concerned about BSE. - Collapse, March 1996. Policy collapsed when vCJD emerged and SEAC judged BSE in food its most probable cause. - Note: the section is very brief. The tightening between 1990 and 1995 and the events of 1996 are not described.

Box 15.3 The costs of BSE (p. 164)#

• Framing. The box says the ramifications are “as yet incomplete”, and that some can be put in money terms while others are “incalculable”. It calls precise total estimates “premature”.
• Public.
• In 1998, UK agriculture department spending on the crisis for 1996–2001 was forecast at £4.2bn (Public Accounts Committee 1999). This is a forecast, not an outturn. Most of it compensated commercial enterprises (farmers for removing over-30-month cattle; slaughtering and rendering), plus research and administration.
• The inquiry cost about £25m.
• Private.
• The export ban cost a £700m a year trade.
• The UK beef market fell about 36% in real terms in the first year, a loss of about £1.15bn in value added (DTZ Pieda 1998).
• Human. vCJD was unknown in scale: “no more than another 100 cases” or “up to a million” (Collinge 1999).
• Notes:
• The chapter gives no count of BSE cases in cattle or of vCJD cases up to 2000.
• Most public money went to private enterprises.
• The implicit contrast with the £1.5m option (p. 158) is not like-for-like (see Limitations).

15.7 Conclusions (pp. 164–166)#

• Concession (p. 164). BSE was always going to be hard, because infected cattle had been eaten before recognition and eradication was very expensive.
• Verdict (pp. 164–165).
• “A great deal” could have been done to cut the risks to consumers, and in the long run to the meat industry and the Exchequer, “especially if” government had openly acknowledged what the evidence did and did not show.
• Instead government claimed to be prudently protecting public health while in practice it “covertly subordinated” health to supporting agricultural sales, with a view also to minimising state intervention and public spending.
• Regulations were “too little and too late” and not properly enforced. Too little was invested in research, and the involvement of independent scientists was “actively discouraged”.
• What genuine precaution would have required, and what might have helped (p. 165). Only items 1–2 are framed as what precaution “would have necessitated”. Items 3–5 are hedged (“might”). 1. “Firstly and most fundamentally”: separating regulation from sponsorship. Since May 1997 government has acknowledged a “fundamental contradiction” in MAFF’s remit, and that recognition “informed” the decision to create the Food Standards Agency (FSA). 2. Acknowledging how little “sound science” was available, and holding open, accountable discussion of the costs and benefits of taking, or failing to take, a wide range of actions. The quotation marks around “sound science” are the authors’. They plausibly signal distance from official usage; the chapter does not say. 3. A Freedom of Information Act “might” start a cultural change in how policy bodies represent scientific evidence. 4. Separating the bodies that provide scientific advice, risk assessment and research from regulatory policy-making “might also have encouraged” more open and robust discussion of risk. 5. The more support for research across many disciplines and institutions with open access to evidence and data, the harder uncertainties would be to conceal, and so the more readily they “might” be reduced.
• Science advisers’ acquiescence (p. 165).
• The authors name one factor that, before March 1996, helped MAFF sustain its narrative. Expert advisers were willing “to acquiesce” in giving advice based on both scientific and non-scientific considerations, which was then presented to the public “as if it was purely scientific”. This suited ministers, who could say they were doing what, and only what, their advisers recommended. It also “flattered the scientists” by presenting them as authoritative and influential.
• Remedy: the advisers’ role should be “more strictly delineated” and open to expert and public scrutiny. Risk assessments are always framed by socio-economic considerations, so democratically accountable ministers should articulate and justify the framing assumptions, and advisers should show how they gathered and interpreted “all the appropriate evidence”.
• Improvement (pp. 165–166).
• Since March 1996, members of SEAC have been “more independent than their predecessors”. On “beef on the bone”, SEAC set out the possible consequences of various courses of action and explicitly identified the decisions ministers would have to take. The chapter praises the form of this advice. It does not discuss the ban that followed.
• BSE policy has become “more precautionary”, partly through greater openness and the FSA.

Box 15.4 Conclusions from the Phillips Inquiry (p. 165)#

• The inquiry. Set up by the Labour government in January 1998 and chaired by Lord Justice Phillips, it reported in October 2000 in 16 volumes, after taking oral and written evidence from more than 600 witnesses. Its terms of reference covered action “up to 20 March 1996”. It was to judge the adequacy of the response “taking into account the state of knowledge at the time”.
• What it focused on. Procedure more than outcome. It was reluctant to judge proportionality, and emphasised communication, expert advice and cooperation between departments.
• Main criticism. Policy was dominated by the political aim of reassuring the public. Yet it found the decisions appropriate, if not always timely, well designed, or properly implemented and enforced.
• Where the authors disagree. They dispute two findings: 1. that government was anxious to act in the interests of human health 2. that MAFF did not favour producers over consumers

They find it “difficult to see” how the evidence available to the Inquiry, some of it summarised in the chapter, is consistent with these findings. The wording is a hedged rebuttal, not a flat statement that the evidence contradicts them. - Phillips’ 160 lessons. The generic ones cover: - advisory committees - in-house expertise - cooperation between animal and human health - enforcement - research to reduce policy-relevant uncertainty - “uncertainty can justify action” - trust - communicating uncertainty - openness

None discusses explicitly what precaution means in practice. - Analytic note: the official inquiry agreed with much of the diagnosis: reassurance dominated, and action was late and poorly enforced. The disagreement is about intent, and about whether the decisions were appropriate. [Background – verify: Phillips reportedly concluded that government “did not lie”, believed the risk to be remote, and that its “campaign of reassurance” was a mistake.]

Table 15.1 (p. 166, “Source: EEA”)#

• Contents. An 11-row timeline running from the 1970s US action to the Phillips Inquiry. The 1996 scrapie-feeding experiment is introduced with “At last”. The Phillips conclusions are judged not “sufficiently rigorous on judging government actions over time”.
• Discrepancies with the text: 1. The 1989 row groups the ruminant feed ban with the slaughter of affected cattle and the SBO ban. This echoes p. 160 (“all in place by the end of 1989”), so it is not strictly an error. But a reader could take 1989 as the feed-ban date, and the ban was announced in June 1988 (pp. 160, 163). 2. The US action is described as banning “scrapie-infected” sheep and goat meat from the “cattle food chain”. Box 15.1 has carcasses “afflicted with, or exposed to, scrapie” kept out of “human or animal foods”.

Other minor errors#

• Hansard is dated 7 June in the text (p. 161) and 8 June in the references (p. 167).
• “Sheppard” appears for Shephard.
• A footnote in Box 15.2 is misnumbered “(19)”. The chapter has only footnotes 11–13, and fn 13 (p. 163) covers the same BASES reports.
• A quotation is marked “emphasis in original” with no visible emphasis (p. 161). Font extraction and a page render confirm the passage has no italics.
• The Welsh CMO quotation on the PHLS opens a quotation mark but never closes it (p. 163), so where the quoted words end is not marked.
• Arithmetic: about 1,200 cases at no more than £1,000 each gives about £1.2m, not the stated £1.5m (p. 158).
• Possible citation mismatch: the Meat Hygiene Division advice on head meat and spinal cords (p. 162) is cited to BSE Inquiry 1999f, whose title covers MRM in 1989. The MRM minutes are cited to 1999e, the SBO account. This may be correct, since the accounts could cover both topics, but it is worth checking.

Case timeline#

Lags (computed from the chapter’s own dates):

What was known when: - 1986–88: the possibility of human risk; unpredictable host range after a species jump; no live test (p. 157). - 1988: official acknowledgement, by analogy, that transmission to humans was possible; ignorance about susceptibility in pigs (pp. 158, 160). - 1988–90: the controls only reduced exposure; MRM and head meat were contamination routes (pp. 160–162). - From 1990: transmission to cats (p. 158). - 1995–97: evidence in humans (p. 158).

The authors’ own lessons and conclusions#

(a) Lessons derived from the evidence 1. Policy-makers’ later claims of an ultra precautionary, science-based approach are “not convincing” against the documentary record (pp. 157, 159, 164). 2. A dual promote-and-protect remit subordinated health to sales, spending and exports, and in the long run failed industry and the Exchequer too (pp. 157, 164). A parallel at EU level is drawn (p. 163). 3. Reassurance undermined precaution. Categorical safety claims made further cheap measures politically dangerous, so they were refused for presentational rather than cost reasons (pp. 161–162). 4. Controls were designed around convenience and commercial value, then presented as total protection (pp. 160–161). 5. Calling a rule unnecessary undermined its enforcement (p. 162). 6. Expert advice was curated to legitimise policy: through membership (p. 162), exclusion of the PHLS (p. 163), and editing of drafts (p. 161). The note-taker adds Southwood’s retrospective “no-goer” remark (p. 160) as possible self-limitation by advisers; the authors do not label it so. 7. Advisers acquiesced in giving mixed scientific and non-scientific advice that was presented as purely scientific. This suited ministers and flattered the scientists, and it helped sustain the narrative (p. 165). 8. Restricting knowledge and delaying research sustained ignorance, which sustained reassurance (pp. 159, 163–164). 9. The strategy was brittle. It collapsed once evidence and outside actors escaped control (p. 164). 10. Cheap early options were forgone. The comparison of £1.5m against billions is implicit; the chapter does not draw it explicitly (pp. 158, 164). 11. The Phillips Inquiry’s conclusions on the government’s concern for health, and on MAFF not favouring producers, are hard to reconcile with its own evidence (p. 165). Separately, the EEA’s Table 15.1 says Phillips’s conclusions “do not seem sufficiently rigorous” (p. 166).

(b) Recommendations and advocacy 1. Separate regulation from sponsorship: “firstly and most fundamentally” (p. 165). 2. Acknowledge the limits of “sound science” and deliberate openly and accountably over a wide range of options, including inaction (p. 165). 3. A Freedom of Information Act “might” change how evidence is represented (p. 165). This and the next two items are hedged. 4. Institutionally separating science advice, risk assessment and research from policy-making “might” encourage more open discussion of risk (p. 165). 5. Fund plural, interdisciplinary, open-data research (p. 165). 6. Assign roles: ministers state and justify framing assumptions; advisers show how they gathered and interpreted all the evidence; both are open to scrutiny (p. 165). 7. Communicate a tolerable level of residual risk rather than zero risk (p. 162). This is framed as a counterfactual. 8. Treat precaution as producing and sharing more information (p. 164). 9. The evidence offered that these reforms work: SEAC on beef on the bone, and the FSA (p. 166). This is early and thin.

Mechanisms and dynamics#

1. Conflict of interest: one body both sponsors and regulates. This is the root diagnosis (p. 157). - Costs to the ministry’s constituency were immediate and concrete: - exports (p. 159) - profits (p. 159) - renderers’ markets (p. 160) - abattoir costs (p. 160) - public-expenditure “offsetting savings” (p. 159) - Health benefits were uncertain, diffuse and deferred. - The same pattern is claimed at EU level, where the internal-market and agriculture DGs held the brief and the European Parliament found health subordinated (p. 163). It is also claimed across countries, with Ireland compared to Austria (p. 163), but that comparison is thin. - The irony: sponsorship failed the sponsored industry (a £700m-a-year export trade lost, and a 36% real fall in the value of the UK beef market in the first year) and the public purse (a forecast £4.2bn in spending, mostly compensation to the industry) (p. 164).

2. The reassurance trap (the chapter’s signature mechanism). The authors describe its elements (pp. 161–164). The eight-step formalisation below is the note-taker’s. It is a self-reinforcing loop: 1. Adopt the institutionally convenient hypothesis (innocuous scrapie, p. 161). 2. Publicly convert “no evidence of risk” into “scientific evidence of safety”, while acknowledging uncertainty in private (pp. 157–158, 160, 161). 3. Fall back on “the controls stop everything” (p. 161). 4. Every further measure now implies the claim was false and invites “why stop there?”. Because zero risk requires eradication, the continuum of options collapses to two poles, and graded precaution becomes unpresentable (p. 161). 5. Cheap measures are refused for presentational reasons: MRM was “agreed not to raise”; head meat was refused because “Nor would it end with spinal cords” (p. 162). 6. Measures that are adopted are described as unnecessary, so enforcement erodes (48%; “window dressing”; p. 162). 7. Activities that would generate knowledge become threats and are avoided: involving the PHLS (p. 163), surveys, experiments and sharing materials (pp. 163–164). Ignorance persists, and “no evidence” stays available. The earlier resistance to notification and information-sharing in 1986–88 (p. 159) came before the public reassurance narrative took hold. It was driven by fear for exports and consumer confidence (mechanism 3), not by the trap itself. 8. The loop collapses when outside evidence cannot be contained (p. 164).

This is path dependence in commitment and communication. An early public position raises the political price of every correction (“house of cards”, p. 161; “policy edifice”, p. 164).

3. Action read as a signal. Measures were judged by what they would communicate to consumers and importers, not by how much risk they removed. Examples: - notifiability and information-sharing (p. 159) - slaughter, seen through its effect on exports (p. 159) - involving the PHLS, which would be “tantamount to admitting” risk (p. 163) - an MRM ban, which would make the nerve-tissue risk “explicit” (p. 162)

The authors’ remedy is to decouple action from alarm by openly tolerating some risk (p. 162). It is untested.

4. Uncertainty and ignorance management. - Uncertainty was clearly understood internally (pp. 157–158, 160, 161). In public it was converted into certainty (p. 161). - Selective analogy (the note-taker’s synthesis; the authors do not name this mechanism). The scrapie analogy was used to reassure. Analogies that alarmed were set aside: other TSEs’ tissue distribution and ewe-to-lamb transmission (p. 160), and the NIBSC link to CJD (p. 158). Known evidence that host range is unpredictable after a species jump (Kimberlin, p. 157) undercut the reassuring analogy. - Ignorance was produced by not looking. Research was delayed, there was no random surveillance, the surveillance body was excluded, and materials were withheld (pp. 161, 163–164). The quotation marks the authors put around “sound science” (p. 165) can be read as marking the irony of claiming a scientific basis while limiting the science. That reading is the note-taker’s.

5. Burden and standard of proof. Decisive action waited for near-direct human evidence (1995–96). Analogy (1988) and cross-species evidence (1990) were not enough (pp. 158, 164). The authors do not use burden-of-proof language, but the pattern is clear. Phillips’ lesson that “uncertainty can justify action” (p. 165) is the stated counter-principle. Note, though, that partial measures were adopted in 1988–89 without proof of harm.

6. Expert advice as legitimation. - The CMO was brought in because ministers wanted his support for a decision not to act (p. 160). - Committees were composed to fit the ministry’s framework (p. 162). - Drafts were edited (p. 161). - Advisers may have limited themselves to what was “politically feasible” (the authors’ paraphrase of Southwood’s retrospective remark, p. 160). - Advice was presented as pure science when it was not, which gave ministers cover and flattered the scientists (p. 165). - The remedy is a clean split: ministers own the framing, advisers own the evidence (p. 165), as SEAC did after 1996 (p. 166).

7. Industry is not monolithic. - Renderers’ markets shaped the ruminant-only feed ban (p. 160). - Exporters’ and abattoirs’ interests shaped the tissue list, the calf exemption and the head-meat decision (pp. 159–160, 162). - But pet-food and meat-products firms moved ahead of the regulator, and ministers acted to avoid being “upstaged” (p. 160). - An adviser’s private advice to industry was more precautionary than public policy (p. 160). - Note-taker’s hypothesis, not the chapter’s: firms exposed to reputational or liability risk can be more precautionary than a regulator shielding a sector. The chapter does not analyse the firms’ motives.

8. Cost logics and who bears the cost. - Short-horizon fiscal and deregulatory thinking: offsetting savings and the sugar beet precedent (p. 159), minimising intervention (p. 164), and the 1979 withdrawal of rendering standards as “unnecessary and excessively restrictive” (p. 158). - The later bill: large and mostly public, paid mainly to compensate private firms (p. 164). - Who bore the risk: - UK consumers: most of the 1,200 early clinical cases and about 30,000 infected cattle were eaten (pp. 158, 161) - other European countries, through feed exports that rose after the domestic ban (p. 163) - future vCJD victims, in unknown numbers (p. 164) - Who was protected early: renderers’ and exporters’ markets (pp. 159–160). A government adviser privately recommended a bovine offal ban in pet food before the public ban on offal in human food (p. 160). The chapter does not say when pet-food manufacturers actually removed offal, and the meat-products industry was also planning to remove it from human food. - The authors present the split of costs between public and private sources as a matter for political judgement, not science (p. 159), whose framing accountable ministers should state and justify openly (p. 165).

9. Implementation and frontline knowledge. - The IEHO spotted a contamination route and proposed a simple fix, which was blocked (p. 162). - Enforcers read official messaging as meaning the rules were cosmetic (p. 162). - Non-compliance is attributed to the narrative. Other causes, such as resources and inspection regimes, are not examined.

10. Information control is brittle. - There was secrecy between 1986 and 1988 (p. 159). - Officials changed their views in February 1988 as media attention grew. The chapter juxtaposes the two and does not claim causation (p. 159). - Control was imperfect, and outside actors acted independently (p. 164). - Collapse was sudden rather than graduated: an immediate export ban and a 36% fall in the market (p. 164).

11. Latency and irreversibility. - Exposure had already occurred before recognition (p. 164). - Delay added irreversible exposure (p. 161). - Partial bans prolonged the epidemic through leakage (p. 160). - Proof took about ten years to arrive (p. 158). - Final harm could not be known in 2000 (p. 164). - The chapter leaves irreversibility implicit rather than theorising it.

12. Governance across levels and trade. - A domestic ban was combined with continued export (p. 163). - Ambiguous legal competence delayed EU action (p. 163). - Member States acted unevenly (p. 163). - Uneven controls and enforcement across Member States left some jurisdictions facing challenges “for many years”, with cases rising in several countries by 2000 (p. 163). The note-taker’s gloss is that the weakest regulatory link shapes exposure across a trading network. Beyond the UK export tonnages, the chapter does not quantify cross-border exposure.

13. Blame avoidance cuts both ways. - Officials’ 1988 argument for slaughter was about avoiding future blame (p. 159). - Ministers sought the CMO’s support for not removing clinical cases (p. 160). - The offal ban came about partly to avoid being “upstaged” (p. 160). - Whether anticipated blame drives precaution or legitimises inaction depends on who is expected to assign blame, and when.

14. Mental models of the people in charge. - The public as irrational. The first CVO feared that “irresponsible or ill-informed publicity” might lead to “hysterical” demands (p. 159). Phillips described information policy as aimed at “sedation” (p. 161). The implication is that information had to be managed, not shared. - Risk as a problem of trade and reputation. The first framing at the top was about exports (p. 159), and the minister’s reasons were money, precedent and exports (p. 159). - Zero risk against acceptable risk. Officials treated “zero risk” as the only viable message, although SEAC said insisting on zero risk was inappropriate (p. 161). The zero-risk frame came from the ministry’s public stance, not from its advisers. - Seeing the logic and looking away. The Meat Hygiene official and the MRM meeting both anticipated that any new control would draw attention to tissues that cannot be removed, such as nerve trunks, lymph nodes and peripheral nervous tissue (p. 162). The documents quoted show they foresaw the logic; they do not say how likely those officials thought infectivity was. The failure was a decision not to confront the trade-off publicly, not a failure of comprehension. - A split between private and public confidence. The clearest pairing: in May 1990 SEAC advised that “no risk” could not be stated categorically, and in June 1990 the minister told the Commons beef was “perfectly safe” (p. 161). Looser pairings, which juxtapose different people, topics or years, are the note-takers’: - Keith Meldrum privately said in 1988 that pig susceptibility was “unknown” (p. 160). In 1992 he was cited for the claim that the controls excluded all contaminated material (p. 161; speaker identified from the reference list). - The Under Secretary acknowledged uncertainty in 1988 (pp. 157–158). - Precaution as a label applied afterwards. Policy-makers later described their approach, in the authors’ paraphrase, as ultra precautionary (p. 157). - Resistance at more than one level. In February 1988 officials urged compensated slaughter and the minister refused (p. 159). In February 1990 ministers wanted a head-meat ban and civil servants blocked it (p. 162). The protective instinct and the resistance to it were not tied to one rank. - Deregulatory ideology in 1979. Leave it to industry (p. 158).

15. Framing and language. - Officials’ words: - “perfectly safe” - “hysterical” and “draconian” - “inflammatory” drafting - “a no-goer” and “pretty revolutionary” - “window dressing” - Phillips’ “sedation” - The authors’ words: - “misrepresentation” - “house of cards” - “covertly” - “sound science” in scare quotes - The title casts reassurance itself as the cause of failure, rather than ignorance or economic interest alone.

16. Learning and reform. - The UK created the Food Standards Agency and made SEAC more independent (pp. 165–166). - The EU moved responsibility to DG SANCO and proposed a European Food Authority (p. 163). - The Phillips Inquiry is a formal act of institutional learning (p. 165), though the authors say it pulled its punches on intent. - As of 2001 it was too early to evaluate the reforms.

Transferable insights (technology-neutral)#

1. When one body both promotes an industry and regulates its hazards, protection tends to give way to promotion, and the arrangement can fail both aims. pp. 157, 159–160, 164–165 (EU level p. 163). Strong for this case, moderate as a generalisation. The case rests on internal minutes and on the government’s own admission in creating the FSA. The cross-jurisdiction claim rests on two examples and data that are not shown.
2. Categorical public reassurance creates a commitment trap. Once “safe” has been asserted, each further protective step implies the claim was false and invites “why not more?”. Cheap, proportionate measures are then refused for presentational reasons. pp. 161–162. Strong. Contemporaneous minutes record presentation as a reason. In the note-taker’s view this is the chapter’s most original and best-documented contribution. One qualification: the head-meat advice also cites “grave consequences for the industry and for the export trade”. Presentation and cost were intertwined, not cleanly separable as the authors’ “not because of their immediate costs” suggests.
3. Absolute-safety rhetoric collapses a continuum of options into “hold the line” or “do everything”. Openly admitting a tolerable residual risk keeps graded responses available. pp. 158, 161–162. Moderate. The collapse is documented; that honesty would have worked is a counterfactual.
4. When protective actions are judged as signals (“acting implies danger”), even near cost-free surveillance and information-sharing get suppressed. pp. 159, 162, 163. Strong within the case. There are several independent, documented instances.
5. Controls built around what is convenient or commercially cheap to remove, rather than where the hazard is, leave residual exposure. The damage is worse when such controls are presented as complete. pp. 160–161. Strong. Documented in the Inquiry’s factual accounts.
6. Partial restrictions that leave a hazardous material circulating in neighbouring uses or markets create leakage routes. They prolong exposure and push it elsewhere. pp. 160, 163. Strong on the facts (export tonnages), moderate on the magnitude. The claim that cross-contamination “greatly prolong[ed]” the epidemic is uncited.
7. Understating why a rule exists erodes compliance and enforcement. p. 162. Moderate. One statistic (48%) and one testimony. Alternative causes of non-compliance are not examined.
8. Controlling who advises and what they may say turns expert advice into legitimation. Presenting value-laden advice as “purely scientific” hides the framing and benefits both politicians and experts. pp. 160–163, 165. Strong on the facts, moderate on the “mutual benefit” interpretation, which is the authors’ reading. The PHLS exclusion and the edited SEAC draft are well documented; the claim that advisers were selected for their fit with the ministry’s framework rests on one official’s retrospective testimony, with no excluded expert named.
9. Advisers pre-filter their advice by what they expect decision-makers to accept. p. 160. Suggestive. One retrospective remark (Southwood), though a telling one.
10. “No evidence of harm” can be partly produced by not looking: not commissioning surveillance and experiments, and not sharing data or materials. Precaution therefore includes deliberately producing and sharing knowledge. pp. 161, 163–164. Moderate. The omissions are well documented; their effect on policy is counterfactual.
11. Analogies with known hazards get used selectively: invoked when they reassure, dropped when they alarm, even when there is evidence that the new hazard may not behave like its analogue. pp. 157–158, 160–161. Moderate. This is my synthesis from the chapter’s evidence, not a mechanism the authors name.
12. Science cannot set the level of protection. Where to act, and who pays, are political choices that accountable decision-makers should make openly, with experts responsible for showing the evidence. pp. 159, 165–166. Asserted. A well-argued normative principle with one early illustration.
13. When harm is delayed and irreversible, the problem is already large when first recognised. Each further delay adds exposure that cannot be undone, and the eventual scale stays unknowable for years. pp. 158, 161, 164. Strong as a dynamic, though the chapter leaves it implicit. The one quantified measure of added exposure (30,000 cattle) rests on a single estimate.
14. A strategy that depends on controlling information is brittle. Leaks, outside actors and accumulating evidence break it, and the correction is sudden and severe. pp. 159–160, 164. Moderate. The chapter does not separate how much of the 1996 shock came from prior reassurance and how much from the vCJD finding itself.
15. Industry is not uniform. Some downstream firms can move ahead of a regulator that is protecting a vulnerable sector, and can force its hand. p. 160. Suggestive. One episode. The chapter does not analyse the firms’ motives; reputational or liability exposure is the note-taker’s conjecture.
16. Frontline practitioners often hold specific knowledge of exposure routes, and it is lost when it conflicts with the official line. p. 162. Suggestive to moderate. One well-documented instance.
17. When decision-makers describe their past conduct as “precautionary” or “science-based”, test the claim against the contemporaneous record. p. 157 against pp. 159–164. Moderate. Clear in this case; a general methodological caution.
18. Minimising short-term cost, including keeping fiscal consistency with unrelated sectors, can reject trivial early costs and incur later costs several orders of magnitude larger, borne mainly by the public purse and the exposed. pp. 158–159, 164. Suggestive. The early measure would not by itself have averted the later costs, so the comparison illustrates a mindset rather than measuring what was forgone.
19. Structural remedies: separate regulation from sponsorship and advice from policy, and add freedom of information and plural, open-data research. pp. 163, 165–166. Asserted or suggestive as of 2001. Advocacy grounded in the diagnosis, with thin early evidence. A prime target for hindsight checking.

Limitations, contestation and bias check#

Advocacy. - The prose is prosecutorial (“misrepresentation”, “covertly”, “knew to be unavailable”). The only counter-view is Phillips. The authors report its main criticism, which agrees with theirs, but dispute two of its conclusions in a sentence without engaging the Inquiry’s reasoning on them (p. 165). - Intent is the most contested claim. There are two levels to it: - Narrow charge, well supported. Policy-makers claimed a certainty they “knew to be unavailable” (p. 161). The documents show they were told certainty was impossible (pp. 157–158, 161), and that the controls were known to be partial (pp. 161–162). - Broader charge, not settled. Health was “covertly subordinated” to agricultural sales (p. 164). The documents do not show that ministers believed the risk was substantial. - A charitable reading fits the same evidence: officials believed the risk was remote (as MAFF vets did in 1986, p. 158) and overstated their confidence to avert what they saw as a disproportionate panic. - [Background – verify: Phillips’ reported findings that government “did not lie” and believed the risk remote support this reading.] - The conclusion that health was “covertly subordinated” (p. 164) is stronger than the evidence strictly establishes. The documented core is well supported: presentation and export concerns repeatedly overrode risk reduction. - Cost and presentation are not cleanly separable. The authors claim measures were refused “not because of their immediate costs” but for presentational reasons (pp. 161–162). Yet their own head-meat evidence cites “grave consequences for the industry and for the export trade” (p. 162), and the MRM minutes do not record the weight given to cost. Presentation was clearly a reason. That it was the reason, rather than cost, is not fully shown.

Thin or secondary evidence. - The Ireland–Austria comparative claim is asserted without data (p. 163), possibly from the authors’ own project. - The claim that the epidemic was “greatly” prolonged has no specific citation (p. 160). - The 30,000-cattle figure is a single estimate (p. 161). [Background – verify: Dealler’s figure is a model projection, and he was a prominent critic of government policy.] - The £12–15bn eradication cost is a press-based maximum (p. 158). - The claim that advisers who did not share the ministry’s framework were excluded (p. 162) rests on one official’s retrospective testimony. No excluded individual is named. - Much of the key evidence is retrospective testimony given at the 1998 Inquiry, after the 1996 crisis. That includes the CVL confidentiality recollection, the “setting up the Committee” remark, the Welsh CMO on the PHLS, and Southwood’s 1996 remark. This testimony is more exposed to hindsight than the contemporaneous minutes. - There are minor internal inconsistencies: - Table 15.1 groups the feed ban under 1989, which could be misread as its date (it was announced June 1988) - £1.5m stated against 1,200 × ≤£1,000 (p. 158) - the US action is described differently in Table 15.1 and Box 15.1 - the Hansard date differs between text and references - “Sheppard” - a misnumbered footnote - an unmarked “emphasis in original”

Counterfactuals and hindsight. - Partly guarded. Decisions are mostly judged against contemporaneous advice and knowledge: SEAC in 1990, NIBSC in 1988, Kimberlin in 1987, the IEHO in 1990, and the CVO’s private admission. The chapter also concedes deep uncertainty and the difficulty of the problem (pp. 157, 164). This is much sounder than judging by outcome alone. - Untested counterfactuals remain. - Would openness about tolerable risk in 1988–90 have allowed stronger measures without an earlier market collapse (p. 162)? The chapter does not weigh the possibility that honesty would simply have brought the 1996 shock forward. - Would the MRM and head-meat measures have materially reduced human exposure? No estimate is given. - £1.5m against £4.2bn is not like-for-like. Removing clinical cases leaves sub-clinical infection untouched, and the authors concede that only eradication reaches zero risk (pp. 158, 161). - Effectiveness not assessed. The chapter never says whether the leaky 1988–89 controls bent the epidemic curve. [Background – verify: UK cases reportedly peaked around 1992–93 and then declined, largely attributed to the 1988 feed ban.] The UK also acted years before the EU (p. 163). “Too little and too late” may understate the partial effectiveness of the measures.

Omissions. - There are no counts of BSE cases in cattle or of vCJD cases up to 2000. - The Southwood Report’s own reassuring conclusion is not discussed. [Background – verify: that human health implications were “most unlikely”, with a caveat about extremely serious consequences if that was wrong.] It complicates the image of advisers who were consistently cautious but overruled. - Named dissenting scientists and disputes about the nature of the agent are absent (though Dealler is cited). - Media and public dynamics are barely covered (p. 159). [Background – verify: the Gummer beefburger photo-call of 1990 is not mentioned.] - Proportionality after 1996 is unexamined. SEAC’s options-based advice on beef on the bone is presented as better practice (p. 166). The chapter praises the form of the advice and does not mention or assess the resulting ban. [Background – verify: the ban (December 1997) was widely criticised at the time as disproportionate.] The over-30-month scheme’s costs are shown but not assessed. - The conditional counterfactuals in Box 15.1 depend on the unresolved question of BSE’s origin (p. 157). [Background – verify: Phillips reportedly favoured a novel bovine source over scrapie; its view on changes to rendering practice needs checking.]

How the precaution framing shapes the conclusions. - Much of the substance concerns honesty, transparency, independence, enforcement and research capacity. These are good-governance lessons that a non-precautionary risk framework, and Phillips itself (p. 165), would also endorse. “Precaution” here is broad and procedural: acknowledge uncertainty, deliberate openly, invest in knowledge, separate roles. - The case strongly supports that procedural precaution. It says little about how strong or proportionate substantive precaution should be. - Selection. BSE is unusually well documented (an inquiry published internal papers) and unusually damning. Lessons from it may be over-weighted relative to cases where reassurance proved justified.

Fair to the chapter. - The authors explicitly reject maximal precaution (pp. 158–159, 164) and accept that science cannot choose the policy (p. 159). - They rest mainly on primary documents from an official inquiry. - They record the institutional learning that had happened by 2000 (pp. 163, 165–166). - The reassurance-trap mechanism is evidenced by officials’ own words and does not depend on hindsight.

Notable quotes#

Short phrases; see the page for context.

Open questions#

1. Intent or belief? Did ministers believe the risk was remote and overstate their confidence (closer to Phillips), or knowingly assert a certainty they thought false (the chapter, p. 161)? What exactly do the Phillips volumes conclude, and how did the authors’ 2005 book respond?
2. Did the structural remedies work? Consider: - the FSA (2000) - [Background – verify: MAFF’s replacement by DEFRA in 2001] - [Background – verify: EFSA in 2002] - [Background – verify: the FOI Act, passed 2000 and in force 2005] - SEAC’s options-based advice

Did they produce better or more precautionary governance, or new tensions between risk assessment and risk management? 3. How did the actual vCJD toll compare with the 2000 range of 100 to a million (p. 164)? [Background – verify: a UK total of 178 is commonly cited, with transmission via blood transfusion and estimates of sub-clinical carriers from appendix surveys.] What does the gap imply for the chapter’s cost argument? 4. How far did the leaky 1988–89 controls reduce the epidemic and human exposure? What explains infection in cattle born after the feed ban? 5. Does systematic comparative work support the Ireland–Austria institutional claim (p. 163)? 6. Would candour about tolerable risk in 1988–90 have been commercially survivable? Do other countries that communicated uncertainty early offer evidence either way? 7. What did post-1996 rendering-deactivation research (p. 158) show about the 1979 counterfactual, and what is the current view on BSE’s origin? 8. Were measures after 1996 proportionate (beef on the bone, the over-30-month scheme)? Can the chapter’s framework recognise over-precaution, or does “more open” simply track “more precautionary”? 9. What were the final public and private costs, UK and EU, against the £4.2bn forecast for 1996–2001 (p. 164)? 10. For cross-chapter synthesis: does the reassurance trap, where categorical safety claims make incremental precaution self-incriminating, recur in other Late Lessons cases?

Audit log#

Independent audit against the text extract (pp. 157–167, read in full) and the PDF (font runs checked on pp. 157–166; pp. 161, 163 and 166 rendered and inspected). Most of the existing notes checked out: quotations, figures, dates, names and page references were accurate, and the italics and “emphasis in original” observations were confirmed. Changes made:

1. Evidence base: replaced “The remaining sources are” (an incomplete list) with “Other sources include” and added the omitted sources (Barclay 1996, Cooke and Martin Inquiry statements, RCEP 1979, BASES, EC White Paper, Nature, Farmers Weekly).
2. Evidence base: the “critical scientist” label for Dealler is not in the chapter; moved it into a [Background – verify] flag.
3. Evidence base: the claim that the Inquiry’s factual accounts were compiled by Inquiry staff from government files is not in the chapter; flagged it as background. Noted that 1999f is a draft factual account.
4. Evidence base: added the caveat that some items presented as contemporaneous (Meldrum June 1988; SEAC May 1990) are cited only to 1998 oral-evidence transcripts.
5. Stance: the chapter names Shephard and Gummer but not their offices; “former agriculture ministers” moved into a [Background – verify] flag (notes and digest).
6. Components: added a note that Boxes 15.1–15.4 are unsigned, so attributing Box 15.4’s rebuttal of Phillips to the chapter authors is probable, not certain (notes and digest).
7. Box 15.1: added the USDA’s first stated motive (preventing spread to other flocks), which the notes had omitted.
8. Box 15.1: noted that the US precedent rests on a secondhand source (Martin 1998, a BSE Inquiry statement).
9. Box 15.1: added a [Background – verify] flag on the RCEP report’s date against the May 1979 election.
10. 15.3 notifiability: corrected “resisted” to the source’s “rejected in part”, and added the authors’ description of notifiability as “an essential tool for disease surveillance” and the Phillips paragraph number.
11. 15.4: the CMO was told because ministers needed his “support”, not “backing”/”cover”. Labelled the “borrow health authority” line as the note-taker’s gloss. Toned down “cover” in the timeline and in mechanisms 6 and 13.
12. 15.5 counterfactual: restored the authors’ hedges (“might have found it easier”; “might have to be tolerated”). The notes had stated the counterfactual more strongly than the source (“could have … without destroying its credibility”).
13. Box 15.2: added the authors’ own scope caveat (“beyond the scope of this case study”) and the “In general” hedge on the Ireland/Austria claim. Page render confirmed that “(19)” is the only in-text marker for fn 13.
14. 15.6: added the omitted clause about MAFF’s “vigorous attempts to sustain the policy”.
15. Box 15.3: added the box’s framing that costs are “as yet incomplete”, partly “incalculable”, and that precise totals would be “premature”.
16. Minor errors: added the unclosed quotation mark in the Welsh CMO quotation (p. 163). Noted the page-render confirmation of the missing emphasis on p. 161.
17. Lessons (a)1: softened “do not survive the documentary record” to the authors’ own “not convincing”.
18. Mechanism 4: labelled “selective analogy” as the note-taker’s synthesis, consistent with insight 11 and the digest.
19. Mechanism 8: “a legitimate political judgement” changed to “a matter for political judgement, not science”. “Legitimate” was not the authors’ word.
20. Mechanism 14: “understood that the agent was probably in tissues that cannot be removed” overstated the documents. Revised it to say that officials anticipated where the logic of further controls would lead.
21. Insight 8: added the qualification that adviser selection rests on one retrospective testimony with no excluded expert named.
22. Digest: widened “fear for exports” to the source’s fear of undermining domestic and international confidence, plus reluctance to spend. Added the notifiability episode.
23. Digest: stated that the link from MAFF’s “not necessary” line to 48% non-compliance is the authors’ inference, resting on one testimony.
24. Digest: “no mention of dissenting scientists” corrected to “no named dissenting scientists”, since general warnings from the wider scientific community are mentioned (pp. 161, 164). Added the transcript-citation caveat to “Mixed evidence base”.
25. Scanned notes and digest for references to specific contemporary technologies or companies not in the source and found none.