LL1-02 digest — Ch2 Fisheries: taking stock#

Late lessons from early warnings (EEA 2001), pp. 17–30 (PDF 17–30). The author, Malcolm MacGarvin, is an ecologist and consultant, and also the report’s Executive editor. There are no panels. Full notes are in notes/LL1-02.md.

Core story#

The chapter argues that fisheries show “history repeating itself” (p. 17).

• Scottish herring and trawling (pp. 17–19).
• In 1865 Bertram showed that nets per boat had nearly quadrupled while catches fell. Huxley and an 1893 committee still saw no decline; the author infers (“presumably”) that landings were equated with stock (p. 18).
• An 1883–85 inquiry’s own research found most complaints against trawling unfounded, though its scientist warned inshore waters “could very easily be trawled out”. The inquiry refused to restrict trawling unless experimental evidence “decisively settled” the issue. Scotland banned inshore trawling anyway (p. 18).
• Between 1898 and 1998, Scottish landings stayed flat while fishers fell from 36,161 to 7,771, and the demersal fleet’s late-1990s return on capital was 0.1% (p. 19).
• Californian sardine (pp. 19–20).
• From the mid-1920s, state scientists warned of overfishing. The federal Bureau conceded the evidence of depletion was strong, but said there was none “that will satisfy everyone”, and preferred “wise use” to “hoarding” (p. 20).
• In 1939 a new Governor replaced the state experts, and assessments moved from “imperative” cuts to “no reason to be concerned”. A catch increase was agreed “to ‘assist the war-effort’”, and the chapter says the stock collapsed that year, 1942.
• Processors who had recovered their investment in a season moved on to Peru’s anchoveta fishery, which also collapsed (p. 20).
• Northern cod (pp. 20–23), the centrepiece.
• After 1977, Canada’s “deliberately conservative” target of about 20% failed. Assessments that converge only with hindsight overestimated the stock and gave “a false sense of assurance” (p. 21).
• The inshore fishers’ Keats Report (1986), built on DFO’s own downplayed analyses, was dismissed as “biased pseudoscience”. The 1988 Alverson Report’s body found fishing mortality “considerably in excess” of target and showed how sensitive stock estimates were to the wide range of plausible fishing-mortality values. In the author’s words, the executive summary “turned on its head” that sensitivity point, using the range to say DFO’s estimates fell “within the range of estimates supported by the data”, “albeit at the lower end”. Alverson himself publicly played down the gap (p. 21). (The Alverson Report is not in the chapter’s reference list.)
• Harris (1990) found fishing mortality “probably” more than double the intended level. He then recommended a 190,000 t catch limit rather than the 125,000 t needed to meet the 20% goal, to avoid “drastic” social and economic repercussions, while warning that “this may contribute to further decline”; it was adopted (pp. 21–22).
• The stock collapsed. The July 1992 moratorium was followed by no recovery by 2000, and costs “in excess of several billion” Canadian dollars (the author’s own estimate; p. 22).
• Governance (pp. 23–26).
• The 1995 FAO Code and UN Agreement made precaution explicit, and US guidance reversed the “burden of proof” (p. 23 n. 3).
• EU/ICES reference points were a “precautionary gloss” on an often unchanged regime (p. 24).
• A 1999 North Sea analysis found the same underestimation of fishing mortality for cod, haddock and whiting (not plaice or sole), a problem known since 1977 (p. 24).
• Since Canada’s precautionary-in-intent regime “still crashed”, the author asks whether current change is enough (p. 25). He favours ecosystem-based, “error-resilient” measures such as no-take zones (pp. 25–26).

Authors’ lessons (p. 26)#

The author’s 14 lessons include: - distinguish the principle from the approach; - set appropriate levels of proof; - separate uncertainty from ignorance; - beware “incredible claims” of scientific soundness; - use historical and lay knowledge; - avoid dominance by one discipline, and elaborate models that explain away failure; - account for real-world conditions (underestimated fishing mortality and technology); - deal with institutional obstacles and secure regulatory independence; - acknowledge value judgements; - keep humility.

He also says cod was “a systems failure” (p. 23), and that there is no single “correct” precautionary approach (p. 25).

Mechanisms#

• Capability masks decline. Rising capability hides decline in output figures.
• Hindsight-only estimates. Estimates that are right only in hindsight give false reassurance.
• Variability as confounder and shield. Natural variability is both.
• Asymmetric proof. Proof standards count the costs of action but not of inaction.
• Positional dissent. Dissent follows interest: first inshore fishers, then offshore.
• Mobile capital. Investors can “mine” the resource and leave.
• Political override. Politics overrides experts, and split jurisdiction breeds euphemism.
• The certainty trap. Scientists gave precise low numbers expecting ranges to be exploited, and it backfired.
• Face-saving. An agency whose authority rests on its science resists revision.
• Model lock-in. Epicycles are added rather than assumptions questioned.
• The rebuilding trap. Depleted resources cannot fund their own recovery.

Transferable insights (technology-neutral)#

1. Growing capability can hide a declining resource. Strong for herring; moderate–strong for sardine (single secondary source, no figures, contested collapse); weaker for cod, where the inshore–offshore link is now questioned (pp. 17–22).
2. Hindsight-convergent methods mislead most during rapid decline, and the bias can recur long after it is identified. Strong for northern cod; moderate for generality, since the North Sea bias appeared in three of five stocks (pp. 21, 24).
3. Demands for certainty, plus fear of range-exploitation, produce false precision. Moderate: one interview study (p. 22).
4. Arbiter institutions defend their numbers, and independent review and outside pressure are needed. Moderate–strong (p. 21).
5. Near-unattainable proof standards make inaction the default. Strong as description, moderate as cause (pp. 18, 20, 25).
6. Mobile capital can rationally deplete and relocate, and place-bound communities pay. Moderate (pp. 19–20).
7. Capability gains can be fully absorbed by depletion. Moderate: endpoint data (p. 19).
8. Political control of advisers can reverse official findings; split jurisdiction breeds stalemate and euphemism. Moderate: the reversal rests on one brief, single-sourced episode (pp. 20, 24).
9. Precaution run through biased models fails, and model-independent set-asides add resilience. Strong for the failure, suggestive for the remedy (pp. 22, 25–26).
10. Precautionary language can relabel unchanged practice. Moderate (pp. 23–24).
11. Frontline knowledge sees early but is positional. Moderate (pp. 17, 20–21).
12. What precaution protects is a value choice. Strong, conceptual (p. 25).
13. Depletion creates a rebuilding trap. Moderate: the counterfactual is unproven (p. 22).

Caveats#

• Author’s position. The author is the report’s Executive editor, part of the editorial team credited with the concluding chapter that reuses this case. He cites his own WWF-UK work for key figures. There are no dissenting panels.
• Single sources. Key episodes rest on single secondary sources (McEvoy; Finlayson).
• Sardine. The environment is treated only as an industry excuse, and Table 2.1 blames “inaction”. External pointers: natural swings are well documented, and the collapse is usually dated to the late 1940s–50s.
• Cod. The chapter’s own evidence (“far worse than even the most pessimistic projections”; cause unknown; inshore–offshore link questioned; p. 22) weakens the counterfactual.
• Lay knowledge. The lesson ignores the chapter’s own counter-cases. In the Keats case, lay observation worked through a commissioned scientific re-analysis.
• Table 2.1 departs from the text in five places: three sharpen the narrative of unheeded warnings (“no action is taken” for 1866–93 despite the Scottish ban; “continuous inaction” causing the 1942 collapse; DFO’s own 1999–2000 concern recast as “dissidents”), one overstates progress (the ecosystem approach “slowly settles”), and one (“claiming it is doing so precautiously”) undercuts the text’s point that Canada’s regime was genuinely precautionary in intent.
• A factual slip. “Then novel” UNCLOS is anachronistic, since the convention dates from 1982.
• Strengths. The chapter is candid in doubting model-based precaution and in declining to find villains.