Crawford Stanley 'Buzz' Holling, an ecologist at the University of British Columbia, published a paper in 1973 titled simply Resilience and Stability of Ecological Systems. After fifteen years studying spruce-budworm outbreaks in eastern Canadian forests, he had concluded that the standard ecological framework — which assumed ecosystems had a single equilibrium they returned to after disturbance — was wrong. Forests had multiple stable states, and what mattered for management was not how stable the current state was but how resilient: how large a disturbance the system could absorb before flipping. The two often pulled in opposite directions: a system optimized to return rapidly to a single setpoint could be brittle against large disturbances. Efficient systems strip away the slack that resilient systems use.
Efficiency is the ratio of useful output to total input. Resilience is the magnitude of disturbance a system can absorb without losing its essential structure or function. Along most relevant axes the two trade against each other, because the slack that absorbs disturbance is what efficiency optimization removes. Holling's distinction sits inside a four-part diagnostic: engineering resilience (return time after small perturbation), ecological resilience (size of perturbation absorbable without regime shift), adaptability, and transformability. The trade-off shows up across every kind of system. Monoculture agriculture maximizes per-hectare yield under benign conditions and crashes under disease: the 1970 Southern Corn Leaf Blight destroyed 15% of the US corn crop because virtually all US hybrid corn shared a single cytoplasmic male-sterile susceptibility. Just-in-time supply chains — the Toyota Production System refined through the 1960s-70s — minimize inventory costs and amplify any disruption through tight coupling; the 2011 Tōhoku earthquake disrupted automotive production worldwide for months. Lean hospital staffing strips surge capacity; COVID-19 revealed this catastrophically. Highly leveraged financial systems operating at thirty-to-one asset-to-equity ratios are efficient on returns and fragile against small losses; the 2008 crisis destroyed several major institutions at exactly those ratios. Nassim Taleb, in The Black Swan (2007) and Antifragile (2012), pushed the framework further with three response types: fragile (loses from disturbance), robust (unchanged), antifragile (gains from disturbance). Antifragile systems — biological evolution under selection pressure, immune-system challenge, muscle response to exercise — actively benefit from variability up to a saturating limit. The features producing resilience are redundancy, diversity, modularity, decoupling, and adaptive capacity — the same features efficiency optimization removes.
The 2020-2022 period delivered the most concentrated lesson in efficiency-resilience trade-offs since 1973. COVID-19 revealed the vulnerability of lean hospital systems and just-in-time medical-supply chains. The 2021 semiconductor shortage exposed how thoroughly the world's most economically critical industry had concentrated its leading-edge fabrication into one company (TSMC) on one island (Taiwan), producing the 2022 US CHIPS Act over the protests of efficiency-minded economists. The March 2021 Ever Given blockage of the Suez Canal held up roughly 12% of global trade for six days. The July 2024 CrowdStrike outage took down millions of Windows machines because a single vendor's update mechanism had become a single point of failure. The 2022-onwards energy-security reaction to Russia's invasion of Ukraine forced Europe to add LNG import capacity at speeds prewar efficiency optimization had ruled out.