After the 2010 Haiti earthquake, international aid organizations rebuilt thousands of structures using the same designs that had collapsed. The structures were rebuilt — but the underlying vulnerabilities to earthquake damage remained unchanged. Within a few years, poorly maintained infrastructure was again at risk. The system had bounced back to its pre-disaster state. It had been restored, but not made more resilient.
This story illustrates the difference between the two most important concepts in resilience thinking for complex systems: restoration (returning to the prior state) and genuine resilience (the capacity to absorb disruption, adapt, and potentially reorganize into a configuration better suited to future conditions). Systems thinking reveals why these two outcomes are not the same, and what it takes to build the latter rather than the former.
Two Kinds of Resilience
Engineering resilience
Engineering resilience (sometimes called robustness) is the capacity of a system to resist disturbance and return quickly to its prior state after being perturbed. It is measured by the speed and completeness of recovery. A bridge is resilient in this sense if it can withstand a given earthquake without damage, or if minor damage can be quickly repaired. An organization is robust in this sense if it can maintain core functions through a disruption and restore normal operations quickly afterward.
Engineering resilience is associated with stability near a single desired state. It is appropriate for systems where the environment is sufficiently stable and predictable that returning to the prior state is always the right goal.
Ecological resilience
Ecological resilience (developed by C.S. Holling in the same tradition as panarchy theory) is a fundamentally different concept: the capacity of a system to absorb disturbance and reorganize while undergoing change so as to retain essentially the same function, structure, identity, and feedbacks. It is measured not by how quickly the system returns to its prior state but by how much disturbance it can absorb before crossing a threshold to a qualitatively different state.
Ecological resilience is associated with multi-stability: complex systems can exist in multiple alternative stable states, and the question of resilience is whether a given disturbance pushes the system across the threshold from one state to another. A resilient forest is not one that returns quickly to its prior configuration after a fire — it is one that can absorb a fire and reorganize into a functional forest ecosystem rather than collapsing into a grassland or scrubland from which forest recovery is very slow or impossible.
Adaptive Capacity: The Third Dimension
Beyond both forms of resilience, researchers in the sustainability sciences have identified a third critical dimension: adaptive capacity. This is the capacity of a system not just to absorb disturbance and maintain or restore its prior state, but to learn from disturbance and reorganize into configurations that are better suited to changed conditions.
Adaptive capacity is what distinguishes systems that can evolve from those that can only recover. A city that experiences a major flood, rebuilds exactly what was there before, and then experiences the same flood damage in the next major storm has resilience in the narrow engineering sense but lacks adaptive capacity. A city that experiences the same flood, learns from it, and redesigns its infrastructure and land use in ways that reduce future flood damage has demonstrated adaptive capacity.
Adaptive capacity is closely related to double-loop learning: learning that changes not just the response to a given challenge but the framework within which responses are generated. Systems with high adaptive capacity can update their own goals, structures, and operating principles in response to learning — rather than simply executing the same responses more or less effectively within an unchanged framework.
The Efficiency-Resilience Trade-Off
One of the most important practical insights of resilience thinking is that efficiency and resilience are in fundamental tension. Highly efficient systems — optimized for a specific, stable operating environment — achieve their efficiency by eliminating redundancy, reducing diversity, tightening coupling, and minimizing slack. These are precisely the properties that provide resilience: redundancy allows alternative pathways when primary pathways fail; diversity provides a portfolio of responses to novel conditions; loose coupling contains disturbances rather than propagating them; slack absorbs variability.
The global supply chains of the 2010s were models of engineering efficiency: just-in-time delivery, single-source suppliers for critical components, tightly coupled logistics networks with minimal inventory buffers. When the COVID-19 pandemic disrupted multiple nodes of these networks simultaneously, the systems’ lack of redundancy, diversity, and buffer capacity produced cascading failures of a kind that more inefficient but more resilient networks would have absorbed. The efficiency that decades of optimization had delivered came at the cost of the resilience that an unpredictable world requires.
Building Resilience in Organizations
Resilience thinking has direct practical implications for organizational design:
- Preserve redundancy in critical systems, even at the cost of efficiency: multiple suppliers for critical inputs, redundant communication channels, overlapping skills within teams.
- Maintain diversity in approaches, personnel, and supply sources: homogeneous systems are more efficient but less robust to novel challenges.
- Build in modularity: systems in which failures in one module are contained rather than propagating across the whole system are inherently more resilient.
- Invest in monitoring for early warning signals of regime shifts: many systems exhibit early warning signals before crossing resilience thresholds, including increased variance, slower recovery from perturbations, and increased spatial correlation. Detecting these signals early allows intervention before the threshold is crossed.
- Create learning loops: organizations that systematically learn from disruptions — through after-action reviews, incident analyses, and structured debriefs — build the adaptive capacity to improve rather than merely recover.
Frequently Asked Questions
Can a system be too resilient?
Yes. A system that is extremely resilient — strongly homeostatic, highly resistant to change — may be unable to adapt when the environment has changed enough that its prior stable state is no longer viable. Excessive resilience can trap a system in a configuration that is no longer appropriate, preventing the adaptive reorganization that changed conditions require. This is why adaptive capacity is distinct from and complementary to resilience: the former enables the latter to be directed toward genuinely better configurations rather than simply restoring prior ones.
Conclusion
Resilience thinking in complex systems goes well beyond the common-sense notion of bouncing back from adversity. True systemic resilience involves the capacity to absorb disruption without crossing into a worse state, the adaptive capacity to reorganize in ways that are better suited to changed conditions, and the learning capacity to update frameworks rather than just improve responses within them. In a world of increasing disruption — from climate change, technological change, pandemic, and geopolitical instability — building this kind of genuine resilience is one of the most important investments any organization, community, or society can make.
