Organizational Resilience as a Dynamic Capability

Defining Resilience: Beyond Bouncing Back

The colloquial understanding of resilience as the ability to "bounce back" is intuitive and, for the purpose of organizational analysis, misleading. Recovery to a prior state is only one of several outcomes the resilience literature describes, and it is arguably the least interesting. Meyer (1982), in one of the foundational empirical studies of how organizations respond to sudden environmental disturbances, distinguished between organizations that merely absorbed a jolt and returned to their previous configuration and those that used the disturbance as an occasion for strategic reorientation. The latter did not bounce back; they bounced forward, emerging with capabilities they had not previously possessed. This distinction between restorative and developmental responses to adversity has persisted through four decades of subsequent research and remains the most important conceptual boundary in the field.

A second boundary separates resilience from robustness. Robustness is the capacity to resist disturbance without changing, the property of a structure engineered to withstand a known load. Resilience, by contrast, presupposes that the disturbance will exceed what resistance alone can absorb and concerns the capacity to reorganize in response. Hamel and Välikangas (2003) framed the distinction in strategic terms, arguing that continued success no longer depends on momentum or on the durability of a fixed business model but on the ability to dynamically reinvent models and strategies as circumstances change. An organization can be robust and brittle simultaneously: highly optimized for a stable environment, efficient under expected conditions, and catastrophically exposed to conditions outside its design envelope. The pursuit of efficiency, in this reading, is not the ally of resilience but frequently its antagonist, because the slack and redundancy that resilience requires are the first casualties of optimization.

Ortiz-de-Mandojana and Bansal (2016) offered a definition that has become widely adopted: resilience is the firm's ability to sense and correct maladaptive tendencies and to cope positively with unexpected situations. Two features of this definition are consequential. First, it locates resilience partly in the pre-disruption period, in the capacity to detect and correct drift before it becomes crisis, rather than exclusively in the response to a shock already underway. Second, it treats resilience as a latent, path-dependent construct that cannot be observed directly and must be inferred from long-term outcomes. In their longitudinal study of 242 firms over 15 years, resilience was assessed not through a survey of managerial perception but through reduced financial volatility, sustained sales growth, and higher survival rates, precisely because the construct is not visible in any single moment and reveals itself only over time and across disturbances.

This latency is the central methodological problem of the field and the reason resilience resists the measurement approaches that work for more observable constructs. An organization that has not recently been tested may appear resilient while harboring latent fragilities, and an organization that has weathered one specific disturbance may have done so through conditions that will not generalize to the next. Any credible account of organizational resilience must therefore explain not merely how organizations behave when disrupted but what latent capability, present before any disruption, determines the range of behaviors available to them when disruption arrives.

The Capability Architecture: Anticipation, Coping, Adaptation

The most useful recent contribution to the definitional problem has been the decomposition of resilience into distinct temporal stages. Duchek (2020) conceptualized organizational resilience as a meta-capability composed of three successive stages, each supported by its own underlying capabilities: anticipation, the capacity to observe and prepare for potential threats before they materialize; coping, the capacity to accept and address problems once a disruption is underway; and adaptation, the capacity to reflect and learn so that the organization is changed by the experience rather than merely returned to its prior state. The value of this decomposition is that it separates capacities that are frequently conflated and that draw on genuinely different organizational competencies. An organization strong in coping but weak in anticipation is one that responds well to crises it should have seen coming; an organization strong in coping but weak in adaptation repeatedly manages the same disturbance without ever ceasing to be vulnerable to it.

The framing of resilience as a meta-capability connects it to the dynamic-capabilities tradition in strategic management. Teece et al. (1997) defined dynamic capabilities as the firm's ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments. Resilience, on this reading, is not a distinct resource that a firm either possesses or lacks but a higher-order capability that governs the reconfiguration of other resources under conditions of disturbance. Lengnick-Hall et al. (2011) developed this position explicitly, proposing that a capacity for resilience is developed by strategically managing human resources to create cognitive, behavioral, and contextual competencies among core employees that, aggregated at the organizational level, make a resilient response possible. Resilience in this account is not stored in a plan or a reserve fund; it is embedded in the knowledge, relationships, and routines through which an organization orients itself, acts, and integrates diverse capabilities under pressure.

The anticipation stage draws heavily on the high-reliability organization literature. Weick and Sutcliffe (2007) documented how organizations operating in unforgiving environments, such as aircraft carriers, nuclear power plants, and emergency medical settings, sustain performance through what they term collective mindfulness: a preoccupation with failure, a reluctance to simplify interpretations, a sensitivity to operations, a commitment to resilience, and a deference to expertise rather than hierarchy in moments of crisis (Weick et al., 1999). What distinguishes these organizations is not that they experience fewer anomalies but that they attend to small anomalies as potential signals of larger problems rather than dismissing them as noise. The capacity to notice weak signals and resist premature closure is a cognitive antecedent of anticipation that most organizations systematically underinvest in, because attending to signals that usually amount to nothing is costly and its payoff is invisible until the rare occasion when the signal mattered.

The coping stage is where sensemaking becomes decisive. Weick's (1993) analysis of the Mann Gulch disaster, in which a wildfire crew was overrun after their situation deviated catastrophically from their expectations, demonstrated that organizational collapse under stress is frequently a collapse of shared meaning rather than of physical capability. When events outrun the interpretive frameworks that coordinate a group, role structures dissolve, communication fragments, and individuals default to self-protective behavior at precisely the moment collective action is most needed. Coping capability therefore depends less on having a plan for the specific disturbance, which by definition is often unforeseen, than on the organization's capacity to maintain shared sensemaking and coordinated action when its plans no longer fit the situation.

Antecedents and the Threat-Rigidity Trap

If resilience is a latent capability, the analytical task is to specify what constitutes it. The literature converges on a small number of antecedent categories. Lengnick-Hall et al. (2011) organized these as cognitive, behavioral, and contextual, and subsequent work has generally preserved this structure while adding a relational dimension that the airline-industry evidence makes unavoidable. The cognitive antecedents concern how the organization attends to and interprets its environment. The behavioral antecedents concern the organization's repertoire of responses and its willingness to act, including the capacity to improvise when established routines do not fit, a competence that is practiced rather than innate and that atrophies in organizations that never deviate from procedure. The structural antecedents concern the slack, redundancy, and distribution of authority that determine how wide a range of responses is physically and organizationally available.

The relational antecedent deserves particular emphasis because it is the one most consistently underestimated. Gittell et al. (2006) studied the U.S. airline industry after the September 11 attacks, an event that affected every major carrier simultaneously and thus offered a natural experiment in differential resilience. Some carriers recovered strongly while others languished, and the difference was not explained by size or by the severity of the initial shock. Carriers that preserved relational reserves, the stock of trust, shared goals, and mutual respect among employee groups built up before the crisis, recovered more successfully, in part because they were able to avoid the deep layoffs that competitors used and that, the study found, inhibited rather than aided recovery over the subsequent four years. Relationships, in this account, function as a reserve in the same sense that capital does: accumulated in good times, drawn upon in bad, and impossible to acquire quickly once the crisis has begun.

The dominant failure mode of organizational resilience is the threat-rigidity response, and understanding it is essential to understanding resilience itself. Staw et al. (1981) documented a robust and counterintuitive pattern: when individuals, groups, and organizations perceive threat, they tend to respond with rigidity rather than flexibility. Information processing narrows, control centralizes, and behavior reverts to well-learned, dominant responses. This tendency is adaptive when the threat calls for a familiar response executed decisively, and profoundly maladaptive when the threat is novel and calls for exactly the flexibility, distributed judgment, and behavioral experimentation that the rigidity response suppresses. The tragedy of threat rigidity is that it activates most strongly under precisely the conditions, severe and unfamiliar disturbance, in which resilience is most needed. Much of what distinguishes resilient organizations is not the presence of some additional capability but the absence of, or the deliberate counteraction of, this default response. Decentralized authority, psychological safety, and the deference-to-expertise principle of high-reliability organizations can all be read as structural and cultural bulwarks against the centralizing, simplifying reflex that threat provokes.

Van der Vegt et al. (2015) situated these organizational antecedents within a wider system, observing that organizational resilience is strongly affected by relationships with other organizations and with the broader environment. A firm's capacity to absorb and adapt to disruption is partly a function of its embeddedness in supply networks, institutional arrangements, and community relationships that either buffer or amplify the disturbances it faces. This systemic view cautions against treating resilience as a purely internal property. An organization may possess considerable internal resilience capacity and nonetheless prove fragile because the network in which it is embedded transmits and compounds shocks faster than internal capability can absorb them.

Measurement and Diagnostic Implications

The latency of resilience creates a genuine measurement problem that no assessment approach fully resolves and that any credible diagnostic must confront honestly. Because resilience is revealed only when an organization is disturbed, and because the specific disturbances an organization will face are largely unknowable in advance, resilience cannot be measured in the way that a currently observable state, such as employee engagement or role clarity, can be. Ortiz-de-Mandojana and Bansal (2016) addressed this by inferring resilience from long-term outcome patterns, an approach that is rigorous for research but unavailable to an organization that wishes to assess its resilience now, before the outcomes that would reveal it have occurred. The diagnostic question is therefore necessarily indirect: not "how resilient is this organization," which cannot be answered until it is tested, but "to what degree are the antecedents of resilience present," which can be assessed.

This reframing makes resilience assessment tractable without pretending the underlying problem away. The antecedent categories identified in the literature, cognitive, behavioral, relational, and structural, are each assessable through a combination of methods even though resilience itself is not directly observable. The cognitive antecedents can be examined through how the organization handles weak signals, near-misses, and anomalies: whether such events are surfaced and examined or suppressed and forgotten. The relational antecedents can be assessed through the trust and coordination-quality measures that organizational research has developed and validated. The structural antecedents, slack, redundancy, and the distribution of decision authority, are among the more directly observable and can be examined through the organization's design and its actual decision practices under ordinary conditions. What an antecedent-based diagnosis provides is not a prediction of how the organization will perform in a future crisis, which no instrument can honestly promise, but an evidence-based account of whether the conditions that make a resilient response possible are being built or eroded.

A second diagnostic implication follows from the stage structure of resilience. Because anticipation, coping, and adaptation draw on different underlying capabilities (Duchek, 2020), an assessment that produces a single resilience score obscures more than it reveals. An organization may be strong in coping, able to mobilize effectively once a crisis is undeniable, while weak in anticipation, unable to detect the signals that would have allowed earlier and cheaper intervention, and weak in adaptation, unable to convert the hard-won lessons of each crisis into durable changes in routine. These are distinct deficits with distinct remedies, and a diagnostic that distinguishes them offers actionable guidance in a way that an undifferentiated index cannot. The most useful resilience diagnostics are therefore multidimensional by design, reporting the organization's relative strength across the stages and antecedents rather than collapsing them into a single number.

A third implication concerns the relationship between resilience and the efficiency pressures that ordinarily govern organizational design. Because the slack and redundancy that resilience requires are visible costs whose benefits are latent and probabilistic, resilience is chronically underfunded relative to its expected value, and the underfunding is invisible until a disturbance exposes it. Diagnostic assessment can serve a specific function here: by making the antecedents of resilience visible and nameable, it converts an abstract and easily deferred concern into a concrete account of specific conditions that are present or absent, strengthening or weakening. This does not resolve the genuine tension between efficiency and resilience, which is a strategic choice that no diagnostic can make for an organization, but it allows the choice to be made deliberately and with evidence rather than by default, which is how the balance is most often struck against resilience.

The field's remaining challenges are substantial. Linnenluecke (2017), in a systematic review of the influential resilience publications across business and management research, documented a literature that is theoretically fragmented, drawing on ecological, engineering, psychological, and strategic traditions that use the same word for related but distinct constructs, and that remains comparatively thin in validated measurement. The practical consequence for organizations is a caution rather than a counsel of despair. Resilience is real, it is consequential, and its antecedents are identifiable and assessable; but any assessment should be understood as a diagnosis of conditions rather than a prediction of outcomes, and any organization claiming to measure resilience directly, rather than the conditions that produce it, is claiming more than the evidence supports. Within that honest boundary, structured assessment of resilience antecedents offers organizations something they otherwise lack: a way to examine, before the next disturbance arrives, whether they are building the capability that disturbances reveal or quietly spending it down.