RESEARCH DOSSIER
EU Digital Operational Resilience Regulation for Insurers
Abstract. The European Union's Digital Operational Resilience Act (DORA) establishes a financial-sector framework for managing information and communication technology (ICT) risk, recording and reporting serious incidents, testing operational resilience, controlling ICT third-party relationships, and overseeing critical technology suppliers. For insurers, its significance arises from the interaction of legally regulated undertakings, operational services that must continue through disruption, and infrastructure shared across firms. This monograph interprets the binding EU instruments, their implementation sequence, the distribution of accountability between boards, insurers, national competent authorities and the European Supervisory Authorities, and early empirical evidence available by November 15, 2026. It distinguishes insurer operational exposure from cyber insurance underwriting risk; examines proportionality, reporting thresholds, testing and contractual obligations; and develops two explicitly illustrative calculations concerning claims-service recovery and correlated vendor outages. The source-critical synthesis identifies three recurring evaluative questions: whether controls preserve actual insurance services, whether information can be transmitted accurately during the incident being reported, and whether governance can identify dependencies extending beyond the legal entity. Early cross-sector incident statistics establish observable supervisory reporting, although they cannot establish insurer-specific causal effects or compliance success. The conclusion specifies operational and supervisory evidence required to move from formal implementation toward measured resilience without proposing a new regulatory theory.
Keywords: Digital Operational Resilience Act; DORA; Regulation (EU) 2022/2554; insurance undertakings; reinsurance; ICT risk management; cyber resilience; operational continuity; major ICT-related incidents; threat-led penetration testing; third-party risk; cloud concentration; Solvency II; insurance supervision; policyholder protection; European Union.
Introduction
Insurance transforms uncertain losses into contractual obligations financed and administered over time. Those obligations now depend on continuously available policy records, actuarial data, premium collection, identity verification, claims triage, payment channels, distribution networks, and external ICT services. A disruption can interrupt the performance of existing promises even where the insurer remains solvent and its contractual obligations are legally unchanged. Digital operational resilience consequently connects the insurer's technical operations to its fundamental economic function: sustaining the capacity to insure, adjudicate and pay.
DORA applies a cross-sector legal structure to this continuity problem. The Digital Operational Resilience Act, formally Regulation (EU) 2022/2554, was adopted on December 14, 2022 and applies from January 17, 2025. Its subject is broader than cyberattack prevention: the statutory definition of digital operational resilience refers to the capacity to maintain reliable operations, including during ICT disruptions. European Insurance and Occupational Pensions Authority (EIOPA) describes the instrument in terms of ICT risk, incidents, testing, third-party management, information sharing and critical provider oversight in its DORA supervisory overview. Treating each of those pillars as part of a single insurer operating system is essential to understanding its effects.
The present research asks three connected questions. First, which legal entities and operational activities fall within the instrument, and how do exclusions and proportionality alter its application? Second, what relationship exists between board accountability, operational resilience, serious-incident reporting, testing, and the contractual control of ICT providers? Third, what can be learned from the first supervisory evidence, including European-wide incident reports and provider designations, without treating cross-sector administrative data as an insurer-specific outcome study?
The method combines close reading of regulations and implementing instruments, dated comparison with Solvency II, analysis of primary supervisory releases, and critical engagement with the published supervisory literature. European Union legal texts are treated as the primary source for legal duties; an agency press release is evidence of an agency's activity and reported observations; academic work is assessed for its stated analytical claim and method. The analysis covers legal rules and publicly available evidence through November 15, 2026. The two quantitative exercises in the final chapters use transparent, hypothetical assumptions and are not estimates of EU insurance-market incidence. National implementation practice is discussed as a supervisory question rather than claimed to have been exhaustively surveyed country by country.
The central analytical distinction concerns the level at which an observation is made. A legal requirement attaches to a financial entity; a disruption propagates through business functions and supply chains; a supervisory dataset may aggregate all financial sectors. Conclusions derived at one level must not be mechanically carried to another. This distinction also guards against equating operational ICT risk borne by an insurer with the separate insured exposure that an insurer might underwrite in its cyber insurance portfolio.
Definitions, History, and Documentary Foundation
Digital operational resilience as a legal object
Article 3 of Regulation (EU) 2022/2554 defines digital operational resilience by reference to the ability of a financial entity to build, assure and review its operational integrity and reliability, directly or through ICT third parties, so that financial services continue with adequate security and quality during disruption. The legislative object therefore includes availability, confidentiality, integrity, authenticity, service continuity and recoverability. It spans ordinary systems failures as well as hostile cyber activity.
Four related terms require precision. An ICT-related incident is an unplanned event or connected series affecting the security of networks or information systems and adversely affecting data or the services provided. A major ICT-related incident is the more consequential legal category determined according to the statutory and delegated classification criteria; an important outage is not automatically a reportable major incident. A critical or important function is assessed by the relevance of a function to the entity's financial performance, soundness, continuity, or regulatory obligations. A critical ICT third-party service provider (CTPP), by contrast, is designated through a Union-level assessment of provider systemic importance. The functional label and the provider designation operate at different legal levels.
An insurer's claims platform can support a critical or important function even where its cloud provider has not been designated a CTPP. Conversely, a provider may be designated critical across the European financial sector while a particular insurer's individual contract supports a noncritical internal function. This difference matters: the legal analysis of an insurer's outsourced service and the Union-level oversight of a provider must be conducted separately.
Legislative development and regulatory architecture
DORA forms part of the development of common EU digital finance regulation and adds uniform ICT operational-resilience duties to the preexisting sectoral prudential framework. The accompanying Directive (EU) 2022/2556 adjusted multiple financial-sector directives, including the insurance regulatory architecture, to accommodate the DORA framework. Its existence helps explain why a sectoral Solvency II requirement cannot be read as replacing the directly applicable DORA requirements.
The relevant insurance baseline is Directive 2009/138/EC, Solvency II, with governance, internal controls and risk management duties concerning all material risks. Solvency II is fundamentally concerned with the insurer's soundness, risk profile and capacity to meet policyholder commitments. DORA adds operationally specific obligations on ICT assets, incidents, tests, providers, documentation and supervisory communication. Their shared concern with effective governance produces cumulative oversight responsibilities. A board may demonstrate strong capital adequacy while still needing to prove the recoverability of a claims system and the enforceability of a critical supplier's access or exit provisions.
The detailed legal program is distributed across delegated and implementing regulations. The 2024 instruments specify ICT risk governance, incident classification, contractual policy, and standardized provider registers. The 2025 instruments add or complete reporting timelines and forms, subcontracting rules, and threat-led penetration-testing requirements. The European Commission maintains an official implementing-acts inventory, which provides a documentary map of these separate texts. Each delegated act should be applied to the issue it regulates; a single undifferentiated citation to DORA obscures important differences in legal basis, reporting time and covered entities.
What constitutes evidence of implementation?
The evidence base has three layers. At the first layer are enacted legal obligations, which demonstrate what is required but contain no proof that an institution has met them. At the second are supervisory publications and aggregate returns, which demonstrate reported observations and regulatory activities. At the third are insurer-level governance records, incident logs, restored-service timestamps, vendor contract clauses and testing results, most of which remain institution-specific and are not available in the public record.
Buttigieg and Brunelli Zimmermann (2024), in their analysis of Europe's supervisory architecture, identify a tension between harmonized regulation and a supervisory structure still distributed across national and European institutions. That problem is particularly pertinent to insurance groups operating across borders: a common legal standard improves comparability, but authorities still need coordinated interpretation, useful data and a credible allocation of response responsibilities. The paper supplies a framework for analyzing coordination difficulties, not an empirical measure of insurer loss reduction.
Insurers' operational risk and cyber insurance underwriting
An insurer can encounter digital disruptions in two economically distinct capacities. As an enterprise, it relies on ICT to administer policies and settle claims. As a risk carrier, it may underwrite policyholder cyber losses, including business interruption and liability. The latter portfolio may accumulate claims when many policyholders depend on a single vendor; the former generates operational disruption, remediation cost, customer harm and potentially other liabilities. Christian Biener, Martin Eling and Jan Hendrik Wirfs (2015), in their empirical analysis of cyber insurability, identify challenges of limited data, dependence and cyber risk characteristics. Their paper concerns cyber insurance risk and cannot be treated as evidence that DORA has changed insurers' own systems performance.
The distinction matters for management responsibility. The insurer's DORA register of ICT providers covers its relevant operational contracts. Its cyber underwriting aggregation model estimates exposures embedded in insured policyholders' contracts. Some vendors may appear in both datasets, yet the legal unit, economic loss measure and available controls differ. A joint dependency review can be analytically useful, but merging datasets without preserving the difference between own operational loss and insured claims would produce misleading estimates.
DORA Scope, Proportionality, and Regulatory Timetable
Undertakings, intermediaries, and legal exclusions
Article 2(1)(n) of DORA's scope provisions covers insurance and reinsurance undertakings, while Article 2(1)(o) separately names insurance, reinsurance and ancillary insurance intermediaries. The starting point is therefore a legal-entity classification, not merely whether a business is commercially described as an insurer, broker, administrator or technology company. The article's exclusions are equally important: Article 2(3)(b) removes insurance and reinsurance undertakings referred to in Article 4 of Solvency II; Article 2(3)(e) excludes intermediaries qualifying as microenterprises or small or medium-sized enterprises under the regulation's definitions.
These are materially different exclusions. The insurer exemption refers to the limited Solvency II perimeter under a specific directive article. The intermediary exemption refers to entity-size classifications defined for DORA. A small insurer cannot simply import the intermediary test, and an intermediary cannot assume that its commercial connection with a large insurer removes its independent scope analysis. The status of branches, groups, service providers and outsourcing relationships must be reconstructed against the relevant legal entity and the precise scope provision.
EIOPA raised a specific transitional concern in its November 2024 opinion on DORA and Solvency II scope. The opinion called for Commission action to avoid disproportionate compliance burdens on small insurance undertakings before the revised Solvency II framework becomes applicable. This is evidence of a recognized scope-transition problem, not a legislative amendment granting a new exemption. Its significance lies in separating current Article 2 applicability from proposed or future adjustments to the insurance prudential perimeter.
A practical evidence chain for determining applicability therefore begins with the entity's registration and authorization, then identifies the legal category in Article 2, tests explicit exemptions, establishes the relevant group or consolidated context, and only then determines which substantive requirements apply. A brand name, corporate-group size or outsourced service classification alone is not a reliable substitute for this sequence.
Proportionality does not eliminate accountability
Article 4 of the DORA regulation requires proportional implementation, taking into account the entity's size and overall risk profile, and the nature, scale and complexity of services, activities and operations. Proportionality shapes the design and intensity of controls within the applicable legal regime. It does not permit an insurer to ignore a critical claims function on the ground that its technology team is small.
Regulatory technique further distinguishes the ordinary ICT risk management framework from the simplified framework of Article 16. That article identifies particular categories of financial entities; its simplified treatment must be read against the enumerated scope and cannot be asserted for all small insurers merely because of their size. Delegated Regulation 2024/1774 contains technical rules for risk management as well as a simplified framework for eligible entities. The correct question is therefore which chapter and provisions apply to the entity, rather than whether one generic IT control manual exists.
An insurer can demonstrate proportionate design through the rationale linking critical functions, business dependency, exposure, available alternatives and testing frequency. Evidence may include a mapped claims-administration system, an assigned accountable executive, risk-based recovery testing and documented supplier review. Replacing this rationale with an undifferentiated checklist loses the meaning of proportionality: different products and customer populations have different tolerances for payment delays, authentication failures and unavailable records.
Temporal precision: applicable, adopted, and observed
DORA was adopted in December 2022 and began applying in January 2025. Instruments specifying particular requirements were published at different times. The ICT risk and simplified-framework RTS, major-incident classification RTS, and critical ICT contract-policy RTS were published in June 2024; the register-of-information ITS followed in December. The reporting-deadline RTS and reporting-template ITS appeared in February 2025; the subcontracting RTS and threat-led testing RTS were published later in 2025.
This staggered development matters to historical assessments. An audit of an insurer's procedures in early 2025 should distinguish duties arising under the already-applicable main regulation from technical specifications becoming operative later. A 2026 assessment should use the operative instruments and their legal dates, while retaining a record of earlier versions when explaining the organization's implementation history. Publication in the Official Journal, entry into force, and commencement of application are distinct legal facts.
Insurance group structures
An insurance group may contain life, non-life and reinsurance undertakings; shared-service companies; distribution subsidiaries; and branches using a central technology platform. The obligation to maintain an appropriate ICT risk program must be traced to entities within the DORA perimeter and, when relevant, consolidated or sub-consolidated requirements. A common group policy is useful only when local legal accountability, critical functions and authority interfaces remain demonstrable.
Consider a hypothetical group with separate life and non-life insurers using one claims-payment provider. Both insurers may be within DORA's scope and share one technical dependency, but their policyholder criticality, reporting significance and contractual arrangements must be assessed at the relevant entity and group levels. Cross-entity coordination may prevent duplicate technical investigation; it must not prevent each entity from meeting its own reporting and governance obligations. The article-specific group contractual-policy rules support consistency across financial entities in an applicable group without erasing their legal identity.
The research implication is that insurer proportionality is best tested at the intersection of a legally applicable entity, the importance of the affected insurance function and the actual technical dependency. No one of these elements determines the other two.
ICT Governance, Incident Classification, and Reporting
Management-body accountability and organizational control
Article 5 of Regulation (EU) 2022/2554 assigns the management body ultimate responsibility for the management of ICT risk and specifies governance responsibilities, including approval, oversight, allocation of resources and maintenance of appropriate knowledge and skills. This structure requires a decision-making process in which technology risks are translated into information the insurer's governing body can assess. Cybersecurity specialists may operate controls, but operational risk acceptance requires a documented allocation of authority and an understanding of consequences for services and clients.
For an insurer, the quality of management reporting depends on the connections it makes. A report that contains patch counts or generic security ratings without identifying policy administration, claims payments, reinsurance settlement and relevant ICT dependencies provides limited evidence that the board understands business impact. Conversely, a report can describe critical services, significant single points of failure, unresolved findings, restoration evidence and contractual exit limits in a manner that allows prioritization. The decisive criterion is traceability from the approved risk tolerance to a business function and a tested operational capability.
Articles 6–15 address ICT risk management through identification, protection and prevention, detection, response and recovery, backup and restoration, learning and communication. The ICT management technical standards further specify policies, procedures, ICT risk assessments and incident management, including attention to threats, vulnerabilities, control effectiveness and residual risk. These obligations point to a cycle of recurring assessment and revision. A one-time regulatory implementation project cannot substitute for the ongoing process because systems, vendors, vulnerabilities, claim volumes and business priorities change.
Accountability can be assessed with five documents, considered together: the board-approved ICT risk tolerance; the inventory linking services and ICT assets; current remediation decisions; evidence of resilience tests and restoration; and an exception or escalation record indicating who accepted residual exposure and why. The documents need not share one software platform, but their decisions must reconcile. A board-approved tolerance that has no connection to an outsourced claims database, or an incident process that cannot retrieve critical supplier contacts during an outage, indicates an implementation gap that is operationally meaningful.
Incident detection and classification
DORA's incident rules require the insurer to record, manage, classify and, where relevant, report ICT-related incidents. Delegated Regulation 2024/1772 establishes materiality thresholds and the combination of criteria used to determine when an incident becomes major. Article 8 ordinarily requires an incident to affect critical services and either meet the special malicious-access threshold identified in Article 9(5)(b) or meet at least two of the other specified materiality thresholds. Article 9 addresses affected clients, counterparties and transactions, duration and downtime, geographical spread, data losses, critical services, economic impact and reputational effect through distinct criteria.
Thresholds depend on the affected service, not simply the insurer's total customer base. For example, Article 9(1) identifies an affected-client threshold above 10 percent of clients using the affected service, or more than 100,000 such clients, among other alternative conditions. These figures are not independent automatic tests for a major incident; they operate within the Article 8 combination rule. A service-specific denominator is consequently indispensable. Comparing impacted claimants with the number of all in-force policies could materially distort classification.
The same delegated regulation addresses repeated events. Article 8(2) treats incidents as one major incident when at least two occur within six months, share the same apparent root cause and collectively meet the major-incident criteria; relevant entities assess recurrence monthly. The provision contains exclusions for microenterprises and Article 16(1) entities. A series of short recurring outages therefore deserves a systematic root-cause review even when no single interruption triggers immediate reporting.
Incident detection must capture both cyber intrusions and operational failures. An insurer should be capable of recognizing failures in policy data integrity, claim allocation, identity services, externally hosted portals, payments and backup restoration. The European Supervisory Authorities' first incident-reporting findings show why an exclusive emphasis on malicious attacks is analytically incomplete: across the reported financial-sector incidents, system failures and external events were prominent drivers, while only about one tenth were categorized as cybersecurity-related. This published sector-level observation is informative about incident composition, not a measured distribution of insurance-only events.
The reporting clock and information quality
Under Article 19 of DORA and the 2025 reporting-deadline regulation, an initial notification is required as early as possible and ordinarily within four hours of classification as major, and no later than 24 hours from awareness of the underlying incident. Where classification happens more than 24 hours after awareness, the act provides a specific four-hour rule from classification. An intermediate report is due no later than 72 hours after the initial notification, with updates when relevant, and a final report no later than one month after the intermediate or latest updated intermediate report. These are legal time limits for the specified reporting stages, not claims-service restoration targets.
The reporting-forms implementing regulation requires a standardized template and allows good-faith estimates where exact information is unavailable at the initial or intermediate stage, with later correction. A useful insurer response protocol therefore maintains separate timestamps for detection, awareness, provisional impact assessment, major classification, supervisory notification, intermediate update, restoration and final assessment. Conflating the timestamp of discovering a technical issue with the timestamp of recognizing a material effect on a critical insurance service can produce an inaccurate legal chronology.
The information problem is recursive: the ICT disruption may impair the insurer's ability to determine which customers, payments, policies and external providers have been affected. A realistic reporting capability accordingly depends on evidence preserved outside the failed component: reliable event timestamps, service maps, contact lists, alternative communication channels, the ability to estimate affected volumes, and a documented mechanism to qualify estimates. Reporting completeness is more credible when the insurer can preserve the distinction between verified counts, estimates and missing information.
A material governance failure can arise from an incident that is detected technically but not classified promptly because the incident-management team lacks the necessary business context. Consider a claims-payment interface failing at 09:00. Its infrastructure monitoring system may record the outage immediately. The materiality evaluation depends on the service's clients, transaction volume, business criticality and other thresholds, all of which may be maintained outside the ICT team. The appropriate response is a coordinated, time-stamped assessment with business owners and compliance, rather than assuming that infrastructure alerts alone determine Article 8 classification.
Client communication and policyholder consequences
DORA's reporting pathway is principally supervisory. Customer-facing implications also matter because insurers provide promised services, often to customers experiencing illness, disaster or financial distress. A delayed claim can cause real hardship even if the insurance liability remains recorded correctly. Resilience governance should therefore track the functional time between the customer's request and meaningful service: a claim submitted, a case triaged, coverage information retrieved, an urgent payment approved, or an annuity disbursement processed.
Article 19 of the DORA framework also addresses communication to clients in circumstances specified by law. The separate rules on insurance conduct, complaints, privacy and contractual claims handling remain relevant. A compliance report that records only regulatory notification timeliness cannot establish whether customers obtained timely services during the event. The insurer needs distinct data on operational disruption, supervisory reporting and policyholder outcomes to answer those different questions.
A robust post-incident examination should compare the planned recovery point and recovery time, actual service restoration, the interval before accumulated work was completed, integrity-check results, the number of affected policyholders, remedial payments where appropriate, and repeat incidents stemming from the same cause. These are appropriate analytical observations rather than claims that DORA itself prescribes identical numeric targets for all insurers. The legal duties support risk-based control; the measurement of customer consequences supplies evidence of whether those controls achieve their purpose.
Resilience Testing, ICT Third Parties, and Concentration Risk
Testing must reach the insurance function
DORA Articles 24–27 establish a digital operational resilience testing program, including risk-based tests and advanced threat-led penetration testing (TLPT) for entities identified under the applicable criteria. Ordinary tests can assess vulnerabilities, business continuity, recovery, network security, application performance and other relevant capacities. The substantive question is whether the testing program demonstrates that critical insurance activities can continue or be recovered within their defined tolerance under realistic conditions.
The TLPT regulatory standards specify entity-identification criteria, test scope and methodology, tester requirements, remediation and supervisory coordination. DORA's general framework envisages advanced TLPT for selected entities at least every three years, subject to the legal criteria and competent-authority processes. It would be incorrect to infer that every insurance intermediary, small insurer or exempt undertaking must automatically conduct a full TLPT exercise. The designation process and entity eligibility must be documented.
A penetration test can provide evidence that attack pathways were identified and addressed, while a recovery exercise shows how a service is restored after a failure. These test different properties. A successful intrusion test cannot by itself show that an insurer can rebuild its policy database from a trusted backup; a backup test cannot show that a malicious actor cannot manipulate claim-payment instructions. Functional resilience requires a sufficiently broad testing portfolio that verifies defensive, restorative and operational capacities.
The design of a useful claims-recovery test can begin with a specific policyholder outcome. The insurer identifies the claims submission and settlement chain, isolates a realistic failure point, simulates the loss of a critical ICT component and observes each affected dependency, including identity, data consistency, supplier assistance and payment authorization. The report then records both restoration of the computer systems and the subsequent clearance of unprocessed claims. A technically available server is not equivalent to a service that has returned to normal performance.
ICT third-party risk is retained by the insurer
Article 28 of the DORA statute establishes ICT third-party risk as an integral element of the financial entity's risk management. Article 30 identifies contractual provisions, with additional requirements for services supporting critical or important functions. These provisions require attention to service specification, locations, security and availability, assistance during incidents, access, audit and inspection, termination and transition. Outsourcing performance does not outsource legal responsibility for the insurer's regulated services.
Delegated Regulation 2024/1773 specifies the content of a policy for ICT contractual arrangements supporting critical or important functions. Article 3 requires governance assignments and annual policy review; Article 4 follows the contract life cycle, including precontract risk assessment, implementation, monitoring and documentation; Article 10 requires documented and periodically reviewed and tested exit plans. An exit plan is credible when it describes transition resources, alternative dependencies, data portability, security, time requirements and compatibility with contractual termination terms.
The contract must be evaluated as part of an operational mechanism. A formal right to an audit means little if an insurer has no procedure for obtaining the relevant findings, assessing their business implications and requiring remediation. A termination clause cannot assure service continuity if the underlying data are not portable or the alternative provider cannot be activated within the necessary period. Likewise, a supplier's certification provides partial assurance but cannot replace a test of the insurer's own application-specific recovery process.
The ICT subcontracting standards address how financial entities assess subcontracting of ICT services supporting critical or important functions. Subcontracting chains introduce added informational distance and may extend across jurisdictions and cloud infrastructure layers. Their relevance depends on whether the subcontracted service effectively supports the insurer's important function. Governance must therefore connect the direct contract to material subcontractors, changes in the chain, access restrictions, location and operational substitutability. It is inadequate to treat the name of the direct vendor as a complete map of the dependency.
The register of information as a dependency dataset
DORA Article 28(3) requires financial entities to maintain a register of contractual arrangements with ICT third-party service providers, with requirements at relevant entity and group levels. The register templates in Implementing Regulation 2024/2956 prescribe structured information, identifiers, contractual relationships, relevant functions and subcontracting information. Data quality dimensions include accuracy, completeness, consistency, integrity, uniformity and validity. This is not merely a list of vendors: it is a relational description of services, contracts, entities and dependencies.
A register's utility can be tested by asking whether the insurer can answer five operational questions without ad hoc reconstruction: which regulated entity uses a supplier; which important functions depend on it; what underlying contract and significant subcontractor chain govern the service; which jurisdictions and information assets are implicated; and what alternative arrangements exist if the service fails. When the answers depend on incompatible names or missing identifiers, a nominally complete supplier inventory can still fail as a tool for incident response.
The European Supervisory Authorities' 2024 register dry-run results provide an early test of this data problem. Nearly 1,000 financial entities submitted experimental registers, but only 6.5 percent of those analyzed passed all quality checks. This was a voluntary or preparatory best-effort exercise across financial sectors, not a probability sample of insurers and not a measure of final compliance after the regime began. Its relevance lies in demonstrating how difficult common data definitions and complete contractual mapping can be when carried into structured reporting.
Provider criticality and systemic accumulation
DORA establishes an EU-level oversight mechanism for ICT third-party providers designated critical. The ESAs' provider oversight explanation describes how EBA, EIOPA and ESMA designate CTPPs and act as Lead Overseers. This authority complements supervision of financial entities. A provider's designation does not release an insurer from the responsibility to manage its own contract, controls, continuity planning and concentration exposure.
The ESAs announced their first provider designations on November 18, 2025, explaining that the assessment drew on financial entities' registers, systemic importance, services supporting critical functions and substitutability. This is institutional evidence that provider concentration had entered operational supervision. It does not establish that all critical dependencies are covered by the first designated set or that every use of a designated provider is individually critical.
Concentration can arise through several channels: direct contracting with a common cloud provider; dependence on a software platform that itself depends on a particular cloud region; shared identity, authentication or communications services; common software updates; and a subcontractor used by many otherwise unrelated vendors. The same underlying technology may therefore affect insurers that appear diversified at the direct-vendor level. Insurer risk maps must examine functional dependencies and plausible common failures, not merely the count of unique contractual counterparties.
The public supervisory record and the academic literature point toward complementary ways to understand concentration. Buttigieg and Brunelli Zimmermann's 2024 supervisory analysis addresses coordination among authorities, whereas the 2026 ESAs cross-border incident findings provide reported evidence of interconnected incidents. Neither supplies the insurer-level dependency parameters needed to quantify aggregate exposure. Such estimation requires contract-level mapping, incident history, correlated failure assumptions and explicit definitions of affected financial services.
Supplier access, exit and geographic reach
An EU insurer may rely on technology supplied and operated outside its home jurisdiction. Its exposure may depend on data-location rules, access rights, a provider's capacity to restore systems, legal enforceability of relevant obligations, and the availability of an effective replacement. The contract-policy technical standards expressly require assessments of the provider's location, where services are delivered and where data are processed or stored. Legal geography and service geography can differ.
A credible exit strategy must confront transition feasibility. If a provider hosts years of policy, claims and accounting records, transition entails data extraction, field mapping, permissions, reconciliation, interoperability testing, retention duties and operational cutover. A contract that states a right to exit without a realistic schedule and tested alternative provides limited operational protection. The purpose of assessing exit is to understand the interval during which insurer services remain vulnerable to a provider's failure or an enforced termination.
For insurers with long-duration liabilities, historical data integrity is particularly important. Incomplete migration can produce delayed claims or valuation errors long after an infrastructure outage is formally resolved. This links DORA's ICT control requirements to the insurer's continuing obligations under insurance and accounting arrangements. The operational risk cannot be resolved simply by recovering a recent customer-facing webpage while leaving historical policy records unreconciled.
Supervisory Coordination, Insurance Conduct, and Comparative Regulation
The allocation of supervisory powers
The responsibility for DORA is distributed among financial entities, national competent authorities, and European institutions. The national authority supervises relevant insurers and receives incident reporting within its mandate. The European Supervisory Authorities develop common technical standards and coordination arrangements and carry out critical-provider oversight through designated Lead Overseers. The European Commission adopts regulatory and implementing instruments under the legislative delegations. The distinction between rulemaking, entity supervision and provider oversight is central to evaluating whether gaps persist in the overall system.
The DORA oversight framework explicitly places critical-provider supervision alongside continuing financial-entity responsibilities. This distribution addresses a structural problem: a shared provider can generate exposure for institutions supervised in different Member States, while no individual insurer's supervisory authority has the same perspective on the provider's entire cross-sector footprint. A European designation and Lead Overseer can facilitate a broader view; the local insurer remains accountable for what it does with the information, recommendations and contractual protections available to it.
Buttigieg and Brunelli Zimmermann (2024) argue in their DORA supervisory study that harmonized substantive rules operate within an institutional architecture that still requires effective supervisory convergence. The insurer-specific consequence is that comparative evaluation should look beyond the wording of a common technical standard and examine whether supervisory requests, risk classifications and reporting interfaces generate comparable operational information. Uniform templates assist this process but cannot replace consistent interpretation of the underlying service and exposure.
National insurance supervisors may have different institutional boundaries, supervisory resources and portfolios. The literature supports recognizing that coordination challenge; it does not support inventing country-by-country rankings of implementation quality. A valid comparative regulatory study would require a dated collection of national guidance, supervisory inspection findings, comparable insurer populations and consistent definitions of sanctions or remediation. Those materials are outside this dossier's full empirical reach.
EIOPA's 2026 supervisory focus
EIOPA designated DORA as one of its supervisory focus areas for 2026. The associated supervisory program emphasized the engagement of administrative, management or supervisory boards, the suitability of ICT risk frameworks, major-incident monitoring, and third-party risk management including registers of information. These choices are instructive because they test the path from formal procedures to operating governance. A supervisor asking whether the board understands the material outsourced claims dependency is evaluating more than whether a policy has been signed.
The 2027–2029 supervisory priorities, published in September 2026, provide context for continuing European attention to digital and operational issues beyond the first implementation period. They are forward-looking supervisory priorities, and must be distinguished from binding DORA legislative provisions in force at the research cutoff. They also cannot be treated as evidence that any specified insurer has implemented every relevant action.
The first cross-sector incident evidence
The European Supervisory Authorities released the first annual DORA incident report on June 3, 2026, reporting on major ICT-related incidents during 2025. The publication states that 3,383 major incidents were reported across EU financial entities, equivalent to approximately 0.18 incidents per financial entity within the relevant reported population; roughly one third had cross-border impact. The same institutional release identifies system failures and external events as leading drivers, and reports that approximately 10 percent of recorded incidents were cybersecurity-related.
Four restrictions govern interpretation. First, the figures aggregate multiple financial sectors, so the number 3,383 is not an insurer incident count. Second, reported major incidents are defined by legal thresholds; their distribution does not describe all technology failures. Third, implementation and reporting practices may have changed during the first calendar year, affecting observed counts and comparability. Fourth, the incident data provide no causal counterfactual establishing how many disruptions DORA prevented. The appropriate scientific conclusion is narrower and still important: a common reporting mechanism has produced a cross-sector administrative record that identifies meaningful patterns of interconnected ICT disruption.
The report's cross-border observation has a direct insurance implication. A shared insurer technology provider can distribute operational effects across multiple jurisdictions even where each insurer's policyholders and contracts are primarily domestic. Comparing the duration and concentration of incidents would require linking provider relationships, specific services and affected insurers at an appropriate level of confidentiality. Public aggregate totals alone cannot identify whether one supplier caused a disproportionate share of insurance losses.
Comparative interaction with Solvency II
Solvency II and DORA occupy complementary levels of prudential governance. The Solvency II legal framework requires effective governance and risk management that covers operational risk alongside underwriting, market, credit and liquidity risks. An Own Risk and Solvency Assessment can evaluate whether operational threats threaten the insurer's overall risk profile and capital position. DORA supplies greater specificity concerning the technological controls, incident reports, tests and ICT supplier relationships through which those threats arise.
The relation is consequential rather than purely terminological. A significant outage might create immediate remediation expenditure, forgone premium collection, delayed claims decisions, additional assistance expense and reputational damage; it could also change future business or affect reinsurance settlements. Only some consequences will translate into recognized balance-sheet effects or capital requirements over a particular horizon. Recording an ICT event therefore requires separate operational, policyholder, accounting and prudential descriptions. The presence of one valuation number does not explain the event's full service consequence.
A sound insurer management system can connect these domains by mapping business services to ICT assets, identifying credible downtime scenarios, recording direct and indirect financial costs, and assessing whether the event changes existing risk tolerances or prudential assumptions. Yet the resulting capital treatment must follow applicable Solvency II accounting and prudential requirements. The exercise does not convert every hour of downtime into a fixed unit of solvency capital.
Insurance consumer protection and conduct
Digital claims management deserves special consideration because insurance services are often most valuable during adverse circumstances. EIOPA's 2026 supervisory focus statement identifies fair consumer treatment in claims management, including digitalization, as a separate area of attention. The juxtaposition is analytically significant. Operational resilience concerns the continued functioning of services, while conduct supervision asks whether service processes produce fair treatment for customers. A technologically restored claims platform may still leave a backlog of urgent cases or inaccurate decisions requiring a separate consumer-outcome assessment.
For a claims outage, relevant evidence includes the number of cases submitted unsuccessfully, alternative intake offered, delayed urgent payments, case-processing errors after restoration, quality of communication and the clearance of backlogs. It may be appropriate to disaggregate outcomes for vulnerable customers and time-critical insurance benefits when legally and ethically supportable. These are analytic indicators that connect resilience to the insurer's insurance function; their use does not imply that DORA requires any one fixed claims-processing benchmark in every Member State.
Insurance groups should retain the distinction between operational resilience and conduct compliance in the structure of their incident dossiers. A supervisory notification records the legally reportable ICT event; a consumer-remediation record documents customer impact and correction; an internal prudential assessment examines financial soundness and risk appetite. The documents may refer to one incident but answer different questions.
International comparison and transferability
DORA is a binding EU framework with a defined financial-entity scope. Comparable concerns exist internationally through supervisory principles for operational risk, governance, outsourcing and business continuity, but different jurisdictions have different legal sources, firm classifications, reporting thresholds and enforcement powers. The IAIS Insurance Core Principles resource is useful for identifying internationally recognized supervisory domains, although it does not enact EU DORA requirements for firms outside the Union. A comparison should therefore distinguish common regulatory objectives from the jurisdiction-specific authority that implements them.
An insurer group operating both inside and outside the EU may use a common cyber-risk methodology, incident classification taxonomy and supplier inventory, while adapting legal reporting and supervisory communications to each jurisdiction. Shared operational evidence can reduce duplicated assessment, but the obligations and deadlines must be governed by the legal entity and its regulator. This is especially important when one technical incident affects a group across jurisdictions: a global incident team can coordinate facts, yet different national duties can attach to distinct legal entities.
The research result is a comparative method, not a global equivalence finding. For each jurisdiction, identify the regulated entity, the operative legal instrument, the triggering event, the required reporting pathway, the provider-responsibility mechanism and the relevant evidence of effective implementation. International harmonization should be assessed through these observable features rather than by matching the broad labels 'cyber resilience' or 'outsourcing governance.'
Comparative Synthesis and Research Conclusions
Three units of analysis and one insurer service chain
The legal and operational evidence supports a three-level reading of DORA. The first unit is the regulated legal entity, which determines whether duties attach, who is accountable and which authority receives reporting. The second is the insurance function, which determines what must continue or recover for the insurer to fulfill its promises. The third is the shared ICT dependency, through which one failure may propagate among systems, legal entities or sectors. This classification is an explanatory synthesis of the existing regime, not a proposed novel doctrine or supervisory model.
The central practical consequence is that each unit requires a distinct data field. A legal entity identifier is needed for scope and reporting. A service or function identifier is needed for business materiality and recovery priorities. A contract and provider identifier is needed to understand external dependencies and common causes. Without all three, an apparently comprehensive program may remain unable to answer which customers will be affected by a supplier event or which entity must notify a supervisor.
The DORA register ITS provides a regulatory basis for structured supplier information, and the major-incident classification RTS requires service-specific impact information. Their conjunction makes it possible, in principle, to connect supplier mapping and incident classification. Whether an individual insurer's information systems achieve that connection is an empirical governance question requiring operational evidence.
Illustrative calculation: recovery time versus backlog clearance
Consider a deliberately hypothetical non-life insurer receiving 1,000 new claims per day at a steady rate. A total intake-and-processing outage lasts 12 hours, creating an expected backlog of 500 claims under a uniform arrival assumption. After technical restoration, the firm processes 1,250 claims per day while 1,000 new claims per day continue to arrive. Its net backlog reduction rate is then 250 claims per day; clearing the initial backlog requires two days after restoration, assuming capacity remains constant, cases are homogeneous and no prior backlog or rework exists.
The arithmetic is elementary: backlog after interruption equals arrival rate multiplied by outage duration; clearance time equals backlog divided by processing capacity less continuing arrival rate. In notation, with arrival rate λ and post-recovery service rate μ greater than λ, B = λT and clearance time = B/(μ − λ). Taking λ = 1,000/day, T = 0.5 day and μ = 1,250/day yields B = 500 and a clearance duration of 500/250 = 2 days.
No parameter in this example derives from an EU insurer dataset. Its scientific value is to expose two different operational outcomes. The time needed to restart the application is 12 hours; the time before normal claims-processing workload is restored extends a further two days. Claim types with different urgency or processing times may yield much longer consequences. Tests of business continuity should therefore report both service restoration and the recovery of business throughput, particularly where policyholders depend on prompt settlement.
The example also shows why a resilience metric should specify its denominator and time horizon. A technical system available for 99.9 percent of a month may nevertheless experience a concentrated outage during a catastrophe, when claim volumes are unusually high. Equal system downtime need not produce equal customer harm. Meaningful assessment combines availability evidence with actual insurance-service volumes and the timing of demand.
Illustrative calculation: a shared provider and correlated outages
A second hypothetical exercise examines common-vendor dependence. Let C denote a shared provider failure during a defined stress period. Suppose P(C) = 0.02. For each of 20 insurers, let the probability of an interruption be 0.80 conditional on C and 0.01 conditional on no C. Assume insurer interruptions are independent conditional on the shared event. These deliberately stylized inputs imply a marginal interruption probability of 0.02 × 0.80 + 0.98 × 0.01 = 0.0258 for each insurer. The expected number of affected insurers is 20 × 0.0258 = 0.516 per stress period.
The expectation alone conceals dependence. For two insurers i and j, conditional independence yields covariance Cov(Xi,Xj) = Var[E(Xi|C)] = P(C)[1 − P(C)](0.80 − 0.01)². Using the stated inputs produces 0.02 × 0.98 × 0.79² = 0.01223236. The variance of one Bernoulli interruption variable is 0.0258 × 0.9742 = 0.02513436, giving a pairwise correlation of approximately 0.487. The illustrative assumption set therefore combines a modest marginal interruption frequency with substantial cross-entity association.
This is an exposition of standard conditional probability, not a newly invented statistical model and not an estimate for any real cloud provider or EU insurer. Changing the probability of the common event, the conditional failure rates, the definition of affected service or the assumption of conditional independence materially changes the outcome. The exercise should not be used for capital calibration without empirical identification, exposure mapping and validation.
The regulatory relevance is that individual providers can appear reliable in isolated service-level statistics even while creating portfolio-wide common-event exposure. The European supervisory designation process is designed to identify important providers at system level. Entity-level DORA controls, including supplier assessment, business continuity and incident reporting, must still evaluate what such common events mean for each insurer's functions. System-level attention and entity-level discipline serve complementary purposes.
Interpreting early evidence without overclaiming effects
The 2024 register dry run and the 2025 incident observations published in 2026 have different denominators and meanings. The register exercise evaluated the quality of preparatory submissions from a participating cross-sector set. The annual incident release aggregated legally classified reported incidents during the first implementation year. Neither should be interpreted as the compliance rate of EU insurers, the frequency of all insurer outages, or a causal estimate of loss reduction from DORA.
Even valid post-implementation time series require careful design. An increase in reported major incidents could reflect improved detection, broader reporting, changed classification or deteriorating operational conditions. A decrease could reflect effective controls, underreporting, threshold effects or differences in the population of reporting entities. A credible empirical test would distinguish reporting propensity from underlying failure frequency and would specify baseline exposure, relevant policyholder service, firm characteristics, time period and legal applicability.
Relevant future research could combine a defined insurer panel with consistent major-incident definitions, outsourcing concentration indices, documented changes in testing and recovery practices, and observed business-service consequences. Potential outputs include incident frequency and severity, affected-client durations, restored-service time, recurrence, third-party involvement and insurer costs. Statistical comparisons should control for firm scale, national reporting architecture, differences in life and non-life service processes, and the timing of technical standards. Data privacy, incident sensitivity and selection bias would remain material constraints.
The institutional significance of insurance obligations
The distinctive importance of DORA for insurers lies in the durability and urgency of the promises administered through digital services. A policy issued years earlier may require accurate records and a payment today. Restoration that preserves database availability but corrupts policy history can impair the quality of a later benefit decision. Supplier concentration also matters over horizons longer than a typical IT project, because migration of accumulated policy and claims information can be difficult and expensive.
This observation supports a coherent management interpretation. Insurer boards need a view of ICT risks in terms of business continuity and policyholder obligations; operational teams need clear recovery and incident responsibilities; compliance teams need legally timed reporting; procurement and risk functions need supplier-control evidence; supervisors need comparable, accurate and timely information. A mature system must keep these views connected without collapsing their distinct legal and operational questions.
Research findings
The documentary analysis supports six specific findings. DORA's insurance perimeter is determined by legal classification and express exemptions, with proportionality applied inside the relevant scope. Board accountability involves demonstrable decisions about technology-supported business functions. Major-incident reporting requires a service-specific threshold analysis and a reliable incident chronology. Testing must evaluate both technical recovery and the actual continuity of insurance operations. Third-party registers and contractual controls are useful when they reveal functional and subcontracting dependencies, including exposure to shared providers. Finally, early EU-wide data establish the existence and structure of supervisory information flows, but publicly available aggregates do not establish insurer-specific causal improvements in resilience.
These findings follow from legal texts, identified supervisory observations, published research and transparent illustrative calculations. They do not claim the creation of a new legal principle, statistical model or general theory. Their contribution is an integrated account of how legal obligations, operational service continuity and provider concentration interact within the insurance business.
Conclusion
DORA makes digital operational resilience a continuing governance responsibility for insurers within its legal scope. Its rules extend across management-body accountability, ICT risk assessment, incident detection and reporting, resilience tests, provider contracts, structured registers, supplier concentration and supervisory coordination. The regulation's operational meaning becomes clear when these requirements are connected to the services insurers owe to policyholders and to the technical dependencies through which they are delivered.
The most consequential interpretive finding is that accountability must be established at the proper unit of analysis. The insurer as a legal entity bears duties; the insurance function defines the service whose failure matters; external ICT providers and subcontractors explain potential propagation and common exposure. Strong oversight requires a reliable mapping among those levels. A signed ICT policy, an incident-notification template or a vendor register considered alone cannot demonstrate the resilience of an insurance service.
The first EU-wide reporting cycle makes both progress and analytical limits visible. The ESAs' published 2025 incident findings, released in June 2026, provide evidence of reported cross-border disruption and the role of system failures and external events. They do not measure the causal effectiveness of DORA for insurers or establish insurer-specific event rates. Further research requires dated, comparable insurer-level evidence linking exposure, controls, tests, incident outcomes and policyholder services.
For insurance science, the governing research problem is consequently measurable continuity of contractual performance under technological disruption. DORA provides a detailed legal architecture; its effectiveness will be judged by the quality of insurer governance, the reliability of information, the realism of tests, the feasibility of supplier transitions and the actual experiences of customers when systems fail. Those criteria provide a disciplined agenda for studying resilience without presuming outcomes that the current evidence cannot support.
References
Biener, Christian, Martin Eling, and Jan Hendrik Wirfs. 2015. “Insurability of Cyber Risk: An Empirical Analysis.” The Geneva Papers on Risk and Insurance—Issues and Practice 40 (1): 131–158. https://doi.org/10.1057/gpp.2014.19.
Buttigieg, Christopher P., and Beatriz Brunelli Zimmermann. 2024. “The Digital Operational Resilience Act: Challenges and Some Reflections on the Adequacy of Europe’s Architecture for Financial Supervision.” ERA Forum 25: 11–28. https://doi.org/10.1007/s12027-024-00793-w.
European Commission. 2024a. Commission Delegated Regulation (EU) 2024/1772 of March 13, 2024, on the Classification of ICT-Related Incidents and Cyber Threats. Official Journal of the European Union, June 25, 2024. https://eur-lex.europa.eu/eli/reg_del/2024/1772/oj/eng.
European Commission. 2024b. Commission Delegated Regulation (EU) 2024/1773 of March 13, 2024, on Policy Regarding Contractual Arrangements for ICT Services Supporting Critical or Important Functions. Official Journal of the European Union, June 25, 2024. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1773.
European Commission. 2024c. Commission Delegated Regulation (EU) 2024/1774 of March 13, 2024, Specifying ICT Risk Management Tools, Methods, Processes, and Policies and the Simplified Framework. Official Journal of the European Union, June 25, 2024. https://eur-lex.europa.eu/eli/reg_del/2024/1774/oj/eng.
European Commission. 2024d. Commission Implementing Regulation (EU) 2024/2956 of November 29, 2024, on Standard Templates for the Register of Information. Official Journal of the European Union, December 2, 2024. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R2956.
European Commission. 2025a. Commission Delegated Regulation (EU) 2025/301 of October 23, 2024, on the Content and Time Limits of Major ICT-Related Incident Reports. Official Journal of the European Union, February 20, 2025. https://eur-lex.europa.eu/eli/reg_del/2025/301/oj/eng.
European Commission. 2025b. Commission Implementing Regulation (EU) 2025/302 of October 23, 2024, on Forms, Templates, and Procedures for Reporting Major ICT-Related Incidents. Official Journal of the European Union, February 20, 2025. https://eur-lex.europa.eu/eli/reg_impl/2025/302/oj.
European Commission. 2025c. Commission Delegated Regulation (EU) 2025/532 of March 24, 2025, on Subcontracting ICT Services Supporting Critical or Important Functions. Official Journal of the European Union, July 2, 2025. https://eur-lex.europa.eu/eli/reg_del/2025/532/oj.
European Commission. 2025d. Commission Delegated Regulation (EU) 2025/1190 of February 13, 2025, on Threat-Led Penetration Testing. Official Journal of the European Union, June 18, 2025. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32025R1190.
European Commission. 2026. “Digital Operational Resilience Regulation: Implementing and Delegated Acts.” Directorate-General for Financial Stability, Financial Services and Capital Markets Union, regulatory documentation inventory, consulted November 15, 2026. https://finance.ec.europa.eu/regulation-and-supervision/financial-services-legislation/implementing-and-delegated-acts/digital-operational-resilience-regulation_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2024a. “ESAs' Dry Run Exercise Shows the Goal of Reporting of Registers of Information under Digital Operational Resilience Act in 2025 within Reach.” December 17, 2024. https://www.eiopa.europa.eu/esas-dry-run-exercise-shows-goal-reporting-registers-information-under-digital-operational-2024-12-17_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2024b. “Opinion on the Scope of DORA in Light of the Review of the Solvency II Framework.” November 14, 2024. https://www.eiopa.europa.eu/publications/eiopa-opinion-scope-dora-light-review-solvency-ii-framework_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2025. “Union-Wide Strategic Supervisory Priorities: Focus Areas for 2026.” October 1, 2025. https://www.eiopa.europa.eu/publications/union-wide-strategic-supervisory-priorities-focus-areas-2026_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2026a. “Digital Operational Resilience Act (DORA).” Official supervisory resource, consulted November 15, 2026. https://www.eiopa.europa.eu/digital-operational-resilience-act-dora_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2026b. “DORA Oversight.” Official supervisory resource, consulted November 15, 2026. https://www.eiopa.europa.eu/digital-operational-resilience-act-dora/dora-oversight_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2026c. “ESAs Publish the First Report on DORA Major ICT-Related Incidents.” June 3, 2026. https://www.eiopa.europa.eu/esas-publish-first-report-dora-major-ict-related-incidents-2026-06-03_en.
European Insurance and Occupational Pensions Authority (EIOPA). 2026d. “Union-Wide Strategic Supervisory Priorities 2027–2029.” September 30, 2026. https://www.eiopa.europa.eu/publications/union-wide-strategic-supervisory-priorities-2027-2029_en.
European Parliament and Council of the European Union. 2009. Directive 2009/138/EC of November 25, 2009, on the Taking-Up and Pursuit of the Business of Insurance and Reinsurance (Solvency II). Official Journal of the European Union. https://eur-lex.europa.eu/eli/dir/2009/138/oj.
European Parliament and Council of the European Union. 2022a. Regulation (EU) 2022/2554 of December 14, 2022, on Digital Operational Resilience for the Financial Sector. Official Journal of the European Union, December 27, 2022. https://eur-lex.europa.eu/eli/reg/2022/2554/oj.
European Parliament and Council of the European Union. 2022b. Directive (EU) 2022/2556 of December 14, 2022, Amending Financial-Services Directives as Regards Digital Operational Resilience for the Financial Sector. Official Journal of the European Union, December 27, 2022. https://eur-lex.europa.eu/eli/dir/2022/2556/oj/eng.
European Securities and Markets Authority (ESMA), European Banking Authority, and European Insurance and Occupational Pensions Authority. 2025. “The European Supervisory Authorities Designate Critical ICT Third-Party Providers under the Digital Operational Resilience Act.” November 18, 2025. https://www.esma.europa.eu/press-news/esma-news/european-supervisory-authorities-designate-critical-ict-third-party-providers.
International Association of Insurance Supervisors (IAIS). 2026. “Insurance Core Principles Online Tool.” Current supervisory principles resource, consulted November 15, 2026. https://www.iais.org/icp-online-tool/.