Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

International Insurance Groups Legal Entity Maps and Control Chains

Abstract. International insurance groups combine legally differentiated institutions under arrangements that allocate authority, financial participation, insurance liabilities, capital support and supervisory responsibility across national borders. This monograph studies the reconstruction of those arrangements from publicly accessible evidence. Comparative institutional analysis integrates the international supervisory framework, selected European Union and New York provisions, accounting standards and legal-entity identifier documentation with dated corporate observations including American International Group's significant-subsidiary hierarchy at December 31, 2024; its June 2024 deconsolidation of Corebridge Financial while retaining a substantial shareholding; Zurich Insurance Group's contractual relationship with Farmers Exchanges in its 2024 reporting; Chubb's 2025 hierarchy and Huatai consolidation; and Swiss Re's 2021 ownership and reporting changes. A worked network calculation distinguishes immediate voting interests from indirect economic participation and tests the sensitivity of a reconstructed chain to omission and threshold choice. The analysis finds that a corporate hierarchy can establish certain reported ownership links, but accounting membership, prudential scope, operational dependence and the identity of the policy obligor require separate evidence. Conditional control assessments and event-dated perimeter reconciliation explain apparent discrepancies between otherwise valid records. The contribution is an integrated, reproducible interpretation of selected institutional evidence rather than a global census or a novel theory. The comparative findings identify where public maps support inference, where supervisory or contractual information is indispensable, and why consolidation cannot be treated as a guarantee of capital availability.

Keywords: Insurance groups; legal entities; ownership chains; voting rights; effective control; accounting consolidation; group supervision; cross-border insurance; legal entity identifiers; prudential perimeter; corporate governance; institutional risk.

Introduction

The international insurance group is an institutional arrangement through which multiple legal persons, branches and associated organizations coordinate the assumption, administration, financing and transfer of risk. Its public identity may be concentrated in a single name while insurance obligations, assets and governance powers remain allocated among particular institutions. Understanding a group therefore requires a reconstruction of relationships as well as identification of the entities carrying them.

The principal research question is how to establish, compare and interpret the legal and economic boundaries of an international insurance group from evidence generated for different purposes. Three dimensions organize the investigation: the network of immediate and indirect ownership, arrangements that establish effective decision-making authority, and the consolidation perimeter specified by a financial or supervisory framework. The explanatory problem is that the same institution can appear within one boundary and outside another without either boundary being erroneous.

A legal-entity map affects financial analysis in several ways. An investor may be interested in consolidated earnings; a policyholder may need to identify the undertaking responsible for a particular claim; a supervisor may need to investigate transmission of loss from a related company; and a market statistician may need to count undertakings consistently. A map designed for only one of these questions can produce misleading results when reused for another. For example, ten companies carrying a common brand might represent ten separate licensed undertakings, ten subsidiaries of one holding company, or a mixture of branches and contractual affiliates. Those possibilities imply different units of observation.

This research uses document-based comparative institutional analysis rather than a representative survey. The legal comparison covers the IAIS international framework, selected Solvency II provisions as displayed before the 2027 reform, and one US state statute. The corporate observations are deliberately dated to 2024. Their selection permits comparison of a disclosed hierarchy, a documented change in accounting control, and a contractual business relationship without share ownership. The empirical results establish the existence and interpretive importance of these mechanisms; they cannot estimate their frequency in the worldwide insurance population.

Financial reporting, prudential analysis and corporate legal identity are considered in their own terms. References to standards identify their subject and relevant version; corporate statements are treated as issuer disclosures rather than independent adjudications of every underlying legal fact. Public information cannot reveal every agreement, beneficial interest or supervisory determination. The conclusions therefore concern demonstrable relationships and the conditions under which further inference is justified.

Concepts, Institutions and Historical Development

The legal person and the insurance enterprise

A legal person is the unit to which a legal system attributes rights, obligations and legal capacity. For the purpose of group research, a legal entity should be represented by its exact legal name, jurisdiction of formation, available registration identifier and effective dates. Authorization to conduct insurance is a further property of an undertaking, determined under the relevant jurisdictional rules. Registration and authorization are separate events; incorporation in a jurisdiction does not by itself establish permission to write any insurance class there.

A subsidiary has its own legal identity. A branch is an establishment of an existing legal entity rather than a further incorporated subsidiary merely because it has a local name or registration. A branch may have a separate supervisory or reporting designation and remains important to the geography of insurance business. The map should therefore connect a branch to the legal person of which it is an establishment and preserve its host-jurisdiction attributes. Assigning the branch an independent equity share would create a fictional ownership relationship.

A holding company may operate predominantly through investments in other companies. An operating insurer may simultaneously issue insurance and hold subsidiaries. Service providers, investment managers, brokers, reinsurers and other organizations may be connected to the same corporate hierarchy. Their inclusion in a legal group does not convert their activities into insurance underwriting. The role of every node must be established independently of its corporate name.

An insurance enterprise as experienced by customers can encompass a broader economic network than a set of consolidated corporations. Agents, distributors, independent claims administrators and external reinsurers may be structurally important to an insurer without being owned by it. Analytical attention to these relationships is warranted by their economic function, while group membership depends on the applicable criterion. This separation also helps prevent the attribution of a contractual counterparty's entire financial exposure to an insurer.

Ownership and the development of consolidation

Corporate ownership structures allow pools of capital to be divided into separately incorporated investments. In a simple share company, a direct holding records a shareholder's entitlement under the relevant class of securities. Voting power records the authority attached to those securities. These quantities can diverge through multiple share classes, voting agreements, intermediate holdings, or rights that are not proportional to equity.

La Porta, Lopez-de-Silanes and Shleifer (1998), in their comparative study of ultimate corporate ownership, demonstrated the importance of tracing ownership beyond the immediate shareholder. Their historical research on large companies in wealthy economies documents separation of economic participation from control through corporate pyramids. Insurance-specific conclusions require insurance-specific records; the study supplies a methodological precedent rather than a sector estimate.

The accounting problem of identifying the reporting group developed alongside the corporate ownership problem. The IASB's IFRS 10 standard history records the integration of earlier consolidation and special-purpose entity guidance into a standard issued in 2011. Its central function is to determine which entities are presented together in consolidated financial statements. The companion disclosure framework extends the reader's view to significant interests beyond the consolidated group. Such reporting responds to the needs of financial statement users; it leaves the legal identity of each constituent undertaking intact.

Insurance supervision developed a distinct concern with the relationship between an insurer's own resources and the behavior of its affiliates. A regulated company may be exposed to risks generated by an unregulated parent, a financing company or an operational service company. Group-wide supervision addresses these relationships while retaining the prudential relevance of individual insurers.

The IAIS's December 2024 ICPs and ComFrame distinguish group-wide supervision arrangements that rely on direct powers over group entities from arrangements that address group risk indirectly through regulated entities. The forms can coexist within different supervisory systems. International coordination seeks comparable outcomes without assuming an identical statutory allocation of authority in each jurisdiction.

The boundaries that must be distinguished

A legal ownership perimeter is the set of entities connected by the specified ownership relation at a particular date. An accounting perimeter is the set included in a particular reporting entity's consolidation under the applicable standards. A prudential perimeter is the scope within which a designated supervisor assesses group risks and applies or coordinates relevant supervisory measures. An operational network describes material service, financing, distribution or risk-transfer dependencies. Each definition identifies a relationship and a purpose.

The identity of the group's head must also be qualified. A top incorporated parent may have no further corporate parent, yet its beneficial ownership can remain dispersed or opaque. An accounting ultimate parent identifies the highest relevant consolidating entity. A supervisory head is determined for the purpose of group-wide oversight. A resolution authority may organize crisis measures around another legal level. These roles can coincide and should be recorded separately even when the names match.

The policy obligor is an additional economically important identity. Insurance liabilities attach to the applicable issuing undertaking and enforceable contractual arrangements. The legal map makes it possible to distinguish the entity named on a policy from an intermediary displaying the group's brand. It is insufficient by itself to establish the terms of a guarantee, the priority of claims or the admissibility of a particular insurance arrangement.

Primary Market Sources and Comparative Research Design

Evidence selection and research units

The study uses a purposive design with observations from public company reporting and authoritative regulatory or standard-setting materials. AIG's Exhibit 21 offers a dated hierarchy with voting percentages attributed to immediate parents. AIG's June 24 current report describes a specific control event and its accounting consequences. Zurich's 2024 annual-report presentation supplies a separate instance in which a material business relationship exists without an ownership stake in the institutions serviced.

The appropriate observational units are legal entities, typed relationships and dated changes, rather than entire corporate groups treated as indivisible observations. For legal comparison the unit is the specified provision or institutional framework. For a disclosed transaction the unit is the event described in the filing. The same group can consequently provide several observations without entering the sample several times as an independent corporation.

The IAIS public IAIG register demonstrates how a sampling frame for large, internationally active groups can be constructed. Identification involves both geographic activity and size: premiums in at least three jurisdictions, foreign gross written premiums representing at least one tenth of total gross written premiums, and either assets of at least USD 50 billion or gross written premiums of at least USD 10 billion on the stipulated three-year averaging basis for size. The framework allows limited supervisory discretion. This register is a frame for identified IAIGs, not a list of all cross-border insurers. The opened register snapshot displayed June 3, 2026; no current total is asserted because retrieved snapshots were not synchronized.

Entity resolution and evidentiary precedence

The preferred identification key combines the legal name and jurisdiction with a stable official identifier where available. The entity identifier is distinct from its corporate parent, ultimate owner and operating brand.

GLEIF's summary of the Regulatory Oversight Committee's Level 2 parent-data policy explains that the relationship records concern accounting consolidation. This is especially useful when a reported child can be associated with its direct and ultimate consolidating parents. The records do not by that fact establish a direct legal-shareholding edge or its percentage. The direct consolidator can also differ from an immediate shareholder that does not itself prepare relevant consolidated accounts.

The accompanying reporting exceptions documentation identifies several reasons why a parent relationship can be absent from published Level 2 records, including non-consolidation and non-public information. Missing data must retain the distinction between an explicit reported exception and the failure of the research process to find a record. Encoding every absence as an independent company would systematically overstate the number of group heads.

A reliable evidence hierarchy is purpose-specific. A corporate registry can establish incorporation and filed officers to the extent its rules provide; a licensing register can establish an authorization as recorded; a subsidiary exhibit can establish issuer-reported links; audited financial reporting can establish the stated accounting treatment; and a supervisor's decision can establish the regulatory treatment within its legal effect. A generic commercial database may aid discovery, but agreement across commercial replicators is weak corroboration if they all copied the same origin.

For each relationship the researcher should preserve the source, document date, relationship effective date or reporting date, exact unit, nature of the assertion, percentage if available, and known limitations. Incorporation jurisdiction and location of insurance business belong in distinct fields. Where only the country of incorporation is reported, an inference about premium location would require additional evidence.

A reproducible comparative procedure

For a given group and date, the investigator first fixes a root and records the purpose of the map. The investigator extracts the names and parent relations directly from the source, preserving the original hierarchy. Each edge receives a type: equity participation, voting rights, consolidation, supervisory inclusion, branch establishment, service relationship, guarantee, or risk transfer. Where the source supplies a percentage it is associated with its stated denominator; where it does not, the percentage remains unknown.

The second stage reconciles alternative sources by date and definition. An ownership link at fiscal year-end and a later transaction disclosure can coexist in a time series. Two sources describing different accounting and regulatory parents can coexist at the same date. Inconsistent claims about the same typed relation at the same effective date require a documented discrepancy and additional evidence.

The third stage tests how outcomes change with classification. The selected measures include the number of edges, number of cross-border edges under an explicit jurisdiction convention, path length, and numerical economic interest only when compatible share rights are supplied. Incomplete lists provide lower bounds on observed nodes and edges rather than estimates of total group size. A missing intermediate node can shorten an observed path, so depth should not be ranked across groups before harmonizing disclosure completeness.

The final stage identifies which claims are supported by the observed documents and which require another source. Each calculated result below follows from an explicitly described input. No sample average, causal regression, coverage estimate or statistically significant industry effect is reported. The small selection is designed for institutional discrimination; its inference is logical and documentary.

Comparative legal scope and temporal controls

The EU comparison uses the rules displayed by EIOPA for Articles 214 and 220. EIOPA's Solvency II framework overview dates the application of the forthcoming amended framework to January 30, 2027. This study does not apply the future regime to the 2024 corporate observations or claim comprehensive coverage of national transposition.

The US comparison uses New York Insurance Law as an example of an enacted state rule. Its application to a particular entity requires the relevant jurisdictional nexus and statutory conditions. A state-law control determination is not inferred from an accounting decision. The IAIS framework provides an additional international supervisory reference, rather than a universally self-executing statute.

Parent Subsidiary Networks

A dated cross-border hierarchy

American International Group's 2024 subsidiary exhibit identifies subsidiaries by jurisdiction of organization and voting securities held by their immediate parent. Its indentation carries relational information. Reading the rows as a flat list would preserve names while destroying the disclosed chains.

A selected transcription illustrates the structure. The following rows reproduce factual relationships and percentages, not the whole exhibit. The observation date is December 31, 2024.

Immediate parent Reported child Child jurisdiction Voting securities (%)
American International Group, Inc. AIG Property Casualty Inc. Delaware 100
AIG Property Casualty Inc. AIG Property Casualty International, LLC Delaware 100
AIG Property Casualty International, LLC AIG International Holdings GmbH Switzerland 100
AIG International Holdings GmbH AIG APAC HOLDINGS PTE. LTD. Singapore 100
AIG APAC HOLDINGS PTE. LTD. AIG Asia Pacific Insurance Pte. Ltd. Singapore 100
AIG APAC HOLDINGS PTE. LTD. AIG Australia Limited Australia 100
AIG International Holdings GmbH AIG Europe Holdings S.a.r.l Luxembourg 100
AIG Europe Holdings S.a.r.l AIG Europe S.A. Luxembourg 100

The eight edges connect nine distinct nodes, with five relationships along the path from the root to AIG Asia Pacific Insurance Pte. Ltd. and five along the path to AIG Europe S.A. (root, property casualty, international, Swiss holding, European holding, European insurer). Both terminal paths therefore have depth five. Australia also has depth five. Under the convention that Delaware is classified as United States for country-level comparisons, four of the eight listed links cross national borders: United States to Switzerland, Switzerland to Singapore, Singapore to Australia and Switzerland to Luxembourg. The remaining four link entities within the same country. The table is a deliberately selected subgraph, so these counts characterize that subgraph alone.

These observations distinguish cross-border legal organization from cross-border underwriting. A five-edge path reaching a Singapore-incorporated organization gives no direct information about where its premiums were written. A Swiss holding intermediary indicates a corporate jurisdiction within the chain; it does not identify the legal issuer of every insurance contract administered by entities in that chain.

Regulatory corroboration of the operating nodes

An ownership schedule and a licensing register answer different questions. Three names from the selected AIG chain can be resolved against supervisory records. APRA (2026) lists AIG Australia Limited among undertakings permitted to undertake new or renewal general insurance business. MAS (2026) records AIG Asia Pacific Insurance Pte. Ltd. as incorporated in Singapore and classified as a direct general insurer. CAA (n.d.) identifies AIG Europe S.A. with legal entity identifier 213800SCCLMKOWSSX732 and enumerates non-life classes 1 through 18.

These observations identify three operating risk carriers in distinct jurisdictions and provide independent evidence of their regulatory roles. They do not retrospectively establish license conditions on December 31, 2024, the effective date of the historical ownership schedule. The temporal separation is analytically consequential: a current register cannot establish an earlier authorization date, and neither a registration nor an authorization proves the current ownership chain. A repeatable study therefore dates its shareholding observations separately from its regulatory observations and records the source-specific uncertainty of each.

Multiplicative participation and branching

For a simple acyclic chain of ordinary shares with proportional cash-flow entitlements, the indirect economic-interest fraction along a single path equals the product of the direct fractions. If a parent has 0.8 of an intermediate company and that intermediate has 0.6 of an insurer, the indirect fraction is 0.48, or 48 percent. This is an arithmetic example, not an observation from AIG. It assumes ordinary proportional rights, no cross-holdings, and a single path; voting control cannot be calculated solely by multiplying those fractions.

For a purely illustrative chain of three 60 percent equity holdings, the economic fraction is 0.6 × 0.6 × 0.6 = 0.216. If each majority can effectively direct its respective investee under the applicable facts and rules, control can persist across the chain while the ultimate economic participation is only 21.6 percent. This example explains why a rule that selects every ultimate economic interest above 50 percent would omit potentially controlled businesses.

For two non-overlapping direct shareholdings in the same company, under a common denomination the stakes can be summed. AIG's exhibit note 3 reports that AIG Seguros Brasil S.A. has one 90.56 percent holding and another 9.44 percent holding by a second group holding company. These reported voting interests total 100 percent. Omitting the footnote would incorrectly leave 9.44 percent unaccounted for. The calculation establishes the disclosed allocation of voting securities; it does not by itself establish class-specific dividend entitlements.

Network calculations therefore need a distinction between path-based participation and the aggregation of distinct holdings. A graph that contains repeated links to the same shares must not add them twice. Circular shareholdings require a different treatment because the same economic claim can re-enter the path. Where equity classes or ownership denominators are unresolved, reporting an exact ultimate percentage would exceed the evidence.

The comparative literature provides a larger historical context. Kartasheva (2021), in Structure and Complexity of Global Insurance Groups, constructed a 2018 asset-ranked sample of 100 insurance groups headquartered in 21 economies. Their approximately 8,000 subsidiaries extended across 117 countries and offshore territories. Geographic diversity, activity diversity, and organizational depth were separate dimensions of the investigation. The sample and reference year differ from the selected individual-company observations in this dossier; the paper's results cannot be extrapolated to the current worldwide population. The distinction supports a measured reading of the AIG reconstruction: a verified path demonstrates cross-border layering, whereas comparative population claims require a defined sampling frame and harmonized dates. The present contribution is a documented case-level interpretation of typed relationships, not a claim to have originated the empirical study of insurance-group complexity.

Corporate purpose and alternative explanations

Intermediate companies can facilitate acquisitions, common governance, regional administration, investment holding, segregation of activities and compliance with local legal arrangements. The existence of an intermediate layer alone does not establish a motive. A graph of national locations provides a way to ask which legal, tax or operational explanations are consistent with the observed arrangement; establishing intent requires transaction documents or other direct evidence.

The alternative hypothesis that a layered ownership chain exists principally for regulatory arbitrage may be relevant in certain cases, but the AIG table cannot adjudicate that question. An observed location is compatible with several purposes. Similarly, the name of a subsidiary is insufficient evidence that it has an active authorization, writes insurance in a particular class or holds a specified amount of capital.

The defensible analytical result is that corporate paths can identify where potential ownership authority traverses jurisdictions, and where additional examination is likely to be necessary. They are a discovery structure for the legal and financial research that follows.

Cross-group comparison: Chubb and Swiss Re

A second listed multinational, Chubb, provides an independent corporate disclosure for comparison. In the 2025 subsidiary exhibit (Chubb Limited, 2026), the Swiss parent is linked to Chubb Group Management and Holdings Ltd. in Bermuda, which holds Chubb Bermuda Insurance Ltd., also in Bermuda. The schedule places Paget Reinsurance Ltd. in the next tier. The three reported ownership fractions on this path are each 100%. Its product is consequently 100%, subject to the interpretation of ownership percentages in the exhibit. Three edges connect four nodes, with one crossing between jurisdictions. The path is illustrative rather than a statement about the entire group.

The same exhibit places Oasis Real Estate Company Ltd. beneath Chubb Bermuda Insurance Ltd. and Scarborough Property Holdings Ltd. beneath Oasis with a reported 40% holding. An unqualified inference of full ownership based on the word subsidiary would therefore fail. A simple product on the selected path yields 40% under the same percentage assumptions. Other paths, agreements, and share-class rights cannot be reconstructed from this row alone. Chubb explicitly acknowledges the omission of other subsidiaries within the stated significance qualification. It follows that a depth or entity-count comparison with the earlier AIG exhibit would be invalid without harmonized dates, disclosure thresholds, and complete populations.

A European prudential disclosure supplies a different type of ownership observation. The Swiss Re International 2021 SFCR (Swiss Re International SE, 2022, pp. 5–6) depicts five legal persons connected by four 100% relationships: Swiss Re Ltd., Swiss Reinsurance Company Ltd., Swiss Re Corporate Solutions Holding Company Ltd., Swiss Re International SE, and Swiss Re Corporate Solutions Insurance China Ltd. The first three were Swiss; the fourth Luxembourgish and the fifth Chinese. The selected path therefore contains two international transitions. This is a historical solvency-report snapshot, not a claim about the group's 2026 structure.

The same SFCR reports a Labuan branch transfer in 2021. A branch is an establishment of a legal person, so a change in the undertaking to which a branch belongs cannot be inferred solely from the presence of the branch jurisdiction in an ownership tree. The transfer evidence establishes a change in assets and liabilities that a legal-person-only hierarchy would miss.

Taken together, the selected AIG, Chubb and Swiss Re paths show the reproducibility of a limited, explicitly dated network reconstruction across different reporting forms. The differences in path length are descriptive observations within selected paths; they provide no basis for ranking the three entire groups by complexity.

Incomplete disclosure and the observed network

AIG states in its exhibit notes that some subsidiaries are omitted because their aggregate significance does not reach the specified subsidiary significance condition, and that substantially all listed subsidiaries are included in its financial consolidation. The group map obtained from the exhibit is therefore a disclosed subset. Treating it as an exhaustive census would erase the distinction between disclosure thresholds and corporate existence.

Suppose a researcher loses the Swiss holding node when importing the eight-edge subgraph and replaces its two reported outward links by inferred links from the US parent and removes the incoming Swiss link. The paths to Singapore and Luxembourg become one edge shorter, while total observed cross-border links change from four to three under the stated country convention. The analyst would seemingly discover a shallower and less geographically complex network entirely because of a data transformation. If the missing node is instead left unresolved, no direct replacement link is justified.

The correct remedy is to distinguish observed edges from transitive reachability. A transitive connection can record that one entity is somewhere above another in the disclosed hierarchy. It must not be exported as a claim identifying the immediate legal shareholder.

Effective Control

Accounting control and decision rights

The accounting concept of control focuses on the capacity to affect an investee's relevant activities and the economic results of that capacity. IASB's IFRS 10 links the obligation to consolidate to control, subject to its specified exceptions. The June 2026 IFRIC Update explains the conjunction of three elements in paragraph 7: decision-making power, exposure to variable outcomes and the ability to influence those outcomes through that power.

An analyst must therefore investigate more than the percentage of shares. Appointment rights, the dispersion or organization of other voters, agreements allocating decisions, restrictions, removal mechanisms and the distinction between a delegated agent and the principal may matter. These factors belong in the evidence record rather than being collapsed into a single continuous ownership percentage.

The 2026 Committee discussion concerns a specified investment-fund fact pattern. Its significance here is methodological: an agent's decision-making authority cannot automatically be attributed to the sole outside investor without assessing the relevant delegation. It does not provide a determination about any insurance group in this study.

AIG and Corebridge Financial

The decisive event in the comparative evidence is AIG's change in accounting treatment of Corebridge. AIG's June 24 Form 8-K states that it held 48.4 percent of Corebridge common stock on June 9, 2024. At that date AIG waived its entitlement to majority representation on Corebridge's governing board and one designated member resigned. AIG reported achieving deconsolidation under US GAAP and applying the fair value option to its retained equity-method investment. It also described the discontinued-operations presentation and corresponding changes to segment information.

The June 2024 quarterly filing independently within the issuer's reporting sequence describes the deconsolidation and the proposed Nippon Life share purchase. The announcement of a future transaction and the completed change in governance are separate events. The ownership change associated with a proposed sale cannot be dated as a completed transfer merely because the agreement was signed.

The comparison isolates an especially important methodological point. A residual ownership interest of 48.4 percent is compatible with the issuer's reported loss of accounting control. A classification that identifies affiliation with the presence of any substantial investment would still place Corebridge inside AIG, while the disclosed accounting presentation changes. A simple majority-share screen would require a prior dated ownership observation before it could identify when that threshold was crossed. The documented event is the change in rights and the reported deconsolidation, not a general conclusion that 48.4 percent always implies or excludes control.

The case also illustrates data revision. Continuing-operations figures can change presentation after a corporate event, including the classification of earlier periods. A time series compiled from archived reports needs to distinguish as-originally-reported values from subsequently reclassified comparisons. Observed changes in group premium or income series may otherwise be attributed to business contraction when the reporting population itself has changed.

The event does not establish whether a particular supervisor treated the retained interest as controlling for its own purposes. No such determination is claimed. The financial reporting standard in this case is US GAAP; the IFRS comparison demonstrates a parallel research problem but is not the governing accounting basis for the issuer's reported decision.

Acquiring control without acquiring all economic interests

Chubb presents a useful counterpoint to AIG's deconsolidation. Its 2024 accounting policy note (Chubb Limited, 2025) identifies July 1, 2023 as the date on which consolidation replaced equity-method treatment for Huatai Group. The reported aggregate shareholdings at December 31, 2023 and 2024 were 76.5% and 85.5%, while consolidation included Huatai's activity and financial position in full from the control date. The amount attributable to noncontrolling interests remains conceptually distinct from the recognition of the controlled entity's assets and liabilities.

The two successive year-end percentages differ by nine percentage points. This subsequent increase is separate from the earlier transition in accounting method. The event history must preserve the control date rather than assigning consolidation to the year-end on which a percentage was observed. The note also identifies primary-beneficiary treatment for relevant variable-interest entities under U.S. GAAP. Together with Corebridge, the observation demonstrates why the accounting perimeter is determined by the applicable control analysis rather than mechanically by the degree of economic participation. No conclusion about the current Huatai structure is drawn.

State-law control and regulatory presumptions

Under New York's Insurance Law Section 1501, control includes direct or indirect power over management and policies through voting interests, qualifying contractual means or other arrangements. A holding of at least ten percent of voting securities gives rise to the statutory presumption, subject to the provisions allowing determinations. Subsections (b) and (c) address circumstances in which the superintendent may find control or determine its absence.

The statutory presumption serves a different institutional purpose from an analyst's choice of majority threshold. A 12 percent voting stake can therefore be a relevant input to a regulatory inquiry even when it does not establish a majority or require consolidation by itself. Likewise an arrangement with a smaller shareholding can merit investigation when the facts indicate controlling influence. Application of the statute depends on the legal nexus and qualifications specified there.

Consider three hypothetical voting stakes: 9, 12 and 51 percent. A binary 50-percent screen selects one; a ten-percent presumption screen selects two. Neither screen adjudicates the underlying rights, exemptions or regulatory determinations. The sensitivity calculation is valuable because it shows that changing the screening definition changes the resulting map even with identical data. It must not be presented as a measured difference between actual supervisory and accounting perimeters.

Contractual coordination without shareholding

Zurich's 2024 annual-report presentation reports that Zurich Insurance Group holds no ownership stake in the Farmers Exchanges. Its wholly owned Farmers Group, Inc., together with specified subsidiaries, supplies services as attorney-in-fact and receives compensation. The reported arrangement establishes a material operational and economic relationship of a different kind from a subsidiary shareholding.

The analytical implication is precise. A map of shares alone omits a connection explicitly relevant to Zurich's business presentation, while an operational map can preserve it as a service relation. The relationship does not establish that Zurich owns the Exchanges, and the cited limited disclosure is insufficient to adjudicate all accounting or supervisory treatment. An assessment of particular liabilities would require the contracts, governing law, legal identity of the obligor and applicable reporting treatment.

This example also cautions against using revenue generated by management services as a substitute for premiums written by the organizations served. Fees and insurance underwriting flows are different economic categories. Even if both appear in a coordinated commercial presentation, their inclusion in an insurer premium market denominator must be justified under the chosen statistical definition.

Governance, agency and local accountability

Decision authority at an intermediate group company can influence underwriting appetite, investment allocation, technology outsourcing or capital strategy. The transmission of a decision to a regulated insurer depends on the powers held by that intermediate company and the legal duties constraining the insurer's own governing bodies. A group policy instruction is evidence of coordination, but enforceability and responsibility require examination of those distinct legal roles.

Agency conflicts can occur between controlling and minority owners, between group managers and an insurer's policyholders, or between operating entities and service providers whose incentives depend on fees. The ownership literature supplies a general explanation for divergence between authority and capital at risk. The insurance context adds the institutional problem of claims obligations and prudential safeguards. The present evidence does not estimate the magnitude of those conflicts or prove that a particular group exploited them.

Consolidation Perimeter

Financial presentation and its boundaries

Financial consolidation organizes statements as if the reporting parent and included subsidiaries constituted one economic reporting unit. The relevant intra-group balances and transactions are removed according to the applicable accounting requirements. This avoids counting an internal receivable and matching payable as independent external assets and obligations of the combined reporting unit.

The consequences of consolidation are visible in an elementary illustrative example. A holding company has an external asset of 100 and an internal receivable of 30 from its insurer subsidiary. The subsidiary has external assets of 150 and the corresponding internal payable of 30. Summing assets without reconciliation gives 280; eliminating the internal receivable yields 250 of aggregated external assets. The corresponding elimination of the internal payable affects aggregated liabilities. The example abstracts from equity investment elimination, valuation adjustments and minority interests; it isolates only one intercompany balance.

It follows that a parent-only balance sheet, the sum of solo-company statements and a consolidated balance sheet can produce different totals for the same organizational network. Such differences may arise from standard accounting operations rather than from reporting errors.

The IASB's IFRS 12 disclosure framework covers interests in subsidiaries and other arrangements including associates and unconsolidated structured entities. Such disclosures help an investigator trace exposures beyond consolidation. They complement the consolidation perimeter rather than providing a universal legal-ownership register.

A further comparison illustrates the role of the reporting root. According to its 2021 consolidation note (Swiss Reinsurance Company Ltd., 2022, p. 12), the mergers of Swiss Re Corporate Solutions Ltd. and Swiss Re Life Capital Ltd. into Swiss Reinsurance Company Ltd. occurred in the third quarter of 2021. The company retrospectively adjusted the comparative 2020 figures at the Swiss Reinsurance Company reporting level. The historical annual report documents a change in the nested legal and accounting structure. It does not imply that the ultimate Swiss Re group acquired those businesses from external parties in 2021. A graph indexed only by the ultimate parent could conceal a change that is material to a subordinate reporting entity; an analysis indexed only by the subordinate parent could mistake an internal reorganization for an external expansion. Both root and effective date are therefore essential research metadata.

Prudential scope and the supervisory head

ICP 23.2 of the IAIS framework identifies accounting consolidation as a practical starting point for determining group supervisory scope. The relevant group-wide supervision guidance also directs consideration of entities outside the accounts where risk or control makes them relevant and calls for recurring reassessment of exclusions and inclusions. Group supervisors coordinate scope with other involved authorities; the prudential head may differ from the level at which resolution powers apply.

A group perimeter can therefore be wider or differently constituted than the set reported in consolidated accounts. Information about a related entity can be relevant even where that entity is subject to another authority's direct supervision. The group-wide analysis examines the exposure of insurance operations, rather than assuming the insurance supervisor has identical direct legal powers over every related undertaking.

Consider a service company supplying claims systems to several insurers. If the service company is outside the financial consolidation, its operational importance may still warrant an assessment of disruption exposure. The finding of risk relevance does not turn the service company into an insurer or imply a standalone license. The legal source for each supervisory power must be separately established.

Similarly, a holding entity may impose financing pressures on subsidiaries through debt and distributions. Consolidation can remove the accounting record of an internal loan while leaving the underlying enforceability, priority, payment timing and liquidity dependence economically important for solo entities. A prudential map should preserve those obligations in its transaction layer.

The selected Solvency II provisions

The EU's Article 214 scope rules permit case-specific exclusions on specified grounds, including information-transfer impediments, negligible relevance, or circumstances in which inclusion would distort the supervisory objective. The provision explicitly addresses the collective significance of undertakings that appear negligible individually. A classification exercise therefore requires both individual and aggregate assessment.

Suppose ten related subsidiaries each represent an analyst's estimated 0.2 percent of a chosen exposure denominator. Their sum is 2 percent if the exposure components are non-overlapping. A rule that discards every company below 0.5 percent would exclude all ten, while a collective test could require further assessment of the combined exposure. The numbers and thresholds are illustrative; Article 214 does not impose these numerical cutoffs. The calculation shows why individually small nodes cannot be deleted solely by an automated filter.

The EU framework also separates the supervisory perimeter from the method of measuring group solvency. Article 220 choice of method establishes accounting consolidation-based method 1 as the default in the specified case, with supervisory discretion to use the alternative method or a combination following consultation where exclusive use of method 1 is inappropriate. Accordingly, inclusion in the group supervisory analysis does not imply that every undertaking's balance sheet receives a uniform numerical treatment.

Capital, fungibility and transferability

The presence of capital in a consolidated balance sheet says little by itself about its availability in a specific insurer experiencing losses. Availability depends on the location of the assets, rules governing eligible capital, liquidity, restrictions on distributions and the enforceability of support arrangements. Consolidation is not a cash-transfer instruction.

For a simple stress illustration, two subsidiaries possess excess resources of 40 and 10 relative to their own chosen requirements. A shock produces an additional need of 25 in the second entity. Consolidated excess of 50 before the shock does not imply that the second entity can obtain the missing 15 immediately. If only 5 can be transferred in time, the local gap remains 10 even though an arithmetic group total suggests ample resources. No real entity or regulatory formula is represented in the example.

The same distinction applies to guarantees and reinsurance. A support contract can change the distribution of losses and potential recovery without moving ownership or necessarily altering the financial consolidation boundary. A risk-transfer map must identify its counterparties, trigger, limits, collateral and legal enforceability as appropriate. Consolidated net exposure and local gross contractual obligations answer different questions.

Empirical research also establishes that internal transfers are a substantive economic relationship rather than a merely formal possibility. Powell, Sommer, and Eckles (2008) examined internal capital markets in insurer groups using recorded transfers and reported evidence consistent with capital allocation toward affiliates with stronger expected performance. The accessible abstract supports this limited description; it does not establish universal transferability, equality of prudential resources, or causal identification for the groups examined here. An internal capital market is therefore properly represented by observed flow and contractual evidence alongside, rather than inferred directly from, a shareholding graph.

The published literature also formalizes the distinction between diversification and legal-entity capital constraints. Asimit, Badescu, Haberman, and Kim (2016) studied efficient risk allocation under Solvency II using an optimization of proportional premium and liability transfers within a non-life group. Their examples vary recovery assumptions and evaluate the implications for required resources. The model uses the QIS5-era specification and assumptions about feasible proportional transfers and regulatory treatment. Its contribution is conditional: modeled benefits depend on the transfer contract and parameters, not just common ownership. It is neither a test of the actual AIG or Chubb networks nor a statement of presently operative capital formulas. The appropriate use of the study here is to separate the observed existence of ownership edges from the assumptions required to estimate the benefits of risk sharing.

Observed capital movements and identification limits

Niehaus (2018) examined internal capital transactions of life insurers using annual statutory statements for 2006–2011. The study's baseline panel regressions in Table 5 used 3,330 observations from 712 companies, excluded inactive or financially invalid observations, normalized the monetary variables by total assets, and clustered standard errors by group. The baseline estimates associate a one-dollar decline in negative performance with 26 cents in additional internal capital contributions, and a one-dollar increase in positive performance with 56 cents in additional internal dividends paid. These are conditional associations from a defined historical population rather than coefficients applicable to every insurer.

The same study identifies a critical disclosure limitation in its data-description section: an affiliate's net annual receipts and payments do not identify the complete bilateral flow matrix. For example, observed net payments of 70 and 30 by two companies, accompanied by a net receipt of 100 by a third, permit both direct payments to the recipient and an alternative in which one paying company first sends its funds through the other. The net values are identical, yet the intermediate funding dependencies differ. These numbers are an illustrative inverse problem, not the study's empirical observations. A capital-flow graph requires transaction-level counterparty evidence, or explicitly identified bounds on plausible paths.

The distinction clarifies why ownership, accounting control, and financing exposure should be reconstructed in separate relation layers. Empirical evidence of internal support establishes that a transfer mechanism operates under particular conditions; it does not establish an enforceable commitment to finance any specific subsidiary in distress. Conversely, accounting elimination of intragroup transfers does not remove their legal or prudential importance at individual-entity level.

Perimeter reconciliation and measurement sensitivity

A comparison of four relevant boundaries shows where contradictions can appear and how to resolve them.

Observed dimension Evidence used Valid inference Further evidence required
AIG 2024 immediate voting links Exhibit 21 and notes Specific reported hierarchy and voting shares Complete ownership census, license or share-class rights
Corebridge June 2024 AIG Form 8-K and quarterly report extract Issuer-reported loss of accounting control and retained stake Any specific regulatory determination
Farmers Exchanges 2024 Zurich annual-report note Service relationship without Zurich equity ownership Exact contractual liabilities and applicable consolidation treatment
IAIS group scope ICP 23 Relevant external risks can be considered Actual supervisor decision for a named entity
EU group solvency Articles 214 and 220 Defined exclusions and choice of calculation method Individual jurisdictional application and current decision

A reproducible reconciliation records a difference as either temporal, semantic, evidentiary or unresolved. A temporal difference reflects a transaction or reporting-date change. A semantic difference reflects different meanings of control or membership. An evidentiary difference reflects distinct disclosure coverage. An unresolved discrepancy occurs when equivalent claims conflict and the records do not permit resolution. This classification is an explanatory convenience drawn from the analysis, not a new formal normative system.

Sensitivity to missing data is especially important. The eight-edge AIG extraction is a transparent lower-level observation, whereas the full corporate group includes additional disclosed and omitted links. The five-edge path lengths in the sample are reproducible, but a group-wide maximum cannot be calculated from the selected subset alone. In a global database, stronger disclosure practices can make one group appear more complex than another even if the underlying legal structures were similar.

Integrated Scholarly Synthesis and Research Frontiers

Implications for insurance-market measurement

A market can be counted by authorized undertakings, economic groups, underwriting branches, distribution networks or contract volume. Each measure is coherent when the statistical population is defined. Combining these units in one denominator produces artificial differences across countries and reporting systems.

For example, a country with two subsidiaries of an international group might report two locally licensed undertakings. A comparative group-based market concentration analysis may combine their eligible local premium volumes under one parent. That operation requires compatible premium definitions and proper treatment of branches, internal reinsurance and market allocation. It also requires a decision about which form of control warrants aggregation. Aggregating every branded or commercially connected business would conflate ownership and service affiliations.

If two insurers hold 20 and 15 percent of the same defined market, and both are verified members of one economic unit for the competition analysis, replacing the two individual shares with one 35 percent share changes their contribution to the Herfindahl-Hirschman sum from 625 to 1,225 percentage-point-squared units. The difference is 600, equal to twice 20 times 15. This is a hypothetical calculation holding all other market shares fixed. It demonstrates the importance of the grouping choice without asserting a concentration level for AIG, Zurich or any jurisdiction.

Group membership may also change without a corresponding change in local contracts. A merger, disposal or deconsolidation can move premiums between statistical group categories even when customer-facing activity continues. Time-series analysis needs transaction-effective dates, a consistent legal population and explicit historical restatement rules.

Risk transmission and supervisory interpretation

Several channels connect legal structure to risk. Ownership can transmit changes in investment value, financing decisions and governance authority. Contracts can transmit losses and cash-flow obligations. Shared operational arrangements can transmit service disruptions. A common brand can create reputational spillovers. The magnitude of each effect depends on distinct information that a map can organize but cannot manufacture.

Consider a hypothetical disturbance at a parent holding company. Its consequences for an insurer may depend on dividend restrictions, existing financial commitments, cross-default terms, and whether external counterparties reassess the insurer because of the parent's problems. A mere path of shareholding edges establishes an institutional connection, not a numerical probability of default. Similarly, common ownership does not reveal the severity of correlated underwriting losses without exposure data.

A multi-layer record is useful because it retains these different transmission pathways side by side. An insurer can have a highly centralized operational dependency with a short service path while being several steps removed along the equity ownership path. The most relevant path depends upon the risk event being studied.

The corporate examples support three narrower propositions. First, disclosed ownership is reconstructed most accurately from the immediate-parent hierarchy and relevant footnotes, rather than from entity-name resemblance. Second, governance changes are capable of moving reporting boundaries while material investment exposure continues, as the Corebridge filing illustrates. Third, economically significant contractual coordination can exist without shareholding, as the Zurich disclosure illustrates. These propositions are sufficient to reject a universal equation of operational affiliation, equity ownership and accounting consolidation for the selected evidence. They do not quantify how often those differences arise.

Reproducibility, revision and data governance

A durable empirical corpus should store snapshots rather than continuously replacing the latest state. A corporate registry name change, a transfer of ownership and a change of consolidating parent are different events. Recording the source date and the legal or accounting effective date enables a later researcher to reconstruct what could have been known at each point. Where the effective date is absent, a report date provides an observation boundary rather than exact proof of transaction timing.

An entity table can hold identifiers, jurisdictions and roles. A relationship table can link two identifiers and retain type, direction, effective interval, quantity, unit, evidence and confidence. An exceptions table can preserve missing parent data and unresolved conflicts. A perimeter table can associate an entity with a particular group, standard, authority and reporting date, including the stated reason for inclusion or exclusion. These are ordinary data-management structures suitable for reproducing the analysis; they do not constitute a proposed new scientific theory.

Confidence should follow the proposition actually checked. A filing can be a strong source for what the issuer reported while remaining insufficient for independently proving an unreported legal agreement. An accounting-parent identifier can be trustworthy within its definition yet unsuitable for measuring beneficial ownership. Data quality therefore has both identification accuracy and semantic suitability.

In future expanded research, samples should be stratified by group-head jurisdiction, legal form, insurance specialization and disclosure regime. A register of large international groups could provide one frame; smaller cross-border insurers would need additional national registers. Researchers should publish inclusion rules, nonresponse and matching methods. Comparisons across countries require an examination of missingness and reporting thresholds before numerical estimates or statistical inference are attempted.

Institutional incentives and unresolved empirical questions

The present observations establish mechanisms, but they do not settle the motives behind international structure. Whether an intermediate holding lowers the cost of capital, improves regional oversight, enables risk transfer, facilitates acquisition, or reflects regulatory and tax differences is a separate causal question. A credible investigation would compare changes around documented corporate transactions against plausible alternatives, using sufficiently harmonized information on capital, costs, operating conditions and legal requirements.

Further research could also investigate the relation between legal network depth and the speed of supervisory coordination during a stress event. Simple path length may be a poor predictor: robust information-sharing arrangements can reduce delays across a complex group, while a seemingly short route can be obstructed by confidentiality or legal barriers. Process evidence, authority powers and data quality must supplement network geometry.

The influence of corporate control on policyholder outcomes is similarly conditional. A common parent may enable coordination and access to resources, or it may transmit financial and operational pressure. Neither effect follows mechanically from the number of subsidiaries. The relevant empirical investigation would identify specific contractual arrangements, safeguards, claims outcomes and counterfactual conditions.

The broader research frontier is therefore the joint reconstruction of ownership, governance, capital, risk transfer and institutional responsibilities with evidence adequate to the chosen outcome. It is possible to strengthen each of these dimensions without collapsing them into one universal group label.

Conclusion

International insurance groups are best understood through dated, purpose-specific reconstructions of their legal entities and relationships. The research has shown that immediate ownership links, decision rights, financial consolidation, prudential scope and operational dependence describe different aspects of a shared institutional environment.

The selected AIG hierarchy establishes the importance of preserving intermediate holding companies, identifying the denominator of reported percentages and reading ownership footnotes. Its eight-edge subgraph demonstrates that geographically distributed control chains can be reconstructed reproducibly while preserving a strict boundary between the observed sample and the complete group. The Brazilian voting-interest note shows how a second holding path can resolve an apparent residual percentage.

AIG's dated Corebridge disclosure establishes why a material economic interest can survive a reported change in accounting control. Zurich's reporting establishes why a commercially relevant service relationship can be absent from an ownership hierarchy. Chubb's Huatai observation provides a converse control transition, and the Swiss Re evidence establishes the importance of identifying the reporting root in an internal reorganization. Together with the legal and accounting comparison, these cases show that apparently different group boundaries often reflect precisely the institutional purpose for which each boundary was established.

The comparison of the IAIS framework, selected Solvency II provisions and New York law demonstrates that supervisory control and scope require attention to the operative rule, jurisdiction, date and factual arrangements. The prudential perimeter addresses group risk and does not imply uniform direct supervisory powers or immediately available group capital. Financial consolidation is an essential source of evidence, alongside entity-specific filings and documented regulatory decisions.

The original analytical contribution consists of the explicit reconstruction of selected ownership paths, transparent sensitivity calculations, case-based comparison of distinct institutional boundaries and integration of the resulting implications for statistical measurement and risk analysis. No new named doctrine or empirical population estimate is asserted. The scientific result is a grounded method of interpretation: a relationship can be used for inference only where its evidence, date, legal meaning and analytical purpose agree.

References

  • International Association of Insurance Supervisors (IAIS). (2024). Insurance Core Principles and Common Framework for the Supervision of Internationally Active Insurance Groups, December 2024 revision. ICP 23, particularly 23.2 and CF 23.2.a; introduction, paragraphs 17–25. ICP 23, particularly 23.2 and CF 23.2.a; introduction, paragraphs 17–25.

  • International Association of Insurance Supervisors (IAIS). (2026). Register of Internationally Active Insurance Groups (IAIGs). Public register, viewed June 3, 2026 snapshot as returned during research. Public register, viewed June 3, 2026 snapshot as returned during research.

  • International Accounting Standards Board (IASB). (2011, as amended). IFRS 10 Consolidated Financial Statements. IFRS Foundation, official standard overview and history. IFRS Foundation, official standard overview and history.

  • International Accounting Standards Board (IASB). (2011, as amended). IFRS 12 Disclosure of Interests in Other Entities. IFRS Foundation, official standard overview. IFRS Foundation, official standard overview.

  • IFRS Interpretations Committee. (2026, June). IFRIC Update June 2026: Control Assessment for a Single-investor Fund (IFRS 10). IFRS Foundation. IFRS Foundation.

  • European Parliament and Council of the European Union. (2009, as amended). Directive 2009/138/EC (Solvency II), Article 214, Scope of group supervision. EIOPA Single Rulebook, version displayed May 3, 2021. EIOPA Single Rulebook, version displayed May 3, 2021.

  • European Parliament and Council of the European Union. (2009, as amended). Directive 2009/138/EC (Solvency II), Article 220, Choice of method. EIOPA Single Rulebook, version displayed May 3, 2021. EIOPA Single Rulebook, version displayed May 3, 2021.

  • European Insurance and Occupational Pensions Authority (EIOPA). (n.d.). Solvency II. Regulatory framework overview. Regulatory framework overview.

  • New York State Legislature. (2023 revision displayed). New York Insurance Law, Article 15, Section 1501: Definitions; determinations. New York State Senate Open Legislation. New York State Senate Open Legislation.

  • Global Legal Entity Identifier Foundation (GLEIF). (n.d.). ROC Policy on Level 2 Data. Summary of Regulatory Oversight Committee policy of March 10, 2016 and fund policy of May 2019. Summary of Regulatory Oversight Committee policy of March 10, 2016 and fund poli.

  • Global Legal Entity Identifier Foundation (GLEIF). (2021). Level 2 Data: Reporting Exceptions Format 2.1. Official documentation, September 2021 revision. Official documentation, September 2021 revision.

  • American International Group, Inc. (2025). Subsidiaries of Registrant as of December 31, 2024. Exhibit 21 to 2024 Form 10-K. U.S. Securities and Exchange Commission, EDGAR, accession 0000005272-25-000012. Securities and Exchange Commission, EDGAR, accession 0000005272-25-000012.

  • American International Group, Inc. (2024, June 24). Current Report on Form 8-K, Item 8.01. U.S. Securities and Exchange Commission, EDGAR, accession 0000005272-24-000068. Securities and Exchange Commission, EDGAR, accession 0000005272-24-000068.

  • American International Group, Inc. (2024). Quarterly Report on Form 10-Q for the quarter ended June 30, 2024. U.S. Securities and Exchange Commission, EDGAR, accession 0000005272-24-000079. Securities and Exchange Commission, EDGAR, accession 0000005272-24-000079.

  • Zurich Insurance Group. (2025). Annual Report 2024, online report and note accompanying Farmers performance presentation. Annual Report 2024, online report and note accompanying Farmers performance pres.

  • La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. (1998). Corporate Ownership Around the World. NBER Working Paper No. 6625, June 1998. DOI: 10.3386/w6625. Working-paper version. Working-paper version.

  • Australian Prudential Regulation Authority (APRA). (2026). List of registered general insurers. Updated September 23, 2026. Source.

  • Monetary Authority of Singapore (MAS). (2026). Financial Institutions Directory: AIG Asia Pacific Insurance Pte. Ltd. Updated April 22, 2026. Source.

  • Commissariat aux Assurances (CAA). (n.d.). Luxembourg non-life insurers. Retrieved November 19, 2026. Source.

  • Powell, L. S., Sommer, D. W., & Eckles, D. L. (2008). The Role of Internal Capital Markets in Financial Intermediaries: Evidence From Insurer Groups. Journal of Risk and Insurance, 75(2), 439–461. DOI: 10.1111/j.1539-6975.2008.00267.x. Source.

  • Kartasheva, A. (2021). Structure and Complexity of Global Insurance Groups. LTI Working Papers No. 10, November 19, 2021. Source.

  • Chubb Limited. (2026). Form 10-K for 2025, Exhibit 21.1: Subsidiaries as of December 31, 2025. U.S. Securities and Exchange Commission, accession 0000896159-26-000005. Source.

  • Chubb Limited. (2025). Form 10-K for 2024, Note 1: General and Summary of Significant Accounting Policies. U.S. Securities and Exchange Commission, accession 0000896159-25-000004. Source.

  • Swiss Re International SE. (2022). Solvency and Financial Condition Report 2021, Section A1, pp. 5–6. Source.

  • Swiss Reinsurance Company Ltd. (2022). Consolidated Annual Report 2021, Note 1, p. 12. Source.

  • Asimit, A. V., Badescu, A. M., Haberman, S., & Kim, E.-S. (2016). Efficient risk allocation within a non-life insurance group under Solvency II Regime. Insurance: Mathematics and Economics, 66, 69–76. DOI: 10.1016/j.insmatheco.2015.10.008. Source.

  • Niehaus, G. (2018). Managing Capital via Internal Capital Market Transactions: The Case of Life Insurers. Journal of Risk and Insurance, 85(1), 69–106. DOI: 10.1111/jori.12143. Source.