Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Insurance Group Risk Concentration and Exposure Mapping

Abstract. Insurance groups accumulate risk across legal entities, underwriting portfolios, investments, reinsurers, territories, business services, and contingent financial obligations. An exposure register organized by individual subsidiary or accounting line can therefore misstate the magnitude, location, and transmission of common shocks. This monograph examines how insurance groups construct decision-useful maps of gross and net exposures, detect shared counterparties and perils, trace intragroup transmission, and incorporate those maps into limits, solvency assessment, and supervisory reporting. It combines a close reading of international insurance-supervisory standards, European group-concentration provisions, U.S. group capital and ORSA guidance, catastrophe modeling standards, and scholarly work on dependence and systemic risk with independent algebraic analysis. A fully disclosed hypothetical group illustrates why intragroup reinsurance must be eliminated at consolidation, why external recoveries remain exposed to counterparty failure, and why legal-entity capital cannot be assumed perfectly transferable. The principal finding is that effective exposure mapping requires simultaneous preservation of transaction-level provenance, economic shock identity, legal ownership, and a consistent consolidation perspective. These linked representations support meaningful scenario analysis and allow supervisors and boards to identify concentrations that disappear in ordinary classifications. Their limits arise from incomplete location data, policy interpretation, dependence estimation, financial contagion, contractual recoverability, and model uncertainty. The work offers a source-grounded analytical synthesis and practical verification tests without proposing a new proprietary theory, regulatory standard, or empirical estimate of an actual insurer.

Keywords: Insurance group; risk concentration; exposure mapping; group supervision; enterprise risk management; accumulation risk; catastrophe exposure; reinsurance counterparty; intragroup transactions; solvency; wrong-way risk; risk aggregation; legal-entity capital; data lineage; stress testing.

Introduction

An insurance group is exposed to the same underlying event through contracts that appear in different administrative systems. A coastal windstorm can create direct indemnity claims in two subsidiaries, assumed reinsurance losses in another, an operational interruption at a shared service center, a decline in investments related to the damaged region, and delayed recoveries from a financially strained retrocessionaire. Each component can be recorded correctly on its own terms while the aggregate picture remains incomplete. The research problem is therefore one of identification, measurement, consolidation, and decision rights across overlapping risk populations.

The IAIS enterprise risk management framework requires integrated consideration of group risks and relationships among risks, while the European Union's Solvency II group provisions explicitly organize supervisory attention around concentrations and intragroup transactions. These instruments establish an appropriate starting point but leave the methodological work of identifying common exposures to the institutions collecting, interpreting, and validating the data. The United States approaches group visibility through a combination of legal-entity supervision, group capital calculation, enterprise-risk reporting, and ORSA; the NAIC group capital explanation expressly situates capital resources and risks inside the group structure.

This study asks three questions. First, what is the object and unit of a group concentration map, and what must remain identifiable when records are joined? Second, how can common counterparties, locations, perils, and contractual protections be aggregated without artificial double counting or unsupported diversification? Third, what does the map reveal about intragroup propagation, legal restrictions, solvency, liquidity, governance, and reporting?

The method is comparative documentary analysis of primary supervisory and professional materials, with targeted scholarly literature and independently derived accounting identities and stress calculations. The legal and supervisory examples are jurisdiction-specific. The mathematics and the portfolio example are illustrative constructions, not estimates from confidential insurer records. No causal performance effect is inferred from a rule, market statistic, or numerical illustration. The scope includes risk identification and information design; detailed capital calibration, reinsurance pricing, and actuarial reserve estimation belong to the associated Science and their separate Research Dossiers.

Definitions, History, and Documentary Foundations

The economic object of concentration

An exposure is a specified financial or operational sensitivity to a defined event, obligation, asset, counterparty, location, or risk driver, measured at a stated date and horizon. Its representation requires a measurement basis: insured value, contractual limit, expected loss, stressed claim, asset market value, net receivable, liquidity demand, or capital requirement. These quantities answer different questions and must retain different names. A limit is a contractual ceiling; a sum insured is an exposure descriptor; a modeled loss is conditional on hazard and vulnerability; a receivable is a claim against a particular obligor; regulatory capital is a modeled or prescribed measure of required financial resources. Conflating them obscures risk.

A concentration exists when several exposures are sufficiently exposed to a common adverse cause, counterparty, physical footprint, economic factor, network dependency, or restriction on resources that aggregation changes the risk of material loss or impaired performance. This definition includes simple single-name concentrations and conditional concentrations in which exposures coalesce only under stress. For insurance groups, concentration can arise on both sides of the balance sheet and in operational systems. The IAIS Insurance Core Principles and ComFrame address ongoing group risk identification, aggregate counterparty exposures, dependencies, and the capacity to maintain diversification under stress (IAIS, 2024).

Risk concentration is conceptually distinct from a large gross amount. A geographically dispersed set of independent risks can have a greater gross insured value and lower tail sensitivity to a single event than a smaller, clustered portfolio. Conversely, a modest concentration in a provider, clearing service, data center, retrocessionaire, or group treasury operation can generate severe loss transmission if many entities depend on its survival. The relevant comparison is therefore the joint loss under specified conditions, the capacity to absorb it, and the ability to move funds where obligations fall due.

Development of group supervisory attention

Insurance supervision historically emphasized individual licensed undertakings and their obligations to policyholders. Corporate groups, cross-border business, assumed reinsurance, investment affiliates, and non-insurance parents add a second analytical perimeter. Recognition of this perimeter does not eliminate local entity obligations: claims remain due from the contracting insurer, and capital availability depends on legal, tax, supervisory, contractual, and timing constraints. The evolution is reflected in the IAIS group-wide supervision standards and the adoption of a global Insurance Capital Standard for internationally active insurance groups in 2024 (IAIS, 2024). The ICS provides a global supervisory measure for its defined entities; it is not a universal business-operating exposure map.

Under Article 244 of Solvency II, significant group risk concentrations are subject to reporting and review, with thresholds and types informed by supervisory judgment and the group structure. Article 245 on intragroup transactions separately addresses transactions that can alter intercompany exposures and financial transmission. The 2015/35 delegated regulation, particularly Articles 376 and 377, specifies relevant direct and indirect exposures, including interconnected counterparties, geography, industries, and catastrophes, alongside loans, guarantees, derivatives, reinsurance, service arrangements, and other intragroup transactions (European Commission, 2015). These are distinct legal categories, not interchangeable formulas.

The EU implementation has a concrete reporting vocabulary. Implementing Regulation 2023/894 specifies supervisory templates including S.37 risk concentration reporting (European Commission, 2023). EIOPA's answer on insurance-policy exposures explains the use of contractual maximum possible payment or treaty capacity in a specified reporting field. Its answer on mitigating techniques clarifies that a mitigation amount attached to one counterparty cannot simply be summed with gross exposures toward another. These answers demonstrate the importance of retaining the counterparty and instrument meaning of every number.

The U.S. structure is different. The NAIC's Group Capital Calculation overview treats group capital as additional supervisory information alongside organizational, enterprise-risk, and ORSA filings; it describes the scope of the group and location of capital. The 2025 ORSA Guidance Manual discusses risk aggregation, double gearing, fungibility, contagion, concentration, liquidity, and the prospective solvency assessment (NAIC, 2025). NAIC model documents and manuals do not automatically impose identical duties in every state: enacted domestic law, filing eligibility, implementation, and supervisory directions control.

Evidence hierarchy and comparative meaning

This monograph uses four kinds of evidence: binding legal texts and adopted supervisory or professional standards; official implementation guidance and public regulatory explanations; research publications that describe mathematical or empirical relationships; and constructed examples whose assumptions are fully visible. Each has a distinct evidentiary capacity. A legal provision establishes a reporting requirement within its scope, not a measured reduction in bankruptcies. An official questionnaire or dashboard measures the observed reporting population under its definitions, not every group worldwide. A mathematical identity shows a relation under stated assumptions, not the suitability of those assumptions for a specific portfolio. An insurer's confidential exposure cannot be established by quoting an industry-wide number.

Contemporary supervisory evidence illustrates why concentration mapping remains active. In its June 2026 financial stability assessment, EIOPA points to sovereign exposures, banking links, private markets, valuations, and geopolitical risks while reporting general sector resilience (EIOPA, 2026). The IAIS Global Insurance Market Report similarly describes emerging supervisory attention to private credit, climate and cyber exposures (IAIS, 2025). These sector observations motivate targeted checks; they do not establish that a particular insurer has a material concentration.

Exposure Mapping Units

The minimum identifiable record

An enterprise map begins with a documented record grain. The most useful primitive is a controlled exposure observation associated with a source transaction or policy and an as-of date. Its minimum fields are: source-system key; legal-entity obligor or owner; contract and coverage; insured object or investment instrument; gross measurement and currency; attachment, deductible, limit and share; relevant peril or financial driver; precise location or multi-location schedule; external and internal counterparty identifiers; effective period and valuation date; source data version; and a trace to assumptions and transformations.

A group may store the information in several linked records rather than one enormous table. Policy, insured object, coverage, catastrophe location, asset, issuer, reinsurer, contract layer, claim, and legal entity are different objects. A location may support many policies; one policy may insure many locations; a treaty can protect multiple portfolios; an affiliate may issue both assets and guarantees. Flattening these relationships into one row risks duplication. A relational key or controlled graph preserves object identity through joins.

The proper unit also depends on the question. For a natural catastrophe, an individual address, building, facility, or segment of infrastructure may be necessary. For liability, the common unit may be a manufacturer, legal theory, product batch, event cohort, or jurisdiction. For cyber, an outsourced platform, software dependency, cloud region, or shared credential provider may be more informative than policyholder headquarters. For sovereign assets, the ultimate issuer, country of risk, currency, and instrument seniority all matter. An enterprise map therefore maintains multiple views of the same ultimate economic risk rather than choosing one permanent geographic or organizational grid.

A reproducible consolidation equation

Let ee index legal entities and jj index identifiable underlying exposure objects. Write xejx_{ej} for the amount attributable to entity ee and object jj on a defined gross basis. If each xejx_{ej} is a distinct external liability or asset slice and ownership overlap is resolved, then the group gross measure in a segment GG is

XGgross=∑e∑jxej1{j∈G}.X_G^{gross}=\sum_e\sum_j x_{ej}1\{j\in G\}.

The qualification about distinct slices is essential. A single assumed risk can appear once in the underlying primary policy and again in an intragroup reinsurance record. Summing those entries without eliminating internal transfer would create a false group-level exposure. By contrast, the risk must appear in each relevant entity view to show legal responsibility.

For a specified event ss, let Ledirect(s)L_e^{direct}(s) be externally originated loss, Cef(s)C_{ef}(s) the payment by affiliate ff to entity ee under an intragroup cover, and Reexternal(s)R_e^{external}(s) a collectible payment from an external risk-transfer counterparty. Ignoring transaction costs and other external exposures, the total consolidated net loss is

Lgroupnet(s)=∑eLedirect(s)−∑eReexternal(s).L_{group}^{net}(s)=\sum_e L_e^{direct}(s)-\sum_e R_e^{external}(s).

Internal payments Cef(s)C_{ef}(s) cancel when all group entities are consolidated, because one subsidiary's recovery is another subsidiary's obligation. This identity is a bookkeeping relation for the event loss, not a statement about each subsidiary's ability to fund its own cash calls. Reinsurance recoveries are contingent and their collectibility must be stressed separately.

For derivatives, investments, guarantees, and operational disruptions, the group adds different loss functions under the same scenario. Denote their consolidated losses A(s)A(s), O(s)O(s), and other expenses K(s)K(s). Total scenario loss is Lgroupnet(s)+A(s)+O(s)+K(s)L_{group}^{net}(s)+A(s)+O(s)+K(s), subject to preventing overlap among the categories. This expression is the basis for a reconciliation, not a universal risk-capital formula.

Data quality as an exposure variable

Exposure concentration is sensitive to missing addresses, aggregate location codes, outdated geocoding, ambiguous named insureds, inaccurate policy shares, mixed currencies, effective-date errors, and incomplete assumptions of historical portfolios. Consider an underwriting dataset reporting 10% of location values only at the postal-code level. If the missing granularity is concentrated in a floodplain or an urban industrial cluster, imputing a geographic midpoint can understate tail accumulation even when the aggregate insured value is correct. A data-quality report should therefore present both the known exposure and an uncertainty envelope under credible alternative spatial assignments.

Joining multiple insurers in a group requires a versioned organization hierarchy and a controlled ultimate-parent identifier. A name match between two corporate entities is weaker evidence than a verified legal identifier, ownership record, or direct link established by transaction documentation. Contract types must be standardized before aggregation: gross sum insured, net retention, treaty maximum, and recoverable asset are not comparable merely because the currency matches.

The BCBS risk-data aggregation principles were developed for banks, especially globally systemically important banks (BCBS, 2013). Their principles of accurate, complete, adaptable, and timely risk data provide a useful comparative information-governance reference, but they do not directly create insurance regulatory obligations. The insurance-specific basis remains the IAIS and applicable national rules.

Maps as controlled projections

A geographical heat map is a projection from the complete exposure population to spatial cells. Its numerical interpretation changes with cell resolution, policy allocation, peril footprint, treatment of multi-site contracts, and whether the quantity represents value, probable claim, or stress capital. A map titled simply "risk" is therefore insufficiently specified. Every display should state its valuation basis, as-of date, grouping criterion, treatment of insurance and reinsurance, currency, scenario assumptions, and exposure population.

A tabular single-name map, an asset-industry network, and a hazard-footprint map each preserve different information. The most reliable group exposure architecture makes these views reproducible from the same versioned records and exposes the joins required to move from one view to another.

Shared Counterparties and Perils

Direct, indirect, and contingent commonality

A group can hold bonds issued by one financial institution, rely on that institution as a derivative counterparty, and expect reinsurance recoveries from a corporate affiliate of the same institution. An entity-by-entity report may show three ordinary lines; a group counterparty map identifies a potential common loss source. The relevant network must resolve direct legal counterparties, their ultimate ownership and guarantees, netting agreements, collateral arrangements, financial-sector exposures, and where contractually relevant, a common funding source.

A one-name concentration ratio can be written qk=Ek/∑jEjq_k=E_k/\sum_j E_j for comparable positive exposures EjE_j. The familiar Herfindahl index H=∑kqk2H=\sum_k q_k^2 summarizes dispersion of that particular measurement base. If allocations of a comparable outward counterparty exposure are 55%, 30%, and 15%, H=0.552+0.302+0.152=0.415H=0.55^2+0.30^2+0.15^2=0.415. This is an illustrative calculation: it expresses concentration of nominal allocation. It neither measures default probability nor establishes the covariance, contractual collectibility, collateral, or recovery timing of those obligations.

A common counterparty can induce wrong-way risk. For example, a reinsurer exposed to the same storm losses may be under greater financial strain precisely when the primary insurer's recovery becomes due. Such association must be represented as a scenario or dependence assumption and supported by counterparty-specific evidence; it cannot be inferred from a group name alone. Similarly, an investment manager and custodian may share a parent but different legally segregated duties. Mapping must distinguish economic affiliation from actual recourse.

For EIOPA S.37 reporting, the 2023 counterparty clarification addresses significant insurance-policy exposures and treaty maximum payments, while its mitigation field clarification prevents adding the same derivative both as a risk-bearing instrument and as an undifferentiated reduction in another counterparty's amount (EIOPA, 2023). A well-structured map therefore records the raw exposure, the instrument that transfers it, the new counterparty created by that transfer, and the conditions under which the transfer is effective.

Common perils and the topology of catastrophe accumulation

Location-sensitive risk has an event footprint, while insurance exposure has a geographic, contractual, and temporal footprint. A windstorm, earthquake, wildfire, flood, or severe convective storm event may cross administrative boundaries and occur simultaneously across many insured objects. Mapping only by country or by one headquarters address can conceal this event-driven dependence.

The 2026 Lloyd's Realistic Disaster Scenarios require market participants to consider defined severe events, including consecutive windstorms, at syndicate and market levels (Lloyd's, 2026). Such scenarios demonstrate an operational distinction: a location map records which contracts are present, whereas an event map specifies the common shock and the loss function applied to each contract. An event footprint may affect property, marine, business interruption, liability, and investments through different pathways.

The Swiss Re sigma analysis for 2025 reports USD 107 billion in global natural catastrophe insured losses for 2025 and attributes a substantial majority to secondary perils, notably wildfires, severe storms, and floods (Swiss Re Institute, 2026). These are industry estimates, not measurements of any particular group's exposure. The scientific implication is methodological: concentration screens designed only around historic peak-peril zones can miss accumulated exposures to frequent, spatially distributed hazards and correlated repair-cost inflation.

The chosen hazard representation must be consistent with current scientific understanding. Location precision matters but is insufficient by itself; vulnerability curves, deductibles, limits, demand surge, conditional business-interruption loss, secondary uncertainty, and unmodeled perils influence the conversion of physical hazard into insured loss. The adopted ASOP No. 38 catastrophe modeling standard calls for evaluating the model's appropriateness, inputs, outputs, validation and limitations (Actuarial Standards Board, 2021). The adopted ASOP No. 56 on modeling adds general model-use and documentation expectations for actuaries operating within its scope (Actuarial Standards Board, 2019).

Dependence is not a label

For losses L1,…,LnL_1,\ldots,L_n, variance aggregation under finite second moments gives

Var(∑iLi)=∑iσi2+2∑i<jρijσiσj.\mathrm{Var}\left(\sum_iL_i\right) =\sum_i\sigma_i^2+2\sum_{i<j}\rho_{ij}\sigma_i\sigma_j.

This identity describes second moments. It does not fully characterize joint tail losses, risk-transfer triggers, or the effect of shared claims inflation. Suppose two modeled exposures have standard deviations 70 and 50, measured in the same monetary unit. At correlation 0.25 their aggregate standard deviation is 702+502+2(0.25)(70)(50)=9150≈95.66\sqrt{70^2+50^2+2(0.25)(70)(50)}=\sqrt{9150}\approx95.66. At correlation 0.85 it becomes 13350≈115.54\sqrt{13350}\approx115.54. The change is entirely due to the assumed dependence; the individual marginal standard deviations are unchanged. These values are calculated examples, not observed market correlations.

The methodological warnings in Embrechts, McNeil, and Straumann (2002) concern the limitations of ordinary correlation as a dependence descriptor. The revised Quantitative Risk Management develops copulas, tail dependence, aggregate risk and model uncertainty more generally (McNeil, Frey, and Embrechts, 2015). A dependence assumption must reflect peril mechanisms and relevant time scales, and model validation must examine the distribution of aggregate outcomes rather than accept a single correlation matrix as proof of diversification.

A practical concentration screen can use several complementary outputs: exposure at risk in the same geographic footprint; aggregate loss under defined scenarios; loss conditional on counterparty failure; selected risk measures over a consistent horizon; contribution to a specified group capital metric; liquidity demand by settlement date; and sensitivity to changing contractual or data assumptions. Those outputs should be labeled and reconciled, not collapsed into a single unqualified score.

Intragroup Transmission

Legal entities and consolidation boundaries

A multinational insurer can operate through regulated subsidiaries, branches, holding companies, captives, internal reinsurers, service companies, and non-insurance businesses. Each may have distinct accounting and supervisory responsibilities. Consolidation should remove internal duplication for the group's overall economic exposure while retaining every internal claim and obligation for legal-entity solvency and liquidity analysis.

The IAIS group-wide ORSA requirements place particular attention on interconnectedness, group-level risk aggregation, capital fungibility, internal transfers, and stresses that can threaten a subsidiary or the group (IAIS, 2024). The 2025 NAIC ORSA guidance likewise addresses intragroup double gearing and restrictions on transferable resources. These considerations make legal boundaries an essential part of the map even after consolidated accounting eliminates internal transactions.

Independently constructed storm and retrocession example

Consider a fictional insurance group with two property insurers A and B, a wholly owned internal reinsurer R, and a holding company H. A single severe storm generates gross contract claims of 120 monetary units in A and 90 in B. The example assumes all amounts are final, simultaneous, and in the same currency; it abstracts from discounting, taxes, reinstatement premiums, claim adjustment costs, and basis risk.

A recovers 60 from R through intragroup reinsurance; B recovers 30 from R. R therefore owes 90 to its affiliates, then receives 50 under an external retrocession contract if that counterparty performs completely and on time. The entity-level event losses after contractual recoveries are A = 60, B = 60, and R = 40. The consolidated group event loss equals 160. The full reconciliation follows.

Exposure layer A B R Consolidated group
Externally originated storm claims 120 90 0 210
Intragroup claims paid or assumed -60 -30 +90 0
External retrocession payment 0 0 -50 -50
Net event loss, full external payment 60 60 40 160
Net event loss, external payment fails 60 60 90 210

The consolidated identity is 210−50=160210-50=160 with performing external retrocession. The internal 90 is a real balance among subsidiaries but cancels on consolidation. If the external retrocessionaire fails completely and there is no effective collateral or substitute recovery, the group loss rises to 210. If recovery is partial, the increase is the unpaid portion of the 50 due under the stress assumptions. These calculations distinguish contractual transfer from collectible transfer.

For each group entity, the payment sequence can still be destabilizing. R may owe its affiliates before it can collect external retrocession; A and B may need claim liquidity before R pays. If R lacks liquid assets, payments may be delayed or require capital support. The group could satisfy a consolidated solvency metric while an individual subsidiary breaches a local requirement or encounters a cash shortfall. Whether H can fund the shortfall depends on actual legal restrictions, asset quality, regulatory permissions, covenants, and settlement timing.

Now suppose, solely for illustration, A holds 30 and B holds 20 in securities issued by the same corporate group as R's external retrocessionaire. The group has a direct 50 exposure to that issuer through investments and a contingent 50 reinsurance recovery arising under this modeled storm. If the storm and issuer distress coincide, up to 100 of distinct contractual amounts merit a common-counterparty stress reconciliation. This does not assert that 100 is the expected loss: asset recovery values, actual guarantors, contract coverage, collateral, and priority of claims must be modeled. Nor should the same 50 reinsurance receivable be counted again as a separate investment without matching instrument identity.

A transmission network

Represent group obligations as directed edges between legal entities. For a scenario ss, let pef(s)p_{ef}(s) be the actual payment by entity ee to entity ff, and be(s)b_e(s) be available liquid resources after the entity's other obligations. A cash shortfall occurs when required outflows exceed cash and permitted inflows at the relevant date. Accounting elimination of internal edges yields zero at group consolidation, but the timing and capacity constraints along those edges determine which subsidiaries can meet obligations.

A useful stress cascade therefore proceeds in dated stages: shock occurrence; direct claims and investment losses; recognition of reinsurance and derivatives; counterparty recoverability; contractual margin or collateral calls; internal liquidity transfers; any regulatory approval or ring-fencing conditions; and remaining legal-entity solvency. Each stage should preserve the assumed event and payment state. A simultaneous aggregate loss cannot describe a failed payment chain unless its cash-flow dependencies are added explicitly.

The risk of group contagion is analyzed in the scholarly Cummins and Weiss (2014) study of systemic risk and the U.S. insurance sector. Their sector-level research motivates distinctions among institutional interconnectedness and broader market transmission; it does not establish that all insurance activity is systemically risky, nor does it yield an observed failure probability for the example group. The independent causal inference here is narrower: group entities may share losses, liquidity claims, or external funding sources through identifiable contractual and financial links.

Fungibility and double gearing

Capital exists where it can legally absorb losses. A consolidated statement of resources does not guarantee that a subsidiary has access to funds held in another jurisdiction or in a separate company. Capital can be legally restricted, economically illiquid, tax-inefficient to move, pledged, dependent on approvals, or already committed to a different obligation. The group map should therefore associate resources with entities, currencies, admissibility, transfer conditions, and time-to-availability.

Double gearing arises when the same economic capital is counted as support at more than one group level without suitable adjustments. Intragroup guarantees and internal financing can also create contingent claims upon the same resource. This is why the NAIC's group capital supervisory discussion and IAIS standards attach importance to group structure as well as the headline number. A group capital ratio and an entity-level liquidity analysis are complementary views of the same enterprise.

Monitoring, Limits, and Validation

Turning the map into decisions

An exposure map becomes operational when each displayed concentration has an accountable owner, a measurement definition, review frequency, threshold or escalation rule, and a documented set of management responses. A board-approved risk appetite creates the qualitative and quantitative boundary within which management selects concentration limits. Limits may be expressed as a percentage of eligible own funds, stress loss, earnings, available liquidity, gross insured value, aggregate line capacity, recovery receivable, or service-provider dependency. These denominators are not interchangeable.

The IAIS group-wide risk management provisions call for ongoing measurement against risk limits and governance consistent with the group's nature and complexity. The European risk-concentration supervisory provision emphasizes reported significance and group-supervisor thresholds; the institution's internal early-warning limit can be more conservative or have a different purpose than the regulatory reporting threshold. A breach means that a defined control has been triggered and must be assessed; it does not by itself establish insolvency.

For example, management could monitor maximum modeled loss from one defined storm footprint, reinsurance recoverables from a common counterparty group, and the proportion of group liquid assets placed with one financial group. Underwriting actions can include capacity caps, price and terms changes, diversified sourcing or additional external risk transfer. Asset actions can include issuer and sector caps, collateral review, and liquidity buffers. Operational actions can include service redundancy or tested recovery procedures. The effect of each action should be recomputed on the same event, entity, and time bases, not inferred from its description.

A compact test protocol

Quality assurance should operate at source, transformation, aggregation, scenario, and decision layers. The following checks form a reproducible audit agenda rather than a new formal regulatory doctrine.

Control Test evidence Failure that the test detects
Population completeness Source-system counts, bound policy totals, signed reconciliation Missing affiliate or portfolio
Exposure identity Contract, object and legal-entity keys Repeated policy or assumed risk counted twice
Location validity Geocoding confidence, peril-specific geospatial overlays Misallocated flood, wind, earthquake or wildfire exposure
Counterparty lineage Legal identifiers, guarantors, ultimate-parent trees, contract links Undetected related obligors or unsupported netting
External versus internal transfer Intercompany elimination and matched recoverable/payable records False group reinsurance credit or double gearing
Valuation basis Currency, date, limit, attachment, contract exclusions Invalid aggregation of unlike monetary measures
Tail dependence Alternative factors, copulas, event catalogs, stress sensitivities Excess diversification credit
Liquidity chronology Contract settlement lags, collateral calls, asset transfer time Apparent solvency with payment-date shortfall
Reproducibility Versioned inputs, model code/configuration, change log Unexplained movement in risk concentrations
Governance response Recorded limit owner, threshold, escalation, action and retest Unacted breach or undisclosed limit override

A strong validation test varies one input or assumption at a time and then varies plausible correlated assumptions. If precise geocoding is uncertain, run an upper and lower exposure allocation; if reinsurance credit depends on one major counterparty, test partial and full nonpayment; if capital transfer requires permission, test the time and amount available under restrictions; if catastrophe models disagree, compare physically meaningful event footprints and assumptions rather than treating one vendor's output as truth. Under the catastrophe modeling ASOP and the general modeling standard, professional model review addresses inputs, intended uses, limitations, and communication of material uncertainty in their applicable professional scope.

Reporting without losing meaning

Supervisory templates, management dashboards, scenario reports, and ORSA narratives serve different audiences and operate on different bases. A controlled mapping process should produce these outputs from traceable upstream facts. The EIOPA S.37 proportional-fund clarification illustrates why the share of an underlying collective investment cannot always be replaced by the fund's entire holding. A separate EIOPA scope clarification underscores that significant group exposures in S.37 cannot simply be restricted to asset fields present in another template (EIOPA, 2023; revised 2025). The reporting perimeter must follow the applicable instruction rather than the convenience of a source export.

A useful management exhibit states gross amount, outward risk transfer, expected collectible recovery under the base case, net retained amount, extreme-loss sensitivity, responsible legal entity, and frequency or timing of update. A supervisor additionally needs confidence that thresholds, related counterparties and intragroup transactions have been identified under the law applicable to the particular group. A public scientific manuscript cannot substitute for confidential institution-specific examination.

Limits of verification

The most material epistemic limits are exposure data that omit risk locations or counterparties, opaque contract exclusions, limited claim experience for rare events, model disagreement, changing property values and vulnerability, nonstationary hazards, endogenous counterparties, and legal constraints that become binding under stress. Parameter confidence intervals are often less important than identifying an omitted common exposure or incorrectly assuming that a contractual protection will pay. A model that cannot provide a timely and explainable stress reconciliation does not become adequate merely because it produces precise decimals.

Model validation should therefore report the scope of what has actually been tested: arithmetic, data joins, consistency with source contracts, sensitivity to alternative scenarios, and comparisons to observed historical events where appropriate. Claims of predictive calibration or real-world solvency benefit require actual independent data and methods not available to this documentary study.

Comparative Synthesis and Research Conclusions

International similarities and differences

The reviewed supervisory systems converge on the necessity of seeing risk beyond the subsidiary ledger, but they operationalize that objective differently. The IAIS principles establish integrated group-wide expectations including ORSA, interdependence, risk limits and capital resources. Solvency II expressly separates significant risk concentration from intragroup transactions and links them to supervisory thresholds and reporting. U.S. group capital and ORSA instruments supplement state legal-entity supervision through group risk assessment and capital location. The IAIS Insurance Capital Standard serves its own internationally active group capital purpose. None of these frameworks authorizes collapsing policy limits, stress losses, recoverables, regulatory capital and liquidity into one unqualified exposure number.

The documents also explain why some maps disagree without either being arithmetically wrong: supervisory group perimeter, accounting consolidation, loss valuation date, insured exposure unit, treatment of look-through holdings, and modeled stress horizon can all differ. Reconciliation requires a documented bridge between those representations. Their coexistence is a fact about distinct analytical questions, not a defect in group reporting.

Findings from the documentary and quantitative analysis

First, source-level identity determines whether the aggregate is scientifically interpretable. The same economic peril can be direct insurance, assumed internal reinsurance, ceded external reinsurance, and a counterparty receivable. These are linked transaction states. The constructed example showed a gross direct loss of 210 and a net group loss of 160 with a collectible external payment of 50. Adding internal assumed reinsurance would have produced an inflated loss; ignoring external nonpayment would have produced an unjustified net benefit. An integrated map preserves the distinction.

Second, concentration is often conditional on a common stress. An exposure to an asset issuer can coexist with contingent reinsurance claims against an affiliate of that issuer, and a geographic event can simultaneously affect underwriting losses, asset values and service continuity. A conventional location map finds only the physical portion. A legal-party network finds the contractual portion. Integrated scenario interrogation is required to determine the interaction.

Third, quantitative aggregation is inseparable from the chosen dependence assumption. The constructed two-risk calculation changed the aggregate standard deviation from approximately 95.66 to 115.54 when correlation moved from 0.25 to 0.85; neither number established a tail quantile. Model credibility rests on the physical and financial mechanism, selected horizon, observed data, uncertainty, and validation, not on the cosmetic sophistication of the dependence model.

Fourth, group resources are located in legal entities. A consolidated group loss calculation cannot answer whether entity A can settle claims before its reinsurer pays or whether capital in H can legally enter R. The same shock must be examined as an aggregate economic loss and as dated legal-entity cash demands. This is consistent with the capital-fungibility and contagion concerns expressed in the IAIS and NAIC materials.

Fifth, risk maps are only as complete as their lineage and exception handling. A correct portfolio total can conceal incorrect hazard placement, connected counterparty identification, outdated protection terms, or wrong supervisory perimeter. Validation must preserve totals and also interrogate grouping assumptions. Testable lineage improves both internal decisions and the consistency of supervisory explanations.

Research implications

An empirical program capable of testing the operational value of exposure mapping would require data on the same insurer groups before and after identifiable control or technology changes, actual counterparty and catastrophe outcomes, relevant business and regulatory confounders, and explicit outcome measures such as identified losses, liquidity shortfalls, limit breaches, or capital actions. The available public standards establish norms and observed reporting frameworks; they do not establish the causal effect of map adoption on firm survival. The potential value of integrated exposure visibility is analytically supported by the examples, but its magnitude is an open empirical question.

Another research direction concerns the behavior of shared service providers and reinsurance networks under simultaneous catastrophe and financial stress. An insurer's external protection can alter the allocation of loss while producing a concentrated claim on a single source of recovery. System-level evidence would need matched contracts, event histories, collateral and settlement data, and legal-entity payment networks. The distinction between economic loss transfer and recoverable risk transfer is central to that inquiry.

Conclusion

Insurance group risk concentration is an enterprise-level property of shared exposures, dependencies, and legal-financial constraints. Its effective mapping begins with identifiable contracts, entities, objects, counterparties, locations, and event definitions; proceeds through explicit gross, intragroup, outward-transfer, and consolidated reconciliations; and culminates in stress-sensitive management and supervisory decisions. The international primary sources establish the legitimacy of group-wide risk identification and supervision, while their differing legal perimeters clarify why a single reporting template cannot serve every purpose.

The independent calculations demonstrate three decisive propositions under stated assumptions: internal reinsurance payments cancel at consolidated group level while remaining economically relevant to subsidiaries; an external claim recovery reduces group loss only to the extent it is collectible; and assumptions about dependence and transferability materially alter the apparent concentration. The resulting analytical picture recognizes geography, counterparty ownership, common shocks, and capital location as simultaneous attributes of insurance risk.

The practical conclusion is to preserve multiple reproducible views of the same underlying risk and require an explicit bridge among them. This allows boards, actuaries, risk functions, and supervisors to understand which risks are shared, where losses will emerge, who must pay, and which resources can respond in time. The research contributes a documented comparative synthesis and independently checked illustrations, while leaving insurer-specific empirical measurement and causal performance testing to datasets and research designs capable of supporting them.

References

  • Actuarial Standards Board. (2019). Actuarial Standard of Practice No. 56: Modeling (revised December 2019, effective October 2020). Official standard.

  • Actuarial Standards Board. (2021). Actuarial Standard of Practice No. 38: Catastrophe Modeling (for All Practice Areas) (revised July 2021, effective December 2021). Official standard.

  • Basel Committee on Banking Supervision. (2013). Principles for effective risk data aggregation and risk reporting (BCBS 239, January 2013). Bank for International Settlements. Official publication.

  • Cummins, J. D., & Weiss, M. A. (2014). Systemic risk and the U.S. insurance sector. Journal of Risk and Insurance, 81(3), 489–528. Publisher and DOI.

  • Embrechts, P., McNeil, A. J., & Straumann, D. (2002). Correlation and dependence in risk management: Properties and pitfalls. In M. A. H. Dempster (Ed.), Risk Management: Value at Risk and Beyond (pp. 176–223). Cambridge University Press. Author bibliography and paper.

  • European Commission. (2015). Commission Delegated Regulation (EU) 2015/35 of 10 October 2014 supplementing Directive 2009/138/EC, especially Articles 376–377. Official Journal of the European Union. Official legal text.

  • European Commission. (2023). Commission Implementing Regulation (EU) 2023/894 of 4 April 2023, supervisory templates and instructions including S.37. Official Journal of the European Union. Official legal text.

  • European Insurance and Occupational Pensions Authority. (2020). Regulatory Q&A 2203: Risk concentration, S.37.01, collective investment undertakings (final response). Official answer.

  • European Insurance and Occupational Pensions Authority. (2023a). Regulatory Q&A 2724: S.37.01 credit and insurance risk mitigation. Official answer.

  • European Insurance and Occupational Pensions Authority. (2023b). Regulatory Q&A 2783: S.37.01 insurance-policy exposures and treaty capacity. Official answer.

  • European Insurance and Occupational Pensions Authority. (2023/2025). Regulatory Q&A 2630: S.37 exposures and S.06 reporting scope (answer revised March 2025). Official answer.

  • European Insurance and Occupational Pensions Authority. (2026). Financial Stability Report: June 2026 (news summary dated June 24, 2026). Official report summary.

  • European Parliament and Council. (2009, consolidated through January 9, 2024). Directive 2009/138/EC on the taking-up and pursuit of the business of Insurance and Reinsurance (Solvency II), Articles 244–245. EUR-Lex consolidated text.

  • International Association of Insurance Supervisors. (2024a). Insurance Core Principles and Common Framework for the Supervision of Internationally Active Insurance Groups (December 2024). Adopted text.

  • International Association of Insurance Supervisors. (2024b). Insurance Capital Standard. Official standard overview.

  • International Association of Insurance Supervisors. (2025). Global Insurance Market Report 2025 and monitoring overview. GIMAR publications.

  • International Association of Insurance Supervisors. (2024–2026 online). Insurance Core Principles and ComFrame: Online Tool, including ICP 8, ICP 16, ICP 23 and ComFrame. Official online tool.

  • Lloyd's. (2026). Realistic Disaster Scenarios: 2026 Scenario Specification and market requirements. Official scenario library.

  • McNeil, A. J., Frey, R., & Embrechts, P. (2015). Quantitative Risk Management: Concepts, Techniques and Tools (revised edition). Princeton University Press. Publisher record.

  • National Association of Insurance Commissioners. (2025). Own Risk and Solvency Assessment (ORSA) Guidance Manual (2025 edition). NAIC official manual.

  • National Association of Insurance Commissioners. (2026). Group Capital Calculation, insurance topics and group supervisory framework (page updated January 5, 2026). NAIC official explanation.

  • Swiss Re Institute. (2026). Natural catastrophes in 2025: The persistent rise of wildfire and storm risk. Sigma, 1/2026. Research publication.