RESEARCH DOSSIER
Insurance Merger Control and Market Concentration Measurement
Abstract. Insurance merger control concerns the competitive consequences of changing control over risk-bearing capacity, underwriting decisions, intermediation, and the acquisition of complementary inputs. Concentration measurement provides a structural representation of those changes only after the market, decision-making enterprise, transaction perimeter, observation period, and activity measure have been identified. This research examines competition and prudential control in the United States, European Union, United Kingdom, and Australia through legislation, supervisory materials, merger decisions, and published economic research. The study develops an independent comparative synthesis of relevant market boundaries, alternative denominators, ownership aggregation, the Herfindahl–Hirschman Index, residual-share bounds, the relationship between statistical concentration and direct rivalry, and the identification of merger-induced changes in price, coverage, and service. Worked examples are synthetic and reproducible; a separate bounded exercise uses published redacted share intervals without reconstructing confidential quantities. The comparative analysis explains why a merger can be competitively harmful within a local risk market despite modest national shares, why multiple brands may represent one underwriting decision maker, and why clearance on competition grounds and authorization under prudential supervision require distinct evidence. The results establish the evidentiary conditions for interpreting structural screens and evaluating entry, efficiencies, counterfactual rivalry, and remedies. No new empirical causal estimate or universal theory is claimed.
Keywords: Insurance merger control; market concentration; Herfindahl–Hirschman Index; market definition; underwriting groups; competition economics; premium measurement; insurance brokerage; prudential supervision; merger remedies; counterfactual; regulatory comparison.
Introduction
A merger is an intervention in the organization of future competition. In insurance the intervention changes the identity of the enterprise that assumes contractual liabilities, allocates scarce underwriting capacity, determines offers, or controls access to an intermediary. This makes insurance merger analysis a problem of institutional organization as well as industrial economics. The same insurance population can be represented by contracts, insured persons, exposures, premiums, reserves, limits, commissions or expected losses; each representation answers a different economic question. Structural concentration is meaningful only when the representation corresponds to the competitive decision of interest.
The central inquiry of this monograph is how market definition, concentration measurement and evidence of competitive effects jointly support the assessment of insurance mergers. Three related questions guide the investigation: which institutions and records establish the competitive unit and transaction boundary; how do alternate measures alter the observed distribution of market participation; and how can the incremental effects of a transaction be separated from changing claims costs, demand, regulation and capacity?
The analysis distinguishes an insurer acquisition from a merger of brokers, a portfolio transfer, a reinsurance transaction and vertical combinations with providers. Their overlap may be economically important, but their denominators and causal mechanisms differ. A broker places risks with insurers and may coordinate access to markets; an underwriter accepts liability; an administrator processes benefits. These functions may coexist within a corporate group. Classification follows function and control rather than branding or consolidated financial revenue alone.
The research employs documentary comparative analysis of specified historical cases, a mathematical audit of concentration measures and a critical examination of published empirical results within the scope of material accessed. The cutoff for discovery is November 19, 2026. Historical rules and decisions are identified by their dates. Accessible public decisions contain redacted data and agency announcements summarize particular findings. The analysis does not reconstruct undisclosed firm shares, estimate a jurisdiction-wide historical panel, or claim to have replicated restricted microdata. Instead it makes explicit what can be inferred from disclosed facts and what would require original observations. This distinction allows institutional and mathematical results to be used without overstating their empirical scope.
Conceptual and Historical Foundations
The economic unit of competition
Insurance supply is the combination of a promise, a price, contract conditions, capital, distribution and claims administration. Competition may concern the probability that a risk is accepted, the size of the deductible, restrictions on covered events, the quality of adjustment or the availability of renewals. A lower quoted premium can accompany an economically inferior contract if the consumer's retained risk increases. Consequently, merger analysis must attend to both price and the content of the protection delivered.
A relevant insurance enterprise is the decision-making unit capable of independently changing those dimensions. Separate incorporated insurers can share a parent, underwriting authority, reinsurance program and centralized strategic decisions. Separate brands may use one risk carrier. Conversely, a broker's placement decisions can be independent of the carrier it uses. Treating every brand as an independent competitor can create an artificially fragmented picture of underlying risk-bearing supply. Combining all entities of a loosely connected commercial network may also erase genuine autonomy. The appropriate aggregation depends on control rights and the activity under assessment.
A change of control can occur without the disappearance of the insured's policy document. Shares, renewal rights, portfolios or underwriting operations may move between owners while policy names continue. Conversely, a transfer of administration alone may leave underwriting competition unchanged but alter service procurement. A precise transaction description must identify which assets, liabilities, rights, distribution arrangements and decision-making functions transfer. The relevant counterfactual is the configuration of independent choices without the transaction, rather than the legal shell count.
Structural enforcement and institutional purpose
The statutory basis in the United States is Clayton Act § 7, which addresses acquisitions capable of materially impairing competition or tending toward monopoly. This prospective legal inquiry differs from the retrospective question of whether a completed combination has demonstrably produced a given price increase. The 2023 enforcement framework remains analytically useful because it sets out structural and competitive factors. Its nonbinding guidance status matters: it describes agency assessment; adjudication remains governed by law and case evidence.
The evolution of structural screens demonstrates the relationship between quantitative indicators and institutional judgments. The U.S. agencies' 2023 Guideline 1 explains the historical evolution of HHI screening, including the higher thresholds adopted in 2010 and the return to earlier concentration thresholds in 2023. An HHI statistic for a fixed market does not change with a regulatory revision, but the policy interpretation of that statistic can change. Comparisons of decisions across decades therefore require contemporaneous screening rules rather than retroactive imposition of a later threshold.
In the European Union, Council Regulation No 139/2004 supplies the statutory framework for assessing concentrations and significant impediments to effective competition. The European Commission's 2004 horizontal guidance treats HHI ranges as indicators in a wider substantive examination. The United Kingdom uses its substantial lessening of competition analysis. Australia's merger regime underwent a procedural change in 2026. These institutions share economic questions about rivalry but differ in statutory wording, evidentiary procedure and the consequences of a particular screen.
Insurance prudential control
Prudential supervision addresses the ability of insurers to honor promises over time. The IAIS's Insurance Core Principles and ComFrame establish a supervisory setting concerned with policyholder protection and group-wide risk. They acquire binding legal force through relevant domestic implementation. National supervisors may evaluate the financial condition of a prospective owner, capital adequacy, governance and continuity of the acquired insurance operation. The NAIC's Form A filing guidance directs a proposed insurance-company acquisition or merger filing to the domestic state in the U.S. state-based system. Its model-law references must be distinguished from the actual law of a given state.
Competition and prudential control respond to different questions even when they concern one acquisition. The transfer of a vulnerable portfolio into a well-capitalized group might improve the enforceability of policyholder claims and simultaneously remove an important competitive constraint in a specialty insurance line. Strong capital ratios do not establish sufficient rivalry; numerous competitors do not establish the solvency of a specific carrier. Neither conclusion may substitute for the other. An institutionally complete transaction analysis preserves both records, identifies the timing of each approval, and reconciles any restrictions on the perimeter of the acquired activities.
Primary Institutional Documents and Comparative Research Design
Comparative legal design
The four principal jurisdictions are selected for contrasting enforcement settings and accessible insurance transaction evidence. They do not constitute a representative random sample of all jurisdictions. The U.S. comparison includes judicial outcomes; the EU comparison includes a conditional decision summary; the British comparison uses public Phase 1 decisional evidence; Australia provides contrasting personal-line assessments. Outcome classifications therefore separate court injunction, conditional clearance, clearance, opposition and abandonment.
The Commission's 2004 Horizontal Merger Guidelines supply historical concentration ranges, including lower-delta conditions in moderately or highly concentrated markets. They are agency screening considerations with qualifications, not a cross-jurisdictional statutory safe harbor. The Commission's 2026 guidelines review records a draft consultation that closed on June 26, 2026. This monograph uses the dated 2004 benchmarks for the historical institutional comparison; it does not infer any adopted replacement from a consultation.
The CMA's Merger Assessment Guidelines (2021) frame case-specific competitive assessment and explicit counterfactual analysis. A decision to clear a merger at the applicable Phase 1 standard differs procedurally from a judgment after a contested trial. The comparative method therefore records both decision outcome and stage. The ACCC's merger regime overview identifies January 1, 2026 as the commencement of Australia's mandatory regime for qualifying acquisitions; the September 2026 legislative amendments are context for interpreting contemporary procedure. A 2025 decision should be described under the procedure then in force.
Legal jurisdictional thresholds answer whether and where a transaction must be notified or can be reviewed. Economic screens answer what the market structure suggests after relevant market identification. A global transaction can exceed notification thresholds while having a small horizontal overlap in one jurisdiction. A smaller transaction can substantially eliminate rivalry in a narrow market. Conflating notification thresholds, capital-control thresholds and HHI thresholds transforms three different tests into a misleading single measure.
Evidence structure and sample construction
An original observation for quantitative merger analysis should be keyed to market, time, activity and enterprise. The enterprise record needs legal name, ultimate decision-making group, changes of ownership, operating brands, underwriting carrier, relevant distribution rights and dates. The market record needs the geographic level, covered product, type of buyer, type of risk, policy issuance basis, period and nominal currency where monetary data are used. The activity record should identify whether the observation is a stock, such as policies in force, or a flow, such as written premium during a reporting year.
Domestic supervisory records are often organized by licensed entity and statutory class. Such records are a useful population frame but require reconciliation to group and competition boundaries. The NAIC's 2024 market-share data release describes countrywide direct written premiums for leading groups and companies, based on filings received as of March 3, 2025. Its all-lines property/casualty top-ten cumulative share of 51.40% was a preliminary reporting snapshot. This is a national descriptive statistic for the reporting population at that date, not a completed HHI for a defined relevant market or a measure of local competition. A cumulative share contains no information about the allocation within the top-ten group without further data.
For every selected case we extract the legal outcome, activity at issue, geographic or purchaser segmentation, measure of competitive significance, evidence of rivalry, alternative suppliers, counterfactual and documented remedy. When an authority reports a share interval, the interval is retained. When an outcome is reported by a litigating agency, the underlying court decision is preferred when available; characterization by a party is not treated as an independently established empirical effect.
The case sample intentionally includes cleared and contested mergers. It illuminates the mechanisms that agencies consider relevant but cannot establish a population probability of clearance or an average merger effect. Public merger files are selected by notification, litigation, disclosure and agency publication practices. A statistical sample of published decisions would require a denominator encompassing investigated and uninvestigated transactions, including changes of procedure, threshold exemptions and withdrawals. No such database has been constructed here.
Evidence matrix
| Case | Date and status | Competitive layer |
|---|---|---|
| Aetna / Humana | 2017 court block | Individual Medicare Advantage; exchange plans |
| Anthem / Cigna | 2017 block affirmed | National accounts; large employers |
| Aon / Willis Towers Watson | 2021 EU conditional clearance; transaction terminated | Commercial risk and reinsurance brokerage |
| Aviva / Direct Line | 2025 CMA clearance | UK personal car and home underwriting |
| IAG / RACQ Insurance | 2025 ACCC non-opposition | Queensland personal lines |
| IAG / RAC Insurance | 2025 opposition; 2026 Phase 2 opposition reported | Western Australian personal lines |
| UnitedHealth / Amedisys | 2025 court-approved settlement | Home health and hospice services |
The cases are treated as separate records rather than interchangeable examples of insurer market power. In particular, the brokerage and provider cases concern adjacent layers. They demonstrate why enterprise identity alone does not define the economic transaction.
Market Definition in Merger Cases
Substitution and contract boundaries
The relevant market is a set of competitive alternatives defined for a particular assessment. The European Commission's 2024 Market Definition Notice emphasizes contemporaneous competitive conditions rather than a firm's commercial classification. An insurance statistical line is a valuable starting point, but the statutory label may contain policies with different customer requirements or risk eligibility. An insurer offering travel cover is not generally a substitute for a homeowner's insurer; within home insurance, a policy that excludes a required hazard may also provide a poor alternative.
The U.S. guidelines' hypothetical monopolist test evaluates whether a hypothetical single controlling supplier would profitably worsen terms. In insurance those terms include premiums, deductibles, limits, exclusions and service. Application requires expected cost and customer response, not only price elasticity. For a product with expected claims of 70, variable expense of 10 and premium of 100, a 5% premium increase raises the simplified contribution from 20 to 25 per retained exposure. Holding costs and risk mix fixed, the critical retained fraction is 20/25, or 80%; a loss of 20% of exposures would eliminate the gain. This is a synthetic one-product illustration, not a measured elasticity. If the departing exposures have lower-than-average expected claims, retained costs rise and the calculation changes. Competitive responses can also occur through brokers, nonrenewal or revised coverage. A market test that assumes unchanged risk composition while significant selective switching is plausible should explicitly model that sensitivity.
Insurance shopping often involves quotes rather than standardized prices. A household may request several offers but accept only one. A large business may invite submissions and award different layers to different insurers. Information about quote overlap, bids, insurer declines, customer switching, broker panels and actual placement is often more directly informative than the count of company names. However, a survey of completed switches excludes consumers who could not find a usable alternative. Quote data exclude people who were deterred before requesting a quote. The selection mechanisms must be recorded.
Aetna and Humana: local demand and institutional boundaries
In the Aetna–Humana memorandum opinion (2017), the District Court treated individual Medicare Advantage plans as an antitrust product market and concluded that the proposed merger was likely to substantially lessen competition across the 364 challenged counties. Structural evidence and direct competition supported that conclusion. It also found that entry and the proposed transfer to Molina would not adequately restore the eliminated rivalry. The DOJ's January 2017 case announcement independently corroborates the result and the rejected proposal to divest approximately 290,000 Medicare Advantage customers.
Medicare Advantage illustrates the need to separate a public entitlement from competing products delivering benefits under its institutional rules. Regulatory constraints do not necessarily remove the dimensions on which participating plans compete. Geographic definition also follows the practical set of offers available to consumers. National participation by multiple groups can coexist with concentrated county-level supply. National enrollment shares are inadequate when local plan choice is the relevant decision.
The same opinion distinguished exchange markets and assessed future competition in three Florida complaint counties. The methodological implication is that a firm's observed exit decision should be examined in context. If participation changes during merger proceedings, a mechanical comparison of then-current insurers could mischaracterize the independent counterfactual. Distinguishing deliberate withdrawal connected with the transaction from economically unavoidable exit is a substantive evidence task.
The Anthem and Cigna matter adds segmentation by purchaser. The DOJ's 2017 appellate outcome announcement reports that the appellate court affirmed the block and identifies district court findings for national accounts based in fourteen states and large employers in Richmond, Virginia. For large employers, a broad geographic network and the ability to administer benefits across locations can be critical procurement requirements. An insurer capable of serving an individual household may not be a credible substitute for an employer demanding integrated multi-location coverage.
Brokerage and specialty risks
The Commission's published Aon–Willis Towers Watson decision summary (decision 2021, summary 2024) distinguishes commercial risk brokerage by category of risk, customer and geographic scope. For the initial transaction, the Commission identified concerns involving large multinational customers and several specialty placements, along with worldwide treaty and facultative non-life reinsurance brokerage. The Commission conditionally cleared the proposal. The U.S. DOJ subsequently reported termination of the merger agreement in July 2021 after its challenge. Conditional clearance in the EU and abandonment of the transaction are different events.
Brokerage market shares can be measured by remuneration, placed premium, client relationships or transaction counts. None is interchangeable with insurer direct written premium. A broker arranging a large placement does not necessarily retain the underwriting exposure. Conversely, a small number of brokers may provide access to specialized global capacity even when underwriting capacity itself is relatively dispersed. The relevant question is which brokers could deliver the required advisory and placement service reliably and at scale.
A multinational customer's workable alternatives may extend globally while a small commercial customer depends on local regulatory access and distribution. A market organized around the location of the office issuing a contract might misrepresent where the customer sources brokerage or where insurers must be licensed. Geography is determined by the economics of substitution combined with institutional restrictions, rather than copied from an administrative reporting unit.
UK personal insurance and distribution
The Aviva–Direct Line decision (2025), especially paragraphs 46–59 and 68–73, examines underwriting of personal car and home insurance in the UK. Its published evidence differentiates volume through live policy counts and value through gross written premiums. It also considers price comparison websites, switching evidence, internal documents and third-party assessments. The CMA's public case register records clearance on July 1, 2025.
In a differentiated insurer market, an individual can compare brands while the relevant underwriting choices sit behind those brands. Price comparison websites affect search costs and distribution, but the visible order of quotes does not automatically establish carrier independence. Investigating ownership and underwriting authority is especially important where several branded offers appear near one another. A count of top-ranked offers may be more concentrated at group level than at brand level, even if the consumer's observed shopping experience appears diverse.
The CMA's use of both volume and value is particularly instructive. An insurer specializing in higher-value vehicles or broad protection may receive more premium for each live policy than its rivals. Its value share then exceeds its volume share without proving a larger number of customers could switch to it. The economic question determines which share is informative. Both should be retained when available; unexplained divergence motivates further inquiry rather than the selection of whichever statistic favors a particular transaction conclusion.
Concentration Methods
Market shares and observation measures
Let x_i be the nonnegative activity of independent enterprise i in the relevant market and reporting period. Let X equal the sum of all eligible enterprises' activity. Its percentage share is s_i = 100 x_i / X, with the shares summing to 100 apart from reporting rounding. A meaningful numerator and denominator must refer to the same product, time period, geography, population and measure. The U.S. agencies' market-share guidance (2023) recognizes that excessive market breadth can understate competitive significance and that the definition of participants depends on the competitive setting.
Direct written premium measures insurance business written before ceded reinsurance is deducted. Gross written premium, as reported in a particular dataset, may include assumed business or other items under that dataset's accounting convention. Earned premium relates to elapsed protection rather than current writing activity. Net premium reflects assumptions and cessions. Policy counts measure contracts but differ from insured lives, vehicles, properties or exposure units, and a single group contract may encompass many people. In reinsurance, premium volume, exposure limit and available capacity may identify different competitive standings. The analysis must establish a consistent reporting convention before summing across enterprises.
In an assumed example, an insurer with 20,000 live policies averaging $1,500 and a rival with 30,000 policies averaging $800 write $30 million and $24 million respectively under a simplified single-period correspondence between policy and premium. Among these two hypothetical suppliers, the first has 40% of policy volume and about 55.56% of premium value. It can have a larger monetary position while covering fewer policies. The example demonstrates arithmetic rather than evidence about any particular market. In an actual portfolio, differences in term, cancellations, earned exposure and new-business timing prevent direct multiplication unless the measurement definitions correspond.
CR measures, entropy and HHI
The concentration ratio CR_k sums the shares of the k largest independent suppliers. It communicates how much of the market is accounted for by the leading firms, but two markets with the same CR_4 can differ significantly inside and outside that group. The Herfindahl–Hirschman Index, with percentage shares, is HHI = sum of s_i squared. A monopoly has HHI 10,000; N equally sized independent firms have HHI 10,000/N. The reciprocal normalized index, 10,000/HHI, is the effective number of equally sized suppliers corresponding to the observed index. It is an interpretation of dispersion, not a count of firms actually present.
Entropy, defined using decimal shares p_i as minus the sum of p_i times ln(p_i), emphasizes the distribution of shares differently. Its exponential is another equivalent-number representation. Measures such as HHI and entropy are useful sensitivity descriptions, but a legal threshold stated in HHI points is not transferable to another measure without a justified mapping. A market's leading-firm stability, entry and shares over time may convey evidence not contained in any one snapshot.
For a pure merger of distinct enterprises A and B, keeping the relevant market and all activity constant, replacing their separate shares a and b by a+b increases the HHI by 2ab. This follows directly from expanding the square. It also clarifies why a merger of two mid-sized rivals can have a more substantial structural increment than the acquisition of a very small firm by a large one. The identity is exact only for the structural combination of fixed shares; it does not forecast future market growth, customer redistribution or divested portfolios.
Synthetic concentration audit
Consider six independent insurance groups with percentage shares 28, 22, 17, 13, 11 and 9. Suppose the first and last propose to merge. All six shares and their squared contributions are specified here; the example does not describe an observed insurance market.
| Group | Initial share | Initial squared share |
|---|---|---|
| A | 28 | 784 |
| B | 22 | 484 |
| C | 17 | 289 |
| D | 13 | 169 |
| E | 11 | 121 |
| F | 9 | 81 |
| Total | 100 | 1928 |
Combining A and F gives shares 37, 22, 17, 13 and 11. The post-merger HHI is 2432, an increase of 504 points. CR_4 changes from 80 to 89. The effective number moves from approximately 5.19 to 4.11. The mechanical delta equals 2 × 28 × 9 = 504. Under the U.S. 2023 guideline, this constructed transaction exceeds both structural-presumption combinations: post-merger HHI above 1800 with delta above 100, and merged share above 30 with delta above 100. That application is an illustration of the published enforcement screen, not a legal opinion on an actual merger. Rebuttal evidence can matter even where a screen is exceeded.
Suppose A has two brands with shares 18 and 10 under one independently controlling parent before the merger. Measuring brands separately subtracts 2 × 18 × 10 = 360 from the initial group HHI, yielding 1568 even though the underlying ownership structure has not changed. Reporting this number as the pre-merger group index would understate concentration. After the actual transaction, counting the existing brands separately could again distort the relevant enterprise measure. A consistent group crosswalk is therefore a prerequisite for longitudinal measurement.
Consider instead a divestiture of four percentage points from the acquired F portfolio to a genuinely new independent entrant, leaving combined A+F with 33 and the entrant with 4. Keeping other shares fixed yields a post-transaction HHI of 2168 and an increment of 240. This sensitivity is computed from share squares, not an assertion that the entrant would be viable. The intervention reduces concentration relative to the unremedied merger, yet remains above the U.S. structural thresholds in the example. A divestiture to an existing supplier would cross-add to that supplier's share; the corresponding HHI effect depends on its starting position. Remedy evaluation must examine independence, capacity and business transfer beyond arithmetic.
Incomplete tails and confidential intervals
A top-k report does not identify the HHI of the full market. If disclosed individual firms have squared-share sum K and omitted firms together hold R percentage points, the omitted contribution is at most R squared (one omitted firm). If the number m of omitted independent enterprises is known and each can have equal share, Cauchy's inequality gives a minimum R squared / m. When m is unknown, K is a conservative lower bound; tighter bounds can use a known maximum individual residual share or ordering. Restrictions on residual firm size should be enforced explicitly before the upper bound is called feasible.
For example, reported independent shares of 30, 25 and 15 imply K = 1750 and R = 30. With six omitted equal firms, the lower feasible index is 1900; the unrestricted upper bound is 2650. If no omitted firm can exceed 15, the maximum is 2200, obtained by allocating the residual to two 15-point firms, or approached if six firms must each have strictly positive shares. The need to identify m, size caps and whether zero is allowed prevents false precision. A CR statistic alone cannot resolve these differences.
The CMA's published 2025 Aviva decision reports a combined share band of 20–30% by both value and volume in UK personal car underwriting, with each merging firm in the 10–20% band. Treat those bounds as closed for a conservative outer envelope, although the publication uses rounded confidential bands rather than exact endpoints. Since a and b are each at least 10 and their sum at most 30, the minimum possible merger delta under independently available interval constraints is 200, at a=b=10, and the maximum is 450, at a=b=15. Using only the two individual bands would give an upper bound of 800, inconsistent with the combined band. This is a constrained synthetic envelope derived from the printed bands, not the CMA's actual HHI or estimated delta. Additional information and exact rounding conventions may narrow it. The remaining independent shares are themselves interval-coded and the residual tail is aggregated. No exact HHI can be responsibly stated from those data.
For the Western Australian case, a combined share interval alone cannot identify the increment because the division between the parties is missing. At a combined share of 60, illustrative pairs of 30 and 30 versus 55 and 5 yield deltas of 1800 and 550 respectively. Both produce the same combined share and different changes in concentration. The comparison reinforces the distinction between combined-market position and merger-induced structural increment. Neither pair is alleged to describe the actual Australian insurers.
Geographical aggregation and time
Consider two equally sized local markets. In the first, independent groups A and B have shares 80 and 20; in the second, groups C and D have shares 80 and 20. Each local HHI is 6800. National pooling gives respective shares 40, 10, 40 and 10 and national HHI of 3400. The simple mean of the local HHIs is 6800. The difference arises because the national calculation first pools observations and then squares shares. If local purchasers cannot switch to the other territory, the national index understates the concentration of their actual alternatives.
For weighted local share vectors p_m and weights w_m summing to one, the weighted mean of squared shares exceeds or equals the sum of squares of the pooled shares by the weighted variance of each group's local shares. This follows by expanding the squared deviation from each enterprise's weighted mean. The inequality is a mathematical relationship, not an assertion that national markets are always incorrectly defined. If buyers can actually substitute nationally, the national market might be appropriate. If local competition is relevant, the weighted local index or its distribution preserves the appropriate level of observation. The choice of weighting by premium, exposures or population should be justified.
The denominator also changes over time. Growth in premium during claims inflation may increase all firms' nominal written volume while leaving exposure shares unchanged. Growth concentrated among high-risk portfolios may make value shares more unequal without a corresponding reduction in customer alternatives. Group acquisitions can mechanically change reported HHI even before a new pricing strategy is adopted. Analysis must identify the dates of legal closing and operational integration and distinguish initial structural impact from subsequent market responses.
Competition Effects
Rivalry, price, quality and risk selection
A concentrated structure creates conditions in which removing a close competitor may allow the remaining enterprise to worsen contractual terms. The mechanism is the internalization of diversion: business lost from one merging firm's product may now flow to another product owned by the same parent. However, heterogeneous consumers and insurance risks make diversion economically asymmetric. Customers departing from a product with high expected claim costs generate different margins than customers with low expected losses. A premium-based diversion estimate that treats all customers as economically identical can therefore misstate the pricing incentive.
Consider an indicative unilateral price-pressure calculation. If one policy has premium 100, expected incremental claims and expenses of 80, and the other merging firm's contribution on a comparable retained policy is 25, with a diversion fraction of 0.30, the simple diverted contribution is 7.5 per lost policy. A per-policy marginal cost saving of 4 would offset part of this incentive in a highly simplified first-order comparison, leaving 3.5 before accounting for own-price response and other factors. This is illustrative algebra, not a calculated UPP for an observed transaction and not a predicted premium change. Contract differences, fixed expenses, risk selection, renewals and equilibrium responses require additional treatment.
Competition can be reduced through narrower coverage or weakened service as well as higher premiums. Insurers choose limits, deductibles, selection rules, underwriting appetite, access to repairers, turnaround times and handling of claims. A premium regression without adjustment for contract characteristics can misattribute a change in product quality to price pressure. The dependent variable should be selected according to the theory of harm: like-for-like quote, policy renewal premium, coverage-adjusted unit price, claims service, probability of receiving an offer, number of eligible insurers, or coverage limits offered.
Coordination is a different mechanism. A merger may simplify the environment for monitoring rivals' behavior and reduce deviation from common patterns. Public rate filings or price-comparison information can have complex effects because they reduce customer search costs and may also increase transparency among suppliers. An empirical coordination assessment should investigate actual strategic responsiveness, available communications evidence, price dispersion and changes in incentives. Parallel premium increases during a catastrophe-related cost shock are not sufficient evidence of coordination.
Contrasting Australian transaction evidence
The ACCC's IAG–RACQ non-opposition announcement of May 2025 concerns home and contents and motor insurance, predominantly in Queensland. It identified continuing competition from other suppliers, assessed RACQI's declining market position and limitations on its future ability to compete, and considered markets for repair services. The authority did not oppose the proposed acquisition. Its assessment depended on the specific target and competitive conditions rather than on a universal principle that mutual insurers should consolidate.
The December 2025 IAG–RAC WA opposition provides a contrast. The ACCC reported estimated combined shares of approximately 55–65% for motor and 50–60% for home and contents insurance in Western Australia, found substantial direct rivalry and considered the target sufficiently well positioned to continue competing. It did not expect remaining suppliers to replace the lost competitive pressure. An official indexed October 2026 Phase 2 opposition announcement reports the ACCC's adverse assessment under the later process; direct page retrieval was unavailable, so only the indexed procedural outcome is included. No subsequent review or appeal outcome is assumed.
The comparison is informative because the acquiring group is common and the activities overlap in broad insurance classification, yet the target's competitive significance and geographic conditions differ. It is not a controlled causal experiment: target strength, market composition, exposure to hazards, available capital, rivals and regulatory procedure all differ. Still, it supplies institutionally direct evidence that target sustainability and effective substitution are central to merger analysis. A general argument about rising reinsurance costs cannot establish that a specific target would exit absent a sale. Capital challenges must be evaluated against available funding, underwriting adjustment, alternative ownership or continued independent operation.
Upstream and downstream effects
Insurance groups both sell coverage and procure services. Merger-induced buyer power may change payments to physicians, repairers, administrators or catastrophe-model suppliers. Lower procurement prices can represent efficiencies when they reduce rents or transaction costs and preserve supply; in other conditions, coercively reduced compensation may diminish entry, investment, quality or capacity. The correct output assessment follows the affected input and downstream consequences.
Vertical combinations can also create incentives to disadvantage outside rivals through access terms, preferential data or control of inputs. The CMA's 2021 guidelines discuss foreclosure ability and the importance of alternative suppliers. A vertically integrated insurer might have an advantage in provider negotiation, but a claim of foreclosure requires evidence of ability, incentive and likely competitive consequences rather than integration by itself.
A relevant adjacent example is UnitedHealth–Amedisys settlement approved in December 2025. The DOJ reported required divestitures of at least 164 home health and hospice locations across nineteen states. That transaction concerns healthcare service competition involving an insurer-affiliated enterprise; the divested facility count is not a market share for health underwriting. The distinction is necessary even when the corporate parent is among the largest insurers. Effects in provider, administration and underwriting markets must be identified on their own evidence.
Empirical evidence and identification
Dafny, Duggan and Ramanarayanan (2012), in Paying a Premium on Your Premium?, use differential local exposure to a national insurer merger to investigate employer-sponsored insurance premiums. Their published abstract reports an estimate that increases in local concentration between 1998 and 2006 left premiums in average markets approximately seven percentage points higher by 2007, and evidence consistent with buyer power in physician labor markets. This result is specific to their period, population and design. The abstract provides the claimed direction, magnitude and identification idea; the full model and proprietary microdata have not been reviewed or replicated for this monograph.
Dafny's earlier Are Health Insurance Markets Competitive? (2010) investigates employer profitability and premium growth in a large-employer panel. The indexed abstract reports that comparatively positive employer profit shocks are associated with later premium increases even within the same plans. This supports the relevance of bargaining and price discrimination in evaluating insurance competition, as reported by the author; it does not establish a merger-specific effect. The two studies together motivate attention to bargaining heterogeneity rather than treating observed premium as a uniform competitive price. Neither result supplies a coefficient transferable to property, life or reinsurance markets without independent validation.
A merger retrospective should identify the outcome variable and treatment independently of subsequent performance. One feasible research design is a difference-in-differences comparison between local markets with relatively high pre-deal overlap and otherwise comparable low-overlap markets. An event-study specification can estimate pre-period differences and subsequent changes relative to an omitted pre-merger reference period. Market fixed effects absorb fixed local factors and time effects absorb common shocks. Coverage changes, claims inflation, regulatory interventions and risk-composition effects remain potential confounders. Researchers should use appropriate clustering at the level of treatment variation and consider spillovers where the same insurance group serves multiple markets.
Overlap measured after closing is endogenous to competitive changes. Pre-deal shares produce a predetermined structural exposure measure but may correlate with different pretrends or target selection. Even an instrument based on differential exposure needs an exclusion restriction: local overlap should affect the outcome through the merger mechanism being estimated, rather than through contemporaneous shocks to the acquired firm's risk segments. Testing pretrends, using placebo merger dates and checking alternative comparison sets assists credibility; no one diagnostic proves the identifying assumptions.
Insurance claims costs and premiums develop over different horizons. A catastrophe in the second year after a merger can alter prices and available coverage even when competitive incentives are unchanged. Controls for geographic hazard, claims severity, labor and materials prices and changes in reinsurance terms must be selected with care. Some variables, such as the firm's post-merger reinsurance strategy, might be mediators through which the merger affects premiums. Controlling for them can remove part of the very effect the analysis seeks to identify. Estimands should distinguish total merger effects from specific mechanisms.
New-business quotes and renewal prices can diverge. An early post-merger effect on pricing to new entrants may coexist with delayed changes to in-force contracts. Policyholder churn may alter loss mix, creating compositional changes in aggregate loss ratios. A balanced panel of unchanged contract types can assist price comparison but excludes consumers denied a comparable renewal. Complementary analyses of offer availability, coverage and exits are required to identify welfare consequences when rationing is part of the mechanism.
Efficiencies, entry and remedies
Claims handling platforms, catastrophe modeling, distribution, systems integration and capital deployment can generate scale benefits. Whether these efficiencies offset a relevant competitive concern depends on magnitude, timing, feasibility, merger specificity and the affected customer population. A projection of consolidated administrative overhead should exclude costs of integration, migration failures, lost staff and duplicate regulatory requirements. A cost saving that does not improve competitive incentives or customer terms may have a different evidentiary significance from an improvement in expected claim service.
Entry must be assessed in the economic market, not merely through the number of licenses issued. A license does not guarantee underwriting appetite, distribution access, reliable reinsurance, actuarial data or adequate risk capital. A potential carrier must offer viable coverage to the relevant buyers on a sufficient scale and within the time relevant to the transaction. Evidence of announced expansion, proven underwriting operations and credible capacity commitments is stronger than an inference from a theoretical absence of legal barriers.
A structural remedy should restore the functional capability that creates rivalry. Transferring contracts may be insufficient if the purchaser lacks distribution, pricing systems, staff, access to data, provider relationships or replacement reinsurance. Portfolio transfer raises prudential questions about administration, continuity and capital. A buyer that already competes in the same market can itself change concentration; the divestiture and acquisition should be assessed together. Behavioral conditions such as access commitments may require monitoring and enforcement over time, with particular attention to incentives to degrade nonprice terms. The specific remedy should address the identified mechanism of harm and its implementation constraints.
Integrated Scientific Synthesis and Future Research
Integrated interpretation of case evidence
The reviewed evidence establishes three distinct layers of judgment. First, market definition identifies which alternatives satisfy the practical needs of buyers or sellers in the relevant geography. Second, structurally measured concentration describes the distribution of activity among independent competitive decisions. Third, the substantive inquiry assesses the removal of actual or potential rivalry, expansion of remaining suppliers, efficiencies, buyer responses and the likely trajectory without the transaction. None of these stages renders the others redundant.
The Aetna and Humana court decision and the Anthem and Cigna appellate outcome demonstrate the economic importance of segmentation by local enrollment or purchaser type. The Aon and Willis Towers Watson summary demonstrates that multinational and specialty brokerage requirements can justify market distinctions even within the same global service industry. The British Aviva and Direct Line determination demonstrates the value of parallel premium and policy-volume shares, qualitative evidence and competitive alternatives. The contrasting Australian judgments show why a target's future autonomous competitive strength matters. The UnitedHealth and Amedisys settlement demonstrates the importance of separating an insurer's corporate identity from the specific service affected by a combination.
These findings are comparative analytical conclusions, not assertions that one set of agencies is systematically stricter than another. To evaluate comparative enforcement tendencies empirically would require standardized transaction characteristics, notification and review selection, exact contemporaneous guidance, case outcomes and a model of procedural differences. The cases were selected for relevance to scientific questions and sufficient public evidence; their distribution of outcomes cannot establish frequency.
The arithmetic examples show that a single national premium share is insufficient to characterize competitive structure. Group aggregation can change the HHI without a change in underlying contracts. Local pooling can mask geographic concentration. An incomplete residual tail generates a range. Published bands impose joint constraints on possible underlying shares, so interval midpoints treated as exact data are inappropriate. A calculated HHI must be accompanied by its definitional metadata. Incomplete evidence warrants bounds or a transparent statement of nonidentification instead of artificial precision.
A reproducible evidence workflow
For an actual transaction, construct a market-by-enterprise observation file with period, product, geography, customer class, source, ownership mapping, measure, amount and observed or imputed status. Retain legal entities and brands in an auxiliary ownership table with effective dates and documented links to underwriting decision makers. Map each policy record only once within an underwriting activity denominator. Treat intermediary and reinsurer roles in distinct tables. Preserve currency and accounting definitions; use consistent conversion periods for cross-country monetary comparisons when necessary.
Before computing concentration, reconcile totals, investigate omitted observations, normalize shares only within a documented population, identify residual mass and test whether residual bounds are feasible. Keep source snapshots and legal decision dates immutable. Compute pre-deal and mechanical post-deal HHI on exactly the same denominator. Calculate the structural change using both direct sums and 2ab, adjusted for any documented divestiture or perimeter change. Calculate CR_4 and a suitable alternative activity measure. Report rounded results only after calculations, retain unrounded source amounts, and document how share intervals constrain inferences.
The next stage assembles evidence of closeness: insurer quote responses, customer switching, procurement records, internal competitor-monitoring documents, licensing and capacity, underwriting restrictions, available distribution and service dimensions. Compare observations across customer and risk segments. Examine the limitations of each source. An internal business document may have stronger evidence of actual competitor monitoring than a statement written in anticipation of merger review, but it may also reflect aspiration rather than operational strategy. Independent third-party evidence and observed decisions can corroborate the interpretation.
A reviewer should challenge the counterfactual independently. Specify what the target would likely offer without acquisition, including expected capital support, business reorientation or alternative ownership. Assess entry and expansion under actual constraints rather than a list of nominal competitors. Examine efficiencies against a realistic alternative, calculate implementation costs and identify benefits to the affected customers. Evaluate each proposed remedy in relation to the original harm and the buyer's ability to compete.
Future research and falsifiability
Three empirical questions follow. The first asks how often group-level concentration differs materially from concentration measured through consumer-facing brands. It can be investigated by linking legal carrier information, ownership data and distribution arrangements to quote and policy datasets. If common-control aggregation does not materially alter many markets, that result should be reported; the direction and magnitude must be determined from data.
The second asks whether mergers involving close rivals produce greater price, coverage or service effects than mergers with identical HHI increments but weaker substitution. A cross-deal design needs consistently measured diversion or bidding overlaps, time-varying risk costs and credible controls for the endogenous selection of partners. A finding that structural measures outperform diversion evidence in certain standardized markets would be informative and should not be excluded by the research design.
The third asks how prudential capacity changes interact with competition over availability. For catastrophe-exposed lines, acquisition by a stronger group might enlarge offered capacity while concentrating control. Observable measures include numbers of quotes, terms of eligible offers, aggregate limits, renewal decisions and claims-paying capacity. A preregistered evaluation should distinguish market-level availability from the financial resilience of individual insurers and examine whether competing groups expand in response.
The contribution of the present research is the integration of these methodological problems with actual comparative cases and transparent numerical examples. It does not originate HHI, the hypothetical monopolist test, causal panel methods or merger theory. Its analytical conclusions are subject to correction by better evidence on the relevant market, legal identity, capacity, or independent counterfactual. Full original insurer microdata, confidential decision schedules, the complete economic papers and some direct legal text versions remain outside this session's verified access. These limits preclude a new cross-country concentration dataset, econometric causal conclusion or claim of independent external scientific review.
Conclusion
Insurance merger control evaluates an acquisition as a change in competitive organization under conditions of contractual risk transfer, regulation and uncertain future liabilities. Relevant market definition must follow available alternatives and purchaser requirements. The competitive enterprise must be identified through control and economic independence, with policy brands, brokerage, underwriting, reinsurance and service procurement kept analytically distinct.
Concentration metrics are reproducible descriptions conditional on those choices. Their interpretation is sensitive to activity measure, geographic pooling, group ownership, omitted enterprises, redaction and time. For a fixed-share merger the HHI increment has an exact algebraic expression; the future HHI and welfare effects depend on customer choices, rival reactions, capacity and remedies. Structural thresholds provide institutionally specific evidentiary screens rather than uniform rules for every insurance market.
The selected decisions establish the value of testing concentration evidence against actual direct rivalry, continuing substitutes and the target's independent prospects. Published economic evidence on employer health insurance shows why the premium and input-market consequences of concentration warrant causal analysis, while the particular estimates cannot simply be exported to unrelated lines or jurisdictions. A credible retrospective requires defined outcomes, comparator markets, risk adjustment and explicit treatment of anticipation, selection and spillovers.
A scientifically adequate merger assessment therefore preserves traceable identities and denominators, supports its legal and empirical assertions with dated sources, distinguishes observation from inference and uses uncertainty bounds where disclosure is incomplete. It treats protection of policyholder promises and preservation of effective rivalry as related institutional responsibilities requiring separate supporting evidence. That is the integrated conclusion of the comparative documentary and mathematical research.
References
United States. 2026. Clayton Act § 7, 15 U.S.C. § 18. U.S. Code, version through Public Law 119-142 (November 10, 2026). Source.
U.S. Department of Justice and Federal Trade Commission. 2023. Merger Guidelines, § 2.1, Guideline 1. Source.
U.S. Department of Justice and Federal Trade Commission. 2023. Merger Guidelines, § 4.3, Market Definition. Source.
U.S. Department of Justice and Federal Trade Commission. 2023. Merger Guidelines, § 4.4, Calculating Market Shares and Concentration. Source.
U.S. Department of Justice and Federal Trade Commission. 2023. 2023 Merger Guidelines (agency landing page). Source.
Council of the European Union. 2004. Council Regulation (EC) No 139/2004 of 20 January 2004 on the Control of Concentrations between Undertakings. Official Journal L 24, January 29, 2004, 1–22. Source.
European Commission. 2004. Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings. Official Journal C 31, February 5, 2004, 5–18. Source.
European Commission. 2026. Review of the Merger Guidelines. Directorate-General for Competition, consultation record. Source.
European Commission. 2024. Commission Notice on the Definition of the Relevant Market for the Purposes of Union Competition Law. C/2024/1645, Official Journal C, February 22, 2024. Source.
Competition and Markets Authority. 2021. Merger Assessment Guidelines. CMA129, March 2021. Source.
Australian Competition and Consumer Commission. 2026. Mergers and Acquisitions. Official merger regime guidance. Source.
International Association of Insurance Supervisors. 2024. Insurance Core Principles and ComFrame. Official standard-setting overview; December 2024 revision. Source.
National Association of Insurance Commissioners. Undated. UCAA Form A: Are You Trying to Acquire or Merge an Insurance Company? Accessed November 19, 2026. Source.
National Association of Insurance Commissioners. 2025. NAIC Releases 2024 Market Share Data. Preliminary March 3, 2025 reporting snapshot. Source.
United States District Court for the District of Columbia. 2017. United States and Plaintiff States v. Aetna Inc. and Humana Inc. Civil Action No. 16-1494 (JDB), Memorandum Opinion, January 23, 2017. Source.
U.S. Department of Justice. 2017. U.S. District Court Blocks Aetna’s Acquisition of Humana. January 23, 2017. Source.
U.S. Department of Justice. 2017. D.C. Circuit Affirms Decision Blocking Anthem’s Acquisition of Cigna. April 28, 2017. Source.
European Commission. 2021. Summary of Commission Decision of 9 July 2021 (Case M.9829 – Aon/Willis Towers Watson), C(2021) 5009. Official Journal C/2024/4363, July 4, 2024. Source.
U.S. Department of Justice. 2021. Attorney General Merrick B. Garland’s Statement on Aon and Willis Towers Watson Decision to Terminate Merger Agreement. July 26, 2021. Source.
Competition and Markets Authority. 2025. Anticipated Acquisition by Aviva plc of Direct Line Insurance Group plc. ME/2239/25, decision July 1, 2025; public text August 7, 2025. Source.
Competition and Markets Authority. 2025. Aviva / Direct Line Merger Inquiry. Public case register, July 1, 2025 clearance. Source.
Australian Competition and Consumer Commission. 2025. IAG’s Proposed Acquisition of RACQ Insurance Not Opposed. May 22, 2025. Source.
Australian Competition and Consumer Commission. 2025. ACCC Opposes IAG’s Proposed Acquisition of RAC WA’s Insurance Arm. December 11, 2025. Source.
Australian Competition and Consumer Commission. 2026. ACCC Opposes IAG’s Proposed Acquisition of RAC Insurance. October 2026. Source.
U.S. Department of Justice. 2025. Court Approves Justice Department’s Settlement in UnitedHealth Group and Amedisys Merger. December 10, 2025. Source.
Dafny, Leemore, Mark Duggan, and Subramaniam Ramanarayanan. 2012. Paying a Premium on Your Premium? Consolidation in the US Health Insurance Industry. American Economic Review 102 (2): 1161–1185. DOI: 10.1257/aer.102.2.1161. Source.
Dafny, Leemore S. 2010. Are Health Insurance Markets Competitive? American Economic Review 100 (4): 1399–1431. DOI: 10.1257/aer.100.4.1399. Source.