Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Cyber Insurance Market Growth Capacity and Competition

Abstract. Commercial cyber insurance has grown into a specialized international market for transferring the financial consequences of digital disruption, data compromise, and related liability. Its development depends on the availability of underwriting capital, the contractual definition of insurable losses, the organization of distribution and reinsurance, and the capacity of buyers to identify and finance residual risk. This monograph examines those relationships through a comparative analysis of regulatory reporting, insurance placement evidence, institutional documents, and scholarly research. The study separates direct written premium, policy counts, renewal rates, limits, claims experience, concentration, and capital at risk. It combines international evidence with a reproducible analysis of United States reporting for 2024–2025, British business survey results, regional renewal indicators, and the insurance industry's exposure to correlated events. The principal conclusion is that premium growth and expansion of effective risk transfer are different phenomena. Competition may broaden coverage and reduce prices while worsening the profitability of particular portfolios; expanding policy counts can reflect small endorsements rather than substantive new protection. The resilience of market supply depends on capital and contract structures that can withstand severe correlated events without abrupt withdrawal. The work presents a descriptive and analytical synthesis, with explicit scope and data limitations, rather than a global causal estimate or new named theory.

Keywords: Cyber insurance; insurance markets; underwriting capacity; premium growth; concentration; market competition; reinsurance; insurance-linked securities; underwriting cycles; protection gap; accumulation risk; insurance economics.

Introduction

The research problem

Cyber insurance is simultaneously a contract market, a risk-bearing market, and a market for specialized services. The contract specifies the events and losses that may give rise to a payment; the insurer allocates capital and arranges reinsurance to support that promise; the intermediary and its service partners participate in distribution, risk appraisal, and recovery. A meaningful account of growth therefore requires an understanding of the economic substance of the promise, its quantity, and the institutional organization through which it is financed.

A larger premium total can result from a rise in price, a wider set of insured businesses, increased limits, an altered mix of exposures, or an accounting change. Conversely, more insurance may be purchased for less premium if renewal prices fall. Competition that expands the number of insurers offering a quote can coexist with dependence on a narrow set of capital providers and cloud infrastructure. These distinctions establish the object of this study: the relationship between market development, effective risk-transfer capacity, and competitive behavior, with particular attention to the underwriting cycle.

The analysis asks how market growth can be measured in a manner consistent with coverage, what determines the supply of credible capacity through the cycle, and how concentration can be interpreted without assuming that visible insurer names correspond to independent sources of risk-bearing capital. It also addresses the implications of apparently abundant capacity for buyers who face restrictions because of their industry, size, or technological dependencies.

Research design

The evidence cutoff is November 18, 2026. The inquiry uses a targeted institutional and scholarly literature review, comparative document analysis, descriptive calculations, and explicitly hypothetical stress illustrations. The principal quantitative source is the 2026 market report (NAIC 2026), specifically its U.S. direct business and domestic reporting series. International supervisory evidence and private placement evidence provide additional perspectives, with separate populations and dates.

No insurer-level confidential exposure inventory, individual policy schedule, transaction database, global firm census, or reinsurer treaty book was obtained. The research therefore does not estimate international causal effects, a universal penetration rate, or an exact global concentration index. Academic results cited from accessible abstracts are limited to their documented questions, methods, or stated findings. Derived statistics below are calculated from published aggregates. Illustrative contracts and scenarios are labeled as such and are not actual quotations or forecasts.

The study distinguishes market observation from interpretation. Where a source reports an association, the analysis examines possible mechanisms rather than converting it into a causal claim. Where an authority summarizes another study, the result is treated as a reported estimate in that authority's document rather than as independently replicated primary evidence. The comparison privileges common units and reporting scopes over the appearance of longer time series.

Conceptual and Historical Foundations

The insured event and the insurance product

An adverse digital event is an occurrence arising from a failure of availability, confidentiality, or integrity, including malicious activity and nonmalicious interruption. Its financial consequences may include restoration, business interruption, investigation, notification, defense, settlement, and other costs. These consequences enter the insured loss distribution only when the relevant policy's insuring agreement, conditions, exclusions, definitions, and limits permit recovery. One event can cause several losses and claims, and a single claim can contain several categories of cost.

First-party protection addresses losses of the insured itself. Third-party liability responds to qualifying obligations toward others. Some products combine both, alongside access to incident-response providers. The coverage may be standalone or incorporated into another policy. Currat, Perry, and Yong (2026) examine these distinctions in Cyber Insurance Unpacked. The existence of a digital dependency is an exposure characteristic; it is not by itself a determination that any policy pays following a failure.

For a simple illustrative occurrence-based layer, let economic loss be X, retention d, and limit u. Assuming all loss components are covered, indemnity is min(max(X − d, 0), u). Real contracts also may involve waiting periods, aggregate limits, sublimits, coinsurance, reporting requirements, or other layers. Therefore, even identical nominal limits can offer different economic protection. A policy count aggregates units of contract administration rather than homogeneous units of risk transfer.

The market considered here includes dedicated cyber insurance and identified affirmative cyber endorsements where the source includes them. Technology errors and omissions coverage, crime insurance, property insurance, and professional liability are relevant where exposures overlap, but are not automatically part of the same statistical aggregate. Silent cyber is a potential source of unmeasured liability, distinct from intentionally priced cyber business. This boundary is necessary for comparing premium series and interpreting reported underwriting results.

Historical development and market formation

The Bank of England's UK protection-gap analysis (2026) traces forms of cover to the 1990s. Early product development was linked to the changing allocation of information-system risk between businesses, service providers, and their insurers. Dedicated products subsequently provided an opportunity to identify and price cyber exposure more explicitly. A specialist underwriting discipline developed because digital incidents combine uncertain frequency, complex causation, heterogeneous firms, and rapidly changing vulnerabilities.

Insurance expansion follows the development of the underlying digital economy but cannot simply be projected from digital spending. A company can increase its technological dependencies while enhancing security; a policy can shift losses from one coverage category to another; technologies can make some events more frequent and other events cheaper to contain. The economic significance of digitalization is that it creates new contingent demands on organizational cash flow, rather than a fixed proportional increase in all insured loss frequencies.

The GAO market study (2021) gives a historical cross-section of the tightening phase. Its observed broker client uptake went from 26% in 2016 to 47% in 2020, and respondents described constrained limits for certain industries alongside rising prices. That evidence records a genuine expansion of demand and pressure on supply for a specified population, but it cannot establish national penetration. The interaction of price increases and limits tightening is especially informative: additional premium did not necessarily imply additional protection for every buyer.

The theoretical significance of this period is that supply and demand can move concurrently. Increased awareness and lender or contractual requirements can shift demand outward at the same time as adverse claims shift capital supply inward. Observed prices then increase while the net effect on coverage sold remains ambiguous. A later soft market does not reverse the historical facts; it changes the contemporaneous balance of willingness to pay, available capital, and underwriting criteria.

Insurability and information

Biener, Eling, and Wirfs (2015), in their empirical insurability analysis, study 994 events drawn from an operational-loss database. Their abstract identifies dependence among losses, information asymmetry, and deficient data as impediments to insurability. The evidence predates much of the modern market and does not measure current pricing. It remains relevant to the structural problem of converting an economic exposure into a sufficiently specified risk distribution.

Insurability is economically conditional. An insurer can offer a contract with a low limit and restrictive wording even when it cannot finance unrestricted catastrophic loss. Some risks may be transferable at one retention and price but unavailable at another. Treating each exposure as categorically insurable or uninsurable erases this contract dependence. For the market researcher, the relevant observation is the set of terms offered and accepted, not simply whether a product is advertised.

A buyer has information about internal controls, operational weaknesses, and loss history that the insurer may observe only imperfectly. Conversely, a specialist insurer may possess experience of threat patterns and claims that a small business does not possess. The underwriting relationship can improve information on both sides. Its economic cost includes the time needed to disclose, validate, and interpret exposure attributes. A lengthy application can reduce misclassification but also create a barrier to buying otherwise suitable coverage.

Information changes across the policy period. An accurate assessment at inception may become outdated following a software migration, acquisition, or change of supplier. An annual snapshot provides an incomplete description of time-varying exposure. Continuous monitoring may improve assessment, but it also raises questions of permission, data quality, interpretation, and whether a temporary signal justifies altering terms. Competition in information and service quality is consequently as important to the formation of a specialized market as the competition in quoted premiums.

Primary Documents and Market Structure

Statistical perimeter and unit of analysis

Direct written premium represents business written directly by insurers during a specified period, subject to the reporting conventions of the source. Gross written premium may include assumed reinsurance depending on the reporting perimeter. Net premium subtracts ceded premium under applicable accounting definitions. Earned premium spreads the recognition of relevant exposure over the period of coverage. An accepted renewal price refers to an individual or sampled transaction. Each is appropriate for a different question.

A premium series can identify the size of a reporting market. Policy counts indicate administrative exposure units. Claims counts indicate reported claim activity, not the aggregate count of cyber incidents. Frequency requires a denominator consistent with the population and period of claims. Loss ratios relate claims or incurred losses to the relevant premium base under stated definitions. Premium cannot be used as a substitute for a limit denominator, nor can claims counts be divided by the number of firms in an unrelated business survey.

Define the market's premium total as the sum of nonoverlapping direct premiums. The number of policies in force at a given date is a stock, whereas annual written premium is a flow. Their ratio is a descriptive premium-to-policy-count indicator only if the reporting populations align. It is not an average renewal quotation. Multiple policies can cover one enterprise, an endorsement can correspond to only a small cyber sublimit, and a policy can cover multiple entities. The counts thus require contract-form information before they can be interpreted as market reach.

Where direct and assumed premiums are combined across organizations, the same obligation may be represented in multiple forms. A reinsurer accepting part of a primary portfolio is providing risk-bearing capacity rather than an additional primary policy to the original buyer. Similarly, a broker placement and an insurer filing may describe the same contract. Cross-source aggregation should avoid treating each reported transaction as new market demand.

Primary distribution and delegated underwriting

Insurers, managing general agents, brokers, managing agents, and reinsurers may perform separate functions. The underwriting entity sets terms and selects or accepts risk; delegated intermediaries may be authorized to quote or bind within a defined authority. The legal carrier remains responsible for the contractual obligations it assumes. Reinsurance can reduce the carrier's net retained losses subject to treaty scope and counterparty performance.

A carrier count measures one dimension of entry. The count of distribution brands measures another. Several intermediaries may share the same insurer paper, and several legal insurers may ultimately depend on common parent capital. The inverse situation is also possible: a major specialist can distribute products across many jurisdictions and still provide distinct independently financed layers through different carriers. These possibilities explain why the measurement of competition requires identification of both underwriting control and the bearer of loss.

Direct distribution may be attractive for standardized exposures. Specialist brokers can help assemble multilayer programs for complex accounts, compare wordings, and negotiate with several risk carriers. The broker provides information and coordination, but the negotiation process also creates search costs and potential agency conflicts. A buyer's access to a distribution platform is therefore different from the existence of credible substitute insurance products.

In a layered placement, one insurer can write the initial tranche after retention, another insurer the next tranche, and still others participate proportionally within a tranche. A primary insurer's term may influence excess wording through follow-form provisions, but coverage need not be identical at all layers. Competition for an excess dollar and competition to lead a complex program may involve different sets of firms. These relationships are relevant both for risk management and for the definition of the economic market.

Admitted and surplus markets

The NAIC financial tables (2026) distinguish domestic admitted business from domestic surplus lines and alien surplus lines. This is an important statistical distinction because the latter participates in the U.S. market without appearing in the domestic-only direct premium total. Market-level comparisons require consistent inclusion of alien business.

For analytical purposes, admitted and surplus channels illustrate institutional differentiation. Authorized local carrier offerings and surplus channels may serve different risk profiles and may operate under different state-specific licensing, placement, rate, form, and guaranty arrangements. This monograph does not claim a uniform legal rule across U.S. states. A surplus placement should be understood as a distinct legal and distribution channel, not as evidence of lower quality or greater capital by definition.

Its economic function often consists in serving accounts requiring unusual terms, specialist underwriting, or additional limits. The price differential between channels may reflect different risk characteristics, cost structures, underwriting restrictions, or coverage terms. A shift toward surplus business cannot be interpreted as an isolated effect of channel regulation without data on the accounts moving and the contracts they acquired.

Limits, exclusions, and visible market supply

Capacity becomes economically meaningful at the level of the peril, the insured entity, the layer, and the period. A line of ten million dollars on an eligible risk does not provide the same protection as ten million dollars of affirmative protection against business interruption at a critical supplier. A single number covering both may disguise the difference.

The market's organization affects the range of feasible wordings. Nieuwesteeg, Visscher, and de Waard (2018) report in their Dutch contract case study a tension between attracting buyers through accessible products and restraining correlated exposures. Their case is historical and local. Its enduring interpretive importance concerns the endogeneity of contract scope: market expansion can affect what insurers are willing to cover, and coverage design affects subsequent expansion.

The buyer's actual insurance product is therefore a bundle: covered events, limits, deductibles, waiting periods, legal conditions, response services, and insurer credit quality. An economically consistent comparison of quotes holds those attributes constant or adjusts explicitly for their differences. A market can achieve lower premiums through less comprehensive coverage, even with no reduction in the price of equivalent protection. Equally, lower premiums with wider coverage represent a greater change in effective price than the headline number suggests.

Capacity Cycles and Underwriting Economics

Determinants of risk-bearing supply

Insurance capacity is the amount of exposure an insurer or group of insurers can and will assume on specified terms. It is constrained by equity, regulatory requirements, internal risk tolerances, reinsurance, liquidity, counterparty quality, estimated loss distributions, and the insurer's competing uses of capital. The maximum theoretical limit supported by capital and the amount the market chooses to offer are distinct.

A simplified premium decomposition is expected covered loss plus expenses plus the cost of risk capital and reinsurance, adjusted for investment effects and target return. This is an accounting-economic diagnostic, not a pricing formula proposed as a new model. Expected loss refers to loss within contracted terms. The capital charge reflects the cost of supporting uncertain outcomes, while reinsurance can reduce the risk that must be retained. The allocation of common expenses and capital is influenced by portfolio composition.

Competition can reduce the return component or the price charged for a perceived exposure, but it does not remove the underlying losses. An insurer may choose to accept lower margins because available reinsurance becomes cheaper, expected loss decreases after changes in controls, or the insurer values access to new business. Alternatively, the insurer may accept a price below the cost suggested by its eventual loss experience because the projected distribution was misspecified. Without transaction information, the mechanism cannot be identified from premium movement alone.

A newly financed underwriting program may enter without a large historical book of late-developing claims. An established insurer may have better experience and data but also legacy exposure. Expansion intentions consequently depend on historical reserves and future expected returns as well as current capital. An apparent advantage of a new participant can disappear when its own book matures.

Frequency, severity, and the changing distribution

Cyber risk comprises several types of loss with different causes and financial characteristics. A firm whose operations are temporarily disrupted may incur restoration and revenue losses without a third-party liability claim. Improper disclosure of personal data can lead to litigation and investigations even where operational disruption is minor. Recovery expenses and legal costs develop over different horizons.

Malavasi et al. (2022) describe in their frequency and severity study regression-based analyses using a U.S. cyber-event dataset and certainty-equivalence methods to investigate insurance viability. The accessible abstract establishes the research method, not a universal distribution or numerical coefficient. For market analysis, the importance of heterogeneity is that a portfolio's average loss may change when the mix of risks changes even if risk within every segment remains constant.

Consider two illustrative segments, one with expected covered loss of USD 1,000 per policy and another with USD 10,000. A portfolio holding 900 policies of the first type and 100 of the second has mean loss of USD 1,900 per policy. If its mix becomes 500 of each, the mean rises to USD 5,500 without any deterioration in segment-specific risk. Changes in premium and loss ratios may therefore reflect business mix as well as underwriting or pricing performance.

Loss data also involve reporting and development. A recently reported incident may have a small initial reserve yet ultimately produce significant defense or interruption expense. Mature calendar years can receive reserve revisions from older policy periods. In an expanding book, current premium may relate to a different set of exposures from the claims included in the calendar-year numerator. A loss-ratio improvement can reflect reserve releases or changed business mix; a deterioration can reflect development of old claims rather than a rise in new incident frequency. Accident-year and underwriting-year analysis may help where the corresponding records are available.

The correlation constraint

Adding independent risks can make total losses proportionally more predictable. Correlation limits this diversification benefit. Suppose 1,000 identical policies each have an annual covered loss of USD 100,000 with probability 0.01, otherwise zero. The expected aggregate loss is USD 1 million. Under independent outcomes, the aggregate variance is 1,000 × 0.01 × 0.99 × (100,000)^2, giving a standard deviation near USD 315,000.

Now suppose instead that all 1,000 covered losses are triggered by the same binary event with probability 0.01. The aggregate is either zero or USD 100 million; expected loss remains USD 1 million, but standard deviation rises to about USD 9.95 million. These are limiting illustrations with identical marginal probabilities, not calibrated estimates of cyber loss. The distinction identifies why capacity must be evaluated at the portfolio level and not simply as a multiple of the expected losses of each account.

Software libraries, cloud infrastructure, managed services, supply chains, and telecommunications services can act as common dependencies. The FSI–IAIS accumulation analysis (Currat, Perry, and Yong 2026, pp. 18–21) examines these channels. Diversification across countries and industries can be less effective when organizations depend on the same underlying service. Conversely, diversification across technical configurations and response capabilities may reduce some correlations even within a geographic market.

A carrier may establish maximum aggregate exposure to an identified dependency, then decline otherwise acceptable risks once the concentration reaches its allocation. This produces a portfolio-level supply constraint: the buyer's own risk has not worsened, but its marginal attractiveness to the insurer has changed. An insurer with a complementary portfolio might still offer protection. The economic value of competition includes the availability of differentiated portfolios, not simply more bidders with the same dependency concentrations.

Reinsurance and alternative capital

Quota share reinsurance can enable a carrier to expand business by ceding a share of premium and losses. Excess-of-loss arrangements provide protection above specified attachments, subject to event and aggregate definitions. Facultative placements can support individual complex risks. The capacity benefit depends on actual contract scope and the economic reliability of the indemnity from the reinsurer.

A direct insurer may write USD 20 million of gross limit and retain a smaller net amount. If a treaty excludes a systemic event covered by the primary policy, however, its net loss for that event can be much larger than the ordinary retention suggests. This is why contract consistency matters to both capital supply and competition. The existence of reinsurance does not demonstrate equivalent protection across all scenarios.

Capital-market instruments can extend the set of parties accepting catastrophe risk. The 2026 FIO annual report estimates USD 1.24 billion of cyber catastrophe bond capacity at the end of 2025 and records USD 450 million of issuance in Q4. Those are capital-market measures and cannot be added to primary written premium to determine the size of the insurance market. The triggers, collateral arrangements, exhaustion conditions, duration, and basis risk determine the value of this capacity to the sponsor.

Consider a hypothetical insurer with USD 80 million of eligible losses after primary policy terms and USD 40 million of recoverable reinsurance. Net underwriting loss is USD 40 million prior to expenses. If the reinsurance does not apply to the event, the net amount becomes USD 80 million. A separate bond paying USD 20 million upon a qualifying external trigger would reduce net loss when the trigger is met but provide nothing when the trigger fails. A gross limit shared by several customers is valuable only to the extent that these financial protections actually respond and remain available.

A complementary financing hypothesis is developed in the cyber insurance supply seminar (Eling 2024). The publicly accessible research abstract describes asymmetric information and uncertainty as barriers to external capital and reports a U.S. identification strategy using the 2017 change in the tax treatment of affiliated foreign reinsurance. This supports treating access to external risk capital as an empirical research question, rather than assuming it follows mechanically from the availability of industry capital. The abstract alone does not establish the validity of the identification assumptions, the magnitude of the effect, or its applicability outside the studied market.

The underwriting cycle

Hard and soft conditions refer to changes in insurance prices, terms, and availability. They arise from the interaction of historical claims, expected loss, reserve development, the cost of capital and reinsurance, entry and exit, and buyers' demand. A hard market may improve prospective underwriting terms yet reduce buyers' access. A soft market may expand access but compress margins if costs grow faster than prices.

The Q3 2026 market overview (Aon 2026b) reports substantial available cyber capacity and continuing competitive pricing alongside concern about interruption, ransomware, prior-year development, and systemic exposure. This evidence is compatible with the possibility that supply is still abundant despite risk concerns. The observation does not establish a subsequent turning point or justify forecasting the direction of the next quarter.

A cycle can differ across account segments. A profitable small-business portfolio may continue to receive competitive terms when large complex risks face narrower offers. Timing may also differ across primary and excess coverage because their exposure to ordinary claims and catastrophe losses differs. A common headline rate index can conceal such divergence. Analysis of the cycle should therefore track terms and renewal outcomes by layer, sector, organization size, controls, and distribution channel.

Underwriting discipline means maintaining adequate prices and coverage boundaries relative to the risks accepted. It remains compatible with lower prices when expected loss, acquisition expense, or capital cost falls. Maintaining discipline during competition requires distinguishing justified improvements in pricing from aggressive expansion under optimistic assumptions. The reported loss ratio is one diagnostic input, not a sufficient test of capital adequacy.

Comparative Measurement of Growth and Competition

United States premium and policy count

The following table uses rounded USD billions for 2024 and 2025 within the stated reporting scope. The source is the 2026 NAIC report (figures 1–3, 6–7). The changes below are calculated by the author from the published amounts before rounding.

Reporting series 2024 2025 Change
U.S. direct written premium including alien surplus lines 9.144 9.464 +3.50%
Domestic direct written premium 7.083 7.459 +5.30%
Domestic admitted direct premium 2.737 2.659 −2.86%
Domestic surplus lines direct premium 4.346 4.800 +10.45%
Implied alien surplus lines direct premium 2.061 2.005 −2.69%
Domestic policies in force (millions) 4.369 5.843 +33.75%

The implied alien amount is calculated by subtracting domestic direct business from the inclusive total in each year. It is a residual of compatible reported series rather than a separately audited foreign-carrier aggregate. All money values are nominal U.S. dollars. The one-year growth rate is (new value divided by old value minus one) multiplied by 100.

Domestic surplus lines contributed about USD 454 million of additional premium, whereas admitted business contracted by approximately USD 78 million. Together these changes yield approximately USD 376 million of domestic growth. Surplus lines thus contributed more than the net domestic increase, with admitted decline offsetting some of that contribution. The residual alien series also contracted slightly, reducing the inclusive growth amount to about USD 320 million. Growth was institutionally uneven.

Dividing domestic direct premium by domestic policies in force gives approximately USD 1,621 in 2024 and USD 1,276 in 2025, a reduction of about 21.3%. This ratio is neither the mean quotation for a fixed-coverage product nor a unit cost per enterprise. Its numerator is annual written premium and its denominator is reported policies in force; composition, inception timing, and multiple placements matter. It demonstrates how strongly interpretations differ when one uses policy counts instead of premium as the measure of expansion.

Coverage mix and effective protection

The source's rounded 2025 product mix assigns endorsements 62.8% of policies but just 3.9% of premium, primary policies 35.9% of policy count and 63.7% of premium, and excess policies 1.4% and 32.5% respectively. Small discrepancies from 100% result from rounding. These are aggregate reporting categories rather than matched policy-level observations.

For product category j, the quotient of its premium share and its policy-count share equals its premium per policy count relative to the whole-market ratio, subject to aligned populations. The quotient is approximately 0.062 for endorsements, 1.77 for primary, and 23.21 for excess. The high excess quotient reflects the economically distinct nature of that book; it is not evidence that any excess product has a unit rate hundreds of times larger than an endorsement. The quotient of primary and endorsement indicators is roughly 28.6.

This calculation has a substantive implication. A rapid expansion in low-premium endorsements can push total policy count upward without a proportional rise in the overall premium series or in the amount of loss transferred. A broader reach of modest coverage can be socially valuable, especially if additional firms gain response services and protection against common incidents. Its value depends on the limits and exclusions rather than its numerical contribution to an administrative count.

A useful follow-up would collect the distribution of per-risk limits, annual aggregate limits, maximum single-event payout, and covered dependent business interruption. If an insurer writes twice as many endorsements with unchanged small sublimits, the number of businesses able to receive some payment may rise substantially while catastrophic coverage expands very little. A separate rise in excess limits could enhance protection for a small number of large enterprises without materially changing overall policy counts. Both changes are legitimate forms of market development.

Concentration from reported group shares

Market concentration measures the distribution of activity across firms, not the concentration of cyber exposures across technologies. The 2023 U.S. Merger Guidelines on concentration define the Herfindahl–Hirschman Index (HHI) as the sum of squared percentage shares. Their structural presumptions also require an appropriately defined relevant market and considerations such as changes in concentration. This study uses HHI descriptively only.

The 2025 domestic top-20 group shares reported in the NAIC table are 8.44, 7.65, 7.29, 4.58, 4.28, 4.05, 3.97, 3.81, 3.60, 3.36, 3.10, 3.10, 2.85, 2.74, 2.69, 2.66, 2.32, 2.23, 2.16, and 1.94 percent. The reported cumulative ratios are CR3 = 23.38%, CR5 = 32.24%, CR10 = 51.03%, and CR20 = 76.82%. They describe aggregate domestic direct premium by reporting group and exclude a separately measured alien segment. Group share rounding is retained throughout the calculation.

Summing the squares of the twenty reported shares yields approximately 361.1 HHI points. The omitted 23.18% is dispersed among smaller groups. Assuming the table is complete for the leading twenty and the ranking is by the same measure, every omitted share is at most about 1.94%. Since the square of a nonnegative share s is no greater than 1.94 × s under this bound, the missing contribution is at most about 45.0 points. The full HHI on this reporting perimeter is therefore bounded approximately between 361 and 406 points, subject to rounding and the ranking assumption. The lower endpoint ignores the positive contribution of omitted firms and is a conservative lower bound, not an attainable exact market estimate.

This approach avoids the common mistake of treating the missing market share as one firm, which would artificially raise measured concentration. An exact HHI requires all shares, and an HHI suitable for competition enforcement requires a defensible relevant product and geographic market. If some nominally distinct entities are controlled by the same economic decision maker, their combined shares are the relevant unit for that question.

Relevant markets and realized choice

The DOJ and FTC's market definition guidance (2023b) provides a basis for distinguishing economically substitutable products and geographic trading areas. Cyber insurance involves significant variation in limits, services, exclusions, size of eligible risks, and capacity. An endorsement attached to a small commercial policy may not constrain the price of a complex multinational primary placement. The underlying direct premium category is a broad statistical aggregate.

The number of insurers participating in a layered program can exceed the number of credible primary leaders. Equally, several brands may share one capital provider. A buyer's alternative quote set should be evaluated through the number of genuinely eligible carriers, willingness to provide a binding limit, equivalence of contract terms, and capacity durability. A seller that advertises a product but declines the account is not a realized alternative.

Aon (2026a), using its own 2025 placement experience, records more than 90 participants in North American cyber placements, with around 80% of that placement premium assigned to the top twenty. The approximate similarity to the domestic aggregate CR20 does not warrant combining the two datasets. The broker chooses a subset of accounts and carriers, whereas the regulatory aggregate covers a different perimeter. Together, the observations are consistent with participation by numerous carriers and substantially uneven market shares; neither establishes the range of adequate offers for a particular account.

For competition analysis, the useful next measurements include bid participation, refusal and withdrawal rates, quote-to-bind conversion, differences between lead and follow capacity, premium per equivalent unit of coverage, and switching costs. A panel of risk profiles followed through renewal would reveal if more bidders actually increase the quantity and quality of offers after accounting for changes in exposure. Public premium rankings cannot provide this evidence alone.

International comparisons

The Global Insurance Market Report 2025 gives 2024 global size estimates of USD 15.3 billion from Munich Re and USD 14.7 billion from Swiss Re. Their then-current 2025 projections were USD 16.3 billion and USD 15.6 billion respectively. These were contemporaneous estimates and projections, not a final audited global total. Differences can arise from coverage definitions, timing, source populations, and treatment of assumed or embedded business. The global figure should remain an estimated range conditioned on the reporting method. The later NAIC global market account (2026, p. 2), citing AM Best's July 2026 market outlook, places 2025 global premium above USD 16 billion. This is a reported third-party estimate; it is not a new NAIC global census. The NAIC narrative links deceleration to a U.S. premium decline, while its own like-for-like data show renewed U.S. growth in 2025 following the 2024 contraction. The numerical statutory series therefore governs the domestic year-on-year comparison.

The IAIS reports that more than nine-tenths of affirmative cyber premium in its individual-insurer monitoring data was written or assumed by groups headquartered in Europe or North America. Headquarters geography is different from the location of the insured risk; the data may include assumed business. This observation supports an analysis of the geographic organization of risk-bearing institutions, not a claim that nine-tenths of insured companies are located in those regions.

Regional rate movements provide a separate view of competitive conditions. Marsh's Q2 2026 IMEA data report cyber renewal reductions of 14% across its India, Middle East, and Africa portfolio, with local results differing sharply. Its observations include new market participation in the Middle East. The rates arise from transactions in the broker's client portfolio rather than a representative universe of businesses, and they refer to rates rather than direct premium market size.

These sources support the existence of regional differentiations in the market. They cannot determine a common worldwide share of organizations insured or relative actuarial price adequacy across countries. Such comparisons require consistent policy forms, exposure quantities, currency conversion, inflation conventions, and geographic allocation bases. A global premium estimate should not be divided by an unrelated estimate of global economic cyber losses to measure the share of those losses reimbursed, because premium is not indemnity.

Survey evidence and unmet demand

The official 2025/2026 British breaches survey (DSIT and Home Office 2026, section 3.3) separates reports of specific cyber insurance from cover incorporated into wider policies. The weighted results by firm size are shown below; split-sample base counts are unweighted and are included to clarify the sample. The survey is not a count of contracts or claims.

Firm size Dedicated cyber policy Cyber cover in wider policy Survey base
Micro 8% 37% 580
Small 15% 40% 241
Medium 24% 36% 154
Large 32% 22% 76
All businesses 10% 37% 1,051

The dedicated-policy share for large firms is four times the microbusiness share in the estimate. The survey's small large-firm base and potentially different mix of exposures counsel against treating that ratio as an exact population parameter or as evidence that firm size causes purchasing behavior. Broader cover and dedicated cover differ in terms and usefulness; summing categories without checking the question design and overlap would be inappropriate.

Among 267 surveyed businesses without such a policy, 39% indicated lack of awareness, 34% that it was not a budget priority, and 19% that it was too expensive. Responses need not be mutually exclusive. These results are informative because demand can remain weak even when capacity is substantial. Lower premiums alone may fail to address problems of information, prioritization, or relevance to the buyer's business.

The study draws an important distinction between the observable absence of insurance and the underlying protection gap. A business may have consciously chosen retention because its exposure is small or it can finance the loss. Another may remain uninsured because it cannot afford available coverage. A third may hold a policy with an inadequate limit. A survey of ownership does not identify the counterfactual value of adequate insurance for each firm.

Integrated Scientific Synthesis and Research Agenda

The relation between demand growth and supply growth

The evidence establishes the importance of studying market development jointly through premium and exposure measures. The U.S. growth calculations show that the change in the number of policies can differ substantially from the change in direct written premium and that institutional substitution among distribution channels affects the total. The product-mix analysis demonstrates that equal units of policy count need not correspond to equal units of protection. These are observed relationships in a specific dataset, not evidence of an invariant international law.

In the illustrative premium accounting identity, total premium is the sum of premiums across policies. In a single homogeneous class, the total can be expressed as policy count times premium per policy. Once classes differ in limits, risks, attachment points, and reporting, the relevant decomposition includes class-specific exposure and unit premium. If a firm buys more coverage while the rate per equivalent unit falls, these two movements can offset each other.

For comparison, consider a homogeneous hypothetical portfolio with 1,000 policies each carrying one unit of limit at a price of 100. In the next year 1,200 policies carry 1.1 units each at a price per unit of 85. Premium grows from 100,000 to 112,200, or 12.2%, while nominal insured limits grow 32% and the equivalent unit rate falls 15%. Alternatively, if all additional policies provide only one-tenth of a unit, the change in total protection would be quite different even with a similar count increase. The examples separate algebraic quantities that empirical datasets must distinguish.

Demand expansion increases the set of exposures that insurers may choose to cover. Supply growth increases the set of financially credible promises they are willing to make. The resulting realized market requires an intersection between the two, mediated by price, information, and terms. A market with latent demand but restricted underwriting will exhibit low realized transactions. A market with ample capacity but insufficient buyer awareness can also exhibit low realized transactions. Policy recommendations based on premium alone risk treating these different situations as one.

Competitive capacity and correlated security exposure

Competition can broaden supply through new underwriting expertise, more efficient distribution, differentiated portfolios, improved services, and access to new capital. Each source of expansion has a different durability. An intermediary can expand distribution quickly without creating a new risk-bearing balance sheet. An additional investment can increase loss-absorbing resources without immediately developing underwriting capability. A reinsurer can support several intermediaries or carriers while simultaneously increasing the aggregate dependence of primary supply on one institution.

Thus the efficient market researcher examines two layers: the market in contracts and the network of risk-bearing resources that supports contracts. This is an ordinary analytical distinction derived from existing insurance economics, not a proposed named framework. The first layer determines the conditions under which buyers obtain a promise. The second determines how the promise withstands a loss event. Both should enter prudential and competition analysis.

An illustrative market may contain ten primary carriers each with USD 50 million of independently retained capacity. If they share one reinsurance facility for a further USD 500 million, aggregate recoverability in a systemic event depends on that facility's aggregate terms, not merely on the sum of the ten marketed limits. If the facility pays a maximum of USD 300 million across all carriers for that event, the combined credible total under those simplified assumptions is USD 800 million rather than USD 1 billion. Counterparty performance and liquidity would require further analysis.

Evidence agenda and identification strategy

A future empirical program should build a consistent annual panel of direct cyber insurance premium, earned premium, written limits, policy counts, counts of eligible enterprises, claims and ultimate incurred losses. The unit would be a policy-year and, where possible, a stable enterprise identifier enabling multiple contracts to be recognized. Product categories would distinguish endorsements, primary standalone layers, excess layers, and relevant overlaps. Reporting would preserve the source's jurisdiction, currency, contract basis, period, and treatment of alien and assumed business.

A renewal panel would contain the expiring and renewed limits, retentions and terms, a fixed measure of exposure, quoted and accepted prices, available carriers, and changes in the underlying technical risk. Matching or fixed effects could help separate buyer-specific variation from general market movements. A credible causal study of competition would also require variation in available alternatives that is not merely a response to the same changes in risk. Entry, exit, or financing events might be considered, but each potential identification event would require evaluation of anticipation, selection, concurrent shocks, and the exclusion restriction before inference.

For prudential work, the portfolio should be evaluated under common but varied scenarios involving service interruption, compromised software, widespread liability, and cross-line exposure. Each scenario would specify the affected population, propagation, downtime, mitigation, policy response, reinstatement, reinsurance recovery, and cash-flow timing. A range of severity calculations is more informative than apparent precision in a single model output. Estimates should be tested against historical incidents where applicable while acknowledging that rare, evolving events may not be represented in historical data.

For protection-gap research, the denominator should be stated before the indicator is calculated. A count-based gap might use the share of exposed enterprises without suitable protection. A loss-based gap would compare modeled or observed economic losses with recoverable insurance indemnity for the same population, period, and event definitions. A limit-adequacy gap would compare available coverage to a justified target or recovery need, with assumptions disclosed. These measures quantify different aspects of unmet need and cannot be converted into each other without additional evidence.

Finally, competition research should collect common ownership, underwriting delegation, binder capacity, use of common reinsurers, quote substitution, and evidence about the distribution of independent capital. The analysis would separately assess competition for the customer-facing product and the financial network's vulnerability to common loss. This agenda supports further empirical work without asserting that existing aggregate evidence already establishes causal effects.

Reproducibility, counterevidence, and interpretive bounds

The numerical calculations in this monograph use explicitly stated published quantities: the two premium totals, their domestic components, the two policy-count values, the product-mix shares, and the ranked group shares. Growth rates use ratios of unrounded reported values. The HHI bound assumes the residual market consists of groups ranked below the reported top twenty. Rounding can move the approximate endpoints slightly. The illustrative examples use independently specified values and can be replicated with ordinary arithmetic.

An alternative interpretation of the observed premium-to-policy-count ratio is that recording practices changed or the composition of the reporting population shifted. No policy-level evidence was obtained to decompose the change into those mechanisms. Another interpretation of price declines is improved expected loss or increased underwriting efficiency rather than greater tolerance for adverse results. These are empirically open possibilities. The evidence does not establish an imminent capacity reversal or a causal effect of competition on profitability.

International sources present inconsistent estimates, and estimates or projections for 2025 should be identified by their date and vintage. A global count of written policies for comparable coverage was not found in the reviewed material. The evidence base is strongest in U.S. statutory reporting and selected European and broker observations. In the British survey, self-reporting may differ from contractual reality. Academic abstracts offer support for the stated literature context, but do not replace detailed replication or an independent systematic review.

The scientific contribution is a comparative synthesis, the derivation of transparent quantitative diagnostics from the 2026 NAIC reporting, a bounded concentration calculation, and the economic interpretation of the interaction of competition and capital with coverage composition. No priority or novel theory is claimed. The conclusions are limited to the evidence and scope described above.

Conclusion

The growth of cyber insurance is best understood through changes in the risks transferred, the contracts providing that transfer, and the capital committed to sustaining the contracts. Aggregate direct premium remains an important measure of market size. Policy counts measure a different aspect of market reach. Rates, effective limits, retentions, loss ratios, and concentration each provide information that these aggregates lack.

The U.S. evidence for 2024–2025 shows a market in which policy counts grew much faster than written premium and surplus lines accounted for the increase in domestic premium despite contraction in admitted business. The concentration calculation also shows the difference between a partial ranking and an exact index. Neither the expansion in counts nor the presence of numerous carriers establishes the breadth of catastrophe protection.

International evidence demonstrates that pricing and access vary across client populations and jurisdictions. The demand for protection can remain weak where available insurance is affordable for some segments but unfamiliar or unsuitable for others. Competition and prudential resilience depend on whether technically and financially differentiated insurers can provide meaningful coverage through periods of elevated claims and stressed reinsurance.

The relevant measure of sustained market development is consequently the durability and economic substance of the protection supplied. Contract comparability, granular risk measurement, transparent allocation of capital, and evidence on buyer outcomes will allow a more complete assessment of how market growth contributes to financial resilience. The research agenda developed here identifies the data and interpretive conditions necessary to make those assessments credible.

References

  • Aon. 2026a. Cyber and E&O: Market Remains Soft but Volatility Grows. Aon article reporting 2025 placement and claims experience. Publication record

  • Aon. 2026b. Q3 2026: Global Insurance Market Overview. Global Insurance Market Insights. Publication record

  • Bank of England. 2026. Mind the Gap: A UK Microprudential Perspective on General Insurance Protection Gaps. Bank Insights. Publication record

  • Biener, Christian, Martin Eling, and Jan Hendrik Wirfs. 2015. Insurability of Cyber Risk: An Empirical Analysis. The Geneva Papers on Risk and Insurance—Issues and Practice 40(1): 131–158. Publication record

  • Currat, Adrien, Joe Perry, and Jeffery Yong. 2026. Cyber Insurance Unpacked: The Corporate Digital Safety Net. FSI Insights on Policy Implementation 75. Basel: Bank for International Settlements, June. Publication record

  • Department for Science, Innovation and Technology and Home Office. 2026. Cyber Security Breaches Survey 2025/2026. London: UK Government. Publication record

  • Eling, Martin (presenter). 2024. The Supply of Cyber Risk Insurance. Research seminar, Insurance Risk and Finance Research Centre, Nanyang Technological University, Singapore, October 29. Seminar abstract

  • International Association of Insurance Supervisors (IAIS). 2025. Global Insurance Market Report 2025. Basel: IAIS. Publication record

  • Lloyd's. 2024. State Backed Cyber Attack Wordings. Market Bulletin Y5433, May 14. London: Lloyd's. Publication record

  • Malavasi, Matteo, Gareth W. Peters, Pavel V. Shevchenko, Stefan Trück, Jiwook Jang, and Georgy Sofronov. 2022. Cyber Risk Frequency, Severity and Insurance Viability. Insurance: Mathematics and Economics 106: 90–114. doi:10.1016/j.insmatheco.2022.05.003. Publication record

  • Marsh. 2026. IMEA Insurance Market Rates. Global Insurance Market Index, Q2 2026 snapshot (July 22, 2026). Publication record

  • National Association of Insurance Commissioners (NAIC). 2026. Report on the Cybersecurity Insurance Market. October 2026. Kansas City, MO: NAIC. Publication record

  • National Institute of Standards and Technology (NIST). 2024. The NIST Cybersecurity Framework (CSF) 2.0. NIST CSWP 29. Gaithersburg, MD: NIST. doi:10.6028/NIST.CSWP.29. Publication record

  • Nieuwesteeg, Bernold, Louis Visscher, and Bob de Waard. 2018. The Law and Economics of Cyber Insurance Contracts: A Case Study. European Review of Private Law 26(3): 371–420. Publication record

  • The Geneva Association. 2026. Strengthening Cyber Resilience Through Insurance. March 31. Zurich: The Geneva Association. Publication record

  • U.S. Department of Justice and Federal Trade Commission. 2023a. Merger Guidelines. Section 2.1, Guideline 1. December 18. Publication record

  • U.S. Department of Justice and Federal Trade Commission. 2023b. Merger Guidelines. Section 4.3, Market Definition. December 18. Publication record

  • U.S. Department of the Treasury, Federal Insurance Office (FIO). 2026. Annual Report on the Insurance Industry. Washington, DC: U.S. Department of the Treasury. Publication record

  • U.S. Government Accountability Office (GAO). 2021. Cyber Insurance: Insurers and Policyholders Face Challenges in an Evolving Market. GAO-21-477. Washington, DC: GAO. Publication record