Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Reinsurance Market Equilibrium and Risk Allocation

Abstract. Reinsurance distributes uncertain insurance liabilities across firms, jurisdictions, investors, and time. The resulting allocation depends on preferences over risk, the probability structure of claims, the scarcity and mobility of risk-bearing capital, information and incentive conditions, contractual frictions, regulation, and the ability of counterparties to perform after adverse events. This monograph reconstructs the classical problem of Pareto-efficient risk sharing in a reinsurance economy, derives marginal-valuation conditions for equilibrium, and examines how actual proportional and excess-of-loss arrangements depart from the frictionless benchmark. It integrates contract design, insurance finance, underwriting cycles, catastrophe risk, capital-market instruments, counterparty exposure, supervisory requirements, and recent international market evidence. A central analytical result is that the price of a reinsurance layer represents an equilibrium compensation for a contingent claim under a specified market structure: expected claims alone cannot determine the observed price, and lower quoted rates alone cannot establish an increase in efficient risk transfer. The study distinguishes mathematical implications from empirically tested findings, evaluates alternative explanations for catastrophe-reinsurance price movements, and develops concrete criteria for assessing risk allocation without introducing a new proprietary theory or doctrine. Its observational basis includes foundational economic research, peer-reviewed studies, international supervisory materials, and OECD reporting through November 15, 2026. Limitations of identification, comparability, data access, and model risk are explicitly stated.

Keywords: Reinsurance; market equilibrium; Pareto efficiency; risk allocation; risk aversion; marginal utility; supply and demand; proportional treaties; excess-of-loss layers; catastrophe risk; cost of capital; solvency; reinsurance cycles; moral hazard; adverse selection; counterparty risk; retrocession; insurance-linked securities; international insurance markets.

Introduction

Reinsurance is a market in contingent obligations. An insurer writing direct policies exchanges premium income for a promise to pay uncertain claims. The reinsurer assumes a specified part of those contingencies against consideration; retrocession can pass portions of the same uncertainty onward. An economically meaningful assessment therefore tracks who bears losses in each state of the world, the capital supporting each promise, the cost of obtaining protection, the expected timing of settlement, and the incentives preserved after cession. The distribution of the gross claim and the distribution of ultimate net loss can be radically different when contracts contain deductibles, limits, exclusions, event definitions, reinstatements, and credit exposure.

The research question concerns equilibrium: what allocation of uncertain losses would agents voluntarily accept at a set of clearing prices, and when does the resulting allocation maximize a suitable welfare objective? This question extends the analysis in Karl Borch's Equilibrium in a Reinsurance Market (1962), published in Econometrica 30(3), 424–444. The historical contribution is the formal study of risk sharing in an exchange economy of insurers. The existence of a mathematical equilibrium and its welfare properties require assumptions about the opportunity set, preferences, beliefs, divisibility, and completeness of trade. Those assumptions matter when the theory is applied to modern loss layers and capital constraints.

There are three separate objects of analysis. The first is an efficient allocation conditional on known probabilities, preferences, resources, and enforceable claims. The second is the market outcome produced by contract negotiation, price formation, and balance-sheet restrictions. The third is the realized distribution of losses and protection after claims, counterparty default, inflation, and model revision. Confusing them causes errors. A price close to an actuarial estimate does not by itself establish efficient allocation; a globally diversified treaty can still concentrate credit exposures; a lower ceded premium can accompany either improved capacity or a change in the attachment point.

The analysis is international in scope. It treats insurer and reinsurer behavior as economic mechanisms and uses jurisdiction-specific legal and reporting rules only where those rules bear on allocation. Supervisory standards are not assumed to have identical legal effect in every state. The IAIS reinsurance supervision framework concerns effective management of ceded risk, while the OECD insurance market evidence documents variation across reporting jurisdictions. Their combination permits an institutional comparison without treating any jurisdiction as a universal template.

Methodologically, this is a critical theoretical and comparative monograph. Equations below are explicit demonstrations under identified assumptions, not newly estimated market coefficients. Empirical findings retain the scope of the underlying studies; no proprietary treaty panel, raw regulatory microdata, or fresh econometric experiment was accessed for this manuscript. The author's original contribution is the integration of equilibrium conditions, contractual payoffs, capital mechanisms, institutional evidence, and identification boundaries into one documented analysis.

Borch Risk-Sharing Equilibrium

The exchange problem

Consider a finite set of states indexed by s, with common strictly positive probabilities p(s), and n firms indexed by i. Firm i enters the risk exchange with random endowment W_i(s), representing assets after its preexisting liabilities in the selected accounting convention. A feasible exchange assigns final state-contingent wealth C_i(s) such that

∑i=1nCi(s)=∑i=1nWi(s)for every state s.\sum_{i=1}^{n} C_i(s)=\sum_{i=1}^{n} W_i(s)\quad \text{for every state }s.

This identity presumes that administrative and transaction costs are absent and no external wealth enters the group. A real reinsurance premium is a transfer among parties; brokerage, taxes, collateral funding and default losses violate this exact zero-friction resource equation unless they are represented as explicit resources or deductions.

Let each firm have differentiable, increasing, strictly concave utility u_i over final wealth. The planner's weighted Pareto problem is

max⁡{Ci(s)}∑i=1nαi∑sp(s)ui(Ci(s)),αi>0,\max_{\{C_i(s)\}} \sum_{i=1}^{n}\alpha_i\sum_s p(s)u_i(C_i(s)), \qquad \alpha_i>0,

subject to the statewise resource constraints. At an interior optimum the first-order conditions are

αiui′(Ci(s))=λ(s)for every i,s.\alpha_i u_i'(C_i(s))=\lambda(s) \quad\text{for every }i,s.

Consequently,

ui′(Ci(s))uj′(Cj(s))=αjαifor every state.\frac{u_i'(C_i(s))}{u_j'(C_j(s))}=\frac{\alpha_j}{\alpha_i} \quad\text{for every state}.

With common probabilities, the ratio of marginal utilities is constant across states. This is the classical efficient-sharing condition. It is not a general assertion that observed reinsurers equalize risk exposures or that every market can transact every state-contingent claim. If beliefs differ, the corresponding Pareto condition is alpha_i p_i(s) u_i'(C_i(s)) = lambda(s), and the probability terms cannot be canceled. If a solvency or transaction constraint binds, its shadow value enters the stationarity equations.

Exponential utility and risk tolerance

For constant absolute risk-aversion utility, write u_i(w) = -exp(-a_i w), with a_i positive. The risk tolerance is T_i = 1/a_i. The efficient-sharing condition implies that each final wealth is an affine function of aggregate wealth W(s):

Ci(s)=Ai+Ti∑jTj W(s),W(s)=∑jWj(s),C_i(s)=A_i+\frac{T_i}{\sum_j T_j}\,W(s), \qquad W(s)=\sum_j W_j(s),

where the constants A_i are set by participation, budget, or wealth-allocation conditions and must sum to zero. Thus the state-varying component of aggregate risk is assigned in proportion to risk tolerance. The result assumes utility is well-defined over the relevant wealth states, no limited-liability kink and no separately imposed capital limit. It demonstrates how preference heterogeneity determines efficient fractions; it cannot yield an observed market share without calibration of preferences and endowments.

Robert Wilson's 1968 study of syndicate risk sharing treats collective decisions and the distribution of uncertain payoffs under heterogeneous preferences. It is a conceptual complement to Borch: a group of agents sharing an outcome has to reconcile both preferences and, potentially, probability assessments. The extension from a syndicate to treaty-based reinsurance requires specifying who can enforce transfers in adverse states.

State prices and competitive decentralization

A complete-markets competitive equilibrium contains state prices q(s) for the contingent delivery of one monetary unit in state s. Individual expected-utility maximization subject to a budget gives

p(s)ui′(Ci(s))=μiq(s)p(s)u_i'(C_i(s))=\mu_i q(s)

for a positive budget multiplier mu_i. Hence the market's relative state prices are proportional to marginal valuation adjusted for state probabilities. Summing feasible individual demands and supplies gives clearing. Welfare follows under the usual convexity, local non-satiation and competitive-completeness assumptions, subject to the initial distribution of endowments. Nothing in this derivation fixes q(s) at p(s): nontrivial marginal utility valuations can place a premium on claims concentrated in bad aggregate states.

A contingent reinsurance payment H(s) has a linear complete-market price P(H) = sum_s q(s)H(s), in a common numeraire and settlement date. The insurer's willingness to pay and the reinsurer's minimum acceptable premium generally differ when trading sets are incomplete, background risks are material, or capital is costly. In this setting the equilibrium price is the result of matching marginal valuations and the feasible supply of capital, not a universal actuarial loading factor.

Hans Bühlmann's economic premium principle (1980) explicitly distinguishes a price derived from a loss distribution alone from one shaped by the surrounding market. His contribution helps explain why a premium rule that assigns the same price to an identical loss distribution in every capital-market state is too restrictive as a general equilibrium description. An isolated actuarial premium may be useful for technical calculation without being the equilibrium transaction price.

Assumptions, existence, and welfare limits

The benchmark has precise boundaries. The agents share a state space and can contract on its realizations. Probability distributions and payoffs are measurable and known within the model. Trade is divisible, promises are performable and settlement can be enforced. Preferences are stable, monotonic and concave. The existence of competitive equilibrium requires suitable regularity and feasibility conditions; the first-order condition alone is necessary under interiority, not a proof that any observed premium vector clears a real market.

Insurance contracts rarely span every loss state. Aggregate catastrophe events contain spatially correlated exposures; lines of business interact through inflation, legal developments and asset returns; reinsurers impose limits and exclusions. Bankruptcy and counterparty credit effects make contractual payments contingent on the solvency of an additional entity. A rigorous application must begin by identifying which parts of the classical opportunity set survive.

Reinsurance Supply and Demand

The cedent's marginal demand

Let L be the insurer's gross claim, H(L) the treaty recovery, P the ceded premium, and A the insurer's capital available to absorb losses. End-of-period cedent wealth can be represented as

Wc=A+G−L+H(L)−P,W_c=A+G-L+H(L)-P,

where G denotes all other net cash flows under a stated horizon and valuation convention. If the insurer chooses a scalar coverage amount h multiplying a payoff H_0(L), its first-order utility condition at an interior optimum is

P′(h)=E[uc′(Wc)H0(L)]E[uc′(Wc)],P'(h)= \frac{E\left[u_c'(W_c)H_0(L)\right]} {E\left[u_c'(W_c)\right]},

when W_c depends on h through h H_0(L) - P(h). This expression gives the cedent's marginal willingness to pay in terminal wealth units. Recoveries in states of low cedent wealth receive more weight. An insolvency boundary, solvency capital charge, tax effect, friction, or risk constraint changes the right-hand side or adds a shadow term.

For the reinsurer with wealth W_r = B - h H_0(L) + P(h) + Z, where Z represents its other net positions, an interior supply condition is

P′(h)=E[ur′(Wr)H0(L)]E[ur′(Wr)].P'(h)= \frac{E\left[u_r'(W_r)H_0(L)\right]} {E\left[u_r'(W_r)\right]}.

A voluntary interior exchange requires both marginal valuations to coincide at the traded position. A market with multiple reinsurers must also clear available risk-bearing commitments under credit, capital, diversification and contract constraints.

Why supply curves become steep

The reinsurer's marginal valuation rises when a proposed layer pays in states that are especially adverse for the rest of its portfolio. Buying more of a poorly diversifying peril can increase marginal capital need even where its expected loss stays constant. Following a major catastrophe, a reduction in net worth can raise the shadow cost of a given exposure; new capital may be supplied at a cost that depends on information, franchise value, investor confidence, and transaction speed. The resulting schedule is locally upward-sloping in a way the unconstrained perfect-capital benchmark suppresses.

A reinsurer can diversify across geographically and temporally distinct risks, but diversification depends on joint distributions rather than geography labels. Several windstorm portfolios may be correlated through shared storm seasons, construction vulnerabilities, inflation, asset values and settlement practices. A worldwide portfolio may contain a common model-error exposure. At high quantiles, dependence and parameter uncertainty can dominate apparent average diversification.

The insurer's demand is equally state-dependent. Higher primary premium growth can increase gross exposure and the value of protection; higher retained earnings can make self-insurance more attractive; regulatory capital relief can raise the value of cession; higher attachment points can reduce treaty volume while leaving cedent exposure to smaller catastrophes. Observed retention is therefore an equilibrium outcome with multiple determinants, not a monotonic readout of reinsurance price.

Premium decomposition

For a one-year indemnity layer, a useful analytical accounting is

P=E[H]+Ccapital+Cexpenses+Cfrictions+Π,P=E[H]+C_{\mathrm{capital}}+C_{\mathrm{expenses}}+ C_{\mathrm{frictions}}+\Pi,

where the residual Pi denotes profit or negotiation effects after the specified components. The identity is only a bookkeeping decomposition once its components are defined. Capital cost is sensitive to loss dependence, funding prices and the marginal portfolio; expenses include contract servicing and claims; frictions may include collateral, brokerage and taxes. Some quantities are jointly determined and cannot be identified from a single quoted premium. It would be incorrect to describe the residual as pure market power without independently estimating expenses, expected losses and financing costs.

The volume of signed limits is not the quantity of risk measured in a common unit. One billion dollars of limit attaching at an extremely remote level transfers a very different distribution of claims from the same nominal limit attaching frequently. Comparable supply and demand analysis should therefore report expected recovery, event exhaustion probability, tail-risk impact and cedent-retained loss profiles in addition to limits.

Participation, bargaining, and imperfect competition

Bilateral treaty negotiation differs from price-taking equilibrium. Reinsurers can possess specialist risk information, capital relationships, claims-adjustment capability and a reputation for performance. Cedents can possess private exposure data or command large diversified business. Search and placement costs make the set of available offers finite. Market concentration, broker matching, collateral conditions and renewal timing can all affect bargaining shares without changing the physical catastrophe distribution.

Theoretical interpretation should therefore separate a competitive state-price result, a constrained competitive result, and a negotiated contract. A bilateral premium somewhere between the cedent's willingness to pay and the reinsurer's reservation price supports a mutually beneficial transaction; it does not prove that the allocation is globally efficient. Consolidation of counterparties can also make renegotiation or post-loss recovery more fragile.

Risk Aversion and Contract Layers

Proportional cession and exposure composition

A quota-share treaty transfers a fixed fraction k of each eligible claim: H(L) = kL, within the treaty definitions. The cedent retains (1-k)L, while the reinsurer assumes kL and receives an agreed share of eligible premiums, possibly net of ceding commission. The aggregate expected recovery is k E[L] if the same loss definition applies throughout; the variance of the cedent's retained claim equals (1-k)^2 Var(L) before adding other exposures, credit effects or premium variability. The treaty shares frequency and severity and may facilitate capacity for growth, but it also transfers relatively ordinary claims and may create a large amount of premium and cash-flow exchange.

Under constant absolute risk aversion with normally distributed aggregate outcomes and identical beliefs, an interior optimal proportional split in the zero-friction benchmark follows from equating marginal risk-bearing costs. For independent background portfolios, the solution can be written in terms of risk tolerance and variance; when backgrounds correlate with L, covariance terms modify the share. A reinsurer can demand more premium for a ceded loss that strongly correlates with the same catastrophe exposure already on its balance sheet. A fraction of gross premium is therefore not an adequate estimate of the effective economic cost without commission, coverage, reinstatement and portfolio information.

A surplus treaty instead allocates individual risks above a specified retained line, frequently using sums insured rather than realized claims as the operational base. Since the fraction ceded varies by risk size, the treaty changes the composition of the retained portfolio. In equilibrium its attractiveness depends on primary underwriting heterogeneity and administrative data quality, not merely on a constant risk-aversion coefficient.

Excess-of-loss contracts

A single-occurrence excess-of-loss layer with attachment a and limit m pays

Ha,m(L)=min⁡{(L−a)+, m},x+=max⁡{x,0}.H_{a,m}(L)=\min\left\{(L-a)^+,\,m\right\}, \qquad x^+=\max\{x,0\}.

The cedent's retained claim, before other terms, is L - H_{a,m}(L). Expected recovery admits the useful tail-integral expression

E[Ha,m(L)]=∫aa+mPr⁡(L>x) dxE[H_{a,m}(L)]=\int_a^{a+m}\Pr(L>x)\,dx

for a nonnegative loss L with a finite expectation. This follows from integrating the positive part of L over the covered interval. It makes the effect of attachment and exhaustion transparent: increasing a removes recovery from less extreme states, while increasing m adds recovery only in states above the old exhaustion point.

For a catastrophe layer, the physical expected loss is often small relative to nominal limit. Yet the marginal valuation may be high when the layer pays in precisely those states in which cedent capital is threatened. The reinsurer faces the corresponding concentration of tail losses. Consequently, the same stochastic layer can be valuable to a capital-constrained cedent and expensive for a reinsurer with positively correlated catastrophe exposures. Capital constraints can keep the two reservation prices apart and prevent the mutually desired extent of transfer.

Multiple risks, events and policy years produce aggregation issues. Per-risk cover differs from per-occurrence cover, annual aggregate cover and stop-loss contracts. Reinstatements may replenish exhausted limits at additional premiums and change the distribution of both protection and post-event payments. Event hours clauses and definitions of occurrence alter which correlated claims aggregate. Layer illustration using a single nonnegative L must not be applied mechanically to a portfolio without specifying the contractual aggregation map.

Optimal insurance design and its assumptions

Artur Raviv's Design of an Optimal Insurance Policy (1979) is a core reference on insurance contract structure under specified preferences and costs. The deductible and coinsurance results associated with this literature are conditional on their cost and information assumptions. They should not be presented as a universal theorem that all equilibrium reinsurance contracts must be excess-of-loss. The insurer-reinsurer relationship differs from household insurance: the buyer is a professional firm with a portfolio, underwriting incentives, leverage and multiple treaty alternatives.

As retention rises, a cedent may preserve incentives and reduce premium expenditure while increasing its capital sensitivity to moderate events. As limit rises, an insurer may improve extreme-event solvency up to the point at which counterparty concentration, basis risk or contract exclusions become important. A treaty optimized for expected utility need not minimize any one observed risk measure, and a treaty optimized for a regulatory risk measure need not maximize shareholder utility. Explicit objective and admissible-contract definitions are therefore necessary.

Risk aversion, tail dependence, and diversification

Variance is a complete sufficient summary of risk only under special distributional and preference structures, such as mean-variance formulations or normal risks with exponential utility. Catastrophe exposures have skewed loss distributions, heavy tails, spatial and temporal dependence, and uncertain parameters. Tail quantiles, expected shortfall and solvency ruin probabilities capture different features. The same expected recovery can produce different balance-sheet protection when losses cluster or counterparties default together.

An economically instructive decomposition follows from conditional expected recovery. If severe industry losses I are observed, a layer's systematic sensitivity may be summarized by the difference between E[H|I severe] and E[H] without claiming this difference is a priced beta. Measuring the conditional distribution, however, requires credible joint models; claims development, inflation, changes in exposure, and incomplete historical catastrophe records can bias tail calibration. The analytical conclusion is structural: the reinsurance value of a layer depends on its co-movement with relevant balance-sheet states, while the physical premium component depends on its marginal loss distribution.

Capital Scarcity and Market Structure

Balance-sheet capacity as a state variable

The capacity constraint is a central bridge between mathematical allocation and market experience. A reinsurer cannot freely expand promised recoveries if it must supply capital and liquid assets that remain credible under adverse events. Even a contract with positive expected profit may be unattractive when it consumes scarce capital at a time when other opportunities are valuable. Formally, if capital required for a portfolio position h is K(h), a constraint K(h) <= B adds a nonnegative multiplier eta to the reinsurer's problem. At a differentiable interior choice with a binding constraint, the marginal offered price contains eta K'(h) in the appropriate cash-flow units, in addition to marginal expected-utility valuation.

The shadow price eta is an economic cost of the binding constraint; it is not automatically a quoted statutory capital rate. Required capital may be calculated using different frameworks, tax effects and internal risk models. The shadow price depends on alternative deployable opportunities, external finance and future franchise income. The mathematical implication is that the same risk may command a higher offer price when risk-bearing resources shrink, even if its expected physical loss remains unchanged.

Ralph Winter's dynamic competitive insurance markets (1994) links dependent insured risks with the cost disadvantage of raising external versus internal equity. The model explains persistence in insurance price cycles by making net worth a state variable. Anne Gron's capacity-constraint evidence (1994) reports an inverse relationship between capacity and the price-payment margin in historical U.S. property-casualty industry series. Her related insurance-cycle study also examines the distinction between capacity constraints and arbitrage-based interpretations. These studies concern their specified periods, countries, data definitions and model tests; they establish neither a permanent pricing coefficient nor an inevitable direction for every contemporary renewal.

Shock, recapitalization, and renewal

A large catastrophe can simultaneously reduce reinsurer capital, increase uncertainty about the loss distribution, expose reserve development problems and raise cedent demand for renewed protection. Because these drivers move together, a simple positive correlation between catastrophe events and treaty prices does not identify a single supply mechanism. Separating them requires data on actual contract coverage, insurer exposures, model revisions, capital transactions and timing.

Kenneth Froot and Paul O'Connell's U.S. catastrophe reinsurance pricing (1999) analyzes detailed contract data and evaluates demand shocks against limited elasticity of capital supply. The study finds supply-side frictions more persuasive in its historical setting. Froot's later catastrophe-risk market examination (2001 published version) emphasizes high premia relative to expected losses and limited purchases of the most extreme coverage, with capital-market imperfections and reinsurer market power among the plausible explanations. The results support scrutiny of capacity and bargaining, but cannot substitute for current event-specific analysis.

An apparently hard market can combine several changes: higher expected modeled losses, tighter availability at certain layers, larger margins for uncertainty, higher funding costs and strategic changes by cedents. A subsequent soft market may feature better capitalization and lower margins even as insurers deliberately retain more loss below higher attachments. Market terminology alone is an unreliable measurement of aggregate risk transfer.

Risk financing and the reinsurance decision

The cedent's economic problem includes alternative capital uses. Keeping more risk may require higher equity, risk reduction, more conservative underwriting, stronger liquidity arrangements, or access to emergency recapitalization. Buying reinsurance exchanges an upfront cost for contingent recovery and potentially reduces a solvency measure, but introduces counterparty and coverage risk. The full net benefit is therefore the avoided cost of retained risk and capital, minus the ceded premium, counterparty exposure and operating frictions, evaluated consistently.

J. David Cummins, Georges Dionne, Robert Gagné and Abdelhakim Nouira report in their costs and benefits of reinsurance study (2009 working-paper record) that ceded protection can raise observed insurer production costs while reducing loss-ratio volatility. This evidence supports a trade-off, not a universal prescription about optimal cession. The welfare assessment depends on the insurer's baseline capacity, risk appetite, claims distribution and how policyholder protection is valued.

Capital raising may also create entry opportunities. High margins encourage new reinsurance capital and alternative structures, but new capacity needs underwriting information, legal and operational systems, ratings or collateral, and distribution. The process can be faster in standardized catastrophe-linked instruments than in bespoke long-tail liabilities, though comparative costs must be shown rather than assumed. Renewal outcomes are jointly determined by the legacy contract stock, new information, and outside capital's opportunity cost.

Market power and specialization

Reinsurance is international yet highly specialized. For common, measurable and standardized risks, competitive offers can be numerous. For highly correlated peak perils, very large limits or unusual liability exposures, the number of credible counterparties is smaller. Economies of scale in catastrophe modeling, rating credibility, broker relationships and claims settlement can create concentrated capacity. Specialists may offer real informational value while possessing bargaining power.

An observed price above the actuarial expectation is not proof of monopoly. Expected losses are estimated under uncertain models; capital has option value; the promised payoff might not be fully collateralized; the reinsurer may bear uncertain settlement and dispute costs. Conversely, invoking expenses and capital constraints cannot automatically dismiss market power. Identification requires a model of all material margins and independent observations on counterparty choice, quoted and signed terms, and changes in available capacity.

International Risk Markets

International capacity and domestic exposure

International reinsurance can combine catastrophe exposures from multiple jurisdictions, shift capital to places facing concentrated risk, and support the underwriting capacity of domestic insurers. Its contribution depends on the joint distribution of those exposures and on the location, liquidity and legal availability of supporting funds. The IAIS guidance on reinsurance identifies geographical diversification and cross-border risk transfer as central features of reinsurance management. The supervisory problem is simultaneous: a cedent can reduce its gross underwriting concentration while increasing recoverable assets owed by a smaller number of external counterparties.

Country-by-country reinsurance data require careful denominator choices. Cession ratios can be measured from premiums, claims, sums insured or modeled capital relief, and the ratios need not move together. Premium-based retention equals net written or earned premiums divided by corresponding gross premiums only after consistent accounting and portfolio scope are established. A rise may reflect changes in rate, business mix, commissions, protection structure or gross volume without an equal percentage shift in loss-bearing exposure.

The OECD's Global Insurance Market Trends 2026, published November 2, 2026, reports that in a comparable set of 52 reporting jurisdictions, 37 showed higher insurer retention ratios and 15 lower ratios between 2024 and 2025; the simple sample average moved from 74.8% to 75.2%. The same publication cites an estimate of USD 785 billion of global reinsurance capital at year-end 2025 and records lower reinsurance pricing in important property catastrophe segments. The capital number is an industry estimate reported by OECD, not an independently reconstructed aggregate in this monograph. These observations coexist: larger global supply and cheaper pricing can accompany higher retained shares, precisely because demand is influenced by capital strategies, attachments, risk appetite and market-specific constraints.

OECD reporting further documents that international aggregates conceal local frictions. Some insurers continued to face difficulty obtaining catastrophe coverage even as broader rates softened, and supervisory respondents associated climate-related loss histories with local capacity conditions. Thus a global price index is insufficient to evaluate whether a particular flood, wildfire, windstorm or emerging cyber exposure is economically transferable at useful terms.

Catastrophe shocks and protection gaps

In catastrophe reinsurance, the insured loss distribution is determined by hazard, exposure, vulnerability, contract terms and claims development. A changing climate can affect hazard distributions while urbanization, insured property values and repair inflation alter exposure and severity. The rate at which physical catastrophe losses become insured claims depends on access to primary insurance and on policy conditions. Reinsurance transfers a portion of existing insured losses; it does not automatically finance uninsured households and businesses.

The OECD's catastrophic financial protection study (2026) evaluates public-private insurance and compensation approaches for large-scale losses. Its discussion clarifies why the potential contribution of private reinsurance varies across perils and jurisdictions. An insurance program may raise participation and improve risk financing while creating fiscal exposures if premiums, pooling, risk reduction and government backing are poorly matched. The presence of a public backstop alters the feasible allocation of extreme losses, which has to be made explicit in any welfare comparison.

The OECD's 2026 survey of sector vulnerabilities reports catastrophe and climate, cyber and macrofinancial risks as recurrent supervisory concerns. These categories do not have interchangeable probability models. A natural hazard may be geographically concentrated but physically modeled, while a cyber shock can transmit through common software or service providers and produce highly correlated claims. As correlation strengthens, a contractual transfer across companies may redistribute exposure within a shared system without creating much additional aggregate diversification.

Alternative reinsurance capital

Catastrophe bonds, collateralized reinsurance, industry loss warranties and other insurance-linked structures widen the potential set of risk bearers. Investor participation can help finance catastrophe layers that would otherwise depend largely on traditional reinsurers. J. David Cummins and Mary Weiss's study of securitized risk transfer (2009) examines the convergence of insurance and financial markets and the economic rationale for hybrid structures. Such instruments must be compared by trigger, basis risk, collateral, maturity, reset terms, costs and recovery timing.

A parametric or industry-index instrument may settle quickly and reduce disputes over individual loss adjustment, but it can leave the cedent with an imperfect hedge. Define an indemnity loss L and index-triggered payment H(I), where I is an observable index. The basis-risk exposure is L-H(I) and its welfare effect depends on the joint distribution of L and I. A lower premium for the indexed instrument might compensate for a materially larger residual tail. Conversely, a collateralized transfer can reduce counterparty credit risk relative to an unsecured promise, subject to collateral-investment, legal and operational risks.

Investor capital is not infinitely elastic. An instrument's yield must compete with returns on other assets and with volatility of investor wealth. Catastrophe events can produce mark-to-market effects, redemptions, opportunity-cost changes and revised model assumptions. The benefit of a wider risk-bearing base rests on correlations with investors' financial portfolios and on transparent triggers. The conceptual test is whether new investors carry incremental exposure, not whether risk has acquired a capital-market label.

Global monitoring and supervisory comparability

The IAIS Global Insurance Market Report 2025 identifies strong insurance-sector solvency and profitability in its monitoring exercise alongside attention to private credit, geoeconomic fragmentation, artificial intelligence, climate and cyber exposures. The report describes the surveyed insurance sector and its supervisory risk landscape; it cannot by itself establish an equilibrium elasticity for global catastrophe-reinsurance supply.

The Insurance Capital Standard, adopted by IAIS in 2024, provides a globally comparable risk-based group capital measure for internationally active insurance groups. Its role is regulatory comparability within the relevant group framework. Individual jurisdictions apply their own legal requirements and supervisory practices; a capital figure from one regulatory regime cannot be substituted automatically for another.

Accounting evidence and economic risk

The IFRS 17 insurance contract standard addresses financial reporting, including reinsurance contracts held, and is effective for annual reporting periods beginning on or after January 1, 2023. It affects the recognition and presentation of insurance-related obligations and recoveries for entities applying the standard. An accounting measurement is not identical to the competitive market price of a state-contingent claim or to cash capital accessible after a catastrophe.

This distinction is material for empirical work. Gross and net premiums, reinsurance contract assets, loss ratios, solvency capital and estimated economic recovery each reflect different conventions. An insurance group may show a large accounting recoverable but still face collection timing, collateral or counterparty risks in extreme states. A dataset joining IFRS 17 reports with non-IFRS reporting regimes must document definitions and conversions before deriving comparative retention or capital-cost statistics.

Market Failure and Institutional Conditions

Asymmetric information and moral hazard

The cedent knows more about underwriting quality, accumulation, reserving and settlement behavior than a distant reinsurer may be able to observe. The reinsurer cannot write a complete contract on every relevant future decision by the cedent. Reinsurance that covers a large share of marginal loss can alter incentives to monitor underlying exposures, improve loss control or negotiate claims settlement. Retention, adjustable premiums, audits, claims cooperation and experience-rating provisions serve partly as incentive mechanisms.

Neil Doherty and Kent Smetters's Moral Hazard in Reinsurance Markets (2005) develops a multiperiod principal-agent model and tests predictions on a U.S. property-liability reinsurance panel. The authors find evidence of loss-sensitive pricing between unaffiliated firms, and stronger use of monitoring for affiliated firms where such monitoring is cheaper. These are substantive institutional findings, but they apply to the empirical period and distinction under study. They do not imply that all affiliated reinsurance transfers genuine independent capital capacity or that present markets have the same monitoring technology.

An insurer may also know more about its own exposure distribution than its reinsurers, allowing adverse selection. Treaty data audits, aggregate exposure reporting, exclusions, claims-control clauses and bargaining over information can respond to such asymmetry, though each has costs. Informational frictions alter both the admissible contract menu and the value of risk transfer. A reported low risk-adjusted price may reflect superior risk quality, understated hazard, inaccessible information or strategic selection; identifying which is a research question.

Counterparty risk, contagion, and performance

Reinsurance recovery is conditional on claims eligibility and on reinsurer performance. A cedent transferring tail exposure may replace underwriting volatility with an asset whose value deteriorates in the same catastrophic state. Multiple cedents using the same reinsurers can generate concentrated credit exposures and feedback effects when large events stress the sector. Diversification of contractual counterparties can be different from diversification of economic ownership, retrocession and correlated collateral.

Cummins and Weiss's insurance systemic-risk analysis (2014) examines linkages among insurers and financial institutions, and related research distinguishes core underwriting risks from noncore financial activities. The study's United States framing matters; a wholesale claim that the global reinsurance network is either inherently stable or necessarily systemic would exceed the evidence. At minimum, risk analysis should map counterparties and the state dependence of their resources.

Under the European Solvency II framework, Article 105 on capital requirements includes counterparty default risk associated with risk-mitigation contracts such as reinsurance. The rule underscores an economic point: cession can reduce one risk component and create another. Collateral, diversification, credit assessment, enforceability and supervisory oversight all affect the net result.

Incentives for catastrophe prevention

Risk transfer can support catastrophe resilience when it allows insurers to finance unexpected losses and continue underwriting. At the same time, inappropriate pricing or guaranteed assistance can weaken incentives to reduce exposure. A welfare analysis therefore includes the behavior of direct policyholders, primary insurers, reinsurers, public agencies and investors. Loss prevention can change the aggregate physical loss distribution; reinsurance alone changes its financial allocation.

Where premiums are risk-sensitive and affordable, risk mitigation can be rewarded through pricing. Where the risk is severely correlated or low-income exposures cannot bear full risk-reflective premiums, participation and distributional objectives can call for explicit public support. The OECD's catastrophe protection assessment recognizes design differences between public-private insurance and post-event assistance. Institutional design consequently belongs in the equilibrium assessment: underwriting conditions and public guarantees alter participation and the allocation of contingent costs.

Jurisdiction, law, and enforceability

Reinsurance obligations depend on treaty wordings, applicable law, arbitration, collateral, sanctions, insolvency rules and cross-border supervision. Credit for reinsurance can depend on the recognition of genuine risk transfer and the financial standing of the reinsurer. The IAIS ICP 13 standard emphasizes insurers' effective reinsurance management rather than presuming all ceded contracts eliminate risk. Different supervisory and legal systems translate that broad standard into distinct obligations.

The institutional inference is that legal performance is part of economic feasibility. State-contingent promises with uncertain collection and dispute resolution are not equivalent to immediately deliverable Arrow-style contingent assets. A model treating promised recoveries as certain can overstate efficient risk transfer and understate the cedent's retained ruin exposure.

Comparative Economic Illustrations

A transparent three-state portfolio

Consider a deliberately synthetic one-period insurer loss distribution, in millions of dollars: loss zero with probability 0.95, loss 20 with probability 0.04 and loss 120 with probability 0.01. There are no expenses, salvage, claim-development errors, investment returns or contractual disputes in the illustration. Gross expected loss is 0.95(0) + 0.04(20) + 0.01(120) = 2.00 million. The figures are an author-constructed arithmetic demonstration, not observations on an actual insurer, catastrophe zone or treaty renewal.

A 25% quota-share treaty pays 0, 5 and 30 in the three states; its expected recovery is 0.50 million and the cedent's retained losses are 0, 15 and 90. A 40-million excess-of-loss layer attaching at 20 million pays 0, 0 and 40; its expected recovery is 0.40 million and the cedent retains 0, 20 and 80. The second treaty transfers less expected claim cost yet offers greater relief in the extreme state. A comparison of quoted premiums cannot identify which is preferable without specifying risk aversion, solvency constraints, counterparty performance and the remaining portfolio.

State Probability Gross loss 25% quota-share recovery 40 excess of 20 recovery Cedent retention under quota share Cedent retention under layer
No loss 95% 0 0 0 0 0
Moderate claim 4% 20 5 0 15 20
Extreme claim 1% 120 30 40 90 80
Expected amount 100% 2.00 0.50 0.40 1.50 1.60

Suppose, purely for illustration, that the premium is 0.65 million for the quota share and 0.70 million for the layer. Expected losses plus premiums then give expected all-in cedent costs of 2.15 million for the quota share and 2.30 million for the layer, versus 2.00 million with no treaty. These differences of 0.15 and 0.30 million are the respective illustrative costs above expected risk recovery. They do not imply the quota share is preferable: a sufficiently high shadow value of extreme-state capital can make the layer attractive despite its higher expected cost and smaller expected recovery. Conversely, a highly capitalized cedent with affordable internal risk-bearing capacity may reject both quotes.

One must not infer a real-world catastrophe premium multiple from these invented numbers. A correctly designed empirical comparison would estimate each contract's expected loss under consistent exposure and modeling assumptions, then compare its economic benefit against alternative capital or reinsurance arrangements. The example demonstrates the distinction between expected-cost ranking and marginal tail-protection ranking, not a universal ordering of treaty types.

Capital scarcity without changing the loss distribution

Let a reinsurer have an available capital budget B for a layer and assume the modeled distribution of H remains unchanged. In a simple constrained problem the capital charge K(h) is increasing in ceded quantity h. If the constraint K(h) <= B binds, reducing B increases the scarcity of admissible risk capacity; the marginal price may therefore rise through the shadow value of capital. The direction is conditional on the constraint being binding and on the market's ability to obtain alternative capital. If new capital enters at an unchanged marginal cost, scarcity can be temporary and price response muted.

The observational counterpart must distinguish supply contraction from demand growth or a revision in expected loss. Consider two renewals that each have a premium of 10% of nominal limit. One attaches at 10 million and the other at 50 million; their expected recoveries can differ by orders of magnitude. Equal rate-on-line is consequently insufficient to infer equal economic price, and a fall in rate-on-line is insufficient to infer a cheaper constant-risk trade. The treaty must be compared on matched exposure, trigger, attachment, limit, reinstatement, modeling vintage and settlement conventions.

A counterparty stress state

Suppose the excess-of-loss layer above pays 40 million in the extreme event, but the reinsurer can fulfill only a fraction r of the contractual recovery in that state. The cedent's extreme retained loss becomes 120 - 40r, ignoring recoveries from collateral or guarantee arrangements. At r=1, the cedent retains 80; at r=0.5, it retains 100; at r=0, it retains 120. The expected realized recovery is now 0.4r million if performance changes only in the extreme state and the stated probabilities remain unchanged.

This elementary scenario shows why promised contractual cession and effective financial transfer differ. A valuation with default probability independent of the loss can misstate the economics when reinsurer impairment is correlated with the very catastrophe that triggers recovery. Collateral, ring-fencing, portfolio diversification and supervisory capital requirements should be evaluated for the extent to which they change the statewise performance distribution.

Cross-border protection and public support

Imagine two otherwise comparable insurers exposed to geographically distinct catastrophes. International pooling could reduce the aggregate capital needed for a given tail criterion if the two loss processes have sufficiently low dependence. If the catastrophes instead share a climate driver, market disruption or correlated capital-market shock, aggregate diversification benefits diminish. Jurisdictional diversification must therefore be assessed through dependence estimates, event footprints and financial exposures rather than by counting countries.

A state-supported catastrophe pool changes the allocation further. Government guarantees can improve the availability of an extreme layer, but they replace some private risk-bearing claims with contingent public obligations. The welfare assessment must therefore include fiscal capacity, risk-reduction incentives and distribution across policyholders and taxpayers. Comparison of private and public-private arrangements requires naming the beneficiary and the states in which the public balance sheet is exposed.

Research Design, Identification, and Robustness

A source-to-claim hierarchy

The scientific corpus used here comprises original published economic articles, journal publisher records and scholarly abstracts, supervisory standards, and international institutional reports. Each supports a different class of claim. A theoretical article supports implications of its formal assumptions; a historical empirical study supports associations or identified tests for its sample; a standard establishes institutional requirements in its stated scope; an international report supports its observed coverage, dates and definitions. No source class licenses unrestricted extrapolation from its evidence.

Borch (1962) and Bühlmann (1980) anchor the efficient allocation and economic pricing discussion. Gron (1994), Winter (1994), Froot and O'Connell (1999), and Froot (2001) explain and investigate capacity and catastrophe-market behavior. Doherty and Smetters (2005) address asymmetric information and monitoring. Cummins and coauthors examine the costs of ceded protection and the expansion of alternative risk-transfer channels. IAIS, IFRS, Solvency II and OECD materials describe the regulatory and international environments. The literature is mutually informative, but its samples, definitions and theoretical assumptions are heterogeneous.

The bibliography does not purport to be an exhaustive census of all reinsurance research. Important historical studies were available only through publisher summaries or indexed records; the text confines claims accordingly. Some IAIS material is a supervisory standard and some OECD content is a synthesis of data supplied by national authorities or commercial providers. This separation prevents false attribution of industry estimates as primary OECD measurements.

Price measurement

An empirical investigation should begin with contract-level terms: cedent, reinsurer, date, line of business, exposure geography, attachment, occurrence or aggregate limit, hours clauses, reinstatement, expected loss, modeled catastrophe return period, premium, commission, and effective collateral. For a homogeneous layer, price-on-line equals premium divided by limit; expected-loss multiple equals premium divided by modeled expected payment. Both must be indexed to the same contract exposure and model vintage. A premium rate on subject premium is a distinct unit of measurement.

The dependent variable might be log premium per unit of model-implied expected loss, allowing for loss-model uncertainty and zero expected values. Alternatively, one could estimate a contract's internal state-price weighting or required capital charge. Both approaches require assumptions not contained in reported premium totals. Expected loss and a shadow capital price are jointly estimated or otherwise inferred, so a regression of premium on expected loss and contemporaneous capital can suffer simultaneity.

Distinguishing economic mechanisms

A useful panel design would separate cedent fixed characteristics, contract-layer characteristics, peril-by-year shocks and reinsurer capital shocks. In principle, variation in reinsurer balance sheets across cedents could help identify a supply effect if it is plausibly independent of the cedent's unobserved changes in demand. In practice, reinsurers specializing in the same peril can experience correlated losses, capital shocks and model revisions. The exclusion restriction therefore needs direct support, not a verbal claim of exogeneity.

Event studies of catastrophe shocks face anticipation, event clustering, exposure migration and nonparallel trends. Difference-in-differences comparisons across strongly and weakly affected regions require testing pretrends and spillovers; premiums and capacity can react before renewal. Instrumental variables based on distant catastrophes require showing that the instruments affect the focal market only through the proposed capital channel. This is difficult when reinsurers share portfolios, retrocession networks and capital-market investors.

To distinguish market power from supply scarcity, a study would need data on offers that were not accepted, market participation, changes in concentration, treaty terms, reinsurer expense and risk models, and alternative-capital prices. Signed contracts alone are selected observations from negotiations. Placement failures and coverage exclusions can conceal rationing from conventional premium indexes. A high realized margin after a benign catastrophe year also cannot identify ex ante pricing power without controlling for the expected distribution at inception.

Dependence and extreme-loss estimation

Tail estimates should be stress-tested for changing exposure, peril frequency and severity, vulnerability, insurance inflation, spatial correlation, loss development and parameter uncertainty. An efficient risk allocation derived from one joint distribution can be fragile to misspecification. Models should distinguish conditional climate scenarios from probabilistic forecasts and identify whether a scenario has a calibrated probability. A catastrophic-loss statistic derived from annual aggregates should not be directly compared with occurrence-level treaty recovery without mapping event aggregation.

For cyber, infectious disease, financial guarantee and geopolitical accumulation, the tail can be affected by rapidly changing network structures and behavioral responses. Historical frequencies alone may be especially uninformative. Robust comparison can report results under alternative dependence assumptions and multiple stress scenarios without assigning unsupported probabilities.

Cross-country comparability

International comparisons should distinguish regulation of primary insurers and reinsurers, IFRS 17 versus other accounting systems, gross versus net reporting, non-life versus life business, direct versus assumed premiums and currency denomination. Reinsurance capital estimates may include traditional equity, dedicated alternative capital and collateralized capacity under different definitions. Aggregating them without reconciliation can double-count claims on capital or obscure how much capacity is truly loss-absorbing.

The 2026 OECD retention comparison is informative on the reported sample, but its simple average across 52 reporting jurisdictions is not a global market-weighted retention rate. Treating it as a worldwide elasticity would be a category error. Further, the comparison across 2024 and 2025 cannot independently distinguish a price effect from changed attachment strategies, business mix or capital regulation. These are substantive empirical limitations, not editorial omissions.

Falsifiable implications and boundary tests

Several explanations yield testable implications without claiming the invention of a new model. Under a binding capital constraint, matched-risk reinsurance prices should respond to plausibly exogenous changes in available reinsurer capital, with larger responses where alternative capacity is scarce. Under dominant expected-loss revision, price changes should track risk-model changes after controlling for quantity and contract terms. Under moral hazard, retention and monitoring provisions should vary with observability and contractual incentives. Under counterparty risk, collateral and credit quality should affect the net benefit of otherwise comparable treaties.

Evidence contrary to these predictions would require reconsidering the mechanisms, not rejecting the analytical framework as a whole. A convincing test needs pre-specified variables, independent measurement, a credible counterfactual and recognition of equilibrium feedback. Observing an association is weaker than identifying its cause. Such distinctions remain necessary even for studies with statistically significant coefficients.

Integrated Analytical Synthesis

The foundational risk-sharing equation and the institutional evidence jointly support an interpretation of reinsurance as a market in scarce and imperfectly enforceable contingent capital. Efficient statewise sharing determines a benchmark marginal valuation. The actual allocation is selected from a narrower feasible set shaped by treaty structure, capital, private information, regulatory accounting, law and settlement capacity. This interpretation integrates established results rather than establishing a separate authorial doctrine.

First, the price of risk has at least three layers of meaning. Its physical expected recovery depends on the loss distribution and contract payoff; its marginal economic value depends on wealth, preferences and background exposures; its transaction price depends on available suppliers, capital costs and bargaining. Claims that market prices exceed expected losses become analytically meaningful only after these quantities are distinguished.

Second, the market's capacity is dynamic. Capital is consumed by catastrophe losses and replenished through retained earnings, external investors and strategic changes to exposure. Demand and supply react to catastrophe information simultaneously. The literature demonstrates why underwriting cycles can arise under financial frictions, while the exact contemporary balance among modeled losses, capital shocks and competition requires new data. An international capacity aggregate can be rising while local availability remains limited.

Third, contract form allocates states rather than nominal limits. Proportional and layered treaties can transfer similar expected loss while offering profoundly different extreme-event protection. Attachment, exhaustion, reinstatement, aggregation and trigger terms must be reconstructed before claims of price comparability or welfare improvement are made. The synthetic portfolio calculation illustrates how a layer can lower the extreme retained loss while transferring a smaller mean loss.

Fourth, the counterparty is part of the risk. A promised recovery that becomes doubtful during a system-wide loss does not replicate a perfectly enforceable state claim. The net economic position requires correlated default modeling, collateral assessment and monitoring of retrocession and group exposures. A policy that rewards gross cession volume without checking recoverability can misread protection.

Fifth, the international system contains multiple overlapping capital channels. Traditional reinsurers, alternative investors, cedent balance sheets and public authorities may all absorb parts of an event. New capacity improves welfare when it brings credible incremental resources and better statewise matching, with incentives and basis risk accounted for. Relabeling an already concentrated exposure or creating an unsecured claim on the same stressed balance sheet may change form more than substance.

Sixth, empirical identification governs the strength of policy conclusions. The research finds evidence consistent with capital constraints, insurance cycles, information frictions and cross-market differences. It does not establish a single timeless premium function or a universal optimal retention ratio. Sound comparison requires matched contracts, state-contingent exposure, measured capital, credible sources and precise reporting definitions.

From these findings follows an operational research judgment: an assessment of reinsurance-market equilibrium is complete only when it states its probability model, participants, outside options, capital and enforcement assumptions, contract payoff map, pricing convention, and welfare criterion. These are explicit dimensions of an analysis, not a new proprietary conceptual system. A result that omits one of them may still be useful, but its interpretation is correspondingly narrower.

Conclusion

Reinsurance equilibria allocate contingent liabilities according to marginal valuations, available risk-bearing capital, contractual possibilities and institutional conditions. In the classical complete-market economy, Pareto-efficient sharing equates weighted marginal utilities across firms and states. Under constant absolute risk aversion and a shared probability structure, the state-varying part of aggregate risk is divided in proportion to agents' risk tolerances. The result identifies a clear mathematical benchmark for studying actual insurance risk markets.

Real reinsurance contracts implement a constrained version of this exchange. Proportional treaties distribute eligible losses broadly; excess-of-loss layers target specified segments of a loss distribution; alternative-capital instruments can expand the set of risk bearers at the cost of trigger, basis, collateral and liquidity considerations. Capital scarcity and costly external financing help explain why premium margins move with the insurance cycle. Information asymmetry, bargaining power, correlated losses and counterparty default further alter both the price and the achieved allocation.

The comparative evidence also confirms that international aggregates must be interpreted carefully. OECD 2026 reporting describes greater reinsurance capital and softer prices in important lines alongside increasing insurer retention in most of the jurisdictions in a specified panel. These observations are mutually compatible in a market in which primary insurers change capital choices, exposure structure and treaty attachments. Institutional standards reinforce the necessity of evaluating ceded risk, counterparty recoverability, accounting treatment and capital recognition in their proper jurisdictions and scope.

The original analytical contribution of this monograph is its source-checked integration of welfare conditions, economic pricing, contract payoffs, balance-sheet capacity, incentive problems and international evidence, including transparent illustrations and explicit rules for interpreting empirical research. The strongest conclusion is methodological and substantive: an economically meaningful reinsurance price or retention ratio derives its interpretation from the contingent loss it covers, the capital and institutions supporting the promised payment, and the outside opportunities of both sides of the exchange. The research makes no claim to have established a newly invented theory, performed original contract-level econometrics, or obtained independent external peer review. The verified manuscript is intended for the separate publication process of Insurance by Bogdanova under Aisentica Development.

References

  • Borch, Karl. 1962. “Equilibrium in a Reinsurance Market.” Econometrica 30 (3): 424–444. Journal record. DOI: 10.2307/1909887.

  • Bühlmann, Hans. 1980. “An Economic Premium Principle.” ASTIN Bulletin 11 (1): 52–60. Casualty Actuarial Society record. DOI: 10.1017/S0515036100006619.

  • Cummins, J. David, Georges Dionne, Robert Gagné, and Abdelhakim Nouira. 2009. “The Costs and Benefits of Reinsurance.” Working-paper record and abstract, Casualty Actuarial Society. Research record.

  • Cummins, J. David, and Mary A. Weiss. 2009. “Convergence of Insurance and Financial Markets: Hybrid and Securitized Risk-Transfer Solutions.” Journal of Risk and Insurance 76 (3): 493–545. Publisher article. DOI: 10.1111/j.1539-6975.2009.01311.x.

  • Cummins, J. David, and Mary A. Weiss. 2014. “Systemic Risk and Regulation of the U.S. Insurance Industry.” In Modernizing Insurance Regulation, edited by John H. Biggs and Matthew P. Richardson. Publisher chapter. DOI: 10.1002/9781118766798.ch7.

  • Doherty, Neil A., and Kent Smetters. 2005. “Moral Hazard in Reinsurance Markets.” Journal of Risk and Insurance 72 (3): 375–391. Bibliographic record.

  • European Insurance and Occupational Pensions Authority. 2024 online rulebook text. “Solvency II Directive: Article 105, Calculation of the Basic Solvency Capital Requirement.” Official rulebook.

  • Froot, Kenneth A. 2001. “The Market for Catastrophe Risk: A Clinical Examination.” Journal of Financial Economics 60 (2–3): 529–571. NBER working paper and publication record. DOI for published article: 10.1016/S0304-405X(01)00052-6.

  • Froot, Kenneth A., and Paul G. J. O'Connell. 1999. “The Pricing of U.S. Catastrophe Reinsurance.” In The Financing of Catastrophe Risk, edited by Kenneth A. Froot. Chicago: University of Chicago Press. NBER book chapter.

  • Gron, Anne. 1994a. “Capacity Constraints and Cycles in Property-Casualty Insurance Markets.” RAND Journal of Economics 25 (1): 110–127. Bibliographic record.

  • Gron, Anne. 1994b. “Evidence of Capacity Constraints in Insurance Markets.” Journal of Law and Economics 37 (2): 349–377. Publisher article. DOI: 10.1086/467317.

  • International Association of Insurance Supervisors. 2024. “ICP 13: Reinsurance and Other Forms of Risk Transfer.” In Insurance Core Principles and Common Framework, revised December 5, 2024. Official ICP 13.

  • International Association of Insurance Supervisors. 2024. “Insurance Capital Standard.” Standard adopted 2024; current official explanatory resource. Official ICS overview.

  • International Association of Insurance Supervisors. 2025. Global Insurance Market Report 2025. Official report announcement, December 2, 2025. IAIS GIMAR release.

  • IFRS Foundation. 2023 onward. “IFRS 17 Insurance Contracts.” Standard and implementation overview, effective for annual periods beginning January 1, 2023. Official IFRS 17 standard overview.

  • OECD. 2026a. Financial Protection Against Catastrophic Risks: Floods, Fires and Other Major Risks. Paris: OECD Publishing. OECD publication.

  • OECD. 2026b. Global Insurance Market Trends 2026. Paris: OECD Publishing, November 2. DOI publication record. DOI: 10.1787/c8be6585-en.

  • OECD. 2026c. “Risks and Vulnerabilities Facing the Insurance Sector.” In Global Insurance Market Trends 2026, chapter 6. OECD online chapter.

  • Raviv, Artur. 1979. “The Design of an Optimal Insurance Policy.” American Economic Review 69 (1): 84–96. Bibliographic record.

  • Wilson, Robert. 1968. “The Theory of Syndicates.” Econometrica 36 (1): 119–132. Journal DOI. DOI: 10.2307/1909607.

  • Winter, Ralph A. 1994. “The Dynamics of Competitive Insurance Markets.” Journal of Financial Intermediation 3 (4): 379–415. Publisher article. DOI: 10.1006/jfin.1994.1011.