Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Reinsurance Cost Allocation in Primary Insurance Pricing

Abstract. Reinsurance modifies the expected losses, cash flows, tail exposure, risk capital, and counterparty risks of a primary insurer. Allocating the cost of that protection to primary insurance products is a distinct actuarial task from determining the price of a reinsurance contract. This monograph examines how ceded premiums, anticipated recoveries, ceding commissions, reinstatement premiums, brokerage, loss-sensitive adjustments, and capital effects can be assigned to underlying products and rating classes without duplicating costs or obscuring the economic transfer. It develops contract-consistent cash-flow identities; derives proportional and excess-of-loss allocations from claim indemnity functions; compares premium-based, expected-recovery, exposure, marginal, and cooperative allocations for shared protection; and analyzes the consequences of aggregation, catastrophe dependence, contractual attachment, counterparty default, and timing. Two fully specified hypothetical examples reconcile quota-share cash flows and shared occurrence excess-of-loss protection, including a Shapley allocation of recoveries and an explicit sensitivity analysis. An international comparative review distinguishes primary pricing, U.S. actuarial practice standards, international supervisory expectations, Solvency II valuation, and IFRS 17 financial reporting. The original analytical synthesis establishes reconciliation, causal exposure relevance, marginal decision usefulness, and transparent treatment of joint benefits as criteria for selecting an allocation method. These criteria organize existing actuarial mathematics and financial reasoning; they are not advanced as a newly established theory. The study concludes that a defensible allocation is a documented economic attribution for a stated decision, not an intrinsically unique division of a shared treaty bill. All numerical data are constructed for illustration and are not insurer observations, market quotations, or regulatory rate indications.

Keywords: Reinsurance cost allocation; primary insurance pricing; ratemaking; quota share; excess-of-loss; ceded premium; expected recoveries; ceding commissions; reinstatement premium; occurrence aggregation; Shapley allocation; marginal cost; capital allocation; catastrophe risk; counterparty risk; treaty cash flows; actuarial governance.

Introduction

The financial question for primary insurance pricing is how purchasing reinsurance changes the prospective cost of writing a given primary portfolio and how that change should appear in prices charged to policyholders. A treaty negotiated at the portfolio level may cover multiple products, territories, policy periods, limits, and peril groups. The cedant pays consideration to the reinsurer; the value obtained appears as contingent indemnification, modified capital requirements, liquidity support, and constraints on the cedant's retained loss distribution. It is therefore possible for the amount of an annual reinsurance invoice to be perfectly known while the proper allocation of its economic cost remains uncertain.

The ASOP 53 prospective-cost standard addresses insurance and reinsurance future-cost estimation within its stated United States professional scope. Its treatment of reinsurance costs as a possible separate component of future cost offers the immediate methodological starting point. The ASOP 29 expense standard specifically identifies premium payments, ceding commissions, anticipated recoveries, retrospective adjustments, and reinstatement premiums when reinsurance cost is reflected as an expense provision. These are not necessarily interchangeable line items. Their economic meanings and timing govern whether a primary pricing indication includes an amount once, twice, or omits it altogether.

The problem has a second dimension: a treaty can benefit several products jointly. An occurrence catastrophe layer may not respond to any single product's claim, yet may respond when claims on several products combine in one occurrence. Allocating that treaty solely by reported premium, by historic ceded claims, or by total sums insured can generate different answers. These methods represent different questions. Premium share describes a revenue base; expected recovery share describes the benefit distribution under a specified stochastic model; marginal contribution describes the effect of adding or removing exposure; and cooperative methods distribute joint benefits subject to an explicit convention. Clark's reinsurance pricing study and Patrik's foundational reinsurance chapter explain why treaty-specific wording and loss data matter. Their distinction between underlying primary insurance and the treaty transaction remains relevant to any allocation exercise.

The analysis proceeds from contractual cash flows to mathematical allocation, then to commissions, capital, illustrations, validation, and international interpretation. All formulas are stated for a clearly defined decision period and valuation basis. The paper distinguishes insurer-level financial accounting from internal product pricing and regulatory rate filings. A cost assignment that is useful for product management may not be a permitted rating factor in a regulated jurisdiction; a recorded accounting expense may not coincide with an economically indicated prospective load. The study uses primary professional standards, official legal and accounting materials, and published actuarial research. Its independent contribution consists of the derivations, counterexamples, comparative judgments, and integrated reconciliation.

Research question, estimand, and scope

The central estimand is the prospective incremental economic cost of a specified reinsurance arrangement for a specified set of primary exposures. Let the decision period be one underwriting year, let the segmentation consist of primary cells indexed by i, and let the treaty attach on a stated subject-loss basis. The object is not merely the reinsurer's quoted rate; it is the change in the cedant's economic cost conditional on the contract and selected counterfactual.

Three practical uses must be identified at the outset. Rate indications seek adequate total prospective revenue for primary policies subject to law and competitive conditions. Product profitability and managerial measurement seek a coherent assignment of a jointly purchased protection program. Reinsurance purchasing seeks the incremental value of changing the program. Each use can employ a different cost allocation even if all rely on the same underlying distribution. The first requires reconciliation with gross losses and expenses; the second requires a declared allocation convention; the third requires marginal comparisons of alternatives.

The analysis concentrates on property/casualty insurance with illustrative quota-share and occurrence excess-of-loss covers. Its mathematical approach also illuminates other risk-transfer contracts when their cash-flow and claim mapping can be defined. Long-duration life, health, retroactive reinsurance, financial reinsurance, and cross-border tax rules introduce additional contractual and reporting mechanics and therefore require separate jurisdiction-specific implementation.

Reinsurance Cover and Net Retention

Primary obligation, treaty obligation, and covered loss

A primary policy creates the insurer's obligation to pay insured claims under its wording and applicable law. Reinsurance creates a separate arrangement under which a reinsurer indemnifies the cedant according to treaty conditions. For a portfolio with random gross covered loss L, define C(L, Z) as the reinsurance recovery determined by losses and any additional state variables Z, including event identification, policy and treaty limits, reinstatements, exclusions, inuring protections, and counterparty payment conditions. Retained contractual loss is N = L − C before disputes, credit loss, and timing. For multiple segments, L = Σ_i L_i, but C need not equal Σ_i c(L_i) when the contract operates on aggregate losses or on occurrence-defined totals.

The time dimension is substantive. An occurrence treaty may respond to an event spanning policy and geographic groups; a claims-made contract may depend on notification; an aggregate cover may use a contract-year loss experience; and a quota share may depend on premium attachment, underwriting year, or risks attaching. Inappropriate matching of the primary losses with the treaty's covered cohort can generate a seemingly plausible cost allocation with no contractual foundation. Claims development, court judgments, trends, inflation, deductibles, limits, and subrogation can change both the total loss and the layer to which it is assigned.

For example, on a single claim X, the ceded amount to a conventional per-risk layer with attachment a and limit h is

c(X;a,h)=min⁡{(X−a)+,h},(x)+=max⁡(x,0).c(X;a,h)=\min\{(X-a)^+,h\},\qquad (x)^+=\max(x,0).

A treaty on an aggregate event instead applies the same function to the contractually aggregated event loss, with any applicable per-policy or occurrence restrictions. A program with inuring lower layers requires explicit ordering. Changing the order of insurance recoveries and policy deductibles may change total cession as well as the corresponding allocation.

Gross, ceded, and net perspectives

A product-level costing model may begin with expected gross primary losses, which can be estimated independently of reinsurance. It may also begin with expected retained losses after reinsurance. Both representations are valid when their remaining components are treated consistently. If expected ceded recovery for product i is E[C_i], expected retained loss is E[L_i] − E[C_i]. If the primary rate also adds an allocation of the full reinsurance premium and other cession charges, that is one coherent net-loss presentation. In a gross-loss presentation, the cost added must instead be net of expected recoveries. Adding the whole reinsurance premium to gross losses without an offset for ceded recoveries generally overstates the prospective cost.

The insurer-level identity clarifies the issue. Let P_R denote the expected reinsurance premium paid, B brokerage and acquisition expense associated with the purchase, R expected additional or reinstatement premium, K ceding commissions and other fixed allowances received, Q expected contingent profit or sliding-scale commissions received, and C the actual recovery random variable. On an undiscounted, consistent timing and contract basis define the expected net economic purchase cost

T=E[PR+B+R−K−Q−C].T=E[P_R+B+R-K-Q-C].

Some terms are deterministic; others depend on realized claims. T is an expected net cost of purchasing risk transfer before adjustments for counterparty payment failure, capital benefits, financing, tax, and other firm-level consequences. It can be negative under particular favorable circumstances. A negative T is not a proof that risk transfer produces a net economic subsidy once all correlated benefits and costs are valued.

Let ℓi=E[Li]\ell_i=E[L_i], let eie_i denote non-reinsurance prospective costs, and let uiu_i denote the target contribution for other risk, capital, and profit components on a chosen basis. If a reconciliation allocates T into amounts tit_i satisfying ∑iti=T\sum_i t_i=T, a gross-loss representation is

Pi∗=ℓi+ei+ui+ti,P_i^{*}=\ell_i+e_i+u_i+t_i,

subject to rate normalization, variable expense loadings, and regulatory constraints. The corresponding retained-loss representation uses ℓi−E[Ci]\ell_i-E[C_i], allocates all reinsurance payments and receipts other than recoveries, and then adds other provisions. These are algebraically equivalent if the same recovery allocation, exposure base, and timing are used.

The ASOP 20 cash-flow standard explicitly addresses gross and net cash flows and recoverables, including ceded reinsurance. Its December 2025 revision became effective for covered actuarial work products issued on or after June 1, 2026. This professional standard makes the valuation date, material assumptions, timing, and purpose of a cash-flow analysis important. It does not require universal discounting of every reinsurance premium or claim.

Economic scope and the counterfactual

The amount T measures purchased protection relative to treating gross losses as the starting estimate. The economic value of the purchase additionally depends on the counterfactual program: retaining all risk, buying a different attachment, changing quota-share percentage, placing a facultative cover, purchasing an insurance-linked security, or reducing exposure. An operational saving achieved by reducing surplus requirements may be attributed to reinsurance only when the same saving is not already embedded in the primary rate's capital provision. The relevant difference is V(with treaty)−V(without treaty)V(\text{with treaty})-V(\text{without treaty}) under a consistent objective, not an unexamined redistribution of all capital costs.

The supervisory interpretation is wider still. The IAIS reinsurance principle treats the effectiveness of risk transfer within risk management and supervision, including risks created by the arrangement. This provides a basis for separately evaluating counterparty exposure, contract enforceability, concentration, and operational execution. Treating every contractual recovery as certain cash collected at the moment the primary claim is paid would suppress important components of value.

Treaty Premium and Recoveries

Premium bases and accounting identities

Treaty premiums can be fixed, deposit-adjustable, subject-premium-rated, loss-sensitive, minimum-premium-based, or contingent on reinstatement. Under a premium-rated contract, changes in gross primary rates and policy mix may alter ceded premium even if underlying expected claims do not change in the same proportion. A separate rate-on-line contract can respond primarily to insured limit and market capacity. There is consequently no general theorem that percentage-of-primary-premium is a fair measure of reinsurance cost.

For a proportional quota share with cession fraction q_i on a group i of eligible policies, let W_i be the expected subject premium and G_i the expected covered gross ultimate claims. If the treaty cedes the same fraction of each covered claim and relevant premium, then PR,i=qiWiP_{R,i}=q_i W_i and E[Ci]=qiE[Gi]E[C_i]=q_iE[G_i]. With ceding allowance rate k_i on ceded premium, E[Ki]=kiqiWiE[K_i]=k_iq_iW_i. The basic cost is

Ti=qiWi(1−ki)−qiE[Gi]T_i=q_iW_i(1-k_i)-q_iE[G_i]

before other adjustments. This identity makes the allocation closely tied to exposure and contract terms. If different products have different cession percentages or commissions, use those parameters rather than a uniform revenue key. If a sliding-scale commission is triggered by portfolio results, E[K]E[K] must be evaluated under the joint distribution; applying a commission percentage evaluated at mean loss ratio generally gives a different answer from the mean commission.

For nonproportional treaties, costs may be fixed per limit purchased while recoveries are contingent on contract-defined exceedances. Clark's treaty-pricing discussion emphasizes the idiosyncratic terms of reinsurance programs; Patrik's reinsurance analysis gives additional historical and technical context. Contract premium is an observed transaction amount when placed, whereas future cost estimation depends on the timing and adjustment provisions expected to operate in the rate period.

Expected recoveries and the stop-loss transform

For a nonnegative ground-up loss X with survival function FˉX(x)=P(X>x)\bar F_X(x)=P(X>x), the expected ceded loss to a layer of width h attaching at a is

E[c(X;a,h)]=∫aa+hFˉX(x) dx.E[c(X;a,h)]=\int_a^{a+h}\bar F_X(x)\,dx.

The derivation follows from c(X;a,h)=(X−a)+−(X−a−h)+c(X;a,h)=(X-a)^+-(X-a-h)^+ and the tail integral identity E[(X−t)+]=∫t∞P(X>x) dxE[(X-t)^+]=\int_t^{\infty}P(X>x)\,dx when the expectation is finite. The identity is directly useful for comparing alternative attachment points and limits. Its differential with respect to attachment is −P(a<X≤a+h)-P(a<X\leq a+h), where the distribution is continuous at the endpoints; differentiation with respect to the upper limit is P(X>a+h)P(X>a+h). Thus, an additional unit of limit generally has a marginal expected recovery equal to the survival probability at the expanded limit, while raising the attachment removes expected recovery over the affected band.

When X is a single loss subject to a primary policy limit m, the relevant variable is min⁡(X,m)\min(X,m), after the contractually proper deductible and loss-expense treatment. For event layers, X is the event aggregate after the treaty's specified inuring reinsurance and applicable loss definitions. A per-risk curve may therefore be unusable for an occurrence treaty without an event aggregation model.

The Swiss Re exposure-curve study by Bernegger (1997) provides a mathematical family for allocating expected losses within layers of a limited loss distribution. In practical use, the exposure curve must be matched to the class of business, policy-limit distribution, original loss cost and relevant severity assumptions. A curve calibrated for property risks cannot automatically be carried across to long-tail liability or multiyear catastrophe aggregates.

Frequency, severity, dependence, and accumulation

For a compound count-and-severity model L=∑j=1NXjL=\sum_{j=1}^{N}X_j, with N independent of identically distributed claim severities, E[L]=E[N]E[X]E[L]=E[N]E[X]. If a per-claim treaty layer applies independently to each claim and conditions support the same representation, E[C]=E[N]E[c(X;a,h)]E[C]=E[N]E[c(X;a,h)]. For an aggregate treaty that responds to the annual sum, E[C]E[C] depends instead on the distribution of L. These are different contractual random variables; using the first formula to price the second can substantially understate or overstate the layer cost.

The dependence structure matters when protection covers more than one primary product. Given correlated product losses LAL_A and LBL_B, an occurrence treaty recovery c(LA+LB)c(L_A+L_B) generally differs from c(LA)+c(LB)c(L_A)+c(L_B). It can respond to simultaneous moderate losses that individually would remain within the retention. Catastrophe occurrence definitions, hours clauses, reinstatement limits, and aggregation across policyholder locations determine how the model maps events into coverage. The ASOP 39 catastrophe guidance addresses the special character of catastrophe provisions in ratemaking and the potential usefulness of noninsurance data and catastrophe models when historical experience is unrepresentative.

Excess-of-loss cost estimates also depend on changes in underlying policy limits. Dynda, Bodoff, and Carvalho (2026), in their limits-drift experience study, analyze distortions when the policy-limit profile shifts over time. The same limitation applies when a cedant uses historical ceded losses to allocate a prospective treaty bill. Historical segment shares may reflect earlier limits, deductibles, portfolio size, and underwriting selection rather than the composition of tomorrow's protection benefit.

Timing, discounting, and collection

An undiscounted cost comparison is suitable only when the intended purpose permits it. Premium may be payable at inception; losses may be recovered over years; commissions may be settled quarterly; reinstatement premiums may follow large events; profit commission may be determined after ultimate development; and the reinsurer may pay after the cedant has paid the policyholder. For nominal stochastic cash flows XtX_t and discount factors dtd_t, a present-value analogue is

TPV=E ⁣[∑tdt(PR,t+Bt+Rt−Kt−Qt−Ct)].T_{\mathrm{PV}}=E\!\left[\sum_t d_t(P_{R,t}+B_t+R_t-K_t-Q_t-C_t)\right].

Cash-flow correlations matter. A reinstatement premium may occur exactly in states in which a large recovery is paid. Discounting each item independently at an inappropriate common rate can distort the comparative value. When estimating the amount expected to be collected, a separate allowance may be necessary for reinsurer default, dispute, collateral, and timing. These adjustments should match the purpose: a pricing model's expected economic shortfall and a solvency balance-sheet valuation follow related but potentially different rules.

Article 81 of Solvency II, accessible through the EIOPA recoverables rulebook, expressly requires consideration of the time difference between reinsurance recoveries and direct payments and adjustment for expected loss due to counterparty default. This is a European prudential valuation rule, not a command to use its exact regulatory measurement as a product-level price load everywhere. Nevertheless, it identifies two economically real dependencies that a primary pricing analysis should examine.

Proportional and Layer Allocation

Allocation requires a specified decision rule

Let AiA_i be the allocated net reinsurance cost assigned to primary product i. A reconciliation requires ∑iAi=T\sum_i A_i=T. This full-allocation constraint is necessary for a complete managerial allocation, yet it does not identify a unique answer. If one shared treaty has N eligible products, infinitely many allocations satisfy the single aggregate constraint. Selection therefore requires an objective that links the allocation to policy risk, treaty use, portfolio decision, and possibly fairness or regulatory acceptability.

Four main allocation families appear in practice and mathematical analysis. A premium-based approach uses eligible subject premium shares; an exposure-based approach uses sums insured, modeled layer-penetration frequency, expected layer losses, or another causally relevant driver; a marginal approach values changes in the treaty purchase or portfolio caused by a small change in exposure; and a cooperative approach distributes joint benefit or cost using an explicit division rule. A single insurer may use exposure-based allocations for routine cost attribution and marginal cost analysis for underwriting expansion decisions. Clear explanation of purpose is more important than a presumption that one convenient driver is universally correct.

For weights w_i with ∑iwi=1\sum_i w_i=1, a simple assignment Ai=wiTA_i=w_iT always reconciles arithmetically. The question is what w measures and whether it remains stable under changes in coverage, limits, business mix, and claims dependence. The distinction between mathematical reconciliation and scientific appropriateness prevents neat bookkeeping from being mistaken for a causal explanation.

Premium, exposure, and expected-recovery weights

Premium allocation sets wi=Wi/∑jWjw_i=W_i/\sum_j W_j, where W is eligible subject premium. This method is transparent and can match the treaty's own premium billing base. It is often weak for a high attachment layer when primary rates include unrelated coverage or market factors. Price changes alone can change allocated reinsurance costs with no associated shift in exposure to the layer. A product with low gross premium and very high catastrophe tail can be undercharged in an allocation based solely on premium.

An expected-recovery allocation uses wi=E[Ci]/E[C]w_i=E[C_i]/E[C] when a defensible decomposition C_i exists and the denominator is positive. Its advantage is that the cost shares follow expected insurance benefits, including concentration in covered layers. Its limitation is that a very low-frequency cover can confer capital or liquidity protection whose value differs from average claim reimbursement. Two segments with equal expected recoveries can generate different tail risk reductions and demand distinct capital charges.

A more refined exposure allocation may use modeled layer loss cost per product, event participation, aggregate-limit consumption, or a risk-driver regression against treaty recoveries. It requires a precise mapping from the primary exposure dataset to the treaty's coverage definition. The ASOP 23 data-quality guidance supports attention to data suitability, review, reliance, and disclosure within its professional scope. The data reconciliation should track exposure period, product codes, geographic concentration, per-policy limits, deductibles, claim development, gross versus net loss treatment, and changes in reinsurance program.

For a zero expected-recovery segment, an expectation-only rule allocates zero cost even if that segment increases the probability of reinstatement, reduces aggregate exhaustion capacity, or interacts with other risks in joint events. In a shared catastrophe program such a result can be incorrect for the intended management decision. The model must represent the relevant joint outcomes before choosing a weighting rule.

Marginal and cooperative allocation

Let T(S)T(S) denote expected net reinsurance cost for a portfolio comprising a set S of products, with the reinsurance program specified consistently for that counterfactual. An incremental cost for adding product i to an existing portfolio S is T(S∪{i})−T(S)T(S\cup\{i\})-T(S). The result can change with S because common fixed charges, coverage thresholds, and diversification modify the marginal effect. Marginal contributions generally do not sum to the full cost of the existing shared contract; forcing them to do so by arbitrary scaling can obscure their economic interpretation.

A cooperative Shapley allocation averages the incremental contribution of each member across all orderings of the members. For a cost or recovery characteristic function v(S), the contribution is

ϕi(v)=∑S⊆N∖{i}∣S∣!(n−∣S∣−1)!n! [v(S∪{i})−v(S)].\phi_i(v)=\sum_{S\subseteq N\setminus\{i\}}\frac{|S|!(n-|S|-1)!}{n!}\,[v(S\cup\{i\})-v(S)].

For two segments A and B, ϕA={v(A)−v(∅)+v(A,B)−v(B)}/2\phi_A=\{v(A)-v(\varnothing)+v(A,B)-v(B)\}/2 and similarly for B. The allocations sum to v(A,B)−v(∅)v(A,B)-v(\varnothing) when the characteristic function is defined consistently. This makes Shapley particularly useful for a shared occurrence recovery whose value arises from joint event losses. It remains a chosen cooperative convention, sensitive to the definition of hypothetical subportfolios and the reinsurance contract used for each subset.

Capital-allocation research provides related arguments but addresses a distinct object. Myers and Read's insurance capital-allocation analysis models contributions to insurer default value; Dhaene, Tsanakas, Valdez, and Vanduffel's optimal allocation principles examine alternative optimization criteria for distributing firm capital. These results cannot be treated as a theorem requiring a particular allocation of a purchased treaty's invoiced premium. Their relevance is to the design and interpretation of capital-sensitive components of the primary pricing decision.

Contract consistency and model boundary

An allocation must reproduce a treaty's actual indemnity function. For example, if an occurrence layer is 5 million in excess of 5 million, a combined 13-million occurrence produces a 5-million recovery, whereas two claims from different independent occurrences producing 9 million and 4 million do not necessarily generate the same response. An analyst who groups losses by calendar day, weather event, legal incident, or catastrophe code without checking the treaty's event definition may assign recoveries to the wrong product or year.

Allocation also requires correct treatment of facultative reinsurance, which may attach to individually identified policies before treaty loss aggregation. Inuring covers, quota shares placed below occurrence protections, and reinstatement terms can alter the gross-to-net mapping. The analyst should first calculate contract-consistent recoveries at the claim or event level and only then allocate the modeled program total. Building one independent calculation for each product and adding its outputs can violate the actual treaty limit.

Claims modeling is subject to estimation and credibility. The ASOP 25 credibility guidance addresses blending subject experience with relevant experience. For a sparse excess layer, five claim-free years do not prove the expected recoveries are zero, just as a single severe event does not automatically establish a persistent high annual loss cost. Calibration may require normalized exposure years, policy-limit changes, modeled hazard, external catastrophe data, and careful documentation of dependency assumptions.

Selection among plausible cost drivers

A defensible allocation procedure begins with the actual problem. To allocate a fixed negotiated bill for financial planning, expected claim benefit may be an intelligible key. To assess the marginal cost of expansion into a coastal territory, prospective change in treaty premium and retained risk is more relevant. To prepare a rate indication, the estimated future net economic cost and the rate plan's legally permissible classification variables are relevant. To evaluate an annual product manager, a stable predeclared allocation can be preferable to an allocation that changes dramatically after an unpredictable event.

This diversity is not methodological weakness. It reflects the different estimands of cost attribution, pricing adequacy, and decision support. The key governance failure is to present a cost allocation chosen for administrative convenience as if it measured causal marginal risk, or to use a marginal analysis as if it were a unique full allocation of the existing program.

Commissions and Participation

Ceding commissions and allowances

In proportional reinsurance, the cedant may receive a ceding commission representing a contribution toward acquisition and administration costs. A basic net-cession comparison deducts this receipt from the cost of ceded premium while separately accounting for expected claims recovered. The incidence on primary products follows the commission's actual contractual base; the commission is not automatically earned at the same rate for every coverage. Where a treaty pays different commission percentages by line, those percentages should be preserved in the product allocation.

The ASOP 29 reinsurance-expense provision identifies commissions and allowances among items to consider, together with premium, recoveries and reinstatement. A ceding allowance should not simultaneously reduce the reinsurance expense and be counted as independent income in a primary pricing model unless the accounting layout explicitly reconciles the two presentations without duplicating economic effect.

A fixed commission K is straightforward. A sliding-scale commission depends on loss experience and contractual thresholds. If K=f(L)K=f(L), then the required prospective quantity is E[f(L)]E[f(L)]. It is generally not f(E[L])f(E[L]). For example, a commission that pays 0 when losses exceed a threshold and 1 million otherwise has expected value equal to 1 million multiplied by the probability of the qualifying outcome. Evaluating the clause only on the mean loss can misstate the expected receipt by as much as the entire commission.

Profit commissions, additional premiums, and retrospective features

Profit commissions, contingent brokerage, experience accounts, loss corridors, and retrospective premium adjustments change the final economic balance and may create nonlinear claims dependence. A profit commission can rise when ceded claims are low, while reinstatement premium rises following particular large losses. Their covariance with recoveries and other cash flows is material. Allocating these contract terms by their mathematical trigger is more defensible than assigning them to a product simply because that product has the largest original premium.

Reinstatement illustrates the distinction. A treaty may permit exhausted limit to be restored for an additional premium, sometimes proportional to the amount reinstated and remaining treaty term. The expected premium depends on the likelihood, timing, and number of exhausting occurrences and the precise reinstatement provision. If the cover offers one reinstatement, a modeled second large event may exhaust all remaining protection. A pricing model that includes expected reinstatement payment while assuming unlimited reinstated cover understates retained tail loss.

The analysis of such contingent payments fits the revised ASOP 20 cash-flow scope, which addresses retrospective premiums, contingent commissions, and the timing and amount of recoverables in scoped property/casualty cash-flow work. Appropriate models should evaluate the joint path of losses, commissions, reinstatements, and settlements, rather than replacing contract-triggered amounts with independent flat percentage loadings.

Premium elasticity and circularity

A quota-share treaty priced as a fixed percentage of subject premium can create an algebraic feedback loop in primary rate development. If increased primary premium mechanically increases reinsurance premium, and a percentage of that premium is added again to the gross premium indication, the relationship should be solved as a simultaneous equation or an appropriate iteration. Ignoring the denominator effect can misstate the indicated rate.

Suppose the variable economic reinsurance charge equals fraction a of primary premium, other variable acquisition expenses equal fraction v of primary premium, and all fixed expected costs total F. With no other profit condition, the required premium is P=F+aP+vPP=F+aP+vP, hence P=F/(1−a−v)P=F/(1-a-v) when a+v<1a+v<1. Here a must be the economically relevant net ceded-premium percentage on the chosen representation, not the gross ceded premium if expected recoveries are already deducted elsewhere. When recoveries depend on underwriting volume and selection, a may itself vary with P; the fixed-point assumption then requires explicit sensitivity analysis.

Tax and legal considerations vary by country, entity, and contract. The mathematical premium gross-up is an internal pricing identity, not a universal statutory formula. A primary insurer should also distinguish costs charged to the cedant from premium taxes or levies collected under its primary regulatory framework. In multilingual, multinational programs, contract interpretation and local admitted-insurance requirements can determine which underwriting entity bears the payment.

Reporting versus product pricing

IFRS 17 establishes presentation and measurement for insurance contracts issued and reinsurance contracts held. The IFRS Foundation's IFRS 17 standard overview describes its risk-adjusted present-value measurement and profit recognition. The AASB 17 reinsurance provisions, which incorporate corresponding IFRS 17 requirements in an Australian standard, reproduce the separation of issued insurance contracts and reinsurance held and the treatment in paragraph 86 of certain commission cash flows as reductions of the premiums paid when presented separately.

This reporting distinction constrains financial statement presentation; it does not identify the economically preferred cost allocation across individual rating cells. Product-pricing models may attribute a treaty cost to a territory or line using a causal exposure key even though accounting groups and coverage units follow IFRS 17 requirements. Conversely, a reinsurance asset recognized in financial reporting is not the same as cash received or the same as a reduction in a primary insurance rate. Reconciliation is required between the management price view and published accounts, with the mapping and timing differences retained.

Capital Benefit and Pricing

Risk transfer and capital relief

Reinsurance can reduce the severity of insurer losses and required risk-bearing capital while introducing counterparty, credit, liquidity, operational, and basis risks. Capital savings may be economically valuable even when expected ceded losses are modest. However, the value depends on the insurer's existing capital position, rating and solvency constraints, available alternatives, asset risks, diversification, and the selected risk measure. It is not valid to assume that every dollar of modeled gross exposure removed reduces required capital by one dollar.

The ASOP 30 capital-cost standard establishes the relevance of capital opportunity cost, investment income, leverage, and reinsurance to profit provisions in its U.S. property/casualty ratemaking scope. In a rate study that has already estimated a gross-risk capital provision, capital relief from reinsurance may enter as a reduction from that provision. If the rate starts with a post-reinsurance capital model, no additional stand-alone saving should be subtracted. Both presentations can be correct; combining them produces double counting.

For a simple one-year approximation with hurdle rate r on eligible risk capital, define KgK_g and KnK_n as comparable gross-program and net-program capital measures. A capital-cost difference is ΔH=r(Kg−Kn)\Delta H=r(K_g-K_n), provided that the capital amounts, financing basis, tax treatment, and period are comparable. If counterparty risk increases post-reinsurance capital, that increase belongs in K_n. A single percentage r is an illustrative economic assumption, not a universal legal rate or a direct formula for Solvency II SCR.

Myers and Read (2001), in Capital Allocation for Insurance Companies, relate capital allocation to marginal default value and dependence among lines and investments. Gründl and Schmeiser (2007), in their critique of capital allocation, argue that allocation is not always necessary for insurance pricing and that net present value analysis may better guide capital-budgeting decisions. Taken together, these studies demonstrate that a risk-capital allocation must be justified by the decision it informs; the existence of a total capital requirement alone does not supply unique product-level price charges.

Risk reduction, opportunity cost, and double-counting tests

The manager should distinguish expected indemnification from risk stabilization. Expected recoveries belong in the claims cash-flow projection. The treaty premium pays for indemnification plus reinsurer expenses, capital, and frictional charges under the market arrangement. A change in the cedant's opportunity cost of holding capital can alter the price required to supply the underlying primary cover. It should not automatically be labeled an expected claim recovery, nor should the reinsurance premium and capital relief both be converted into arbitrary identical percentage multipliers.

A useful identity expresses expected total primary economic cost under the treaty as

Fnet=E[L]+E[Eprimary]+T+Hnet,F_{\mathrm{net}}=E[L]+E[E_{\mathrm{primary}}]+T+H_{\mathrm{net}},

where HnetH_{\mathrm{net}} is a post-treaty capital and contingency provision on the same basis. The comparison without the treaty is

Fgross=E[L]+E[Eprimary]+Hgross.F_{\mathrm{gross}}=E[L]+E[E_{\mathrm{primary}}]+H_{\mathrm{gross}}.

Subtracting yields Fnet−Fgross=T−(Hgross−Hnet)F_{\mathrm{net}}-F_{\mathrm{gross}}=T-(H_{\mathrm{gross}}-H_{\mathrm{net}}). This simple equation makes clear that treaty economic net cost and capital benefits are related but distinct terms. If the gross and net risk loads contain incompatible expense and profit components, the comparison needs further alignment before it becomes decision-useful.

The optimal capital-allocation study by Dhaene et al. (2012) provides examples of different risk-allocation objectives yielding different allocation methods. A modeler must state whether the allocation measures economic capital, regulatory capital, expected shortfall contributions, contractual reinsurance benefits, or managerial profitability. Their measures can be correlated but are not interchangeable. In particular, an Euler risk contribution can become unstable or undefined at nondifferentiable portfolio points; a Shapley value depends on a coalition model and may require many portfolio recalculations.

Reinsurer credit, basis, and collateral

A cedant exposed to a large expected recovery from an unsecured reinsurer bears nonperformance risk. Security provisions, collateral, trust arrangements, credit quality, disputes, and timing may change expected realized recoveries. A capital model may therefore record less relief for weakly secured protection. A pure contractual indemnity model describes the reinsurer's promise; an economic collection model describes the cedant's attainable payment, and both have their uses.

The Solvency II recoverables provision explicitly incorporates counterparty default into recoverables valuation within the EU system. In a U.S. or other jurisdiction, statutory recognition of ceded balances, collateral requirements, and capital credit differ. An internationally applicable research conclusion must separate the physical stochastic loss process, the contract response, the insurer's accounting basis, and the regulator's recognition rules.

Reinsurance basis risk can arise from exclusions, aggregate event definitions, settlement timing, policy-limit mismatch, and unmodeled coverage conditions. A reinsurer may also deny or delay a claim. Therefore, an allocation based on full modeled contractual recovery should disclose whether an adjustment for collectible recovery and related liquidity cost has been made. Treating expected contractual and expected collected recovery as synonyms can create a fictitious premium benefit.

International comparability and governance

The IAIS framework for reinsurance provides supervisory principles on risk transfer, while the European Solvency II Directive contains a distinct legislative capital and counterparty-default framework. U.S. actuarial standards such as ASOP 29 and ASOP 53 describe professional conduct for covered assignments rather than directly legislating a common worldwide price. IFRS 17 prescribes accounting presentation and measurement to entities within its scope. These documents inform different components of the same business problem.

Governance should identify the contract version, period, segmentation, modeled gross loss distribution, ceded indemnity, valuation basis, commission treatment, capital methodology, and reconciling allocation weights. It should also document the actual primary pricing purpose, responsible reviewer, material assumptions supplied by brokers or external model providers, and the degree of data verification. External peer review can strengthen confidence when commissioned, but the internal scientific and editorial validation of the present monograph is a separately documented process.

Case Tests and Reconciliation

Case A: two primary products under proportional cover

Consider two invented primary products, property A and casualty B, with expected eligible subject premiums of $22 million and $16 million and expected gross covered ultimate claims of $12 million and $8 million. A quota-share agreement cedes 25% of A and 10% of B; both products receive a 30% ceding allowance on their ceded premium. For this example, the eligible premium base and ultimate-loss scope coincide, and the commissions and recoveries are assumed fully collectible on an undiscounted basis. There are no additional brokerage charges, retrospective adjustments, or changing exposure conditions.

Quantity, $ million Property A Casualty B Total
Eligible subject premium 22.000 16.000 38.000
Gross expected ultimate claims 12.000 8.000 20.000
Ceded proportion 25% 10% Varies
Ceded premium 5.500 1.600 7.100
Ceding commission received 1.650 0.480 2.130
Expected ceded recovery 3.000 0.800 3.800
Net reinsurance cost 0.850 0.320 1.170
Expected retained claims 9.000 7.200 16.200

The aggregate identity is 7.100−2.130−3.800=1.1707.100-2.130-3.800=1.170 million. Both product amounts derive from their own cessions and commission bases, and 0.850+0.320=1.1700.850+0.320=1.170. The economic claims-plus-reinsurance position is 16.200+(7.100−2.130)=21.17016.200+(7.100-2.130)=21.170 million, equal to gross expected claims 20.00020.000 plus net reinsurance cost 1.1701.170. This equality is a critical double-counting test. An indication that began with the $20 million gross loss estimate and then added $7.1 million without subtracting commissions and ceded recoveries would make the wrong economic comparison.

A crude premium allocation would assign A a share 22/3822/38 and B 16/3816/38 of the $1.17 million net cost: $0.677368 million and $0.492632 million. That result reconciles arithmetically but ignores the deliberately unequal treaty cessions. It overstates B's allocated cost relative to the treaty-consistent product cash-flow attribution and understates A's. This example demonstrates why even a simple proportional treaty should not be allocated by gross premium alone when policy terms differ.

The experience is hypothetical. No conclusion about an actual property or casualty market rate can be drawn from it. If products had different claim settlement patterns or reinsurance commission terms, a present-value analysis could produce a different allocation; if ceding percentages varied within products, a finer granularity would be necessary.

Case B: shared occurrence excess-of-loss protection

Consider a separate invented reinsurance purchase: one shared $5-million occurrence layer attaching at $5 million of eligible event loss from products A and B combined. To make the mathematical issue transparent, use four equally weighted event-loss states in which the single occurrence's product losses are as displayed. No second occurrence, reinstatement exhaustion, legal dispute, or inuring lower treaty changes the modeled recoveries in these four states. Monetary amounts below are millions of dollars.

Equally weighted state A loss B loss Combined loss Treaty recovery
One 3 0 3 0
Two 8 0 8 3
Three 9 4 13 5
Four 1 7 8 3
Arithmetic expectation 5.25 2.75 8.00 2.75

The gross expected loss here is $8 million, and the treaty's expected contractual recovery is $2.75 million. When expected claims are retained, their aggregate is $5.25 million. For state three, the $13-million combined loss exhausts the $5-million layer; treating A and B independently would produce $4 million from A and $0 from B, missing the contract's fifth million of joint-event benefit. State four similarly yields a combined $3-million recovery, compared with $0 from A alone and $2 million from B alone.

For the two-product coalition, define v(S)v(S) as the treaty-layer recovery evaluated on the losses of products in S under the same state, preserving attachment and limit. With v(∅)=0v(\varnothing)=0, the two-player Shapley allocation equals one half of the standalone recovery plus one half of the increment when added to the other product. The resulting state allocations are:

State A Shapley recovery B Shapley recovery Check
One 0.0 0.0 0.0
Two 3.0 0.0 3.0
Three 4.5 0.5 5.0
Four 0.5 2.5 3.0
Expected amount 2.0 0.75 2.75

The arithmetic is explicit. In state three, v(A)=4v(A)=4, v(B)=0v(B)=0, and v(A,B)=5v(A,B)=5. Therefore A receives (4+5−0)/2=4.5(4+5-0)/2=4.5, and B receives (0+5−4)/2=0.5(0+5-4)/2=0.5. In state four, v(A)=0v(A)=0, v(B)=2v(B)=2, and v(A,B)=3v(A,B)=3, producing A's 0.5 and B's 2.5. Every state's allocated total equals the treaty recovery, and the expected allocated recoveries sum to $2.75 million. This calculation illustrates the cooperative attribution of an event layer; it is not a universal replacement for other reasonable keys.

Assume separately that the prospective cedant cost of this treaty consists of a $3.60-million fixed premium, $0.18 million brokerage, and an expected $0.12 million contractually possible additional premium. The example treats the last item as an externally supplied scenario-based expected value and does not infer it from the simplified four states. The expected all-in payment is $3.90 million, and the expected net economic reinsurance cost is 3.90−2.75=1.153.90-2.75=1.15 million, before capital, financing, or counterparty adjustments.

For a transparent illustrative cost split, allocate the $3.90-million payment by the modeled Shapley recovery shares 2/2.752/2.75 and 0.75/2.750.75/2.75. The allocated paid costs are $2.836364 million and $1.063636 million; after deducting allocated expected recoveries, A receives a net treaty cost of approximately $0.836364 million and B $0.313636 million. Their sum equals $1.15 million. This is a recoveries-based bill allocation using Shapley to attribute joint-event benefits. It is not a Shapley calculation of the purchased premium itself, nor is it proof that the cost of reinsurer capital follows expected recoveries.

As a contrasting premium-share rule, suppose eligible primary premium shares are 70% A and 30% B. Allocating the $3.90-million payment in these proportions gives $2.73 million and $1.17 million. Deducting the unchanged expected Shapley recoveries gives net allocated costs of $0.73 million and $0.42 million. Both methods reconcile to $1.15 million. Their product-level results differ because they answer different attribution questions. The primary premium split is introduced solely as an assumption for this comparison; it is not derived from the four event states.

Capital relief and a complete double-counting test

For a separate illustrative sensitivity applied to Case B, assume a consistent gross-program risk capital requirement of $15 million and a net-program requirement of $12 million after all treaty-related risks, with a one-year 8% capital-cost rate. The gross and net capital provisions are $1.20 million and $0.96 million; the modeled saving is $0.24 million. The treaty therefore changes total economic cost relative to no reinsurance by 1.15−0.24=0.911.15-0.24=0.91 million, if all other provisions are unchanged. If the gross-price capital provision has already been reduced to $0.96 million in the net-reinsurance rate, subtracting an additional $0.24 million from that net-reinsurance indication would count capital relief twice.

For an illustrative product split of the $3-million capital reduction, assign $2.5 million to A and $0.5 million to B only as an assumed actuarial capital attribution. The corresponding capital savings are $0.20 million and $0.04 million. Under the expected-recovery allocation above, A's resulting net impact on the gross-based price is approximately 0.836364−0.200000=0.6363640.836364-0.200000=0.636364 million and B's 0.313636−0.040000=0.2736360.313636-0.040000=0.273636 million. Their sum equals $0.91 million. Under the premium-based treaty allocation, analogous impacts would be $0.53 million and $0.38 million. All allocations in this paragraph are dependent on the assumed capital attribution, not an inferred optimal division.

A pricing actuary must additionally decide how rate indications translate into charged premiums. A segment's new price may not move dollar-for-dollar with allocated treaty cost because variable expenses, taxes, approved rating factors, market constraints, portfolio selection, and regulatory processes matter. The example calculates cost contributions, not issued policy prices or an insurer's approved filing.

Stress tests: the allocation changes with assumptions

Stress testing should vary both the amount of expected recovery and the segment shares. Consider a more dangerous catastrophe mix that increases the frequency of simultaneous A/B losses. The occurrence-layer expectation may rise even when individual product mean losses change little, because the probability of penetrating the shared attachment increases. A fixed premium-share allocation would largely miss this interaction; an event-conditioned cooperative allocation changes with the new joint distribution.

A second stress changes underlying policy limits. If B raises its insured limits while primary premium increases only modestly, its contribution to the upper tail and the treaty's exhaustion distribution can increase more than its share of revenue. The 2026 limits-drift paper explains why past limit profiles can mislead prospective excess-layer estimates. A third stress decreases the reinsurer's credit quality, producing lower economically collectible recovery and potentially higher net capital; the cost assignment should then reflect the segment and treaty exposures generating the recovery asset, together with aggregate credit concentration.

A fourth stress changes the objective. A management accounting rule might preserve a historically agreed allocation for comparability while a marginal price test evaluates changes in the loss distribution from one additional policy. Neither result supersedes the other without specifying the purpose. The model documentation should report the source of each difference rather than smoothing disagreements into one unexplained average.

Reconciliation protocol and internal validity

The mathematical checks in the two cases are independent of market data. They verify that treaty-consistent recoveries are bounded by the stated layer, that allocated recoveries sum to treaty recoveries state by state, that gross minus ceded equals retained expected losses, that commission and premium balances reproduce treaty cost, and that capital savings are counted once. The implied assumptions—one valuation period, consistent monetary units, no double-recovery, fully collectible indemnities, and declared allocation rules—are part of the result.

A reproducible implementation can follow this sequence. First freeze the underlying exposure, claims, treaty wording, effective dates, and inuring structure. Next calculate gross claims under a common economic scenario set and compute contractual treaty recoveries with event aggregation, limits, and all contingencies. Then separately calculate expected paid premium, commissions, reinstatement, and recoveries and reconcile total cost. Choose and document the allocation objective, and allocate each cash-flow category with an appropriate driver. Project post-reinsurance capital and credit exposures on a matching basis, then verify whether the initial gross-price capital load has already been adjusted. Finally reconcile the segment totals to the treaty ledger and the intended pricing measurement and retain a separate accounting bridge for reporting under any relevant accounting standard.

Useful validations include nonnegativity and contractual upper bounds for individual recoveries; equality of gross, ceded, and net loss totals; no treaty reimbursement greater than contractual covered losses; correct claim-versus-occurrence unit; correct treatment of policy and event limits; explicit cash-flow timing; evidence of the relevant limit distribution; concentration and counterparty stress; and documentation of material model disagreement. The ASOP 53 disclosure provisions support recording the methods and material assumptions used to estimate reinsurance costs in the standard's professional scope.

Original analytical synthesis

Five conclusions follow jointly from the foregoing derivations and source comparison. First, primary pricing uses an economic net-cession cost even when the reinsurance treaty is contracted on a gross-premium basis; the expected recovery offset is a necessary part of the identity. Second, a shared excess-of-loss treaty has an event-level benefit function, and product-level recoveries must be constructed from the treaty's actual aggregation, not from independent product layers. Third, full attribution of the treaty invoice and marginal valuation of a product's additional risk are different estimands. Fourth, the capital effect must be obtained from a consistent comparison of gross and net risk-bearing positions and appears exactly once in the primary cost identity. Fifth, auditability depends on the joint preservation of contract, data, source, allocation method, and reconciled cash-flow totals.

These are conclusions of the present analytical comparison using established actuarial mathematics, financial reasoning, and cited research. The Shapley calculation is an application of a known cooperative allocation method to a specified illustrative event function. The gross-to-net bridge is an accounting identity under the stated assumptions. Neither is described as a newly discovered insurance theorem or an independent authorial doctrine.

The most robust practical result is the separation of three questions in a primary pricing study: which contingent cash flows the treaty actually transfers; how those cash flows and shared costs are assigned to the primary risks that generate them; and how the resulting retained-risk position changes capital and prospective prices. A study that answers these questions separately and then reconciles them provides more decision-useful evidence than a single unexamined percentage surcharge.

Conclusion

Reinsurance cost allocation in primary insurance pricing is the actuarial attribution of the prospective net economic cost and risk consequences of a purchased risk-transfer arrangement to the underlying primary exposures. The analysis starts with the contract's indemnity function and the corresponding gross loss distribution. It then measures premiums, expected indemnification, commissions, contingent payments, brokerage, collection, and timing on a common basis. An allocation rule translates those amounts to products or rating classes according to the decision being supported, while the insurer's capital model determines the separate effect on the cost of retained risk.

Proportional treaties permit relatively direct attribution when ceded percentages, premium bases, recoveries, and commission terms are identifiable by product. Shared excess-of-loss covers require joint-event modeling, particularly where the same event involves multiple products or catastrophe concentrations. Premium shares, modeled expected recovery shares, marginal contributions, and cooperative allocations can all reconcile to the treaty total while assigning different costs to products. Their appropriateness depends on stated objectives and data quality.

The illustrated calculations demonstrate complete gross-to-net reconciliation and identify the central double-counting traps. A ceded premium cannot be added to gross expected loss without an appropriate recovery offset; a ceding allowance cannot be credited twice; contractual indemnity and collected recovery can differ; and capital relief cannot be deducted again after net capital requirements have already been reflected. Regulatory supervision, accounting presentation, professional actuarial guidance, and managerial pricing use related information for different purposes and must retain their respective boundaries.

The study's scientific result is an integrated, verifiable chain from primary exposure to treaty cash flow, from treaty cash flow to shared allocation, and from shared allocation to the primary prospective rate indication. Its reproducibility rests on explicit mathematical conditions, full reconciliation, contemporaneous source attribution, and transparent limitations. Further empirical work would require actual insurer treaty contracts, claim bordereaux, scenario simulations, historical limit distributions, recovery-credit assessments, and jurisdiction-specific rate data. Such empirical fitting and external independent peer review have not been represented as performed in this monograph.

References